10.1 The toolkit is not a body plan #
10.1.1 Conserved genes, divergent uses #
Animals that look nothing like each other use a shared set of signalling pathways β Hedgehog, Wnt, TGF-Ξ²/BMP, FGF, Notch, receptor tyrosine kinases β and a shared set of transcription-factor families: Hox, Pax, Sox, Fox, T-box, basic helix-loop-helix. This is the developmental toolkit. It was discovered because a sequence cloned in the fly often had a vertebrate homologue that did a recognisably similar biochemical job. Mouse Pax6 can start an eye programme in a fly disc. The homeodomain of a Hox protein binds DNA in either animal. Conservation of the part is real and is the reason a fly chapter is in a course that also teaches limbs.
Conservation of the part is not conservation of the picture. Hox proteins assign identity to a region. They do not contain a description of a vertebra or a haltere. Where and when the gene is transcribed, and which other factors sit on those enhancers, decide the outcome. The same BMP antagonism that lets frog ectoderm become neural, via chordin, lets fly ectoderm escape Dpp, via Sog. In the frog the neural side is dorsal and BMP is ventral. In the fly the nerve cord is ventral and Dpp is dorsal. The antagonism is homologous as a pathway. The orientation is inverted relative to the mouth. That inversion revived an old anatomical suggestion, Geoffroy Saint-Hilaire’s, that a vertebrate’s dorsal side corresponds to an arthropod’s ventral side. The molecular map is a developmental fact. Where the inversion sits on a tree, and whether the common ancestor was already inverted relative to one of the living groups, is a phylogenetic question.
Co-option is the ordinary way the toolkit makes new form. A gene used in one place is wired, by a new enhancer, into another. Nothing in that sentence requires the two places to be the same organ inherited from a common ancestor.
10.1.2 Distal-less, Pax6, and false homology #
Distal-less (Dll) is required for the distal part of the Drosophila leg, and it is expressed there. It is also expressed at the distal tips of outgrowths in many other animals: vertebrate limb buds, polychaete parapodia, onychophoran lobopods, echinoderm tube feet. The survey is real. The bad inference is that insect legs, bird wings, and tube feet are therefore homologous organs. Homology of organs means inheritance of that structure from a common ancestor that had it. The last common bilaterian may have used Distal-less in some outgrowth, or in the nervous system, where the gene also works, without having possessed a jointed limb, a tetrapod limb, and a tube foot as the same body part. Later lineages can co-opt a “make a distal tip” cassette independently.
The careful term is deep homology: a shared genetic circuit underlying structures that may not themselves be homologous. Deep homology of the Distal-less outgrowth circuit is a good description of the expression data. Homology of all animal appendages is not. The same caution applies to Pax6. A shared retinal-determination network, and homologous opsins, support homology of a photoreceptive cell type and of the specification cassette. They do not mean that a camera eye and a compound eye are one organ that was merely scaled. The optics were elaborated separately on a shared molecular starting point. Saying “all eyes are homologous because of Pax6” and saying “eyes have nothing in common” are both wrong. The cassette is ancient. The organs are a mixture of inheritance and convergence, and only a tree can sort a particular pair.
10.2 Regulatory evolution #
10.2.1 Modular enhancers #
A protein used in many times and places is a bad thing to mutate. Change the coding sequence of a Hox protein or of a signalling ligand and every context that uses it feels the change. Pleiotropy makes most coding changes deleterious. Expression, by contrast, is modular. The even-skipped stripe 2 enhancer from Chapter 4 drives one stripe because it binds one combination of activators and repressors. Other enhancers of the same gene drive the other stripes. A mutation in the stripe 2 element can move, weaken, or delete stripe 2 and leave the protein, and the other stripes, intact. Cis-regulatory elements are the units evolution can edit without rewriting the gene.
Two cases make the argument concrete. Pelvic spines were lost repeatedly in freshwater sticklebacks. The coding sequence of Pitx1, a transcription factor required for hindlimb and pelvis, is intact. A pelvis-specific enhancer is deleted. Other tissues that need Pitx1 still have their enhancers, so the fish lives. A mouse that has lost Pitx1 coding function has a severe hindlimb reduction and also craniofacial defects, and it dies as a neonate. The enhancer deletion achieves the skeletal change the environment favours and avoids the pleiotropy that a broken protein would bring. In Drosophila sechellia, larval trichomes were lost through several small changes in enhancers of shavenbaby, which encodes a transcription factor that runs the trichome programme. The protein did not have to change. Several regulatory mutations of small effect added up. That is the population-genetic texture of the same idea: form often moves by many cis-acting steps.
Snakes show a correlated shift rather than a finished causal proof. An expanded domain of thoracic Hox expression, including Hoxc6 and Hoxc8, accompanies the long rib-bearing trunk and the absence of the forelimb at its ancestral somite position. The axis was extended by a change in where a code is deployed. These notes do not claim that one enhancer is the entire genetic cause of a snake. They claim that shifting a Hox boundary is the right kind of developmental cause, and that Chapter 5’s vertebral code is the code being shifted.
10.2.2 When the protein itself changes #
“Often” is not “always”. Insect Ultrabithorax represses Distal-less in the abdomen, which is part of why insects lack abdominal limbs. A brine-shrimp Ubx protein does not repress Distal-less in the same way, and the difference maps in part to the insect protein sequence, not only to where the gene is expressed. A coding change contributed to a major difference in body plan. It could do so because the new activity was still selective: this protein, in this tissue, gained a repression it did not previously have. Coding evolution is most plausible when it changes a specific interaction rather than destroying the protein for every target. Heterotypy, a change in the kind of molecule, is the name sometimes given to that route. The honest summary is the one the subject can defend. A change in regulation is the less pleiotropic and therefore the more frequent route to a change in form. It is not the only route, and Distal-less itself is on both sides of the evidence: its deployment is co-opted by enhancers, and its repression by Ubx was altered by a protein.
10.3 Heterotopy, heterochrony, and the hand-off #
10.3.1 Place and time #
Heterotopy is a change in the place where a developmental event happens. The eyespot on a butterfly wing is the example to keep. Distal-less, Hedgehog, and Engrailed, a cassette familiar from the wing margin and from appendage tips, are redeployed at a new focus in the middle of the wing surface, and concentric rings of pigment form around that focus. The eyespot is not a leg growing in the wrong place. It is an old circuit co-opted to a new address. That is heterotopy, and it is the same operation as a new enhancer.
Heterochrony is a change in timing or rate. The axolotl is a paedomorphic salamander: it becomes sexually mature while keeping larval features, external gills and a tail fin, because the ancestral metamorphic schedule does not run. The block is endocrine, at the level of thyroid-hormone signalling, and metamorphosis can be started experimentally by thyroxine. The genes for a terrestrial adult body are still there. Their schedule has changed. The webbing of a bat wing is a related shift that Chapter 7 already described mechanistically. Interdigital apoptosis is delayed or locally prevented, because Gremlin stays on, while the digits themselves prolong their outgrowth. Timing and amount both change. Heterometry, a change in the amount of a signal, is the fourth word in the set. Beak depth and width in Darwin’s finches are associated with how much BMP4 is expressed in the beak primordium; beak length is associated with a different modulator, calmodulin. BMP4 is not the sole cause of every beak, and later work made the network larger. It is a genuine example of form tracking the dose of a pathway this course has already met.
10.3.2 What these notes will not decide #
Whether two structures are homologous, which state was ancestral, and whether a particular enhancer allele spread by selection or by drift, are questions that need a phylogeny and a population. Evolutionary Biology owns that frame. Developmental biology supplies the mechanism that makes a variant morphologically possible: a toolkit gene can be redeployed because its enhancers are modular, a circuit can be shared without the organs being the same organ, and a shift in place or time is enough to change a body without inventing a new protein class. Bringing a Distal-less in situ hybridisation to a debate about homology, as if the stain were a phylogenetic character by itself, is the mistake this chapter is for.
10.4 Key terms, summary, and self-check #
Key terms. Toolkit; co-option; deep homology; modular enhancer; pleiotropy; heterotopy; heterochrony; heterometry; heterotypy; paedomorphosis.
Summary. Animals share signalling pathways and transcription-factor families, and they reuse them in different places. Distal-less at many appendage tips is deep homology of an outgrowth circuit, not a proof that all appendages are one organ; Pax6 supports a shared photoreceptor programme, not the homology of camera and compound eyes. Because enhancers are modular, a pelvis-specific deletion of a Pitx1 enhancer can remove spines without the lethality of losing Pitx1 everywhere, though coding changes such as the insect Ubx repression domain also alter form. Heterotopy moves a process in space, as in a butterfly eyespot; heterochrony moves it in time, as in the axolotl’s unfinished metamorphosis. The tree on which those changes are ancestral or derived is the neighbouring subject’s problem.
Self-check.
1. Why does Distal-less expression in a fly leg and a vertebrate limb bud not show that those appendages are homologous organs? What does deep homology mean instead?
2. Why can a stickleback lose pelvic spines by an enhancer deletion and survive, when a mouse that loses Pitx1 protein does not?
3. Give one case in which a change in protein sequence, rather than only in cis-regulation, altered which appendages form.
4. Classify the axolotl’s retained gills, and the butterfly eyespot, as heterochrony or heterotopy.
5. The fly expresses Dpp dorsally and the frog expresses BMP ventrally, with chordin/Sog on the opposite side. Which fact is developmental, and which question about that inversion belongs to Evolutionary Biology?
—
# Closing: how the chapters fit together
A fertilised egg is already polarised by what the mother put into it, and the rest of development is the embryo interpreting, subdividing, and physically executing that start. Chapter 1 supplies the tests. An isolation says “specified”; a graft that cannot respecify says “determined”; a graft that changes the host says “induced”. Every later experiment is one of those three. Chapter 2 is why the tests have something to find before the zygotic genome is on: localised maternal products, and a calcium restart that gets from metaphase II to cleavage. Chapter 3 turns yolk into a cleavage geometry, cleavage into germ layers, and dorsal ectoderm into a tube plus a crest. Without the germ-layer table, organogenesis is a list of gadgets.
Chapter 4 is the clearest cascade in the subject, and it is not a template to copy onto a frog cell. The syncytium is why Bicoid and Dorsal can be transcription-factor morphogens. The moment the fly cellularises, it uses Hedgehog and Wingless like everyone else. Chapter 5 is the cellular version: a Nieuwkoop centre induces an organiser, the organiser is a source of BMP antagonists, early Wnt dorsalises, later Wnt posteriorises, leftward ciliary flow biases Nodal and Pitx2, and a Notch clock cuts somites where an FGF wavefront allows it. Chapter 6 names the five pathways those chapters have already used, so that limb, eye, gut, meristem, and eyespot are read as re-deployments rather than as new biochemistry.
Chapter 7 is two organs done in full and two morphogenetic behaviours, migration and death, that are not “another signalling centre”. Chapter 8 is what patterning is for. Cells keep the genome, lock a regulatory state, and, in a few animals, re-read positional values after an injury. Ageing is fenced off on purpose. Chapter 9 asks the same four questions of a plant β axis, cell type, shape, stem cells β under the constraint that nothing migrates, and answers them with auxin transport, a WUSβCLV loop, a quiescent centre, and a floral combination code. Chapter 10 refuses the last easy mistake: a shared tool is not a shared organ, and a tree is required before a stain becomes an ancestor.
Revision checklist #
1. Define fate, specification, and determination by the operation that demonstrates each, and give one autonomous and one conditional specification.
2. State the fast block, the cortical reaction, and the role of calcium, and explain why a maternal-effect gene ignores a wild-type allele from the father.
3. Match yolk to cleavage type, name the mid-blastula transition, and write the germ-layer table including crest, placode, somite, and lateral plate.
4. Order the Drosophila anterior cascade and say why it requires a syncytium; put high nuclear Dorsal on the ventral side with Twist and Snail, and Dpp on the dorsal side.
5. Say what a grafted dorsal lip does, what the Nieuwkoop centre is, and why head formation needs Wnt antagonists as well as chordin, noggin, and follistatin.
6. Draw leftβright asymmetry as flow, Nodal, Lefty, and Pitx2, and somitogenesis as a Notch clock plus an FGF/Wnt wavefront.
7. For each of the five pathways, give receptor class and transcriptional output, and define a morphogen so that a one-threshold inducer fails the definition.
8. Give the limb’s three centres and the fate of the progress-zone model; give the eye as one field split by Shh, then cup, lens, and Pax6.
9. State genomic equivalence and the limit of MyoD; contrast hydra, neoblasts, and the salamander blastema with mammalian scarring; say why ageing is not a final stage.
10. Predict an ABCDE mutant; state the WUSβCLV and WOX5 loops; and explain deep homology of Distal-less without calling every appendage the same organ.
Selected sources #
These notes are an original teaching summary, not a transcript of any textbook.
- Gilbert SF, Barresi MJ. Developmental Biology. Sinauer Associates / Oxford University Press.
- Wolpert L, Tickle C, Martinez Arias A. Principles of Development. Oxford University Press.
- Slack JMW. Essential Developmental Biology. Wiley-Blackwell.
- Leyser O, Day S. Mechanisms in Plant Development. Blackwell Publishing.
- Spemann H, Mangold H. Γber Induktion von Embryonalanlagen durch Implantation artfremder Organisatoren. Archiv fΓΌr Mikroskopische Anatomie und Entwicklungsmechanik 100:599β638 (1924).
- NΓΌsslein-Volhard C, Wieschaus E. Mutations affecting segment number and polarity in Drosophila. Nature 287:795β801 (1980).
—
—
# Part C. Visual Figures & Scientific Diagrams Gallery
Comprehensive high-resolution visual reference repository for Developmental Biology from the B10logy Visual Atlas.
Amphibian Egg
—
# Part C. Visual Figures & Scientific Diagrams Gallery
Essential high-resolution reference diagrams for Developmental Biology from the B10logy Visual Atlas.