Evolutionary biology can be long on theory but short on evidence. So
biologists were delighted when they thought they had a rare chance to
conduct an experimental test of a popular and appealing theory involving
a kind of molecular arms race between the sexes. Not only was the test
compelling but there was every indication that its results would turn
out to support the theory.
But when the experiment, reported in the current issue of Nature
Genetics, was completed, the scientists got a rude surprise. The results
were not what they expected. Now, as the dust starts to settle,
researchers are asking whether the theory is wrong or whether the
experiment is flawed.
The evolutionary theory drew on a notion that there could be a
battle between the sexes even in the womb. One popular theory, advanced
by Dr. David Haig, an evolutionary biologist at Harvard University, said
that if a species is polygamous -- several males can be the fathers of
one litter -- each male will want to be sure his progeny survive, even
at the expense of the other offspring and the mother. And so, the theory
says, it will be in a male's interests to carry genes that make his
fetuses grow very big, very quickly. The mother, on the other hand, will
want to keep as many of her embryos alive as possible, so she will want
to pass on genes that counteract the male's genes.
But the conflict will not arise, theoretically, if the species is
monogamous, since it is then in the interest of both the mother and the
father to give all fetuses an equal chance.
And so, Dr. Haig's theory says, there can be a sort of genetic tug of
war in polygamous species. Males will develop genes that make their
offspring grow quickly. And females will develop genes to counteract
those male genes.
What made the theory particularly appealing was that it could explain
a perplexing phenomenon known as genomic imprinting. Certain genes only
function if they are inherited from the father, and the mother's copies
of those genes are forever silenced. Other genes only function if they
are inherited from the mother, and the father's copies are turned off.
About 100 to 200 genes, 1 to 2 percent of the total in the mammalian
genome, are thought to be imprinted, said Dr. Shirley M. Tilghman, a
Howard Hughes professor of molecular biology at Princeton University.
The idea of imprinting seemed paradoxical.
"We usually assume that organisms should do the sensible thing," said
Dr. Laurence D. Hurst, a professor of evolutionary genetics at the
University of Bath in England. "If you have two copies of a gene, you
should be able to express both so that if one misfires, you have a
backup."
The challenge for biologists, Dr. Hurst said, was to explain why,
with imprinting, "the organism is doing what at first glance is a stupid
thing."
Over the years, more than a dozen hypotheses to explain imprinting
have been proposed. But Dr. Haig's gained widest acceptance. He argues
that imprinting is the way that males of polygamous species attempt to
make their fetuses grow big. The males will imprint a gene, chemically
modifying it over the course of evolution, so that only the male's copy
of the gene is expressed. And that will be a gene the encourages fetal
growth. Females will respond by imprinting genes that will slow fetal
growth. Dr. Haig's theory also predicts that monogamous species would
not imprint their genes.
The genetic arms race model proposed by Dr. Haig "best explains all
the experimental data on imprinting," Dr. Tilghman said.
Dr. Tilghman, who studies imprinting, decided to test Dr. Haig's
model directly. What she needed was a monogamous mammal and a closely
related polygamous one. Then she could ask whether genes that were
imprinted in the polygamous species were not imprinted in the species
that was monogamous.
"The hardest job was finding a monogamous mammal," she said.
Her search led to a 1965 paper by Dr. Wallace Dawson of the
University of South Carolina, who had discovered and maintained two
species of mice, one of which, Peromyscus polionotus, was monogamous,
while the other, Peromyscus maniculatus, was polygamous.
Dr. Dawson had mated the monogamous and the polygamous species of
Peromyscus and the results fit exactly with the predictions of Dr.
Haig's hypothesis. When the mother was monogamous and the father
polygamous, the theory predicts that the father's imprinted genes would
run into no opposition from the mother and the embryos should be large.
In fact, they are huge -- so big that they cannot be delivered and the
mother and her fetuses die unless scientists deliver the baby mice by
Caesarean section. But when the mother is polygamous and the father is
monogamous, the theory predicts small embryos. And in that case the
embryos were runts whose growth was dramatically stunted. Even the
placentas were affected, with those from the giant mice six times the
size of those from the runts.
"We thought, 'This is perfect,' " Dr. Tilghman said. "We will get
these mice and find out that the monogamous mice do not imprint and the
polygamous ones do. We will have done everything but prove Haig's
hypothesis."
Instead, to her great surprise, Dr. Tilghman found that both the
monogamous and the polygamous mice imprinted their genes.
Dr. Tilghman and her colleagues started by testing a gene called as
H19. That gene, she said, was always "perfectly imprinted" whenever
researchers examined it. A copy of the gene inherited from the mother
was always turned on full blast during fetal life and the copy inherited
from the father was shut down. If Dr. Haig's theory explained
imprinting, it would be expected that H19 would not be imprinted in the
monogamous mice. But it was. And, as with the polygamous mice, only the
copy from the mother was expressed.
Dr. Tilghman and her colleagues went on to look at seven more genes,
but to no avail. The imprinting patterns were not what she expected; the
predictions of Dr. Haig's theory were not fulfilled.
The debate over how to interpret these results has now begun.
Dr. Tilghman says that her data do not necessarily contradict Dr.
Haig's hypothesis because, for example, the monogamous mouse species may
have become that way fairly recently in evolutionary time and may have
retained imprinting as a remnant of its polygamous past.
Imprinting might also have the subsidiary effect of spurring
speciation, Dr. Tilghman said. Maybe the first thing that happens when
species diverge is that they start modifying genes -- imprinting them --
and eventually, as a consequence, they cannot mate. This idea would not
rule out imprinting as a reflection of a battle between the sexes, Dr.
Tilghman said. But it would say that the consequences of the battle are
not predictable and when it occurs independently, species might
diverge.
Dr. Hurst, a skeptic, said that the ready defense of Dr. Haig's
hypothesis perfectly illustrates what he sees as the problem with
evolutionary biology: no one says, before the fact, what experimental
results will cause them to reject a hypothesis. That way, he said, if
the data support a hypothesis they are embraced and if they do not
support it they are explained away. "It's a double standard," he said.
But Dr. Haig sees the matter differently.
"My interpretation would not be that the theory was wrong, but that
the theory doesn't explain everything," Dr. Haig said.
He said he welcomed Dr. Tilghman's experiments.
"I'm incredibly grateful to her," Dr. Haig said, "and I'm glad that
she came up with questions to which we don't have ready answers. That's
the purpose of science."