Evolutionary Developmental Biology (Evo-Devo): Past, Present, and Future

I go back to the late nineteenth century when we find the origins of evo–devo in the research of individuals in England (largely Trinity College, Cambridge) and in Continental Europe. These evolutionary morphologists/evolutionary embryologists were attracted to this research following the publications of The Origin of Species by Charles Darwin (Darwin 1859) and Ernst Haeckel’s theory that ontogeny recapitulates phylogeny (Haeckel 1866). Paradoxically, the first published study testing Darwin’s theory using embryos and larvae—Fritz Müller’s study of crustacean life histories (Müller 1864)—showed that ontogeny could be used to understand patterns of evolutionary history (phylogeny) and that mechanisms could be sought in ontogeny. So varied were crustacean life history strategies found to be that Müller found he could use the details and varieties of life history stages to construct a phylogeny of crustacean relationships. Haeckel took exactly the opposite position. Haeckel theorized that phylogeny explains ontogeny and erected his Biogenetic Law on this basis.

Frustration with reconstructing evolutionary trees from embryonic sequences, the rise of experimental and physiological approaches to embryonic development in the 1880s, and the rediscovery of Mendelian genetics in 1900 all cast evolutionary embryology into a backwater from which it would take a century to resurface. Mendel’s principles of segregation and assortment coupled with studies on the fruit fly Drosophila provided a powerful foundation upon which the new science of genetics was built. Publication of Dobzhansky’s (1937) influential book, “Genetics and the Origin of Species,” provided a basis for understanding evolution through population genetics, the mathematical models for which have been developed in the 1920.

By the middle of the twentieth century, maintenance of the features of organisms, variation in those features, and the origin of new features all seemed explicable by a fusion of Mendelian and population genetics. Paradoxically, the use of Drosophila as the model organism for genetics eliminated the roles of embryonic development and of the environment from evolutionary discussion and theory; inbred laboratory organisms display none of the variation and adaptability seen in nature.

Despite the valiant attempts of a handful of individual researchers through the first six decades of the twentieth century, linking embryos and evolution did not make a comeback until Stephen J. Gould’s book Ontogeny and Phylogeny was published in 1977. Weighing in at 501 pages, this was no easy read, especially as the first half was a detailed history of the way in which development and evolution (ontogeny and phylogeny—the terms proposed by Haeckel) had been related in the past and when the past for Gould began at 450 BC. This historical tour de force was important; Gould reminded us, for indeed we had forgotten that we had been seeking relationships between development and evolution for millennia.

The second half of Gould’s book was equally important because it revisited a concept conceived by Haeckel in the 1860s, elaborated by Gavin de Beer in the 1930s (de Beer 1930), but then all but abandoned. Haeckel and de Beer’s concept is summed up in the word “heterochrony”—changes in the timing of developmental processes between a descendant and its ancestors (see Zelditch 2001; Willmore 2010). De Beer’s insight was to take studies showing that genes affected the rates of physiological processes in insects (Goldschmidt 1918) and to see that the rates of development must change during evolution. Here was a way in which comparative embryology could be investigated in an evolutionary (phylogenetic) framework). Heterochrony as a developmental mechanism operating during individual development could be selected for and so was important in evolutionary change; heterochrony was an “evolutionary developmental mechanism” (Hall and Olson 2003).

Julian Huxley, one of the founders of the modern synthesis of evolution, appreciated the importance for evolution of the study of genes in development—“a study of genes during development is as essential for an understanding of evolution as are the study of mutation and selection” (Huxley 1942, p. 8)—but even Huxley neglected to incorporate development into evolutionary theory. Other founders of the modern synthesis such as Ernst Mayr spoke of an internal biology (to which development belonged) and an external biology (evolution, ecological interactions) as if the two were separate and non-overlapping/interacting one to the other (Mayr 1982, 1997).

Evo–devo exploded as heterochrony was found everywhere. Along the way, heterochrony became such a pervasive term that it lost some of its explanatory power; any change in timing became heterochrony, whether evolutionarily relevant or not (Zelditch 2001, especially pp. vii–ix).

The next major impetus to evo–devo was not the resurrection of a previously known evolutionary developmental mechanisms but the discovery that all animals (subsequently shown for all plants and fungi too) share genes that contain a 180-bp sequence known as the homeobox and that these genes, known as homeobox, homeotic, or Hox genes, are responsible for determining that animals have an anterior and a posterior, a dorsal and a ventral side, and specific regions (often as repeated segments) along the body axis—head at one end tail at the other, thorax in front of abdomen, wings on a specific pair of segments, and so forth (Lewis 1978; Gehring 1985, 1998; Averof 1997; Grenier et al. 1997; Carroll 2008).

“Master” genes, also known as developmental or regulatory genes, were discovered. One of the best understood of such genes is the paired-box protein gene known as Pax-6 in vertebrates and as eyeless (ey) in Drosophila. As determined from its DNA sequence, orthologues of Pax-6 are present throughout the Animal Kingdom. As determined from functional studies, the role of Pax-6 in anteriorizing the embryo and in the formation of anterior sensory structures also is conserved across the Animal Kingdom. Although best known as the major gene controlling eye development, Pax-6 functions in organisms that lack eyes, reflecting its ancient developmental role. Indeed, ey from fruit flies can initiate eyes in frogs and Pax-6 from frogs can initiate eyes in Drosophila (Dahl et al. 1997; Suga et al. 2010). Here is an astonishing and previously unthought-of genetic conservation across animals whose morphology varied enormously.

Research programs in evo–devo are comparative and experimental (Hall 1999; Raff and Love 2004). Increasingly, they involve what Laubichler (2007) calls “evolutionary developmental genetics.” In pursuit of their analysis of the integration of development, evolution, and ecology, Collins et al. (2007) analyzed three model systems—breeding of the domestic dog Canis familiaris, research on a sea anemone Nematostella vectensis, and research into horn development/evolution in dung beetles in the genus Onthophagus—and the development/evolution of vertebrate limbs and mammalian teeth. In a very different approach, McCain (2010) identified trends in evo–devo between 1996 and 2008 by analyzing linkages between the literature in what she identified as three core journals of evo–devo [Evolution & Development; Development, Genes & Evolution; and the Journal of Experimental Zoology (Part B)]. Her analysis is worth a close examination, as is an earlier study in which the research of Conrad Waddington was used to trace the rise of evo–devo (McCain 2009).

Critical to our ability to answer these questions and evaluate these model systems are a sound basis to ensure that we are comparing the same organisms/processes/genes—the central biological concept of homology (Hall 1994, 2012)—and our ability to determine the direction of evolutionary change because of the development of robust phylogenetic trees of relationships (Valentine 2004; Erwin et al. 2011). We have the model systems and we have the methodology but still lack the required theory. In their analysis of the status of modeling in evo–devo, Collins et al. (2007) concluded that “these models are mostly diagrammatic and functional; very few analytical and predictive models exist within EvoDevo” (p. 373). In part, they see this deficiency residing in evo–devo’s reliance on developmental genetics rather than evolutionary theory, a position that is slowly changing.

Molecular genetics has revolutionized evo–devo over the past two decades. Integrating our expanding molecular understanding with mechanisms operating at the cell or other levels (tissues, organs, whole organism, organism–environment interactions; Box 1) has been and remains a major goal and a challenge for evo–devo. Essentially, it involves opening the black box between genotype and phenotype, taking out what is found in the box and how it fits together and then determining how to put the contents back in the box (Hall 1999, 2003a, b; West-Eberhard 2003; Carroll et al. 2005). One of the items found in the black box is known as epigenetics.

Many of the controls on gene regulation and function are subsumed under the term “epigenetics,” a term coined by the British geneticist and embryologist Conrad Waddington for the causal factors that control gene action during development (Waddington 1940; and see the papers in Hall and Laubichler 2009). Hall (1992) defined epigenetics as “the sum of the genetic and non-genetic factors acting upon cells to control selectively the gene expression that produces increasing phenotypic complexity during development” (1992, p. 89). To this, I would only add “and evolution” at the end.

Epigenetics is still used with this original meaning but has increasingly come to be applied at the molecular level for heritable changes to the DNA other than changes to the nucleotide bases. Such changes include methylation, imprinting, and regulation of chromatin—which have been known for some time (Biémont 2010; Hallgrímsson and Hall 2011; Moazed 2011; Molaro et al. 2011)—and regulation of genes by small RNA molecules, especially miRNAs (Kosik 2009; Hallgrímsson and Hall 2011). This now very well-characterized second inheritance system was appreciated in its infancy by John Maynard Smith, a leading twentieth century evolutionary theorist: “There is a second inheritance system—an epigenetic inheritance system—in addition to the system based on DNA sequence that links sexual generations (Maynard Smith 1989, p. 11).

The heritable aspect of epigenetics has shown us that organisms do not start their lives as naked nuclear DNA. They possess DNA in their mitochondria, epigenetic “marks” in their nuclear DNA, and they inherit mRNA and proteins that were produced under the control of their mother’s DNA and deposited into the egg cytoplasm. Epigenetics provides another, but not the only other, means by which heritable information operates in organismal development. The ability to learn behaviors, interact with environments, and construct niches (Laland et al. 2008; Gissis and Jablonka 2011) are three further mechanisms introduced below.

Integrated studies using molecular biology, molecular genetics, developmental biology, phylogenetics, paleontology, and molecular paleobiology are revealing previously unimagined information on how features change during evolution (Erwin and Wing 2000; Hall 2002; Wilkins 2002, 2007; Carroll et al. 2005; Peterson et al. 2007; Raff 2007; Erwin et al. 2011).

One instance is exemplified by the rise of paleobiology as a discipline that brought evolutionary theory back into paleontology and incorporated developmental, phylogenetic, and environmental approaches into a biological perspective of fossils. The journal Paleobiology celebrated its 25th anniversary in 2000 with a special issue, Deep Time: Paleobiology’s Perspective, published as a supplement to volume 26. An edited volume of 26 essays by leading paleontologists and published in 2009 demonstrates the breadth and depth of insight achieved by paleobiology as “The Paleobiological Revolution” (Sepkoski and Ruse 2009).

We have long appreciated that development is responsible for introducing variation at the level of the individual (Darwin 1859; Thomson 1988; Hallgrímsson and Hall 2005; Salazar-Ciudad 2006; Fusco and Minelli 2010). Much of the research in evo–devo, however, has been conducted on features of organisms that characterize particular groups and whose evolutionary origin was a major departure (transition is the term often used for the change, novelty, or innovation for the character) from prior characters (Brylski and Hall 1988a, b; Hall and Kerney 2012). Important examples include feathers and the origin of birds/flight (Prum and Brush 2002; Xu et al. 2011), flowers and the origin of land plants (Niklas 1997; Leliaert et al. 2011), wings and the origin of insects (Carroll et al. 1995; Abouheif and Wray 2002), and shells and the origin of turtles (Gilbert et al. 2001; Rieppel 2001; Willmore 2010). Indeed, some researchers see the major contribution of evo–devo (and its major contribution to expanding the modern synthesis of evolution) as being to provide a theoretical basis from developmental biology for the origin of novelties (Müller and Wagner 1991; Müller 2007a, b Pigliucci and Müller 2010).

New groups of organisms, as represented by new classes or orders, can arise slowly through the gradual accumulation of new characters or can arise more rapidly through key innovations (the origination of wings or lungs, for example) or through coordinated changes in different characters as seen in the origin of lungs, middle ear ossicles, and the transformation of fins to limbs at the origin of the tetrapods (Thomson 1988; Shubin et al. 1997). This is how microevolutionary changes within species are linked to macroevolutionary changes, as reflected in levels of classification: “the careful analyses of the differences in pathways between organisms of known phylogenetic relationship” (Thomson 1988, p. 138).

The list of mechanisms summarized in Box 1 implies that evolutionary developmental mechanisms are not all found in the genes, although all have a genetic basis. This is because new mechanisms emerge as development proceeds. Evolutionary developmental mechanisms may be genetic, cellular, developmental, physiological, hormonal, or any combination of these. Embryonic development is hierarchical, with new properties and mechanisms emerging as development unfolds, each dependent on the stage/processes preceding them. The single cell that is the fertilized egg cannot show any of the cell-to-cell interactions that characterize the multicellular embryo, some of which come about because cells take up new positions in the embryo through active migration. The recent elucidation of gene networks is providing the regulatory link between the genotype and cellular modules (Davidson 2006; Davidson and Erwin 2006; Wagner et al. 2007; Wilkins 2007). Embryonic inductions, tissue, organ, and functional interactions link cellular activity to the phenotype. The onset of embryonic movement ushers in a new type of process, which is interactions between developing tissues and organs such as bones and muscles (Box 1; and see Hall 1999, 2003a, b; Wilkins 2002; Hall and Olson 2003).

Some of the most striking information has come from the discovery that different morphological types of some single species (often referred to as morphs) arise following interactions with individuals of other species. These range from species that live in societies with different life history forms [chemical signaling between individuals determines the balance of workers and soldiers in ant colonies (Nijhout 1999); chemicals produced by predatory tadpole shrimp induce the formation of a protective helmet and enlargement of the head in the water shrimp Daphnia (Petrusek et al. 2009)]; the density of tadpoles and/or the amount of food promoting the development of large cannibalistic tadpoles in such species as the New Mexico spadefoot toad Scaphiopus multiplicatus (Pfennig 1992; Ledón-Rettig and Pfenning 2011); moth larvae-mimicking leaves of catkins in response to seasonal levels of tannin produced by oak trees when either catkins or leaves are on the trees (Greene 1989, 1999); and interactions of bacterial viral and eukaryotic species in the human gut with one another and the role of this microbiota in maintaining the health of their human hosts (Clemente et al. 2012).

Of course, it is impossible to tell what the future of evo–devo will be or what evo–devo will bring to the future. Nevertheless, given the past dramatic history of links between development and evolution, it is interesting to speculate on the future.

It could be business as usual, with evo–devo continuing to inform us of how changes in development relate to changes in evolution. Or evo–devo may change.