Pharyngula

Yesterday, I pointed out that Jonathan Wells was grossly ignorant of basic ideas in evo-devo. This isn’t too surprising; he’s a creationist, he has an agenda to destroy evolutionary biology, and he’s going to rail against evolution…same ol’, same ol’. That’s nothing, though. Wells and his fellows at the Discovery Institute have an even more radical goal of fighting natural, material explanations of many other phenomena, and his latest screed at the DI house organ is against natural explanations of development. Not evolution, not evo-devo, just plain basic developmental biology—apparently, he wants to imply that the development of the embryo requires the intervention of a Designer, or as he refers to that busy being in this essay, a postmaster.

He starts off by setting up a problem in a reasonable way—I actually say something very much like this on the first day of my development class.

The many different kinds of cells in an animal or plant develop from a single fertilized egg cell. Humans, for example, consist of cells that form bone, skin, muscle, digestive organs, nerves and many other tissues. Such cells are so different from each other in form and function that an untrained observer might conclude that they represent different species.

Yet all of these cells contain the same DNA, a fact long known to embryologists as “genomic equivalence.” As the fertilized egg divides, it bequeaths a complete set of DNA (its “genome”) to all of its descendants – with a few minor exceptions, such as red blood cells, which have no DNA at all. But if bone, skin, muscle, digestive and nerve cells all have the same DNA, why are they so different? Why don’t nerve cells secrete juices that digest the brain? Part of the answer is that although brain cells have the genes for digestive juices, those genes are turned off in nerves. As an embryo develops, its cells go through a phase called “differentiation” that turns some genes on and leaves others turned off.

But this does not solve the problem, since it begs the question of why two cells with the same DNA would differentiate in two distinct ways.

This is developmental biology 101. We describe this central issue, and then, basically, the rest of the term is spent explaining exactly how this works: we tell the students about maternal factors, environmental effects, cell signaling, single transduction, gene repression and activation, all this wonderfully fun stuff that explains how cells with the same DNA would differentiate in many distinct ways.

Wells does not do that. He raises this introductory question that just about all of us developmental biologists use in our courses, and then falls flat and acts as if it is a complete mystery how this problem is resolved. It’s as if he signed up for a development course in his academic career, caught the first few sentences by the instructor, and then fell asleep for the rest of the term. The closest example I know of to this performance is the infamous Darwin quote-mine, in which Darwin rhetorically poses an interesting question…

To suppose that the eye, with all its inimitable contrivances for adjusting the focus to different distances, for admitting different amounts of light, and for the correction of spherical and chromatic aberration, could have been formed by natural selection, seems, I freely confess, absurd in the highest possible degree.

…and then proceeds to answer it in detail. The creationists are fond of quoting the problem and leaving out the answer, and Wells has done exactly the same thing, mentioning a basic problem in the field and then omitting the fact that molecular genetics not only has the answer in principle, but has also worked out the details in many systems.

Here’s a brief primer in the answer.

Let’s imagine a very simple developmental system in which we have only two genes: a black gene and a red gene. When the black gene is expressed, it makes a gene product, the black protein, that binds and shuts off the red gene. Similarly, if the red gene is expressed, it makes a protein that binds and shuts off the black gene. Furthermore, let’s say that in the absence of these proteins, both genes will be expressed, but the red gene is more strongly/rapidly turned on. Now lets assume we’ve got a single celled egg with this strand of DNA, and it’s also pre-loaded with the black protein by its mother, as in this cartoon.

i-6b2d7522d09b8d5b979f02d48c1a2aab-dif_prob.gif

The red gene is turned off by all that black protein, and the black gene will just be making more black protein: when the cell divides into two, the pattern will continue, and all the progeny cells will continue to have the same pattern of gene expression. Oh, no! How can we ever do anything interesting?

Here’s one way:

i-b52a8cd44e31aa024e7d34843aac238d-dif_distrib.gif

We start with the same system, black gene active and red gene repressed, but we have a second step there, called “differential distribution of determinants”. Somehow, the black protein is localized to one part of the dividing egg, so that when it divides in the third step, no black repressor is present in one of the daughter cells. In the fourth step, the red gene is active, makes the red protein, and shuts off the black gene. Look! Two cells with exactly the same genes, but with different patterns of gene expression!

At this point, you should be thinking, “Hey, he cheated—he said ‘somehow’ the protein is localized! Isn’t the ‘somehow’ what we want explained?”

You’d be right. The ‘somehow’ of localization is the real question, but I want you to be in the same state of mind as Jonathan Wells. Wells looks at this problem, and thinks correctly that step 2 sure looks like an interesting problem, incorrectly claims (or more cynically, intentionally misrepresents the state of our knowledge) that the differential distribution is unexplained, and worse, starts babbling. He makes an interesting analogy that each cell has its own zip code defined as a pattern of gene expression, and that we need to assign unique zip codes to each cell…but then, typical for an IDist, decides that this implies a Postmaster and Slots and a Wall (metaphor run amuck! Watch out! I’m surprised he didn’t postulate a Tape Dispenser and a Stamp Vending Machine.)

Yet the existence of “cellular zip codes” still doesn’t solve the problem either (and the authors of the article don’t claim that it does). If the human body were the United States and cells were postal envelopes, each would start out bearing every zip code in the country on its face. Only after the postmaster had stuck each envelope into one of many slots on the wall to direct it to its final destination would a particular zip code be highlighted. Obviously, the postmaster and the array of slots play a major role in determining where each letter goes.

If the DNA corresponds to zip codes that are originally the same on every envelope, where in the embryo are the postmaster and the slots? What is it that highlights one zip code but not others? Where is the all-important developmental information that directs cells to different parts of the body and tells them where they are and how to differentiate?

A fellow just has to laugh at the feeble imagination of the IDists: a postmaster? You’ve got to be kidding me. Why postulate an anthropomorphic entity in your analogy when we’ve got a wealth of examples of natural processes that sort out cell identities? Here are some examples of the ‘somehows’ to explain step 2 of my cartoon—these are all known, functional mechanisms for setting up differential gene activity in developing cells, and none of them require handwaving about some unknown entity monitoring every embryo and intelligently tagging cells.

  • Maternal factors. The egg is not a uniform smear of factors. When produced in the ovary, various proteins that specify patterns of gene activity are non-uniformly distributed, so that the dividing cells receive different portions of these maternal components. The best known example now is probably bicoid in the fly, but this principle has been known for a century or two; frogs also have visibly striated eggs, and Conklin in 1905 described a detailed pattern in ascidians.

  • Sperm entry. One obvious symmetry breaking event in diploid organisms is fertilization: a single sperm enters the egg, and its entry can trigger a cascade of membrane and cytoplasmic events localized to one side of the zygote. This is a very common signal, used in amphibians, for instance, that can be used to generate differential gene activity.

  • Environmental influences. Many animals use gravity as a cue, and the mechanism is easy enough to understand: denser proteins will sink through the cytoplasm to localize to the bottom of the egg. Another cue is light; the alga Fucus uses light to induce calcium fluxes on one side of the embryo that lead to differentiation of the rhizoid. Mammalian embryos are oriented relative to the site of implantation in the uterus.

  • Chance. Especially when dealing with small numbers of molecules, chance fluctuations in the distribution of the determinants can lead to small quantitative differences in gene expression, which can be coupled to feedback mechanisms that amplify differences (for example, as I described for Notch) which then produce stronger qualitative differences in cell identity.

These describe ways of breaking symmetry early in development to establish initial polarity, but later events are also explainable: they become more complicated in detail, but the problems are simpler in principle. Once you’ve got two cell types, additional cell types can be generated by, for instance, cell signaling. A cell that is in the middle of a patch of the red cells in my cartoon, for example, is in a different environment than a cell at the boundary between the red and the black; a cell that senses both red cells and black cells as neighbors might be triggered to activate a third gene. Another mechanism used in collections of cells is the gradient—a cell can sense its location along an axis by measuring the concentration of a particular signaling molecule, reading a kind of molecular vernier that specifies identity. As development unfolds, the cellular environment becomes a richer and more complicated source of diverse signals that are used to further increase the diversity of potential differentiated cell types. (Again, you might look back at my article on Notch for examples).

Wells tries to turn his misconceptions into an indictment of “neo-Darwinism” in his conclusion, but he’s missing the point.

By focusing attention on DNA as the supposed source of raw materials for evolution, neo-Darwinism has systematically downplayed the nature and location of developmental information elsewhere in the embryo. Obviously, there is more to embryo development than is dreamt of in neo-Darwinian philosophy.

Errrm, this has long been a complaint about the neo-Darwinian synthesis by developmental biologists—it neglected development almost completely. There is a groundswell of activity to correct that omission, driven by the evidence collected by the evo-devo approach, and (perhaps less persuasively) by the philosophy of Developmental Systems Theory. Scientists have turned to the study of natural mechanisms to answer these questions; they are not going to waste time calling up the hacks at the Discovery Institute for their bad ideas. And, most importantly…

No postmasters are involved.

Comments

  1. #1 Torbjörn Larsson
    January 26, 2007

    Tangentially, I’ve heard the converse of the 2LoT argument a couple of times; development proves that 2LoT doesn’t apply to living beings.

    Oh, it applies alright, it’s that development shows that it isn’t a problem for life.

    But more to the point, it is an absurd thought from the beginning to apply the basic law of 2LoT here. Never mind that Earth is an open system with plenty of energy from the sun passing through, so 2LoT mean little for any subsystem – how do you define the system and the entropy for a living being and its environment in the first place? That is like trying to figure out how to do my private economy solely on some general principle of national economy.