mostly appear translucent. They unfurl their delicate strands of DNA,
the better to expose genes that direct the workings of cells. But
in the cells of female mammals, which have two X chromosomes, one
sister of the pair remains shut off in a corner, tightly bundled and
The dark sister sits subdued, not directing the
production of thousands of proteins the way its voluble counterpart
does--and indeed as the other 22 pairs of human chromosomes do.
It didn't start out that way. Early in development,
when the female embryo is still a clump of undifferentiated cells,
the two X chromosomes are indistinguishable. But as the tiny female
embryo begins to develop form, fate intervenes for one of the sisters.
A crucial switch, known as the X inactivation center, turns on and
a signal sweeps down the chromosome. Genes linked with that chromosome's
X inactivation center become silent--never to be read into proteins.
As the cells of the embryo divide, the inactivated
X in each cell is duplicated like all the other chromosomes, but its
descendants remain silent even after being copied for many cell generations.
Unity of Sex
when all is said and done, is a method for mixing up genetic material.
Sex is such a good idea that species ranging from ferns to frogs
have converged on near-universal methods of sexual reproduction.
Considering how diverse the different sexually
reproducing species are, how much they share in common is remarkable.
Sexually reproducing species almost invariably have two genders--males
and females, of course--which usually exist in equal proportions,
and where opposites mate. In almost all sexually reproducing species,
too, gametes come in two forms--fat eggs and tiny, mobile sperm.
Usually, an organism inherits one complete set of genes from each
parent. When the time comes to make its own gametes, the individual
packages one complete gene set, shuffled from both parents, into
each of its gametes.
Sexually reproducing species widely rely on
sex chromosomes to determine sexual fate. Some exceptions do exist--some
reptiles, for example, rely on environmental cues instead. The
alligator is one such reptile. The temperature an alligator is
exposed to during a critical window of development determines
whether it will wind up a male or a female. More commonly, however,
heredity determines sex, which seems a more reliable way of producing
even proportions of the sexes.
Like sexual reproduction, sex chromosome systems
have evolved independently but in common ways. For example, flies,
worms, birds, and mammals (to restrict the examples to animals)
all have sex chromosomes that share basic characteristics, yet
their sex chromosomes are not related at the level of molecules
(they use different genetic pathways to determine sex).
Some themes guide the evolution and logic of
sex chromosome systems. One such these is that chromosomes come
in sister pairs, one inherited from the mother and one from the
father. Sex chromosomes, for example the X and the Y in mammals,
often look starkly different, yet are thought to derive from once-identical
sister chromosomes. Generally, one sex chromosome (the X in mammals)
looks like a standard-issue, medium-sized chromosome with thousands
of genes, while the other (the Y in mammals) tends to be shrunken,
clogged with junk DNA and almost bereft of genes. For example,
the human Y may house only two dozen genes, whereas an average
human chromosome houses one hundred times that number.
A general characteristic of sex chromosomes
is that they do not recombine as other pairs of chromosomes do,
except at their tips. Recombination is a process where sister
chromosomes align, intermingle, and trade fragments of genetic
material before segregating into egg or sperm cells.
Unrelated sex chromosome systems also invented
methods of dosage compensation, which balance the dose of gene
products derived from the sex chromosomes between males and females.
Genes act in complex concert with each other, and the levels of
gene products have been fine-tuned over the course of evolution.
In many species, females (XX) inherit two X chromosomes while
males (XY) inherit just one. Fruit flies that confronted this
dilemma have solved it by elevating expression from the male's
single X chromosome. Mammals, meanwhile, inactivate the second
X chromosome in female cells--or, more precisely, they permit
only one X chromosome to remain active in a cell.
Despite differences in the specifics of how
sex and sex chromosomes have evolved, various lineages, whether
worms or flies or buffalo, have confronted similar problems and
come up with similar solutions.
The inactive X isn't altogether silent, however.
In humans, at least 16 genes are known to escape X inactivation. These
genes are read into proteins whether they dwell on the inactive or
active sister. But many fewer genes seem to escape X inactivation
on the mouse X.
What came first--inactivation, as in mice, or escape
from inactivation, as in humans? How did these genes behave in the
mammalian ancestor? Did regulation switch during the evolution of
primates or rodents? Tracing X inactivation affords a rare opportunity
to reconstruct how a fundamental process in development evolved.
Sex chromosomes, like sex, have arisen independently
many times, in different kinds of organisms, yet have evolved in parallel
ways (see sidebar).
MANY measures, the inactivated X is distinct from its sister and from
other working chromosomes. For one thing, it just looks different
under a microscope--condensed and segregated against the nuclear wall,
instead of fanned out throughout the nucleus like the other chromosomes.
Also, the inactivated X is chemically altered.
Methyl groups tacked onto it appear to cement inactivation. The methyl
groups serve as an inheritable marker, ensuring that copies made from
the inactive X are also inactive.
During development, X inactivation of a chromosome
seems an indiscriminate process. If a fragment of an X chromosome
with the inactivation center (or even the inactivation center alone)
happens to become attached to another chromosome, the inactivating
force can roll down the other chromosome too, silencing many genes
with no previous experience of X inactivation. X inactivation spreading
into other chromosomes in this way usually kills a cell.
IS THE view at the "forest" level. But at the "tree" level of individual
genes, X inactivation appears less thorough, at least in humans. Some
of the genes that evade inactivation are clustered at the ends of
the chromosome, but others are sprinkled throughout the chromosome's
body, among inactivated genes.
The set of genes on the mouse X chromosome matches
the set on the human X, so the activity of genes can be compared between
the two species. And the mouse X chromosome appears to be more thoroughly
X-inactivated than the human X. Nearly all the genes known to escape
X inactivation in humans succumb to X inactivation in mice; only two
genes escape X inactivation in mice as well as in humans. No additional
genes that escape inactivation are known in mice.
Again, the question arises, what came first? Were
genes that escape X inactivation in humans active in the common ancestor
of mammals as well, or were they already subject to X inactivation,
as in mice? When did regulation change? To extrapolate ancestral conditions,
our lab undertook a comparative study of X inactivation, looking at
a host of mammalian species.
Humans and mice are about as distantly related
to each other as any two placental mammals can be. The placental mammals
are thought to have begun diverging from one another about 100 million
years ago. The various Orders--primates, rodents, carnivores, rabbits
and hares, horses and their relatives, etc.--separated quickly, perhaps
over the course of 30 million years. The precise sequence of their
branching has been difficult to decode and remains unresolved. Still,
surveying X inactivation in a range of placental mammals can reveal
whether inactivation or escape is more likely to represent the ancestral
Whether one or two copies of a gene are switched
on in any given female cell cannot be observed simply. Directly observing
gene activity in a wide range of species would have been impossible,
so our lab turned to DNA methylation as a way to assay X inactivation.
In every case that we and others have examined, methylation of an
X-linked gene in a region overlapping its start site correlates with
We studied three genes. Close relatives--all the
primates for instance, or subsets of rodents--always went together
in activating or inactivating each of the genes. So switching between
X inactivation and escape from inactivation does not appear to occur
often. But knowing the X inactivation status of one gene proved a
poor predictor of the X inactivation status of another gene in the
same species or Order.
Although X inactivation acts like a tidal wave
during development, during evolution X inactivation may have advanced
much more daintily--gene by gene or patch by patch. X inactivation
may appear as sweeping as it does only because the patchwork acquisition
of inactivation has become so complete that inactivation now appears
Y: THE DRIVING
of the evolution of X inactivation without speaking of the history
of the X and Y chromosomes leaves out critical context, something
like cataloguing the rarefied structures of orchids without considering
the habits of insects which evolved in concert with the flowers.
The X and Y chromosomes hardly look like a pair.
Textbooks have considered the mammalian X and Y fully differentiated
from each other.
Yet the X and the Y descended from a pair of identical
sister chromosomes. Early in the evolution of mammals, before monotremes
(echidnas and the platypus) and marsupials (opossum, kangaroos and
other Australian mammals) diverged from the placental mammals about
130 to 150 million years ago, the sex chromosomes arose. We can reconstruct
how this happened. Probably, a gene variant appeared that influenced
sex determination. In mammals the gene on the Y that triggers male
development was discovered a decade ago.
This evolutionary history applies generally when
sex chromosomes arise. A sex-affecting variant appears on a chromosome
that contains a haphazard collection of genes, most of which have
nothing to do with sex determination or differentiation.
Once a sex-determining gene appeared, recombination
became restricted around that gene, as if to ensure that it remained
confined to the male line. Eventually, the region of suppressed recombination
expanded to include nearly the whole Y. How recombination between
sex chromosomes becomes suppressed and how the suppression spreads
remain some of the most mysterious steps in sex chromosome evolution.
Now, the human X and Y chromosomes recombine only at their very tips.
Genetic recombination happens to be a vigorous
tonic for chromosomes, good for organisms, and the main point of sexual
reproduction. When recombination is suppressed, genetic integrity
comes tumbling down. The X chromosome can recombine along its whole
length with a sister X whenever it passes through a female. The Y
chromosome, however, never recombines along most of its length. Without
recombination, genetic disintegration follows. Exactly why recombination
is so useful remains to be plumbed. Though we may not know exactly
how recombination exerts its cleansing powers, we know that without
it DNA rearrangements accumulate, genes decay, and useless bits of
Mostly, cells scrupulously guard genetic integrity,
but when DNA becomes useless, repetitive and devoid of genes, cells
can toss it out without suffering any damage. Over time, the Y apparently
lost nearly all the genes it once shared with the X. The mammalian
Y is so degenerate that until recently many researchers believed that
it did nothing except determine sex.
When the Y degenerates, a male keeps only one copy
of each of the thousands of genes it once shared with the X. A female
still has two copies of these genes, intact on her X chromosomes.
To balance dosage of gene products, it benefits the male to amplify
expression of X-linked genes. Fruit flies stop at that to achieve
dosage compensation. Mammals have gone one step further. Expression
from the X was probably amplified in both males and females. Then
it became in the females' interest to inactivate one of the Xs.
Ultimately, these two things evolve in tandem:
the Y loses its similarity with the X, and X inactivation spreads.
But by what steps did (and does) this process occur during evolution?
Such a question would probably be impossible to answer for mammals
if the X and Y were fully differentiated as previously supposed. The
genes that escape X inactivation were not considered evolutionary
intermediates. They seemed to be flukes; perhaps their dosage doesn't
But their dosage does matter. Why else would most
of the genes that escape X inactivation in humans also have conserved
Y cousins (or homologs)? The homologs on the Y--and this has been
proven in the case of one gene--appear functionally interchangeable
with their cousins on the X.
Genes that escape X inactivation and have Y homologs
are caught in intermediate stages of evolution. Finding trapped intermediates
lets us reconstruct the pathway by which mammalian sex chromosomes
have evolved, just as trapped chemical intermediates can permit the
reconstruction of a biochemical pathway.
Several Y genes that have decayed, or whose function
has become limited, still have X homologs that escape X inactivation.
But no cases are known in which a gene is subject to X inactivation
yet has a Y homolog that remains conserved in structure and widely
expressed. In other words, Y degeneration or divergence appears invariably
to precede the expansion of X inactivation. Decay or divergence of
genes on the Y drives the acquisition of X inactivation, not the other
Yet, the story of the Y is not solely a story of
decay. Genes shared with the X have tended to be lost, but about half
of the genes on the human Y arrived there relatively recently. Chromosomes
are labile enough that genes can be transferred piecemeal from different
places. Once something lands on the Y, genes tend to be amplified
in copy number and rearranged. The new residents of the Y may be most
likely to survive on the rogue chromosome if they confer male-specific
advantages. Indeed, some of the newcomers to the human Y appear to
help in sperm formation. This is yet another example of similar, independent
evolutionary trajectories. The collection of genes on the human Y
is not related to the collection on the fruitfly Y, but in each case,
the Y appears to be a bastion of male-fertility factors.
may hardly seem like a personal subject, but the X and the Y show
a dramatic range of character: the two sister X chromosomes, one vivacious
and airy, the other silent and bundled into itself; the deadbeat-father
Y shirking its responsibilities and leaving more and more responsibility
to the X; the single-mother X making do by evolving complex adaptations.
Think of it as a poignant tale of fallibility--and of compensation.