How to Build a Dinosaur (17 page)

It was enough to make a theoretical physicist throw up his hands. And it was the fly that provided the answers. Bithorax was the first of the single-gene mutations in the fly that turned out to be so important in development. The mutation made the hind wings look like the front wings. Others came quickly. “A rather spectacular mutant, antennapedia, causes the development of legs in place of the antennae on the head,” Carroll wrote. A number of these genes were discovered. They were called homeotic genes, and in each case a mutation turned one body part into another part, antennae into legs, as with antennapedia
,
or hind wings into forewings, as with bithorax. Another characteristic of these genes is that all applied to modular body parts, building blocks that when varied in size, shape, or number produced a creature whose form was different. A fly or other insect could have different numbers of wings, or legs, all made from basic, repeatable body parts.

These were obviously very powerful genes governing the overall pattern of the fly body, and, as it turned out, not just the fly body. They were the master genes of evo-devo. There were eight genes in two clusters, the Antennapedia Complex and the Bithorax Complex, five for the front half of the fly and three for the back half. Carroll writes, “Even more provocative, the relative order of the genes in these two clusters corresponded to the relative order of the body parts they affected.” In other words the physical arrangement of the genes along the chromosome put them in the same sequence—head to tail—as the parts of the fly they governed. The genes shared a stretch of DNA called the homeobox and they were called homeobox, or HOX, genes. In each of these genes the homeobox contained the code for a stretch of protein that was designed to latch on to other genes in order to turn them on and off. Similar homeobox stretches were found in genes and proteins in frogs, birds, and mammals, meaning that throughout the animal kingdom these HOX genes and proteins were turning genetic switches on and off during the development of the embryo. The HOX genes were master coordinators of development. And they were so similar that it meant they had remained the same over the course of five hundred million years of evolution. That is how far back one would have to go to find the common ancestor of fruit flies and mammals. These master genes were clearly essential to life.

As research continued, in the ’80s and ’90s other master controlling genes were discovered, genes for building essential organs like limbs and hearts. Genes to control patterns of growth. The way all these genes in what is called the “genetic tool kit” function is by producing proteins that switch other genes on and off (transcription factors). They also produce proteins to travel to other cells and set off sequences of gene activation that alter how the cells behave, how they move, and at what stage of development, and what the rate of growth should be.

The whole structure of control is not fully understood, and some stretches of DNA have been found to produce small snippets of RNA that are never translated into proteins. These micro RNAs also turn genes on and off, genes that may then produce either RNA or proteins to control other genes, and so on down the line. The potential combinations boggle the mind, but if one were able to map out every instruction, every gene activation and chemical event, in order and location, one would then have the instructions for building a worm, or a fly.

Sonic hedgehog, for example, is the name for both a gene and its protein. The protein is a transcription factor that affects growth. You can take a developing embryo and add or inhibit sonic hedgehog without actually changing the genes, and you will turn on or off the growth of a forelimb, or a tail. The odd name is a result of laboratory humor among fruit fly geneticists. One version of the gene causes fruit fly embryos to be covered with spikes so they look like a hedgehog. Sonic hedgehog comes from the cartoon character. Groucho and smurf are other such genes, also death executioner Blc-2.

One of the important families of growth factors observed in developing embryos is the bone morphogenetic protein family, BMP. Different kinds of BMP, identified by number, control genes that cause growth of bone cells. But what is the control for turning on the control gene? As Carroll describes it, “there are separate switches for BMP5 expression in ribs, limbs, fingertips, the outer ear, the inner ear, vertebrae, thyroid cartilage, nasal sinuses, the sternum, and more.”

Each switch has different sequences of DNA within it, to which different proteins bind. “An average-size switch is usually several hundred base pairs of DNA long. Within this span there may be anywhere from a half dozen to twenty or more signature sequences for several different proteins.” Carroll estimates the different combinations possible with five hundred DNA binding proteins that can work together in pairs or large numbers to activate sequences in switches. There are “12,500,000 different three-way combinations and over 6 billion different four-way combinations.”

Perhaps even more intriguing is that, “There is no ‘masterbuilder’ in the embryo,” as Lewis Wolpert wrote in
The Triumph of the Embryo.
The cells talk to each other. “There is no central government but rather, a number of small self-governing regions.” And one event determines the next, writes Wolpert, “There are thus no genes for ‘arm’ or ‘leg’ as such, but specific genes which become active during their formation. The complexity of development is due to the cascade of effects.”

Imagine the development of an embryo as a self-conducting symphony in which the sound of the bassoons triggers the tympani. The bassoons are triggered by the violins, but depending on what the violins play, and when and how loud, the bassoons may play differently, which will affect the tympani. And if the tympani play long enough, that stops the violins.

Embryologists have watched every stage of growth of organisms like the fly and the chicken, and have mapped where cells go to become a brain or a liver, and what chemicals are present in the cells when they proliferate or change. They have looked at limb growth in great detail and noted when the buds that become digits first appear and how many grow and which ones do not grow. They have watched the death of cells functioning to sculpt shapes that then continue to grow.

And it has become clear that this is how the forms of animals change during evolution. A mutation in a master or signaling gene, or a change in a switch, or switches, extends the fingers in a bat’s wing and makes the webbing grow. For each new shape and form, there is no new suite of genes that provide a whole new set of detailed instructions for a wing instead of a limb. Changes in regulation reverberate through the system of switches and feedback loops to create new forms.

Charles Darwin’s notion of natural selection remains as the most powerful, most fully understood force of nature. It “selects.” Some changes in development will be useful, while others will be fatal. But on the evolutionary voyage from dinosaur to falcon, what happens is not that a whole new set of falcon genes is developed for beak, wings, and eyes. Instead the instructions for limbs, feathers, eyes, and tail are changed so that the same building blocks of the vertebrate body are put together in different ways.

The hope for applying knowledge of development to evolution and, for our purposes, to find a way back through the extinction barrier, is to link microevolution to macroevolution. If we can tie development, recorded down to the specific gene and its protein product, to the gross anatomy of fossils, we will have a whole new level of understanding about the evolution of form. This kind of work is in its early stages, but there are some good examples.

Feathers are one feature, highly pertinent to both avian and nonavian dinosaurs, for which this has been done in elegant and satisfying detail by Richard O. Prum of Yale and several colleagues.

Their work is all the more interesting because, in the absence of evidence from genetics and developmental biology, a theory of feather evolution had been developed that seemed to make sense but turned out to be impossible. “According to this scenario,” Prum and Alan H. Brush wrote in
Scientific American
in March 2003, “scales became feathers by first elongating, then growing fringed edges, and finally producing hooked and grooved barbules.”

To understand why this couldn’t have happened, it’s necessary first to understand the structure of that lovely feather floating in the wind, or contributing to the fluffiness of your pillow. Feathers are essentially long tubes with branches. The branches also have branches, and those branches again have something like branches, except that the last twiglike extensions are hooks or barbules that hold the feather together.

There are two different sorts of feathers. One is the blue jay or pigeon feather you may find on the ground, the turkey feather you can buy if you tie flies to catch trout. The other is found in great numbers as the down in your sleeping bag, comforter, or winter coat. The first is pennaceous and the second is—and this has to be one of the great words of biology—plumulaceous. The pennaceous feathers have the branching described above, while the plumulaceous feathers have very little main stem and instead a tangle of lesser branches, with barbules that link together, forming the air-trapping matrix that keeps birds warm, and people as well, in their sleeping bags and puffy mountaineering coats.

The first step in understanding what feathers are and how they evolved was achieved simply by tracking embryonic growth at a microscopic level. Feathers grow out of the skin or epidermis, the outer layer of cells in the developing embryo. Part of the skin starts to thicken, and then to grow out into a tube, while around the growing tube a cylinder of cells form the follicle. The follicle keeps generating a kind of cell that produces keratin, the substance in fingernails and hair. The new cells at the bottom push the old cells at the top, “eventually creating the entire feather in an elaborate choreography that is one of the wonders of nature,” Prum writes.

One aspect of the growth is indeed wonderful, and complex. As the hollow tube grows, something happens with the part of the follicle called the collar, which is the source of the growth of keratin-producing cells that push the central, hollow shaft of the feather out from the skin. It begins producing ridges on the central shaft that grow in a helix on the tube, turning into the main branches as the feather grows. Then the barbules grow from these branches. All of this happens at once, which gives a hint of the mystery and wonder in the way organisms grow from one cell to a complex creature, with so many cells forming so many and such complex patterns, all timed to occur at the right moment and directed to the right place. The feather is just one small example of this sort of change in concert.

Prum and other colleagues proposed that in the course of this development they could see the way feathers had evolved. Primitive structures, like those they identified in the early stages of feather development, must have appeared first in evolution. Animals must have existed that had only these tubelike structures. Only later did the feathers that let birds fly emerge.

In other words, feather evolution, like feather growth in the embryo, proceeded by discrete steps. And one step had to be completed before the next one could occur. Each step depended on what had gone before. The final product, the feathers that enable the flight of falcons and swallows, came long after the feather first evolved. And since those first feathers had absolutely no connection to flying, feathers had to have evolved for some other purpose. The feather has been one of the features creationists have long pointed to as an impossibility for evolution. How could feathers, a truly novel development, not just a longer arm or a thicker skull, evolve on their own and just happen to be useful for flying? Prum and colleagues showed exactly how that could and did happen. Features emerged that served one purpose, and as other features were built on them they changed into the structures we see today.

First to evolve were simple tubes, hollow cylinders, then barbs that formed tufts on the tubes. In the next stage feathers became tubes with branches that had tufts, or barbules. There was one more step, which was for the barbules to change shape to have hooks at the end. These hooks are what allow a feather to close and feel as if it is one piece, repelling water, or pushing on air. After the stage of the hooking barbules, the change that produced true flight feathers could have taken place. This was an asymmetric feather, with more on one side of the central tube.

Prum and John F. Fallon and Matthew Harris at the University of Wisconsin-Madison went deeper into development, using techniques to observe which genes were active at which stages and in which locations in the growing feather. They found two well-known genes and the proteins they coded for. Sonic hedgehog and one of the bone morphogenetic proteins, BMP2, were present in different places and different concentrations promoting growth (sonic hedgehog) and the differentiation of new kinds of cells (BMP2). BMP2 was also limiting cell proliferation.

First they would appear where the feather germ was starting, later at the beginning of the ridges that turned into the first branches. The two proteins directed the growth of the feather, and did it in stages, just as Prum and colleagues were suggesting, with each stage possible only because of the one before it. Without the feather germ there could be no ridges or branches or tufts. The general picture they saw in development and proposed in evolution was that first came the central shaft, the hollow tube. Then came downy tufts. Finally came the helical ridges, organized branches, and barbules that made modern feathers, the sort that can be found on a starling or on
Archaeopteryx
.

On a cold night when you crawl into the warm cave under a down comforter, you are taking advantage of millions of years of evolution, mediated by two genes and the proteins they code for—sonic hedgehog and bone morphogenetic protein .