Gene regulation and evolution

In 1975, Mary-Claire King and Allan Wilson observed that humans and chimpanzees share approximately 99 percent of their protein-coding DNA, yet differ dramatically in anatomy, cognition, and behaviour. They proposed that the evolution of form and function depends not primarily on changes in genes themselves but on changes in how, when, and where those genes are expressed.¹ In the decades since, research in comparative genomics, evolutionary developmental biology, and functional genomics has confirmed that regulatory evolution is a central mechanism of phenotypic diversification across the tree of life.

The regulatory genome

Gene expression in eukaryotes is controlled by a complex hierarchy of regulatory DNA sequences. Promoters, located immediately upstream of coding sequences, recruit RNA polymerase and basal transcription factors to initiate transcription. Enhancers, which can be located thousands or even millions of base pairs from their target genes, bind transcription factors and modulate the timing, location, and intensity of gene expression. Silencers repress transcription, while insulators prevent enhancers from activating inappropriate genes.²

The FANTOM5 consortium has catalogued hundreds of thousands of active enhancers across human cell types, revealing that the regulatory genome is far larger and more complex than the protein-coding genome. While approximately 20,000 protein-coding genes account for roughly 1.5 percent of the human genome, regulatory elements may comprise 5 to 10 percent or more of total genomic DNA.¹³ The chromatin landscape adds an additional layer of regulation, with histone modifications and chromatin accessibility determining which regulatory elements are active in any given cell type or developmental stage.¹⁴

Modularity and evolvability

A key property of cis-regulatory elements is their modularity. Individual enhancers typically control gene expression in a specific tissue, developmental stage, or environmental context, and a single gene may be regulated by dozens of independent enhancers. This modular architecture means that a mutation in one enhancer can alter gene expression in one tissue without affecting the gene's function in other tissues where different enhancers drive expression.⁹

Sean Carroll has argued that this modularity makes cis-regulatory evolution the primary mechanism for morphological change because it minimises pleiotropic effects. A mutation in a protein-coding gene alters the protein everywhere it is expressed, potentially disrupting multiple functions simultaneously. By contrast, a mutation in a tissue-specific enhancer changes expression only in that tissue, leaving all other functions intact. This asymmetry means that regulatory mutations have a higher probability of being beneficial or neutral rather than deleterious, giving them a selective advantage as substrates for evolutionary change.^{9, 10}

Regulatory evolution in action

Some of the most compelling evidence for regulatory evolution comes from convergent evolution studies where the same regulatory changes have occurred independently in different lineages. In threespine sticklebacks, freshwater populations have repeatedly lost pelvic spines through deletion of a tissue-specific enhancer of the Pitx1 gene. The coding sequence of Pitx1 remains intact and continues to function normally in other tissues, but the loss of its pelvic enhancer eliminates expression specifically in the developing pelvic region, producing a heritable anatomical change through a purely regulatory mechanism.¹² Genome-wide studies of stickleback adaptation have since revealed that regulatory changes at dozens of loci underlie the repeated evolution of freshwater phenotypes, making threespine sticklebacks one of the best-characterised systems for studying the genetic basis of adaptive evolution.⁵

Similarly, changes in pigmentation across vertebrates have been traced to regulatory mutations in the Kitlg (Kit ligand) gene. In sticklebacks adapting to different aquatic environments and in human populations adapting to different UV environments, independent cis-regulatory changes at the same locus have produced convergent shifts in skin and scale pigmentation without altering the Kitlg protein itself.³ These examples illustrate a broader pattern: surveys of the genetic basis of morphological evolution find that cis-regulatory changes account for a disproportionately large fraction of cases, particularly for traits involving morphology and pigmentation.⁴

Human regulatory evolution

The divergence between humans and chimpanzees provides a natural experiment in regulatory evolution. Despite sharing approximately 98.7 percent of their DNA, the two species differ extensively in brain size, facial morphology, hand anatomy, and cognitive abilities. Comparative genomic analyses have identified hundreds of human-specific regulatory changes, including both gains and losses of enhancer activity.⁸ Even within Homo sapiens, regulatory variation between human populations contributes to phenotypic diversity: genome-wide mapping of active regulatory elements across European, African, and East Asian populations has revealed population-specific differences in enhancer and promoter activity that correlate with known phenotypic variation.⁷

McLean and colleagues identified 510 sequences that were conserved across mammals but deleted specifically in the human lineage. Most of these human-specific deletions removed regulatory elements rather than coding sequences. One deletion eliminated an enhancer near the androgen receptor gene that in chimpanzees drives expression of sensory vibrissae (whiskers) and penile spines, structures absent in humans. Another deletion removed an enhancer of the tumour suppressor gene GADD45G that normally limits cell division in the developing brain, potentially contributing to the expansion of the human forebrain.⁸

Enhancer evolution across species

Large-scale comparative studies have revealed that enhancer activity evolves rapidly across mammalian species. Villar and colleagues mapped active enhancers in liver tissue across 20 mammalian species and found that while a core set of enhancers is conserved across mammals, a substantial fraction of enhancers are lineage-specific, having been gained or lost during mammalian diversification. Promoter activity, by contrast, was highly conserved, indicating that the regulatory architecture of genes is hierarchically organised with rapidly evolving enhancers modulating the output of conserved core promoters.⁶

Despite rapid turnover in enhancer sequences, the underlying regulatory logic may be deeply conserved. Wong and colleagues demonstrated that the combinatorial grammar of transcription factor binding sites within enhancers is shared across animal phyla spanning more than 600 million years of divergence, even when the enhancer sequences themselves have diverged beyond recognisable homology. This deep conservation of the regulatory code suggests that the rules governing enhancer function are ancient and that regulatory evolution operates by rewiring conserved transcription factor interactions rather than inventing entirely new regulatory mechanisms.¹⁵

Transposable elements and regulatory innovation

Transposable elements, which comprise approximately 45 percent of the human genome, have emerged as an important source of new regulatory elements. When transposons insert near genes, they can carry transcription factor binding sites that create novel enhancers or alter existing regulatory landscapes. Over evolutionary time, some transposon-derived sequences have been co-opted (or exapted) as functional regulatory elements, a process sometimes called regulatory domestication.¹⁶

Chuong and colleagues showed that endogenous retroviruses, a class of transposable elements, were co-opted to rewire the innate immune regulatory network in mammals. Ancient retroviral insertions near immune genes provided binding sites for interferon-responsive transcription factors, bringing those genes under the control of innate immune signalling pathways. This mechanism of regulatory innovation through transposon co-option represents a form of evolutionary tinkering in which pre-existing mobile genetic elements are repurposed to generate new patterns of gene expression.¹⁶

Regulatory versus coding evolution

The relative contributions of regulatory and coding mutations to evolutionary change remain debated. Stern and Orgogozo surveyed hundreds of cases where the genetic basis of morphological evolution was known and found that cis-regulatory changes were responsible in a majority of cases involving morphological traits, while coding changes predominated for physiological and biochemical adaptations such as toxin resistance, antibiotic resistance, and metabolic shifts.⁴

This pattern makes biological sense. Morphological traits are typically controlled by genes expressed in multiple tissues during development, where regulatory mutations offer tissue-specific changes with minimal pleiotropic cost. Biochemical traits, by contrast, often depend on the specific properties of a protein, such as its binding affinity for a substrate, where only a coding change can produce the required functional modification.^{4, 11} The emerging consensus is that both regulatory and coding evolution contribute to adaptation, but their relative importance depends on the nature of the trait under selection and the pleiotropic constraints on the underlying gene.

Conserved non-coding elements

Comparative genomics has revealed that a substantial fraction of the non-coding genome is under evolutionary constraint, indicating functional importance. Lindblad-Toh and colleagues aligned the genomes of 29 mammalian species and identified approximately 3.6 million constrained non-coding elements, totalling roughly 4.2% of the human genome—nearly three times the amount of DNA under constraint in protein-coding exons.¹⁷ Many of these conserved non-coding elements function as enhancers, silencers, or insulators, and their deep conservation across tens of millions of years of mammalian evolution indicates that they are essential for normal development and physiology.

The existence of ultraconserved non-coding elements—sequences hundreds of base pairs long that are identical in humans, mice, and rats despite over 80 million years of independent evolution—suggests that some regulatory functions are so critical that even single nucleotide changes are lethal or severely deleterious. These ultraconserved regions are strongly enriched near genes involved in embryonic development, neural differentiation, and transcriptional regulation, consistent with the hypothesis that the most conserved regulatory elements control the most fundamental aspects of body plan specification.¹⁷

Paradoxically, the rapid turnover of enhancers coexists with this deep conservation. A relatively small core set of regulatory elements is shared across all mammals and appears to be indispensable, while a much larger set of lineage-specific enhancers arises and disappears on shorter evolutionary timescales. This two-tiered regulatory architecture—a conserved core supplemented by rapidly evolving peripheral elements—provides a mechanism for maintaining essential body plan features while permitting the fine-tuning of gene expression patterns that underlies species-specific adaptations.^{6, 15, 17}