The article “How Beach Life Favors Blond Mice” talks about the genetics

25 Jun The article “How Beach Life Favors Blond Mice” talks about the genetics

Question
The article “How Beach Life Favors Blond Mice” talks about the genetics behind the evolution of coat color in mice on beaches in the Florida panhandle. What is the selective force thought to drive the change in coat color? What genes are involved in this evolutionary change? Why is it thought that starting with a dark coat color, the mutations leading to the light “beach” coat would have to happen in a certain order? What violations of Hardy-Weinberg equilibrium would be involved in going from the dark inland coat to the light coat?

How Beach Life Favors Blond Mice
Sand Hills of Nebraska (Science, 28 August,
p. 1095). “We’re finally at the point where
we can start to identify the genes responsible
for phenotypic variation,” says Hoekstra.
And while working in Arizona, she says she
In June, at a meeting in Cold Spring Harpicked up far too many “presents” bulging with bor, New York, Hoekstra described the third
an angry rattlesnake. Fortunately, this trap of the three genes responsible for coat-color
weighs too little to have a snake inside, and no variation in Peromyscus mice and laid out her
deadly spiders are expected.
view of the order in which mutations leading
In a line of about 100 traps, Hoekstra to paler mice occurred. “We’re trying to
retrieves eight mice; her companions turn up reconstruct the evolutionary path, genetic step
four more, not a bad take for a full-moon by genetic step,” she says. “Understanding
night, when mice tend to be less active.
how characters evolve is a critical question,
The mice are part of a project started 6 years and she is bringing a significant contribution,”
ago to figure out the genetic changes that says developmental geneticist Claude Desplan
underlie adaptations these animals make to the of New York University. He adds that her work
world around them. Biologists have long mar- demonstrates that “one can really identify
veled at how oldfield mice living on beaches evolving traits.”
are much paler than those living inland, and
Hoekstra and her team are part of a
Hoekstra is searching for pigment genes genomics explosion in natural history studresponsible for the color variation.
ies. “This is an example of work
She’s combining molecular, devel… merging the ‘green’ and
opmental, genetic, and ecological
‘white’ side of biology, in which
approaches, including putting sciencemag.org
we learn about trait evolution
thousands of clay decoys on
from the biochemical levels
Podcast interview
beaches to test the effects of coat
within cells to how those traits
with author
color on predation risk and map- Elizabeth Pennisi.
are selected for or against in natping genes and testing pigment
ural populations,” says Hans
protein function in cell cultures. “We’re attack- Ellegren, an evolutionary biologist at Upping the system from all sides,” says Hoekstra.
sala University in Sweden. Mark McKone, a
On this trip, Hoekstra and her team are biologist at Carleton College in Northfield,
looking not just at coat-color variation but Minnesota, agrees: The work “could be a
also at variation in burrow-building. Most model for how to approach evolution in the
deer mice build short, shallow burrows; old- postgenomic period,” when genetic inforfield mice go for deeper, longer ones. Back mation and tools are more readily available.
in the lab, Harvard graduate student Evan
Kingsley is trying to pin down the genetics New tools, classic model
of tail length: Mice in forests have longer Hoekstra’s team represents the latest genertails. Recently, Hoekstra postdoc Catherine ation of researchers tracking down genes
Linnen described a genetic change under- that underlie so-called quantitative traits
lying light-colored deer mice that match the such as height or body mass, which—

FREEPORT, FLORIDA—It’s a hot, sticky
July night here in western Florida, but to
Hopi Hoekstra, it feels like Christmas Eve.
Hoekstra, a Harvard University evolutionary
biologist, and her field crew have set out
more than 400 small metal boxes, throwing a
handful of sunflower seeds into each box
before setting it on the ground, usually next
to a mound of sand representing the debris
from a mouse burrow. When she inspects
these live animal traps the following morning, she says it will be like “unwrapping
presents.” Her eagerness is palpable.
“You’re going to be blown away by this
field,” graduate student Jesse Weber had told
Hoekstra when they first drove down a sand
road into the Lafayette Creek Wildlife Management Area, a 13-square-kilometer expanse
of overgrown fields kept open in part by controlled burns. Never before had Weber and
Harvard postdoc Vera Domingues seen such a
dense concentration of burrows dug by the
oldfield mice, Peromyscus polionotus, that
they study.
By 7:30 the next morning, Hoekstra,
Domingues, Weber, and Harvard undergraduate Diane Brimmer are making their
way from trap to trap, sidestepping fire ant
hills, prickly pear, and thorny vines while keeping an eye out for pygmy rattlers. Typically, the
trapdoors are still ajar, and at most a grasshopper or two jumps out into Hoekstra’s face as
she empties the sunflower seeds. But three traps
down the line, the door is closed and Hoekstra
senses something inside. At past field sites,
she’s had to worry about lethal spiders crawling
in, positioned to nab any unsuspecting hand.

1330

11 SEPTEMBER 2009

Online

VOL 325

SCIENCE

Published by AAAS

www.sciencemag.org

CREDITS (LEFT TO RIGHT): SHAWN CAREY/MIGRATION PRODUCTIONS; J. B. MILLER/FLORIDA PARK SERVICE

A young evolutionary biologist tackles the genetic complexity of a
classic case of adaptation in mice

Downloaded from www.sciencemag.org on October 6, 2009

NEWSFOCUS

Lighten up. Several genes transformed mainland
mice (left) into paler beach mice that blend in better
with their environment.

unlike, say, eye color—vary by degree and
are influenced by multiple genes. It is
painstaking work.
Researchers home in on such genes
through intensive breeding studies combined with careful analysis of trait characteristics: spots, stripes, and so on for coat color;
depth, length, and angle for burrowing
behavior. They correlate the traits with specific markers in genetic maps to pinpoint
stretches of DNA known as quantitative trait
loci (QTLs) that contain the genes of interest. “This is done well in insects but is much
more difficult in mammals,” says Desplan.
Over the past 20 years, several studies have
identified QTLs in mammals, but few have
managed to narrow the search to specific

Her animal of choice is a textbook case of
adaptation. Peromyscus mice are distant relatives of house mice. For more than a century,
researchers had observed them in the wild,
describing their looks and behaviors. In 1909,
light-colored P. polionotus were discovered
on Florida’s barrier islands, a sharp contrast
to dark-brown, gray-bellied mainland mice of
the same species. Some 6000 years ago, dark
oldfield mice moved into these newly formed
beaches and islands. Today, eight subspecies
of these light-colored P. polionotus exist on
Florida’s coasts.
In the late 1920s, natural historian Francis
Sumner guaranteed P. polionotus a place in
the textbooks when he drove from Florida’s
Gulf Coast inland 150 kilometers collecting mice in eight places along the way, noting a correlation between soil and mouse
color. When he started, he was convinced
that humidity caused the variation in color.

light areas of their bodies, traits duly noted for
each individual. This variation indicated that
more than one gene was involved, but because
the second generation still contained some
mice that looked like the parents, Hoekstra
knew that relatively few genes were important. “It wasn’t one, it wasn’t 100,” Hoekstra
recalls. So she decided to go after them all.
Weber and Cynthia Steiner, now at the
San Diego Zoo Institute for Conservation
Research in California, developed and
applied a set of more than 100 microsatellite
markers, small pieces of variable DNA
located across the genome. They correlated
the markers with the presence or absence of
the various color pattern traits. That work
yielded three hot spots—QTLs—that seemed
to determine what the mice looked like.
The researchers looked at the sequences of
the house mouse and rat genomes for pigmentrelated genes at those locations and found

DISTRIBUTION OF BEACH AND MAINLAND MICE

Mainland
mouse

Lafayette Creek
mice

Santa Rosa Island
beach mouse

Anastasia Island
beach mouse

Alabama
beach mouse

LOCATION

Perdido Key
beach mouse

Choctawhatchee
beach mouse

St. Andrew
beach mouse

Pallid
beach mouse*
*extinct subspecies

Southeastern
beach mouse

CREDIT: ADAPTED FROM C. STEINER ET AL., MOL. BIOL. EVOL. 26, 35 (2009), FIG. 1

Mouse of a different color. Mice from different locales have evolved site-specific coat colors, except those at Lafayette Creek, which have a variety of pelt patterns.

genes, let alone identify mutations that
result in changes such as coat color.
The discovery in 2005 by David Kingsley
of Stanford University in Palo Alto, California, and colleagues that a change in the
ectodysplasin gene led to the loss of armor in
freshwater sticklebacks (Science, 25 March
2005, p. 1928) “got the field excited,” says
Hoekstra. It was the first QTL study using
natural populations to come up with a gene
that was not already suspected to be
involved and, later, to pin down its mutation.
Hoekstra hopes to go into more detail with
her mouse studies. Whereas Kingsley
focused on the gene with the biggest effect,
she is searching for several genes. “If we
identify multiple genes and understand the
interactions between those genes, we can
also learn something new about evolutionary processes,” she explains.

By the project’s end, he was more convinced that genetics caused the differences,
driven by selection for camouflage. “It’s
one of the best studies of intraspecific variation,” says Hoekstra.
Giants in evolutionary biology, including
Ernst Mayr, Theodosius Dobzhansky, John
Maynard Smith, J. B. S. Haldane, and Sewall
Wright, have cited the work as a classic
example of adaptation. Others followed
Sumner, looking at various aspects of beach
mice ecology, but they were unable to pin
down the genetics. Hoekstra saw an opportunity: “We now have the molecular tools to
answer the questions that they were asking
more than a half-century ago.”
She and her colleagues bred dark and light
mice, then generated 800 second-generation
offspring. These hybrid mice differed in their
stripes and splotches and the extent of dark or

www.sciencemag.org

SCIENCE

VOL 325

Published by AAAS

Downloaded from www.sciencemag.org on October 6, 2009

NEWSFOCUS

promising candidates. One was Mc1r, which
codes for a receptor protein in pigmentproducing cells. Hoekstra was at first skeptical. In her studies of black pocket mice on volcanic rock in Arizona, one version of that gene
was responsible for the black mice and another
for light mice; it was not clear how the gene
might play a role in determining fine details
such as nose blazes and tail stripes.
But not only did they prove that Mc1r was
involved, they also found a single-base
change that led to an amino acid mutation that
dampened receptor activity (Science, 15 July
2005, p. 374; 7 July 2006, p. 101). A second
candidate gene, Agouti, panned out as well.
In this case, the versions of the gene in dark
and light mice were identical; yet the gene in
beach mice was much more active, leading
to much more messenger RNA and presumably protein that reduced dark-pigment pro-

11 SEPTEMBER 2009

1331

NEWSFOCUS

Mouse maven. Hopi Hoekstra combines
molecular and field expertise to study
the genetics of wild mice.

1332

laboratory mice, made for dirty-blond mice.
Corin was also active in the hair follicles of
oldfield mice, Hoekstra reported in June at
“Evolution: The Molecular Landscape” in
Cold Spring Harbor. The gene in light and
dark mice was almost the same, but it was
much more active in light mice. Thus, as
with Agouti, a change in regulation may be
key to the change in coat color.
In the simplest scenario, the effect of
these genes would be additive: Two “light”
versions of the variable genes would lead to a
paler mouse than one version would, and the

11 SEPTEMBER 2009

VOL 325

SCIENCE

Published by AAAS

palest mice would have “light” versions of all
three. But that’s not the case with Agouti,
Corin, and Mc1r. These genes have epistatic
interactions: A “dark” Agouti version counters any lightening effect of a “light” Corin
or Mc1R, for example.
These epistatic effects can dictate the
order in which alleles in a population must
pop up in order to be selected for and spread.
“You need to have the agouti allele first,”
says Hoekstra, because the “light” versions
of Corin or Mc1r would be invisible to selection if only the “dark” agouti were present.

www.sciencemag.org

CREDIT: E. PENNISI/SCIENCE

duction, particularly in the cheeks, tail, and
eyebrows, Hoekstra, Weber, and Steiner
reported in 2007.
They had a false start with the third
region identified in the QTL studies. Harvard graduate student Emily Jacobs-Palmer
eventually ruled out several pigmentation
genes, including a promising one called Kit
that turned out to lie outside the QTL. Then
last year, Bruce Morgan of Harvard Medical
School in Boston and his colleagues
reported that mutating a gene called Corin,
which was expressed in the hair follicles of

Downloaded from www.sciencemag.org on October 6, 2009

Park. She studied the biomechanics of invertebrates
throughout the school year. During that time, James
Patton, curator of mammals at the Berkeley Museum of
Vertebrate Zoology, got her hooked on four-legged furry
creatures by taking her to trap gophers in Arizona. And
before starting graduate school, she spent 3 months as
shipboard mammalogist on a joint Japanese, Russian,
and American expedition to collect animals in the
Kuril Islands off Russia.
Her Ph.D. dissertation at the University of Washington, Seattle, involved months of fieldwork in the Andes
tracking down a sex chromosome polymorphism in mice.
Some females seemed to have both a big and a small X,
which later proved to be a Y chromosome, even though
these females were completely fertile, producing more
young than the typical female with two X chromosomes.
“This was an oddball system,” Hoekstra recalls. Afterward, “I got interested in more general questions.”
Fascinated by the genetics underlying adaptation,
Melding Mammals and Molecules to Track Evolution
she spent her postdoc trapping black mice on ancient
Arizona volcanoes and tracking down the gene responsiSelf-described as a bubbly California girl, Hopi Hoekstra entered the Uni- ble for the change. In these field studies, she developed a yen for her
versity of California, Berkeley (UCB), not thinking about being a scientist. camp meal of choice: cold SpaghettiOs and mini meat balls straight from
Her goal was to become the U.S. ambassador to the Netherlands—both her the can, with a Miller Light.
parents are Dutch—and an accomplished collegiate volleyball player. Then
She considers herself a molecular person: “We’re interested in the molshe got her first summer job: Dressed in white, she hiked the Berkeley Hills ecules that are important to the organism,” she says. Yet she also knows
just east of campus, a tick target for researchers assessing where and when just how much cornmeal it takes when skinning a mouse to ensure the pelt
hikers were most susceptible to attacks by Lyme disease–transmitting ticks. won’t be greasy and that shrews have fragile skin that’s hard to pull off.
“It still makes me itch just to think about it,” she says.
The breadth of projects include an analysis of shrew venom proteins
But the experience made Hoekstra itch for more fieldwork and, even- and a collaboration on a genetic study of mice in Bulgaria that seem to
tually, a life as a biologist. Two years ago, she moved from the University cooperate to build large mounds that they coinhabit to get through
of California, San Diego, to Cambridge, Massachusetts, as a Harvard Uni- harsh winters.
versity evolutionary biologist. She is also currently curator of mammals at
“Being able to be a molecular biologist and be comfortable with the
Harvard’s Museum of Comparative Zoology. Although only in her mid- whole organism—few people do that as well as Hopi, and that’s where
30s, “Hopi has rapidly made herself a name in the evolutionary biology progress [in the field] will be made,” says Mark McKone, a biologist at
community,” says Hans Ellegren of Uppsala University in Sweden. Her Carleton College in Northfield, Minnesota. “When you put [her research]
honors include a young investigator award from the Arnold and Mabel together, it’s more than the sum of its parts.”
Beckman Foundation and prizes from her professional societies and her
Hoekstra doesn’t get out into the field much anymore. Instead, she
universities. “She’s just about one of the deepest thinkers in the area,” lives vicariously through her students and postdocs, with the goal of spendsays Carlos Bustamante of Cornell University, who adds that her beach ing time at least once with each of them in the field. “When they have a
mice experiments “are beautifully thought out and designed.”
really good day, they call and leave a message,” she says, or send a photo
She traces her professional roots back to her UCB experience, where from their phones, such as an image of 44 traps stacked up against a brick
she managed to do research almost year-round, even as an undergraduate. wall, signaling that their trapping yielded a bonanza. “They just send a picOne summer, she analyzed pack rat middens in Yellowstone National ture [without words] because they know I know what it means.”
–E.P.

CREDIT: E. PENNISI/SCIENCE

Burrowing in
Weber has taken on an even more challenging
project: using these mice to look at the genetics underlying burrowing behavior. “It’s pathbreaking work on the evolution of behavior in
a natural environment,” says field biologist
Peter Grant of Princeton University. “QTL

studies are widespread in general but rare in
behavior studies of organisms in nature.”
Unlike coat color, almost nothing is
known about genes that might guide burrowing. Yet oldfield mice and their sister species,
deer mice, differ dramatically and, it seems,
consistently in the burrows they build. The
latter tend to knock off their digging less than
10 centimeters down. Oldfield mice shovel
down 1 meter, even 2, hollow out a nest chamber, and then excavate an escape tunnel that
tends to shoot directly back up to just below

Here in Freeport, he’s doing some
ground-truthing. He catches the mice in the
burrows so he can correlate their DNA with
the tunnels’ dimensions. He picks what looks
like a freshly dug hole, shovels out some dirt,
then drops to his knees to scoop the sand and
clay away with his hands until he sees a
round, light-colored spot in the wall of the
hole. His f inger easily pokes through it,
revealing it to be a plug of sand blocking the
burrow tunnel. Alternating between shoveling and scooping and probing the tunnel with
a long, flexible, plastic tube
(sprinkler tubing), he excavates
the tunnel, eventually breaking
Bagging burrows. The beach mice
field crew measures a mouse burrow
into a widened area filled with
after making a cast of its tunnels.
nesting material. “This nest is
gigantic,” he says.
He confers with Hoekstra
about where she should stand in
anticipation of mice emerging
from the invisible escape hatch.
She shifts to the right a halfmeter, then bends her legs
slightly, hands on her knees. She
looks like the volleyball player
she used to be, expecting a
serve, except she’s looking
down, not up.
Weber pokes the tubing in a
little farther. Suddenly, two heads
pop up about 20 centimeters to
Hoekstra’s right. She dives to
clamp her gloved hands over the
heads. But as she peeks through
her f ingers, one dashes out
between her legs, and the other
heads full speed in the opposite
direction. Both she and Weber
pursue that one, darting from
the surface. The mice plug up the burrow bush to bush after the mouse until finally
about 15 centimeters from the entrance, seal- Weber has it in hand. The other is long gone.
ing themselves safely in underground.
While Weber measures the size and
Back in the lab, Weber has filled 10 boxes, shape of the burrow, Hoekstra measures the
each 122 cm by 152.5 cm by 92.5 cm tall, sacrificed mouse, then dissects out its liver
with 1.5 tons of premium playground sand. to save for DNA tests, removes the skin to
He has crossed oldfield with deer mice, then mount the pelt for future studies of the color
crossed their offspring back with either par- pattern, and saves the skeleton for the
ent, and he’s looking at what sorts of burrows museum’s collections. The sun sets bright
these backcrossed progeny dig. The distribu- red in front of her, and the full moon is a big
tion of burrow sizes in this second generation white ball in the sky behind her.
will provide a rough indication of how many
Weber and Hoekstra seem tired but congenes are involved in determining burrow- tent. The burrows they’ve dug up were
ing behavior. Weber squirts household insu- deeper and longer than usual; shoveling
lating foam from a spray can down the bur- heavy, wet sand was tough going. They’ve
rows. The foam expands to fill the nest and been up since before dawn and have an
passageways and hardens to provide a three- evening of setting traps ahead of them.
dimensional model of the burrow. So far he’s “But once in a while, it’s good if it’s hard,”
tested 200 mice and has partially filled the Hoekstra says. “Then you appreciate it
attic of the Museum of Comparative Zoology when it’s easy.”
with casts of their burrows.
–ELIZABETH PENNISI

www.sciencemag.org

SCIENCE

VOL 325

Published by AAAS

11 SEPTEMBER 2009

Downloaded from www.sciencemag.org on October 6, 2009

By backcrossing the second-generation
mice with their parents and f iguring out
which version of each of the three genes the
offspring had, Hoekstra’s team was able to
tease out the interactions among the genes.
The light-mouse version of Corin lightens
the coat only when the light-mouse versions
of both of the other genes are also present,
Hoekstra reported. Thus, it is likely that
genetic change in Corin occurred after the
changes to Mc1r and Agouti.
Meanwhile, Domingues and graduate
student Lynne Mullen are trying
to track down the exact base
changes involved in the Agouti
and Corin regulatory regions.
Working with postdoc Brant
Peterson, they are figuring out a
way to sequence 200,000-base
chunks surrounding each of these
genes in multiple individuals.
They plan to scan for differences
that correlate with coat color patterns. “We will probably see lots
of differences,” says Hoekstra.
“The question is, ‘What are the
important ones?’ ”
The work Domingues is doing
here might help answer that
question. The landscape is dotted
with spots of white sand sparsely
broken up by vegetation amid
fields solidly covered with low
bush and plants, and in a few
places, meter-tall trees have
taken hold. When local fish and
wildlife managers first directed
her to this spot, Domingues
expected the mice to be uniformly dark, but quite a few had
beachlike features.
Hoekstra and Domingues eagerly discuss
the pelage of each catch. How far a dark
stripe extends down the tail, the expanse of
white on the cheeks, the presence of a nose
blaze all matter, as they signal something
interesting going on in the genetics of these
supposed-to-be-dark mainland mice.
Domingues plans to try to pin down the
genes—and mutations—involved in all the
variation she sees, using the three genes
implicated in beach mouse paleness as a
jumping-off point.

1333