Positive selection of co-opted mobile genetic elements in a mammalian gene: If you can't beat them, join them.

The proopiomelanocortin (Pomc) gene encodes a prepropeptide with essential functions in the response to stress and energy balance, which is expressed in the pituitary and hypothalamus of vertebrate animals. Neuronal expression of Pomc is controlled by two distal enhancers named nPE1 and nPE2. Using transgenic mice, we observed that both enhancers drive identical expression patterns in the mammalian hypothalamus, starting at embryonic day 10.5, when endogenous Pomc expression commences. This overlapping enhancer activity is maintained throughout hypothalamic development and into adulthood. We also found that nPE1 and nPE2 were exapted as neuronal enhancers into the POMC locus after the sequential insertion of two unrelated retroposons. Thus, nPE1 and nPE2 are functional analogs and represent an authentic first example of convergent molecular evolution of cell-specific transcriptional enhancers. In this Commentary we discuss the following questions that remain unanswered: (1) how does transcriptional control of POMC operate in hypothalamic neurons of non-mammalian vertebrates? (2) What evolutionary forces are maintaining two discrete neuronal POMC enhancers under purifying selection for the last ~100 million years in all placental mammals? (3) What is the contribution of MaLRs to genome evolution?

The proopiomelanocortin gene (POMC) and its encoded peptides constitute a formidable theater of operations used by multiple evolutionary forces that shaped an adaptive polyfunctional toolkit that has increased the fitness of all vertebrate animals when exposed to environmental stress. Several unrelated peptide-coding sequences were recruited more than 400 million years ago (MYA) to assemble a modular prohormone, reminiscent of the polycystronic organization of prokaryotic genes. Targeted endopeptidase cleavage of the POMC prohormone gives rise to cell-specific combinations of the bioactive melanocortins α-, β- and γ-MSH, ACTH and the opioid peptide β-endorphin (Fig. 1A), which together coordinate the stress-response program. POMC is mainly expressed in pituitary corticotrophs and melanotrophs and in a group of neurons present in the arcuate nucleus of the hypothalamus.

Figure 1.

(A) Schematic of the mouse POMC gene. The exapted neuronal enhancers nPE1 and nPE2, and the distal pituitary enhancer are color-boxed. Pomc exons are in black boxes. Coding sequences for the POMC prohormone and the peptides obtained after endoproteoly‎tic cleavage are indicated in colored boxes. The asterisk next to β-MSH denotes that this peptide is not released in mice. (B) Schematic of the transgenes nPE1Pomc–tomato and nPE2Pomc-EGFP transgenes described in Ref. 6. (C) Expression of tomato and EGFP in the developing hypothalamus of e13.5 compound transgenic mouse embryos showis extensive colocalization of both reportersin neurons.

The tissue-specific control of POMC expression also follows a modular organization with an array of transcriptional enhancers, particularly identified in mammals, that are scattered along the POMC 5′ flanking region. Pituitary specific POMC expression depends on the concerted action of proximal sequences located within 400 bp upstream of the transcription start site and a distal enhancer module, whereas neuronal expression is independently controlled by cis-regulatory elements located further upstream., A combination of genome comparison and expression studies in transgenic mice led us to the identification of two mammalian conserved neuronal enhancers, nPE1 and nPE2, that are able to independently drive reporter gene expression to POMC hypothalamic neurons when placed upstream of a minimal heterologous promoter. We also demonstrated that only the concurrent removal of nPE1 and nPE2 from transgenic constructs leads to the loss of reporter gene expression in POMC hypothalamic neurons. More recently, we found that nPE1 and nPE2 were exapted as neuronal enhancers into the POMC locus after the sequential nearby insertion of two unrelated retroposons. nPE2 originated from the earlier exaptation of a CORE-SINE retroposon in the lineage leading to mammals sometime before the Prototheria/Metatheria split (~170 MYA) whereas nPE1 is a placental novelty that originated from a Mammalian-apparent LTR (MaLR) retroposon after the Metatheria/Eutheria split (~150 MYA) and before the wide mammalian radiation that occurred 90 MYA. Thus, POMC represents one example of probably thousands of vertebrate genes that evolved following a “foreign legion” strategy through the recruitment of unrelated components that, once assembled into a functional gene, acquired a unified “esprit de corps” that contributes to improve fitness (Fig. 1A).

A special case is reserved for teleost fishes, because they have two paralog copies of POMC (α and β) that originated after a whole-genome duplication that occurred 320 MYA in the teleost lineage. POMCα and POMCβ underwent a subfunctionalization process during teleost evolution through the partitioning of their expression domains and respective peptide repertoires which has maintained both paralogs under purifying selection. POMCα retained its expression in the nucleus lateralis tuberis (homolog fish structure to the arcuate nucleus), but how the two paralogs regulate their individual expression patterns and coordinate their physiological functions remains to be studied.

To determine whether nPE1 and nPE2 control neuronal POMC expression in overlapping or complementary spatio-temporal domains within the developing and adult arcuate nucleus we generated transgenic mice expressing the fluorescent markers tomato and EGFP under the transcriptional control of nPE1 or nPE2, respectively (Fig. 1B). Analysis of compound transgenic mice demonstrated co-expression of both reporter genes in neurons along the entire arcuate nucleus from the earliest onset of POMC expression at e10.5. Similar results were obtained at all later embryonic and postnatal stages examined (Fig. 1C). Altogether, these results demonstrated that nPE1 and nPE2 are two functional analogs and represent an authentic first example of convergent molecular evolution of cell-specific enhancers.

Although our studies have provided several key insights about the cis-acting elements controlling mammalian POMC expression, several questions remain unanswered:

(1) How does transcriptional control of POMC operate in hypothalamic neurons of non-mammalian vertebrates? POMC is expressed in neurons of the mediobasal hypothalamus of all jawed vertebrates examined to date including mammalian, avian, reptile, amphibian and teleost fish species. However, the neuronal enhancers nPE1 and nPE2 have been identified only in mammals, consistent with the evolutionary origin and fate of the exapted CORE-SINE (nPE2) and MaLR (nPE1) retroposons. Our efforts to identify nPE1 or nPE2 orthologs in non-mammalian Classes have been unproductive. Although the existence of an ancestral neuronal enhancer(s) of POMC seems probable, phylogenetic footprinting analyses to identify alternative conserved non-coding regions in the vicinities of POMC have consistently failed. However, the proposition that each vertebrate Class independently acquired convergent molecular mechanisms to transactivate POMC in arcuate neurons does not seem parsimonious. Thus, based on the existing sequence comparison data among and within vertebrate Classes together with what we have learned about the evolutionary origin of nPE1 and nPE2 we favor the hypothesis that an ancestral neuronal POMC enhancer existed in the last common forerunner to all extant vertebrates. This ancestral enhancer that may still exist in a particular lineage has been dynamically reshuffled following lineage-specific mechanisms, which could have involved, as in mammals, the exaptation of mobile genetic elements of distinct origin. POMC is one of many genes in which protein-coding sequences have been under greater selective constraint during longer evolutionary periods than transcriptional enhancers which remain conserved only within specific branches. We believe that this differential mechanism between coding and non-coding sequences is key to illuminate the genetic basis of developmental repatterning and animal diversity.

(2) What evolutionary forces are maintaining two discrete neuronal POMC enhancers under purifying selection for the last ~100 million years in all placental mammals? The adaptive value of having two functionally overlapping enhancers, instead of just one, has been proposed for developmental genes in light of the canalization theory originally elaborated by Waddington and Schmalhausen who put forward the idea that the precise and stable spatio-temporal phenotypes normally observed during embryogenesis are assured by duplicated pathways that reduce variability by buffering environmental perturbations. Recently, two groups independently demonstrated for the D. melanogaster genes snail and shavenbaby that overlapping (shadow) enhancers provide adaptive robustness to overcome suboptimal environmental conditions during fly development., Although Pomc expression starts at embryonic day 10.5 in the mouse, it cannot be considered a classical developmental brain gene because all known functions of POMC-derived peptides are exerted after birth and POMC knockout mice or humans do not exhibit overt central developmental defects.,

Thus, nPE1 and nPE2 probably constitute the first example of overlapping (shadow) enhancers acting in a gene primarily involved in postnatal development and adult physiology. For a gene like Pomc, the adaptive value of having two apparently redundant enhancers may be related to the probability of increasing transcriptional rate and/or minimizing its variance. These two, not mutually exclusive, hypotheses may be illustrated by a double scull boat that not only has the capacity to propel faster but also to keep on going in the event that one rower wears out (acquisition of a deleterious mutation). Overlapping enhancers might have occurred as a compensatory evolutionary mechanism during the unicellular to multicellular transition to circumvent the increased vulnerability to inactivating mutations imposed by the emergence of numerous cell-specific enhancers controlling expression of every gene in different cell types and developmental trajectories. Our expression studies in transgenic mice showing that nPE1 and nPE2 drive reporter gene expression to an identical group of neurons along the entire space-time dimension of the mouse arcuate nucleus do not support the possibility that the two enhancers underwent a subfunctionalization fission process, as observed for enhancer paralogs of duplicated genes.,, To the contrary, the independent recruitment of two evolutionary unrelated enhancers to jointly subserve neuronal Pomc expression may be viewed as a case of superfunctionalization, a term recently coined for the duplication of genes that control circadian rhythms in prokaryotes., Different from these reported cases, nPE1 and nPE2 illustrate superfunctionalization of transcriptional mammalian enhancers. Mutant mice carrying targeted deletions of nPE1 and/or nPE2 will constitute a valuable tool to experimentally test the functional importance of each enhancer in the control of neuronal Pomc expression and the adaptive value of enhancer superfunctionalization in adult mice.

(3) What is the contribution of MaLRs to genome evolution? Non-autonomous retroposons are intracellular parasites capable of colonizing their host genomes in association with autonomous retrotransposon partners that provide the necessary “copy and paste” enzymatic machinery. Host genomes, in turn, developed insightful strategies throughout evolution to prevent massive transposon insertions that would otherwise impair their genetic information flow. After a long-lasting evolutionary arms race, mobile elements and genomes have reached a tolerable cease-fire. However, beneath the apparently calm surface of a Cold War in which negotiations and betrayals are routine, some transposable elements silently defect to play active roles in the host genome camp. This mechanism called exaptation has been observed in a handful of retroposons that, after acquiring selectable mutations during the course of evolution, became transcriptional enhancers of mammalian genes. The neuronal POMC enhancers nPE1 and nPE2 are two of the five known and functionally verified cases of earlier retroposons that were exapted into transcriptional enhancers., The incipient list also includes an LF-SINE exapted into a developing brain enhancer of Isl1, an AmnSINE1 that evolved into a brain enhancer of Fgf8 and the more recent discovery of another AmnSINE1 exapted as an enhancer of the transcription factor Satb2 which participates in cortical brain development. Although these examples are sparse, we believe they represent the tip of an emerging iceberg that will buttress the theory developed in the past two decades that mobile genetic elements contributed extensively to gene and genome evolution as proposed many years ago.-

Because the mechanisms developed by each phylogenetic branch to inactivate and/or eliminate retroposons have been unequally effective on different types of transposable elements, particular retroposon families have been able to resist genomic surveillance, retain transpositional activity and, therefore, colonize their host genomes in large copy numbers. For example, members of the CORE-SINE MAR1 family are very abundant, and even still active, in marsupial genomes., Similarly, MaLR THE1B retroposons are present in high copy number in primate genomes suggesting that they were particularly active until recently. The species-specific or order-specific abundance of particular families of retroposons in combination with the development of a bioinformatic algorithm that we named “in silico paleogenomics” has helped us to identify the evolutionary origin of nPE1 and nPE2. This method, based on the Blast algorithm, incorporates an evolutionary model to the search strategy that allows the identification of lower identity homolog sequences such as mobile elements that were co-opted by the host genome after suffering numerous mutations and becoming DNA relics from the once active transposable elements., We believe that the use of in silico paleogenomics, or other similar strategies, will allow the identification of the origin of many transcriptional enhancers that were exapted throughout evolution to provide adaptive complexity and richness to the regulation of gene expression.