TURNED ON
A revolution in the field of evolution?
by H. ALLEN ORR
Issue of 2005-10-24
Posted 2005-10-17

If you read much popular science, you?d be forgiven for 
thinking that biology has become something of a banana 
republic. A seemingly endless series of books and 
newspaper articles reports that biology is being roiled by 
any number of revolutions. Take your pick: genomics, 
proteomics, medical genetics, and, now, something called 
evo devo. Some of the revolutionary rhetoric is surely 
hype, but these are, undeniably, exciting times in 
biology. Entire genomes are being decoded at an astounding 
rate (nearly three hundred species have been done, and 
more than seven hundred others are in the works), and new 
high-tech approaches to old problems seem to appear by the 
week. The result of all this has been some genuinely 
surprising scientific findings. And some of the biggest 
have come from the new science of evo devo.

Evo devo, punk-band-inspired slang for ?evolutionary 
developmental biology,? holds the promise of a radical new 
way to look at life?s evolution. Its central thesis is 
simple. Organisms show two kinds of change through time: 
during the lifetime of a single animal (you don?t look 
much like the egg you started as) and during the 
evolutionary history of a biological lineage (you don?t 
look much like your three-and-a-half-billion-year-old 
ancestor). Evo devo?s key claim is that the first kind of 
change can provide important insights into the second.

The notion that there?s a connection between evolution and 
development?the growth of an organism from a single cell, 
through an embryo, to an adult?is both natural and old. It 
was especially popular in the nineteenth century; Ernst 
Haeckel?s law of recapitulation?the proposition that 
?ontogeny recapitulates phylogeny??captured the spirit of 
the age. In Haeckel?s view, the growth of an animal plays 
out, in speeded-up form, the whole history of life: an egg 
represents our original single-celled ancestor, an embryo 
of four cells represents a primitive creature, and so on. 
The idea that embryology, as developmental biology was 
then known, reveals important truths about evolution also 
played a big part in the first real revolution in 
evolutionary thought: Darwinism. Darwin himself believed 
that embryology provided powerful evidence of life?s 
descent from a common ancestor, evidence that he laid out 
in loving detail in ?The Origin of Species.? Why, Darwin 
asked, are the embryos of various animals so much more 
similar than the adults? Why do rudimentary legs appear 
during the development of some snakes? And why do whale 
fetuses have teeth, when, as adults, whales ?have not a 
tooth in their heads?? Darwin argued that odd facts like 
these are far more readily explained by his theory of 
evolution than by special creation.

In the early twentieth century, though, the new science of 
genetics nearly killed the old science of embryology. In 
science, as in everything else, there are opportunity 
costs. And the cost of working out genetics?of explaining 
how genes are passed from one generation to the next?was 
that many biologists abandoned embryology. The result was 
a biology that was heavy on genes but light on embryos.

By the second revolution in evolutionary thought?the 
so-called Modern Synthesis of the nineteen-thirties and 
forties, when Darwin?s theory of evolution was fused with 
Mendel?s theory of genetics?development was left out in 
the cold. The resulting evolutionary theory took a 
mathematical turn, emphasizing how genes grow more or less 
common under the influence of quantifiable forces like 
natural selection. Evolutionists might talk about genes, 
or about fitness, or about equations that linked the two, 
but they almost never talked about embryos. And so it was 
through most of the twentieth century.

But no longer. Although animal development has always 
seemed one of the most mysterious of biological 
phenomena?the emergence of an elaborate adult from a 
formless egg seems almost magical?geneticists and 
molecular biologists in the nineteen-eighties and nineties 
worked hard to demystify the process, and made dizzyingly 
rapid progress. By the start of the new millennium, 
scientists could explain the complex convolutions of 
embryos in stunning molecular detail. Today, there?s much 
excited talk among evolutionary biologists about bringing 
this new understanding of development to bear on 
evolution. Indeed, several scientists, including, most 
prominently, the late paleontologist Stephen Jay Gould, 
have argued that the incipient science of evo devo might 
represent the next great revolution in the history of 
evolutionary thought.

Sean Carroll, a professor of biology at the University of 
WisconsinMadison, and a leading evo-devo researcher, 
surveys the state of the field for a general audience in 
his book ?Endless Forms Most Beautiful? (Norton; $25.95). 
Though Carroll?s book, like much popular science, can be a 
bit breathless (we?re treated to ?brave biologists? whose 
work ?gave birth to an exciting new vista?), he does an 
admirable job of explaining science that was, until now, 
entombed in technical journals. So just what are these big 
findings? Not surprisingly, they all hinge on how DNA 
builds organisms.

DNA is the stuff that gets passed from parent to child, 
and one of its main jobs is to tell cells how to make 
proteins, the molecules that we?re basically made of. A 
stretch of DNA that makes a protein is called a gene. 
Biologists usually say that a gene ?codes for? a protein, 
by which they mean that the gene?s DNA sequence instructs 
a cell to build a specific kind of protein; a slightly 
different DNA sequence might yield a slightly different 
protein. One of our genes, for example, codes for insulin, 
a protein that helps cells absorb glucose. Not all our DNA 
is given over to genes, though; there are also long 
stretches of so-called noncoding DNA, which sit between 
genes. So as we move along a string of DNA, we might first 
come to a gene, and then to a stretch of noncoding DNA, 
and then to another gene, and so on. Despite the relative 
fame of genes and the relative obscurity of noncoding DNA, 
more than ninety-five per cent of our DNA is noncoding.

Given that nearly all our cells carry the same DNA, you 
might wonder why your eyes look different from your 
pancreas. The main reason is that genes don?t necessarily 
make their proteins in every cell. Instead, a particular 
gene in a particular cell might or might not be 
?expressed? (that is, switched on, and so making its 
protein). Part of the reason that your pancreas differs 
from your eye is that the insulin gene, among others, is 
expressed only in your pancreas and not in your eye. How 
particular genes get switched on or off plays an important 
part in the evo-devo story.

Evo devo?s first big finding is that all animals are built 
from essentially the same genes. In the past few decades, 
biologists collected thousands of genetic mutations that 
disrupt the normal course of an embryo?s development. 
(Most of this work involved the humble fruit fly.) By 
characterizing how particular mutations derail the growth 
of embryos, biologists were able to figure out which genes 
control key steps in animal development. In one of the 
biggest breakthroughs, biologists worked out how fruit-fly 
embryos decide which of their ends should be the head, 
which the tail, and what should go between. Part of the 
answer involves what are called Hox genes. Different Hox 
genes get expressed in different parts of a fly?s body, 
and each Hox gene tells that body part what appendage it 
should grow. A Hox gene expressed in the head, for 
example, might tell the head to grow antennae, while a Hox 
gene expressed in the body might tell the body to grow 
legs. If you tinker experimentally with Hox genes, you get 
the stuff of B movies: mutant flies, for example, that 
have legs, not antennae, growing out of their heads.

But the truly surprising thing about Hox genes turns out 
to be evolutionary. All animals have Hox genes, and nearly 
all animals use their Hox genes to determine which 
appendage should go where along the axis that runs from 
head to tail. Given that the major animal groups, among 
them arthropods (now including insects), mollusks 
(snails), annelids (worms), and chordates (human beings), 
were in place at the start of the Cambrian period, Hox 
genes must be at least half a billion years old.

What?s more, plenty of important genes turn out to be this 
old. We now know that several hundred genes, including the 
Hox genes, are needed to lay out an animal?s basic design. 
Carroll calls these the ?tool kit? genes, and they?re the 
central characters in his story. Nearly all tool-kit genes 
are present in all animals, and they do much the same 
thing in all animals. The same gene, for example, that 
triggers eye development in fruit flies also triggers eye 
development in mice. Indeed, genetically engineered flies 
will happily build eyes if supplied only with the mouse 
gene. (They build fly eyes, not mouse eyes.) Similarly, a 
gene that affects pigmentation in birds like the chicken 
and the bananaquit also affects pigmentation in mammals 
like the jaguar and you. Indeed, changes in bird-plumage 
color often involve the same gene that causes red hair in 
humans. This surprising genetic conservatism across nearly 
all animals is evo devo?s key empirical finding: swans, 
swallowtails, and socialites are all built from the same 
genes.

Why, then, do different creatures look so different? How 
do penguins and people emerge from the same genes? Evo 
devo?s answer to this question represents its second big 
finding. Different animal designs reflect the use of the 
same old genes, but expressed at different times and in 
different places in the organism. As embryos, penguins 
might express one combination of genes in their limbs (and 
the result is wings), while people might express another 
(and the result is arms).

The basis of this selective expression involves that part 
of the DNA which is noncoding. Most genes, like most light 
fixtures, have ?switches? near them. These switches, which 
are made of DNA, affect only whether a gene is on in a 
particular cell at a particular time; they do not change 
the actual protein coded by a gene. One switch might 
specify whether a gene should be on in the pancreas and 
another whether it should be on in adults. What?s special 
about many of those tool-kit genes is that they make 
proteins that toggle these switches. If a tool-kit protein 
finds and binds to a switch, it insures, through a complex 
molecular choreography, that a certain gene is expressed 
(or, in some cases, not expressed). In effect, tool-kit 
proteins act like molecular fingers, reaching out and 
physically turning on or off the switches that sit next to 
genes. And, just as a human finger might turn many 
different light switches on or off, so a single tool-kit 
protein might switch many genes on or off. In the end, the 
logic of animal development involves a long cascade: 
tool-kit genes effectively switch other genes on or off, 
some of which then switch yet other genes on or off, and 
so it continues throughout the assembly of an adult. If 
all goes well, each of the possibly trillions of cells in 
an animal?s body will express just the right genes: 
insulin in your pancreas, not in your eye.

The real excitement about evo devo, however, has to do 
with its third claim. Carroll and others have taken the 
next, and by far the most radical, step and argue that 
evolution is mostly a matter of throwing these switches. 
There?s certainly some evidence for this idea. In one 
well-known example, a group of researchers identified a 
small region of DNA that causes a striking difference 
between two kinds of stickleback fish. The kind that lives 
in the Pacific Ocean has spiky spines on its pelvis; lake 
stickleback, which evolved at the end of the last Ice Age, 
lack these spines, probably because they?re no longer 
needed to defend against oceanic predators. This 
difference in anatomy is genetic?ocean fish grow spines no 
matter what kind of water they?re reared in?and it appears 
to be due to a single gene, Pitx1. Remarkably, though, 
Pitx1 produces precisely the same protein in both ocean 
and lake fish. So why do they look different? Because, out 
of all the tissues that normally express Pitx1?head, 
trunk, pelvis, and tail?the pelvis of lake fish stopped 
expressing it. Ten thousand years ago or so, a mutation 
occurred at a DNA switch near Pitx1 and pelvic spines 
disappeared from lake fish.

Evo devo?s emphasis on switch-throwing represents a 
profound departure from evolutionary biology?s long 
obsession with genes. Animal evolution works not so much 
by changing genes, Carroll maintains, but by changing when 
and where a conserved set of genes is expressed. In the 
lingo, evolution is regulatory (involving patterns of gene 
expression), not structural (involving the precise 
proteins coded by genes). You can think of this 
distinction in terms of those light switches. Imagine two 
houses that were built from the same blueprint and that 
were initially identical. But now, years later, we notice 
that they look different at night. In one, the first floor 
is bright and the second floor dim; in the other, the 
opposite is true. This difference could have arisen in two 
ways. Maybe the houses now feature different lights; the 
owners of the first house might, for instance, have 
replaced bulbs on the first floor with brighter ones; the 
other owners might have done the same thing on the second 
floor. But maybe?and this is the evo-devo picture?the 
owners of the first house have switched on most 
first-floor lights and switched off most second-floor 
lights; the owners of the second house might have done the 
reverse. Evo devo tells us that animal species look 
different not because their structural bits and pieces 
have changed but because they switch on and off the same 
old bits and pieces in different combinations. Roughly 
speaking, then, penguins and people differ for the same 
reason your pancreas and eye differ: they?re expressing 
different genes.

In the evo-devo view, animals still adapt to their 
environments by Darwin?s natural selection, but adaptation 
involves a different kind of genetic change from what 
biologists used to assume. As a result, evolutionists may 
need to abandon what Carroll calls their ?protein-centric 
perspective? and look instead to the poorly understood 
noncoding DNA that sits between genes. Evo devo?s 
advocates argue that switch-throwing also makes good 
evolutionary sense. If a gene itself were to change, its 
altered protein would show up in every tissue that 
expresses the gene, and the new protein probably wouldn?t 
work well in some tissue. If a switch were to change, 
though, the same old protein would show up in all the 
usual tissues except one. This kind of ?modular? change is 
far less likely to wreak havoc on an organism and so is 
much more likely to be used by evolution. Indeed, 
evolution likes modular change for the same reason 
engineers do: while changing all the metal in an airplane 
to some new alloy would be asking for trouble, switching 
the metal in the bathroom door handle to the new alloy 
would probably go fine.

The idea that animal diversity reflects switch-throwing 
might also help to explain how so many different kinds of 
animals emerge from so few genes. Vertebrates, including 
human beings, carry only around twenty-five thousand genes 
(before the genome was decoded, scientists typically 
guessed that there would be around a hundred thousand); 
most other animals carry even fewer. How does nature 
squeeze so many kinds of creatures out of such a modest 
genetic endowment? The answer, according to evo devo, is 
combinatoric explosion. Although thousands of genes might 
not sound like enough to explain the almost unimaginable 
diversity of what are, after all, the most sophisticated 
machines in the known universe, the number of combinations 
in which these genes can be expressed (gene 1 with gene 2, 
but not with genes 3, 4, 5, etc.) is nearly boundless. In 
other words, there are many species for the same reason 
that there are many sentences: you might know only a few 
thousand words, but you probably don?t lose sleep over the 
prospect of running out of sentences.

What Carroll calls the ?Evo Devo Revolution? hasn?t been 
greeted with universal enthusiasm. The main worry that 
skeptics raise is that the empirical evidence for evo 
devo?s claims remains somewhat limited. Studies like those 
in the stickleback are notoriously hard to perform, and 
there are very few cases where we can point to the genetic 
changes that underlie the evolution of animal form. 
Although experiments so far suggest that evolutionary 
changes at switches are fairly common, it?s not yet clear 
whether they?re the norm. Indeed, the biologists Marc 
Kirschner and John Gerhart argue in their new book, ?The 
Plausibility of Life? (Yale; $30), that the sort of 
regulatory changes that Carroll and others play up are 
less than universal. While they?re generally sympathetic 
to evo devo, Kirschner and Gerhart discuss several cases 
where evolution has involved good old-fashioned change at 
genes. And Carroll himself acknowledges important 
exceptions to his view.

One of the biggest involves our own lineage, the 
vertebrates. We now know that the rise of the vertebrates, 
some five hundred and twenty million years ago, was marked 
by two separate doublings of the entire genome. 
Vertebrates, in other words, suddenly found themselves 
with four times as many genes as their predecessors. 
Natural selection seized upon this new material and drove 
the evolution of novel functions at many of these genes. 
Evo devo?s claim that evolution is a matter of throwing 
switches at the same old genes fails conspicuously here. 
Vertebrates are who they are?a large group of 
spectacularly complex beasts?because of evolution at all 
those thousands of new genes. Though Carroll concedes 
this, he maintains that vertebrate evolution since those 
genomic doublings involved switches. Still, the rise of 
vertebrates is a big case to lose. In the end, it seems 
fair to say that, while most evolutionists believe that 
switches provide an important substrate for evolution?and 
I, for one, suspect the evo-devo view may prove right?some 
remain unconvinced that switches are the mainstay of 
evolution.

A second concern is more philosophical. Even if all the 
scientific claims made by evo devo are justified, is it 
really a revolution? Naturally, it depends on what you 
mean by ?revolution.? If you mean major progress, 
breakthroughs that add considerably to our understanding, 
who could argue? But some scientists seem to have in mind 
something more like a paradigm shift of the sort 
popularized by Thomas Kuhn in ?The Structure of Scientific 
Revolutions.? In such a shift, an old and familiar way of 
looking at things gets swept aside by a radically new one. 
The Darwinian revolution was certainly something like 
this: one could not look at biology in the same way after 
Darwin. And the Modern Synthesis, the second defining 
moment in the history of evolutionary biology, came close: 
it?s hard to think of evolution in the same way once you 
know that it has to obey the laws of Mendel?s genetics.

But evo devo doesn?t seem quite in this league. The ideas 
that Carroll champions wouldn?t undermine other, more 
traditional, ways of thinking of evolution. Even in the 
evo-devo account, change in animal form still involves 
change in DNA sequence and change in Darwinian 
fitness?just as evolutionary biologists assumed in all 
their equations over the past half-century. So it?s hard 
to see how evo devo could supplant these other approaches. 
The point isn?t that evo devo is not important; it surely 
is. For the first time, we have a good understanding of 
how particular changes in DNA cause particular changes in 
embryos, which, in turn, cause particular changes in 
species. The point is that not all significant science 
turns our world upside down. Despite the nearly 
irresistible romance of the scientific revolution, the 
history of evolutionary biology might end up looking a lot 
more like evo devo?s own history of the animal kingdom: a 
few radical innovations early on, followed by some 
intensely interesting tinkering.