Guest post: A new link between ontogeny and phylogeny

Jeremy Purvis writes: I once had an English professor in college who began class by writing the phrase “ontogeny recapitulates phylogeny” on the blackboard. He then began to explain—in English-professor terms—how embryos pass through stages of development that closely resemble successive stages in their own evolution. Though I can’t remember exactly, I believe this introduction was used as some sort of cross-disciplinary metaphor for one of Shakespeare’s plays, in which a certain scene in the play recapitulated the entire plotline. Obviously, I learned very little biology or literature from that experience.

I did learn a little biology, however, from a recent paper (Kalinka et al. 2010.  Gene expression divergence recapitulates the developmental hourglass model.  Nature 468, 811-4, PMID 21150996). In this paper, the authors provide some hard evidence to support a longstanding observation in developmental biology: organisms in the same phylum tend to look remarkably similar during mid-development. This interval of morphological conservation is known as the phylotypic period, and can be thought of as the waist of an hourglass: embryos of different species look most different from each other during early and late development, and — oddly — least different in the middle. The notion that ontogeny recapitulates phylogeny was discredited decades ago  — though the compelling rhythm of the phrase keeps it alive in English classes —  but the idea that there is some deep similarity (on a morphological level) between embryos of related species remains.  But how could it be true that the morphologies of embryos are more similar in the middle of development than at either end? Could it be that this stage of embryonic development is particularly complex and hard to change for some reason?  Or, is the phylotypic period some kind of mass hallucination — are the similarities superficial and unimportant?

One way that the “hourglass” shape of development could be explained is if a conserved set of genes, with strong morphological effects, are activated at this period and cause similar shape changes in the embryos of related species.  To test directly for evidence of this kind of gene-expression-driven morphological patterning, Kalinka et al. carried out a fairly simple set of experiments and calculations. Expression profiles for over 3,000 genes were compared at different times during embryogenesis in six species of Drosophila, which are separated in evolutionary time by up to 40 million years. This allowed the authors to calculate the divergence (basically, the lack of correlation) in gene expression among the different flies and look for the time points during development when the species-to-species gene expression patterns were most similar. As a first result, the authors showed that time expression profiles for all pairs in this 6-membered group could recapitulate the known phylogenetic relationships to a striking degree. That is,  if two flies are closely related (as measured by sequence homology) their gene expression patterns were also very similar.  .If you build a phylogenetic tree solely on the basis of similarities between gene expression profiles, the tree has the same topology to the “real” (sequence-based) tree, although the lengths of the branches are different.  This result hinted at an important link between embryogenesis and phylogenetic relationship; perhaps I shouldn’t be so hard on my old English professor.

The key finding in the paper, however, came after plotting this divergence across all genes as a function of time. Gene expression divergence over time did in fact follow an hourglass pattern.  They actually plot divergence against developmental stage, and find a neat curve that looks like half of an hourglass.  The pinch in the curve — the point of maximal conservation — occurs during a period that corresponds to the extended germband stage, which, for flies, is generally regarded as the phylotypic period. In a related calculation, the variance at each time point was used to calculate the force of stabilizing selection on each gene. Think of stabilizing selection as evolutionary pruning, in which natural selection tweaks genomes to favor some sort of optimum in an emergent phenotype. As expected, the strongest selective constraint was found during the phylotypic period, suggesting that natural selection is especially picky during this critical phase.  These two results together go a long way towards confirming that the phylotypic period is real — not just an illusion — and that there is something very special about this phase of embryonic development.

Why is embryonic development so tightly constrained during the phylotypic period? Rudolf Raff has proposed that “an increase in the number of global interactions between genes and developmental processes during the phylotypic period renders any evolutionary modification highly deleterious due to their damaging side-effects.” This is an attractive idea that would explain the global inflexibility in expression that was observed during this key phase of development. As is sometimes the case, however, studies such as this one do not answer open questions in biology so much as they help sharpen and define ones that already exist. How would we test Raff’s proposal, and move toward understanding the conservation of morphology and gene expression patterns in the phylotypic stage of development? A good scientist once told me: the first step is to observe; the next step is to perturb. One approach would be to perform an RNAi screen to measure the sensitivity of embryonic development to various gene knockdowns. If the global interactions model is correct, we would expect a greater overall sensitivity to RNAi during the phenotypic period. This observe-then-perturb approach to experimentation actually works a bit like natural selection: small changes at the molecular level give rise to measurable changes in an emergent property that is selected for a functional goal. For evolution, the presumed goal is to constrain mid-stage development; for biologists, the goal is to gain a better understanding of the phylotypic period. I suppose you could say, in some sense, that the effort to understand biology through experimental perturbation is the scientist’s ultimate opportunity to recapitulate phylogeny.

Kalinka AT, Varga KM, Gerrard DT, Preibisch S, Corcoran DL, Jarrells J, Ohler U, Bergman CM, & Tomancak P (2010). Gene expression divergence recapitulates the developmental hourglass model. Nature, 468 (7325), 811-4 PMID: 21150996

See also:

Duboule, D. Temporal colinearity and the phylotypic progression: a basis for the stability of a vertebrate Bauplan and the evolution of morphologies through heterochrony. Dev. Suppl.135–142 (1994).

Raff, R. A. The Shape of Life: Genes, Development and the Evolution of Animal Form. (Univ Chicago Press, 1996).