In Chap. 5 (Vertebrates) and Chap. 6 (Human) we pointed out that in all vertebrates in a middle period of their embryonic development a similar basic body construction, a conserved ‘Bauplan’ (construction plan, body plan), can be recognized. Conserved characteristics are: dorsal neural tube, notochord, somites, branchial pharynx with gill pouches, cartilaginous pharyngeal arches (gill arches), from 4 to 6 gill arteries, and a ventral heart. In their subsequent ontogeny these basal characteristics develop quite differently in class-, family- and eventually in species-specific ways, reflecting the phylogeny.

Vertebrates of all taxonomic groups transitorily pass developmental stages reminiscent of their common ancestor. Temporarily it appears as if the tiny human embryo (1–20 mm in size) would be about to develop into a fish, subsequently into a salamander, a reptile, a monkey and finally into a human being (see Figs. 6.​24 and 6.​9). Such observations prompted anatomists and zoologists of the eighteenth and nineteenth century such as Carl Ernst von Baer and Ernst Haeckel to emphasize these astonishing similarities and to formulate ideas and hypotheses which culminated in Haeckels much debated “basic biogenetic law” – which in fact is a rule – stating ontogeny be an abbreviated form of phylogeny (see Sect. 6.​6). However different the very early and the very late developmental stages may be, in the middle period a phylotypic stage (Figs. 5.​9 and 6.​8) can be recognized in which the embryos are amazingly similar to each other (Fig. 6.​17). The concept of phylotypic stages has first been applied to vertebrates but phylotypic stages can also be defined in other groups. In insects, for example, the stage of the extended germ band is considered to represent the phylotypic stage. In molluscs it is the veliger larva, in echinoderms the gastrula or the pluteus larva can be considered to be phylotypic though, or because, these pre-metamorphic stages display bilateral symmetry thus representing the ancestral primordial trait of the phylum.

Back to vertebrates: Already the fathers of developmental biology not only emphasized common traits but also pointed out that it is just the earliest phase of development in which development proceeds highly differently and in class-specific patterns, being in mammals very different compared to amphibians. For instance in human embryogenesis, very early preparations are made for developing the placenta, even before the embryo proper is formed. In the blastocyst the outer layer becomes an extraembryonic precursor of the placenta, called the trophoblast (Greek: trophein = to nourish) as it serves to take up nutrients provided by the uterus of the mother. Accordingly, the embryo does not arise from the entire ball-shaped embryo, called the blastula in amphibians and blastocyst in mammals; instead in the mammalian blastocyst the embryo actually forms from a small area in the interior (see Fig. 6.​5). Such differences reveal adaptations to habitats as different as the pond and the womb.

In all vertebrates a transitory notochord is formed prior to the vertebral column. This apparent detour has a practical function. The notochord assists in mechanically stretching the head-tail axis and it is an indispensable emitter of development-organizing signals. It orchestrates the formation and subdivision of the spinal cord (Fig. 16.​1), summons the cells of the somites to migrate and to assemble around the notochord where they form the vertebral bodies (Fig. 10.​6). The notochord even assists in determining the position of inner organs such as liver and pancreas. Ontogenetic inputs and presetting cannot arbitrarily be modified; the processes are too complex and must take place in a sequence of progressive steps.

Another example of constraints: In all vertebrates capable of flight, pterosaurians, birds, and bats, wings could only be developed at the expense of the forelimbs. The gift of grace of another extra pair of extremities remained reserved for mythological figures such as dragons and winged unicorns, and the celestial angels. The range of attainable morphologies is confined.

Fossil records provide evidence that the basic traits of all recent body plans already existed in the Cambrian period about 540–480 million years ago. They thus arose from a putative common ancestor that already existed previously. Fossils in the form of imprints resembling recent medusae are known from the Precambrian (Vendium, 640 million years ago) but fossil records do not provide information on how these constructions once arose, simply because residues of soft tissue do not persist or merely persist as imprints in rock. Clues to the earliest history of life can only be deduced from the developmental pathways and the inventory of conserved genes of recent animals.

The remarkable congruency of basic developmental mechanisms and gene expression patterns in organisms belonging to different taxa gave a new lease of life to an area of research that became known as “evo-devo” biology and increasingly attracts attention. Though awareness that ontogeny partly recapitulates phylogeny enabled and stimulated synergies between developmental and evolutionary biology for a long time, only the comparative analysis of gene expression patterns and regulatory networks merged these complementary disciplines on a broad level.

With entire genomes sequenced from a yearly increasing number of animals that belong to different taxa it is becoming apparent that type and complexity of a body plan does not correlate with the total number of genes. Evolution generates duplications of key developmental genes, change of function of the duplicated copies, and altered integration into regulatory networks. For example changes in the body plan may be caused be the acquisition of new binding sites for transcription factors (as in case of lens proteins, see below) or by an altered time span or time sequence by which signalling molecules, receptors or other proteins are produced and active.

Comparative analyses including many different species show that a molecular “toolkit” has existed since the early phase of history. Already at the level of the Cnidaria this toolkit contains almost the entire stock of instruments that enabled the triploblastic animals to develop an amazing diversity of body plans. (But of course, as this basal toolkit as such did not per se give rise to triploblasty in cnidarians, the kit must contain additional tools in the triploblastic clades.) According to the hypothesis of a common toolkit some genes are expected to fulfil similar tasks in morphologically different organisms. There is hope that by detecting such conserved genes and identifying their function it might be possible to define an ancestral “zootype” – an ancestral prototype animal expected to be similar to the common ancestor of all multicellular animals, the Metazoa.

In contrast to classic developmental biology with only few but intensely investigated model organisms evo-devo biology analyses a broad spectrum of recent taxa and species that are positioned close to branching points in the phylogenetic tree or whose position is a matter of debate. Alternatively, or in addition, the function of developmentally relevant proteins is compared across the entire animal kingdom. By comparing the expression patterns of key marker genes and the regulatory interactions of gene products, it is possible to identify traits of common ancestors at different evolutionary levels. In parallel, in comparing gene sequences by means of computer programs (algorithms), phylogenetic relationships can be deduced, compared and matched with existing pedigrees construed on the basis of morphological or biochemical traits (Fig. 22.1).

Functionally highly conserved gene products are, as a rule, integrated in conserved molecular mechanisms of regulation and control similar structures. On the other hand, molecular biology also reveals convergences by showing that the morphology and function of a structure can be attained and ensured by different ways. Thus, in the course of evolution, different molecules may have been utilized to generate, for example, a refractive eye lens.

Before we consider the procedure and results of evo-devo research by discussing actual examples we first summarize the basic assumptions underlying the theory of evolution and discuss problems the researcher is confronted with, whether he/she is trained in molecular biology or/and in zoology.

Evolutionary changes have their origin in changes in the genome. Relationships and changes in morphology and biochemistry should be reflected in the genome and in the spectrum of the encoded proteins. These, in turn, are the building blocks which bring about the changing morphological traits in ontogeny and the eventual lasting phenotype. Thus, the discovery of the Hox genes and their putative proto-Hox precursors in all multicellular animals (Metazoa) investigated so far pointed to a parallel molecular and morphological developmental sequence correlated with the increasing complexity of body plans (see Sect. 22.4).

Homology in Morphology. When morphologists speak of homology they mean affiliation of a structure to an evolutionary developmental sequence. Homologous anatomical structures in various organisms derive from a common original stem structure. Criteria for homology in the sense of traditional morphology/anatomy are:

Homology in Molecular Biology When molecular biologists and biochemists speak of homology they mean high sequence identity in genes or polypeptides, respectively. Nobody doubts that genes displaying sequence identity over many hundreds or thousands of nucleotides derive from a common primordial gene. In genes or proteins, respectively, with low sequence similarity but similar function, the three-dimensional folding structure of the protein might have been preserved; but in this case convergence must be also taken into consideration. Moreover, recombination of DNA modules by exon shuffling may cause partial homology to several different genes or proteins (see below Sect. 22.3.1) and thus make reconstruction of a phylogenetic tree a difficult task.

Does homology at the level of genes not per se imply homology at the level of the phenotype? Unfortunately it does not! Four examples will provide proof.

In general terms, homology means: two entities have a common origin in the history of life. For example in various vertebrates, say zebrafish, grass frog, and hummingbird, the spinal column and the forelimbs (pectoral fin, foreleg, and wing) are homologous to each other. In a restricted sense one speaks of orthologous structures, meaning structures derived from a common ancestor and found in different species. Yet, within one and the same individual the somites and vertebral bodies are also homologous repetitions of the same basic theme. Likewise, in tetrapods forelimbs and hindlimbs are variations of a basic structure – one speaks of serial homology or of paralogous structures.

In parallel the molecular geneticist finds homologous genes in different species and taxa. The Hox gene labial, for instance, exists in Drosophila and the mouse, and likewise the gene Antennapedia. Both genes are members of the Hox cluster and arose from a common primordial gene, that was serially duplicated, and in annelids, arthropods and vertebrates is found in the anterior part of the Hox cluster (Figs. 22.2 and 22.3). One speaks of

Genetic relationships between organisms can reliably only be reconstructed when orthologous genes are compared! To reconstruct a phylogenetic tree of the Hox cluster, for example, the labial gene of a given species must be compared with the labial gene of other species, or the Antennapedia genes with Antennapedia genes of other species, but not labial genes with Antennapedia genes. (A comparison of paralogues can be used to reconstruct the branching routes within the Hox cluster). Orthologous key genes with conserved functions, in particular orthologous master- or selector genes, point to common principles in the control mechanisms.

If natural relationships and evolutionary changes are reflected in the genomes we may expect that genes exist which vary to some extent from species to species but are conserved on the whole. On the other hand, we also expect novelties, for humans are neither hummingbirds nor insects.

Gene synthesis de novo is not known up to date (at the utmost genes might arise from non-coding sequences, see framed text in Sect. 22.3.3). So how can novelties emerge?

Basic mechanisms creating something new may be comprehensible when we consider how conserved gene clusters with several paralogs and how multidomain proteins arose. Based on the crystalline examples above, we may also envision how non-homologous proteins can be used to optimise the function of homologous structures.

Sequence analysis reveals in many proteins a modular composition with similarity to different functional domains of other proteins. In particular, a modular composition is evident in multimodal growth factor receptors, which combine ligand-binding and enzyme activity. A growth factor receptor of the FGFR family (see Fig. 11.​8) for example possesses an extracellular ligand-binding domain, a transmembrane domain, various intracellular docking sites, and a kinase domain. The partial sequence similarity to receptors of other types and families suggests that these proteins are composed of modules. In the course of replication and recombination new genes were created in evolution by recombining segments of already existing genes. A plausible and widely accepted hypothesis proposes exon–shuffling as a mechanism of recombination. As a rule, genes of eukaryotes are interrupted by several interspersed introns. Exon-intron boundaries frequently correlate with those of protein domains. Through recombination of initially separate exons highly complex multimodal protein might have arisen (a model of proteins consisting of segments derived from recombined DNA segments are the antibodies, see Fig. 18.​8). According to recent reports about the origin of the growth factor binding receptor tyrosine kinases (RTK) first events of recombination took place already in the common ancestor of multicellular animals and their sister group, the choanoflagellates. Such processes warn us about using partial sequences in genes and proteins as sole indicators of homology. On the other hand comparing partial sequences found in extant organisms enable reconstruction of the evolution of modular proteins using algorithms of bioinformatics, and to correlate the result with phylogenetic trees constructed based on morphological, biochemical and genetic criteria.

In the genome of all species investigated so far DNA sequences of apparently foreign origin are found. Frequently foreign (heterologous) DNA is transferred by viruses from one organism into another organism, also irrespective of borders between species or even phyla. Such sequences persist in the form of transposons or sequences partially reminiscent of transposons. Their viral origin is evident or suggested by characteristic subsequences. Even in non-functional remnant sequences partial sequences coding for viral envelope proteins (gag, env) or other typical viral sequences (pol, ORF) or characteristic palindromic flanking sequences remain traceable testifying their viral origin. According to quantitative estimations transposons and related sequences sum up to about 45 % of the human genome. They are commonly looked upon as junk DNA, but among the sequences imported into our genome sadly there also are cancer-generating oncogenes (Chap. 19). Beneficial for humans is, on the other hand, the likely cooption of transposons in lymphocytes. The many short sequences framed by palindromic sequences are used to program the variable, antigen-binding parts of B-cell receptors and hence of antibodies by somatic recombination (Fig. 18.​8). The same applies for the variable domains of the T-cell receptors. Apparently, transposons were put into service of the adaptive immune systems during evolution of the vertebrates. In biotechnology, transposons are regularly used as vectors to produce transgenic animals.

When the first genomes were sequenced, an astonishing number of conserved genes was found to be shared by the fly Drosophila, the round worm Caenorhabditis elegans and man. On the other hand in the human genome genes were detected that also were found in other mammals but not in the fly or the worm. These genes therefore were counted as mammalian-specific. Yet, many of the allegedly mammalian or vertebrate-specific genes were subsequently identified in the genomes of various members of the phylum Cnidaria (Acropora, Nematostella, Hydra, Hydractinia) as well as in the genome of the annelid Platynereis dumerilii. Apparently these genes were already present in the common ancestor of the Cnidaria and Bilateria but were lost in the branch leading to the Ecdysozoa including Drosophila and C. elegans.

Once more it must be emphasized that phylogenetic relationships can be reconstructed reliably only by comparing multiple genes from many different taxa. These restrictions must be taken into account in any evo-devo projects and reconstructions of phylogenetic trees.

In many animal groups the WNT signalling system is concerned with the establishment of the primary body axis, or in bilaterally symmetric animals, the establishment of one of the three body axes, the longitudinal anterior-posterior axis. Signalling centre is the posterior pole (Fig. 22.6). Moreover, the system is concerned with the control of stem cells in all animals investigated.

If one wants to understand how body plans developed from the early to the complex animals it is reasonable to look at animal taxa that early in evolution diverged from the main lineage of the tree leading to bilateral symmetric animals. It is obvious that presence of a conserved gene in representatives of early side branches and also in the Bilateria lineage indicates its presence in the inventory of the last common ancestor that existed before the branches diverged. Among the taxa that branched off early in evolution are the Porifera (sponges), the Placozoa represented by the tiny two-layered Trichoplax, and the Cnidaria with Clytia, Hydra, Hydractinia, the coral Acropora, and the sea anemone Nematostella. The majority of multicellular animals share an indirect development with a simple motile ciliated larva. This larva has an anterior end pointing in the direction of forward movement, and has an elongated body. Whether this longitudinal axis is called antero-posterior axis, oral-aboral axis, or head-tail axis is a matter of convention and tradition. It is this main longitudinal body axis that is established under the regime of the canonical WNT/β-catenin system (Fig. 9.​5).

The Wnt/β-catenin signalling pathway (see Fig. 11.​6), once detected as a signalling system that is required for correct outgrowth of wings in Drosophila, turned out to be essential for the establishment of an embryonic body axis in all animals investigated with respect to the genetic basis of axis formation, except in Drosophila itself. WNT signalling molecules are expressed early in embryogenesis – in part by maternal, in part by zygotic genes. In previous Chapters we referred to the examples of the hydroid Hydractinia (Fig. 9.​5), sea urchin (Fig. 4.​6) and Xenopus (Figs. 5.​18, 5.​19, and 10.​4) and found a common principle: The WNT system directs and patterns the body axis that extends from the blastopore to the opposite pole (Fig. 22.7). The head-tail axis (a-p axis) in vertebrates can only be implemented correctly when WNT is emitted from the posterior pole at the blastopore, and on the other hand at the anterior pole anti-WNT factors neutralize the WNT signals thus enabling head formation. Working in the opposite direction, WNT and anti-WNT gradients share responsibility in the subdivision of the body into head, trunk, and tail. However, when we compare representatives of different taxa we see an important difference. While in triploblastic animals WNT is expressed with maximum intensity at the posterior pole and excluded in the head region, in Cnidaria we see the obviously inverse situation: Wnt is expressed at the head pole. Such an inversion is not detected when we take the blastopore as point of reference and invert the polyp upside down: wnt expression indicates that Hydra is standing on its head.

Unexpectedly, already early in phylogeny an ample collection of WNT signalling molecules occurred through repeated duplication of the wnt gene. Already in Cnidaria 11 wnts are present! The successive and overlapping expression of all these WNT ligands along the body axis in Nematostella points to a potential function in the subdivision of the body column into functionally different zones (Fig. 22.8). The canonical WNT system apparently is a system that was acquired early in the history of life for specification and regionalization of the primary body axis. In triploblastic animals this system was supplemented or replaced (Drosophila) by HOX transcription factors.