CONTENTS Volume 339 Issue 6115

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 5
COVER
Head of a juvenile Nile crocodile (Crocodylus niloticus).
The irregular polygonal scales on the face and jaws of
crocodiles are not controlled genetically. Instead, the patterns
are generated through a self-organizational process in which
the stiff skin cracks in a tensional stress field as the crocodile
grows. See page 78.
Photo: Michel C. Milinkovitch and Adrien Debry
DEPARTMENTS
9 This Week in Science
11 Editors’ Choice
13 Science Staff
99 New Products
100 Science Careers
page 22
EDITORIAL
10 Reflecting on Goals for Science
Bruce Alberts
NEWS OF THE WEEK
14 A roundup of the week’s top stories
NEWS & ANALYSIS
16 H5N1 Researchers Ready
as Moratorium Nears End
17 China Partners With Gay Groups
on HIV Screening
19 Budget Crunch to Shrink
Science Programs at Chicago’s
Field Museum
20 Materials Research Society
Fall Meeting and Exhibit
The Power of Friction
A Boost for Lithium-Sulfur Batteries
Space Solar Cells With a Down-to-Earth Cost
NEWS FOCUS
22 Death of a Star
25 The Immune System’s Compact
Genomic Counterpart
28 Europe’s ¤2 Billion Bet on the Future
LETTERS
30 NextGen VOICES
BOOKS ET AL.
34 Failed Evidence
D. A. Harris, reviewed by W. C. Thompson
35 Sick Societies
D. Stuckler and K. Siegel, Eds.,
reviewed by L. P. Fried
POLICY FORUM
36 Threats from India’s Himalaya Dams
R. E. Grumbine and M. K. Pandit
PERSPECTIVES
38 Getting Past Polyproline Pauses
A. R. Buskirk and R. Green
>> Reports pp. 82 and 85
39 Entropies of Adsorbed Molecules
Exceed Expectations
J. F. Weaver
40 Science, New Media, and the Public
D. Brossard and D. A. Scheufele
42 Negative Temperatures?
L. D. Carr
>> Report p. 52
43 Bioinspired Oxidation Catalysts
M. Largeron and M.-B. Fleury
44 How Was Early Earth Kept Warm?
J. F. Kasting
>> Report p. 64
CONTENTS continued >>
Explore our rich online offerings,
including multimedia, news, and our
two research journals—Science Signaling
and Science Translational Medicine—
at WWW.SCIENCEMAG.ORG
page 34
Published by AAAS
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 7
CONTENTS
page 91
pages 42 & 52
REPORTS
46 A Stringent Limit on a Drifting
Proton-to-Electron Mass Ratio from
Alcohol in the Early Universe
J. Bagdonaite et al.
The proton-to-electron mass ratio inferred
from methanol lines in a distant galaxy
is in accord with the laboratory value.
49 Alignment of Magnetized Accretion
Disks and Relativistic Jets with
Spinning Black Holes
J. C. McKinney et al.
Numerical simulations reveal a previously
unknown mechanism to align accretion disks
and jets with rotating black holes.
52 Negative Absolute Temperature for
Motional Degrees of Freedom
S. Braun et al.
A gas of potassium atoms in an optical
lattice displays an inverted population
of energy levels.
>> Perspective p. 42
55 Current-Driven Spin Dynamics of
Artificially Constructed Quantum Magnets
A. A. Khajetoorians et al.
Spin-polarized scanning tunneling microscopy
is used to exert the spin-transfer torque on a
small atomic cluster.
59 Strategic Redox Relay Enables a Scalable
Synthesis of Ouabagenin, a Bioactive
Cardenolide
H. Renata et al.
A synthesis showcases multiple creative
indirect methods of selectively hydroxylating
saturated carbon centers.
64 Hydrogen-Nitrogen Greenhouse Warming
in Earth’s Early Atmosphere
R. Wordsworth and R. Pierrehumbert
Absorption of solar radiation associated
with H
2
-N
2
collisions may have been key
to warming the atmosphere of early Earth.
>> Perspective p. 44
67 Highly Variable El Niño–Southern
Oscillation Throughout the Holocene
K. M. Cobb et al.
Coral records show the El Niño–Southern
Oscillation may be less sensitive to past
climate forcing than previously thought.
70 The Spatial and Temporal Origin of
Chandelier Cells in Mouse Neocortex
H. Taniguchi et al.
Chandelier interneurons migrate along
well-defined paths as they mingle with
pyramidal neurons to build circuits.
74 An Update of Wallace’s Zoogeographic
Regions of the World
B. G. Holt et al.
Mapping the geographic distribution and
phylogenetic relationships of 21,037
vertebrate species yields 11 realms.
78 Crocodile Head Scales Are Not
Developmental Units But Emerge
from Physical Cracking
M. C. Milinkovitch et al.
Tension-induced cracks generate random
polygonal scales in the keratinized skin
on crocodile heads and jaws.
82 Translation Elongation Factor EF-P
Alleviates Ribosome Stalling at
Polyproline Stretches
S. Ude et al.
85 EF-P Is Essential for Rapid Synthesis of
Proteins Containing Consecutive Proline
Residues
L. K. Doerfel et al.
A universally conserved translation factor
facilitates synthesis of peptides that would
otherwise cause ribosome stalling.
>> Perspective p. 38
88 Para-Aminosalicylic Acid Acts as an
Alternative Substrate of Folate Metabolism
in Mycobacterium tuberculosis
S. Chakraborty et al.
A drug used against tuberculosis for the past
50 years is a metabolic poison and not an
enzyme inhibitor.
91 Dynamic Persistence of Antibiotic-Stressed
Mycobacteria
Y. Wakamoto et al.
Bacterial cells that divide before a pulse
of catalase expression survive the application
of the antibiotic isoniazid.
96 The End of History Illusion
J. Quoidbach et al.
Even though we know that we’ve changed over
the years, we believe, mistakenly, that we
won’t change much in the future.
SCIENCE (ISSN 0036-8075) is published weekly on Friday, except the last week in December, by the American Association for the Advancement of Science, 1200
New York Avenue, NW, Washington, DC 20005. Periodicals Mail postage (publication No. 484460) paid at Washington, DC, and additional mailing offices. Copyright © 2013
by the American Association for the Advancement of Science. The title SCIENCE is a registered trademark of the AAAS. Domestic individual membership and subscription (51 issues):
$149 ($74 allocated to subscription). Domestic institutional subscription (51 issues): $990; Foreign postage extra: Mexico, Caribbean (surface mail) $55; other countries (air assist
delivery) $85. First class, airmail, student, and emeritus rates on request. Canadian rates with GST available upon request, GST #1254 88122. Publications Mail Agreement Number
1069624. Printed in the U.S.A.
Change of address: Allow 4 weeks, giving old and new addresses and 8-digit account number. Postmaster: Send change of address to AAAS, P.O. Box 96178, Washington, DC
20090–6178. Single-copy sales: $10.00 current issue, $15.00 back issue prepaid includes surface postage; bulk rates on request. Authorization to photocopy material for
internal or personal use under circumstances not falling within the fair use provisions of the Copyright Act is granted by AAAS to libraries and other users registered with the Copyright
Clearance Center (CCC) Transactional Reporting Service, provided that $30.00 per article is paid directly to CCC, 222 Rosewood Drive, Danvers, MA 01923. The identification code for
Science is 0036-8075. Science is indexed in the Reader’s Guide to Periodical Literature and in several specialized indexes.
Published by AAAS
9
EDITED BY STELLA HURTLEY
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

R
A
L
F

K
A
E
H
L
E
R

(
K
I
P
A
C
/
S
L
A
C
)

A
N
D

J
O
N
A
T
H
A
N

M
C
K
I
N
N
E
Y

(
U
M
D
)
;

C
E
N
T
E
R

F
O
R

M
A
C
R
O
E
C
O
L
O
G
Y
,

E
V
O
L
U
T
I
O
N

A
N
D

C
L
I
M
A
T
E
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013
Negative Is Hotter
A common-sense perception of temperature tells
us that the lower the temperature, the colder
it is. However, below absolute zero, there is a
netherworld of negative temperatures, which
are, counterintuitively, hotter than positive tem-
peratures. Usually, such states are achieved in
the laboratory and are characterized by a higher
occupation of high-energy versus low-energy
states. This is most easily done for systems that
have a finite spectrum of energy states, bounded
from above and below. Braun et al. (p. 52; see
the Perspective by Carr) achieved negative tem-
perature for a system in which its spectrum was
only bounded on one side. Starting with a gas of
39
K bosonic atoms with repulsive interactions in
a dipole trap and an optical lattice, a final state
with negative temperature was reached where
the atoms attract each other.
Placing the O’s
in Ouabagenin
Whereas enzymes are remarkably
adept at selectively oxidizing
saturated carbon centers, these
reactions seriously challenge
chemists. In a 19-step synthe-
sis of ouabagenin, Renata
et al. (p. 59) showcase a
range of creative indirect
methods to install the six
hydroxyl groups in the
steroid’s framework. These
include a solid-state photo-
chemical transformation, as well
as dehydrogenation sequences that
place olefins in proper position for
oxygenation. The route also yields several
intermediates poised for elaboration to distinct
analogs for exploratory medicinal chemistry.
Faint Young Sun
During the first ~billion years after the forma-
tion of Earth, the Sun was significantly less
luminous than today, delivering considerably
less solar energy to our planet. Nevertheless,
the geological record shows that Earth harbored
liquid water and was not frozen—as would be
expected on the basis of the reduced solar out-
put. The presence of liquid water on Earth dur-
ing this time is referred to as the “faint young
Sun” paradox. Wordsworth and Pierrehumbert
(p. 64; see the Perspective by Kasting) suggest
a new explanation for this mystery, invoking
absorption of solar radiation owing to collisions
between atmospheric H
2
and N
2
as a source of
heating. This mechanism could have supplied
enough extra warming to keep surface tempera-
tures above 0°C.
Building Cross Connectivity
Chandelier cells innervate the initial segment
of axons from pyramidal neurons and are thus
placed to regulate pyramidal cell circuits in the
brain. Chandelier cells of mice are marked by
expression of the NKX2.1 transcription factor.
Taniguchi et al. (p. 70, published online 22
November) followed the development of these
neurons and found that chandelier cells originate
from the ventral germinal zone. The nascent cells
migrate and integrate with cortical neurons fol-
lowing specific developmental pathways.
Next-Generation
Biogeography
In 1876, Alfred Russel Wallace mapped the zoo-
geographical regions of the world,
based on the distributions and
taxonomic relationships of
broadly defined mam-
malian families. Wal-
lace’s classification of
zoogeographical re-
gions became a cor-
nerstone of modern
biogeography and a
reference for a wide
variety of biological
disciplines, including
global biodiversity and
conservation sciences. Holt
et al. (p. 74, published online
20 December) present a next-gen-
eration map of wallacean zoogeographic regions,
incorporating phylogenetic data on >20,000
vertebrate species to discern and characterize their
natural biogeographic patterns.
All About Noise
How individual cells behave within a larger “aver-
age” population can be surprising. Wakamoto et
al. (p. 91) developed a method for investigating
the consequences of phenotypic variability in
single mycobacterial cells exposed to the pro-
drug, isoniazid. Isoniazid needs to be activated by
bacterial catalase. In the isoniazid–mycobacterium
system, random fluctuations in catalase activity
were important for cell survival. Because catalase
is essential, it cannot be ablated; however, cata-
lase activity pulsed randomly in the mycobacteria.
Thus, a subpopulation of individual cells manage
to avoid being killed by the activated antibiotic.
Older and Wiser
Do we ever stop growing up? Quoidbach et al.
(p. 96) elicited estimates of people’s personal-
ity, values, and choices and compared how
much, for instance, 33-year-olds believed that
they would change in the next 10 years with
how much 43-year-olds reported that they had
changed in the past 10 years. For groups span-
ning 18 to 68 years of age, people of all ages
described more change in the past 10 years
than they would have predicted 10 years ago.
Black Hole Alignment
In interpreting observations of black holes,
it is important to assess the extent to which
the accretion disks around black holes and
the jets that they produce align with the spin
of the black hole. Based on numerical simu-
lations of magnetohydrodynamic fluid flows
around rotating black holes, McKinney et al.
(p. 49, published online 15 November) de-
scribe a previously unknown mechanism that
acts to align the accretion disk and jet with
the spin axis of black holes. Unlike previ-
ous mechanisms, this effect applies to thick
disks, like the one thought to exist around
the massive black hole at the center of
our galaxy.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Reflecting on Goals for Science
SINCE BECOMING THE EDITOR-IN-CHIEF OF SCIENCE IN 2008, ONE OF MY MAJOR GOALS HAS BEEN
to explore how our widely read publication can help to improve science education at all lev-
els—from the very first years of schooling through graduate school. Other goals have been the
promotion of important standards in the scientific community and publishing new research
results of the highest quality. Here, I highlight some efforts designed to promote these aims.
Science publishes many editorials on the subject of education. Donald Kennedy, my pre-
decessor as Editor-in-Chief, introduced our Education Forum. This has been complemented
by special issues on education: Education and Technology (2009); Science, Language,
and Literacy (2010); and Investing Early in Education (2011). The next education special
issue—my last—will be Grand Challenges in Science Education (April 2013). And Science
is now publishing peer-reviewed research in the science of education.
Over the course of the past 4 years, Science has also held two
contests designed to help spread good science education across the
globe. The first highlighted the best English-language, open-access
science education resources on the Web. All of the winners of this
Science Prize for Online Resources for Education (SPORE) have
published essays in Science, and these have now been compiled into
a free 60-page electronic booklet available on our special education
Web site.* The second contest, the Science Prize for Inquiry-Based
Instruction (IBI), focused on college-level coursework. The 2011
winners published their essays in 2012, and the 2012 winners will
be published throughout 2013. They cover a very broad range of
science, with titles that include Discovering Nanoscience, Engag-
ing Students in Earthquakes via Real-Time Data and Decisions,
Engaging Undergraduates in Global Health Technology Innovation,
and Student-Directed Discovery of the Plant Microbiome and Its
Products. Each IBI winner also provides a supplement with instructions for teaching their
science inquiry module. All of these resources are made freely available,

and it is our hope
that their open-access publication will encourage the spread of outstanding science teach-
ing, just as the publication of the best scientific research in Science stimulates new science.
With regard to standards for scientific research, Science has recently completed an exper-
iment designed to improve the process of peer review for research in all of the scientific
disciplines we publish. As a result, in 2013 we will be moving to a system that provides an
opportunity for each reviewer to comment on the other reviews obtained for the same sub-
mitted manuscript, just prior to the decision that is made by a small team of our editors as to
whether to reject, accept only with revision, or accept a manuscript for publication. In this
way, we hope to minimize the amount of revision required for manuscripts that are judged to
be nearly suitable for publication.
Finally, Science’s Senior Editorial Board has increased our awareness of outstanding
scientific findings that are very difficult to convey clearly in a short format. Some con-
tain numerous micrographs that require high-resolution displays, while others require more
space due to a complexity of analysis. Therefore, in a new offering, selected papers will
appear online with the length and integration of methods and data that are most appropriate
for reporting the research. These papers will be represented in the print magazine by special
structured abstracts, but they will be listed as research articles in the print Table of Contents,
being indexed and cited the same way as the shorter papers that appear in full in print. They
will also appear in full in all digital versions of Science, such as those designed for mobile
devices. A similar approach is planned for selected review articles.
Change is good, and for this and many other reasons, we look forward to a new year
at Science.
C
R
E
D
I
T
S
:

(
T
O
P
)

T
O
M

K
O
C
H
E
L
;

(
L
E
F
T
)

I
S
T
O
C
K
P
H
O
T
O
.
C
O
M
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 10
EDITORIAL
10.1126/science.1234351
– Bruce Alberts
Bruce Alberts is Editor-
in-Chief of Science.
*www.sciencemag.org/site/extra/education

http://scim.ag/Te3EmY
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

11
EDITORS’CHOICE
C
R
E
D
I
T
S
:

T
H
I
N
K
S
T
O
C
K
EDITED BY KRISTEN MUELLER AND MARIA CRUZ
Continued on page 12
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013
a key to understanding those topics. It is also
known that volcanic activity can be influenced
by surface mass loading and the resulting
isostatic adjustment of the underlying Earth
and that glacial cycles change the distribution
of water and ice on the crust. Therefore, the
possibility that volcanism might vary in accor-
dance with Milankovich periodicities (changes
in the amount and distribution of solar energy
incident on Earth due to variations in its orbital
configuration) in response to glacial cycling has
been an intriguing one. Kutterolf et al. present
the most comprehensive data set yet to address
this idea, developed from an extensive collec-
tion of marine sediment cores from around the
Pacific Ocean basin. From these cores, they
were able to show that the frequency of circum-
Pacific volcanism varied with a 41,000-year
period, the obliquity band of Milankovich
cycles, and that changes in volcanic activity
lagged slightly behind glacial unloading, con-
sistent with the idea that eruptions are forced
by the mass distribution variations attendant
with deglaciations. — HJS
Geology 10.1130/G33419.1 (2012).
E C O N O MI C S
Marshmallows and Rösti(graben)
Does the language we speak influence how we
think? Chen adds to the lengthy and continu-
ing discussion of this question by linking
language to future-oriented behaviors, such
as a child who resists the temptation to eat
one marshmallow right now so as to be given
two marshmallows to enjoy just a few minutes
later. He does so by noting that the marking
of future tense is obligatory in some lan-
guages (French) and suggests that this
induces a cognitive representation
of the future as being distinct from
the present; in other words, tomor-
row is less a continuation of
today and more a new day
altogether. This leads to the
expectation that countries
in which so-called strong
future-time reference
languages predominate
would exhibit lower rates of
future-oriented behavior, such as saving
and exercise. Looking across countries in the
World Values Survey confirms this prediction,
even after controlling for various geographic,
cultural, and institutional factors. Furthermore,
looking within countries, such as Switzerland,
C H E MI S T R Y
Clusters Couple Chloroarenes
In the classic Ullmann coupling reaction
reported more than a century ago, iodine-
substituted aromatic rings were coupled at high
temperatures by using copper as a reducing
agent to form a carbon-carbon bond. Further
work extended this coupling reaction to more
conventional chloroarenes by using coreduc-
tants and precious-metal catalysts, but reaction
temperatures still tended to be high. Dhital et
al. now show that a wide variety of chloroarenes
can be coupled at ambient temperatures (25°
to 45°C). The reactions proceeded under basic
conditions (in a mixed organic-aqueous solvent)
in the presence of gold-palladium nanocluster
catalysts. Neither pure gold nor pure palladium
clusters could catalyze the reaction, indicat-
ing that alloying of the metals was critical for
reactivity. Density functional theory calculations
indicated that the critical difference for the
alloy clusters is that they favor activation of the
carbon-chlorine bond through the dissociative
adsorption of the arene chlorides. — PDS
J. Am. Chem. Soc. 10.1021/ja390606k (2012).
G E O L O G Y
Feeling the Pressure
Volcanism has a substantial impact on climate,
the global carbon cycle, the evolution of land-
forms, and a host of other important processes;
establishing how it has varied in the past is thus
MI C R O B I O L O G Y
Bacterial Breeze
Mount Bachelor in Oregon, USA, is at 2.8 km above sea level and is a useful high point
from which to sample trans-Pacific dust plumes in the upper troposphere and lower
stratosphere. Using this observatory, Smith et al. investigated what living matter gets
transported in the ~64-teragram annual aerosol load from Asia. Airborne bacterial
numbers and species were measured in two major plume events occurring in April and
May 2011, from which some Gram-positive organisms were recovered and cultured.
Atmospheric modeling revealed that the air masses lifted and swept through a storm
loop from locations near China, Korea, and Japan, and sequencing detected ~2800
bacterial species (operational taxonomic units) from a broad range of phyla. A few
marine archaeans were also identified, but what was notable was the preponderance
of spore-forming species capable of surviving extreme conditions. The work offers an
indication of the potential role microbes play in cloud nucleation and precipitation
in large-scale events, as well as their potential to be important air pollutants. — CA
Appl. Environ. Microbiol. 10.1128/AEM.03029-12 (2012).
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

EDITORS’CHOICE
C
R
E
D
I
T
:

M
A
J
U
M
D
E
R

E
T

A
L
.
,

C
H
E
M
.

M
A
T
E
R
.

2
4
,

4
6
4
7

(
2
0
1
2
)
of assembly and disassembly alternately allow
secretory granules to bind and then move
through the cortex en route to Ca
2+
-stimulated
fusion with the plasma membrane. By examin-
ing secretion from single cells, the oscillations
could be observed to increase the secretion
rate, as also predicted by a mathematical
model. This oscillatory dynamics may thus allow
cells to separate the two opposing roles of the
actin cortex (barrier and carrier) temporally and
thereby increase secretion efficiency. — SMH
Nat. Cell Biol. 14, 1261 (2012).
C H E MI S T R Y
A Washable MOF
Metal organic framework (MOF) materials are
highly porous materials that consist of metal
atoms linked together with organic ligands.
They are useful for capturing, storing, or filter-
ing gases, as well as for catalysis and sensing
applications.
Most MOF
materials are
unstable when
exposed to water,
thus preventing
their use as aqueous
sieves, where their
uniform porosity with
pores of just
the right
size would
otherwise be very
helpful. Majumder
et al. have developed
a family of water-stable
MOF materials based on
either magnesium ions or
certain first-row transition
metal ions—namely, nickel, cobalt, or manga-
nese—that are linked together with perylene
tetracarboxylate (PTC) ligands. The reaction be-
tween the potassium salt of PTC and the specific
metal acetate was performed in water and could
readily be done on milligram scales. The authors
noted that if either the acetate or potassium
salts were varied, only an amorphous material
was obtained. Ni-PTC was able to extract methyl
viologen, a known toxic herbicide, from parts
per million aqueous solutions. The material also
absorbed methylene blue, a dye with a similar
hydrodynamic radius, although two larger
molecules were clearly excluded from the MOF.
Ni-PTC also showed 300:1 selectivity of CO
2

versus N
2
, with a high binding enthalpy for the
CO
2
, suggesting possible uses for gas capture
applications. — MSL
Chem. Mater. 24, 4647 (2012).
that feature both strong and weak future-time
reference language speakers reveals that the
German-speaking Swiss save at more than
twice the rate of their fellows on the other side
of the linguistic divide. — GJC
Am. Econ. Rev., in press (2013);
http://cowles.econ.yale.edu/P/cd/d18a/d1820.pdf.
MI C R O B I O L O G Y
Exit Essentials
The bacteria Salmonella enterica is a major
cause of food poisoning. Salmonella invades
host cells by injecting these cells with virulence
factors by means of a molecular machine called
a type 3 secretion system, encoded by the
Salmonella pathogenicity island 2 (T3SS-SPI-2);
replicates, and then rapidly disperses systemi-
cally. It is important to check the dynamics and
mechanisms of spread in intact animals because
new vaccine candidates using strains mutated
in T3SS-SPI-2 are being developed for use
in humans. Grant et al. found that mutant S.
enterica, in contrast to wild type, replicated to
high numbers within individual spleen and
liver cells and formed only a few new foci
of infection. Further experiments showed
that the mutant bacteria were trapped
in the initial cell because they were
held in check by intracellular oxidase
activity that generates free radicals. Unex-
pectedly, the pathogen appeared to require
T3SS-SPI-2 to exit cells and spread through
an organ. This implies that a net bacterial
cell count alone will not tell you whether
a Salmonella infection has successfully
established and disseminated. — CA
PLoS Pathog. 8, e1003070 (2012).
C E L L B I O L O G Y
Waves of Separation
During exocytosis of secretory granules, the
actin cortex has two opposing roles: It can act
as a mechanical barrier impeding access to
the plasma membrane, yet it can also act as a
carrier that facilitates secretion. The mechanism
by which cells resolve this apparent paradox
has been unclear. Working with cultured mast
cells, Wollman and Meyer found that cells use
phase-shifted oscillations of Ca
2+
and F-actin
assembly to create a cyclic secretory engine
that enhances the rate of exocytosis of secretory
granules. The cellwide cortical actin oscillations
were initiated by Ca
2+
and phosphatidylinositol
4,5-bisphosphate oscillations that promoted
oscillations of N-WASP recruitment (an actin
nucleator), thus triggering waves of cortical
F-actin assembly and disassembly. These waves
Continued from page 11
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org
17050 Montebello Rd, Cupertino, CA 95014
Email: AAASInfo@betchartexpeditions.com
www.betchartexpeditions.com
For a detailed brochure,
please call (800) 252-4910
All prices are per person twin share + air
Edging stunning 18,000 ft. Himalayan
peaks, you will meet farmers, herb
doctors, shopkeepers, and monks.
Visit ancient towns of Weishan, Dali,
and Lijiang. Discover the Naxi, who
developed a distinctive cultural tradi-
tion in stunning valleys. $3,995 + air.
BETCHART EXPEDITIONS Inc.
Discover the RichCultures
&Heritage of Yunnan!
BACKROADS CHINA
March 21–April 3, 2013
AAAS is here –
preparing minority students
for careers in science.
Part of AAAS’s mission is to diversify and
strengthen the scientific work force. To help
achieve this goal, AAAS partners with NSF
to present the Historically Black Colleges
and Universities Undergraduate Program,
a conference where students from HBCUs
get experience presenting their research,
networking with peers, meeting with
representatives from graduate schools,
and learning about career opportunities.
If you’re not yet a AAAS member, join us.
Together we can make a difference.
To learn more, visit
aaas.org/plusyou/hbcuup
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 14
NEWS OF THE WEEK
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

R
O
Y

K
A
L
T
S
C
H
M
I
D
T
/
L
B
N
L
;

J
A
C
Q
U
E
L
I
N
E

M
C
B
R
I
D
E
/
L
L
N
L
;

N
A
S
A
/
B
I
L
L

I
N
G
A
L
L
S
;

N
A
S
A
/
J
P
L
-
C
A
L
T
E
C
H
/
M
S
S
S
;

2
0
1
1

C
E
R
N
,

F
O
R

T
H
E

B
E
N
E
F
I
T

O
F

T
H
E

C
M
S

C
O
L
L
A
B
O
R
A
T
I
O
N
;

A
L
M
A

(
E
S
O
/
N
A
O
J
/
N
R
A
O
)
,

C
A
R
L
O
S

P
A
D
I
L
L
A

(
N
R
A
O
/
A
U
I
/
N
S
F
)
;

E
.
U
.
Geneva, Switzerland 3
Is It the Higgs?
This spring, physicists working with the
world’s largest atom smasher, the Large
Hadron Collider in Switzerland, have a shot
at determining whether the particle they dis-
covered last summer is something other than
the long-sought Higgs boson. They may have
measurements of the particle’s spin and
parity, which for the Higgs should have the
values 0 and +1, respectively. Hungry for
new physics, many physicists would be hap-
pier if the measurements defy expectations.
Karachi, Pakistan 4
Setback for Polio Eradication?
Following targeted killings of nine polio
workers in Pakistan in late December, the
country’s polio program is in turmoil, and
door-to-door vaccination campaigns have
been suspended. The number of cases in
Pakistan, one of just three remaining polio-
endemic countries, seems certain to rise.
But how severe the setback will be for other
nations is still unclear—after Nigeria’s 2003
ban on polio vaccinations, the virus spread to
some 20 other countries across the globe.
Washington, D.C. 5
Will Obama’s Team Re-Up?
Is there still fire in
my belly?
That’s what Energy
Secretary Steven Chu,
presidential science
adviser John Holdren,
and NASA Administra-
tor Charles Bolden may
be asking themselves
as their boss, President
Barack Obama, prepares
for his second inaugu-
ration this month. The
three science adminis-
trators were forced to
grow thick skins dur-
ing the president’s first
term, defending several
controversial policies
of the administration.
Last month, two other
members of Obama’s
self-described “dream
team” announced their departures: National
Oceanic and Atmospheric Administration
Administrator Jane Lubchenco and Environ-
mental Protection Agency chief Lisa Jackson.
Atacama Desert, Chile 6
ALMA Fully Online
In March, the world’s largest astronomical
observatory, the Atacama Large Millimeter/
submillimeter Array (ALMA), will be
declared officially complete. A collabora-
tion between Europe, Japan, and the United
States, the $1 billion ALMA is the first
global observatory. Its 66 dish antennas will
focus on a little-studied part of the spectrum
between microwaves and the infrared—a
universe of cold, dim objects such as stars
and planetary systems coalescing from gas
clouds and the formation of the universe’s
earliest galaxies.
Washington, D.C. 1
Curiosity’s Big Year
Curiosity may have wowed the world with
its perfect landing on Mars, but in 2013 the
rover is going to have to deliver the scien-
tific goods. The
$2.5 billion mis-
sion should return
details of the story
of dramatic global
change on ancient
Mars, but the pres-
sure is on for more
headline-grabbing
news: Will the
rover find organic
matter lingering from ancient life? And is
something—martian microbes, perhaps—
even now churning out methane?
Brussels 2
E.U.’s Big Budget Decisions
E.U. leaders will try again in 2013 to agree
on budget numbers for the next funding
period, which runs from 2014 through 2020.
With pressure from some
countries to cut the budget
and demands from others
to keep agriculture and
other subsidies, science
spending was caught in the
middle—and got squeezed
in recent proposals. Failure to
hammer out a deal will delay, for example,
the proposed €80 billion Horizon 2020
research funding program.
3, 11
7
2
4
1, 5, 14
6
Chu
Wit
c
m
i
A LOOK AHEAD FOR 2013
Holdren
Bolden
AROUND THE WORLD
10
8
9
13
12
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 15
NEWS
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

N
I
H
/
W
I
K
I
M
E
D
I
A

C
O
M
M
O
N
S
;

I
S
R
O
;

J
E
A
N
-
M
A
R
C

A
Y
R
A
U
L
T
/
W
I
K
I
M
E
D
I
A

C
O
M
M
O
N
S
;

E
S
A
;

C
L
I
M
A
T
E

C
H
A
N
G
E

2
0
0
7
:

T
H
E

P
H
Y
S
I
C
A
L

S
C
I
E
N
C
E

B
A
S
I
S
,

F
I
G

1
0
.
8
,

C
A
M
B
R
I
D
G
E

U
N
I
V
E
R
S
I
T
Y

P
R
E
S
S
/
I
P
C
C
;

P
H
O
T
O
S
.
C
O
M
Iperó, Brazil 7
Nuclear Reactor Ground Breaking
Construction begins this year on the
$450 million Brazilian Multipurpose Reac-
tor. Brazil’s four research reactors are all
more than 30 years old; the new reactor,
built by the Ministry of Science, Technol-
ogy, and Innovation, is slated for comple-
tion by 2017, and will greatly boost the
production of radioisotopes for treating
diseases. Scientific goals include extend-
ing the life span and increasing the safety of
nuclear power plants, as well as conducting
a variety of types of material research.
Kourou, French Guiana 8
Milky Way Monitor Lifts Off
An unprecedented stellar census will begin
later this year when the European Space
Agency launches Gaia, a mission to mea-
sure the exact position, velocity, and other
characteristics of 1 billion of the Milky
Way’s stars (1% of the total). Gaia will
illuminate the origins, structure, and evolu-
tion of our galaxy—and, hopefully, it will
discover huge numbers of minor asteroids
around our sun and tens of thousands of
extrasolar planetary systems.
Beijing 9
Stem Cell Guidelines
China’s Ministry of Health is expected to
issue long-awaited guidelines governing
stem cell research in 2013. Untested stem
cell therapies have proliferated at some
Chinese hospitals, prompting government
crackdowns that hamper not only profiteer-
ing charlatans but also scientists planning
clinical trials. Last January, the ministry
halted approvals for clinical research involv-
ing stem cells. Researchers hope the guide-
lines will provide clarity in a country once
seen as a hot spot for stem cell research.
Paris 10
French Research, Post-Sarkozy
After a 5-month nation-
wide consultation, the
government of François
Hollande, France’s first
socialist president in 17
years, will chart a new
course for research and
higher education and
possibly dial back some
innovations of the last
decade. Hollande’s polit-
ical base has urged him to amend a program
to create top research clusters, as well as the
National Research Agency and the much-
criticized evaluation system. Improving stu-
dents’ success rates will be a key target.
Geneva, Switzerland 11
Gauging the Global Greenhouse
The Nobel-winning Intergovernmental
Panel on Climate Change will deliver its
next assessment of the physical science of
climate change in September, but block-
buster developments that could bolster
greenhouse gas
mitigation are
unlikely, and pro-
nouncements on
weather extremes
and sea-level
rise won’t be
barn burners.
Past reports
have already
answered the big
questions: The
world is warm-
ing, humans are behind most of that, and
climate is sensitive enough to greenhouse
gases that 2100 looks grim.
Kumamoto Prefecture, Japan 12
U.N. Mercury Treaty Nears Finish
In October, representatives of 120 coun-
tries are expected to gather in Kumamoto
Prefecture, a region known for an infamous
example of industrial mercury contami-
nation, to finalize a U.N. treaty aimed at
reducing the industrial use and trade of the
potent neurotoxin. The mercury control
effort has been under way since 2009.
Bangalore, India 13
Mars-Bound Mangalyaan
India’s mission to Mars, an
orbiting satellite dubbed
Mangalyaan,
is slated to
launch in
November. The
$100 million satel-
lite will carry a hand-
ful of instruments, including
cameras and spectrometers, that are
intended to scan for signs of a biosphere
on the Red Planet—but the Indian Space
Research Organisation notes that the tiny
satellite is “more of a technological mis-
sion than a science mission.” It may also
kick off a Mars-bound space race between
India and China.
Washington, D.C. 14
Permanent Cliff Dwellers
The Obama administration has promised to
protect research from deep cuts in negotia-
tions with Congress on reducing the federal
deficit. As Science went to press, it looked
like the U.S. scientific community might
temporarily avoid the fiscal cliff. But threats
of budgetary catastrophe promise to be a
recurring theme in 2013. The next test may
come in March, when a 6-month freeze on
agency budgets expires.
Hollande
A human embryonic stem cell.
Predicted 4° to 5°C
warming (red) by 2100.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 16
NEWS & ANALYSIS
C
R
E
D
I
T
:

D
.

M
A
L
A
K
O
F
F
/
S
C
I
E
N
C
E
Researchers who study the H5N1 avian influ-
enza virus will soon be able to do some sci-
ence that’s been off-limits for nearly a year.
U.S. government officials say they expect
to put the finishing touches this month on
new rules designed to help funding agencies
identify and regulate especially problematic
H5N1 studies before they begin (Science,
7 December 2012, p. 1271). Leading influ-
enza scientists say that move will enable
them to lift a year-old, self-imposed mora-
torium on certain kinds of potentially dan-
gerous experiments (Science, 21 December
2012, p. 1533).
The moratorium “has now achieved its
purpose” by giving governments time to
act, says one of its 39 signers, virologist
Ron Fouchier of Erasmus MC in Rotter-
dam, the Netherlands. “It is time to get back
to work.”
The two developments would essentially
end a long and bruising controversy over
the risks and benefits of H5N1 research.
The controversy began in late 2011, after
two research teams—including one led by
Fouchier—showed how to alter the virus,
which normally infects birds, so that it can
move between mammals. Fearing that such
“gain-of-function” experiments could enable
terrorists or a lab accident to start a deadly
human pandemic, critics demanded stricter
controls on science that could be used for
good and evil. The issue has been especially
sensitive for the U.S. government, because its
National Institutes of Health (NIH) funded
the two studies and is one of the world’s big-
gest funders of H5N1 research.
The controversy has had an impact far
beyond influenza laboratories, however. In
March, it prompted U.S. officials to impose
new rules that require systematic reviews of
federally funded studies involving 15 “high-
consequence” pathogens and toxins, includ-
ing H5N1, that could be used as bioterror
weapons. In November, NIH issued draft
guidelines designed to help the agency decide
which types of H5N1 gain-of-function stud-
ies would need more stringent reviews.
Other nations, meanwhile, also moved to
tighten laboratory safety requirements and
reviews of such dual-use research. The con-
flict even threatened to blow apart a frag-
ile global agreement on sharing samples of
influenza viruses. The experience “polar-
ized the science community in a way that
is fairly rare,” says Amy Patterson, an NIH
administrator within the director’s office
who has helped lead efforts to design the
new rules.
Anthony Fauci, director of the National
Institute of Allergy and Infectious Dis-
eases (NIAID), defends the need for what
he admits was an “arduous, highly charged”
process that included nearly a dozen major
public meetings around the world. The sci-
entific community “was fundamentally
closed” to hearing the fears and concerns
of the public and policymakers, says Fauci,
who played a major backstage role in devel-
oping the moratorium and reassuring poli-
ticians that science administrators could
agree upon a safe path forward. “We’re
starting to get to a place where many of the
objections have dissipated.”
Some outside observers, however, are
skeptical that the process has informed the
public or made the world safer. “The mor-
atorium was a communication strategy,”
says Peter Sandman, a risk communication
expert in Princeton, New Jersey, who has
worked extensively on pandemic prepared-
ness. “It was never a goal of the moratorium
to educate people that there might be a sig-
nificant risk of releasing a bioengineered
H5N1 virus with pandemic potential from
a laboratory. … To the contrary, the morato-
rium aimed to buy time to persuade people
that that risk was negligible.”
Those risks were a major topic of con-
versation at a meeting last month at NIH
that attracted some 200 scientists, policy-
makers, and nonprof it advocates from
around the world. One agenda item was
whether scientists can actually get informa-
tion from H5N1 gain-of-function studies
that would be useful for heading off a pan-
demic. Although many speakers said yes,
researchers from places where H5N1 is a
simmering problem, such as Indonesia and
Vietnam, said they put a higher priority on
characterizing naturally emerging viruses
than creating new ones.
Thomas Inglesby, head of the Center for
Biosecurity of the University of Pittsburgh
Medical Center, argued that scientists had
overstated the benefits. “We should con-
tinue the moratorium,” he said. And “if we
do decide to proceed, all should acknowl-
edge the extraordinary risk.”
Meeting participants were also asked
to suggest improvements to the new NIH
funding guidelines. They spell out seven
criteria that a gain-of-function study would
have to meet to be eligible for NIH support.
A study that failed to meet even one of the
criteria would get a more intensive review
by NIH’s parent, the U.S. Department of
Health and Human Services (HHS). If pub-
lishing the results might pose a danger to
public safety, the draft guidelines suggest
that the study be transferred to an agency
that does classified research, such as the
Department of Defense.
That idea was widely panned at the meet-
ing, however, with Fauci and other speak-
ers arguing that secret work on H5N1 could
H5N1 Researchers Ready
As Moratorium Nears End
AVI AN I NFLUENZA
Flu views. Scientists debate whether to resume con-
troversial studies of the H5N1 avian influenza virus.
NEWS & ANALYSIS
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 17
C
R
E
D
I
T
S
:

D
.

M
A
L
A
K
O
F
F
/
S
C
I
E
N
C
E

(
2
)
undermine international collaboration and
fuel suspicions that the United States was
conducting research on bioweapons. “If
we fund something, it should be with the
assumption that it will be published,” Fauci
said. “NIH doesn’t do classified research.”
Attendees also pointed to criteria they
felt were worded poorly. “As written, a lot
of H5N1 studies that were not intended to
come to review will get reviewed,” predicted
Nancy Cox, a virologist and head of influ-
enza research at the Centers for Disease
Control and Prevention in Atlanta. Richard
Webby, an H5N1 researcher at St. Jude Chil-
dren’s Research Hospital in Memphis, Ten-
nessee, observed that, “if you read this con-
servatively, [HHS is] going to have to review
75% of H5N1 studies in the NIH portfolio.”
One suggestion that gained traction was
the need to rewrite a draft criterion that
asked scientists to provide “evidence” that
the H5N1 virus they wanted to create “could
be produced through a natural evolution-
ary process in the foreseeable future.” Such
crystal ball-gazing is very difficult, several
speakers argued.
Instead, they proposed that NIH limit the
extra, HHS-level reviews to a subset of stud-
ies that would create especially dangerous
H5N1 viruses. They had in mind viruses that
could infect mammals and be spread through
the air, by droplets of saliva for instance. A
“growing chorus” of researchers endorsed
that idea, noted Patterson, who will revise the
draft guidelines after the 10 January deadline
for public comments has passed.
Fauci made it clear that he wants to move
quickly to finalize the rules. “I’m sensitive
to the fact that this can’t be drawn out over a
long time,” he said.
The new policy isn’t expected to affect
many studies initially. Gain-of-function
research accounts for less than $10 million—
or fewer than 1%—of NIAID’s overall influ-
enza research spending, Fauci said, and
fewer than 10% of its H5N1 grants.
H5N1 researchers who don’t rely on NIH
funding could lift the moratorium before the
United States has finalized its policy, both
Fauci and Fouchier said. Those scientists have
been “patiently waiting,” for NIH to complete
its work, Fauci said, and some are now “going
to go ahead with their experiments if their
country and funder allow it,” he predicted.
That development should help the H5N1
influenza community refocus on important
issues such as understanding how poten-
tial pandemic viruses work and developing
better vaccines and surveillance methods,
Fouchier said. “It is sobering to see how
much time has been spent on addressing the
challenges [posed by H5N1 research] and
not on the scientific opportunities,” he said.
–DAVID MALAKOFF
HONG KONG—“One-night stand?” asks
an ad on the homepage of Gztz.org. Offer-
ing one of China’s largest online platforms
for gay and bisexual men, the organization
boasts 200,000 registered members—30,000
of them in and around Guangzhou, where
it operates a community center. On its clut-
tered home page, users can choose to chat
with other members, browse relevant news,
or peruse a list of bathhouses and sex stores
that serve gay men. But it’s the provocatively
titled ad—which actually urges men to visit
the community center for free, confidential
HIV screenings—that occupies the most
prominent spot.
That’s because one of Gztz.org’s key
roles is to not to match up gay men, but to
protect their health. It is among a handful
of community-based organizations (CBOs)
across China that offer HIV testing, usually
in collaboration with local officials. Govern-
ment health workers draw blood and do lab
testing, while Gztz.org volunteers provide
counseling to those with positive tests and
advise on treatment options.
Since 2008, Gztz.org has performed
about 11,000 HIV screening tests. Roughly
500 of them came back with positive pre-
liminary results. To receive confirmatory
testing, a member must register his national
identification card number. That is where
many men tend to fall through the cracks.
But Gztz.org co-founder Roger Meng says
that his volunteers have convinced more
than 90% of men with preliminary positives
to go in for further testing.
This is a major achievement for Meng’s
group. Their successful collaboration with
government health workers also hints at an
improvement in the Chinese government’s
rocky relationship with men who have sex
with men (MSM), long subject to discrimi-
nation in the workplace and persecution by
police. The country’s current HIV/AIDS
strategy includes reaching out to CBOs (the
preferred term in China for such unofficial
organizations). In November, Li Keqiang,
who is expected to take over from Wen
Jiabao as premier in 2013, underscored that
point by meeting with representatives of sev-
eral such organizations in Beijing, includ-
ing those that work with MSM. Health offi-
cials, meanwhile, pledged to make it easier
for civil society groups doing HIV preven-
tion to operate in China. “We perceived that
we needed to have a more supportive policy,
with more funding and a more supportive
NEWS
China Partners With Gay Groups on HIV Screening
PUBLI C HEALTH
Painful process. Amy Patterson (left) has helped lead the NIH effort to devise new H5N1 research rules, with
input from scientists and public health experts including Ilaria Capua from Italy, Joseph Sriyal Malik Peiris from
Hong Kong, and Adel Mahmoud from the United States (right to left in right photo).
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 18
NEWS&ANALYSIS
C
R
E
D
I
T
:

©

J
I
A
N
A
N

Y
U
/
R
E
U
T
E
R
S
/
C
O
R
B
I
S
environment, to encourage NGOs [nongov-
ernmental organizations] to participate in
HIV and AIDS activities,” says Wu Zunyou,
director of the National Center for AIDS/
STD Control and Prevention of the Chinese
Center for Disease Control and Prevention
(China CDC).
The results delivered by groups such as
Gztz.org so far suggest that is a wise strat-
egy. A Joint United Nations Programme
on HIV/AIDS–funded review of Gztz.org’s
activities and 15 similar HIV testing mod-
els in China, released by the coalition China
Male Tongzhi Health Forum in November,
reported that of 808 men who screened posi-
tive over a period of 6 to 18 months, 89%
returned for confirmatory testing. A further
95% of those who were confirmed as HIV-
positive agreed to take a CD4 test, which
measures the person’s immune health and
can guide whether immediate treatment is
needed. Gay and bisexual men know, Meng
says, that at organizations like his they will
be spared the accusing looks and moraliz-
ing that they sometimes encounter at other
clinics. “People feel much more comfortable
than with the traditional model, where the
test taker is a nurse who has nothing to do
with the gay community.”
China has an estimated 780,000 cases of
HIV/AIDS, and according to Wu, MSM are
the risk group with the most rapidly increas-
ing infection rate. MSM accounted for
roughly 50% of new cases reported in pro-
vincial capital cities in 2012 so far, he says.
Reaching these men has long been a goal for
the government, which first allocated funds
for collaborating on HIV prevention with
CBOs in 2002. The Global Fund to Fight
AIDS, Tuberculosis and Malaria began sep-
arately providing generous grants for work-
ing with CBOs soon after. In 2011, how-
ever, the fund halted payments to China out
of concern that the money was not reaching
its intended targets (Science, 10 June 2011,
p. 1253). Although funding was reinstated
several months later, the rebuff stung; Chi-
nese health officials pledged to clean up the
spending process. A renewed commitment
to working with groups like Meng’s came
soon after.
But although more Chinese MSM are
seeking testing, only an estimated 47% have
ever been tested. A survey of 500 MSM in
Beijing, published in Sexually Transmit-
ted Infections in December 2011, found
that an alarming 86% of those with HIV
did not know they were infected. In focus
groups conducted with 49 MSM in Nan-
jing last February, one of the two most cited
reasons men gave for not getting tested was
the fear of being outed or discriminated
against by doctors. (Chongyi Wei, a public
health researcher at the University of Pitts-
burgh’s Graduate School of Public Health
in Pennsylvania, and colleagues conducted
the interviews but haven’t yet published
their results.) Confidentiality is a very deli-
cate issue: Between 17% and 33% of MSM
in China are married. And many men are
unaware that HIV testing is free and readily
available, Wei says: “We still haven’t gotten
out the message completely.”
Mistrust of government health workers
complicates treatment as well, as too often
those initially diagnosed with HIV do not
return for care when they need it. “If we
want to have a long-term strategy, we need
to retain people,” says Frank Y. Wong, a
global health scholar at Emory University in
Atlanta. “If you cannot retain them, then for-
get about treatment.”
The government is betting that CBOs are
the missing link. “We have tried to provide
outreach services to deliver health educa-
tion, condoms, and counseling and encour-
age people to come to STD clinics for test-
ing, but there is still a big gap,” says Chen
Xiangsheng, vice director of the National
Center for Sexually Transmitted Diseases
Control in Nanjing. “CBOs may play some
role to help facilitate the process.” A study
published online in AIDS Care on 15 Octo-
ber 2012 suggests that such groups may
facilitate treatment, too. Eric J. Nehl of
Emory University and colleagues found
that in a survey of 605 MSM in Chengdu
conducted by a local group, 63% of HIV-
positive men were willing to be contacted
by the organization for follow-up after
3 months.
Today, the Global Fund’s money for HIV
prevention is drying up, and China is on a
list of countries that starting in 2013 will no
longer be issued grants. Even so, the Chi-
nese government is committed to funding
CBOs, Wu says, and is now developing offi-
cial guidelines for working with them. Ear-
lier this year, he joined a health ministry del-
egation to the United States, Australia, and
Japan to see how NGOs figure in those coun-
tries’ HIV/AIDS prevention strategies.
In China, two key challenges are to
ensure that NGOs such as Gztz.org develop
sophisticated outreach plans and properly
administer funds. “NGOs in China are still
quite young,” Chen says. Success of the
civil society groups ultimately depends on
the cooperation of the provincial centers
for disease control. In some cities, local
CDC workers remain resistant to involv-
ing the groups in HIV testing, according
to the China Male Tongzhi Health Forum
review. The ability of groups—and govern-
ment officials—to make inroads with MSM
is further limited by homosexuality’s persis-
tent stigma. At the Hong Kong workshop,
some praised a groundbreaking social mar-
keting campaign by the U.S. Centers for
Disease Control and Prevention targeting
African-American gay and bisexual men.
Called “Testing Makes Us Stronger,” the
campaign was developed in consultation
with community leaders and features prom-
inently placed billboards, along with televi-
sion and radio ads, in six cities. Such a far-
reaching effort is not yet possible in China,
Wei says, because “it may be perceived as
promoting homosexuality.” And yet, efforts
that are highly targeted will likely miss men
who don’t socialize within the gay commu-
nity. Meng estimates the greater Guang-
zhou area is home to some 200,000 MSM—
meaning his group only reaches 15% of its
target demographic. Wei says: “That’s the
biggest barrier in a Chinese context.”
–MARA HVISTENDAHL
Outreach, old style. A wall slogan
in Fuyang warns of HIV risks in 2006;
health workers now use the Web to
reach people who are at risk.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 19
NEWS&ANALYSIS
The Field Museum of Natural History in Chi-
cago, Illinois, is planning significant layoffs
to its world-class faculty in a cost-cutting
move that threatens to erode its international
reputation. The news has shaken the Field
Museum’s scientists, and left researchers at
other museums puzzled and dismayed.
“It feels like we’re watching the Field
Museum publicly commit suicide,” says
Michael Donoghue, a botanist at Yale Univer-
sity and the former director of the Yale Pea-
body Museum of Natural History. “They’ve
basically announced that they’re going to
drain the scientific lifeblood out of the place.
And that means they’re no longer going to be
a player on the stage of the great museums of
the world. It’s a really dumb move. There has
to be an alternative.”
Last month, the Field Museum announced
that it needs to shave $5 million from its over-
all annual budget of about $70 million to
stave off a fiscal crisis. Museum President
Richard Lariviere says that the Field Muse-
um’s science departments will likely lose
$3 million from a $10-million-a-year budget.
That account supports 27 curators and 116
other staff scientists. Some will necessarily
lose their positions, Lariviere says, although
“how many has yet to be determined.”
Lariviere, a classical Indian scholar who
became president in October, says he reached
the decision reluctantly and that there are no
better options. “The fact that there isn’t more
money for science isn’t a point of satisfaction
for anyone,” he says. “But to deny the finan-
cial reality, and to continue to spend more
money than comes in, is to doom this place.
Our goal is to grow the museum, but to do
that we first need to balance the budget.”
Created in 1893, the Field Museum is a
leading research center in the natural and
anthropological sciences. It houses some
25 million specimens of fossils, plants, and
animals, as well as millions of cultural items
from around the world. Many of the speci-
mens were collected by the museum’s cura-
tors, who also maintain labs and obtain out-
side funding to do research on the collec-
tions. Last year, for example, Field Museum
scientists discovered more than 200 plants
and animals and published numerous articles
in peer-reviewed journals.
“Our scientific colleagues acknowledge
that we’re in the top ranks of world research-
ers,” says Peter Makovicky, a geologist and
chair of the museum’s geology department.
“So when the administration makes an
announcement to the world that it’s willing to
sacrifice that, it is worrying.”
At Lariviere’s request, the museum has
formed a special committee to review the
museum’s finances and decide if a “financial
exigency” exists. Such a declaration would
allow the museum to lay off its tenured sci-
entists (most of the curators have tenure)
along with other professional staff members.
But because Lariviere told local media that
curatorial positions would be cut, museum
staff members believe “that the administra-
tion already has a foregone conclusion about
the committee’s decision,” says Makovicky, a
committee member. “But we haven’t met yet
to discuss the exigency issue.”
Lariviere tells Science that “it is up to the
committee to answer the question” about
whether the museum is in financial exigency.
“I hope that they find it is not the case,” he
says. “But I don’t think that they will.”
The current fiscal crisis has several ele-
ments. Its $300 million endowment, which
provides roughly one-fifth of its operating
budget, was hammered in the 2008 reces-
sion. Although it has bounced back, it has not
grown as fast as had been anticipated before
the crash. Attendance has remained flat for
decades, and revenues from the city of Chi-
cago and from private sources are down. In
addition, the administration has revised the
museum’s accounting practices, so that the
$12 million a year spent on maintaining the
museum’s 90-year-old building is now part
of the operating budget.
Another major component is the
$7-million-a-year cost of servicing debt for
projects such as the $90 million Collections
Resource Center, a state-of-the-art facility for
storing and studying specimens that opened
in 2005. “It’s a very impressive [facility],”
says James Hanken, a herpetologist and the
director of Harvard University’s Museum of
Comparative Zoology in Cambridge, Massa-
chusetts. “But they made a financial gamble
when they built it, and they lost.”
Over the years, these factors have contrib-
uted to a growing imbalance between reve-
nues and expenses. In 2008, museum offi-
cials imposed several cost-cutting measures,
including a hiring freeze; slashing its house-
keeping, security, operations, and exhibi-
tions staff; and trimming exhibition budgets.
The museum has also reduced its scientific
staff through attrition. In 2004, the museum
employed 40 curators compared with the
current 27; among those losses are the two
herpetologists once employed to maintain its
internationally acclaimed collection of rep-
tile and amphibian specimens.
Lariviere says that a smaller staff is still
capable of doing world-class research. “It’s
absurd to think that we can’t do science
because we don’t have a herpetologist. The
very notion that these cuts mean the end
of science at the museum … doesn’t make
sense.” He says the collections will remain
available to outside researchers. “The
enhancement of the collection lies in the sci-
entific studies done on them, but the Field
Museum staff scientists aren’t the only ones
who do this,” he adds.
Outside scientists say that layoffs could
do permanent damage to the museum’s rep-
utation as a first-rate scientific institution.
“An inevitable decline will follow from those
cuts,” says Neil Shubin, a paleontologist at
the University of Chicago. “They’ll lose
what is unique about the Field: Its scientific
authority, which comes from the eminence
of their scientists and collections. It will take
years to replace, if it can ever be done.”
–VIRGINIA MORELL
Ruffled feathers. Scientists say proposed layoffs at
the Field Museum are “a really dumb move.”
C
R
E
D
I
T
:

J
O
H
N

W
E
I
N
S
T
E
I
N
/
T
H
E

F
I
E
L
D

M
U
S
E
U
M
Budget Crunch to Shrink Science
Programs at Chicago’s Field Museum
PUBLI C COLLECTI ONS
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 20
C
R
E
D
I
T
:

C
O
U
R
T
E
S
Y

O
F

G
U
A
N
G

Z
H
U
,

Z
O
N
G
-
H
O
N
G

L
I
N
,

A
N
D

Z
H
O
N
G

L
.

W
A
N
G
/
G
E
O
R
G
I
A

T
E
C
H
When you’re on the move and unable to plug in your portable
devices to recharge them, you could use a tiny generator to harvest
some of the energy of movement and turn it into electric charge.
Engineers have created such devices, but so far they haven’t pro-
duced enough power to drive most small portable electronics. At
the meeting, however, researchers led by Zhong Lin Wang at the
Georgia Institute of Technology (Georgia Tech) in Atlanta reported
a new type of generator that uses everyday static electricity to juice
things up considerably.
Since 2005, Wang and his colleagues have been working on tiny
generators, or nanogenerators, most of which rely on the piezoelec-
tric effect, which converts mechanical strain into an electric voltage
across certain materials. Connect the material to a circuit and the
voltage will drive a current. Despite improvements by Wang’s team
and others, such piezoelectric devices have generated mere micro-
watts of power. Until, that is, they wrapped one in plastic.
While investigating one piezo generator, Feng-Ru Fan, one of
Wang’s graduate students, put a layer of plastic known as PMMA on
top of it for protection. The generator in turn was sitting on a differ-
ent plastic called Kapton. When Fan measured the device’s perfor-
mance, he consistently found it was generating a higher voltage than
expected. Fan and Wang spent months investigating what was hap-
pening, before concluding that the added power was coming from
static electricity. When the PMMA and Kapton come into contact
and rub against one another, friction generates electrical charges.
The two plastics have a different affinity for electrons, with the Kap-
ton eager to grab electrons and the PMMA able to give them up.
Wang and Fan realized that if these two plastics were incorpo-
MEETINGBRIEFS>>
MATERIALS RESEARCH SOCIETY FALL MEETING AND EXHIBIT | NOVEMBER 25–30 | BOSTON
The Power of Friction
They weren’t ready in time for this holiday
season’s gadgets, but better lithium-ion bat-
teries could be in the offing. Today’s lithium-
ion cells are already slightly better than those
produced a few years ago. But new innova-
tions unveiled at the meeting could see five-
fold improvements in battery performance.
Like all batteries, today’s lithium-ion
rechargeables work by shuttling electri-
cal charges back and forth between two
electrodes—a positively charged cathode
and a negatively charged anode. When the
battery is fully charged, positively charged
lithium ions are nestled in a matrix of nega-
tively charged graphitic carbon at the anode.
When the switch on a toy or tool is turned
on, electrons are pulled out of the graphite
and sent through an external circuit to per-
form work before being injected back into
the material in the cathode, typically an
alloy such as lithium cobalt oxide (LiCoO
2
).
Shifting electrons from the anode to the
cathode causes the lithium ions to migrate
through an ion-conducting electrolyte to
the positively charged electrode. When the
rechargeable is plugged into a socket, the
applied voltage drives electrons back out of
the cathode into the anode; the lithium ions
then detach from the cathode and migrate
back to the anode to team up with the elec-
trons in the graphite again.
One problem is that LiCoO
2
cathodes
can’t hold on to very many lithium ions,
which keeps the battery’s overall electrical
storage capacity low. Researchers have long
wanted to replace the LiCoO
2
with sulfur,
each atom of which can grab nearly 10 times
the number of lithium ions. When other
considerations are taken into account, this
should give lithium-sulfur batteries about
five times the capacity of current lithium-
ion cells.
But sulfur has its problems. The first task
of any electrode is to be a good conductor,
allowing electrons to shuttle in and out eas-
ily. Sulfur is a mediocre conductor. Also,
because so many lithium ions can bind to
the sulfur, this causes the cathode material to
swell and shrink repeatedly during charging
and discharging. Ultimately, this causes it to
crack and break apart. Unwanted side reac-
tions involving lithium and sulfur can also
create a family of byproducts called polysul-
fides that can poison lithium batteries.
At the meeting, Yi Cui, a materials sci-
entist at Stanford University in Palo Alto,
California, reported a possible way around
sulfur’s problems. Cui and his team encap-
sulated tiny nanoparticles of sulfur inside
a shell of titanium dioxide (TiO
2
), leav-
ing extra space inside each shell. They then
packed their coated nanoparticles together
to form a cathode. When they ran their bat-
tery, they found that TiO
2
’s high conductiv-
ity made it easy to shuttle electrons in and
out. During discharge, the lithium ions read-
ily penetrated the TiO
2
shells and bound to
sulfur atoms in the nanoparticles. And even
though the sulfide nanoparticles repeatedly
swelled and shrank inside their shells as
the batteries were charged and discharged,
Light tap. A simple repeated tap
of the foot on a nanogenertor is
enough to power 600 LED bulbs.
A Boost for Lithium-Sulfur Batteries
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 21
NEWS&ANALYSIS
grow other semiconductor alloys. These
top layers are grown to perfectly match the
atomic arrangement of atoms in the underly-
ing GaAs, which gives them good electronic
properties. Manufacturers then use a pro-
cess known as epitaxial liftoff to remove the
topping layers and recover the GaAs wafer
so that it can be reused. Unfortunately, the
epitaxial liftoff usually causes some minor
damage to the GaAs wafer, so it can only
be used a few times before engineers must
replace it with a fresh wafer. As a result,
costs for GaAs solar cells remain high.
Three years ago, Stephen Forrest, a mate-
rials scientist at the University of Michigan,
Ann Arbor, came up with a way to solve
this problem for a related semiconductor
alloy called indium phosphide (InP). His
team’s strategy was to add two additional
very chemically different layers to the sand-
wichlike stack of materials. The added lay-
ers allowed the materials grown above to
continue to match the exact lattice
spacing of InP, but because they
were chemically distinct, they could
be etched away selectively without
damaging the underlying InP wafer
at all, enabling it to be reused again
and again.
Still, Forrest says that interest in
the approach was muted because InP
is not as good of a solar cell mate-
rial as GaAs. At the meeting, Forrest
presented his group’s latest achieve-
ments in extending their epitaxial
liftoff approach to GaAs. As before,
the two extra chemically distinct
sacrificial layers allowed Forrest’s
student Kyusang Lee and electri-
cal engineer Jeramy Zimmerman
to grow thin layers of high-quality
GaAs on top of a GaAs wafer and
then remove the top layer while not
damaging the underlying wafer at all. More-
over, additional steps also allowed them to
bond the final GaAs layer to a clear plastic
substrate, giving them a flexible solar cell
with more than 22% efficiency.
If cheap solar concentrators are added
to focus more light onto the cell, Forrest
says he believes that they should be able
to convert more than 30% of the energy in
sunlight into electricity. If that’s the case,
Forrest says his calculations show that they
can reduce the cost of power from the cells
to less than $1 per watt, roughly the current
cost for silicon-based solar cells. If they can
muster further improvements, that price
could drop close to grid parity—the holy
grail for solar power.
–ROBERT F. SERVICE
The joke among electrical engineers is that
gallium arsenide (GaAs) is the semiconduc-
tor of the future and always will be. Nowhere
is that more true than with commercial solar
cells that use semiconductors to convert sun-
light to electricity—a market where profit
margins are razor thin. GaAs has been used
for decades to make ultrahigh-efficiency
solar cells for spacecraft. But the out-of-
this-world cost of GaAs itself makes these
too expensive for mass-market uses. Now,
researchers at the University of Michigan
may have found a way not only to drop the
cost of producing GaAs cells, but also to
drop the cost of the power they produce to
near that of grid power from fossil fuels.
Rao Tatavarti, a condensed matter
physicist at MicroLink Devices in Niles,
Illinois, says the new work is “a good
approach.” However, he cautions, the work
remains an early proof-of-concept, and
scaling up advances in GaAs solar cells has
long proven challenging. “In principle it’s a
good idea. But it can be a costly process to
do it on a large scale,” Tatavarti cautions.
The idea for that process isn’t dramati-
cally different from what others have been
trying to do for a long time. Crystals of
GaAs are typically grown in 200-millimeter-
diameter cylinders that are then sliced into
thin wafers. Other materials are then layered
on the wafers and patterned to make elec-
tronic devices or solar cells. But this tends to
use too much of the expensive GaAs.
More recently, groups around the world
have used GaAs as a substrate on which to
C
R
E
D
I
T
:

K
Y
U
S
A
N
G

L
E
E
/
U
N
I
V
E
R
S
I
T
Y

O
F

M
I
C
H
I
G
A
N
Recycled. A new strategy for reusing gallium arsenide wafers
may allow solar cells made from this top performer to be cheap
and flexible.
Space Solar Cells With
A Down-to-Earth Cost
the small size of the particles allowed these
changes without cracking.
Finally, the TiO
2
also constrained the
polysulfides, so these byproducts were
unable to escape and poison the rest of the
cell. At the meeting, Cui reported that the
new batteries have a capacity of about 800
to 1000 milliamp-hours per gram, roughly
six times that of the current devices on the
market. And Cui said his team charged and
discharged the battery more than 1000 times
with negligible drop off in performance.
Pooi See Lee, a materials scientist at
Nanyang Technological University in Sin-
gapore, says that Cui’s work represents
“big progress” for lithium-sulfur batteries.
In previous work, Cui’s team encapsulated
silicon nanoparticles in either a carbon or
polymer coating for use as a high capacity
anode, which can potentially give lithium-
ion batteries another 10-fold power boost.
Now, Cui says, his group is working to put
the two nanoparticle electrodes together to
see if they can produce the battery Christ-
mas presents have been waiting for.
rated into flexible materials in a shoe or a
piece of clothing, for example, the mechani-
cal forces of walking or tapping a foot would
cause the negatively and positively charged
surfaces to separate and create an electrical
potential. Hook up a wire between the two
surfaces and electrons will flow through the
wire from the electron-rich Kapton to the
electron-poor PMMA.
Initially they created a maximum poten-
tial of 5 volts. But at the meeting, Wang
reported that by patterning the surface of
the two materials to increase the amount of
friction they could create a potential of up to
1000 volts and a current density of 128 milli-
watts per cubic centimeter. A 6-centimeter-
by-6-centimeter device powered by foot tap-
ping was able to recharge a cell phone bat-
tery or power some 600 LED lights. “I was
very impressed with the power output,” says
Seung-Wuk Lee, a bioengineer and nano-
generator expert at the University of Cali-
fornia, Berkeley. The Georgia Tech results
are also reported in the 12 December 2012
issue of Nano Letters.
Wang says he is already considering
making square meter–sized devices with
up to 200 layers of nanogenerators stacked
atop one another for use in harvesting ocean
power. Such a device may be able to produce
as much as 40 kilowatts of power per cubic
meter, which could make nanogenerators a
large-scale power source.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 22
C
R
E
D
I
T
:

C
O
U
R
T
E
S
Y

O
F

B
.

J
.

F
U
L
T
O
N
,

T
H
E

B
Y
R
N
E

O
B
S
E
R
V
A
T
O
R
Y

A
T

S
E
D
G
W
I
C
K

R
E
S
E
R
V
E

A
N
D

T
H
E

P
A
L
O
M
A
R

T
R
A
N
S
I
E
N
T

F
A
C
T
O
R
Y
PETER NUGENT DROVE TO WORK ON THE
morning of 24 August 2011, still oblivious to
the faraway cosmic explosion that would con-
sume him for weeks ahead. Walking toward
the entrance of the National Energy Research
Scientific Computing Center (NERSC) in
Oakland, California, he stopped to watch
news vans covering protests against a shut-
down of cell phone service on the Bay Area
Rapid Transit system. Then Nugent, a theoret-
ical astrophysicist who could be mistaken for
a football player if he came thundering down
a hall, turned his attention away from earthly
matters and dived back into the otherworldly
pursuit of astrophysics.
That morning, he had an urgent task: fixing
a glitch in a digital pipeline that feeds astro-
nomical images from a 1.2-meter-diameter
survey telescope on Mount Palo-
mar to the computers at NERSC.
The pipeline had crashed the
night before, leaving thousands
of pictures waiting to be uploaded
into a database, where they would
be scanned by software designed
to identify potentially interesting
events such as a gamma ray burst. Shortly after
noon, when the pipeline had been restored and
all the images uploaded and analyzed, Nugent
sat down to see what the system had picked
from the previous night’s observations as can-
didates worthy of follow-up. He knew that
his buddy, Joshua Bloom, was likely doing
exactly that a few kilometers away at the Uni-
versity of California (UC), Berkeley, where
Bloom is an associate professor of astronomy.
Bloom—who had helped write the soft-
ware for the automated search—was in
fact looking through the top picks of the
night, having logged in at Berkeley. And as
they often did, the two began chatting over
Gmail while going through the results. At
12:40 p.m., Bloom messaged Nugent that he
had found an object that looked like a super-
nova that had gone off 400 million light-years
away. “That’s great,” Nugent responded.
Then, seconds later, Nugent spotted another
candidate that looked like a supernova in the
Pinwheel galaxy, only 21 million light-years
away. Because this one was so much closer
to Earth, it was potentially of greater value
to astronomers. “I see your $20 and raise
you $100,” Nugent joked. “Dang,” Bloom
replied, checking it out.
Detecting a newly exploding star or
gamma ray burst tends to quicken the pulse
of astronomers. Because of the transient
nature of stellar explosions—
they can fade away within hours
to days—the moment of discov-
ery marks the beginning of a race
against time to collect data about
the phenomenon. And so, within
minutes of finding the object that
he had labeled PTF11kly, Nugent
instant-messaged another colleague named
Mark Sullivan at the University of Oxford
in the United Kingdom, asking if Sullivan
could arrange for a telescope to start observ-
ing the object immediately.
Luckily, Sullivan was still in his office
even though the local time was past 9 p.m.
“Reckon it’s real?” Sullivan asked, to which
Nugent replied: “It is.” Sullivan e-mailed
operators of the Liverpool Telescope, a
2-meter-diameter robotic instrument on La
Palma in the Canary Islands off the coast
of Spain. Within an hour, the telescope had
begun taking images and spectra of the super-
nova, and around 2:30 p.m. California time,
Nugent and Sullivan were looking at the first
spectroscopic results, puzzling over the kind
of supernova it was. After a few minutes of
studying its spectral features, Nugent excit-
edly messaged Sullivan that the object was a
type Ia—a class of supernova that shines with
such predictable luminosity that astrophysi-
cists use it as a standard candle for measuring
cosmic distances. Following convention, the
researchers named the object SN 2011fe.
Because of their usefulness to cosmol-
ogy, type Ias are valuable finds. The one
that Nugent and his colleagues had discov-
ered was even more valuable because it had
been detected just 11 hours after the super-
nova went off, making it the youngest type Ia
discovered to date and allowing astrono-
mers to study the explosion from an earlier
stage in its progression than any type Ia seen
before. Nugent sent out an Astronomer’s
Telegram on the Web encouraging observers
around the world to follow up on the object,
which the researchers would later label “the
supernova of a generation” and “an instant
cosmic classic.”
Nugent had already called a colleague at
UC Berkeley—a quiet, 42-year-old astro-
physicist named Weidong Li—to start look-
ing at archival images of the Pinwheel galaxy
in an effort to determine the supernova’s pro-
genitor. Nugent knew that Li’s expertise in
astrometry, the measurement of the position
and motion of stars, would be instrumental
in tracing SN 2011fe’s history. Over the next
several weeks, Nugent, Bloom, and Li would
devote themselves night and day to studying
the supernova’s present and past.
Death of a Star
The discovery of a nearby supernova has brought astrophysicists
closer to understanding a class of stellar explosions. Along with that
success came an unexpected tragedy.
NEWSFOCUS
Online
sciencemag.org
Podcast interview
with author Yudhijit
Bhattacharjee (http://
scim.ag/pod_6115).
Burning bright.
The supernova SN 2011fe
(above), first detected in
August 2011, has become
one of the most closely
studied supernovae in
history.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 23
NEWSFOCUS
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

Y
.

B
H
A
T
T
A
C
H
A
R
J
E
E
/
S
C
I
E
N
C
E
;

A
L
E
X

F
I
L
I
P
P
E
N
K
O
Along with the work of dozens of others,
the efforts of these three researchers would
lead to a new understanding of how type Ias
originate and unfold. But not all three would
be around to celebrate the insights gained
from SN 2011fe, which would end up gen-
erating dozens of research papers. Months
later, one of their lives would come to a tragic
and unexpected end.
Core compression
Nugent, Bloom, and Li had known each
other for several years before SN 2011fe lit
up the sky. They had come to astronomy by
very different paths.
Nugent, the son of a lawyer, got hooked
on space as a kid following NASA’s lunar
program. His interest turned to astronomy
when his grandfather gave him a telescope
at the age of 12. In college, he briefly consid-
ered becoming an English major but hated
rewriting, so he chose physics instead. Later,
while exploring graduate schools, he ran into
David Branch—a supernova expert at the
University of Oklahoma—who quickly con-
vinced him that supernovae were the most
interesting things to study in astronomy. In
1996, he joined Lawrence Berkeley National
Laboratory for a postdoctoral fellowship that
turned into a staff position.
Li’s beginnings were a world apart from
Nugent’s. Born to a farming couple in a Chi-
nese mountain village, he was the first person
in his district to go to college. Like Nugent,
he became fascinated by supernovae, which
he would later describe as “the glorious
explosive stage of stellar
evolution.” After earning
his doctorate from Beijing
Normal University in 1995,
Li began working at the Bei-
jing Astronomical Observa-
tory, where he set up China’s
first systematic supernovae
search using a telescope at
the observatory’s Xinglong
station. Within a year, the
survey had found six new
supernovae, thanks in part to
scheduling software that Li
had written to specify what
parts of the sky the telescope
should observe when.
The success brought Li to the attention of
supernova researchers elsewhere, including
Alex Filippenko at UC Berkeley, who was
looking for somebody to lead a supernova
search he was initiating with a robotic tele-
scope at Lick Observatory in Mount Hamil-
ton, California. In 1997, Filippenko hired Li
as a postdoc.
Nugent met Li not long after he arrived at
Berkeley, at a meeting of Filippenko’s research
group, where Li passed around sweets he’d
brought from China. Nugent didn’t care for
the sweets (“Bean paste is still bean paste,”
he would later joke) but was impressed by the
smiling and soft-spoken Li.
Within months, Li had the Lick search
up and running. The very next year, the sur-
vey yielded a rich haul of 20 nearby super-
novae, marking the beginning of what was
to become a consistently productive run. Li’s
role became pivotal to Filippenko’s group.
“If Weidong were to be run
over by a truck,” Filippenko
would remark at confer-
ences, “my whole group
would fall apart.”
The youngest of the
three—Bloom—came to
UC Berkeley as an assis-
tant professor in 2005 after
getting a Ph.D. from the
California Institute of Tech-
nology (Caltech) in Pasa-
dena with work on gamma
ray bursts. He was 30, with
sparkling eyes, a blond goa-
tee, and a knack for deadpan
humor that he deployed to
spice up his talks and rib colleagues.
Bloom and Nugent became friends, and
in 2008 they began working together on the
Palomar Transient Factory (PTF)—a project
led by Bloom’s doctoral adviser at Caltech,
Shrinivas Kulkarni. The project was an auto-
mated search for transient phenomena includ-
ing gamma ray bursts and supernovae.
Astronomers had been conducting such
automated surveys for more than a decade, but
Bloom wanted to take them a step further. Until
then, automated searches—such as the one
that Li was in charge of at Mount Hamilton—
used computers to schedule observations,
control the telescope, and scan images of the
sky for possible new supernovae or gamma
ray bursts by comparing the images with older
reference images. A human being, however,
still had to inspect each candidate to deter-
mine whether it was a real astrophysical object
or something spurious like a speckle.
Bloom and his colleagues developed
algorithms to distinguish between fake and
real candidates and to determine what kind
of object a candidate might be: a gamma
ray burst, a nova, a variable star, or some-
thing else. By inspecting various features of
a candidate—such as its brightness and the
brightness of its host galaxy—the algorithms
could make a probabilistic statement about the
candidate, for example, classifying it as a
supernova with 80% probability. The algo-
rithms, called Realbogus and Oarical, gave
PTF the ability to sift through several kilo-
bytes of astronomical data within hours and
classify thousands of new candidates from
every night’s observations.
Explosion
Death comes to mortals and stars alike. For
stars that end as supernovae, however, it
brings ultimate glory: a flash of splendor often
more brilliant than the combined brightness
of an entire galaxy. Type Ia supernovae like
the one that set astronomers’ hearts racing on
24 August 2011, make up a special class of
such stellar explosions. What makes them
special is that all of them produce nearly the
same brightness.
Astronomers think type Ias arise in binary
star systems in which a small, dense star
known as a white dwarf has been steadily
accumulating material dumped onto it by a
companion star. If the white dwarf happens to
be composed entirely of carbon and oxygen,
something extraordinary happens after it has
gained enough material to approach 1.4 solar
masses. The stage is set for a runaway thermo-
nuclear reaction in which carbon and oxy-
gen atoms fuse into nickel. The nickel decays
radioactively into cobalt, which then decays
radioactively to iron, powering the super-
nova’s incandescence.
Although this theoretical model is gener-
ally accepted, astronomers have been look-
ing for empirical evidence to confirm many
of the details. A fundamental question about
the progenitor system is whether the star that
explodes is indeed a carbon-oxygen white
Partners in discovery. Peter Nugent (left) and Joshua
Bloom—and many others—followed SN 2011fe
for months.
Analyze this. Li’s work helped under-
stand the supernova’s antecedents.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 24
NEWSFOCUS
C
R
E
D
I
T
:

S
H
R
I
P
A
T
H
I

H
A
D
I
G
A
L
,

P
I
R
A
T
E

T
E
A
M
dwarf. Another question is what kind of star
the companion has to be in order for the
exploding star to result in a type Ia.
Those questions were on the minds of
Nugent and his colleagues on the afternoon
of 24 August as SN 2011fe burned in the
sky, continuing its ascent in brightness. After
calling Li—who began looking at archived
images of the Pinwheel galaxy—Nugent sent
a text message to Caltech astronomer Richard
Ellis urging him to start observing the super-
nova with the Hubble Space Telescope. But
the Hubble was booked; observations had to
wait a few days until it became free.
As the sun was setting over the West Coast,
Nugent drove to the astronomy department on
the UC Berkeley campus and walked down to
a room in the building’s basement from where
astronomers can conduct remote observa-
tions at Lick Observatory and the Keck tele-
scopes at Mauna Kea in Hawaii. A couple of
Filippenko’s students were at the controls
of the Lick Observatory. Geoff Marcy, the
planet-hunting Berkeley astronomer, and a
student were preparing to use one of the Keck
instruments. Drop everything, Nugent told
them: “You need to observe this at Lick—
and you need to observe this at Keck.” For
the next several hours, both telescopes held
SN 2011fe in their gaze, obtaining spectra.
The days that followed turned into a blur
for Nugent, Li, and Bloom as they analyzed
data pouring in from various telescopes on
the ground and in space. In the spectrum of
the supernova, Nugent found the signatures
of carbon and oxygen, suggesting that the
exploding star was a carbon-oxygen white
dwarf as models predicted. According to one
leading model, the companion star should
have been a red giant—a large, bloated star
nearing the end of its life, with a reddish enve-
lope and a relatively cool surface temperature.
But Li’s scrutiny of archival images of the gal-
axy found no such star near where the super-
nova was now blazing, effectively ruling out a
red giant as the companion.
Nugent’s calculations agreed. If the com-
panion had been a red giant or other large star,
they showed, the outermost shell of material
ejected by the supernova would have slammed
into that star within a day or so of the explo-
sion. Because telescopes had seen no sign of
such a collision, Nugent and others concluded
that the companion star had been considerably
smaller than a red giant—probably a star of
sunlike size still in the middle of its life.
For the next 2 weeks, as SN 2011fe got
brighter in the sky, Li and Nugent burned
the midnight oil to prepare their results for
publication. By 9 September, a day before
the supernova reached peak brightness—
allowing thousands of amateur astronomers
to view it with backyard telescopes—both
Li and Nugent had submitted their papers to
Nature. Adam Riess, an astronomer at Johns
Hopkins University in Baltimore, Maryland,
who years earlier had lost a lopsided game of
table tennis to Li at Berkeley, sent Li a con-
gratulatory note: “I’ll bet no one has sub-
mitted a paper on a supernova by the time it
reached peak! You are even faster in your sci-
ence work than in ping pong! I am in awe.”
Three days later, Bloom learned of a piece
of data that would help confirm the nature
of the supernova’s progenitor. A 0.4-meter-
diameter robotic telescope on the Mediter-
ranean island of Mallorca had imaged the
Pinwheel galaxy on the same night that the
supernova was detected, some 7.5 hours
before the detection was made at Palomar.
The images from Mallorca showed no super-
nova in the patch of sky where the star had
exploded, even though 4 hours had elapsed
since the moment of explosion. To have
remained undetectable for so long, Bloom
and his colleagues calculated, the exploding
star must have been at most 2% the diameter
of the sun—a white dwarf.
Brightness falls
Through the fall, SN 2011fe dropped in
brightness as its nickel decayed to cobalt, and
its cobalt to iron. As December approached,
Nugent, Bloom, and Li e-mailed back and
forth to finalize press releases their insti-
tutions were drafting to announce the two
Nature papers, which were due out in the jour-
nal’s 15 December issue.
In the second week of December, Nugent
flew to Stockholm to participate in the festivi-
ties related to the Nobel prizes. Several astron-
omers, including Nugent and Filippenko, had
been invited to celebrate the physics Nobel,
which was being awarded for the discovery
of the accelerating universe. In some sense,
the prize was a celebration of type Ia super-
novae, whose usefulness in measuring cosmic
distances was the foundation of the discovery.
On the evening of 12 December,
Filippenko returned to his hotel room after
dinner and checked his e-mail. Among the
dozens of messages in his inbox was one from
Li. The subject line said: farewell. Filippenko,
who had always worried about losing Li to
a competing research group, clicked on the
message with trepidation.
“Dear Alex,” the message began. “Please
find a seat to sit down before reading this
email. I am sure you will be shocked beyond
belief. By now, I should have already com-
mitted suicide.”
Stunned, Filippenko called Berkeley,
where it was midafternoon. It was too late, he
learned. Li had already killed himself.
In his suicide note, Li indicated that he
had taken his life because of a personal fam-
ily difficulty. Li apologized to Bloom and
Nugent for the inconvenient timing of his
death—days before the publication of the
Nature papers. Li’s cell phone number was
on the embargoed press releases issued the
week before; reporters had been calling Li’s
number without getting a response.
Nugent had just returned to Berkeley from
Stockholm and was pulling into his driveway
when he got Filippenko’s e-mail bearing the
sad news. Bloom got the message sitting in a
hotel room in Hong Kong.
Devastated, they combed through months
of their e-mail exchanges with Li, looking
for clues to depression that they might have
picked up on. They found nothing. Li’s death
would likely remain a mystery to them. All
they knew for certain was that a shining star
had dimmed. –YUDHIJIT BHATTACHARJEE
Remote. The PIRATE observatory on Mallorca island
provided key evidence for determining the star
system that gave rise to SN 2011fe.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 25
NEWSFOCUS
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

A
.

A
.

A
R
A
V
I
N

E
T

A
L
.
,

M
O
L
E
C
U
L
A
R

C
E
L
L

3
1

(
2
6

S
E
P
T
E
M
B
E
R

2
0
0
8
)

©

2
0
0
8

E
L
S
E
V
I
E
R

I
N
C
.
;

A
D
A
P
T
E
D

F
R
O
M

M
.

P
.

B
A
G
I
J
N

E
T

A
L
.
,

S
C
I
E
N
C
E

3
3
7

(
8

A
U
G
U
S
T

2
0
1
2
)
Parasitic DNA has infiltrated our genome and
threatens our future. As in most other animals,
much of the human genome derives from
self-serving DNA strands known as transpo-
sons. These genetic gypsies often jump to new
chromosome locations, sometimes disabling
genes and even triggering cancer. In the germ
line—sperm and eggs and the cells that spawn
them—a transposon hopping to a new posi-
tion can lead to sterility, a disaster from a
Darwinian point of view. “Failure to control
transposons in most animals is the surest path
to extinction,” says biochemical geneticist
Phillip Zamore of the University of Massa-
chusetts Medical School in Worcester.
For that reason, a specialized group of
RNA molecules known as piRNAs (pro-
nounced “pie-RNAs”) are the superheroes
of animal genomes. Discovered in the past
decade, piRNAs team up with certain proteins
to shackle transposons in animal germline
cells. Together, these protein-RNA combos
create a molecular defense that scientists
liken to an immune system for the genome.
Like our immune system, piRNAs and their
partners can tell friend from foe, mobilize a
response, and adapt to new invaders. Simi-
larly, our genome guardians have a memory,
a record of past threats.
“The complexity of this [piRNA] path-
way has exploded during evolution,” says
Julius Brennecke, a developmental geneticist
at the Institute of Molecular Biotechnology in
Vienna. The number of piRNA varieties that
humans produce isn’t clear, but the total could
be in the millions. “It’s not often that you dis-
cover something that is so abun-
dant and that was missed for so
long,” Zamore says. “It’s the per-
fect scientific problem.”
Researchers intrigued by this
problem have begun to sketch out
details of how these small RNAs
keep transposons in check. “We
are starting to learn what’s in the
[piRNA] toolbox,” says molec-
ular biologist Ramesh Pillai of
the European Molecular Biology
Laboratory in Grenoble, France.
But biologists still don’t know
how cells manufacture this type
of RNA, or what piRNAs might
do outside the cells of the germ line. “In
mammals, the transposon-silencing function
is just a small piece of what they do, but it’s
the only piece we understand,” Zamore says.
One recent study, for instance, raises the pos-
sibility that these molecules are important for
learning. Researchers are also stumped as to
why mice generate hundreds of thousands,
even millions, of piRNA varieties that have no
known transposon targets. “These are exciting
times” in the field, Pillai says.
Discovery of a new RNA
When researchers first detected piRNAs in
2001, they were just beginning to grasp the
importance of so-called small RNAs. These
molecules, which are typically between 18 and
40 nucleotides long and don’t code for pro-
teins, were proving ubiquitous. “Small RNAs
have been harnessed by almost every single
life form we know,” Pillai says. Organisms
deploy some small RNAs to turn down the
activity of their own genes, albeit indirectly.
Before a cell synthesizes the protein encoded
by a gene, it first makes an RNA version of
the gene, known as messenger RNA (mRNA).
The best-known types of small RNAs—
small interfering RNAs (siRNAs) and
microRNAs—target these mRNAs, destroy-
ing them or preventing the cell from translat-
ing them into proteins. Many organisms also
enlist small RNAs to defend against patho-
gens. In plants and nematodes, for instance,
small RNAs help destroy viral RNA.
Eleven years ago, Alexei Aravin, then a
graduate student at Moscow State Univer-
sity, and colleagues discovered several small
RNAs that shut down a transposonlike gene
in fruit flies. At the time, the only hint that
the molecules belonged to an unrecognized
group of RNAs was that they were slightly
longer than siRNAs, says Aravin, who is
now a molecular biologist at the Califor-
nia Institute of Technology in Pasadena. In a
follow-up fruit fly study 2 years later, how-
ever, he and colleagues identified more than
170 unique small RNAs that target trans-
posons, suggesting that the insects have a
specialized type of RNA for this function.
Aravin would soon come across this new
class of small RNAs again, though by follow-
ing a different research tack. He and other
biologists were looking into the workings of
Piwi proteins, which studies had indicated are
necessary for fertility in several kinds of ani-
mals. Piwi proteins are part of the Argonaute
family. siRNAs and microRNAs work by con-
sorting with non-Piwi Argonaute proteins
that slice up RNA molecules. Some research-
ers speculated that Piwis also functioned by
partnering with RNAs. “It was only logical
to imagine that these similar family members
would also bind to small RNAs,” says molec-
ular biologist Gregory Hannon of the Cold
Spring Harbor Laboratory in New York. In
2006, Aravin and colleagues, Hannon and co-
workers, and two other groups independently
confirmed this hypothesis, uncovering thou-
sands of small RNAs that collaborate with
Piwi proteins in mice. Researchers realized
that these small RNAs resembled the ones
Aravin and colleagues had initially identified
in fruit flies and declared that all of them were
an RNA family unto themselves, the Piwi-
interacting RNAs, or piRNAs.
piRNAs differ from microRNAs and
siRNAs in several ways (see table, p. 27).
As Zamore’s team first reported in 2006 in
Science (21 July 2006, p. 320), cells don’t
need the enzyme Dicer to make piRNAs.
However, Dicer is essential for the matu-
The Immune System’s Compact
Genomic Counterpart
Small but powerful, piRNAs protect the genome and may have other functions as well
CELL BI OLOGY
Guarding the germ line. DNA (blue) and two kinds
of protective Piwi proteins (red, green) are visible in
this section of a mouse testis.
control piRNA
Shut down. Glowing egg cells stand out in worms whose piRNAs
can’t silence a foreign DNA sequence (left panels). But when the
piRNAs recognize and shut down the sequence, the egg cells are
dark (right panels).
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 26
NEWSFOCUS
C
R
E
D
I
T
:

C
.

B
I
C
K
E
L
/
S
C
I
E
N
C
E
ration of microRNAs and siRNAs. Also,
unlike siRNAs and microRNAs, piRNAs
are exclusive to animals, occurring even in
ancient groups such as sponges.
The DNA sequences that code for
piRNAs are bunched in a few so-called
piRNA clusters. One of the field’s big myster-
ies is how these clusters give rise to piRNAs,
notes molecular geneticist Eric Miska of the
University of Cambridge in the United King-
dom. “piRNA biogenesis is still very enig-
matic.” Cells likely make an RNA copy of
an entire cluster and then dissect it, hewing
the fragments into piRNAs. But the details of
this processing remain obscure. “More than
10 proteins are involved, but we know very
little about what steps they are
doing,” Aravin says.
Detecting danger
Although the workings of the
piRNA system differ from those
of our immune system, these
two defenses face many of the
same challenges. Their first job
is detecting danger. piRNA clus-
ters are crucial for this function.
They contain partial and complete
transposon sequences, and they
serve as the memory banks for the
piRNA system. “It’s the way ani-
mals write down which transpo-
sons have invaded their genome,”
Zamore says. Each piRNA targets
transposons that contain a match-
ing sequence to its own RNA
sequence. By making piRNAs
that correspond to the transposon
sequences stored in the clusters,
animals can keep these selfish
strands in check.
But what if an animal has to
contend with a transposon that it
hasn’t encountered before? The
piRNA system relies on a nifty
trick in these situations. “It makes
use of the only thing that [selfish] genetic ele-
ments have in common—they move around
the genome,” says molecular geneticist René
Ketting of the Institute of Molecular Biology
in Mainz, Germany.
As a new transposon migrates from loca-
tion to location, it should eventually land in
a piRNA cluster. When that happens, the
transposon becomes part of the memory
bank, and the animal will begin producing
complementary, or matching, piRNAs to
thwart the genomic interloper. Each piRNA
cluster “is kind of a trap,” Pillai says. “Once
a transposon falls in, you have immunity.”
Thanks to their genomic immune sys-
tem, animals can recover from “infection” by
a new transposon, much as you get over the
flu because your immune system defeats the
influenza virus. For example, in a study pub-
lished in the 23 December 2011 issue of Cell,
molecular geneticist William Theurkauf of the
University of Massachusetts Medical School,
Zamore, and colleagues followed what hap-
pened to young female flies that inherited a
transposon called the P element, which they
hadn’t tangled with before. At first, the trans-
poson got the jump on the insects. They were
infertile and produced scant piRNAs that
had any ability to control the P element. The
genomic invader also unleashed other trans-
posons that had been lurking in the flies’
chromosomes. But as the flies grew older,
they began to rein in the P element, crank-
ing out piRNAs that targeted it. Moreover,
the researchers found that other transposons
released by the P element began falling into
piRNA clusters, presumably allowing the flies
to make piRNAs to counter them as well. As
a result, the flies regained some of their egg-
producing capability.
But how does a fly or another animal tell a
transposon from its own DNA? If the immune
system mistakes “self ” for microbial invad-
ers, its responses can trigger autoimmune dis-
eases. One way that piRNAs avoid triggering
genomic autoimmunity is their specificity;
they key on transposons with complementary
sequences. But a study of nematodes indicates
that the piRNA system might deploy a second
mechanism to prevent self-directed attacks,
suggests molecular geneticist Craig Mello
of the University of Massachusetts Medical
School. He shared the 2006 Nobel Prize in
physiology or medicine for discovering RNA
interference: the ability of small RNAs to shut
down gene activity.
To move, transposons often make an RNA
copy, or transcript, of themselves that’s con-
verted back to DNA in a new place. In the
6 July 2012 issue of Cell, Mello’s team pro-
posed a novel way that piRNAs can avoid
mistaking this transposon RNA for a cell’s
vital RNA, such as messenger
RNAs. “People thought that pi-
RNAs would target ‘aberrant
RNA,’ ” that is, any sequence that
differed from the animal’s own
RNA sequences, Mello says. He
has a different take: “Our findings
suggest that a foreign sequence is
recognized as foreign because it’s
never been expressed”—used to
make protein.
The researchers drew this con-
clusion after equipping nema-
todes with a fragment of worm
DNA that also included a foreign
sequence—instructions for mak-
ing the fluorescent protein GFP.
Mello and colleagues observed
a curious pattern in the resulting
mutant nematodes. In some of
the worms, piRNAs ignored the
inserted DNA, treating it as if it
were a normal gene. Those worms
made GFP and lit up.
But other worms remained
dark because they reacted to the
introduced DNA sequence as if
it were a transposon and shut it
down, preventing the produc-
tion of GFP. These differences
remained steady from generation to genera-
tion, Mello notes. “The ones that are on stay
on, and the ones that are off stay off.”
According to Mello, why some inserted
DNA sequences are initially expressed and
others are silenced is probably a matter of
chance. But if the DNA snippet is accepted
and used to make proteins, the animal there-
after treats it as “self,” Mello suggests. He and
his colleagues hypothesize that worms have a
molecular pathway that keeps track of which
DNA sequences have been active and prevents
piRNAs and Piwi proteins from interfering
with them. The researchers haven’t pinpointed
which molecules perform this job, Mello says,
Transposon
transcript
DNA with
PIWI
protein
Initial
piRNAs
PIWI
protein
piRNA cluster
transcript
T
PIWI PIW
in
I
ein i
No game for transposons. In the ping-pong loop, one Piwi protein and a
piRNA slice up transposon RNA with a matching sequence. The transposon frag-
ments then join a different Piwi protein to produce more matching piRNAs.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 27
NEWSFOCUS
but they suspect an Argonaute protein called
CSR-1 is the ringleader. Pillai describes this
potential recognition mechanism as “an inter-
esting idea and plausible,” adding that the
Mello group’s paper is “the only one which
might explain the available data.”
Taking on transposons
Once immune cells meet an intruder, they
counterattack. piRNAs do the same, using
a variety of measures against transposons.
Some piRNAs dive into the fray. They track
down transposon RNAs, and the Piwi proteins
they bring along slice up the rogue strands.
Some piRNAs let others do the dirty work.
In the 3 August 2012 issue of Science (p. 574),
Miska and colleagues described how piRNAs
boost their power by enlisting siRNAs to sti-
fle transposons. The reason, Miska suggests,
might be that although piRNAs come in many
varieties—more than 16,000 in nematodes—
each germline cell harbors just a few copies of
each one. “They can’t do much on their own,”
he says. In contrast, siRNAs are plentiful.
Other animals bolster their piRNAs
directly, relying on what’s called the ping-
pong amplification loop. Hannon’s team and
a group led by molecular biologist Mikiko
Siomi, now at Keio University School of
Medicine in Tokyo, independently described
this mechanism in flies in 2007, but mice and
zebrafish also take advantage of a similar pro-
cess. In the ping-pong loop, piRNAs and Piwi
proteins slice up transposon RNA. The result-
ing fragments undergo modification and join
with other Piwi proteins to cut up RNA tran-
scripts of piRNA clusters, thus making new
piRNAs (see diagram, p. 26). The loop “only
amplifies the useful piRNAs,” those with a
target available in the cell, says Brennecke, a
co-author on one of the papers.
piRNAs can fight back even if a transpo-
son remains quiescent and hides as a stretch
of genomic DNA. Researchers have found
that in mice, piRNAs spur germline cells
to affix methyl groups to transposon DNA,
preventing its transcription into RNA and
thereby blocking the rogue strand’s move-
ment to a new location in the genome.
Fruit fly piRNAs stymie transposon tran-
scription using a similar mechanism that
involves molecular modification of histones,
the protein spools around which DNA coils,
Brennecke and colleagues reported in the
21 November 2012 issue of Cell.
These strategies lead to long-term pro-
tection, and in that regard, piRNAs have our
immune system beat. After you’ve recovered
from an infectious disease, you often will be
immune to the pathogen that caused it for
life—but your children and grandchildren
won’t be. Animals’ genomic guardians, by
contrast, can suppress some transposons for
multiple generations, Mello’s and Miska’s
teams revealed in the 6 July 2012 issue of
Cell. For example, this genomic resistance
lasts for at least 20 generations in nematodes,
Miska and colleagues showed. Persistent
protection makes sense: Transposons in the
germ line can reactivate each generation, so
locking them down long-term is beneficial.
More than defense?
piRNAs may do more than thwart transpo-
sons. Some scientists suspect that they, like
siRNAs and microRNAs, help adjust gene
expression. For example, Mello and col-
leagues suggest that the targets of many nem-
atode piRNAs are some of the worm’s own
genes, not transposons. They have shown that
about 1000 of the roughly 20,000 nematode
genes are under piRNA control. Many of
the genes are normally turned off but switch
on in worms that lack one kind of Piwi pro-
tein, Mello and colleagues reported in the
6 July 2012 issue of Cell. One possibility,
Mello says, is that these genes perform func-
tions that are useful in certain environmen-
tal conditions, such as when the worms are
under stress. When times are tough, a cell in
the germ line of a parent worm might rein in
piRNAs, allowing the genes to switch on and
helping the offspring cope with adversity.
The idea that piRNAs are tweaking
gene activity gets mixed reviews from other
researchers. Some remain skeptical that pi-
RNAs ever silence genes. Even if they accept
that possibility, other scientists question
whether animals other than nematodes avail
themselves of this gene-controlling mecha-
nism. Few transposons trouble nematodes, so
the worms might have the freedom to divert
their piRNAs to new roles. “In most systems,
the evidence favors transposons being the tar-
gets” of piRNAs, and not genes, Ketting says.
Also unclear is whether piRNAs function
in nongermline cells. Most scientists have
dismissed the possibility, as siRNAs quash
transposons in these cells. Moreover, only
germline cells seem to make Piwi proteins,
piRNAs’ collaborators.
But a discovery from neuroscientist Eric
Kandel of Columbia University and col-
leagues suggests that piRNAs are active in the
central nervous system, helping create mem-
ories. In the 27 April 2012 issue of Cell, the
team reported that they had identified piRNAs
in neurons from the sea slug Aplysia. The
piRNAs help block the production of a pro-
tein called CREB2, which inhibits memory
formation in these animals. Testing piRNAs’
role in learning in other creatures shouldn’t
be difficult, Ketting says. If they do have a
role, deleting Piwi proteins in animals such as
mice or flies should cause memory lapses.
The question that has researchers scratch-
ing their heads involves the pachytene
piRNAs, which are named for the stage of
meiosis—the process that produces eggs
and sperm—in which they appear. Mammals
generate a huge number of different pachy-
tene piRNAs—one recent study estimated
the total for mice at more than 800,000,
but Zamore says that value is almost cer-
tainly too low. Yet the sequences of the
pachytene piRNAs do not match those of
any transposons, suggesting that they aren’t
targeting the rogue strands. “What these
pachytene piRNAs are doing—nobody
knows,” Aravin says.
As piRNA researchers delve into such
mysteries, some also wonder if these genomic
superheroes sometimes take the day off to
help a species adapt. We and other animals
are alive today because, over hundreds of mil-
lions of years, piRNAs helped our recent and
distant ancestors tamp down transposons. But
transposons aren’t necessarily all bad. They
also create genetic variation in the germ line
that is the raw material for natural selection.
A few researchers speculate that when con-
ditions are rough, animals might inhibit their
piRNAs to unleash transposons and trigger
more mutations, speeding up their evolution.
That idea is “a very attractive hypothesis,”
Zamore says. “I’d like to think of a way to test
it experimentally.” –MITCH LESLIE
siRNA microRNA piRNA
Length 21–24 nucleotides 20–25 nucleotides 21–31 nucleotides
Organization Double-stranded Single-stranded Single-stranded
Requires Dicer for maturation? Yes Yes No
Found in Animals, plants, Animals, plants, Only animals
fungi, protists protists
Function Controlling Controlling Blocking
gene expression, gene expression transposons
blocking transposons
COMPARING DIFFERENT KINDS OF SMALL RNAS
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 28
C
R
E
D
I
T
:

F
U
T
U
R
I
C
T
As the global financial crisis erupted in 2008,
Dirk Helbing, a physicist and mathematician
turned sociology professor, had an idea: What
if all the data on human behavior could be
integrated into a computer model that could
predict such financial calamities—or, for that
matter, revolutions, wars, the effects of over-
fishing, or epidemics? “Everybody saw that
the models we had weren’t working. And we
have so much more data available today,”
Helbing says.
A researcher at the Swiss Federal Institute
of Technology in Zurich, Helbing envisioned
a model that would help predict the societal
implications of political decisions such as
using bioethanol as fuel, banning high-fre-
quency trading, or letting Greece drop out of
the Eurozone. To make that possible, he says,
a “planetary nervous system” would collect
data on an unprecedented scale and develop
“socially aware software” to analyze it.
That vision may sound like science fiction
or even a dystopian nightmare. But it’s the
kind of idea that the European Commission,
the European Union’s executive body, might
just deem worth €1 billion. This month, the
commission will pick two winners—from a
shortlist of six—in its Future and Emerging
Technologies Flagship Initiative, a new fund-
ing program that aims to reward bold initia-
tives that attempt to reach a “visionary goal”
and provide “novel benefits for society.”
In 2011, a scientific panel picked the six
proposals based on a very short outline of
their idea from a list of 21, and then ranked
them (see table); since then, the candidate
teams, each comprising many dozens of
labs and companies, have spent 18 months
and €1.5 million to develop a research plan
as well as a Web site with splashy videos. In
December, a new group of experts evaluated
these proposals, and in late January, the com-
mission will announce which two get the go-
ahead. It isn’t obligated to pick the top two of
the experts’ unpublished ranking.
The original list of 30 projects included
some off-the wall ideas, such as Matrix
redone, which would allow people to learn
martial arts or ride a motorbike in a virtual
reality world inspired by the sci-fi flick The
Matrix; another, the Personal Fabricator Net-
work, would allow everyone to manufacture
their own smart devices at home.
But even some of the six shortlisted plans
aren’t run-of-the-mill. There are two other
modeling projects: a controversial plan to
simulate the entire human brain, and another
to build a “virtual patient” on which to test
therapies. There’s also a project to build
robots that might help you dress or make
coffee, one aiming to use tiny wearable sen-
sors to monitor your body and environment
for health risks, and a plan to develop new,
graphene-based electronics. The two win-
ners stand to receive €1 billion each over the
next 10 years.
But some European scientists worry that
the ambitions are just too grandiose—the
computer models in particular have come
under fire—and the funding too much. Aver-
age research projects might get €300,000 in
funding, says Peter König, a computational
neuroscientist at the University of Osnabrück
in Germany, and “nobody has been able to
explain to me why replacing 3000 good proj-
ects by one project is a good idea.”
And although a billion euros can free
researchers from many burdens, largesse at
this scale can also lead to waste, says Stefan
Hornbostel, a sociologist at Humboldt Uni-
versity in Berlin who heads the Institute for
Research Information and Quality Assur-
ance. “There is a danger that it does not fol-
low the burning scientific questions anymore,
but rather a bureaucratic logic that this money
needs to be spent.”
Others praise the commission for daring
to dream. Ernst-Ludwig Winnacker, who held
various positions in German and European
science funding agencies and currently heads
the Human Frontier Science Program, invokes
Lewis and Clark, who explored the American
West in the early 19th century. “They had
no idea what they would find,” he says—but
30 years later, the gold rush started. To stay
ahead, the European Union should fund pio-
neering research rather than areas where the
gold rush has already set in, he argues.
Whether all of the money will materialize
remains to be seen. The two winning teams
will together receive a total of €108 million in
the first 2.5 years; after that, this would go up
to €100 million annually each. But half is sup-
posed to come from member states or research
organizations, and how much they will actu-
ally give is unclear. The German ministry of
research, for instance, says it will wait for
the final selection before deciding what con-
tributions it can make, but it is skeptical of
the whole concept of “Flagships.” Germany
prefers using “tried and tested instruments
… instead of new complex funding mecha-
nisms,” the ministry said in a statement that
was sent to Science.
Extravagant claims
There’s no doubt that the projects are ambi-
tious. IT Future of Medicine, for instance,
aims to collate a patient’s genome, blood
parameters, medical history, and other data
into a “virtual patient,” on which any therapy
can be tried in silico. “Everything else that
is complicated or dangerous we model on
NEWSFOCUS
Europe’s €2 Billion Bet on the Future
This month, the European Union will pick two futuristic research proposals
and shower them with up to €1 billion each. But will it be money well spent?
RESEARCH FUNDI NG
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 29
NEWSFOCUS
C
R
E
D
I
T
S

(
T
O
P

T
O

B
O
T
T
O
M
)
:

F
.

G
U
I
N
E
A
;

I
S
T
O
C
K
P
H
O
T
O
;

©
C
O
T
E
S
Y
S
,

T
U
M
/
K
U
R
T

F
U
C
H
S
computers, whether designing a car or train-
ing pilots,” says project coordinator Hans
Lehrach, a director at the Max Planck Institute
for Molecular Genetics in Berlin. “But we still
crash patients in real life.”
When the original proposal list was whit-
tled down, the virtual patient ended up in
fifth place; the top spot went to FuturICT,
Helbing’s world model. But scientists have
criticized that project for underestimating the
world’s complexity. If nobody saw the Arab
Spring coming, why assume that a computer
model can predict a revolution? “People have
tried to build such huge models of the world
before, for climate, democracy, or violence,”
Hornbostel says. “But even on a theoreti-
cal level, it is completely unclear how many
of the basic components interact.” “We are
not building a crystal ball,” Helbing retorts.
“This is a tool to understand causal relation-
ships, to see what makes some systems stable
and others unstable.”
The Human Brain Project has faced vocal
opposition, too (Science, 11 November 2011,
p. 748). The idea is to model everything
known about neurons—including genetics,
chemistry, and electrical signaling—and then
stitch virtual neurons together into networks
resembling the human brain. But there are big
knowledge gaps at each stage of the process,
and many neuroscientists doubt that anything
can be learned from such a model; some argue
that it would divert money from other impor-
tant areas in neuroscience.
And the Flagship program’s review pro-
cesses may not detect weaknesses, König
says. Small research grants are critically eval-
uated and sometimes accepted only after sub-
stantial changes; with a billion euros up for
grabs, the review should be orders of magni-
tude stricter, he says. “But the opposite is the
case. Getting more money is actually easier.”
And although the projects will be evaluated
independently once under way, König says
that they are essentially too big to fail. “These
projects are so huge and there are so many
people involved that everybody wants them to
be a success,” he says. “And everybody wants
them to look like a success even if they aren’t.”
But Ewan Birney of the European Bioin-
formatics Institute in Hinxton, U.K., who’s
involved in the IT Future of Medicine proj-
ect, believes that the models’ critics may be
shortsighted. “People tend to overestimate
what they can do in 2 years and underestimate
what they can do in 10 years,” Birney says.
Emotional robots
Qualms about the models’ feasibility might
give the more practical competitors an edge.
One of those is a coordinated research effort
into graphene, a relatively new material made
of carbon atoms arranged in a honeycomb
lattice that conducts both light and electric-
ity. Andre Geim and Konstantin Novoselov,
two Europeans who won a physics Nobel
Prize for their work on graphene in 2010, are
involved in the project. Jari Kinaret of Chalm-
ers University of Technology in Gothenburg,
Sweden, who coordinates Graphene, says
that flexible electronic devices, such as an
e-reader that you could roll up, are among the
low-hanging fruits of this proposal.
Replacing silicon transistors in chips
with graphene will be more difficult because
its electrical properties make it very hard
to switch such a transistor off completely.
“Some of the things we promise are going to
happen, some of them are not going to hap-
pen,” Kinaret says. But Flagship’s long-term
funding is vital to cross the “valley of death”
between research and application, he says.
Now, Europe does more graphene research
than other parts of the world, but it obtains
fewer patents.
Another project called RoboCom aims
to make robots affordable and pioneer their
use in natural disasters, surgery, and house-
keeping; its researchers also want robots to
become more sensitive and more emotional.
The sixth project aims to develop tiny sen-
sors called “guardian angels.” Integrated into
clothes, they could monitor your heart rate,
blood alcohol level, or fatigue; sensors in cars
could help avoid crashes. They would run on
energy harvested from the sun, vibrations, or
temperature changes.
A feasible model?
Winnacker points out that the selection com-
mittee doesn’t have to pick projects that seem
closer to application. “I would always favor
the risky projects,” he says. “In the other
areas, the breakthroughs may have happened
already.” But Marja Makarow, vice president
for research of the Academy of Finland and a
former chief executive of the European Sci-
ence Foundation, wonders if the top-down
approach of the Flagship initiative can deliver
“disruptive technologies” at all. Such technol-
ogies usually emerge from high-class, basic
research, she argues: “Nobody asked for the
World Wide Web or light-emitting diodes.”
For now, the six candidates are sitting
tight and awaiting who gets the nod. The
final choice “will show just how much risk
the E.U. is willing to take,” says Wolfgang
Boch, who coordinates the initiative at the
European Commission in Brussels. But in
the end, the point isn’t just to develop new
robots or models, Boch says. “The Flagships
weren’t created to launch two projects, but to
see whether this is a feasible funding model,”
he says. And that’s a question that—at least
today—no model in the world can predict.
–KAI KUPFERSCHMIDT
<< The Human Brain Project. Supercomputers would simulate
and help people understand the human brain. Key argument: Only
a model can bring together everything we know about neuroscience.
Guardian Angels. A network of energy-harvesting sensors that can
monitor people’s health status, scan the environment for dangers,
and provide advice “to increase the happiness people experience.”
<< Graphene. Better batteries, lighter planes, and flexible elec-
tronics are some of graphene’s promises. “Disruptive science” is
hard to do piecemeal, proponents say; a Flagship grant would
allow for coordination.
<< RoboCom. Inspired by animals, its goal is to develop “robot
companions” better able to respond to human needs. Engineering
these machines would also help understand the design principles of
biological bodies and brains.
IT Future of Medicine. Aims to use individual medical data to
build a personalized computer model for 500 million Europeans. The
approach is currently pioneered in cancer treatment.
E.U. Flagships: The Candidates at a Glance
FuturICT. Proposes massive data mining to build a planetary scale
simulator freely available for use. Promises “historic breakthroughs”
in “revealing the hidden laws and processes underlying societies.”
1
2
3
4
5
6
Global forecast. An image from the FuturICT pro-
posal, which aims to help predict major events.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 30
LETTERS
edited by Jennifer Sills
LETTERS I BOOKS I POLICY FORUM I EDUCATION FORUM I PERSPECTIVES
34
Disregarding science No more stalling
38
COMMENTARY
C
R
E
D
I
T
S
:

P
H
O
T
O
S

C
O
U
R
T
E
S
Y

O
F

T
H
E

A
U
T
H
O
R
S
NextGen Speaks
MY FELLOW AMERICANS, WE AS A NATION AND
as a global community stand opposed to the
greatest threat our world has ever known…:
global warming.… As president, I hereby
announce a government-wide effort to
develop and implement a compre-
hensive science policy that will
allow us to overcome global
warming. I therefore call upon
Congress to triple the budgets
specifically designated for clean
energy research across our basic
research agencies. To fund this work, and to
unleash the power of the market to reduce
greenhouse gas emissions, I also insist that
Congress finally do what must be done: pass
a comprehensive carbon tax. My fellow citi-
zens, we have dithered too long and thus far
failed to confront what amounts to an existen-
tial crisis for our planet. I therefore ask all of
you to join with me so that we might save our
green Earth, our only home, not only for our-
selves, but also for all posterity!
BENJAMIN H. KRINSKY
Committee on Evolutionary Biology, University of Chicago,
Chicago, IL 60637, USA. E-mail: krinsky@uchicago.edu
CANADA’S GREATEST CHALLENGE TODAY IS THE
efficient and sustainable production and uti-
lization of energy. Throughout all human
history, it is our mastery of various forms of
energy that has driven our civilization, mul-
tiplied our capabilities, and raised our stan-
dard life at all levels. We substituted manual
labor with animal labor. Our steam engines
revolutionized transportation. The advent of
electricity separated production from con-
sumption and has modernized everyday life.
Fossil fuels have increased our
industrial capabilities expo-
nentially. As we have contin-
ued to expand and improve
our energy-generation capa-
bilities, we have also learned
of the monumental price these technologies
exact on our environment and consequently
our home planet. We cannot go back to the
dark ages of history, but we must change if
we wish to move forward. As your leader, I
will expand our existing energy research and
development programs and institute new
ones to seek out and implement even greater
alternatives. Our generation will give the next
one the greatest gift we can ever offer, a better
future than the present we live.
NEILSON NGUYEN
Department of Chemical and Physical Sciences, University
of Toronto, Mississauga, Mississauga, ON L5L 1C6, Canada.
E-mail: neilson.nguyen@utoronto.ca
THE BIGGEST CHALLENGE FACING GHANA IS
the high incidence of infectious diseases. My
first action as President of Ghana will be to
elevate the Health Ministry to the level of the
Ghana Armed Forces.… To combat infec-
tious diseases, I will use a four-prong battle
strategy. (i) Street-level: I will train an army
of youth equipped with public health and
environmental engineering expertise to iden-
tify and eliminate disease-causing
conditions. (ii) Town-level:
I will organize town coun-
cils with Army Captains as
heads, with a specific man-
date to source local material,
local private funding, and a
local workforce to redesign towns.
(iii) Institutional-level: I will establish one
world-class institution for research, devel-
opment, and manufacturing to operate every
phase of the health service value chain. (iv)
National-level: I will constitute a government
of people (home and abroad) who are accom-
plished early- to mid-career research scien-
Results: Experiments in Governing
If you had just been elected to your nation’s highest office, what would you say in your inaugural
address? What is the greatest challenge facing your country, and how would you use science
to address it? In October, we asked young scientists to envision this scenario and send us their
speeches. We heard from more than 200 readers, with a range of concerns spanning from health
to electricity to politics. Common themes included global warming, energy, sustainability, and
biodiversity. A sample of the best responses can be found below. To allow for as many voices as
possible, in some cases we have printed excerpts of longer submissions (indicated by ellipses) and
lightly copyedited original text for clarity. To read the complete versions, as well as many more,
go to http://scim.ag/NextGen5Results.
Submit Now: Science Communication’s Future
Add your voice to Science! Our new NextGen VOICES survey is now open:
Ideally, how will scientists share their results with each other and the public in 50 years?
To submit, go to http://scim.ag/NextGen6
Deadline for submissions is 15 February. A selection of the best responses will be published in the
5 April issue of Science. Submissions should be 250 words or less. Anonymous submissions will
not be considered. Please submit only once.
NextGenVOICES
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 31
39 40
Sizing up
surface entropy
Science
communication online
C
R
E
D
I
T
S
:

(
I
L
L
U
S
T
R
A
T
I
O
N
)

J
O
E

S
U
T
L
I
F
F
/
W
W
W
.
C
D
A
D
/
J
O
E
;

P
H
O
T
O
S

C
O
U
R
T
E
S
Y

O
F

T
H
E

A
U
T
H
O
R
S
tists, engineers, technologists, and military
leaders who understand the concept of lead-
ership with empathy and a sense of urgency.
This system will eliminate the huge loss of
productivity due to the high-level infectious
disease burden and drastically reduce poverty
and squalor in Ghana.
PATRICK KOBINA ARTHUR
Department of Biochemistry, Cell, and Molecular Biology,
University of Ghana, Legon-Accra, Ghana. E-mail:
parthur@ug.edu.gh
WE STAND TODAY AT A CRITICAL JUNCTURE
of world history. The biggest challenge
Switzerland faces is extremism. Despite
enormous efforts, we have failed to elimi-
nate this evil. We have tried to address hos-
tility with force and aggression, which has
only resulted in hurting our cause further. I
believe that we need to use our
insight from cognitive neurosci-
ence to better understand the
underpinnings and dynamics of
hostile behaviors in the extrem-
ists…. I would urge our scien-
tific community to come forward
and develop a psychosocial model, through
which we could intervene early enough to
prevent a war-child from thinking of carry-
ing a weapon.
ALI JAWAID
Brain Research Institute, University of Zurich/Swiss Federal
Institute of Technology, 8057 Zurich, Switzerland. E-mail:
alijawaid84@gmail.com
MY FELLOW DANES! I STAND BEFORE YOU TO
tell you that we are facing a crisis….Our edu-
cation system is failing the brightest of our
children. We are teaching
them to stifle their curiosity
to fit our standardized cur-
riculums…. That is why
I’m announcing optional
nationwide contests acces-
sible from any PC or tablet,
targeted at stimulating ambition in children
unchallenged by our current system. New
test subjects and themes will be announced
every Monday, centering on subjects such
as functional programming, applied math,
information gathering, and spoken language
as gauged by voice recognition software.
Tests will be published every Friday and will
stay open for 6 hours. By Sunday night, stu-
dents ranking in the top half will be able to
see their national, regional, and local posi-
tion, and may choose to post this ranking to
their social networks. This new program will
teach our children to learn to learn efficiently
and focused, while keeping them engaged
through competition with their peers.
ERIK ESMANN POULSEN
Department of Political Science and Government, Aarhus
University, Aarhus, DK-8000 Aarhus C, Denmark. E-mail:
eaep84@gmail.com
NEPAL’S ANCIENT TECHNOLOGIES, SUCH AS
pagoda architecture, agriculture, sculpture,
mining skill, and Ayurvedic medicine…
are disappearing with the older generation.
Today, Nepal is one of the poorest coun-
tries…. Contrary to the thoughts of many
Nepalese, investment in science and research
is not the waste of money. We have many
sloppy rivers, and with science we can use
them to make electricity and to irrigate the
plains. Mountains are continu-
ally getting breeze and sun-
light, and science may turn
them into usable energy.
The budget for science and
research can support practi-
cal applications of Nepal’s traditions, such as
standardization of traditional herbal drugs,
and modernization of pagoda architecture
and mining technology, which will be fruitful
at low cost. This helps sustainable industries
as well as higher education, and will amelio-
rate the economy of Nepal.
BISHNU P. MARASINI
Nepal Academy of Science and Technology, Khumaltar,
Lalitpur, Nepal. E-mail: bishnu.marasini@gmail.com
THE BIGGEST CHALLENGE FACING FINLAND
today is the widening gap between the pros-
perous, well-off individuals and…
those who cannot find their place
in society…. I will use a multi-
disciplinary approach includ-
ing psychological, medical,
genetic, social, and econom-
ics sciences to develop suitable
preventions, interventions, and ways
to stop the intergenerational transmission of
malfunctioning and poverty. …[T]he earlier a
person’s life course is changed, the easier and
more cost-effective it will be….
PÄIVI MERJONEN
Unit of Personality, Work, and Health Psychology, Insti-
tute of Behavioural Sciences, University of Helsinki,
00014, Finland and Department of Biological Psychology,
VU University 1081 BT, Amsterdam, Netherlands. E-mail:
paivi.merjonen@helsinki.fi
TODAY AS I STAND BEFORE YOU AS YOUR
appointed leader, [Pakistan has] an electric-
ity shortage,…Dengue breakouts, unsani-
tary conditions…, and a shortage of clean
drinking water and many basic health facili-
ties.… We have huge resources for electric-
ity production. We can build a good, clean
water delivery system, with
the best filtering systems. We
will research the control and
prevention of diseases and
pandemics and produce effi-
cient, cheap medicines. No
more half-baked, superstitious
ideas and methods. We will use the most
advanced and approved scientific methods to
tackle our problems, and the great minds of
this country will deliver our salvation…!
HARIS RIAZ
Department of Electrical Engineering, Pakistan Institute
of Engineering and Applied Sciences, Islamabad, Capital,
44000, Pakistan. E-mail: haris_rzz@yahoo.com
SUSTAINABLE DEVELOPMENT HAS BEEN THE
biggest challenge facing China. Once, the
technology was our hurdle, but today, it is
the policies. A major difficulty in achieving
sustainable development goals is that most
of the policy-makers are nonscientists who
lack a realistic understanding of technology’s
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 32
LETTERS
advantages and pitfalls. We must remove this
barrier, so that the benefits of our science flow
to society. I have made science an important
part of my agenda, and in the next decade,
a new program is due to launch, streamlin-
ing bureaucracy and strengthening our sci-
ence base. This program will provide sup-
port for promising students to pursue higher
education in elite scientific institutions and
universities around the world. When they
return to our country, I want to
hire many of these young sci-
entists to work as the techni-
cal and scientific assistants for
the policy-makers. I strongly
encourage them to participate
in the discussions of environmen-
tal, economic, and social issues, in which sci-
entific thinking should come into play. More
essential, policy-makers need to think like
scientists, so that they can base sustainable
development on science. I am committed to
supporting the scientists and policy-makers
who pursue this goal collectively. They will
add to the policy landscape diversity, and
bring a whole new dimension to the sustain-
able development.
JIANG ZHAO
School of Automation Science and Electrical Engineering,
Beijing University, Beijing, 100191, China. E-mail: jzhao@
asee.buaa.edu.cn
TODAY, WE START A NEW ERA IN THE DOMINICAN
Republic…. By the end of our term in office,
we plan to have created discussion in class-
rooms and conference halls instead of courts;
we want people in libraries instead of jail; we
need Academies of Science instead
of gangs. We want to proclaim
that we no longer have a gov-
ernment to control people but
one that serves to promote
welfare and respect….
LUIS B. GÓMEZ LUCIANO
Molecular and Biological Agricultural Sciences, Taiwan Inter-
national Graduate Program, Academia Sinica, Taipei, 115,
Taiwan and Consejo para el Desarrollo de Fondo Grande,
Inc., Fondo Grande, Loma de Cabrera, 63000, Dominican
Republic. E-mail: luisgomezluciano@gmail.com

…THE IDEA OF EXCHANGING FAVORS HAS
been prospering in the Brazilian politics for a
long time. The corruption stamp has marked
several Brazilian governments. Indirectly,
it kills more than cancer or AIDS. If these
billions of reais (Brazilian currency) were
invested in health, education, or science, the
progress of this country would be faster than
one could imagine. Basic scientific edu-
cation for our youngsters, who will be the
future politicians, should be the main strat-
egy for a long-term action.
Ideally, every high school
should be directly associ-
ated with a research labo-
ratory, whose masters and
Ph.D. students are responsible
for giving tutorials for the young appren-
tices. If we can implement the Max Perutz
certainty that “in science, truth always
wins,” maybe it will force national leaders to
recognize that everybody plays on the same
team and in the same World Cup.
GUILHERME MARTINS SANTOS
Molecular Pharmacology Laboratory, Department of Phar-
maceutical Sciences, Faculty of Health Sciences, Univer-
sity of Brazil, Brasília, CEP 70910-900, Brazil. E-mail:
gsantos@unb.br
PEOPLE OF AUSTRALIA, WHILE WE ARE MAK-
ing great leaps and bounds in improving
our livelihoods through access to affordable
technology, plentiful energy, and abundant
food, we are unwittingly making trade-offs
resulting in the degradation of our natural
environment and the important goods and
services it provides. As a consequence, cli-
mate change is getting worse, there is less
water, biodiversity is disappearing, land is
less productive, and people are richer yet
less satisfied. Historically, gov-
ernments have set up indi-
vidual departments to deal
with these discrete issues—
for example, environmen-
tal protection or economic
development. In so doing,
silos often inadvertently develop, resulting
in suboptimal communication which ulti-
mately leads to poor decision-making that
determines the future of complex and inter-
connected systems. Thankfully, this age of
primitive policy development is behind us;
I am honored to announce the formation
of the Department of Systems Decision
Making. Encompassing health, environ-
ment, economics, and industry, this depart-
ment will consider all dynamic relationships
that exist between natural, socioeconomic,
and political systems….
ADRIAN WARD
International Energy Centre, Brisbane, QLD 4000, Austra-
lia and School of Geography, Planning, and Environmen-
tal Management, University of Queensland, Brisbane, QLD
4072, Australia. E-mail: a.ward@uq.edu.au
SCIENCE IS SENDING US BLARING SIGNALS
that there are urgent problems facing the
United States and every nation, such as the
interrelated issues of climate change, over-
population, and energy generation. All of
these problems will require new scientific
and multinational solutions. Yet…only 20%
of the world’s population—those from the
developed world—produces three-quarters
of peer-reviewed scientific papers. We are
doing both our country and our world a mas-
sive disservice by so severely underutilizing
all of the potential brainpower
our global civilization has
to offer. We need more pro-
grams that facilitate scien-
tific collaborations and fos-
ter scientific capacity build-
ing between developing and
developed countries. We need to provide
more ways for scientists in our own country
to venture into the developing world to do
research alongside scientists there….
CHRISTA A. HASENKOPF
Cooperative Institute for Research in the Environmental Sci-
ences, University of Colorado, Boulder, Boulder, CO 80309,
USA and National University of Mongolia, Ulaanbaatar,
Mongolia. E-mail: christa.hasenkopf@colorado.edu
…WHILE THERE ARE EXAMPLES OF WORLD-
class excellence in our science education
system, all too frequently we as U.S. citizens
lack easy access to real science and real sci-
entists. This is why, as President, I will create
the Presidential Publicity Platform. Through
an ongoing weekly format, I will take the
time to sit down and speak with
some of our nation’s scientists
about topics central to our
future as a country. In these
online fireside chats, we will
create a forum for the pub-
lic to experience science and
ask their own questions. …I believe that the
strength of our nation has always resided
in its ideas, and in America’s affection for
the pragmatic. It is time for us to renew our
acquaintance with science, to renew our
faith in its recommendations, and to renew
once more the promise of our shared future.
SETH M. STEVENSON
Science Department, Hufford Junior High, Joliet School Dis-
trict 86, Joliet IL 60441, USA. E-mail: sethmstevenson@
gmail.com
Letters to the Editor
Letters (~300 words) discuss material published in
Science in the past 3 months or matters of gen-
eral interest. Letters are not acknowledged upon
receipt. Whether published in full or in part, Let-
ters are subject to editing for clarity and space.
Letters submitted, published, or posted elsewhere,
in print or online, will be disqualified. To submit a
Letter, go to www.submit2science.org.
C
R
E
D
I
T
S
:

(
G
U
I
L
H
E
R
M
E

M
A
R
T
I
N
S

S
A
N
T
O
S
)

K
E
N
I
A

R
I
B
E
I
R
O
;

(
S
E
T
H

M
.

S
T
E
V
E
N
S
O
N
)

A
N
D
R
E
A

S
T
E
V
E
N
S
O
N

W
O
N
/
B
I
O
M
O
D
A
L

L
L
C
;

O
T
H
E
R

P
H
O
T
O
S

C
O
U
R
T
E
S
Y

O
F

T
H
E

A
U
T
H
O
R
S

Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

BOOKS ET AL.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 34
A
lthough science has long been recog-
nized as our most reliable pathway to
truth, people are sometimes reluctant
to accept scientific evidence,
particularly when it challenges
established practices or cher-
ished beliefs. In Failed Evi-
dence: Why Law Enforcement
Resists Science, David A. Har-
ris accuses police and prosecu-
tors of unwarranted skepticism
toward science and tries to
explain their perspective. His
provocative book will inter-
est those concerned broadly with rejection of
science as well as those interested in the U.S.
criminal justice system.
Like Naomi Oreskes and Erik Conway in
their Merchants of Doubt ( 1), Harris (Univer-
sity of Pittsburgh School of Law) provides
rich examples of the pigheadedness and tor-
tured logic of people confronted with incon-
venient scientific findings. He also explores
the underlying reasons for their resistance.
Whereas Oreskes and Conway focus on the
ideological and financial motives of those
who have challenged scientific evidence (on
the dangers of tobacco, acid rain, and global
warming), Harris focuses on cognitive bias
and the pressures and incentives of the insti-
tutional environment.
DNA evidence has proven the innocence
of over 300 mistakenly convicted men and,
in the process, demonstrated the fallibility
of our criminal justice system. Harris dis-
cusses three problems that have contributed
to such miscarriages of justice: mistaken eye-
witnesses, false confessions, and misleading
forensic science evidence. Scientific research
has shown ways to reduce each of these prob-
lems, Harris claims, but the criminal justice
system has failed to adopt promising science-
based reforms. The reason, he argues, is ada-
mant opposition by police and prosecutors
who refuse to accept what the research shows.
For example, a large body of psychologi-
cal research has shown that changes in line-
up procedures can improve the ability of
eyewitnesses to choose the actual perpetra-
tor from a lineup of possible suspects. Sim-
ple alterations such as presenting the possi-
ble suspects sequentially rather than simul-
taneously reduce the rate of false but not of
correct identifications. The
rate of false identifications
also declines when the offi-
cers who interact with the
eyewitness during the proce-
dure are “blind”—when they
do not know which person in
the lineup is the real suspect.
Yet police and prosecutors
have resisted calls to conduct
lineups in a blind, sequential
manner. In a telling chapter, “In their own
words,” Harris reviews various explanations
that police and prosecutors offer for opposing
these reforms and finds them illogical, convo-
luted, and unconvincing.
The heart of the book is Harris’s effort
to explain “the real reasons” that police and
prosecutors resist change. In a particularly
interesting chapter, Harris argues that the
problem stems, in part, from psychological
biases arising from cognitive dissonance,
group polarization, loss aversion, and status
threats. Although fascinating, his effort to
weave psychological theory into an explana-
tion for skepticism toward science seems ad
hoc and speculative. He is on firmer ground
in a subsequent chapter that focuses on insti-
tutional and political barriers to change. Here
Harris draws on his deep knowledge of the
criminal justice system to identify various
institutional incentives and cultural factors
that make it easier for police and prosecutors
to see dangers rather than benefits in science-
based proposals for reform.
Harris provides a list of changes that he
would like to see in procedures for eyewit-
ness identification, interrogation of sus-
pects, and evaluation of forensic science. In
his closing chapters, he offers a variety of
practical suggestions to reformers regard-
ing how those might be achieved. He finds
“reasons for hope” in improvements that
have been instituted in a number of states.
He emphasizes the need to gain the sup-
port of police and prosecutors for a reform
agenda and claims that this is possible. How-
ever, cynical readers will note that many of
the examples he cites are reforms that were
imposed on police and prosecutors by court
rulings or executive order rather than ones
they accepted willingly.
For readers interested in the phenomenon
of unwarranted skepticism toward science,
the book’s chief value lies in Harris’s detailed
account of social and psychological factors
that cause police and prosecutors to resist
science-based reforms. Merchants of Doubt
located the roots of such skepticism in the
coordinated efforts of ideologues committed
to a particular economic vision. Failed Evi-
dence traces it to more mundane sources—
Rejecting the Evidence
SCIENCE AND JUSTICE
William C. Thompson
The reviewer is at the Department of Criminology, Law and
Society and School of Law, University of California, Irvine,
CA 92697, USA. E-mail: william.thompson@uci.edu
Failed Evidence
Why Law Enforcement
Resists Science
by David A. Harris
New York University Press,
New York, 2012. 270 pp. $35.
ISBN 9780814790557.
Problematic procedure. Traditional simultaneous presentation of suspects in police lineups allows the witness
to choose who looks most like the person remembered (a relative judgment). Sequential presentation forces the
witness to decide whether the individual on view is the one remembered. C
R
E
D
I
T
:

B
E
T
T
M
A
N
N
/
C
O
R
B
I
S
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

BOOKS ET AL.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 35
A
round the world, countries are con-
fronted with rising rates of obe-
sity and chronic diseases. Globally,
four chronic diseases now account for three
of every five deaths: heart disease, com-
mon cancers, respiratory disease, and diabe-
tes. Although thought to be associated with
higher standards of living, they now affect
more poor people than wealthy ones. Dietary
energy consumption is rising while physical
activity is declining; contents and quantities
of diets have changed along with healthiness
of available food. Simultaneous transitions
in key risk factors indicate that the causative
circumstances differ markedly from those
even 25 years ago. Diet and physical activ-
ity now account for 40% and 10% of deaths
from these diseases, with tobacco exposure
accounting for 17%. Also changing are the
“causes of causes”: political, industrial, phil-
anthropic, and governance factors as well as
the demands on health systems (on both the
public health and the medical-care delivery
sides). The changed circumstances of our
lives appear to be acting in concert to cre-
ate an emergent property of ill health. How
do we make sense of the changes and alter
approaches established in a different world to
secure our well-being in the new one?
Contributors to Sick Societies: Respond-
ing to the Global Challenge of Chronic Dis-
ease lay out the case above and then demon-
Countering Chronic
Diseases
HEALTH
Linda P. Fried
The reviewer is at the Mailman School of Public Health,
Columbia University, 722 West 168th Street, 14th Floor,
New York, NY 10032, USA. E-mail: lpfried@columbia.edu
strate that the solutions lie in prevention at
the population level. They cite published data
that 80% of heart disease, stroke, and type 2
diabetes and 40% of cancers could be avoided
through healthy diet, regular physical activ-
ity, and avoidance of tobacco use. The vol-
ume focuses on identifying why although
“preventing diseases may be biologically fea-
sible” it does not appear to be happening. The
book boldly tackles this critical question—
and takes few prisoners in doing so.
The ambitious book attempts to lay out
all the factors driving the complex system
that led to the ill health of our
recent creation. The contribu-
tors examine both causes and
obstacles to solutions. The open-
ing chapters offer a useful sum-
mary of the case of how and why
the causes of our ill health are
changing. The basics of know-
ing how much disease exists, and
where, are problematic, and this
comprehensive overview rests
on essential epidemiology. The
researcher-authors call for an
understanding not just of how the risk factors
for chronic disease have changed in the past
few decades globally but also of the severe
challenge that the upsurge in chronic diseases
poses to low- and middle-income countries
(countries simultaneously still grappling with
infectious diseases, safe water, and adequate
food along with preventing infant and mater-
nal mortality). They then offer extensive
descriptions of the players (governmental,
corporate, civil society, and philanthropic)
in global health governance and the 20th-
century approaches we continue to follow.
Subsequent contributions complement this
global perspective with analyses of the fac-
tors at play in different countries, showing
how regional variations in the drivers of dis-
ease must also be attended to.
The book makes a strong case for the
importance of addressing the underly-
ing drivers of the risk factors, the causes of
causes: globalized food sources, marketing
of unhealthy products, adverse incentives in
food pricing, urbanization and urban envi-
ronments that provide inadequate access to
healthy food and little opportunity for physi-
cal activity—compounded by environmen-
tally adverse changes. The shifts in risk fac-
tors are made palpable through a history of
food and the food industry and analyses of
governance, power, and drivers of decision-
making that affect global health.
The authors do an excellent job laying
out the many factors at play in the growing
epidemic of chronic diseases. They demon-
strate that the circumstances underlying these
diseases are, in many cases, beyond the con-
trol of individuals and that institutions and
leaders outside of health systems and sci-
ence have to be part of the solutions. How-
ever, the book is limited and limiting in sup-
porting the achievement of these solutions. It
undercuts that goal with its accusatory tone
toward all players (from scientists to leaders
of industry), simplistic labeling of inaction as
the result of vested interests and conflicts of
interest, dismissal of the credibility of initial
corporate steps in the direction of solutions,
and stress on scientists’ lack of
impact in creating change from
evidence. The book’s bluntness
in such assertions may serve to
take elephants from the ceiling
and place them on the table for
discussion. But it might have
moved us closer to solutions
had it laid out a vision for how
the different sectors must work
together, based on evidence
as to the causes of causes. In
particular, although contribu-
tors mention examples of approaches (such
as Finland’s North Karelia project and New
York City’s public health system) that have
led to declines of chronic diseases, these are
not analyzed in the depth required to reveal
the intersectoral responsibility and collab-
oration needed to create such successes.
Efforts in North Karelia have produced a
75% decline in cardiovascular mortality over
23 years; prevention is responsible for half of
the dramatic decline in cardiovascular mor-
tality in the United States in the late 20th cen-
tury. Such existing models offer a start for
envisioning (and analyzing the limitations
of) generalizable alternatives to our current
atomized situation of siloed sectors that the
authors describe.
Ultimately, the authors clearly argue that
stemming epidemic chronic diseases, both
regionally and globally, will require popu-
lation-wide measures and prevention. That
places the responsibility, appropriately, on
public health science and public health solu-
tions, with implementation resting on con-
certed action by all societal sectors. The
book falls short of elucidating a path to
ensuring a system that can reach that goal
while also taking appropriate care of those
who are sick. Nonetheless, Sick Societies
effectively articulates the need for consid-
erably greater investment (by both govern-
ments and industry) in research to identify
the levers for prevention if we are to reverse
the upsurge in chronic diseases.
Sick Societies
Responding to the
Global Challenge of
Chronic Disease
David Stuckler and
Karen Siegel, Eds.
Oxford University Press,
Oxford, 2011. 372 pp.
Paper, $75, £39.99.
ISBN 9780199574407.
10.1126/science.1230826
in criminal justice, a mindset that arises from
the roles that police and prosecutors play in a
complex institutional structure and the incen-
tives that structure provides. Where we stand
is said to depend on where we sit. Harris sug-
gests this may be particularly true when we
assess scientific evidence.
References
1. N. Oreskes, E. M. Conway, Merchants of Doubt: How a
Handful of Scientists Obscured the Truth on Issues from
Tobacco Smoke to Global Warming (Bloomsbury, New
York, 2010); reviewed in (2).
2. P. Kitcher, Science 328, 1230 (2010).
10.1126/science.1231183
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 36
POLICYFORUM
T
o meet surging domestic energy
demand, provide power to the largest
population in the world that lacks elec-
tricity (> 400 million people), and reduce rap-
idly growing CO
2
emissions, the Government
of India (GOI) has embarked on a fast-track
dam-building program. Over the next several
decades, the GOI aims to construct 292 dams
throughout the Indian Himalaya, doubling
current hydropower capacity and contribut-
ing ~6% to projected national energy needs
by 2030 ( 1). With the use of coal set to expand,
India’s total carbon emissions are projected to
more than double by 2030 ( 2). New dams can
play a dual role, helping to limit emissions
while providing power to needy people. But
major problems loom. We discuss approaches
to these as Himalaya hydropower expands.
Biodiversity Impacts
Studies have recognized dam building as
the most substantial human impact on riv-
erine ecosystems ( 3). But most studies of
ecological effects of river regulation have
been carried out in the developed northern
hemisphere; such work is largely unknown
in India. We assessed impacts and trends of
land-use changes from proposed dam build-
ing on terrestrial biodiversity in the Indian
Himalaya ( 4). Of 292 proposed dams, the
study assessed 132 for which public data
were available, ranging from 7 to 11,000
MW in size. Of these, 90% would be run-of-
the-river dams without storage reservoirs,
yet this would not change the impacts high-
lighted below.
If all dams are constructed as proposed, in
28 of 32 major river valleys, the Indian Hima-
laya would have one of the highest average
dam densities in the world, with one dam for
every 32 km of river channel. Proposed loca-
tions of dams correlate with zones of species
richness for angiosperms, birds, fishes, and
butterflies. In the Indian Himalaya, subtropi-
cal and temperate forests are most vulnerable
to species losses driven by land-use changes
( 5), yet 88% of proposed dams are located in
these ecosystems (see the figure). Over half
of the dams would be in dense, relatively
undisturbed forests. Forest loss due to direct
submergence and habitat degradation from
dam building could lead to loss of 22 angio-
sperm and 7 vertebrate taxa by 2025 ( 4). This
conservative estimate did not consider effects
of habitat fragmentation or isolation on mul-
tiple endemic species, general infrastructure
development (such as smaller hydro projects
and roads), or climate change.
Weak Laws and Practices
Since 1994, India has had a national environ-
mental impact assessment (EIA) law with
specific provisions to address threatened and
endangered species, protected areas, and
other biodiversity concerns. Each state where
proposed water resource development proj-
ects are located carries out initial planning
with GOI oversight ( 6). States then imple-
ment EIA provisions after federal approvals.
However, there are problems with using inad-
equate baseline data, monitoring of compli-
ance with EIA protocols, and weak enforce-
ment of sanctions when compliance is poor ( 7,
8). For example, states have approved projects
before mandatory public hearings were held
( 9). The EIA process has already proven to
be inadequate to address cumulative impacts
from single dams ( 10). More important, for
dam construction at the scale proposed in the
Indian Himalaya, no cumulative effects anal-
yses are legally required in EIAs, including
cascade effects of multiple hydropower proj-
ects in a single river basin. Recent reviews
have recommended reductions in dams even
without including analysis of sediment load
changes, road construction, climate change,
and livelihood impacts ( 11, 12). The GOI has
not followed these recommendations. Given
this history and the scale of proposed hydro-
power development, it is doubtful that EIA
law as currently implemented can adequately
address nearly 300 new dam proposals.
Because of high population density and
the GOI’s historical commitment to hydro-
power, dams have displaced Indian citizens
for decades. GOI figures on how many people
have been displaced are difficult to access.
Estimates range from 16.4 to 40 million; only
China has displaced more people by dams
( 13). Although a national resettlement stat-
ute has existed since 2004, Indian resettle-
ment law suffers from fragmented division
of responsibilities between state and federal
regulators. There are no binding provisions to
address cumulative social impacts. With mil-
Threats from India’s Himalaya Dams
ECOLOGY
R. Edward Grumbine
1
* and Maharaj K. Pandit
2

Ecosystem damage and population
resettlements loom, owing to poor
planning and impact assessments.
*Author for correspondence. E-mail: ed.grumbine@gmail.com
1
Key Laboratory of Biodiversity and Biogeography, Kunming
Institute of Botany, Chinese Academy of Sciences, Kunming,
650204, China.
2
Department of Environmental Studies,
Centre for Interdisciplinary Studies of Mountain and Hill
Environment, University of Delhi, Delhi 110007, India.
China
Regional borders disputed by China
Regional borders disputed by China
Regional borders disputed by Pakistan
India
India
Bhutan
Nepal
P
a
k
i
s
t
a
n
Bangladesh
M
y
a
n
m
a
r
International boundary
River valley boundary
River/water body
Dam sites
Dam sites (in species-rich zones)
Dense forest
Open forest
Scrub
100 0
100
200
300 50
Kilometers
N
Distribution of proposed dams in the Himalayan states of India. Spatial data from Landsat Enhanced The-
matic Mapper Plus (www.landcover.org/data) (30-m resolution) to delineate forest and nonforest land cover, ana-
lyzed using Erdas Imagine 9.1 software. Dam site locations are from ( 4). Borders reflect the authors’ original map
of Indian government proposals. References to disputed regions are added by Science to reflect a United Nations
map of Southeast Asia (www.un.org/Depts/Cartographic/map/profile/Souteast-Asia.pdf). Science does not have a
position regarding boundaries and names shown on this map.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 37
POLICYFORUM
lions of people losing land rights and resettled
without adequate compensation, it is easy to
understand why demonstrations against dams
are an ongoing feature of Indian civil society.
Improving Planning and Implementation
Most of India’s proposed Himalaya hydro-
power projects are not yet built; the country
does not have to sacrifice its biodiversity nor
inequitably resettle people affected by dams.
Yet even before the massive national black-
out in July 2012, the GOI has been under
intense pressure to increase electricity sup-
plies. Solutions lie with integrating national
energy planning and specific hydropower
sector reforms.
First, despite the availability of system-
atic planning frameworks, the GOI has not
conducted a countrywide review of future
energy and water requirements that consid-
ers alternatives beyond hydropower; such a
study might reduce the assumed need for so
many dams ( 14, 15). For example, India loses
20 to 30% of total power generated ( 16), an
amount greater than all current hydropower
production. Reducing these losses, mainly
due to poor grid transmission and theft,
could help meet GOI hydropower targets out
to 2030, likely at less cost. India’s ongoing
Ganga River Basin Assessment and recently
released draft National Water Policy are steps
in the right direction, but they say little about
hydropower in the Himalaya and make no
link between protecting forests and provi-
sions of water resources ( 17, 18).
A second action would be improving
India’s EIA process. Conflict between fed-
eral and state decision-making, leading to
lack of institutional accountability, is a prob-
lem throughout all stages of EIA implemen-
tation. A first step would be to ensure profes-
sional standards are followed for certifica-
tion of private consultants who prepare EIA
reports. Another improvement would be for
the GOI, states, and private developers to
embrace new protocols released by the Inter-
national Hydropower Association ( 19). These
standards could help identify problems early
in the EIA process, include more stakeholder
public participation, and better target national
and state parties to be held accountable for
implementation. EIA revision must include
comprehensive cumulative impacts and river
basin assessments. This would aid project
evaluation and increase India’s capacity to
engage in transboundary river negotiations.
Third, resettlement regulation reform
would improve India’s management of dam
development. A roadmap for using social
assessment data in project decision-making
needs to be prepared. Rules governing how
public participation is integrated into reset-
tlement processes need to be tightened ( 20).
Resettlement compensation must be reformu-
lated. Historically, the GOI has approached
resettlement as a law-and-order issue, not
as one about loss of homelands and cultural
traditions. Although the scale of potential
new downstream resettlement across India is
sobering ( 21), there are no data publicly avail-
able on how many citizens face resettlement
because of Himalayan hydro projects. Regu-
lators need clear resettlement rules before
projects multiply.
Transboundary Politics of Hydropower
Beyond specific project concerns, there are
political reasons to strengthen India’s energy
planning, EIA, and resettlement laws. The
two most important basins for hydropower
development in the country, the Brahmapu-
tra and the Indus, are transboundary rivers
that begin on the Tibetan Plateau (China) and
flow into Bangladesh and Pakistan, respec-
tively. Every neighbor of India with undevel-
oped hydropower sites is building or planning
to build multiple dams, totaling at minimum
129 projects ( 22). There has been little coor-
dination between nations; India is not unique
as it appears ready to expedite environmen-
tal review of hydropower projects on these
rivers to gain “prior appropriation” of water
resources before neighbors develop dams
( 23). Without negotiations to create inte-
grated transboundary river basin planning,
it is unlikely that any single nation’s devel-
opment can be optimized. This concern is
underlined when projected climate change
impacts on Indian Himalaya rivers are con-
sidered. Out to 2050 (well within the aver-
age lifetime of dams being built today), mod-
eled decreases in mean upstream water sup-
ply from the Brahmaputra and Indus are 19.6
and 8.4%, respectively ( 24). Decreases of this
magnitude may lead to reduction in a river’s
capacity to produce electricity, exacerbating
regional political tensions over water-based
energy production.
In this context, improved assessment
of hydropower development in the Indian
Himalaya assumes international signifi-
cance. Given the large number of regional
hydropower projects, it is essential to encour-
age transboundary river basin manage-
ment throughout the Indian Himalaya and
beyond ( 25). Multilateral energy partnerships
between countries may eventually replace
current state-focused development behavior.
For India’s proposed Himalaya dams, at mini-
mum, it would be desirable to prioritize proj-
ects located in degraded forests, whereas con-
struction around biodiversity-rich, dense for-
ests and sites with thorny resettlement issues
should be subject to integrated scientific and
social review before final development deci-
sions are made.
References and Notes
1. Government of India, Hydropower Policy (Ministry of
Power, New Delhi, 2008); http://powermin.nic.in/whats_
new/pdf/new_hydro_policy.pdf.
2. British Petroleum, BP Energy Outlook 2030 (BP, London,
2012).
3. M. Dynesius, C. Nilsson, Science 266, 753 (1994).
4. M. K. Pandit, R. E. Grumbine, Conserv. Biol. 26, 1061
(2012).
5. M. Pandit, N. S. Sodhi, L. P. Koh, A. Bhaskar, B. W. Brook,
Biodivers. Conserv. 16, 153 (2007).
6. N. Choudhury, Sustainable Dam Development in India
(German Development Institute, Bonn, 2010).
7. J. Singh, Curr. Sci. 90, 784 (2006).
8. M. Menon, K. Kohli, Econ. Polit. Wkly. 44, 20 (2009).
9. T. Sharma, Mega dams: Campaigning against the plans
of the Indian government (2012); www.opendemocracy.
net/openindia/tanmoy-sharma/mega-dams-campaign-
ing-against-plans-of-indian-government.
10. O. Nandimath, Oxford Handbook of Environmental Deci-
sion Making in India: An EIA Model (Oxford Univ. Press,
New Delhi, 2009).
11. Centre for Interdisciplinary Studies of Mountain & Hill
Environment, Carrying Capacity Study of Teesta Basin in
Sikkim (Ministry of Environment & Forests, Government
of India and NHPC, New Delhi & Faridabad, India, 2007).
12. Wildlife Institute of India, Assessment of Cumulative
Impacts of Hydroelectric Projects on Aquatic and Ter-
restrial Biodiversity in Alaknanda and Bhagirathi Basins,
Uttarakhand (Wildlife Institute of India, Dehradun,
2012).
13. N. S. Negi, S. Ganguly, Development projects vs. inter-
nally displaced populations in India: A literature based
appraisal (Working paper No. 103, Center on Migration,
Citizenship, and Development, Univ. Bielefeld, Bielefeld,
Germany, 2011); http://pub.uni-bielefeld.de/publica-
tion/2535157.
14. World Wildlife Fund, Climate Solutions: WWFs Vision for
2050 (World Wildlife Fund, Gland, Switzerland, 2007).
15. P. Sreedharan et al., Electricity Grid Evolution in India
(Energy and Environmental Economics Inc., San Fran-
cisco, 2011); available online.
16. A. Subramanian, Can India’s power problems be solved?
(2012); www.piie.com/blogs/realtime/?p=3051.
17. Press Information Bureau, Government of India,
Third Meeting of the National Ganga River Basin
Authority, (2012); http://pib.nic.in/newsite/erelease.
aspx?relid=82308.
18. Ministry of Water Resources, Draft National Water Policy
2012, (Government of India, New Delhi, India, 2012);
http://mowr.gov.in/.
19. J. Tollefson, Nature 474, 430 (2011).
20. T. Rajaram, A. Das, J. Environ. Manage. 92, 140 (2011).
21. B. Richter et al., Water Altern. 3, 14 (2010).
22. S. Dharmadikhary, Mountains of Concrete: Dam Build-
ing in the Himalayas (International Rivers, Berkeley, CA,
2008).
23. T. Hennig, in Hydropower-Practice and Application,
H. Samadi-Boroujeni, Ed. (InTech, Shanghai, China,
2012), pp. 293–320.
24. W. W. Immerzeel, L. P. van Beek, M. F. Bierkens, Science
328, 1382 (2010).
25. C. Ebinger, Energy and Security in South Asia: Coopera-
tion or Conflict? (Brookings Institution Press, Washing-
ton, DC, 2011).
Acknowledgments: Authors acknowledge support from the
Chinese Academy of Sciences (grant 2010T1S2), University of
Delhi, and University Scholars Programme, National University
of Singapore (R-377-000-035-112; Ministry of Education–
Academic Research Fund).
10.1126/science.1227211
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 38
PERSPECTIVES
T
he ribosome is a versa-
tile enzyme, but it can-
not synthesize all pro-
teins equally well—certain
combinations of amino acids
pose problems. Ribosome
stalling by fairly long amino
acid motifs can regulate gene
expression in a variety of
organisms, from bacteria to
humans ( 1, 2). In these cases,
the nascent peptide interacts
with the peptide exit channel
of the ribosome to induce a conformation
that prevents peptide bond formation. Two
papers in this issue, by Doerfel et al. on page
85 ( 3) and Ude et al. on page 82 ( 4), reveal
that stalling is common and fundamental:
Short, proline (Pro)–rich motifs impede pro-
tein synthesis, and stalling is alleviated by a
poorly understood elongation factor, EF-P.
These studies offer a model of the biologi-
cal function and mechanism of EF-P through
a convergence of biochemical and genetic
methods.
EF-P was identified biochemically nearly
40 years ago ( 5); its name reflects the ability
to enhance the peptidyl-transferase (P) activ-
ity of the ribosome, which links amino acids
during the translation of messenger RNA
(mRNA). While clues to its function trickled
in over the years, a compelling biological role
for EF-P remained elusive. Following up on
the original biochemical findings, Doerfel et
al. took a broader look at the substrate speci-
ficity of EF-P; in testing all 20 amino acids,
it became clear that its stimulatory effects on
catalysis with Pro were unique. Pro is both
a poor donor and acceptor of peptide ( 6– 8).
Indeed, Doerfel et al. found that the synthesis
of peptides with consecutive Pro residues is
severely inhibited, possibly because Pro is a
secondary amine (an imino acid) or because
of the conformational constraints imposed by
its cyclic side chain. Their biochemical stud-
ies revealed that EF-P supports rapid and
robust synthesis of polyproline sequences.
Ude et al. arrived at the same conclu-
sion through genetic analysis of cad genes
involved in acid resistance in Escherichia
coli. Using transposon mutagenesis, they
identified yjeK as essential for robust expres-
sion of cadA and cadC. YjeK is one of two
proteins that make essential covalent modifi-
cations of EF-P ( 9, 10); deletion of the other,
YjeA, or of EF-P itself results in loss of CadC
function as well. It turns out that the cadC
gene has a string of three prolines in its cod-
ing region; the authors pinpoint this sequence
as the site of EF-P action. Deletion or muta-
tion of the polyproline motif makes CadC
synthesis independent of EF-P.
Additional findings from Doerfel et al.
and Ude et al. further delineate the scope of
EF-P activity. One observation is that three
proline residues are required for robust stall-
ing, and longer polyproline sequences do not
appear to have any additional effect. Of more
than 10 examples tested in these two studies,
ribosome pausing was detected at the poly-
proline motif whether it is engineered or nat-
urally occurring. This suggests that the phe-
nomenon is context-independent. The amino
acid—not the transfer RNA (tRNA) that it is
tethered to during translation—is proposed to
be the critical player, as different Pro codons
elicited similar results. Also, the role of EF-P
appears to be limited to polyproline stretches;
other translational intermediates that can be
detected in vitro were not resolved by EF-P.
Structural studies show that the three
domains of EF-P mimic the structure of
tRNA ( 11). This shape allows EF-P to inter-
act with the ribosome between the canoni-
cal E and P sites on the ribosome where
tRNAs bind. One domain of EF-P contacts
the small ribosomal subunit near the mRNA,
and another domain contacts the large sub-
unit near the peptidyl-transferase center ( 12).
Of particular interest is a conserved loop in
domain I of EF-P, corresponding to the 3′
end of the tRNA, where the amino acid is
tethered. A conserved lysine residue in this
loop, Lys
34
, lies close to the ribosomal pep-
tidyl-transferase center and to the peptidyl-
tRNA linkage. This Lys residue is modified
with β-lysine by the enzymes YjeK and YjeA
( 9, 10), which creates a long, flexible cofac-
tor that could reach into the ribosome’s active
site and alter its conformation or the chemis-
try of peptide bond formation. Biochemical
and genetic data show that this modification
is essential for EF-P activity ( 10).
From these varied approaches, a clarify-
ing model of EF-P function emerges (see the
figure). A ribosome stutters upon encounter-
ing a string of Pro codons in mRNA; spe-
cifically, when two consecutive Pro residues
are found at the end of a nascent peptide
chain, the rate of peptidyl transfer to subse-
quent Pro-tRNA in the A site of the ribosome
decreases, presumably because the geome-
try of peptidyl-transfer reaction is perturbed.
EF-P enters the ribosome through the empty
E site and binds close to the peptidyl-tRNA,
which allows the β-lysyl moiety on Lys
34
to
reach into the active site of the ribosome and
restore productive catalysis. Following pep-
tide bond formation, EF-P dissociates and
protein synthesis resumes as usual.
The importance of this mechanism is
underscored by the universal conservation of
EF-P and its orthologs. Although some bac-
teria can tolerate deletion of the efp gene, its
loss induces pleiotropic phenotypes, includ-
ing loss of virulence in Salmonella ( 10). The
findings of Doerfel et al. and Ude et al. sug-
gest that these phenotypes can be traced to
defects in the translation of proteins contain-
ing polyproline motifs (roughly 100 E. coli
proteins contain such a motif). In eukaryotes,
Getting Past Polyproline Pauses
BIOCHEMISTRY
Allen R. Buskirk
1
and Rachel Green
2

A protein that associates with the ribosome
alleviates stalling and stimulates translation
of proline-rich motifs.
No pausing at prolines. (Left) A
ribosome stalls during translation
when a peptidyl-tRNA ending in
two proline residues is in the P site
of the ribosome and a tRNA carry-
ing a proline residue has docked
into the A site. (Right) The univer-
sal translation factor EF-P relieves
stalling through interactions with
the catalytic center of the ribo-
some. After peptide bond forma-
tion, EF-P dissociates and protein
synthesis resumes.
C
R
E
D
I
T
:

A
D
A
P
T
E
D

F
R
O
M

D
.

W
I
L
S
O
N
1
Department of Chemistry and Biochemistry, Brigham
Young University, Provo, UT 84602, USA.
2
Howard Hughes
Medical Institute, Department of Molecular Biology and
Genetics, Johns Hopkins University School of Medicine, Bal-
timore, MD 21205, USA. E-mail: ragreen@jhmi.edu
Polyproline-
stalled ribosome
EF-P stimulates peptide-
bond formation
mRNA
tRNA
Growing
polypeptide
chain
Proline
EF-P
Ribosome
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 39
PERSPECTIVES
Entropies of Adsorbed Molecules
Exceed Expectations
CHEMISTRY
Jason F. Weaver
Unexpectedly large entropies of adsorbed
molecules, predicted from their gas-phase
values, lead to better models of the rates
of surface reactions.
the orthologous protein, eukaryotic trans-
lation initiation factor 5A (eIF5A), likely
plays a similar function ( 13). Like EF-P, it is
modified at a Lys residue, but with a unique
hypusine moiety derived from spermidine
( 14). eIF5A (and the enzymes that modify it)
are essential in eukaryotes, perhaps because
eukaryotes have many proteins rich in poly-
proline stretches. Differences in the modify-
ing enzymes for EF-P and eIF5A suggest that
this pathway may be a viable target for future
antimicrobials.
This model of EF-P function raises addi-
tional questions. Mechanistically, it is unclear
how polyproline stretches stall ribosomes and
how EF-P alleviates stalling. Even though
modification of Lys
34
is essential for EF-P
activity, 70% of bacterial species lack homo-
logs of YjeA and YjeK ( 15). It is also unknown
how EF-P and its modifying enzymes are reg-
ulated. Perhaps their expression is coupled
to regulation of specific genes in the cell,
controlling the copy number of target pro-
teins in response to various cellular inputs.
More broadly, the findings of Doerfel et al.
and Ude et al. call to mind other motifs that
trigger translational stalling (e.g., positively
charged amino acids, structured mRNA, and
rare codons), raising the possibility that the
cell has evolved special mechanisms for deal-
ing with them as well. The ribosome profiling
technique ( 16) will undoubtedly be instru-
mental in sorting out these stalling events and
the mechanisms for their resolution.
References
1. K. Ito, S. Chiba, K. Pogliano, Biochem. Biophys. Res.
Commun. 393, 1 (2010).
2. K. Yanagitani et al., Science 331, 586 (2011).
3. L. K. Doerfel et al., Science 339, 85 (2013);
10.1126/science.1229017.
4. S. Ude et al., Science 339, 82 (2013);
10.1126/science.1228985.
5. B. R. Glick, M. C. Ganoza, Proc. Natl. Acad. Sci. U.S.A.
72, 4257 (1975).
6. D. R. Tanner et al., J. Biol. Chem. 284, 34809 (2009).
7. I. Wohlgemuth et al., J. Biol. Chem. 283, 32229 (2008).
8. M. Y. Pavlov et al., Proc. Natl. Acad. Sci. US.A. 106, 50
(2009).
9. T. Yanagisawa, T. Sumida, R. Ishii, C. Takemoto, S.
Yokoyama, Nat. Struct. Mol. Biol. 17, 1136 (2010).
10. W. W. Navarre et al., Mol. Cell 39, 209 (2010).
11. K. Hanawa-Suetsugu et al., Proc. Natl. Acad. Sci. U.S.A.
101, 9595 (2004).
12. G. Blaha, R. E. Stanley, T. A. Steitz, Science 325, 966
(2009).
13. P. Saini, D. E. Eyler, R. Green, T. E. Dever, Nature 459,
118 (2009).
14. M. H. Park et al., Amino Acids 38, 491 (2010).
15. M. Bailly, V. de Crécy-Lagard, Biol. Direct 5, 3 (2010).
16. N. T. Ingolia et al., Science 324, 218 (2009).
T
he ability to accurately predict the
rates of chemical reactions at surfaces
is essential to improving technologi-
cal applications that rely on molecule-surface
interactions, such as designing new catalysts
for use in chemical synthesis and power-gen-
eration applications. Such accurate predic-
tions depend on knowledge of both the entro-
pies and enthalpies of the reacting species.
Although advances in molecular modeling
have increased the accuracy for the enthal-
pies of adsorbed molecules, reliable meth-
ods for calculating the entropies of adsorbed
molecules have been lacking; even estimat-
ing their magnitudes has remained elusive.
Campbell and Sellers ( 1) now show that the
entropies of adsorbed molecules on crystal-
line surfaces are surprisingly large and lin-
early correlate with the entropies of gaseous
molecules. Their finding provides a simple
yet reliable way to calculate the entropies
and affords new insights into the motions of
adsorbed molecules. An immediate impact of
the Campbell-Sellers correlation is that it will
greatly improve the accuracy of calculated
desorption rates of adsorbed molecules.
Any description of a surface reaction
needs to consider how molecules bind to
the surface, and a common way to probe the
kinetics of this process is thermal desorp-
tion—heating the surface in vacuum and
monitoring the desorption rate. The rate coef-
ficient for desorption k is typically described
by an Arrhenius relation k = A exp(–E/RT),
where A is the kinetic prefactor, E is the acti-
vation energy for reaction, R is the gas con-
stant, and T is absolute temperature. The pre-
factor scales exponentially with the differ-
ence in entropy between the initial reactant
molecule and a transition structure that the
molecule adopts along the reaction path.
Adsorption is often a nonactivated pro-
cess; that is, the molecule encounters no
potential barrier as it moves toward the sur-
face. Thus, the transition structure for desorp-
tion may be taken as the gaseous molecule
moving away from the surface. To predict
a desorption prefactor for this typical case,
both the entropy of the gas-phase molecule
and that of the adsorbed molecule are needed.
The former is readily calculated, but estimat-
ing the entropy of an adsorbed molecule is
difficult because the center-of-mass motions
of the molecule within the potential field of
the surface are not well understood at tem-
peratures where the desorption rate becomes
appreciable.
Simple models can provide bounds on
the values of the entropies of adsorbed mol-
ecules. A lower bound on the entropy is
obtained by assuming that the adsorbed mol-
ecule remains localized in a potential well
and that the molecular motions parallel to
the surface are highly restricted. The upper
bound is given by the so-called ideal two-
dimensional gas model, in which the mole-
cule-surface potential is flat within the sur-
face plane—the molecule can freely translate
and rotate across the surface. Unfortunately,
the limiting values of the desorption prefac-
Department of Chemical Engineering, University of Florida,
Gainesville, FL 32611, USA. E-mail: weaver@che.ufl.edu
Getting ready to leave. The illustration depicts an
adsorbed methane molecule moving across a typical
potential energy landscape encountered at the sur-
face of a crystalline solid. The Campbell-Sellers cor-
relation reveals that the molecule readily traverses
the in-plane barriers to motion at temperatures
where desorption becomes important, and thereby
executes nearly free translational motion (black
arrows) and rotational motion (red arrow) within the
surface plane.
10.1126/science.1233338
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 40
PERSPECTIVES
N
ine in 10 internet users in the United
States turn to search engines to find
information ( 1), and 60% of the U.S.
public seeking information about specific sci-
entific issues lists the Internet as their primary
source of information ( 2). This has created a
new urgency for scientists to pay attention to
these trends and to the emerging scholarly lit-
erature about communicating science in this
brave new “online” world.
Among the U.S. public, time spent on the
World Wide Web has been linked to more
positive attitudes toward science, even when
controlling for use of traditional mass media
such as newspapers and television ( 3). For
instance, frequent Web users are more likely
to report in surveys that they support basic sci-
entific research even if it may not have imme-
diate societal benefits. Research suggests
that the availability of science news from
the Internet may inform U.S. audiences with
different educational backgrounds. In other
words, online science sources may be help-
ing to narrow knowledge gaps caused partly
by science coverage in traditional media that
tends to be tailored to highly educated audi-
ences ( 4). Unfortunately, equivalent data for
other countries is not available yet.
Recent communication research, how-
ever, has also identified at least three areas
in which the new realities of an online infor-
mation environment will increasingly force
scientists and social scientists to rethink the
interface between the science community
and the public. One area is science journal-
ism. The rise of online media since the late
1990s has come at the expense of traditional
mass print and broadcast media. Less space
has been allocated for scientific issues, even
to the complete elimination of science cover-
age in some outlets ( 5). Today, audiences turn
more and more to blogs and other online-only
media sources for information about specific
scientific issues and much less to online ver-
sions of traditional news outlets. Almost half
of Americans currently rely on nontraditional
online sources, and only 12% turn to science
news from online content provided by tradi-
tional print newspapers and magazines ( 2).
Another area of concern is the trend among
online information providers to select and
prioritize content by using algorithms and/or
audience metrics, such as how often an online
Science, New Media,
and the Public
SOCIAL SCIENCE
Dominique Brossard and Dietram A. Scheufele
A better understanding is needed about
how the online environment affects the
communication of science information
to the public.
Department of Life Sciences Communication, University of
Wisconsin–Madison, 1545 Observatory Drive, Madison, WI
53706–1215, USA. E-mail: dbrossard@wisc.edu
tor generally span a very wide range that also
grows with increasing molecular size and
temperature. For example, the limiting mod-
els predict prefactors that differ by a factor of
about 250 for methane desorption near 65 K,
with this difference increasing to nearly 5 ×
10
4
for n-butane desorption at 170 K ( 2).
Without a reliable estimate of the entro-
pic contribution, the calculated desorption
rate can have an uncertainty of several orders
of magnitude. Previously, the available data
were insufficient to establish a trend in the
entropies of adsorbed molecules. To address
this issue, Campbell and Sellers calculated
the entropies of numerous adsorbed mol-
ecules by evaluating reported data obtained
from measurements of equilibrium adsorp-
tion isotherms and thermal desorption rates.
They found that the standard entropies of the
adsorbed molecules S
ad
0
are linearly corre-
lated with the standard entropies of the gas-
eous molecules S
g
0
, and that the relation S
ad
0

= 0.70S
g
0
– 3.3R accurately fits a large set of
data, spanning entropy values over a range of
~50R. The data set includes different classes
of molecules and surfaces, such as n-alkanes,
methanol, and several small molecules
adsorbed on magnesium oxide, the close-
packed (111) crystal face of platinum, and
graphite surfaces. The desorption prefac-
tors computed from the Campbell-Sellers
equation reduce the maximum error to 50.
The large slope of the Campbell-Sellers
relation reveals that the entropies of adsorbed
molecules are quite high and approach those
of two-dimensional gas molecules. The
implication is that the adsorbed molecules
move nearly freely within the surface plane at
temperatures where desorption first becomes
important. The Campbell-Sellers relation
establishes that, in general, adsorbed mol-
ecules readily overcome in-plane barriers to
motion once the molecules acquire nearly
enough energy to surmount the larger energy
barrier for desorption (see the figure).
Although this finding is physically reason-
able and perhaps obvious in retrospect, Camp-
bell and Sellers have actually demonstrated
and quantified the large entropies of adsorbed
molecules. Thus, modeling the in-plane, cen-
ter-of-mass motions as localized vibrations
substantially overestimates desorption prefac-
tors for many adsorbed molecules. This idea
was recently confirmed for the dissociation
of molecular propane on a palladium oxide
surface ( 3) and has also been demonstrated
by molecular dynamics simulations of alkane
desorption ( 4). Remarkably, the Campbell-
Sellers equation also applies to molecules
present in adsorbate islands, indicating that
even densely packed molecules have high
entropies near the onset of desorption.
Theoretical work may be able to provide
a general framework from which to calcu-
late entropies on the basis of molecule and
surface properties. The Campbell-Sellers
equation will provide essential guidance
to such efforts, as it reveals that the mole-
cule-surface potential continues to slightly
hinder the motions of adsorbed molecules
even as desorption becomes important.
Key challenges in theoretical modeling
will be to develop methods that properly
and efficiently describe the weakly hin-
dered motions of various types of molecules
adsorbed on different surfaces. Although
more challenging, this situation is analogous
to determining the torsional states available
to many polyatomic molecules. The findings
of Campbell and Sellers indeed represent
an important advance in the understanding
and quantification of the entropies of a wide
range of adsorbed molecules, and may well
be broadly applicable.
References
1. C. T. Campbell, J. R. V. Sellers, J. Am. Chem. Soc. 134,
18109 (2012).
2. S. L. Tait et al., J. Chem. Phys. 125, 234308 (2006).
3. A. Antony, A. Asthagiri, J. F. Weaver, Phys. Chem. Chem.
Phys. 14, 12202 (2012).
4. K. A. Fichthorn, R. A. Miron, Phys. Rev. Lett. 89, 196103
(2002).
10.1126/science.1231552
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 41
PERSPECTIVES
story is “clicked on” (viewed), for-
warded to others via email, or posted
on social media. Researchers are only
beginning to understand how the
nearly 5 billion Web searches through
the search engine Google everyday
can shape the way we make sense of
all the new information we encoun-
ter ( 6– 8). Some of that work shows
that there are often clear discrepan-
cies between what people search for
online, which specific areas are sug-
gested to them by search engines,
and what people ultimately find. As
a result, someone’s initial question
about a scientific topic, the search
results offered by a search engine, and
the algorithms that a search provider
uses to tailor retrieved content to a
search may all be linked in a self-rein-
forcing informational spiral in which
search queries and the resulting Web
traffic drive algorithms and vice versa
( 7). This raises an interesting paradox when it
comes to relatively new scientific topics, such
as nanotechnology, that are still unfamiliar to
many people: Is the World Wide Web opening
up a new world of easily accessible scientific
information to lay audiences with just a few
clicks? Or are we moving toward an online sci-
ence communication environment in which
knowledge gain and opinion formation are
increasingly shaped by how search engines
present results, direct traffic, and ultimately
narrow our informational choices? Critical
discussions about these developments have
mostly been restricted to the political arena,
with a focus on how people search and find
information in electoral settings ( 8). There is
a real urgency for the scientific community to
pay closer (empirical) attention to these new
challenges for communicating science.
How society debates emerging technolo-
gies is also dramatically changing because of
the social nature of Web 2.0—a second gen-
eration of the Web that allows people to pro-
duce and debate information online. Social
networks—both online and offline—play an
important role in shaping how information
and influence spreads among citizens ( 9).
But online social networks and social media,
in particular, may also have more latent and
potentially more powerful indirect effects. In
the current media environment, for example,
science stories usually are not presented in
isolation but instead are embedded in a host
of cues about their accuracy, importance,
or popularity. These cues that accompany
nearly all online news stories include viewer
“tweets”—very short messages posted
through the Twitter microblogging service—
about a topic in the news crawl on televi-
sion; reader comments on blog posts; or the
number of “likes” on Facebook, the massive
social media networking Web site. Such cues
may add meaning beyond what the author of
the original story intended to convey.
A recent conference presented an exami-
nation of the effects of these unintended influ-
ences of Web 2.0 environments empirically
by manipulating only the tone of the com-
ments (civil or uncivil) that followed an online
science news story in a national survey exper-
iment ( 10). All participants were exposed to
the same, balanced news item (covering nan-
otechnology as an emerging technology) and
to a set of comments following the story that
were consistent in terms of content but dif-
fered in tone. Disturbingly, readers’ interpre-
tations of potential risks associated with the
technology described in the news article dif-
fered significantly depending only on the tone
of the manipulated reader comments posted
with the story. Exposure to uncivil comments
(which included name calling and other non–
content-specific expressions of incivility)
polarized the views among proponents and
opponents of the technology with respect to
its potential risks. In other words, just the tone
of the comments following balanced science
stories in Web 2.0 environments can signifi-
cantly alter how audiences think about the
technology itself.
Online environments are providing audi-
ences with great opportunities to connect
with science, but social scientists are only
beginning to understand the nature of these
connections and their potential outcomes on
how audiences all make sense of complex
scientific issues. Moreover, new tools of data
collection and analysis (often captured under
the “big data” label) have created opportu-
nities for interdisciplinary collaborations
among computer science, computational lin-
guistics, and other social sciences. These col-
laborations will hopefully help in the analy-
sis of audience-media interactions in Web
2.0 environments in more generalizable real-
world settings.
A world in which one in seven people
actively use Facebook ( 11), and more than
340 million tweets are being posted everyday
( 12) is not the future of science communica-
tion any more. It is today’s reality. Scientists
and social scientists must explore outcomes
of online interactions about science in much
greater detail. This work will have to be based
on rigorous empirical social science rather
than guesswork and anecdotal evidence about
how to communicate complex and sometimes
controversial science in these new informa-
tion environments. Without applied research
on how to best communicate science online,
we risk creating a future where the dynam-
ics of online communication systems have
a stronger impact on public views about sci-
ence than the specific research that we as sci-
entists are trying to communicate.
References and Notes
1. K. Purcell, J. Brenner, L. Rainie, Search Engine Use 2012
(Pew Internet & American Life Project, Washington, DC,
2012).
2. National Science Board, Science and Engineering Indica-
tors 2012 (National Science Foundation, Washington,
DC, 2012).
3. A. D. Dudo et al., Communic. Res. 38, 754 (2011).
4. M. A. Cacciatore, D. A. Scheufele, E. A. Corley, Public
Underst. Sci.; 10.1177/0963662512447606 (2012).
5. C. Russell, in Science and the Media, D. Kennedy,
G. Overholser, Eds. (American Academy of Arts and
Sciences, Cambridge, MA, 2010), pp. 13–43.
6. E. Segev, A. Baram-Tsabari, Public Underst. Sci. 21, 813
(2012).
7. P. Ladwig et al., Mater. Today 13, 52 (2010).
8. E. Pariser, The Filter Bubble: How the New Personalized
Web Is Changing What We Read and How We Think
(Penguin, New York, 2012).
9. S. Aral, D. Walker, Science 337, 337 (2012).
10. P. Ladwig et al., abstract, The Annual Conference of
the Association for Education in Journalism and Mass
Communication, St. Louis, MO, August 2011;
www.aejmc.com/home/2011/06/ctec-2011-abstracts.
11. Facebook Inc., Key Facts (2012); http://newsroom.fb.com/
Key-Facts.
12. Twitter Inc. Twitter Turns Six (2012); http://blog.twitter.
com/2012/03/twitter-turns-six.html.
Acknowledgments: This material is based on work supported
by grants from NSF to Arizona State University (grant no. SES-
0531194) and the University of Wisconsin–Madison (grant no.
SES-DMR-0832760). Any opinions, findings, and conclusions
or recommendations expressed in this material are those of the
authors and do not necessarily reflect the views of NSF.
10.1126/science.1232329 P
H
O
T
O

C
R
E
D
I
T
:

I
S
T
O
C
K
P
H
O
T
O
.
C
O
M
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 42
PERSPECTIVES
U
ltracold quantum gases present an
exquisitely tunable quantum sys-
tem. Applications include preci-
sion measurement ( 1), quantum simulations
for advanced materials design ( 2), and new
regimes of chemistry ( 3). Typically trapped
in a combination of magnetic fields and laser
beams, strongly isolated from the environ-
ment in an ultrahigh vacuum, and cooled to
temperatures less than a microdegree above
absolute zero, they are the coldest known
material in the universe. The interactions
between atoms in the gas can be tuned over
seven orders of magnitude and from repul-
sive to attractive ( 4). The addition of stand-
ing waves made from interfering lasers at
optical wavelengths gives rise to an optical
lattice, a crystal of light, periodic just like
the usual crystals made of matter. On page
52 of this issue, Braun et al. ( 5) use these
special features of ultracold quantum gases
to produce a thermodynamic oddity—nega-
tive temperature.
Temperature is casually associated with
hot and cold. How can something be “colder”
than absolute zero? The answer lies in a more
precise notion of temperature. Temperature is
a single-parameter curve fit to a probability
distribution. Given a large number of parti-
cles, we can say each of them has a probabil-
ity to have some energy, P(E). Most will be in
low-energy states and a few in higher-energy
states. This probability distribution can be fit
very well with an exponential falling away to
zero. Of course, the actual distribution may
be very noisy, but an exponential fit is still
a good approximation (see the figure, panel
A). Negative temperature means most parti-
cles are in a high-energy state, with a few in a
low-energy state, so that the exponential rises
instead of falls (see the figure, panel E).
To create negative temperature, Braun et
al. had to produce an upper bound in energy,
so particles could pile up in high-energy rather
than low-energy states. In their experiment,
there are three important kinds of energy:
kinetic energy, or the energy of motion in the
optical lattice; potential energy, due to mag-
netic fields trapping the gas; and interaction
energy, due to interactions between the atoms
in their gas. The lattice naturally gives an
upper bound to kinetic energy via the forma-
tion of a band gap, a sort of energetic barrier
to higher-energy states. The potential energy
was made negative by the clever use of an
anti-trap on top of the lattice, taking the shape
of an upside-down parabola. Finally, the
interactions were tuned to be attractive (neg-
ative). Thus, all three energies had an upper
bound and, in principle, the atoms could pile
up in high-energy states.
Braun et al. convinced their gas to undergo
such a strange inversion using a quantum
phase transition ( 6), an extension of the well-
known thermodynamic concept of phase tran-
sitions to a regime in which the temperature
is so low that it plays no role in the change
of phase. In this case, they worked with two
phases, superfluid and Mott insulator. In a
superfluid, the gas flows freely without vis-
cosity and is coherent, like a laser, but made
of matter instead of light. In a Mott insula-
tor, the atoms freeze into a regular pattern and
become incompressible, similar to a solid.
Braun et al. first make their atoms repulsive
in a superfluid phase. They tune them to a
Mott insulating phase by simply turning up
the intensity of the optical lattice lasers, mak-
ing the lattice deeper. Then they tune the inte-
reactions to be attractive and at the same time
turn their trap upside down to be an anti-trap.
Finally, they melt the Mott insulator to obtain
an attractive superfluid. These anti-traps have
been used before, to create a self-propagat-
ing pulse of atoms that does not disperse (a
bright soliton) from attractive gases in one
dimension ( 7, 8). However, attractive quan-
tum gases in two and three dimensions can
implode, rather spectacularly ( 9). This ten-
dency to implode is called negative pressure.
The negative temperature is precisely what
stabilizes the gas against negative pressure
and implosion; the Mott insulator serves as
a bridge state between positive temperature
and pressure, and negative temperature and
pressure (see the figure, panels B to D).
Braun et al.’s exploration of negative
Negative Temperatures?
PHYSICS
Lincoln D. Carr
A cloud of potassium atoms is tuned
to negative temperatures via a quantum
phase transition.
Positive temperature
A E
B C D
Negative temperature
Repulsive superfluid Mott insulator Attractive superfluid
P
r
o
b
a
b
i
l
i
t
y

P
(
E
)
P
r
o
b
a
b
i
l
i
t
y

P
(
E
)
Energy E Energy E
Schematic experimental distribution
Low energy, positive pressure High energy, negative pressure
High energy
Quantum phase transition
Superfluid amplitude
Atoms Superfluid amplitude
envelope
Trap flipped to anti-trap, interactions
from repulsive to attractive
Optical lattice
Anti-trap
Low energy
Trap
Low energy, incompressible
One-parameter thermal fit
k
B
T > 0 k
B
T < 0
Low energy E
min High energy E
max
Less than zero. (A) Temperature is a one-parameter fit: As the energy gets large, the probability that an atom
will have that energy falls away exponentially. A quantum phase transition from a repulsive superfluid (B) to
a Mott insulator (C) provides a bridge to an attractive superfluid (D), resulting in negative pressure balanced
by negative temperature (E).
Department of Physics, Colorado School of Mines, Boulder,
CO 80401, USA. E-mail: lcarr@mines.edu
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 43
PERSPECTIVES
Bioinspired Oxidation Catalysts
CHEMISTRY
Martine Largeron and Maurice-Bernard Fleury
Metalloenzyme-like catalytic systems
oxidize amines to imines under
environmentally friendly conditions.
temperature is part of a general theme of
pushing the limits of thermodynamics and
quantum mechanics with ultracold quan-
tum gases. Quantum phase transitions have
been explored intimately in these systems
( 2, 10). There is a special class of dynami-
cal systems, called integrable, that never
truly develop a temperature because their
properties are different from those of a sys-
tem in equilibrium with a thermal environ-
ment. Classically, integrability is opposed
to chaos; chaotic dynamical systems ther-
malize and become thermodynamic. The
borderline between integrability and chaos
is described by a famous and beautiful
theory, called Kolmogorov-Arnold-Moser
(KAM) theory ( 11). To date, we do not
know whether there is a KAM theory for
quantum mechanics. It is now believed that
a whole new concept is needed to deal with
near-integrable quantum systems, casually
called prethermalization, in which physical
quantities after relaxation are described by
the fancy name “generalized Gibbs ensem-
ble” ( 12– 15).
Thermodynamics is at the heart of chem-
istry, engineering, and many biological ques-
tions. In ultracold quantum gases, the basic
concepts of thermodynamics, positive or neg-
ative temperature, or whether a temperature
concept is even relevant, are under intense
and profound exploration.
References
1. J. K. Stockton, K. Takase, M. A. Kasevich, Phys. Rev. Lett.
107, 133001 (2011).
2. M. Lewenstein et al., Adv. Phys. 56, 243 (2007).
3. L. D. Carr, D. Demille, R. Krems, J. Ye, New J. Phys. 11,
055049 (2009).
4. S. E. Pollack et al., Phys. Rev. Lett. 102, 090402 (2009).
5. S. Braun et al., Science 339, 52 (2013).
6. L. D. Carr, Ed., Understanding Quantum Phase Transi-
tions (Taylor and Francis, New York, 2010).
7. L. D. Carr, Y. Castin, Phys. Rev. A 66, 063602 (2002).
8. L. Khaykovich et al., Science 296, 1290 (2002).
9. E. A. Donley et al., Nature 412, 295 (2001).
10. M. Greiner, O. Mandel, T. Esslinger, T. W. Hänsch, I. Bloch,
Nature 415, 39 (2002).
11. M. Tabor, Chaos and Integrability in Nonlinear Dynamics:
An Introduction (Wiley, New York, 1989).
12. M. Rigol, V. Dunjko, V. Yurovsky, M. Olshanii, Phys. Rev.
Lett. 98, 050405 (2007).
13. A. C. Cassidy, C. W. Clark, M. Rigol, Phys. Rev. Lett. 106,
140405 (2011).
14. P. Calabrese, F. H. Essler, M. Fagotti, Phys. Rev. Lett. 106,
227203 (2011).
15. M. A. Cazalilla, A. Iucci, M. C. Chung, Phys. Rev. E. 85,
011133 (2012).
10.1126/science.1232558
I
mines are key intermediates in the synthe-
sis of fine chemicals and numerous bio-
logically active compounds. They have
traditionally been prepared through conden-
sation of amines with carbonyl compounds,
but the latter are extremely active and thus
difficult to handle. A powerful alternative
strategy involves coupling primary alcohols
and amines through catalytic alcohol activa-
tion by temporary oxidation to an aldehyde
( 1). However, with few exceptions ( 2), these
aerobic oxidative reactions require high reac-
tion temperatures and catalysts that contain
expensive and rare metals. Furthermore, this
approach is challenging because imines can
readily undergo hydrogenation ( 3). Recently
developed metalloenzyme-like catalytic sys-
tems allow the aerobic oxidation of amines to
imines under very mild conditions. They are
environmentally friendly because they avoid
the use of oxidants, energy-consuming pro-
cessing steps, and undesirable reaction media.
Efficient catalytic methods exist for the
oxidation of secondary amines (R
1
CH
2
NHR
2
)
to imines (R
1
CH=NR
2
) ( 4), but until recently,
little attention was given to the oxidation
of primary amines (RCH
2
NH
2
), probably
because the generated imines (RCH=NH, in
which a second α-amino hydrogen is avail-
able) are generally intermediate products
that are rapidly dehydrogenated to nitriles
(RC≡N) ( 5). Green processes have also been
developed that use biocompatible transition-
metal catalysts, with dioxygen or air as the
sole oxidant. However, most of these methods
have limitations. For example, a solvent-free
copper-catalyzed synthesis of imines from
primary amines uses air as a benign oxidant
but requires high reaction temperatures ( 6). A
simple copper/TEMPO (2,2,6,6-tetramethyl-
1-piperidinyloxyl) system catalyzes the aer-
obic oxidation of amines to imines at room
temperature, but is efficient only for benzylic
amines ( 7).
Naturally occurring metalloenzymes have
long been recognized as attractive catalysts
for aerobic oxidations because they can oper-
ate under mild conditions with complete che-
Faculté des Sciences Pharmaceutiques et Biologiques, UMR
8638 CNRS–Université Paris Descartes, 75270 Paris Cedex
06, France. E-mail: martine.largeron@parisdescartes.fr;
maurice.fleury@parisdescartes.fr
H
2
N H
2
N R
O O
O
O
O
O
HN
O
O
O
HN
HO
HO
OH
NH
3
R R R N O
H
2
O
2
Organic cofactor
Cu metal cofactor
H
2
O, O
2
CH
3
CN, O
2
Cu metal co-catalyst
CH
3
OH, air/O
2
Immobilized Pt/Ir nanoclusters
CHCI
3
:H
2
O (9:1), O
2
1
Copper amine oxidases Biomimetic catalytic systems
2
3
R
Biomimetic success. (Left) Copper amine oxidase enzymes catalyze the formation of aldehydes from amines.
(Right) Catalyst systems developed to mimic these natural enzymes enable the aerobic oxidation of amines
to imines under mild conditions.
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 44
PERSPECTIVES
moselectivity. Simulation of the function of
these enzymes has led to the discovery of
many biomimetic oxidation catalysts ( 8, 9).
Among metalloenzymes, copper amine
oxidases (CuAOs) couple the oxidation of pri-
mary amines to aldehydes with the reduction
of dioxygen to hydrogen peroxide through
the synergistic action of the topaquinone
(TPQ) organic cofactor and copper (see the
figure, left) ( 10). Quinone models that mimic
the active site of copper amine oxidases have
yielded important insights into the mecha-
nism by which these enzymes operate ( 11).
These insights have been put to use in the
development of biomimetic catalytic systems
for the aerobic oxidation of primary amines
(see the figure, right). Building on biochemi-
cal model studies, Wendlandt and Stahl have
shown that the TPQ analog 4-tert-butyl-2-
hydroxybenzoquinone 1 is an efficient biomi-
metic organocatalyst for the chemoselective
oxidation of primary amines to imines ( 12).
They obtained imines in high yields (80 to
95%) from diverse benzylic amines at room
temperature under 1 atm of molecular oxygen.
This quinone model failed to oxidize aliphatic
primary amines, but the exclusive selectivity
for primary benzylic amines allowed selective
formation of heterocoupled imines.
Copper amine oxidases have also inspired
a biomimetic homogeneous catalytic system
for the aerobic oxidation of primary amines
to imines, based on the synergistic combi-
nation of copper and an o-iminoquinone
organocatalyst 2 first discovered from elec-
trochemical investigations ( 13). Low catalyst
loadings (2 mol% of 2 and 0.2 mol% of Cu)
are sufficient to activate the α-C−H bond of
benzylic and aliphatic amines, which were
converted to alkylated imines under ambient
conditions ( 14).
Yuan et al. have reported a heterogeneous
catalytic system for the aerobic oxidation of
amines that also shares some characteristics
with copper amine oxidases ( 15). Their coop-
erative catalytic system consists of a hetero-
geneous Pt/Ir bimetallic nanocluster catalyst
immobilized on a styrene-based copolymer
support and 4-tert-butyl-o-quinone 3 as the
redox-active organic cofactor. The two cata-
lysts act together to reduce the energy of the
transition state to a degree that neither cata-
lyst can accomplish alone. With this system,
the authors were able to dehydrogenate pri-
mary benzylic amines to imines in high iso-
lated yields at room temperature under 1 atm
of molecular oxygen. The catalytic system
can also convert secondary amines to imines,
and the heterogeneous catalyst could be
recovered easily and reused up to five times
without loss of activity (81 to 87% for first to
fifth uses).
Many challenges remain, including the
development of biomimetic catalytic systems
that operate effectively at room temperature
with ambient air rather than pure molecular
oxygen. The development of recyclable het-
erogeneous nanocluster catalysts that contain
biocompatible rather than rare and precious
metals would also be welcome. From a more
general viewpoint, mimicking the function
of amine oxidase enzymes would provide
environmentally friendly organic synthesis
because air is the cheaper and less polluting
stoichiometric oxidant. Although challeng-
ing, this strategy would allow the chemistry
of C –H bond activation to be extended to the
functionalization of the α-position of primary
amines under mild conditions.
References and Notes
1. B. Gnanaprakasam, J. Zhang, D. Milstein, Angew. Chem.
Int. Ed. 49, 1468 (2010).
2. H. Tian, X. Yu, Q. Li, J. Wang, Q. Xu, Adv. Synth. Catal.
354, 2671 (2012).
3. A. J. A. Watson, J. M. J. Williams, Science 329, 635 (2010).
4. J. S. M. Samec, A. H. Ell, J.-E. Bäckvall, Chem. Eur. J. 11,
2327 (2005).
5. K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 42,
1480 (2003).
6. R. D. Patil, S. Adimurthy, Adv. Synth. Catal. 353, 1695
(2011).
7. Z. Hu, F. M. Kerton, Org. Biomol. Chem. 10, 1618 (2012).
8. L. Que, Jr., W. B. Tolman, Nature 455, 333 (2008).
9. S.-I. Murahashi, Angew. Chem. Int. Ed. Engl. 34, 2443
(1995).
10. M. Mure, S. A. Mills, J. P. Klinman, Biochemistry 41, 9269
(2002).
11. M. Mure, Acc. Chem. Res. 37, 131 (2004).
12. A. E. Wendlandt, S. S. Stahl, Org. Lett. 14, 2850 (2012).
13. M. Largeron, A. Chiaroni, M.-B. Fleury, Chem. Eur. J. 14,
996 (2008).
14. M. Largeron, M.-B. Fleury, Angew. Chem. Int. Ed. 51,
5409 (2012).
15. H. Yuan, W.-J. Yoo, H. Miyamura, S. Kobayashi, J. Am.
Chem. Soc. 134, 13970 (2012).
Acknowledgment: We thank CNRS for financial support.
10.1126/science.1232220
How Was Early Earth Kept Warm?
ATMOSPHERIC SCIENCE
James F. Kasting
Greenhouse warming by molecular hydrogen
may have contributed to warming the surface
of the early Earth.
W
hy do some gases cause green-
house warming, whereas others
do not? H
2
O is a greenhouse gas
because it has a permanent electric dipole
moment (a charge separation within the
molecule) that allows it to interact strongly
with electromagnetic radiation. CO
2
also
has an electric dipole moment, but it has to
bend or stretch asymmetrically to create it,
because, unlike H
2
O, it is a linear molecule.
N
2
and O
2
are not normally considered to be
greenhouse gases, because these symmetric,
diatomic molecules have no electric dipole
moment and cannot bend or stretch to create
one. But as Wordsworth and Pierrehumbert
show on page 64 of this issue ( 1), N
2
and
molecular hydrogen (H
2
) can be greenhouse
gases under the right conditions; H
2
may
have been important for Earth’s Archean cli-
mate (before 2.5 billion years ago).
Researchers who study the outer planets
have long known that N
2
and H
2
can con-
tribute to the greenhouse effect. They do
so through collision-induced absorption,
whereby collisions between molecules allow
them to absorb radiation that excites them to
higher rotational states. The physics is com-
plex ( 2), but the implications for the green-
house effect are straightforward if one con-
siders that it is the rotational states of H
2
that
are being excited.
Going through the math ( 3) shows why
H
2
can be an effective greenhouse gas on
Earth and why N
2
or O
2
cannot. For all
three gases, the absorption is spread over
a range of wavelengths corresponding to
the populated j states. But an N atom is 14
times as heavy as an H atom, and the dis-
tance between the atoms is also greater in N
2

than in H
2
; hence, the moment of inertia of
N
2
is much larger, and its rotational energy
levels are more closely spaced. As a result,
the rotational constant B

is ~2 cm
–1
for N
2

and ~60 cm
–1
for H
2
( 2). This means that N
2

absorbs only at wavenumbers (1/λ) <500
cm
–1
, where the energy radiated from Earth’s
surface is relatively small and where H
2
O
already absorbs strongly [see Fig. 1A in ( 1)].
What matters most for Earth’s climate
is the region from 830 to 1250 cm
–1
, where
H
2
O and CO
2
both absorb poorly. Colli-
Geosciences Department, Pennsylvania State University,
University Park, PA 16802, USA. E-mail: jfk4@psu.edu
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 45
PERSPECTIVES
sions between molecules could, in principle,
excite N
2
or O
2
to high enough rotational
states to absorb in this region. However,
to absorb at ~1000 cm
–1
, N
2
would need to
be in the j = 250 rotational state, where it
would be spinning so fast that it would fall
apart. In contrast, H
2
, with its higher rota-
tional constant, only needs to be in the j =
7 state to absorb at 1000 cm
–1
. On Saturn’s
moon Titan, where H
2
absorption has been
well studied ( 4), the surface temperature is
about –180°C, so only the lowest rotational
energy levels of H
2
are filled and absorption
is confined to long wavelengths. But at room
temperature, these rotational levels are pop-
ulated, and H
2
does absorb at 1000 cm
–1
.
Wordsworth and Pierrehumbert now
show that the presence of 10% H
2
in Earth’s
Archean atmosphere could conceivably
have increased surface temperatures by as
much as 10° to 15°C. The warming is stron-
gest if N
2
was also more abundant in the
early atmosphere, as other researchers have
postulated. But was H
2
really as important
for Earth’s early climate as they suggest?
Well, yes and no. The Sun was ~25% less
bright at that time, and Earth’s surface was
not frozen, so some warming mechanism
was clearly needed. The warming could
have been provided entirely by higher CO
2

concentrations, which are predicted to result
from feedbacks in the carbonate-silicate
cycle ( 5). But some researchers have argued
that atmospheric CO
2
concentrations in the
Archean were low, on the basis of sugges-
tions of high rates of seafloor weathering ( 6)
or on constraints from geochemical indica-
tors such as paleosols ( 7, 8) or banded iron
formations (BIFs) ( 9). The BIF constraints
assume unlimited availability of reductants
in the Archean ocean, which was probably
not the case ( 10). The paleosol constraints
are more meaningful, but the error bars in
these analyses may be larger than the authors
of ( 7, 8) acknowledge. Furthermore, addi-
tional warming mechanisms, including CH
4

( 11), pressure broadening of CO
2
and H
2
O
by high N
2
concentrations ( 12), and albedo
feedbacks ( 9) appear capable of making up
for any shortfall in warming by CO
2
(see
the figure).
Most important, high H
2
levels are
unlikely to have existed for long once life
evolved, because methanogens are pre-
dicted to have consumed it efficiently ( 13).
Words worth and Pierrehumbert suggest that
H
2
consumption by methanogens may have
been limited by nutrient availability, but this
seems unlikely because primary productiv-
ity in these purely anaerobic biospheres is
predicted to have been two to three orders of
magnitude lower than today ( 13).
Nevertheless, the realization that H
2

can warm terrestrial planet climates could
be important for the prebiotic Earth, early
Mars, and young Earth-like exoplanets.
Large amounts of H
2
in a planet’s atmo-
sphere could allow it to remain habitable out
to as far as 10 AU around a star like the Sun
(1 AU = 1 astronomical unit = mean Earth-
Sun distance) ( 14). These numbers apply to a
hypothetical planet with three times Earth’s
mass and a 40-bar H
2
atmosphere—a type of
object that may or may not exist. But rocky
planets with atmospheres composed of N
2
,
CO
2
, and H
2
might still remain warm at dis-
tances well beyond the traditional, ~1.7-AU
boundary of the habitable zone. The evolu-
tion of organisms like methanogens could
conceivably destabilize the climates of such
planets. But we will, nevertheless, need to
keep this H
2
greenhouse warming mecha-
nism in mind as we decipher our own solar
system’s history and search for other habit-
able planets like our own.
References and Notes
1. R. Wordsworth, R. Pierrehumbert, Science 339, 64
(2013).
2. A. Borysow, L. Frommhold, Astrophys. J. 303, 495
(1986).
3. In the simplest approximation, a molecule like N
2
or H
2

can be treated as a rigid rotator, for which the energy
levels are spaced according to E
i
= Bj(j + 1) where B = h
h
2
/2I is the rotational constant, j is the quantum number,
I = µR
2
is the moment of inertia of the molecule, h is
Planck’s constant h/2π, R is the intermolecular distance,
and µ is the reduced mass m
1
m
2
/(m
1
+ m
2
) = m/2, where
m is the mass of an individual N or H atom. The energy
difference between successive levels is ∆ E = E
j+1
– E
j
=
2B(j + 1). Given that j only changes by one unit when a
photon is absorbed, the frequency ν or wavelength λ at
which the absorption takes place can be calculated from
∆E = hν = hc/λ where c is the speed of light. Spectros-
copists often express this energy difference in terms of
reciprocal wavelength (1/λ) using the modified rotational
constant B
_
= B/hc.
4. C. P. McKay, J. B. Pollack, R. Courtin, Science 253, 1118
(1991).
5. J. F. Kasting, Precambrian Res. 34, 205 (1987).
6. N. H. Sleep, K. Zahnle, J. Geophys. Res. 106, 1373
(2001).
7. N. D. Sheldon, Precambrian Res. 147, 148 (2006).
8. S. G. Driese et al., Precambrian Res. 189, 1 (2011).
9. M. T. Rosing, D. K. Bird, N. H. Sleep, C. J. Bjerrum, Nature
464, 744 (2010).
10. N. Dauphas, J. F. Kasting, Nature 474, E1 (2011).
11. J. D. Haqq-Misra, S. D. Domagal-Goldman, P. J. Kasting,
J. F. Kasting, Astrobiol. 8, 1127 (2008).
12. C. Goldblatt et al., Nat. Geosci. 2, 891 (2009).
13. P. Kharecha, J. Kasting, J. Siefert, Geobiol. 3, 53 (2005).
14. R. Pierrehumbert, E. Gaidos, Astrophys. J. 734, L13
(2011).
10.1126/science.1232662
300
290
280
270
260
10
–4
10
–3
10
–2
10
–1
pCO
2
(bar)
15°C
0°C
Paleosol constraints
Higher H
2
(?)
Higher N
2
(?)
CH
4
Albedo changes (?)
fCH
4
= 1000 ppmv
fCH
4
= 0
S
u
r
f
a
c
e

t
e
m
p
e
r
a
t
u
r
e

(
K
)
Warming up the planet in the Archean. Around 2.8 billion years ago, the solar luminosity was ~0.8 times
that of today ( 11), yet Earth’s surface was not frozen. If the atmosphere contained not only CO
2
but also 1000
parts per million by volume (ppmv) methane—a plausible amount for the Archean ( 13)—then this meth-
ane would have produced a 12°C warming (vertical arrow). A further 4° to 5°C of warming could have been
caused by pressure broadening of CO
2
and H
2
O lines by N
2
concentrations that were two to three times as
high as today’s ( 12). Albedo changes may have led to a further warming ( 9), bringing mean surface tempera-
tures up to today’s value of ~15°C. Alternatively, as suggested by Wordsworth and Pierrehumbert, 10% H
2
,
in combination with higher N
2
concentrations, could have kept the mean surface temperature above freezing,
even if CO
2
levels were similar to today (300 ppmv). Shaded area: estimated range of atmospheric CO
2
con-
centrations from a 2.7-billion-year-old paleosol ( 8). fCH
4
, methane mole fraction. [Figure adapted from ( 11)]
Published by AAAS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

A Stringent Limit on a Drifting
Proton-to-Electron Mass Ratio
from Alcohol in the Early Universe
Julija Bagdonaite,
1
Paul Jansen,
1
Christian Henkel,
2,3
Hendrick L. Bethlem,
1
Karl M. Menten,
2
Wim Ubachs
1
*
The standard model of physics is built on the fundamental constants of nature, but it does not
provide an explanation for their values, nor require their constancy over space and time.
Here we set a limit on a possible cosmological variation of the proton-to-electron mass ratio m
by comparing transitions in methanol observed in the early universe with those measured in
the laboratory. From radio-astronomical observations of PKS1830-211, we deduced a constraint
of ∆m/m = (0.0 T 1.0) × 10
−7
at redshift z = 0.89, corresponding to a look-back time of 7 billion
years. This is consistent with a null result.
T
he standard model of particle physics, the
theory describing symmetries and forces
of nature at the deepest level, does not
provide an intrinsic explanation for the values of
the fundamental coupling constants, nor does it
prohibit that the fundamental constants depend
on time and space. In contrast, Einstein’s equiv-
alence principle, a basic assumption of general
relativity, assumes that the laws of nature, and
hence the fundamental constants are independent
of a local reference system. Some cosmological
scenarios aimed at explaining the fine-tuning
between fundamental constants sketch an evolv-
ing mechanism, where minimally varying con-
stants are crucial for reaching the present state
of complexity in the universe (1). Theoretical ap-
proaches involving additional scalar fields have
imposed bounds on varying constants through
tests of the weak equivalence principle (2). In the
past decade the search for small variations of
dimensionless fundamental constants over cosmo-
logical time scales has become an active exper-
imental endeavor, in particular because accurate
measurements of spectral lines of atoms at high
redshift have provided indication for a possible
variation of the fine structure constant a, either
temporally (3, 4) or spatially (5, 6).
A second dimensionless fundamental constant
m, representing the proton-to-electron mass ratio
m
p
/m
e
, probes the cosmological evolution of the
nuclear versus the electroweak sector in the stan-
dard model. A search for a possible drift of m has
been made operational by comparing observa-
tions of spectral lines of the hydrogen molecule
(H
2
) in distant galaxies with accurate laboratory
measurements (7). These investigations, based on
observations with the world’s largest optical
telescopes, have yielded a limit at the level of
∆m/m < 10
−5
for look-back times of 12 billion
years (8, 9).
Inversion transitions of ammonia (NH
3
) were
found to be ~100 times more sensitive to m-variation
than H
2
transitions (10, 11). Astronomical obser-
vations of NH
3
, in the microwave or radio range
of the electromagnetic spectrum, led to stringent
1s constraints at the level of (1.0 T 4.7) × 10
−7
(12) and (–3.5 T 1.2) × 10
−7
(13). This has shifted
the paradigm for probing m-variation from optical
to radio astronomy. Here we use the extreme
sensitivity of methanol (CH
3
OH) (14, 15) to probe
the variation of the proton-to-electron mass ratio
m over cosmic time.
Methanol (Fig. 1A) is the simplest alcohol
and consists of a hydroxyl group attached to a
methyl group. The C-O bond is flexible, allow-
ing the hydroxyl group to rotate with respect to
the methyl group. This so-called internal rotation
is strongly hindered by the repulsion between the
hydrogen atoms of the different groups, result-
ing in a threefold barrier (Fig. 1B). If the barrier
were infinitely high, the levels in the torsional
well would be degenerate. Quantum mechanical
tunneling through the barriers lifts this degen-
eracy, resulting in three levels that are labeled
according to symmetry: A, E1, and E2 (16). Be-
cause the symmetry of the nuclear wave function
is preserved in radiative transitions as well as
in (nonreactive) collisions, the A and E levels
of methanol can be regarded as belonging to
two separate chemical species
The sensitivity coefficient, K
m
, of a transition
with frequency n is defined by ∆n/n = K
m
× ∆m/m.
The frequencies of pure rotational transitions, such
as the transitions indicated by the red and orange
arrow in Fig. 1C, are inversely proportional to
the reduced mass of methanol and hence to
the proton-to-electron mass ratio. Consequently,
these have sensitivity coefficients equal to –1.
The frequencies of pure torsional transitions—
which are not allowed in methanol—depend ex-
ponentially on the reduced moments of inertia of
REPORTS
1
Department of Physics and Astronomy, VU University Amster-
dam, De Boelelaan 1081, 1081 HV Amsterdam, Netherlands.
2
Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69,
53121 Bonn, Germany.
3
Astronomy Department, King Abdulaziz
University, Post Office Box 80203, Jeddah, Saudi Arabia.
*To whom correspondence should be addressed. E-mail:
w.m.g.ubachs@vu.nl
Fig. 1. (A) Pictorial rep-
resentation of the metha-
nol molecule. (B) Potential
energy as a function of the
dihedral angle between
the OH group and one of
the CH bonds in the methyl
group. V
3
denotes the bar-
rier height. The horizontal
lines represent the energy
for the levels in the torsion-
vibrational ground state,
n
t
= 0, and first excited
state, n
t
= 1. (C) Energy
level structure of the torsion-
rotation ground state of
methanol. Each level is
labeled according to its
torsional symmetry Ts, to-
tal angular momentum J,
and its projection K on
the molecule-fixed axis.
The energy-level structure
of methanol resembles that
of a prolate symmetric top,
with the difference that each K manifold is offset depending on its torsional symmetry. Levels of A-symmetry
with |K| > 0 are split by the slight asymmetry of the molecule. Levels of E1 and E2 symmetry in the K = 0
manifold cannot be distinguished and are labeled as E. The four transitions observed in this study are
indicated by the four arrows. The transitions indicated by the red and orange arrows are pure rotational
transitions and have a sensitivity coefficient of –1. The transitions indicated by the blue and green arrows
are mixed torsion-rotation transitions and have sensitivity coefficients of –32.8 and –7.4, respectively.
π/3 2π/3 π 4π/3 5π/3 2π
Torsional angle (rad)
300
400 400
E
n
e
r
g
y

(
c
m
-
1
)
ν
t
= 0

V
3
ν
t
= 1
200
100
0
0
130
140
J Ts
0 A
A
A
E2
1
2
1
E2 2
0 E
A 3
E 1
A
±
1
E2 3
2 E
A
±
2
A 4 E1 1
J
K = 0
|K| = 1
+
+
+
+
+
Ts
E
n
e
r
g
y

(
c
m
-
1
)
135
145
C
A
B
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 46

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

the methyl and hydroxyl groups and have a
sensitivity coefficient of –2.5. The sensitivity
of mixed transitions—i.e., transitions in which
both the internal and overall rotation state is
changed—is strongly enhanced. The sensitivity
coefficients for different transitions in methanol
range from –42 to +53. These enhancements
occur generally in every internal rotor molecule,
but because of a number of favorable prop-
erties, the effect is exceptionally large in meth-
anol (17).
Methanol is abundantly present in the uni-
verse, and more than a 1000 lines have been
recorded in our galaxy (18). So far, searches for
methanol absorption in far-distant galaxies have
yielded detection only in the gravitational lens
system PKS1830-211 (19). A limit on ∆m/m (20)
has been previously derived on the basis of two
methanol lines. We present a comprehensive
study of radio astronomical observations of four
methanol lines in PKS1830-211, including the
two previously observed, with improved signal-
to-noise ratio.
The background source of this system, PKS1830-
211, is a high redshift (z = 2.507) blazar, which is
radio loud and time variable and is viewed as two
spotlike features and an Einstein ring, which
result from gravitational lensing by the interven-
ing face-on spiral galaxy (21, 22). The redshift of
the main molecular absorptions from the galaxy
is z = 0.88582 (19, 23), corresponding to a look-
back time of 7.0 billion years, or half the age of
the universe (24). More than 30 different molec-
ular species were detected in the lensing galaxy
of PKS1830-211 (19). Molecular absorption is
mostly detected toward one of the two blazar
images (the southwestern), whereas the other
image (the northeastern) shows weaker and
fewer molecular lines at a slightly different red-
shift but stronger neutral hydrogen absorption
[e.g., (19, 25)].
The CH
3
OH lines were recorded with the
100-m single-dish Effelsberg radio telescope,
using the 5-, 1.3-, and 1-cm receivers. Preliminary
detections were performed during the course of
2011, and subsequently systematic observations
were performed in a narrow time slot. The data
were registered onto a local standard of rest ve-
locity scale, which was centered at z = 0.88582.
The two blazar images and the Einstein ring are
unresolved and PKS1830-211 is effectively treated
as a point source, which is an assumption under-
lying the present study.
The recorded spectra are shown in Fig. 2.
For a single transition, the spectra taken on var-
ious days were averaged together, weighting the
individual scans by their integration time. The
lines were calibrated by the total continuum so
that their strength is expressed as line-to-continuum
flux density ratio. The profiles, devoid of under-
lying structure, were fitted as a single Gaussian
(Table 1). The accuracy of the position measure-
ments is at the level of 1 to 4% of the line width.
The velocities between different transitions are
interrelated via
V/c = –K
m
∆m/m
where c is the speed of light, and ∆m/m rep-
resents the deviation from the current laboratory
value of m, defined so that a positive sign indi-
cates a larger m in the high-redshift–absorbing
Fig. 2. Methanol absorp-
tionlines onalocal standard
of rest (LSR) velocity scale
relative to z = 0.88582,
observed toward PKS1830-
211. The transitions and
their approximate observed
frequencies are indicated
in each panel. The color-
ing of corresponding tran-
sitions matches that of
Fig. 1. The top spectrum
in each panel is a time-
weighted average of the
individual spectra, which
are displayed below the
combined one. For each
spectrum, the position of
a fitted Gaussian (depicted
as light green curves) is
shown in the graph at
the right. Residuals are
shown at the top of each
combined spectrum with
dashed lines indicating
T1s offsets. In the case
of ~25-GHz observations,
two proximate methanol
transitions were recorded.
They are separated by
27.494 km/s; the fitted
positions of the weaker
line are corrected to bring
the measurements on a
common scale. The lines
are calibrated by the to-
tal continuum. The black
square (upper panel) and
the black diamond (lower panel) represent the single line observations from Ellingsen et al. (20) and
Muller et al. (19), respectively. The line positions, originally reported on a heliocentric velocity scale, were
transformed to the LSR scale via V
LSR
– V
HEL
= 12.432 km/s.
K
µ
= -32.8
0
-1
0
-2.5
0
-0.1
L
i
n
e
-
t
o
-
C
o
n
t
i
n
u
u
m

R
a
t
i
o

[
%
]
-120 -60 0 60 120 6 8 10 12
Relative Velocity [km s
-1
]
07-03-2012
06-03-2012
Combined
K
µ
= -7.4 Residuals
08-12-2011
05-04-2012
06-04-2012
10-12-2011
09-12-2011
Combined
Residuals
K
µ
= -1 Residuals
Combined
28-02-2012
25-02-2012
23-02-2012
2
-1
1
0
E
0
0
1
0
E, 0
0
1
0
A
+
3
-1
2
0
E
~6 GHz
~25 GHz
~32 GHz
Table 1. A summary of the relevant parameters and results. Laboratory data: lower and upper energy
level quantum numbers; laboratory frequencies, n, of the four relevant methanol absorption lines; and
their uncertainties, fractional uncertainties, and uncertainties in terms of Doppler shift, ∆v
D
, in km/s.
Calculations: the sensitivity coefficients, K
m
. Observations: the measured local standard of rest velocities of
the lines (relative to z = 0.88582) and the line widths with their 1s uncertainties. Assuming a molecular
hydrogen density of 2 × 10
3
cm
−3
and a kinetic temperature of 80 K (12, 29), full width at half maximum
linewidths as fitted to observations and a T
CMB
= 2.728(1 + z) = 5.145 K for the temperature of the
cosmic microwave background (CMB) radiation at z = 0.88582, the optical depths t yield a total column
density of 2.0 × 10
14
cm
−2
from a large velocity gradient radiative transfer model (30).
Laboratory data Calc. Observations
Line
JW
KW
–J'
K'
Ts
n
(GHz)
∆n/n
∆v
D
(km/s)
Ref. K
m
Position
(km/s)
Width
(km/s)
t
3
-1
–2
0
E 12.178597 (4) 3 × 10
−7
0.1 (31) –32.8 9.06 T 0.67 16.4 T 1.4 0.0024
0
0
–1
0
A
+
48.3724558 (7) 2 × 10
−8
0.006 (32) –1 8.40 T 0.10 10.8 T 0.2 0.045
0
0
–1
0
E 48.376892 (10) 2 × 10
−7
0.06 (33) –1 9.12 T 0.30 14.6 T 0.6 0.016
2
-1
–1
0
E 60.531489 (10) 2 × 10
−7
0.06 (33) –7.4 9.83 T 0.43 17.0 T 0.9 0.028
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 47
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

galaxy (i.e., ∆m = m
z
– m
lab
). Therefore, to de-
termine the fractional change in m, the peak po-
sitions of the four transitions are plotted (in V/c)
versus K
m
, and a (dashed) line is fitted to the
data (Fig. 3). Because the A and E levels of
methanol can be regarded as belonging to two
separate species, the data were analyzed in two
different ways: First, only the three transitions
from E levels were fitted, then the A transition
was added to the sample. The analysis of the E
transitions results in ∆m/m = (–0.1 T 7.6) × 10
−8
,
which is consistent with a nonvarying m at the
level of 1.5×10
−7
(95% confidence level). The
reduced chi-squared, c
n
2
, which is a measure of
the quality of the fit, is ~2.0 (26). The fit on all
four transitions has a much larger c
n
2
of 6.4,
which might be attributed to segregation is-
sues (see below), and it delivers ∆m/m = (11.0 T
6.8) × 10
−8
.
The upper limit derived here is statistically
more constraining than that derived in previous
tests in the radio-domain (11–13, 19, 20). More-
over, compared to the methods used in previous
studies, the methanol method is more robust
against systematic effects. In particular, it is much
less sensitive to the assumption that all absorbing
species reside in the same physical location and
hence are at the same redshift. Spatial segregation
of different absorbers may mimic or hide a var-
iation of m. This is the limiting systematic error
for tests based on the comparison between
different molecular species, such as the compar-
ison of ammonia with various rotational lines in
HCO
+
, HCN, CS, and so forth (11–13). The mo-
lecular survey in PKS1830-211 suggests that seg-
regation effects are prominent among different
species (19). For instance, a single methanol line
was found to be displaced from the average ab-
sorption velocity by more than 3 km/s (19). Our
test is based exclusively on a single molecular
species. However, as discussed above, the E and
A type methanol should be considered as dif-
ferent species and thus may undergo spatial seg-
regation effects. In the combined spectrum,
the 0
0
–1
0
A
+
and 0
0
–1
0
E transitions, falling in
close proximity in a single scan of the receiver,
are separated by 0.72 T 0.32 km/s. Moreover,
the linewidths of the E lines are markedly
larger than that of the A line (Table 1). Because
this is suggestive of a spatial segregation of the E
and A symmetry methanol molecules, we adopt
a fiducial limit on ∆m/m from the fit of only E
transitions.
Another source of systematic error is the
known variability of the lensed object PKS1830-
211. The absorption strength of radio lines was
found to vary strongly, by a factor of >6 in a
time span of 3 years, and this was ascribed to
the intensity changes in the background con-
tinuum source (27). This phenomenon might
cause a varying alignment through parts of the
absorbing spiral and therewith absorption
through varying Doppler components over time.
Hence this variability may affect the derivation
of a m-constraint from radio-observations. For
this reason we adopted a measurement strategy
to explicitly address the source variability issue.
Spectra of the anchor lines (the middle panel in
Fig. 2) were recorded in December 2011 and
April 2012, whereas the two strongly shifting
lines have been observed in-between in Febru-
ary and March 2012 (28). The strong (0
0
–1
0
A
+
)
line in the combined spectrum from December
2011 is positioned at 8.32 T 0.10 km/s, and at
8.80 T 0.24 km/s in the spectrum from April
2012. The difference between them is 0.48 T
0.26 km/s, possibly indicative of a small sys-
tematic shift due to variability. We have assessed
this possible systematic effect as caused by time
variability in two models (26) and have chosen
the one producing the largest uncertainty (∆m/m
of 7.0 × 10
−8
) to give a conservative estimate.
Thus, we obtain a limit on varying m to be ∆m/m =
(–0.1 T 7.6
stat
T 7.0
sys
) × 10
−8
or, if the statistical
and systematic uncertainties are added in quadra-
ture, a limit of ∆m/m = (0.0 T 1.0) × 10
−7
.
References and Notes
1. L. Smolin, Physica A 340, 705 (2004).
2. J. D. Barrow, J. Magueijo, Phys. Rev. D Part. Fields
Gravit. Cosmol. 72, 043521 (2005).
3. J. K. Webb, V. V. Flambaum, C. W. Churchill, M. J. Drinkwater,
J. D. Barrow, Phys. Rev. Lett. 82, 884 (1999).
4. M. T. Murphy, J. K. Webb, V. V. Flambaum, Mon. Not. R.
Astron. Soc. 345, 609 (2003).
5. J. K. Webb et al., Phys. Rev. Lett. 107, 191101 (2011).
6. J. A. King et al., Mon. Not. R. Astron. Soc. 422, 3370
(2012).
7. E. Reinhold et al., Phys. Rev. Lett. 96, 151101 (2006).
8. A. Malec et al., Mon. Not. R. Astron. Soc. 403, 1541
(2010).
9. F. van Weerdenburg, M. T. Murphy, A. L. Malec, L. Kaper,
W. Ubachs, Phys. Rev. Lett. 106, 180802 (2011).
10. J. van Veldhoven et al., Eur. Phys. J. D 31, 337 (2004).
11. M. T. Murphy, V. V. Flambaum, S. Muller, C. Henkel,
Science 320, 1611 (2008).
12. C. Henkel et al., Astron. Astrophys. 500, 725 (2009).
13. N. Kanekar, Astrophys. J. Lett. 728, L12 (2011).
14. P. Jansen, L.-H. Xu, I. Kleiner, W. Ubachs, H. L. Bethlem,
Phys. Rev. Lett. 106, 100801 (2011).
15. S. A. Levshakov, M. G. Kozlov, D. Reimers, Astrophys. J.
738, 26 (2011).
16. C. C. Lin, J. D. Swalen, Rev. Mod. Phys. 31, 841 (1959).
17. P. Jansen, I. Kleiner, L.-H. Xu, W. Ubachs, H. L. Bethlem,
Phys. Rev. A 84, 062505 (2011).
18. F. J. Lovas, J. Phys. Chem. Ref. Data 33, 177 (2004).
19. S. Muller et al., Astron. Astrophys. 535, A103 (2011).
20. S. P. Ellingsen, M. A. Voronkov, S. L. Breen, J. E. J. Lovell,
Astrophys. J. Lett. 747, L7 (2012).
21. D. L. Jauncey et al., Nature 352, 132 (1991).
22. C. Lidman et al., Astrophys. J. 514, L57 (1999).
23. T. Wiklind, F. Combes, Nature 379, 139 (1996).
24. Adopting a standard L-cosmology with H
0
= 73 km s
−1
Mpc
−1
, W
m
= 0.28, W
L
= 0.72.
25. J. N. Chengalur, A. G. de Bruyn, D. Narasimha, Astron.
Astrophys. 343, L79 (1999).
26. See supplementary materials on Science Online.
27. S. Muller, M. Guélin, Astron. Astrophys. 491, 739 (2008).
28. Each line required a change of receiver, for which
reason the time intervals could not be shorter.
29. K. M. Menten et al., Astron. Astrophys. 492, 725
(2008).
30. S. Leurini et al., Astron. Astrophys. 422, 573 (2004).
31. S. M. Breckenridge, S. G. Kukolich, Astrophys. J. 438,
504 (1995).
32. J. E. M. Heuvel, A. Dymanus, J. Mol. Spectrosc. 45, 282
(1973).
33. H. S. P. Müller, K. M. Menten, H. Mäder, Astron. Astrophys.
428, 1019 (2004).
Acknowledgments: This work is supported by the Foundation for
Fundamental Research on Matter program “Broken Mirrors &
Drifting Constants.” H.L.B. acknowledges support from the
Netherlands Organization for Scientific Research via a VIDI
grant and by the European Research Council via a Starting Grant.
We thank the staff of the Effelsberg radio telescope
for their hospitality and support. The raw data of the
radioastronomical observations are available upon request
from the Max Planck Institute for Radio Astronomy at Bonn
(kmenten@mpifr-bonn.mpg.de).
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1224898/DC1
Supplementary Text
Table S1
Reference (34)
18 May 2012; accepted 16 November 2012
Published online 13 December 2012;
10.1126/science.1224898
Fig. 3. The positions of
the four observed meth-
anol lines (represented
by V/c with respect to z =
0.88582) are plotted ver-
sus their sensitivity coef-
ficients, K
m
. The bold blue
horizontal line represents
the fiducial result of a fit
to the E-type lines, where-
as the dashed line repre-
sents a fit to all four lines.
A positive slope of the fitted
line implies that m had a
smaller value in the early
universe than is measured
in the laboratory. The blue-
shaded surface is a den-
sity plot of simulated data
points from the blue fitted
line and reflects the con-
fidence bands of the fit.
Color-coding of the data points is the same as in Figs. 1 and 2.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 48
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Alignment of Magnetized Accretion
Disks and Relativistic Jets with
Spinning Black Holes
Jonathan C. McKinney,
1,2
* Alexander Tchekhovskoy,
3
Roger D. Blandford
1
Accreting black holes (BHs) produce intense radiation and powerful relativistic jets, which are
affected by the BH’s spin magnitude and direction. Although thin disks might align with the
BH spin axis via the Bardeen-Petterson effect, this does not apply to jet systems with thick
disks. We used fully three-dimensional general relativistic magnetohydrodynamical simulations
to study accreting BHs with various spin vectors and disk thicknesses and with magnetic flux
reaching saturation. Our simulations reveal a “magneto-spin alignment” mechanism that causes
magnetized disks and jets to align with the BH spin near BHs and to reorient with the outer
disk farther away. This mechanism has implications for the evolution of BH mass and spin, BH
feedback on host galaxies, and resolved BH images for the accreting BHs in SgrA* and M87.
A
strophysical black holes (BHs) operate
as engines that convert the gravitational
binding energy of accreting plasmas into
intense radiation (1) and release BH spin energy
(2–4) into powerful relativistic jets (5, 6). Rel-
ativistic jets from accreting BHs are commonly
observed to emerge from active galactic nuclei
(AGN) or quasars, x-ray binaries that behave as
microquasars, and gamma-ray burst (GRB)
events. GRB jets allow one to probe the earliest
epochs of star formation, whereas radiation and
jets from AGN play a direct dynamical role via
feedback that suppresses star formation in their
host galaxies (7).
BHs are also intrinsically interesting because
they act as laboratories for probing Einstein’s
general relativity theory and for testing theories
about accreting BHs and jets. Astrophysical BHs
are characterized primarily by their mass (M) and
dimensionless spin angular momentum ( j). BHs
have been measured to have masses of tens to
billions of solar masses M

; the mass of the BH
in M87 is ~6 × 10
9
M

(8). Spins have also been
measured and span over many of the possible
values, including near the maximal value of j ~ 1
in the BH x-ray binary GRS1915+105 with
M ~ 14 M

and in the AGN MCG-6-30-15 with
M ~ 3 × 10
6
M

(9, 10). Structures on a few BH
event horizon length scales have recently been
resolved by Earth-sized radio telescope interfer-
ometry for SgrA* (11, 12) and M87 (13, 14).
Although the BH’s present angular momen-
tum axis is set by the history of plasma accretion
and mergers with other BHs, the gas being cur-
rently supplied (at mass accretion rate M
:
) to the
BH can have an arbitrarily different angular mo-
mentum axis. This relative tilt, given by q
tilt
for
the angle between the BHspin axis and the disk’s
angular momentum axis at large distances, influ-
ences the intensity of the radiation via changes in
the gravitational potential felt by the plasma; it
also influences what an observer at different
viewing angles sees as a result of disk warping
and jet bending.
One mechanism known to possibly affect the
orientation of a disk or jet is the Bardeen-Petterson
(BP) effect (15–18), where Lense-Thirring (LT)
forces induced by the BHframe-dragging cause a
misaligned disk to precess and warp until a local
viscosity aligns a very thin disk out to some dis-
tance [estimated to be out to r ~ 10r
g
to 10
5
r
g
,
where r
g
is a gravitational radius (19), depending
on assumptions] from the BH. Whereas the vis-
cosity has been thought to result from turbulence
driven by the magneto-rotational instability (MRI)
that amplifies weak small-scale (<~H, the disk
height) magnetic fields (20), magnetohydrody-
namical (MHD) simulations of weakly magnet-
ized disks have not yet produced any BPalignment
effect (21). The BP effect and LT precession re-
main commonly invoked mechanisms to under-
stand how tilt affects the evolution of BH mass
and spin (22, 23), how merging BHs are affected
by any nearby plasma (24), and how disks and
their jets are oriented (25, 26).
Large-scale electromagnetic (EM) fields might
also affect the jet’s and disk’s orientations via ex-
ternal confinement forces (27, 28). Estimates based
on the presumption that large-scale magnetic fields
are weaker than turbulent disk fields suggested
that EM forces are insufficient to align the disk
with the BH(27) or the BHwith the disk (29, 30).
Simulations without disks have given ambiguous
results for the EM jet direction. For a uniform ver-
tical magnetic field and no disk, the jet is directed
along the magnetic field direction rather than along
the BH’s tilted spin axis (31), whereas isolated mag-
netic threads tend to align with the BH spin axis
when there is no disk to restrict their motion (32).
We have used fully three-dimensional (3D)
general relativistic (GR) MHD simulations (33)
1
Kavli Institute for Particle Astrophysics and Cosmology, Stanford
University, Stanford, CA 94309, USA.
2
Department of Physics
and Joint Space-Science Institute, University of Maryland, Col-
lege Park, MD 20742, USA.
3
Center for Theoretical Science,
Princeton University, Princeton, NJ 08544, USA.
*To whom correspondence should be addressed. E-mail:
jcm@umd.edu
Table 1. Tilted black hole disk-jet systems. The simulation models are listed by model name, which
identifies the approximate BH spin of j (values following “A”), something about the magnetic field choices
(values following “B” and “N”) (28), and the initial relative tilt (q
tilt,0
in radians) between the BH spin axis
and the disk rotation axis (values following “T”; if T is absent, q
tilt,0
= 0). Successive columns give the
dimensionless BH spin ( j); the evolved quasi–steady-state value of the disk height/radius ratio H/R
between r ~ 20r
g
and r ~ 30r
g
; the initial relative tilt (q
tilt,0
) between the disk + jet (having the same tilt
initially) and the BH spin axis; the evolved relative tilt between the BH spin axis and the disk and jet,
respectively, at r = 4r
g
; and the same measurements at r = 30r
g
. If the relative tilt is 0.0, the disk or jet
remained aligned with the BH spin axis; if the tilt is equal to the initial tilt, the disk or jet was unaffected.
Model name BH j Disk H/R Initial tilt
Disk tilt
r = 4r
g
Jet tilt
r = 4r
g
Disk tilt
r = 30r
g
Jet tilt
r = 30r
g
A0.94BfN40 0.9375 0.6 0 0 0 0 0
A0.94BfN40T0.35 0.9375 0.6 0.35 0.0 0.0 0.2 0.2
A0.94BfN40T0.7 0.9375 0.6 0.70 0.0 0.0 0.4 0.3
A0.94BfN40T1.5708 0.9375 0.6 1.5708 0.0 0.1 0.5 0.7
A-0.9N100 –0.9 0.3 0 0 0 0 0
A-0.9N100T0.15 –0.9 0.3 0.15 0.1 0.1 0.1 0.2
A-0.9N100T0.3 –0.9 0.3 0.30 0.2 0.2 0.2 0.2
A-0.9N100T0.6 –0.9 0.3 0.60 0.2 0.3 0.4 0.3
A-0.9N100T1.5708 –0.9 0.3 1.5708 0.2 0.4 0.9 0.8
A0.9N100 0.9 0.3 0 0 0 0 0
A0.9N100T0.15 0.9 0.3 0.15 0.0 0.0 0.1 0.1
A0.9N100T0.3 0.9 0.3 0.30 0.1 0.1 0.2 0.2
A0.9N100T0.6 0.9 0.3 0.60 0.1 0.1 0.3 0.3
A0.9N100T1.5708 0.9 0.3 1.5708 0.2 0.3 0.7 0.6
A0.9N100 0.99 0.3 0 0 0 0 0
A0.99N100T0.15 0.99 0.3 0.15 0.0 0.1 0.1 0.1
A0.99N100T0.3 0.99 0.3 0.30 0.1 0.1 0.2 0.2
A0.99N100T0.6 0.99 0.3 0.60 0.1 0.1 0.3 0.4
A0.99N100T1.5708 0.99 0.3 1.5708 0.1 0.1 0.6 0.6
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 49
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

of accreting BHs to show that near the BH, both
the disk angular momentum and jet direction
reorient and align with the BH’s spin axis. Our
simulations were designed so that the magnetic
field built up to a natural saturation strength with,
roughly, the disk’s thermal + ram + gravitational
forces balancing the disk’s and jet’s magnetic
forces such that the trapped large-scale magnetic
field threading the BH and disk became strong
relative to the disk’s turbulent field. The saturated
field strength has been shown to be (i) indepen-
dent of the strength of the initial magnetic field
when the surrounding medium has a sufficient
supply of magnetic flux, (ii) weakly dependent on
BHspin, and (iii) proportionally dependent on disk
thickness (4, 28, 34). We considered various BH
spins (35), BH tilts, and disks with a quasi–
steady-state height/radius ratio (H/R) of ~0.6
for thick disks and ~0.3 for thinner disks. Nu-
merical convergence of our results was determined
on the basis of convergence quality measures for
how well the MRI and turbulence were resolved
(table S2) as well as by explicit convergence test-
ing (33).
Let us motivate these MHD simulations by
estimating whether EM forces are expected to
dominate LT forces on the rotating heavy disk.
Imagine a toy model with a flat heavy disk tilted
and pushed up against the magnetized jet gen-
erated directly by the rotating BH. For a magnetic
field B bending on scale r, the EMtorque per unit
area is t
EM
~ rB
r
B
f
/4 (33) for a jet magnetic field
that has both radial (B
r
) and toroidal (B
f
~ rB
r

F
)
components and rotates with an angular frequen-
cy of Ω
F
[where r
g

F
/c ~ j/8 for j ~ 1 (28)]. The
radial field is written in terms of a dimensionless
magnetic flux given by
U ≈
0:7F
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4pr
2
g
M
:
c
q ð1Þ
for magnetic flux F ~ 4pr
2
B
r
for B in Gaussian
units (r
g
and c reintroduced for dimensional
clarity) that is consistent with measurements in
our previous work (28). This gives
t
EM
e
r
2
g
8pr
2
M
:
W
F
U
2
ð2Þ
Meanwhile, the LTtorque per unit area is t
LT
~Ω
LT
L
with LT precession rate Ω
LT
~ 2j/r
3
, disk angular
momentum per unit area L ~ Srv
f
, and disk sur-
face density S
e
M
:
=ð2prv
r
Þ. This gives
t
LT
e
jcr
2
g
M
:
v
f
pr
3
v
r
ð3Þ
The ratio of the EMto LTtorques for j ~ 1 is then
t
EM
=t
LT
e
U
2
rv
r
=ð64r
g
v
f
Þ, with r
g
reintroduced
for dimensional clarity. Far beyond the horizon,
t
EM
t
LT
e
1
64
U
2
ra
eff
r
g
H
R

2
ð4Þ
for an effective viscosity a
eff
≡ v
r/
[(H/R)
2
v
f
].
Over the horizon and in the jet, U
e
10 for
our thinner disk models and U
e
17 for our
thick disk models (28). Also, for both thicknesses,
(r/r
g
)a
eff
~ 15 and roughly constant with radius,
and v
r
/v
f
~ 1 near the horizon (28). So, at all dis-
tances, the jet’s EM forces lead to t
EM
/t
LT
~ 2
for our thinner disk models and t
EM
/t
LT
> ~5 for
our thick disk models. Hence, we expect EMforces
to dominate LT forces for both our thinner and
thick disk models [including for small spins (33)].
EM alignment forces are effective when they
are larger than forces associated with the newly ac-
creted rotating plasma with torque per unit area of
t
acc
e
M
:
v
f
=ð2prÞ. Therefore, t
EM
=t
acc
e
U
2
W
F
=
ð4rv
f
Þ, and when these torques are equal one
obtains an implicit equation for a “magneto-spin
alignment” radius of
r
msa
e
W
F
r
2
g
U
2
4v
f
ð5Þ
(with r
g
reintroduced), within which EM forces
can torque the accreting dense material. For suf-
ficiently small values of U or j, no alignment can
occur. We obtain r
msa
> ~30r
g
for our thinner and
thick disk models that are sub-Keplerian by a
factor of 0.5 to 0.1, respectively (28), although
accurate estimates require performing more sim-
ulations or accounting for more physics that could
lead to much different r
msa
(33).
Our self-consistent fully 3D general rela-
tivistic magnetohydrodynamic (GRMHD) simu-
lations started with a disk around an untilted BH
where the BH spin axis, disk rotational axis, and
emergent jet’s direction all pointed in the vertical
(z) direction. As the simulation proceeded, the
mass and magnetic flux readily advected from
large distances onto the BH. The magnetic flux
versus radius saturated on the BH and within the
disk near the BH after magnetic forces balanced
the disk’s thermal + ram + gravitational forces.
Magnetic braking causes such disks to become
even more sub-Keplerian than weakly magnetized
thick disks (28), which means that the classical
thin disk innermost stable circular orbit position
is even less applicable than for weakly magne-
tized thick disks. The simulations were evolved
for a long time period so that the disk reached a
quasi-stationary magnetically saturated state out
to about r ~ 40r
g
(28, 33).
Then, the BH spin axis was instantly tilted by
an angle of q
tilt,0
(see Table 1 for tilts used for
different spins and disk thicknesses). The tilted
disk-jet systemunderwent a violent rearrangement
for the larger tilts. The frame-dragging forces
caused the nearly split-monopole BH magneto-
sphere to align with the BHspin axis, as expected
because the misaligned angular momentum was
radiated away as part of the electromagnetic
Fig. 1. 3D snapshot for an evolved model with j = 0.99, initial relative tilt q
tilt,0
≈ 90°, and disk thickness
H/R ~ 0.3. The rotating BH sits at the center of the box of size r = −40r
g
to r = +40r
g
in each dimension.
The snapshot shows the disk near the BH (yellow isosurface, which is mostly flat in the figure plane), the
highly magnetized jet region (blue isosurface, with magnetic energy per unit rest-mass energy equal to
about 70), the rotational axis of the disk both initially and at large distances (orange cylinder), outer disk
(green-yellow volume rendering, more aligned with disk rotational axis at large distances), magnetic field
vectors (like iron filings on that surface) for a cross section of the jet (cyan vectors), and jet magnetic field
lines (white lines) that trace from the BH out to large distances. The disk and jet near the BH are aligned
with the BH spin axis and point mostly in and out of the figure plane, whereas at larger distances the jet
points roughly halfway between the BH spin axis and the disk’s rotational axis (pointing along the orange
cylinder).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 50
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

outflowon Alfvén time scales. The magnetic torque
then caused the heavy disk to lose its misaligned
component of angular momentumand so reorient
with the BH’s rotating magnetosphere. The time
scale for alignment seems to be roughly the
Alfvén crossing time near the heavy disk and
BH. This “magneto-spin alignment” occurs be-
cause the magnetic field built up to a natural
saturation strength on the horizon where U
e
10
(depending on the thickness and tilt), which led
to the BHmagnetosphere’s forces dominating the
disk dynamics and LT forces.
All simulations (including with zero tilt) were
then further evolved in time until all the tilted
simulations reached a new quasi-stationary state
out to r ~ 40r
g
. This ensured that any differences
at later times in the disk and jet between tilted and
untilted simulations were due to the BH tilt. This
also ensured that each of the tilted and untilted
simulations reached their own quasi–steady state
values for magnetic flux near the BH, magnetic
flux in the disk, mass accretion rate, etc. We then
measured the evolved relative tilt between the
BH spin axis and the disk and jet at r = 4r
g
and
r = 30r
g
(Table 1). For all our models, the disk
and jet aligned with the BH spin axis near the
BH, whereas the disk axis and jet direction de-
viated at larger distances. Such deviations are ex-
pected because the jet interacted with circulation
with stronger mass inflows at larger distances
(33). Despite the tilts and jet deviations, the BH’s
efficiency (defined as the ratio of energy out to
energy in) was roughly 100% for j > ~0.9 (table
S1), where more tilt led to reduced efficiency due
to more spatially and temporally irregular mass
inflow.
Our most extreme case of a tilted BH accre-
tion disk and jet systemis the j = 0.99 model with
a full tilt of q
tilt,0
= 1.5708 ≈ 90° and disk thick-
ness H/R ~ 0.3. Even in this extremely tilted case,
the evolved disk and jet near the BHaligned with
the BH spin axis (Fig. 1, fig. S1, and movie S1).
The jet’s magnetic field wound around the per-
sistent relativistic jet, and the magnetic field was
well ordered even for this highly tilted case. The
jet was not symmetric around the jet axis, and
instead there was a broad wing (with opening
half-angle of about 25° by r = 40r
g
) and a narrow
wing (with opening half-angle of about 5° by r =
40r
g
). At large distances from the BH, the jet
drilled its way through the disk material and grad-
ually got pushed away from the disk (Fig. 2 and
movie S2). The magnetic field wound around with
a pitch angle of about 45° near the BH and a
smaller pitch angle at larger distances. By r ~
300r
g
, the jet had become parallel with (but offset
from) the outer disk rotational axis, and so the jet
and counterjet were also offset.
Thus, our simulations have revealed a “magneto-
spin alignment” mechanism that aligns the disk
and jet axes with the BH spin axis near the BH
once the magnetic field has saturated on the BH
and within the disk (36). Unlike the BP effect, the
mechanism actually works best for thick disks,
and so the magneto-spin alignment mechanism
should control jet systems where thick disks [due
to accretion at either very low (37) or very high
rates when H/R > ~0.5] are expected. For ex-
ample, for SgrA* and M87, if the BHs rotate
sufficiently rapidly (33), then we expect the photon
spectra, temporal behaviors, and resolved images
of their jets and disks to be affected by nonzero
relative tilts as a result of disk warping and jet
bending near the BH.
Tidal disruption flare events such as Swift
J164449.3+573451 are thought to be produced
by very high accretion rates onto BHs, which
launch fairly persistent jets that dissipate and
give the observed emission (38). Our results sug-
gest that the inner disk and inner jet are both
aligned with the BH spin axis, but the observed
jet dissipating at large distances need not point
along the BHspin axis. EMforces do not direct-
ly cause any precession (27), so the lack of LT
precession-induced variability does not alone im-
ply that the jet is necessarily driven by the BH
spin power (26). Further, quasi-periodic oscilla-
tions (39) and long-term dips seen in this sys-
tem’s light curve might be explained by oscillations
in the disk-jet magnetospheric interface (28) or
by periods of magnetic flux accumulation and re-
jection by the BH (4) (both occurring for untilted
systems) rather than by LT precession.
Jet dissipation and emission (e.g., in blazars)
might be due to the jet ramming into the disk
until the jet aligns with the disk rotational axis at
large distances. Measurements of BH spin in
AGN and x-ray binaries might be affected by
assumptions about the alignment between the
disk, BH spin, and jet (9, 10). The cosmological
evolution of BH mass and spin and AGN feed-
back for accretion at high rates might be affected
by the higher BH spin-down rates and jet effi-
ciencies relative to standard thin disk spin-down
rates and radiative efficiencies (28) and also by
how the jet aligns the disk material before LT
torques can be effective, thus possibly leading to
less change in the BH spin direction relative to
the BP effect.
References and Notes
1. N. I. Shakura, R. A. Sunyaev, Astron. Astrophys. 24, 337
(1973).
2. R. D. Blandford, R. L. Znajek, Mon. Not. R. Astron. Soc.
179, 433 (1977).
3. S. Koide, K. Shibata, T. Kudoh, D. L. Meier, Science 295,
1688 (2002).
4. A. Tchekhovskoy, R. Narayan, J. C. McKinney, Mon. Not.
R. Astron. Soc. 418, L79 (2011).
5. R. V. E. Lovelace, Nature 262, 649 (1976).
6. M. C. Begelman, M. J. Rees, R. D. Blandford, Nature 279,
770 (1979).
7. T. Di Matteo, V. Springel, L. Hernquist, Nature 433, 604
(2005).
8. K. Gebhardt et al., Astrophys. J. 729, 119 (2011).
9. C. S. Reynolds, A. C. Fabian, Astrophys. J. 675, 1048 (2008).
10. J. E. McClintock et al., http://arxiv.org/abs/0911.5408
(2009).
11. S. S. Doeleman et al., Nature 455, 78 (2008).
12. V. L. Fish et al., Astrophys. J. 727, L36 (2011).
13. K. Hada et al., Nature 477, 185 (2011).
14. S. S. Doeleman et al., Science 338, 355 (2012).
15. J. M. Bardeen, J. A. Petterson, Astrophys. J. 195, L65
(1975).
16. S. P. Hatchett, M. C. Begelman, C. L. Sarazin, Astrophys.
J. 247, 677 (1981).
Fig. 2. 3D snapshot similar to Fig. 1 but showing the outer disk (density isosurface in purple) at large
distances from the BH in a box of size r = −350r
g
to r = +350r
g
in each dimension. The jet (blue
isosurface, here with magnetic energy per unit rest-mass energy equal to ~4, still corresponding to the jet
spine) is aligned with the BHspin near the BHbut gradually gets pushed by the disk material and becomes
parallel to (but offset from) the disk rotational axis at large distances. The strong interaction between the
jet and disk has left an asymmetry or warp in the disk density at large radii.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 51
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

17. S. Kumar, J. E. Pringle, Mon. Not. R. Astron. Soc. 213,
435 (1985).
18. R. P. Nelson, J. C. B. Papaloizou, Mon. Not. R. Astron.
Soc. 315, 570 (2000).
19. We typically set GM = c = 1, where c is the speed of light,
G is the gravitational constant, and M is the mass of
the BH, so that r
g
≡ GM/c
2
= 1. For dimensional clarity,
these constants are sometimes reintroduced.
20. S. A. Balbus, J. F. Hawley, Astrophys. J. 376, 214
(1991).
21. P. C. Fragile, O. M. Blaes, P. Anninos, J. D. Salmonson,
Astrophys. J. 668, 417 (2007).
22. M. C. Begelman, R. D. Blandford, M. J. Rees, Nature 287,
307 (1980).
23. A. Perego, M. Dotti, M. Colpi, M. Volonteri, Mon. Not. R.
Astron. Soc. 399, 2249 (2009).
24. T. Bogdanović, C. S. Reynolds, M. C. Miller, Astrophys. J.
661, L147 (2007).
25. P. Natarajan, P. J. Armitage, Mon. Not. R. Astron. Soc.
309, 961 (1999).
26. N. Stone, A. Loeb, Phys. Rev. Lett. 108, 061302
(2012).
27. A. R. King, J. P. Lasota, Astron. Astrophys. 58, 175
(1977).
28. J. C. McKinney, A. Tchekhovskoy, R. D. Blandford,
Mon. Not. R. Astron. Soc. 423, 3083 (2012).
29. H. Kim, H. K. Lee, C. H. Lee, J. Cosmol. Astropart. Phys.
2003, 1 (2003).
30. A. R. King, in Magnetic Fields in the Universe: From
Laboratory and Stars to Primordial Structures,
E. M. de Gouveia dal Pino, G. Lugones, A. Lazarian,
Eds. (American Institute of Physics, Melville, NY, 2005),
pp. 175–182.
31. C. Palenzuela, T. Garrett, L. Lehner, S. L. Liebling,
Phys. Rev. D 82, 044045 (2010).
32. V. Semenov, S. Dyadechkin, B. Punsly, Science 305, 978
(2004).
33. See supplementary materials on Science Online.
34. A. Tchekhovskoy, J. C. McKinney, Mon. Not. R. Astron. Soc.
423, L55 (2012).
35. Spins of j ~ 0.9 give mid-range BH rotation rates. See
the Physical Models section in the supplement.
36. The magnetic field built up via direct magnetic
flux advection, but the buildup might also occur via
dynamo generation, as seen in our prior untilted
simulations that showed emergent large-scale
dipolar flux patches.
37. R. Narayan, I. Yi, R. Mahadevan, Nature 374, 623
(1995).
38. J. S. Bloom et al., Science 333, 203 (2011).
39. R. C. Reis et al., Science 337, 949 (2012).
Acknowledgments: J.C.M. thanks R. Narayan, J. Dexter,
and P. C. Fragile for useful discussions, and R. Kaehler at
KIPAC (SLAC/Stanford) for the artistic rendering in fig. S1.
Supported by NASA Fermi grant NNX11AO21G ( J.C.M.), a
Princeton Center for Theoretical Science fellowship (A.T.),
and NSF Extreme Science and Engineering Discovery
Environment resources provided by the Texas Advanced
Computing Center (Lonestar/Ranch) and the National
Institute for Computational Sciences (Kraken) under awards
TG-PHY120005 ( J.C.M.) and TG-AST100040 (A.T.) and
provided by NASA Advanced Supercomputing (Pleiades) for
the Fermi grant. GRMHD simulation data are contained in
Table 1 and tables S1 and S2. A.T. is a Princeton Center for
Theoretical Science Fellow.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1230811/DC1
Materials and Methods
Fig. S1
Tables S1 and S2
Movies S1 and S2
References (40–98)
27 September 2012; accepted 7 November 2012
Published online 15 November 2012;
10.1126/science.1230811
Negative Absolute Temperature for
Motional Degrees of Freedom
S. Braun,
1,2
J. P. Ronzheimer,
1,2
M. Schreiber,
1,2
S. S. Hodgman,
1,2
T. Rom,
1,2
I. Bloch,
1,2
U. Schneider
1,2
*
Absolute temperature is usually bound to be positive. Under special conditions, however,
negative temperatures—in which high-energy states are more occupied than low-energy
states—are also possible. Such states have been demonstrated in localized systems with finite,
discrete spectra. Here, we prepared a negative temperature state for motional degrees of
freedom. By tailoring the Bose-Hubbard Hamiltonian, we created an attractively interacting
ensemble of ultracold bosons at negative temperature that is stable against collapse for
arbitrary atom numbers. The quasimomentum distribution develops sharp peaks at the upper
band edge, revealing thermal equilibrium and bosonic coherence over several lattice sites.
Negative temperatures imply negative pressures and open up new parameter regimes for
cold atoms, enabling fundamentally new many-body states.
A
bsolute temperature T is one of the cen-
tral concepts of statistical mechanics and
is a measure of, for example, the amount
of disordered motion in a classical ideal gas. There-
fore, nothing can be colder than T = 0, where
classical particles would be at rest. In a thermal
state of such an ideal gas, the probability P
i
for a
particle to occupy a state i with kinetic energy E
i
is proportional to the Boltzmann factor
P
i
º e
−Ei =kBT
ð1Þ
where k
B
is Boltzmann’s constant. An ensemble
at positive temperature is described by an occu-
pation distribution that decreases exponentially
with energy. If we were to extend this formula to
negative absolute temperatures, exponentially in-
creasing distributions would result. Because the
distribution needs to be normalizable, at positive
temperatures a lower bound in energy is re-
quired, as the probabilities P
i
would diverge for
E
i
→ –∞. Negative temperatures, on the other
hand, demand an upper bound in energy (1, 2). In
daily life, negative temperatures are absent, be-
cause kinetic energy in most systems, including
particles in free space, only provides a lower en-
ergy bound. Even in lattice systems, where kinet-
ic energy is split into distinct bands, implementing
an upper energy bound for motional degrees of
freedom is challenging, because potential and in-
teraction energy need to be limited as well (3, 4).
So far, negative temperatures have been realized
in localized spin systems (5–7), where the finite,
discrete spectrum naturally provides both lower
and upper energy bounds. Here, we were able to
realize a negative temperature state for motional
degrees of freedom.
In Fig. 1A, we schematically show the rela-
tion between entropy S and energy E for a ther-
mal system possessing both lower and upper
energy bounds. Starting at minimumenergy, where
only the ground state is populated, an increase in
energy leads to an occupation of a larger number
of states and therefore an increase in entropy. As
the temperature approaches infinity, all states be-
come equally populated and the entropy reaches
its maximum possible value S
max
. However,
the energy can be increased even further if high-
energy states are more populated than low-energy
ones. In this regime, the entropy decreases with
energy, which, according to the thermodynamic
definition of temperature (8) (1/T = ∂S/∂E), re-
sults in negative temperatures. The temperature is
discontinuous at maximumentropy, jumping from
positive to negative infinity. This is a consequence
of the historic definition of temperature. A con-
tinuous and monotonically increasing tempera-
ture scale would be given by −b = −1/k
B
T, also
emphasizing that negative temperature states are
hotter than positive temperature states, i.e., in
thermal contact, heat would flow from a negative
to a positive temperature system.
Because negative temperature systems can ab-
sorb entropy while releasing energy, they give
rise to several counterintuitive effects, such as
Carnot engines with an efficiency greater than
unity (4). Through a stability analysis for thermo-
dynamic equilibrium, we showed that negative
temperature states of motional degrees of free-
domnecessarily possess negative pressure (9) and
are thus of fundamental interest to the description
of dark energy in cosmology, where negative pres-
sure is required to account for the accelerating
expansion of the universe (10).
Cold atoms in optical lattices are an ideal
system to create negative temperature states be-
cause of the isolation from the environment and
independent control of all relevant parameters
(11). Bosonic atoms in the lowest band of a
1
Fakultät für Physik, Ludwig-Maximilians-Universität München,
Schellingstraße 4, 80799 Munich, Germany.
2
Max-Planck-Institut
für Quantenoptik, Hans-Kopfermann-Straße 1, 85748 Garching,
Germany.
*To whom correspondence should be addressed. E-mail:
ulrich.schneider@lmu.de.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 52
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

sufficiently deep optical lattice are described by
the Bose-Hubbard Hamiltonian (12)
H ¼ −J ∑
〈i;j〉
b
%

i
b
%
j
þ
U
2

i
n
%
i
ðn
%
i
− 1Þ þ V∑
i
r
2
i
n
%
i
ð2Þ
Here, J is the tunneling matrix element be-
tween neighboring lattice sites 〈i, j〉, and b
%
i
and
b
%

i
are the annihilation and creation operator,
respectively, for a boson on site i, U is the on-site
interaction energy, n
%
i
¼ b
%

i
b
%
i
is the local number
operator, and V º w
2
describes the external har-
monic confinement, with r
i
denoting the posi-
tion of site i with respect to the trap center and w
the trap frequency.
In Fig. 1B, we show how lower and upper
bounds can be realized for the three terms in the
Hubbard Hamiltonian. The restriction to a single
band naturally provides lower and upper bounds
for the kinetic energy E
kin
, but the interaction
term E
int
presents a challenge: Because in prin-
ciple all bosons could occupy the same lattice
site, the interaction energy can diverge in the
thermodynamic limit. For repulsive interactions
(U > 0), the interaction energy is only bounded
from below but not from above, thereby limiting
the system to positive temperatures; in contrast,
for attractive interactions (U < 0), only an upper
bound for the interaction energy is established,
rendering positive temperature ensembles unsta-
ble. The situation is different for the Fermi-Hubbard
model, where the Pauli principle enforces an up-
per limit on the interaction energy per atom of
U/2 and thereby allows negative temperatures
even in the repulsive case (13, 14). Similarly, a
trapping potential V > 0 only provides a lower
bound for the potential energy E
pot
, whereas an
Fig. 1. Negative absolute temperature in optical lattices. (A) Sketch of entropy
as a function of energy in a canonical ensemble possessing both lower (E
min
) and
upper (E
max
) energy bounds. (Insets) Sample occupation distributions of single-
particle states for positive, infinite, and negative temperature, assuming a weakly
interacting ensemble. (B) Energy bounds of the three terms of the 2D Bose-
Hubbard Hamiltonian: kinetic (E
kin
), interaction (E
int
), and potential (E
pot
) energy.
(C) Measured momentum distributions (TOF images) for positive (left) and neg-
ative (right) temperature states. Both images are averages of about 20 shots;
both optical densities (OD) are individually scaled. The contour plots below show
the tight-binding dispersion relation; momenta with large occupation are high-
lighted. The white square in the center indicates the first Brillouin zone.
Fig. 2. Experimental sequence and TOF images. (A) Top to bottom: lattice
depth, horizontal trap frequency, and scattering length as a function of
time. Blue indicates the sequence for positive, red for negative temper-
ature of the final state. (B) TOF images of the atomic cloud at various times
t in the sequence. Blue borders indicate positive, red negative temper-
atures. The initial picture in a shallow lattice at t = 6.8 ms is taken once for
a scattering length of a = 309(5) a
0
(top) as in the sequence, and once for
a = 33(1) a
0
(bottom; OD rescaled by a factor of 0.25), comparable to the
final images. All images are averages of about 20 individual shots. See
also Fig. 1C.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 53
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

anti-trapping potential V < 0 creates an upper
bound. Therefore, stable negative temperature states
with bosons can exist only for attractive interac-
tions and an anti-trapping potential.
To bridge the transition between positive and
negative temperatures, we used the n = 1 Mott
insulator (15) close to the atomic limit (|U|/J →∞),
which can be approximated by a product of Fock
statesjY〉 ¼ ∏
i
b
%

i
j0〉. Because this state is a many-
body eigenstate in both the repulsive and the at-
tractive case, it allows us to switch between these
regimes, ideally without producing entropy. The
employed sequence (Fig. 2A) is based on a pro-
posal by Rapp et al. (4), building on previous ideas
by Mosk (3). It essentially consists of loading a
repulsively interacting Bose-Einstein condensate
(BEC) into the deep Mott insulating regime (I in
Fig. 2A), switching U and V to negative values
(II), and finally melting the Mott insulator again
by reducing |U|/J (III). For comparison, we also
created a final positive temperature state with an
analog sequence.
The experiment started with a BECof 1.1(2) ×
10
5 39
K atoms in a pure dipole trap with hori-
zontal trap frequency w
dip
(V > 0) at positive
temperature (T > 0) and a scattering length of a =
309(5) a
0
, with a
0
the Bohr radius. We ramped
up a three-dimensional (3D) optical lattice (I) with
simple cubic symmetry to a depth of V
lat
= 22(1)
E
r
. Here, E
r
¼ h
2
=ð2ml
2
lat
Þ is the recoil energy
with Planck’s constant h, the atomic mass m,
and the lattice wavelength l
lat
= 736.65 nm. The
blue-detuned optical lattice provides an overall
anti-trapping potential with a formally imaginary
horizontal trap frequency w
lat
that reduces the
confinement of the dipole trap, giving an effective
horizontal trap frequencyw
hor
¼ ðw
2
dip
þ w
2
lat
Þ
1=2
.
Once the atoms are in the deep Mott insulating
regime where tunneling can essentially be ne-
glected [tunneling time t = h/(2pJ ) = 10(2) ms],
we set the desired attractive (repulsive) interac-
tions (II) to prepare a final negative (positive)
temperature state using a Feshbach resonance
(16). Simultaneously, we decreased the horizon-
tal confinement to an overall anti-trapping (trap-
ping) potential by reducing w
dip
. Subsequently,
we lowered the horizontal lattice depths (III), yield-
ing a final value of U/J = −2.1(1) [U/J = +1.9(1)],
and probed the resulting momentum distribution
by absorption imaging after 7 ms time-of-flight
(TOF). The whole sequence was experimentally
optimized to maximize the visibility of the final
negative temperature state. We chose a 2D geom-
etry for the final state to enable strong anti-trapping
potentials and to avoid detrimental effects due to
gravity (9).
In Fig. 2B, we showTOF images of the cloud
for various times t in the sequence, indicated in
Fig. 2A. During the initial lattice ramp [at V
lat
=
6.1(1)E
r
], interference peaks of the superfluid
in the lattice can be observed (t = 6.8 ms) (Fig.
2B, top). Because quantum depletion caused by
the strong repulsive interactions already reduces
the visibility of the interference peaks in this
image (17), we also showthe initial superfluid for
identical lattice and dipole ramps, but at a scat-
tering length of a = 33(1) a
0
(t = 6.8 ms) (Fig. 2B,
bottom). The interference peaks are lost as the
Mott insulating regime is entered (t = 25 ms). In
the deep lattice, only weak nearest-neighbor cor-
relations are expected, resulting in similar images
for both repulsive and attractive interactions (t =
28 ms). After reducing the horizontal lattice depths
back into the superfluid regime, the coherence
of the atomic sample emerges again. For positive
temperatures, the final image at t = 30.5 ms is
comparable, albeit somewhat heated, to the ini-
tial one at t = 6.8 ms, whereas for attractive in-
teractions, sharp peaks show up in the corners of
the first Brillouin zone, indicating macroscopic
occupation of maximumkinetic energy. The spon-
taneous development of these sharp interference
peaks is a striking signature of a stable negative
temperature state for motional degrees of freedom.
In principle, the system can enter the negative
temperature regime following one of two routes:
It either stays close to thermal equilibrium during
the entire sequence or, alternatively, relaxes toward
a thermal distribution during lattice ramp-down.
Either way demonstrates the thermodynamic sta-
bility of this negative temperature state.
To examine the degree of thermalization in
the final states, we used band-mapped (18) im-
ages and extracted the kinetic energy distribution,
assuming a noninteracting lattice dispersion rela-
tion E
kin
(q
x
,q
y
). The result is shown in Fig. 3,
displaying very good agreement with a fitted
Bose-Einstein distribution. The fitted tempera-
tures of T = −2.2J/k
B
and T = 2.7J/k
B
for the two
cases only represent upper bounds for the ab-
solute values |T| of the average temperature be-
cause the fits neglect the inhomogeneous filling
of the sample (9). Both temperatures are slight-
ly larger than the critical temperature |T
BKT
| ≈
1.8J/k
B
(19) for the superfluid transition in an
infinite 2Dsystembut lie belowthe condensation
temperature |T
C
| = 3.4(2)J/k
B
of noninteracting
Fig. 3. Occupation distributions. The occupation
of the kinetic energies within the first Brillouin
zone is plotted for the final positive (blue) and neg-
ative (red) temperature states. Points show exper-
imental data extracted from band-mapped pictures.
Solid lines are fits to a noninteracting Bose-Einstein
distributionassumingahomogeneous system. (Insets)
Top row: Symmetrized positive (left) and negative
(right) temperature images of the quasimomentum
distribution in the horizontal plane. Bottom row:
Fitted distributions for the two cases. All distribu-
tions are broadened by the in situ cloud size (9).
Fig. 4. Stability of the positive (blue) and negative (red) temperature states. Main figure: Visibility
V = (n
b
− n
r
)/(n
b
+ n
r
) extracted from the atom numbers in the black (n
b
) and red (n
r
) boxes (indicated in
the TOF images) plotted versus hold time in the final state for various horizontal trap frequencies. Dark
red, |w
hor
|/2p = 43(1) Hz anti-trapping; medium red, 22(3) Hz anti-trapping; light red, 42(3) Hz trapping;
blue, 45(3) Hz trapping. (Inset) Coherence lifetimes t extracted from exponential fits (solid lines in main
figure). The statistical error bars from the fits are smaller than the data points. The color scale of the
images is identical to Fig. 2B (see also fig. S3).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 54
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

bosons in a 2D harmonic trap for the given av-
erage density (9).
Ideally, entropy is produced during the se-
quence only in the superfluid/normal shell around
the interim Mott insulator: While ramping to the
deep lattice, the atoms in this shell localize to
individual lattice sites and can subsequently be
described as a |T| = ∞ system (14). Numerical
calculations have shown that the total entropy
produced in this process can be small (4), because
most of the atoms are located in the Mott insu-
lating core. We attribute the observed additional
heating during the sequence to nonadiabaticities
during lattice ramp-down and residual double oc-
cupancies in the interim Mott insulator.
In principle, the coherence length of the atomic
sample can be extracted from the interference
pattern recorded after a long TOF (20). How-
ever, the experiment was limited to finite TOF,
where the momentum distribution is convolved
with the initial spatial distribution. By comparing
the measured TOF images with theoretically ex-
pected distributions, we were able to extract a
coherence length in the final negative tempera-
ture state of three to five lattice constants (9).
To demonstrate the stability of the observed
negative temperature state, Fig. 4 shows the vis-
ibility of the interference pattern as a function of
hold time in the final lattice. The resulting life-
time of the coherence in the final negative tem-
perature state crucially depends on the horizontal
trap frequencies (inset): Lifetimes exceed t =
600 ms for an optimally chosen anti-trapping po-
tential, but an increasingly fast loss of coherence
is visible for less anti-trapping geometries. In the
case of trapping potentials, the ensemble can even
return to metastable positive temperatures, giving
rise to the small negative visibilities observed af-
ter longer hold times (fig. S4). The loss of coher-
ence probably originates froma mismatch between
the attractive mean field and the external poten-
tial, which acts as an effective potential and leads
to fast dephasing between lattice sites.
The high stability of the negative temperature
state for the optimally chosen anti-trapping poten-
tial indicates that the final chemical potential is
matched throughout the sample such that no glob-
al redistribution of atoms is necessary. The re-
maining slow decay of coherence is not specific
to the negative temperature state because we also
observe comparable heating for the correspond-
ing positive temperature case (blue data in Fig. 4),
as well as the initial superfluid in the lattice. It
probably originates from three-body losses and
light-assisted collisions. In contrast to metastable
excited states (21), this isolated negative tem-
perature ensemble is intrinsically stable and
cannot decay into states at lower kinetic energies.
It represents a stable bosonic ensemble at attract-
ive interactions for arbitrary atom numbers; the
negative temperature stabilizes the system against
mean-field collapse that is driven by the negative
pressure.
Negative temperature states can be exploited
to investigate the Mott insulator transition (22) as
well as the renormalization of Hubbard parameters
(23, 24) for attractive interactions. As the stabil-
ity of the attractive gas relies on the bounded
kinetic energy in the Hubbard model, it naturally
allows a controlled study of the transition from
stable to unstable by lowering the lattice depth,
thereby connecting this regime with the study
of collapsing BECs (25), which is also of interest
for cosmology (26). Negative temperatures also
considerably enhance the parameter space acces-
sible for quantum simulations in optical lattices,
because they enable the study of newmany-body
systems whenever the bands are not symmetric
with respect to the inversion of kinetic energy.
This is the case, for example, in triangular or
Kagomé lattices, where in current implementa-
tions (27) the interesting flat band is the highest
of three sub-bands. In fermionic systems, nega-
tive temperatures enable, for example, the study
of the attractive three-component model with
symmetric interactions [SU(3)] describing color
superfluidity and trion (baryon) formation using
repulsive
173
Yb (28), where low losses and sym-
metric interactions are expected but magnetic
Feshbach resonances are absent.
References and Notes
1. N. F. Ramsey, Phys. Rev. 103, 20 (1956).
2. M. J. Klein, Phys. Rev. 104, 589 (1956).
3. A. P. Mosk, Phys. Rev. Lett. 95, 040403 (2005).
4. A. Rapp, S. Mandt, A. Rosch, Phys. Rev. Lett. 105,
220405 (2010).
5. E. M. Purcell, R. V. Pound, Phys. Rev. 81, 279 (1951).
6. A. S. Oja, O. V. Lounasmaa, Rev. Mod. Phys. 69, 1
(1997).
7. P. Medley, D. M. Weld, H. Miyake, D. E. Pritchard,
W. Ketterle, Phys. Rev. Lett. 106, 195301 (2011).
8. K. Huang, Statistical Mechanics (Wiley, New York, ed. 2,
1987).
9. See supplementary materials on Science Online.
10. J. A. Frieman, M. S. Turner, D. Huterer, Annu. Rev.
Astron. Astrophys. 46, 385 (2008).
11. I. Bloch, J. Dalibard, W. Zwerger, Rev. Mod. Phys. 80,
885 (2008).
12. D. Jaksch, C. Bruder, J. I. Cirac, C. W. Gardiner, P. Zoller,
Phys. Rev. Lett. 81, 3108 (1998).
13. N. Tsuji, T. Oka, P. Werner, H. Aoki, Phys. Rev. Lett. 106,
236401 (2011).
14. U. Schneider et al., Nat. Phys. 8, 213 (2012).
15. M. Greiner, O. Mandel, T. Esslinger, T. W. Hänsch,
I. Bloch, Nature 415, 39 (2002).
16. M. Zaccanti et al., Nat. Phys. 5, 586 (2009).
17. K. Xu et al., Phys. Rev. Lett. 96, 180405 (2006).
18. A. Kastberg, W. D. Phillips, S. L. Rolston, R. J. C. Spreeuw,
P. S. Jessen, Phys. Rev. Lett. 74, 1542 (1995).
19. B. Capogrosso-Sansone, S. G. Söyler, N. Prokof'ev,
B. Svistunov, Phys. Rev. A 77, 015602 (2008).
20. F. Gerbier et al., Phys. Rev. Lett. 101, 155303
(2008).
21. E. Haller et al., Science 325, 1224 (2009).
22. M. J. Mark et al., Phys. Rev. Lett. 108, 215302 (2012).
23. S. Will et al., Nature 465, 197 (2010).
24. J. Heinze et al., Phys. Rev. Lett. 107, 135303 (2011).
25. E. A. Donley et al., Nature 412, 295 (2001).
26. T. Fukuyama, M. Morikawa, T. Tatekawa, J. Cosmol.
Astropart. Phys. 2008, 033 (2008).
27. G.-B. Jo et al., Phys. Rev. Lett. 108, 045305 (2012).
28. Á. Rapp, Phys. Rev. A 85, 043612 (2012).
Acknowledgments: We thank A. Rapp, A. Rosch, S. Mandt,
and W. Hofstetter for helpful discussions and D. Garbe
for technical assistance. We acknowledge financial
support by the Deutsche Forschungsgemeinschaft (FOR801,
Deutsch-Israelisches Kooperationsprojekt Quantum phases of
ultracold atoms in optical lattices), the U.S. Defense Advanced
Research Projects Agency Optical Lattice Emulator program,
and Nanosystems Initiative Munich.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/52/DC1
Supplementary Text
Figs. S1 to S4
References (29–38)
23 July 2012; accepted 6 November 2012
10.1126/science.1227831
Current-Driven Spin Dynamics
of Artificially Constructed
Quantum Magnets
Alexander Ako Khajetoorians,
1
* Benjamin Baxevanis,
2
Christoph Hübner,
2
Tobias Schlenk,
1
Stefan Krause,
1
Tim Oliver Wehling,
3,4
Samir Lounis,
5
Alexander Lichtenstein,
2
Daniela Pfannkuche,
2
Jens Wiebe,
1
* Roland Wiesendanger
1
The future of nanoscale spin-based technologies hinges on a fundamental understanding and
dynamic control of atomic-scale magnets. The role of the substrate conduction electrons on
the dynamics of supported atomic magnets is still a question of interest lacking experimental
insight. We characterized the temperature-dependent dynamical response of artificially constructed
magnets, composed of a few exchange-coupled atomic spins adsorbed on a metallic substrate,
to spin-polarized currents driven and read out by a magnetic scanning tunneling microscope
tip. The dynamics, reflected by two-state spin noise, is quantified by a model that considers the
interplay between quantum tunneling and sequential spin transitions driven by electron spin-flip
processes and accounts for an observed spin-transfer torque effect.
F
or magnetic storage technology (1), where
magnets represent bits of information(2, 3),
effective manipulation of the magnetization
without a magnetic field is of crucial importance.
All-electrical manipulation offers technological
advantages, such as highly localized bit control
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 55
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

free of moving parts and compatibility with stan-
dard semiconductor technology (4, 5). Integral
to current-induced switching is the spin-transfer
torque (4–6) (STT) effect, where angular mo-
mentum transfer from a flux of incident spin-
polarized electrons can exert “torque” on the
magnetization, causing it to switch in a prefer-
ential direction (7). STT can efficiently reverse
the magnetization of magnetic layers (8, 9), drive
domain walls (10) and vortices, and be combined
with giant magneto-resistance technology to gener-
ate high-frequency electrical oscillators (11, 12).
The ultimate goal of such spin-based technologies
would be the total electrical control of an atomic
“bit.” Factors complicating the achievement of
this goal include the presence of a substrate, which
can dramatically modify both the magnetic an-
isotropy and the moment at the single atom level
(13), and spin quantization, which can play an
enhanced role in the magnetization dynamics,
for example, via quantum tunneling (14).
For small magnets that are sufficiently de-
coupled from an electron bath, such as molecular
magnets (15), STTcan be ascribed to asymmetric
spin pumping induced by spin-polarized trans-
port electrons that favorably excite particular spin
states, via transfer of angular momentum, over or
through an anisotropy barrier (16, 17). A gener-
alized Anderson model taking into account the
coupling of the electron baths to the quantum
spin adequately describes the dynamics of such
isolated spins (16–19). For small magnets that
are directly adsorbed on a metallic substrate
where the moment strongly couples to the sub-
strate conduction electrons (20, 21), it is still
questionable whether such a model is applica-
ble and what the exact role is of the substrate
conduction electrons on the spin dynamics of
the magnet (22).
By using spin-polarized scanning tunneling
microscopy (SP-STM), we demonstrate that it
is possible to characterize the real-time current-
driven dynamics of tailored magnets composed
of a few direct-exchange-coupled Fe atoms on
a metallic copper single-crystal surface. STM
topographs of single Fe atoms on the surface and
subsequent assembly of a five-atom Fe magnet
by tip-induced atomic manipulation (23, 24)
are shown in Fig. 1, A and B. Density functional
theory (DFT) (24) revealed two stable geomet-
ric configurations, “pyramid” (Fig. 1C) and “flat”
(Fig. 1D), of such five-atom Fe magnets on the
copper surface. The total magnetic moment of
each configuration, including orbital contributions,
is m
J
≈ 15.2m
B
(where m
B
is a Bohr magneton)
and m
J
≈ 14.7m
B
, respectively, which averages to
~3m
B
per atom, close to the magnetic moment of
a single Fe adatom (~3.5m
B
) (20, 25). The large
total magnetic moment results from a strong fer-
romagnetic exchange coupling between the con-
stituent Fe atoms to the total angular momentum
(26). We thus treat the magnet as a single total
angular momentum, J (21), which is related to
the total magnetic moment via m
J
= gm
B
J {where
g is the g factor [assumed g ≈ 2 (20)]}; we will
refer to the total angular momentum as the spin
for short. A considerable out-of-plane magnetic
anisotropy (24) leads to a schematic level di-
agram of the spin states (Fig. 1E).
At modest current values using a magneti-
cally coated tip that has out-of-plane spin sen-
sitivity (24), spin-based telegraph noise (27) can
be read out on top of the assembled magnet in
constant-current mode (Fig. 1F). It represents full
reversal of the magnetization between two de-
generate ground states, labeled |T15/2〉 in Fig. 1E,
aligned parallel/antiparallel to the surface normal.
We refer to these up and down states as “1” and
“0.” Although all T|J
z
| states contribute to the
spin dynamics, the limited time resolution of
our experiment allows us to resolve the signal
from the magnet only in one of the two degen-
erate ground states, because they are the longest
lived. Therefore, it is possible to extract the oc-
cupational lifetimes t
T
of the 1 and 0 states as
well as the mean lifetime t* = (t
+
+ t

)/2 from
the measured spin noise, for a given tunneling
current I
t
, bias voltage V
s
, out-of-plane magnetic
field B, and temperature T (24). The histogramof
the two-state noise (Fig. 1, G and H), at B = 0 T,
shows a considerable asymmetry in the state-
resolved occupational lifetimes t
T
, indicating a
preferential state. This preferential state is reversed
1
Institute of Applied Physics, Hamburg University, Jungiusstrasse
11, 20355 Hamburg, Germany.
2
I. Institute of Theoretical Physics,
Hamburg University, Jungiusstrasse 9, 20355 Hamburg,
Germany.
3
Institute of Theoretical Physics, Bremen University,
Otto-Hahn-Allee 1, 28359 Bremen, Germany.
4
Bremen Center
for Computational Materials Science, Bremen University, Am
Fallturm 1a, 28359 Bremen, Germany.
5
Peter Grünberg Institute
and Institute for Advanced Simulation, Forschungszentrum Jülich
and JARA (Jülich Aachen Research Alliance), 52425 Jülich, Germany.
*To whom correspondence should be addressed. E-mail:
akhajeto@physnet.uni-hamburg.de (A.A.K.); jwiebe@physnet.
uni-hamburg.de (J.W.)
Fig. 1. Constant-current STM images of single Fe atoms on the surface of Cu(111) (A) before and (B)
after construction of a five-atom Fe magnet. V
s
= –10 mV, I
t
= 0.6 nA. (C and D) DFT calculations of
the relaxed geometry of the two possible configurations (pyramid and flat, respectively) of the five-
atom Fe magnet (red spheres) on Cu(111) (blue spheres) constructed in (B). (E) Energy-dependent |J
z

states for a five-atom Fe magnet with total spin J = 15/2 and a magnetic anisotropy barrier of D
b
. The
J
%
z
states that have a finite overlap across the barrier are indicated by the same line (solid or dashed). The
arrows indicate three of the many transitions meditated by nonequilibrium-induced QT, as well as
excitation and relaxation via te

and se

. (F) Spin-dependent telegraph signal in the tip height Z
measured on a typical five-atom magnet where ∆Z ≈ 4 pm denotes switching between the two de-
generate ground states 0 (|a

〉) and 1 (|a
+
〉), which are color-coded (blue and red) (T = 0.3 K, I
t
= 1 nA,
V
s
= −10 mV). (G and H) Voltage polarity–dependent histograms (y scale from 0 to 1), at B = 0 T, of
the state-dependent lifetimes t
T
of |a
T
〉 illustrating favorability of the state 1 for positive bias (and
0 for negative bias). The height of each bin represents the probability (%) of finding the labeled
state at a given time.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 56
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

when the bias polarity, and hence the spin-polarized
current direction, is reversed. The polarity-dependent
asymmetry in the state of the magnet, in the ab-
sence of a magnetic field, results froma consider-
able STTinduced by the spin-polarized tunneling
electrons.
The mean switching frequency (n* = 1/t*)
has a linear dependence on I
t
(Fig. 2A), indi-
cating that the total number of tunneling elec-
trons q (= I
t
× t*) needed to reverse the magnetic
state is independent of the current (Fig. 2B); for
a five-atom magnet, ~5 × 10
9
incident tunneling
electrons are needed to reverse the magnetic
state. Similar behavior is observed as the mag-
net size is increased from five to six atoms. The
fact that q is independent of I
t
is a hallmark of
an inelastic spin excitation (17, 20, 21, 28, 29),
where each tunneling electron has a fixed prob-
ability of exciting the spin eigenstates of the
magnet (30). Also the observed asymmetry be-
tween the state-dependent lifetimes (t
T
) (B = 0),
resulting from STT, is independent of the mag-
nitude of I
t
(Fig. 2, A and B, histograms). If
Joule heating was important here, increasing I
t
would reduce this asymmetry because heating
would symmetrically reverse the magnetization,
like seen for larger-scale Fe islands (27). There-
fore, Joule heating can be ignored in the chosen
range of both I
t
and V
s
in our experiment. This is
most likely the result of the comparatively small
bias voltage V
s
used here, which is too small to
strongly couple the spin to phonons (31).
The dependence of the lifetime t* on the
applied bias V
s
(Fig. 2C) for several selected five-
atom magnets exhibits a stronger than exponen-
tial increase in t* as the energy of the tunneling
electrons (eV
s
) is lowered. Both current- and
voltage-dependent studies of t* suggest that, in
the absence of current or for very small tunneling
electron energy, these magnets remain stable for
extremely long times. The exact current and en-
ergy dependence of t* depends intricately on the
spin-dependent energy landscape of the magnet
(Fig. 1E), which is determined by the magnetic
anisotropy and the total spin of the system.
In order to quantify the dynamics and link the
experimentally observed lifetimes to the magnetic
properties of the system, we used the following
quantumimpurity model (24). The spin eigenstates
|a〉 of the magnet in an out-of-plane magnetic field
B are derived from a “giant spin” Hamiltonian
H
%
spin
¼gm
B
BJ
%
z
þDJ
%2
z
þ EðJ
%
2
x
−J
%
2
y
Þ (15), where
J
%

¼ ðJ
%
x
; J
%
y
; J
%
z
Þ is the vector spin operator of
the magnet, and D and E are the axial and trans-
verse magnetic anisotropy, respectively, which
determine the preferential orientation of the spin.
A master equation (24) characterizes the dynam-
ics by considering all possible transition proba-
bilities between the various spin eigenstates |a〉
(Fig. 1E) induced both by sequential transitions
over the anisotropy barrier [driven by spin-flip
scattering of tunneling electrons (te

) and substrate
conduction electrons (se

)] and by quantum tun-
neling (QT) through the anisotropy barrier.
Model calculations (Fig. 2C) show excellent
agreement with the experimental data for mag-
netic anisotropy values of D≈ –0.1 meVand E ≈
0.02 meV, confirming the uniaxial out-of plane
anisotropy with a weak in-plane anisotropy pre-
dicted by ab initio methods (fig. S3). Because E
is small, J
%
z
eigenstates are approximately the ei-
genstates of H
%
spin
(Fig. 1E). Our model calcu-
lations show that t*(V ) is very sensitive to small
variations in the anisotropy, as seen experimen-
tally for different magnets. Moreover, as V
s
is
decreased, t* dramatically increases, spanning
three decades. This divergent behavior of t* is
seen as the energy approaches the calculated first
ν
Fig. 2. (A) Current dependence of the mean switching frequency for five (black) and six (gray)
atom Fe magnets (V
s
= –10 mV, T = 0.3 K). Each data point represents an average value taken
from 1 to 20 different magnets. All studied magnets of a particular size showed nearly identical
values of n* at a given I
t
. Solid lines are linear fits to the measured data. (B) Total number of
tunneling electrons, q = I
t
× t*, extracted from (A), needed to switch a given magnet composed of
the indicated number of atoms. Solid lines are fits to a constant q indicated in the figure. (Right)
Histograms (y scale from 0 to 1) illustrating that the asymmetry in the two states 0 and 1 is
independent of current (B = 0 T). The height of each bin represents the probability (%) of finding
the labeled state at a given time. (C) Bias voltage dependence (V
s
) of t* for five selected five-atom
magnets (symbols). Solid lines are model calculations for a quantum magnet having a spin of J =
15/2 and the indicated anisotropy values D and E in meV. The vertical dashed lines indicate the
energy difference between the ground state and the first excited state ∆
01
resulting from the model
(I
t
=3 nA, T = 0.3 K). Error bars indicate SE.
Fig. 3. (A) Temperature dependence of the charge needed to reverse a
five-atom magnet for various I
t
as listed. Solid lines indicate the model
calculations for a quantum magnet of spin J = 15/2 with the anisotropy
values D = −0.09 meV and E = 0.02 meV. The indicated regions illustrate
the temperature range where QT and se

driven transitions are impor-
tant. (B) Magnet-size dependency of q(T) for five (I
t
= 0.4 nA), six (I
t
=
2.5 nA), and seven (I
t
= 26 nA) atom Fe magnets (V
s
= –10 mV, B = 0 T).
(Insets) q plotted logarithmically as a function of 1/T. Error bars, SE.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 57
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

transition energy, ∆
01
(Fig. 2C, dashed lines), for
a given magnet. No switching was observed be-
low |V
s
| < 2 mV for currents I
t
≤ 3 nA up to
maximum observation times of 2 hours.
To further investigate the role of QTand spin-
flip scattering of (te

) and (se

) on the dynamics
of the magnet, we varied the temperature. To remove
the current dependency, we plotted the switching
charge q(T) rather than the lifetime (Fig. 3). The
required charge q to switch a five-atom magnet
decreases as the temperature increases, reflecting
the decrease of t*(T) (Fig. 3A). The slope is rather
small for the low-temperature range T ≤ 1.5 K,
but q falls off more strongly for T ≥ 1.5 K,
resulting in a shoulder in q(T) at T ≈ 1.5 K. For
T ≥ 2 K, an exponential tail indicates Arrhenius
behavior, similar to what was observed for larger-
scale Fe islands (27). The Arrhenius behavior for
T ≥ 2 K is evident from the log plot in Fig. 3A
(inset), whereas a plateaulike regime appears for
T ≤ 1.5 K. Such plateaus have been attributed to
QT of the magnetization through the anisotropy
barrier (19); however, our results show a devi-
ation from a constant transition rate in the low-
temperature regime, indicating that QT alone
cannot account for the observed temperature
dependence of t*.
In order to link the temperature dependence
of q(T) to the different spin-transition pro-
cesses, we used the quantum impurity model
with the D and E values extracted from the fits
in Fig. 2C and fit the T dependence of q for
different I
t
(24) (solid lines in Fig. 3A), yielding
good agreement with experiment. On closer in-
spection, the model reveals that this temperature
dependence results from the following two spin
transition mechanisms that contribute to the
switching of the magnet in addition to te

-driven
sequential transitions: (i) sequential transitions
of the magnet induced by spin-flip scattering of
thermally excited substrate electron-hole pairs
generated by the broadening of the Fermi func-
tion. This process, which is analogous to “damp-
ing” (32), dominates at higher temperatures. (ii)
Nonequilibrium QT resulting from transverse
anisotropy (E ≠ 0). Note that the equilibrium
QT, which fully reverses the magnetic state from
one ground state to the other without any addi-
tional spin-flip processes (14), is blocked because
for a half-integer spin the states across the barrier,
which have the same | J
z
|, have zero overlap. How-
ever, in nonequilibrium, a two-step process where
a te

or se

electron changes J
z
can lead to QT
because there is a finite overlap of states across
the barrier for particular values of ∆J
z
≠ 0 re-
sulting from E ≠ 0 (Fig. 1E). The most dominant
QT channels responsible for reversal are closer
to the top of the barrier, because the energy bar-
rier is smaller for these states than for states near
the bottomof the barrier. Therefore, QTprocesses
are mainly preceded by sequential spin transitions
that drive the system into higher excited states,
resulting in a weak temperature dependence of
the whole QT process. The temperature-dependent
interplay between the two transition effects (i) and
(ii) leads to a shoulder in q(T); at higher temper-
ature, the well-known Arrhenius behavior is re-
covered. Moreover, the shape of the shoulder
depends on the size of the transverse anisotropy E
(fig. S2). Thus, the shoulder serves as a delinea-
tion of the temperature range where QT plays a
prominent role in the magnetization dynamics.
Therefore, the underlying mechanism of how the
magnetization is fully reversed can be understood
within the quantum impurity model when account-
ing for both substrate conduction electrons and
nonequilibrium-induced QT.
We explored the effects of an external out-of-
plane magnetic field, B, on the observed dynam-
ics. With increasing B, the asymmetry in the spin
noise increases; that is, the lifetime t
+
of one of
the ground states |a
+
〉 becomes larger at the ex-
pense of the lifetime t

of the other ground state
|a

〉 (Fig. 4, A and B). This is a result of the
Zeeman energy, gm
B
B J
%
z
, favoring the ground
state of the magnet, which has a magnetization
pointing parallel to B; t
+
indeed shows a mono-
tonic increase (and t

a monotonic decrease) with
increasing B. The balance t
+
= t

occurs at a
nonzero magnetic field, |B
STT
| ≈ 0.1 T. The
reason for the observed asymmetry at B = 0 T is
a result of the spin polarization of the tunneling
current created by the magnetic tip. The spin-
polarized current favors t e

sequential transi-
tions over the barrier in one particular direction
as opposed to the other (17), analogous to STT.
As a proof of the STT effect, we reversed the
direction of I
t
by changing the bias polarity,
resulting in the expected sign change of B
STT
(Fig. 4B).
The observed asymmetry between the life-
times t
+
(B) and t

(B) can be further quan-
tified by the nonequilibrium magnetization
M(B) = [t
+
(B) − t

(B)]/[t
+
(B) + t

(B)] (Fig. 4C),
which saturates toward T1 at a comparatively
large magnetic field B > 0.5 T for the five-atom
magnet. M(B) calculated from the quantum
impurity model assuming the anisotropy values
extracted fromthe voltage-dependent t* (Fig. 2C)
is in good agreement with experiments for both
bias polarities. The spin polarization of the tip,
assumed to be about 10% in the calculation, ac-
counts for the observed polarity-dependent STT
effect, B
STT
, of the nonequilibrium magnetization
curve. The nonzero transverse anisotropy is re-
sponsible for the large magnetic field needed to
saturate the magnetization as compared with
the single Fe atom limit (20). The presence of
finite E in combination with a large spin (J )
leads to a line shape of M(B) that is nearly inde-
pendent of T for temperatures belowthe Arrhenius
regime (Fig. 4D), in contrast to atoms that exhibit
no transverse anisotropy (13).
The observed STT behavior persists for
larger magnets, with the overall saturation field
Fig. 4. (A) Evolution of the spin-dependent telegraph noise in increasing magnetic field for a five-atom
magnet (V
s
= −10 mV, I
t
= 1 nA, T = 0.3 K). (B) State-resolved lifetime (t
T
) of a five-atom magnet for
negative and positive bias polarity in red and blue, respectively, at V
s
as indicated. The vertical lines and
arrow indicate the strength of the magnetic field, B
STT
, needed to compensate the spin-transfer torque
effect. (C) Nonequilibrium magnetization curve M(B) of five- and six-atom magnets for negative and
positive bias polarity, as indicated. (D) Temperature dependence of M(B) for both positive and negative
bias polarities. Solid and dashed lines in (C) and (D) indicate calculations from the quantum impurity
model with a spin J = 15/2 using the anisotropy values D = −0.09 meV and E = 0.02 meV and assuming a
tip spin polarization of 10%. Error bars, SE.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 58
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

dropping as a result of the increased total spin
as manifested by the sharper line shape in M(B)
(Fig. 4C). As the size of the magnet increases,
similar behavior is seen in q(I
t
), but the overall
value of q needed to reverse the larger magnet
increases (Fig. 2, A and B). The general shape
of q(T) (Fig. 3B) persists for larger magnets,
proving that the underlying mechanisms that
govern the dynamics remain the same. Whereas
D and E change as the size of the magnet in-
creases, thus affecting t*(I
t
,V
s
), the strong in-
crease in q needed to switch the magnet as the
magnet gets larger is most likely a consequence
of the increased spin of the magnet. Increasing
the spin results in an increased number of se-
quential transitions needed to reverse the magne-
tization, requiring a higher total charge to reverse
the magnetic state.
Although the relaxation of the magnets
studied here is far from purely quantum, namely
the quantum phase is destroyed, it is surprising
that, for such a strongly hybridized spin coupled
to an electron bath, quantum effects are indeed
necessary to fully describe the dynamics of the
system. The strength of the STT studied here is
determined solely by the total spin polarization of
the tip independent of the total current (24). This
can be seen in the nearly constant asymmetry in
Fig. 2, Aand B, and fig. S4 regardless of the mean
switching frequency n*, illustrating the quantum
nature of STT in atomic-scale magnets (16). Our
work brings to light fundamental processes of in-
terest for future magnetic memory devices that are
scaled to atomic dimensions.
References and Notes
1. J. Åkerman, Science 308, 508 (2005).
2. M. N. Baibich et al., Phys. Rev. Lett. 61, 2472 (1988).
3. B. Dieny et al., Phys. Rev. B 43, 1297 (1991).
4. J. Slonczewski, J. Magn. Magn. Mater. 159, L1 (1996).
5. L. Berger, Phys. Rev. B 54, 9353 (1996).
6. D. Ralph, M. Stiles, J. Magn. Magn. Mater. 320, 1190
(2008).
7. A. Brataas, A. D. Kent, H. Ohno, Nat. Mater. 11, 372
(2012).
8. M. Tsoi et al., Phys. Rev. Lett. 80, 4281 (1998).
9. E. B. Myers, D. C. Ralph, J. A. Katine, R. N. Louie,
R. A. Buhrman, Science 285, 867 (1999).
10. S. S. P. Parkin, M. Hayashi, L. Thomas, Science 320,
190 (2008).
11. M. Tsoi et al., Nature 406, 46 (2000).
12. S. I. Kiselev et al., Nature 425, 380 (2003).
13. F. Meier, L. Zhou, J. Wiebe, R. Wiesendanger, Science
320, 82 (2008).
14. W. Wernsdorfer, R. Sessoli, Science 284, 133 (1999).
15. D. Gatteschi, R. Sessoli, Molecular Nanomagnets
(Oxford Univ. Press, Oxford, ed. 1, 2006).
16. F. Delgado, J. J. Palacios, J. Fernández-Rossier, Phys. Rev.
Lett. 104, 026601 (2010).
17. S. Loth et al., Nat. Phys. 6, 340 (2010).
18. M. Misiorny, J. Barnaś, Phys. Rev. B 76, 054448
(2007).
19. S. Loth, S. Baumann, C. P. Lutz, D. M. Eigler,
A. J. Heinrich, Science 335, 196 (2012).
20. A. A. Khajetoorians et al., Phys. Rev. Lett. 106, 037205
(2011).
21. T. Schuh et al., Phys. Rev. B 84, 104401 (2011).
22. D. L. Mills, P. Lederer, Phys. Rev. 160, 590 (1967).
23. D. M. Eigler, E. K. Schweizer, Nature 344, 524 (1990).
24. Materials and methods are available as supplementary
material on Science Online.
25. A. A. Khajetoorians et al., Nat. Phys. 8, 497 (2012).
26. A. Bergman, L. Nordström, A. Burlamaqui Klautau,
S. Frota-Pessôa, O. Eriksson, Phys. Rev. B 75, 224425
(2007).
27. S. Krause, L. Berbil-Bautista, G. Herzog, M. Bode,
R. Wiesendanger, Science 317, 1537 (2007).
28. C. F. Hirjibehedin, C. P. Lutz, A. J. Heinrich, Science 312,
1021 (2006); 10.1126/science.1125398.
29. J. Fransson, Nano Lett. 9, 2414 (2009).
30. N. Lorente, J.-P. Gauyacq, Phys. Rev. Lett. 103, 176601
(2009).
31. H. Gawronski, M. Mehlhorn, K. Morgenstern, Science
319, 930 (2008).
32. T. Gilbert, IEEE Trans. Magn. 40, 3443 (2004).
Acknowledgments: Financial support from the European
Research Council (ERC) Advanced Grant “FURORE”;
Bundesministerium für Bildung und Forschung; the Deutsche
Forschungsgemeinschaft via the SFB668, SPP 1285 (B.B.),
FOR1346, and the Graduiertenkolleg 1286 “Functional
Metal-Semiconductor Hybrid Systems”; the Cluster of
Excellence “Nanospintronics” funded by the Forschungs-und
Wissenschaftsstiftung Hamburg; and HGF-YIG Programme
VH-NG-717 (Functional nanoscale structure probe and
simulation laboratory–Funsilab, S.L.) is gratefully
acknowledged.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/55/DC1
Materials and Methods
Supplementary Text
Figs. S1 to S4
References (33–48)
7 August 2012; accepted 12 November 2012
10.1126/science.1228519
Strategic Redox Relay Enables A
Scalable Synthesis of Ouabagenin, A
Bioactive Cardenolide
Hans Renata, Qianghui Zhou, Phil S. Baran*
Here, we report on a scalable route to the polyhydroxylated steroid ouabagenin with an unusual
take on the age-old practice of steroid semisynthesis. The incorporation of both redox and
stereochemical relays during the design of this synthesis resulted in efficient access to more
than 500 milligrams of a key precursor toward ouabagenin—and ultimately ouabagenin itself—and
the discovery of innovative methods for carbon-hydrogen (C-H) and carbon-carbon activation and
carbon-oxygen bond homolysis. Given the medicinal relevance of the cardenolides in the treatment
of congestive heart failure, a variety of ouabagenin analogs could potentially be generated from the
key intermediate as a means of addressing the narrow therapeutic index of these molecules. This
synthesis also showcases an approach to bypass the historically challenging problem of selective
C-H oxidation of saturated carbon centers in a controlled fashion.
I
n the realm of terpene synthesis, nature has
evolved a highly efficient biosynthetic system
to achieve chemo- and stereoselective oxi-
dations far beyond the capabilities of chemical
synthesis in the laboratory. For example, the com-
plex steroid ouabagenin is thought to have arisen
from progesterone (Fig. 1A) through a series of
direct, chemoselective oxidations employing a
multitude of hydroxylase enzymes (1). A step-
by-step emulation of this oxidation sequence is
difficult, if not impossible, to execute in a labo-
ratory setting. As a way to circumvent this direct
oxidation problem, a plethora of directed oxi-
dation methods have been invented (2), whereby
a template is typically appended to effect site-
selective oxidation(s) on a molecular framework.
Historically, another approach, which relies on
the indirect use of preexisting functional groups
of the framework in the absence of a directing
template, has also been widely employed, as evi-
denced by the wealth of synthesis literature on
terpene functionalizations (3, 4). Use of this
approach has enabled semisyntheses of highly
complex steroids such as digitoxin (5), batracho-
toxin (6), dihydroconessine (7), cephalostatin 1
(8), cortistatin A (9), and withanolide A (10) and
has laid a foundation for pharmaceutical research
on commercial semisynthetic steroid medicines
such as finasteride, dexamethasone, and proges-
tin. Despite these achievements, the scalable syn-
thesis of polyhydroxylated steroids (more than
five hydroxyl moieties on the tetracyclic skele-
ton) (Fig. 1B) is unknown. Here, we report the
accomplishment of such a feat using a quasibio-
mimetic oxidation approach, which relies on
the strategic interplay of two relay elements: (i)
redox relay, defined as rapid (one or two steps)
transfer of redox information from one site to
another within a framework, and (ii) oxidative
stereochemical relay, defined as transfer of stereo-
chemical information during an oxidative pro-
cess. It is the interplay of both strategies, rather
than their isolated use, that distinguishes this
synthesis from conventional strategies.
Ouabagenin (1) was specifically chosen as a
target molecule to showcase the application of
this quasibiomimetic oxidation strategy because
of its highly oxidized molecular framework and
its biological relevance as a positive inotrope (11).
Department of Chemistry, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, CA 92037, USA.
*To whom correspondence should be addressed. E-mail:
pbaran@scripps.edu
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 59
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Ouabagenin belongs to a unique class of steroids,
known as cardenolides, possessing both cis A/B
and C/D ring fusions with an angular hydroxyl
moiety at C14, as well as a b-oriented butenolide
substituent at C17. The high oxidation level of its
framework poses a formidable challenge for syn-
thesis efforts and, in addition, the predominantly
b orientation of its polyhydroxylation pattern
renders ouabagenin an able ligand to inorganic
species (12), including borosilicate glassware.
Its parent glycoside, ouabain, along with digoxin
and digitoxin, is used as a treatment for conges-
tive heart failure, a progressive condition that
currently affects approximately two million peo-
ple in the United States alone (13). The use of
cardiotonic steroids, however, is complicated by
an extremely narrow therapeutic index, and pa-
tients are often treated with 60% of the toxic
dose. Although structure-activity relationship
studies have been conducted on cardenolides
and related bufadienolides (14), they have been
limited to hydroxylated analogs, and studies on
other heterocyclic analogs are relatively rare.
Thus, a de novo synthesis that would allow
versatile topological diversification could lead
to the development of further analogs potentially
possessing a wider therapeutic window as safer
alternatives (15). A previous synthesis of ouaba-
genin by Deslongchamps and colleagues, although
an elegant accomplishment, proceeded in 41 steps
from Hajos-Parrish ketone in 0.21% overall yield,
producing 7 mg of their most advanced interme-
diate, which was intercepted by a relay synthesis
from degradation of authentic ouabain to arrive
at ouabagenin in 8 steps (16). In addition, several
synthetic studies on this molecule have been
disclosed (17).
In our synthesis planning (Fig. 1C), we en-
visioned that the butenolide moiety would be
appended late in the synthesis, revealing the oua-
bagenin ketonic core, so-called ouabageninone
(2), in a retrosynthetic sense. Three relay events,
systematically applied during the planning stage,
formed the basis of our approach. The first re-
quired installation of the tertiary alcohol at C14
on pentaol 3 through relay of the redox in-
formation coded in the C17 ketone. Oxidative
stereochemical relay—another modality of this
strategy—would leverage the primary alcohol
at C19 to correctly install the requisite oxidation
state at the C1 and C5 positions. Last, discon-
nection of the hydroxyl group at C19 inspired
the invention of a redox relay from the C11
ketone functionality of 5. The requisite hydroxyl
moieties at C3 and C11 would be generated
from stereocontrolled reductions of the respec-
tive carbonyl groups. In light of the affinity of the
fully functionalized A ring for inorganic species
and common laboratory glassware, strategic pro-
tection would be applied on the more advanced
intermediates for their ease of handling. Given
the need for a scalable and economically viable
route to the cardenolides and derivatives thereof,
cortisone acetate (6, ~US$1/gram, from OChem
Incorporated) was chosen as an ideal starting mate-
rial. The realization of this strategy enabled the
scalable synthesis of ouabagenin, where all steps
up to step M (Fig. 2) have been conducted on a
gram scale.
Cortisone acetate was converted to adreno-
sterone 5 (also commercially available, but more
expensive) employing a modified protocol (see
supplementary materials) that proceeded in 86%
yield (Fig. 2). After ketalization and recrystalli-
zation, the first redox-relay event of the route was
realized in the Norrish type II photochemical
functionalization of the angular C19 methyl group
by the C11 ketone moiety (18) because we were
unable to reproduce a porphyrin-catalyzed, di-
rect C19 methyl hydroxylation report (19). Initial
experimentation revealed that although desired
cyclobutanol 7 could be obtained by convention-
al solution photochemistry, the reaction proceeded
in only modest yield (see supplementary mate-
rials) and was plagued by competitive formation
B
C
A
Quasi-Biomimetic
Oxidation
Me
OH
O
HO
HO
HO
HO
HO
Me
HO
HO
HO
HO
HO
O
14
17
19
O
HO
Me
O
O
1
5
11
Me
O Me
O
O
19
cortisone
acetate (6)
(US$1.2/gram)
ouabagenin (1)
HO
OH
Me
OH
O
O
H
H
HO
HO
HO
Me
Me
O
O
H
H
H
OH
O
AcO
Redox-Relay
[C
17
C
14
]
Stereochemical
Relay
[C
19
C
1
;
C
19
C
5
]
Redox-Relay
[C
11
C
19
]
2
3
4
5
Me
Me
OH
HO OH
OH
OH
O
Me
HO
OH
O O
Me
Me
H
H
H
O
Me
OH
H
Me
H
HO
HO
HO
H
HO
Me
Me
HO
OH
Me
OH
O
O
H
H
HO
HO
HO
Polyhydroxylated Steroids
Biosynthesis
Me
O
Me
H
H
H
O
Me
progesterone ouabagenin (1)
Hydroxylase
Enzymes
herbasterol withangulatin G
Fig. 1. (A) Biosynthesis of the highly complex steroid ouabagenin. Structure of a typical hydroxylase
enzyme is shown. Image from the Research Collaboratory for Structural Bioinformatics Protein Data Base
(PDB) (www.pdb.org) of PDB ID 1FAH (41). (B) Selected examples of polyhydroxylated steroids. (C)
Retrosynthetic analysis of ouabagenin.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 60
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

of side products, most notably fromNorrish type I
cleavage of the C9-C11 bond of the steroidal
framework. We therefore explored solid-state irra-
diation of ketalized 5 in aqueous suspension—
drawing inspiration from Garcia-Garibay’s work
on synthesis using solid-state Norrish type I photo-
chemistry (20)—which, to our delight, not only
improved the yield for the cyclobutanol forma-
tion but also led to notable suppression of side
products. This improvement in chemoselectivity
comes with a trade-off in the rate of the reaction
(3 to 5 days to ~90% conversion, 2.5-g scale),
presumably due to the limited exposure of solid
surface area in a typical photoreaction vessel. The
use of a flow reactor should improve the reaction
efficiency and allow this reaction to be conducted
on an even larger scale because continuous oper-
ation will allow more uniform exposure of solid
surface area and closer positioning to the light
source (21).
Although the crystal structure analysis of 7
revealed weaker bond strength of the C11-C19
bond relative to the C9-C11 bond, initial attempts
to effect oxidative fragmentation (18, 22) of the
C11-C19 bond were met with failure; for exam-
ple, use of (diacetoxyiodo)benzene resulted in
the undesired fragmentation of the C9-C11 bond.
Eventually, iodide 8 could be obtained in ex-
cellent yield by the use of Barluenga’s reagent
(23), but the cost of the reagent proved to be a
substantial drawback for large-scale operations.
A more economical alternative was finally re-
alized by using N-iodosuccinimide as the oxidant
(24). This transformation is likely to proceed via
the intermediacy of a transient hypoiodite spe-
cies, which undergoes a chemoselective homolysis
of the C11-C19 bond, followed by recombination
with an iodine radical. Selective deketalization
of C3 and hydrolysis of the C19 iodide moiety
furnished enone alcohol 9, which was primed
for the next series of relay events, where the C19
hydroxyl moiety would serve to facilitate the in-
troduction of additional hydroxyl moieties on
C1 and C5. In addition, we surmised that the
angular disposition of the C19 hydroxyl moiety
would also translate to a diastereoselective
stereochemical relay to the b face of the A ring.
Success was eventually realized by epoxidation
of the enone moiety that—in contrast to epoxida-
tion literature on simpler C19 methyl substrates
(25)—proceeded with complete facial selectiv-
ity, likely through the formation of a hydrogen-
bonding network under the protic conditions of
the reaction. Dehydrogenation of the C1-C2 bond
of the resulting epoxide with selenium dioxide,
followed by another iteration of a directed ep-
oxidation event, provided diepoxide 11 in good
yield. Facial selectivity of the two epoxidation
events was confirmed by x-ray analysis after
deketalization.
Reductive opening of diepoxide 11 proved
to be the most difficult transformation to secure
on scale. A gamut of conditions was surveyed
(26), only to give a mixture of A ring enones as
products. Triol 12 was only accessible via treat-
ment with in situ generated aluminum amalgam.
Preliminary trials of the use of this reagent in a
mixture of organic solvent and water were beset
by unsatisfactory conversion and observation of
partial reduction of the epoxides (only one ep-
oxide was cleaved, and the other remained in-
tact). After extensive optimization, we found that
HO
Me
O
O
O
O
Me
O
O
O
HO
O
O
Me
O
OH
HO
HO
B h
C NIS, Li
2
CO
3
Me
O
O
O
(68%+
12%
recovered sm)
(85%)
H H
H H
H H H H
H
H H
H
O O
Me
Me
O
O
O
Me
O
O
O
H H
H
Me
Me
O
B
Et
K Li, NH
3
L PPTS,
Me
2
CO
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
O
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
O
O O
2
, Co(acac)
2
O
B
Et
OH
O
OH
(55% over 2 steps)
(69% over
2 steps)
O
O
H Al-Hg, H
2
O
O
O
O
O
Me
I
O
O
Me
HO
O
D TiCl
4
;
AgOAc
(71% overall)
H H
H H
H
H
O
O E H
2
O
2
F SeO
2
(50% over
3 steps)
I PPTS,
Me
2
CO
J LiBEt
3
H
(56%)
(63% over
2 steps)
M TMSOTf; Pd
II
19
19
11
19
19 1
2
5 5
1
3
11
5
11
14 15
14
17
[solid state]
[gram scale]
[gram scale]
[gram scale]
[C19 directed
epoxidation]
[gram scale]
[single diastereomer]
F
3
C
F F
F
F F
[stereocontrolled
hydration]
["on-water"]
[gram scale]
7 8
12 11 9
13
N SiO
2
, DIPEA
14 15 16 (protected
ouabageninone)
P HCl
(PFT)
[gram scale]
[gram scale]
essential additive
(86%)
[gram scale]
O
O
Me
O
O
HO
H H
H
O
5
[C19 directed
epoxidation]
[gram scale]
G H
2
O
2
[single diastereomer]
10
[15 steps overall]
[8 column puriÞcations]
[600 mg of 16 prepared]
bond a = 1.567 Å, bond b = 1.538 Å
PhI(OAc)
2
= bond "b" cleavage; NIS = bond "a" cleavage
[reagent controlled chemo- and regioselective C-C fragmentation]
a
b
Me
Me
O
H
H
O
H
adrenosterone (5)
O
A H
+
,
HO
OH
(81%)
11'
Fig. 2. Construction of the ouabagenin core. Reagents and conditions: (A)
p-TsOH [0.1 equivalent (equiv.)], ethylene glycol (30 equiv.), PhCH
3
, 135°C,
6 hours, 81%. (B) hn, SDS solution, 120 hours, 68%. (C) hn, N-iodosuccinimide
(3 equiv.), Li
2
CO
3
(3.5 equiv.), MeOH, PhCH
3
, 23°C, 20 min, 85%. (D) TiCl
4
(1 M in CH
2
Cl
2
, 1 equiv.), CH
2
Cl
2
, –10°C, 15 min; AgOAc (1.5 equiv.), tetra-
hydrofuran (THF), 50°C, 2 hours, 71% overall. (E) H
2
O
2
[35 weight percent
(wt %) in H
2
O, 6 equiv.], 3 M NaOH (1 equiv.), MeOH, 0°C, 75 min. (F) SeO
2
(1.1 equiv.), PhCl, 90°C, 10 hours. (G) H
2
O
2
(35 wt % in H
2
O, 6 equiv.), 3 M
NaOH (1 equiv.), MeOH, 0°C, 75 min, 50% (over three steps). (H) Al-Hg,
saturated NaHCO
3
, –5°C, 1 hour, 56%. (I) Pyridinium p-toluenesulfonate
(0.2 equiv.), CaSO
4
(2.5 equiv.), Me
2
CO, 23°C, 20 hours. (J) LiBEt
3
H (1 M in
THF, 1.1equiv.), THF, –78°C, 30min, 63%(over two steps). (K) Li (60 equiv.), NH
3
,
THF, –78°C, 30 min. (L) Pyridiniump-toluenesulfonate (1.5 equiv.), Me
2
CO, 70°C,
16 hours, 69%(over two steps). (M) TMSOTf (3 equiv.), Et
3
N(4 equiv.), CH
2
Cl
2
, 0 to
23°C, 30 min; Pd(OAc)
2
(1.2 equiv.), MeCN, 23°C, 3hours, thenFeCl
3
(1equiv.),
0°C, 10 min. (N) SiO
2
, iPr
2
EtN (55 equiv.), C
7
F
8
, 45 min, 55% (over two steps).
(O) Co(acac)
2
(0.2 equiv.), PhSiH
3
(3 equiv.), O
2
, dioxane, 23°C, 3 hours, 86%.
(P) Concentrated (conc.) HCl (1 equiv.), CH
2
Cl
2
, 23°C, 1 hour, 87%.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 61
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

the use of “on-water” (aqueous suspension) con-
ditions (27) was critical to obtaining a satisfactory
yield of triol 12. Specifically, a saturated NaHCO
3
solution was found to be the optimal medium for
this transformation (see supplementary materials).
After acetonide formation, the ketone moiety on
C3 was reduced with LiBEt
3
H, which effected a
concomitant formation of the ethyl boronic ester
of the two hydroxyl moieties on C1 and C5, there-
by bypassing the need for an additional protec-
tion step. The desired a configuration of the
hydroxyl group on C11 could be accessed by em-
ploying a thermodynamic reduction (Li/NH
3
).
Last, mild deketalization of C17 set the stage for
the final redox-relay event.
The C17 ketone moiety so revealed rendered
the C15-C16 methylene subunits amenable to de-
hydrogenation to furnish the conjugated enone.
Initial attempts at olefin isomerization on this
enone (28), however, were hampered by low
conversion and epimerization of the C14 stereo-
center, which represented a dead end because this
epimer could not be converted to 15. Ultimately,
we found that the use of fluorinated solvents, in
particular perfluorotoluene, aided the conversion
to 15, while suppressing the epimerization of the
Fig. 3. Completing the synthe-
sis of ouabagenin. Reagents and
conditions: (A) N
2
H
4
(10 equiv.),
Et
3
N (10 equiv.), 4:1 CH
2
Cl
2
:EtOH,
50°C, 5 hours; I
2
(3 equiv.), Et
3
N
(4 equiv.), THF, 10 min. (B) 17
(4 equiv.), [Ph
2
PO
2
][NBu
4
] (4
equiv.), Pd(PPh
3
)
4
(0.15 equiv.),
Cu(thiophene-2-carboxylate) (3
equiv.), N,N´-dimethylformamide
(DMF), 23°C, 2 hours, 42% (over
two steps). (C) CoCl
2
·6H
2
O (2.5
equiv.), NaBH
4
(5 equiv.), EtOH, 0
to 23°C, 20 min. (D) 2-tert-Butyl-
1,1,3,3,-tetramethylguanidine
(1.5 equiv.), C
6
H
6
, 100°C, 10 min,
70% (over two steps). (E) Conc.
HCl (2equiv.), MeOH, 23°C, 30min,
90%.
Me
O
B
Me
HO
OH
HO
HO
H
H
OH
O
O
ouabagenin (1)
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
OH
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
OH
O
O
C [Co
2
B]
OH
Me
O
O
O
H
H
Me
Et
OH
O
OH
17
A N
2
H
4
; I
2
, Et
3
N
17
O
O
Bu
3
Sn
17
B 17, CuTC,
Pd(PPh
3
)
4
O
O
E HCl
Me
OH
O
O via
16 18
19
20
(42% over
2 steps)
(70% over
2 steps)
17
16
(90%)
Me
2
N
Me
2
N
NtBu
D
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
O
Me
O
O
O
H
H
Me
Me
O
B
Et
OH
O
Fe
2
(ox)
3
,
Selectßuor,
NaBH
4
F
15 21
51%
Me
HO
OH
HO
HO
H
H
OH
O
O
ouabagenin (1)
OH
Me
Me
O
H
H
O
H
adrenosterone (5)
O
Me
Me
O
H
H
OH
H
testosterone
Me
Me
HO
H
H
H
OH
O
O
digitoxigenin
Me
Me
O
H
H
H
11-oxo-progesterone
O
O
Me
Me
HO
H
O
O
HO
N
Me
OH
batrachotoxinin A
• 20 steps;
• 3-step installation
of C14-OH;
• 2-step installation
of C19-OH
•15 steps; • 7-step
installation
of C14-OH (indirect
functionalization)
A B
Me
O
via
H
OH
Me
O
via
AcO
O
Me
via
• >40 steps; • 6-step
installation
of C14-OH (indirect
functionalization)
Br
Me
HO
OH
HO
H
H
OH
O
O
strophanthidol
Me
Me
AcO
H
H
H
pregnenolone acetate
• 24 steps; • 8-step
installation of C14
and C19-OH (indirect
functionalization)
O
Me
AcO
O
Br
via
= redox "donor"
= redox "acceptor"
[concise synthesis achieved
through redox-relay]
19
14
11
17
19
14
Me
Fig. 4. (A) Late-stage functionalization of deconjugated enone. Reagents and
conditions: Fe
2
(oxalate)
3
(4 equiv.), Selectfluor (4 equiv.), NaBH
4
(6.4 equiv.),
3:4:2 MeCN:THF:H
2
O, 0°C, 20 min, 51%. (B) Comparison to previous semi-
syntheses of related cardenolides.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 62
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

C14 center (see supplementary materials). For-
mation of p-p complexes has been suggested (29)
in the observed catalyst activity enhancement of
olefin metathesis in fluorinated aromatic solvents,
and this phenomenon between perfluorotoluene
and our enone systemcould also play a role in the
rate enhancement of our isomerization reaction,
although other effects—such as dielectric con-
stant and silicon-fluorine interactions—cannot be
ruled out. Mukaiyama hydration (30) of the re-
sulting olefin, although straightforward, neces-
sitated the use of dioxane as solvent to obtain a
satisfactory diastereomeric ratio of hydration
products (dr = 8:1, in favor of 16). Success of
this transformation marked the completion of
strategically protected ouabageninone, the ke-
tonic core of the target molecule that would
enable not only the synthesis of ouabagenin but
also the versatile access to analogs varied at C17,
as well as the related bufadienolide family of
natural products. As a testament to the scalability
of this 15-step sequence, >500 mg of this mol-
ecule has been synthesized to date. Although it is
definitely possible to convert the total amount
of 16 that we have procured to date to more
than 100 mg of the natural product, we are plan-
ning to use a substantial portion of this material
as a branching point to access various analogs at
C17 and profile their therapeutic indexes.
Attachment of the butenolide subunit (Fig. 3)
was accomplished by first converting ketone 16
to the corresponding vinyl iodide using Barton’s
protocol (31) and then subjecting the product to
a Stille cross-coupling with known stannane 17
(32). The use of Fürstner’s conditions (33) proved
to be critical to deliver dienoate 18 in a syntheti-
cally useful yield. Direct reduction of the C16-C17
olefin of dienoate 18 turned out to be formidable
because many of the conditions tried led to reduc-
tion from the convex face of the molecule. This
problemwas circumventedbyfirst treatingdienoate
18 with in situ generated Co
2
B (34) to afford
exclusively tetrasubstituted olefin 19 (no reduc-
tion of C16-C17 or butenolide moiety was ob-
served). Amyriad of bases were tried to isomerize
this olefin into conjugation, only to produce the
wrong stereochemistry of the newly generated
chiral center at C17. Ultimately, we found that
heating 19 in the presence of Barton’s base (35)
delivered enoate 20 with a predominantly correct
disposition of the butenolide moiety (dr = 3:1).
Lastly, unmasking of all the hydroxyl groups
under acidic condition delivered synthetic oua-
bagenin (1).
Salient features of the current route include
(i) the application of solid-state Norrish type II
photochemistry in natural product synthesis;
(ii) chemoselective, reagent-controlled cyclo-
butanol fragmentation using an inexpensive
reagent, N-iodosuccinimide; (iii) implementa-
tion of an “on-water” epoxide fragmentation; (iv)
a highly selective olefin isomerization promoted
by fluorous media; (v) a highly diastereoselective
Mukaiyama hydration to furnish the requisite cis
C/Dring junction; (vi) a chemoselective dienoate
reduction with Co
2
B; and (vii) robustness and
scalability of the route in an academic setting
(20 steps fromandrenosterone, 0.56%overall yield,
77% average yield per step). Scalable complex
molecule synthesis (gram scale or above) not
only demonstrates feasibility but also ensures the
robustness of the route, delivers useful quantities
to biological collaborators, and lowers the barrier
for pharmaceutical development. Despite being a
semisynthetic approach, the current route is ame-
nable to scaffold diversification: Preliminary
studies suggest that opposite stereochemistry
of both the C3 and C11 centers is readily acces-
sible in a controlled fashion; the isolated olefin
on 15 affords a platform for versatile function-
alizations, including the introduction of hetero-
atoms, such as fluorine, with radical-based
methods (showcased in Fig. 4A) (36); last, the
late-stage Stille coupling represents yet another
potential point of diversification by way of attach-
ment of different heterocyclic domains at C17.
The synthesis of the highly oxidized natural
product ouabagenin represents a proof of con-
cept for the development of a retrosynthetic strat-
egy centered solely on the synergistic union of
redox and stereochemical relays. As illustrated
in Fig. 4B, previous approaches to the semi-
synthesis of cardenolides (5, 6, 37) and related
steroids (batrachatoxin) used indirect means to
install crucial oxidized functionality at C19 and
C14. In all three cases, this resulted in lengthy
routes, compromising both the brevity and
scalability of the syntheses. In contrast, by using
a transform-based approach to “look ahead” (38),
the C11 and C17 ketones were viewed as “redox
donors” for the requisite oxidations at C19 and
C14, respectively. This logic enabled a dramatic
increase of complexity to be imparted on a mini-
mally oxidized, readily available steroid. In a
similar vein to synthesis designs predicated on
minimizing protecting group chemistry (39) or
maximizing aspects of synthesis economy (40),
the primary purpose of this work was to generate
opportunities to innovate. Although the ouaba-
genin case study reported here is only a single
example, we anticipate that syntheses that in-
corporate such a strategic interplay will result in
more efficient routes to terpenes.
References and Notes
1. R. Tschesche, Proc. R. Soc. Lond. B Biol. Sci. 180, 187
(1972).
2. G. Rousseau, B. Breit, Angew. Chem. Int. Ed. 50, 2450 (2011).
3. J. R. Hanson, Nat. Prod. Rep. 22, 104 (2005).
4. Y. Ishihara, P. S. Baran, Synlett 12, 1733 (2010).
5. K. Wiesner, T. Y. R. Tsai, Pure Appl. Chem. 58, 799 (1986).
6. R. Imhof, M. E. Gössinger, W. Graf, H. Berner,
L. Berner-Fenz, H. Wehrli, Helv. Chim. Acta 55, 1151 (1972).
7. E. J. Corey, W. R. Hertler, J. Am. Chem. Soc. 81, 5209
(1959).
8. K. C. Fortner, D. Kato, Y. Tanaka, M. D. Shair, J. Am.
Chem. Soc. 132, 275 (2010) and references cited therein.
9. J. Shi et al., J. Am. Chem. Soc. 133, 8014 (2011).
10. C. K. Jana et al., Angew. Chem. Int. Ed. 50, 8407
(2011).
11. A. Szent-Gyorgyi, Chemical Physiology of Contraction in
Body and Heart Muscle, (Academic Press, New York,
1953), pp. 86–91.
12. A. Kawamura et al., Proc. Natl. Acad. Sci. U.S.A. 96,
6654 (1999).
13. D. D. Schocken, M. I. Arrieta, P. E. Leaverton, E. A. Ross,
J. Am. Coll. Cardiol. 20, 301 (1992).
14. Y. Kamano et al., Bioorg. Med. Chem. 6, 1103 (1998).
15. P. W. Erhardt, J. Med. Chem. 30, 231 (1987).
16. H. Zhang, M. S. Reddy, S. Phoenix, P. Deslongchamps,
Angew. Chem. Int. Ed. 47, 1272 (2008).
17. B. Heasley, Chem. Eur. J. 18, 3092 (2012).
18. H. Wehrli, M. S. Heller, K. Schaffner, O. Jäger,
Helv. Chim. Acta 44, 2162 (1961).
19. B. Vijayarahavan, S. M. S. Chauhan, Tetrahedron Lett. 31,
6223 (1990).
20. D. Ng, Z. Yang, M. A. Garcia-Garibay, Org. Lett. 6,
645 (2004).
21. N. G. Anderson, Org. Process Res. Dev. 16, 852 (2012).
22. K. Wietzerbin, J. Bernadou, B. Meunier, Eur. J.
Inorg. Chem. 2000, 1391 (2000).
23. J. Barluenga, F. González-Bobes, M. C. Murguía,
S. R. Ananthoju, J. M. González, Chem. Eur. J. 10, 4206
(2004).
24. C. E. McDonald, H. Holcomb, T. Leathers,
F. Ampadu-Nyarko, J. Frommer Jr., Tetrahedron Lett. 31,
6283 (1990).
25. S. Hrycko, P. Morand, F. L. Lee, E. J. Gabe, J. Org. Chem.
53, 1515 (1988).
26. R. C. Larock, Comprehensive Organic Transformations
(Wiley, New York, 1989), pp. 505–508.
27. S. Narayan et al., Angew. Chem. Int. Ed. 44, 3275
(2005).
28. U. Egner et al., Tetrahedron 55, 11267 (1999).
29. C. Samojłowicz, M. Bieniek, A. Zarecki, R. Kadyrov,
K. Grela, Chem. Commun. 2008, 6282 (2008).
30. S. Isayama, T. Mukaiyama, Chem. Lett. 18, 1071
(1989).
31. D. H. R. Barton, R. E. O’Brien, S. Sternhell, J. Chem. Soc.
1962, 470 (1962).
32. G. J. Hollingworth, G. Perkins, J. Sweeney, J. Chem. Soc.
Perkins Trans. 1996, 1913 (1996).
33. A. Fürstner et al., Chem. Commun. 2008, 2873 (2008).
34. S.-K. Chung, J. Org. Chem. 44, 1014 (1979).
35. D. H. R. Barton, J. D. Elliott, S. D. Géro, J. Chem. Soc.,
Perkin Trans. 1 1982, 2085 (1982).
36. T. J. Barker, D. L. Boger, J. Am. Chem. Soc. 134, 13588
(2012).
37. E. Yoshii, T. Oribe, K. Tumura, T. Koizumi, J. Org. Chem.
43, 3946 (1978).
38. E. J. Corey, X.-M. Cheng, The Logic of Chemical Synthesis
(Wiley, New York, 1989), p. 17.
39. P. S. Baran, T. J. Maimone, J. M. Richter, Nature 446,
404 (2007).
40. T. Newhouse, P. S. Baran, R. W. Hoffmann, Chem. Soc.
Rev. 38, 3010 (2009).
41. H. Yeom, S. G. Sligar, H. Li, T. L. Poulos, A. J. Fulco,
Biochemistry 34, 14733 (1995).
Acknowledgments: Financial support for this work was
provided by LEO Pharma and an unrestricted grant from TEVA.
Early studies were supported by the NIH/NCI (CA134785).
Bristol-Myers Squibb supported a graduate fellowship to
H.R., and the Shanghai Institute of Organic Chemistry
supported a postdoctoral fellowship to Q.Z. We are grateful to
A. Rheingold and C. E. Moore (University of California, San
Diego) for x-ray crystallographic analysis and G. Siuzdak (the
Scripps Research Institute) for mass spectrometry assistance.
We thank J. Felding, H. Bladh, and R. Shenvi for valuable
discussions. Metrical parameters for the structures of
compounds 7 and 11′ are available free of charge from the
Cambridge Crystallographic Data Centre under reference
numbers CCDC-909412 and 909413, respectively.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/59/DC1
Materials and Methods
Figs. S1 and S2
Tables S1 to S17
Reference (42)
24 September 2012; accepted 9 November 2012
10.1126/science.1230631
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 63
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Hydrogen-Nitrogen Greenhouse
Warming in Earth's Early Atmosphere
Robin Wordsworth* and Raymond Pierrehumbert
Understanding how Earth has sustained surface liquid water throughout its history remains a key
challenge, given that the Sun’s luminosity was much lower in the past. Here we show that with
an atmospheric composition consistent with the most recent constraints, the early Earth would have
been significantly warmed by H
2
-N
2
collision–induced absorption. With two to three times the
present-day atmospheric mass of N
2
and a H
2
mixing ratio of 0.1, H
2
-N
2
warming would be
sufficient to raise global mean surface temperatures above 0°C under 75% of present-day solar
flux, with CO
2
levels only 2 to 25 times the present-day values. Depending on their time of emergence
and diversification, early methanogens may have caused global cooling via the conversion of H
2
and CO
2
to CH
4
, with potentially observable consequences in the geological record.
O
ne of the most durable questions about
Earth’s early climate arises fromthe faint
young Sun effect: Because progressive
accumulation of He in a star’s core causes its lu-
minosity to increase with age (1), the solar energy
incident on Earth was significantly lower [∼75%
of present-day values 3.8 billion years ago (Ga)]
during the Hadean and Archean eras (2). Because
geological evidence shows that Earth was not in
a globally glaciated, snowball state throughout
this time (3), additional mechanisms must have
been present to warm the climate.
Previous explanations for this altered climate
have included increased atmospheric ammonia
or CH
4
, a decreased surface albedo, and changes
in the distribution of clouds (2, 4, 5). However,
all of these mechanisms have subsequently been
shown to suffer important defects (6, 7). Increased
atmospheric CO
2
is one plausible solution be-
cause of the climate buffer provided by the crust-
al carbonate-silicate cycle (8), but efficient mantle
CO
2
cycling in the Hadean and Archean prob-
ably reduced the strength of this feedback (9).
Interpretation of the geological record is difficult,
but analyses of mid- to late Archean paleosols
suggest between ~10 and 50 times the present
atmospheric level (PAL) of CO
2
(7, 10). Even
lower values (around three times PAL) have been
derived from analysis of magnetite and siderite
equilibria in Archean banded iron formations, al-
though the underlying assumptions behind these
limits have been questioned (5, 7).
One recently proposed mechanism for coun-
teracting the faint young Sun is pressure-induced
broadening of the absorption lines of existing
greenhouse gases due to increased atmospheric
N
2
, which alone should cause a warming of be-
tween 3 and 8 K(11). Recent attempts to estimate
atmospheric density fromfossil raindrop imprints
suggest a value near that of the present day around
2.7 Ga, with an upper limit of double the present
value (12). However, the mantle has a large N
inventory that is correlated with radiogenic
40
Ar
but not primordial
36
Ar, indicating that it orig-
inated from subducted crust, where it had most
likely been fixed biologically. It is therefore like-
ly that on the pre-biotic/early Archean Earth, the
atmospheric N
2
content was around two to three
times the present-day value (11).
Hydrogen, the most abundant gas in the so-
lar system, has previously been ignored in the
Archean climate budget, presumably because it
was long thought to be a minor constituent
even in the early atmosphere (13). After the initial
loss of Earth’s primordial hydrogen envelope,
the mixing ratio of H
2
in the atmosphere was
primarily determined by a balance between
outgassing, surface/ocean chemistry, and escape
to space. It was long believed that the escape of
H
2
in the Archean was rapid, with diffusion from
the lower atmosphere to the exobase the limiting
factor, as it is today. However, recent numerical
calculations imply that the rate of hydrodynamic
H
2
escape on the early Earth was more strongly
constrained by the adiabatic cooling of the es-
caping gas, given a limited extreme ultraviolet
(XUV) energy input (14–16). As a result, H
2
could have been a major constituent (up to ∼30%
by volume) of the Archean atmosphere unless
surface or ocean biogeochemistry continuously
removed it.
A single H
2
molecule is homonuclear and
diatomic and hence has no direct vibrational or
rotational absorption bands. Nonetheless, mo-
Department of Geological Sciences, University of Chicago,
Chicago, IL 60637, USA.
*To whom correspondence should be addressed. E-mail:
rwordsworth@uchicago.edu
A
B
Fig. 1. (A) Infrared absorption at the surface for an Archean atmosphere with 3 × PAL N
2
, 10% H
2
,
1700 ppm CO
2
(∼10 × PAL), 5 ppm CH
4
, surface temperature 273 K, and relative humidity 0.77. k
v
,
absorptivity per unit of length; v, wavenumber. (B) Corresponding outgoing longwave radiation
(OLR), assuming line absorption only (gray line) and with CIA included (red line). F
v
, flux per unit
wavenumber. The black line shows the blackbody emission at 273 K, and the numbers in the inset
show the integrated OLR for each case. See the supplementary materials for calculation details. Given
high atmospheric H
2
and N
2
levels, H
2
-N
2
absorption dominates in the 750 to 1200 cm
−1
spectral
region, where the bands of the other greenhouse gases are relatively weak.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 64
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

lecular hydrogen interacts strongly with infra-
red radiation via collision-induced absorption
(CIA), the strength of which scales with the
product of the densities of the two interacting
gases. CIA has been well studied for the gas
giant planets and Titan, where it dominates ra-
diative transfer in the middle and lower portions
of the atmosphere (17, 18). On early Earth, N
2
and H
2
may both have been abundant in the
atmosphere, so interacting pairs of N
2
-N
2
, H
2
-N
2
,
and H
2
-H
2
should all be considered as potential
contributors to greenhouse warming.
Figure 1A shows the H
2
and N
2
CIA bands
alongside absorption bands of the other main
greenhouse gases in the primitive atmosphere.
Although N
2
-N
2
dimer absorption is strong in
the 0 to 300 cm
−1
part of the spectrum, it is over-
whelmed by H
2
O absorption and is too far from
the peak of blackbody emission at terrestrial tem-
peratures to have a significant effect. H
2
-H
2
ab-
sorbs over a much broader spectral range, but
the dependence of CIA on the squared density
of the gas means that even for the upper-limit
estimates of H
2
mixing ratios, its overall effect is
small. The N
2
-H
2
band, however, absorbs strong-
ly, and it peaks near the critical 750 to 1200 cm
−1
window region, where the contribution of all
the other gases to opacity is small.
To investigate the climatic effects of this
previously neglected opacity, we performed one-
dimensional (1D) radiative-convective simula-
tions for a range of plausible Archean atmospheric
compositions. We used a correlated-k, two-stream
radiative transfer scheme (19) with 80 bands in the
infrared and 36 bands in the visible. Temperature-
pressure profiles were created over 80 vertical
levels, with moist adiabatic adjustment assumed
when the temperature profile became unstable to
convection. A solar flux of 75% of the present-
day level (1024.5 W m
−2
) was assumed, and the
surface albedo was set to 0.23 (20).
Figure 1B shows the resulting decrease in
outgoing longwave radiation (OLR) in each band
due to the presence of H
2
for an atmosphere
with the same composition as shown in Fig. 1A.
The OLR is significantly reduced in the 750 to
1200 cm
−1
windowregion and around 500 cm
−1
,
whereas minor increases occur around the 15-mm
CO
2
absorption band because of the decrease in
atmospheric lapse rate as H
2
levels increase. Be-
cause the H
2
-N
2
absorption spectrum overlaps
with those of water vapor and CO
2
, radiative
forcing fromH
2
-N
2
is highest when these gases
are scarce. The radiative forcing also increases
strongly with the atmospheric N
2
content because
of the dependence of CIA on the density of each
interactant. At the PAL of N
2
, H
2
-N
2
CIA pro-
vides only a few watts per square meter, whereas
for temperatures of 280 K or less in the “high”
warming scenario with PALCO
2
, radiative forcing
exceeds 24 W m
−2
(Table 1 and Fig. 2B) (21).
To determine the increase in equilibrium sur-
face temperature under this forcing, we next ran
the radiative-convective model in time-stepping
mode. Figure 2, C to E, shows that when atmo-
spheric N
2
is elevated, higher H
2
mixing ratios
cause significant increases in surface temper-
ature. At PAL N
2
, H
2
-N
2
CIAcauses at most 2 to
3 Kof warming, but this rises to 10 to 15 Kas N
2
levels are increased. Given a H
2
mixing ratio of
10%and 2 × PAL N
2
, mean surface temperatures
exceed the melting point of water for CO
2
levels
∼25 × PAL. At 3 × PAL N
2
, only ∼2 × PAL CO
2
is required.
To what extent could H
2
-N
2
warming have
influenced the evolution of Earth’s early climate?
We can begin by considering the limiting case
where no significant biological alteration of the
atmosphere occurred. Then, the H
2
abundance
in the early atmosphere can be determined by a
Table 1. Definitions of the atmospheric scenarios used in Figs. 2 and 3.
Scenario N
2
mass column (kg m
−2
) H
2
volume mixing ratio (mol/mol)
Present-day 7.80 × 10
3
0.0
Moderate 1.56 × 10
4
0.05
High 2.34 × 10
4
0.1
A
B
C
D
E
Fig. 2. Radiative forcing due to the presence of H for (A) moderate
and (B) high H
2
-N
2
warming scenarios (Table 1), as a function of surface
temperature and atmospheric CO
2
partial pressure. The calculated sur-
face temperature in thermal equilibrium as a function of atmospheric
H
2
volume mixing ratio and CO
2
partial pressure, for a reduced solar flux
F = 0.75 present-day solar flux (F
0
) and total atmospheric N
2
amounts
(C) one, (D) two, and (E) three times that of Earth today. Results are
shown assuming no atmospheric CH
4
and a surface albedo of 0.23.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 65
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

simple balance between volcanic outgassing and
escape to space, based on atmospheric redox
balance in the absence of significant organic burial
(22). Abiotic CO
2
levels can then be estimated by
assuming that the carbonate-silicate temperature-
weathering feedback (8) operated on 1- to 10-
million-year time scales, driving atmospheric CO
2
downward to a level dependent on the solar flux
and the amount of H
2
and N
2
in the atmosphere
(23). Figure 3 shows estimated CO
2
concentra-
tions under 75% of the present-day solar flux for
various CO
2
outgassing rates, assuming no bio-
logical alteration of the atmosphere. Seafloor
weathering is neglected in the calculation (9),
so the CO
2
levels shown are most likely over-
estimated. Even so, it is clear that high H
2
and
N
2
could easily have reduced atmospheric CO
2
to below 100 × PAL even under 75% solar flux.
After the emergence of life, the extent of
biological alteration of the atmosphere was de-
pendent on the energy source and net produc-
tivity of the dominant ecosystem. Of particular
interest in the context of H
2
-rich atmospheres are
methanogens, which feed on the chemical energy
released by combining CO
2
and H
2
in the
reaction CO
2
+ 4H
2
→ CH
4
+ 2H
2
O. Phyloge-
netic analyses suggest that these organisms
evolved at some point in the mid-Archean, with
hyperthermophilic methanogenesis preceding the
mesophilic adaptation by at least several hun-
dred million years (24, 25). Although in small
quantities CH
4
causes warming, its potential as a
greenhouse gas is limited by the separation of its
main absorption band fromthe peak of blackbody
absorption (Fig. 1A) and its conversion to
photochemical haze for [CH
4
]/[CO
2
] ratios
greater than ∼0.1, which tends to have a
cooling effect (6).
For methanogens to have increased surface
temperatures after their emergence, therefore,
the biological conversion of CO
2
and H
2
to CH
4
must have balanced CH
4
photolysis due to the
solar UV flux at relatively low atmospheric CH
4
levels. Estimates of the CH
4
photolysis rate un-
der the elevated UV/Lyman-a fluxes expected
in the Archean are in the range 2 × 10
11
to 5 ×
10
11
cm
−2
s
−1
for hydrogen-rich atmospheres
(26). For comparison, the present-day biogenic
CH
4
flux is on the order of 1 × 10
11
cm
−2
s
−1
(26). Ecosystemmodels that assume no nutrient
or climate constraints for biological productivity
yield steady-state [CH
4
]/[H
2
] ratios of around
0.03 if H
2
is initially abundant, implying an
extremely hazy atmosphere and little warming
from H
2
-N
2
CIA (26). However, in the early
Archean, the reduced continental area and in-
creased ocean volume likely made nutrient and
habitat limitations significantly more stringent
than they are today (27).
If habitat limitations were not important,
emergent methanogens in an early climate dom-
inated by H
2
-N
2
and CO
2
warming would have
continued to consume H
2
and CO
2
rapidly until
global cooling became the limiting factor on
biological productivity (28). In the most drastic
scenario, global glaciation would result, followed
by a buildup of CO
2
and H
2
levels until the cli-
mate warmed again. Earth was ice-free through-
out most of the Archean, but geological evidence
exists for an early glaciation event 2.8 to 2.9 Ga
(29), which may have been caused by the rise of
methanogenesis. Understanding the detailed ef-
fects of the early biosphere on climate requires
3D climate simulations coupled with models of
the (local) dependence of ecosystem productivity
on surface temperature and nutrient availability.
This is an important topic for future study.
Methanogens are often assumed to have been
an integral part of the early Archean ecosystem
because a CH
4
greenhouse was believed neces-
sary to solve the faint young Sun problem. In
contrast, our results show that an early climate
dominated by abiotic H
2
-N
2
and CO
2
warming
is consistent with both observational and theoret-
ical limits on atmospheric CO
2
levels and phylo-
genetic analyses suggesting later diversification
of the Archaea. H
2
-N
2
warming is also likely to
be important in the search for biosignatures on
super-Earth exoplanets, whose higher masses im-
ply lower energy-limited hydrogen escape rates
and larger typical atmospheric N
2
inventories. Be-
cause incident XUV flux is a function of orbital
distance, H
2
-N
2
warming may be of particular
importance to the habitability of terrestrial exo-
planets that are far from their host stars.
References and Notes
1. D. O. Gough, Sol. Phys. 74, 21 (1981).
2. C. Sagan, G. Mullen, Science 177, 52 (1972).
3. W. H. Peck, J. W. Valley, S. A. Wilde, C. M. Graham,
Geochim. Cosmochim. Acta 65, 4215 (2001).
4. A. A. Pavlov, L. L. Brown, J. F. Kasting, J. Geophys. Res.
106, 23267 (2001).
5. M. T. Rosing, D. K. Bird, N. H. Sleep, C. J. Bjerrum,
Nature 464, 744 (2010).
6. J. D. Haqq-Misra, S. D. Domagal-Goldman, P. J. Kasting,
J. F. Kasting, Astrobiology 8, 1127 (2008).
7. G. Feulner, Rev. Geophys. 50, RG2006 (2012).
8. J. C. G. Walker, P. B. Hays, J. F. Kasting, J. Geophys. Res.
86, 9776 (1981).
9. N. H. Sleep, K. Zahnle, J. Geophys. Res. 106, 1373
(2001).
10. S. G. Driese et al., Precambrian Res. 189, 1 (2011).
11. C. Goldblatt et al., Nat. Geosci. 2, 891 (2009).
12. S. M. Som, D. C. Catling, J. P. Harnmeijer, P. M. Polivka,
R. Buick, Nature 484, 359 (2012).
13. For exosolar planets, it has been suggested that H
2
could enable planetary habitability in certain cases, but
only in quantities of 1 bar or more due to direct H
2
-H
2
CIA (30–32). This mechanism was considered and
dismissed for the early Earth in (2).
14. A. J. Watson, T. M. Donahue, J. C. G. Walker, Icarus 48,
150 (1981).
15. F. Tian, O. B. Toon, A. A. Pavlov, H. De Sterck, Science
308, 1014 (2005).
16. The calculations in (15) did not take nonthermal escape
or heating due to heavier gases into account, leading to
some claims that they underestimated the total H
2
escape
(33). However, nonthermal escape is only dominant on
the present-day Earth and should have been limited in
the Archean (34), and infrared emission from radiatively
active neutral species such as CO
2
causes significant
cooling of the thermosphere. A final assessment of the
hydrogen escape rates from Earth’s early atmosphere
awaits development of 3D multicomponent hydrodynamic
escape models with coupled radiative transfer and
chemistry.
17. A. Borysow, L. Frommhold, Astrophys. J. 303, 495
(1986).
18. C. P. McKay, J. B. Pollack, R. Courtin, Science 253,
1118 (1991).
19. R. Wordsworth et al., Astron. Astrophys. 522, A22+ (2010).
20. Materials and methods are available as supplementary
materials on Science Online.
21. Here we use the term PAL to refer to present-day
atmospheric partial pressure for CO
2
but present-day
atmospheric mass for N
2
. For all simulations, we first fix
the mixing ratios of each gas and the total mass column
of N
2
, M
N
2
. We then calculate the total atmospheric
pressure as p
tot
¼ gM
N2
m=m
N2
f
N2
, where g is gravity,
m is the mean molar mass, and f
N
2
and m
N
2
are the
volume mixing ratio and molar mass of N
2
, respectively.
The partial pressure of CO
2
is then simply p
tot
f
CO
2
.
22. J. F. Kasting, D. E. Canfield, The Global Oxygen Cycle
(Wiley, Chichester, UK, 2012), pp. 93–104.
23. To produce Fig. 3 we expressed the CO
2
carbonate-silicate weathering rate as
W
W0
¼
g
g
0
½1 þ a
P
ðT
surf
− T
0
ފ
a
ð
pCO
2
pCO
2
0
Þ
b
expð
Tsurf − T0
TU
Þ (35).
We then set W equal to the assumed CO
2
outgassing rate
Fig. 3. Atmospheric CO
2
partial pressure, assuming
an abiotic carbonate-silicate
weathering feedback, no
seafloor weathering, and
a solar flux F =0.75F
0
. The
H
2
and N
2
levels for the
present-day, moderate, and
high scenarios are given in
Table 1. Results are shown
for three values of the CO
2
outgassing rate G
CO
2
as a
function of the present-
day rate.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 66
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

G
CO
2
and calculated T
surf
as a function of the CO
2
partial
pressure p
CO
2
using results from the radiative-convective
model. See the supplementary materials for a definition of
all terms and justification of the method.
24. C. H. House, B. Runnegar, S. T. Fitz-Gibbon, Geobiology
1, 15 (2003).
25. C. E. Blank, Geobiology 7, 495 (2009).
26. P. Kharecha, J. Kasting, J. Siefert, Geobiology 3, 53 (2005).
27. Modern ocean net primary productivity is concentrated in
shallow regions close to continents (26), but in the early
Archean, the continental crust volume was about three times
lower and the ocean volume may have been up to 25%
greater than today (36, 37). In the deep ocean away from
submarine vents, rates of H
2
supply would be reduced by
∼10
3
as compared to the mixed layer, implying a decrease
of biological productivity there by a similar factor (26).
28. Indirectly, abundant methanogenesis could also have
caused global cooling via drawdown of atmospheric N
2
.
In atmospheres with 1000 parts per million (ppm) CH
4
,
atmospheric HCN production rates via photolysis can
reach 1 × 10
10
cm
−2
s
−1
(38). Without a mechanism to
reform N
2
, this could cause the removal of the entire
present-day atmospheric N
2
inventory on a time scale on
the order of 100 million years.
29. G. M. Young, V. von Brunn, D. J. C. Gold, W. E. L. Minter,
J. Geol. 106, 523 (1998).
30. D. J. Stevenson, Nature 400, 32 (1999).
31. R. Pierrehumbert, E. Gaidos, The Astrophysical Journal
Letters 734, L13 (2011).
32. R. Wordsworth, Icarus 219, 267 (2012).
33. D. C. Catling, Science 311, 38 (2006).
34. F. Tian, O. B. Toon, A. A. Pavlov, Science 311, 38 (2006).
35. D. S. Abbot, N. B. Cowan, F. J. Ciesla, Astrophys. J. 178,
756 (2012).
36. B. Dhuime, C. J. Hawkesworth, P. A. Cawood, C. D. Storey,
Science 335, 1334 (2012).
37. E. C. Pope, D. K. Bird, M. T. Rosing, Proc. Natl. Acad.
Sci. U.S.A. 109, 4371 (2012).
38. F. Tian, J. F. Kasting, K. Zahnle, Earth Planet. Sci. Lett.
308, 417 (2011).
Acknowledgments: Climate calculations were performed on
the iDataPlex computational cluster of the Université
de Paris 6, France. R.W. thanks S. Lewis, J. Waldbauer,
C. Goldblatt, N. Dauphas, M. Coleman, and D. Abbot for
insightful discussions.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/64/DC1
Materials and Methods
References
7 June 2012; accepted 5 November 2012
10.1126/science.1225759
Highly Variable El Niño–Southern
Oscillation Throughout the Holocene
Kim M. Cobb,
1
* Niko Westphal,
2
† Hussein R. Sayani,
1
Jordan T. Watson,
2
Emanuele Di Lorenzo,
1
H. Cheng,
3,4
R. L. Edwards,
4
Christopher D. Charles
2
The El Niño–Southern Oscillation (ENSO) drives large changes in global climate patterns from
year to year, yet its sensitivity to continued anthropogenic greenhouse forcing is uncertain. We
analyzed fossil coral reconstructions of ENSO spanning the past 7000 years from the Northern
Line Islands, located in the center of action for ENSO. The corals document highly variable ENSO
activity, with no evidence for a systematic trend in ENSO variance, which is contrary to some
models that exhibit a response to insolation forcing over this same period. Twentieth-century
ENSO variance is significantly higher than average fossil coral ENSO variance but is not unprecedented.
Our results suggest that forced changes in ENSO, whether natural or anthropogenic, may be
difficult to detect against a background of large internal variability.
T
he relative strength of the El Niño–Southern
Oscillation (ENSO) phenomenon remains
one of the most prominent uncertainties
in general circulation model (GCM) projections
of future climate change (1). ENSO is respon-
sible for much of the interannual temperature
and precipitation variability across the globe and
thus influences energy and water use, ecosystem
dynamics, and human health. Consequently, the
broad range of ENSO projections represents a
fundamental challenge for the development of
meaningful climate change adaptation strategies.
The uncertainty that arises from comparison of
various models is compounded by the fact that
instrumental records of ENSO are not sufficient-
ly long to test the accuracy of any given model
performance; these records are simply not long
enough to provide robust estimates of natural
ENSO variability.
Paleo-ENSO records can provide the neces-
sary tests of GCM performance by tracking the
response of ENSO to a variety of past natural cli-
mate forcings. One prominent example of this
approach comes fromthe mid-Holocene [~6 thou-
sand years (ky) ago] (2), an interval when both
GCM simulations (3–6) and paleo-ENSO recon-
structions (7–10) showreduced ENSOvariability,
associated with changes in insolation forcing (11).
Such data-model agreement seemingly gives
credence to the GCMs’ abilities to simulate forced
changes in ENSO. However, the paleoclimate
evidence for a mid-Holocene reduction of ENSO
variability is limited—it comprises millennia-long
lake or marine sediment records that lack the res-
olution required to resolve ENSOexplicitly (8–10),
along with several fossil coral sequences that are
seasonally resolved but short (7). Furthermore,
the majority of these records rely on ENSO
precipitation responses that may have changed
through time.
We analyzed the ENSO variability contained
in a collection of monthly resolved, uranium/
thorium (U/Th)–dated fossil coral records span-
ning the past 7 ky from the central tropical Pa-
cific. The archive roughly triples the amount of
fossil coral data available to assess ENSO evo-
lution through this time interval. The fossil corals
come fromChristmas (2°N, 157°W) and Fanning
(4°N, 160°W) Islands, located in the Northern
Line Islands, which are the site of numerous
high-fidelity coral-based reconstructions of 20th-
century ENSOspanning the past century (12–14)
and millennium (15). These records allow for
quantitative estimates of ENSOvariance through
time—estimates that can be used to gauge the
magnitude of potential forced changes in ENSO
variance, both natural and anthropogenic, with
respect to natural ENSO variability. Such es-
timates of natural ENSO variability also pro-
vide particularly valuable tests of the long-term
ENSO variability exhibited in millennia-long
GCM simulations.
Previous work has demonstrated the accuracy
and reproducibility of paleo-ENSO reconstruc-
tions constructed by using modern and fossil
corals fromPalmyra Island (6°N, 162°W), which
is just north of Christmas and Fanning Islands
(15). U/Th ages for the 17 fossil coral sequences
(7 fromFanning and 10 fromChristmas) presented
in this study range from 1.3 to 6.9 ky (table S1).
Scanning electron microscopy photos reveal evi-
dence for extremely heterogeneous levels of dia-
genesis, with trace to moderate alteration (<5%
by weight) sometimes visible in the same coral
(fig. S1 and table S2). Like the modern corals, the
fossil coral cores were processed following stan-
dard procedures and sampled at 1-mm intervals
for oxygen isotopic (d
18
O) analyses [long-term
reproducibility of d
18
O is T 0.07 per mil (1s)]
(16). The fossil corals’ growth rates range from 8
to 20 mm/year, and the records range in length
from 19 to 81 years (fig. S2 and table S2).
Northern Line Island coral d
18
O records re-
flect changes in sea-surface temperature (SST) as
well as the d
18
O of seawater (the latter linearly
correlated to sea-surface salinity) (17), both of
which change during ENSO extremes. El Niño
events bring warmer and rainier conditions to the
Northern Line Islands, conditions which construc-
tively act to decrease coral d
18
O (12–14). The
opposite climatic effects occur during cool La
Niña extremes, driving an increase in coral d
18
O.
Modern Line Islands coral d
18
Orecords are high-
ly correlated to the regional-scale NIÑO3.4 index
(defined as the average of instrumental SSTs from
5°S to 5°N, and from 120° to 170°W) (Fig. 1). As
1
School of Earth and Atmospheric Sciences, Georgia Institute of
Technology, Atlanta, GA 30332, USA.
2
Scripps Institution of
Oceanography, La Jolla, CA 92037, USA.
3
Institute of Global
Environmental Change, Xi’an Jiaotong University, Xi’an 710049,
China.
4
Department of Earth Sciences, University of Minnesota,
Minneapolis, MN 55455, USA.
*To whom correspondence should be addressed. E-mail:
kcobb@eas.gatech.edu
†Present address: Department of Earth Sciences, Eidgenössische
Technische Hochschule, CH-8092 Zürich, Switzerland.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 67
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

expected, the magnitude of ENSO-related coral
d
18
O anomalies scale with the islands’ proximity
to the equator. Therefore, in order to assess relative
changes in ENSO variance across the entire Line
Islands fossil coral collection, we benchmark paleo-
ENSOvariance in a given fossil coral against late
20th century ENSOvariance recorded in the mod-
ern coral d
18
O records from the same island.
Our records reveal that ENSO strength has
varied substantially over the past 7 ky (Fig. 2A
and table S2). In some sequences, interannual
variability is statistically insignificant, with re-
ductions of up to –70% (relative to the late 20th
century) in 2- to 7-year variance. Such dramatic
reductions in ENSO variance occur sporadically
throughout the new collection of fossil corals, in-
dependently of coral age or preservation. The
coral d
18
O reconstructions are broadly consistent
with the range of ENSO variance changes con-
tained in existing fossil coral d
18
O reconstruc-
tions (7, 18, 19). The expanded collection of
fossil coral records shows no systematic shifts in
ENSOvariance through the mid- to late Holocene.
For example, ENSO variance estimates from the
mid-Holocene are statistically indistinguishable
from those of the past millennium (Fig. 3, A and
B). In fact, the only time period that stands out is
the 20th century, when ENSO variance estimates
are significantly higher than ENSO variance es-
timates fromthe Line Islands fossil corals (Fig. 3,
C and D), as assessed with a Monte Carlo–based
approach that uses 10,000 pairs of pseudocoral
time series to assess the significance of observed
differences in 2- to 7-year variance (16). How-
ever, the high late 20th century ENSO variance
is not unprecedented; early 17th century ENSO
variance is the highest of the entire Line Islands
1
8
O

A
n
o
m
a
l
y

(

)

S
S
T

A
n
o
m
a
l
y

(
°
C
)

2
1
0
-1
-2
Year
-0.4
-0.2
0
0.2
0.4
1900 1920 1940 1960 1980 2000
Fig. 1. ENSO variability in Line Islands modern coral d
18
O records and the NIÑO3.4 index [Extended
Reconstructed Sea Surface Temperatures, version 3 (ERSST v.3b)] (29). Coral d
18
Odata are fromChristmas
[spliced record constructed from data in (12, 14, 18, 30); (fig. S3)]; Fanning [spliced record constructed
fromdata in (14) and a new Fanning modern d
18
O record (16)] (fig. S4); and Palmyra (13). Each monthly
resolved time series has been 2- to 7-year bandpass filtered to highlight ENSO-related variability.
Correlation coefficients computed between the bandpassed coral records and the NIÑO3.4 SST index
are –0.92, –0.85, and –0.82 for Christmas, Fanning, and Palmyra, respectively.
A B C D
E
Fig. 2. Overview of new and published ENSO reconstructions
spanning the past 7.5 ky. (A) Relative ENSO variance changes in
fossil coral d
18
O time series, calculated from sliding 30-year win-
dows (with a 5-year step) of the SD of the 2- to 7-year band,
plotted as percent differences from 1968–1998 C.E. intervals of
corresponding modern coral d
18
O time series from each site
[modern benchmark intervals for non–Line Island corals differ by
T5 years (table S3)]. Coral time series with 20 years < length < 30 years are plotted with cross-filled circles. Coral time series with 10 years < length < 20 years
are plotted with open circles. Fossil coral d
18
O data are from Christmas Island (green) [this study and (31)]; micro-atolls (pink) (18); Fanning Island (blue) (this
study); Palmyra (red) (15); and Papua New Guinea [yellow (7) and gray (19)]. (B) Relative ENSO variance changes over the 20th century as recorded in modern
coral d
18
O records and the NIÑO3.4 SST index (ERSSTv.3b) (29), calculated from sliding 30-year windows (with a 5-year step) of the SD of the 2- to 7-year
bandpassed time series, plotted as percent differences from the 1968–1998 intervals of the records [modern benchmark intervals for non–Line Island corals
differ by T5 years (table S3)]. Palmyra fossil coral NB9 (15) is replotted here (at date 1926 C.E.) for comparison with the Palmyra modern coral ENSO strength
during the early 20th century. Modern coral d
18
O data are from Palmyra (red) (13); Fanning, (blue) (fig S4); Christmas Island (green) (fig. S3); and Papua New
Guinea [(yellow (7) and gray (19)]. (C) Relative ENSO variance changes in the NCAR CCSM4 1 ky preindustrial control run (23), calculated from sliding 30-year
windows (with a 5-year step) of the SD of the 2- to 7-year bandpassed annual NIÑO3.4 time series, plotted as percent differences from the average SD of the
entire 2- to 7-year bandpassed model NIÑO3.4 time series. (D) Relative ENSOvariance changes in the GFDL CM2.1 2 ky preindustrial control run (25), calculated
fromsliding 30-year windows (with a 5-year step) of the SD of the 2- to 7-year bandpassed monthly NIÑO3.4 time series, plotted as percent differences fromthe
average SD of the entire 2- to 7-year bandpassed model NIÑO3.4 time series. (E) Records fromEcuadorian Lake Pallcacocha (red; plotted as an 80-point smooth
of the raw red intensity record) (8) and Galapagos Lake El Junco (black; plotted as percent sand fraction) (10).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 68
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

reconstruction. The choice of a late 20th century
benchmark makes the fossil coral ENSOstrengths
appear anomalously weak, when in fact it is the
late 20th century ENSO strength that is anoma-
lously strong.
The coral-based ENSOvariance estimates dif-
fer appreciably from several lower-resolution re-
cords fromthe eastern equatorial Pacific that have
been used to infer sizable mid-Holocene reduc-
tions in ENSO variance (Fig. 2E). These archives
include two lake sediment records that record
fewer El Niño–related flood events in the mid-
Holocene (8, 10, 20) as well as a marine sediment
record that exhibits a drop in foraminifera pop-
ulation variance (reflective of a decrease in annual
and/or interannual SSTextremes) during the mid-
Holocene (9). Differences in the attributes of the
various archives (climate sensitivities, chronol-
ogy, resolution, and continuity) make it difficult
to fully reconcile the divergent views of the Hol-
ocene evolution of ENSO. However, instrumental
observations suggest that it is possible for east-
ern Pacific precipitation anomalies to be muted
while central Pacific ENSO variance is relatively
strong (21), and the various mid-Holocene ENSO
proxies may reflect such geographic complexities.
Whatever their origin, the differences underscore
the need for rigorous comparison among differ-
ent archives, ideally facilitated by translation of
tropical Pacific climate variability simulated by
models into explicit expectations for proxies (22).
The expanded coral archive does allow for a
quantitative assessment of the long-term ENSO
variability exhibited by coupled GCMs. The broad
range of ENSO variance changes exhibited by
the fossil coral database is also characteristic of
ENSO variance changes in long, unforced integra-
tions of the National Center for Atmospheric Re-
search (NCAR)–Community Climate SystemModel
version 4 (CCSM4) (23) and National Oceanic and
Atmospheric Administration (NOAA)–Geophysical
FluidDynamics Laboratory(GFDL)–Climate Model
2.1 (CM2.1) (24, 25) fully coupled GCMs (Fig. 2,
C and D, respectively). In the model simulations,
NIÑO3.4 variance changes by T40% (2s); some
centuries-long epochs exhibit large changes in
ENSO strength, whereas other epochs feature a
relatively stable ENSO. The ENSO variance esti-
mates from the Line Islands fossil corals exhibit a
similar range of variability, fluctuating by T36%
(2s) about an average of –42% (with respect to
the late 20th century coral benchmark). In fact,
the fossil coral ENSOvariance distribution is con-
sistent with the ENSO variance distribution in
models (Fig. 3F), as assessed by subsampling the
model time series to match the fossil coral lengths
by using a Monte Carlo–based approach (16).
The actual Line Islands fossil coral distribution is
contained by the spread of the 10,000 pseudocoral
realizations, falling well within the 2s limits of
the model distribution.
We conclude that the model data provide a
reasonably satisfactory representation of the ENSO
variance distribution in the real fossil coral data.
The convergence between data-based and model-
based estimates of long-term, intrinsic ENSOvar-
iability suggests that at least some of the current
generation of models may accurately simulate
long-termchanges in ENSOvariance, even with-
out the inclusion of solar and volcanic natural
climate forcings in the models. This data-model
agreement suggests that either (i) the models over-
estimate unforced ENSO variability or (ii) the
natural radiative forcing had little influence on
paleo-ENSO variance. In any case, one main im-
plication of both the coral- and model-based esti-
mates of intrinsic ENSO variability is that the
detection (and attribution) of any changes in ENSO
properties would require very long time series
spanning many centuries, to the extent that de-
tection of such changes is even possible (26).
Taken together, the Line Islands fossil coral
data suggest that much of the observed differences
A B
C
D
E F
Fig. 3. Comparison of probability density functions of ENSO strength in Line Islands coral data from
different time periods, and in model output. (A) Distributions of relative ENSO variance changes in Line
Islands corals in the 6- to 7-ky interval (black, n = 193 years), as compared with the 0- to 1-ky interval
(white, n = 475 years). (B) Significance test of calculated difference between the SD of 2- to 7-year
variance in 6- to 7-ky Line Islands corals versus 0- to 1-ky Line Islands corals, as benchmarked using
differences calculated from 10,000 “pseudocoral” data sets constructed from the 2-ky-long GFDL CM2.1
NIÑO3.4 time series (16). Red bar indicates the calculated difference between the mean SDs of the 6- to
7-ky and 0- to 1-ky intervals, as compared with the threshold for pseudocoral SD differences significant at
the 95% level (black bar). (C) Distribution of relative ENSO variance changes in the Line Islands modern
coral data (white; the three islands’ relative ENSO variance changes as plotted in Fig. 2B have been
averaged together to form a composite Line Islands modern coral; n = 112 years) and fossil coral data
(black; compiled from Fig. 2A; n = 990 years). The red arrow marks the late-20th-century ENSO variance
benchmark (0%by definition) for the Line Islands corals. (D) Same as (B), but for the modern versus fossil
coral distributions in (C). (E) Distribution of relative ENSO variance changes in the unforced NCAR
CCSM4
23
(black) and GFDL CM2.1
25
(white) control runs, compiled from Fig. 2, C and D, respectively. (F)
Plot of distribution of relative changes of ENSO variance in actual Line Islands fossil coral data (red), as
compared with the distribution of 10,000 “pseudo fossil coral” databases compiled by extracting 990 years
of GFDL model output in segment lengths corresponding to the actual fossil coral lengths (gray, envelope
of individual pseudocoral ensemble distributions; black, pseudocoral mean with 1s bars), and the dis-
tribution of relative changes of ENSO variance fromthe full 2-ky-long NIÑO3.4 time series fromthe GFDL
model (blue).
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 69
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

in ENSOvariance over the past 7 ky reflect strong
internal variability. The fact that we detect no dis-
cernible influence of orbitally induced insola-
tion forcing on ENSO is noteworthy, given that
the effect of insolation forcing on global mon-
soon circulations is well documented (27, 28).
Relatively robust 20th-century ENSO variability
may reflect a sensitivity to anthropogenic green-
house forcing, but definitive proof of such an
effect requires much longer data sets than are
currently available, given the large range of nat-
ural ENSO variability implied by the available
fossil coral data.
References and Notes
1. M. Collins et al., Nat. Geosci. 3, 391 (2010).
2. J. Brown, A. W. Tudhope, M. Collins, H. V. McGregor,
Paleoceanography 23, PA3202 (2008).
3. B. L. Otto-Bliesner, E. C. Brady, S. I. Shin,
Z. Y. Liu, C. Shields, Geophys. Res. Lett. 30,
2198 (2003).
4. Z. Y. Liu, J. Kutzbach, L. X. Wu, Geophys. Res. Lett. 27,
2265 (2000).
5. A. Timmermann, S. J. Lorenz, S. I. An, A. Clement,
S. P. Xie, J. Clim. 20, 4147 (2007).
6. W. Zheng, P. Braconnot, E. Guilyardi, U. Merkel, Y. Yu,
Clim. Dyn. 30, 745 (2008).
7. A. W. Tudhope et al., Science 291, 1511 (2001).
8. C. M. Moy, G. O. Seltzer, D. T. Rodbell, D. M. Anderson,
Nature 420, 162 (2002).
9. A. Koutavas, P. B. deMenocal, G. C. Olive, J. Lynch-Stieglitz,
Geology 34, 993 (2006).
10. J. L. Conroy, J. T. Overpeck, J. E. Cole, T. M. Shanahan,
M. Steinitz-Kannan, Quat. Sci. Rev. 27, 1166
(2008).
11. A. C. Clement, R. Seager, M. A. Cane, Paleoceanography
15, 731 (2000).
12. M. N. Evans, R. G. Fairbanks, J. L. Rubenstone,
J. Geophys. Res. 104, 13409 (1999).
13. K. M. Cobb, C. D. Charles, D. E. Hunter, Geophys. Res. Lett.
28, 2209 (2001).
14. I. S. Nurhati, K. M. Cobb, C. D. Charles, R. B. Dunbar,
Geophys. Res. Lett. 36, L21606 (2009).
15. K. M. Cobb, C. D. Charles, H. Cheng, R. L. Edwards,
Nature 424, 271 (2003).
16. Materials and methods are available as supplementary
materials on Science Online.
17. R. G. Fairbanks et al., Coral Reefs 16, S93 (1997).
18. C. D. Woodroffe, M. R. Beech, M. K. Gagan,
Geophys. Res. Lett. 30, 1358 (2003).
19. H. V. McGregor, M. K. Gagan, Geophys. Res. Lett. 31,
L11204 (2004).
20. D. T. Rodbell et al., Science 283, 516 (1999).
21. K. Ashok, S. K. Behera, S. A. Rao, H. Weng, T. Yamagata,
J. Geophys. Res. 112, C11007 (2007).
22. D. M. Thompson, T. R. Ault, M. N. Evans, J. E. Cole,
J. Emile-Geay, Geophys. Res. Lett. 38, L14706 (2011).
23. C. Deser et al., J. Clim. 25, 2622 (2012).
24. A. T. Wittenberg, A. Rosati, N. C. Lau, J. J. Ploshay,
J. Clim. 19, 698 (2006).
25. A. T. Wittenberg, Geophys. Res. Lett. 36, L12702
(2009).
26. S. Stevenson et al., J. Clim. 25, 2129 (2012).
27. X. Wang et al., Geophys. Res. Lett. 34, L23701 (2007).
28. Y. J. Wang et al., Nature 451, 1090 (2008).
29. T. M. Smith, R. W. Reynolds, T. C. Peterson, J. Lawrimore,
J. Clim. 21, 2283 (2008).
30. H. V. McGregor, M. J. Fischer, M. K. Gagan, D. Fink,
C. D. Woodroffe, Geochim. Cosmochim. Acta 75, 3930
(2011).
31. L. K. Zaunbrecher et al., Paleoceanography 25, PA4212
(2010).
Acknowledgments: The authors acknowledge field support
from the Norwegian Cruise Lines, the National Geographic
Waitt program, and the Palmyra Atoll Research Consortium.
The research was funded by NSF–Division of Ocean Sciences
award 0752091 to K.M.C. and C.D.C., National Natural Science
Foundation of China award 41230524 to H.C., NSF–Division of
Atmospheric Sciences award 1103403 to R.L.E. and H.C., and
U.S. Department of Energy (DOE) award DOE-SC0005597 to
E.D.L. A. Wittenberg and C. Ammann generously provided the
NIÑO3.4 time series from the long control runs of the
NOAA-GFDL-CM2.1 and NCAR-CCSM4 models, respectively.
Data availability: All data and metadata are archived at NCDC
(ftp://ftp.ncdc.noaa.gov/pub/data/paleo/coral/east_pacific/line-
islands2013.txt).
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/67/DC1
Materials and Methods
Figs. S1 to S4
Tables S1 to S3
References
1 August 2012; accepted 9 November 2012
10.1126/science.1228246
The Spatial and Temporal Origin of
Chandelier Cells in Mouse Neocortex
Hiroki Taniguchi,* Jiangteng Lu, Z. Josh Huang†
Diverse g-aminobutyric acid–releasing interneurons regulate the functional organization of
cortical circuits and derive from multiple embryonic sources. It remains unclear to what extent
embryonic origin influences interneuron specification and cortical integration because of
difficulties in tracking defined cell types. Here, we followed the developmental trajectory of
chandelier cells (ChCs), the most distinct interneurons that innervate the axon initial segment of
pyramidal neurons and control action potential initiation. ChCs mainly derive from the ventral
germinal zone of the lateral ventricle during late gestation and require the homeodomain protein
Nkx2.1 for their specification. They migrate with stereotyped routes and schedule and achieve
specific laminar distribution in the cortex. The developmental specification of this bona fide
interneuron type likely contributes to the assembly of a cortical circuit motif.
A
fundamental issue in understanding cor-
tical circuitry concerns the origin of the
identity and diversity of g-aminobutyric
acid–releasing interneurons—basic components
of inhibitory circuit organization and assembly.
Diverse interneurons regulate the delicate balan-
ce and dynamic operation of cortical networks
(1) and are generated from the medial and caudal
ganglionic eminence (MGE and CGE) of the
basal ganglia and the preoptic area (2). However,
there has been continued debate on what con-
stitutes an interneuron type (3) and to what extent
the phenotypic descriptions that are used to em-
pirically distinguish cell populations reflect the
intrinsic biological processes, such as their de-
velopmental specification. This is in part due to
the difficulty in tracking the developmental his-
tory of any distinct interneurons, fromtheir origin
to integration into cortical networks.
Among cortical interneurons, the chandelier
cell (ChC) (4) displays exceptional stereotypy
and specificity: They innervate the axon initial
segment (AIS), site of action potential initiation
of pyramidal neurons (PyNs) (5) and likely rep-
resent a bona fide interneuron type. A single
ChC innervates hundreds of PyNs and may exert
powerful control over the spiking of a PyN en-
semble; thus, together they might constitute a
“basic motif” of cortical circuits. However, since
their discovery nearly four decades ago, the de-
velopmental origin and cortical organization of
ChCs have remained elusive.
The homeodomain protein NKX2.1 is spe-
cifically expressed by MGE progenitors (6).
NKX2.1 regulates the formation of the MGE
(6) and the specification of cortical interneurons
(7). We found that after the MGEmorphologically
flattened by about embryonic day 15 (~E15),
NKX2.1 expression continued at the ventral ger-
minal zone (VGZ) of the lateral ventricle (Fig. 1,
A to C), likely a remnant or extension of MGE,
and persisted into the first postnatal week. Along
the rostral-caudal axis of VGZ, NKX2.1
+
cells
were restricted toward the middle-caudal region
(Fig. 1, A and C); they expressed progenitor and
mitotic markers (Fig. 1C and fig. S1) and in-
corporated 5-bromo-2′-deoxyuridine (BrdU) (8).
We generated an Nkx2.1
CreER
knock-in mouse (8),
with inducible site-specific Cre recombinase (CreER),
to achieve a genetic fate mapping of the NKX2.1
+
VGZ cells (Fig. 1B), focusing on late gestation
and neonatal stages.
A tdTomato red fluorescent protein (RFP)
reporter (9) induced by tamoxifen on E17 in
Nkx2.1
CreER
;Ai9 embryos labeled radial glialike
progenitors along the middle-caudal regions of
VGZ (Fig. 1E). VGZ-derived neurons migrated
along the lateral wall of the ventricle, reaching
the cortex by E18 to postnatal day 0 (P0) (Fig. 1,
C, E, and J). Emerging from the dorsal-lateral
side of the ventricle wall, migrating neurons split
Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724,
USA.
*Present address: Max Planck Florida Institute, One Max Planck
Way, Jupiter, FL 33458, USA.
†To whom correspondence should be addressed. E-mail:
huangj@cshl.edu
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 70
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

into both medial and lateral streams to populate
the ventricular zone of the cortex by P1 (Fig. 1, G
and J). In addition to this route, neurons gener-
ated from the middle-caudal VGZ must also mi-
grate anteriorally, along the ventricle wall and/or
within the VZ of cortex, to populate rostral cor-
tical areas (Fig. 1, A and D). After entering the
cortex, migrating neurons first passed through
the cortical plate, reaching layer1 (L1) by P2 to P3
(Fig. 1H, fig. S2, and movie S1). These cells re-
mained and likely spread in L1 for several more
days before descending back into the cortex,
forming a dense band of cells at the L2-L1 border
(Fig. 1, I and J). There was also a significant pop-
ulation of neurons reaching deep layers between
P3 and P7 (Fig. 1, H to J), although their migra-
tion route and schedule are less clear.
When we assessed the identity of E17 VGZ-
derived neurons after the third postnatal week, a
large majority of RFP
+
cells in L2 turned out to
be ChCs [91.0% in medial prefrontal cortex
(mPFC), 91.8% in cingulate cortex, 71.6% in
somatosensory cortex], which extend highly
characteristic axon arbors with dense “cartridges”
of vertically oriented strings of synaptic boutons
(Fig. 2). Among the RFP
+
nonChCs, some were
parvalbumin
+
(PV
+
) and likely fast-spiking bas-
ket cells, and others were unidentified, as they
were negative for somatostatin, calretinin, and
vasointestinal peptide. Patch-clamp recording of
L2 ChCs in mPFC revealed distinct intrinsic
properties compared with fast-spiking basket in-
terneurons (fig. S3 and tables S1 and S2). Paired
recording of two nearby ChCs (within 100 mm)
demonstrated weak yet extensive (14 of 16 ChC
pairs, or 90%) electrical coupling. In addition, a
single spike in an L2 ChC occasionally also gen-
erated disynaptic glutamatergic inputs both to the
nearby ChC and back to itself (fig. S4). These
results substantiate recent studies (10, 11) and
further suggest that nearby ChCs might activate
each other by firing an intermediate population of
PyNs, at least in cortical slices in vitro. Through
electrical coupling and disynaptic excitation,
nearby ChCs might form a coherent network.
Therefore, in addition to ganglionic eminence
and preoptic area, VGZ of the lateral ventricle
represents a novel source of cortical interneurons
and mostly generates chandelier cells. These
VGZ-derived ChCs showed a striking laminar
pattern, with particular enrichment in upper L2
Fig. 1. Nkx2.1
+
pro-
genitors in the embry-
onic VGZ are a novel
source of cortical inter-
neurons. (A) NKX2.1
expression(red) inmiddle-
caudal regions of the
late embryonic VGZ. Ar-
rows indicate the mi-
gration of VGZ-derived
cells toward the dorsal
and rostral cortex. (B)
A coronal section from
a region indicated in
(A). Red dots represent
VGZ-derived cells mi-
grating toward the cor-
tex. Genetic fate mapping
strategy is depicted to the
right. In an Nkx2.1
CreER
;
Ai9 mouse, Cre-mediated
recombination is induced
by tamoxifen and acti-
vates RFP expression. Str,
striatum. (C) Immunofluo-
rescence showing NKX2.1
protein (arrows) in the
middle-to-caudal region
(middle and right panels)
but not rostral region(open
arrowinleft panel) of E17
VGZ. (D) Nkx2.1
CreER
;Ai9
mice induced at E17 and
examined at E18. RFP-
labeled progenitors were
abundant in the middle-
caudal regions (solid ar-
rows) but not rostral region (open arrow) of VGZ. RFP
+
migrating cells
(arrowheads) were found not only in midcaudal sections but also in the rostral
section. (E) Coronal section of an E18 brain induced by low-dose tamoxifen.
More sparse VGZ progenitors (arrow) and postmitotic cells (arrowheads in
inset) are labeled, giving a clearer view of the migratory stream along the
lateral wall of the ventricle. (F) Colocalization (arrow) of NKX2.1 and Nestin
in VGZ progenitors shown by double immunostaining. (G) Coronal section of
a P0 brain induced at E17. Migrating cells emerge fromthe dorsolateral wall
of the ventricle (star) and split into lateral and medial streams (arrowheads)
within the cortical subventricular zone. (H) Coronal section of a P2 brain
induced at E17. Migrating cells, each with a characteristic leading process
(arrowheads in insets) pass through cortical plate into L1 (arrow). (Insets)
Tangentially migrating cells in the subventricular and intermediate zone
(yellow arrowheads, H1) and radially migrating cells in the cortical plate
(light gray arrowheads, H2). (I) Coronal section of a P7 brain induced at E17.
RFP
+
cells descend from L1 into the cortex and settle at the L1-L2 border
(arrow), forming dense clusters, and start to differentiate (arrowheads in I1
and I2). Deep-layer RFP
+
cells (arrowheads in I3) also settle and show signs of
differentiation by this stage but their migration route is unclear. (J) Diagram
depicting the migration route and schedule of VGZ-derived cortical cells. Scale
bars: 500 mmin (C), (D), and (G to I); 200 mmin (E); 50 mmin (F) and insets of
(E), (H), and (I); and 25 mm in inset of (F).
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 71
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

and in L5 and L6 (Fig. 2, A and B). Our in-
duction procedure did not label all ChCs, given
the consideration that the efficiency of CreER
recombination is not 100% and the fact that we
observed more ChCs with two doses of tamox-
ifen induction (e.g., at E17 and P0); we were
unable to assess the percentage of ChCs that were
labeled. Nonetheless, multiple experiments showed
that ChCs were significantly more abundant in the
frontal cortex (cingulate and pre- and infralimbic)
compared with sensory cortex (primary visual and
auditory) (Fig. 2, A and C, and fig. S5).
The somata of L2 ChCs often pressed against
the L2-L1 border (Fig. 2, E and F). They ex-
tended multiple dendrites to L1 and L2. Their
axons exhibited the classic chandelier-like mor-
phology, with local arbors extending a dense
array of vertical strings of synaptic boutons in
L2/3. Colabeling with AIS markers demon-
strated that ChC terminals not only aligned to the
AIS of PyNs but in fact targeted the distal portion
of AIS (Fig. 2F) (the first terminal of ChC
cartridge is 10.6 T 0.4 mm distal from the prox-
imal end of the AIS), which accumulates the low-
threshold sodiumchannel Nav1.6 and is the exact
site of spike initiation (12). This result unequiv-
ocally demonstrates the exquisite specificity of
ChC synapses suggested previously by vesicular
g-aminobutyric acid (GABA) transporter staining
of putative ChC terminals in human cortex (13).
L5 ChCs extended straight apical dendrites into
supragranular layers and basal dendrites within
L5; their axon arbors elaborated both below
and above the soma, forming dense cartridges
at AIS of local PyNs (Fig. 2, A and G, and fig.
S6). L6 also contained ChCs, which innervated
PyNs with characteristic strings of boutons (Fig.
2H), but these cartridges appeared in multiple
orientations that likely reflected the horizontal or
inverted orientations of L6 PyN AISs. Given
their distinct laminar location, and dendrite and
axon arborization, ChCs in layers 2, 5, and 6 are
likely recruited by distinct cortical and subcortical
inputs and regulate different populations of PyNs.
Therefore, we have identified three subgroups of
ChCs, or axo-axonic cells (5), in mouse
neocortex.
Awidely held notion is that ChCs are a subset
of PV
+
interneurons, a majority of which are fast-
spiking basket cells. However, we found that
only a subset of ChCs tested PV
+
by immuno-
staining, and this fraction varied from ~50% in
barrel cortex to ~15% in mPFC (Fig. 2D and fig.
S7). This difference in PV levels among ChCs
might reflect subtle specializations or variations
in physiological states.
During the assembly of the multilayered neo-
cortex, PyNs (14) and interneurons (15) are in-
tegrated in an “inside-out” manner, with later-born
cells migrating past earlier-born populations to
occupy more superficial laminae. MGE-derived
interneurons that share birth times with PyNs pref-
erentially populate the same cortical layer (16).
Fig. 2. Areal and laminar distribution of ChCs in neocortex. (A) Coronal
section showing the overall distribution of RFP
+
neurons at P28 after E17
induction of an Nkx2.1
CreER
;Ai9 mouse. (B) Laminar distribution of RFP
+
cells
in cingulate (CC), motor (MC), and somatosensory (SSC) cortices. (C) L2 RFP
+
cells are more abundant in medial frontal cortices (mPFC, CC) compared with
sensory cortices (SSC); ~70 and 90% of RFP
+
cells are ChCs in SSC and mPFC,
respectively. (D) Coexpression of PV in L2 RFP
+
cells varies greatly across
cortical areas. (E) Colabeling of pyramidal cell AIS [phosphorylated inhibitor
of nuclear factor kB (IkB), green] and an L2 ChC. Terminal cartridges align
with AISs and target the distal portion of an AIS (arrowheads in inset). (F)
An L2 ChC with characteristic chandelier-like axon arbor and terminal car-
tridges bearing strings of synaptic boutons (arrowheads in inset). Note the
prominent L1 dendrites (arrow). (G) An L5 ChC with characteristic chandelier-
like axon arbor and terminal cartridges (arrowheads in inset). Note the rel-
atively straight apical dendrites (arrow). (H) An L6 ChC with dense local axon
arbor and terminal cartridges (arrowheads in inset); its dendrites (arrows)
extend above and below the soma. Scale bars: 500 mmin (A), 50 mmin (E) and
(F to H), and 5 mm in insets of (E) and (F to H).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 72
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Fig. 3. Temporal profile of ChC generation in the NKX2.1 lineage.
(A) Scheme of ChC birth dating. In an Nkx2.1
CreER
;Ai9 mouse,
NKX2.1 lineage cells are labeled by tamoxifen (Tmx) induction at
E17 (red arrowhead), paired with a single BrdU injection at the
indicated date (blue arrowheads). Mice were analyzed for colocaliza-
tion after P21. (B) The temporal profile of L2 ChC generation in the
cingulate cortex, with a peak at E16. (C) A BrdU-labeled L2 ChC. (D
and E) L5 and L6 ChCs are also born in the late embryonic stage. A
ChC in L5 (D) and L6 (E) labeled by BrdU at E17, Tmx-induced at
E17, and analyzed at P28. Insets are single confocal sections. Scale
bars: 50 mm in (C) to (E) and 10 mm in insets.
Fig. 4. NKX2.1 in VGZ progenitors is
necessary for the specification of ChCs.
(A) Scheme of transplantation exper-
iment. Nkx2.1
CreER
;Ai9 donors are in-
duced at E16 with tamoxifen (Tmx);
RFP-labeled cells fromVGZ are dissected
at E17 and transplanted to the somato-
sensory cortex of P3 wild-type host pups
and analyzed at P24. (B) The transplanta-
tion of RFP
+
VGZ cells to a wild-type
brain. (C) Coronal section of a P24 host
cortex in which ChCs differentiated from
transplanted RFP
+
VGZ cells. L2 (arrows)
and L5 ChCs (arrowheads) are readily
identified. The laminar distribution of
transplanted ChCs closely resembles
that of endogenous ChCs, and L2 ChCs
extend characteristic L1 dendrites and
local axon arbors. Inset, transplanted
ChC axon terminals (arrows) inner-
vate PyNs at AIS (phosphorylated IkB,
green). (D) High-magnification view of
a transplanted L5 ChC with character-
istic dendritic and axon morphology.
(E) Scheme of transplantation from
Nkx2.1
CreER/fix
(Nkx2.1
KO
);Ai9 embryos.
Animals are induced three times from
E15 to E17, and RFP
+
VGZ cells are
transplanted into P1 wild-type pups.
In the same set of experiments, Nkx2.1
CreER/+
(Nkx2.1
+/−
);Ai9 mice were used as controls. (F) Coronal
section of a P28 host brain transplanted with Nkx2.1
+/−
VGZ cells. ChCs accumulate in L2-L1 border, as
well as in L5. Higher-magnification view of a boxed region shows dense vertically oriented axon
terminals (F1), with characteristic strings of boutons (F2). (G) Neurons derived from Nkx2.1
KO
VGZ
cells accumulate in L2 and L5, but fail to differentiate into ChCs. Higher-magnification view of the boxed region shows a lack of L1 dendrites and vertically
oriented axon terminals (G1), with almost a complete absence of terminal cartridges, (G2). Scale bars: 500 mmin (C), (F), (G); 100 mmin (F1) and (G1); 10 mmin
(F2) and (G2).
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 73
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

To examine the temporal profile of ChC produc-
tion and their correlation to laminar deployment,
we injected a single pulse of BrdU into pregnant
Nkx2.1
CreER
;Ai9 females at successive days be-
tween E15 and P1 to label mitotic progenitors,
each paired with a pulse of tamoxifen at E17 to
label NKX2.1
+
cells (Fig. 3A). We first quanti-
fied the fraction of L2 ChCs (identified by mor-
phology) in mPFCthat were also BrdU
+
. Although
there was ChC production by E15, consistent
with a previous study that suggested ventral MGE
as a possible source at this time (17), the peak of
ChC generation was at E16, when MGE has
morphologically disappeared and NKX2.1 expres-
sion has appeared at VGZ (Fig. 3, Band C). ChC
genesis then diminished but persisted until the
end of gestation. This result significantly extends
the time course of interneuron generation from
NKX2.1 progenitors and suggests that ChCs may
be the last cohort, generated at a time when PyN
neurogenesis has largely completed. More sur-
prisingly, we found that L5 and L6 ChCs were
also BrdU
+
after E17 induction (Fig. 3, Dand E),
which indicated that they were generated long
after L5 and L6 PyNs, and as late as L2 ChCs.
Therefore, the laminar deployment of ChCs does
not follow the inside-out sequence, further dis-
tinguishing them from MGE-derived interneurons.
NKX2.1 expression during late gestation
also included the preoptic area and striatum(6).
To prove VGZ as the source of ChCs, we la-
beled NKX2.1
+
cells at E16, then transplanted
the RFP
+
VGZ progenitors to the somatosensory
cortex of P3 wild-type hosts (Fig. 4, A to B).
After 3 weeks, these exogenous progenitors not
only differentiated into neurons that spread to
the medial frontal areas and settled into appro-
priate laminae (e.g., L2 and L5) but further ma-
tured into quintessential ChCs (Fig. 4, C and D).
These results indicate that NKX2.1
+
progenitors
in the late embryonic VGZare the main source of
ChCs. They further demonstrate that the identity
of a ChC is determined by its spatial and tem-
poral origin (i.e., lineage and birth time) and, once
specified, a cell autonomous program can unfold
in ectopic locations, even without proper migra-
tion, to direct differentiation.
To examine the role of NKX2.1 in ChC spec-
ification, we deleted Nkx2.1 in VGZ progeni-
tors using a conditional strategy (fig. S8) and
transplanted NKX2.1-deficient VGZ cells to the
somatosensory cortex of wild-type pups (Fig. 4E).
These Nkx2.1
KO
progenitors gave rise to neurons
that accumulated in L2 and L5 after 3 weeks (Fig.
4G), a laminar pattern that resembled those of
endogenous ChCs. However, Nkx2.1
KO
lineage
cells did not differentiate into ChCs, as indicated
by the lack of L1 dendrites and almost absence of
cartridgelike axon terminals (Fig. 4, F and G).
Together with previous studies (6, 7), our results
suggest that NKX2.1 may regulate the appropri-
ate temporal competence of progenitors, which
likely undergo sequential changes with cell divi-
sion. They further indicate that NKX2.1 expression
in VGZ progenitors is necessary to complete the
specification of a distinct, and probably the last,
cohort in this lineage—the ChCs.
A recent study demonstrated that progeni-
tors below the ventral wall of the lateral ventricle
(i.e., VGZ) of human infants give rise to a medial
migratory stream destined to the ventral mPFC
(18). Despite species differences in the develop-
mental timing of corticogenesis, this study and
our findings raise the possibility that the NKX2.1
+
progenitors inVGZandtheir extendedneurogenesis
might have evolved, since rodents, to enrich and
diversify cortical interneurons, including ChCs.
Studies in numerous systems (19) have dem-
onstrated that the specification of neuronal iden-
tities early in development exerts strong influences
in their subsequent positioning, connectivity, and
function, but, to what extent this principle applies
to the assembly of cortical circuits has been
unclear. Here, we discovered that young chande-
lier cells, once specified through their lineage and
birth time in the VGZ, migrate with a stereotyped
route and achieve distinct laminar patterns before
innervating a subdomain of PyN AIS. Therefore,
interneurons with a distinct identity are likely
endowed with cell-intrinsic programs that con-
tribute to their subsequent integration into their
destined cortical networks. Deficiencies in ChCs
have been implicated in brain disorders, includ-
ing schizophrenia (20). Genetic targeting of ChCs
establishes an entry point that integrates studies of
fate specification, laminar deployment, connectiv-
ity, and network dynamics in the context of cor-
tical circuit assemblyand function. This may provide
a probe to circuit pathogenesis in models of neu-
ropsychiatric disorders.
References and Notes
1. H. Markram et al., Nat. Rev. Neurosci. 5, 793 (2004).
2. D. M. Gelman, O. Marín, Eur. J. Neurosci. 31, 2136 (2010).
3. G. A. Ascoli et al., Nat. Rev. Neurosci. 9, 557 (2008).
4. J. Szentágothai, M. A. Arbib, Neurosci. Res. Program Bull.
12, 305 (1974).
5. P. Somogyi, Brain Res. 136, 345 (1977).
6. L. Sussel, O. Marin, S. Kimura, J. L. Rubenstein,
Development 126, 3359 (1999).
7. S. J. Butt et al., Neuron 59, 722 (2008).
8. H. Taniguchi et al., Neuron 71, 995 (2011).
9. L. Madisen et al., Nat. Neurosci. 13, 133 (2010).
10. J. Szabadics et al., Science 311, 233 (2006).
11. A. Woodruff, Q. Xu, S. A. Anderson, R. Yuste, Front.
Neural Circuits 3, 15 (2009).
12. W. Hu et al., Nat. Neurosci. 12, 996 (2009).
13. M. C. Inda, J. DeFelipe, A. Muñoz, Proc. Natl. Acad. Sci.
U.S.A. 103, 2920 (2006).
14. J. B. Angevine Jr., R. L. Sidman, Nature 192, 766
(1961).
15. A. Fairén, A. Cobas, M. Fonseca, J. Comp. Neurol. 251,
67 (1986).
16. M. W. Miller, Brain Res. 355, 187 (1985).
17. M. Inan, J. Welagen, S. A. Anderson, Cereb. Cortex 22,
820 (2012).
18. N. Sanai et al., Nature 478, 382 (2011).
19. T. M. Jessell, Nat. Rev. Genet. 1, 20 (2000).
20. D. A. Lewis, Dev. Neurobiol. 71, 118 (2011).
Acknowledgments: We would like to thank G. Miyoshi for
technical advice of fate-mapping experiments and helpful
discussion. Support by NIH R01 MH094705. H.T. is a
Precursory Research for Embryonic Science and Technology
investigator supported by Japan Science and Technology
Agency and was supported by a Brain and Behavior Research
Foundation (NARSAD) Postdoctoral Fellowship. J.L. is in part
supported by a Patterson Foundation Postdoctoral Fellowship.
Z.J.H. is supported by the Simons Foundation and a NARSAD
Distinguished Investigator award.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1227622/DC1
Materials and Methods
Figs. S1 to S8
Tables S1 and S2
References (21, 22)
Movie S1
18 July 2012; accepted 1 November 2012
Published online 22 November 2012;
10.1126/science.1227622
An Update of Wallace’s
Zoogeographic Regions of the World
Ben G. Holt,
1
* Jean-Philippe Lessard,
1
*† Michael K. Borregaard,
1
Susanne A. Fritz,
1,2
Miguel B. Araújo,
1,3,4
Dimitar Dimitrov,
5
Pierre-Henri Fabre,
5
Catherine H. Graham,
6
Gary R. Graves,
1,7
Knud A. Jønsson,
5
David Nogués-Bravo,
1
Zhiheng Wang,
1
Robert J. Whittaker,
1,8
Jon Fjeldså,
5
Carsten Rahbek
1
Modern attempts to produce biogeographic maps focus on the distribution of species, and the
maps are typically drawn without phylogenetic considerations. Here, we generate a global map
of zoogeographic regions by combining data on the distributions and phylogenetic relationships of
21,037 species of amphibians, birds, and mammals. We identify 20 distinct zoogeographic regions,
which are grouped into 11 larger realms. We document the lack of support for several regions
previously defined based on distributional data and show that spatial turnover in the phylogenetic
composition of vertebrate assemblages is higher in the Southern than in the Northern Hemisphere. We
further show that the integration of phylogenetic information provides valuable insight on historical
relationships among regions, permitting the identification of evolutionarily unique regions of the world.
B
iogeographic and bioclimatic regions are
the fundamental units of comparison in
many broad-scale ecological and evolu-
tionary studies (1, 2) and provide an essential tool
for conservation planning (3, 4). In 1876, Alfred
Russel Wallace published the first map of global
terrestrial zoogeographic regions (5), which later
became the cornerstone of modern biogeography
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 74
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

(3). Using existing knowledge of his time (6),
mostly on the distributions and taxonomic rela-
tionships of broadly defined vertebrate families,
Wallace divided the world into six terrestrial
zoogeographic units largely delineated by what we
now know as the continental plates. Despite rely-
ing on limited information and lacking a statistical
basis, Wallace’s original map is still in use today.
Wallace’s original zoogeographic regional-
ization scheme considered ancestral relationships
among species, but subsequent schemes generally
used data only on the contemporary distribu-
tions of species without explicitly considering
phylogenetic relationships (7–9). Phylogenetic
trees contain essential information on the evolu-
tionary relationships of species and have be-
come increasingly available in recent decades,
permitting the delineation of biogeographic re-
gions as originally envisioned by Wallace. The
opportunity now exists to use phylogenetic in-
formation for grouping assemblages of species
into biogeographic units on a global scale. In ad-
dition to permitting a sound delimitation of bio-
geographic regions, phylogenetic information
allows quantifying phylogenetic affinities among
regions (e.g., 10). Newly developed statistical
frameworks facilitate the transparent character-
ization of large biogeographic data sets while min-
imizing the need for subjective decisions (11).
Here, we delineated the terrestrial zoogeo-
graphic realms and regions of the world (12) by
integrating data on the global distributions and
phylogenetic relationships of the world’s am-
phibians (6110 species), nonpelagic birds (10,074
species), and nonmarine mammals (4853 species),
a total of 21,037 vertebrates species [see (13) for
details]. Pairwise phylogenetic beta diversity (pb)
metrics were used to quantify change in phyloge-
netic composition among species assemblages
across the globe. Analyses of combined taxa pb
values identified a total of 20 zoogeographic re-
gions, nested within 11 larger realms, and quan-
tified phylogenetic relatedness among all pairs of
realms and regions (Fig. 1, figs. S1 and S2, and
tables S1 and S2). We also used pb to quantify
the uniqueness of regions, with the Australian
(mean pb = 0.68), Madagascan (mean pb = 0.63),
and South American (mean pb = 0.61) regions
being the most phylogenetically distinct assem-
blages of vertebrates (Fig. 2). These evolutiona-
rily unique regions harbor radiations of species
from several clades that are either restricted to a
given region or found in only a few regions.
Our combined taxa map (Fig. 1) contrasts
with some previously published global zoogeo-
graphic maps derived exclusively from data on
the distribution of vertebrate species (8, 9, 11).
The key discrepancy between our classification
of zoogeographic regions and these previous
classifications is the lack of support for previ-
ous Palearctic boundaries, which restricted this
biogeographic region to the higher latitudes of
the Eastern Hemisphere. The regions of central
and eastern Siberia are phylogenetically more
similar to the arctic parts of the Nearctic region,
as traditionally defined, than to other parts of
the Palearctic (fig. S2). As a result, our newly
defined Palearctic realm extends across the
arctic and into the northern part of the Western
Hemisphere (Fig. 1 and fig. S1). These results
bear similarities with the zoogeographic map of
(11) derived from data on the global distribution
of mammal families. In addition, our results sug-
gest that the Saharo-Arabian realm is interme-
diate between the Afrotropical and Sino-Japanese
realms [see the nonmetric multidimensional scaling
(NMDS) plot in fig. S2]. Finally, we newly define
the Panamanian, Sino-Japanese, and Oceanian
realms [but see the Oceanian realm of Udvardy
in (14) derived from data on plants].
Our classification of vertebrate assemblages
into zoogeographic units exhibits some interest-
ing similarities with Wallace’s original classi-
fication, as well as some important differences
(fig. S3). For example, Wallace classified islands
east of Borneo and Bali in his Australian region
(fig. S3), which is analogous to our Oceanian and
Australian realms combined (Fig. 1 and fig. S1).
In contrast, we find that at least some of these
islands (e.g., Sulawesi) belong to our Oriental
realm, which spans Sundaland, Indochina, and
India (Fig. 1 and fig. S1). Moreover, our Ocean-
ian realmis separate fromthe Australian realmand
includes New Guinea together with the Pacific
Islands (14), whereas Wallace lumped these
two biogeographic units into the Australian re-
gion. Wallace further argued that the Makassar
1
Center for Macroecology, Evolution, and Climate, Depart-
ment of Biology, University of Copenhagen, 2100 Copenhagen
Ø, Denmark.
2
Biodiversity and Climate Research Centre (BiK-F)
and Senckenberg Gesellschaft für Naturforschung, Sencken-
berganlage 25, 60325 Frankfurt, Germany.
3
Department of Bio-
geography and Global Change, National Museum of Natural
Sciences, Consejo Superior de Investigaciones Científicas, Calle de
José Gutiérrez Abascal, 2, 28006 Madrid, Spain.
4
Centro de
Investigação em Biodiversidade e Recursos Genéticos, Universi-
dade de Évora, Largo dos Colegiais, 7000 Évora, Portugal.
5
Center for Macroecology, Evolution, and Climate, Natural History
Museum of Denmark, University of Copenhagen, 2100
CopenhagenØ, Denmark.
6
Department of Ecology andEvolution,
Stony Brook University, Stony Brook, NY 11794–5245, USA.
7
Department of Vertebrate Zoology, MRC-116, National Mu-
seum of Natural History, Smithsonian Institution, Post Office
Box 37012, Washington, DC 20013–7012, USA.
8
Biodiversity
Research Group, School of Geography and the Environment,
Oxford University Centre for the Environment, South Parks Road,
Oxford OX1 3QY, UK.
*These authors contributed equally to this work.
†To whom correspondence should be addressed. E-mail:
jplessard@bio.ku.dk
Fig. 1. Map of the terrestrial zoogeographic realms and regions of the world.
Zoogeographic realms and regions are the product of analytical clustering of
phylogenetic turnover of assemblages of species, including 21,037 species of
amphibians, nonpelagic birds, and nonmarine mammals worldwide. Dashed
lines delineate the 20 zoogeographic regions identified in this study. Thick
lines group these regions into 11 broad-scale realms, which are named. Color
differences depict the amount of phylogenetic turnover among realms. (For more
details on relationships among realms, see the dendrogram and NMDS plot in
fig. S1.) Dotted regions have no species records, and Antarctica is not included in
the analyses.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 75
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Strait between Borneo and Sulawesi, nowknown
as “Wallace’s Line” (15), was a major barrier to
dispersal that greatly inhibited exchanges between
the Australian and Asian land masses. Much
debate subsequently arose regarding the precise
location of the principal faunal divide between
Wallace’s Oriental and Australian realms (15)
(see fig. S3 for an illustration of Wallace’s original
line). Our combined taxa analyses lend the stron-
gest support to the hypothesis of Weber (16), who
positioned this boundary east of Sulawesi, corre-
sponding to the zoogeographic boundary sepa-
rating our Oriental and Oceanian realms (Fig. 1
and fig. S1). However, our taxon-specific geo-
graphic delineation for birds is more consistent
with Wallace’s line than Weber’s line (Fig. 3A
and figs. S3 and S4A).
The delineation of and relationships among
our zoogeographic regions differ among taxa
(Fig. 3 and fig. S4), and we find more regions for
mammals (n = 34 regions) than for amphibians
or birds (both n = 19 regions). A comparison
of pb matrices across the three vertebrate taxa
reveals that amphibian assemblages located in
the northeastern Arctico-Siberian, southern Afri-
can, and Madagascan regions are more phyloge-
netically distinct than those of birds or mammals
for the same regions (fig. S5). Moreover, the
Australian region harbors more phylogenetically
distinct assemblages of amphibians and mam-
mals relative to birds (fig. S5). Using a partial
Mantel test [see (13) for details on this analysis],
which accounts for geographic distances among
species assemblages (17), we find that global pb
values for birds and mammals are more strongly
correlated (r = 0.68, P < 0.001) than for am-
phibians and birds (r = 0.39, P < 0.001) or am-
phibians and mammals (r = 0.43, P < 0.001).
These results might partly reflect a major episode
of diversification early in the evolutionary history
of amphibians (18). Alternatively, differences in
spatial patterns of phylogenetic turnover among
vertebrate classes might result from lower dis-
persal ability (19) and greater sensitivity of
amphibians to environmental conditions (20). In-
terestingly, previous comparative studies docu-
mented similar incongruence in the diversity and
distribution of amphibians relative to that of birds
and mammals (21, 22).
The contrast between our zoogeographic
regions with regions based only on distribution-
al data (Fig. 4) demonstrates the consequences
of incorporating phylogenetic information in the
delineation of zoogeographic units. Relative to
expectations based on turnover of species, spa-
tial turnover in the phylogenetic composition of
assemblages of species is generally low in the
Fig. 2. Map of evolutionary uniqueness for terrestrial zoogeographic
regions of the world based on data for 21,037 species of vertebrates.
Evolutionary uniqueness is calculated as the mean of pairwise pb values
between the focal region and all other regions. Colors indicate the degree
to which each region differs from all other regions based on mean pairwise
pb. Regions colored in dark red are the most evolutionarily unique. Dotted
regions have no species records, and Antarctica is not included in the
analyses.
Fig. 3. Maps of terrestrial
zoogeographic regions of
the world based on data
for (A) amphibians (6110
species), (B) birds (10,074
species), and (C) nonma-
rine mammals (4853 spe-
cies). Color differences
depict the relative amount
of phylogenetic turnover
among regions within each
taxonomic clade. (For more
details on relationships
among regions, see the
dendrogram and NMDS
plots in fig. S4, A to C.)
Dotted regions have no
species records, and Ant-
arctica is not included in
the analyses.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 76
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Northern Hemisphere, whereas the opposite is
true in the Southern Hemisphere (Fig. 4A). In
particular, amphibians exhibit low spatial turn-
over in phylogenetic composition relative to their
turnover in the composition of species between
the North American and Eurasian regions (Fig. 4B;
also compare fig. S4A with fig. S6A). Higher
phylogenetic uniqueness in the Southern than in
the Northern Hemisphere is consistent with long-
term isolation having left a pervasive signature
on species assemblages, where oceanic barriers
have limited dispersal between continents (23, 24).
In the Northern Hemisphere, the newly defined
boundaries of the Palearctic realm might reflect
the continuous presence of nonglaciated tundra
in eastern Siberia and Beringia (25), whereas the
subtle differences in the phylogenetic compo-
sition of assemblages over the Northern Hemi-
sphere as a whole might be a consequence of a
high degree of connectivity and range dynamics.
Low rates of extinctions resulting from greater
climatic stability in the Southern Hemisphere could
also have contributed to this pattern by allowing
species that belong to ancient clades to persist
through time (26, 27).
Our maps of zoogeographic realms and re-
gions provide a broad overview of the distribution
of the world’s amphibians, birds, and nonma-
rine mammals, allowing the identification of ge-
ographic areas harboring distinct evolutionary
histories [see (28) for links to downloadable maps
of zoogeographic realms and regions for projec-
tion in GIS (geographic information systems)
mapping software and Google Earth]. These
maps reflect major advances made in recent dec-
ades regarding our knowledge of the distribu-
tion and phylogeny of vertebrates and can be
used to elucidate the forces and historical events
responsible for the formation of the biogeo-
graphic realms and regions we recognize today.
Our delineation of the zoogeographic realms
and regions of the world, and especially that of
the realms, appears robust to the type and quality
of distributional and phylogenetic data used [see
(13) for details]. Inclusion of additional phyloge-
netic information on branch length or improved
resolution of the phylogenetic trees has the
potential to facilitate a finer delineation of regions
within our realms. The inclusion of data (when
they become available) on reptiles, invertebrates,
and/or plants may also affect the boundaries of
our realms and regions and the relationships
among them. Nevertheless, the maps presented
here delineate robust zoogeographic units for
vertebrates that can be scaled within specific con-
tinents and/or taxonomic clades. Due to these
qualities, our analytical approach and zoogeo-
graphic maps provide a baseline for a wide variety
of comparative ecological, biogeographic, evolu-
tionary, and conservation-based studies (3, 22, 29).
References and Notes
1. M. D. Crisp et al., Nature 458, 754 (2009).
2. S. R. Loarie et al., Nature 462, 1052 (2009).
3. R. J. Ladle, R. J. Whittaker, Conservation Biogeography
(Wiley-Blackwell, Chichester, UK, 2011).
4. M. V. Lomolino, B. R. Riddle, R. J. Whittaker, J. H. Brown,
Biogeography (Sinauer Associates, Sunderland, MA, ed. 4, 2010).
5. A. R. Wallace, The Geographical Distribution of Animals
(Cambridge Univ. Press, Cambridge, 1876).
6. P. L. Sclater, J. Proc. Linn. Soc. Lond. 2, 130 (1858).
7. D. M. Olson et al., Bioscience 51, 933 (2001).
8. Ş. Procheş, S. Ramdhani, Bioscience 62, 260 (2012).
9. C. H. Smith, J. Biogeogr. 10, 455 (1983).
10. C. H. Graham, P. V. A. Fine, Ecol. Lett. 11, 1265 (2008).
11. H. Kreft, W. Jetz, J. Biogeogr. 37, 2029 (2010).
12. A. R. Wallace and his contemporary W. L. Sclater used
the terms “region” to denote six main zoogeographical
units at a global scale and “subregion” to denote finer
scale subdivisions. Wallace’s and Sclater’s regions and
subregions are roughly equivalent to the realms and
regions proposed here. The work of Sclater was published
in The Geographical Journal (1894–1897).
13. Materials and methods are available as supplementary
materials on Science Online.
14. M. D. F. Udvardy, A Classification of the Biogeographic
Provinces of the World, Occasional Paper no. 18
(International Union for Conservation of Nature, Morges,
Switzerland, 1975).
15. E. Mayr, Q. Rev. Biol. 19, 1 (1944).
16. M. Weber, Der Indo-Australische Archipel und die
Geschichte Seiner Tierwelt (Gustav Fischer Verlag, Jena,
Germany, 1902).
17. N. Mantel, Cancer Res. 27, 209 (1967).
18. K. Roelants et al., Proc. Natl. Acad. Sci. U.S.A. 104, 887
(2007).
19. M. Munguía, C. Rahbek, T. F. Rangel, J. A. F. Diniz-Filho,
M. B. Araújo, PLoS ONE 7, e34420 (2012).
20. L. B. Buckley, W. Jetz, Proc. Natl. Acad. Sci. U.S.A. 105,
17836 (2008).
21. S. A. Fritz, C. Rahbek, J. Biogeogr. 39, 1373 (2012).
22. R. Grenyer et al., Nature 444, 93 (2006).
23. J. Cracraft, Proc. Biol. Sci. 268, 459 (2001).
24. P. Upchurch, Trends Ecol. Evol. 23, 229 (2008).
25. J. R. M. Allen et al., Quat. Sci. Rev. 29, 2604 (2010).
26. J. Fjeldså, E. Lambin, B. Mertens, Ecography 22, 63 (1999).
27. B. Sandel et al., Science 334, 660 (2011).
28. Downloadable maps of zoogeographic realms and regions
for visualization in GIS and Google Earth can be found at
http://macroecology.ku.dk/resources/wallace.
29. K. A. Wilson, M. F. McBride, M. Bode, H. P. Possingham,
Nature 440, 337 (2006).
Acknowledgments: We thank the Danish National Research
Foundation for its support of the Center for Macroecology,
Fig. 4. Combined (A) and taxon-specific (B to D) maps illustrating the
degree of phylogenetic turnover relative to the turnover of species observed
among zoogeographic regions based on data for species of (A) amphibians,
(B) birds, and (C) nonmarine mammals. The color scale depicts the degree to
which faunal turnover between the regional assignment of the focal grid cell
and the regional assignment of all other grid cells results from differences in
pb relative to beta diversity. Red colors indicate regions with a high degree
of phylogenetic differentiation relative to compositional differentiation,
whereas blue colors indicate the opposite. Dotted regions have no species’
records, and Antarctica is not included in the analyses.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 77
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Evolution, and Climate. B.G.H. also thanks the Marie Curie
Actions under the Seventh Framework Programme
(PIEF-GA-2009-252888). M.B.A. also thanks the
Spanish Research Council (CSIC) for support, and S.A.F.
thanks the Landes-Offensive zur Entwicklung
Wissenschaftlich-Ökonomischer Exzellenz program of
Hesse’s Ministry of Higher Education, Research, and the
Arts. We thank L. Hansen for help with data and reference
compilations. We thank the International Union for Conservation
of Nature and Natural Resources for making the amphibian
and mammal range data available. Data are archived at
http://macroecology.ku.dk/resources/wallace.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/74/DC1
Materials and Methods
Figs. S1 to S11
Tables S1 to S5
Appendices S1 and S2
References (30–729)
1 August 2012; accepted 15 November 2012
10.1126/science.1228282
Crocodile Head Scales Are Not
Developmental Units But Emerge
from Physical Cracking
Michel C. Milinkovitch,
1
* Liana Manukyan,
1
Adrien Debry,
1
Nicolas Di-Poï,
1
Samuel Martin,
2
Daljit Singh,
3
Dominique Lambert,
4
Matthias Zwicker
3
Various lineages of amniotes display keratinized skin appendages (feathers, hairs, and scales) that
differentiate in the embryo from genetically controlled developmental units whose spatial
organization is patterned by reaction-diffusion mechanisms (RDMs). We show that, contrary to skin
appendages in other amniotes (as well as body scales in crocodiles), face and jaws scales of
crocodiles are random polygonal domains of highly keratinized skin, rather than genetically
controlled elements, and emerge from a physical self-organizing stochastic process distinct from
RDMs: cracking of the developing skin in a stress field. We suggest that the rapid growth of the
crocodile embryonic facial and jaw skeleton, combined with the development of a very keratinized
skin, generates the mechanical stress that causes cracking.
A
mniotes exhibit a keratinized epidermis
preventing water loss and skin append-
ages that play major roles in thermoregu-
lation, photoprotection, camouflage, behavioral
display, and defense against predators. Whereas
mammals and birds evolved hairs and feathers,
respectively, reptiles developed various types of
scales. Although their developmental processes
share some signaling pathways, it is unclear
whether mammalian hairs, avian feathers and
feet scales, and reptilian scales are homologous
or if some of them evolved convergently (1). In
birds and mammals, a reaction-diffusion mech-
anism (RDM) (2) generates a spatial pattern of
placodes that develop and differentiate into fol-
licular organs with a dermal papilla and cycling
growth of an elongated keratinized epidermal
structure (hairs or feathers) (3). However, scales
in reptiles do not form true follicles and might
not develop from placodes (4). Instead, reptilian
scales originate in the embryo fromregular dermo-
epidermal elevations (1). Whereas the regular
spatial organization of scales on the largest por-
tion of the reptilian body is determined by a RDM,
additional positional cues are likely involved in
the development of the scale plates present on
the head of many snakes and lizards. These head
scales form a predictable symmetrical pattern
(Fig. 1A) and provide mechanical protection.
The face and jaws of crocodilians are covered
by polygonal scales (hereafter called “head scales”)
that are strictly adjoining and nonoverlapping,
but these polygons are irregular and their spatial
distribution seems largely random (Fig. 1, B
and C). Using high-resolution three-dimensional
(3D) geometry and texture reconstructions (5–7),
1
Laboratory of Artificial and Natural Evolution (LANE), Depart-
ment of Genetics and Evolution, University of Geneva, Sciences
III, 30, Quai Ernest-Ansermet, 1211 Geneva, Switzerland.
2
La
Ferme aux Crocodiles, Pierrelatte, France.
3
Computer Graphics
Group, University of Bern, Switzerland.
4
Department of Math-
ematics and Namur Center for Complex Systems, University of
Namur, Belgium.
*To whom correspondence should be addressed. E-mail:
michel.milinkovitch@unige.ch
Fig. 1. Spatial distribu-
tion of head scales. (A)
Headscales inmost snakes
(here, a corn snake) are
polygons (two upper pan-
els) with stereotyped spa-
tial distribution (two lower
panels): left (yellow) and
right (red) scale edges
overlap when reflected
across the sagittal plane
(blue). (B) Polygonal head
scales in crocodiles have
a largely random spatial
distribution without sym-
metrical correspondence
between left and right.
(C) Head scales from dif-
ferent individuals have
different distributions of
scales’ sizes and localiza-
tions (blue and red edges
fromtopand bottomcroc-
odiles, respectively).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 78
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

as well as developmental biology techniques,
we show that crocodilians’ head scales are not
genetically controlled developmental units and
that their spatial patterning is generated through
physical cracking of a living tissue in a stress
field. This phenomenon might not involve any
specific genetic instruction besides those asso-
ciated with cell proliferation and general physical
parameters such as skin stiffness and thickness.
By marking and analyzing various features
directly on 3D models of multiple Nile croco-
dile (Crocodylus niloticus) individuals (Fig. 1 and
movie S1), we show that spatial distribution of
head scales is largely random. First, reflection of
the network of scales’ edges across the sagittal
plane indicates high variability between the left
and right head pattern (Fig. 1B and fig. S1A).
Second, nonrigid alignment (8) of head geom-
etries from different individuals indicates a
similarly large variability in scale patterns in
terms of polygons’ sizes and localizations (Fig. 1C
and fig. S1B).
This combination of order and chaos in the
distribution of head scales is reminiscent of the
topological assemblage of soap foams (9, 10).
Recent studies used the 2D foam model for
studying self-organizing principles and stochas-
tic processes shaping epithelial topology dur-
ing growth and homeostasis (11–13) because the
causal cell-surface mechanics is comparable to
the physics of foam formation (14). Similarly,
the pattern of crocodile head scales could result
from energy minimization of contact surfaces
among genetically determined elements (scales).
However, two other mechanisms could gener-
ate random distributions of polygonal elements:
(i) a RDMpatterning the spatial organization of
genetically determined developmental units, as
for mammalian hairs or avian feathers, and (ii)
cracking of a material layer causing its fracture
into adjacent polygonal domains (15).
Although stochastic patterns generated by
these processes share some universal mathe-
matical properties (see supplementary materials),
foams and crack patterns are generated by very
different physical phenomena that may be iden-
tified on the basis of other statistical features.
First, crocodile head scales do not show a good
fit to the area distribution function expected for
foams (fig. S3). Second, a fundamental differ-
ence between foams and crack patterns is that the
latter can exhibit incomplete edges (15), of which
many are observed on the head of crocodiles
(Fig. 2A).
Another key feature is the angle among edges
at nodes. In foams, edges are circular arcs in-
tersecting only three at a time with an angle of
120°, as imposed by the three instantaneous
equal length-tension force vectors acting at a
node. This rule is observed in all types of foams,
including animal epithelia (12, 16), although
the distribution of angles can be widened due to
local stress generated by cell division and growth.
On the other hand, crack patterns can gener-
ate various angle distributions. Nonhierarchical
cracking arises when fractures propagate simul-
taneously (Fig. 2B), and junctions tend to form
at 120° (17, 18). Furthermore, when a crack front
splits, or when multiple cracks are nucleated
from a single point, the junctions among edges
also tend to be 120°. Conversely, crack patterns
can be hierarchical (17, 19); that is, fractures are
formed successively, and propagating cracks
will tend to join previous cracks at a 90° angle.
Indeed, the local stress perpendicular to a crack
is relaxed and concentrates at the tip of the crack
(explaining its propagation), but the stress com-
ponent parallel to the crack is not affected. Hence,
as cracks propagate perpendicularly to the di-
rection of the maximum stress component, a
secondary crack can turn when it approaches
an older one and tends to join it at 90°. Simi-
larly, if a crack starts on the side of an older
crack, it will initially tend to propagate at a
right angle (17). Multiple examples of 90° con-
nections and incomplete edges reorienting their
propagation front are visible on the crocodiles
face and jaws (Fig. 2C). We also observe “lad-
dering” patterns (17) of paired parallel primary
fractures with perpendicular multiple secondary
cracks (Fig. 2D) and internal edges connecting
perpendicularly to the border of the network
(Fig. 2E). The distribution function of edge an-
gles is bimodal in many crocodiles analyzed
(fig. S4A), suggesting either that hierarchical
and nonhierarchical cracking processes coexist
or that head scale networks undergo a “matu-
ration” process (20–22) (see supplementary text).
Dome pressure receptors (DPRs) are pig-
mented round submillimetric sensory organs
(Fig. 2F), distributed on the crocodile face and
jaws, that detect surface pressure waves, allow-
ing crocodiles to swiftly orient, even in darkness,
toward a prey perturbing the water-air interface
(23). The dome shape of DPRs is due to a modi-
fied epidermis and the presence of a pocket of
various cell types in the outermost portion of
the dermis (Fig. 2F). We marked the localiza-
tion of DPRs on the 3D models of all scanned
individuals (orange dots, Fig. 2, G and H). Many
of the cracks that have stopped their course did
so close to a DPR (Fig. 2G and fig. S4C). Given
that the most frequent cause for fracture arrest
is when the crack front meets a heterogeneity in
the system (15), it is likely that the modified
skin thickness and composition at and around
DPRs constitute such heterogeneities. In ad-
dition, the course of many edges avoids DPRs
(Fig. 2H and fig. S4C).
The overall distribution of DPRs seems rather
homogeneous except where the density is in-
creased near the teeth and decreased at the back
of the jaws and on the top of the face (fig. S5).
Fig. 2. Signatures of cracking. (A) Many scale edges on crocodiles’ head are unconnected at one or both ends.
(B) Three incomplete cracks interact symmetrically. (C) Edges reorienting (arrows) and connecting with angles
close to 90°. (D) “Laddering” between parallel primary cracks. (E) 90° network border connections. (F) DPRs
are pigmented sensory organs (left) with a modified epidermis (right, section at embryonic stage E70, that is, at
70 days of egg incubation) and a pocket of branched nerves (white arrowheads). (G) Incomplete cracks
stopping in close proximity to a DPR (orange dots). (H) Crack propagation avoids DPRs.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 79
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Different crocodile individuals differ by as much
as 21% and 48% in their total number of DPRs
and crack edges, respectively. Remarkably, these
two interindividual variations are inversely cor-
related: Crocodiles with fewer DPRs have more
crack edges (fig. S4D). Given that the develop-
ment of DPRs precedes cracking, this correla-
tion suggests that DPRs constrain cracking, as
already implied by Fig. 2, G and H. Despite the
fact that the distributions of cracks and DPRs
both have a strong stochastic component, the
constraining effect of DPRs on cracking is no-
ticeable: The edges tend to travel along the zones
of DPRs lowest local density (fig. S4E).
The archetypal cracking process in physics
is due to shrinkage [through removal of a dif-
fusing quantity, either heat or a liquid (20)] of a
material layer adherent to a nonshrinking sub-
strate (15, 17), such that a stress field builds up
and causes fractures when the stress exceeds a
threshold characteristic of the material. Croc-
odiles have a particularly thick and rigid skin
due to the presence of a highly collagenous
dermis and an epidermis rich in b-keratins (24).
The skin covering their head shows a yet thicker
(about 2×) and more keratinized epidermis. We
suggest that the rapid growth of the crocodile
embryonic facial and jawskeleton (relative to the
size of the neurocranium), combined with the
development of a very keratinized skin, gener-
ates the mechanical stress that causes cracking.
Here, it is not the cracking layer that shrinks
but the underlying substrate layer that grows. It
explains that first-order cracks (fig. S6) tend to
traverse the width of the face because the head
is growing longitudinally faster than in other
directions.
In snakes and lizards, scales are develop-
mental units: Each scale differentiates and grows
from a primordium that can be identified by in
situ hybridization with probes targeting genes
belonging to signaling pathways involved in
early skin appendage development (1, 4). The
large head scales form a predictable pattern
following positional cues, such that the identity
of adult snake head scales can be recognized
while they develop from primordia in the em-
bryo (Fig. 3A). In crocodiles, all postcranial
scales follow that same principle of develop-
ment (Fig. 3B): Spatial distribution of primordia
is established, then each primordium differen-
tiates, first into a symmetrical elevation and sec-
ond as an oriented asymmetrical scale overlapping
with more posterior scales (Fig. 3C).
However, crocodile head scales do not form
from scale primordia or further developmental
stages. Instead, a pattern of DPRs primordia is
generated on the face and jaws: The dome shape
of DPRs has already started to form before
any scale appears (Fig. 3D). Afterward, grooves
progressively appear, propagate, and intercon-
nect (while avoiding DPRs) to forma continuous
network across the developing skin (Fig. 4A).
The process generates polygonal domains of
skin, each containing a randomnumber of DPRs.
Therefore, scales on the face and jaws of croc-
odiles (i) are not serial homologs of scales else-
where on the body and (ii) are not even genetically
controlled developmental units. Instead, they emerge
from physical cracking.
During a typical cracking process, fractures
are nucleated at the upper surface but quickly
spread downward and affect the whole thick-
ness of the material layer (19). The developing
skin on the crocodile’s head similarly reacts to
the stress field as it develops deep groves that
can reach the stiff underlying tissues (Fig. 4B).
Our analyses indicate that cell proliferation in
the epidermis layer is vastly increased in the
deepest region of the skin grooves correspond-
ing to cracks (Fig. 4C), suggesting that a heal-
ing process allows the skin layer to maintain its
continuous covering. The local biological pro-
cess (cell proliferation) might be driven by the
purely physical parameter (mechanical stress) as
follows: In zones of highest stress, local bulging
is nucleated. The local stress component per-
pendicular to the bulge is relaxed and concen-
trates at its tip, explaining the propagation of
both the stress and proliferation maxima (hence,
the corresponding propagation of the bulge). In
a manner entirely analogous to true physical
cracking, the bulge front would propagate per-
pendicularly to the direction of maximum stress
components, explaining the topology of the re-
sulting random polygonal domains of skin. The
role of proliferation reinforces the above sugges-
tion that crack patterns in crocodiles might expe-
rience maturation (20–22), explaining the observed
mixture of hierarchical and nonhierarchical fea-
tures (see supplementary text and fig. S4A).
We have shown that the irregular polygonal
domains of skin on the crocodile face and jaws
are produced by cracking, a mechanism that is
distinct from those generating scales on the
postcranial portion of the crocodile body, as
well as on the body and head of all other rep-
Fig. 3. Crocodile head scales are not developmental units. (A) In snakes, each body scale (ventral, v;
latero-dorsal, ld) differentiates from a primordium (Shh gene probe for in situ hybridization, corn
snake embryo); head scales also develop from primordia, with positional cues determining scale
identity (la, labial scales; r, rostral; n, nasal; in, internasal; pf, prefrontal; pro, preocular; so, supraocular;
pto, postocular). (B) Postcranial scales (zoom on trunk, Ctnnb1 probe) also develop from primordia that
(C) differentiate into symmetrical, then oriented asymmetrical and overlapping, scales. (D) Crocodile head
scales never form scale primordia [nor developmental stages shown in (C)] but, instead, develop a
pattern of DPRs (one DPR circled with dotted line; dome shape visible at E45) before any scale
appears (probe: Ctnnb1).
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 80
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

tiles. This cracking process is primarily physical.
However, it does not mean that genetically con-
trolled parameters are irrelevant. For example,
although a crack pattern is visible in all croc-
odilian species, spatial distribution varies consid-
erably, possibly because of species-specific skull
geometry and growth but also skin composition
and thickness. Given that these parameters, as
well as cell proliferation, are genetically con-
trolled, the variation of head crack patterns among
crocodilian species is likely due to an interplay
between physically and genetically controlled
parameters. Our study suggests that, besides
RDM, a larger set of physical self-organizational
processes contribute to the production of the
enormous diversity of patterns observed in living
systems.
References and Notes
1. C. Chang et al., Int. J. Dev. Biol. 53, 813 (2009).
2. A. M. Turing, Philos. Trans. R. Soc. London Ser. B 237, 37
(1952).
3. M. R. Schneider, R. Schmidt-Ullrich, R. Paus, Curr. Biol.
19, R132 (2009).
4. D. Dhouailly, J. Anat. 214, 587 (2009).
5. K. N. Snavely, S. M. Seitz, R. Szeliski, ACM Transactions
on Graphics (Proc. SIGGRAPH 2006) 25, 835 (2006).
6. Y. Furukawa, J. Ponce, Int. J. Comput. Vis. 84, 257
(2009).
7. M. Kazhdan, M. Bolitho, H. Hoppe, Eurographics
Symposium on Geometry Processing 2006, 61
(2006).
8. H. Li, B. Adams, L. J. Guibas, M. Pauly, ACM Transactions
on Graphics (Proc. SIGGRAPH Asia 2009) 28,
175 (2009).
9. D. Weaire, S. Hutzler, The Physics of Foams (Oxford Univ.
Press, Oxford, 1999).
10. D. Weaire, N. Rivier, Contemp. Phys. 25, 59 (1984).
11. T. Lecuit, P. F. Lenne, Nat. Rev. Mol. Cell Biol. 8, 633
(2007).
12. M. C. Gibson, A. B. Patel, R. Nagpal, N. Perrimon, Nature
442, 1038 (2006).
13. D. A. W. Thompson, On Growth and Form (Cambridge
Univ. Press, Cambridge, 1917).
14. R. Farhadifar, J. C. Röper, B. Aigouy, S. Eaton, F. Jülicher,
Curr. Biol. 17, 2095 (2007).
15. H. J. Herrmann, S. P. Roux, Eds. Statistical Models for the
Fracture of Disordered Media: Random Materials and
Processes (North-Holland, Elsevier, Amsterdam, 1990).
16. T. Hayashi, R. W. Carthew, Nature 431, 647 (2004).
17. K. A. Shorlin, J. R. de Bruyn, M. Graham, S. W. Morris,
Phys. Rev. E Stat. Phys. Plasmas Fluids Relat. Interdiscip.
Topics 61, 6950 (2000).
18. E. A. Jagla, A. G. Rojo, Phys. Rev. E Stat. Nonlin. Soft
Matter Phys. 65, 026203 (2002).
19. S. Bohn, J. Platkiewicz, B. Andreotti, M. Adda-Bedia,
Y. Couder, Phys. Rev. E Stat. Nonlin. Soft Matter Phys.
71, 046215 (2005).
20. L. Goehring, L. Mahadevan, S. W. Morris, Proc. Natl.
Acad. Sci. U.S.A. 106, 387 (2009).
21. L. Goehring, S. W. Morris, Europhys. Lett. 69, 739
(2005).
22. L. Goehring, R. Conroy, A. Akhter, W. J. Clegg,
A. F. Routh, Soft Matter 6, 3562 (2010).
23. D. Soares, Nature 417, 241 (2002).
24. L. Alibardi, L. W. Knapp, R. H. Sawyer, J. Submicrosc.
Cytol. Pathol. 38, 175 (2006).
Acknowledgments: This work was supported by the University
of Geneva, the Swiss National Science Foundation, and the
Schmidheiny Foundation. A. Tzika helped with in situs. H. Li
assisted with nonrigid registration. We thank R. Pellet for
assistance in mechanics design and A. Roux, M. Gonzalez-Gaitan,
B. Chopard, U. Schibler, and anonymous reviewers for
useful comments and suggestions.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1226265/DC1
Material and Methods
Supplementary Text
Figs. S1 to S6
Table S1
Movie S1
References (25–33)
18 June 2012; accepted 29 October 2012
Published online 29 November 2012;
10.1126/science.1226265
Fig. 4. Crocodile head
skin cracks during devel-
opment. (A) There is no
sign of cracking at E45
(but DPRs primordia are
already developed, Fig.
3D), then primary cracks
(arrowheads) appear on
the sides of the upper
jaws and progress toward
the top of the face (dotted
line). At E65, primarycracks
reached the top of the
head and are followed
by secondary cracks in
other orientations (ar-
rows). (B) Three sequen-
tial skin sections along
primary (pc) and second-
ary (sc) cracks (ep, epi-
dermis; de, dermis; bo,
bone tissue). (C) Antibody
to pan cadherin stains the
wholeepidermis, antibody
to proliferating cell nucle-
ar antigen(PCNA) indicates
increased proliferation (ar-
rows), and terminal deox-
ynucleotidyl transferase–
mediated deoxyuridine
triphosphate nick end la-
beling (TUNEL) assay indi-
cates absence of apoptosis
in cracks.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 81
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Translation Elongation Factor
EF-P Alleviates Ribosome Stalling
at Polyproline Stretches
Susanne Ude,
1,2
* Jürgen Lassak,
1,2
* Agata L. Starosta,
1,3
Tobias Kraxenberger,
1,2

Daniel N. Wilson,
1,3
‡ Kirsten Jung
1,2

Translation elongation factor P (EF-P) is critical for virulence in bacteria. EF-P is present in all
bacteria and orthologous to archaeal and eukaryotic initiation factor 5A, yet the biological function
has so far remained enigmatic. Here, we demonstrate that EF-P is an elongation factor that
enhances translation of polyproline-containing proteins: In the absence of EF-P, ribosomes stall
at polyproline stretches, whereas the presence of EF-P alleviates the translational stalling.
Moreover, we demonstrate the physiological relevance of EF-P to fine-tune the expression of the
polyproline-containing pH receptor CadC to levels necessary for an appropriate stress response.
Bacterial, archaeal, and eukaryotic cells have hundreds to thousands of polyproline-containing
proteins of diverse function, suggesting that EF-P and a/eIF-5A are critical for copy-number
adjustment of multiple pathways across all kingdoms of life.
T
he ubiquitous elongation factor P (EF-P)
(1) is orthologous to archaeal and eukary-
otic initiation factor 5A (a/eIF-5A) (2),
binds to the ribosome, and stimulates peptide-
bond formation (3–5). Lys
34
(K34) of Escherich-
ia coli EF-P is lysinylated and hydroxylated by
YjeK, YjeA, and YfcM (6–10) (fig. S1A), which
enhances the ability of EF-P to stimulate peptide-
bond formation (9). Inactivation of efp, yjeA, or
yjeKresults in defects in bacterial growth, fitness,
and membrane integrity (6, 11, 12), as well as
stress resistance and virulence (7, 13–16).
Our functional insights into the mechanismof
EF-P arose through our efforts to discover new
regulatory players of the Cad module, a lysine-
dependent acid-resistance system (17, 18) (fig.
S1B). The membrane-integrated sensor and tran-
scriptional regulator CadC regulates induction of
the cadBAoperon encoding the lysine/cadaverine
antiporter CadB and the lysine decarboxylase
CadA. Transposon mutagenesis identified a mini-
Tn10Kan
R
insertion in yjeK that results in a neg-
ative CadA phenotype (19) (fig. S1, C and D).
Because YjeK is involved in the EF-P modifica-
tion pathway, we tested whether deletions of yjeA
and efp also influence the Cad module by mea-
suring CadA activity in the corresponding mu-
tants. All deletion mutants—yjeA

, yjeK

, and
efp

—lacked CadAactivity, unless the respective
gene was provided in trans (Fig. 1A). However, the
CadA negative phenotype could not be rescued
1
Center for Integrated Protein Science Munich (CiPSM), Ludwig-
Maximilians-Universität München, Munich, Germany.
2
Depart-
ment of Biology I, Microbiology, Ludwig-Maximilians-Universität
München, Großhaderner Strasse 2-4, 82152 Martinsried,
Germany.
3
Gene Center and Department of Biochemistry, Ludwig-
Maximilians-Universität München, Feodor-Lynen-Str. 25, 81377
Munich, Germany.
*These authors contributed equally to this work.
†Present address: AMSilk GmbH, Am Klopferspitz 19, 82152
Martinsried, Germany.
‡To whom correspondence should be addressed. E-mail:
jung@lmu.de (K.J.); wilson@genzentrum.de (D.N.W.)
CadC
+ EF-P
CadC-PPPIP/AAAIS
+ EF-P
55
70
40
35
25
15
10
1 2 3 4 5 6 7 8 9 10 11 12 13 14
FL
Time (min)
55
70
40
35
25
15
10
CadC
RNase A - + - +
15 16 17 18
CadC
1-125
5 10 15 30 5 10 15 30 5 10 30 5 10 30
A
DNA-binding TM pH-sensing
fMet
CadC
Gln
C
a
d
C
C
a
d
C
P
P
P
I
P
/




A
A
A
I
S
P
1
2
1
A
C
a
d
C
-
-
+
+
CadC
B
D E F
FL
C
- EF-P - EF-P
efp efp
lysine + H cadaverine + CO
+
2
CadA
R
e
l
.

C
a
d
A

a
c
t
i
v
i
t
y

(
m
u
t
a
n
t

/

w
i
l
d
t
y
p
e
)
0
0.2
1.2
1.0
0.8
0.6
0.4
+

y
j
e
A
+

y
j
e
K
+

e
f
p
+

e
f
p
-



K
3
4
A
∆yjeA ∆efp ∆yjeK
T
L
0
2
T
L
3
0
T
L
6
6
T
L
1
0
0
T
L
1
0
8
T
L
1
3
1
T
L
1
5
8
R
e
l
.

β
-
g
a
l
.

a
c
t
i
v
i
t
y
(
e
f
p

/
e
f
p


)
-

+
0
0.2
1.2
1.0
0.8
0.6
0.4
1.4
efp
efp
+
-
T
L
1
5
8
-


P
1
2
1
A
T
L
1
5
8
-
P
P
P
I
P
/
A
A
A
I
S
T
L
1
5
8


β
-
g
a
l
.

a
c
t
i
v
i
t
y

(
M
U
)
0
200
600
1000
1400
1800
(kDa)
(kDa)
- - -
100 0 11 66 90 (%)
TL30
TL66
TL100
TL108
Gln
TL02 LacZ
TL131 PPPIP
120
TL158
124
PPPIP
Fig. 1. Lysinylated EF-P alleviates ribosome stalling at polyproline stretches in
CadC. (A) CadA activity was monitored in yjeA

, yjeK

, and efp

strains bearing
empty vectors (–) or complemented with the corresponding genes in trans.
Error bars indicate the standard deviation of the mean. (B) Schematic depicting
the domain structure of full-length CadC in comparison with CadC’-LacZ hy-
brids. (C) The b-galactosidase activities of CadC’-LacZ (TL02-TL158) hybrids
from (B) were determined in the efp

and efp
+
strains. (D) The b-galactosidase
activities of TL158, TL158-P121A, and TL158-PPPIP/AAAIS were determined in
efp
+
and efp

strains. (E) The protein levels of full-length CadC, CadC-P121A,
and CadC-PPPIP/AAAIS in efp
+
and efp

strains were determined by Western
blotting. Wild-type CadC in the efp
+
strain was assigned to 100%. (F) Auto-
radiographs of SDS–polyacrylamide gels indicating [
35
S]Met-labeled in vitro
translation time courses of CadC and CadC-PPPIP/AAAIS in the absence and
presence of modified EF-P. Control reactions of CadC and truncated CadC
1-125
(– RNaseA) were subjected to RNaseA treatment (+ RNaseA). Positions of full-
length CadC (FL), peptidyl-tRNA, and free peptide are arrowed.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 82
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

by an unmodified EF-P–K34Avariant. Whereas
EF-P was dispensable for the translation of CadA,
it was required for translation of CadC (fig. S1, E
to H), such that the absence of EF-P significantly
diminished CadC protein levels (fig. S1G).
The influence of EF-P on CadC translation
was monitored by generating translational fu-
sions (TL) between CadC variants of different
lengths (2 to 158 amino acids) and LacZ and
then expressing them in efp

and efp
+
strains
(Fig. 1B). EF-P did not influence translation
when the second amino acid of CadC was varied
(fig. S2, A and B), nor was any effect observed
for the CadC-LacZ hybrids TL02, TL30, TL66,
TL100, and TL108 (Fig. 1C). By contrast, the
b-galactosidase activities of TL131 and TL158
were decreased in the efp

strain (Fig. 1C), sug-
gesting that EF-P–dependent translation of CadC
relies on a motif located between amino acids
108 and 158. In CadC, this region contains a
cluster of prolines (Pro
120
-Pro
121
-Pro
122
-Ile
123
-
Pro
124
) (Fig. 1B). Proline is a particularly poor
donor and acceptor during peptide-bond forma-
tion (20–22), and therefore EF-P might support
translation of CadCby overcoming the detrimen-
tal effect that the proline cluster has on the
ribosomal peptidyltransferase activity. We inter-
rupted the proline cluster (PPPIP) by generating
CadC variants with a single Pro
121
→Ala
121
mu-
tation (P121A→PAPIP) or with all four prolines
mutated (PPPIP→AAAIS, where S indicates Ser)
and found that EF-P had little influence on trans-
lation of both TL158-P121A and TL158-PPPIP/
AAAIS (Fig. 1D). Western blotting indicated that
the wild-type CadC level was significantly re-
duced in the efp

(11%) compared with the efp
+
strain (assigned to 100%), whereas the protein
levels of the CadC-P121A and CadC-PPPIP/
AAAIS variants remained higher (66 and 90%,
respectively) (Fig. 1E).
We used an in vitro transcription-translation
system (23) to monitor translation of wild-type
CadC via incorporation of radiolabeled [
35
S]-
methionine (Fig. 1F, lanes 1 to 4). Synthesis of
full-length (FL) CadC (~60 kD) was observed
after 10 min (lane 2) and increased with longer
incubation (lanes 3 and 4). A prominent band of
35 kD persisted for 15 min (lanes 1 to 3) before
becoming reduced or absent at 30 min (lane 4),
consistent with a translation intermediate that even-
tually elongates to form the FL CadC product.
Furthermore, the 35-kD band was absent when
modified EF-P was included in the reaction (Fig.
1F, lanes 5 to 8). The size of the 35-kD product
matches that of a CadCnascent polypeptide chain
translated up to the proline cluster (~15 kD) but
still remaining attached to a tRNA (~20 kD). By
contrast, when translation of CadC-PPPIP/AAAIS
was tested (Fig. 1F, lanes 9 to 14), the 35-kD
band was absent, regardless of the presence or
absence of EF-P. In vitro translation performed
using CadC truncated directly after the proline
cluster (CadC
1-125
) revealed a product of the same
size as in wild-type CadC (compare lanes 15 and
17 in Fig. 1F). Moreover, both the 10-min 35-kD
CadCintermediate (lane 15) and 35-kDCadC
1-125
product (lane 17) were reduced to 15 kD via
subsequent ribonuclease A (RNase A) treatment
that degrades the tRNA(Fig. 1F, lanes 16 and 18).
Collectively, these data suggest that the ribosome
stalls at the proline cluster during CadC transla-
tion in the absence of EF-P and that the presence
of EF-P alleviates the translational stalling.
To define the minimal requirements under
which stalling occurs, we generated five cadC’-lacZ
55
70
40
35
25
15
10
Fluc-3xPro
+ EF-P
1 2 3 4 5 6 7 8
0 10 20 30
0
100
200
300
400
control
EF-P (unmod)
EF-P
EF-P-K34A
R
e
l
.

F
l
u
c
-
3
x
P
r
o

a
c
t
i
v
i
t
y

(
%
)
Time (min)
5 10 30 60 5 10 30 60
Fluc
3xPro
1-175
Fluc-3xPro
Ala Arg Asn Asp Cys Glu Gln Gly His Ile Leu Lys Met Phe Pro Ser Thr Trp Tyr Val
500
2000
1500
1000
0
5X:
RNase A
A
D
E
B
C
- EF-P
n
o
P
r
o
5
x
P
r
o
4
x
P
r
o
3
x
P
r
o
2
x
P
r
o
1
x
P
r
o
3000
2000
1000
0
β
-
g
a
l
.

a
c
t
i
v
i
t
y

(
M
U
)
n:
efp
efp
+
-
efp
efp
+
-
Time (min)
(kDa)
FL
0 5 10 30 15
Time (min)
TL30 LacZ
n x Pro
TL30 LacZ
5X
FL
Fluc-3xPro luminescence
EF-P
144
128
9 10 11 12
- + - +
β
-
g
a
l
.

a
c
t
i
v
i
t
y

(
M
U
)
Fig. 2. Three consecutive prolines are sufficient for ribosome stalling. The
b-galactosidase activities (A) of CadC
1-30
-LacZ (TL30) hybrids bearing up
to five prolines and (B) of CadC
1-30
-LacZ (TL30) hybrids bearing five
consecutive repeats of each of the 20 proteinogenic amino acids were
measured in efp
+
and efp

strains. Error bars indicate standard deviation
of the mean. (C) In vitro translation time courses of Fluc-3xPro and
respective controls performed as in Fig. 1F. (D) Luminescence of in vitro
translated Fluc-3xPro was monitored over time (microtiter plate, left) and
quantified (right) in the absence of EF-P (■) or presence of unmodified
EF-P (●), 128 (green

) and 144 (blue

) modified EF-P, or EF-P-K34A
(

). The 100% luminescence is defined as the luminescence produced by
Fluc-3xPro after a 30-min incubation in the absence of EF-P.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 83
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

reporter fusions encoding truncated CadC
1-30
(TL30) followed by one to five consecutive pro-
lines and the LacZ reporter. The b-galactosidase
activities of the five constructs in the efp
+
and
efp

strains revealed that, when three or more
consecutive prolines were inserted, translation be-
came strictly dependent on EF-P (Fig. 2A). More-
over, the EF-P dependency was not stronger with
more than three prolines, indicating that three
consecutive prolines are sufficient to confer EF-P
dependence. The amino acid specificity of EF-P
dependent translation was tested by using a series
of reporter fusions where a five–amino acid re-
peat of each of the 20 proteinogenic amino acids
was individually inserted after the truncated CadC
of the TL30 CadC-LacZ reporter. EF-P depen-
dency was only observed in the presence of pro-
line repeats (Fig. 2B). Moreover, the amino acid
proline in the nascent chain, rather than the type
of proline codon in the mRNA, influences EF-P
dependency (fig. S2C), and the context of the three
prolines is of negligible importance (fig. S2D).
We introduced three consecutive prolines into
a Firefly luciferase (Fluc) reporter, generating Fluc-
3xPro. In vitro translation of Fluc-3xPro performed
in the absence of EF-P led to an additional band
of ~37 kD at 5 to 10 min (Fig. 2C). This was
consistent with a peptidyl-tRNA stalled at the
proline cluster (~17 kD), which was confirmed
by translation of Fluc-3xPro truncated directly
after the third proline codon and subsequent RNase
A treatment (Fig. 2C). When EF-P was present,
the 37-kD band was absent, and the yield of full-
length Fluc-3xPro was higher at the 10- and 30-min
time points. To assess this quantitatively, we per-
formed the same reactions and monitored lumi-
nescence (Fig. 2D). When modified EF-P was
present, we observed a 3.5-fold increase in lumi-
nescence compared with the control without
EF-P, whereas no luminescence increase was ob-
served when unmodified EF-P or EF-P-K34Awas
used (Fig. 2D), thus demonstrating the lysinylation
modification to be essential for EF-P activity.
Activation of wild-type CadC requires two
stimuli, low pH and exogenous lysine, to induce
cadBA expression (Fig. 3, A and B), whereas a
single-stimulus induction was observed for the
EF-P–independent CadC-PPPIP/AAAIS variant
(Fig. 3B). The copy number of CadC proteins in
the wild type was 3 to 5 molecules per cell, which
decreased to 0 or 1 molecules per cell in efp

cells
and increased to 11 to 14 molecules per cell for
CadC-PPPIP/AAAIS (Fig. 3C). Consistently, we
observed a strong correlation between the cellular
CadC levels and stress response (fig. S2E). Acti-
vation of CadC is dependent on its cosensor LysP
(Fig. 3A). Overproduction of LysPprevents cadBA
expression under all conditions (24), indicating
that the copy number of LysP is also crucial for
the dual-stimulus response. We propose a balance
model for the Cad-module: In wild-type cells un-
der noninducing conditions, the CadC/LysP ratio
is balanced (Fig. 3D, top); however, in response
to the two stimuli, lowpHand lysine, the balance
is shifted in favor of activated CadC, leading to
cadBA expression (Fig. 3B, top). In cells lacking
EF-P, the CadC/LysP ratio is imbalanced to dis-
favor CadC, and therefore cadBA expression is
significantly reduced (Fig. 3, B and D, middle).
In cells harboring the EF-P–independent CadC-
PPPIP/AAAIS, the CadC/LysP ratio is imbalanced
in favor of CadC, leading to a loss in sensitivity
of the stimulus-dependent cadBA expression
(Fig. 3, B and D, bottom). Thus, EF-P represents
a previously unknown mechanism to fine-tune
the protein copy number of CadC.
The E. coli proteome contains ~100 proteins
with at least one polyproline stretch (table S1).
We chose seven proteins, AmiB, FlhC, Flk, NlpD,
RzoR, TonB, and UvrB, differing in function,
cellular localization, proline number, and location
of the polyproline stretch (table S2) and fused
them to lacZ. The b-galactosidase activity of all
tested hybrid proteins was reduced in the efp

mutant (Fig. 3E), indicating the requirement of
EF-P for their translation. In vitro translation
assays also demonstrated translational stalling at
the polyproline stretches within FlhC, Flk, and
RzoR, which could be alleviated by EF-P (Fig. 3F).
These findings suggest that translation of most, if
not all, polyproline-containing proteins is depen-
dent on EF-P. Moreover, the observation that efp,
yjeA, or yjeKmutants attenuate virulence (7, 13–16)
can nowbe rationalized because important recep-
tors, such as EnvZ and PhoR, which regulate path-
ogenicity (25, 26), contain polyproline stretches.
Genomic analyses reveal an abundance of poly-
proline stretches not only in bacteria but also in
archaea and eukaryotes, suggesting that EF-P and
a/eIF-5A alleviate translational stalling for multi-
ple cellular pathways across the three domains
of life.
Fig. 3. EF-P influences
the stress response of CadC
and translation of other
polyproline-containing pro-
teins. (A) Schematic of the
CadCstress response. (B) CadC
activation of the P
cadBA
pro-
moter was tested in a report-
er strain(P
cadBA
::lacZ) bearing
CadC or CadC-PPPIP/AAAIS in
efp
+
orefp

backgrounds. cadBA
expression (b-galactosidase
activity) was monitored in the
presence of no stimulus or
one (pH = 5.8 or 10 mM
lysine) or two stimuli (pH =
5.8 and 10 mMlysine). Error
bars indicate standard devia-
tion of the mean. (C) Mole-
cule numbers per cell of CadC
and CadC-PPPIP/AAAIS in the
cells from (B) were analyzed
by Western blotting. (D) A
balance model illustrating the
impact of alterations in the
copy number of CadC and
LysP onthe stimulus response.
(E) The b-galactosidase activ-
ities of translation fusions of the polyproline-containing E. coli proteins AmiB, FlhC, Flk, NlpD, RzoR, TonB, and UvrB with LacZ were determined in efp
+
and efp

strains. (F) In vitro translation time courses of FlhC, Flk, and RzoR, with respective controls performed as in Fig. 1F.
Ø
CadC LysP
PcadBA
lacZ
no lysine neutral pH
CadC LysP
PcadBA
lacZ
lysine low pH
+
+
A
non-inducing condition
inducing condition
FlhC

FlhC
- + - - - -
+ + - - - -
Time (min):
55
70
40
35
25
15
10
Flk

RNase A
Flk
EF-P
- + - - - -
+ + - - - -
RzoR

RzoR
1-18
- + - - - -
+ + - - - -
5 10 5 10 10 10 5 10 5 10 10 10 5 10 5 10 10 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
(kDa)
FL
FL
FL
1-10
1-53 E
F
0
1000
2000
3000
4000
5000
AmiB FlhC Flk NlpD RzoR TonB UvrB
efp
efp
+
-
β
-
g
a
l
.

a
c
t
i
v
i
t
y

(
M
U
)
B C D
11-14
3-5
0-1
cadBA expression (MU)
no stimulus
one stimulus (low pH)
two stimuli
(low pH and lysine)
one stimulus (lysine)
CadC
CadC
efp
efp
CadC-
PPPIP/AAAIS
efp
+
+
-
2000 4000 6000 8000 0
CadC LysP
CadC per cell
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 84
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

References and Notes
1. M. Bailly, V. de Crécy-Lagard, Biol. Direct 5, 3 (2010).
2. N. C. Kyrpides, C. R. Woese, Proc. Natl. Acad. Sci. U.S.A.
95, 224 (1998).
3. B. R. Glick, M. C. Ganoza, Proc. Natl. Acad. Sci. U.S.A.
72, 4257 (1975).
4. B. R. Glick, S. Chládek, M. C. Ganoza, Eur. J. Biochem.
97, 23 (1979).
5. G. Blaha, R. E. Stanley, T. A. Steitz, Science 325, 966
(2009).
6. T. Yanagisawa, T. Sumida, R. Ishii, C. Takemoto,
S. Yokoyama, Nat. Struct. Mol. Biol. 17, 1136 (2010).
7. W. W. Navarre et al., Mol. Cell 39, 209 (2010).
8. H. Roy et al., Nat. Chem. Biol. 7, 667 (2011).
9. J. H. Park et al., J. Biol. Chem. 287, 2579 (2012).
10. L. Peil et al., Nat. Chem. Biol. 8, 695 (2012).
11. E. de Crécy et al., Appl. Microbiol. Biotechnol. 77,
489 (2007).
12. S. B. Zou et al., J. Bacteriol. 194, 413 (2012).
13. K. Kaniga, M. S. Compton, R. Curtiss 3rd, P. Sundaram,
Infect. Immun. 66, 5599 (1998).
14. S. M. Bearson, B. L. Bearson, M. A. Rasmussen,
Appl. Environ. Microbiol. 72, 2829 (2006).
15. W. T. Peng, L. M. Banta, T. C. Charles, E. W. Nester,
J. Bacteriol. 183, 36 (2001).
16. S. M. Bearson, B. L. Bearson, B. W. Brunelle,
V. K. Sharma, I. S. Lee, Foodborne Pathog. Dis.
8, 725 (2011).
17. S. Y. Meng, G. N. Bennett, J. Bacteriol. 174, 2670
(1992).
18. G. Fritz et al., J. Mol. Biol. 393, 272 (2009).
19. Material and methods are available as supplementary
materials on Science Online.
20. H. Muto, K. Ito, Biochem. Biophys. Res. Commun. 366,
1043 (2008).
21. M. Y. Pavlov et al., Proc. Natl. Acad. Sci. U.S.A. 106, 50
(2009).
22. I. Wohlgemuth, S. Brenner, M. Beringer, M. V. Rodnina,
J. Biol. Chem. 283, 32229 (2008).
23. Y. Shimizu et al., Nat. Biotechnol. 19, 751 (2001).
24. M. N. Neely, E. R. Olson, J. Bacteriol. 178, 5522
(1996).
25. J. Garmendia, C. R. Beuzón, J. Ruiz-Albert, D. W. Holden,
Microbiology 149, 2385 (2003).
26. R. L. Lucas et al., J. Bacteriol. 182, 1872 (2000).
Acknowledgments: This work was supported by the
Excellence Cluster “Center for integrated Protein Science”
Munich (Exc114/1), the European Molecular Biology
Organization young investigator program (to D.N.W.) and
grants JU270/5-3 (K.J.) and WI3285/1-1 (D.N.W.) from the
Deutsche Forschungsgemeinschaft. A.L.S. is funded by an
AXA Postdoctoral fellowship.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1228985/DC1
Materials and Methods
Figs. S1 to S2
Tables S1 to S5
References (27–55)
17 August 2012; accepted 29 October 2012
Published online 13 December 2012;
10.1126/science.1228985
EF-P Is Essential for Rapid Synthesis
of Proteins Containing Consecutive
Proline Residues
Lili K. Doerfel,
1
* Ingo Wohlgemuth,
1,2
* Christina Kothe,
1
Frank Peske,
1
Henning Urlaub,
2,3
Marina V. Rodnina
1

Elongation factor P (EF-P) is a translation factor of unknown function that has been implicated
in a great variety of cellular processes. Here, we show that EF-P prevents ribosome from stalling
during synthesis of proteins containing consecutive prolines, such as PPG, PPP, or longer proline
strings, in natural and engineered model proteins. EF-P promotes peptide-bond formation and
stabilizes the peptidyl–transfer RNA in the catalytic center of the ribosome. EF-P is posttranslationally
modified by a hydroxylated b-lysine attached to a lysine residue. The modification enhances the
catalytic proficiency of the factor mainly by increasing its affinity to the ribosome. We propose
that EF-P and its eukaryotic homolog, eIF5A, are essential for the synthesis of a subset of proteins
containing proline stretches in all cells.
B
acterial elongation factor P (EF-P) and its
archaeal/eukaryotic homolog, initiation
factor 5A (a/eIF5A), are universally con-
served proteins with unknown cellular function.
EF-P was reported to modulate cell viability,
growth, virulence, motility, and sensitivity to low
osmolarity, detergents, and antibiotics (1, 2).
The EF-P binding site on the ribosome is located
at the interface of the 30S and 50S ribosomal
subunits between the binding sites for peptidyl-
tRNA (P site) and the exiting tRNA (E site). The
N-terminal domain of EF-P interacts with the ac-
ceptor stem of the P site–bound tRNA near the
peptidyl transferase center (3). Both EF-P and
eIF5A are posttranslationally modified (4–8).
In Escherichia coli EF-P, Lys
34
is modified by
the action of three enzymes, YjeK, YjeA, and
YcfM (5, 7–9). The proposed cellular function
of EF-P/eIF5A is to optimize the ribosomes for
more productive interactions with tRNA and re-
lease factors (3, 10–13). However, the magnitude
of the reported effects appeared too small (less
than twofold) for a universal function. We clarify
the functional role of EF-P by investigating the
role of the factor in the translation of various
mRNA sequences.
The activity of EF-P in accelerating peptide
bond formation was originally identified by its
ability to increase the yield of formylmethionyl-
puromycin (f Met-Pmn) synthesis (12, 13). Because
reaction rates vary by ~1000-fold for different
C-terminal amino acids in the peptidyl-tRNA
(14), we examined whether EF-P may specifical-
ly accelerate the product formation with poorly
reactive peptidyl-tRNAs, such as those containing
C-terminal proline residues (Fig. 1Aand fig. S1).
The formation of the tripeptide fMetPro-Pmn was
accelerated by almost 90-fold in the presence of
EF-P, whereas in the other cases the reaction was
stimulated by less than fivefold. To explore the
effect of the amino acid following a proline resi-
due, we tested Gly- and Pro-tRNAs that are known
to be slow in peptide bond formation and can
contribute to ribosome stalling (15–17). Whereas
EF-P enhanced peptide-bond formation by less
than twofold for most combinations, large effects
were observed for the formation of fMetProGly
(fMPG, eightfold) and fMetProPro (16-fold) (Fig.
1B). When longer strings of Pro and Gly residues,
such as fMPPG, fMPPGF, or fMPPPF, were ex-
amined, practically no correct product was formed
in the absence of EF-P, whereas the addition of
EF-P rescued the rapid production of the respec-
tive peptide (Fig. 1C). The low yield of product
formation in the absence of EF-P can be explained
by the loss of peptidyl-tRNA from the stalled ribo-
somes (e.g., fMPP-tRNA; fig. S2A). These results
suggested a specific effect of EF-P in accelerating
the reaction with poorly reactive substrates that
otherwise cause ribosome stalling.
To verify whether proline-containing sequences
as such induce ribosome stalling that can be res-
cued by EF-P, we engineered PG, PP, PPG, and
PPP sequences into an N-terminal fragment of
protein PrmC, which originally does not contain
such sequences and is rapidly synthesized inde-
pendent of the presence of EF-P (Fig. 2). Intro-
ducing a glycine residue after Pro
20
resulted in the
ribosome pausing, as shown by the accumulation
of a peptide of about 20 amino acids; however,
pausing was transient and not affected by EF-P.
When a second proline residue preceding Pro
20
was engineered, the ribosome paused at the ProPro
sequence, and EF-P reduced the pause time from
20 to 10 s. When PPG or PPP sequences were
introduced, the ribosomes were stalled, and es-
sentially no full-length product was produced. The
ribosomes stalled upon translation of the PPG se-
quence that contained a ProPro-ending peptidyl-
tRNA in the P site and a Gly-tRNA in the A site
(fig. S2B). Addition of EF-P rescued the synthesis
of full-length product by preventing or alleviating
ribosome stalling (Fig. 2 and fig. S2C).
1
Department of Physical Biochemistry, Max Planck Institute
for Biophysical Chemistry, 37077 Goettingen, Germany.
2
Bio-
analytical Mass Spectrometry Group, Max Planck Institute for
Biophysical Chemistry, 37077 Goettingen, Germany.
3
Bio-
analytics, Department of Clinical Chemistry, University Medical
Center Goettingen, 37075 Goettingen, Germany.
*These authors contributed equally to this work.
†To whom correspondence should be addressed. E-mail:
rodnina@mpibpc.mpg.de
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 85
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Fig. 1. EF-P promotes the synthesis of Pro- and
Gly-containing peptides on the ribosome. (A) Rates
(k
obs
) of peptide-bond formation between fMet-
tRNA
fMet
or fMetX-tRNA
X
and Pmn, where X stands
for different amino acids, as indicated, in the ab-
sence (white) or presence (black) of EF-P (3 mM). Error
bars indicate standard deviations. Single-letter ab-
breviations for the amino acid residues are as fol-
lows: D, Asp; E, Glu; F, Phe; G, Gly; K, Lys; M, Met;
P, Pro; Q, Gln; R, Arg; V, Val; and W, Trp. (B) Rates
of peptide-bond formation between fMet-tRNA
(fM) or fMetPro-tRNA (fMP) in the P site and Gly-
tRNA
Gly
, Phe-tRNA
Phe
, or Pro-tRNA
Pro
(G, F, and P,
respectively) in the A site. (C) Formation of model
oligopeptides in a reconstituted translation system
in the absence (open) and presence (solid) of EF-P.
Fig. 2. EF-P prevents ribosome stalling on PPGand
PPP sequences engineered into PrmC. Translation
of the N-terminal domain of PrmC (75 amino acids)
with wild-type (WT) or mutant sequences containing
PG, PP, PPG, or PPP in the absence or presence of
EF-P. Peptides were separated by SDS–polyacrylamide
gel electrophoresis (SDS-PAGE) and visualized by
the fluorescence of BODIPY-FL attached to the N
terminus of the peptides. M1 and M2 are peptide
markers for PrmC fragments of the indicated num-
ber of amino acids. Single-letter abbreviations for
the amino acid residues are as follows: A, Ala; C, Cys;
D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L,
Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T,
Thr; V, Val; W, Trp; and Y, Tyr.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 86
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

The sequences PPG and PPP, as well as lon-
ger proline stretches, are found in a large number
of cellular proteins, raising the question of wheth-
er they require EF-P for their synthesis. We tested
the effect of EF-P on the synthesis of some of
these proteins, such as TonB, YafD, Rz1, and
AmiB(Fig. 3). In the absence of EF-P, translation
was stalled at the consecutive prolines, and the
addition of EF-P alleviated ribosome stalling, re-
sulting in the rapid synthesis of the respective
full-length peptide. When the Pro stretches were
particularly long, as in AmiBor Rz1, very little full-
length product was synthesized in the absence of
EF-P, because translation was halted at the stalling
site. In all cases, the addition of EF-P resulted in the
efficient synthesis of full-length protein. Transient
pauses that occurred at sequences other than PPP
and PPG were not alleviated by EF-P (Fig. 3).
Posttranslational modification of EF-P ap-
pears crucial for the factor’s function (9, 18).
Translation of the fMPPG sequence, which was
abolished in the absence of EF-P, was fully re-
stored with EF-P carrying the b-lysine modifica-
tion at Lys
34
; further hydroxylation did not alter
the activity, whereas the absence of modifications
reduced the activity of the factor (Fig. 4). A sim-
ilar effect was observed for the translation of
the PPG sequence in PrmC (fig. S4). Analysis of
the reaction kinetics indicated that the modifica-
tion increased the affinity of EF-P binding to the
ribosome 30-fold (Fig. 4). The maximum rate of
fMPPG synthesis decreased ~fourfold with un-
modified compared with modified EF-P, resulting
in a >100-fold difference in k
cat
/K
M
(where k
cat
is the maximum reaction rate obtained by hyper-
bolic fitting and K
M
is the Michaelis constant).
The low catalytic proficiency of unmodified EF-P
may explain why deletions of the yjeA or yjeK
genes that code for the modification enzymes
lead to phenotypes that are similar to, or only
somewhat milder than, the deletion of efp, the
gene coding for EF-P (1, 2). On the other hand,
even in the absence of active EF-P, small amounts
of proteins with proline stretches are slowly
formed (Fig. 3); this may be sufficient to support
the viability of strains in which the efp, yieA, and
yieK genes are deleted.
Our data indicate that EF-P is a translation
factor that promotes the synthesis of proteins
containing PPG, PPP, and longer polyproline
stretches by preventing ribosome stalling during
Fig. 3. EF-P alleviates PPP- and PPG-induced stalling during synthesis of natural proteins. Translation
products of TonB (239 amino acids), YafD (75 amino acids from the N terminus), Rz1 (62 amino acids),
and AmiB (159 amino acids from the N terminus) were separated by SDS-PAGE. M3 and M4 are peptide
markers containing TonB fragments of the indicated lengths. M5, M6, and M7 are peptide markers of the
indicated lengths of YafD, RZ1, and AmiB sequences, respectively.
Fig. 4. Contribution of Lys
34
modi-
fications to EF-P activity. (A) Time
courses of fMPPG synthesis. The
formation of the tetrapeptide was
measured in the absence of EF-P
(open circles), in the presence of
unmodified EF-P (open squares),
overexpressed lysinylated but un-
hydroxylated EF-P (solid triangles),
and lysinylated/hydroxylated overexpressed EF-P (solid circles) or native EF-P (open
triangles). The extent of the respective modification was verified by mass spec-
trometry (fig. S3). (B) Yield of fMPPG peptide as a function of EF-P concentration
with unmodified (open circles; K
M
= 2.4 T 0.5 mM) or fully modified (solid circles;
K
M
= 0.08 T 0.02 mM) EF-P. K
M
is the EF-P concentration at which 50% of the
maximumyield is reached. (C) EF-P concentration dependence of fMPPG synthesis
rate with unmodified EF-P (open circles; k
cat
= 0.12 T 0.03 s
−1
) and fully modified
EF-P (solid circles; k
cat
= 0.65 T 0.02 s
−1
). Error bars indicate standard deviations.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 87
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

the formation of proline-proline or proline-glycine
peptide bonds. Thus, EF-P (and likely eIF5A) can
be considered a third universal translation elon-
gation factor that acts on the ribosome, in ad-
dition to the two factors, EF-Tu and EF-G, that
catalyze the mRNA decoding and tRNA trans-
location steps, respectively. In contrast to the
latter two factors, which act in each elongation
cycle irrespective of the amino acid incorporated,
EF-P and eIF5A are responsible for the elonga-
tion of a subset of nascent proteins and are the
only factors identified so far that augment the
peptidyl transferase activity of the ribosome.
The pyrrolidine ring of proline restricts the
number of accessible conformations for this res-
idue and may impose structural constraints on
the positioning of the amino acid in the peptidyl
transferase center, resulting in slow peptide-bond
formation and thereby ribosome stalling. EF-P
facilitates peptide-bond formation by stabilizing
peptidyl-tRNA on the ribosome and possibly by
promoting optimal positioning of the substrates.
The cellular fate of the nascent peptide on stalled
ribosomes is likely to depend on its length: Shorter
peptides may be lost because peptidyl-tRNA drop
off the ribosome (16, 19), whereas ribosomes stalled
after the synthesis of longer peptides might be
rescued through SsrA-, ArfA-, or YaeJ-dependent
pathways (20–23).
Of more than 4000 annotated E. coli pro-
teins, about 270 contain motifs of three or more
consecutive prolines or PPG motifs (table S1).
Among those, proteins belonging to the basal
transcription-translation machinery are under-
represented, whereas metabolic enzymes, trans-
porters, and regulatory transcription factors are
frequent. The observed pleiotropic effects of efp,
yjeA, or yjeK deletions (1) can be rationalized by
the impaired synthesis of proteins with proline
stretches. For example, decreased synthesis of
TonB, a protein fromthe EF-P interactome (24)
that supplies energy for the function of TonB-
dependent transporters, may result in a cumula-
tive inhibition of transport processes (25). The
severe effects of efp, yjeA, or yjeK deletions on
cell motility may be explained by the failure to
support the coordinated expression of flagellar
proteins FlhC, FliF, and Flk, all containing PPP
and PPG motifs, whereas the deficiency of EF-P
mutants in using g-glutamyl-glycine as a nitrogen
source (1) may be caused by impaired synthesis
of the corresponding catabolic enzyme, g-glutamyl-
transpeptidase, which contains a PPP motif. Last-
ly, EF-P may affect the translation of virulence
proteins, such as protein EspF, which is a key
player during the infection of eukaryotic hosts by
enterohaemorrhagic and enteropathogenic strains
of E. coli (26, 27). EspF contains several proline
runs, including PPPP sequences, in its function-
ally important Src homology 3 binding domains
(27). Given that the posttranslational modifica-
tions of eukaryotic eIF5Aand bacterial EF-P are
different, the occurrence of proline runs in many
pathogenic proteins makes EF-P and its modifi-
cation enzymes promising new targets for de-
veloping highly specific, potent antimicrobials.
References and Notes
1. S. B. Zou et al., J. Bacteriol. 194, 413 (2012).
2. S. B. Zou, H. Roy, M. Ibba, W. W. Navarre, Virulence 2,
147 (2011).
3. G. Blaha, R. E. Stanley, T. A. Steitz, Science 325, 966 (2009).
4. E. C. Wolff, K. R. Kang, Y. S. Kim, M. H. Park,
Amino Acids 33, 341 (2007).
5. T. Yanagisawa, T. Sumida, R. Ishii, C. Takemoto,
S. Yokoyama, Nat. Struct. Mol. Biol. 17, 1136 (2010).
6. M. Bailly, V. de Crécy-Lagard, Biol. Direct 5, 3 (2010).
7. H. Roy et al., Nat. Chem. Biol. 7, 667 (2011).
8. L. Peil et al., Nat. Chem. Biol. 8, 695 (2012).
9. W. W. Navarre et al., Mol. Cell 39, 209 (2010).
10. B. R. Glick, S. Chládek, M. C. Ganoza, Eur. J. Biochem.
97, 23 (1979).
11. B. R. Glick, M. C. Ganoza, Proc. Natl. Acad. Sci. U.S.A.
72, 4257 (1975).
12. P. Saini, D. E. Eyler, R. Green, T. E. Dever, Nature 459,
118 (2009).
13. A. P. Gregio, V. P. Cano, J. S. Avaca, S. R. Valentini,
C. F. Zanelli, Biochem. Biophys. Res. Commun. 380,
785 (2009).
14. I. Wohlgemuth, S. Brenner, M. Beringer, M. V. Rodnina,
J. Biol. Chem. 283, 32229 (2008).
15. M. Johansson et al., Proc. Natl. Acad. Sci. U.S.A. 108,
79 (2011).
16. M. Y. Pavlov et al., Proc. Natl. Acad. Sci. U.S.A. 106,
50 (2009).
17. D. R. Tanner, D. A. Cariello, C. J. Woolstenhulme,
M. A. Broadbent, A. R. Buskirk, J. Biol. Chem. 284,
34809 (2009).
18. J. H. Park et al., J. Biol. Chem. 287, 2579 (2012).
19. V. Heurgué-Hamard et al., EMBO J. 17, 808 (1998).
20. Y. Chadani, K. Ono, K. Kutsukake, T. Abo, Mol. Microbiol.
80, 772 (2011).
21. Y. Chadani et al., Mol. Microbiol. 78, 796 (2010).
22. C. S. Hayes, B. Bose, R. T. Sauer, J. Biol. Chem. 277,
33825 (2002).
23. T. Sunohara, T. Abo, T. Inada, H. Aiba, RNA 8, 1416 (2002).
24. L. Salwinski et al., Nucleic Acids Res. 32, D449 (2004).
25. N. Noinaj, M. Guillier, T. J. Barnard, S. K. Buchanan,
Annu. Rev. Microbiol. 64, 43 (2010).
26. N. A. Sallee et al., Nature 454, 1005 (2008).
27. A. Holmes, S. Mühlen, A. J. Roe, P. Dean, Infect. Immun.
78, 4445 (2010).
Acknowledgments: We thank W. Wintermeyer for critically
reading the manuscript; W. Holtkamp and J. Mittelstaet for the
translation assay and M1 and M2 peptide markers; D. Görlich
and C. Enke for the antibody against EF-P; and O. Geintzer,
S. Kappler, T. Wiles, M. Zimmermann, T. Uhlendorf, A. Bursy,
and U. Plessmann for expert technical assistance. The work was
funded by the Max Planck Society and by grants from the
Deutsche Forschungsgemeinschaft.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1229017/DC1
Materials and Methods
Figs. S1 to S4
Table S1
References (28–33)
20 August 2012; accepted 26 October 2012
Published online 13 December 2012;
10.1126/science.1229017
Para-Aminosalicylic Acid Acts as an
Alternative Substrate of Folate Metabolism
in Mycobacterium tuberculosis
Sumit Chakraborty,
1,2
Todd Gruber,
3
Clifton E. Barry III,
3
Helena I. Boshoff,
3
Kyu Y. Rhee
1,2
*
Folate biosynthesis is an established anti-infective target, and the antifolate para-aminosalicylic
acid (PAS) was one of the first anti-infectives introduced into clinical practice on the basis of
target-based drug discovery. Fifty years later, PAS continues to be used to treat tuberculosis. PAS
is assumed to inhibit dihydropteroate synthase (DHPS) in Mycobacterium tuberculosis by mimicking
the substrate p-aminobenzoate (PABA). However, we found that sulfonamide inhibitors of DHPS
inhibited growth of M. tuberculosis only weakly because of their intracellular metabolism. In
contrast, PAS served as a replacement substrate for DHPS. Products of PAS metabolism at this
and subsequent steps in folate metabolism inhibited those enzymes, competing with their
substrates. PAS is thus a prodrug that blocks growth of M. tuberculosis when its active forms
are generated by enzymes in the pathway they poison.
D
omagk’s discovery of the sulfonamide
Prontosil, a synthetic prodrug inhibitor
of dihydropteroate synthase (DHPS),
led to the first commercially available antibiotic
and established folate biosynthesis as the first
biologically selective target for anti-infectives.
However, sulfonamides had little activity against
Mycobacterium tuberculosis, the single leading
cause of death from bacterial infection (1). Shortly
after Waksman’s and Schatz’s discovery of strepto-
mycin, Lehmann developed para-aminosalicylic
acid (PAS), a close structural analog of the folate
precursor p-aminobenzoate (PABA) (2). PAS be-
came the second drug for tuberculosis (TB). Hitch-
ings subsequently developed antibacterial inhibitors
of dihydrofolate reductase, validating a second
enzyme of folate metabolism as a therapeutically
selective target (3). In parallel, studies of PAS in
combination with streptomycin and eventually
other anti-TB drugs laid the foundation for mod-
ern TB chemotherapy.
1
Department of Medicine, Weill Cornell Medical College, New
York, NY 10065, USA.
2
Department of Microbiology and Im-
munology, Weill Cornell Medical College, New York, NY 10065,
USA.
3
Tuberculosis Research Section, Laboratory of Clinical
Infectious Diseases, National Institute for Allergy and Infec-
tious Diseases (NIAID), National Institutes of Health (NIH),
Bethesda, MD 20892, USA.
*To whom correspondence should be addressed. E-mail:
kyr9001@med.cornell.edu
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 88
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Current TB chemotherapy regimens are longer
and more complex than for any other bacterial in-
fection, resulting in substantial rates of treatment
noncompliance, treatment failure, and emergence
of multidrug resistant (MDR), extensively drug
resistant (XDR) and totally drug resistant (TDR)
TB. Given that fresh approaches are needed (4),
an improved understanding of the mode of ac-
tion of drugs whose effectiveness has been es-
tablished in clinical practice seems imperative.
Much has been learned from the genetic basis of
drug resistance, cell-free assays with candidate tar-
get enzymes, measurements of rates of synthesis
of the major classes of macromolecules, and tran-
scriptional perturbations and adaptations. How-
ever, there have been few, if any, applications of
metabolomics to intact bacteria during the pre-
lethal phase of drug exposure. Metabolomics is
the most direct readout of the biochemical state
of cells with a simultaneous report of uptake and
modifications of a drug and has the potential to
reveal the pharmacokinetics and pharmaco-
dynamics that take place within the pathogen.
The presumption that PAS functions as a
competitive inhibitor of DHPS, the second en-
zyme in folate biosynthesis (fig. S1) (5), seemed
to be supported by the discovery that some
Fig. 1. (A) Concentration-
dependent accumulation
of drug within M. tuber-
culosis after incubation
of filter-laden M. tubercu-
losis cultures atop drug-
impregnated Middlebrook
7H9–based agar media for
18 hours. Values on the
x axis denote the rela-
tive drug concentration ex-
pressed in multiples of
the drug’s MIC in plate agar [or, for dapsone (APS) (22) and sulfanilamide (SNL), molar equivalent
of sulfamethoxazole (SMX) doses used]. The y axis values denote relative molar abundance per unit
cell biomass as reported by residual protein content of each sample. (B) Concentration-dependent
accumulation of PABA in M. tuberculosis after incubation as in (A); all values are the average of
experimental triplicates T SEM.
Fig. 2. (A) Time-dependent accumulation of PAS, PABA, N-methyl-PAS
(N-Me-PAS), and N,N-dimethyl PAS [N-(Me)
2
-PAS] in M. tuberculosis after
incubation with PAS, as in Fig. 1B. (B) Extracted ion chromatograms with
accompanying mass spectra (insets) showing concentration-dependent for-
mation of PAS-derived pteroate and folate analogs (structures shown above
each plot, with PAS core highlighted in red) after exposure of M. tuberculosis
to PAS for 18 hours. (C) Chemical and kinetic competence of PABA and PAS as
in vitro substrates for purified recombinant M. tuberculosis DHPS (top table)
and FolC (bottom). The y axis values in (B) denote relative ion intensity per
unit cell biomass, and values are the average of experimental triplicates T SEM.
RT, retention time.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 89
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

PAS-resistant strains harbor mutations in the folate-
consuming enzyme thymidylate synthase A(thyA)
(6). The puzzle is that PAS is a modest in vitro
inhibitor of DHPS [apparent inhibition constant
(K
i
app
) = 1 mM], whereas more potent inhibitors
of M. tuberculosis DHPS lack activity against in-
tact M. tuberculosis (7).
Adapting a filter culture system to support
metabolomic profiling of M. tuberculosis in re-
sponse to drug treatment (8), we established that
PAS exhibited the same minimal inhibitory con-
centration (MIC) (1 mg/ml) in a thymidine-free
liquid media and on agar-supported filters and ac-
cumulated in M. tuberculosis in a concentration-
dependent manner (Fig. 1A). To ensure that a
functionally substantial proportion of the cell-
associated PAS was cytosolic, we monitored intra-
cellular PABA levels as a reporter of PAS binding
to DHPS because it is assumed to displace PABA
and slow the consumption of PABA. As negative
controls, we incubated M. tuberculosis with equi-
molar concentrations of three in vitro inhibitors
of M. tuberculosis DHPS, with inhibition con-
stants (K
i
) ranging from 10 to 90 nM but weak
(sulfamethoxazole; MIC = 50 mg/ml) or no (sul-
fanilamide, dapsone) whole-cell activity, expect-
ing to confirm their failure to accumulate in M.
tuberculosis (7, 9). However, we observed dose-
dependent accumulations of all three compounds
(Fig. 1A) and of PABA by 18 hours (Fig. 1B), as
well as changes in the levels of 154 endogenous
metabolites (fig. S2), confirming that they accu-
mulated in M. tuberculosis cytosol. The specific
patterns of changes in metabolite levels were high-
ly similar for sulfamethoxazole, sulfanilamide, and
dapsone but less so for PAS.
PAS resulted in a linear dose-dependent ac-
cumulation of PABA across the concentration
range tested, whereas the three sulfonamides that
are more potent DHPS inhibitors yielded dose-
dependent accumulations of PABA that plateaued
at levels lower than those achieved by PAS (Fig.
1B). The lower plateau in PABA accumulation
seen with the sulfonamides than with PAS sug-
gested that M. tuberculosis may inactivate the
sulfonamides while less effectively disposing
of PAS. Indeed, we observed multiple mass spe-
cies corresponding to inactive biotransformation
products of sulfamethoxazole, sulfanilamide,
and dapsone that accumulated in a dose-dependent
manner (fig. S3 and table S1). Because both sul-
famethoxazole and dapsone resulted in the same
degree of PABA accumulation, these results also
indicated that DHPS inhibition alone cannot ex-
plain the growth inhibitory activity of sulfameth-
oxazole. Next, we focused on the fate of PAS.
Examining the PAS and PABA levels in
M. tuberculosis over 48 hours, we discovered
two distinct but kinetically matched patterns of ac-
cumulation (Fig. 2A). PAS accumulation peaked
~6 hours after drug exposure and fell over the next
42 hours. In contrast, PABA exhibited an early
phase of accumulation that slowed by 6 hours
and then began to accelerate again, until it pla-
teaued after about 18 hours after drug exposure.
The dissociation between PAS and PABA lev-
els during the second phase of PABA accumu-
lation suggested that PAS may have undergone
enzymatic biotransformation to an additional
species also capable of inhibiting DHPS. We
sought such products by identifying chromato-
graphically resolved ion families that were only
found in PAS-treated samples and whose abun-
dance increased in a dose-dependent manner.
Using authentic chemical standards, we confirmed
that PAS underwent sequential mono- and di-
methylation at the N4 position (Fig. 2A and table
S1) (10). N-mono-methyl PAS was equipotent to
PAS in antimycobacterial activity, but N-di-methyl
PAS was inactive (11).
We then sought to understand how PAS and
N-mono-methyl PAS might affect the M. tuber-
culosis folate pathway by identifying biotransfor-
mation products corresponding to the sequential
turnover of PAS and its N-methyl derivative by
M. tuberculosis DHPS (FolP1) and dihydrofolate
synthase/folyl-polyglutamyl synthase (FolC) (Fig.
2B and fig. S4) (12–14). In vitro reactions using
the purified recombinant enzymes confirmed that
PAS was a kinetically and chemically competent
substrate for M. tuberculosis FolP1 and could
give rise to a dihydrofolate analog (Fig. 2C).
Most reduced folate species, such as dihydro-
folate and tetrahydrofolate, exhibit half-lives (t
1/2
) <
1 hour (14), rendering it difficult to quantify such
PAS-derived folate analogs in M. tuberculosis.
As a surrogate, we sought to characterize the func-
tional impact of their formation by monitoring the
time- and dose-dependent effects of PAS on four
folate-dependent pathways in M. tuberculosis, the
synthesis of glycine, methionine, thymidine tri-
phosphate, and purines, which are critical for the
synthesis of protein, DNA, and RNA. Endogenous
levels of deoxythymidine monophosphate (dTMP)
and 5-formamido-1-(5-phospho-D-ribosyl)imidazole-
4-carboxamide (FAICAR) were below the limit
of detection, but we succeeded in monitoring lev-
els of serine, glycine, homocysteine, methio-
nine, deoxyuridine monophosphate (dUMP), and
Fig. 3. (A) Heatmap representation of six key folate-dependent metabolites in M. tuberculosis after
treatment with bioequivalent doses of PAS, SMX, APS, or SNL as in Fig. 1A. Numbers denote drug
concentration expressed in relation to MIC or, for APS and SNL, molar equivalent of SMX doses used.
Color intensities correspond to relative levels of metabolites and are expressed as the log
2
-transformed
ratio of the normalized signal intensity in drug-treated cells at each concentration to the normalized
signal intensity in the drug-free sample and visually scaled as indicated in the top right inset. Signal
intensities were normalized to cell protein biomass at each concentration. (B) Time-dependent effects of
PAS on folate-dependent metabolite levels shown in (A). (C) Heatmap representation demonstrating a
time-dependent metabolic rescue of the PAS-induced effects with a 10-fold molar excess of exogenous
PABA. All values are the average of experimental triplicates T SEM. Labels and abbreviations are as in
(A) and Fig. 1B.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 90
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

5-amino-1-(5-phospho-D-ribosyl)imidazole-4-
carboxamide (AICAR) as chemically stable, pre-
viously validated metabolic surrogates of these
pathways (Fig. 3A and fig. S5) (15, 16). Studies
in Escherichia coli have supported a view of
folate biosynthesis in which inhibition at vari-
ous points in the folate pathway leads to a gen-
eralized depletion of folate intermediates and
folate-dependent products downstream of the
site of inhibition (15). In contrast, treatment with
PASresulted in selective increases in serine, homo-
cysteine, AICAR, and dUMP levels that kineti-
cally followed the accumulation of PABA(Fig. 3,
A and B). These changes differed from those ob-
served with sulfamethoxazole, sulfanilamide, and
dapsone (Fig. 3A), which inhibit M. tuberculosis
DHPS, and also those reported for E. coli treated
with an inhibitor of its dihydrofolate reductase
(15, 16). Moreover, co-incubation of M. tubercu-
losis with exogenous PABA reversed the changes
seen with PAS but did not affect the uptake of
PAS or basal levels of the monitored folate-
dependent reporters (Fig. 3C). Thus, PAS exerts
its antimycobacterial activity through its effects
on M. tuberculosis folate metabolismdownstream
of DHPS.
The specific site (or sites) of PAS action on
M. tuberculosis folate pathway that result in
growth inhibition remain to be elucidated. How-
ever, studies of bacterial and human thymidylate
synthases, including those of Mtb, and AICAR
transformylases have reported potent inhibition by
dihydrofolate- and methotrexate-related species
highly similar to the PAS-derived folate analogs
identified here (17–21).
Enzyme inhibition is a cornerstone of ra-
tional drug development, but observing growth
inhibition upon exposure of a cell to an enzyme
inhibitor does not suffice to reveal the compound’s
mode of action. Without the ability to monitor
directly the accumulation of a drug, its biotrans-
formation, and its intracellular effects, a full under-
standing of mode of action cannot be achieved.
The metabolomic approach used here showed
that PAS, despite its in vitro activity as a competi-
tive inhibitor of DHPS, did not inhibit growth of
M. tuberculosis by inhibiting DHPS. Sulfona-
mides that are more potent DHPS inhibitors than
PAS but poor antimycobacterial agents do not
owe their lack of efficacy to inadequate uptake,
as had been inferred. Instead, they are inactivated
by bacterial metabolism and inhibit DHPS only
weakly in situ. PAS poisons folate-dependent path-
ways not only by serving as a replacement sub-
strate for DHPS but also by the products of that
reaction serving as replacement substrates and/or
inhibitors of subsequent enzymes.
This discovery revises our understanding of
an old drug (PAS) and old target (DHPS) and
underscores that catalysis by, rather than inhibi-
tion of, an enzyme can be exploited for drug
development.
References and Notes
1. M. Raviglione et al., Lancet 379, 1902 (2012).
2. E. R. Long, The Chemistry and Chemotherapy of
Tuberculosis (Williams & Wilkins, Baltimore, MD, 1958),
pp. 347–352.
3. G. B. Elion, G. H. Hitchings, Adv. Chemother. 2, 91 (1965).
4. C. Nathan, Cell Host Microbe 5, 220 (2009).
5. J. Rengarajan et al., Mol. Microbiol. 53, 275 (2004).
6. V. Mathys et al., Antimicrob. Agents Chemother. 53,
2100 (2009).
7. V. Nopponpunth, W. Sirawaraporn, P. J. Greene,
D. V. Santi, J. Bacteriol. 181, 6814 (1999).
8. L. P. de Carvalho et al., Chem. Biol. 17, 1122 (2010).
9. T. S. Huang et al., J. Antimicrob. Chemother. 67, 633
(2012).
10. D. L. Schuessler, T. Parish, PLoS ONE 7, e34471
(2012).
11. K. G. Rosdahl, Svensk Kemisk Tidskrift 60, 12 (1948).
12. M. Gengenbacher, T. Xu, P. Niyomrattanakit, G. Spraggon,
T. Dick, FEMS Microbiol. Lett. 287, 128 (2008).
13. S. T. Cole et al., Nature 393, 537 (1998).
14. W. Lu, Y. K. Kwon, J. D. Rabinowitz, J. Am. Soc. Mass
Spectrom. 18, 898 (2007).
15. Y. K. Kwon, M. B. Higgins, J. D. Rabinowitz, ACS Chem.
Biol. 5, 787 (2010).
16. J. R. Wei et al., Proc. Natl. Acad. Sci. U.S.A. 108,
4176 (2011).
17. C. J. Allegra, J. C. Drake, J. Jolivet, B. A. Chabner,
Proc. Natl. Acad. Sci. U.S.A. 82, 4881 (1985).
18. C. J. Allegra, R. L. Fine, J. C. Drake, B. A. Chabner, J. Biol.
Chem. 261, 6478 (1986).
19. H. T. Spencer, J. E. Villafranca, J. R. Appleman,
Biochemistry 36, 4212 (1997).
20. A. S. Fivian-Hughes, J. Houghton, E. O. Davis,
Microbiology 158, 308 (2012).
21. J. H. Hunter, R. Gujjar, C. K. Pang, P. K. Rathod,
PLoS ONE 3, e2237 (2008).
22. M. J. Brauer et al., Proc. Natl. Acad. Sci. U.S.A. 103,
19302 (2006).
Acknowledgments: We thank C. Nathan for critical discussions
and reading of the manuscript, D. Garboczi for assistance
in preparing purified FolP1 and FolC, R. Lee and M. Serono for
the generous gifts of pure folate intermediates and pteroate
analogs, S. Fischer for expert mass spectrometric support, a
Burroughs Wellcome Career Award in the Biomedical Sciences
to K.Y.R., the William Randolph Hearst Foundation, and the Bill
and Melinda Gates TB Drug Accelerator Program (OPP1024050)
for support. This work was supported in part by the Intramural
Research Program of NIAID, NIH. Primary data are deposited in
Dryad; the provisional doi is 10.5061/dryad.vh36n.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/88/DC1
Materials and Methods
Figs. S1 to S5
Table S1
References (23–26)
17 August 2012; accepted 18 October 2012
Published online 1 November 2012;
10.1126/science.1228980
Dynamic Persistence of
Antibiotic-Stressed Mycobacteria
Yuichi Wakamoto,
1
*† Neeraj Dhar,
1
* Remy Chait,
2
‡ Katrin Schneider,
1
François Signorino-Gelo,
1
Stanislas Leibler,
2,3
John D. McKinney
1
§
Exposure of an isogenic bacterial population to a cidal antibiotic typically fails to eliminate a small
fraction of refractory cells. Historically, fractional killing has been attributed to infrequently
dividing or nondividing “persisters.” Using microfluidic cultures and time-lapse microscopy, we
found that Mycobacterium smegmatis persists by dividing in the presence of the drug isoniazid
(INH). Although persistence in these studies was characterized by stable numbers of cells, this
apparent stability was actually a dynamic state of balanced division and death. Single cells
expressed catalase-peroxidase (KatG), which activates INH, in stochastic pulses that were negatively
correlated with cell survival. These behaviors may reflect epigenetic effects, because KatG pulsing
and death were correlated between sibling cells. Selection of lineages characterized by infrequent
KatG pulsing could allow nonresponsive adaptation during prolonged drug exposure.
M
utations and genetic exchange are im-
portant drivers of microbial diversity,
but these events are rare. Often cells
display phenotypic diversity that is not due to
genetic variations (1, 2). This diversity is critical
for microbial persistence in fluctuating environ-
ments because it ensures that some individuals
may survive a lethal stress that would otherwise
extinguish the population. For example, fraction-
al killing of bacterial populations by antibiotics
is attributed to phenotypic variants called “per-
sisters” (3–5). Recent work has also uncovered
a similar role for persisters in fractional killing
of cancer cells by cytotoxic drugs (6, 7).
Microbial persistence was identified by Bigger
in early studies on penicillin, which inhibits bac-
terial cell wall biogenesis (8). The ineffectiveness
of penicillin against stationary-phase bacteria
(9, 10) suggested that persisters might constitute
a subpopulation of nondividing cells. This con-
cept was confirmed when preexisting subpop-
ulations of infrequently dividing or nondividing
1
School of Life Sciences, Swiss Federal Institute of Technology
in Lausanne (EPFL), 1015 Lausanne, Switzerland.
2
Laboratory
of Living Matter, The Rockefeller University, New York, NY
10021, USA.
3
Simons Center for Systems Biology and School
of Natural Sciences, Institute for Advanced Study, Princeton,
NJ 08540, USA.
*These authors contributed equally to this work.
†Present address: Research Center for Complex Systems Biol-
ogy, University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo
153-8902, Japan, and PRESTO, Japan Science and Technology
Agency ( JST), 4-1-8 Honcho, Kawaguchi, Saitama 332-0012,
Japan.
‡Present address: Institute of Science and Technology Austria,
A-3400 Klosterneuberg, Austria.
§To whom correspondence should be addressed. E-mail:
john.mckinney@epfl.ch
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 91
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Escherichia coli cells were shown to be refractory
to ampicillin (11). Stochastic growth rate switch-
ing might be an evolved strategy for survival in
fluctuating environments (12).
Reversible adoption of a slowly growing or
nongrowing state is regarded as a general strategy
for persistence. This assumption has been the ra-
tionale for persister-enrichment methods (13, 14),
evaluation of antibiotic therapies (15), evolutionary
models of genetic resistance (16) and social be-
haviors (17), and drug discovery strategies (18, 19).
Here, we show that persistence is not correlated
with single-cell growth rates in Mycobacterium
smegmatis exposed to isoniazid (INH), which in-
hibits cell wall biogenesis (20). Instead, cell fate is
linked to single-cell dynamics of the INH-activating
enzyme catalase-peroxidase (KatG), which is ex-
pressed in stochastic pulses that are negatively
correlated with survival. KatG pulsing and death
are positively correlated between sibling cells,
which may reflect epigenetic effects.
Fractional killing is characterized by multi-
phasic kinetics of population decay. We confirmed
that INHkilling of M. smegmatis was multiphasic,
comprising delay (d), killing (k), and persistence
( p) phases (fig. S1). We measured INH killing
kinetics at the single-cell level by culturing the
bacteria in a microfluidic device (fig. S2) and track-
ing individual cells by time-lapse microscopy. In
antibiotic-free medium, cells grew and divided to
formmicrocolonies (movie S1). Within ~30 min of
adding INH to the flow medium, all cells mark-
edly slowed their rates of growth and division.
Cell lysis began after ~6 hours and continued
throughout 144 hours of INH exposure (Fig. 1A
and movie S2). When INH was withdrawn,
one of the intact cells in Fig. 1A resumed rapid
growth and division; a second exposure to INH
killed the progeny with kinetics similar to those
of the first exposure (Fig. 1A and movie S2). We
conclude that persistence was due to reversible
phenotypic tolerance rather than stable genetic
mutations.
We tracked the fates of individual lineages
by generating pedigrees for 153 progenitor cells
that were present at the time of INH addition
(Fig. 1B and fig. S3A). After 72 or 144 hours of
INHexposure, 114 or 136 (respectively) of the 153
lineages were extinguished because the founder
cell and all of its progeny were lysed. Unexpect-
edly, divisions continued under INH exposure;
129 of 153 progenitor cells divided at least once.
Figure 1C depicts representative trajectories of
an individual that was killed and an individual
that continued dividing throughout the entire
period of INH exposure. In separate experiments
we confirmed that growth, division, and death
Fig. 1. Multiphasic dynam-
ics of INH-mediated killing.
Bacteria expressing GFP were
imaged on fluorescence and
phase channels for 212hours
at 15-min intervals and ex-
posed to INH (50 mg/ml) at
12 to 156 hours and at 188
to 212 hours. This experi-
ment was repeated10times.
(A) Representative image se-
ries. Numbers indicate hours
elapsed. Scalebar, 5mm. 7H9,
Middlebrook 7H9 medium.
(B) Pedigree tree of cells de-
scended from a single pro-
genitor. Broken lines indicate
INH addition (12 hours) or
withdrawal (156 hours). Bi-
furcations indicate divisions.
End points indicate deaths.
Grayscale in the background
corresponds to slope of pop-
ulation decay curve shown
in (D). (C) Cell size was mea-
sured at 15-min intervals:
(i) nonpersistent lineage, (ii)
persistent lineage. (D to F)
Population dynamics re-
constructed from single-
cell measurements. Broken
lines indicate INH addition
(12hours). (D) Summednum-
bers of surviving cells. (E)
Summed sizes (areas) of sur-
viving cells in arbitrary units.
(F) Mean cell size (black line)
and SD (red lines).
0
2
4
6
8
10
2
4
6
8
0
10
0 20 40 60 80 100 120 140 160
10
100
1000
0 50 100 150
0
2
4
6
8
10
12
0 50 100 150
10
100
1000
0 50 100 150
Time (h) Time (h) Time (h)
M
e
a
n

c
e
l
l

s
i
z
e

(
µ
m
2
)
S
u
m
m
e
d

c
e
l
l

s
i
z
e
s

(
a
.
u
.
)
C
e
l
l

n
u
m
b
e
r
A
Time (h) Time (h)
0 20 40 60 80 100 120 140 160
C
e
l
l

s
i
z
e

(
µ
m
2
)
(i) Non-persistent lineage
(ii) Persistent lineage
B C
D E F
(d)
(k)
(p)
(d)
(k)
(p)
0 h 12 h
7H9 7H9→INH
24 h
INH
36 h
INH
72 h
INH INH→7H9
156 h
7H9→INH
188 h
INH
212 h
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 92
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

continued for at least 240 hours under INH ex-
posure (fig. S3B).
Reconstruction of population behavior from
single-cell data revealed that INH-mediated kill-
ing of microfluidic cultures followed first-order
kinetics [y = a exp(–bt), where a is the initial cell
number of each phase and y is the cell number
at time t after entering each phase], with rate
constants b = 3.50 (T0.02) × 10
−2
hour
−1
during
the k phase and b = 0.38 (T0.02) × 10
−2
hour
−1
during the p phase (Fig. 1D). During the d phase,
cell numbers increased (Fig. 1D) while aggregate
cell area stayed the same (Fig. 1E), indicating that
INH inhibits growth more rapidly than it inhibits
division. Reductive divisions caused a sharp but
transient decline in mean cell size (Fig. 1F, black
line). Single-cell size variation increased with time
(Fig. 1F, red lines), which suggested that INH
might perturb mechanisms coupling division to
cell size.
We tested the hypothesis that persisters are
preexisting, either slowly growing or nongrowing,
individuals by comparing the pre-INHgrowth rates
of single cells. We found that single-cell growth
was size-dependent, such that larger cells tended
to growfaster (Fig. 2A). The frequency distribution
of cell elongation rates was unimodal and could
be viewed as a Gaussian distribution with an addi-
tional left tail (P=0.28, Kolmogorov-Smirnov test
for the Gaussian null hypothesis) (Fig. 2B). For
cells exposed to INH for 72 or 144 hours, the
mean pre-INH elongation rates of persistent and
nonpersistent individuals were not significantly
different (P = 0.17 or 0.20, respectively, Welch’s t
test) (Fig. 2C). We conclude that single-cell growth
rates and cell fates are not correlated in INH-
stressed M. smegmatis, in contrast to ampicillin-
stressed E. coli (11).
Single-cell analysis revealed that the transi-
tion from fast (k phase) to slow ( p phase) killing
was due to a decreasing death rate rather than an
increasing division rate (Fig. 2D). Consequently,
the p phase is a dynamic state in which the rates
of cell division and death are balanced, resulting
in a deceptively static number of viable cells.
When bacteria divide, each newborn cell in-
herits a “new” pole from the last division and an
“old” pole from a previous division (Fig. 3A). Ac-
cording to a recent report, old-pole siblings elongate
faster and are more susceptible to antibiotics than
new-pole siblings (21). Although we did not ob-
serve a significant difference in elongation rates
of old- and new-pole siblings (P= 0.16, Welch’s t
test) (Fig. 3B), pole age could potentially influ-
ence survival for other reasons.
We tested whether the fates of sibling cells
were correlated by counting the number of obser-
vations of each of the four possible combinations
of cell fates for 1764 sibling pairs (Fig. 3C). A
cell “died” if it underwent lysis and “survived” if
it divided to produce two daughter cells. The num-
bers of events in each category (Fig. 3D) were used
to calculate the conditional survival probabilities
of old- and new-pole cells depending on the fate
of the paired sibling. We found that the survival
of sibling pairs was positively correlated (P= 2 ×
10
−7
, c
2
test) (Fig. 3E), suggesting the possible
contribution of epigenetic effects. This analysis
also revealed a weak positive association between
survival and inheritance of the old pole (P= 0.027,
c
2
test) (Fig. 3E); this association is the opposite
of that reported by Aldridge et al. (21).
INH is a prodrug that requires activation by
bacterial catalase-peroxidase (KatG) (fig. S4A)
(20). We investigated the relationship between
KatG levels and INH killing with the use of a
strain that expressed katG from an anhydrote-
tracycline (ATc)–inducible promoter. Cells grown
with increasing concentrations of ATc expressed
increasing levels of katG mRNA (fig. S4B) and
protein (fig. S4C). ATc-induced cultures were
exposed to INH and the fraction of surviving
colony-forming units was measured. These exper-
iments revealed a sensitive scaling relationship
between KatG levels and INH killing, such that
a factor of ~2 increase in KatG(fig. S4C) resulted
in a factor of ~1000 enhancement of killing at
48 hours (fig. S4D). This effect was not apparent
at 24 hours, indicating that overexpression of
KatG enhanced killing only during the p phase
(fig. S4D).
The sensitivity of INHkilling to small changes
in KatG expression suggested that cell-to-cell
variation in KatG levels might influence cell
fate under INHexposure. We tested this hypothe-
sis in a strain in which chromosomal katG was
replaced by a katG::dsRed2 fusion gene. The
wild-type and katG::dsRed2 strains were equally
sensitive to INH (fig. S5). Unexpectedly, single
cells expressed KatG-DsRed2 in short-lived sto-
chastic pulses (Fig. 4A and movie S3), separated
by intervals in which KatG-DsRed2 fluorescence
barely exceeded the detection threshold (Fig. 4B).
KatG-DsRed2 pulsing was independent of INH,
although cells were brighter in the presence of
the drug (movie S3). Native KatG protein was
less stable than green fluorescent protein (GFP),
which we used as an internal control (Fig. 4C).
These results suggest that active degradation
contributed to the sharp decline in KatG-DsRed2
fluorescence after peak intensity in pulsing cells.
We hypothesized that pulsing cells might be
more vulnerable to INH killing as a consequence
0
0.02
0.04
0.06
0.08
0 20 40 60 80 100 120 140
Death rate
Division rate
D
0
0.1
0.2
0.3
0.4
0.5
0 0.2 0.4
B
0
0.5
1
1.5
2
0 2 4 6 8
0
0.2
0.4
0.1 0.2 0.3 0
0
0.1
0.2
0.3
0.4
0.5
C 72 hours INH exposure 144 hours INH exposure
F
r
e
q
u
e
n
c
y
C
e
l
l

s
i
z
e

i
n
c
r
e
m
e
n
t

d
u
r
i
n
g
t
h
e

n
e
x
t

o
n
e

h
o
u
r

i
n
t
e
r
v
a
l

A
(
t
)

(
µ
m
2
)
F
r
e
q
u
e
n
c
y
Frequency
Elongation rate (h
-1
)
Cell size A(t) (µm
2
)
P
r
e
-
e
x
p
o
s
u
r
e
e
l
o
n
g
a
t
i
o
n

r
a
t
e

(
h
-
1
)
Non-persistent
n = 179
Persistent
n = 41
Non-persistent
n = 112
Persistent
n = 14
R
a
t
e

(
h
-
1
)
Time under INH (h)
n = 312
n = 312
A
P = 0.17 P = 0.20
k p d
Fig. 2. Persistence of single cells under INH exposure. (A) Relationship between cell size A(t) and size
increment DA(t), where Dt = 1 hour. Circles, single-cell measurements; squares, means T SE of DA(t) in
size windows n ≤ A(t) < (n + 1) mm
2
, where n is a positive integer. Histograms are frequency distributions
of A(t) (top) and DA(t) (right). (B) Single-cell elongation rates were calculated by measuring cell length
every 15 min from birth until next division and fitting to exponential curves. (C) Single-cell elongation
rates and cell fates under INH exposure. A “persistent” cell produced one or more progeny that survived; a
“nonpersistent” cell failed to produce any surviving progeny. Circles, single-cell elongation rates during
the 1-hour interval before INH exposure; squares, means T SD. (D) Rates of division and death under INH
exposure. Divisions (squares) and deaths (circles) were counted at 15-min intervals, normalized by total
cell number, and averaged over 10 hours T SE (see supplementary materials).
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 93
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

of increased activation of INH. We tested this
idea by comparing the survival probabilities of
pulsing (K = 1) and nonpulsing (K = 0) cells. This
analysis confirmed that pulsing cells were less
likely to survive (P=7.9 ×10
−5
, c
2
test) (Fig. 4D).
We also found that KatG-DsRed2 pulsing was
positively correlated between sibling pairs (P =
7.2 × 10
−8
, c
2
test) (Fig. 4E), which might explain
why cell fate was also correlated between siblings
(Fig. 3E).
Historically, fractional killing has been attri-
buted to subpopulations of nonreplicating bacte-
ria on the assumption that antibiotics targeting
growth processes should not damage growth-
arrested cells (8). Recently, this interpretation has
been challenged by the observation that growth
arrest per se is not sufficient to explain the re-
fractoriness of stationary-phase Pseudomonas
bacteria to antibiotics (22). In populations of ex-
ponentially growing mycobacteria, we found no
correlation between single-cell growth rates and
cell fates, inasmuch as slowly growing individ-
uals were as likely to die (or persist) as rapidly
growing individuals. These observations do not
rule out the possibility that in other systems, per-
sistence may be linked to single-cell growth rates,
as in E. coli exposed to ampicillin (11).
Although bacterial numbers are relatively
constant during the p phase of INH killing, we
found that this stability masks a dynamic state
of balanced division and death. Ongoing division
could promote the emergence of resistant genetic
variants because the rate of spontaneous mutation
increases in proportion to the division rate (23).
It has been argued that resistant mutants are un-
likely to arise de novo under antibiotic exposure
on the assumption that persistent cells are non-
replicating (16, 24). Our discovery that persistent
cells may grow and divide in the presence of an
antibiotic necessitates a reexamination of how
persistence might enhance the generation of re-
sistant mutants.
The link between stochastic expression of katG
and the fate of single cells under antibiotic stress
might provide insight into the generality of per-
sistence. In principle, stochastic expression of any
factor that facilitates or opposes the action of an
antibiotic could influence the fate of single cells.
The additive contribution of independently fluc-
tuating factors might explain why the link be-
tween KatG-DsRed2 pulsing and death, although
significant, was not absolute. Consistent with
this view, a recent study showed that expression
of E. coli HipA toxin from a “noisy” promoter
generates supra-threshold fluctuations that pro-
tect a fraction of cells from killing by ampicillin
(25). These observations raise the possibility that
natural cell-to-cell variation in HipAlevels might
influence survival probability under antibiotic
pressure.
Because KatG activates INH, the observed
low frequency of KatG pulsing should increase
cell-to-cell variation of INH activation (fig. S6A)
relative to other modes of KatG expression, such
as high-frequency pulsing (fig. S6B) or constitu-
tive expression (fig. S6C). This prediction is con-
sistent with our observation that pulsing and
nonpulsing individuals have different survival prob-
abilities. Because low-frequency, high-amplitude
pulsing should broaden the single-cell frequen-
cy distribution of activated INH concentrations
(fig. S6D), this regime is more likely to generate
lineages that persist because they fail to activate
INH to lethal levels.
Although the mechanistic basis of KatG
pulsing is unknown, we envisage several possi-
bilities. First, recent work in E. coli has estab-
lished that single-cell gene expression is inherently
stochastic and “bursty” (26). Burst frequency and
amplitude can be tuned by nucleotide changes
within the nucleic acid sequences that specify
the efficiencies of transcription and translation
(27). Thus, KatGpulsing may simply reflect low-
frequency, high-amplitude stochastic bursting.
Second, recent work in Bacillus subtilis demon-
strated that stochastic pulses of transcription are
generated by a network architecture comprising
a noise-triggered ultrasensitive switch coupled
to mixed (positive and negative) feedback loops
operating on different time scales (28). Similarly,
KatG pulsing might reflect the operation of net-
work motifs that randomly switch the katG pro-
moter between “on” and “off” states. Third, in
many bacteria, transcription of katG is induced
by oxidative stress (29). Thus, stochastic produc-
tion of reactive oxygen species during respiratory
metabolism (30) could trigger transient pulses of
katG transcription. In each of these hypothetical
scenarios, instability of KatG would contribute
to a sharp rise of KatG levels at the initiation of
a transcriptional pulse and a sharp fall of KatG
levels at the termination of such a pulse.
INH-mediated selection of cell lineages char-
acterized by infrequent pulsing of KatG could
result in gradual adaptation of the population as
less-fit phenotypic variants are eliminated. It is
also plausible that antibiotics or other extrinsic
stimuli might trigger adaptive responses; indeed,
environmental factors have been shown to mod-
ify the efficiency of antibiotic killing in several
systems (31–38). However, sense-and-respond
adaptive strategies are expensive because the
requisite molecular machinery must be expressed
constitutively (39). In contrast, nonresponsive,
selection-mediated adaptation could provide a
simple mechanism for cell populations to adapt
to stresses that they had never encountered in
A
C D
E
? 0 0 1 0 1 0 2 0 0 2 1 0 0 1 ?
0 0 1 ? 0 1 0 ?
0 ? 0 ?
? ?
4
3
2
1
G
e
n
e
r
a
t
i
o
n
Progenitor
B
0
0.1
0.2
0.3
0.4
0.5
P(S
o
=1 | S
n
=0)
0.32 ± 0.03
P(S
n
=1 | S
o
=0)
0.28 ± 0.03
P(S
o
=1 | S
n
=1)
0.44 ± 0.04
P(S
n
=1 | S
o
=1)
0.40 ± 0.04
P(S
o
=1)
0.36±0.02
Mother
cell
S
n
=0 S
o
=1
P(S
n
=1)
0.32±0.03
Mother
cell
S
o
=0
P(S
n
=1)
0.32±0.03
Mother
cell
Old-pole New-pole
E
l
o
n
g
a
t
i
o
n

r
a
t
e

(
h
-
1
)
S
n
=1
(Survival)
S
n
=0
(Death)
S
o
=1
(Survival)
S
o
=0
(Death)
Old-pole
New-pole
Total
Total
252 316
378 818
568
1196
630 1134 1764 (S
o
,S
n
)=(1,1)
Old New
(S
o
,S
n
)=(1,0)
Old New
(S
o
,S
n
)=(0,1)
Old New
(S
o
,S
n
)=(0,0)
Old New
Mother
cell
P(S
o
=1)
0.36±0.02
S
n
=1
Fig. 3. Positively correlated survival of sibling pairs. (A) Each individual inherits a “new” pole from the
most recent division (age 0) and an “old” pole from a previous division (age ≥1). In time-lapse
experiments, pole ages can be assigned from the third generation onward (white cells), excluding the
poles of unknown age inherited fromthe progenitor cell (gray cells). (B) Elongation rates of old- and new-
pole siblings. Circles, elongation rates of single cells pre-INH; squares, means T SD, which were 0.22 T
0.06 hour
−1
for old-pole cells (n = 112) and 0.21 T 0.07 hour
−1
for new-pole cells (n = 112). (C) Four
possible fates for a sibling pair. S = 1 if the cell survives until next division; S = 0 if the cell dies before
dividing; S
o
, old-pole sibling; S
n
, new-pole sibling. (D) Number of single-cell observations of the four fate
outcomes (n = 1764 cells). (E) Sibling pair correlation of cell fates. P(S = 1), total survival probability;
P(S
o
= 1), survival probability of old-pole cells; P(S
n
= 1), survival probability of new-pole cells. Numbers
indicate survival probabilities T 95% confidence intervals.
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 94
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

their evolutionary histories, or stresses that were
encountered too infrequently to repay the invest-
ment in responsive mechanisms.
References and Notes
1. A. Eldar, M. B. Elowitz, Nature 467, 167 (2010).
2. G. Balázsi, A. van Oudenaarden, J. J. Collins, Cell 144,
910 (2011).
3. N. Dhar, J. D. McKinney, Curr. Opin. Microbiol. 10, 30
(2007).
4. K. Lewis, Annu. Rev. Microbiol. 64, 357 (2010).
5. N. Q. Balaban, Curr. Opin. Genet. Dev. 21, 768 (2011).
6. A. Brock, H. Chang, S. Huang, Nat. Rev. Genet. 10,
336 (2009).
7. S. V. Sharma et al., Cell 141, 69 (2010).
8. J. Bigger, Lancet 244, 497 (1944).
9. G. Hobby, K. Meyer, E. Chaffee, Proc. Soc. Exp. Biol. Med.
50, 281 (1942).
10. G. Hobby, M. Dawson, Proc. Soc. Exp. Biol. Med. 56,
181 (1944).
11. N. Q. Balaban, J. Merrin, R. Chait, L. Kowalik, S. Leibler,
Science 305, 1622 (2004).
12. E. Kussell, R. Kishony, N. Q. Balaban, S. Leibler, Genetics
169, 1807 (2005).
13. D. Shah et al., BMC Microbiol. 6, 53 (2006).
14. I. Keren, S. Minami, E. Rubin, K. Lewis, MBiol. 2, e00100
(2011).
15. N. G. Cogan, J. Theor. Biol. 238, 694 (2006).
16. B. R. Levin, D. E. Rozen, Nat. Rev. Microbiol. 4, 556 (2006).
17. A. Gardner, S. A. West, A. S. Griffin, PLoS ONE 2, e752
(2007).
18. A. R. Coates, Y. Hu, Trends Pharmacol. Sci. 29, 143
(2008).
19. J. G. Hurdle, A. J. O’Neill, I. Chopra, R. E. Lee, Nat. Rev.
Microbiol. 9, 62 (2011).
20. C. Vilchèze, W. R. Jacobs Jr., Annu. Rev. Microbiol. 61, 35
(2007).
21. B. B. Aldridge et al., Science 335, 100 (2012).
22. D. Nguyen et al., Science 334, 982 (2011).
23. H. E. Kubitschek, H. E. Bendigkeit, Mutat. Res. 106, 113
(1964).
24. M. Lipsitch, B. R. Levin, Antimicrob. Agents Chemother.
41, 363 (1997).
25. E. Rotem et al., Proc. Natl. Acad. Sci. U.S.A. 107,
12541 (2010).
26. G. W. Li, X. S. Xie, Nature 475, 308 (2011).
27. H. Maamar, A. Raj, D. Dubnau, Science 317, 526 (2007).
28. J. C. W. Locke, J. W. Young, M. Fontes, M. J. Hernández
Jiménez, M. B. Elowitz, Science 334, 366 (2011).
29. D. R. Sherman et al., Proc. Natl. Acad. Sci. U.S.A. 92,
6625 (1995).
30. S. Iuchi, L. Weiner, J. Biochem. 120, 1055 (1996).
31. C. Miller et al., Science 305, 1629 (2004).
32. N. Dhar, J. D. McKinney, Proc. Natl. Acad. Sci. U.S.A.
107, 12275 (2010).
33. T. Dörr, M. Vulić, K. Lewis, PLoS Biol. 8, e1000317 (2010).
34. K. N. Adams et al., Cell 145, 39 (2011).
35. K. R. Allison, M. P. Brynildsen, J. J. Collins, Nature 473,
216 (2011).
36. S. H. Baek, A. H. Li, C. M. Sassetti, PLoS Biol. 9,
e1001065 (2011).
37. Y. Wu, M. Vulic, I. Keren, K. Lewis, Antimicrob. Agents
Chemother. 56, 4922 (2012).
38. S. S. Grant, B. B. Kaufmann, N. S. Chand, N. Haseley,
D. T. Hung, Proc. Natl. Acad. Sci. U.S.A. 109, 12147 (2012).
39. E. Kussell, S. Leibler, Science 309, 2075 (2005).
Acknowledgments: Supported by fellowships from the
Charles H. Revson Foundation and the Heiser Program for
Research in Leprosy and Tuberculosis of the New York
Community Trust (N.D.); a Harvey L. Karp Discovery Fellowship
and the JST PRESTO Program (Y.W.); and the Bill and
Melinda Gates Foundation, NIH grant HL088906, and Swiss
National Science Foundation grant 310030_135639 ( J.D.M.).
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/91/DC1
Materials and Methods
Figs. S1 to S6
Movies S1 to S3
References (40–43)
6 September 2012; accepted 15 November 2012
10.1126/science.1229858
Fig. 4. Single-cell KatG
pulsing and cell fate. Bacte-
ria expressing KatG-DsRed
were imaged on fluores-
cence and phase channels
for 88 hours at 15-min in-
tervals and exposed to INH
(50mg/ml) at 16to88hours.
This experiment was repeated
five times. (A) Representa-
tive image series. Numbers
indicate hours under INHex-
posure. Scale bar, 5 mm. (B)
Single-cell pedigrees (upper
panels) and time traces of
KatG-DsRed fluorescence in
arbitrary units (lower pan-
els) for nonpersistent (left)
and persistent (right) line-
ages. Black points in upper
panels indicate pulse peaks.
(C) KatGstability. Cells were
pulse-labeled with [
35
S]Met
and [
35
S]Cys and chased
A
2.0 h 4.0 h 5.5 h 7.0 h
B
E
P(K
o
=1 | K
n
=0)
0.43 ± 0.11
n = 84
P(K
n
=1 | K
o
=0)
0.41 ± 0.11
n = 81
P(K
o
=1 | K
n
=1)
0.78 ± 0.07
n = 149
P(K
n
=1 | K
o
=1)
0.76 ± 0.07
n = 152
P(K
o
=1)
0.65±0.06
Mother
cell
K
n
=0 K
o
=1
P(K
n
=1)
0.64±0.06
Mother
cell
K
o
=0
P(K
n
=1)
0.64±0.06
Mother
cell
Mother
cell
P(K
o
=1)
0.65±0.06
K
n
=1
D
0 0.5 1.0
Chase time (h)
C
62
49
28
Survival
probability
Conditional
survival probability
P(S=1)
0.21 ± 0.04
n = 488
P(S=1 | K=1)
0.16 ± 0.04
P(S=1 | K=0)
0.31 ± 0.07
Yes
K=1
n = 313
No
K=0
n = 175
Pulse
Non-persistent lineage Persistent lineage
L
a
b
e
l
e
d

p
r
o
t
e
i
n

(
%
)
Chase time (h)
INH INH INH INH
0
2
4
6
8
0 10 20 30 40 50
Time under INH (h)
0 10 20 30 40 50
Time under INH (h)
with unlabeled Met and Cys. Left: Pull-downs of
cell lysates with antibody to KatG (upper panel) or
GFP-Trap (lower panel) were analyzed by SDS–
polyacrylamide gel electrophoresis. Right: Band
intensities were quantified by scanning auto-
radiograms. Half-lives were ~0.7 hours for KatG
versus ~5.7 hours for GFP. This experiment was re-
peated twice. (D) Negatively correlated KatG-DsRed
pulsing and cell survival. K = 1, pulse; K = 0, no
pulse. Values are survival probabilities T 95%con-
fidence intervals. (E) Sibling pair correlation of
KatG-DsRed pulsing. K
o
= 1, old-pole cell pulsed;
K
n
= 1, new-pole cell pulsed (pulsing indicated
by red outlining). Values are pulse probabilities T
95% confidence intervals.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 95
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

The End of History Illusion
Jordi Quoidbach,
1,2
Daniel T. Gilbert,
2
* Timothy D. Wilson
3
We measured the personalities, values, and preferences of more than 19,000 people who ranged in
age from 18 to 68 and asked them to report how much they had changed in the past decade and/or
to predict how much they would change in the next decade. Young people, middle-aged people, and
older people all believed they had changed a lot in the past but would change relatively little in the future.
People, it seems, regard the present as a watershed moment at which they have finally become the
person they will be for the rest of their lives. This “end of history illusion” had practical consequences,
leading people to overpay for future opportunities to indulge their current preferences.
A
t every stage of life, people make de-
cisions that profoundly influence the
lives of the people they will become—
and when they finally become those people,
they aren’t always thrilled about it. Young adults
pay to remove the tattoos that teenagers paid
to get, middle-aged adults rush to divorce the
people whom young adults rushed to marry,
and older adults visit health spas to lose what
middle-aged adults visited restaurants to gain.
Why do people so often make decisions that their
future selves regret?
One possibility is that people have a funda-
mental misconception about their future selves.
Time is a powerful force that transforms people’s
preferences, reshapes their values, and alters
their personalities, and we suspect that people
generally underestimate the magnitude of those
changes. In other words, people may believe
that who they are today is pretty much who they
will be tomorrow, despite the fact that it isn’t
who they were yesterday. In the studies we de-
scribe here, we showed that people expect to
change little in the future, despite knowing that
they have changed a lot in the past, and that this
tendency bedevils their decision-making. We call
this tendency to underestimate the magnitude
of future change the “end of history illusion.”
To investigate this phenomenon, we asked
samples of people who varied widely in age to
predict how much they would change over the
next 10 years, we asked similar samples to re-
port how much they had changed over the past
10 years, and we compared the predictions of
people aged a years to the reports of people aged
a +10 years. We expected people aged a years
to predict less change over the next 10 years
than people aged a +10 years reported over the
past 10 years. We used this strategy to study how
much people thought they would change in the
domains of personality (a person’s character-
istic patterns of behavior), core values (a person’s
ideals and principles), and preferences (a person’s
likes and dislikes).
In study 1, we sought to determine whether
people underestimate the extent to which their
personalities will change in the future. We re-
cruited a sample of 7519 adults ranging in age
from 18 to 68 years [mean (M) = 40 years, stan-
dard deviation (SD) = 11.3 years, 80% women]
through the Web site of a popular television show
and asked them to complete the Ten Item Per-
sonality Inventory (1), which is a standard mea-
sure of the five trait dimensions that underlie
human personality (i.e., conscientiousness, agree-
ableness, emotional stability, openness to experi-
ence, and extraversion). Participants were then
randomly assigned either to the reporter condi-
tion (and were asked to complete the measure as
they would have completed it 10 years earlier) or
the predictor condition (and were asked to com-
plete the measure as they thought they would
complete it 10 years hence). We then computed
the absolute value of the difference between par-
ticipants’ ratings of their current personality and
their reported or predicted personality and aver-
aged these across the five traits to create a mea-
sure of reported or predicted change in personality.
Additional methodological details about study
1 can be found in supplementary text 1 to 3.
We analyzed these measures by first assigning
a value to each of the 41 10-year periods between
ages 18 and 68. We called this variable “decade.”
For each decade, we compared the predictions of
predictors aged a to the reports of reporters aged
a + 10 years. So, for example, when decade = 1,
we compared 18-year-old predictors and 28-year-
old reporters; when decade = 2, we compared
19-year-old predictors and 29-year-old reporters;
and so on. We did not collect data from reporters
who were younger than 28 years, because in our
sample there were no predictors younger than
18 years with whom to compare them, and we
did not collect data from predictors who were
older than 58 years, because in our sample there
were no reporters older than 68 years with whom
to compare them.
We entered participants’ reported or predicted
changes in personality into a multiple regression
analysis with three predictor variables: decade
(coded 1 through 41), condition (coded 1 for pre-
dictors and –1 for reporters), and a “decade X
condition” interaction. First, the analysis revealed
an effect of decade [beta coefficient (b) = –0.13,
P<0.001], indicating that the older the participants
were, the less personality change they reported or
predicted. This finding is consistent with a large
body of research showing that personality becomes
more stable as people age (2). Second, the anal-
ysis revealed the expected effect of condition (b =
–0.14, P < 0.001). The top panel of Fig. 1 shows
this end of history illusion: Predictors aged a
predicted that they would change less over the
next decade than reporters aged a + 10 years
reported having changed over the same decade.
Finally, there was no decade X condition inter-
action (b = 0.01, P = 0.68), indicating that the
magnitude of the end of history illusion did not
change across decades. Next, we conducted follow-
up studies to answer three questions.
First, is it possible that the discrepancy be-
tween participants’ reports and predictions in
study 1 was due entirely to the erroneous mem-
ory of reporters, who may have overestimated
how much they had changed in the past 10 years,
rather than to the erroneous predictions of pre-
dictors, who may have underestimated howmuch
they would change in the next 10 years? To in-
vestigate this possibility, we compared the mag-
nitudes of the predicted and reported personality
changes in our sample to the magnitude of ac-
tual personality change observed in an indepen-
dent sample of 3808 adults ranging from 20 to
75 years old (M = 47.2 years, SD = 12.4 years,
55% women), whose personalities had been mea-
sured as part of the MacArthur Foundation Sur-
vey of Midlife Development in the United States
(MIDUS). These adults completed the MIDUS
Big Five scale (3) for the first time in 1995–1996
(MIDUS 1) and for a second time in 2004–2006
(MIDUS 2). The MIDUS Big Five scale has good
construct validity and correlates with other sim-
ilar scales (4, 5). Because the personality mea-
sures used in the MIDUS study and in our study
were scored on different scales, direct compari-
son of the data was not possible. To estimate
the magnitudes of actual, reported, and predicted
personality change, we computed intraclass cor-
relations (ICC-A1), which account for both ab-
solute and rank-based change (6). Specifically,
we computed (i) the ICCbetween the two admin-
istrations of the personality test in the MIDUS
sample, which was 0.52; (ii) the ICC between cur-
rent and reported personality for participants in
our sample, which was 0.51; and (iii) the ICC be-
tween current and predicted personality for par-
ticipants in our sample, which was 0.65. Larger
ICCs, of course, indicate less personality change.
As inspection of these ICCs reveals, the magnitude
of reported personality change in our sample was
almost identical to the magnitude of actual per-
sonality change in the MIDUS sample, suggesting
that participants in our sample were relatively ac-
curate when reporting the amount of change they
had experienced in the past. However, the magni-
tude of actual personality change in the MIDUS
sample was substantially larger than the magni-
tude of predicted personality change in our sam-
ple, suggesting that participants in our sample
were relatively inaccurate when predicting the
amount of change they would experience in the
future. In short, it seems likely that the discrepan-
cy between the reported and predicted personality
1
National Fund for Scientific Research, Brussels, Belgium.
2
Department of Psychology, Harvard University, Cambridge,
MA 02138, USA.
3
Department of Psychology, University of
Virginia, Charlottesville, VA 22904–4400, USA.
*To whom correspondence should be addressed. E-mail:
gilbert@wjh.harvard.edu
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 96
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

changes of participants in study 1 is due at least
in part to errors of prediction and not merely to
errors of memory. Study 3 provides further support
for this claim.
Second, is it possible that reporters and pre-
dictors in study 1 interpreted the scales differ-
ently, so that words such as “conscientious” or
“agreeable” meant one thing to reporters and
another thing to predictors? To investigate this
possibility, we replicated study 1 with an indepen-
dent sample of 613 adults (M = 40.5 years, SD =
8.4 years, 86.6% women) recruited through the
same Web site and using a design in which each
participant was assigned to both the reporter and
the predictor conditions, thus ensuring that any
idiosyncratic interpretation of the scales would
influence both conditions equally. This design
required that we restrict our sample to participants
aged 28 to 58. Because participants contributed
data to both conditions, we performed a multi-
level version of the analysis described in study
1. The analysis revealed the expected effect of
condition (b = –7.69, P = 0.001), indicating
that predictors aged a years predicted that they
would change less over the next decade than re-
porters aged a + 10 years reported having changed
over the same decade—even though the reports
and predictions were made by the same partic-
ipants. This finding suggests that idiosyncratic
interpretations of the scale are not the cause of
the effects seen in study 1.
Third, is it possible that predictors in study
1 knew that they would change over the next
10 years, but because they did not know exactly
how they would change, they did not feel confi-
dent predicting specific changes? To investigate
this possibility, we replicated study 1 with an inde-
pendent sample of 1163 adults (M = 38.4 years,
SD = 12.1 years, 78% women) recruited through
the same Web site. Instead of being asked to
report or predict their specific personality traits,
these participants were simply asked to report
how much they felt they had “changed as a
person over the last 10 years” and how much they
thought they would “change as a person over the
next 10 years.” Because some participants con-
tributed data to both conditions, we performed a
multilevel version of the analysis described in
study 1. The analysis revealed the expected effect
of condition (b = –0.74, P = 0.007), indicating
that predictors aged a years predicted that they
would change less over the next decade than
reporters aged a + 10 years reported having changed
over the same decade. This finding suggests that
a lack of specific knowledge about how one
might change in the future was not the cause of
the effects seen in study 1.
In study 2, we sought to determine whether
the end of history illusion was limited to the do-
main of personality, and so we repeated our pro-
cedure in the domain of core values. We recruited
a new sample of 2717 adults ranging in age from
18 to 68 years (M = 38.6 years, SD = 10.6 years,
82% women) through the same Web site and
asked them to indicate the importance of each
of 10 basic values (such as hedonism, success,
security, etc.) that were taken from the Schwartz
Value Inventory (7). Otherwise, the design was
identical to that of study 1.
We performed a regression analysis similar to
the one performed in study 1. First, the analysis
revealed an effect of decade (b = –0.23, P <
0.0010, indicating that the older participants
were, the less change in their core values they
reported or predicted. Second, the analysis re-
vealed the expected effect of condition (b =
–0.46, P < 0.001). The middle panel of Fig. 1
shows this end of history illusion: Predictors
aged a years predicted that they would change
less over the next decade than reporters aged
a + 10 years reported having changed over the
Fig. 1. Standardized predicted and reported changes across decades in study 1 (top panel), study 2 (middle panel), and study 3 (bottom panel). The graph
shows moving averages smoothed with a 4-year Gaussian filter. Additional information about this figure can be found in supplementary text 4.
www.sciencemag.org SCIENCE VOL 339 4 JANUARY 2013 97
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

same decade. Finally, the analysis revealed a dec-
ade × condition interaction (b = 0.08, P < 0.001).
Although the magnitude of the end of history
illusion decreased as participants got older, it was
nonetheless present even in the oldest group of
participants (aged 50 and up) (b = –0.34, P <
0.001). Further discussion of this decade × con-
dition interaction can be found in supplementary
text 5.
The foregoing studies show that people ex-
pect to experience less change in their person-
alities and core values over the next decade than
people a decade older report having experienced
over the past decade. The analysis presented in
study 1 suggests that this discrepancy represents,
at least in part, an error of prediction and is not
merely an error of memory. To provide further
support for this claim, in study 3 we examined
the end of history illusion in a domain in which
memory was likely to be highly reliable. Rather
than asking reporters to remember how extra-
verted they had been or how much they had once
valued honesty, we asked themto remember sim-
ple facts about their strongest preferences, such
as the name of their favorite musical band or
the name of their best friend. We reasoned that
if participants remembered having a different
best friend 10 years ago but expected to have
the same best friend 10 years from now, then this
was probably not due to a pervasive tendency
for people of all ages to actually keep their best
friends but mistakenly remember changing them.
To test this hypothesis, we recruited a new
sample of 7130 adults ranging from 18 to 68 years
old (M=40.2 years, SD=11.1 years, 80%women)
through the same Web site and asked them to
report their favorite type of music, their favorite
type of vacation, their favorite type of food, their
favorite hobby, and the name of their best friend.
Participants were then randomly assigned either
to the reporter condition (and were asked to report
whether each of their current preferences was
the same as or different than it was 10 years ago) or
the predictor condition (and were asked to predict
whether each of their current preferences would be
the same or different 10 years fromnow). We then
counted the number of items on which participants
responded “different” and used this as a measure
of reported or predicted changes in preference.
We performed a regression analysis similar
to the ones performed in studies 1 and 2. First,
the analysis revealed an effect of decade (b =
–0.14, P < 0.001). The older participants were,
the less change in preferences they reported or
predicted. Second, the analysis revealed the ex-
pected effect of condition (b = –0.19, P < 0.001).
The bottom panel in Fig. 1 shows this end of
history illusion: Predictors aged a years pre-
dicted that their preferences would change less
over the next decade than reporters aged a + 10
years reported that their preferences had changed
over the same decade. Finally, the analysis re-
vealed a decade × condition interaction (b =
0.07, P < 0.001). Although the magnitude of
the end of history illusion decreased as partic-
ipants got older, it was nonetheless present even
in the oldest group of participants (aged 50 and
up) (b = –0.08, P < 0.01). Further discussion of this
decade × condition interaction can be found in
supplementary text 5, and additional details about
study 3 can be found in supplementary text 6.
The foregoing studies suggest that people
underestimate the extent to which their person-
alities, values, and preferences will change in the
future. In study 4, we sought to show that this
end of history illusion can have practical conse-
quences. Specifically, we sought to show that be-
cause people overestimate the stability of their
current preferences, they will overpay for future
opportunities to indulge them.
In study 4, we recruited a new sample of 170
adults ranging from 18 to 64 years old (M =
34.9 years, SD = 10.6 years, 52% women) through
the Amazon Mechanical Turk Web site (8, 9).
Some participants were randomly assigned to
the “future concert” condition. These participants
were asked to name their current favorite musical
band and then to report the maximum amount
of money they thought they would be willing
to pay today in order to see that band perform in
10 years. Other participants were randomly as-
signed to the “present concert” condition. These
participants were asked to name the musical band
that was their favorite 10 years ago and then to
report the amount of money that they thought
they would be willing to pay today to see that
band perform in the coming week.
We performed a regression analysis similar to
the ones performed in studies 1, 2, and 3. First,
the analysis revealed the expected effect of con-
dition (b = 0.16, P < 0.05). Participants aged a
years thought they would pay 61% more to
see their current favorite band perform 10 years
in the future (M = $129) than participants aged
a + 10 years thought they would pay to see their
once-favorite band perform in the present (M =
$80). The analysis revealed no effect of decade
(b = –0.06, P = 0.41), indicating that the price
participants thought they would pay did not vary
with age, and no decade × condition interaction
(b = 0.01, P = 0.94), indicating that willingness
to pay more for a future concert than a present
concert did not diminish in magnitude as partic-
ipants got older. In short, participants substantial-
ly overpaid for a future opportunity to indulge a
current preference.
Across six studies of more than 19,000 par-
ticipants, we found consistent evidence to indi-
cate that people underestimate how much they
will change in the future, and that doing so can
lead to suboptimal decisions. Although these data
cannot tell us what causes the end of history
illusion, two possibilities seem likely. First, most
people believe that their personalities are attract-
ive, their values admirable, and their preferences
wise (10); and having reached that exalted state,
they may be reluctant to entertain the possibility
of change. People also like to believe that they
know themselves well (11), and the possibility
of future change may threaten that belief. In short,
people are motivated to think well of themselves
and to feel secure in that understanding, and the
end of history illusion may help them accomplish
these goals.
Second, there is at least one important differ-
ence between the cognitive processes that allow
people to look forward and backward in time
(12). Prospection is a constructive process, ret-
rospection is a reconstructive process, and con-
structing new things is typically more difficult
than reconstructing old ones (13, 14). The reason
this matters is that people often draw inferences
from the ease with which they can remember or
imagine (15, 16). If people find it difficult to
imagine the ways in which their traits, values, or
preferences will change in the future, they may
assume that such changes are unlikely. In short,
people may confuse the difficulty of imagining
personal change with the unlikelihood of change
itself.
Although the magnitude of this end of his-
tory illusion in some of our studies was greater
for younger people than for older people, it was
nonetheless evident at every stage of adult life
that we could analyze. Both teenagers and grand-
parents seem to believe that the pace of personal
change has slowed to a crawl and that they have
recently become the people they will remain.
History, it seems, is always ending today.
References and Notes
1. S. D. Gosling, P. J. Rentfrow, W. B. Swann Jr., J. Res. Pers.
37, 504 (2003).
2. B. W. Roberts, K. E. Walton, W. Viechtbauer, Psychol. Bull.
132, 1 (2006).
3. M. E. Lachman, S. L. Weaver, The Midlife Development
Inventory (MIDI) Personality Scales: Scale Construction
and Scoring (Brandeis University, Waltham, MA, 1997).
4. D. K. Mroczek, C. M. Kolarz, J. Pers. Soc. Psychol. 75,
1333 (1998).
5. K. M. Prenda, M. E. Lachman, Psychol. Aging 16, 206 (2001).
6. K. O. McGraw, S. P. Wong, Psychol. Methods 1, 30 (1996).
7. S. H. Schwartz, in Advances in Experimental Social
Psychology, M. P. Zanna, Ed. (Academic Press, Orlando,
FL, 1992), vol. 25, pp. 1–65.
8. G. Paolacci, J. Chandler, P. G. Ipeirotis, Judgm. Decis. Mak.
5, 411 (2010).
9. M. Buhrmester, T. Kwang, S. D. Gosling, Perspect.
Psychol. Sci. 6, 3 (2011).
10. C. Sedikides, M. D. Alicke, in The Oxford Handbook of
Human Motivation, R. M. Ryan, Ed. (Oxford Univ. Press,
Oxford), pp. 303–322.
11. W. B. Swann Jr., in Handbook of Theories of Social
Psychology, P. Van Lang, A. Kruglanski, E. T. Higgins, Eds.
(Sage, London, 2012), pp. 23–42.
12. D. R. Addis, A. T. Wong, D. L. Schacter, Neuropsychologia
45, 1363 (2007).
13. M. D. Robinson, G. L. Clore, Psychol. Bull. 128, 934 (2002).
14. M. Ross, Psychol. Rev. 96, 341 (1989).
15. N. Schwarz et al., J. Pers. Soc. Psychol. 61, 195 (1991).
16. A. Tversky, D. Kahneman, Cognit. Psychol. 5, 207 (1973).
Acknowledgments: We acknowledge the support of Research
Grant BCS-0722132 from NSF to D.T.G. and T.D.W. Raw
data from all studies are on deposit at the Inter-university
Consortium for Political and Social Research (deposit no. 32668)
and can be accessed at www.icpsr.umich.edu.
Supplementary Materials
www.sciencemag.org/cgi/content/full/339/6115/96/DC1
Supplementary Text
Table S1
24 August 2012; accepted 16 November 2012
10.1126/science.1229294
4 JANUARY 2013 VOL 339 SCIENCE www.sciencemag.org 98
REPORTS

o
n

J
a
n
u
a
r
y

3
,

2
0
1
3
w
w
w
.
s
c
i
e
n
c
e
m
a
g
.
o
r
g
D
o
w
n
l
o
a
d
e
d

f
r
o
m

What can a RabMAb
®
do for you?
Discover more at abcam.com
Rabbit Monoclonal Antibodies (RabMAbs)
offer multiple advantages to bring you the
highest quality antibody possible.
#4
#2
#6
#7
High Affinity
Fully Validated
Novel Epitopes
Multiple Applications
#2
Fully Validated Fully Validated
#5
#5 High Specificity
High Specificity
offer high
specificity which is essential in
detecting post-translational
modifications such as
phosphorylation:
RabMAbs
®
WB on NIH/3T3 using anti-Akt 1
Phospho (pT450) RabMAb
(ab108266). (A) untreated or (B)
treated with Lambda Phospha-
tase. Akt1 (ab32510) as a pan
control
Try a RabMAb and discover more at abcam.com/rabmabs-advantages
Akt1 RabMAb
(ab32510)
R
a
b
MA
b
s ®
H
i
g
h
Q
u
a
lity A
n
t
i
b
o
d
i
e
s
CELL SCIENCES INC • 480 NEPONSET STREET, BUILDING 12A, CANTON, MA 02021 • INFO@CELLSCIENCES.COM
TOLL FREE: (888) 769-1246 • TEL: (781) 828-0610 • FAX: (781) 828-0542 • WEB: WWW.CELLSCIENCES.COM
w
w
w
.
c
e
l
l
s
c
i
e
n
c
e
s
.
c
o
m
w
w
w
.
c
e
l
l
s
c
i
e
n
c
e
s
.
c
o
m
cell sciences
Æ
cytokine center
Æ
LIST OF PROTEINS
4-1BBL
4-1BB Receptor
6 Ckine
ACAD8
ACAT2
gAcrp30/Adipolean
Activin A
ACY1
ADAT1
Adiponectin
ADRP
AITRL
Akt1
Alpha-Feto Protein (AFP)
Alpha-Galactosidase A
Angiopoietin-1 (Ang-1)
Angiopoietin-2 (Ang-2)
Angiostatin K1-3
Annexin-V
apo-SAA
Apoliprotein A-1
Apoliprotein E2
Apoliprotein E3
Apoliprotein E4
APRIL
Artemin
ATF2
Aurora A
Aurora B
BAFF
BAFF Receptor
BCA-1 / BLC / CXCL13
BCMA
BD-1
BD-2
BD-3
BDNF
Bivalirudin
BMP-2
BMP-4
BMP-7
BMP-13
sBMPR-1A
Brain Natriuretic Protein
BRAK
Breast Tumor Antigen
C5a
C5L2 Peptide
C-10
C-Reactive Protein
C-Src
Calbindin D-9K
Calbindin D-28K
Calbindin D-29K
Calmodulin
Calcitonin Acetate
Carbonic Anhydrase III
Carcino-embryonicAntigen
Cardiotrophin-1
Caspase-3
Caspase-6
CD4
CD14
CD22
CD40 Ligand / TRAP
CD95 / sFas Ligand
CD105 / Endoglin
CHIPS
CNTF
Collagen
CREB
CTACK/CCL27
CTGF
CTGFL/WISP-2
CTLA-4/Fc
CXCL16
Cytokeratin 8
DEP-1
Desmopressin
Disulfide Oxidoreductase
E-selectin
ECGF
EGF
Elafin/SKALP
EMAP-II
ENA-78
Endostatin
Enteropeptidase
Eotaxin
Eotaxin-2
Eotaxin-3 (TSC)
EPHB2
EPHB4
Eptifbatide
Erk-2
Erythropoietin (EPO)
Exodus-2
Fas Ligand
Fas Receptor
FGF-1 (acidic)
FGF-2 (basic)
FGF-4
FGF-5
FGF-6
FGF-7/ KGF
FGF-8
FGF-9
FGF-10
FGF-16
FGF-17
FGF-18
FGF-19
FGF-20
sFGFR-1 (IIIc) / Fc Chimera
sFGFR-2 (IIIc) / Fc Chimera
sFGFR-3 / Fc Chimera
sFGFR-4 / Fc Chimera
sFlt-1 (native)
sFlt-1 (D3)
sFlt-1 (D4)
sFlt-1 (D5)
sFlt-1 (D7)
Flt3-Ligand
sFlt-4
sFlt-4/ Fc Chimera
Follistatin
FSH
Fractalkine/ CX3C
G-CSF
α-Galactosidase A
Galectin-1
Galectin-3
Gastrointestinal CA
GCP-2
GDF-3
GDF-9
GDF-11
GDNF
GLP-1
Glucagon
Goserelin
GM-CSF
GPBB
GROα
GROβ
GROγ
GRO/MGSA
Growth Hormone
Growth Hormone BP
GST-p21/WAF-1
HB-EGF
HCC-1
HGF
Histdyl-tRNA synthetase
Histrelin
HRG1-β1
I-309
I-TAC
IFN-α
IFN-α A
IFN-α 2a
IFN-α 2b
IFN-β
IFN-γ
IFN-Omega
IGF-I
IGF-II
proIGF-II
IGFBP-1
IGFBP-2
IGFBP-3
IGFBP-4
IGFBP-4
IGFBP-5
IGFBP-6
IGFBP-7
IL-1α
IL-1β
Browse our web site of recombinant
proteins, including cytokines, growth
factors, chemokines and neurotrophins.
Daily shipping and competitive pricing are
offered. Bulk quantities of many proteins
available. Cell Sciences also carries
corresponding antibodies and ELISA kits.
IL-2
IL-3
IL-4
sIL-4 Receptor
IL-5
IL-6
sIL-6 Receptor
IL-7
IL-8 (72 a.a.)
IL-8 (77 a.a.)
IL-9
IL-10
IL-11
IL-12
IL-13
IL-13 analog
IL-15
IL-16 (121 a.a.)
IL-16 (130 a.a.)
IL-17
IL-17B
IL-17D
IL-17E
IL-17F
IL-19
IL-20
IL-22
IL-31
Insulin
IP-10
JE
JNK2a1
JNK2a2
KC / CXCL1
KGF
L-asparaginase
LAG-1
LALF Peptide
LAR-PTP
LC-1
LBP
LD-78β
LDH
LEC/NCC-4
Leptin
LIGHT
LIX
LKM
LL-37
Lymphotactin
sLYVE-1
M-CSF
MCP-1 (MCAF)
MCP-2
MCP-3
MCP-4
MCP-5
MDC (67 a.a.)
MDC (69 a.a.)
MDH
MEC
Mek-1
MIA
Midkine
MIG / CXCL9
MIP-1α / CCL3
MIP-1β / CCL4
MIP-3 / CCL23
MIP-3α / CCL20
MIP-3β / CCL19
MIP-4 (PARC) / CCL18
MIP-5 / CCL15
MMP-3
MMP-7
MMP-13
Myostatin
Nanog
NAP-2
Neurturin
NFAT-1
beta-NGF
NOGGIN
NOV
NP-1
NT-1/BCSF-3
NT-3
NT-4
Ocreotide
Oncostatin M
Osteoprotegerin (OPG)
OTOR
Oxytocin
p38-α
Parathyroid Hormone
PDGF-AA
PDGF-AB
PDGF-BB
Persephin
PF-4
PIGF-1
PIGF-2
PKA α-subunit
PKC-α
PKC-γ
Pleiotrophin
PLGF-1
Polymyxin B (PMB)
PRAS40
PRL-1
PRL-2
PRL-3
Prokineticin-2
Prolactin
Protirelin
PTHrP
PTP1B
PTP-IA2
PTP-MEG2
PTP-PEST
sRANK
sRANKL
RANTES
RELM-α
RELM-β
Resistin
RPTPβ
RPTPγ
RPTPμ
SCF
SCGF-α
SCGF-β
SDF-1α
SDF-1β
Secretin
SF20
SHP-2
STAT1
c-Src
TACI
TARC
TC-PTP
TECK
TFF2
TGF-α
TGF-β1
TGF-β2
TGF-β3
Thymosin α1
sTIE-1/Fc Chimera
sTIE-2/Fc Chimera
TL-1A
TNF-α
TNF-β
sTNFR1
sTNFR2
TPO
TRAIL/Apo2L
sTRAIL R-1 (DR4)
sTRAIL R-2 (DR5)
TSH
TSLP
TWEAK
TWEAK Receptor
Urokinase
VEGF121
VEGF145
VEGF165
VEGF-C
VEGF-C I525
EG-VEGF
VEGF-E
HB-VEGF-E
sVEGFR-1
sVEGFR-2
sVEGFR-3
WISP-1
WISP-2
WISP-3
WNT-1
BLItz brilliantly packs the power of Dip and
Read™ label-free analysis into a personal assay
system. Give BLItz a drop of your sample and it
does the rest!
• Protein presence/absence in seconds
• Binding kinetics assays at your bench
• Protein quantitation in seconds
• Develop immunoassays in minutes
• Easily analyze crude samples
Cleverly priced under $20K so you can have
your own little genius.
Want to try BLItz in your lab? Visit
BlitzMeNow.com or call 855.BLITZ.ME.
Little Genius
MiSeq
®
. Next-generation sequencing for all you seek.
You want amazing accuracy and performance on your benchtop. Illumina
delivers—yet again. The MiSeq personal sequencer is the only fully integrated,
truly end-to-end benchtop solution around. It’s just one more example of
why Illumina solutions generate a remarkable 90% of all the world’s sequencing
data. Discover what’s possible.
www.illumina.com/miseq
I seek to get there frst.
Discover the
Source
Reagent Proteinsisthesource
With over 5,000 recombinant proteins available,
Reagent Proteins provides seamless access to the
highest quality reagent, pre-clinical and cGMPgrade
proteins for research purposes.
reagentprotei ns. com phone 858 · 352 · 4400 fax 858 · 352 · 4602
This award is funded by an endowment established through a bequest fromMartin Wachtel and is presented by Science Translational Medicine.
Nomination packages must include the following materials, written in English:
» A letter describing the nominee’s significant contributions to cancer research,
including supporting publications. The letter should explain how the candidate’s
research promises to make a lasting impact on the cancer field. (Applicants may
nominate themselves.)
» A letter of support from another scientist who knows the applicant’s work.
» The nominee’s curriculum vitae.
» A Perspective (1500 words maximum, 1 figure) written by the nominee describing
their research project and explaining how it advances our understanding of cancer.
The research described in the Perspective must be in the field of cancer and the
nominee must have performed or directed the work within the last 10 years.
Call for
Nominations
Due by Friday
February 1st, 2013
This new annual award will honor an early-career cancer researcher who has performed outstanding work
in the field of cancer research. Nominees must have received their PhD or MD within the last 10 years. The
award winner will be invited to deliver a public lecture on his or her research and will receive a cash award
of $25,000. The award winner’s Perspective will be published in Science Translational Medicine.
is pleased
to announce the
Submit nominations to wachtelprize@aaas.org and put “Wachtel Award Nomination” in the subject line.
For more information see: http://www.aaas.org/aboutaaas/awards/wachtel/
AAAS MARTIN AND ROSE WACHTEL
CANCER RESEARCH AWARD
YEARS OF
APPLIED SCIENCE
& TECHNOLOGY
20
ADVANCING
February 11-15 • Moscone North Convention Center • San Francisco, CA
Molecular Med
TRI-CON2013
Premier Sponsors:
TriConference.com
February 11-12
Concurrent Symposia
Emerging Molecular Diagnostics
Partnering Forum
February 12
Afternoon & Dinner Short Courses
February 13-15
Conference Programs
Diagnostics Channel
Therapeutics Channel
Clinical Channel
Informatics Channel
Cancer Channel
Organized by
Cambridge Healthtech Institute
Register by January 11th & SAVE!
Make Sure to Mention Keycode L35 when Registering!
Newly offered instrumentation, apparatus, and laboratory materials of interest to researchers in all disciplines in academic, industrial, and governmental organizations are
featured in this space. Emphasis is given to purpose, chief characteristics, and availability of products and materials. Endorsement by Science or AAAS of any products or
materials mentioned is not implied. Additional information may be obtained from the manufacturer or supplier.
99
Produced by the Science/AAAS Custom Publishing Ofhce New Products: Protein Analysis
Life Science Technologies
LABEL-FREE PROTEIN ANALYSIS
BLItz is a revolutionary new platform that will enable scientistsófor the first timeóto
conduct label-free analyses of proteins using small sample volumes easily and cost-efficiently
at their lab bench. With BLItz, researchers will be able to perform specific protein quanti-
tation, kinetics, and affinity analyses of protein interactions with just a drop of sample in a
simple ëdrop, read, doneí format. BLItz is smallótaking up less surface area than a tablet
personal computeróso it can fit easily in an individual scientistís workspace and is simple to
use. It requires only 4 µL of sampleó15 to 20 times less volume than needed for traditional,
SPR-based label-free assays or microplate-based ELISA assays. Additionally, its Dip and Read
assays utilize disposable, ready-to-use biosensors to uniquely provide results in seconds to
minutes. BLItz utilizes BioLayer Interferometry (BLI) technology, enabling real-time kinet-
ics, affinity, and quantitation measurements in high throughput applications.
FORTEBIO
For info: 855-254-8963 www.blitzmenow.com
Electronically submit your new product description or product literature information! Go to www.sciencemag.org/products/newproducts.dtl for more information.
www.sciencemag.org/products SCIENCE VOL 339 4 JANUARY 2013
PROTEIN DETECTION TEST
ProReveal Test is a rapid in situ fluorescence-based method of detecting
proteins on surgical instruments. The patented, ultrasensitive ProRev-
eal Test combines an integrated ProReveal Protein Test Kit and a Pro-
Reveal Viewer with easy-to-use touchscreen and software. In just two
simple spray and see steps, users can detect nanogram levels of protein
contamination, no matter where it is located on their entire surgical
instrument, in less than five minutes. The ProReveal Test aligns with the
protein contamination testing recommendations in the United King-
domís new NHS CFPP-01-01 consultation document. The ProReveal
Test uses very sensitive, yet cost-effective, fluorescence technology that
will tell you exactly where and how much protein is on your surgical
instrumentóa significant quality enhancement from the current test-
ing available.
SYNOPTICS HEALTH
For info: +44-(0)-1223-727114 www.synopticshealth.com
IMMUNOAFFINITY SAMPLE PREPARATION
Mass Spectrometric Immunoassay (MSIA) pipette tip technology deliv-
ers a simplified way to enrich target proteins for downstream mass spec-
trometric analysis. The new MSIA pipette tips provide scientists with an
easier immunoaffinity approach that yields better and more reproduc-
ible qualified results. MSIA tips feature a 300 µL pipette tip embedded
with a proprietary, highly porous immunoaffinity column that provides
increased sensitivity with improved reproducibility and sample through-
put. Benefits include: enhanced antibody/antigen-binding kinetics that
allow enrichment down to very low levels (femtomoles) for downstream
mass spectrometric protein/peptide quantification; improved signal-to-
noise ratios due to increased specificity and target recovery as compared
to conventional bead-based methods; fast and flexible sample throughput
with manual multichannel or fully automated 96-well processing in less
than 30 minutes; and customizable assays that meet different biomarker
research needs.
THERMO FISHER SCIENTIFIC
For info: 800-995-2787 www.thermoscientific.com/msi
PROTEIN EXPRESSION REPORTER KITS
A new suite of X-MAN Reporter Kits has been designed for investiga-
tion of protein expression levels, protein localization within a cell, and
protein-protein interaction within a molecular pathway, at natural levels
of protein and promoter activity. The kits include genetically defined re-
porter cell lines generated using Horizonís GENESIS gene editing tech-
nology with Promegaís luminescent and fluorescent reporter technolo-
gies. X-MAN NanoLuc Reporter cell lines incorporate Promegaís new
NanoLuc Luciferase reporter directly into the native loci of genes of
interest, permitting evaluation of genes and proteins at physiologically
relevant levels. X-MAN HaloTag Reporter cell lines facilitate the intro-
duction of Promegaís HaloTag as an N- or C-terminal fusion to genes
encoding proteins of interest at their native genetic loci. X-MAN Glo
Reporter Kits include X-MAN cell lines pretransfected with Promegaís
Firefly Luciferase for in vivo imaging and VivoGlo luciferin, provided as
isogenic cell lines for experimental control.
HORIZON DISCOVERY
For info: +44-(0)-1223-655580 www.horizondiscovery.com
MONOCLONAL ANTIBODIES
Fourteen new human monoclonal antibodies from Trillium Diagnostics
have been added to Beckman Coulterís flow cytometry portfolio. The
analyte specific reagents (ASRs) provide high specificity, convenience,
and flexibility in flow cytometry. A new cell permeabilization reagent,
Intra-Cell, is also being offered. The new monoclonals include hu-
man anti-hemoglobin F, which recognizes fetal hemoglobin molecules;
CD163 markers for anti-inflammatory signal and disease research; and
CD64 markers, which react with Fc gamma 1 receptors. Characterized,
quality controlled, convenient, liquid reagents allow users to spend less
time in sample preparation and more time on experiments. Rounding
out the line is a monoclonal for myeloid nuclear differentiation antigen
(MNDA). All antibodies are available unconjugated or conjugated to a
selection of fluorochromes. The new antibodies are supplied in 1.0 mL of
buffer, are ready to use, and facilitate flexibility in experimental design.
BECKMAN COULTER
For info: 800-742-2345 www.beckmancoulter.com
AS WE SEE IT,
OUR JOB IS
TO SIMPLIFY YOURS.
www.tecan.com
The Americas: +1 919 361 5200 Europe: +49 79 5194 170 Asia: +81 44 556 7311
info@tecan.com
© 2013 Tecan Trading AG, Switzerland, all rights reserved. For disclaimer and trademarks please visit www.tecan.com
Tecan. No matter what your everyday challenges or the goals for your lab … we understand.
Visit us at SLAS—booth 701—to learn more.
TouchTools

3.0
Enjoy simple workflow
programming and an easier
user experience through a next
generation touchscreen interface.
Tecan Academy
Get users up to speed faster and
ensure consistent instrument
operation with on-demand
online training and certification.
Infinite
®
200 PRO Multimode Reader
With Integrated Incubator
Enhance throughput in cell-based assays with an
integrated systemthat controls sample conditions
with a built-in Gas Control Module (GCM

).
Optimizing efficiency in your lab is about to get a little easier with the latest innovations from Tecan. From cutting-edge
instrumentation to a better way to train, we know what it takes to help you streamline workflows and boost productivity.
PRODUCTIVITY—WE UNDERSTAND.
Cancer Biology
www.promega.com/pathwaybiology
To get a FREE sample of any one of these reagents, visit:
The complexity of cancer systems biology requires
innovave tools for interrogang the signaling pathways
responsible for oncological transformaon.
Promegaís integrated tools for reporter
gene analysis assure biologically
relevant results in cancer research.
FuGENEÆ HD
The next generaon
transfecon reagent,
effecve on almost every
cell type with virtually
no cell toxicity
ONE-Gloô+Tox
Mulplexed reporter
gene analysis with off-
target toxicity detecon
in the same well
New! NanoLucô
and pGL4 Tox Vectors
Introducing NanoLuc - the
brightest, smallest, luciferase
available - plus a new line
of pGL4 response element
vectors for mapping
oncological pathways
Assistant Professor, Harvard
Medical School, based at Joslin
Diabetes Center Boston. Area:
islet biology, with preference to
in vivo and/or in vitro beta cell
replenishment.
Requirements: PhD or MD, out-
standing publication record; inde-
pendent funding desirable.
Please send CV including 3 refer-
ences to:
S. Bonner-Weir
Section of Islet Cell and
Regenerative Biology
Joslin Diabetes Center
1 Joslin Place
Boston, MA02215
Minority and female candidates
are encouraged to apply.
Academic Leader and Senior Investigator, Sanford Orthopedics & Sports Medicine
Sanford Orthopedics & Sports Medicine and Sanford Research are recruiting a Senior Investiga-
tor and Academic Leader for our large, vibrant and accomplished orthopedics and sports medicine
group. It is anticipated that a request will be made to appoint the successful candidate to the rank
of Professor in the Department of Surgery of the Sanford School of Medicine of The University of
South Dakota, Sioux Falls, SD.
We are looking for an experienced academic orthopedic surgical leader with a track record of research
accomplishment and administrative expertise in clinical and/or basic research. The successful can-
didate will help develop a broad based research program throughout the enterprise and will be a key
leader in the anticipated development of an orthopedic surgery residency and fellowship program
in sports medicine. The candidate should hold the MD, MD/PhD or equivalent degree and have a
demonstrated record of relevant research accomplishments as evidenced by training, publications
and peer reviewed funding.
Significant institutional support will be provided through Sanford Health and Sanford Research. In
addition, a comprehensive compensation package will be tailored to the individual’s qualifications.
Sanford Health includes 35 hospitals, 140 clinic locations and nearly 1,200 physicians in 70 specialty
areas of medicine across North and South Dakota, Minnesota, Iowa and Nebraska. The system is
experiencing dynamic growth and development in conjunction with Denny Sanford’s nearly $700
million in gifts, the largest ever to a health care organization in America. These gifts are making
possible the implementation of the several initiatives which support our Centers of Excellence,
including Orthopedics and Sports Medicine.
Candidates are invited to submit a cover letter, curriculum vitae, and a brief description of research
plans to:
H. Eugene Hoyme, MD, Chief Academic Officer, Sanford Health
President and Senior Scientist, Sanford Research/USD
Professor of Pedicatrics, Sanford School of Medicine of The University of South Dakota
The Sanford Center, 2301 E. 60
th
Street North, Sioux Falls, SD 57104-0589
gene.hoyme@sanfordhealth.org
Sanford Health is an Equal Opportunity/Affirmative Action Employer.
Download your free copy today at
ScienceCareers.org/booklets
Brought to you by the
AAAS/Science Business Office
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
Tracy Holmes
Worldwide Associate Director
Science Careers
Phone: +44 (0) 1223 326525
THE AMERICAS
E-mail: advertise@sciencecareers.org
Fax: 202-289-6742
Tina Burks
East Coast/West Coast/South America
Phone: 202-326-6577
Allyson Rosen
Midwest/Canada/Corporate
Phone: 202-326-6578
Marci Gallun
Sales Administrator
Phone: 202-326-6582
Online Job Posting Questions
Phone: 202-312-6375
EUROPE & REST OF WORLD
E-mail: ads@science-int.co.uk
Fax: +44 (0) 1223 326532
Ben Asfaha
Phone: +44 (0)1223 326529
Lucy Nelson
Phone: +44 (0)1223 326527
Kelly Grace
Phone: +44 (0) 1223 326528
JAPAN
Yuri Kobayashi
Phone: +81-50-3696-5100
E-mail: ykobayas@aaas.org
CHINA &TAIWAN
Ruolei Wu
Phone: +86-1367-1015-294
E-mail: rwu@aaas.org
For full advertising details, go to
ScienceCareers.org and click
For Employers, or call one of
our representatives.
All ads submitted for publication must comply
with applicable U.S. and non-U.S. laws. Science
reserves the right to refuse any advertisement
at its sole discretion for any reason, including
without limitation for offensive language or
inappropriate content, and all advertising is
subject to publisher approval. Science encour-
ages our readers to alert us to any ads that
they feel may be discriminatory or offensive.
Science Careers
Advertising
There’s only one
FACULTY POSITIONS
ABOUT KAUST
- EPIGENETICS
- ADVANCED BIOIMAGING AND PHOTONICS
The Biological and Environmental Sciences and Engineering (BESE) Division at King Abdullah University of Science and
Technology (KAUST) invites exceptional applicants to apply for Faculty positions in the research areas listed below.
- ENVIRONMENTAL GENOMICS
- MARINE BIOMOLECULAR ENGINEERING
- MICROBIOLOGY OF EXTREME MARINE ENVIRONMENTS
In collaboration with the Red Sea Research Center:
- ENVIRONMENTAL TOXICOLOGY
In collaboration with the Water Desalination and Reuse Research Center:
- HALOPHYTE PHYSIOLOGY AND DOMESTICATION
- PLANT EPIGENETICS AND SIGNALING
In collaboration with the Center for Desert Agriculture:
BIOLOGICAL AND ENVIRONMENTAL SCIENCES AND ENGINEERING DIVISION
KAUST is an international, graduate research university dedicated to advancing science and technology through
interdisciplinary research, education, and innovation. Located on the shores of the Red Sea in Saudi Arabia, KAUST
offers superb research facilities, generous assured research funding, and internationally competitive salaries,
attracting top international faculty, scientists, engineers, and students to conduct fundamental and goal-oriented
research to address the world’s pressing scientific and technological challenges related to the sustainability of
water, food, energy, and the environment.
All candidates should have the ability to pursue a high impact research program and have a commitment to teaching
at the graduate level. For more information and to apply for the positions listed above, please visit
http://bese.kaust.edu.sa and click on the employment tab. Applications received by February 1, 2013 will receive
full consideration and positions will remain open until filled.
Assistant Professor: Basic biological processes at the
interface between human health and the environment
The Institute of Marine and Environmental Technology (IMET) of the University
System of Maryland seeks applications for the position of Assistant Professor
(tenure-track). The successful candidate will have a track-record of outstanding
research on basic biological processes at the interface between human health
and the environment. Potential focus areas include but are not limited to
epigenetic effects of toxicants, ecotoxicology, glycobiology of biofilms
and host-pathogen interactions. Candidates with an interest in practical and
biomedical applications of their fundamental research, including interactions
with industry, are encouraged to apply. A demonstrated ability to establish a
well-funded research program is expected for the successful candidate, with
preference for candidates with active K99, RO1 or equivalent funded research
programs. Primary appointment is at the University of Maryland Baltimore,
with a secondary appointment at the University of Maryland Center for
Environmental Science. The primary focus of the appointee will be on research;
teaching at the graduate level is encouraged.
IMET’s mission is to study the biology of coastal marine biosystems and ensure
their sustainable use, as well as exploit marine-derived systems to improve
human health (www.imet.usmd.edu). IMET brings together faculty members
from three major USM research institutions - the University of Maryland
Baltimore (UMB), the University of Maryland Baltimore County (UMBC),
and the University of Maryland Center for Environmental Science (UMCES)
- in a dedicated state-of-the-art research facility located in Baltimore's Inner
Harbor.
A Ph.D. or equivalent terminal degree is required in molecular biology,
microbiology, biochemistry, bioengineering or related fields and candidates
should have postdoctoral experience and a strong publication record. Applicants
should send the following in a single PDF file to Russell Hill, Director, IMET;
email to hill@umces.edu: (1) detailed curriculum vitae, (2) statement of
research interests and goals and (3) names and contact details of three to five
references. To receive full consideration, applications should be submitted by
March 1, 2013.
The USM is an Equal Opportunity, Affirmative Action Employer.
Director Cardiovascular Research Center
The Cardiovascular Research Center (CVRC) at Duke University
Medical Center invites applications from outstanding basic or early-
translational cardiovascular researchers for its Director.
The Duke CVRC, established in 2010, promotes and stimulates basic
discoveries that will advance understanding of the molecular pathogenesis
of cardiovascular diseases. The Center provides a framework of support
for our membership of more than 70 faculty members spanning six
academic departments at Duke University and Duke-NUS in Singapore,
including core resources, pilot initiative funds, training programs, and a
seminar series.
The candidate should have an internationally recognized track
record in cutting edge laboratory research in basic/early translational
cardiovascular science. The successful applicant will be provided with
substantial and sustainable resources to direct the recruitment of new
scientists and physician-scientists into the Division of Cardiology/
Department of Medicine; to partner with other Departments (clinical and
basic science) to recruit investigators that will add to the CVRC mission;
and to support research and collaborative efforts within the CVRC, such
as pilot projects and cores.
High priority will be placed on creativity and an outstanding track record
in laboratory research. Demonstration of mentoring and administrative
experience will be highly valued. A generous start up package and
compensation commensurate with the position will be provided.
Interested individuals should submit a CV and statement of research
interest as PDF files directly to: DUKECVRC@dm.duke.edu. The
deadline for receipt of applications is March 1, 2013.
Duke University Medical Center is an
Equal Opportunity/Affirmative Action Employer.
The University of Texas Health Science Center at Tyler invites applications
fromextramurally funded, outstanding scientist/administrators to serve as
Chair of Cellular and Molecular Biology. Administrative experience
required. Administrative responsibilities include oversight/development
of a productive faculty group and of a biotechnology master’s graduate
program. Track record of successful research and current extramural
funding required. Resources to support this recruitment include ample
laboratory space, additional faculty positions to support growth of research
programs, an endowment and unique UT System resources.
Tyler is located midway between Dallas and Shreveport amidst the
picturesque lakes and piney woods of East Texas. The mission of the
basic and clinical research at the UT Health Science Center focuses
on lung injury/repair, pulmonary infectious diseases, coagulation,
immunology and oncology. A new $40 million academic building
housing the Oncology Program opened in November 2011. A complete
listing and description of current faculty research interests can be found
online via UTHSCT’s website http://www.uthct.edu/Research.
The successful candidate will bring or establish a dynamic independent
research program in a discipline related to the mission of UTHSCT.
Preference given to programs in either lung injury and repair,
aspects of matrix biology or oncology. Associate and Full Professor
level candidates will be considered. Applicants should submit their
curriculum vitae, a statement of future administrative and research plans
and the names of three references to: Dr. Anna Kurdowska, Faculty
Search Committee Chair, University of Texas Health Science Center
at Tyler, 11937 US Highway 271, Tyler, Texas 75708-3154, or by email
to anna.kurdowska @uthct.edu.
is an EEO/AA Employer M/F/V/D. This position is
security sensitive and subject to Texas Education Code 51.215 which
authorizes the employer to obtain criminal history information.
MOLECULAR BIOLOGIST/BIOCHEMIST
2-year Visiting Faculty Position
Biology Department
The Biology Department at Williams College, a premier liberal arts college
with a long-standing tradition of excellence in the sciences, invites applica-
tions for a 2-year visiting assistant professor to begin July 2013. We are
seeking a broadly trained MOLECULARBIOLOGIST/BIOCHEMIST
who will teach immunology and other upper-level courses in his/her area
of specialty. Normally, faculty members teach one course and two labora-
tory sections (or the equivalent) each semester.
Aresearch programthat involves talented undergraduates is expected, and
a limited amount of internal funds for research are available. APh.D. and
postdoctoral experience are required. All offers of employment are contin-
gent upon completion of a background check. Applicants should submit
curriculum vitae, brief statements of teaching and research interests, and
arrange for three letters of recommendation to be sent by January 31,
2013 to: biochemsearch@williams.edu, care of Steven Swoap, Chair,
Department of Biology, Williams College.
Williams College is a coeducational liberal arts institution located in the
Berkshire Hills of western Massachusetts with easy access to the culturally
rich cities of Albany, Boston, and New York City. The College is com-
mitted to building and supporting a diverse population of approximately
2,000 students, and to fostering an inclusive faculty, staff and curriculum.
Williams has built its reputation on outstanding teaching and scholarship
and on the academic excellence of its students. Please visit the Williams
College website (http://www.williams.edu). Beyond meeting fully its
legal obligations for non-discrimination,
Williams College is committed to building a diverse and inclusive
community where members from all backgrounds can live, learn,
and thrive.
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
The Person: The ideal candidate will have considerable research experience demonstrating a strong understanding of computational, technological, and biological
issues. In addition, candidates should possess recognized research management and leadership abilities. Candidates with substantial training and experience in
computation/informatics, biomedical technology, and biomedical research will be considered. A strong understanding of the role of computation, informatics,
and technology in uncovering biological principles and in advancing research on human health and disease is desired. This individual will report to the Director,
National Institute of General Medical Sciences (NIGMS), but will also have access to the Director, National Institutes of Health (NIH), in coordinating activities
across NIH and among federal agencies.
The Challenge: A significant challenge for the biomedical research community is the integration of the vast amount of accumulating scientific data in order to
develop predictive understanding of basic biological processes. The ability to meet this challenge will be critically dependent on advances in bioinformatics and
computational biology and on discovery and deployment of newtechnologies. The Division of Biomedical Technology, Bioinformatics, and Computational Biology
(BBCB) is responsible for stimulating and funding research in these areas of importance for NIGMS. The Division supports research on bioinformatics, databases,
and data mining; modeling of complex biological systems; algorithmic development and software engineering; mathematical biology, high-performance computing,
molecular imaging, structural biology, and proteomics, among other areas. In addition, the Division is responsible for managing the NIH Biomedical Information
Science and Technology Initiative (BISTI), an agency-wide effort to stimulate and coordinate use of computer science and technology to address problems in
biology and medicine. The Division also actively collaborates with other NIH components and federal agencies in developing policies in these areas. NIGMS is
seeking a leader in this field to direct the Division and the BISTI efforts, and to coordinate the work of both with other interested federal agencies and the broader
scientific community. Information about the Division and BISTI is available at: http://www.nigms.nih.gov/About/Overview/bbcb.htm and http://bisti.nih.gov.
Position Requirements: Candidates must have an M.D., Ph.D., or equivalent degree in a field relevant to the position. Please see the official vacancy announce-
ment for qualification requirements and what to submit. The position will be filled under a Title 42 excepted service appointment.
Salary/Benefits: Salary is competitive and will be commensurate with the experience of the candidate. A recruitment or relocation bonus may be available,
and relocation expenses may be paid. A full package of Federal Civil Service benefits is available, including: retirement, health and life insurance, long term
care insurance, leave, and a Thrift Savings Plan (401K equivalent). The successful candidate is subject to a background investigation and financial disclosure
requirements.
How to Apply: The official vacancy announcement is available at: http://www.nigms.nih.gov/About/Job_Vacancies/
Information about the NIGMS can be found at http://www.nigms.nih.gov/About/.
NIGMS will begin accepting applications on October 26, 2012, and plans to have the position open for at least 30 days, but will not to close the application
process until a candidate has been selected.
You may contact Krystal Kelly with questions about this vacancy at kellykry@od.nih.gov or 301-594-3827.
HHS and NIH are Equal Opportunity Employers
Director, Division of Biomedical Technology, Bioinformatics, and Computational Biology
National Institute of General Medical Sciences
National Institutes of Health
Department of Health and Human Services
Department of Biomedical Sciences
Chair and Professor
Applications are invited for the position of inaugural Chair and Professor in the Department of Biomedical Sciences at the Quillen College of
Medicine, East Tennessee State University. In early 2012 the Departments of Anatomy and Cell Biology, Biochemistry and Molecular Biology,
Microbiology, Pharmacology and Physiology were merged to form the Department of Biomedical Sciences. The department actively participates
in a thriving graduate program awarding Ph.D. degrees in Biomedical Sciences. Areas of research focus within the Department and the College
of Medicine include cardiovascular physiology, infectious disease, immunology, inflammation, neuroscience, cancer and aging. Candidates with
active research programs in these areas are particularly encouraged to apply.
We seek a dynamic and energetic individual to provide leadership and direction for the newly created department. The Chair will be expected to
implement a vision for the future of the department. The successful candidate must have an earned doctorate in an appropriate biomedical discipline
and qualifications commensurate with appointment at the rank of Professor. Astrong record of scholarly activity in a medical school environment
and continuous extramural funding for research at the national level are required. In addition, the Chair must possess the skills to facilitate and
promote fundamental and translational research within the department. Administrative responsibilities will include oversight of faculty recruit-
ment as well as development, mentoring, evaluation, promotion and retention of faculty. The Chair will also be responsible for finance and budget,
strategic planning and resource allocation within the department. The successful candidate must have a passion for educational excellence at the
graduate and medical student levels. In addition, the Chair must be familiar with LCME requirements and program direction as they relate to the
first two years of medical student education.
Review of applications will begin March 1, 2013 and will continue until the position is filled. Applications should include a cover letter, a com-
plete and current CV, names and addresses of three references as well as brief statements of educational, research and leadership philosophy.
Candidates for this position must apply online at https://jobs.etsu.edu. Inquiries about this position may be directed to Dr. David Williams
(williamd@etsu.edu, telephone 423-439-6363).
The Quillen College of Medicine is a community-based medical school located in the beautiful mountains of northeast Tennessee. Quillen College
of Medicine serves southern Appalachia and the healthcare needs of over one million people. Our campus is located in the Tri-Cities area on the
grounds of the James H. Quillen VA Medical Center. The Tri-Cities area is a highly rated place to live in North America. The low cost of living,
access to health care and education along with the beautiful mountainous region, parks, lakes and national forests make it an outstanding place to
live and work.
ETSU is an Equal Opportunity/Affirmative Action Employer that employs only U.S. citizens and resident aliens authorized to work in the U.S.
Women and minorities are encouraged to apply. ETSU is a smoke and drug free environment.
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
ALBERT M. MATTOCKS PROFESSORSHIP
DEPARTMENT OF PHARMACEUTICAL SCIENCES
The University of Michigan College of Pharmacy invites applications
for an Endowed Faculty Position, with tenure, at the rank of Professor
in the Department of Pharmaceutical Sciences. Applicants must have an
outstanding, well-funded, research program and an ability to develop
programmatic collaborative research and funding across the pharmaceuti-
cal, biomedical, chemical and engineering sciences. Preference will be
given to those scientists with a significant research interest in one or
moreof the following areas:
• Solid-state chemistry, formulation and drug delivery
• Formulation and delivery of biomacromolecules
• Interfacial and colloid chemistry in drug delivery
• Polymer chemistry and biomaterials
• Drug delivery for cellular therapies
Applicants must demonstrate a strong record of leadership in research
and professional activities, and a strong commitment to the educational
objectives of the graduate and professional programs at the College.
Applicants should submit a pdf file of their curriculum vitae, future
research plans, and three representative publications to Patrina
Hardy (thardy@umich.edu). For more information on the posi-
tion, you may contact the Search Committee Chair (David Smith:
smithb@umich.edu) or the Department Chair (Steven Schwendeman:
schwende@umich.edu).
The University of Michigan is an Equal Opportunity/Affirmative
Action Employer, which is committed to recruiting, supporting and
fostering a diverse community.
Mason seeks a Vice President for
Research and Economic Development,
for more information, visit
jobs.gmu.edu
A message from
the president,
scan here....
Nominations, including self-nominations, will be accepted for
this position. The nomination reviewwill begin on February 28,
2013 and will remain open until the position is flled. Nomina-
tion letters must refer to the proposed candidate’s specifc qual-
ifcations based on the duties outlined above, and should be
sent via email to Dr. Jim Olds at jolds@gmu.edu. For full con-
sideration, applicants must complete and submit the online ap-
plication for position FA784z at http://jobs.gmu.edu/; upload
a cover letter, resume, and list of three professional references
with contact information.
George Mason University is an affrmative action/equal opportunity
employer. Women and minority faculty members are encouraged to apply.
The Center of Excellence in Infectious Disease Research at the Paul L.
Foster School of Medicine, Texas Tech University Health Sciences
Center, El Paso, Texas is seeking highly qualified candidates for tenure
track faculty positions at the Assistant or Associate Professor level. The
Center is primarily interested in investigators with research interests in
infectious diseases and immunology. The position reports to the
Co-Directors of the Center of Excellence for Infectious Diseases.
Minimum Qualifications:
 M.D. or Ph.D. degree in a field related to infectious diseases or
immunology,
 3 years of postdoctoral experience, strong publication record
Experience in emerging infectious diseases and in the use of latest
technologies is highly preferred. Investigators with funded grant
support are particularly encouraged to apply and will be considered for
the rank of Associate Professor. The selected candidate should have an
interest in assisting with the development of a new Graduate School of
Biomedical Sciences, a willingness to start something new and to help a
new institution to grow and develop. Good communication skills and
demonstrated ability to work in a collegial environment are essential.
A competitive salary, start-up package and comprehensive benefits are
available. Interested candidates must apply online at
https://jobs.texastech.edu/postings/50140 Requisition # 87477
For further information, potential applicants may send inquiries to:
Manjunath Swamy, M.D. manjunath.swamy@ttuhsc.edu or Premlata
Shankar, M.D. premlata.shankar@ttuhsc.edu, Co-directors of the
Center of Excellence for Infectious Diseases.
Texas Tech University Health Sciences Center is an Equal Opportunity/
Affirmative Action Employer
*AAAS follows CAN-SPAM and European Safe Harbor guidelines for protecting your privacy.
We will never sell your e-mail address and you can opt-out of receiving e-mails at any time.
STAY INFORMED!
STAYCONNECTED!
Get more from your
AAAS membership
Are you currently registered to receive e-mails fromAAAS
and Science? E-mail is the primary way that AAAS communicates
with our members about AAAS programs, newmember benefits,
invitations to special events, and, of course, the latest news and
research being published in Science.
Sign up today to ensure that you are getting the most out of your
membership and Science subscription.
*
To get started visit:
promo.aaas.org/stayconnected You’ll need your
AAAS Member number. Find it above your name
on your Science mailing label.
Don’t miss a thing. Sign up for e-mail
communications fromAAAS today!
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
MMV 2013 11th Call for
Proposals:
MMV welcomes projects in the hit-to-
lead and lead optimization stages for
new families of molecules specifcally
addressing the key priorities of the
malaria eradication agenda: transmission
blocking via the human host, and
prevention of P. vivax relapse through
killing of liver stage hypnozoites or
reactivating them so as to be killed in
the blood stages.
In addition, proposals are sought for
novel chemical series with a long
half-life (ideally >10h in rodents) and
confrmed in vivo effcacy that could
have potential for well tolerated P.
falciparumchemoprophylaxis or asexual
blood stage treatment in humans. Any
proposals based on existing chemotypes
must clearly address known issues.
Please note that early target validation
falls outside of our mandate.
Finally, we also welcome proposals on
novel translational study concepts that
enable early confirmation of relapse
prevention, transmission blocking and
prophylactic activity in humans.
Templates for the 3-page Letter of
Interest to apply for the above Call can
be found at www.mmv.org.
Deadline for receipt in MMV offce:
12 noon CET March 15th, 2013
Send your 3 page letter of interest,
electronically to:
proposals@mmv.org
Neurobiologist / Neuropharmacologist.
Senior research scientist to establish novel
drug discovery concepts in neuro-
degenerative diseases
Roche, Switzerland
“Make your mark.
Improve lives.”
Who we are
At Roche, 80,000 people across 150 countries are pushing back the
frontiers of healthcare. Working together, we’ve become one of the
world’s leading research-focused healthcare groups. Our success is
built on innovation, curiosity and diversity.
The headquarters in Basel is one of Roche’s largest sites, over 8,000
people from approximately 80 countries work at Roche Basel. Favored
by its geographic location in the heart of Europe, the Basel area is
one of the most dynamic economic regions in Switzerland — a great
place to live and work.
The Position
We are looking for a highly motivated senior research scientist to
join our laboratories in preclinical neuroscience to expand and
strengthen our drug discovery capabilities, particularly related to
pathogenic mechanisms in neurodegenerative diseases such as
Alzheimer’s and Parkinson’s Disease.
• As a senior research scientists and integrated member of project
teams you are providing expert scientific input to drug discovery
programs.
• You are expected to conduct laboratory research in the evaluation
and profiling of drug candidates including in vitro and in vivo
pharmacology and mechanistic studies.
• As a laboratory head you will be guiding a small research team.
Who you are
You’re someone who wants to influence your own development.
You’re looking for a company where you have the opportunity to
pursue your interests across functions and geographies. Where a job
title is not considered the final definition of who you are, but the
starting point.
• You have a PhD or equivalent University degree in neurobiology,
neuropharmacology and/or cell biology with profound knowledge
of molecular mechanisms in neurodegenerative diseases.
• A proven record of 3–5 year research experience in drug disco-
very processes and preclinical neuropharmacology illustrates your
scientific achievements.
• You are familiar with cellular disease models (eg. iPSCs) and
state-of-the-art animal models of neurodegeneration to study
disease pathways and functional deficits.
• A scientific background and expertise in neuro-immunology and/
or protein misfolding & clearance is an advantage.
• You are used to work in a multi-disciplinary environment and you
have excellent communication skills in English.
Job ID No.: 00406156
Contact Line: H.R. Loetscher, Phone: +41 61 688 3259
The next step is yours. To apply online for this position visit
www.careers.roche.ch
Roche is an Equal Opportunity Employer
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
AAAS is here – helping scientists achieve career success.
Every month, over 400,000 students and scientists visit ScienceCareers.org in search of the information, advice, and
opportunities they need to take the next step in their careers.
A complete career resource, free to the public, Science Careers offers a suite of tools and services developed specifically
for scientists. With hundreds of career development articles, webinars and downloadable booklets filled with practical
advice, a community forum providing answers to career questions, and thousands of job listings in academia, govern-
ment, and industry, Science Careers has helped countless individuals prepare themselves for successful careers.
As a AAAS member, your dues help AAAS make this service freely available to the scientific community. If you’re not
a member, join us. Together we can make a difference.
To learn more, visit aaas.org/plusyou/sciencecareers
Established in 1952, Beihang University is one of top research-oriented universities in China, focusing on fundamental cutting edge
research and high-level education. One of the frst universities funded by China’s “211” and “985” programs, it has seven national key
laboratories and twenty-fve provincial and ministerial key laboratories. Today, Beihang University is known for outstanding research
and education in aviation and aerospace sciences, power and energy engineering, materials science, mechanics, information science,
transportation science, instrumentation science, management science and law.
Beihang University is on a clear path to become a world-class university in many engineering and science disciplines. As part of
Beihang’s further pursuit for excellence in research and education, we have expanded our global search for the best research talent to
join our International Research Institute for Multidisciplinary Science (IRIMS). Five independent international research centers (IRC)
were established recently under the name of IRIMS. As the core part of IRIMS, IRCs are devoted to establish a world-class, advanced
and multidisciplinary research platform.
Beihang University invites applications for full-time Professors, Associate Professors and excellent scientists. Preference will be given
to candidates whose research emphasis demonstrates the potential to complement and advance the IRIMS existing research strengths.
Successful candidates will be provided competitive salaries and start-up funds.
Positions Available
Position offered by the Recruitment Program of Global Experts (1000 Plan Professorship) •
Position offered by the Chang Jiang Scholars Program •
Position offered by the Recruitment Program of Global Young Experts (1000 Plan Professorship for Young Talents) •
Position offered by Beihang University’s Zhuoyue Program of Professors •
Position offered by Beihang University’s Zhuoyue Program of Associate Professors. •
Interested individuals should send curriculumvitae by email to rscrcb@buaa.edu.cn, with “FacultyApplication” in the title. For more
information, please visit the university’s Human Resource Department website http://rsc.buaa.edu.cn/, or contact us by email rscrcb@
buaa.edu.cn or by telephone 86-010-82317779.
Faculty Positions Available in
Beihang University, China
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
POSITIONS OPEN
FACULTY POSITIONS IN VIROLOGY
The Department of Molecular Biology and Mi-
crobiology at Tufts University School of Medicine.
Boston, Massachusetts (website: http://medicine.
tufts.edu/Education/Academic-Departments/
Basic-Science-Departments/Molecular-Biology-and-
Microbiology) is undergoing a major expansion and
will be adding tenure-track faculty positions. Posi-
tions are available at the ASSISTANT, ASSOCIATE,
and FULL PROFESSOR levels. The successful can-
didates will build on the core strengths of the Depart-
ment and focus on the study of viruses, including, but
not limited to, structural virology, virus-cell interac-
tion, viral pathogenesis, host-virus evolution, and global
health and emerging viruses. We are seeking candidates
who use innovative approaches to investigate problems
at molecular, cellular, and organismal levels. We offer
generous startup packages, newly renovated laboratory
space, and a highly collaborative environment with
opportunities for both basic and translational research.
A joint appointment at Tufts Medical Center is pos-
sible for individuals with clinical training.
Applicants should hold a Ph.D. and/or M.D. de-
gree, and have an excellent track record of research
productivity. Successful candidates will be expected
to establish and direct a thriving, high-quality indepen-
dent research program and to contribute to graduate
and medical education. Please submit electronic ap-
plications including curriculum vitae, a statement of
research interests and the names and contact infor-
mation for at least three references to website: https://
academicjobsonline.org/ajo/jobs/2448. Review
of applications will begin January 1 and continue on
a rolling basis thru March 31, 2013. Inquiries, but
not application materials, may be directed to e-mail:
mbmfacultyrecruitment@tufts.edu.
Tufts University is an Equal Opportunity/Affirmative Action
Employer. We are committed to increasing the diversity of our
faculty. Applications from women and members of underserved
minority groups are encouraged.
PHARMACEUTICAL SCIENCES
FACULTY POSITION
The University of Tennessee College of Pharmacy
The Department of Pharmaceutical Sciences in the
College of Pharmacy at the University of Tennessee
Health Science Center in Memphis, Tennessee, is seek-
ing applications for a 12-month full-time, tenure-track
faculty position at the FULL, ASSOCIATE, or AS-
SISTANT PROFESSOR level that is state supported.
The successful candidate is expected to devote greater
than a 60% effort to research. The applicant should
have core expertise in medicinal chemistry, pharma-
cology, pharmaceutics, chemical biology, biomedical
engineering, structural biology, or a related discipline
focused on drug discovery and development. Candi-
dates with a well-funded program in cancer, diabetes,
antibiotics, drug delivery, nano-medicine, nanotechnol-
ogy, bio-imaging, or lipid research are highly encour-
aged to apply. The successful candidate is expected to
have a Ph.D. or equivalent degree, the ability to ac-
quire sustained external, investigator-initiated funding,
including National Institutes of Health principal inves-
tigator funding, a commitment to excellence in teach-
ing, and excellent oral and written communication skills.
Applications will be processed until the position is filled.
Please submit curriculum vitae, summary of research
interests, contact information, and three letters of ref-
erence to: John Buolamwini, Ph.D. Professor and
Vice Chair, Chair of Faculty Search Committee, De-
partment of Pharmaceutical Sciences, 847 Monroe
Ave, Suite 327, Memphis, TN 38163. The Univer-
sity of Tennessee Health Science Center is located in
Memphis, TN, an economically vibrant center, with
a metropolitan population of more than 1.3 million,
reflecting the richness along the bluffs of the mighty
Mississippi River. The College of Pharmacy is located
in a new, 187,000 square-foot building on the Health
Science Center complex.
The University of Tennessee Health Science Center is an
Equal Opportunity/Affirmative Action Employer.
POSITIONS OPEN
FACULTY OPPORTUNITY
Assistant/Associate Professor of Microbiology
The Department of Microbiology invites applica-
tions for faculty positions in the areas of virus-host
interactions, immunology, emerging infectious diseases,
and novel vaccine strategies. We are seeking individuals
who will develop well-funded, active, independent re-
search programs. Successful candidates will join a mul-
tidisciplinary department focusing on molecular and
genetic aspects of infectious diseases, host-pathogen
interactions, and immunological host responses to in-
fection. They will also participate in teaching at the grad-
uate and medical school levels.
For consideration, please send your curriculum vitae,
selected reprints, names, and addresses of three refer-
ences, and a statement of research interest to:
Chair, Search Committee
Department of Microbiology, Box 1124
Mount Sinai School of Medicine
One Gustave L. Levy Place
New York, NY 10029-6574
You may also e-mail: nina.umerah@mssm.edu. Please
visit the department on our website: http://www.
mssm.edu/departments-and-institutes/microbiology.
We are an Equal Opportunity Employer fostering diversity in the
workplace.
STAFF SCIENTIST II
The Department of Biochemistry & Molecular Bi-
ology at the Medical University of South Carolina is
seeking qualified candidates for the position of Staff
Scientist II and Nuclear Magnetic Resonance Manager.
The successful candidate will be responsible for the
management and maintenance of three magnets lo-
cated in the Drug Discovery Building and their ac-
cessory equipment such as cryogenically cooled probes.
He/she will also initiate and develop with colleagues a
strategy for nuclear magnetic resonance and crystal-
lography facilities. He/she will provide highly spe-
cialized technical and scientific assistance to faculty,
postdocs, and graduate students to further their research
and prepare structural biology data for publication and
grant applications. Computer software and hardware
maintenance, upgrades, diagnosis of faults and installa-
tion of structural calculation and analysis software(s) are
paramount responsibilities. Duties also include train-
ing of faculty, postdoctoral researchers, and students
on all aspects of structure determination, molecular
dynamics, and drug discovery. Establishing external
collaborations and serving as a point of contact for
equipment manufacturers to organize complex repairs
to systems in order to maintain optimal performance is
required.
A Ph.D. in Biochemistry or related sciences is required
for this position, as well as two years of post-doctorate
experience and evidence of scholarly activities. Individ-
uals must have expertise in the following areas: electronic
engineering, NMR focusing on small molecule and pro-
tein work and protein/RNA structure determination,
validation software, computer networking and Linux
based software such as scientific python programming.
Qualified candidates should apply online at website:
http://www.jobs.musc.edu.
For more information, please contact: Mrs. Karen
Fulghum, Business Manager, Department of Bio-
chemistry &Molecular Biology, MUSC 173 Ashley
Avenue, BSB504, Charleston, SC 29425. E-mail:
fulghumk@musc.edu.
We are an Equal Opportunity Employer, supporting workplace
diversity of minorities/females/veterans/persons with disabilities.
POSITIONS OPEN
BIOLOGY INSTRUCTION
Assistant Professor
The Department of Biological Sciences at Clemson
University invites applications for two tenure-track As-
sistant Professor faculty positions to begin in August
2013. The Department seeks faculty with strong com-
mitments to the teaching, undergraduate research, and
educational outreach missions of Clemson Universi-
ty. Applicants must have a Ph.D. in the biological
sciences. Evidence of excellence in teaching is essential.
Experience in the development of introductory biol-
ogy curricula and experience and interest in inquiry-
based approaches in biology instruction are desired.
The successful candidates will be expected to partici-
pate in the development and teaching of our introduc-
tory biology courses and the advising of undergraduate
students. In addition, successful candidates will be ex-
pected to develop a research program on undergradu-
ate instruction or undergraduate research training. The
Department of Biological Sciences has outstanding
outreach education programs that provide research
opportunities through the nationally funded SC Life
and DNA Learning Center programs. Further informa-
tion about departmental resources, programs, and fac-
ulty research interests are available at website: http://
www.clemson.edu/cafls/departments/biosci/.
In a single PDF file, compile a cover letter, curriculum
vitae, a statement of teaching interests and philoso-
phy, a statement of research interests in undergrad-
uate instruction or research training, and the names
of three references. In addition, provide PDF files of up
to three publications. Send all application documents
to: Robert J. Kosinski (e-mail: rjksn@clemson.edu),
Department of Biological Sciences, 132 Long Hall,
Clemson University, Clemson, SC 29634-0314. For
full consideration please submit your application no
later than February 15, 2013. Review of applications
will continue until the positions are filled.
Clemson University is an Affirmative Action/Equal Opportu-
nity Employer and does not discriminate against any individual or
group of individuals on the basis of age, color, disability, gender,
national origin, race, religion, sexual orientation, veteran status or
genetic information.
PROFESSOR in
Systems Engineering (Tenured)
The Department of Systems Engineering, Faculty of
Engineering, Shizuoka University, invites applications
for a tenured professor broadly in the field of either In-
formation Sciences or Systems Engineering. The Schools
of Engineering will be reformed on April 2013 and the
prospective professor will be affiliated to Department
Mathematical and Systems Engineering. We seek out-
standing investigators who will build active research
programs, teach effectively at the undergraduate and
graduate levels at an earliest day after July 1, 2013. Ph.D
is necessary. Deadline for application: March 4, 2013.
For further information, please see the position descrip-
tion at website: http://www.shizuoka.ac.jp/recruit/
2012/20121205_eng_en.pdf. Shizuoka University is an
Equal Opportunity Employer.
FACULTY POSITIONS - MEDICAL SCHOOL
The Saint James School of Medicine, an international
medical school (website: http://www.sjsm.org),
invites applications from candidates with teaching and/or
research experience in any of the basic medical sciences
for its Caribbean campuses. Faculty positions are cur-
rently available in Anatomy, Physiology, Pathology,
Microbiology, Biochemistry, Physical Diagnosis, Clin-
ical Correlation of Basic Sciences, and Pharmacology.
Applicants must be M.D., D.O., and/or Ph.D.
Teaching experience in the U.S. system is desirable
but not required. Retired persons are encouraged to
apply. Attractive salary and benefits. Submit curriculum
vitae to e-mail: mjansen@mail.sjsm.org or by mail to:
HRDS Inc., 1480 Renaissance Drive. Suite 300,
Park Ridge, IL 60068.
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
4 JANUARY 2013 VOL 339 SCIENCE www.sciencecareers.org 108
There’s only one
Galileo Galilei
Career advice I Job postings I Job Alerts I Career Forum I Crafting resumes/CVs I Preparing for interviews
For your career in science, there’s only one
Careers
ScienceCareers.org
orn in 1564, Galileo Galilei once contemplated a career in the priesthood. It’s perhaps fortunate
for science that upon the urging of his father, he instead decided to enroll at the University of
Pisa. His career in science began with medicine and from there he subsequently went on to become
a philosopher, physicist, mathematician, and astronomer, for which he is perhaps best known. His
astronomical observations and subsequent improvements to telescopes built his reputation as a
leading scientist of his time, but also led him to probe subject matter counter to prevailing dogma.
His expressed views on the Earth’s movement around the sun caused him to be declared suspect
of heresy, which for some time led to a ban on the reprinting of his works.
Galileo’s career changed science for all of us and he was without doubt a leading light in the
scientific revolution, which is perhaps why Albert Einstein called himthe father of modern science.
Want to challenge the status quo and make the Earth move? At Science we are here to help you
in your own scientific career with expert career advice, forums, job postings, and more —all for free.
For your career in science, there’s only one Science. Visit Science today at ScienceCareers.org.
B
POSITIONS OPEN
FACULTY POSITION in
Integrative Biology
The Department of Biological Sciences at Clemson
University invites applications for a tenure-track fac-
ulty position in integrative organismal biology at the
ASSISTANT PROFESSOR level, to begin August
2013. Postdoctoral experience is required. We are seek-
ing a broadly trained biologist whose research utilizes
multidisciplinary approaches to improve understand-
ing of organismal adaptation and function of inver-
tebrates. Specific areas of research are open but should
emphasize comparative, experimental, or evolutionary
approaches. The successful candidate will be expected
to interact with faculty having diverse interests rang-
ing from organismal biology, ecology, and evolution
to cell, developmental, and molecular biology, thereby
supporting University emphasis areas in Sustainable
Environment and Biomedicine and Biotechnology. The
successful candidate will also be expected to establish
innovative, externally funded research programs of na-
tional distinction, and to be an excellent teacher. Teach-
ing responsibilities include one upper level undergraduate
course per semester in our animal diversity and/or func-
tional biology core areas and graduate course(s) in one_s
specialty. Applications should include a cover letter,
a curriculum vitae, a statement of current and planned
research, and a statement of teaching philosophy and
interests merged into a single PDF file; three letters
of reference are also required and up to three reprints
may be included. Review of applications will begin
February 15, 2013 and will continue until the position
is filled. Please send the application PDF document
and reprints electronically to e-mail intbiosearch@
clemson.edu. References should be instructed to send
letters directly to this same e-mail address with the
applicant_s name in the subject line. Further informa-
tion about this position, departmental resources, pro-
grams, and faculty research interests are available at
website: http://www.clemson.edu/biosci.
Clemson University is an Affirmative Action/Equal Oppor-
tunity Employer and does not discriminate against any individual
or group of individuals on the basis of age, color, disability, gender,
national origin, race, religion, sexual orientation, veteran status or
genetic information.
TENURE-TRACK ASSISTANT PROFESSOR
IN CELL BIOLOGY
Clemson University
The Department of Biological Sciences at Clemson
University is seeking a tenure-track Assistant Professor
starting August 2013 for its expanding teaching and
research emphasis in cell biology. Candidates that fo-
cus on all aspects of cell biology will be considered, but
those with interests in developmental biology, immu-
nology, physiology, and endocrinology, especially as
related to human diseases, are invited to apply.
The successful candidate will develop and maintain a
strong, independent, and extramurally funded research
program and contribute to both undergraduate and
graduate (M.S.-Ph.D.) education in Biology and/or Mi-
crobiology. Teaching responsibilities include one up-
per level undergraduate course per semester in our cell
biology and/or microbiology cores areas and graduate
course(s) in one_s specialty.
Qualifications Required: Applicants must have a
Ph.D. in Cell Biology, Developmental Biology, Immu-
nology, Physiology, or a related discipline, and at least
two years of relevant postdoctoral research experience.
Application Procedure: Applications should be sent
as a single PDF file containing: a cover letter, curric-
ulum vitae, a statement of research interests, copies of
up to three publications, and contact information for
three references, and sent to e-mail: cellbiosearch@
clemson.edu. Applications will be reviewed begin-
ning February 15, 2013 and will continue until the
position is filled. More information on the department,
programs, and faculty research interests can be found
at website: http://www.clemson.edu/biosci.
Clemson University is an Affirmative Action/Equal Oppor-
tunity Employer and does not discriminate against any individual
or group of individuals on the basis of age, color, disability, gender,
national origin, race, religion, sexual orientation, veteran status or
genetic information.
POSITIONS OPEN
SENIOR-LEVEL FACULTY POSITIONS
University of California, Riverside
Bourns College of Engineering
The Bourns College of Engineering at the
University of California, Riverside expects to hire
two outstanding senior faculty to provide leader-
ship in sustainable energy engineering begin-
ning in the 2013/2014 academic year. Highly
motivated and recognized individuals are sought
to lead and conduct cutting-edge research in gen-
eration, storage, distribution, and/or materials
development for sustainable energy systems.
Both positions will be interdisciplinary with the
successful candidate affiliated with one or more
engineering departments in the College, the Ma-
terials Science and Engineering Program, and the
Winston Chung Global Energy Center. Specific
areas of interest are provided at website: http://
www.engr.ucr.edu/facultysearch/. Equal Em-
ployment Opportunity /Affirmative Action Employer.
VERTEBRATE ANATOMY
Saint Anselm College
Manchester, NH
A one-year, non-tenure-track faculty position begin-
ning in August 2013. Teaching assignments include
Human Anatomy and Physiology (with lab) and Com-
parative Anatomy (with lab) and a third course. Submit
a cover letter and curriculum vitae at website: http://
www.anselm.edu/hr and three letters of recommen-
dation should be submitted to Dr. Donald Rhodes,
e-mail: drhodes@anselm.edu by February 15, 2013.
Your
career
is our
cause.
• Job Postings
• Job Alerts
• Resume/CV
Database
• Career Advice
• Career Forum
www.
sciencecareers.org
Get help
fromthe
experts.
POSITIONS OPEN
DEAN
Weill Cornell Graduate School of
Medical Sciences
The Weill Cornell Graduate School of Medical Sci-
ences is a unique partnership between the Weill Cornell
Medical College and the Sloan-Kettering Institute for
Cancer Research of Memorial Sloan-Kettering Cancer
Center in New York City. The Graduate School builds
on the combined strengths of these two institutions,
and is composed of faculty from the Weill Cornell
Medical College and The Sloan-Kettering Institute.
Currently the Graduate School enrolls È450 Ph.D.
and M.D.-Ph.D. students, and the combined faculty
numbers È300. The Dean of the Graduate School is
the chief academic officer responsible for the aca-
demic programs in the basic and health sciences that
are offered by the Graduate School. The two institu-
tions are now searching for a new dean, to take over
from the current Dean who is stepping down after 15
years of service to the Graduate School.
Candidates for the position of Dean of the Grad-
uate School of Medical Sciences should be interna-
tionally recognized leaders in research and graduate
education, with significant experience in the admin-
istration of such programs. The individual should hold
an M.D. and/or Ph.D. degree and have demonstrated
leadership qualities, and be committed to the recruit-
ment of a diverse student body in terms of gender, race
and ethnicity.
The Dean will provide dynamic leadership, with ma-
jor recapitalization, refocus, and enrichment of the
Graduate School_s missions in research education and
training. The successful candidate will shape and im-
plement a strategic vision for an institution that will
train the future leaders in basic and translational sci-
ence, who are prepared to lead the way in elucidating
the biological mechanisms underlying the genesis of
disease and thereby identifying novel cures. To this end,
the Graduate School will interact synergistically with
an incomparable set of partners on the tri-institutional
campus on the Upper East Side of Manhattan—including
its sister colleges at Cornell University as well as its
neighbors and affiliates: NewYork-Presbyterian Hos-
pital, the Hospital for Special Surgery, the Memorial
Sloan-Kettering Cancer Center and The Rockefeller
University.
The Dean of the Weill Cornell Graduate School of
Medical Sciences reports to the Provost for Medical
Affairs at Cornell University, and has responsibility for
personnel, facilities, finance, planning, and develop-
ment. The Dean oversees the office of the Graduate
School, which has responsibility for all matters relating
to student admission, progression, and graduation, in-
cluding curriculum and outcomes analysis.
We appreciate receiving recommendations or nom-
inations of potential candidates. Candidates who choose
to self-nominate should submit a cover letter in which
they indicate their interest, experience, and qualifica-
tions for the position, their curriculum vitae, and the
names and contact information of three references to:
Professor Olaf S. Andersen, c/o e-mail: mad2033@
med.cornell.edu.
Weill Cornell Medical College is an Equal Opportunity/
Affirmative Action Educator and Employer.
FAST TRACK IN LIFE SCIENCES
This international programof the University of Zurich
is open to outstanding Bachelor students, who wish
to obtain a Ph.D. degree in the Life Sciences. Appli-
cants must hold a Bachelor degree in Biology or Bio-
chemistry (possibly also in Bioinformatics or applied
Mathematics). Financial support is provided in accord-
ance with the guidelines of the Swiss National Science
Foundation for Ph.D. student salaries. After success-
fully completing the Master_s program, students may
enter one of the Ph.D. Programs at the Life Science
Zurich Graduate School. Applications for the fall se-
mester are welcome by February 28, 2013 website:
http://www.biologie.uzh.ch/Studium/Studiengang/
FastTrack.html.
o
n
l
i
n
e

@
s
c
i
e
n
c
e
c
a
r
e
e
r
s
.
o
r
g
4 JANUARY 2013 VOL 339 SCIENCE www.sciencecareers.org 110

Master your semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master your semester with Scribd & The New York Times

Cancel anytime.