ET AL.

28 January 2011 | $10

Bioconnect.Making connections in bio logical.Your Favorite Gene powered by Ingenuity has always allowed you to make connections to yourbiology through dynamic interaction networks and pathways. Based on your feedback, YourFavorite Gene now includes a biologically relevant literature search, new gene regulation viewers,expression study results, and identifies biochemical compounds related to your gene. Spend lesstime searching multiple sites and find it all on Your Favorite Gene, and best of all, it’s free.wherebiobegins.com/yfgSigma and Sigma-Aldrich are registered trademarks belonging to Sigma-Aldrich Co. and its affiliate Sigma-Aldrich Biotechnology, L.P.Ingenuity is a registered trademark belonging to Ingenuity Systems.

R&D Systems Tools for Cell Biology ResearchProteinsAntibodiesELISAsAssay ServicesMultiAnalyte ProfilingR&D Systems Quantikine ® ELISAsThe Most Referenced ImmunoassaysA direct measure of product quality is the frequency of citations in the scientific literature. R&D Systems has morethan 20 years of experience designing, testing, and optimizing the most cited ELISA kits in the world. Find out whyscientists trust R&D Systems ELISAs more than any other brand.Activity AssaysStem CellsELISpot KitsFlow CytometryCell SelectionR&D Systems is the Most Referenced ELISA ManufacturerR&D Systems41.8%17.3%2.5%2.1%1.6%1.4%Approximately 42% of Referenced Immunoassays are Developed and Manufactured byR&D Systems. A survey of 860 manuscripts from 44 journals was conducted to compare thenumber of citations specifying the use of R&D Systems ELISAs to the number citing ELISAs fromother commercial sources. A total of 433 ELISA citations referencing immunoassays from 66different vendors were identified in the survey.7.9%3.7%NEW Quantikine ELISA Kits α1-Acid Glycoprotein Angiopoietin-like 3 CathepsinV Clusterin Dkk-1 EGF R/ErbB1 EG-VEGF/PK1 Fetuin A FGF-21 Galectin-3 Gas 6 GDF-15 IL-17A/F Heterodimer IL-19 Lipocalin-2/NGAL MBL Proprotein Convertase 9/PCSK9 Periostin/OSF-2 Progranulin ST2/IL-1 R4 Thrombomodulin/CD141 Tie-1 TIM-1/KIM-1For more information visit our website at www.RnDSystems.com/go/ELISAFor research use only. Not for use in diagnostic procedures.R&D Systems, Inc. www.RnDSystems.comR&D Systems Europe, Ltd. www.RnDSystems.co.ukR&D Systems China Co., Ltd. www.RnDSystemsChina.com.cn

CONTENTS Volume 331 Issue 6016EDITORIAL377 New Views on News and ResearchColin Norman and Stewart WillsNEWS OF THE WEEK382 A roundup of the week’s top storiesNEWS & ANALYSIS386 Collins Sparks Furor WithSupposed NIH Reshuffling387 Did Modern Humans TravelOut of Africa Via Arabia?>> Report p. 453388 Despite Sensitivities, ScientistsSeek to Solve Haiti’s Cholera Riddle389 Pressure Growing to Set a Dateto Destroy Remaining Smallpox Stocks390 Last-Ditch Effort to Save Embattled Ape391 Telling Time Without Turning On GenesNEWS FOCUS392 A New View of the Birth of Homo sapiensThe Species Problem395 Going the DistanceTreading Air>> Science PodcastLETTERS398 Recognizing Scientists and TechnologistsR. M. WhiteGenetics-Based Field StudiesPrioritize SafetyA. A. JamesOrigins of BiodiversityV. RullResponseC. Hoorn et al.400 CORRECTIONS AND CLARIFICATIONSBOOKS ET AL.401 Honeybee DemocracyT. D. Seeley, reviewed by L. Chittka and A. Mesoudi402 A Rich BioethicsA. Briggle, reviewed by T. Lewens403 High SocietyM. Jay et al., curatorsEDUCATION FORUM404 Defeating Creationism in the Courtroom,But Not in the ClassroomM. B. Berkman and E. PlutzerPERSPECTIVES406 Cloud Computing—What’s in It for Meas a Scientist?A. Fox407 The Genomic View of BacterialDiversificationM. C. Enright and B. G. Spratt>> Research Article p. 430409 A Tail of DivisionA. F. Cowman and C. J. Tonkin>> Report p. 473410 Why Starving Cells Eat ThemselvesD. G. Hardie>> Report p. 456411 Chemical Kinetics Under TestM. H. Alexander>> Report p. 448SCIENCE PRIZE ESSAY413 Penguins and Polar Bears IntegratesScience and LiteracyJ. Fries-Gaither and K. LightleASSOCIATION AFFAIRS416 Presidential Address:Life on the River of ScienceP. Agrepage 395page 403CONTENTS continued >>COVERFalse-colored scanning electron micrograph of the bacterialpathogen Streptococcus pneumoniae (typical length,~1 micrometer), a major cause of pneumonia and meningitis.On page 430, Croucher et al. describe the use of whole-genomesequencing to track the global spread of a rapidly recombiningS. pneumoniae lineage as it adapts to antibiotic use andvaccine introduction.Image: Dennis Kunkel Microscopy, Inc./Visuals Unlimited, Inc.DEPARTMENTS375 This Week in Science378 Editors’ Choice380 Science Staff422 AAAS News & Notes481 New Products482 Science Careerswww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 369

the difference between discover and do overGibco ® . Because every little thing matters.Fact: your research will only be as accurate, as efficient, as groundbreaking as the cell cultureyou’re working with. That’s why you’re settling for nothing less than Gibco ® media. Consideringit could be the difference between publishing your results and reading someone else’s, whyexperiment with your experiment?invitrogen.com/discovergibcodownload the free mobile app at http://gettag.mobi

CONTENTSREVIEW426 The Newest Synthesis: Understandingthe Interplay of Evolutionary andEcological DynamicsT. W. SchoenerRESEARCH ARTICLES430 Rapid Pneumococcal Evolution inResponse to Clinical InterventionsN. J. Croucher et al.Streptococcus pneumoniae evades vaccinesand drugs by high levels of recombinationand rapid adaptation.>> Perspective p. 407435 The Genetic Landscape of the ChildhoodCancer MedulloblastomaD. W. Parsons et al.Genomic analysis of a childhood cancerreveals markedly fewer mutations thanwhat is typically seen in adult cancers.REPORTS439 Rotational Symmetry Breaking inthe Hidden-Order Phase of URu 2 Si 2R. Okazaki et al.The mysterious hidden-order phaseof a heavy-fermion compound may bean electronic nematic state.442 High-Gain Backward Lasing in AirA. Dogariu et al.Focused ultraviolet light creates a distant“lasing spark” in air that can then be usedfor remote detection.445 Magnetic Bistability of Moleculesin Homogeneous Solution atRoom TemperatureS. Venkataramani et al.A photoresponsive ligand is used to reversiblymodulate the magnetic properties of a nickelcompound.448 Kinetic Isotope Effects for the Reactionsof Muonic Helium and Muonium with H 2D. G. Fleming et al.Calculated reaction rates for two hydrogenisotopes, one 36 times heavier than theother, agree with experiments at 500 kelvin.>> Perspective p. 411450 Enhanced Modern Heat Transfer to theArctic by Warm Atlantic WaterR. F. Spielhagen et al.Water flow from the Atlantic Ocean intothe Arctic through the Fram Strait is warmerthan at any time in the past 2000 years.453 The Southern Route “Out of Africa”:Evidence for an Early Expansion ofModern Humans into ArabiaS. J. Armitage et al.Artifacts in eastern Arabia dating to 100,000years ago imply that modern humans leftAfrica early, as climate fluctuated.>> News & Analysis p. 387456 Phosphorylation of ULK1 (hATG1) byAMP-Activated Protein Kinase ConnectsEnergy Sensing to MitophagyD. F. Egan et al.A protein kinase links energy stores to controlof autophagy.>> Perspective p. 410461 Effects of Experimental SeaweedDeposition on Lizard and Ant Predationin an Island Food WebJ. Piovia-Scott et al.In a simple island ecosystem, a rare pulseof resource input alters the interactionbetween predator species.463 Metagenomic Discovery ofBiomass-Degrading Genes andGenomes from Cow RumenM. Hess et al.Metagenomic sequencing of biomass-degradingmicrobes from cow rumen reveals newcarbohydrate-active enzymes.468 Cleavage of NIK by the API2-MALT1 FusionOncoprotein Leads to Noncanonical NF-kBActivationS. Rosebeck et al.An oncogenic fusion protein promoteslymphomagenesis by activating anoncanonical signaling pathway.473 Intramembrane Cleavage of AMA1 TriggersToxoplasma to Switch from an Invasiveto a Replicative ModeJ. M. Santos et al.Membrane proteins govern a change frominvasion to replication of an intracellularparasite.>> Perspective p. 409477 Big and Mighty: Preverbal InfantsMentally Represent Social DominanceL. Thomsen et al.Infants use size as a predictor in judgingwhich contestant will win.>> Science PodcastCONTENTS continued >>page 413pages 387 & 453pages 409 & 473www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 371

Next-gen GWAS.NOW.It’s a content revolution.Up to 50% more coverage of common and rarevariants than all other arrays. Maximum power forany population.The Omni family of microarrays can propel yourstudies into true next-gen GWAS. With a clearlydefined product path to the future. Immediateutility. Future flexibility.Get on the path to next-gen GWAS.Now is the time. Get started atwww.illumina.com/GWAS

CONTENTSSCIENCEONLINESCIENCE LABSOnline Article RebootWe’re rethinking the format and functionality ofour online articles. Check out our working prototypeand share your feedback at labs.sciencemag.orgSCIENCEXPRESSwww.sciencexpress.orgMicrotubule Stabilization Reduces Scarringand Causes Axon Regeneration AfterSpinal Cord InjuryF. Hellal et al.Taxol stimulates the capacity of axons togrow after spinal cord injury.10.1126/science.1201148Structures of SAS-6 SuggestIts Organization in CentriolesM. van Breugel et al.Self-assembly of a centriolar protein maycontribute to organizing the cartwheel-like huband establishing centriole symmetry.10.1126/science.1199325The Magnitude and Duration of LateOrdovician–Early Silurian GlaciationS. Finnegan et al.Carbonate isotopes reveal a link between pastocean temperatures and mass extinction.10.1126/science.1200803Linking Policy on Climate and FoodH. C. J. Godfray et al.10.1126/science.1202899SCIENCENOWwww.sciencenow.orgHighlights From Our Daily News CoverageFrédéric Chopin’s ‘Madness’ DiagnosedNew study argues that composer sufferedfrom a type of epilepsy.http://scim.ag/mad-chopinThe World’s Smallest FarmersSingle-celled amoebas plant bacteria for food.http://scim.ag/tiny-farmersWhen Trees Attack, Fungus Can ParryGenetic analysis reveals why the pine beetlehas ravaged so many trees.http://scim.ag/GrosmanniaSCIENCESIGNALINGwww.sciencesignaling.orgThe Signal Transduction Knowledge Environment25 January issue: http://scim.ag/ss012511RESEARCH ARTICLE: Genome-Wide RNAi ScreenReveals Disease-Associated Genes That AreCommon to Hedgehog and Wnt SignalingL. S. Jacob et al.Loss of the kinase and tumor suppressor Stk11 (Lkb1)has opposite effects on Hh and Wnt signaling.RESEARCH ARTICLE: V 2 Receptor–MediatedAutocrine Role of Somatodendritic Releaseof AVP in Rat Vasopressin Neurons UnderHypo-Osmotic ConditionsK. Sato et al.The antidiuretic hormone arginine vasopressin(AVP) is also an autocrine signal that functionsin cell volume regulation.RESEARCH ARTICLE: Rho and Rho-Kinase Activityin Adipocytes Contributes to a Vicious Cycle inObesity That May Involve Mechanical StretchY. Hara et al.Mechanical stretch activates Rho-kinase in adipocytes,promoting obesity and obesity-related complications.PODCASTResearchers discuss how mechanical stretch regulatessignaling in adipocytes and in hypothalamic neurons.MEETING REPORT: Dependence Receptors—From Basic Research to Drug DevelopmentP. Mehlen and D. E. BredesenPERSPECTIVE: Control of DNMT1 Abundancein Epigenetic Inheritance by Acetylation,Ubiquitylation, and the Histone CodeC. BronnerJOURNAL CLUB: The Role of L(u)ck inT Cell TriggeringR. BerrySCIENCECAREERSwww.sciencecareers.org/career_magazineFree Career Resources for ScientistsScience Blogging and TenureV. RaperCan pretenured scientists blog about sciencewithout damaging their careers?http://scim.ag/dFDgmlA Loyal Fan of Women’s Health ResearchK. HedePhysician-scientist Rebecca Jackson’s enthusiasmfor research is matched only by her passionfor Ohio State football.http://scim.ag/hPH8NJExperimental Error: Lies, Damned Lies,and SeminarsA. RubenFor all the naive and gullible graduate studentsout there, here is a handy guide to what thosespeakers are really saying.http://scim.ag/ezFYgKSCIENCETRANSLATIONAL MEDICINEwww.sciencetranslationalmedicine.orgIntegrating Medicine and Science26 January issue: http://scim.ag/stm012611COMMENTARY: Between Confidentiality andScientific Exchange—The Place of Publicationin Drug Discovery and Pharmaceutical ResearchM. ClozelA combined preclinical and clinical publication mightbe the ideal approach in translational medicine.RESEARCH ARTICLE: UroporphyrinogenDecarboxylase Is a Radiosensitizing Targetfor Head and Neck CancerE. Ito et al.Uroporphyrinogen decarboxylase inhibitionsensitizes head and neck cancer to bothradiotherapy and chemotherapy.RESEARCH ARTICLE: Manipulating theBioenergetics of Alloreactive T Cells CausesTheir Selective Apoptosis and ArrestsGraft-Versus-Host DiseaseE. Gatza et al.Bioenergetic properties differentiate alloreactiveT cells from other proliferating cells and can beexploited to arrest GVHD in mice.SCIENCEPODCASTwww.sciencemag.org/multimedia/podcastFree Weekly ShowDownload the 28 January Science Podcast to hearabout infants’ awareness of social dominance,studying bird metabolism in wind tunnels,your letters to Science, and more.SCIENCEINSIDERnews.sciencemag.org/scienceinsiderScience Policy News and AnalysisSCIENCE (ISSN 0036-8075) is published weekly on Friday, except the lastweek in December, by the American Association for the Advancement ofScience, 1200 New York Avenue, NW, Washington, DC 20005. Periodicals Mailpostage (publication No. 484460) paid at Washington, DC, and additional mailingoffices. Copyright © 2011 by the American Association for the Advancement ofScience. The title SCIENCE is a registered trademark of the AAAS. Domestic individualmembership 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 availableupon 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 accountnumber. 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 prepaidincludes surface postage; bulk rates on request. Authorization to photocopymaterial for internal or personal use under circumstances not falling within the fairuse provisions of the Copyright Act is granted by AAAS to libraries and other usersregistered with the Copyright Clearance Center (CCC) Transactional Reporting Service,provided that $25.00 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 theReader’s Guide to Periodical Literature and in several specialized indexes.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 373

UNDERSTANDING CHANGENew tools to advance epigenetics researchFor over 35 years, New England Biolabs has been committed to understanding the mechanisms of restriction andmethylation of DNA. This expertise in enzymology has recently led to the development of a suite of validatedproducts for epigenetics research. These unique solutions to study DNA methylation are designed to address someof the challenges of the current methods. EpiMark validated reagents simplify epigenetics research and expand thepotential for biomarker discovery.Simplify DNA methylation analysis with MspJIPlantHela (Maize) Yeast– + – + – + MspJIEpiMark validated products include:• Newly discovered methylation-dependent restriction enzymes• A novel kit for 5-hmC and 5-mC analysis and quantitation• Methyltransferases• Histones• Genomic DNAs32 bpTo learn how these products can help you to betterunderstand epigenetic changes, visit neb.com/epigenetics.MspJI recognizes methylated and hydroxymethylated DNA and cleavesout 32 bp fragments for sequencing analysis. Overnight digestion of1 µg of genomic DNA from various sources with or without MspJI isshown. Note: Yeast DNA does not contain methylated DNA, thereforeno 32-mer is detected.CLONING & MAPPINGDNA AMPLIFICATION& PCRRNA ANALYSISPROTEIN EXPRESSION& ANALYSISGENE EXPRESSION& CELLULAR ANALYSISwww.neb.com

EDITED BY STELLA HURTLEYCREDITS (TOP TO BOTTOM): JONAS LØVAAS GJERSTAD; OKAZAKI ET AL.Digesting GrassIdentification of additional enzymes that candegrade cellulose efficiently should help inthe development of biofuels on an industrialscale. Uncultured microorganisms living incow rumen are highly effective at degradingplant cell walls. Hess et al. (p. 463) usedmetagenomics and single-genome sequencingto assemble draft genomes from microbesadhering to rumen-incubated switchgrassto identify nearly 28,000 genes relatedto known biomass-degrading families. Ninetycandidate carbohydrate-degrading enzymeswere synthesized and their activity analyzedagainst 10 different substrates, including thebiofuel crops miscanthus and switchgrass.The data set greatly expands the repertoire offull-length genes available for use in industrialbiotechnology.Nickels High and LowCertain transition metal complexes undergo aphenomenon termed spin crossover, in whicha rise in temperature shifts the proportion ofvalence electrons that share the same spinorientation, and thus drive magnetic response.However, the collective shift across an ensembleof molecules in solution tends to be gradual andhas been difficult to harness in switching applicationssuch as contrast variation in magneticresonance imaging. Venkataramani et al.(p. 445) present an alternative magnetic switchingscheme, in which a photoresponsive ligandis used to modulate the electronic structureof a nickel complex. Isomerizations driven bydifferent wavelengths of light selectively pushthe ligand onto or off the nickel ion, therebyinducing a high or low spin state. Up to 75%of the ensemble could thus be switched rapidlybetween magnetically active and passive statesat room temperature.Ecology ↔ EvolutionEcology and evolution are fundamentally connected—ecologicalprocesses drive patterns ofnatural selection and thus shape evolutionarytrajectories. The possibility that evolutionarychange could also influence ecologicalprocesses is much less clear; it has often beenassumed that evolutionary change takes toolong to impact ecology. Schoener (p. 426)reviews how evolutionary change can in fact beextremely rapid, occurring over ecological timescales. Rapid evolutionary changes can thusgive feedback on ecology and impact ecologicalprocesses such as population growth, communitystructure, and productivity.The Pneumococcus GamePneumonia and meningitis caused by Streptococcuspneumoniae have accounted for millionsof deaths through the ages. The Spanish strainis a globally occurring, multiply antibioticresistantlineage that has been exceptionallytroublesome. Isolates of this lineage have beencollected since 1984 from diverse geographicorigins. Croucher et al. (p. 430, see the cover;see the Perspective by Enright and Spratt) havenow used high-throughput sequencing technologyto dissect serotype 23F’s detailed evolution.The bacterium appears to respond to publichealth measures with extraordinary genetic agility,using a variety of recombination and basesubstitutions that can turn over three-quartersof the genome, enabling the pathogen to evadethe effects of vaccines and antibiotics.Decoding aChildhood CancerThe identification of recurrent genetic changesin human tumors can provide importantmechanistic insights into how tumors ariseand, ideally, prompt new ideas for effectivetherapies. To date, this “cancer genomics”strategy has been applied only to adult cancers.Parsons et al. (p. 435, published online 16December) now catalog the genetic alterationspresent in medulloblastoma, a brain tumor thatmainly affects children. Interestingly, there were5 to 10 times fewer genetic alterations in thesetumors compared with solid tumors that typicallyaffect adults. Among the most frequentlymutated genes were two coding for enzymesthat methylate histones, as well as genes affectingsignaling pathways critical for normal braindevelopment.UnderstandingMysterious OrderThe heavy fermion compound URu 2 Si 2 enters anordered phase below 17.5 kelvin but, despite25 years of extensive study, the exact nature ofthis phase remains a mystery. Using very pure,small specimens, Okazaki et al. (p. 439) foundthat the magnetic susceptibility in the orderedphase breaks the fourfoldrotational symmetryof the URu 2 Si 2 crystal,which suggests that thehidden order is an electronicnematic state.LaserBackfireStandoff detection is an important method tomonitor or probe regions that are otherwiseinaccessible to direct sampling techniques.Dogariu et al. (p. 442) report the observationof high-gain infrared lasing in air from the focalregion of an ultraviolet laser. The backwarddirectedinfrared laser light could then be usedas a remote spectroscopic tool that samplesthe air on its return path back to the senderposition. The technique should find a range ofapplications—from the detection of greenhousegas emissions and pollution to the detection ofpotentially hazardous and explosive materials.Reaction of HydrogenIsotopes, Great and SmallKinetic isotope effects reveal mechanisticinsight into chemical reactions. Larger isotoperatios often lead to larger changes in rate,and, commonly, the largest ratio is 2—forthe substitution of deuterium for hydrogen.Fleming et al. (p. 448; see the Perspective byAlexander) compared results for the simplestchemical reaction, of an H atom with H 2, fortwo hydrogen isotopes created with positive andnegative muons, which provide an unprecedentedlylarge mass ratio of 36. At 500 kelvin, therelative reaction rates measured agreed withthose calculated using variational transitionstate theory.Continued on page 376www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011375

This Week in ScienceContinued from page 375Warmest in 2000 YearsAlthough the climate warming of the past century has occurred nearly everywhere, the Arctic shows thegreatest temperature increase. Most of the heat transported to the Arctic comes from the Atlantic Ocean,which has warmed greatly over the past 150 years. How did the temperature of Arctic inflow from theAtlantic vary before anthropogenic climate warming began? Spielhagen et al. (p. 450) present a recordof oceanic temperature variations for the last 2000 years, which shows that Atlantic water entering theArctic through the Fram Strait is warmer than it has been in two millennia. This unprecedented warmingis likely to represent a key factor in the apparent transition toward an ice-free Arctic Ocean.Before the Mediterranean ExodusModern humans originated in Africa and then spread to Eurasia and beyond. The timing and locationsof their emigration have been uncertain; genetic and archaeological data support an exodus along theMediterranean by 60,000 years ago, but earlier attempts may have occurred, for example, in responseto the massive Toba volcanic eruption about 75,000 years ago. Armitage et al. (p. 453; see the newsstory by Lawler) now describe artifacts from about 100,000 years ago found in eastern Arabia, indicatingthat modern humans were already there by then. This location would have allowed access tothe Fertile Crescent and India as sea level dropped. The findings suggest that there may indeed havebeen an early exodus of modern humans from Africa.Predatory SynergismSynergies between species, in whichthe combined effect of multiplespecies is greater than the sumof their individual effects, linkbiodiversity to ecosystem function,but these synergies may be sensitiveto external influences. Piovia-Scottet al. (p. 461) investigated how asynergy between two very different predator species in the Bahamas was affected by the influx ofresources from the adjacent ocean. Adding seaweed to entire small islands (simulating the naturalmarine-to-terrestrial energy transfer that occurs during storms) eliminated the synergistic effect oftwo predators—lizards and ants—on arthropod herbivores and their food plants, increasing theoverall rate of herbivory.Parasite Replication TriggerApicomplexan parasites, including Toxoplasma gondii and the malaria parasite Plasmodium falciparum,actively invade host cells. Little is known about the signals that govern initiation of replicationonce the parasite is intracellular. Apical membrane antigen 1 (AMA1) is released onto the parasitesurface during invasion and cleaved by intramembrane proteolysis, mediated by a rhomboid-likeserine protease, ROM4. Santos et al. (p. 473, published online 23 December; see the Perspective byCowman and Tonkin) used conditional expression of Toxoplasma ROM4 to show that ROM4 activityis not essential for invasion, but instead is required for subsequent replication of the intracellularparasite. Furthermore, transgenic expression of the cleaved cytoplasmic tail of AMA1 alone—eitherfrom the Toxoplasma AMA1 or from its P. falciparum ortholog—completely restored the replicativecapacity of the intracellular parasites. Thus, intramembrane cleavage of AMA1 is required to triggerparasite replication within the host cell.Bigger Beats SmallerSocial cooperation has been the subject of intense experimental investigation, both in humans andin other animals, and recent studies have detailed the developmental origins of these representationsin human children and infants. Thomsen et al. (p. 477) now document the appearance ofsocial dominance awareness in infants before the end of the first year. Infants were found to becapable of representing the conflicting goals of two cartoon characters, appearing to use size as acue to predict which character will win.CREDIT: PIOVIA-SCOTT ET AL.28 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

EDITORIALCREDITS: (TOP LEFT) MICHAEL W. HICKS; (BOTTOM LEFT) MARTYN GREENColin Norman is theNews Editor of Science.E-mail: cnorman@aaas.org.Stewart Wills is theEditorial Director,Web & New Media, ofScience. E-mail: swills@aaas.org.New Views on News and ResearchTHIS WEEK, SCIENCE UNVEILS TWO PROJECTS AIMED AT MAKING OUR CONTENT MORE USEFULand accessible amid today's information glut: a revamp of the News section to provide afaster launch into the week's top stories, and a prototype of some new ideas for our onlinearticles, on which we seek your feedback.For 20 years, Science’s News section has included a page called Random Samples, aneclectic mix of brief items about science and the scientific community. Reader surveys haveconsistently shown that it is one of the most-read features in the magazine. Starting withthis issue (see p. 382), we are expanding and refocusing the type of coverage that has madeRandom Samples so popular. Each week, the News section will open with three or four pagesof brief items under the heading News of the Week. The aim is to give busy readers a quicktake on the important events in the world of science. It will include a roundup of the topstories in science policy, written from our far-flung bureaus: Washington,DC; Cambridge, UK; Paris; Berlin; Tokyo; Beijing; NewDelhi; and São Paulo. We will bring you news of scientists makingheadlines and present the bottom line on newly published researchpapers. In the tradition of Random Samples, we will also includeitems that are offbeat, whimsical, or just plain interesting—quotes,factoids, striking images. News of the Week will be edited by LaurenSchenkman, who has been editing Random Samples since the tragicdeath last year of the page’s long-time editor Constance (Tancy)Holden. The items will be drawn in part from Science’s daily onlinecoverage of research findings, ScienceNOW, and our science policyblog ScienceInsider (http://news.sciencemag.org).News of the Week is designed to provide a quick overview ofthe week’s top events; the rest of the News section focuses beyondthe headlines. In a section renamed News & Analysis, our awardwinningnews team will draw on expert opinion to put the news in context. And the NewsFocus section will continue to take a broad look at the trends, personalities, and eventsshaping the world of science. At a time when traditional journalism is under severe pressure,and science reporting is declining in many major media outlets, our goal—and that of ourpublisher, the American Association for the Advancement of Science—is to provide the bestcoverage across the sciences and throughout the interface of science with society.On a different note, we're also reaching out to our Web readers this week for theirfresh, direct input to help us rethink the format and functionality of our online articles.Recently, a staff team from across Science has been exploring a variety of new ideas forWeb article content, including a tabbed interface, summary material to help put the articlein context, more visible ties to related content, a different treatment for figures andsupport ing online data, and a variety of other possibilities that we hope will save userstime, make online articles more readable and functional, and open up new possibilities forfurther exploration.We've rolled these ideas into an initial prototype that is posted in a beta version athttp://labs.sciencemag.org/. We encourage readers to try out the prototype and leave ustheir thoughts. What do you like and dislike about this first model? What's missing thatwould help you work more efficiently with the content; and what's distracting? What's at thetop of your own "wish list" for more effectively interacting with scientific content? The prototypeis very much "release 0.1"; we plan to incorporate new ideas and adjustments basedon your feedback. This process will help us to zero in on a new article interface that's bettertuned to user needs in the crowded and rapidly changing environment for communicatingscientific information. Stop by for a look, give us your thoughts and ideas, and become apart of this new undertaking.– Colin Norman and Stewart Wills10.1126/science.1203015www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011377

EDITORS’CHOICEEDITED BY KRISTEN MUELLER AND JAKE YESTONSOCIOLOGYComplex ContributionsIt is well established that academic achievementis related to socioeconomic status (SES), but itis not yet clear at what age this effect emerges.Tucker-Drob et al. examined whether the effectsof SES on mental ability could be observed in infancy.They analyzed 750 pairs of twins from theEarly Childhood Longitudinal Study, Birth Cohortdata file. At 10 months and 2 years of age, thechildren’s mental ability was assessed usingtests that included pulling a string to ring a bell,putting cubes in a cup and matching pictures,among other tasks. SES was determined fromthe education and occupations of the parentsand family income.At 10 months of age, SES was not relatedto mental ability. Between 10 months and 2years of age, however, the authors observed thatincreased SES was associated with larger gainsin mental ability. At 2 years of age, the influenceof genetics on cognitive development was higherin children of higher SES, whereas genetics hadvery little effect on the mental ability of childrenraised in low-SES homes. These findings do notindicate any differences in intrinsic intelligence,but may provide support for efforts to provideenrichment for young children from disadvantagedbackgrounds. — BJPsychol. Sci. 22, 125 (2011).SIGNALINGSignaling Across PathwaysThe signaling pathways that regulate cellfunction are obvious targets for therapeuticintervention, but such strategies are complicatedby our incomplete understanding of thesepathways and also the links between them.Jacob et al. used an RNA interference screen inmouse cells to find new components of the Wntand Hedgehog signaling pathways. Both are keyregulatory pathways in development that arealso linked to various diseases, including cancer.Loss of the protein kinase Lkb1 inhibitedHedgehog pathway–dependent gene expressionthrough effects on the primary cilium, whereHedgehog signaling is organized. Loss of Lkb1also disrupted Wnt signaling, but through aThe tremendous rumble of thunderclouds that sends children under their bedcoversand provides the electrifying atmosphere indicative of an impending storm also holdsfascination for scientists who wish to understand the mechanism of lightning and theassociated buildup and release of energy. Experiments have become more sophisticatedsince Benjamin Franklin’s kite-flying forays to probe lightning directly. Of recent interestare terrestrial gamma-ray flashes (TGFs)—flashes of high-energy radiation that occurpredominantly at the upper regions of clouds in the tropics and appear to originatedeep in the atmosphere at an altitude of 10 to 20 km. Since space-based satellitesfirst observed these TGFs in the 1990s, the formation mechanisms have tended to bedescribed in terms of runaway electron acceleration models resulting from the largeelectric fields thunderclouds discharge. Tavani et al. report on a systematic study ofobservations across a wide energy range (350 keV to 100 MeV) by the Italian SpaceAgency’s AGILE observatory that raise questions about those models. They argue thatthe emission spectrum at the very highest energies cannot be reconciled with existingmodels and that other mechanism(s) must be at play. Although no definitive model isprovided, the detailed data set should prove useful for those trying to understand theenergy dynamics of our atmosphere. — ISOPhys. Rev. Lett. 106, 18501 (2011).distinct mechanism. Similar results were observedwhen expression of Lkb1 was inhibited in zebrafishembryos. Similar expanded approaches thusappear necessary tounderstand the functionsof key signalingcomponents and howthey might be managedto affect complexdiseases. — LBRSci. Signal. 4, ra4(2011).ASTROPHYSICSTuning in to Bright SourcesThe Fermi space telescope has revealed thousandsof new sources of gamma-ray emission; however,the nature of hundreds of them remains a mystery.To check whether some of the unidentifiedsources could be pulsars, Ransom et al. searchedfor radio pulsations in 25 sources in the FermiLarge Area Telescope Bright Source List—a catalogof the brightest sources detected in the earlymonths of the Fermi mission. These 25 sourcesCREDITS (TOP TO BOTTOM): ISTOCKPHOTO.COM; JACOB ET AL., SCI. SIGNAL. 4, RA4 (2011)37828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

CREDIT: BEAU LOTTOwere selected to be nonvariable and not associatedwith known pulsars, black holes at the centersof galaxies, or x-ray sources that had previouslybeen searched for radio pulsations. Radio observationswith the Green Bank Telescope revealedthree pulsars with millisecond periods, all of whichare in binary systems where a neutron star anda companion star orbit one another about theircommon center of mass. The new pulsars haveproperties consistent with those of typical radiomillisecond pulsars. Their detection suggests thatmost, if not all, radio millisecond pulsars producegamma rays, and that their radio and gamma-raybeams are comparable in size. It is thus possiblethat millisecond pulsars contribute to the diffuseisotropic gamma-ray background. — MJCAstrophys. J. 727, L16 (2011).CHEMISTRYCarbonic Acid AloftThe chemistry of carbon dioxide in water playsa remarkably diverse series of roles in our dailylives—modulating acidity in blood and oceanwater, lending soda its sparkle and bread itsfluff. It’s all the more remarkable, then, that theadduct of the two molecules, carbonic acid, spentcenturies eluding characterization. Its deprotonatedconjugate bicarbonate (HCO 3−) is reactivebut easily isolable; in contrast, it was only recentlyestablished that the HOC(O)OH molecule persistsfor any length of time in solution before breakingapart. Bernard et al. have now shown thatcarbonic acid is also at least kinetically stable inthe gas phase as well. By subliming the solid acidand then capturing it in a frozen argon matrix,the authors were able to detect the vibrationalsignatures of two conformational isomers (witha W-shaped one predominating) as well as ahydrogen-bonded dimer. The findings raise theprospect of finding the molecule in comet tails orother planetary atmospheres. — JSYAngew. Chem. Int. Ed. 10.1002/anie.201004729 (2010).SIGNALINGWorking Against the ClockOur circadian rhythms keep us in tune withthe day. Some of the molecular signals thatimplement and regulate the circadian rhythmhave been identified, but the complexity of thevarious systems affected by circadian rhythms isnot well understood. Many people work againsttheir circadian clock, whether it be a scientist whohops seven time zones eastward and still hopes tobe awake at a conference session with the lightsdimmed, or someone who works the night shiftand needs cognitive acuity and physical dexterityduring the hours when most of us are asleep. ToEDITORS’CHOICEfurther understand how disruption of the circadianrhythm affects us, Karatsoreos et al. studiedmice kept in an unnaturally short day/night cycle.With a cycle of 20 rather than 24 hours, themice showed a variety of disrupted responses.They gained weight, had elevated insulin levels,and demonstrated reduced cognitive flexibility.Neurons in the brain’s cortex showed reducedcomplexity. Study of these mice may help us understand,for example, the unexpected incidenceof obesity among shift workers. — PJHProc. Natl. Acad. Sci. U.S.A. 108, 10.1073/pnas.1018375108 (2011).EDUCATIONChild ScientistsWhat would happen if, instead of consultingprevious literature, scientists asked children foradvice on designing experiments? In the case ofthe Blackawton Bees, 8- to 10-year-old childrencapitalized on their own curiosity and observationsto devise questions, propose a hypothesis,design experiments, and perform data analysisin an original study examining how bees perceiveand remember their surroundings. Besidesdiscovering that bees use both color and spatialanalysis in deciding which color of flower to foragefrom, it served as an example of real scienceand engaged the students. This is evident in thepublished paper, written by the students, thatcontains statements such as “Before doing theseexperiments we did not really think about bees,”and “This experiment is important, because noone in history, including adults, has done thisexperiment before.” In this way, science educationbecame more of a process of contributingto asking questions and devising strategies toanswer those questions instead of a passiveclassroom lesson. Afterward, the students cameto the same conclusion that every scientist hascome to at one point in their career: “Science iscool and fun because you get to do stuff that noone has ever done before.” — MMBiol. Lett. 10.1098/rsbl.2010.1056 (2010).Call forPapersScienceSignalingScience Signaling, from thepublisher of Science, AAAS,features top-notch, peerreviewed,original researchweekly. Submit your manuscriptsin the following areasof cellular regulation:• Biochemistry• Bioinformatics• Cell Biology• Development• Immunology• Microbiology• Molecular Biology• Neuroscience• Pharmacology• Physiology andMedicine• Systems BiologySubscribing to Science Signalingensures that you and your labhave the latest cell signalingresources. For more informationvisit www.ScienceSignaling.orgChief Scientific EditorMichael B. Yaffe, M.D., Ph.D.Associate Professor, Department of BiologyMassachusetts Institute of TechnologyEditorNancy R. Gough, Ph.D.AAASSubmit your research at:www.sciencesignaling.org/about/help/research.dtlwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011

www.sciencemag.org1200 New York Avenue, NWWashington, DC 20005Editorial: 202-326-6550, FAX 202-289-7562News: 202-326-6581, FAX 202-371-9227Bateman House, 82-88 Hills RoadCambridge, UK CB2 1LQ+44 (0) 1223 326500, FAX +44 (0) 1223 326501SUBSCRIPTION SERVICES For change of address, missing issues, neworders and renewals, and payment questions: 866-434-AAAS (2227)or 202-326-6417, FAX 202-842-1065. Mailing addresses: AAAS, P.O.Box 96178, Washington, DC 20090-6178 or AAAS Member Services,1200 New York Avenue, NW, Washington, DC 20005INSTITUTIONAL SITE LICENSES please call 202-326-6755 for anyquestions or informationREPRINTS: Author Inquiries 800-635-7181Commercial Inquiries 803-359-4578PERMISSIONS 202-326-7074, FAX 202-682-0816MEMBER BENEFITS AAAS/Barnes&Noble.com bookstore www.aaas.org/bn;AAAS Online Store www.apisource.com/aaas/ code MKB6; AAASTravels: Betchart Expeditions 800-252-4910; Apple Store www.apple.com/eppstore/aaas; Bank of America MasterCard 1-800-833-6262priority code FAA3YU; Cold Spring Harbor Laboratory PressPublications www.cshlpress.com/affiliates/aaas.htm; GEICO AutoInsurance www.geico.com/landingpage/go51.htm?logo=17624;Hertz 800-654-2200 CDP#343457; Office Depot https://bsd.officedepot.com/portalLogin.do; Seabury & Smith Life Insurance 800-424-9883; Subaru VIP Program 202-326-6417; VIP Moving Serviceswww.vipmayfl ower.com/domestic/index.html; Other Benefi ts: AAASMember Services 202-326-6417 or www.aaasmember.org.science_editors@aaas.org (for general editorial queries)science_letters@aaas.org (for queries about letters)science_reviews@aaas.org (for returning manuscript reviews)science_bookrevs@aaas.org (for book review queries)Published by the American Association for the Advancement of Science(AAAS), Science serves its readers as a forum for the presentation anddiscussion of important issues related to the advancement of science,including the presentation of minority or conflicting points of view,rather than by publishing only material on which a consensus has beenreached. Accordingly, all articles published in Science—includingeditorials, news and comment, and book reviews—are signed and reflectthe individual views of the authors and not official points of view adoptedby AAAS or the institutions with which the authors are affiliated.AAAS was founded in 1848 and incorporated in 1874. Its mission is toadvance science, engineering, and innovation throughout the world forthe benefit of all people. The goals of the association are to: enhancecommunication among scientists, engineers, and the public; promote anddefend the integrity of science and its use; strengthen support for thescience and technology enterprise; provide a voice for science on societalissues; promote the responsible use of science in public policy; strengthenand diversify the science and technology workforce; foster education inscience and technology for everyone; increase public engagement withscience and technology; and advance international cooperation in science.INFORMATION FOR AUTHORSSee pages 352 and 353 of the 15 January 2010 issue oraccess www.sciencemag.org/about/authorsEDITOR-IN-CHIEF Bruce AlbertsEXECUTIVE EDITORNEWS EDITORMonica M. Bradford Colin NormanMANAGING EDITOR, RESEARCH JOURNALS Katrina L. KelnerDEPUTY EDITORS R. Brooks Hanson, Barbara R. Jasny, AndrewM. SugdenEDITORIAL SENIOR EDITORS/COMMENTARY Lisa D. Chong, Brad Wible; SENIOREDITORS Gilbert J. Chin, Pamela J. Hines, Paula A. Kiberstis (Boston), MarcS. Lavine (Toronto), Beverly A. Purnell, L. Bryan Ray, Guy Riddihough,H. Jesse Smith, Phillip D. Szuromi (Tennessee), Valda Vinson, JakeS. Yeston; ASSOCIATE EDITORS Kristen L. Mueller, Jelena Stajic, SachaVignieri, Nicholas S. Wigginton, Laura M. Zahn (San Diego); BOOK REVIEWEDITOR Sherman J. Suter; ASSOCIATE LETTERS EDITOR Jennifer Sills; EDITORIALMANAGER Cara Tate; SENIOR COPY EDITORS Jeffrey E. Cook, Cynthia Howe, HarryJach, Lauren Kmec, Barbara P. Ordway, Trista Wagoner; COPY EDITOR ChrisFiliatreau; EDITORIAL COORDINATORS Carolyn Kyle, Beverly Shields; PUBLICATIONSASSISTANTS Ramatoulaye Diop, Joi S. Granger, Emily Guise, JeffreyHearn, Michael Hicks, Lisa Johnson, Scott Miller, Jerry Richardson,Jennifer A. Seibert, Brian White, Anita Wynn; EDITORIAL ASSISTANTS EmilyC. Horton, Patricia M. Moore, Miriam Weinberg; EXECUTIVE ASSISTANTAlison Crawford; ADMINISTRATIVE SUPPORT Maryrose Madrid; EDITORIALFELLOW Melissa R. McCartneyEDITORIAL DIRECTOR, WEB AND NEW MEDIA Stewart Wills; SENIOR WEBEDITOR Tara S. Marathe; WEB EDITOR Robert Frederick; RESEARCH ASSOCIATECorinna Cohn; WEB DEVELOPMENT MANAGER Martyn Green; WEB DEVELOPERAndrew Whitesell; INTERN Sophia CaiNEWS DEPUTY NEWS EDITORS Robert Coontz, David Grimm (Online), EliotMarshall, Jeffrey Mervis, Leslie Roberts; CONTRIBUTING EDITORS ElizabethCulotta, Polly Shulman; NEWS WRITERS Yudhijit Bhattacharjee, Adrian Cho,Jennifer Couzin, Jocelyn Kaiser, Richard A. Kerr, Eli Kintisch, GregMiller, Elizabeth Pennisi, Lauren Schenkman, Robert F. Service(Pacifi c NW), Erik Stokstad; WEB DEVELOPER Daniel Berger; INTERN SaraReardon; CONTRIBUTING CORRESPONDENTS Jon Cohen (San Diego, CA),Daniel Ferber, Ann Gibbons, Sam Kean, Andrew Lawler, MitchLeslie, Charles C. Mann, Virginia Morell, Gary Taubes; COPYEDITORS Linda B. Felaco, Melvin Gatling, Melissa Raimondi; ADMINISTRATIVESUPPORT Scherraine Mack; BUREAUS San Diego, CA: 760-942-3252,FAX 760-942-4979; Pacific Northwest: 503-963-1940PRODUCTION DIRECTOR Wendy K. Shank; ASSISTANT MANAGER RebeccaDoshi; SENIOR SPECIALISTS Steve Forrester, Chris Redwood, AnthonyRosen; PREFLIGHT DIRECTOR David M. Tompkins; MANAGER Marcus Spiegler;SPECIALIST Jason HillmanART DIRECTOR Yael Fitzpatrick; ASSOCIATE ART DIRECTOR Laura Creveling;SENIOR ILLUSTRATORS Chris Bickel, Katharine Sutliff; ILLUSTRATOR YanaHammond; SENIOR ART ASSOCIATES Holly Bishop, Preston Huey,Nayomi Kevitiyagala; ART ASSOCIATES Kay Engman, Matthew Twombly;PHOTO EDITOR Leslie BlizardSCIENCE INTERNATIONALEUROPE (science@science-int.co.uk) EDITORIAL: INTERNATIONAL MANAGINGEDITOR Andrew M. Sugden; SENIOR EDITOR/COMMENTARY Julia Fahrenkamp-Uppenbrink; SENIOR EDITORS Caroline Ash, Stella M. Hurtley, Ian S.Osborne, Peter Stern; ASSOCIATE EDITOR Maria Cruz; LOCUM EDITOR HelenPickersgill; EDITORIAL SUPPORT Samantha Hogg, Alice Whaley;ADMINISTRATIVE SUPPORT John Cannell, Janet Clements, LouiseHartwell; NEWS: EUROPE NEWS EDITOR John Travis; DEPUTY NEWS EDITOR DanielClery; CONTRIBUTINGCORRESPONDENTS Michael Balter (Paris), JohnBohannon (Vienna), Martin Enserink (Amsterdam and Paris), GretchenVogel (Berlin); INTERN Jennifer CarpenterLATIN AMERICA CONTRIBUTING CORRESPONDENT Antonio RegaladoASIA Japan Office: Asca Corporation, Tomoko Furusawa, Rustic Bldg. 7F,77 Tenjin-cho, Shinjuku-ku, Tokyo 162-0808, Japan; +81 36802 4616, FAX +81 3 6802 4615, inquiry@sciencemag.jp;ASIA NEWS EDITOR Richard Stone (Beijing: rstone@aaas.org); CONTRIBUTINGCORRESPONDENTS Dennis Normile [Japan: +81 (0) 3 3391 0630, FAX+81 (0) 3 5936 3531; dnormile@gol.com]; Hao Xin [China:cindyhao@gmail.com]; Pallava Bagla [South Asia: +91 (0) 11 22712896; pbagla@vsnl.com]EXECUTIVE PUBLISHER Alan I. LeshnerPUBLISHER Beth RosnerFULFILLMENT SYSTEMS AND OPERATIONS (membership@aaas.org);DIRECTOR Waylon Butler; CUSTOMER SERVICE SUPERVISOR Pat Butler;SPECIALISTS Latoya Casteel, LaVonda Crawford, Vicki Linton,April Marshall; DATA ENTRY SUPERVISOR Cynthia Johnson; SPECIALISTSShirlene Hall, Tarrika Hill, William JonesBUSINESS OPERATIONS AND ADMINISTRATION DIRECTOR Deborah Rivera-Wienhold; BUSINESS SYSTEMS AND FINANCIAL ANALYSIS DIRECTOR Randy Yi; MANAGER,BUSINESS ANALYSIS Eric Knott; MANAGER, BUSINESS OPERATIONS Jessica Tierney;FINANCIAL ANALYSTS Priti Pamnani, Celeste Troxler; RIGHTS AND PERMISSIONS:ADMINISTRATOR Emilie David; ASSOCIATE Elizabeth Sandler; MARKETING DIRECTORIan King; MARKETING MANAGERS Allison Pritchard, Alison Chandler, JulianneWielga; MARKETING ASSOCIATES Aimee Aponte, Mary Ellen Crowley, WendyWise; SENIOR MARKETING EXECUTIVE Jennifer Reeves; DIRECTOR, SITE LICENSINGTom Ryan; DIRECTOR, CORPORATE RELATIONS Eileen Bernadette Moran; PUBLISHERRELATIONS, eRESOURCES SPECIALIST Kiki Forsythe; SENIOR PUBLISHER RELATIONSSPECIALIST Catherine Holland; PUBLISHER RELATIONS, EAST COAST Phillip Smith;PUBLISHER RELATIONS, WEST COAST Philip Tsolakidis; FULFILLMENT SUPERVISORIquo Edim; FULFILLMENT COORDINATOR Carrie MacDonald, Destiny Pinson; MAR-KETING MANAGER Christina Schlecht; MARKETING ASSOCIATE Laura Tutino; ELECTRONICMEDIA: MANAGER Lizabeth Harman; PROJECT MANAGER Trista Snyder;ASSISTANT MANAGER Lisa Stanford; SENIOR PRODUCTION SPECIALISTS Ryan Atkins,Christopher Coleman, COMPUTER SPECIALIST Walter Jones, Kai Zhang;PRODUCTION SPECIALISTS Antoinette Hodal, Nichele Johnston, Kimberly Oster;DIRECTOR, WEB AND NEW MEDIA Will CollinsADVERTISING DIRECTOR, WORLDWIDE AD SALES Bill MoranCOMMERCIAL EDITOR Sean Sanders: 202-326-6430ASSISTANT COMMERCIAL EDITOR Tianna Hicklin 202-326-6463P RODUCT (science_advertising@aaas.org); MIDWEST RickBongiovanni: 330-405-7080, FAX 330-405-7081; EAST COAST/E. CANADA Laurie Faraday: 508-747-9395, FAX 617-507-8189;WEST COAST/W. CANADA Lynne Stickrod: 415-931-9782, FAX 415-520-6940; UK/EUROPE/ASIA Roger Goncalves: TEL/FAX +41 43 243 1358;JAPAN ASCA Corporation, Nanako Ide +81 (0) 3 6802 4616, FAX+81 (0) 3 6802 4615; ads@sciencemag.jp; SENIOR TRAFFIC ASSOCIATEDeiandra SimmsWORLDWIDE ASSOCIATE DIRECTOR OF SCIENCE CAREERS Tracy Holmes: +44 (0)1223 326525, FAX +44 (0) 1223 326532CLASSIFIED (advertise@sciencecareers.org); U.S.: MIDWEST/WEST COAST/SOUTH CENTRAL/CANADA Tina Burks: 202-326-6577; EAST COAST/INDUSTRYElizabeth Early: 202-326-6578; SALES ADMINISTRATOR: Marci GallunSALES COORDINATORS Rohan Edmonson, Shirley Young; EUROPE/ROW SALES: Susanne Kharraz, Dan Pennington, Alex Palmer; SALESASSISTANT Lisa Patterson; JAPAN ASCA Corporation, Jie Chin +81 (0) 36802 4616, FAX +81 (0) 3 6802 4615; careerads@sciencemag.jp;ADVERTISING SUPPORT MANAGER Karen Foote: 202-326-6740; ADVERTISINGPRODUCTION OPERATIONS MANAGER Deborah Tompkins; SENIOR PRODUCTIONSPECIALIST/GRAPHIC DESIGNER Amy Hardcastle; PRODUCTION SPECIALISTYuse Lajiminmuhip; SENIOR TRAFFIC ASSOCIATE Christine HallAAAS BOARD OF DIRECTORS RETIRING PRESIDENT, CHAIR Peter C. Agre;PRESIDENT Alice Huang; PRESIDENT-ELECT Nina Fedoroff; TREASURER DavidE. Shaw; CHIEF EXECUTIVE OFFICER Alan I. Leshner; BOARD Linda P. B.Katehi, Nancy Knowlton, Stephen Mayo, Cherry A. Murray, Julia M.Phillips, Sue V. Rosser, David D. Sabatini, Thomas A. WoolseySENIOR EDITORIAL BOARDCori Bargmann, The Rockefeller Univ.John I. Brauman, Chair, Stanford Univ.Richard Losick, Harvard Univ.Michael S. Turner, University of ChicagoBOARD OF REVIEWING EDITORSAdriano Aguzzi, Univ. Hospital ZürichTakuzo Aida, Univ. of Tokyo.Sonia Altizer, Univ. of GeorgiaDavid Altshuler, Broad InstituteRichard Amasino, Univ. of Wisconsin, MadisonAngelika Amon, MITKathryn Anderson, Memorial Sloan-Kettering Cancer CenterSiv G. E. Andersson, Uppsala Univ.Peter Andolfatto, Princeton Univ.Meinrat O. Andreae, Max Planck Inst., MainzJohn A. Bargh, Yale Univ.Ben Barres, Stanford Medical SchoolMarisa Bartolomei, Univ. of Penn. School of Med.Jordi Bascompte, Estación Biológica de Doñana, CSICFacundo Batista, London Research Inst.Ray H. Baughman, Univ. of Texas, DallasDavid Baum, Univ. of WisconsinYasmine Belkaid, NIAID, NIHStephen J. Benkovic, Penn State Univ.Gregory C. Beroza, Stanford Univ.Ton Bisseling, Wageningen Univ.Mina Bissell, Lawrence Berkeley National LabPeer Bork, EMBLPaul M. Brakefield, Leiden Univ.Christian Büchel, Universitätsklinikum Hamburg-EppendorfJoseph A. Burns, Cornell Univ.William P. Butz, Population Reference BureauGyorgy Buzsaki, Rutgers Univ.Mats Carlsson, Univ. of OsloMildred Cho, Stanford Univ.David Clapham, Children’s Hospital, BostonDavid Clary, Oxford UniversityJ. M. Claverie, CNRS, MarseilleJonathan D. Cohen, Princeton Univ.Andrew Cossins, Univ. of LiverpoolAlan Cowman, Walter & Eliza Hall Inst.Robert H. Crabtree, Yale Univ.Wolfgang Cramer, Potsdam Inst. for Climate Impact ResearchF. Fleming Crim, Univ. of WisconsinJeff L. Dangl, Univ. of North CarolinaTom Daniel, Univ. of WashingtonStanislas Dehaene, Collège de FranceEmmanouil T. Dermitzakis, Univ. of Geneva Medical SchoolRobert Desimone, MITClaude Desplan, New York Univ.Ap Dijksterhuis, Radboud Univ. of NijmegenDennis Discher, Univ. of PennsylvaniaScott C. Doney, Woods Hole Oceanographic Inst.Jennifer A. Doudna, Univ. of California, BerkeleyJulian Downward, Cancer Research UKBruce Dunn, Univ. of California, Los AngelesChristopher Dye, WHOMichael B. Elowitz, Calif. Inst. of TechnologyGerhard Ertl, Fritz-Haber-Institut, BerlinBarry Everitt, Univ. of CambridgePaul G. Falkowski, Rutgers Univ.Ernst Fehr, Univ. of ZurichTom Fenchel, Univ. of CopenhagenAlain Fischer, INSERMWulfram Gerstner, EPFL LausanneKarl-Heinz Glassmeier, Inst. for Geophysics &Extraterrestrial PhysicsDiane Griffin, Johns Hopkins Bloomberg School ofPublic HealthChristian Haass, Ludwig Maximilians Univ.Steven Hahn, Fred Hutchinson Cancer Research CenterGregory J. Hannon, Cold Spring Harbor Lab.Dennis L. Hartmann, Univ. of WashingtonChris Hawkesworth, Univ. of St AndrewsMartin Heimann, Max Planck Inst., JenaJames A. Hendler, Rensselaer Polytechnic Inst.Janet G. Hering, Swiss Fed. Inst. of AquaticScience & TechnologyRay Hilborn, Univ. of WashingtonMichael E. Himmel, National Renewable Energy Lab.Kei Hirose, Tokyo Inst. of TechnologyOve Hoegh-Guldberg, Univ. of QueenslandDavid Holden, Imperial CollegeLora Hooper, UT Southwestern Medical Ctr at DallasJeffrey A. Hubbell, EPFL LausanneSteven Jacobsen, Univ. of California, Los AngelesPeter Jonas, Universität FreiburgBarbara B. Kahn, Harvard Medical SchoolDaniel Kahne, Harvard Univ.Bernhard Keimer, Max Planck Inst., StuttgartRobert Kingston, Harvard Medical SchoolHanna Kokko, Univ. of HelsinkiAlberto R. Kornblihtt, Univ. of Buenos AiresLeonid Kruglyak, Princeton Univ.Lee Kump, Penn State Univ.Mitchell A. Lazar, Univ. of PennsylvaniaDavid Lazer, Harvard Univ.Virginia Lee, Univ. of PennsylvaniaOttoline Leyser, Univ. of New YorkOlle Lindvall, Univ. Hospital, LundMarcia C. Linn, Univ. of California, BerkeleyJohn Lis, Cornell Univ.Richard Losick, Harvard Univ.Jonathan Losos, Harvard Univ.Ke Lu, Chinese Acad. of SciencesLaura Machesky, CRUK Beatson Inst. for Cancer ResearchAndrew P. MacKenzie, Univ. of St AndrewsAnne Magurran, Univ. of St AndrewsOscar Marin, CSIC & Univ. Miguel HernándezCharles Marshall, Univ. of California, BerkeleyMartin M. Matzuk, Baylor College of MedicineGrahma Medley, Univ. of WarwickYasushi Miyashita, Univ. of TokyoRichard Morris, Univ. of EdinburghEdvard Moser, Norwegian Univ. of Science and TechnologySean Munro, MRC Lab. of Molecular BiologyNaoto Nagaosa, Univ. of TokyoJames Nelson, Stanford Univ. School of Med.Timothy W. Nilsen, Case Western Reserve Univ.Pär Nordlund, Karolinska Inst.Helga Nowotny, European Research Advisory BoardStuart H. Orkin, Dana-Farber Cancer Inst.Christine Ortiz, MITElinor Ostrom, Indiana Univ.Andrew Oswald, Univ. of WarwickJonathan T. Overpeck, Univ. of ArizonaP. David Pearson, Univ. of California, BerkeleyReginald M. Penner, Univ. of California, IrvineJohn H. J. Petrini, Memorial Sloan-Kettering Cancer CenterSimon Phillpot, Univ. of FloridaPhilippe Poulin, CNRSColin Renfrew, Univ. of CambridgeTrevor Robbins, Univ. of CambridgeBarbara A. Romanowicz, Univ. of California, BerkeleyJens Rostrup-Nielsen, Haldor TopsoeEdward M. Rubin, Lawrence Berkeley National LabShimon Sakaguchi, Kyoto Univ.Jürgen Sandkühler, Medical Univ. of ViennaRandy Seeley, Univ. of CincinnatiChristine Seidman, Harvard Medical SchoolVladimir Shalaev, Purdue Univ.Joseph Silk, Univ. of OxfordDavor Solter, Inst. of Medical Biology, SingaporeAllan C. Spradling, Carnegie Institution of WashingtonJonathan Sprent, Garvan Inst. of Medical ResearchElsbeth Stern, ETH ZürichYoshiko Takahashi, Nara Inst. of Science and TechnologyJurg Tschopp, Univ. of LausanneHerbert Virgin, Washington Univ.Bert Vogelstein, Johns Hopkins Univ.Cynthia Volkert, Univ. of GottingenBruce D. Walker, Harvard Medical SchoolIan Walmsley, Univ. of OxfordChristopher A. Walsh, Harvard Medical SchoolDavid A. Wardle, Swedish Univ. of Agric SciencesColin Watts, Univ. of DundeeDetlef Weigel, Max Planck Inst., TübingenJonathan Weissman, Univ. of California, San FranciscoSue Wessler, Univ. of GeorgiaIan A. Wilson, The Scripps Res. Inst.Timothy D. Wilson, Univ. of VirginiaXiaoliang Sunney Xie, Harvard Univ.John R. Yates III, The Scripps Res. Inst.Jan Zaanen, Leiden Univ.Mayana Zatz, University of Sao PaoloJonathan Zehr, Ocean SciencesHuda Zoghbi, Baylor College of MedicineMaria Zuber, MITBOOK REVIEW BOARDJohn Aldrich, Duke Univ.David Bloom, Harvard Univ.Angela Creager, Princeton Univ.Richard Shweder, Univ. of ChicagoEd Wasserman, DuPontLewis Wolpert, Univ. College London380 28 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

1,000sOF GRANTSMILLIONSIN FUNDINGGrantsNet.The first comprehensivescience grants database.GrantsNet is expanding its listings ofsome 900 funding programs from privatefoundations and not-for-profit organizationsto include 400 to 500 new entries from thegrants.gov site. This provides the firstcomprehensive database of fundingopportunities to research scientists andadministrators, career counselors, financialaid specialists, and undergraduate andgraduate students. For listings, go towww.grantsnet.orgNational Academy of SciencesCongratulates theRECIPIENTS OF ITS2O11 AWARDSSince 1886 the National Academy of Sciences has presentedawards to honor outstanding contributions to science and toencourage new and innovative research.NAS Public Welfare MedalIsmail SerageldinBibliotheca AlexandrinaArctowski Medal (Solar Physics)John W. HarveyNational Solar ObservatoryArthur L. Day Prize and Lectureship (Physics of the Earth)R. Lawrence EdwardsUniversity of MinnesotaRichard Lounsbery Award (Biology and Medicine)Bonnie L. BasslerPrinceton UniversityNAS Award in Chemical SciencesStephen J. BenkovicPennsylvania State UniversityNAS Award for Chemistry in Service to SocietyPaul J. ReiderAmgen, Inc.NAS Award for the Industrial Application of Science(Agricultural Science)H. Boyd WoodruffSoil Microbiology Associates, Inc.NAS Award for Initiatives in Research (Optical Science)Martin T. ZanniUniversity of Wisconsin-MadisonNAS Award in Molecular BiologyJames M. BergerUniversity of California, BerkeleyNAS Award for Scientific Reviewing (Economics)Thomas J. SargentNew York UniversityTroland Research Awards (Psychology)Elizabeth A. BuffaloEmory University School of MedicineJoshua B. TenenbaumMassachusetts Institute of TechnologySelman A. Waksman Award in MicrobiologyCarol A. GrossUniversity of California, San FranciscoNominations for the 2012 awards will be accepted throughSeptember 15, 2011. For details visit:www.nasonline.org/awards

NEWS OF THE WEEKCairo, Egypt 4Dead Queen Still Kicking Up Dust652, 7AROUND THE WORLDBasel, Switzerland 1Melanoma Drug Extends LifeThe drug company Roche announced thatan experimental drug increased the survivalof patients with advanced skin cancerwhose tumors had a specific mutation in agene called B-RAF. RG7204, developed bythe biotech company Plexxikon, targets aprotein encoded by the mutated gene thatdrives cancer growth in about half of allpatients with malignant melanoma. About40,000 people worldwide die annually fromthis aggressive cancer, most less than a yearafter being diagnosed.In the closely watched phase III trial,patients with the B-RAF mutation whoreceived the drug lived significantly longerand their tumors grew more slowlycompared with patients who received astandard chemotherapy drug, Roche officialsannounced in a press release. But thecompany did not disclose how much longerpatients lived on average. Full results willbe reported later this year at a meeting.Patients receiving the standard drug willnow be offered the option of switching toRG7204.Washington, D.C. 2JPL Scientists SubjectTo Background ChecksThe U.S. Supreme Court has upheld the useof background checks by the governmenton scientists and other workers at NASA’sJet Propulsion Laboratory (JPL). The decisioncaps a legal battle begun in 2007, when28 scientists and engineers at JPL, which isowned by NASA but operated by the CaliforniaInstitute of Technology in Pasadena,filed suit against new screening procedures14announced by NASA. They argued thatquestionnaires asking about drug use, counseling,and “trustworthiness” were too intrusiveand harmed the open environment atJPL. As contract employees, they said, theyshould be subject to less scrutiny than governmentemployees who work with classifiedmaterial. An appeals court agreed, but theSupreme Court unanimously disagreed.NASA will now have to decide whetherto reinstate the policy, which has been onhold since 2007. http://scim.ag/NASA-checksTokyo, Japan 3Solar Sail Mission Sails OnThe IKAROS solar sailmission was formallyextended, having passedall performance testsplanned for its original6-month life. Launchedby the Japan AerospaceExploration Agency(JAXA) last year on 21 May (Science, 7 May2010, p. 677), IKAROS, with its 20-meterdiagonal, 0.0075-millimeter-thick polyimidesail, became the first craft propelled throughspace by the pressure of photons in sunlight.“We achieved complete success,” says missionmanager Osamu Mori of JAXA’s Instituteof Space and Astronautical Science inSagamihara, near Tokyo. JAXA will fundthe mission for another year so scientists cantry to pull off some advanced navigationaltricks, such as varying the sail’s angle to thesun. Such experiments will help in planningmore ambitious solar sail missions,says Mori. Separately, NASA announced lastweek that its solar sail mission, NanoSail-D,which was thought to be malfunctioning,successfully deployed its sail and is workingas planned.3Egypt’s SupremeCouncil of Antiquitiesdemanded the returnof a famous bust ofQueen Nefertiti thathas been in Berlin’sNeues Museum sinceits discovery in 1912 byGerman archaeologists.Egypt has been tryingto get her back sincethe 1920s—AdolfHitler refused to sendher back in the 1930s—but now councilchief Zahi Hawass has made the request inwriting to the Prussian Cultural HeritageFoundation, which oversees the museum.Egypt maintains that the discoverer misledEgyptian authorities after its discovery, butthe initial German reaction was dismissive.Galápagos Islands, Ecuador 5Fight Rages On AgainstInvasive RatsConservationists have stepped up their waragainst alien rats in the Galápagos. Officialswith Ecuador’s Galápagos National Parkannounced that with conservationists fromvarious nonprofit organizations, they hadbegun carpet-bombing the archipelago’ssmaller islands with rat poison systematicallyreleased from a helicopter. Rats firstarrived as stowaways in Western sailingships and are a problem because they eatnative tortoise and bird eggs. Although conservationistshave been killing rats for yearsusing bait and traps, the new strategy aimsfor “100% eradication” from nine islandsand islets, including Jervis and Beagleislands, the officials announced in a statement.To protect a native bird that mightNOTED>The Kafkaesque bureaucracy of theFramework Programme, the EuropeanUnion’s multibillion-euro researchfunding system, is getting simpler. Aspecific E.U.-approved accounting systemwill no longer be required, smallbusiness partners can now be paid aflat rate, and multiple sets of rulesand procedures will be eliminated.http://scim.ag/no-kafkaCREDITS (TOP TO BOTTOM): WIKIMEDIA COMMONS; © JAXA38228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSCREDITS (TOP TO BOTTOM): JAPAN PRIZE FOUNDATION (4); IIKKA HANSKI/UNIVERSITY OF HELSINKI; (SOURCE) ERCotherwise eat the poisoned rats, conservationistscaptured 20 Galápagos hawks andplan to keep them in captivity for 2 months.They said that “mitigation steps will betaken” to protect the sole endemic rodent, amouse found on Santiago Island.Houston, Texas 6Cancer Center ReceivesBig Thank-You GiftWhen MD Anderson Cancer Center inHouston, Texas, admitted some cancerpatients from the United Arab Emirates(UAE) for treatment, officials there hadno idea it would lead to the largest gift inthe hospital’s history. Now, the Khalifabin Zayed Al Nahyan Charity Foundation,formed by (and named after) the UAE’s president,has announced it is giving MD Anderson$150 million. The money will be usedfor a new building and work in pancreaticcancer and personalized medicine.Some $25 million will go to study biomarkers,genetic patterns in tumors thatcould guide treatment for individual patients.“There’s been a lot of stops and starts” in thebiomarkers field, with many not panning out,admits Raymond DuBois, the cancer center’sprovost and executive vice president. “We’reat a critical stage now,” and this infusion ofmoney, he says, will surely help.Washington, D.C. 7A Call for Stem CellResearchers to ShareResearchers working with stem cellsshould follow the example of their colleaguesin genetic sequencing and clinicalresearch, setting up global networksfor sharing data, materials, and intellectualproperty, according to a report releasedby an international consortium on stemcells and ethics. The Hinxton Group recommendedsetting up a publicly availableglobal stem cell registry that would includea cell line’s characteristics and informationon how it was derived. Stem cell banks andcell repositories should be expanded andshould coordinate their work, the reportsays, and funding agencies and journalsshould make data and material sharingmandatory. A database of stem cell–relatedpatents is also urgently needed, the groupsays, to help scientists deal with the thornythicket of intellectual property that hasgrown along with the hot field.http://scim.ag/hinxton-groupNEWSMAKERSJapan PrizesTOKYO—A breakthrough drug and acomputer operating system shared thelaurels of one of science’s top honors,the Japan Prize.The Bioscienceand Medical Scienceprize goesto TadamitsuKishimoto andToshio Hirano, bothof Osaka University,for identifying theprotein interleukinClockwise from topleft: Kishimoto, HiranoRitchie, and Thompson.Number of grantees5550454035302520151050United KingdomGermanyFranceERC GRANTEES2010 • 26 NationalitiesMaleFemaleItalyIsraelNetherlands6, elucidating itsimmune functions,and determining itsrole in rheumatoidarthritis. The pairultimately helped develop a drug that treatsthe debilitating disorder by blocking theprotein’s activity.Kenneth Thompson and Dennis Ritchietake home the Information and Communicationsprize for developing the UNIXcomputer operating system in the 1960s and’70s while at Bell Laboratories in MurrayHill, New Jersey. UNIX set new standardsfor simplicity and ease of adapting softwareto different computing platforms. Thesource code was also distributed to users sothat they could contribute improvements,marking the beginning of the open systemsconcept. Thompson is now at Google Inc.;Ritchie is retired.The winners in each category will share$600,000 and receive commemorativemedals at an April ceremony here.http://scim.ag/Japan-prizeSpainSwitzerlandSwedenUnited StatesNationality of granteesTHEY SAID ITThis week, our Facebook fans chimed in ona story of a dino with just one finger (seep. 384). Some of our favorite comments:“ Thanks for flipping us theprehistoric bird!”“ It’s ok as long as the dinodoesn’t have to do mathhomework.”“ And that was how the binarysystem was created, right?”http://www.facebook.com/ScienceNOWFinn Bags Swedish PrizeThis year’s Crafoord Prize has gone to Finnishecologist Ilkka Hanski of the Universityof Helsinki for his contributions to understandingthe impact of habitat fragmentationon species’ survival. Hanski has spent muchof his 30-year career assessing the risk oflocal extinctions in environments subjectto growing human influence. The prize,awarded annually by the Royal >>ERC Still Looking for WomenThe European Research Council (ERC) announcedthat 266 “advanced scientists” won its latest roundof funding, with British and German researchersgrabbing the most grants. But just 9.4% of thewinners are women, an imbalance seen in previousERC funding rounds that is a source of continuingfrustration to the organization.DenmarkNorwayBelgiumFinlandAustriaHungaryGreecePortugalIrelandRussian FederationBulgariaCanadaCzech Republicwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 383MaltaNew ZealandTunisia

NEWS OF THE WEEK>>NEWSMAKERSSwedish Academy of Sciences to researchersin fields not covered by the Nobels,recognizes Hanski for developing modelsto help conservationists manage naturalenvironments.Hanski says he has not yet decidedwhat he will do with the 4 million kronor(approximately $600,000) that come withthe prize, but he suspects that he will buya plot of forest in Finland to save it fromfuture development. http://scim.ag/crafoordCanadian-Born Geologist Tappedfor Key E.U. Research PostExperimental volcanologist Donald BruceDingwell was born and raised in Canada, buthe may soon be playing a big role in Europeanresearch funding. Science has learnedthat he has been chosen as the next secretarygeneral of the European Research Council(ERC), the E.U.’s funding agency for individualbasic researchers. But negotiationsabout his contract are still ongoing, cautionsERC President Helga Nowotny.Dingwell is currently director of theDepartment of Earth and EnvironmentalSciences at Ludwig Maximilians UniversityMunich in Germany. As secretary general, hewould be a liaison to ERC’s executive agency,a managing body ultimately controlled by theEuropean Commission, on behalf of ERC’sscientific council, which sets strategy. ERCpreviously decided to scrap the position ofsecretary general but reversed that decisionpending an overhaul of its organizationalstructure. http://scim.ag/dingwellFINDINGSStar Light, Star BrightThis infrared image from NASA’s SpitzerSpace Telescope shows what may be the mostluminous group of superstars in the entiregalaxy. Such groups—named OB associationsfor the O and B spectral types of theirhot, blue suns—sculpt vast regions of spacethrough their radiation and supernova explosions.The newly discovered Dragonfish associationin the Southern Cross is 32,000 lightyearsdistant and harbors about 400 hot, blue,luminous stars, Mubdi Rahman of the Universityof Toronto in Canada and colleagueswill report in an upcoming issue of The AstrophysicalJournal Letters. The stars’ extremeultraviolet radiation strips electrons from protons,thereby ionizing interstellar hydrogengas and setting it aglow.Let’s Stay TogetherParting is such sweet sorrow that Sumatranand Bornean orangutans may have separatedinto distinct species more than half a millionyears later than previously assumed.Researchers have completed a draftsequence of the orangutan genome derivedSingle-Digit Dinofrom 11 individualson the islandsof Sumatra andBorneo, the onlyplaces where ourendangered, orangehairedrelatives livein the wild. A comparisonof the twospecies’ DNA suggeststhey separatedjust 400,000 yearsago, revising previousestimates of atleast 1 million yearsago. (They were physically separated atleast 21,000 years ago, when land bridgesbetween the two islands disappeared.) “Mostprevious studies used small sets of markersand a limited amount of DNA sequence,”says Devin Locke, a structural geneticist atWashington University School of Medicinein St. Louis, Missouri, and the lead author ofthe study, which appears online this weekin Nature.Orangutans lead a more sedentary lifestylethan other great apes, and their DNAsuggests they may evolve more slowly,too. Key drivers of evolution are stretchesof DNA called retrotransposons that jumparound the genome, creating new genes oraltering regulation of existing ones. The newdata reveal that retrotransposons known asAlu elements have moved around the orangutangenome much less than they have in thehuman and chimpanzee genomes (the onlyother two great apes to have been sequenced).http://scim.ag/orang-genesMeat-eating dinosaurs were very good at finding food, thus their evolutionarysuccess over some 165 million years. But during their time on Earth, theykept losing something that might seem important: their fingers. The earliestcarnivorous dinosaurs had five fingers, although only four were actuallyfunctional. Many later meat eaters had three, and evolution left the mightyTyrannosaurus rex with only two. Now researchers have unearthed the firstknown dinosaur with only one finger. The new single-digit species, namedLinhenykus monodactylus, was found in a roughly 80-million-year-old rockformation in Inner Mongolia, lead author Xing Xu of the Chinese Academyof Sciences Institute of Vertebrate Paleontology and Paleoanthropologyin Beijing and colleagues report online in the Proceedings of the NationalAcademy of Sciences.Linhenykus, which was probably about a meter tall, belongs to a familyof dinosaurs called alvarezsauroids. The team suggests that the single clawlikedigit was an adaptation for digging, perhaps for insects such as termites.CREDITS (TOP TO BOTTOM): MUBDI RAHMAN ET AL./SPITZER SPACE TELESCOPE GLIMPSE SURVEY; THINKSTOCK; © JULIUS T. CSOTONYI (WWW.CSOTONYI.COM)38428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSCREDITS (TOP TO BOTTOM): ALICE CHUNG AND INGMAR RIEDEL-KRUSE; SCOTT SOLOMONBY THE NUMBERS40% The percentage of U.S. highschool seniors who scored below thebasic level of achievement in scienceon the 2009 National Assessment ofEducational Progress. Only 1% of12th-graders scored at the advancedlevel on the test, a quadrennial measureof reading, math, and sciencethat is known as the Nation’s ReportCard (http://scim.ag/NAEP-2009)12 million The average extent,in square kilometers, of Arctic seaicecover during December 2010,according to the World MeteorologicalOrganization. That makes itthe lowest on record for December,about 1.35 million square kilo metersbelow the 1979–2000 average.12% The percentage of first-yearU.S. college students who, whenasked about seven common types ofquantitative reasoning activities, saidthey had never done any of them,according to the 2010 NationalSurvey of Student Engagement.Frédéric Chopin’s‘Madness’ DiagnosedIn 1848, Polish composer and piano virtuosoFrédéric Chopin was performing at a housein Paris when he suddenly stopped in themiddle of a piece and left the stage. He laterwrote to a friend that he had seen creaturescrawling out of his piano. Chopin is widelyviewed as a tortured artist, but a new papersuggests his eccentricities might have beendue to epilepsy.Radiologist Manuel Vásquez Carunchoand neurologist Francisco Brañas Fernándezof Xeral-Calde Hospital in Lugo, Spain,studied Chopin’s writings and that of friendsand pupils for descriptions of his hallucinationsand wild behavior. Only a handful ofneurological disorders produce the phantasmagoriathat tormented Chopin, who didn’tabuse drugs or alcohol. The authors rule outschizophrenia and other common psychosesbecause Chopin’s hallucinations were visual,Random SamplePetri ArcadeLong before he became a bioengineer,Ingmar Riedel-Kruse wasa typical 12-year-old video gameafi cionado. “I wasted a year of mylife doing that—’wasted’ in a goodsense,” he recalls. He soon movedon from playing with computers toprogramming them and, eventually,to biotechnology research. About ayear ago, while reading a Wikipedia article about the history of video games, he had an inspiration:Why not start gaming with actual, live microbes?After “lots of trial and error,” Riedel-Kruse says, he and his colleagues at Stanford Universitydeveloped a basic game console that nudges paramecia around a microfluidic chamberwith chemical gradients or mild electric fields. A microscope camera pipes images of thewriggling protozoans into a laptop game window. Games include soccerlike “Ciliaball”; “Pac-mecium,” in which paramecia gobble virtual yeast dots while avoiding a cartoon zebra fishlarva; and “Pond Pong,” in which two players bat the microbes back and forth by releasingchemicals from a needle tip. Descriptions of the whole microarcade, including games involvingyeast and DNA, appear this month in the journal Lab on a Chip.Riedel-Kruse says putting biotic games on the Internet could give researchers a way tocrowdsource real-time biology experiments to online players. He’s also working on bringingthem into schools to inspire future scientists—after all, it worked for him.not auditory, and because he lacked othertelltale symptoms such as eye problems ormigraines. His short hallucinatory episodesare a hallmark of temporal lobe epilepsy, theteam reports online in Medical Humanities.Other researchers call the proposal interestingbut perhaps too subtle. The authorsthemselves admit it is difficult to be conclusivewithout the ability to observe Chopinhimself. However, Caruncho points out, testimoniesfrom witnesses are key in diagnosingepilepsy even today.http://scim.ag/mad-chopinThe World’s Smallest FarmersIt’s too bad they don’t make microscopicoveralls. A study published online last weekin Nature finds that the single-celled organismDictyostelium discoideum harvests bacterialike farmers harvest crops.An individual D. discoideum, or “Dicty,”amoeba cell can live independently, slurpingup bacteria in the soil. When thefood is gone, it joins with its comrades toform a tiny sluglike organism that wrigglesto greener pastures. Once there, theslug becomes a stalk with a fruiting body(pictured)—a tiny globe on top that releasesspores, each spawning a single amoeba.Debra Brock, a graduate student inecology and evolutionary biology at RiceUniversity in Houston, Texas, was studyingspores from wild Dicty amoebae whenshe saw something she’d never seen before:bacteria in the fruiting body. To find outwhether the bacteria were just an infection,she gave the spores antibiotics, then placedthem on a fresh patch of bacteria. The sporesthat had originally harbored bacteria pickedup the bugs again, indicating that they werecollecting bacteria. Other experimentsshowed that the amoebas “planted” their newenvironments with bacteria and harvestedthem. Several animals are known to farm;some ant species tend fungi, for example.But researchers say it’s surprising to find thebehavior in such a simple organism.http://scim.ag/tiny-farmerswww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 385

NEWS & ANALYSISMover and shaker. NIH Director Francis Collins sayshe is “not apologetic” about a plan to dismantle onecenter to create a new one.U.S. SCIENCE POLICYCollins Sparks Furor WithProposed NIH ReshufflingA plan to create a new center aimed at developingdrugs at the U.S. National Institutes ofHealth (NIH) has many biomedical researchersin an uproar. Since NIH Director FrancisCollins endorsed the plan in mid-December,it has drawn a flood of concerned and sometimesangry comments, mainly because creatinga new NIH center would entail breakingup another center. Questions last week froma congressional committee have cast uncertaintyover Collins’s plan to launch the newcenter by next October.Collins defended the proposal last week: “Iwill not be apologetic for wanting to see thescientific opportunities that relate to translationapproached in a very bold way,” he toldScience. “And of course change is always distressing,especially if people aren’t quite surewhere it’s going. So I understand the anxietythat currently exists.”The reorganization grew out of a 7 December2010 recommendation by an NIH advisoryboard, called the Scientific ManagementReview Board (SMRB), to create a centeraimed at speeding the development of biomedicaldiscoveries into treatments (Science,10 December 2010, p. 1462). Dubbed theNational Center for Advancing TranslationalSciences (NCATS), it would fold together$632 million in NIH programs focused onclinical research and drug discovery anddevelopment that are now at other institutes. Itwould also house the Cures Acceleration Network,a drug development program created byCongress last year that, if fully funded, wouldbring NCATS’s budget up to $1.1 billion.The translational center has gotten a mixedresponse. Although some have questionedwhether drug development fits NIH’s mission,The New York Times gave it mostly positivecoverage last week. Other worries, however,stem from NIH’s decision to also disbandits National Center for Research Resources.NCRR’s largest program, the $490 millionClinical and Translational Science Awards,would move into NCATS. NIH officials wantto move the remaining 60% of NCRR, includinggrants for minority institutions and supportfor large shared instruments and animalmodels, to other institutes.Since mid-December, this proposed reorganizationhas triggered more than 1200comments on an NIH feedback site. Manyscientists worry that splitting up NCRRcomponents would be harmful—for example,putting the comparative medicine program’sprimate models in one institute andmodels such as zebrafish in another. Complaintscontinued last week when NIH DeputyDirector Lawrence Tabak posted onlinea “straw model” showing how the piecesof NCRR would be distributed. Some havea designated home, but most have beenassigned to an “interim infrastructure unit”in the NIH director’s office—which, as somepointed out, looks much like NCRR. Collinssays much of the interim unit would eventuallygo to the National Institute of GeneralMedical Sciences (NIGMS), NIH’s basicresearch institute.Critics within and outside NIH areincensed by how decisions have been made.“The SMRB process simply backed into aforegone conclusion,” and NCRR staff havereceived “an astounding level of … disrespect,”said one commenter on the feedbacksite. The feeling is that “Dr. Collinsis accustomed to getting what he wants,”says biochemist Mark Lively of Wake ForestUniversity School of Medicine in NorthCarolina, who is a member of the NCRRadvisory council.In last week’s interview with Science,Collins said statutory requirements compelledhim to notify Congress of his planto abolish NCRR before gathering publicinput. NIH’s parent organization, theDepartment of Health and Human Services(HHS), sent those letters on 14 January. Hesaid, “In my best of all worlds, I would havewanted … a lot more discussion.”Collins also said he wasn’t influenced by alaw limiting NIH to no more than the current27 institutes and centers. He could have askedCongress to allow NIH to exceed the cap temporarilyuntil it merges its alcoholism anddrug abuse institutes next year, but “I didn’teven ask.” Many NCRR programs fit morelogically in other institutes, he says.A task force co-chaired by Tabak plans tofinish the plan for NCRR by March so it canbe part of NIH’s 2012 budget request. Meanwhile,Congress has 180 days to object to thereorganization. Last week, John Bartrum,a staffer on the House of Representativesappropriations subcommittee that overseesHHS’s budget, sent HHS and NIH a long listof questions. Bartrum’s e-mail says “we havenot taken any position” on either NCATSor NCRR, but it asks for an explanation ofNCATS’s mission, details of its budget, andhow NIH decided to eliminate NCRR.One concern is that some of the programsNCATS would house are now funded by a potof money in the NIH director’s office for projectsthat are supposed to end after a few years.It’s not clear how NCATS would maintainfunding for them at a time when the overallNIH budget is expected to stay flat. But NIHofficials said that contrary to a statementin The New York Times story, NIH has “noplans to ‘cannibalize’ ” other parts of NIHto support NCATS. –JOCELYN KAISERCREDIT: GETTY IMAGES38628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSCREDITS: A. LAWLER/SCIENCE; HANS-PETER UERPMANNHUMAN EVOLUTIONDid Modern Humans Travel Out of Africa Via Arabia?JEBEL FAYA, UNITED ARAB EMIRATES—Thebarren desert and hills here seem whollyinhospitable, with sparse rain and sandy soilsupporting only a few nomadic Bedouin. Butthings were different 125,000 years ago, whenthe desert was savanna, with plentiful waterand game, and under the protection of a rockoverhang, a group of hominids whiled awaytheir time making stone tools. A Germanledteam argues on page 453 that these toolswere made by modern humans who mayhave crossed directly from Africa as part of amigration spreading across Europe, Asia, andAustralia. Although most researchers agreethat our species came out of Africa in one ormore waves (see p. 392), those dates are morethan 50,000 years earlier than most believeour ancestors left the continent.The audacious claim bySimon Armitage of Royal Holloway,University of London, andcolleagues is sparking interestand controversy. “This is reallyquite spectacular,” says archaeologistMichael Petraglia ofthe University of Oxford in theUnited Kingdom, who has previouslyargued that Homo sapiensleft Africa before the massiveeruption of an Indonesian volcano74,000 years ago, a catastrophethought to have left much ofAsia unlivable for early humans(Science, 5 March 2010, p. 1187).“It breaks the back of the currentconsensus view.” But others, suchas archaeologist Paul Mellarsof the University of Cambridgein the United Kingdom, say thatalthough the discovery is importantand well dated, the conclusionsare flawed. “I’m totallyunpersuaded,” he says. “There’s not a scrap ofevidence here that these were made by modernhumans, nor that they came from Africa.”The debate centers on a collection ofstone tools found here at Jebel Faya, a longlimestone mountain an hour’s drive from thebustling urban center of Sharjah and 55 kilometersfrom the Persian Gulf. A rock shelterindents the mountain’s end, a few metersabove a desolate plain where only camelsgraze today. The overhang is modest,but it has sheltered humans for millennia,say excavators Hans-Peter and MargaretheRocky road. Hans-Peter Uerpmannsurveys the Jebel Faya site, where earlyhumans left stone tools behind likethis one (inset).Uerpmann of the University of Tübingen inGermany. They began digging here in 2003,uncovering artifacts from the Iron, Bronze,and Neolithic periods before hitting materialfrom the Middle Paleolithic era, roughly300,000 to 30,000 years ago. Using singlegrainoptically stimulated luminescence,which measures how much time has passedsince materials were last exposed to light, theteam dated the oldest set of artifacts, includingstone hand axes, blades, and scrapers, toabout 125,000 years ago.Arabia and its fierce deserts havelong been seen more as obstaclesthan conduits to human migration,and most archaeology ogyhere has focused on historicaltimes. Recent studies, however, show wetterperiods such as one that began around130,000 years ago. And a spate of findingsin the past 25 years show that hominins werein the region during the Middle Paleolithic.Early H. sapiens skulls and tools from Skhuland Qafzeh caves in Israel are now dated to100,000 to 130,000 years ago, for example.Co-author Anthony Marks of SouthernMethodist University in Dallas, Texas, saysthe combination of artifacts from Jebel Faya,such as two-sided blades and small hand axes,is remarkably similar to assemblages madeduring this period in East Africa, when ourown species was the only known hominin onthat continent. Other hominins, such as theNeandertals who populated Europe and northAsia, did not use this combination of toolsand were not likely to have been in Arabia, hesays. That makes the African origin likely “byprocess of elimination.”Marks says the tools don’t resemblethose from Israel or the Aterian tools fromthe same era in North Africa (Science,7 January, p. 20). He suggests that H. sapiensmay have left Africa in different waves, withthe Arabian tools representing a migrationlaunched from East Africa.Petraglia agrees that it’s likely thatH. sapiens made the tools and that theycame from Africa. “This is out of the hab-itatrange of Neandertals,” he notes. “Sothey make a really strong and plausibleargument.” The team believesthat these early modern humansmay have even pushed on acrossthe Persian Gulf, perhaps to India,Indonesia, and eventually Australia.Petraglia claims evidence of earlyH. sapiens in India both before andafterthe Indonesian eruption, though othersdispute that assertion.Mellars, in contrast, sees no evidencethat the Jebel Faya artifacts are of an EastAfrican style. He says one of the bifacialsis stout rather than narrow like those commonin Africa and adds that the authors havenot ruled out Neandertals and even H. erectusas the toolmakers. “Everything hinges onwhether that material is explicitly African—and I don’t see that.”Other researchers are enthusiastic aboutthe Jebel Faya discovery but cautious aboutthe conclusions. Archaeologist Mark Beech, avisiting fellow at the University of York in theUnited Kingdom who has worked extensivelyin the United Arab Emirates, praises the paperbut adds: “One site does not confirm the outof-Africa-via-Arabiahypothesis.”Hans-Peter Uerpmann agrees, saying thatfossil bones are needed “before we can beabsolutely sure” that the tools were made byH. sapiens. Other researchers are hot on thetrail: Petraglia leaves this month to continuework in Saudi Arabia, and other archaeologistsplan to comb Arabian caves and sandsfor signs that our ancestors passed this way.–ANDREW LAWLERwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 387

NEWS&ANALYSISPointing fingers. Haitian protesters have accusedNepalese peacekeepers of introducing cholera intothe country.EPIDEMIOLOGYDespite Sensitivities, Scientists SeekTo Solve Haiti’s Cholera RiddleWhen French epidemiologist RenaudPiarroux came back from a 3-week missionto Haiti in late November, he faced adilemma. Piarroux, who had been invitedby the Haitian government to investigate thecountry’s explosive cholera outbreak, wasconvinced that the bacterium had been introducedby Nepalese soldiers taking part in theUnited Nations Stabilization Mission in Haiti(MINUSTAH)—and he wanted to present hisevidence. At the same time, he was leery ofexacerbating tensions in Haiti, where angrymobs had already demanded the departureof the Nepalese. Even after his confidentialreport for the Haitian government was leakedto the press, Piarroux didn’t talk to reportersuntil he got permission from French authorities.“I was very concerned,” he says.He’s not alone. Several cholera expertstold Science that nailing the source of the outbreakcould potentially embarrass the UnitedNations, distract from the day-to-day fight tocontrol the outbreak, and even lead to violence.So their passion for traditional shoeleatherepidemiology has been tempered bydiplomatic and strategic concerns.Indeed, prominent cholera scientistsdeclined to discuss the issue with Scienceor would only speak off the record. The U.S.Centers for Disease Control and Prevention(CDC) in Atlanta is investigating the source,but a spokesperson referred questions about itto a panel charged by U.N. Secretary-GeneralBan Ki-moon with investigating the outbreak.(The panel’s chair, Alejandro Cravioto of theInternational Centre for Diarrhoeal DiseaseResearch, Bangladesh, in Dhaka says none ofits four members will speak to the press untiltheir job is finished in April.) Meanwhile, arecent editorial in The Lancet Infectious Diseasescautioned against “apportioning blame”and declared the hunt for the source “a matterof scientific curiosity for the future.”Yet several scientists say that attitude iswrong-headed. Knowing how the outbreakstarted is very important, says MatthewCHOLERA IN HAITILathèmeLa ThèmeMINUSTAH baseMeilleArtiboniteTrou ChouchouDétourMirebalaisMeilleFirst cholera casesArtibonitePort au PrinceHaitiGround zero? Because the first cholera casesoccurred in Meille, a French epidemiologist thinksthe Nepalese base is the source.Waldor of Harvard Medical School in Boston,because it could help prevent outbreaksfrom happening elsewhere. Besides, “if anepidemic killed 4000 people in Europe or theU.S., we would want to know exactly whereit came from,” says Piarroux. “So why notwhen the same happens in Haiti?”Scientific results published in the past2 months support an Asian origin for theoutbreak but say nothing about Nepal. InNovember, CDC scientists reported inMorbidity and Mortality Weekly Report thatpulsed-field gel electrophoresis—a widelyused method to type microbial strains—showed that Haiti’s cholera was indistinguishablefrom strains “found in countriesin South Asia and elsewhere.”Sequencing the strain’s entire genome andcomparing it with that of other strains canprovide more detailed information, and a firststab at that came on 6 January in a paper byWaldor and 15 other researchers in The NewEngland Journal of Medicine. The groupsequenced two samples from Haiti, two fromBangladesh collected during 2002 and 2008outbreaks, and a Peruvian strain from 1991.They found that the Haitian strains wereclosely related to those from Bangladesh butnot to the one from Peru.Like CDC, the group is now sequencinga much larger number of samples, includingone taken in Nepal about 5 years ago, saysWaldor, who’s also trying to get his handson a sample from the outbreak that occurredin Nepal last fall, the same time the peacekeepersdeparted for Haiti. But whetheradditional samples will provide a definitiveanswer is not clear, he says. Vibrio choleraedoesn’t evolve quickly, so there may not beenough differences between strains from differentcountries to tell them apart.But for Piarroux, a researcher at the Universityof the Mediterranean in Marseille whopreviously investigated several African choleraoutbreaks, there’s enough circumstantialevidence to clinch the case. Working withepidemiologists of Haiti’s Ministry of PublicHealth and Population, he found that thevery first wave of cholera cases occurred ina village in central Haiti called Meille, wheremany inhabitants collected their drinkingwater from a stream, also named Meille, justdownriver from the Nepalese MINUSTAHcamp. During the first days of the epidemic,the Haitians noticed a pipe carrying a nauseatingliquid from a septic tank inside the camp tothe stream. The pipe—which was mentionedCREDITS (TOP TO BOTTOM): REUTERS; ADAPTED FROM R. PIARROUX38828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWS&ANALYSISCREDIT: APin a Haitian report of which Science obtaineda copy—was later removed, Piarroux says inhis own report. (A paper he wrote about theoutbreak is currently under review.)Cholera can travel the world inside people’sintestines, even unnoticed, becausemany infected people have few or no symptoms.But Piarroux does not believe that’swhat happened in Haiti. Based on the numberof people who got sick in the first waveof the outbreak farther downriver, he calculatedthat the Nepalese must have had dozensof patients and dumped hundreds of liters ofcontaminated stool. As he wrote in his leakedreport, there was a “massive contamination”of the Artibonite River that suggests a largeoutbreak inside the camp. A spokespersonfor MINUSTAH did not respond to e-mailedquestions, but the United Nations has deniedthat soldiers were sick.Waldor says he finds Piarroux’s evidence“suggestive but not absolutely conclusive.”There are ways to collect more evidence—forinstance, by testing the Nepalese soldiers forantibodies against V. cholerae. “It may be gettinglate for that,” Waldor cautions, becauseantibodies peak after about a month.Some scientists don’t believe foreignersintroduced cholera at all, despite the molecularclues. Most prominent among them isRita Colwell, a veteran microbiologist atthe University of Maryland, College Park.Colwell believes that most cholera outbreaksare caused by bacteria that lurk locally andproliferate when conditions are favorable—in this case, perhaps a climate event calledLa Niña. To Waldor, the idea that a microbeso closely resembling a South Asian strainwould emerge in Haitian waters is “franklyabsurd.” (Colwell e-mailed Science that shewas “in Bangladesh working on cholera andunable to respond” to questions.)If the Nepalese introduced cholera, saysWaldor, several measures could be consideredto prevent a repeat. Aid workers or peacekeepersfrom cholera-endemic countries whoare sent to cholera-free but vulnerable placeslike Haiti could be screened in advance, forinstance, or given prophylactic antibiotics.But Harvard cholera scientist EdwardRyan counters that testing thousands ofsoldiers using rectal swabs would be timeconsumingand costly—and that testing isn’tvery accurate. Prescribing antibiotics wouldpose problems, such as adverse reactionsand cause drug resistance in V. cholerae andother microbes. What’s more, says CDC epidemiologistEric Mintz, businesspeople, visitingrelatives, and tourists would also haveto be tested. “It wouldn’t be very practical,”Mintz says.–MARTIN ENSERINKERADICATIONPressure Growing to Set a Date toDestroy Remaining Smallpox StocksThe fate of the world’s last stocks of the deadlysmallpox virus is once again being debated.Last week, the executive board of the WorldHealth Organization (WHO) began discussingwhat research remains to be done with thelive virus. In May, WHO’s governing body, theWorld Health Assembly (WHA), will decidewhether to set a firm deadline for the stock’sdestruction. Contrary to some media reportsthat the executive board recommended keepingthe stocks, no conclusions were reached,says a WHO spokesperson.Smallpox, or variola, killed hundreds ofmillions of people before it was declared eradicatedin 1980. Only two labs still hold tightlyDoomed? The United States and Russia insist livesmallpox virus is still needed for research.secured stocks of the virus: VECTOR nearNovosibirsk, Russia, and the U.S. Centers forDisease Control and Prevention in Atlanta.A deadline for destroying these stocks firstset by WHO in 1990 has been postponedseveral times after the United States andRussia argued that research should continueon vaccines, antiviral drugs, and diagnostics(Science, 25 January 2002, p. 598).Since 2006, support has grown fordestruction, says Jonathan Tucker, a bioweaponsexpert at Darmstadt University of Technologyin Germany. Developing countries thathad suffered severe outbreaks have arguedthat several research goals had already beenachieved, and the others seem “unrealistic,”says Tucker. In response, in 2007 WHA calledfor a review of research since 1999.That panel of smallpox experts weighedin last October. Considerable progress hasbeen made, they noted: 48 strains of the virushave been sequenced, and scientists havealso developed new diagnostics, a safer versionof the standard smallpox vaccine, newvaccines for immune-compromised people,and two new candidate antiviral drugs. Butprogress has been slow on developing a nonhumanprimate model for smallpox infection,the panel said.WHO then asked another advisory groupof infectious-disease experts outside the poxvirusfield to review the scientific report; theyrecommended a limited scope of research.Although research should continue on vaccines,live virus is needed only to test drugsin vitro. Nor is a nonhuman primate modelworth further study, they said, suggesting thatregulatory authorities base their decisions onanimal studies of closely related viruses suchas monkeypox and cowpox.Both reports fed into a report from WHO’ssmallpox advisory committee that was formallypresented at last week’s meeting of the34-member WHO Executive Board, whereRussia, the United States, and—in a breakwith the past—some African countries reportedlyexpressed support for retaining stocks forresearch. But South Africa also urged that afirm date be set for destruction, sources toldScience. There was no final agreement, saysWHO spokesman Gregory Härtl; the “substantivediscussion on the eventual destructionof smallpox stocks will take place in Mayat the World Health Assembly.”Tucker expects other countries to join thecall for a destruction date. In May, the ThirdWorld Network (TWN), a Malaysian advocacygroup, will also lobby for limiting theresearch allowed with live virus, as the nonpoxvirusWHO advisory group suggested.That group “represents the broader publichealth view,” says Edward Hammond, a consultantfor TWN in Austin.Tucker also supports destruction. But ina commentary this month in Biosecurity andBioterrorism, he suggested a compromise toavoid a “diplomatic train wreck.” The UnitedStates and Russia, which now hold 451 vialsand 120 vials, respectively, should destroyall but 10 vials each. Countries should agreeto make it a crime to synthesize the variolavirus, and the United States and Russia shouldensure that antiviral drugs and vaccines arereadily available to developing countries.–JOCELYN KAISERwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 389

NEWS&ANALYSISPicking up the scent. Ranger Wang Jin Qianglistens for the gibbon’s morning call.PRIMATOLOGYLast-Ditch Effort to SaveEmbattled ApeBAWANGLING, CHINA—Dawn has brokenover the rainforest, and time is of the essence.Wang Jin Qiang climbs the steep tropical terrainwith the ease of a mountain goat, pausingonly to let less fit companions catch theirbreath and tend to cuts and scratches from rattanthorns. After several bursts of brisk hiking,the ranger stops near a magnificent tulip oakwith buttress roots that would turn a cathedralgreen with envy.“Shhh!” he says, finger to lips.A breeze has picked up. Almost lost amidthe rustling of leaves is a melodious whoop.“That’s their morning call,” Wang whispers.We strain to listen for a minute, then Wang’swalkie-talkie crackles to life. It’s anotherranger who has spent the past week campedout at a backcountry station solely to observewhat may be the world’s most endangered primate,the Hainan gibbon. After some hushedchatter in local dialect, Wang says to us, “Hespotted them. We’re close, let’s go.”By the latest tally, there are only22 Hainan gibbons—one family with11 members, another with seven members,and four loners—remaining in their last refuge,Bawangling National NatureReserve on southern China’sHainan Island. Here, rangers andscientists hope to prevent the firstprimate extinction in recordedhistory. “This is the most likelyprimate to go extinct. We have tosave it,” says Long Yongcheng,chief scientist for The NatureConservancy’s China programand the go-to person for China’sprimatology community.Long and others are guardedlyoptimistic. Bawangling’s gibbonpopulation dwindled to singledigits a quarter-century ago, butgovernment protection and high fecundityhave helped the species recapture some lostground. Experts can only speculate whetherthe genetic bottleneck the gibbon is squeezingthrough will cripple its survival odds.“We’re down to a scary number,” says conservationbiologist Chan Pui Lok Bosco, head ofKadoorie Conservation China (KCC), a nonprofitin Hong Kong. But other species undersimilar genetic duress have pulled through.Half a century ago, an estimated 2000 gibbonsroamed the rugged interior of Hainan,now known for its beach resorts. Villagers hadavidly hunted the gibbons to use their bodyparts in traditional medicine. But the big blowcame in the 1960s, when much of Hainan’slowland rainforest—the gibbon’s preferredhabitat—was converted to rubber plantations.The gibbons retreated to higher terrain, wheresubpopulations guttered and winked out. Bythe 1990s, a swath of land in Bawanglingsome 600 to 900 meters above sea level hadbecome their last redoubt.Realizing that something precious wasabout to be lost, provincial authorities in the1990s banned logging in Bawangling andRare sighting. Hainan gibbons, with baby, in Bawangling reserve.cracked down on poaching, which carries a10-year prison sentence. By then most villagershad sworn off hunting. Years ago, a localman and his entire family died suddenly afterhe had shot a gibbon, Chan says. A fortunateconsequence, he says, is that in the Bawanglingarea “hunting is taboo.”A new push to save the gibbon began in2003, when Hainan’s forestry departmentinvited Kadoorie to conduct a gibbon censusand devise a conservation action plan inBawangling. Chan and his colleagues confirmedthat there are only two families left andthat they are confined to suboptimal habitat.With few gibbon food trees at higher elevations,the gibbons must roam far to forage.For that reason, Chan says, “they have thelargest home range of any gibbon species inthe world.” Hainan gibbons are arboreal creatures;none has been observed to come downto the ground. To not interfere with the gibbons’diet, KCC trained locals to collect seedsof fallen fruit, such as the gibbon’s apparentfavorite, tao lan (Pouteria annamensis). KCChas since planted more than 80,000 food treesin degraded lowland in Bawangling, saysChan. And starting in 2005, KCC has sponsoredfour pairs of rangers—two for eachgibbon family—to spend 5-day shifts in thebackcountry monitoring the rare ape.After 5 years of intense study, the prognosisis uncertain. Researchers are collectingdroppings and hair to probe the genetic bottleneck.On the plus side, three females are nowpregnant, and the gibbons are managing toavoid human encounters.The gibbons are clearly skilled at that.After hearing their faint morning chorus, ittook only 10 minutes for Wang and companyto reach the area where the other ranger hadcalled in the sighting. By the time we arrived,the gibbons were gone. We followed througha singular land: jewel orchids withexquisite leaves, the ground puncturedby inch-wide holes made byforest crabs before hibernating,towering sugar palms, and a standof trees clad with strange leavesthat came to life—the beatingwings of thousands of iridescentblue and black butterflies.The gibbons, however, eludedus. That’s a survival skill. As longas they steer clear of their largerbrainedcousins, continue to makebabies, and have access to enoughfood, says Wang, “I think theyhave a chance.” –RICHARD STONECREDITS (TOP TO BOTTOM): R. STONE/SCIENCE; CHEN QING/BAWANGLING NATIONAL NATURE RESERVE39028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWS&ANALYSISCREDITS (TOP TO BOTTOM): CORBIS; BRIGITTE HALLIDAY, PROCESSCELL BIOLOGYTelling Time Without Turning On GenesOur planet’s spinning on its axis has a profoundeffect on all living organisms. Creaturesfrom bacteria to humans have internal timekeepersthat keep them in sync with Earth’s24-hour day-and-night cycle. Researchershave identified a host of genes and proteinsthat help track time, controlling daily cyclesof sleep and wakefulness, hunger, and metabolismin dozens of organisms and cell types.Most of the timekeepers discovered so farhave depended on gene transcription, theprocess in which cells use the informationstored in genes to make proteins, to drivethe complex feedback loops of moleculesthat make up a cell’s internal clock.Now researchers have found evidence—in people and in algae—for a eukaryotic circadianclock that works independently ofgene activity. The clock is present in humanred blood cells, which lack a nucleus and sodon’t make any new proteins, and a eukaryoticalga called Ostreococcus tauri. Thisprotein-based timekeeper, says AndrewMillar, a chronobiologist at the Universityof Edinburgh in the United Kingdom and anauthor on one of the two studies, might representan evolutionarily ancient way of keepingcellular time.There had been previous evidence thatsome circadian clocks could run independentlyof gene transcription. In 2005,researchers showed that the clock proteinsfrom cyanobacteria could run on theirown without input from the cell’s nucleus(Science, 15 April 2005, p. 414). Butthe new papers, published this week inNature, are the first to demonstrate a transcription-freeclock in eukaryotic cells.Other work had hinted that eukaryotesmight not always rely on transcription forcircadian timing, but it had been difficult toprove the case because chemicals that blockprotein synthesis have many side effects oncells. John O’Neill and Akhilesh Reddy, bothneuroscientists at the University of Cambridgein the United Kingdom, wondered whetherthey could use red blood cells as a naturallyoccurring example of a transcription-free cell.Reddy had shown in previous studies thata protein called peroxiredoxin (PRX), anantioxidant that helps to protect cells againstfree radicals, was involved in circadian cyclesin liver cells. So he and O’Neill isolated redblood cells from three healthy volunteers,kept the cells at a constant temperature andin complete darkness for 60 hours, and thentook samples of the cells and measuredthe oxidation status of PRX every 4 hours.Sure enough, the proteins showed a roughly24-hour cycle of oxidation.The ability to react to environmental stimuliis a key characteristic of circadian clocks,so the researchers tested whether they could“entrain” the red blood cells by exposing themto 12-hour cycles of higher and lowertemperatures. (They reasoned thatAncient timekeeper? Human red blood cells (top) andthe alga Ostreococcus have a similar circadian clock.red blood cells might react to the body’s dailyvariations in temperature, which is highest inthe evening and lowest in the morning.) ThePRX rhythms responded to the temperaturecues, suggesting that this clock could react toenvironmental signals.Before moving to Cambridge, O’Neillhad worked in Millar’s lab with Ostreococcus,looking for rhythm keepers in the cellsthat worked independently of transcription.The alga has a characteristic that made it especiallyuseful for the studies: After several daysin constant darkness, it shuts down its proteinmakingmachinery. However, when the lightsare turned back on, the cells show evidencethat they are still keeping time, even in theabsence of new proteins. No one had pinneddown the molecules involved in this alga’smystery clock.The researchers decided to check whetherthe alga’s PRX protein, which is very similarto the human one, undergoes a daily cyclelike the one in red blood cells. They found thatunder normal 12-hour dark-light cycles, theoxidation of the PRX protein rose and fell in a24-hour pattern. The pattern was still evidentwhen the algae were kept in constant darkness,even when chemicals were added thatblock protein manufacture.“We couldn’t believe it. We did [theexperiments] again and checked”whether they might have made somemistake, Reddy says. The similarity intwo such different organisms was a surprisebecause so far the clock genes identifiedin unrelated organisms have verylittle in common with each other—those inplants, for example, are completely differentfrom those in animals. “No one has beenable to show such a similarity across suchdiverse organisms,” Reddy says.It is not yet clear whether the PRX proteinsare a “hand” on the cellular clock—a readout—orwhether they are part of the machinerythat actually helps the cells keep time.And no one is suggesting that the transcription-basedclock is unimportant in human, oreven alga, cells; the researchers are workingto find out exactly how the two clockscomplement or back each other up. JosephBass, a chronobiologist at NorthwesternUniversity in Evanston, Illinois, notes thatthe PRX proteins play a role in cell metabolismand energy use, which fits in with recentstudies in humans that have found strong linksbetween disrupted circadian patterns and metabolicdisorders.The PRX proteins will give scientists anew tool to probe the workings of circadianclocks across a range of organisms, saysMartha Merrow, a chronobiologist at theUniversity of Groningen in the Netherlands.And Reddy suggests that the new work mayhelp scientists finally pin down the circadianmechanism in organisms, such as yeast andCaenorhabditis elegans, for which no clockrelatedgenes have been identified. “Becausebacteria, plants, and humans have these completelynonoverlapping [clock] genes, peoplehave assumed that clocks had evolved completelyindependently,” Reddy says. “Theymight all be linked through a common mechanismafter all.” –GRETCHEN VOGELwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 391

NEWSFOCUSA New ViewOf the Birth ofHomo sapiensNew genomic data are settling an oldargument about how our species evolvedFOR 27 YEARS, CHRIS STRINGER ANDMilford Wolpoff have been at odds aboutwhere and how our species was born.Stringer, a paleoanthropologist at the NaturalHistory Museum in London, held thatmodern humans came out of Africa, spreadaround the world, and replaced, rather thanmated with, the archaic humans they met.But Wolpoff, of the University of Michigan,Ann Arbor, argued that a single, worldwidespecies of human, including archaic formsoutside of Africa, met, mingled and hadoffspring, and so produced Homo sapiens.The battle has been long andbitter: When reviewing a manuscriptin the 1980s, Wolpoffscribbled “Stringer’s desperateargument” under a chart;in a 1996 book, Stringer wrotethat “attention to inconvenientdetails has never been part ofthe Wolpoff style.” At one tensemeeting, the pair presentedopposing views in rival sessionson the same day—and Wolpoffdidn’t invite Stringer to themeeting’s press conference. “Itwas difficult for a long time,”recalls Stringer.Then, in the past year, geneticists announcedthe nearly complete nucleargenomes of two different archaic humans:Neandertals, and their enigmatic easterncousins from southern Siberia. These dataprovide a much higher resolution view ofour past, much as a new telescope allowsastronomers to see farther back in timein the universe. When compared with thegenomes of living people, the ancientgenomes allow anthropologists to thoroughlytest the competing models of humanorigins for the first time.Going back in time. A researcher extracts DNA from a fossil.The DNA data suggest not one butat least two instances of interbreedingbetween archaic and modern humans, raisingthe question of whether H. sapiens at thatpoint was a distinct species (see sidebar,p. 394). And so they appear to refute the completereplacement aspect of the Out of Africamodel. “[Modern humans] are certainly comingout of Africa, but we’re finding evidenceof low levels of admixture wherever youlook,” says evolutionary geneticist MichaelHammer of the University of Arizona in Tucson.Stringer admits: “The story has undoubtedlygot a whole lot more complicated.”But the genomic data don’t prove theclassic multiregionalism model correcteither. They suggest only a small amountof interbreeding, presumably at the marginswhere invading moderns met archaic groupsthat were the worldwide descendants ofH. erectus, the human ancestor that leftAfrica 1.8 million years ago. “I have latelytaken to talking about the best model asreplacement with hybridization, … [or]‘leaky replacement,’ ” says paleogeneticistSvante Pääbo of the Max Planck Institute forEvolutionary Anthropology in Leipzig, leadauthor of the two nuclear genome studies.The new picture most resembles so-calledCREDITS: MAX PLANCK INSTITUTE FOR EVOLUTIONARY ANTHROPOLOGY39228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSFOCUSFIGURE SOURCES: CHRIS STRINGER, MILFORD WOLPOFF (LEFT); ERIC DURAND, MONTGOMERY SLATKIN (RIGHT)Ancient abode. A fi nger and molar (inset) of anew type of human were found in Denisova Cave,Siberia.assimilation models, which got relativelylittle attention over the years. “This meansso much,” says Fred Smith of Illinois StateUniversity in Normal, who proposed such amodel. “I just thought ‘Hallelujah! No matterwhat anybody else says, I was as close tocorrect as anybody.’ ”Evolving modelsStringer and others first proposed Africa asthe birthplace of modern humans back in themid-1980s. The same year, researchers publisheda landmark study that traced the maternallyinherited mitochondrial DNA (mtDNA)of all living people to a femaleancestor that lived in Africa about200,000 years ago, dubbed mitochondrialEve. She caught theattention of the popular press,landing on the cover of Newsweekand Time.Additional studies of livingpeople—from Y chromosomes tosnippets of nuclear DNA to theentire mtDNA genome—consistentlyfound that Africans werethe most diverse genetically. Thissuggests that modern humansarose in Africa, where they hadmore time to accumulate mutationsthan on other continents(Science, 17 November 2006,p. 1068). Meanwhile, ancientDNA technology also took off.Pääbo’s group sequenced firsta few bits of Neandertal mitochondrialDNA in 1997, thenAfricansthe entire mitochondrial genomes of severalNeandertals—and found them to be distinctfrom those of living people. So ancientDNA, too, argued against the idea of mixingbetween Neandertals and moderns. Over theyears the replacement model became the leadingtheory, with only a stubborn few, includingWolpoff, holding to multiregionalism.Yet there were a few dissenting notes. A fewstudies of individual genes found evidence ofmigration from Asia into Africa, rather thanvice versa. Population geneticists warned thatcomplete replacement was unlikely, giventhe distribution of alleles in living humans.And a few paleoanthropologists proposedmiddle-of-the-road models. Smith, a formerstudent of Wolpoff’s, suggested that most ofour ancestors arose in Africa but interbred withlocal populations as they spread out aroundthe globe, with archaic people contributing toabout 10% of living people’s genomes. At theUniversity of Hamburg in Germany, GunterBrauer similarly proposed replacement withhybridization, but with a trivial amount ofinterbreeding. But neither model got muchtraction; they were either ignored or lumpedin with multiregionalism. “Assimilation gotkicked so much,” recalls Smith.Over time, the two more extreme modelsmoved toward the middle, with mostmultiregionalists recognizing that the chiefancestors of modern humans arose in Africa.“The broad line of evolution is pretty clear:Our ancestors came out of Africa,” saysbiological anthropologist John Relethfordof the State University of New York Collegeat Oneonta. “But what happens next is kindof complex.”MultiregionalContinuityEuropeansAsiansSpread of Homo erectusthroughout the worldAfrican origin for Homo erectusAustraliansScenarios of Modern Human OriginsYears agoModernhumans50,000100,000Homoerectus1,800,000AfricansOut of AfricaEuropeansAsiansX X XGenes from the pastThen in May 2010 came the Neandertals’complete nuclear genome, sequenced fromthe bones of three female Neandertals wholived in Croatia more than 38,000 years ago.Pääbo’s international team found that a smallamount—1% to 4%—of the nuclear DNAof Europeans and Asians, but not of Africans,can be traced to Neandertals. The mostlikely model to explain this, Pääbo says, wasthat early modern humans arose in Africabut interbred with Neandertals in the MiddleEast or Arabia before spreading into Asiaand Europe, about 50,000 to 80,000 years ago(Science, 7 May 2010, pp. 680, 710).Seven months later, on 23 December,the team published in Nature the completenuclear genome of a girl’s pinky fingerfrom Denisova Cave in the Altai Mountainsof southern Siberia. To their surprise,the genome was neither a Neandertal’s nora modern human’s, yet the girl was alive atthe same time, dating to at least 30,000 yearsago and probably older than 50,000 years.Her DNA was most like a Neandertal’s, buther people were a distinct group that had longbeen separated from Neandertals.By comparing parts of the Denisovangenome directly with the same segments ofDNA in 53 populations of living people, theteam found that the Denisovans shared 4%to 6% of their DNA with Melanesians fromPapua New Guinea and the BougainvilleIslands. Those segments were not found inNeandertals or other living humans.The most likely scenario for how all thishappened is that after Neandertal and Denisovanpopulations split about 200,000 years ago,Changing views. Two models of modern human origins (left) are being challenged by new insights based on ancientDNA (right), which suggest some limited interbreeding between modern and archaic populations.AustraliansSpread of Homo erectusthroughout the worldAfrican origin for Homo erectusTimein years0about100,000about500,000AfricanLeaky ReplacementEuropeanAsianMelanesianModern HumanLineageInterbreedingNeandertalNeandertalLineageDenisovanmodern humans interbred with Neandertals asthey left Africa in the past 100,000 years. ThusNeandertals left their mark in the genomes ofliving Asians and Europeans, says co-authorMontgomery Slatkin, a population geneticistat the University of California, Berkeley.Later, a subset of this group of moderns—who carried some Neandertal DNA—headedeast toward Melanesia and interbred with theDenisovans in Asia on the way. As a result,Melanesians inherited DNA from both Neandertalsand Denisovans, with as much as 8% oftheir DNA coming from archaic people, saysco-author David Reich, a population geneticistat Harvard Medical School in Boston.This means H. sapiens mixed it up withat least two different archaic peoples, in atleast two distinct times and places. To some,that’s starting to sound a lot like multiregionalism.“It’s hard to explain how good I feelwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 393

NEWSFOCUSThe Species Problemabout this,” says Wolpoff, who says that seeingcomplete replacement falsified twice in1 year was beyond his wildest expectations.“It was a good year.”And yet the interbreeding with archaichumans seems limited—from 1% to 8% ofsome living people’s genomes. Stringer andmany others don’t consider it full-scale multiregionalcontinuity. “I think interbreedingwas at a low level,” says Slatkin, who saysthat if there had been a great deal of admixture,the genetic data would have revealed italready. Low levels of interbreeding suggestthat either archaic people mated with modernsonly rarely—or their hybrid offspring had lowfitness and so produced few viable offspring,says population geneticist Laurent Excoffierof the University of Bern in Switzerland.In any case, Reich notes that at least 90%of our genomes are inherited from Africanancestors whoreplaced the archaicpeople on other continentsbut hybridizedwith them aroundthe margins. And thatscenario most closelybacks the assimilationmodels proposed bySmith and Brauer.Of course, it’s possiblethat future data willNeandertal RangeAFRICAoverturn today’s “leaky replacement” model.Slatkin says he cannot rule out an alternativeexplanation for the data: The “archaic”DNA thought to have come from mating withNeandertals could instead stem from a veryancient ancestor that we shared with Neandertals.Most modern humans retained thosearchaic sequences, but Africans lost them. ButOur ancestors had sex with at least two kinds of archaic humans at two different times and places—and those liaisons produced surviving children, according to the latest ancient DNA research (seemain text, p. 392). But were the participants in these prehistoric encounters members of separatespecies? Doesn’t a species, by definition, breed only with others of that species?These are the questions paleogeneticist Svante Pääbo dodged twice last year. His team publishedtwo papers proposing that both Neandertals and mysterious humans from Denisova Cavein Siberia interbred with ancient modern humans. But the researchers avoided the thorny questionof species designation and simply referred to Neandertals, Denisovans, and modern humansas “populations.” “I think discussion of what is a species and what is a subspecies is a sterile academicendeavor,” says Pääbo, who works at the Max Planck Institute for Evolutionary Anthropologyin Leipzig, Germany.The question of how to define a species has divided researchers for centuries. Darwin’s words inOn the Origin of Species still hold: “No one definition has satisfied all naturalists.” However, manyscientists use the biological species concept proposed by Ernst Mayr: “groups of actually or potentiallyinterbreeding natural populations, which are reproductively isolated from other such groups.”The draft versions of the Neandertal and Denisovan nuclear genomes show low levels of interbreedingbetween each of them and modern humans. Apply Mayr’s definition strictly, and all threemust be considered Homo sapiens. “They mated with each other. We’ll call them the same species,”says molecular anthropologist John Hawks of the University of Wisconsin, Madison.But that’s a minority view among paleoanthropologists. Many consider Neandertals a speciesseparate from modern humans because the anatomical and developmental differences are “anorder of magnitude higher than anything we can observe between extant human populations,”says Jean-Jacques Hublin, a co-author of Pääbo’s at Max Planck. In the real world, he says, Mayr’sconcept doesn’t hold up: “There are about 330 closely related species of mammals that interbreed,and at least a third of them can produce fertile hybrids.”There’s also no agreed-upon yardstick for how much morphologic or genetic difference separatesspecies. That’s why Pääbo’s team avoided the species question a second time with respect tothe Denisovans. These hominins are known only from a scrap of bone, a single tooth, and theirDNA. They are genetically closest to Neandertals. The genetic distance between Denisovans andNeandertals, in fact, is only 9% larger than that between a living Frenchman and a living San Bushmanin Africa, both of whom belong to H. sapiens. But so far Neandertals seem to have low geneticdiversity, based on the DNA of six Neandertals from Russia to Spain. To Pääbo’s team, that makesthe difference from the Denisovans significant.Also, the Denisovan tooth doesn’t look much like that of a Neandertal. So the team considersthem a distinct population but declined to name a new species. “Why take a stand on it when it willonly lead to discussions and no one will have the final word?” asks Pääbo.–A.G.EURASIAContact zones. Modern humans fromAfrica interbred with Neandertals (pink).Then one group mixed with Denisovans(green) on the way to Melanesia.DenisovansMelanesiansOCEANIAAUSTRALIASlatkin says this “doesn’t seem very plausible,”because it requires modern human populationswith the archaic DNA and those withoutit to have been partially isolated from eachother in Africa for hundreds of thousands ofyears. And it seems even less probable thatMelanesians and Denisovans are the onlygroups that retained a second set of archaicDNA motifs from a common ancestor sharedby all modern humans, Neandertals and Denisovans.If those explanations do prove true,replacement would not be falsified.In the wake of the big genome studies,other researchers such as Hammer are scrutinizingDNA from more living humansto further test the model. Researchers arealso trying to pinpoint when admixturehappened, which has significant consequences.At just what point did we evolvefrom archaic humans to become “modern”humans? “There are still archaic [genetic]features floating around until amazinglyrecently, until 40,000 years ago,” says Hammer.He wonders whether the process ofbecoming modern took longer and was morecomplex than once thought. “There’s no lineyou can draw and say everything after thisis modern. That’s the elephant in the room.”Meanwhile, paleoanthropologists aresearching for fossils in Asia that might belongto the enigmatic Denisovan population—andmight yield more ancient DNA. PaleoanthropologistRussell Ciochon of the University ofIowa in Iowa City and Wolpoff say there areseveral known, ambiguous fossils in Asia thatmight be candidates for early Denisovans. “Ibelieve things were going on in Asia that wejust don’t know about,” says Ciochon. “Beforethis paper on the Denisovans, we didn’t haveany insight into this. Now, with this nucleargenome, I find myself talking about ‘the Denisovans.’It’s already had an impact.”As for Stringer and Wolpoff, both nowin their 60s, their battle has mellowed. Theirviews, while still distinct, have convergedsomewhat, and they shared a beer at a Neandertalmeeting last year. “The reason we geton well now,” says Stringer, “is we both thinkwe’ve been proved right.” –ANN GIBBONS39428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSFOCUSWind tunnel champ. Marcel Klaassen watchesBlue, a bird that flew 16 hours and went nowhere.CREDIT: MAGNUS ELANDERPHYSIOLOGYGoing the DistanceOne challenge to studying migration energetics:Not every bird will fly for hours in a wind tunnelLONDON, CANADA—In a dimly lit room,Chris Guglielmo reaches into a cloth bagand gently pulls out a tiny bird. He checksthat its feathers and wings are intact, thenputs his hand inside a Plexiglas tunnel andreleases it. All eyes in the room are mesmerizedby the blur of beating wings asthis yellow-rumped warbler flits about thechamber. An 8-meter-per-second headwindkeeps the bird in place as it begins whatGuglielmo hopes will be a multihour flight.Moments later, the warbler perches on anet at the upper end of the tunnel. It can becoaxed back into the air, but it’s clear thatit prefers to stay put. “I’m a little bummed,but it’s how it goes,” says Guglielmo’s graduatestudent, Alex Gerson, as he recapturesthe bird.A $9.3 million grant has boughtGuglielmo’s group at the University ofWestern Ontario here a state-of-the-art birdresearch facility, complete with aviaries, surgeryroom, a bird-sized magnetic resonanceimaging (MRI) machine, and a $1.5 millionwind tunnel, the only one in the world inwhich temperature, humidity, and barometricpressure can be controlled. A year old, thetunnel presents researchers with an unprecedentedopportunity to probe the mysteriesof migration in exquisite detail.Guglielmo and Gerson want to understandfuel and water use by migrants. Inthe field, they follow birds that stop over tostock up during the fall and spring migration.What the birds eat can affect their abilityto reach their final destination, and conservationistsare keen to know which dietswork best—and possibly provide them. “It’snot enough, if you care about wild birds, togo out and count them and study their habitat,”says Guglielmo. “You mustknow more about the mechanisms,”and that’s where the windtunnel comes in. “It’s providingus with a tool to answer questionsthat we could only answerindirectly [before],” says ScottMcWilliams, a physiological ecologist at theUniversity of Rhode Island, Providence.Wind tunnels provide a controlled environmentin which researchers can takebefore-and-after physiological measurementsto assess the toll flight takes on fat stores, proteinmass, water, and even immunologicalfunction when birds fly long distances. Inaddition, high-speed cameras and techniquesfor visualizing air flow enable Guglielmo’scolleagues to assess the biomechanics of flappingwings, giving a better handle on the linkbetween aerodynamics and migration.Onlinesciencemag.orgPodcast interviewwith authorElizabeth Pennisi.But money can’t buy birds that are keento fly, and Guglielmo and his team have hadmixed success finding willing avian partners.An early project using starlings didwell, despite taking place when the tunnelwas not quite finished. It netted “Super,” abird that always cooperates and will evenfly into the wind tunnel on its own accord.But a study involving robins took monthsto identify five somewhat cooperativefliers; switching to Swainson’s thrushesworked better. One immunological projectinvolving a shorebird called a ruff isstranded because the birds show no inclinationto take to the air. And Guglielmo hasjust started testing warblers to see if highproteinor high-carbohydrate diets makea difference in energy use during flight.“There’s a lot of trial and error on what speciesof birds to use and how to find individualsthat will fly,” says Guglielmo.Eat, drink, and flyWhen Guglielmo gets discouraged, hethinks about Blue, a wind-tunnel recordholder. In the mid-1990s, field ecologistsMarcel Klaassen and Åke Lindström wantedto learn more about how stopovers influencesubsequent migration. “In the wild, you canstudy a bird only a few days and then it’sgone, literally gone,” explains Lindström,based at Lund University in Sweden. Untilthat time, researchers had only tested nonmigratorybirds, such as homing pigeonsand small parrots, in short wind-tunnelflights, “so we didn’t know if what we foundout was relevant” to migratory birds, hesays. Lund had a newly built wind tunnel,one that later served as inspirationfor Guglielmo’s.On the advice of a windtunnelexpert, Lindström andKlaassen were trying to trainthrush nightingales to fly from astick for a food reward, but thebirds were not cooperating. Getting one tostay in the air for even just a minute “wascause for celebration,” Lindström recalls.That all changed when someone suggestedremoving the stick.Blue looked back for his perch, wavereda little, then settled down to flying. In onetrial, Blue kept going for 16 hours, showingno signs of slowing down even whenthe researchers decided it was time to gohome to bed. Over the next 2 months, Bluemade seven 12-hour flights, yielding datathat led to five research papers on fuel use.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 395

NEWSFOCUSTreading AirMigration test bed. Researchers cancontrol this wind tunnel’s temperature,humidity, and air pressure.Wind tunnels have come a long way since formerDuPont president and amateur ornithologistCrawford Greenewalt built one of the first50 years ago to study hummingbirds. A fanat the end of his 46-centimeter “tunnel” kepthummingbirds stationary while he filmed themflying; those high-speed images impressedamateurs and academics alike.In 1994, Swedish biologists greatlyimproved on the concept, borrowing fromdesigns used in engineering. Air was recirculatedthrough a 20-meter-long rectangularchamber. Just upstream of the 2-meter-longtest section, the tunnel widens and then narrows,which greatly reduces turbulence andmakes it easier to assess how power requirementsfor fl ight vary with speed. In this setting,Anders Hedenström of Lund University inSweden and his colleagues adopted an engineeringtechnology that uses fog and lasers tovisualize airflow. They documented for the firsttime what happens to air as its moves acrossa real bird’s wing. The tunnel can be tilted toallow birds to glide or “climb” to higher altitudes.Over the years Hedenström and his colleagueshave studied fl ight in 15 species, aswell as in bats.A copy of the Lund tunnel was built at theMax Planck Institute for Ornithology in Seewiesen,Germany, in 1999. There, 150 birds from25 species have been evaluated.When Chris Guglielmo started at the Universityof Western Ontario in London, Canada,in 2005, he realized those tunnels were toofar away for him to include wind-tunnel studiesroutinely in his research. Others on campuswanted better bird-care facilities for theirresearch, so he teamed up with nine colleaguesand proposed the Advanced Facility for AvianResearch. Much to their university president’ssurprise, the Canada Foundation for Innovationapproved this grant application in 2007.The wind tunnel is the centerpiece of anew building that includes indoor and outdooraviaries, some of which have small poolsof water for shorebirds and waterfowl, acousticchambers, behavior-observation rooms, andenvironmental chambers for controlling lightcycles, as well as temperature and humidity.Made of steel, two stories high and 12 meterslong, the tunnel looks quite imposing. Birds flyin a 1-meter-high by 1.5-meter-wide octagonalPlexiglas cylinder, almost completely coveredwith blankets to discourage the birds fromgetting too close to the walls. Pressure, temperature,and humidity in the tunnel can allbe modifi ed to simulate different conditions,including high altitude.The whole flight section is encased in asteel room, with air locks. Once the room andwind tunnel are locked down, “in a few minutes,you can go up to over 23,000 feet,” saysGuglielmo. “There’s a lot of theory about whathappens when birds fly at [high] altitudes, butnowhere in the world had we had a place to testthat empirically.”Guglielmo and his colleagues are still developingprotocols for low-pressure research, butthey have already simulated weather fronts.Scott MacDougall-Shackleton, also of WesternOntario, and Guglielmo put the birds into thetunnel at night and just before the lights go on,they cause a drop or rise in air pressure and temperatureand observe how that affects the birds’behavior when they wake up. It seems that whenthe birds sense bad weather coming on, they aremuch quicker to start feeding, says Guglielmo.The tunnel also doubles as a studio forunderstanding the mechanics of bird flight.Working with Roi Gurka of Ben Gurion Universityin Beer-Sheva, Israel, Gregory Kopp andgraduate student Adam Kirchhefer of WesternOntario have filled the tunnel with oil dropletsand set up a laser field in front of a high-speedvideo camera that captures the movement ofthe droplets—and the air—as a bird fl ies. Asecond camera is trained on the bird itself. Inparticular, they are studying the movement ofair at the wing tips. For these studies, a particularlycooperative starling named Super wearsan orange-tinted mask to protect its eyes fromthe laser. “A few good seconds of video will keepyou going for a long time,” says Guglielmo. Anda sample size of one or two is sufficient. “It’s adifferent ball game when you want to look atmigration,” where one needs multiple birdswilling to fly for hours at a time. –E.P.The work showed that the estimated powerrequirements for migratory flight were lessthan those for nonmigratory flights and suggestedthat birds burn protein as well as faton their long journeys.Searching for eager fliers like Blue,Guglielmo knew that the wind tunnel itselfwas not the problem. Already, McWilliamshad had a successful run for his experimenton the effects of fatty acid composition anddietary oxidants on exercise performance,even as the tunnel was being finished. In thesummer of 2009, McWilliams’s team captured120 starlings, feeding some of themfood rich in olive oil (a monounsaturatedfat) and others a diet high in canola oil (apolyunsaturated fat). His earlier experimentsshowed that migrating birds willchoose fruits, such as viburnum or bayberry,that are high in polyunsaturated fats,even though metabolizing polyunsaturatedfats generates more oxidative stress thanburning monounsaturated fats. McWilliamswanted to determine whether birds eatingthe canola oil diet flew more efficiently,burning less fat per hour, than birds eatingolive oil, making the extra stress worth it.The researchers look at how birds fueltheir flights by determining the change inthe ratio of fat to lean mass, and the rate atwhich the birds expend energy. Researchershave traditionally assessed fat contentby checking the color and condition ofCREDIT: SCOTT MACDOUGALL-SHACKLETON39628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

NEWSFOCUSTaking measure. Briefly immobilized,this bird undergoes a magnetic resonancescan of its body composition.check how the state of the immune systemaffects flight. There is some evidence thatbirds with infections don’t fly as well, andit is important to quantify the handicap,particularly for epidemiologists trying topredict how far avian-borne diseases willspread. The tradeoff that may exist betweenimmunocompetence and flight performance“is very difficult to study in a natural setting,”she points out.The wind tunnel is too small for flyingducks and geese, so Nebel wants touse ruffs as stand-ins. However, the ruffs,which come from a captive colony, showedno inclination to migrate, despite havingput on weight and being used to captivity.Nebel has tried hatching, raising, and trainingyoung. Even so, at 2 months old, theseyoungsters fly no more than a few minutes ata time. Her last hope is that as they mature,the birds will develop larger flight musclesthat will let them stay in the air longer.CREDIT: E. PENNISI/SCIENCEthe belly. A more accurate measurementusing x-rays requires the birds to be anesthetized.The Western Ontario lab has a farmore convenient tool, however: a custombuiltquantitative magnetic resonance scannerthat can produce an accurate readoutof the lean mass, fat mass, and body watercontent quickly and noninvasively. It takesjust 2 minutes, with the bird temporarilyimmobilized in a Plexiglas tube. “It’s nomore difficult than weighing a bird,” saysGuglielmo. “It’s really a new way to dothese experiments.”To assess how much energy the birdsexpend, the researchers inject birds justprior to flight with water containing heavierisotopes of hydrogen and oxygen. Theythen take blood samples before and afterflight and use the change in the relativeamounts of these stable isotopes to estimatecarbon dioxide production, a proxy forenergy expenditure.McWilliams was eager to begin testingthe effects of the different diets on flightenergetics, but construction had fallenbehind schedule, and the experiment wasdelayed for weeks. By September 2009,he couldn’t wait any longer. “It was a littletouch and go,” he recalls. In between cyclesof construction, he and his colleaguestrained birds to fly in the wind tunnel andthen, for the experiment, let each bird fly aslong as it wanted, hoping for hourlong orlonger flights. McWilliams is still analyzingthe data, but he says The Nature Conservancyis eager to know the answer so it canbetter manage the preserves used by migratingbirds.Hit or missGuglielmo and Gerson knew from experiencethat finding the right species for theirexperiments, and the right individuals, washit or miss. Gerson is tackling another aspectof flight energetics: water balance. Studiesof birds that fly across deserts indicate thatto make up for water loss, the birds breakdown their organs, harvesting water from themetabolism of proteins in muscles or the gut,for example. Gerson has been testing thisidea by examining the change in body water,fat mass, and lean mass in birds flying in thewind tunnel.American robins, common migrants ofoptimal size, seemed ideal. “But they didn’twork out so well,” says Guglielmo. Still, heand Gerson did learn something from robins:Hand-raised birds were no more likelyto cooperate than wild-caught birds that weretrained. So Gerson started trying whateverbirds he could get his hands on, red-eyed vireos,gray catbirds, yellow-rumped warblers.Their conclusion: “Fat birds and calm birdsfly,” quips Guglielmo.Swainson’s thrushes seemed to workwell. One went more than 5 hours nonstop,the equivalent of about 180 kilometers, usingabout 0.25 grams of fat per hour. Gerson hasbeen testing them in 10% and 80% humidity,checking weight loss and using the MRI scannerto measure changes in water content andlean and fat mass before and after the flights.Preliminary results show that “if the bird fliesin dry conditions, it will burn more of its leanmass to liberate more water,” he says.Fat and calm is no guarantee of success,however. Silke Nebel, a postdoc, wants toFlying without a netGuglielmo settled on the yellow-rumpedwarbler because it is small, and the species isknown to adapt well to captivity. The yellowrumpedwarbler varies its diet seasonally,feasting on insects in the spring but turningmore toward fruits in the fall. Guglielmo islooking at how diet affects body compositionand fuel use during flight.For these experiments, he mixes up pansof casein- or sugar-rich agar and, after themixtures harden, runs the gel through a foodmill to make small, appetizing “worms.”Guglielmo calls the high-protein fare hisAtkins diet for birds; those on it weighabout 2 grams less than birds feasting onhigh carbohydrates.When Guglielmo first let the warblers tryout the tunnel, they didn’t show much inclinationto keep flying and instead kept landingon the upwind net. In earlier experiments, theresearchers had tried everything—makingthat end of the tunnel dark, flashing strobelights, even spraying water jets to discouragethe birds from perching there. This night,they are pulling out all stops and taking thenet down altogether. They tested a bird for20 minutes and it didn’t head upwind toofar, so now they are ready to see if the warblerswill fly long distances. The next night,one bird lasted 45 minutes before landing onthe tunnel floor—an okay flight time but notideal. But the evening after that was a differentstory. “Purple-black,” named for thecolor tag on its leg, “found his groove andflew for 6 hours,” Guglielmo e-mailed thenext day. “We had a great night.”–ELIZABETH PENNISIwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 397

COMMENTARYLETTERSedited by Jennifer SillsDrugs across cultures403Running your programson cloud computers406LETTERS I BOOKS I POLICY FORUM I EDUCATION FORUM I PERSPECTIVESRecognizing Scientistsand TechnologistsON 17 NOVEMBER 2010, PRESIDENT OBAMA PRESENTED THENational Medals of Science and the National Medals of Technologyand Innovation. These medals are the highest honor that thenation can bestow in science and technology, yet they are rarely tioned by the popular media. Because Congress does not appropriate fundsmen-to implement the “outreach” of these medals, for many years the only national recognition wasa private award ceremony with the President.In 1991, George Rathmann, one of the founders of the biotech industry, facilitated the formationof what is now the National Science and Technology Medals Foundation. The missionof the Foundation is to promote the National Medal Laureates as role models for students andthereby encourage interest in science and math. To accomplish this goal, the Foundation hostsa banquet in conjunction with the White House ceremony. This banquet features videos highlightingthe technical accomplishments of the Laureates, which then become the basis for storiesthat appear throughout the country.Over the years, the Foundation has accumulated a wealth of electronic material on theLaureates, including biographies, interviews, and descriptions of their accomplishments (1).This recognition not only is a way to recognize the Laureates’ enormous efforts, but also servesto focus our attention on the seminal ideas in science, mathematics, and engineering. The storiesbehind these accomplishments often provide inspiration to others, which is essential topromote further achievements.ROBERT M. WHITEMaterials Science and Engineering, Stanford University, Stanford, CA 94305, USA. E-mail: RMWhite@stanford.eduReference1. National Science and Technology Medals Foundation (www.nationalmedals.org).Genetics-Based FieldStudies Prioritize SafetyM. ENSERINK’S NEWS OF THE WEEK STORY ONthe open release trials of genetically modifiedmosquitoes in the Cayman Islands (“GMmosquito trial alarms opponents, strains tiesin Gates-funded project,” 19 November 2010,p. 1030) highlights the growing pains associatedwith bringing new technologies out ofthe laboratory into the field. Unlike for vaccines,drugs, and insecticides, no industrywidestandards are yet in place to guide eitherpublic or private efforts in the developmentof these technologies. However, it is importantfor the public to know that the scientistsworking on these new technologies areaggressively supporting the formulation ofbest practices for their safe, efficient, ethical,and regulated application, and are reachingout to experts from a range of relevant disciplinesfor advice and counsel. A series ofpublications document the evolution of thisprocess (1–5). Indeed, efforts are currentlyunder way to develop a guidance frameworkfor quality standards to assess safety and efficacyand to address regulatory, legal, social,and cultural issues, as recommended by aninternational consultation held at the WorldHealth Organization in 2009 (5). Thus,although we have not achieved harmonizedinternational standards, as hastaken decades for other technologies,we are much closer than most peoplerealize. We recognize the need to ensurethat our enthusiasm for the promise of theseapproaches as powerful public health toolsdoes not outstrip our responsibility to applyscientifically validated and socially acceptableproduct development practices. The tragedywould be if this important but complexbirthing process were to stifle creativity in thedevelopment of not only genetics-based solutions,but all truly novel approaches that seekto reduce the serious health threat of diseasessuch as malaria and dengue fever. We hopethat debates over specific circumstances donot cloud the urgent need for the developmentand deployment of new tools to mitigate thesedisease scourges.ANTHONY A. JAMESDepartments of Microbiology and Molecular Genetics, andMolecular Biology and Biochemistry, University of California,Irvine, CA 92697–3900, USA. E-mail: aajames@uci.eduReferences1. M. Benedict et al., Vector-Borne Zoonotic Dis. 8, 127(2008).2. J. V. Lavery, L. C. Harrington, T. W. Scott, Am. J. Trop.Med. Hyg. 79, 312 (2008).3. J. V. Lavery et al., Trends Parasitol. 26, 279(2010).4. J. Mumford, Asian Pac. J. Mol. Biol. Biotechnol. 17, 91(2009).5. WHO/TDR, “Progress and prospects for the use ofgenetically-modified mosquitoes to inhibit diseasetransmission,” Report on planning meeting 1: Technicalconsultation on current status and planning for futuredevelopment of genetically-modified mosquitoes formalaria and dengue control (WHO/TDR publications10.2471/TDR.10.978-924-1599238, 2010).Origins of BiodiversityTHE ORIGIN OF THE HIGH NEOTROPICAL BIOdiversityhas been a controversial topicsince Darwin. The debate has focusedon the relative influences of the climatechanges during the Pleistocene (the pastCREDIT: WIKIMEDIA COMMONS39828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

CREDIT: LUIS CARILLO, VENEZUELAParasitereplication~2.6 million years) and the tectonic andgeographical reorganizations that occurredbefore the Pleistocene (1). In their Review“Amazonia through time: Andean uplift,climate change, landscape evolution, andbiodiversity” (12 November 2010, p. 927),C. Hoorn et al. conclude that the biodiversitypatterns of the Amazon basin werelargely shaped before the Pleistocene, facilitatedby the Andean uplift. The authors dismissPleistocene diversification by arguingthat, in the Neotropics, the refuge hypothesis(proposing that species diversifiedin isolated forests during glacial periods)has already been abandoned, and that fossilsand molecular phylogenetics supportmostly pre-Pleistocene diversification.However, Pleistocene diversificationcould have resulted from a variety of mechanismsother than isolated forest remnants(2–4). Furthermore, Hoorn et al. cite mymeta-analysis (5) in support of the pre-Pleistocene diversification, yet the conclusionsI drew from that study contradict thoseof Hoorn et al.’s Review. I concluded thatabout half of the dated extant neotropical speciesoriginated during the Pleistocene and theother half before it, and that speciation proceededin a continuous fashion with no evidentbursts (5). In addition, phylogenetic evidenceprovided by Hoorn et al. is based onthe dating of complexes of extant species (thecrown clades) that in fact records the age ofthe oldest species within each group (6) butnot necessarily the age of all the extant species,which should be necessarily younger.This overestimates pre-Pleistocene diversification.Earth’s biodiversity gradients arethe result of a long and complex history ofevolutionary trends, mediated by ecologicalprocesses and governed by external forces,in which not only speciation but also extinctionshould be considered, especially in extratropicalareas (7).The topic requires the synergy of manydisciplines, in a wide range of spatial andtemporal scales. Pleistocene speciation is onemore element and should not be neglected;after all, we ourselves are a Pleistocene speciesbarely 200,000 years old.VALENTÍ RULLSPORE PrizeEssay409 413Palynology and Paleoecology Laboratory, Botanical Instituteof Barcelona (CSIC-ICUB), 08038 Barcelona, Spain.E-mail: vrull@ibb.csic.esReferences1. M. B. Bush, H. Hooghiemstra, in Climate Change andBiodiversity, T. E. Lovejoy, L. Hannah, Eds. (Yale Univ.Press, New Haven, CT, 2005), pp. 125–137.2. M. B. Bush, J. Biogeogr. 21, 5 (1994).3. B. P. Noonan, P. Gaucher, Mol. Ecol. 14, 3017 (2005).4. V. Rull, J. Biogeogr. 31, 1 (2005).5. V. Rull, Mol. Ecol. 17, 2722 (2008).6. M. J. Benton, Biol. Rev. 75, 633 (2000).7. M. S. McGlone, Glob. Ecol. Biogeogr. Lett. 5, 309 (1996).ResponseIN OUR REVIEW, WE LINK THE OUTSTANDINGspecies richness in northern South Americato the cataclysmic changes induced byAndean mountain building. Evidence forthis is the correlation between sedimentaryrecords, the paleontological record, datedmolecular phylogenies, and present speciesdistributions. Our conclusions contradictthe hypothesis that has dominated for morethan 40 years: that the outstanding levels ofNeotropical species richness and current distributionpatterns were mainly produced byQuaternary climatic fluctuations (1, 2), i.e.,in the past 2.6 million years. All evidencein our meta-analysis pointstoward an older origin ofAmazonian biodiversity.Rull argues that we ignoreQuaternary evidence on speciation,in part by erroneouslyreferring to his previous metaanalysis(3) as evidence for pre-Quaternary diversification.Rull’s finding that about halfof all extant species analyzedoriginated during Quaternarytimes (3) is not surprising.Assuming the average specieslongevity is some 100,000to a couple of million years(3–5), at any point in time wewould expect to find that mostspecies originated in the pastfew million years. Rull’s evidencethat extant species originatedrecently does not contradictthe idea that the totalnumber of species was just as high (and formost organism groups higher) before theQuaternary, even if the species that existedthen have since become extinct. Moreover, ifPleistocene glaciations had indeed producedmost of the species richness observed today—as implied in the original formulation of the“refuge theory” (1)—this would unrealisticallyimply that all previous diversity was producedby entirely different mechanisms. Thisrealization severely undermines the role ofglaciation dynamics in accounting for Neotropicalspecies richness.Rull’s suggestion that we overestimatedpre-Quaternary diversification by usinggenera instead of species as taxonomic unitsin our meta-analysis is misleading. Extinctionis more likely to affect older lineagesthan younger ones—simply because speciesthat have arisen recently have had lesstime to go extinct (6)—meaning that Pre-Quaternary speciation events were probablyunder estimated in Rull’s meta-analysis (3).Stochastic diversification models (6)can correct for the effect of backgroundextinction in diversification rate estimates,but these models have provenun realistic because of their over simplifiedassumptions (7) and sensitivity to incompletetaxon sampling (8), a commonfeature in Neotropical phylogenies. Estimatesof crown ages of genera are, arguably,less sensitive to incomplete taxonsampling, because in most species-level phylogenies,sampling is aimed to cover the geographicand morphological variation withina genus. This should lead to more robust ageestimation of deeper nodes even when manyspecies are missing.Angel Falls, Venezuela.Debate continues about whenand where neotropicalbiodiversity developed.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 399

LETTERSThe data we assembled show that theblueprint of present Amazonia was laid outin pre-Quaternary times, but they also havethe potential to provide us clues on how therainforest may react to future global warming.It is also clear that Amazonian biotawithstood large geodynamic (9) and climaticfluctuations but that humans, the youngproduct of Quaternary evolution, pose thebiggest threat to this wealth of biodiversity.C. HOORN, 1 F. P. WESSELINGH, 2 H. TER STEEGE, 3M. A. BERMUDEZ, 4 A. MORA, 5 J. SEVINK, 1I. SANMARTÍN, 6 A. SANCHEZ-MESEGUER, 6C. L. ANDERSON, 6 J. P. FIGUEIREDO, 7C. JARAMILLO, 8 D. RIFF, 9 F. R. NEGRI, 10H. HOOGHIEMSTRA, 1 J. LUNDBERG, 11 T. STADLER, 12T. SÄRKINEN, 13 A. ANTONELLI 14,15 *Letters to the EditorLetters (~300 words) discuss material published inScience in the past 3 months or matters of generalinterest. Letters are not acknowledged uponreceipt. Whether published in full or in part, Lettersare subject to editing for clarity and space.Letters submitted, published, or posted elsewhere,in print or online, will be disqualified. To submit aLetter, go to www.submit2science.org.1Paleoecology and Landscape Ecology, Institute for Biodiversityand Ecosystem Dynamics, University of Amsterdam,1098 XH Amsterdam, Netherlands. 2 Nederlands Centrumvoor Biodiversiteit Naturalis, 2300 RA Leiden, Netherlands.3 Institute of Environmental Biology, Departmentof Biology, Faculty of Science, Utrecht University, 3584 CHUtrecht, Netherlands. 4 Laboratorios de Termocronología yGeomatemáticas, Escuela de Geología, Minas y Geofísica, Facultadde Ingeniería, Universidad Central de Venezuela, 1053,Caracas, Venezuela. 5 ECOPETROL, Instituto Colombiano delPetroleo, Piedecuesta, Santander, Colombia. 6 Real JardinBotanico, CSIC, 28014 Madrid, Spain. 7 Petroleo Brasileiro SA(Petrobras), CEP 20.031-170, Rio de Janeiro, Brazil. 8 SmithsonianTropical Research Institute, Balboa, Republic of Panama.9 Instituto de Biologia, Universidade Federal de Uberlândia,Campus Umuarama, Bairro Umuarama, Uberlândia,CEP 38400-902, Minas Gerais, Brazil. 10 Laboratório de Paleontologia,Campus Floresta, Universidade Federal do Acre,Estrada do Canela Fina, Cruzeiro do Sul, Acre, CEP 69980-000, AC, Brazil. 11 Department of Ichthyology, Academy ofNatural Sciences, Philadelphia, PA 19103, USA. 12 Instituteof Integrative Biology, ETH Zürich, 8092 Zürich, Switzerland.13Department of Plant Sciences, University of Oxford, OxfordOX1 3RB, UK. 14 Gothenburg Botanical Garden, Carl SkottsbergsGata 22A, 413 19 Göteborg, Sweden. 15 Department ofPlant and Environmental Sciences, University of Gothenburg,Carl Skottsbergs Gata 22B, 413 19 Göteborg, Sweden.*To whom correspondence should be addressed. E-mail:alexandre.antonelli@vgregion.seReferences1. J. Haffer, Science 165, 131 (1969).2. V. Rull, J. Biogeogr. 32, 921 (2005).3. V. Rull, Mol. Ecol. 17, 2722 (2008).4. M. Foote, D. M. Raup, Paleobiology 22, 121 (1996).5. R. T. Pennington et al., Proc. Natl. Acad. Sci. U.S.A. 107,13783 (2010).6. S. Nee et al., Philos. Trans. R. Soc. London B 344, 77(1994).7. D. L. Rabosky, Evolution 64, 1816 (2010).8. N. Cusimano, S. S. Renner, Syst. Biol. 59, 458 (2010).9. G. E. Shephard et al., Nat. Geosci. 3, 870 (2010).CORRECTIONS AND CLARIFICATIONSPerspectives: “The feeding habits of ammonites” by K.Tanabe (7 January, p. 37). The legend should read as followswith corrected genus names: “(Top) Polyptychoceras sp. withBaculites-like lower and upper jaws… (Bottom) Anagaudryceraslimatum (a lytoceratid) with a nautilus-like lower jaw.”Policy Forum: “Boosting CITES” by J. Phelps et al. (24December 2010, p. 1752). The heading for the fourthpotential solution was missing. “A Peer-Review Process”should have appeared before the paragraph on the secondpage that begins, “CITES shortcomings may be overlookedbecause the convention lacks internal and external checksand balances.” The header has been added in the HTMLversion online.News Focus: “Will homebody researchers turn Japaninto a scientific backwater?” by D. Normile (10 December2010, p. 1475). There is a shift in the increments on thevertical axis of the graph indicating the number of individualsmaking overseas visits. From 0 up to 10,000 individuals,the graph uses increments of 2000; above 20,000, ituses increments of 20,000.News of the Week: “U.N. biodiversity summit yields welcomeand unexpected progress” by D. Normile (5 November2010, p. 742). The name of Alison Stattersfield, headof science for BirdLife International, was misspelled.Learn how current eventsare impacting your work.ScienceInsider, the new policy blog from the journalScience, is your source for breaking news and instantanalysis from the nexus of politics and science.Produced by an international team of science journalists,ScienceInsider offers hard-hitting coverage on a range ofissues including climate change, bioterrorism, researchfunding, and more.Before research happens at the bench, science policy isformulated in the halls of government. Make sure youunderstand how current events are impacting your work.Read ScienceInsider today.www.ScienceInsider.orgBreaking news and analysis fromthe world of science policy40028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

BOOKS ET AL.CREDIT: JOHN G. SEELEY/COURTESY THOMAS D. SEELEYBEHAVIORInsect Swarm IntelligenceLars Chittka and Alex MesoudiThe swarming behavior of honeybeesconstitutes one of the most astoundingphenomena in group decision-makingamong animals: When a new queen is raisedin the honeybee hive, her predecessor departsalong with approximately 10,000 workers.Rather than en masse erratically searchingaround for a new locationHoneybee Democracyby Thomas D. SeeleyPrinceton University Press,Princeton, NJ, 2010. 279 pp.$29.95. ISBN 9780691147215.The reviewers are at the Research Centre for Psychology,School of Biological and Chemical Sciences, Queen Mary,University of London, Mile End Road, London E1 4NS, UK.E-mail: l.chittka@qmul.ac.uk(as a group of vertebratesmight), the bee swarm doessomething remarkable. Itacts, in a sense, like a singlebeing: pausing, collectinginformation, carefully consideringits options, and thenmaking a unanimous decisionabout where to move. The consensus cantake days to reach. Several hundred scouts fanout across a territory of up to 70 km 2 seekinga potential home, such as a hollow treewith a knothole entrance. Successful scoutsreturn to the swarm and advertise the locationof their discovery, but there is initiallymuch disagreement among scouts and differencein the quality of sites found. The swarmmust come to an agreement, however. Individualbees won’t survive on their own, norwill groups of bees without their queen. If thepolling about where to move takes too long,the swarm risks exposure to severe weatherconditions that may spell the end of the wholeendeavor. And if the group chooses a poorqualitysite, the colony will not survive thewinter. When Martin Lindauer, Nobel laureateKarl von Frisch’s most successful student,first described the debate among scoutsto his mentor, von Frisch exclaimed: “Congratulations!You have witnessed an ideal parliamentarydebate; your bees can evidentlychange their decision when other scouts haveto announce a better nesting site” ( 1). Havingdedicated decades to the study of thisunique, pluralistic decision-making process,Tom Seeley offers an engaging and fascinatingaccount of it in Honeybee Democracy.Seeley writes with infectious enthusiasm,and indeed there is much to be enthusiasticabout. When he took over the study of collectivebee decision-making from Lindauerin the 1970s, many mysteries remained.How is a cohesive relocation of several thousandworker bees and a reluctant queen thathasn’t seen daylight since her nuptial flight atleast a year earlier initiated? How do scoutsexplore the landscape and evaluate the suitabilityof cavities for nesting sites? Back atthe swarm, how do they convey the quality oftheir discoveries? How is agreement reachedin the absence of any top-down,centralized moderation of thedebate? What ensures that theconsensus converges on thebest option? What is the signalfor lift-off? Once the swarm isairborne, how can some hundredinformed scouts guide a“school-bus sized cloud,” containingseveral thousand individuals, over adistance of several kilometersto an inconspicuous knothole?Incorporating findings frominnumerable ingenious experimentsby the author, HoneybeeDemocracy includes answers toall the above questions.Some ingredients of Seeley’sapproach are worth highlighting.He strongly advocatesstarting an investigationwith the inductive (bottom-up)approach used by von Frischand Lindauer before him. Carefullyobserving your studyorganism in its natural setting,taking everything in, youget to know your study organismthoroughly from manyangles—and let unexpectedor inexplicable phenomenapop out for you. Only then doyou develop testable hypothesesand rigorous experiments to zero in onhow particular processes might be explained.Surprisingly, even though many great behavioralbiologists have adhered to this philosophy,it is now fashionable to use instead atop-down, hypothesis-driven approach: startingwith what you and pretty much everyoneelse expects and either confirming or rejectingthat. It is hard to see how one would everexplore genuinely new territory in this way.Seeley’s simple message is, keep your eyesopen for the unexpected.Another ingredient of the author’s successis the elegance and simplicity of his experiments.In times when researchers are oftenPondering collective intelligence. Tom Seeley during a 1974pilot study of a swarm choosing its home.assessed not by intellectual contributionor productivity but instead by the amountof funding they secure, Seeley shows thatcutting-edge science can be produced with“equipment obtainable from the local shoppingmall” (as Francis Ratnieks observes onthe book’s dust jacket). Indeed, many of theexperimental procedures he used are simplelike sushi (and equally exquisite). Somerequire considerable courage, however: onesimultaneously exposed Seeley to the risk offalling from a tree, being attacked by angrybees, and being killed by cyanide gas. Onesenses that he carried out experiments in thegolden age before the blight of “health andsafety” and “risk assessment.” The book isalso pleasantly free of any pretense of anapplied justification for the work. Althoughat one point Seeley inadvertently defuseda tense cold war confrontation between theUnited States and the Soviet Union by pointingout that what was thought to be chemicalweapons residue was actually bee excrement,his work was inspired entirely by thequest for scientific discovery. Funders andpolicy-makers today need reminding thatpure, blue-skies research is a component ofany healthy society.Nonetheless, the work might have considerableimplications beyond the questionof how honeybee swarms move house. Seeleyargues that human groups—juries, committees,governments—could learn fromhow bees make decisions. He suggests minimizinga leader’s influence, allowing eachgroup member to contribute their opinionsin an independent and unbiased manner, andonly reaching a single group decision once awww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 401

BOOKS ET AL.democratic quorum has been reached. Seeleyeven put these rules into practice as head ofCornell University’s Department of Neurobiologyand Behavior by minimizing his owninfluence, encouraging minority views, andmaking decisions through secret ballot.Yet we should not ignore the disanalogiesbetween human and bee decision-making.Human groups are frequently not united bycommon interest in the way that honeybeeswarms are united by shared kinship. Theformer often comprise conflicting factionseach fighting for their own self-interest. Andwhen human groups do act as cohesive units,they are often too cohesive, with their membersrarely acting as independent decisionmakerslike honeybee scouts. Conformityprevents dissenting views and conflictingevidence from being considered, often withdisastrous consequences—such as whenNASA ignored warnings that a componenton the Challenger shuttle was faulty and wentahead with its doomed launch ( 2). Moreover,whereas honeybee swarms are cooperativeby virtue of shared kinship, groups ofpeople are cooperative partly through conformity( 3). Eliminating conform ity mayeliminate decision-making errors, but it mayalso reduce the cohesiveness that maintainshuman groups in the first place.It is to Seeley’s credit that he stimulatessuch a wide-ranging debate over the similaritiesand differences in group decisionmakingamong species. In addition, HoneybeeDemocracy offers wonderful testament to hiscareer of careful investigation of a remarkablenatural phenomenon. The breadth and depthof the studies reported in it should inspire allstudents of animal behavior.References1. M. Lindauer, in Experimental Behavioral Ecology, B. Hölldobler,M. Lindauer, Eds. (G. Fisher, Stuttgart, 1985), pp.5–7.2. J. K. Esser, J. S. Lindoerfer, J. Behav. Decis. Making 2,167 (1989).3. R. Boyd, P. J. Richerson, Culture and the EvolutionaryProcess (Univ. Chicago Press, Chicago, 1985).10.1126/science.1199780BIOETHICSAn Embarrassment of RichesTim LewensLeon Kass—recently retired fromthe University of Chicago—chairedthe President’s Council on Bioethics(henceforth the Kass council) from itsestablishment by President George W. Bushin November 2001 until Kass stepped downin 2005. Kass’s own academic work is sometimesregarded by scientists and ethicists asreactionary. He is known, forexample, for his defense ofwhat he calls “the wisdom ofrepugnance.” Conservativethinkers take this work toexplain why we should heed by Adam Brigglethe instinctive aversion thatmany people feel when confrontedby new technologicaldevelopments, whether thoseinvolve the overhaul of traditionalparenting structures inthe face of new reproductive technologies orthe pharmaceutical enhancement of humanbodies and minds. For others, there is no wisdomin repugnance itself: to claim otherwiseis simply to offer a flimsy legitimation of irrationaldistaste in the face of progress.The Kass council’s reports, even morethan Kass’s own work, became, in AdamBriggle’s words, “a lightning rod for politicalcontroversy.” In particular, the councilattracted criticism from many that its membershiphad been stacked to reflect Bush’sThe reviewer is at the Department of History and Philosophyof Science, Cambridge University, Free School Lane, CambridgeCB2 3RH, UK. E-mail: tml1000@cam.ac.ukA Rich BioethicsPublic Policy, Biotechnology,and the Kass CouncilUniversity of Notre Dame Press,Notre Dame, IN, 2010. 230 pp.Paper, $30. ISBN 9780268022211.Studies in Medical Ethics.own conservative views and that it was insufficientlyattentive to the existence of disagreementamong its own members. In hisbrief and breezy A Rich Bioethics, Briggle (aphilosopher at the University of North Texas)sets out to give an account of the council’sfundamental conception of bioethics and toevaluate its performance against that conception.Lest one think the bookis an unadulterated apologia,let me say that Briggle’s verdicton the Kass council isnot uniformly positive. Heargues, for example, thatKass’s own appointment aschairman should have beenconducted in a more openmanner. He also argues thatsome of the council’s reportswere “noticeably deficientwhen it comes to conveying dissent.” Briggleaccepts that these factors justify some of thepolitical criticism Kass’s council received.However, his main goal is to vindicate thecouncil’s underlying vision of a “rich bioethics,”and he argues that in the main the councilwas true to that vision.What, then, is rich bioethics? It has anumber of strands, some very sensible, somefar more questionable. It seeks to articulatethe concerns that diverse constituencies feelwhen presented with technological innovations.It also involves a rather more sophisticatedrecognition that we cannot simply setsociety up in such a way that different individuals,with different ethical outlooks, canall have equally good chances of living in theways that they wish. The decisions that governmentmakes, Briggle argues, cannot bethoroughly neutral across varying conceptionsof the good life. This, again, is quitecorrect: Outlawing the use of cognitiveenhancingdrugs affects the ability of thosewith a commitment to a certain vision of selfimprovementto achieve their goals. If governmentpermits such drugs to be used, broadsocietal changes (in terms of norms of performance,for example) negatively affect thelives of those whose commitment to a certainconception of “genuine” achievement makesthem oppose the use of the drugs. For that reason,we need to be aware of the ways in whichregulatory choices—even apparently “liberal”policies, which allow everyone a supposedlypersonal sphere of action—can promotesome conceptions of a good life whiledisadvantaging others.Briggle’s rich bioethics becomes morequestionable once we move beyond thesefoundational points. First, he seems to implythat while a bioethics council can articulateand explain the ethical standpoints of diversegroups, it is not appropriate for such a councilto suggest pragmatic ways in which theethical concerns of all might be reasonably,or partially, met. Yet even if one is convincedby Briggle’s arguments that there is no whollyethically neutral framework for the regulationof new technologies, it does not followthat a properly rich bioethics should refrainfrom defending pragmatic political solutionsto the problem of reasonable conflict amongthose stances. Second, Briggle believes (if Iunderstand him correctly) that the Kass councilwas the first to appreciate the importanceof rich bioethics. But once we realize thatbioethics can be at the same time decisive inits policy recommendations, mindful of the40228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

BOOKS ET AL.importance of balancing diverse ethical concerns,and rich in its perception of the natureof those concerns, we will surely agree thatmany other national ethics councils (in theUnited States and elsewhere) have been justas rich in their treatment of pressing issues asKass’s was.This brings me to a final point. Briggle’sdefense of the Kass council is most strainedwhere he defends the propriety of appealingto species’ natures as ethical yardsticks. Hethinks that such natures exist and are reliablyinterpretable for all species: “The nature ofDanaus plexippus [the monarch butterfly]is seamlessly descriptive and normative, asit defines what constitutes full flourishingwithin the pattern of that kind.” If D. plexippushas a nature that defines what constitutesits flourishing, so Homo sapiens has a naturethat defines what constitutes our flourishing.This neo-Aristotelian view can then bewheeled in to justify a conservative oppositionto new technologies that purport to violateor alter that nature.Briggle moves far too quickly to justifythis very contentious ethical framework. Ifhuman nature is really “seamlessly descriptiveand normative,” then one might thinkthat by investigating human nature—i.e.,by studying biology and psychology—theanswers to fundamental ethical questionswill be made plain. And yet it seems thevery essence of ethical issues that people canunderstand the biological and psychologicalfacts—relating to cognitive enhancement,say—while disagreeing about what should bedone. Briggle would counter this by remindingus of his skepticism of the existence of anystrict distinction between descriptive factsand ethical values. He is also skeptical of theview that ethical judgment is simply a matterof expressing one’s preferences. Brigglethinks this expressivist view cannot be reconciledwith the fact that we can persuade others,on rational grounds, to change their ethicalviews. But for those who think that ethicalstandpoints can be sensitive to reasons, andeven for those who deny the fact-value distinction,Briggle’s notion of “right reasoningfrom nature” remains unclear, unless it simplymeans careful reasoning that takes intoaccount the best available evidence.Such reasoning is the sort of thing thateveryone thinks they are doing, whether theyare “transhumanists” (generally bullish aboutthe ability of technology to improve our typicalhuman capacities) or instead are the sortsof thoughtful conservatives who worry thatvaluable human institutions and practices willbe defaced by such “progress.” At the end ofBriggle’s book, one is left with a worry that theKass council had no special claim to riches.10.1126/science.1198263CREDIT: WELLCOME LIBRARY, LONDONDRUGS“A Sin, a Crime, a Vice, or a Disease?”People appear to have an almostuniversalreliance on chemical agentsto bring relief to the anxious state ofbeing human. Such mind-altering drugs arethe stuff of the latest exhibition at the WellcomeCollection in London, High Society,which continues through 27 February 2011.The exhibition offers a sober and neutralreflection on the often-controversial subject.Curators Mike Jay, CarolineFisher, and Emily Sargentguide visitors throughHigh Societyan eclectic collation of artifacts,literary matter, andartworks—one typical ofthe wide-ranging approachwe have learned to expect ofthis excellent museum.The curators organizedthe exhibition around fivetopics that cover production,trade, self-experimentation,collective experiences, andsocial questions having to dowith drug control. In additionto expected displaysof tools required for using by Mike Jayand abusing drugs and lessexpectedart installations,there are some extraordinarytreasures. Lying among thelarge leather-bound booksthat the Wellcome seems toMike Jay, Caroline Fisher,and Emily Sargent, curatorsWellcome Collection, London.Through 27 February 2011.www.wellcomecollection.org/whats-on/exhibitions/high-society.aspxHigh SocietyMind-Altering Drugsin History and Culture/The Central Role of Mind-Altering Drugs in History,Science, and CultureThames and Hudson, London,2010. 192 pp. £18.95. ISBN9780500251720. Inner Traditions,Rochester, VT. Paper,$19.95. ISBN 9781594773938.own in profusion are two tiny clay tablets,dating from before the 6th century BCE.The Babylonian one describes the medicinalplant “qunnabu” (probably cannabis)used for gynecological conditions. In anothercase lies a sheet of bluish paper covered inthin writing and attendant ink blots: SamuelT. Coleridge’s manuscript of the poetic fragment,Kubla Khan, which he claimed resultedfrom a laudanum-induceddream. Many other fascinatingthings have been packed intothe small exhibit space, makingit worth devoting at least acouple of hours to examiningthem. Finally, to lull us into aninteresting fog of introspection,the floor plan has visitorsexit through a large cubicle projectinga panorama of opiumpoppies drifting in a breeze ofsoporific music.Many of the items on displayare also illustrated in the bookthat accompanies the exhibition.There, cultural historianJay explores medical, religious,social, and economic rolespsychoactive substances playaround the world. His examplesinclude coffee, kava, betel nuts,ecstasy, hashish, opium, andtobacco. He ranges from theA popular pick-me-up. Dudley Hardy’s ad (circa1916) for “a marvelous restorative.”use of botanicals in antiquity through the selfexperimentationof early scientists to possibledesigner neurochemicals. And he discussesthe far-reaching consequences of internationaltrade in intoxicants, prohibitions, andthe “war on drugs.” Whether or not they areable to visit the exhibition, readers will findthe book provides an excellent overview ofthe cultural history of drugs. –Caroline Ash10.1126/science.1202766www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 403

EDUCATIONFORUMSCIENCE EDUCATIONDefeating Creationism in theCourtroom, But Not in the ClassroomMichael B. Berkman and Eric Plutzer *Sixty percent of U.S. high school biology teachersare not advocates for either evolutionary biologyor nonscientific alternatives.Just over 5 years ago, the scientific communityturned its attention to a courtroomin Harrisburg, Pennsylvania. Eleven parentssued their Dover, Pennsylvania, schoolboard to overturn a policy explicitly legitimizingintelligent design creationism. Thecase, Kitzmiller v. Dover, followed a familiarscript: Local citizens wanted their religiousvalues validated by the science curriculum;prominent academics testified to the scientificconsensus on evolution; and creationistslost decisively. Intelligent design was notscience, held the court, but rather an effort toadvance a religious view via public schools,a violation of the U.S. Constitution’s EstablishmentClause ( 1). Many scientists cheeredthe decision, agreeing with the court that theschool board displayed “breathtaking inanity”[p. 765 ( 1)]. We suggest that the cheeringwas premature and the victory incomplete.Systematic Undermining of ScienceCreationism has lost every major U.S. federalcourt case for the past 40 years, andstate curricular standards have improved ( 2).But considerable research suggests that supportersof evolution, scientific methods, andreason itself are losing battles in America’sclassrooms, where instruction in evolutionarybiology “has been absent, cursory, orfraught with misinformation” [p. 21 ( 3), and(4)]. Extending this research, we have beeninvestigating the evolution-creationism battlein state governments ( 5) and the nation’sclassrooms ( 2, 6). Central to this research isthe National Survey of High School Biologyteachers, based on a nationally representativeprobability sample of 926 public highschool biology instructors ( 2, 6) (see the figure).[See supporting online material (SOM)for details.] The data reveal a pervasive reluctanceof teachers to forthrightly explain evolutionarybiology.The data further expose a cycle of ignorancein which community antievolutionattitudes are perpetuated by teaching thatreinforces local community sentiment. Forexample, we ranked school districts fromDepartment of Political Science, The Pennsylvania StateUniversity, University Park, PA 16802, USA.*Author for correspondence. E-mail: Plutzer@psu.eduleast to most socially conservative, and in the15% most socially conservative school districts,nearly 4 in 10 teachers personally donot accept human evolution (compared with11% in the least conservative districts) and,consequently, devote only minimal time toevolutionary biology in their classes [tableA8.2 in ( 2)]. The next generation of adults isthus predisposed to share the antievolutionviews of their parents.More promising data suggest that America’shigh schools contain thousands of outstanding,effective educators of evolutionarybiology. We estimate that 28% of allbiology teachers consistently implement themajor recommendations and conclusionsof the National Research Council ( 7): Theyunabashedly introduce evidence that evolutionhas occurred and craft lesson plans sothat evolution is a theme that unifies disparatetopics in biology ( 2).At the opposite extreme are 13% of theteachers surveyed who explicitly advocatecreationism or intelligent design by spendingat least 1 hour of class time presentingit in a positive light (an additional 5% ofteachers report that they endorse creationismin passing or when answering studentquestions). The boldness and confidence ofthis minority should not be underestimated.Although 29% percent of all other teachersreport having been “nervous at an openhouse event or meeting with parents,” only19% of advocates of creationism report this(χ 2 = 5.1, P = 0.024).Advocate ofEvolutionary biology(28% of all teachers)Neither(60% of all teachers)Creationism(13% of all teachers)7%10%26%33%37%Percentage of each group who completed a course on evolutionPercentage of each group rating themselves exceptionalSome advocates of creationism insistedthat they—not bench scientists—are theones practicing proper science: A Minnesotateacher commented, “I don’t teach the theoryof evolution in my life science classes, nor doI teach the Big Bang Theory in my [E]arth[S]cience classes…. We do not have time todo something that is at best poor science.”Others rejected the possibility that scientificmethods can shed light on the origin of species.An Illinois teacher responded, “I amalways amazed at how evolution and creationismare treated as if they are right orwrong. They are both belief systems that cannever be truly or fully proved or discredited.”The Cautious 60%But if mainstream science and the moderncreationist movement each have their classroomallies, they still account for only about40% of all high school biology teachers.What of the majority of teachers, the “cautious60%,” who are neither strong advocatesfor evolutionary biology nor explicit endorsersof nonscientific alternatives?Our data show that these teachers understandablywant to avoid controversy. Oftenthey have not taken a course in evolution andthey lack confidence in their ability to defendit (see the figure, see SOM for details). Theirstrategies for avoiding controversy are varied,but three were especially common and eachhas the effect of undermining science ( 8).Some teach evolutionary biology as thoughit only applies to molecular biology—completelyignoring macroevolutionof species. At56%Self-reports of qualifications of teachers, classified by approach toteaching evolution. Based on responses from 926 U.S. public high schoolbiology teachers. See SOM for survey details.best, this approach sacrificesa rich understandingof the diversity ofspecies. At worst it lendscredence to the creationistclaim that there is noevidence for one speciesgiving rise to others.Others defend theteaching of evolutionas a necessary evil,using state examinationrequirements as a convenientmeans to disassociatethemselves from40428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

EDUCATIONFORUMthe very material they are expected to teach.These examinations have only been recentlyintroduced in most states. Yet, many teacherstold us that they tell students that it doesnot matter if they really “believe” in evolution,so long as they know it for the test. OneMichigan teacher tells students that they needto understand evolution because the biologycurriculum “is organized as if evolution istrue” [emphasis added].Finally, a sizable number of teachersexpose their students to all positions—scientificor not. Students should make uptheir own minds, explained a Pennsylvaniateacher, “based on their own beliefs andresearch. Not on what a textbook or on what ateacher says.” Many of these teachers mighthave great confidence in their students’ability to learn by exploration. But doesa 15-year-old student really have enoughinformation to reject thousands of peerreviewedscientific papers? This approachtells students that well-established conceptslike common ancestry can be debated in thesame way we debate personal opinions.The cautious 60% may play a far moreimportant role in hindering scientific literacyin the United States than the smaller numberof explicit creationists. The strategies ofemphasizing microevolution, justifying thecurriculum on the basis of state-wide tests, or“teaching the controversy” all undermine thelegitimacy of findings that are well establishedby the combination of peer review and replication.These teachers fail to explain the natureof scientific inquiry, undermine the authorityof established experts, and legitimize creationistarguments, even if unintentionally.Courts, Standards, Preservice TeachersBiology will be the only high school scienceclass for 21 to 25% of U.S. high school graduates,and more high school students take generalbiology than any other science course ( 9).But many are not afforded a sound scienceeducation, which is problematic in a democracydependent on meaningful citizen inputon highly technical, but consequential, publicpolicies. Research suggests several waysthat scientists and scientific organizations canaddress this situation.First, continued participation in federallaw suits is essential, as federal courts havebeen shown to limit effectively the ability ofstate and local governments to endorse nonscientificalternatives to evolution ( 5). Likewise,the active role of scientists and scientificorganizations has improved curricularstandards in many states, and such reformhas the potential to be especially effectivein states having high-stakes science tests(2). But change due to improved standardsis likely to be slow, because standards havethe greatest impact on the newest teachers—those who were socialized in an era of standards-basededucation and who take standardsand testing for granted ( 2). In addition,further improvements in state standards maybe difficult because public opinion has beenremarkably immune to outreach and publicscience efforts over the past three decades(10).We suggest that increased focus be placedon preservice teachers (i.e., those preparingto be, but not yet, teachers). Teachers who areadvocates for evolutionary biology are morelikely to have completed a course in evolutionthan teachers who are ambivalent about evolutionor who teach creationism (see the figure).Indeed, completing an evolution courseis a powerful predictor of the classroom timedevoted to evolution ( 6, 11) and the likelihoodthat teachers will integrate evolutioninto their class as a unifying theme ( 2). Manynonresearch institutions lack the resources tooffer a stand-alone evolution course regularly,however, and such institutions educate manyhigh school science teachers. Requiring anevolution course for all preservice biologyteachers, as well as provision of resourcesto provide such a course, would likely leadto meaningful improvement in secondaryschool science instruction.In addition to their relative lack of evolutioncoursework, teachers in the ambivalentmiddle 60% also resemble those who endorsecreationism in that few believe that they havean exceptional understanding of evolutionarybiology (see the figure). Yet, unlike creationists,few of these ambivalent teachers hold ayoung-Earth belief system (e.g., that the universeis only about 10,000 years old) thatwould prevent them from becoming strongadvocates for evolutionary biology. Therefore,improving the instruction they receivein evolution as undergraduates is essential.Outreach efforts such as webinars, guestspeakers, and refresher courses—the types ofefforts currently aimed at secondary schoolteachers—could be tailored and targeted forboth preservice teachers and for biology andscience education professors at teaching-orientedcolleges. This two-pronged effort mayhelp increase the percentage of new teacherswho accept and embrace the findings ofevolutionary biology. Better understanding ofthe field should provide them with more confidenceto teach evolution forthrightly, evenin communities where public opinion is sympatheticto creationism.More effectively integrating evolution intothe education of preservice biology teachersmay also have the indirect effect of encouragingstudents who cannot accept evolution as amatter of faith to pursue other careers. Effectiveprograms directed at preservice teacherscan therefore both reduce the number ofevolution deniers in the nation’s classrooms,increase the number who would gladly accepthelp in teaching evolution, and increase thenumber of cautious teachers who are neverthelesswilling to embrace rigorous standards.This would reduce the supply of teacherswho are especially attractive to the mostconservative school districts, weakening thecycle of ignorance.Outreach efforts primarily benefit teacherswho want to be helped, so expanding thecorps of science teachers who want to behelped is critical. Thus, focusing on the preservicestage may be “the most effective wayfor scientists to help to improve the understandingof evolution” [p. 332 ( 12)]. Bettertrainedteachers will be able to more effectivelytake advantage of details in their textbooksand supplementary material publishedby the National Academy of Sciences and toput aside fear of reactions and pressures frommembers of their communities. It would alsomake them more critical advocates for highqualitystandards and textbooks. Combinedwith continued successes in courtrooms andthe halls of state government, this approachoffers our best chance of increasing the scienceliteracy of future generations.References and Notes1. Kitzmiller v. Dover, 400 F. Suppl. 2d 707 (M.D. Pa 2005).2. M. B. Berkman, E. Plutzer, Evolution, Creationism, andthe Battle to Control America’s Classrooms (CambridgeUniv. Press, Cambridge, 2010).3. M. L. Rutledge, M. A. Mitchell, Am. Biol. Teach. 64, 21(2002).4. R. Moore, Bioscience 52, 378 (2002).5. M. B. Berkman, E. Plutzer, Perspect. Polit. 7, 485 (2009).6. M. B. Berkman et al., PLoS Biol. 6, e124 (2008).7. National Research Council, National Science EducationStandards (National Academy Press, Washington, DC,1996).8. These three strategy categories are based on answers toopen-ended questions, which were answered voluntarily.This limits our ability to quantify how broadly such strategiesare used (a goal for future work).9. National Center for Education Statistics, U.S. Departmentof Education, National Assessment of Educational Progress,2005 High School Transcript Study (HSTS),http://nces.ed.gov/nationsreportcard/hsts/.10. E. Plutzer, M. B. Berkman, Public Opin. Q. 72, 540 (2008).11. L. A. Donnelly, W. J. Boone, J. Res. Sci. Teach. 44, 236(2007).12. G. Branch, E. C. Scott, J. Rosenau, Annu. Rev. GenomicsHum. Genet. 11, 317 (2010).13. Supported by the U.S. NSF, the Spencer Foundation,and the John Templeton Foundation. All conclusions arethe responsibility of the authors and do not necessarilyreflect the views of the any funding organization.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/404/DC110.1126/science.1198902www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 405

PERSPECTIVESCOMPUTER SCIENCECloud Computing—What’s in Itfor Me as a Scientist?Armando FoxComputational tasks that are inherentlyparallel, from simulations to studentassignments, can be run faster on the datacenter resources of public clouds.Many scientists would loveaccess to large-scale computationalresources butfind that the programming demands ofusing a supercomputer—as well as thecost and queuing time—are too daunting.Privately owned cloud computers—largedata centers filled with computersthat mainly run their company’ssoftware—are now becoming availableto outside users, including scientistsand educators. Companies are leasingtheir computing resources on demandfrom a large shared pool to individualswho run their own software on a payas-you-gobasis. This approach is anexample of cost associativity ( 1): 1000computers used for 1 hour costs thesame as one computer used for 1000hours. If your problem can be computedin a way that takes advantage ofparallel processing, you can now getthe answer 1000 times as fast for thesame amount of money.Although companies had long beenoperating “private clouds” that runprograms such as Google Search orMicrosoft Hotmail, Amazon was thefirst to let outside users run software on theircomputers. For example, Amazon’s ElasticCompute Cloud (EC2), announced in late2007, allows anyone with a credit card to useany number of computers in Amazon’s datacenters for 8.5 cents per computer-hour withno minimum or maximum purchase and nocontract. Such an arrangement is possiblebecause these “warehouse-scale” data centers(~50,000 servers, see the figure) are fiveto seven times cheaper to build and operatethan smaller facilities (~1000 servers) ( 2), interms of network, storage, and administratorcosts. When operational costs, which includehuman administrators, power, and networking,are considered, the charges for cloudcomputing to outside users is price-competitivewith using in-house facilities.Cost-associativity enables new capabilities.For example, a researcher in our laboratorycreated an automated classifier to detectDepartment of Computer Science, University of California,Berkeley, CA 94720, USA. E-mail: fox@cs.berkeley.eduCommodity computing. This view of aMicrosoft data center is a tiny fraction of theservers available for cloud computing.spam on the popular social-communicationsite Twitter.com. Training the classifier tookabout 270 hours on a typical desktop workstation.The training program could be parallelized:The problem could be broken intopieces and run at the same time, only occasionallysharing results between differentpieces. The same task took 3 hours usingabout 100 servers in Amazon’s cloud (about$250 in usage fees). Many universities havealso begun to use cloud computing for education,where cost-associativity is a great fit forsemester courses. Lots of computing demandcould be provided around assignment deadlines(more than even the biggest schoolscould provide), and when demand is less(between deadlines), no outside resourcesneed to be purchased.Initially, cloud-computing hardware wasconfigured primarily for its earliest adopters—Web-basedapplications—and earlyattempts to run scientific applications on thecloud gave discouraging results ( 3, 4). Newhardware is now configured for better performanceon scientific applications.For example, Amazon’s recently added“cluster computing instances,” pricedat $1.60 per computer-hour, run scientificbenchmarks 8.5 times as fast asthe original cloud hardware, accordingto experiments at the National EnergyResearch Scientific Computing Laboratoryat Lawrence Berkeley NationalLaboratory ( 5).Cloud computing works best whena problem can be broken down into alarge number of relatively independenttasks, each running on its owncomputer. Software frameworks likeGoogle’s MapReduce ( 6) and its opensourceequivalent Hadoop ( 7) providea data-parallel “building block” forexpressing such computations (muchlike a Web design framework allowsyou to “build” a Web site by filling inthe relevant information and functionsyou want). Critically, these frameworksalso hide the complex softwaremachinery that handles inevitable transientmachine failures when hundreds of machinesin a cloud environment work on a problemsimultaneously. Many of the “success stories”of science in the cloud have embracedHadoop, and other popular tools such as thestatistical package R now feature librariesthat integrate with it. However, many problemscannot be easily expressed in terms ofmap and reduce tasks (mapping parcels outthe work, and reduce collates the results).Even when they can, the programming effortrequired may be substantial.Most desktop software is not writtento take advantage of cloud computing andrequires modification before it could harnesscloud resources and run faster. However,the popular packages MATLAB andMathematica are now available in versionsthat can “farm out” work to a public cloud.Cloud vendors including Amazon and IBMare working with independent software vendorson cloud-friendly versions of popularscientific software.CREDIT: MICROSOFT CORPORATION40628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

PERSPECTIVESWhat about software designed for parallelsupercomputers? Programs written tothe MPI (message-passing interface) standardrun in the cloud, but the performance ofsuch programs is sensitive to whether all themachines run in tight lockstep when workinga problem (a characteristic of supercomputersbut one that is not necessarily part of thecloud-computing architecture). Still, cloudtechnology suppliers such as Intel, AdvancedMicro Devices, and VMware are addinghardware and software features necessaryto improve MPI performance in the cloud, amove that highlights new “buying power” inthe scientific computing community.Many scientific computing problems donot require supercomputer performance butwould benefit greatly from modest parallelism—say,tens or hundreds of computers inthe cloud. The total “time to answer” may stillbe quicker, according to Foster ( 8), even forlarger problems because a cloud-based supercomputercan be provisioned and running inminutes rather than hours or days, withoutwaiting in a queue.Some problems are so data intensive thatthey demand large-scale computing. TheLarge Synoptic Survey Telescope in Chilemay generate up to 30 terabytes of datadaily. At typical long-haul network speeds ofaround 20 megabits per second ( 9), 30 terabyteswould take more than 4 months and$3000 in network charges to copy to Amazon’scloud. Many cloud providers now allowusers to ship crates of hard drives to be physicallyincorporated into the cloud infrastructure,an idea championed by the late Jim Gray.The lure of improved performance hasalready drawn scientists and engineers touse cloud computing ( 10) in research on anumber of topics, including large-populationgenetic risk analysis, information retrieval,and particle physics. Cloud computing installationsare also being established by academic-industrialconsortia [see, for example,(11–13)] to further encourage adoption ofcloud computing by scientists.References and Notes1. M. Armbrust et al., Commun. ACM 53, 50 (2010).2. J. Hamilton, Presentation at 2nd Large-Scale DistributedSystems and Middleware (LADIS. Inf. Serv.) Workshop, 15to 17 September 2008, White Plains, NY.3. C. Evangelinos, C. N. Hill, First ACM Workshop on CloudComputing and Its Applications (CCA’08), 22 to 23 October2008, Chicago, IL.4. E. Walker, login Magazine 33(5), 18 (2008). Availableonline at www.usenix.org/publications/login/2008-10/openpdfs/walker.pdf.5. www.lbl.gov/cs/Archive/news071310.html6. J. Dean, S. Ghemawat, Proceedings of the 6th USENIXSymposium on Operating Systems Design and Implementation(OSDI ’04), 5 to 8 December 2004, San Diego, CA.7. http://hadoop.apache.org8. I. Foster, What’s faster—A supercomputer or EC2?http://ianfoster.typepad.com/blog/2009/08/whatsfastera-supercomputer-or-ec2.html.9. http://aws.amazon.com/solutions/case-studies10. S. Garfinkel, Technical Report TR-08-07, Harvard University(2007). Available online at http://simson.net/clips/academic/2007.Harvard.S3.pdf.11. http://labs.yahoo.com/Cloud_Computing12. www.nsf.gov/news/news_summ.jsp?cntn_id=11468613. https://opencirrus.org14. Much of what is reported here builds on a survey ofBerkeley cloud-computing research (1) supported bythe National Science Foundation (grant CNS-0509559),Sun Microsystems, Google, Microsoft, Amazon WebServices, Cisco Systems, Cloudera, eBay, Facebook,Fujitsu, Hewlett-Packard, Intel, Network Appliance, SAPComputer Solutions, VMWare, and Yahoo! Inc., as wellas matching funds from the State of California MICROProgram (grants 06-152, 07-010, 06-148, 07-012,06-146, 07-009, 06-147, 07-013, 06-149, 06-150, and07-008) and the University of California Industry/UniversityCooperative Research Program (UC Discovery) grantCOM07-10240. I also thank M. Marr, J. Shalf, G. Bell, andJ. Simons for their comments and suggestions.10.1126/science.1198981GENOMICSThe Genomic View of BacterialDiversificationBy comparing whole genome sequences,researchers reconstruct the evolution andglobal spread of an antibiotic-resistant strain.1Biocontrol Ltd., Sharnbrook, Bedfordshire MK44 1LQ, UK.2Imperial College, London SW7 2AZ, UK. E-mail: b.spratt@imperial.ac.ukMark C. Enright 1 and Brian G. Spratt 2Bacteria have unusual and variable sexlives, ranging from near-celibacy toevident promiscuity. “Sex” in bacteriainvolves a recipient bacterium replacing,by homologous recombination, smallregions of its genome with correspondingregions from other strains ( 1). Studies overthe past 20 years have demonstrated that inrelatively promiscuous bacterial species,frequent recombination can drive rapiddiversification of strains. In more celibatespecies, diversification is much slower anddepends on the slow accumulation of pointmutations ( 1, 2). Sorting out the roles thatthese processes play in the pace and patternof bacterial evolution, however, has provenproblematic. On page 430 of this issue,Croucher et al. ( 3) demonstrate a powerfulnew approach to the problem. By comparingthe genomes of many isolates of a singlestrain of a bacterium that causes pneumonia,they were able to rapidly obtain acomprehensive evaluation of the role thatrecombination, point mutation and othergenetic processes played in its diversification.The approach offers insight into howthis pathogenic strain can rapidly evolveresistance to antibiotics and evade futurevaccines, and promises to help researchersbetter understand how this resistant strainmay have spread globally.It has been difficult to estimate the ratesand patterns of recombinational replacementsby comparing the genomes of distantlyrelated strains of bacterial species,particularly when recombination has beenfrequent. Similarly, it has been problematicto identify those parts of the genome thathave not been involved in recombinationand can provide reliable information aboutphylogeny (the bacteria’s evolutionary history).Most of these problems can be overcome,however, by comparing the genomesof isolates that have a very recent commonancestor, such as isolates of an antibioticresistantstrain of a bacterial pathogen,which must have emerged within the antibioticera. These isolates will have accumulateda relatively small number of mutationsand recombinational replacements, allowingresearchers to more easily determinethe number and size of the replacements.The ability to construct a reliable phylogeny,by identifying variation due to pointmutations, is also greatly simplified. Differencesamong isolates can be mapped ontoa tree and correlated with genomic differences,such as shifts in antibiotic resistanceprofiles or changes in virulence or antigenstargeted by newly introduced vaccines.Sequence data can provide informationabout the dynamics and geographic spreadof pathogens. Such an analysis of veryrecent events, however, requires a pathogenthat evolves rapidly enough to enable suffi-www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 407

PERSPECTIVESPMEN1Spain 1984Diversification of PMEN1 clonecient mutations to accumulate over the timescale of interest ( 4)—for example, duringthe rapid global spread of a new influenzapandemic ( 5). This requirement has largelylimited the use of powerful new analyticalmethods to RNA viruses, which have veryhigh mutation rates (4–7). Mutation rates inbacteria have been considered to be far toolow to apply these new “phylodynamic” and“phylogeographic” approaches to reconstructingthe geographic spread of a pathogenicstrain over a matter of a few years ordecades. Several research groups, however,have recently found that short-term mutationrates in various bacterial pathogensare much higher than previously believed(about 10 −6 substitutions per site per year)(8–11). Although these rates are still muchlower than those of RNA viruses, the muchlarger size of bacterial genomes, combinedwith the relatively high mutation rates,should allow researchers to apply the analyticalmethods pioneered for RNA virusesto bacterial pathogens, if whole genomesequences are used.The ease with which investigators canUK 1990USA 2000Vietnam 2005History of an antibiotic-resistant strain. Since its origin around 1970, the multiply antibiotic-resistantPMEN1 (pneumococcal molecular epidemiology network 1) clone of S. pneumoniae has diversified rapidly.PMEN1 is usually identified by multilocus sequence typing (MLST), which characterizes isolates using thesequences of seven genes. Most isolates of PMEN1 retain the ancestral sequences (central blue circle, left),but many isolates have arisen that differ in the sequence of one of the seven genes (around central bluecircle, left) and two differ at two of the genes. Analyzing the genome sequences of 240 isolates of PMEN1provides a comprehensive picture of this rapid genetic diversification of the antibiotic-resistant strain over40 years. The genomes of four isolates of PMEN1 are represented (blue circles, right). Point mutations (singlesmall white circles) and recombinational replacements (linked colored circles) are shown.now obtain the whole genomes of bacterialpathogens is opening up a number ofquestions that previously were impossibleor difficult to address. One of these is howvirulent or highly drug-resistant strains ofbacterial pathogens spread within hospitalsor between hospitals and nursing homeswithin a region. Not surprisingly, the firstapplication of the genome sequencing oflarge numbers of isolates of a single straininvolved a study of the spread of a prevalentmethicillin-resistant Staphylococcus aureus(MRSA) strain ( 11), and the preliminaryresults are encouraging.Croucher et al. used large-scale genomesequencing to explore the evolutionarydynamics of a single multidrug-resistantstrain of the community-acquired pathogen,Streptococcus pneumoniae. By sequencing240 isolates, obtained from 22 countriesbetween 1984 and 2008, they were able totrack how the pathogen evolved over 40years as it spread from continent to continent(see the figure). In this species, recombinationis frequent: 74% of the genomehad received a recombinational replacementin at least one isolate. The genomewideapproach provides the first indicationof the broad size distribution of theserecombinational replacements, with 5%replacing more than 30 kb of the genome.Some of these replacements are responsiblefor changes in the bacteria’s capsular serotypeand involve replacement of the largecapsular polysaccharide biosynthetic locus.These changes almost certainly reflectselection by the human immune responseor, recently, by the introduction of conjugatevaccines, which target a small subsetof the 90+ known polysaccharide capsules.Several chromosomal regions were frequentlyreplaced by recombination, includingthe pspA and pspC genes that encodesurface antigens being developed as vaccinecandidates.By removing the recombinationalreplacements from the genomic data set, theresearchers identified polymorphisms thatthey used to construct a phylogeny, allowingthem to map the dynamics of the acquisitionand loss of antibiotic resistance and of capsularserotype changes. Phylogeographicanalysis supported a view that this multiresistantstrain originated in Europe about 40years ago, which is consistent with its initialrecognition in Spain in the 1970s.Overall, the large-scale genomicapproach dramatically shows the remarkableability of this pathogen to evolve oververy short time scales by frequent recombination(and other processes). The consequencesof such rapid evolution for theemergence and spread of antibiotic resistanceand the evasion of vaccines targetedat single-protein antigens are clear.Another striking indication of the abilityof S. pneumoniae to evolve rapidly isprovided by a recent study that sequencedthe genomes of six isolates obtained over7 months from the nasopharynx of a singlechild ( 12). During this period, the genomeof a strain that colonized the child receivedan estimated 23 recombinational replacements,many of which were derived froma co-colonizing strain, resulting in thereplacement of 7.8% of the genome ( 12).Previous studies have already reachedmany of Croucher et al.’s conclusions(13, 14). Suggesting that we knew all thisbefore, however, misses the importanceof their study, in which a single experimentprovided more information than hasbeen achieved over 15 years of research. Incontrast to previous work, which involvedthe analysis of a few genes per isolate,Croucher et al.’s genomic approach providesthe complete story of how bacteria40828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

PERSPECTIVESdiversify over the very short time scales thatare of crucial importance for understandingand predicting the response of pathogens tonew antibiotics and vaccines. It representsa new and powerful approach to obtaininga mass of information about the populationand evolutionary biology of any bacterial(or archaeal) species at a modest and rapidlyfalling cost. Exciting times lie ahead for theintegration of bacterial genome data, epidemiology,and evolution.References1. J. M. Smith, C. G. Dowson, B. G. Spratt, Nature 349, 29(1991).2. E. J. Feil, B. G. Spratt, Annu. Rev. Microbiol. 55, 561(2001).3. N. J. Croucher et al., Science 311, 430 (2011).4. O. G. Pybus, A. Rambaut, Nat. Rev. Genet. 10, 540(2009).5. C. Fraser et al., WHO Rapid Pandemic Assessment Collaboration,Science 324, 1557 (2009).6. E. C. Holmes, B. T. Grenfell, PLOS Comput. Biol. 5,e1000505 (2009).7. P. Lemey, A. Rambaut, A. J. Drummond, M. A. Suchard,PLOS Comput. Biol. 5, e1000520 (2009).8. L. Feng et al., PLoS ONE 3, e4053 (2008).9. G. Morelli et al., PLoS Genet. 6, e1001036(2010).10. D. J. Wilson et al., Mol. Biol. Evol. 26, 385 (2009).11. S. R. Harris et al., Science 327, 469 (2010).12. N. L. Hiller et al., PLoS Pathog. 6, e1001108(2010).13. E. J. Feil, J. M. Smith, M. C. Enright, B. G. Spratt,Genetics 154, 1439 (2000).14. T. J. Coffey et al., Mol. Microbiol. 27, 73 (1998).15. B.G.S. acknowledges funding from the Wellcome Trust.10.1126/science.1201690MICROBIOLOGYA Tail of DivisionAlan F. Cowman and Christopher J. TonkinA protein that controls the entry of Toxoplasmagondii into a host cell also plays a role incontrolling the parasite’s replication.CREDIT: Y. HAMMOND/SCIENCEAcommon mechanism of pathogensurvival is to invade and replicatewithin the protective niche of a hostcell. This is true for apicomplexan parasites, agroup of medically and agriculturally importantpathogens that includes the malariacausingPlasmodium spp., and Toxoplasmagondii, the causative agent of toxoplasmosis.An apicomplexan parasite penetrates thehost cell plasma membrane and replicatesonly after invasion is complete. Althoughthere is some understanding ofhow these pathogens recognizeand invade host cells, little hasbeen known about the molecularsteps that initiate replication.On page 473 of this issue, Santoset al. ( 1) show that invasionand replication are inextricablylinked through the function of akey apicomplexan parasite protein,Apical Membrane Antigen1 (AMA1) ( 2–4).Experiments with the moregenetically tractable T. gondiihave shown that AMA1 isreleased from the parasite’smicroneme organelles onto itssurface. This is required for thesubsequent release of lipid-richrhoptry organelles into the hostcell to form the parasitophorousvacuole that harbors and protectsthe pathogen from degradation( 2). AMA1 functions inWalter and Eliza Hall Institute of MedicalResearch, Melbourne, Victoria 3052, Australia,and Department of Medical Biology,University of Melbourne, Melbourne, Victoria3052, Australia. E-mail: cowman@wehi.edu.au; tonkin@wehi.edu.auinvasion by binding to proteins (the RONcomplex) that the parasite inserts into thehost cell plasma membrane (see the figure)(5). The AMA1-RON complex is associatedwith the “moving junction,” a contact point ofthe host cell and parasite plasma membranesthrough which the parasite gains traction topull itself into the host cell.AMA1 has a single transmembranedomain and a short, highly conserved cytoplasmictail. It is thought that the cytoplasmicHost cell invasion Parasite replication Host cell lysisParasitic RONcomplexRON4RON8RON5RON2ROM4AMA1CleavageToxoplasmagondiiHost cell cytoplasmReleaseNucleartranslocationReplicative switch(transcription)InvasionSignaltransductioncascadeRhoptriesMicronemesNucleustail could inform the parasite of contact witha host cell, thus triggering rhoptry releaseand activation of the parasite’s actomyosininvasion motor ( 2, 4, 6, 7). Indeed, the cytoplasmictail of AMA1 in Plasmodium playsa role in the invasion of erythrocytes ( 4),and its phosphorylation in P. falciparum isimportant in this function ( 6).Cleavage of AMA1 is thought to berequired for the parasite to release the hostcell plasma membrane and complete invasion.In both T. gondii and Plasmodium,serine proteases calledrhomboids ( 8) cleave AMA1(and other parasite proteins thatfunction in adhesion to the hostcell) distal to the membrane, aprerequisite for successful invasion( 9, 10). Indeed, the absenceof Rhomboid 4 (ROM4) in T.gondii prevents shedding of theparasite’s surface proteins, anda decrease in ROM4 functioninterferes with the apical-poste-MovingjunctionLinking invasion and replication.To invade a host cell, AMA1 protein issecreted by the parasite T. gondii andinserted into the pathogen’s cell surfacemembrane. It binds to the parasite-derivedRON complex (RON2,RON5, RON4, and RON8) that hasbeen inserted into the host cell plasmamembrane. The protease ROM4 islocated in the parasite’s plasma membrane,where it cleaves AMA1 in thetransmembrane region, thus releasingthe AMA1 tail into the cytoplasm.The AMA1 cytoplasmic tail may interactwith factors that trigger a signalingcascade, and/or itself translocate intothe nucleus, to initiate the replicativetranscriptional program.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 409

PERSPECTIVESrior distribution of cell surface proteins thatinteract with host cell surface proteins, thusimpairing the invasion process ( 9).Santos et al. reveal another critical functionof AMA1. They show that cleavage ofAMA1 by ROM4 in T. gondii releases thecytoplasmic tail of AMA1, which then providesa signal for the invading parasite tobegin replication inside the host cell. Thefinding indicates that invasion and replicationin apicomplexan parasites are intimatelylinked. The authors demonstrate thisimportant finding by constructing a T. gondiiparasite that expressed a catalytically inactiveform of ROM4 (ddROM4 S-A) ( 11). Themutant protein acted as a dominant negativeand inhibited host cell invasion, confirmingother work ( 9). However, the main effect ofddROM4 S-Aexpression was to block parasitereplication. These parasites were blocked latein their cell division cycle (during S phase).Santos et al. hypothesized that the AMA1cytoplasmic tail released by ROM4 initiatedasexual replication. Expression of the cytoplasmictail rescued the cell division phenotypein parasites expressing the inactiveform of ROM4, thereby allowing parasites todevelop normally. Expression of the AMA1cytoplasmic tail of P. falciparum also rescuedthe aberrant cell division phenotype, showingthat this effect of AMA1 is a general functionin apicomplexan parasites. A serine residue inthe AMA1 cytoplasmic tail of P. falciparum isrequired for invasion and is phosphorylated(4, 6). Santos et al. found that mutation of thisserine did not affect parasite cell division,indicating that the cytoplasmic tail of AMA1has dual functions in invasion and initiationof the replicative phase of its life cycle.AMA1 and ROM4 are both on the surfaceof extracellular parasites and at this stage, replicationis suppressed. This suggests that alevel of control must exist to govern the switchto replication only upon invasion. Could theactivity of ROM4 be regulated to preventAMA1 cleavage? This seems unlikely, asROM4 activity in extracellular parasites hasbeen inferred previously ( 9). Furthermore,Santos et al. also found that the AMA1 cytoplasmictail alone cannot induce replication ofextracellular parasites, suggesting that otherregulatory mechanisms must exist.How does AMA1 promote replication?Possibilities include initiation of a signalingcascade, or directly influencing transcriptionmachinery upon translocation into thenucleus. The identity of factors that associatewith the AMA1 tail to drive two processes—replication and invasion—will be of greatinterest in the future.References1. J. M. Santos, D. J. P. Ferguson, M. J. Blackman,D. Soldati-Favre, Science 331, 473 (2011); 10.1126/science.1199284.2. J. Mital, M. Meissner, D. Soldati, G. E. Ward, Mol. Biol.Cell 16, 4341 (2005).3. T. Triglia et al., Mol. Microbiol. 38, 706 (2000).4. M. Treeck et al., PLoS Pathog. 5, e1000322 (2009).5. S. Besteiro, A. Michelin, J. Poncet, J. F. Dubremetz,M. Lebrun, PLoS Pathog. 5, e1000309 (2009).6. K. Leykauf et al., PLoS Pathog. 6, e1000941 (2010).7. D. Richard et al., J. Biol. Chem. 285, 14815 (2010).8. S. Urban, M. Freeman, Mol. Cell 11, 1425 (2003).9. J. S. Buguliskis, F. Brossier, J. Shuman, L. D. Sibley, PLoSPathog. 6, e1000858 (2010).10. R. A. O’Donnell et al., J. Cell Biol. 174, 1023 (2006).11. A. Herm-Götz et al., Nat. Methods 4, 1003 (2007).10.1126/science.1201692CELL BIOLOGYWhy Starving Cells Eat ThemselvesD. Grahame HardieCollege of Life Sciences, University of Dundee, ScotlandDD1 5EH, UK. E-mail: d.g.hardie@dundee.ac.ukIn the 1950s, electron microscopistsobserved regions of cytoplasm and organellesbeing engulfed by a double membraneto form vesicles that then fused withlysosomes ( 1). This process, called autophagy(“self-eating”), was correctly deduced tobe a mechanism to degrade and recycle thecontents of nutrient-starved cells ( 2). Sincethen, signaling pathways that sense nutrientdeprivation and energy stress have been identified,but how these interact with autophagyhas not been clear. On page 456 in this issue,Egan et al. ( 3) show that the adenosine monophosphate–activatedprotein kinase (AMPK),an energy sensor that is also activated by glucosestarvation, triggers autophagy by phosphorylatingULK1, establishing a directmolecular link between the two processes.The autophagy gene ATG1 (originallyAPG1) was identified in the yeast Saccharomycescerevisiae through mutations thatled to failure of autophagic vesicles to accumulatewhen cells were starved ( 4). The twoorthologs of Atg1 in mammals, ULK1 andULK2, are the catalytic subunits of multisubunitprotein kinases (also containing ATG13and FIP200) that initiate the autophagic cascade.A key cellular nutrient sensor is thetarget of rapamycin complex–1 (TORC1),which is activated by amino acids via theRag guanosine triphosphatases (GTPases)(5). Conversely, amino acid starvation inhibitsTORC1 and promotes autophagy, as doesthe TORC1 inhibitor rapamycin. TORC1 isinhibited by AMPK ( 6, 7), and it has beengenerally assumed that this is how energystress or glucose starvation promotes autophagy.However, Egan et al. now show thatAMPK also achieves this by directly phosphorylatingULK1.In mammals, increases in the ADP:ATPand/or AMP:ATP ratios, which indicate anenergy deficit, are the signals that activateAMPK ( 8). The yeast ortholog is activated byglucose starvation, although whether changesin adenine nucleotides are the key activatingsignals ( 9) remains unclear.To identify new targets for AMPK, Eganet al. looked for the conserved sequencemotif that AMPK recognizes in its substrates(7, 10). Phosphorylation by AMPK often createsbinding sites for 14-3-3 proteins, so theauthors also searched for proteins that boundIn cells starved of nutrients, the process ofautophagy is induced to recycle surplus cellularcomponents to support cell survival.to 14-3-3 proteins when AMPK was activated.This two-part screen identified four potentialsites in the central domain of ULK1 (Ser 467 ,Ser 555 , Thr 574 , and Ser 637 ), of which two (Ser 555and Thr 574 ) are conserved in ULK2. Supportingthis connection, a screen for proteins thatinteract with ULK1 or ULK2 had identifiedsix of the seven subunit isoforms of AMPK(11), suggesting that AMPK and ULK1 orULK2 form stable complexes.Using mass spectrometry and antibodiesagainst specific phosphorylation sites,Egan et al. show that three of the ULK1 siteswere phosphorylated by AMPK in cell-freeassays, and that Ser 555 was phosphorylatedin wild-type, but not AMPK-deficient, cellstreated with an AMPK activator. To determinewhether phosphorylation at one ormore of the sites is required to induce autophagy,the authors generated a “4A” mutant inwhich all four serine and threonine residueswere changed to alanine. Wild-type ULK1or the 4A mutant was expressed in culturedhuman cells in which the amount of endogenousULK1 had been reduced by RNA interference,or in ULK1-deficient mouse embryofibroblasts. These approaches revealed thatvarious markers of autophagy were reduced41028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

PERSPECTIVESEnergy stress Low nutrients( ATP) ( Glucose)ADP/AMPAMPKLKBIULK1(ATG13, FIP200)AutophagyNutrientsCell survivalTSCI/2in cells expressing the 4A mutant, eitherunder normal culture conditions or duringnutrient starvation. Egan et al. claim that theSer 555 and Thr 574 sites are conserved in thenematode Caenorhabditis elegans orthologUnc-51, although this can be debated ( 12).They do provide evidence that the C. elegansortholog of AMPK (Aak-2) triggers autophagy,although the role of phosphorylation ofUnc-51 was not addressed.Although further work is required todelineate the roles of the individual phosphorylationsites, the findings of Egan etal. suggest that the AMPK signaling pathwayuses a “belt and braces” approach toactivate autophagy in mammals, using bothphosphorylation of ULK1 and inhibition ofXInsulinAktHigh nutrients( Glucose) ( Amino acids)Rag GTPasesRheb-GTPXTORC1 (TOR, Raptor)XCell growthPathways linking nutrients and autophagy. Under conditions of energy stress or glucosestarvation, mammalian cells activate AMPK, which then phosphorylates and activates the ULK1complex. AMPK also inactivates TORC1 by phosphorylating Raptor as well as by phosphorylatingand activating TSC2 (converting Rheb to its inactive, GDP form). Signaling pathways that areblocked by AMPK activation under these conditions are marked with a cross. Glucose availability,acting in metazoans via insulin and Akt, activates TORC1 by phosphorylation and inactivation ofTSC2. Amino acid availability activates TORC1 via the Rag GTPases. Thus, the ULK1 complex triggersautophagy during nutrient starvation and energy stress, whereas TORC1 inhibits autophagyand stimulates cell growth in the presence of nutrients and sufficient energy.TORC1 (see the figure). AMPKinhibits TORC1 by phosphorylatingits upstream regulatorTSC2 ( 6) and the TORC1subunit, Raptor ( 9). However,some of these mechanisms arenot conserved in lower eukaryotes.Thus, there is no orthologof TSC2 in S. cerevisiae,and although there is a Raptorortholog (Kog1), because of alow degree of sequence conservationit is not clear whetherthe AMPK sites are conserved. Indeed,there is little current evidence that the yeastAMPK ortholog (the SNF1 complex) actsupstream of TOR, although there is evidencethat it regulates autophagy. In snf1 mutantsthe increase in autophagy induced whencells enter stationary phase due to glucosedepletion is defective, whereas both ATG1and ATG13 suppress the phenotype of lowglycogen accumulation observed during stationaryphase ( 13). Thus, the “belt” of activationof ULK1 (Atg1) by AMPK (SNF1) mayhave evolved early, whereas the “braces” ofinhibition of TORC1 by AMPK may havearisen later. One can only speculate as towhy this might have happened, althoughthe role of the AMPK complex itself alsoappears to have changed. In yeast, the SNF1complex is probably mainly concerned withthe response to glucose starvation, but formammalian cells the ability of AMPK tosense energy status may have become moreimportant. Perhaps inhibition of TORC1 byAMPK evolved as a mechanism to ensurethat autophagy remains active and growthrepressed during energy stress, even thoughamino acids are available.References and Notes1. S. L. Clark Jr., J. Biophys. Biochem. Cytol. 3, 349 (1957).2. C. de Duve, R. Wattiaux, Annu. Rev. Physiol. 28, 435(1966).3. D. F. Egan et al., Science 331, 456 (2011); 10.1126/science.1196371.4. M. Tsukada, Y. Ohsumi, FEBS Lett. 333, 169 (1993).5. Y. Sancak et al., Science 320, 1496 (2008).6. K. Inoki, T. Zhu, K. L. Guan, Cell 115, 577 (2003).7. D. M. Gwinn et al., Mol. Cell 30, 214 (2008).8. B. Xiao et al., Nature 449, 496 (2007).9. W. A. Wilson, S. A. Hawley, D. G. Hardie, Curr. Biol. 6,1426 (1996).10. J. W. Scott, D. G. Norman, S. A. Hawley, L. Kontogiannis,D. G. Hardie, J. Mol. Biol. 317, 309 (2002).11. C. Behrends, M. E. Sowa, S. P. Gygi, J. W. Harper, Nature466, 68 (2010).12. In an alignment by this author, correspondence of Ser 555with Thr 484 was not convincing, while Thr 532 in C. elegansaligned with Ser 637 rather than Thr 574 .13. Z. Wang, W. A. Wilson, M. A. Fujino, P. J. Roach, Mol. Cell.Biol. 21, 5742 (2001).10.1126/science.1201691CHEMISTRYChemical Kinetics Under TestMillard H. AlexanderThe dependence of reaction rates onthe isotopic identity of the reactantsand products, called the “kinetic isotopeeffect” ( 1–3), is a manifestation of therole quantum zero-point energy plays inchemical kinetics and is a consequence ofthe Born-Oppenheimer (BO) separation ofelectronic and nuclear motion in molecules(4, 5). On page 448 of this issue, FlemingDepartment of Chemistry and Biochemistry and Institute forPhysical Science and Technology, University of Maryland, CollegePark, MD 20742–2021, USA. E-mail: mha@umd.eduet al. ( 6) use muon chemistry to probe therange of nuclear masses over which thisapproximation is valid for the chemicalreaction H + H 2→ H 2+ H ( 7).The BO approximation makes possiblethe practical application of quantummechanics to all of molecular science. As thearrangement of the nuclei changes, the BOapproximation postulates that the electronswill remain in a particular quantum state. Thepractical implication is that any molecule,consisting of a number of nuclei and, usually,a larger number of electrons, can be treatedA wide range of hydrogen and helium isotopesserve as a fundamental test of isotope effectsand tunneling in chemical reactions.quantum mechanically in two stages. First,for a given nuclear arrangement, one determinesthe energy of the electrons, usually intheir ground electronic state. The gradient ofthis electronic energy (along with the electrostaticrepulsion between the nuclei) thendetermines the forces on the nuclei. This twostageseparation enables the understanding ofmolecular spectroscopy as well as the simulationof the molecular dynamics in systemsranging from the simple to the complex.One direct consequence of the BOapproximation is the understanding of anwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 411

PERSPECTIVESEnergy (cm –1 )50004000300020001000011.5important quantum effect in chemical kinetics:the kinetic isotope effect ( 1–3). Considerthe simplest diatomic molecule H 2and itsisotop omer HD (hydrogen deuteride). Withinthe BO approximation, both H 2and HD havethe same electronic Hamiltonian and, thus,the same potential energy curves. However,because their nuclear reduced masses are sodifferent [0.50 atomic mass units (amu) forH 2and 0.66 amu for HD], the vibrationalspacing differs dramatically. In particular,the quantum zero-point energy of H 2is 33%larger than that of HD.Especially in the case of H transfer andexchange, the dependence of zero-pointenergy on the isotopic identity of the reactantsand/or products in a chemical reaction canlead to large differences in equilibrium constantsand chemical reaction rate constants(1–3). The isotopic variation of the zero-pointenergy will affect not only the reaction exothermicitybut also the height of the transitionstate—the activation energy. The magnitudeof the rate constant is proportional to the densityof vibration-rotation states of the reactioncomplex at the transition state, which alsodepends on the isotopic masses ( 1).A subtle consequence of differing isotopicmasses on chemical reaction rates can be2R(H 2H 3) (Å)Mapping out a reaction. Surface mesh plot of the potential energy surface for the collinear H 1+ H 2H 3→H 1H 2+ H 3reaction as a function of the two bond distances, R, superimposed on a conventional contour plotof the potential energy surface. The minimum-energy reaction path is indicated by the solid orange line. Thezero-point energy in the reactant and product arrangements, and at the barrier, is indicated by straight blacklines. The three corresponding green slices indicate the range of coordinate space made classically accessibleby zero-point energy.2.52.5seen in terms of the potential energy surfacefor the H + H 2reaction (see the figure). Theappreciable zero-point motion of the light Hatom implies that the reaction occurs not justalong the reaction path, which is the minimum-energypathway depicted in the figure,but also over a sizable region of configurationspace around this minimum path, asindicated by the green slices. Isotopic substitutionwill control the width of the slice ofthe potential energy surfaces sampled duringthe reaction.Fleming et al. report the study of H + H 2reactivity over the widest possible range ofisotopic masses. First, they have succeededin replacing the H atom with muonium (anelectron orbiting a positive muon, µ + ). Withinthe BO approximation, this behaves as avery light H atom with a mass of only 0.11amu. On the other end of the scale, they havereplaced one of the electrons in the 4 He atomwith a negative muon, µ – . Because µ – is 207times heavier than the electron, it orbits theHe nucleus at a very close range, effectivelyscreening one of the protons. Thus, the 4 Heµsystem behaves chemically as a very heavy Hatom with mass of 4.1 amu.The rate constants from these uniqueexperiments, combining accelerator physics33211.5R(H 1H 2) (Å)and chemical kinetics, are compared with fullquantum-mechanical reactive-scattering calculations( 8), based on a very sophisticatedBO H 3potential energy surface ( 9), and withthe result of a variational extension of transitionstate theory ( 10), in its latest form ( 11).Excellent agreement is found with experimentover the entire range, in which themass of the H atom is varied from 0.11 amu,through 1 amu (the mass of the normal Hatom), and then to 4.1 amu. The tests reportedby Fleming et al. over a large range of isotopicmasses thus verifies the BO separationof electronic and nuclear motion in the paradigmH + H 2reaction. Further, the comparisonbetween the exact quantum treatment ofthe nuclear dynamics and (modified) transitionstate theory has shown that the latteraccurately describes the subtle variation ofchemical reactivity with the isotopic mass ofthe reactant, at least in the H + H 2reaction.The quantitative agreement betweentheory and experiment seen by Fleming etal. confirms our ability to predict with highaccuracy the fine details of potential energysurfaces for small triatomic reactions and tomodel the details of the reaction dynamics.However, a breakdown in any approximationis always possible. Recent work ( 12, 13)has delimited the small, but measurable, BObreakdown in the simplest chemical reactionbeyond H + H 2, namely F + H 2→ FH + Hand its analog Cl + H 2→ ClH + H. One wonderswhether it would be possible to find a triatomicreaction with a barrier in which thedependence of the rate constants (or, in moredetail, the cross sections) on the isotopic massof the reactants deviates from the predictionsof transition state theory.References and Notes1. I. W. M. Smith, Kinetics and Dynamics of Elementary GasReactions (Butterworths, London, 1980).2. R. E. Weston Jr., Science 158, 332 (1967).3. J. Bigeleisen, M. G. Mayer, J. Chem. Phys. 15, 261(1947).4. M. Born, R. Oppenheimer, Ann. Phys. 389, 457 (1927).5. http://en.wikipedia.org/wiki/Born–Oppenheimer_approximation.6. D. G. Fleming et al., Science 331, 448 (2011).7. J. V. Michael, J. R. Fisher, J. M. Bowman, Q. Sun, Science249, 269 (1990).8. Y. Sun, D. J. Kouri, D. G. Truhlar, D. W. Schwenke, Phys.Rev. A 41, 4857 (1990).9. S. L. Mielke, B. C. Garrett, K. A. Peterson, J. Chem. Phys.116, 4142 (2002).10. D. G. Truhlar, B. C. Garrett, Acc. Chem. Res. 13, 440(1980).11. B. C. Garrett, D. G. Truhlar, R. S. Grev, A. W. Magnuson,J. Phys. Chem. 84, 1730 (1980).12. L. Che et al., Science 317, 1061 (2007).13. X. Wang et al., Science 322, 573 (2008).14. Supported by the U.S. National Science Foundation(grant CHE-0848110) and by the U.S. Department ofEnergy (grant DES0002323).10.1126/science.120150941228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

ESSAYSPORE* SERIES WINNERPenguins and Polar BearsIntegrates Science and LiteracyElementary teachers find resources about theArctic and Antarctica along with practical ideasfor enhancing children’s literacy skills in thisonline project.CREDIT: JOSH LANDIS/NATIONAL SCIENCE FOUNDATIONJessica Fries-Gaither and Kimberly LightleScience is often not included in elementaryschool curriculum despite the recognizedimportance of early developmentof science concepts and skills (1–3).In addition, there are few science resourcesavailable online that are focused on elementaryschool students. This combined state ofaffairs inspired the creation of an online magazinefor elementary school teachers and theirstudents: Beyond Penguins and Polar Bears.Launched in March of 2008, Beyond Penguinsand Polar Bears (http://beyondpenguins.nsdl.org)consists of 20 thematic issuesrelating elementary science concepts to thereal-world context of the polar regions (see thefirst figure). Beyond Penguins and Polar Bearspresents science content for topics such asrocks and minerals, the water cycle, seasons,states and changes of matter, plants, the indigenouspeoples of the Arctic, polar researchand explorers, and climate change in the contextof life in the Arctic and Antarctica.The online magazine format provides professionaldevelopment content and instructionalresources with a focus on integratingscience and literacy. The strategy was initiallyconceived as a “Trojan horse” approach,using the reading and writing that elementaryteachers were comfortable with as a vehiclefor increased science instruction.As the project progressed, it becameapparent that the overlap between scienceand literacy was much richer than originallyanticipated. Inquiry-based instructionand hands-on experiences help studentsdevelop the background knowledge neededto comprehend science texts. These experiencesalso build vocabulary and provide anauthentic purpose for reading and writing(4). Reading about science concepts complementsand extends the knowledge that studentsgain from inquiry and hands-on experiencesand can substitute for direct experiencewhen that is not possible (5). Writingand discussion around scientific conceptsCollege of Education and Human Ecology, School of Teachingand Learning, The Ohio State University, 1929 KennyRoad, Suite 400, Columbus, OH 43210, USA.*SPORE, Science Prize for Online Resources in Education;www.sciencemag.org/special/spore/.Author for correspondence. E-mail: fries-gaither.1@osu.eduBeacon Valley field camp, Antarctica.allow students to share their knowledge asmembers of a scientific community in thesame way practicing scientists do (6). Evidencesuggests that such a combined curricularapproach can benefit the developmentof students’ reading engagement and comprehension,academic language, and writtenand oral discourse abilities (7–9). There isalso growing recognition that students’ abilityto engage in the authentic practice andprocess of science requires the developmentand use of the structures of logical argumentationand the ability to read and write ininformational text genres (10, 11).The content of each issue of BeyondPenguins and Polar Bears is organized intofive departments:• Professional Learning contains resourcesfor teachers, including science and literacycontent knowledge, student misconceptionsin science, teaching and assessmentstrategies, and information about creatingan equitable classroom environment.• Science and Literacy includes lessonplans and a virtual bookshelf of children’sliterature.• Across the Curriculum includes ideas forintegrating other disciplines—such asmath, social studies, and art—into a sciencelesson or unit.• In the Field: Scientists at Work profilespolar researchers.• Polar News and Notes provides updateson news, research, and opportunities forteachers and students.Beyond Penguins and Polar Bears alsoincludes original content. A monthly nonfictionarticle, called a feature story, is written forstudents and is available in three grade bands,kindergarten to grade 1 (K to 1) and grades 2to 3 and 4 to 5. Teachers can print full-colorbooks with illustrations, access electronicversions with recorded narration, or printa text-only version of each story. Unit planshelp teachers assemble the resources availablein a given issue into a cohesive instructionalframework based on the 5E learning cycle(engage, explore, explain, elaborate, and evaluate),a model used to plan inquiry lessons. Allresources are correlated to the National ScienceEducation Standards and/or the NationalCouncil of Teachers of English Language Artsand International Reading Association’s Standardsfor the English Language Arts.Pilot testing of project materials with19 teachers and 173 K to 5 students inwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 413

ESSAYColumbus, Ohio, andCharlotte, North Carolina,was conducted overa 2-year period. Teachersand students were differentin every year of testing.Teachers were askedto include resources fromBeyond Penguins andPolar Bears as either a supplementto or a replacement for the existingcurriculum, to keep a journal of the Website and resource usage, and to participatein classroom observations, interviews, andpre- and post-teaching questionnaires.Data that were collected assessed teacherusage of reform-based teaching practices,including the integration of science andliteracy instruction, utilization of BeyondPenguins and Polar Bears, and changes inclassroom practices following integrationof the resources therein. Teachers reportedchanges, such as an increase in their likelihoodto provide opportunities to read aboutscience and to have their students writeto communicate scientific results. Teachersalso reported that an increased contentknowledge about the polar regions and sciencein general also increased their confidencein teaching science (12).Student questionnaires assessed attitudesregarding science. Statistically significantchanges were seen in third grade studentswho were less likely to agree with the statementsthat “science was mostly memorizingfacts” and that “science is more for boys thangirls.” They were also more likely to agreethat writing is important in science (13).In subsequent pilot testing, another groupAbout the authorsNorthern lights, Greenland.of third grade students was significantlymore likely to agree with the statements “Ilike science,” “I am good at science,” and “Iunderstand more of what goes on in science”(12). Over the entire pilot testing period,students were more likely to agree withthe statements “I am good at science” and“writing is important in science.” Studentsin fourth and fifth grade did not show significantchanges in their agreement with thesestatements after exposure to project materials.Continued evaluation efforts include acloser look at why third grade might be acritical point in developing students’ understandingand attitudes about science.Beyond Penguins and Polar Bears capitalizeson social media, such as Facebook(Beyond Penguins) and Twitter (@beyondpenguins),to connect teachers to project content.Monthly Web seminars provide opportunitiesfor virtual professional developmentin science and literacy instruction andexposure to project resources. Archives anddescriptions of upcoming seminars are availableon the home page and at http://bit.ly/BPPBseminars. Such seminars have receivedpositive evaluations from participants, whohave included classroom teachers, curriculumspecialists, and other education professionals.Jessica Fries-Gaither is an education resource specialist at TheOhio State University and project director of Beyond Penguins andPolar Bears.Kimberly Lightle, Ph.D., is a science educatorat The Ohio State University and principalinvestigator of Beyond Penguins andPolar Bears. Beyond Penguins and PolarBears was developed by an interdisciplinaryteam from The Ohio State University College of Education andHuman Ecology; the Ohio Resource Center for Mathematics, Science,and Reading; the Byrd Polar Research Center; the ColumbusCenter for Science and Industry (COSI); the Upper Arlington PublicLibrary; and the National Science Digital Library (NSDL). Expertsin science and education fi elds, including climatologist Andy Monaghan of the NationalCenter for Atmospheric Research (NCAR), paleomammologist Ross MacPhee of the AmericanMuseum of Natural History (AMNH), and Mark McCaffrey of the Cooperative Institute forResearch in Environmental Sciences (CIRES), served as guest columnists.Although our goal was to provide anonline resource that could be easily accessedacross the country and around the world, wealso realized that many teachers still preferprinted versions of resources. We thereforeprovide a “print-on-demand” option (seehttp://bit.ly/printondemand).The Arctic and Antarctica are remote,beautiful, unique, and fragile (see the secondfigure). Through Beyond Penguins andPolar Bears, we hope that we have broughtthese far-off places a little closer to elementaryteachers and students and have inspiredexcellent science instruction, too.References and Notes1. J. McMurrer, Instructional Time in Elementary Schools: ACloser Look at Changes for Specifi c Subjects (Center onEducation Policy, Washington, DC, 2008); www.cep-dc.org/displayDocument.cfm?DocumentID=309.2. R. A. Duschl, H. A. Schweingruber, A. W. Shouse, Eds.,Taking Science to School: Learning and Teaching Sciencein Grades K–8 (National Academies Press, Washington,DC, 2007); www.nap.edu/catalog.php?record_id=11625.3. R. Dorph et al., The Status of Science Education in BayArea Elementary Schools (Lawrence Hall of Science, Univ.of California, Berkeley, Berkeley, CA, 2007);www.lawrencehallofscience.org/rea/bayareastudy/.4. G. N. Cervetti, C. A. Jaynes, E. H. Hiebert, in ReadingMore, Reading Better, E. H. Hiebert, Ed. (Guilford, NewYork, 2009), chap. 4; http://scienceandliteracy.org/sites/scienceandliteracy.org/files/biblio/cervettihiebertjaynes_2009_knowledgethrureading_pd_93843.pdf.5. A. S. Palincsar, S. J. Magnusson, in Cognition and Instruction:Twenty-Five Years of Progress, S. Carver and D.Klahr, Eds. (Erlbaum, Mahwah, NJ, 2001), pp. 151–194.6. J. L. Tilson, Connect 21, 5 (2007); http://scienceandliteracy.org/sites/scienceandliteracy.org/files/biblio/tilson_discoursecirclesconnect_pdf_47687.pdf.7. J. Barber, K. N. Catz, paper presented at the 87th AnnualMeeting of the American Educational Research Association,San Francisco, CA, 6 April 2006.8. G. Cervetti, P. D. Pearson, M. A. Bravo, J. Barber, Readingand Writing in the Service of Inquiry-Based Science(Regents of the University of California, 2005);http://scienceandliteracy.org/sites/scienceandliteracy.org/files/biblio/cervetti_pearson_bravo_barber_2005_pdf_12873.pdf.9. H. Patrick, P. Mantzicopoulos, A. Samarapungavan, YoungChild. 64, 32 (2009).10. N. K. Duke, V. S. Bennett-Armistead, Reading andWriting Informational Text in the Primary Grades:Research-Based Practices (Scholastic, New York, 2003).11. J. F. Osborne, Camb. J. Educ. 32, 203 (2002).12. S. B. Woodruff, K. L. Morio, Y. Li, Evaluation of BeyondPenguins and Polar Bears: Integrating Literacy and IPYin the K–5 Classroom; Year 2 Report 2008–2009 (Ohio’sEvaluation & Assessment Center for Mathematics andScience Education, Oxford, OH, 2009); http://issuu.com/dlatosu/docs/beyondpenguins_evaluationreport_year2.13. S. B. Woodruff, K. L. Morio, Y. Li, Evaluation of BeyondPenguins and Polar Bears: Integrating Literacy and IPYin the K-5 Classroom; Annual Report 2009–2010 (Ohio’sEvaluation & Assessment Center for Mathematics andScience Education, Oxford, OH, 2010); http://issuu.com/dlatosu/docs/beyondpenguins_evaluationreport_year3.14. Beyond Penguins contributes to the National ScienceDigital Library and is supported by the NSF’s Division ofResearch on Learning in Formal and Informal Settingsunder grant 0733024. Any opinions, findings, conclusions,or recommendations expressed on this Web site are those ofthe authors and do not necessarily reflect the views of NSF.10.1126/science.1196976CREDITS: (TOP) NICK RUSSILL/FLICKR; (BOTTOM) MARGAUZ BALDRIDGE/OHIO STATE UNIVERSITY41428 JANUARY 2011 VOL331 SCIENCE www.sciencemag.org

INTRODUCINGAAAS MemberCentralThe exclusive new website for the AAAS member community.AAAS MemberCentral is a new website focused on helping you — the scientists, engineers,educators, students, policymakers, and concerned citizens who make up the AAAScommunity — connect like never before.On MemberCentral you can contribute to discussion groups or blogs, participate in awebinar, or share photos of your field research. You can exchange ideas, learn aboutyour fellow members, and gain fresh insights into issues that matter to you the most.MemberCentral is also an easy access point for a wide variety of other AAAS membershipbenefits, like discounts on cars and books, travel opportunities, and more.Experience MemberCentral for yourself. Visit MemberCentral.aaas.org today and log inusing your Science online username and password.MemberCentral.aaas.org

ASSOCIATION AFFAIRSPRESIDENTIAL ADDRESSLife on the River of SciencePeter AgreTHE YEAR 2010 MARKED THE CENTENNIALof Mark Twain’s death. More than any otherAmerican author, Twain exemplified the useof personal anecdotes to illustrate events inour nation’s history. With this in mind, I willattempt to share my experiences in science,beginning one half-century ago, with a viewfor how we as individuals are part of the greatriver that science has become.It is my hope to stimulate young scientistsand inform the nonscientific public thatachieving success in science involves severalfeatures. But if my experience is representative,the most important features includebasic curiosity, the will to take chances,and the generous attention of family membersand teachers. Probably for most individualswho became scientists during thistime, guarantees of financial success werenot considered, and with no large fortuneto lose, it is no surprise that many scientistscame from America’s large middle class.Although it occurred more than 50 yearsago, I distinctly recall Wernher von Braundiscussing rocketry and space travel. In anunforgettable Disney program, Universityof California Berkeley Nobel laureate GlennSeaborg provided a remarkable demonstrationof the chemical chain reaction. Holdinga spring-loaded mousetrap with a ping-pongball on it, he sprung the trap and the ballflew. The camera then panned a room withthe floor covered with activated mousetraps,each with a ping-pong ball on it. Seaborgtossed in one ping-pong ball, and within secondsthe room was a cloud of flying mousetrapsand ping-pong balls.Science and Growing Up in the 1950sWhile I was a child in the 1950s, “science”was a word familiar to American schoolchildren.Just a decade earlier, the end of WorldWar II, brought about by dropping atomicbombs on Hiroshima and Nagasaki, Japan,made everyone aware that we had enteredthe nuclear age. Because the polio epidemichad touched tens of thousands of Americanfamilies, the name of vaccine pioneer JonasSalk was universally recognized.The widespread introduction of televisioninto American homes in the 1950sallowed American children to see the specialmagic of science by watching Don Herbertas “Mr. Wizard” on Saturday mornings. OnSunday evenings, children viewed the Disneyshow, whose format reflected the fourkingdoms of the Disneyland amusementpark, one of which, Tomorrowland, focusedon science.Peter Agre is university professor and director of the JohnsHopkins Malaria Research Institute at the Johns HopkinsBloomberg School of Public Health. He served as the presidentof the American Association for the Advancement ofScience (AAAS) from February 2009 to February 2010.This article is adapted from the Presidential Address hedelivered at the AAAS Annual Meeting in San Diego on 18February 2010.Seal River enters Hudson Bay, 2004Of course, actual hands-on sciencetrumped even television, and having a fatherwho was head of the chemistry departmentat St. Olaf College, a small liberal arts schoolin the rolling farmlands of southern Minnesota,gave me opportunities available to fewchildren. My brothers and I marveled as weadded a drop of a colorless solution to a beakercontaining another colorless solutionthat instantly turned brilliant pink. Additionof a drop from a third solution causedthe pink color to disappear again. What wefirst viewed as “magic” became understandablewhen we learned about alkali, acid, andindicator dyes. Marked by the experience,I recall our third-grade teacher asking usto draw ourselves as adults performing ourlife’s work. I proudly drew a picture of achemist holding test tubes, since I wanted tobe like my father—a chemist.On a Saturday afternoon in October 1957,Dad came across the meadow to where mybrothers and I were playing ball to bring ushome for dinner. As we walked, he spokeof the breaking news story on the radio: thelaunching of Sputnik. We scanned the skyand failed to see the satellite, but it was noless real. More than any other single event,the launching of Sputnik began a remarkablerenewal of the already strong American supportfor science. The motivation was based onthe national humiliation of being beaten intospace by our adversary, the Soviet Union, butthe outcome was very positive.The outpouring of funds for science andscience education affected us directly. Dadwrote a National Science Foundation Fellowshipthat allowed us to move to Berkeley, California,for a sabbatical year at the Universityof California. Perhaps resembling the Norwegianequivalent of the Beverly Hillbillies,we packed our old Chevrolet station wagonfor the drive across the country. We arrivedin Berkeley, a forward-thinking, multiculturalcommunity markedly unlike our quiet farmingcommunity in southern Minnesota.A Family’s Scientific HeroHero figures are important in the developmentof a child. During our year in California, webecame familiar with Dad’s new colleagues,including a chemist from Caltech with whomDad served on the American Chemical SocietyEducation Committee.Linus Pauling had an exuberant personality,and we got to know him when he stayed atour house. Eating cornflakes at the breakfasttable with the tall, grinning Pauling, wearingCREDIT: BOB FRENCH41628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

2010CREDIT: AP PHOTOa black beret over his curly white hair, wassimply unforgettable. Dad always beamedabout Pauling, who had received the 1954Nobel Prize for Chemistry for elucidatingthe nature of the chemical bond and solvingstructures of proteins. That he discoveredhemoglobin S, the molecular cause of sicklecell anemia, was no less astonishing.In addition to his brilliant laboratoryresearch, Pauling also had a unique role as ascience activist. Well known in the 1950s forhis opposition to U.S. nuclear weapons development,Pauling’s public visibility caused theU.S. State Department to revoke his passport.Falsely accused of being a communist,Pauling frequently provoked the right wingof American politics during and after theMcCarthy era.Lecturing prodigiously around the world,Pauling often presented technical lecturesabout chemistry during the day and educationallectures to the public at night.The public lectures focused on the dangerof thermonuclear weapons testing, madeinternationally famous due to publicationof his bestseller entitled No More War (1).A remarkable raconteur, Pauling conveyednuclear testing information in terms that allcould understand, saying that the dangersthat radioactive fallout held for the health ofinnocent people had become an internationalcrisis. Pauling seemingly used every possibleopportunity to speak out.In April 1962, Pauling was one of 49Nobel laureates invited to a black tie dinnerat the White House hosted by President Kennedy—anevent Kennedy later recalled as“the most extraordinary collection of talent,of human knowledge, that has ever been gatheredtogether at the White House, with thepossible exception of when Thomas Jeffersondined alone.” Not about to miss an opportunityto publicly press for an end to nucleartesting, Pauling in shirtsleeves joined a proteston the sidewalks around the White Housebearing a sign: “Mr. Kennedy, Mr. Macmillan,WE HAVE NO RIGHT TO TEST.”We viewed this at home on the eveningnews. Apparently Kennedy was well awareof Pauling’s protest and was reported to havesmiled when meeting Pauling in the receptionline. Known for his wit, Kennedy remarkedthat he understood that Pauling had beenaround the White House already (2). Buthe recognized the significance of Pauling’sprotests, and just 6 weeks before his assassination,the president signed the Limited TestBan Treaty, preventing nuclear arms testingin the atmosphere, an event that markedlyLinus Pauling, White House, 1962reduced the tension of the nuclear arms race.As Kennedy was being buried in ArlingtonCemetery, the Pauling family was preparingfor his trip to Oslo to receive his secondNobel: the 1962 Peace Prize.Others have frequently commented onPauling’s exceptional personality. Roughly 9years ago at a reception in Stockholm, I hadthe opportunity to meet James Watson. Havingreread his classic The Double Helix (3)on a family wilderness canoe trip the summerbefore, I complimented Watson on the firstsentence—a grabber that draws the reader’sattention: “I have never seen Francis Crickin a modest mood.” Watson nodded when Iexplained that Linus Pauling had been a familyfriend, then grinned broadly and added, “Ihave also never seen Linus Pauling in a modestmood.” To which I thought to myself, “Doheroes really need to be modest?”Medicine as a Scientific CareerMy path to science came not from a pressingambition to become a prominent researcherbut from a desire to pursue a career in medicine.This was in part a cultural emphasislearned early at home. As much as our fatheremphasized math and science, our mother,a farm girl who never had theopportunity to attend college,was self-taught through a loveof reading. And she read to usevery night from the children’sBible as well as great bookslike the Laura Ingalls Wilderseries. I remember snugglingwith my siblings on the sofaduring cold Minnesota eveningsas Mother read.Although I was just a smallchild, it became obvious to methat two of my five siblingswere already manifesting lifelongdisabilities: a brotherdiagnosed with mental retardationand motor skill dysfunctionand a sister with a variantof Tourette syndrome and alack of impulse control. Whilemy parents did not dwell on itexcessively, we were neverthelessreminded that all were notequivalently blessed, and wewere always encouraged to useour talents for the well-beingof those less fortunate.Also strongly encouragedin our Minnesota NorwegianLutheran community was asense of responsibility for those in the developingworld. Medical missionaries werewidely respected throughout our community.Minnesota Governor Al Quie’s sister, a nurse,and her husband, a surgeon, spent their entirecareers attending to the health needs of ruralCameroon. Even our congressman, WalterJudd, had spent a decade as a physician servingthe poor in rural China.Having the opportunity to attend AugsburgCollege in Minneapolis, where Fathertaught after leaving St. Olaf, I majored inchemistry and received outstanding lecturesand laboratory experiences. Several ofmy Augsburg classmates were from medicalmissionary families and had grown up inrural Africa, India, and Asia. This group projectedsincere altruism without any hint ofself-promotion.Through an introduction to RichardVarco, a 1955 Lasker Award–winning pioneerin heart surgery, I received the opportunityto undertake summer research at theUniversity of Minnesota Heart Hospital.The experience was pivotal in solidifyingmy interest in medicine. All five of the summerstudents I worked with have risen toleadership positions in leading institutionswww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 417

ASSOCIATION AFFAIRSof academic medicine. Two of my youngerbrothers must have reached the same conclusion,since both have gone on to becomemedical doctors.Accepted at medical school and havingfinished my college coursework earlyin the winter of 1970, I had several monthsfor an elective experience. I chose to use thetime backpacking throughout east, southeast,and southern Asia and the Middle East.This allowed me the chance to view exoticplaces but also the opportunity to visit sitesof medical research in Thailand and Indiathat intrigued me greatly. Although “goingnative” often brings unwanted experienceswith traveler’s diarrhea and other maladies,I survived the experience excited about theprospect of global health research at JohnsHopkins University, renowned for its internationalprograms.Early Research ExperiencesProbably the most important experiencefor any young scientist is the opportunity tojoin an exciting laboratory. Although algorithmsmay exist for matching students withlabs, my experience suggested that personalcontact may be much more informative. Asa first-year medical student at Johns Hopkins,I became close personal friends withVann Bennett, a fellow medical studentwho had been a Stanford wrestler and a bigwavesurfer in Hawaii. Vann and I shared alove for vigorous physical exercise, such asbicycle camping trips throughout the Appalachians.Vann’s fascination with biochemistrywas something that I grew to admire, andI envied his special bravado—often skippingclass assignments to pursue his independentresearch in a laboratory.Intrigued by our medical school lecturesin microbiology, I became fascinated by therecent advances in understanding the molecularbasis of cholera, a horrible diarrheal diseasethat was sweeping through Asia, killingtens of thousands of infants and small children.I began working on a related project,the molecular basis of the well-known maladyE. coli traveler’s diarrhea, in the lab ofR. Bradley Sack, a young infectious diseaseresearcher at Hopkins. With Brad’s support,I made steady progress, studying the toxinby injecting crude mixtures into the lumen ofligated segments of rabbit small intestine thatthen became swollen with secreted diarrhealfluid. It soon became clear that purificationof the toxin would require more sophisticatedtechnical expertise.It was through Vann that I came to join thelab led by Pedro Cuatrecasas, a brilliant Spanish-bornphysician scientist who had trainedat the National Institutes of Health (NIH)with Christian Anfinsen, who had shared the1972 Nobel Prize in Chemistry for establishingthat the primary sequence of a protein,staphylococcal nuclease, determined thestructure and catalytic activity of the enzyme.Although only in his 30s, Pedro hadalready achieved great recognition for pioneeringthe technique of “affinity chromatography”to biological problems (4). Thisnovel concept used an insoluble matrix ofpolymeric beads to which small moleculeshad been covalently attached, and throughwhich complicated biological mixtures suchas cytoplasm or solubilized membranescould be filtered. Using affinity chromatography,success had been achieved in isolating,for the first time, receptors for hormonessuch as insulin, epidermal growthfactor, and estrogen. For this work, Pedrowas to subsequently share the 1987 WolfFoundation Prize in Medicine. Despite thefame he achieved from affinity chromatography,Pedro alwaysencouraged us to seekimportant biologicalproblems, rather thansimply seeking problemsto apply our technicalrepertoire to.After purification ofcholera toxin in 1970and recognition thatganglioside membranelipids served as receptorsites, Pedro enteredthe field to characterizethe membrane eventsof cholera toxin activationof membrane adenylate cyclase. Thisalso made the Cuatrecasas lab well suitedfor studying the elusive E. coli enterotoxin.Using ganglioside affinity gels we had prepared,we succeeded in adsorbing the enterotoxinactivity to the affinity columns and wereable to elute highly enriched subunits of theenterotoxin from the column, allowing us toraise antibodies and develop an immunoassayto diagnose E. coli enterotoxic diarrhea fromcrude stool samples (5).The prospect of using our new techniqueto screen individuals in the field in Bangladesh,where diarrheal diseases were rampant,was our long-term objective. As is not uncommon,scaling up a prototype lab test for clinicaltesting was not simple. Having committedmy entire senior year of medical school to theproject, it was still not ready for field testing,so I decided to remain an extra year as a postdoctoralfellow in Pedro’s lab. This was consideredhighly questionable by the medicalstudent advisors, as well as most of my medicalschool classmates, since graduating andbecoming resident physicians are equivalentto the Holy Grail for medical students.But for me, remaining associated with thecolorful members of Pedro’s team was muchmore exciting. The group included severalthat had bypassed lucrative careers in clinicalmedicine to follow their passions for science.Looking back years later, I think it was clearlythe people in Pedro’s lab that made scienceirresistibly appealing to me: Vann; GianfredoPuca, a debonair Neopolitan film actor whorevolutionized the understanding of femininityby purifying the estrogen receptor; IgnacioSandoval, a Spanish leftist whose loveof biology was only exceeded by his hatredof the Franco regime; Naji Sahyoun, a loyalPalestinian; and Marvin Siegel, the son of anorthodox rabbi, became lifelong best friendsafter working closely together in the lab. AndXenopus oocytes placed in hypo-osmolar solution: controloocyte (left) and oocyte expressing the 28-kD protein (right)even more importantly, I had met my futurewife, Mary, who worked in a neurovirologylaboratory at Johns Hopkins, and our livestook on a special tone resembling La Bohèmewith a science orientation.Pathway to a DiscoveryThe reason one lab makes a discovery whereothers have failed has been an issue longdebated in science. It may well be that each labis unique in its own way, but this seems overlysimplistic, and it may be that a few generalizationsare reasonable. So I will summarize,with a little editorializing, how our lab cameto discover the protein now known as aquaporin-1(AQP1), the first defined molecularwater channel—using our story to communicateseven “pearls” for younger scientists.CREDIT: ADAPTED FROM G. M. PRESTON, T. P. CARROLL, W. B. GUGGINO, P. AGRE, SCIENCE 256, 385 (1992)41828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

2010CREDIT: PETER AGRE1. Be lucky by keeping your eyes wideopen. After joining the faculty at JohnsHopkins in 1984, my small laboratorywas working on human erythrocytemembranes. We set out to isolate the core32-kD membrane protein of the Rhesusblood group antigens. In addition to the32-kD Rh, a 28-kD membrane proteincopurified. We subsequently realized thatthe smaller polypeptide was a contaminantand was unrelated to Rh.2. Consult the wisdom of your colleagues.The nature of the 28-kD protein wasintriguing. Unlike most proteins, the28-kD polypeptide stained poorly withCoomassie blue, so had not been noticedpreviously. When it was purified tohomogeneity, we realized that the 28-kDmembrane protein was extremely abundantand seemed to be a homotetramericprotein, suggesting that it might be amembrane channel. N-terminal proteinsequencing provided sufficient informationfor cDNA cloning that showed thatthe new protein was related to a seriesof functionally undefined proteins fromdiverse organisms, including plants. Ispoke with numerous colleagues aboutthe possible role for the 28-kD protein,until one had an insightful idea.3. Respect but don’t overcompartmentalizework and family life. Despite the needto share responsibilities for the raising ofchildren and contributing to domesticduties, it is important to remain flexible.The conversation that helped elucidatethe function of the new 28-kD proteinoccurred at the end of one of our familycamping trips. While returning from theEverglades, we stopped in Chapel Hill,North Carolina, so that Mary could visitfriends, the children would have a chanceto play with their former playmates, andI could drop in on some of my scientificfriends. This was to become just one ofmany examples where our family life catalyzeda scientific advance.4. Great ideas come when you least expectthem. What began as a casual visit withmy former professor John Parker, ahematologist and membrane physiologistat the University of North CarolinaSchool of Medicine, became a majorevent. I shared with Parker what I knewabout the 28-kD protein. John quicklyrecognized that the protein might besomething long sought by physiologists:the membrane channel for water.Workers in multiple labs worldwide hadRural Zambia, 2008generated indirect evidence supportingthe existence of water channels, but nonehad isolated the putative channel andestablished its molecular identity.5. Team up with those bearing neededexpertise. Parker also suggested that Icontact Bill Guggino when I returned toHopkins. Guggino, a membrane-channelphysiologist, used techniques in his labfor the expression and study of membraneion channels. This allowed ourpostdoctoral fellow Gregory Prestonto perform a simple assay demonstratingthat frog eggs expressing the newprotein became osmotically active andexploded when placed in distilled water(6). This began a long series of collaborationswith an expanding group ofaquaporin scientists around the world,including renal physiologists at NIH,structural biologists in Basel and Kyoto,microscopists in Aarhus, neuroscientistsin Oslo, eye researchers in London andTokyo, and others. Together, many scientistsprovided new insight into the fundamentalprocesses involved in renal concentration,vision, skin integrity, brainedema, thermal stress, arsenic clearance,and obesity, as well as the adaptation ofplants to drought (7).6. Don’t take yourself too seriously. In thebeginning, it was difficult not to takeevery advance in the aquaporin field personallyand view newcomers and competitorswith great suspicion. It soonbecame obvious that we were simply asmall section of a new field of science.When the Nobel call eventually camefrom the Royal Swedish Academy ofScience, it was my 78-year-old motherwho put it in proper context, “That’s verynice, but don’t let it go to his head.” Herintent was to communicate that awardsare nice, but doing something useful forothers is really important.7. Science is about new adventures. Theidea of working in global health hadbeen my original career ambition. Withthe recognition of aquaporins in theparasite causing malaria, we cautiouslyentered the field. In many ways, thismarked a new direction for our lab, butone that we were eager to follow. Andwith a new direction came a new opportunity:directorship of the Johns HopkinsMalaria Research Institute (JHMRI) atthe Bloomberg School of Public Health(8). The new role has been a challenge,but it has been a delight to be part of aprogram with as much global importanceas malaria research. The JHMRIlab programs in Baltimore and fieldworkin rural Zambia are both stimulatingand rewarding. And the ease with whichmalaria crosses borders, killing hundredsof thousands of African children,is a clear reminder of the internationalaspects of science.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 419

ASSOCIATION AFFAIRSScience as a Tool for InternationalDiplomacyScience has always been one of the most internationalof human endeavors, and this trendis certainly increasing. Every year, thousandsof young scientists come to the United Statesfrom abroad to undertake scientific educationand research. Thus, science provides a uniqueapproach to advancing good will towardAmerica in the international arena.It is no secret that the U.S. government isviewed negatively in the Muslim world, especiallyafter the military intrusions in Iraq andAfghanistan. The Zogby Poll [(9); see tablebelow] reveals a clear bimodal response.The great majority of citizens of five moderateArab nations held distinctly unfavorableviews of the United States in general. Inmarked contrast, these same individuals heldfavorable views of U.S. science and technology.This provides an opportunity to use ourbackground as nongovernment scientists toreinforce the positive: that U.S.-generatedscience and technology may improve the livesof people all around the world.U.S. in General U.S. Science and TechnologyCountry Favorable Unfavorable Favorable UnfavorableMorocco 11% 88 90 08Saudi Arabia 04 94 48 51Jordan 15 78 83 13Lebanon 20 69 52 46UAE 14 73 84 12[From Impressions of America (9)]University of Havana, 2009Having the chance as president of AAASto participate in some outstanding programs,one stood out as particularly appealing: theAAAS Center for Science Diplomacy. Thepotential to establish contact and engage withscientists in countries considered adversarialto the U.S. government is an opportunity forscience to serve as a unique bridge. Foundedin 2008, the center is directed by VaughanTurekian, an atmospheric geochemist andinternational policy expert, with special advisorNorman Neureiter, a chemist with extensivepolicy experience. I was greatly pleasedto participate as senior scientist in a series oftrips abroad (10).Recognizing that some of our visits wereto countries where there are serious intergovernmentaltensions related to a wide range ofissues such as proliferation, human rights, andeconomic openness, each visit was undertakenwith an independent nongovernmentalorganization and with private funding, in mostcases from the Richard Lounsbery Foundation.Efforts were made to inform appropriateU.S. authorities of such visits, but it was to beplain in every case that we served as representativesof the U.S. scientific and research community,not the U.S. government.Cuba. Together with members of the NewAmerica Foundation, a trip to Havana, Cuba,was made in November 2009. Our trip, thefirst AAAS visit since 1997, included a visitto the Finlay Center for Vaccines Researchand Production, wherewe observed the Cubanefforts to prevent infectiousdiseases such as type Bmeningococcal meningitis.Endemic before the Revolution,malaria has been eliminatedfrom Cuba, despite aheavy malaria presence inHaiti, just to the east. Cubanefforts to provide universalprenatal health care have succeeded in raisingthe average life span to 78+ years, equivalentto that in the United tates.The University of Havana generates alarge number of science graduates, but laboratoryopportunities are limited. Certainly theinvestment of funds in laboratories to trainyoung scientists could be mutually beneficialto Cuba and the United States. Potential scientificcollaborations could be rapidly undertakenonce the five-decade political standoffbetween our governments is resolved.Democratic People’s Republic ofKorea (DPRK; North Korea). The AAASCenter for Science Diplomacy worked formore than a year with the U.S. CivilianResearch Development Foundation to gainan invitation to visit Pyongyang in December2009. Located only ~500 miles east ofBeijing, Pyongyang exists in a world apartfrom the hustle and thickly polluted air ofthe vibrant Chinese capital. Having only amodest power grid, Pyongyang has crystalclearair and is impeccablyneat but frigid due to a lack ofinterior heating.It was obvious from ourvisit that young North Koreanshave great prowess withcomputer systems—alsomade clear from ongoingcollaboration between KimChaek University in Pyongyangand Syracuse Universityin New York. DPRK agriculturaland health sciences werefar less advanced and representareas where the UnitedStates could be helpful to the DPRK.After we had spent 1 week under thewatchful eyes and gracious hospitality ofthe State Academy of Science, our hostsexhibited a warmhearted response whenpresented with the necktie I had wornduring my Nobel lecture in Stockholm.Bridge across Zambezi, Livingstone,Zambia, 2008CREDITS: PETER AGRE42028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

2010reveal their views of Americans. After mylecture to outstanding high-school sciencestudents at the University of Delhi, I was presentedwith a painting of the beautiful multiarmedHindu goddess Saraswati, who iswidely worshiped throughout India as thebearer of enlightenment. Our internationalfriends can tell us things we in the UnitedStates need to hear.Pyongyang University of Science and Technology, 2009Final ReflectionScience is the medium of our life’s work.Whether we are frustrated or joyful, wealways know that each new day in the labbrings an opportunity to make profoundadvancements in the understanding of naturethat may improve the well-being of others.One of the talented young people fromChina in our lab always took the optimisticapproach, summarized by the Mandarin characterfor “crisis.” It is actually two characters:wei, meaning “time of danger,” and ji, meaning“time of opportunity.”For me, the crisis of living in today’sworld includes dangers such as increasingmicrobial drug resistance, the obesity epidemic,environmental damage, the needfor sustainable energy supplies, and hostilerelations between countries. It seemsclear to me that solutions to each will comethrough the great opportunities provided byscience. And with that in mind, I am optimisticabout the future.CREDIT: PETER AGREEmblazoned on the shield of the Johns HopkinsDepartment of Medicine is the wordAequanimitas. The term means “imperturbability.”The tie, bearing this word, was givenwith our wish that it be worn by the firstDPRK scientist who wins a Nobel. Althoughthe political distance between our governmentshas not decreased in the months afterour trip, it should be kept in mind that wellintentionedscientists reside in the DPRK.When the U.S. and DPRK governmentsfinally establish an accord, it is expected thatthe pursuit of peaceful areas of science maybe a bridging mechanism.Myanmar (Burma). Another country ofcurrent difficulty due to its longstanding militaryjunta is Myanmar, a nation of 59 millionpeople with rich natural resources. Togetherwith the U.S. Collection Humanitarian Corps,a visit was organized in April 2010. Of severalministries visited, the Ministry of Healthmay have been most important, because ofthe heavy infestation of malaria throughoutthe countryside that is problematic to neighboringcountries. Since our visit, nationalelections have been held, and the nationalleader and 1991 Nobel Peace Prize LaureateAung Sang Suu Kyi has been released fromconfinement in her home. Although internationallyrecognized democracy has not yetreturned to Burma, it is possible that continuedscientific dialogue may contribute to thetransparent and ongoing interactions with theinternational community that serve as a preconditionfor any sustainable and effectivesystem of governance .India. The world’s largest democracy,India and the United States have had longstandingscience engagement. It was a privilegeto attend the 10th anniversary of theIndo-U.S. Science and Technology Forum inDecember 2010. This included visits to theNational Institute of Malaria Research, otherlaboratories, and university campuses, anda public lecture on science. Despite India’smany social and economic problems, an airof optimism is readily apparent in India, andscience seems a large part of it.Given our longstanding mutual support,Indian scientists shared observations thatWei Ji – Mandarin for ”crisis“References1. L. Pauling, No More War (Dodd, Mead, New York, 1958).2. T. Hager, Force of Nature—The Life of Linus Pauling(Simon & Schuster, New York, 1995), pp. 537–538.3. J. D. Watson, The Double Helix: A Personal Account of theDiscovery of the Structure of DNA (1968; reprinted byAtheneum, Simon & Schuster, New York, 1980).4. P. Cuatrecasas, Annu. Rev. Biochem. 40, 259 (1971).5. R. B. Sack, Annu. Rev. Microbiol. 29, 333 (1975).6. G. M. Preston, T. P. Carroll, W. B. Guggino, P. Agre,Science 256, 385 (1992).7. P. Agre, Le Prix Nobel 2003 (Nobel Foundation,Stockholm, Sweden, 2004).8. Johns Hopkins Bloomberg School of Public HealthMagazine (Baltimore, MD, winter, 2011).9. J. Zogby, Impressions of America: How Arabs ViewAmerica, How Arabs Learn About America (ZogbyInternational, Washington, DC, 2004).10. P. Agre, V. Turekian, Sci. Transl. Med. 2, 1 (2010).10.1126/science.1202341www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 421

AAASNEWS&NOTESEDITED BY EDWARD W. LEMPINENCREDIT: DOUG MILLS/THE NEW YORK TIMES/REDUXASSOCIATION AFFAIRSNina Fedoroff: 21st-Century ChallengesRequire Global Focus by ScientistsNina Fedoroff is starting her new yearin Saudi Arabia, focused on some lifelongresearch and policy passions.She’s establishing a new center for desertagriculture at King Abdullah Universityof Science and Technology,where scientists from many countriesand diverse disciplines will tackle oneof the most significant challenges ofthe 21st century: how to feed a planetin the face of diminished arable landand shrinking freshwater supply.For 3 years as science adviser at theU.S. State Department and USAID,Fedoroff urged researchers to collaborateon “truly global problems thatdo not respect political boundaries orpolitical positions.” As the incomingAAAS president, she will encouragethe association to further expand itsinternational engagement and support a futurefor science that looks much like the work ofher new center.AAAS is “a wonderful interface betweenscience and society, and this is something thatis important throughout the world,” Fedoroffsaid in a recent interview. “And it’s never beenmore important for scientists to work togetheron the big issues confronting the world: food,energy, and water.”Her own experiences—from workingclosely with molecular biologists in the formerSoviet Union to mentoring a Brazilianpostdoctoral fellow—have convinced herthat science is well suited to bringing nationstogether in this enterprise.Science is an evidence-based explorationof nature, where “it doesn’t matter whatlanguage the science is done in—the criteriaof excellence are really the same,” she said.“That makes it possible for scientists of manypolitical persuasions and many religiousbeliefs to talk to each other.”AAAS is already a leader in supportingscience diplomacy and brokering internationalcollaborations, Fedoroff noted, citingprograms supported by its Center forScience Diplomacy and Center for Science,Technology, and Sustainable Development,among others. And she is happy that AAASIncoming president. Nina V. Fedoroff will succeed Alice S. Huangas AAAS president when the association’s Annual Meeting closes on21 February. Huang will begin a 1-year term as AAAS Board Chair.is a partner of the Global Knowledge Initiative,a nongovernmental organization foundedby Fedoroff and others that facilitates internationalscientific partnerships.But AAAS can do even more to help“jump the gap between countries that are welladvanced in science and those that are not,”she said. “Every country seeks to be a knowledgesociety today, and I think one of themost important things that scientists can dois to participate in making connections withscientists in less-developed countries to provideknow-how, collaboration, and sometimeseven materials.”Fedoroff received the 2006 NationalMedal of Science from U.S. President GeorgeW. Bush for her pioneering research in thefields of plant genetics, plant responses toenvironmental stress, and genetically modifiedcrops. She received her Ph.D. in molecularbiology from Rockefeller University in1972. In addition to her role as visiting professorat King Abdullah University, she is anEvan Pugh Professor of Pennsylvania StateUniversity and a member of the external facultyof the Santa Fe Institute.She served on the AAAS Board of Directorsfrom 2000 to 2004 and was elected aAAAS Fellow in 2010.She will succeed Alice S. Huang as presidentwhen the AAAS Annual Meetingcloses on 21 February; Huang willbegin a 1-year term as chairperson ofthe AAAS Board.As science adviser to the StateDepartment, Fedoroff was instrumentalin bringing more scientists to workin U.S. embassies, overseeing thedeployment of the first U.S. scienceenvoys, and developing and maintainingscience and technology agreementswith many countries.With the new U.S. Congress inplace this month, Fedoroff said it is difficultto predict the fortunes of domesticand international science programsin the near future. Targeted programsin energy and climate change researchmay not disappear, she suggested, “butI think that the budget for basic scienceprobably won’t expand a whole lot.”“Some of our champions of science areretiring from Congress, and I think it’s goingto be...even more important for scientists tocome to Washington and make themselvesheard,” particularly on the scientific consensusregarding global climate change, Fedoroffsaid. “I think they have to come, they have totestify, they have to write, they have to lecture—anythingthey can do to get the messageacross to the public.”Fedoroff thinks that researchers especiallyneed to sharpen one message: How scientistsdo their jobs. They work “by carefully constructingevidence and testing hypotheses,”she said, drawing a conclusion only when theweight of the evidence supports that conclusion.People in government or business, shesaid, “often don’t understand that science isn’tjust another point of view or opinion.”Fedoroff sees the AAAS presidency asanother chance in her wide-ranging career toshare science’s potential. “I remember sittingin my study 20, 25 years ago and wishing thatI could have a larger voice,” she said. “I’mdelighted to have that larger voice, becauseI think that one’s accomplishments, one’s 15minutes of fame, need to be put to good use,and I can’t think of a better use than communicatingabout science.” –Becky Ham42228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

ASSOCIATION AFFAIRSAAAS Members Elected as FellowsIn December 2010, the AAAS Council elected 503 members as Fellows of AAAS. Theseindividuals will be recognized for their contributions to science and technology at the FellowsForum to be held on 19 February 2011 during the AAAS Annual Meeting in Washington, D.C.Presented by section affiliation, they are:Section on Agriculture, Food, and RenewableResourcesMarcus Alley, Virginia Tech • Clifton A. Baile,Univ. of Georgia • Jerome F. Baker, Sigma Xi •Warwick M. Bayly, Washington State Univ. • PaulM. Bertsch, Univ. of Kentucky • Bryony C. Bonning,Iowa State Univ. • Kenneth J. Boote, Univ. of Florida• Claude E. Boyd, Auburn Univ. • Ingrid C. Burke,Univ. of Wyoming • Arun K. Chatterjee, Univ. ofMissouri-Columbia • Martin B. Dickman,Texas A&M Univ. • Steven Fales, Iowa StateUniv. • Bradley W. Fenwick, Univ. of Tennessee,Knoxville • Stanton Gelvin, Purdue Univ. •Scott A. Jackson, Purdue Univ. • Jiming Jiang, Univ.of Wisconsin-Madison • Phyllis E. Johnson, Univ. ofNorth Dakota • William C. Koskinen, USDA-ARS• Robert I. Krieger, Univ. of California, Riverside• Thomas A. Miller, Univ. of California, Riverside• David Neale, Univ. of California, Davis • YakovPachepsky, USDA-ARS • John Ryan, InternationalCenter for Agricultural Research in the Dry Areas,Syria • Eugene Sander, Univ. of Arizona • Johan Six,Univ. of California, Davis • Laosheng Wu, Univ. ofCalifornia, Riverside • Scott R. Yates, USDA-ARS• Frank G. Zalom, Univ. of California, Davis •Yong-Guan Zhu, Chinese Academy of SciencesSection on AnthropologyRichard A. Diehl, Univ. of Alabama • AgustínFuentes, Univ. of Notre Dame • Richard L. Jantz,Univ. of Tennessee, Knoxville • Michelle Lampl,Emory Univ. • Paul W. Leslie, Univ. of North Carolina,Chapel Hill • Fiona B. Marshall, WashingtonUniv., St. Louis • Anne C. Stone, Arizona StateUniv. • Samuel D. Stout, Ohio State Univ.Section on AstronomyDale P. Cruikshank, NASA Ames ResearchCenter • Wendy Freedman, The Observatoriesof the Carnegie Institution for Science • Lee W.Hartmann, Univ. of Michigan • Bradley M.Peterson, Ohio State Univ. • Marc HowardPinsonneault, Ohio State Univ. • Michael Werner,Jet Propulsion Laboratory • Aleksander Wolszczan,Pennsylvania State Univ.Section on Atmospheric and HydrosphericSciencesSusan K. Avery, Woods Hole Oceanographic Institution• Wallace S. Broecker, Columbia Univ. •Scott C. Doney, Woods Hole Oceanographic Institution• Catherine Gautier, Univ. of California,Santa Barbara • Roberto César Izaurralde, PacificNorthwest National Laboratory • Sydney Levitus,National Oceanic and Atmospheric Administration• Michael Oppenheimer, Princeton Univ. • ClaireL. Parkinson, NASA Goddard Space Flight Center• Peter Schlosser, Columbia Univ.Section on Biological SciencesFrederick W. Alt, Children’s Hospital Boston/Immune Disease Institute • James C. Alwine,Univ. of Pennsylvania • Brenda Andrews, Univ.of Toronto, Canada • Charles K. Barlowe, DartmouthMedical School • Franklin G. Berger, Univ.of South Carolina • Kerry S. Bloom, Univ. of NorthCarolina, Chapel Hill • Eduardo Blumwald, Univ.of California, Davis • Carolyn Hovde Bohach,Univ. of Idaho • Joerg Bohlmann, Univ. of BritishColumbia, Canada • Charles M. Boone, Univ.of Toronto, Canada • Michael Breitenbach, Univ. ofSalzburg, Austria • Roger Brent, Fred HutchinsonCancer Research Center • Donald A. Bryant, PennsylvaniaState Univ. • Peter M. J. Burgers, WashingtonUniv. School of Medicine, St. Louis • Thomas P.Burris, The Scripps Research Institute • Ragan M.Callaway, Univ. of Montana • Diane R. Campbell,Univ. of California, Irvine • Blanche Capel, DukeUniv. School of Medicine • Maria E. Cardenas, DukeUniv. Medical Center • Kevin R. Carman, LouisianaState Univ. • David A. Caron, Univ. of SouthernCalifornia • Ta-Yuan Chang, Dartmouth MedicalSchool • Daniel G. Colley, Univ. of Georgia • SusanG. Conard, George Mason Univ. • Michael M. Cox,Univ. of Wisconsin-Madison • Valeria Culotta,Johns Hopkins Univ. • David C. Culver, AmericanUniv. • Alan G. Darvill, Univ. of Georgia • Eric A.Davidson, Woods Hole Research Center • VictorL. Davidson, Univ. of Mississippi Medical Center• Edward F. DeLong, Massachusetts Institute ofTechnology • Lynda Delph, Indiana Univ. • BenitoO. de Lumen, Univ. of California, Berkeley • RikDerynck, Univ. of California, San Francisco • MarkW. Dewhirst, Duke Univ. Medical Center • RobertoDocampo, Univ. of Georgia • Carl J. Douglas, Univ.of British Columbia, Canada • Michael Doyle,Univ. of Georgia • Gary Norman Drews, Univ. ofUtah • Katrina Jane Edwards, Univ. of SouthernCalifornia • Robert N. Eisenman, Fred HutchinsonCancer Research Center • Jorge C. Escalante-Semerena, Univ. of Wisconsin-Madison • MaryCynthia Farach-Carson, Rice Univ. • AsgerallyFazleabas, Michigan State Univ. • Nina V. Fedoroff,King Abdullah Univ. of Science and Technology,Saudi Arabia • Susan J. Fisher, Univ. of California,San Francisco • Thomas D. Fox, Cornell Univ. •William (Ned) Friedman, Univ. of Colorado, Boulder• Jiří Friml, Ghent Univ., Belgium • Xiang-Dong Fu, Univ. of California, San Diego • VadimGladyshev, Brigham and Women’s Hospital/Harvard Medical School • Jessica Gurevitch, StateUniv. of New York at Stony Brook • Kasturi Haldar,Univ. of Notre Dame • Roger Hangarter, IndianaUniv. • Paul M. Hasegawa, Purdue Univ. • EdwardHawrot, Brown Univ. • Mien-Chie Hung, M. D.Anderson Cancer Center • Roger William Innes,Indiana Univ. • S. Michal Jazwinski, Tulane Univ.Health Sciences Center • Chris A. Kaiser, MassachusettsInstitute of Technology • Jack D. Keene,Duke Univ. Medical Center • Roberto Kolter, HarvardMedical School • Allan E. Konopka, PacificNorthwest National Laboratory • Sally Kornbluth,Duke Univ. School of Medicine • Don R. Levitan,Florida State Univ. • Charles Liarakos, NationalScience Foundation • Haifan Lin, Yale Univ. Schoolof Medicine • Jerry B. Lingrel, Univ. of Cincinnati• Jianguo Liu, Michigan State Univ. • David M.Lodge, Univ. of Notre Dame • Guillermina Lozano,M.D. Anderson Cancer Center • Hong Ma, PennsylvaniaState Univ. • Yasunori Machida, NagoyaUniv., Japan • Manuela Martins-Green, Univ. ofCalifornia, Riverside • James M. Mason, NationalInstitute of Environmental Health Sciences, NIH• Gerald E. McClearn, Pennsylvania State Univ. •C. Robertson McClung, Dartmouth College • G.Stanley McKnight, Univ. of Washington • WilliamF. Morgan, Pacific Northwest National Laboratory• Naoko K. Nishizawa, The Univ. of Tokyo,Japan • Magnus Nordborg, Gregor Mendel Institute,Austria • Peter J. Novick, Univ. of California,San Diego • Stephen Nowicki, Duke Univ. •Janet M. Oliver, Univ. of New Mexico Health SciencesCenter • Terry L. Orr-Weaver, WhiteheadInstitute for Biomedical Research • KatherineW. Osteryoung, Michigan State Univ. • BernhardPalsson, Univ. of California, San Diego • MaryC. Pearl, Garrison Institute • Stuart W. Peltz, PTCTherapeutics • Ann Marie Pendergast, Duke Univ.Medical Center • Nora I. Perrone-Bizzozero,Univ. of New Mexico School of Medicine • PeterPhilippsen, Biozentrum, Univ. of Basel, Switzerland• Andrew Pomiankowski, Univ. College London •Huntington Potter, Univ. of South Florida College ofMedicine • Rudolf A. Raff, Indiana Univ. • FrederickA. Rainey, Louisiana State Univ. • Raymond Reeves,Washington State Univ. • Susan M. Rosenberg,Baylor College of Medicine • Patrick S. Schnable,Iowa State Univ. • Saul J. Silverstein, ColumbiaUniv. • David K. Skelly, Yale Univ. • Pamela L. C.Small, Univ. of Tennessee, Knoxville • Michael J.Smerdon, Washington State Univ. • James M.Sodetz, Univ. of South Carolina • Abraham L.Sonenshein, Tufts Univ. School of Medicine •Gustavo A. Stolovitzky, IBM T.J. Watson ResearchCenter • Craig V. Sullivan, North Carolina StateUniv. • Heven Sze, Univ. of Maryland, CollegePark • Michael Thomashow, Michigan State Univ.• David W. Threadgill, North Carolina State Univ. •Barbara J. Turpin, Rutgers, The State Univ. of NewJersey • Johannes Christian Vogel, The Naturalwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 423

History Museum, London • Andreas Wagner, Univ.of Zurich, Switzerland • Detlef Weigel, Max PlanckInstitute for Developmental Biology, Germany •Jonathan Wendel, Iowa State Univ. • Susan R.Wente, Vanderbilt Univ. Medical Center • Gary M.Wessel, Brown Univ. • William T. Wickner, DartmouthMedical School • Rod A. Wing, Univ. of Arizona• John C. Wingfield, Univ. of California, Davis• Ronny Clifford Woodruff, Bowling Green StateUniv. • Hong Wu, Univ. of California, Los Angeles •John D. York, Duke Univ. Medical Center • PatriciaClaire Zambryski, Univ. of California, Berkeley •Kang Zhang, Univ. of California San DiegoSection on ChemistryNicholas L. Abbott, Univ. of Wisconsin-Madison• I. Jonathan Amster, Univ. of Georgia • Scott L.Anderson, Univ. of Utah • Brian C. Benicewicz,Univ. of South Carolina • Veronica M. Bierbaum,Univ. of Colorado, Boulder • Helen E. Blackwell,Univ. of Wisconsin-Madison • John J. Boland, TrinityCollege Dublin, Ireland • J. Martin BollingerJr., Pennsylvania State Univ. • Bruce Branchaud,Life Technologies • Charles T. Campbell, Univ.of Washington • Michael Chan, Ohio StateUniv. • Christine S. Chow, Wayne State Univ. •R. Graham Cooks, Purdue Univ. • Charles S.Craik, Univ. of California, San Francisco • Paul S.Cremer, Texas A&M Univ. • Hong jie Dai, StanfordUniv. • Liem Dang, Pacific Northwest NationalLaboratory • Sheila S. David, Univ. of California,Davis • Steven E. Ealick, Cornell Univ. • MarkEdiger, Univ. of Wisconsin-Madison • Stephen W.Fesik, Vanderbilt Univ. School of Medicine • PaulFrederick Fitzpatrick, Univ. of Texas Health ScienceCenter, San Antonio • Craig J. Forsyth, Ohio StateUniv. • Michael J. Frisch, Gaussian, Inc. • BruceGanem, Cornell Univ. • Marc M. Greenberg, JohnsHopkins Univ. • Mihal Gross, U.S. Office of NavalResearch • Lizbeth Hedstrom, Brandeis Univ. •Janice Hicks, National Science Foundation •Rustem Ismagilov, Univ. of Chicago • Jeffrey N.Johnston, Vanderbilt Univ. • Prashant Kamat,Univ. of Notre Dame • Ehud Keinan, The ScrippsResearch Institute • Julie A. Leary, Univ. of California,Davis • Nancy E. Levinger, Colorado State Univ.• Jun Li, Tsinghua Univ., Beijing • Wenbin Lin, Univ.of North Carolina, Chapel Hill • Todd L. Lowary,Univ. of Alberta, Canada • Alenka Luzar, VirginiaCommonwealth Univ. • David Gilbert Lynn, EmoryUniv. • David MacMillan, Princeton Univ. • AnnaMapp, Univ. of Michigan • Adam Matzger, Univ.of Michigan • Robin L. McCarley, Louisiana StateUniv. • Saundra Yancy McGuire, Louisiana StateUniv. • Scott A. McLuckey, Purdue Univ. • MarvinJ. Miller, Univ. of Notre Dame • John Montgomery,Univ. of Michigan • Umit S. Ozkan, Ohio StateUniv. • Hongkun Park, Harvard Univ. • Dehua Pei,Ohio State Univ. • Joseph W. Perry, Georgia Instituteof Technology • Erwin Poliakoff, LouisianaState Univ. • Coleen Pugh, Univ. of Akron • DanielL. Reger, Univ. of South Carolina • William S.Rees Jr., Los Alamos National Laboratory •Nigel Richards, Univ. of Florida • Thomas Rizzo,École Polytechnique Fédérale de Lausanne, Switzerland• Steven Rokita, Univ. of Maryland,College Park • Melanie Sanford, Univ. of Michigan• Gregory K. Schenter, Pacific Northwest NationalLaboratory • William F. Schneider, Univ. of NotreDame • Anthony S. Serianni, Univ. of Notre Dame• Paul B. Shepson, Purdue Univ. • Sheldon G.Shore, Ohio State Univ. • Bradley D. Smith, Univ.of Notre Dame • Lloyd Smith, Univ. of Wisconsin-Madison • Shannon Stahl, Univ. of Wisconsin-Madison • Ruth E. Stark, City College of NewYork • Oliver Steinbock, Florida State Univ. •Michael P. Stone, Vanderbilt Univ. • Nong jian(NJ) Tao, Arizona State Univ. • SankaranThayumanavan, Univ. of Massachusetts, Amherst• Patricia A. Thiel, Iowa State Univ. • John G.Verkade, Iowa State Univ. • Eric Wickstrom,Thomas Jefferson Univ. • Ziling Xue, Univ. ofTennessee, KnoxvilleSection on Dentistry and Oral Health SciencesYang Chai, Univ. of Southern California • BjornR. Olsen, Harvard School of Dental MedicineSection on EducationRobert J. Beichner, North Carolina State Univ. • IdaChow, Society for Developmental Biology • BruceA. Fuchs, National Institutes of Health • DavidL. Haury, Ohio State Univ. • Thomas A. Holme,Iowa State Univ. • Hedy Moscovici, CaliforniaState Univ., Dominguez Hills • Terry S. Woodin,National Science FoundationSection on EngineeringSos Agaian, Univ. of Texas at San Antonio • J.Stewart Aitchison, Univ. of Toronto, Canada •James M. Anderson, Case Western Reserve Univ. •Bahman Anvari, Univ. of California, Riverside• Nasser Ashgriz, Univ. of Toronto, Canada •Gilda A. Barabino, Georgia Institute of Technology• Rena Bizios, Univ. of Texas at San Antonio •Leonard J. Bond, Pacific Northwest National Laboratory• Joan F. Brennecke, Univ. of Notre Dame •Shih-Fu Chang, Columbia Univ. • Shaochen Chen,Univ. of California, San Diego • Robert L. Clark Jr.,Univ. of Rochester • Michael W. Deem, Rice Univ. •Stephen P. DeWeerth, Georgia Institute of Technology• Duane B. Dimos, Sandia National Laboratories• Francis J. Doyle III, Univ. of California, SantaBarbara • Vinayak Dravid, Northwestern Univ. •Elizabeth Edwards, Univ. of Toronto, Canada • JayA. Farrell, Univ. of California, Riverside • W. KentFuchs, Cornell Univ. • Don P. Giddens, GeorgiaInstitute of Technology • Ahmed Hassanein, PurdueUniv. • Arthur H. Heuer, Case Western Reserve Univ.• Kristina M. Johnson, U.S. Department of Energy• Jimmy (Chih-Ming) Kao, National Sun-Yat-SenUniv., Taiwan • Mostafa Kaveh, Univ. of Minnesota• Ashok Kumar, Univ. of South Florida • Kelvin H.Lee, Delaware Biotechnology Institute • Russell J.Lefevre, Univ. of North Dakota • Edward J. Maginn,Univ. of Notre Dame • Bhubaneswar Mishra, NewYork Univ. • Salvatore D. Morgera, Univ. of SouthFlorida • Rajakkannu Mutharasan, Drexel Univ. •Jun Nogami, Univ. of Toronto, Canada • Ramana M.Pidaparti, Virginia Commonwealth Univ. • MichaelW. Plesniak, George Washington Univ. • Charles H.Reilly, Univ. of Central Florida • J. Paul Santerre,Univ. of Toronto, Canada • Molly S. Shoichet, Univ.of Toronto, Canada • Mrityunjay Singh, Ohio AerospaceInstitute • Rajiv K. Singh, Univ. of Florida• Alok Sinha, Pennsylvania State Univ. • AndrewJ. Steckl, Univ. of Cincinnati • Kirankumar V.Topudurti, U.S. Army Corps of Engineers • RobertJ. Trew, North Carolina State Univ. • Arvind Varma,Purdue Univ. • Albert Z.H. Wang, Univ. of California,Riverside • Kon-Well Wang, Univ. of Michigan• Richard C. Willson, Univ. of Houston • EugeneWong, Univ. of California, Berkeley • ZhuominZhang, Georgia Institute of TechnologySection on General Interest in Science andEngineeringMariette DiChristina, Scientifi c American • EricJolly, Science Museum of Minnesota • CristineRussell, Harvard Univ. • Erika Shugart, MarianKoshland Science Museum of the National Academyof Sciences • Morris A. (Bud) Ward, MorrisA. Ward, Inc.Section on Geology and GeographyCraig D. Allen, U.S. Geological Survey • Joel D.Blum, Univ. of Michigan • Andrew G. Fountain,Portland State Univ. • Philip Froelich, Florida StateUniv. • Timothy D. Herbert, Brown Univ. • Susan M.Kidwell, Univ. of Chicago • Larry Donald McKay,Univ. of Tennessee, Knoxville • John F. Mustard,Brown Univ. • Karl F. Nordstrom, Rutgers, The StateUniv. of New Jersey • Ellen Thomas, Yale Univ.Section on History and Philosophy of ScienceDavis Baird, Clark Univ. • Jed Z. Buchwald, CaliforniaInstitute of Technology • John Dupré, Univ.of Exeter, United Kingdom • Thomas J. Nickles,Univ. of Nevada, RenoSection on Industrial Science and TechnologyStanley R. Bull, National Renewable EnergyLaboratorySection on Information, Computing, andCommunicationSrinivas Aluru, Iowa State Univ. • Victor Bahl,Microsoft Research • David R. Boggs, ConsultingElectrical Engineer • Geoffrey Charles Bowker,Univ. of Pittsburgh • John M. Carroll, PennsylvaniaState Univ. • J. J. Garcia-Lunes-Aceves, Univ. ofCalifornia, Santa Cruz/Palo Alto Research Center• Venu Govindaraju, Univ. at Buffalo, State Univ.of New York • Hamid Jafarkhani, Univ. of California,Irvine • Farnam Jahanian, Univ. of Michigan• Phokion G. Kolaitis, Univ. of California, Santa42428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

Cruz • C. C. Jay Kuo, Univ. of Southern California• Dinesh Manocha, Univ. of North Carolina,Chapel Hill • Hanan Samet, Univ. of Maryland,College Park • Manuela M. Veloso, CarnegieMellon Univ. • Barry Wessler, Wessler ConsultingSection on Linguistics and Language SciencesKaren Emmorey, San Diego State Univ. • DavidPesetsky, Massachusetts Institute of Technology •Eric Potsdam, Univ. of FloridaSection on MathematicsDouglas N. Arnold, Univ. of Minnesota • H.T. Banks, North Carolina State Univ. • DonaldBurkholder, Univ. of Illinois, Urbana-Champaign• James Carlson, Clay Mathematics Institute• Raúl E. Curto, Univ. of Iowa • Charles W.Groetsch, The Citadel • James (Mac) Hyman,Tulane Univ. • Philip C. Kutzko, Univ. of Iowa• Yousef Saad, Univ. of Minnesota • KennethStephenson, Univ. of Tennessee, KnoxvilleSection on Medical SciencesAbdu F. Azad, Univ. of Maryland School of Medicine• Jorge L. Benach, Stony Brook Univ. • BarryR. Bloom, Harvard School of Public Health • SusanBuchbinder, San Francisco Department of PublicHealth • Lewis C. Cantley, Beth Israel DeaconessMedical Center • R. John Collier, Harvard MedicalSchool • P. Michael Conn, Oregon Health & ScienceUniv./Oregon National Primate Research Center• Jason G. Cyster, Univ. of California, San Francisco• Tamara Doering, Washington Univ. Schoolof Medicine, St. Louis • Gideon Dreyfuss, Univ. ofPennsylvania School of Medicine • Jorge E. Galán,Yale Univ. School of Medicine • Pierluigi Gambetti,Case Western Reserve Univ. • Alfred L. George Jr.,Vanderbilt Univ. • Kun-Liang Guan, Univ. of California,San Diego • Thomas Alan Hamilton, ClevelandClinic Lerner College of Medicine, Case WesternReserve Univ. • Jae Ung Jung, Univ. of SouthernCalifornia • Agnes B. Kane, Brown Univ. • JeannieT. Lee, Massachusetts General Hospital • Se-JinLee, Johns Hopkins Medical School • Linheng Li,Stowers Institute for Medical Research • MalcolmA. Martin, National Institute of Allergy and InfectiousDiseases, NIH • Terry F. McElwain, WashingtonState Univ. • William A. Muller, NorthwesternUniv. Feinberg School of Medicine • ThomasB. Nutman, National Institute of Allergy and InfectiousDiseases, NIH • Stuart H. Orkin, Dana-Farber Cancer Institute • Jing-Hsiung James Ou,Univ. of Southern California Keck School of Medicine• David A. Relman, Stanford Univ. • BertL. Semler, Univ. of California, Irvine • Charles N.Serhan, Brigham and Women’s Hospital • AlanSher, National Institute of Allergy and InfectiousDiseases, NIH • Yang Shi, Harvard Medical School• Patrick F. Sullivan, Univ. of North Carolina, ChapelHill • Cornelis P. Terhorst, Beth Israel DeaconessMedical Center, Harvard Medical School •Stefanie Vogel, Univ. of Maryland School ofMedicine • Qing Kenneth Wang, Cleveland Clinic •Raymond M. Welsh, Univ. of Massachusetts MedicalSchool • Eileen White, The Cancer Institute ofNew JerseySection on NeuroscienceAllan Basbaum, Univ. of California, San Francisco• Edward M. Callaway, The Salk Institute for BiologicalStudies • Marie-Françoise Chesselet, DavidGeffen School of Medicine at UCLA • JacquelineN. Crawley, National Institute of Mental Health,NIH • Robert B. Darnell, Rockefeller Univ. •Michael E. Greenberg, Harvard Medical School• John C. Gore, Vanderbilt Univ. • Lori L. Isom,Univ. of Michigan • Yishi Jin, Univ. of California,San Diego • Frank M. LaFerla, Univ. of California,Irvine • Mark P. Mattson, National Institute onAging, NIH • Freda D. Miller, Hospital for SickChildren/University of Toronto, Canada • EdvardMoser, Kavli Institute for Systems Neuroscience andCentre for the Biology of Memory, Norwegian Univ.of Science and Technology • Mark M. Rasenick, Univ.of Illinois, Chicago • Dan H. Sanes, New York Univ.• Jerry Silver, Case Western Reserve Univ. • RogerD. Traub, IBM T.J. Watson Research Center • Li-Huei Tsai, Massachusetts Institute of Technology •Robert Vassar, Northwestern Univ. Feinberg Schoolof Medicine • Victor Wilson, Rockefeller Univ.Section on Pharmaceutical SciencesPhilip E. Bourne, Univ. of California, San Diego• Ram Ganapathi, Cleveland Clinic • Barry Gold,Univ. of Pittsburgh • Roy S. Wu, National CancerInstitute, NIH • Chung S. Yang, Rutgers, TheState Univ. of New JerseySection on PhysicsEva Y. Andrei, Rutgers, The State Univ. of NewJersey • John A. Becker, Retired, Lawrence LivermoreNational Laboratory • John M. Carpenter,Argonne National Laboratory • Chia-Ling Chien,Johns Hopkins Univ. • Andrew Cleland, Univ. ofCalifornia, Santa Barbara • David Thomas Delpy,UK Engineering and Physical Sciences ResearchCouncil • Steven B. Dierker, Brookhaven NationalLaboratory • Charles S. Fadley, Univ. of California,Davis • Arthur C. Gossard, Univ. of California, SantaBarbara • Richard L. Greene, Univ. of Maryland,College Park • Karl A. Gschneidner Jr., Iowa StateUniv. • William Morris Hartmann, Michigan StateUniv. • Ann Heinson, Univ. of California, Riverside• Walter F. Henning, Argonne National Laboratory• Joseph Incandela, Univ. of California, SantaBarbara • Abolhassan Jawahery, Univ. of Maryland,College Park • Dmitri Kharzeev, BrookhavenNational Laboratory • Walter Kutschera, Univ.of Vienna, Austria • Laurence S. Littenberg,Brookhaven National Laboratory • Paul B.Mackenzie, Fermi National Accelerator Laboratory• Allen Paine Mills Jr., Univ. of California, Riverside• N. Phuan Ong, Princeton Univ. • MichaelChristopher Payne, Cambridge Univ., UnitedKingdom • Tilman Pfau, Universität Stuttgart, Germany• Simon R. Phillpot, Univ. of Florida • KarinM. Rabe, Rutgers, The State Univ. of New Jersey• Kennedy Reed, Lawrence Livermore NationalLaboratory • Per Arne Rikvold, Florida State Univ.• Frances M. Ross, IBM T. J. Watson Research Center• Angel Rubio, Univ. of the Basque Country,Spain • Edward Seidel, National Science Foundation• Bradley M. Sherrill, Michigan State Univ. •Soren P. Sorensen, Univ. of Tennessee, Knoxville •Joseph A. Stroscio, National Institute of Standardsand Technology • Chris G. Van de Walle, Univ. ofCalifornia, Santa Barbara • Victor GeorgievichVeselago, Russian Academy of Sciences • JohnP. Wikswo, Vanderbilt Univ. • John F. Wilkerson,Univ. of North Carolina, Chapel HillSection on PsychologyBrian Baird, U.S. House of Representatives • JohnA. Bargh, Yale Univ. • Lawrence Barsalou, EmoryUniv. • Sian Beilock, Univ. of Chicago • JeanDecety, Univ. of Chicago • Dedre Gentner, NorthwesternUniv. • Charles F. Mactutus, Univ. of SouthCarolina • George R. Mangun, Univ. of California,Davis • John E. Richards, Univ. of South Carolina• Patrick E. Shrout, New York Univ. • Barbara A.Spellman, Univ. of Virginia • Gregory L. Stuart,Univ. of Tennessee, Knoxville • John C. Trueswell,Univ. of Pennsylvania • Sandra R. Waxman, NorthwesternUniv.Section on Social, Economic, andPolitical ScienceRichard P. Appelbaum, Univ. of California, SantaBarbara • Virginia Cain, Centers for Disease Controland Prevention • Lynda Carlson, National ScienceFoundation • Edward G. Carmines, IndianaUniv. • Nicholas A. Christakis, Harvard MedicalSchool • David Collier, Univ. of California, Berkeley• Eileen M. Crimmins, Univ. of Southern California• Jon A. Krosnick, Stanford Univ. • AmanUllah, Univ. of California, RiversideSection on Societal Impacts of Science andEngineeringBruce Bimber, Univ. of California, Santa Barbara• Arlene Blum, Green Science Policy Institute •William Boone Bonvillian, Washington, D.C.Office of Massachusetts Institute of Technology• Beth A. Fischer, Univ. of Pittsburgh • Juan E.Gilbert, Clemson Univ. • Lora Lee Martin, CaliforniaCouncil on Science and Technology • ValerieThomas, Georgia Institute of TechnologySection on StatisticsTimothy G. Gregoire, Yale Univ. • John P. Lehoczky,Carnegie Mellon Univ. • Marepalli Bhaskara Rao,Univ. of Cincinnati • James L. Rosenberger, PennsylvaniaState Univ. • Jane-Ling Wang, Univ. ofCalifornia, Davis • Larry Alan Wasserman, CarnegieMellon Univ. • Linda J. Young, Univ. of Florida• Hongyu Zhao, Yale School of Public Healthwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 425

REVIEWThe Newest Synthesis: Understandingthe Interplay of Evolutionary andEcological DynamicsThomas W. SchoenerThe effect of ecological change on evolution has long been a focus of scientific research. Thereverse—how evolutionary dynamics affect ecological traits—has only recently captured our attention,however, with the realization that evolution can occur over ecological time scales. This newlyhighlighted causal direction and the implied feedback loop—eco-evolutionary dynamics—isinvigorating both ecologists and evolutionists and blurring the distinction between them. Despite somerecent relevant studies, the importance of the evolution-to-ecology pathway across systems is stillunknown. Only an extensive research effort involving multiple experimental approaches—particularlylong-term field experiments—over a variety of ecological communities will provide the answer.Three-quarters of a century ago, the NewSynthesis—an integration of paleontology,systematics, morphology, and genetics—captured the imaginations of evolutionary biologists.Ecology, then less well developed, played arelatively small role in this synthesis, even thoughthe major influence of ecological processes on evolutionarytrajectories was recognized. The dynamicaleffect of evolution on ecology, however, hasonly recently become widely appreciated. Here Ireview the emerging field of eco-evolutionary dynamics,whose major precept is that both directionsof effect—ecology to evolution and evolution toecology—are substantial. The general argument isas follows: Many studies have documented thatecological change affects evolution; indeed, naturalselection is where ecology and evolution meet. Suchstudies of “evolution in action,” i.e., over observabletime scales, show that evolution can be very rapid.This opens the possibility that evolutionary dynamicscan also affect ecology; specifically, evolutioncan occur so quickly that ecological and evolutionarychange may be commensurate in time and mayinteract in a feedback loop. Despite the potential fora directional influence of evolution on ecology, conceptualbolstering from mathematical theory, andsome recent empirical studies, we still don’t knowif the evolution-ecology pathway is frequent andstrong enough in nature to be broadly important.Ecology Affects EvolutionMany studies have shown that evolution is shapedby ecology. The most detailed of these concernongoing observations of natural selection (1)drivingphenotypic response to changing environmentalconditions. Awell-known example is foundin work on a Galápagos ground finch, Geospizafortis. In this species, larger beaks dominated thepopulation after dry years when large seeds weremore available. Correspondingly, after wet yearsthe direction of selection reversed, favoring smallerbeaks that were more appropriate for handlingthe small seeds produced in the wet environment.These results demonstrate an immediately observableadaptive response to selection favoringanimals with the beak best suited for handlingavailable food (2). A second set of examplesinvolves life histories and morphologies of fish.Intense harvest resulted in Atlantic cod (Gadusmorhua) evolving earlier maturation and a smalleradult body size (3). A similar pattern was observedamong guppies (Poecilia reticulata) after experimentalpredator introduction (4). In these cases,rapid adaptation for smaller body size and anearlier age at reproduction occurred in responseABTotalpopulationRate of changeBody andbeak size1.50.5-0.58005002000.50.2-0.119751980 1985 1990 1975 1980 1985 1990Year1976 1978 1980 1982 1984 1986 1988 1990Yearto intense predation. Finally, steelhead and rainbowtrout are both members of the species Oncorhynchusmykiss; however, steelhead migrate,whereas rainbows do not. In an already classic casenot only of adaptation but also of rapid speciation,steelhead trout introduced (in 1910) to a stream isolatedabove a waterfall evolved a rainbow trout nonmigratorylife-style. Rainbow-like steelhead laterable to move across the falls were reproductivelyisolated from the ancestral steelhead population(5). In short, here adaptation to local ecologicalconditions resulted in not only a change in behavior,but also adaptive divergence and subsequentinability to breed with the ancestral population.Evolution Can Be Fast…By definition, observations of evolution in actiondocument extremely rapid evolution. The numberof reliable studies in which natural selectionhas been observed has increased markedly sinceEndler reviewed them in his classic 1986 book,Natural Selection in the Wild.Themostcomprehensivediscussions are by Reznick and colleagues(4, 6), who found 47 studies demonstrating or implyingrapid evolution for a variety of traits (morphological,physiological, life-history, phenological,and behavioral). Most examples involve colonizationevents, i.e., invasion, or local changes in semiisolatedpopulations across a varying environment(7). Although the examples are all from nature (ratherthan the laboratory), humans, by deliberate or accidentalintroduction or other environmental modification,have spurred many of them: Human predatoryactivities are especially effective (8). Indeed, withoutrecent and severe anthropogenic actions, the list4000200000.80.40.0Mediumground finchG. fortisTotal seeds(mg m -2 )Fractionlarge seedsBeakdepthEvolution and Ecology, University of California, Davis, OneShields Avenue, Davis, CA 95616, USA. E-mail: twschoener@ucdavis.eduFig. 1. (A) Dynamics for evolutionary and ecological traits in G. fortis. (B) The annual ecological (solidline) and evolutionary (dashed line) contributions to the total rate of change (per year) in populationgrowth rate. [From Hairston et al.(11)]42628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REVIEWof cases of rapid evolution would be much shorter.In view of such studies, our concept of evolutionarytime is also changing rapidly. As Fussmann et al.(9) highlight: “By rapid evolution we mean evolutionoccurring on time scales tractable in laboratorystudies of population dynamics—up to about 1000generations but typically many fewer.” This wasmuch shorter than the half-million years for evolutionarytime given by the ecologist Slobodkin inhis classic 1961 book, Growth and Regulation ofAnimal Populations (10); his ecological time was“on the order of ten times the generation time of thespecies involved.” Hairston et al.(11) go the farthestin defining “rapid evolution” as “a genetic changerapid enough to have a measurable impact on simultaneousecological change,” thus coming full circle.These reinterpretations of how quickly evolutioncan occur open the door to the possibilitythat evolutionary dynamics can affect ecologicaldynamics—in principle. We have just seen that, becauseevolution can be so fast (4, 6, 8), evolutionaryand ecological time can be commensurate. Indeed,natural selection and population dynamics areboth affected by births and deaths of individuals(12–14). In fact, in the founding paper on selectiongradients (15), mortality in house sparrows (Passerdomesticus) after a winter storm provided an illustrativeexample for the computations. To paraphraseKokkoandLόpez-Sepulcre(16):Assessingfitnessisultimately counting offspring that transmit genesto future generations. Therefore, population dynamics(an ecological entity) can depend on thefitness of population members (an evolutionaryentity). Similarly, an individual’s fitness can dependon the ecological trait of population density,including densities of conspecifics, heterospecificcompetitors, prey, or predators. Although tightlyrelated, however, the link between natural selectionand population dynamics is not entirely straightforward.Specifically, the measure of fitness usedwhen calculating selection [e.g., (15)] is relative toother members of the population (i.e., how manymore offspring does individual A produce relative toindividualB,ortherestofthepopulation).Incontrast,in population dynamics we are concerned with theabsolute output of an individual and how thatcontributes to population properties.As summarized, that ecology affects evolutionis the cornerstone of the natural selectionconcept. But can the reverse also happen to asubstantial extent? More precisely, what is thecontribution to population growth of evolutionaryfactors, such as genetically based phenotypicchange, relative to ecological factors such aschanges in the predation or resource regime orabiotic environmental change? The seminal 2005paper by Hairston et al. (11), in which what hascome to be called the “Geber method” was introduced,first tackled this question. Using aclever technique based on analysis of variance,Hairston et al. partitioned the factors causing yearto-yearvariation in population growth rate into,roughly speaking, evolutionary and ecologicalcomponents. In the finch G. fortis discussed above(2), evolutionary contributions (via beak shapeAAbundance Abundance Abundance AbundanceAbundance3002001505030020016012080140Soay sheepBighornsheepat RamMountainRoe deerMountain goatsand body size) exceeded ecological contributions(via seed density and fraction of large seeds) by afactor of 2.2 (Fig. 1). For Hairston et al.’sdataonthe copepod Onychodiaptomus sanguineus, theevolutionary contribution (via life history: whetherdiapausing or immediately hatching eggs areproduced) was one-fourth the ecological contribution(via fish predation), which they call “lessthan in the finch example, but nevertheless substantial.”They also applied a related method toAbrams and Matsuda’s (17) model of predatorpreydynamics, showing that evolutionary effectsare substantial but not dominant, being 63% ofecological effects. Ezard et al.(18) recently usedthe Hairston et al. methodology to partition populationchange for five ungulate species. Theycharacterize the evolutionary (morphological variables)versus ecological (vegetational and climaticvariables) contributions as “statistically indistinguishable”(Fig. 2). Indeed, the best supportedmodel includes both evolutionary and ecologicalfactors, as well as the interaction between them.BMean absolute change0.1100Bighornsheep at60 Sheep River01970 1980 1990 2000 2010 SSSK BHRM RDTF MGCR BHSRYearPopulation0.50.40.30.2Soay sheepFig. 2. (A) Population numbers for the five ungulate species in Ezard et al.’s (18) study. Time series areshown for the periods of each population analyzed. Species are (from top to bottom) Soay sheep, bighornsheep at Ram Mountain, roe deer, mountain goats, and bighorn sheep at Sheep River. (B) Mean absolutechange of environment (blue bars) and phenotype (white bars) with 95% parametric confidence intervals,calculated following Hairston et al. (11). The change quantifies the effects on population growth ofenvironmental and phenotypic changes, respectively. Study populations follow the order in the left panel.In retrospect, we can ask why the correspondenceof ecological and evolutionary time wasnot recognized [e.g., (10)]. Kinnison and Hairston(19) suggest two major reasons. First, fitnessgains made in the course of selection might notmuch influence the equilibrium population size,e.g., the carrying capacity, because of strong negativedensity-dependent regulation. Saccheri andHanski (20) conclude that explicit demonstrationsof this are few, despite theoretical plausibility.Second, the rate of evolution in the paleontologicalrecord is “ponderous,” and this is inconsistent withrapid evolution observed in contemporary studies.Thompson (21) suggests this is because episodesof one-way directional selection are interspersedwith episodes of stasis or even episodes of selectionin the other direction. Hendry and Kinnison(22), building on Gingerich (23), combined datafrom various studies and showed that the longerthe observation period, the weaker the evolution—thus a kind of bias exists [the same trend wasfound later by Hoekstra et al. (24) for studies ofwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 427

REVIEWselection measured by survival]. To explore theissue farther, Hairston et al. (11) supplementedHendry and Kinnison’s data with per-generationdata for two species—the finch and the copepoddiscussed above—and found that short timeintervals could give very low rates of evolutionas well as high ones [Fig. 3; see also (6, 25)]. Sothere is a bias, but it is mainly that the variance ofthe rates decreases with length of the observationperiod. Further, that selection can frequently changedirection has recently become well establishedthanks to Siepielski et al.’s (26) reviewof89studies incorporating 5519 estimates of selection.As a result of these and other studies, a frequentcorrespondence of ecological and evolutionarytime is now widely accepted. For example,Carroll et al.(27)say“the real distinction betweenmicroevolution [short-term] and macroevolution[long-term] may lie only in the degree to whichthe factors causing evolution are fluctuating orare gradually and persistently directional, andnot in the ecological significance of that evolution.”Indeed, the entire issue has been raised almostto the level of a scientific manifesto, whereopposition is considered misguided and fruitless.Thus, Kinnison and Hairston (19)couldsay:“The notion that evolutionary processes are vanishinglyslow, and that contemporary evolutionis exceptional, does not match current scientificknowledge and is counterproductive.”…But Few Empirical Examples Exist ofEvolution Affecting Ecological DynamicsFour recent reviews (28–31) have shown that geneticvariation within a species can affect its ecologicalcommunity. It is now even becoming possibleto generalize about what ecological units are mostaffected: Bailey et al. (30) showed that effects ofwithin-plant-species genetic variation are strongestat the individual level (phytochemistry, physiology,morphology) and weakest at the community (speciesrichness, total abundance, community composition)and ecosystem (carbon accumulation,productivity, soil-nutrient dynamics) levels.These reviews generally deal with static properties,or what Losos (32) has called “retrospectivestudies, how present day ecological processes canbe understood as the outcome of historical events.”In contrast, to understand the ongoing effects ofevolution on population dynamics, we need studiesover multiple generations in real time. How manymight there be? An attempt to determine this numberprecisely was made in 2007 by Fussmann et al.(9), who isolated “the handful of [studies] that…come close to providing empirical support for ecoevolutionarycommunity dynamics.” Their criteriawere as follows: (i) Does the study document changeof abundance of multiple populations over severalgenerations? (ii) Is there a record of genetic frequenciesand their changes over time? (iii) Is there aplausible mechanistic link between ecological andevolutionary dynamics? (iv) Is there a control reportingecological dynamics in the absence ofevolution? No single study fulfilled all criteria, andonly eight studies came close to qualifying. Ofthese, all satisfy criterion 1, but five fail at criterion2 or 4, and three meet neither (33). Six studieswere performed in the laboratory and two in thefield (the latter were observational, not experimental,and include the finch study discussed above).Is the lack of field-experimental studies adeficiency? Two properties of field experimentsargue that it is. First, field conditions are natural,incorporating most of the background of thephenomena we are trying to understand. Second,experimental design is rigorous in that it employsOnychodiaptomussanguineusRate of evolution(log 10haldanes)0--1-2-3-4-5GeospizafortisGeospiza max. rateMean rateMin. raterandom selection of arenas for the various treatments.Four field-experimental studies of howevolution affects ecological properties haveappeared since the review of Fussmann et al.—would any meet their criteria? Harmon et al.(34)showed that combinations of two species ofsticklebacks Gasterosteus aculeatus (from differenthabitats and with unique morphologies thathad evolved repeatedly from a common ancestor)differentially affected community-ecologicalproperties (primary production, dissolved organicmatter, prey species diversity). Johnson et al. (35)experimentally generated selection on biomass,life history, and herbivore resistance of the primrose,Oenothera biennis. The resulting evolutionarychanges affected abundance and diversity ofthe associated arthropods. Palkovacs et al. (36)Onychodiaptomus max. rateMean rateMin. rate0 0.5 1 1.5 2 2.5Interval (log 10generations)Fig. 3. Relation between rate of phenotypic evolution and numberof generations over which measurements were made. Original figure(small dots) is from Hendry and Kinnison (22) with new calculationsadded for the finch G. fortis and the copepod O. sanguineus,showingthe slowest, most rapid, and the average rates of evolution (pergeneration) over the 30- and 10-year, respectively, periods of study.[From Hairston et al.(11)]used combinations of guppy and killifish (Rivulushartii) populations with different evolutionaryhistories to simulate invasion, one-species evolution,and coevolution. They found responses inecosystem properties such as algal biomass anddecomposition. In some cases, the evolutionarytreatments had larger effects than the ecological(invasion) treatment. Bassar et al. (37), also usingguppies, showed that different phenotypes were associatedwith different values of ecosystem structures(biomasses of various groups) and functions(productivity, nutrient flux, leafdecompositionrates), much asHarmon et al. showed. Althoughall experimental and field [ormesocosm (38)] based, none ofthese studies was dynamic inFussmann et al.’s sense (with thepartial exception of Johnson et al.,which ran for 2 years). Rather,most fall into Losos’s(32) “retrospective”category—how ecologycan result from historical events.Nine studies appearing afterFussmann et al.’s review wentto press are multigenerational.Three were field-observational andon mammals: Pelletier et al. (39)on Soay sheep, Ozgul et al.(40)onyellow-bellied marmots (Marmotaflaviventris), and Ezard et al.’sungulate study (18) discussed above.The six others were experimentalbut were conducted in the laboratory:Kerr et al. (41), Bull et al.(42), and Brockhurst et al. (43)on bacteria and phage; Lennonand Martiny (44) on cyanobacteriaand viruses; terHorst et al.(45) on protozoans and mosquitoes;and Becks et al.(46)onalgaeand rotifers. To elaborate on one,Kerr et al. devised a metapopulationusing plates on which thebacterial host Escherichia coli inhabitedmedia-filled wells. Thepathogenic phage always wipedout the bacteria within each welland then went extinct, resulting in an empty wellthat could be recolonized by bacteria. The setupselected over time for “prudent” phage if migrationbetween wells was restricted, whereas “rapacious”phage was selected under unrestrictedmigration. The evolution of the prudent variety isreminiscent of Myxoma virus and rabbits, one ofFussmann et al.’s qualifying examples: Less virulentvirus strains were selected for. These laboratorystudies provide valuable insights into themechanics of the eco-evolutionary feedback process,and the observational studies analyze longtermfield data in fundamentally new ways. Still,no field-experimental study has yet been published.An example of the form such a study can take isgiven in Fig. 4, which sketches an ongoing studyusing two species of island lizards.42828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REVIEWA B CSmall lizardperches onground, rocks,and low invegetation,diminishingarthropodsthere.Before introduction oflarge lizardPopulationsize of smalllizard reduced,and survivorsescape intohigher, thinnervegetation.Feedback Loops Between Evolution andEcology: Will They Be Important or Trivial?Dobzhansky notoriously said in 1964: “Nothingin biology makes sense except in the light of evolution.”This was supplanted half a century laterby Grant and Grant’s(2): “Nothing in evolutionarybiology makes sense except in the light of ecology.”Pelletier et al. (12) quickly followed with“Nothing in evolution or ecology makes senseexcept in the light of the other,” and this sentimentis pretty much where we are today. If ecology affectsevolution (long supported) and evolutionaffects ecology (becoming increasingly supported),then what? The transformed ecology might affectevolution, and so on, back and forth in a feedbackloop. Kokko and Lόpez-Sepulcre (16) call this“ecogenetic feedback”: “If density influences everyone’sreproductive prospects to the same extent,one has merely restated the ecological concept ofdensity dependence. But if density variation hasa differential effect on individual fitness dependingon…phenotype, we have a feedback loop. Inthis loop, individual behavior or life history, influencedby genes, has an effect on population dynamics…andthe resulting change in populationdynamics in turn…[may] differentially favour…[certain] genotypes…in the population….”The idea of feedback between evolution andecology is not itself new. One of Fussmann et al.’sexamples, Pimentel (47), developed his geneticfeedback concept and tested it in the lab half a centuryago. Taper and Case’s 1985 character displacementtheory (48) modeled coevolution of competingspecies’ phenotypes and food niches while allowingfood resources to change dynamically. Chitty alsoused a feedback idea to explain cycling selectionbetween dispersers and homebodies in rodents (49);however, little supporting evidence was found andexplicit modeling gave the opposite result—naturalselection diminished cyclical behavior (50, 51).Shortly after introductionSmall lizardadapts byevolving shorterlimbs for higher,thinner vegetation,diminishingarthropods thereand increasingits population size.A few years laterFig. 4. Experimental approach used in ongoing study of eco-evolutionary dynamics in Caribbeanlizards. (A) Before introduction of the large predatory lizard Leiocephalus carinatus (curly-tailed lizard),the smaller Anolis sagrei isfoundbothonthegroundandtrunksoftrees.(B) Introduction of predatorsmay result in reduced population size and change in habitat use by A. sagrei. (C) Rapid evolutionof A. sagrei could be precipitated by change in selection pressures, and this may have amplifyingeffects on the system in the form of altered arthropod densities and distributions, as well as increasedpopulation size of A. sagrei [as argued in Strauss et al. (54)].Given this long history, what then is new aboutthe emerging field of eco-evolutionary dynamics?Palkovacs et al. (36) say:“it remains a frontier toexperimentally examine the ecosystem effects ofdynamically evolving (and coevolving) populationsin the wild…one potentially critical element thatcan only be captured using dynamic experiments isthe eco-evolutionary feedback….” Herein lies thenovelty current researchers are exploring. Becauseour investigations to reveal the role of such feedbackare just beginning, it is hard to guess at the outcome.There are large open questions, the most importantof which, according to Thompson (21), “is whetherthe persistence of interactions and the stability ofcommunities truly rely upon ongoing rapid evolution…orwhether such rapid evolution is ecologicallytrivial.” It is a question well worthy of thelargeresearcheffortitwilltaketolearntheanswer.References and Notes1. Strictly speaking, observing natural selection is not the sameas observing evolution. Evolution is defined as a change ingene frequency, and that selection resulted in geneticchange is not always demonstrated. Among Reznick et al.’s(4) 47 studies, 30 provide “reasonable evidence for agenetic basis” such as estimates of trait heritability,reciprocal transplants, and common-garden experiments.2. P. Grant, R. Grant, How and Why Species Multiply:The Radiation of Darwin's Finches (Princeton Univ. Press,Princeton, NJ, 2008).3. E. M. Olsen et al., Nature 428, 932 (2004).4. D. Reznick, H. Rodd, L. Nunney, in Evolutionary ConservationBiology, R. Ferrière, U. Dieckmann, D. Couvet, Eds.(Cambridge Univ. Press, Cambridge, 2004).5. D. E. Pearse et al., J. Hered. 100, 515 (2009).6. D. N. Reznick, C. K. Ghalambor, Genetica 112–113,183 (2001).7. The second of these cases is in the context of ametapopulation, defined as a set of populations connectedby some but not overwhelming immigration between them.8. C. T. Darimont et al., Proc. Natl. Acad. Sci. U.S.A. 106,952 (2009).9. G. F. Fussmann, M. Loreau, P. A. Abrams, Funct. Ecol. 21,465 (2007).10. L. B. Slobodkin, Growth and Regulation of AnimalPopulations (Holt, Rinehart and Winston, New York, 1961).11. N. G. Hairston Jr., S. P. Ellner, M. A. Geber, T. Yoshida,J. A. Fox, Ecol. Lett. 8, 1114 (2005).12. F. Pelletier, D. Garant, A. P. Hendry, Philos. Trans. R. Soc. B364, 1483 (2009).13. T. Coulson et al., Proc. Biol. Sci. 273, 547 (2006).14. T. Coulson, S. Tuljapurkar, D. Z. Childs, J. Anim. Ecol. 79,1226 (2010).15. R. Lande, S. J. Arnold, Evolution 37, 1210 (1983).16. H. Kokko, A. López-Sepulcre, Ecol. Lett. 10, 773 (2007).17. P. Abrams, H. Matsuda, Evolution 51, 1742 (1997).18. T. H. G. Ezard, S. D. Côté, F. Pelletier, Philos. Trans.R. Soc. B 364, 1491 (2009).19. M. J. Kinnison, N. G. Hairston Jr., Funct. Ecol. 21,444 (2007).20. I. Saccheri, I. Hanski, Trends Ecol. Evol. 21, 341 (2006).21. J. N. Thompson, Trends Ecol. Evol. 13, 329 (1998).22. A. P. Hendry, M. T. Kinnison, Evolution 53, 1637 (1999).23. P. D. Gingerich, Science 222, 159 (1983).24. H. E. Hoekstra et al., Proc. Natl. Acad. Sci. U.S.A. 98,9157 (2001).25. S. Estes, S. J. Arnold, Am. Nat. 169, 227 (2007).26. A. M. Siepielski, J. D. DiBattista, S. M. Carlson, Ecol. Lett.12, 1261 (2009).27. S. P. Carroll, A. P. Hendry, D. N. Reznick, C. W. Fox,Funct. Ecol. 21, 387 (2007).28. T. G. Whitham et al., Nat. Rev. Genet. 7, 510 (2006).29. A. R. Hughes, B. D. Inouye, M. T. J. Johnson,N. Underwood, M. Vellend, Ecol. Lett. 11, 609 (2008).30. J. K. Bailey et al., Philos.Trans.R.Soc.B364, 1607 (2009).31. J. R. Haloin, S. Y. Strauss, Ann. N.Y. Acad. Sci. 1133,87 (2008).32. J. B. Losos, Annu. Rev. Ecol. Syst. 25, 467 (1994).33. The study with rotifers and green algae in their tablefulfills all criteria if one adds a separate genetic study.Also, the yeast-water-flea study has recently beensupplemented by observations of epidemics being cutshort by rapid evolution of resistance (52).34. L. J. Harmon et al., Nature 458, 1167 (2009).35. M. T. J. Johnson, M. Vellend, J. R. Stinchcombe,Philos. Trans. R. Soc. B 364, 1593 (2009).36. E. P. Palkovacs et al., Philos. Trans. R. Soc. B 364,1617 (2009).37. R. D. Bassar et al., Proc. Natl. Acad. Sci. U.S.A. 107,3616 (2010).38. Mesocosm experiments use replicate arenas, usuallyplaced in a field setting, that are intermediate in sizebetween microcosms (inevitably done in the laboratory)and typical field plots.39. F. Pelletier, T. Clutton-Brock, J. Pemberton, S. Tuljapurkar,T. Coulson, Science 315, 1571 (2007).40. A. Ozgul et al., Nature 466, 482 (2010).41. B. Kerr, C. Neuhauser, B. J. M. Bohannan, A. M. Dean,Nature 442, 75 (2006).42. J. J. Bull, J. Millstein, J. Orcutt, H. A. Wichman, Am. Nat.167, E39 (2006).43. M. A. Brockhurst, A. Buckling, V. Poullain, M. E. Hochberg,Evolution 61, 1238 (2007).44. J. T. Lennon, J. B. H. Martiny, Ecol. Lett. 11, 1178 (2008).45. C. P. terHorst, T. E. Miller, D. R. Levitan, Ecology 91,629 (2010).46. L. Becks, S. P. Ellner, L. E. Jones, N. G. Hairston Jr.,Ecol. Lett. 13, 989 (2010).47. D. Pimentel, Am. Nat. 95, 65 (1961).48. M. L. Taper, T. J. Case, Ecology 66, 355 (1985).49. D. Chitty, Can. J. Zool. 38, 99 (1960).50. N.C.Stenseth,R.A.Ims,inThe Biology of Lemmings,N.C.Stenseth,R.A.Ims,Eds.(AcademicPress,London,1993), p. 62.51. However, in general, selection may stabilize or destabilizepopulation cycles (53).52. M. A. Duffy, S. R. Hall, C. E. Cáceres, A. R. Ives,Ecology 90, 1441 (2009).53. P. A. Abrams, Annu. Rev. Ecol. Syst. 31, 79 (2000).54. S. Y. Strauss, J. A. Lau, T. W. Schoener, P. Tiffin,Ecol. Lett. 11, 199 (2008).55. I thank S. P. Ellner, G. F. Fussmann, N. G. Hairston Jr.,and J. B. Losos for very helpful comments. Supported bygrants from NSF.10.1126/science.1193954www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 429

RESEARCH ARTICLESRapid Pneumococcal Evolution inResponse to Clinical InterventionsNicholas J. Croucher, 1 Simon R. Harris, 1 Christophe Fraser, 2 Michael A. Quail, 1 John Burton, 1Mark van der Linden, 3 Lesley McGee, 4 Anne von Gottberg, 5 Jae Hoon Song, 6 Kwan Soo Ko, 7Bruno Pichon, 8 Stephen Baker, 9 Christopher M. Parry, 9 Lotte M. Lambertsen, 10 Dea Shahinas, 11Dylan R. Pillai, 11 Timothy J. Mitchell, 12 Gordon Dougan, 1 Alexander Tomasz, 13Keith P. Klugman, 4,5,14 Julian Parkhill, 1 William P. Hanage, 2,15 Stephen D. Bentley 1 *Epidemiological studies of the naturally transformable bacterial pathogen Streptococcuspneumoniae have previously been confounded by high rates of recombination. Sequencing240 isolates of the PMEN1 (Spain 23F -1) multidrug-resistant lineage enabled base substitutionsto be distinguished from polymorphisms arising through horizontal sequence transfer. More than700 recombinations were detected, with genes encoding major antigens frequently affected.Among these were 10 capsule-switching events, one of which accompanied a population shiftas vaccine-escape serotype 19A isolates emerged in the USA after the introduction of theconjugate polysaccharide vaccine. The evolution of resistance to fluoroquinolones, rifampicin,and macrolides was observed to occur on multiple occasions. This study details how genomicplasticity within lineages of recombinogenic bacteria can permit adaptation to clinicalinterventions over remarkably short time scales.Streptococcus pneumoniae is a highly recombinogenichuman nasopharyngealcommensal and respiratory pathogen estimatedto be responsible for a global burden ofalmost 15 million cases of invasive disease in2000 (1). Since the 1970s, the susceptibility of thepneumococcal population to antibiotics has decreased,largely as a consequence of the emergenceand spread of a few multidrug-resistant clones (2).1 The Wellcome Trust Sanger Institute, Wellcome Trust GenomeCampus, Hinxton, Cambridge CB10 1SA, UK. 2 Department ofInfectious Disease Epidemiology, Imperial College, St Mary'sCampus, Norfolk Place, London W2 1PG, UK. 3 Institute forMedical Microbiology, National Reference Center for Streptococci,University Hospital, RWTH Aachen, Pauwelsstrasse 30,52074 Aachen, Germany. 4 Respiratory Diseases Branch, Centersfor Disease Control and Prevention, Atlanta, GA 30333,USA. 5 Respiratory and Meningeal Pathogens Research Unit,National Institute for Communicable Diseases of the NationalHealth Laboratory Service and University of Witwatersrand,Johannesburg 2000, South Africa. 6 Samsung Medical Centre,Sungkyunkwan University School of Medicine and Asia PacificFoundation for Infectious Disease, Seoul, South Korea. 7 Departmentof Molecular Cell Biology, Sungkyunkwan UniversitySchool of Medicine, Suwon 440-746, South Korea. 8 Respiratoryand Systemic Infection Laboratory, Health Protection AgencyCentre for Infections, London NW9 5HT, UK. 9 The Hospital forTropical Diseases, Wellcome Trust Major Overseas Programme,Oxford University Clinical Research Unit, Ho Chi Minh City,Vietnam. 10 Department of Microbiological Surveillance andResearch, Statens Serum Institut, 2300 Copenhagen S, Denmark.11 Department of Laboratory Medicine and Pathobiology, Universityof Toronto and Ontario Agency for Health Protection andPromotion, Toronto, Ontario, M5G 1V2, Canada. 12 Institute ofInfection, Immunity and Inflammation, University of Glasgow,Glasgow G12 8TA, UK. 13 Laboratory of Microbiology, The RockefellerUniversity, 1230 York Avenue, New York, NY 10021, USA.14 Hubert Department of Global Health, Rollins School of PublicHealth. and Division of Infectious Diseases, School of Medicine,Emory University, Atlanta, GA 30322, USA. 15 Department ofEpidemiology, Harvard School of Public Health, 677 HuntingtonAvenue, Boston, MA 02115, USA.*To whom correspondence should be addressed. E-mailsdb@sanger.ac.uk.The first recognized example was PneumococcalMolecular Epidemiology Network clone 1 (PMEN1),an S. pneumoniae lineage typically identified asbeing sequence type (ST) 81 and serotype 23F, aswell as exhibiting resistance to multiple antibiotics,including penicillin. The genome sequence of thefirst identified member of the clone, isolated in ahospital in Barcelona in 1984, revealed that it hadacquired a Tn5252-type integrative and conjugativeelement (ICE) that carries a linearized chloramphenicolresistance plasmid and a Tn916-type elementwith a tetM tetracycline-resistance gene (3).This lineage was subsequently found to bepresent in Africa, Asia, and America (4–8) and,by the late 1990s, was estimated to be causingalmost 40% of penicillin-resistant pneumococcaldisease in the USA (9). Following the introductionof a heptavalent conjugate polysaccharidevaccine (PCV7) in many countries since 2000,which includes capsule type 23F as one of its sevenantigens, a decrease in the frequency of serotype23F invasive disease and carriage has beenobserved (10). However, this has been accompaniedby a rise in disease caused by nonvaccine-serotypepneumococci, such as the multidrug-resistantserotype 19A strains now common in the USA(11). Based on evidence from multilocus sequencetyping (MLST) from America (12) and Europe(13), some of these are thought to include capsuleswitch variants of PMEN1 lineage.To study how this lineage has evolved as ithas spread, we used Illumina sequencing of multiplexedgenomic DNA libraries to characterize aglobal collection of 240 PMEN1 strains isolatedbetween 1984 and 2008. Strains were identifiedeither by using MLST or on the basis of serotype,drug-resistance profile, and targeted polymerasechain reaction (14). Selected isolates were distributedamong Europe (seven countries, 81 strains);South Africa (37 strains); America (six countries,54 strains); and Asia (eight countries, 68 strains)(table S1) and included a variety of drug-resistanceprofiles, as well as five serotypes distinct fromthe ancestral 23F: 19F (also included in PCV7),19A, 6A, 15B, and 3.Construction of the phylogeny. Sequencereads were mapped against the complete referencechromosome of S. pneumoniae ATCC700669 (3) and, by using the criteria described inHarris et al.(15), 39,107 polymorphic sites wereidentified within the PMEN1 lineage. Maximumlikelihood analysis produced a phylogeny with ahigh proportion of homoplasic sites (23%) and aweak correlation between the date of a strain’sisolation and its distance from the root of the tree(Pearson correlation, N = 222, R 2 = 0.05, p =0.001) (fig. S1), which suggested that variationwas primarily arising through incorporation ofimported DNA and not through steady accumulationof base substitutions. As these strains areclosely related, sequences acquired by recombinationcould be identified as loci with a highdensity of polymorphisms. These events were reconstructedonto the phylogeny and, by using aniterative algorithm (14), an alignment and treebased on vertically inherited base substitutionsalone was generated.From this analysis (Fig. 1), a total of 57,736single-nucleotide polymorphisms (SNPs) wereidentified, 50,720 (88%) of which were introducedby 702 recombination events. This gives aFig. 1. Phylogeography and sequence variation of PMEN1. (A) Global phylogeny of PMEN1. The maximumlikelihood tree, constructed using substitutions outside of recombination events, is colored according tolocation, as reconstructed through the phylogeny by using parsimony. Shaded boxes and dashed linesindicate isolates that have switched capsule type from the ancestral 23F serotype. †Independent switches tothe same serotype are distinguished by annotation with daggers. Specific clades referred to in the text aremarked on the tree: A (South Africa), I (International), V (Vietnam), S (Spain 19A), and U (USA 19A). (B)Recombinations detected in PMEN1. The panel shows the chromosomal locations of the putative recombinationevents detected in each terminal taxon. Red blocks are recombinations predicted to have occurred on aninternal branch and, therefore, are shared by multiple isolates through common descent. Blue blocks arerecombinations predicted to occur on terminal branches and hence are present in only one strain. The greenblocks indicate recombinations predicted to have occurred along the branch to the outgroup (S. pneumoniaeBM4200), used to root the tree. (C) Biological relevance of recombination. The heat map shows the densityof independent recombination events within PMEN1 in relation to the annotation of the reference genome.All regions that have undergone 10 or more recombination events are marked and annotated (Tn916 isencompassed within ICESpn23FST81).43028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

psrPIRESEARCH ARTICLESpspCProphage φMM1-2008ICESpn23FST81pspAcps locus110000002000000AWestern Europe South AfricaEastern EuropeSoutheast AsiaNorth America Central and SouthAmericaAVUS19F †19F15B319A ††CNumber of recombinationsaffecting a baseB6A1419A19F ††19A †0 > 101 2 3 4 5 6 7 8 9www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 431

RESEARCH ARTICLESper site r/m ratio (the relative likelihood that apolymorphism was introduced through recombinationrather than point mutation) of 7.2, lessthan the previously calculated value of ~66 fromMLST data (16). By removing recombinationevents from the phylogeny, the number of homoplasicsites is reduced by 97%, and the tree hassignificantly shortened branches, such that rootto-tipdistance more strongly correlates with dateof isolation (R 2 = 0.46, p =

other European isolates. The estimated times oforigin for clades U (1996; 95% credible interval1992–1999) and S (1998; 95% credible interval1996–1999) both predate the introduction ofPCV7, and accordingly a third 19A switch, fromCanada, was isolated in 1994. Hence, it appearsthat these changes in serotype after vaccine introductionresult from an expansion of preexistingcapsular variants, which were relatively uncommonand not part of the predominant population,and would have therefore been difficult to detectbefore the existence of the selection pressureexerted by the vaccine.Seven further serotype-switching events canbe detected in the data (Fig. 2), including threeswitches to serotype 19F. The polyphyletic natureof these 19F isolates is supported by the variationobserved between the acquired cps loci, as is alsothe case for the 19A isolates (fig. S5). The previouslyknown switches to serotypes 3, 6A, and15B are only found to occur once each in thephylogeny, and in addition, a single Korean samplethat had not been typed was identified as aserotype 14 variant by mapping reads to knowncps loci (20). The recombination events leadingto these switches ranged from 21,780 bp to39,182 bp in size, with a mean of 28.2 kb. Only35 homologous recombinations of an equivalentsize or larger occur elsewhere in the genome;most such events are much smaller (fig. S3),ABS. pneumoniaeATCC 700669Spain, 1984which makes it surprising that serotype switchingoccurs with such frequency and which indicates arole for balancing selection at this locus. Additionally,the span of these events appears to belimited by the flanking penicillin-binding proteingenes, the sequences of which are crucial in determiningb-lactam resistance in pneumococci(21). Only the recombination causing the switchto serotype 3 affects one of these, and it introducesjust a single SNP into the pbpX CDS, which doesnot appear to compromise the strain’s penicillinresistance (table S1). Hence, the positioning ofthese two genes may hinder the transfer of capsulebiosynthesis operons from penicillin-sensitiveto penicillin-resistant pneumococci via larger recombinations,although size constraints alone couldalso cause such a distribution.Resistance to non–b-lactam antibiotics. Thestrong selection pressures exerted by antibioticson the PMEN1 lineage are manifest as multipleexamples of geographically disparate isolatesconverging on common resistance mechanisms.Single base substitutions causing reduced susceptibilityto some classes of antibiotics haveoccurred multiple times throughout the phylogeny,as observed in S. aureus (15)andSalmonellaTyphi (22) populations, including mutations inparC, parE, andgyrA, which cause increasedresistance to fluoroquinolone antibiotics (23), andchanges in rpoB causing resistance to rifampicinOmega element Tn917 Mega elementOmega repressoraph3’Omega repressorermBTransposonermBResolvaseorf23orf21orf20orf18Transposaseorf16orf15UmuC-type proteinUmuD-type proteinmelorf14orf131 17517meftetM(24). The S79F, S79Y, and D83N mutations (25)in parC are estimated to occur nine, three, andfive times, respectively, in PMEN1; additionally,D435N in the adjacent parE gene is found tohappen three times. The S81F and S81Y substitutions,in the same position of gyrA, are found fourand two times, respectively. None of these mutationsare predicted to have been introduced byrecombination, whereas changes at position H499of rpoB causing rifampicin resistance are introducedtwice by horizontal transfer and three timesby means of base substitution.Resistance to macrolide antibiotics tends notto derive from SNPs, but from acquisition of CDSsfacilitating one of the two common resistancemechanisms: methylation of the target ribosomalRNA by erm genes and removal of the drug fromthe cell by the macrolide efflux (mef )-type effluxpumps. Both can be found in the PMEN1 population,and in all cases, the genes appear to beintegrated into the Tn916 transposon (Fig. 3). Theyare carried by three different elements. Tn917,consistingof an ermB gene with an associated transposonand resolvase, inserts into open readingframe orf9 of Tn916 (26). A second has been characterizedas the macrolide efflux genetic assembly(mega) element (27), which carries a mef/mel effluxpump system and, in PMEN1, inserts upstream oforf9. A third element (henceforth referred to as anomega element, for omega and multidrug-resistanceorf9orf8Integrase(pseudogene)RESEARCH ARTICLESS. pneumoniae9409France, 2002S. pneumoniaeH034800032UK, 2003S. pneumoniae11930SpainS. pneumoniae11928SpainS. pneumoniaeMalM6Malaysia, 19991 227651 203701 256451 227931 23032S. pneumoniae23771Russia, 20041 25880Fig. 3. Acquisition of macrolide-resistance cassettes. The three full-lengthresistance cassettes are shown in (A): the omega element, which carries anaph3′ aminoglycoside-resistance gene and an ermB macrolide-resistance gene;Tn917, which carries just the ermB methylase; and the mega element, whichcarries the mel/mef macrolide efflux system. (B) A comparison of the differentTn916 variants in the PMEN1 lineage. Red bands between the sequencesindicate BLASTN matches. The omega element is shaded green when present atfull length and shaded gray where present as a remnant resulting from arecombination between the omega repressor–encoding genes, which concomitantlyleads to the fusion of an omega transcriptional repressor domain to the3′ of the orf20 CDS. Tn917 is boxed in purple, with the two parts of orf9,intowhich it inserts, indicated on either side. The mega element is boxed in orange.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 433

RESEARCH ARTICLESencoding genetic assembly) carries both an ermBgene and an aminoglycoside phosphotransferase,with the latter flanked by direct repeats of omegatranscriptional repressor genes, and is found justdownstream of orf20.Rather than a single acquisition of these elementsoccurring, and the resulting clones spreadingand replacing macrolide-sensitive isolates, allthree elements appear to have been acquiredmultiple times across the phylogeny (fig. S10).The mega element is predominantly shared byisolates in clade I, although the ermB-encodingomega element appears to have been subsequentlyacquired on two occasions, and Tn917 has entirelysuperseded the mega element in one isolate.This is congruent with the known advantages oftarget methylation over drug efflux as a broaderspectrumresistance mechanism (28). In mostinstances of the omega element, only the ermBencodingpart remains; the aminoglycosidephosphotransferase appears to have been deletedthrough a recombination between the omegaencodinggenes, which leaves only an omegadomain–encoding open reading frame fused toorf20 as a scar. This implies that the benefit of theaminoglycoside-resistance element may have notbeen sufficient to maintain it on the ICE.Components of the accessory genome.Other than the insertion of these cassettes, theICE itself is otherwise relatively unchangedthroughout the population. In two cases, the 5′region of the element up to, and including, thelantibiotic synthesis machinery is deleted, whereasthe self-immunity genes are retained (fig. S6).This deletion, which also removes the integratedchloramphenicol-resistance plasmid, is analogousto that observed in the pneumococcal pathogenicityisland–1 of the PMEN1 lineage, in whichall that remains are the immunity genes from aonce-intact lantibiotic synthesis machinery (3). Intwo other cases, the ICE has been supplanted byalternative transposons, both of which are similarcomposites of Tn5252-andTn916-type elements:In S. pneumoniae 11876, a wholesale replacementat the same locus entails the gain of an omegaelement at the expense of losing resistance tochloramphenicol (fig. S7), whereas, in isolate11930, the new ICE inserts elsewhere in thechromosome and carries two ermB genes, as wellas a chloramphenicol acetyltransferase (fig. S8).The only other identified conjugative element wasan ICESt1-type transposon shared by isolates8140 and 8143 (fig. S9), and the only extrachromosomalelement present in the data set wasthe plasmid pSpn1(29), found in isolate SA8.The accessory genome is primarily composedof prophage sequence (fig. S11), with little evidenceof much variation in the complement of metabolicgenes. Viral sequences appear to be a transient featureof the pneumococcal chromosome (fig. S12),with few persisting long enough to be detected inrelated isolates. Four of the new prophage thatcould be assembled were found to insert into thecompetence pilus structural gene comYC, whichlies within an operon shown to be essential forcompetence in S. pneumoniae (30). In two caseswhere such phage appear to be shared throughcommon descent by pairs of isolates, no recombinationevents can be detected that are unique toeither member of the pair, consistent with a nonfunctionalcompetence system in these isolates.Furthermore, assaying the competence of availablelysogenic strains in vitro also suggested thatthese phage insertions abrogate the ability of theirhost to take up exogenous DNA (fig. S13).Discussion. The ability to distinguish verticallyacquired substitutions from horizontallyacquired sequences is crucial to successfully reconstructingphylogenies for recombinogenicorganisms such as S. pneumoniae. Phylogeniesare in turn essential for detailed studies of eventssuch as intercontinental transmission, capsuletype switching, and antibiotic-resistance acquisition.Although current epidemiological typingmethods have indicated that recombination isfrequent among the pneumococcal population,they cannot sufficiently account for its impact onrelations between strains at such high resolution.Only the availability of such a sample of wholegenomesequences makes it possible to adequatelyreconstruct the natural history of a lineage. Thebase substitutions used to construct the phylogenyhave accumulated over about 40 years andoccur, on average, once every 15 weeks. Recombinationshappen at a rate about 1/10th as fast butintroduce a mean of 72 SNPs each. The responsesto the different anthropogenic selectionpressures acting on this variation are distinct. Theapparently weak selection by aminoglycosidesand chloramphenicol has led to the occasionaldeletion of loci encoding resistance to these antibiotics.By contrast, resistance to macrolide antibioticshas been acquired frequently throughoutthe phylogeny, with selection strong enough todrive supplementation or replacement of the resistanceafforded by the mef efflux pump with thebroader-range resistance provided by ermBmediatedtarget modification. The response tovaccine selection is different and involves the depletionof the resident population before it canrespond to the selection pressure and therebyopens the niche to isolates that already expressednonvaccine serotypes. This is likely to reflect thehigh host population coverage of PCV7 in theUSA, as opposed to macrolides or other antibiotics,and the relative likelihood of the recombinationevents that underlie these responses.Over a few decades, this single pneumococcallineage has acquired drug resistance and theability to evade vaccine pressure multiple times,demonstrating the remarkable adaptability of recombinogenicbacteria such as the pneumococcus.PMEN1 is, nevertheless, only one lineage ofthis pathogen. Our relative ignorance of the forcesthat affect bacterial evolution over the long term isillustrated by BM4200 (31), a multidrug-resistantserotype 23F isolate of ST1010 sequenced as theoutgroup for this analysis (Fig. 1). This isolatedates to 1978 but, despite its apparent similarityto PMEN1 strains, has been found very rarelysince then. Hence, this phenotype is not sufficientto guarantee success, an observation supported bythe continued presence of successful, but susceptible,pneumococci in the population (32, 33).Improved understanding of the interplay betweenecology and adaptation in other lineages throughfurther focused sequencing programs may provecrucial to the future control of this, and other,diverse bacterial pathogens.References and Notes1. K. L. O’Brien et al.; Hib and Pneumococcal Global Burdenof Disease Study Team, Lancet 374, 893 (2009).2. L. McGee et al., J. Clin. Microbiol. 39, 2565 (2001).3. N. J. Croucher et al., J. Bacteriol. 191, 1480 (2009).4. R. Muñoz et al., J. Infect. Dis. 164, 302 (1991).5. C. M. Parry et al., Antimicrob. Agents Chemother. 46,3512 (2002).6. K. P. Klugman et al., Eur. J. Clin. Microbiol. Infect. Dis.13, 171 (1994).7. L. McGee, K. P. Klugman, D. Friedland, H. J. Lee,Microb. Drug Resist. 3, 253 (1997).8. A. Tarasi, Y. Chong, K. Lee, A. Tomasz, Microb. DrugResist. 3, 105 (1997).9. A. Corso, E. P. Severina, V. F. Petruk, Y. R. Mauriz,A. Tomasz, Microb. Drug Resist. 4, 325 (1998).10. R. Dagan, K. P. Klugman, Lancet Infect. Dis. 8, 785 (2008).11. M. R. Moore et al., J. Infect. Dis. 197, 1016 (2008).12. C. Muñoz-Almagro et al., Clin. Infect. Dis. 46, 174 (2008).13. C. Ardanuy et al., J. Antimicrob. Chemother. 64, 507 (2009).14. Materials and methods are available as supportingmaterial on Science Online.15. S. R. Harris et al., Science 327, 469 (2010).16. E. J. Feil, J. M. Smith, M. C. Enright, B. G. Spratt,Genetics 154, 1439 (2000).17. H. Ochman, S. Elwyn, N. A. Moran, Proc. Natl. Acad. Sci.U.S.A. 96, 12638 (1999).18. T. L. McCool et al., Infect. Immun. 71, 5724 (2003).19. L. Rose et al., J. Infect. Dis. 198, 375 (2008).20. S. D. Bentley et al., PLoS Genet. 2, e31 (2006).21. K. Trzciński, C. M. Thompson, M. Lipsitch, J. Bacteriol.186, 3447 (2004).22. K. E. Holt et al., Nat. Genet. 40, 987 (2008).23. M. W. Pletz et al., Emerg. Infect. Dis. 12, 1462 (2006).24. M. J. Ferrándiz et al.; Spanish Pneumococcal InfectionStudy Network, Antimicrob. Agents Chemother. 49,2237 (2005).25. Single-letter abbreviations for the amino acid residuesare 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.26. J. H. Shaw, D. B. Clewell, J. Bacteriol. 164, 782 (1985).27. M. Del Grosso, R. Camilli, F. Iannelli, G. Pozzi,A. Pantosti, Antimicrob. Agents Chemother. 50, 3361 (2006).28. M. Del Grosso, J. G. Northwood, D. J. Farrell, A. Pantosti,Antimicrob. Agents Chemother. 51, 4184 (2007).29. P. Romero et al., Plasmid 58, 51 (2007).30. E. V. Pestova, D. A. Morrison, J. Bacteriol. 180, 2701 (1998).31. A. Buu-Hoï, T. Horodniceanu, J. Bacteriol. 143, 313 (1980).32. W. P. Hanage, C. Fraser, J. Tang, T. R. Connor,J. Corander, Science 324, 1454 (2009).33. C. Colijn et al., J. R. Soc. Interface 7, 905 (2010).34. We thank the participating surveillance networks, listedin table S1, and the core informatics, library-making,and sequencing teams at the Wellcome Trust SangerInstitute. Attending authors were grateful for theopportunity to discuss this project at the Permafrostconference. Sequence accession codes are given in tablesS1 and S2. This work was funded by the Wellcome Trust.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/430/DC1Materials and MethodsFigs. S1 to S13Tables S1 to S5References1 October 2010; accepted 8 December 201010.1126/science.119854543428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

The Genetic Landscape of theChildhood Cancer MedulloblastomaD. Williams Parsons, 1,2 * Meng Li, 1 * Xiaosong Zhang, 1 * Siân Jones, 1 * Rebecca J. Leary, 1 *Jimmy Cheng-Ho Lin, 1 Simina M. Boca, 3 Hannah Carter, 4 Josue Samayoa, 4 Chetan Bettegowda, 1,5Gary L. Gallia, 5 George I. Jallo, 5 Zev A. Binder, 5 Yuri Nikolsky, 6 James Hartigan, 7 Doug R. Smith, 7Daniela S. Gerhard, 8 Daniel W. Fults, 9 Scott VandenBerg, 10 Mitchel S. Berger, 11Suely Kazue Nagahashi Marie, 12 Sueli Mieko Oba Shinjo, 12 Carlos Clara, 13 Peter C. Phillips, 14Jane E. Minturn, 14 Jaclyn A. Biegel, 14 Alexander R. Judkins, 16 † Adam C. Resnick, 15 Phillip B. Storm, 15Tom Curran, 16 Yiping He, 17 B. Ahmed Rasheed, 17 Henry S. Friedman, 17 Stephen T. Keir, 17Roger McLendon, 17 Paul A. Northcott, 18 Michael D. Taylor, 18 Peter C. Burger, 19 Gregory J. Riggins, 1,5Rachel Karchin, 4 Giovanni Parmigiani, 20 Darell D. Bigner, 17 Hai Yan, 17 Nick Papadopoulos, 1Bert Vogelstein, 1 ‡ Kenneth W. Kinzler, 1 ‡ Victor E. Velculescu 1 ‡Medulloblastoma (MB) is the most common malignant brain tumor of children. To identify thegenetic alterations in this tumor type, we searched for copy number alterations using high-densitymicroarrays and sequenced all known protein-coding genes and microRNA genes using Sangersequencing in a set of 22 MBs. We found that, on average, each tumor had 11 gene alterations,fewer by a factor of 5 to 10 than in the adult solid tumors that have been sequenced to date. Inaddition to alterations in the Hedgehog and Wnt pathways, our analysis led to the discovery of genesnot previously known to be altered in MBs. Most notably, inactivating mutations of the histone-lysineN-methyltransferase genes MLL2 or MLL3 were identified in 16% of MB patients. These resultsdemonstrate key differences between the genetic landscapes of adult and childhood cancers, highlightdysregulation of developmental pathways as an important mechanism underlying MBs, and identifyaroleforaspecifictypeofhistonemethylationinhumantumorigenesis.Medulloblastomas (MBs) originate in thecerebellum, have a propensity to disseminatethroughout the central nervoussystem, and are diagnosed in approximately1 in 200,000 children less than 15 years old eachyear (1). Although aggressive multimodal therapyhas improved the prognosis for children with MB,a substantial proportion of patients are currentlyincurable (2). Moreover, survivors often suffer considerabletreatment-related morbidities, includingneurocognitive deficits related to radiation therapy.New insights into the pathogenesis of thesetumors are therefore sorely needed. Gene-basedresearch has identified two subgroups of MBs,one associated with mutated genes within theHedgehog pathway and the other associatedwith altered Wnt pathway genes (3, 4). Amplificationsof MYC and the transcription factorOTX2 (5–7), mutations in TP53 (8), and a numberof chromosomal alterations have also beenidentified in MBs. These discoveries have helpeddefine the pathogenesis of MB and have improvedour ability to identify patients who mightbenefit from therapies targeting these pathways.However, most MB patients do not have alterationsin these genes, and the compendium ofgenetic alterations causing MB is unknown.The determination of the human genome sequenceand improvements in sequencing and bioinformatictechnologies have recently permittedgenome-wide analyses of human cancers. To date,the sequences of all protein-encoding genes havebeen reported in more than 80 human cancers(9–20), representing a variety of adult tumors. Inthis study, we provide a comprehensive sequenceanalysis of a solid tumor of childhood. Our datapoint to a major genetic difference between adultand childhood solid tumors and provide new informationto guide further research on this disease.Sequencing strategy. In the first stage of ouranalysis, which we have called the “discoveryscreen,” 457,814 primers (table S1) were used toamplify and sequence 225,752 protein-coding exons,adjacent intronic splice donor and acceptorsites,andmicroRNA(miRNA)genesin22pediatricMB samples (17 samples extracted directlyfrom primary tumors, 4 samples passaged in nudemice as xenografts, and 1 cell line) (tables S2 andS3). Seven metastatic MBs were selected for inclusionin the discovery screen to ensure that highstagetumors were well represented in the study.One matched normal blood sample was sequencedas a control. These analyses corresponded to 50,191transcripts representing at least 21,039 proteinencodinggenes present in the Ensembl, ConsensusCoding Sequences (CCDS), and RefSeq databasesand 715 miRNA genes from the miRBase database.A total of 404,438 primers were described in ourprevious publications, and an additional 53,376primers were newly designed to amplify technicallychallenging genomic regions, miRNAs, or newlydiscovered Ensembl genes (table S1). The data wereassembled for each amplified region and evaluatedusing stringent quality-control criteria, resulting inthe successful amplification and sequencing of 96%of targeted amplicons and 95% of targeted bases inthe 22 tumors. A total of 735 megabases (Mb) oftumor sequence data were generated in this manner.After automated and manual curation of the sequencetraces, regions containing potential sequencealterations (single base mutations and small insertionsand deletions) not present in the referencegenome or single-nucleotide polymorphism (SNP)databases were reamplified in both the tumor andRESEARCH ARTICLESmatched normal tissue DNA and analyzed eitherthrough sequencing by synthesis on an IlluminaGAII instrument or by conventional Sanger sequencing(21). This process allowed us to confirm thepresence of the mutation in the tumor sample anddetermine whether the alteration was somatic (i.e.,tumor-specific). Additionally, mutations identifiedin the four xenograft samples were confirmed tobe present in the corresponding primary tumors.Analysis of sequence and copy number alterations.A total of 225 somatic mutations wereidentified in this manner (Table 1 and table S4). Ofthese, 199 (88%) were point mutations and theremainder were small insertions, duplications, ordeletions, ranging from 1 to 48 base pairs in length.Of the point mutations, 148 (74%) were predictedto result in nonsynonymous changes, 42 (21%)were predicted to be synonymous, and 9 (5%) werelocated at canonical splice site residues that werelikely to alter normal splicing. Of the 225 somaticmutations, 36 (16%) were predicted to prematurelytruncate the encoded protein, either through newlygenerated nonsense mutations or through insertions,duplications, or deletions leading to a changein reading frame. The mutation spectrum observedfor MB was similar to those seen in pancreatic,1 Ludwig Center for Cancer Genetics and Therapeutics andHoward Hughes Medical Institute, Johns Hopkins KimmelCancer Center, Baltimore, MD 21231, USA. 2 Texas Children’sCancer Center and Departments of Pediatrics and Molecularand Human Genetics, Baylor College of Medicine, Houston TX77030, USA.3 Department of Biostatistics, Johns HopkinsBloomberg School of Public Health, Baltimore, MD 21205,USA. 4 Department of Biomedical Engineering, Institute forComputational Medicine, Johns Hopkins Medical Institutions,Baltimore, MD 21218, USA. 5 Department of Neurosurgery,Johns Hopkins University School of Medicine, Baltimore, MD21231, USA. 6 GeneGo,Inc.,St.Joseph,St.Joseph,MI49085,USA. 7 Beckman Coulter Genomics, Inc., Danvers, MA 01923,USA. 8 Office of Cancer Genomics, National Cancer Institute,National Institutes of Health, Department of Health andHuman Services, Bethesda, MD 20892, USA. 9 Department ofNeurosurgery, University of Utah School of Medicine, Salt LakeCity, UT 84132, USA. 10 Department of Pathology, Division ofNeuropathology, University of California–San Diego, San Diego,CA, USA. 11 Department of Neurological Surgery, University ofCalifornia–San Francisco, San Francisco, CA 94143, USA. 12 Departmentof Neurology, School of Medicine of University of Sao Paulo,Sao Paulo, Brazil. 13 Pio XII Foundation, Barretos Cancer Hospital,Sao Paulo, Brazil. 14 Department of Pediatrics, The Children’sHospital of Philadelphia, Philadelphia, PA 19104, USA. 15 Divisionof Neurosurgery, The Children’s Hospital of Philadelphia,Philadelphia, PA 19104, USA. 16 Department of Pathology andLaboratory Medicine, The Children’s Hospital of Philadelphia,Philadelphia, PA 19104, USA. 17 The Preston Robert Tisch BrainTumor Center, Pediatric Brain Tumor Foundation Institute,Department of Pathology, and Department of Surgery, DukeUniversity Medical Center, Durham, NC 27710, USA. 18 Divisionof Neurosurgery and Program in Developmental and Stem CellBiology, Hospital for Sick Children, University of Toronto, Toronto,Ontario M5G1L7, Canada. 19 Department of Pathology, JohnsHopkins University School of Medicine, Baltimore, MD 21231,USA. 20 Department of Biostatistics and Computational Biology,Dana-Farber Cancer Institute, and Department of Biostatistics,Harvard School of Public Health, Boston, MA, USA.*These authors contributed equally to this work.†Present address: Department of Pathology and LaboratoryMedicine, Children’s Hospital of Los Angeles, Los Angeles,CA 90027, USA.‡To whom correspondence should be addressed: velculescu@jhmi.edu (V.E.V.); kinzlke@jhmi.edu (K.W.K.); vogelbe@gmail.com (B.V.)www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 435

RESEARCH ARTICLEScolorectal, glial, and other malignancies (22), with5′-CG to 5′-TA transitions observed more commonlythan other substitutions (Table 1). Such transitionsare generally associated with endogenousprocesses, such as deamination of 5-methylcytosineresidues, rather than exposure to exogenous carcinogens(23).The distribution of somatic mutations among the22 MBs is illustrated in Fig. 1. Two key differenceswere observed in this cancer as compared to thetypical adult solid tumor. First, the average numberof nonsilent somatic mutations (nonsynonymousmissense, nonsense, indels, or splice site alterations)per MB patient was only 8.3, which is fewer by afactor of 5 to 10 than the average number of alterationsdetected in the previously studied solidtumor types (Table 1). Second, the proportion ofnonsense mutations was more than twice as high asexpected given the mutation spectra observed in thistumor type (P

RESEARCH ARTICLEStion and length of the gene. The lower the passengerprobability score, the less likely it is that mutationsin the specific gene represent passengers. Passengerprobability scores of the candidate cancer genes(CAN-genes) identified in MB are listed in Table 2.At the individual mutation level, the Cancer-Specific High-Throughput Annotation of SomaticMutations (CHASM) score is a metric reflecting thelikelihood that a missense mutation alters the normalfunction of the respective protein and provides aselective advantage to the tumor cell (25). TheCHASM score is based on 73 biochemical features,including conservation of the wild-type amino acidand the mutation’s predicted effects on secondarystructure. The CHASM score for each mutation observedin this study and the associated P value arelisted in table S4. Nonsense mutations, as well assmall insertions or deletions that disrupt the readingframe, are likely to disrupt function and are assigneda score of 0.001 in this table. About 36% of theevaluated mutations in MB were predicted to disruptgene function using this approach, a proportionhigher than observed in the adult tumor typesanalyzed to date (21).Finally, we evaluated the discovery screenmutational data (including both sequence and copynumber alterations) at a higher “gene set” level.There is now abundant evidence that alterationsof driver genes can be productively organizedaccording to the biochemical pathways and biologicalprocesses through which they act. The numberof gene sets that define these pathways andTable 2. Medulloblastoma CAN-genes.*GeneNumber ofmutationsNumber ofamplificationsprocesses is much less than the number of genesand can provide clarity to lists of genes identifiedthrough mutational analyses. In the current study,we used a recently described approach that scoreseach gene set at the patient rather than the genelevel and is more powerful than conventional geneorientedapproaches (21, 26). The most statisticallysignificant pathways and biologic processes highlightedby this gene-set analysis are depicted intable S7. Of these, two—the Hedgehog and Wntsignaling pathways—have been previously shownto play a critical role in MB development. In theHedgehog pathway, PTCH1 was mutated in 15 of88 (17%) tumors, and in the Wnt pathway, CTNNB1was mutated in 11 of 88 (13%) tumors (table S4).Notably, however, the pathways most highlyenriched for genetic alterations had not previouslybeen implicated in MB. These involved genes responsiblefor chromatin remodeling and transcriptionalregulation, particularly the histone-lysineN-methyltransferase MLL2. Eighteen of the 88(20%) tumors harbored a mutation in a gene withinthese pathways or in a related gene member: thehistone-lysine-N methyltransferases MLL2 (mutatedin 12 tumors) and MLL3 (3 tumors); the SWI/SNF–related matrix-associated actin-dependent regulatorof chromatin members SMARCA4 (3 tumors) andARID1A (1 tumor); and the histone lysine demethylaseKDM6B (1 tumor). The mutations in thesegenes could be clearly distinguished from passengeralterations. In MLL2, for example, 8 of the 12 mutations(67%) were predicted to truncate the encodedNumber ofdeletionsPassengerprobabilityPTCH1 22 / 88 0 / 23 0 / 23

RESEARCH ARTICLESto those observed in earlier studies. We alsoidentified amplifications of MYC and OTX2,bothpreviously implicated in MB (5–7).The ability to investigate the sequence of allcoding genes in MBs has also revealed mutatedgenes not previously implicated in MBs (table S4).Among these, MLL2 and MLL3 were of greatestinterest, because the frequency of inactivating mutationsunequivocally establishes them as MB tumorsuppressor genes. This genetic evidence is consistentwith functional studies showing that knock-out ofmurine MLL3 results in ureteral epithelial cancers(29). These genes are large and have been reportedin the Catalogue of Somatic Mutations in Cancer(COSMIC) database to be altered in occasional cancers,but not at a sufficiently high frequency to distinguishthem from passenger alterations (and withno evidence of a high fraction of inactivating mutations)(30). Interestingly, inactivating germline mutationsof MLL2 have recently been identified as acause of Kabuki syndrome, a multiple malformationdisorder without known cancer predisposition (31).The general role of genes controlling histonemethylation has become increasingly recognized asa common feature of human cancers. For example,inactivating mutations of the histone H3K27 demethylasegene UTX have been observed in multiplemyelomas, esophageal cancers, and renal cell cancers(32). In addition, a small fraction of renal cellcancers contain mutations in the histone methyltransferasegene SETD2 and the histone demethylasegene JARID1C (33), and the histonemethyltransferase gene EZH2 has been found tobe mutated in non-Hodgkin’s lymphomas (34).Most recently, frequent mutations of the chromatinremodeling gene ARID1A have been discovered inovarian clear cell carcinomas (20, 35); of note, oneARID1A mutation was discovered in our MB patients(table S4). A link between histone methylationgenes (although not MLL2 or MLL3) andMBhasalso previously been hypothesized based on theobservation that copy number alterations affectingchromosomal regions containing histone methyltransferasesor demethylases occur in a subset ofMBs (36).The mechanism(s) through which MLL genescontribute to tumorigenesis are not known, but someclues can be gleaned from the literature. The MLLfamily of histone H3K4 trimethylases includes sevengenes (MLL1, MLL2, MLL3, MLL4, MLL5, SET1A,and SET1B) (37). MLL family genes have beenshown to regulate HOX gene expression (38, 39),and an attractive possibility is that they normallydown-regulate OTX2, an MB oncogene (6, 7, 40).Another possibility is suggested by the observationthat b-catenin brings MLL complexes to the enhancersof genes regulating the Wnt pathway, therebyactivating their expression (41). A third possibilityis that MLL family genes are important for transcriptionalregulation of normal brain development anddifferentiation (42) and that their disruption maylead to aberrant proliferation of precursor cells.The identification of MLL2 and MLL3 as frequentlyinactivated MB genes supports the conceptthat MB is fundamentally characterized by dysregulationof core developmental pathways (43).Although alterations of classic cancer genes (e.g.,TP53, MYC, andPTEN) were also identified inthese childhood tumors, our sequence analysisdemonstrated that mutations of genes involved innormal developmental processes, such as MLLfamily genes and Hedgehog and Wnt pathwaygenes, were much more frequent. The fact that arelatively small number of somatic mutations issufficient for MB pathogenesis as compared toadult solid tumors provides further evidence thatthe temporally restricted subversion of normalcerebellar development is critical in the developmentof these tumors. This is consistent with theobservation that the incidence of MB decreasesmarkedly after childhood, with the tumors becomingquite rare after the age of 40 years (1). Itwill be interesting to determine whether geneticalterations in developmental pathways are a keyfeature of all childhood malignancies.The development of an improved classificationsystem for MB that could be used to guide targetedrisk-adapted therapy to patients is a primary goal ofcurrent MB research. The designation of specifichistologic subtypes of MB has proven to be of someprognostic value. For example, large-cell/anaplasticMBs, which are aggressive tumors often associatedwith MYC amplification, carry a relativelypoor prognosis (44), whereas desmoplastic MBs,which frequently have alterations of PTCH1 orother Hedgehog pathway genes (4), are moreeasily treatable. However, molecular studies haverevealed that these histologic subtypes are biologicallyheterogeneous (3); in addition, most MBsare of the classic subtype and do not have definingmolecular alterations. Our results add an additionallayer of complexity to these classifications.Although activation of the Wnt and Hedgehogpathways are generally considered to define twoMB subtypes (3), our data revealed that thesegroups overlap, because two adult MBs werefound to contain mutations of both PTCH1 andCTNNB1 (tables S2 and S4). Similarly, MLL2/MLL3mutations do not appear exclusive to any knownsubset of MBs: Mutations were identified in bothpediatric and adult MBs and were found in allhistologic subtypes (although they were most commonin large-cell/anaplastic MBs) (Table 3 andtables S9 and S10). In addition, the frequency ofMLL2/MLL3 mutationswasobservedtobesimilarin PTCH1-orCTNNB1-mutated MBs (4/24, 17%)as compared to MBs without mutations in PTCH1or CTNNB1 (10/64, 16%). Further studies of thesegenes in larger number of MBs that have beenanalyzed for pathologic subtypes will be needed toclarify the molecular classification of this tumor.We conclude that each MB is driven by a smallnumber of driver mutations, and in our cohort, theTable 3. Characteristics of medulloblastomas with mutations in MLL2-related genes.*TumorIDMLL2mutationMLL3mutationSMARCA4mutationARID1AmutationKDM6BmutationPatient age(years)MB subtypePTCH1mutationCTNNB1mutationMB104X Nonsense Nonsense 8 ClassicMB108C Missense Unknown UnknownMB115PT Missense 5 ClassicMB118PT Frameshift 9 Classic YesMB124PT Frameshift 9 Large cell/anaplasticMB126PT Missense 11 Large cell/anaplasticMB127PT Frameshift 10 UnknownMB129PT Nonsense Unknown UnknownMB130PT Frameshift Unknown Unknown YesMB135PT Frameshift Unknown UnknownMB205PT Nonsense 11 ClassicMB216PT Missense 33 Large cell/anaplasticMB231PT Missense 18 Classic YesMB245PT Frameshift 9 ClassicMB246PT Missense Missense 7 Classic YesMB249PT Nonsense 10 Classic YesMB251PT Frameshift 26 Large cell/anaplastic YesMB253PT Nonsense 32 Nodular/desmoplastic*All genes reported in the table were determined to be wild type unless otherwise indicated.43828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

gene set most highly enriched for alterations includedMLL2. However, there are several limitationsto our study. Although in a few cases wehave identified two or three bona fide cancer genesthat are mutated in individual MBs, other casesshow no mutations of any known cancer gene andonly one alteration of any gene (Fig. 1 and tableS4). Several explanations for the relative absenceof genetic alterations in occasional MBs can beoffered. First, despite the use of classic Sanger sequencing,a small fraction of the exome cannot beexamined, either because of a very high GC contentor of homology to highly related genes. Second, itis possible that mutations in the noncoding regionsof the genome could occur, and these would notbe detected. Third, copy-neutral genetic translocations,not evaluated in our study, could be present inthose tumors with very few point mutations, amplifications,or homozygous deletions. Fourth, it ispossible that low copy number gains or loss-ofheterozygosity(LOH) of specific regions containinghistone-modifying genes could mimic theintragenic mutations that we observed (36). Finally,it is possible that heritable epigenetic alterationsare responsible for initiating some MBs. The lastexplanation, involving covalent changes in chromatinproteins and DNA, is intriguing given thenew data on MLL2 in this tumor type. It shouldthus be informative to characterize the methylationstatus of histones and DNA in MBs with andwithout MLL2/MLL3 gene alterations, as well asto determine the expression changes resulting fromthese gene mutations. These data highlight the importantconnection between genetic alterations inthe cancer genome and epigenetic pathways andprovide potentially new avenues for research anddisease management in MB patients.References and Notes1. F. Giangaspero et al., in WHO Classification of theCentral Nervous System, H. O. D. N. Louis, O. D. Wiestler,W. K. Cavenee, Eds. (WHO Press, Lyon, France, 2007).2. W. R. Polkinghorn, N. J. Tarbell, Nat. Clin. Pract. Oncol.4, 295 (2007).3. P. A. Northcott et al., J. Clin. Oncol., published online7 September 2010; 10.1200/JCO.2009.27.43244. M. C. Thompson et al., J. Clin. Oncol. 24, 1924 (2006).5. S. H. Bigner et al., Cancer Res. 50, 2347 (1990).6. K. Boon, C. G. Eberhart, G. J. Riggins, Cancer Res. 65,703 (2005).7. C. Di et al., Cancer Res. 65, 919 (2005).8. R. L. Saylors 3rd et al., Cancer Res. 51, 4721 (1991).9. L. Ding et al., Nature 464, 999 (2010).10. W. Lee et al., Nature 465, 473 (2010).11. E. R. Mardis et al., J. Med. 361, 1058 (2009).12. E. D. Pleasance et al., Nature 463, 191 (2009).13. E. D. Pleasance et al., Nature 463, 184 (2010).14. S. P. Shah et al., Nature 461, 809 (2009).15. T. J. Ley et al., Nature 456, 66 (2008).16. T. Sjoblom et al., Science 314, 268 (2006).17. L. D. Wood et al., Science 318, 1108 (2007).18. S. Jones et al., Science 321, 1801 (2008).19. D. W. Parsons et al., Science 321, 1807 (2008).20. S. Jones et al., Science 330, 228 (2010).21. Materials and methods are available as supportingmaterial on Science Online.22. C. Greenman et al., Nature 446, 153 (2007).23. T. Soussi, C. Béroud, Hum. Mutat. 21, 192 (2003).24. R. J. Leary et al., Proc. Natl. Acad. Sci. U.S.A. 105, 16224(2008).25. H. Carter et al., Cancer Res. 69, 6660 (2009).26. S. M. Boca et al., Genome Biol. 11, R112 (2010).27. P. A. Northcott, J. T. Rutka, M. D. Taylor, Neurosurg. Focus28, E6 (2010).28. N. Beerenwinkel et al., PLOS Comput. Biol. 3, e225(2007).29. J. Lee et al., Proc. Natl. Acad. Sci. U.S.A. 106, 8513(2009).30. S. A. Forbes et al., Nucleic Acids Res. 38, D652 (2010).31. S. B. Ng et al., Nat. Genet. 42, 790 (2010).32. G. van Haaften et al., Nat. Genet. 41, 521 (2009).33. G. L. Dalgliesh et al., Nature 463, 360 (2010).34. R. D. Morin et al., Nat. Genet. 42, 181 (2010).35. K. C. Wiegand et al., N. Engl. J. Med. 363, 1532(2010).36. P. A. Northcott et al., Nat. Genet. 41, 465 (2009).37. M. Vermeulen, H. T. M. Timmers, Epigenomics 2, 395(2010).38. K. I. Ansari, S. S. Mandal, FEBS J. 277, 1790 (2010).39. K. Agger et al., Nature 449, 731 (2007).40. D. C. Adamson et al., Cancer Res. 70, 181 (2010).41. J.Sierra,T.Yoshida,C.A.Joazeiro,K.A.Jones,Genes Dev.20, 586 (2006).42. D. A. Lim et al., Nature 458, 529 (2009).43. R. J. Gilbertson, D. W. Ellison, Annu. Rev. Pathol. 3,341 (2008).44. C. G. Eberhart et al., Cancer 94, 552 (2002).45. We thank J. Ptak, N. Silliman, L. Dobbyn, and M. Whalenfor technical assistance with sequencing analyses, andM. Ehinger, D. Satterfield, E. Lipp, D. Lister, andM. J. Dougherty for help with sample preparation anddata collection. This project has been funded in part by theNational Cancer Institute, National Institutes of Health,under contract HHSN261200800001E. The content of thispublication does not necessarily reflect the views or policiesof the Department of Health and Human Services, nordoes mention of trade names, commercial products, ororganizations imply endorsement by the U.S. government.This work was supported by the Virginia and D. K. LudwigFundforCancerResearch,Alex’s LemonadeStandFoundation, the American Brain Tumor Association, theBrain Tumor Research Fund at Johns Hopkins, the HoglundFoundation, the Ready or Not Foundation, the Children’sBrain Tumor Foundation, the Pediatric Brain TumorFoundation Institute, the David and Barbara B. HirschhornFoundation, American Association for Cancer ResearchStand Up To Cancer Dream Team Translational CancerResearch Grant, Johns Hopkins Sommer Scholar Program,NIH grants CA121113, CA096832, CA057345, CA118822,CA135877, and GM074906-01A1/B7BSCW, NSF grantDBI 0845275, and DOD NDSEG Fellowship 32 CFR 168a.D.W.P. is a Graham Cancer Research Scholar at TexasChildren’s Cancer Center. Under licensing agreementsbetween the Johns Hopkins University and BeckmanCoulter, B.V., K.W.K., and V.E.V. are entitled to a share ofroyalties received by the university on sales of productsrelated to research described in this paper. N.P., B.V.,K.W.K., and V.E.V are cofounders of Inostics and PersonalGenome Diagnostics and are members of their ScientificAdvisory Boards. N.P., B.V., K.W.K., and V.E.V. own Inosticsand Personal Genome Diagnostics stock, which is subjectto certain restrictions under university policy. The terms ofthese arrangements are managed by Johns HopkinsUniversity in accordance with its conflict-of-interest policies.Supporting Online Materialwww.sciencemag.org/cgi/content/full/science.1198056/DC1Materials and MethodsTables S1 to S10References21 September 2010; accepted 8 December 2010Published online 16 December 2010;10.1126/science.1198056Rotational Symmetry Breakingin the Hidden-Order Phase of URu 2 Si 2R. Okazaki, 1 * T. Shibauchi, 1 † H. J. Shi, 1 Y. Haga, 2 T. D. Matsuda, 2 E. Yamamoto, 2Y. Onuki, 2,3 H. Ikeda, 1 Y. Matsuda 1A second-order phase transition is characterized by spontaneous symmetry breaking. The natureof the broken symmetry in the so-called “hidden-order” phase transition in the heavy-fermioncompound URu 2 Si 2 , at transition temperature T h = 17.5 K, has posed a long-standing mystery.We report the emergence of an in-plane anisotropy of the magnetic susceptibility below T h , whichbreaks the four-fold rotational symmetry of the tetragonal URu 2 Si 2 . Two-fold oscillations in themagnetic torque under in-plane field rotation were sensitively detected in small pure crystals.Our findings suggest that the hidden-order phase is an electronic “nematic” phase, atranslationally invariant metallic phase with spontaneous breaking of rotational symmetry.Asecond-order phase transition generallycauses a change in symmetry, such as rotational,gauge, or time-reversal symmetry.REPORTSAn order parameter can then be introduced todescribe the low-temperature ordered phase witha reduced symmetry. The heavy-fermion compoundURu 2 Si 2 undergoes a second-order phasetransition at T h = 17.5 K, which is accompaniedby large anomalies in thermodynamic and transportproperties (1–3). Because the nature of theassociated order parameter has not been elucidated,the low-temperature phase is referred toas the hidden-order phase. It is characterized byseveral remarkable features. No structural phasetransition is observed at T h . A tiny magneticmoment appears (M 0 ≈ 0.03m B ,wherem B is theBohr magneton) below T h (4), but it is far too1 Department of Physics, Kyoto University, Kyoto 606-8502,Japan. 2 Advanced Science Research Center, Japan AtomicEnergy Agency, Tokai 319-1195, Japan. 3 Graduate Schoolof Science, Osaka University, Toyonaka, Osaka 560-0043,Japan.*Present address: Department of Physics, Nagoya University,Nagoya 464-8602, Japan.†To whom correspondence should be addressed. E-mail:shibauchi@scphys.kyoto-u.ac.jpwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 439

REPORTSsmall to explain the large entropy released duringthe transition and seems to have an extrinsicorigin (5–7). An electronic excitation gap is formedon a large portion of the Fermi surface and mostof the carriers (~90%) disappear (8–10); a gap isalso formed in the incommensurate magneticexcitation spectrum (11).The nature of the hidden order cannot be determinedwithout understanding which symmetryis being broken. Several microscopic models, includingmultipole ordering (12–16), spin-densitywave formation (17–19), orbital currents (20), andhelicity order (21), have been proposed. However,despite intensive experimental and theoreticalstudies, this remains an open question.The measurement of the magnetic torque t =m 0 MV × H has a high sensitivity for detectingmagnetic anisotropy. Here V isthesamplevolume,M is the induced magnetization, H is themagnetic field, and m 0 is the permeability ofvacuum. In particular, torque measurements performedfor a range of directions of H spanningthe tetragonal ab plane in URu 2 Si 2 provide astringent test of whether the hidden-orderparameter breaks the crystal four-fold symmetry.In such a geometry, t is a periodic function ofdouble the azimuthal angle f measured from thea axis:t 2f ¼ 1 2 m 0H 2 V½(c aa − c bb )sin 2f − 2c ab cos 2fŠð1Þwhere the susceptibility tensor c ij is given byM i = ∑ j c ij H j (Fig. 1). In a system with tetragonalsymmetry, t 2f should be zero because c aa = c bband c ab = 0. Finite values of t 2f may appear if anew electronic or magnetic state emerges thatbreaks the tetragonal symmetry. In such a case,rotational symmetry breaking is revealed by c aa ≠c bb or c ab ≠ 0, depending on the orthorhombicitydirection.Figure 2A depicts the torque measured infield H, the orientation of which is varied withina plane that includes the c axis (Fig. 2A, inset).Both below and above T h , the curves are perfectlysinusoidal and can be fitted with t(T, H, q)=A 2q (T, H) sin 2q, where A is the amplitude, q isthe polar angle, and T is absolute temperature.In our carefully selected crystals (22), no hysteresisis observed, indicating no detectable ferromagneticimpurities (the hysteresis componentis less than 0.01% of the total torque). In thisgeometry, the difference Dc ca between the c axisand in-plane susceptibilities yields a two-foldoscillation term t 2q (q, T, H) with respect to qrotation,t 2q ¼ 1 2 m 0H 2 VDc ca sin 2q ð2ÞThe H-linear dependence of A 2q (H, T )/Hwith a negligible y intercept (Fig. 2B) indicatesa field-independent magnetic susceptibility—that is, a purely paramagnetic response. This alsoAURuSicabB_[110]χ aa - χ abb [010][110]χ aa + χ aba [100]Fig. 1. (A) The tetragonal crystal structure of URu 2 Si 2 .(B) U atom arrangement in the ab plane. Therelevant axes and the susceptibility components for the c aa = c bb case are also shown. (C)Schematicsofin-plane f-scan measurements. The magnetic field H was applied in the ab planewithhighalignmentprecision (within 0.02°).τ (10 -12 N⋅m)300200100τ (10 -12 N⋅m)τ 2φ (10 -12 N⋅m)τ 4φ (10 -12 N⋅m)0-100-200-3004 D20-2-4420-2-4420-2-40AHθ30McT = 6 K0 90 180 270φ (degrees)60 90 120θ (degrees)ET = 8 K0 90 180 270φ (degrees)VPiezoresistorCantileverSingle crystalF1504.2 K15 K17 K20 KT = 10 K1800 90 180 270φ (degrees)GCbA 2θ /(µ 0 H) (10 -12 N⋅m/T)χ, ∆χ (10 -3 cm 3 /mol)HφMT = 14 K0 90 180 270φ (degrees)a5040B302010 T = 4.2 K00 1 2 3µ 0 H (T)108 C6 χ, H || c4∆χ ca2 χ, H ⊥ c00 5 10 15 20Temperature (K)H4T = 18 K0 90 180 270φ (degrees)Fig. 2. Magnetic torque curves measured at |m 0 H|=4T.(A)Torquet(q) as a function of the polar angleq (lower inset) measured at several temperatures. Torque curves measured by rotating H in clockwise(dotted line) and anticlockwise (solid line) directions coincide. Upper inset illustrates the experimentalconfiguration for t(q) measurements.(B) Magnetic field dependence of the amplitude of the two-foldoscillation of the torque divided by the field, A 2q /(m 0 H), at T =4.2K.(C) Leftaxis:Temperaturedependence of the susceptibility c for H⊥c (green) and H//c (blue). Their difference Dc ca is shown inblack. Right axis: Temperature dependence of A 2q (red circles). (D to H) Upper panels show raw torquecurves t(f) as a function of the azimuthal angle f at several temperatures. Middle and lower panels showthe two-fold cos 2f and four-fold sin 4f components of the torque curves, respectively, obtained fromFourier analysis of the raw torque curves.c400300200100025A 2θ (10 -12 N⋅m)44028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

einforces the absence of ferromagnetic impurities.To check the consistency of the data, we alsomeasured the susceptibility of a different singlecrystal of URu 2 Si 2 in the same batch by meansof a superconducting quantum interference device(SQUID) magnetometer (Fig. 2C). The amplitudesof Dc ca obtained in the two types ofmeasurements coincide in both the hidden-orderand paramagnetic states.Having established the evidence of the purelyparamagnetic response in our single crystals,we now consider the in-plane torque measuredin H rotating within the ab plane. To excludetwo-fold oscillations appearing as a result of misalignment,H is precisely applied in the ab planewithin an error of less than 0.02° by controllingtwo superconducting magnets and a rotating stage(22) (Fig. 1C). Shown in Fig. 2, D to H, is thetemperature evolution of the torque t(f) at4Tin the hidden-order phase. All torque curves areperfectly reversible with respect to the fieldrotation direction; t(f) can be decomposed ast = t 2f + t 4f + t 6f + …, where t 2nf = A 2nf sin2n(f – f 0 ) is a term with 2n-fold symmetrywith n =1,2,… Inthemiddleandlowerpanelsof Fig. 2, D to H, the two-fold and four-foldcomponents obtained from the Fourier analysisare displayed.Our results show that the two-fold oscillationis distinctly observed in the hidden-orderstate, whereas it is absent in the paramagneticphase at T = 18 K slightly above T h (middlepanels of Fig. 2, D to H). Indeed, the torquecurves shown in the upper panels become asymmetricwith respect to 90° rotations below T h .We note that the observed four-fold oscillationst 4f (and higher-order terms) arise primarily fromthe nonlinear susceptibilities (23). The presenceof the two-fold oscillation, which follows thefunctional form t 2f = A 2f cos 2f, clearly demonstratesthat c ab ≠ 0, whereas c aa = c bb . Figure3A depicts the temperature dependence of|A 2f |/V, which indicates sizable in-plane anisotropy2c ab /c aa at low temperatures. The two-fold amplitudebecomes nonzero precisely at T h =17.5K(Fig. 3A, inset). This result, together with the absenceof hysteresis in the torque curves, rulesout the possibility that very tiny ferromagneticimpurities are the origin of the two-fold symmetry(22). Moreover, |A 2f | grows rapidly with decreasingtemperature in a manner quite differentfrom Dc ca (T) (Fig. 2C). This indicates that amisalignment of H from the ab plane is unlikelyto be the origin of the two-fold symmetry.On the basis of this reasoning, we conclude thatthe amplitude of two-fold oscillations is amanifestation of intrinsic in-plane anisotropy ofthe susceptibility (Fig. 1B):c½110Š ¼c aa þ c ab ≠ c½110Š¼c aa − c abð3ÞThe in-plane anisotropy that sets in preciselyat T h indicates that the rotational symmetry isbroken in the hidden-order phase. The temperaturedependence of |A 2f | º c ab near T h is fittedwith c ab º c(T h – T) +c´(T h – T) 2 much betterthan with a purely quadratic fit (Fig. 3A, inset).The leading T-linear term near T h (Fig. 3B) isnaturally expected from the standard Landautheory of second-order transition when c ab is describedbythesquaredtermofanorderparameterh º (T h – T) 1/2 . Thus, the observed in-planeanisotropy 2c ab (T)/c aa implies that the hiddenorderparameter h breaks the four-fold tetragonalsymmetry.Why has such an in-plane magnetic anisotropynot been reported before? To address this, wemeasured several samples with different sizes. Inmillimeter-sized crystals, we observed no differencebetween c½110Š(T)andc½110ŠðTÞ, but insamples with a smaller volume V, a nonzero2c ab º |A 2f |/V appeared (Fig. 3A). This mayimply that the hidden-order phase forms domainswith different preferred directions in theab plane, which may be a natural consequenceREPORTSof the tetragonal crystal structure (with possiblesmall undetected orthorhombicity in thedomain). A domain size on the order of tens ofmicrometers would explain both our results andthe difficulties in observing this effect. The factthat the torque curves remain unchanged forfield-cooling conditions at different field angles(fig. S2) implies that the formation of such domainsis predominantly determined and stronglypinned by the underlying crystal conditions, suchas internal stress or disorder.We now comment on the relevance of thepresent observation to other experimental results.The nuclear magnetic resonance (NMR) spectrahave been reported to exhibit an anomalousbroadening in the hidden-order phase for H//ab(5, 6, 24). This phenomenon has been discussedin terms of the orbital currents associated withthe breaking of time-reversal symmetry (20).The present results provide an alternative explanationto this: Finite values of c ab give rise tothe broadening of the NMR spectrum, which isestimated to be ~1 Oe at 4 T (22), roughly thesame as the reported values (6). We also notethat some of the nuclear quadrupole frequencyn Q measurements (25), which are sensitive tothe local charge density, report a small changein the slope of n Q (T) at the Ru site below T h .As for the crystal structure, no discernible latticedistortions have been reported. This may be aresult of the weak coupling between the latticeand electronic excitations relevant to the two-foldsymmetry. The symmetry of superconductivity,which appears at T c = 1.4 K and is embeddedin the hidden-order phase, should also be restrictedby the found rotational asymmetry. Recently, achiral d-wave superconducting state of the formsin k z2 c sin k x þ k ya þ i sin k x − k ya ð4Þ22has been proposed (10). The rotational symmetrybreaking implies that the states includingFig. 3. (A) Temperature dependenceof two-fold oscillation amplitudedivided by the sample volume|A 2f |/V measured for sample 7(blue circles) and sample 8 (redcircles). The normalized in-plane susceptibilityanisotropy 2c ab /c aa =(c[110] − c[110])/c[100] is evaluatedin the right axis. SQUID datafor the large crystal, sample 0, arealso shown (gray circles) in thesame scale. Inset: the data for sample8 near T h = 17.5 K are fitted toc(T h – T) +c´(T h – T) 2 (solid line)and the quadratic dependence º(T h – T) 2 (dashed line). (B) Temperaturevariation of |A 2f |/V for sample8 on a log-log scale. The black andblue dotted lines represent linearand quadratic temperature dependence,respectively.|A 2φ |/V (N/m 2 )A20015010050005|A 2φ |/V (N/m 2 )302010012c(T h - T) + c'(T h - T) 2c = 1.93 ± 0.41c' = 0.31 ± 0.0616∝ (T h - T) 2T hTemperature (K)20URu 2 Si 2 single crystals#8 (46×46×13 µm 3 )#7 (100×100×22 µm 3 )#0 (1.3×0.28×2.3 mm 3 )10864|2χ ab (T)|/χ aa (T = 20 K) (%)|A 2φ |/V (N/m 2 )10010654T h65243210 15 20012 3 4 5 6 7 8 2 3 4 5 6 7 8250.1 1Temperature (K)1-T/T h43232B∝ (1 - T/T h ) 2 ∝ (1 - T/T h )www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 441

REPORTSsin [(k x + k y )/2]a and sin [(k x – k y )/2]a havedifferent superconducting transition temperaturesT c1 and T c2 (< T c1 ), which results in a phasetransition inside the superconducting phase atT c2 . A recently reported anomaly in the lowercritical field measurements appears to supportsuch exotic superconducting states (26).Our results, together with the absence of a largemagnetic ordered moment, impose constraints ontheoretical models of the hidden order in URu 2 Si 2 .The breaking of the four-fold rotational symmetryin a tetragonal crystal structure is in general a hallmarkof electronic nematic phases (27). Electronicstates that break crystal rotational symmetry withoutshowing ordered moments have also beensuggested in various strongly correlated electronsystems, including Sr 3 Ru 2 O 7 (28), high-T c cuprates(29), and frustrated magnets (30), which are discussedin terms of stripe or nematic orders. Ourfinding of a directional electronic state in a heavyfermionmaterial is another example of such exoticorder in correlated matter.References and Notes1. T. T. M. Palstra et al., Phys. Rev. Lett. 55, 2727 (1985).2. M. B. Maple et al., Phys. Rev. Lett. 56, 185 (1986).3. A. P. Ramirez et al., Phys. Rev. Lett. 68, 2680 (1992).4. C. Broholm et al., Phys. Rev. Lett. 58, 1467 (1987).5. K. Matsuda, Y. Kohori, T. Kohara, K. Kuwahara,H. Amitsuka, Phys. Rev. Lett. 87, 087203 (2001).6. S. Takagi et al., J. Phys. Soc. Jpn. 76, 033708 (2007).7. H. Amitsuka et al., J. Magn. Magn. Mater. 310, 214 (2007).8. J. Schoenes, C. Schönenberger, J. J. M. Franse,A. A. Menovsky, Phys. Rev. B 35, 5375 (1987).9. K. Behnia et al., Phys. Rev. Lett. 94, 156405 (2005).10. Y. Kasahara et al., Phys. Rev. Lett. 99, 116402 (2007).11. C. R. Wiebe et al., Nat. Phys. 3, 96 (2007).12. P. Santini, G. Amoretti, Phys.Rev.Lett.73, 1027 (1994).13. A. Kiss, P. Fazekas, Phys. Rev. B 71, 054415 (2005).14. K. Haule, G. Kotliar, Nat. Phys. 5, 796 (2009).15. F. Cricchio, F. Bultmark, O. Grånäs, L. Nordström,Phys. Rev. Lett. 103, 107202 (2009).16. H. Harima, K. Miyake, J. Flouquet, J. Phys. Soc. Jpn. 79,033705 (2010).17. H. Ikeda, Y. Ohashi, Phys. Rev. Lett. 81, 3723 (1998).18. V. P. Mineev, M. E. Zhitomirsky, Phys. Rev. B 72, 014432(2005).19. S. Elgazzar, J. Rusz, M. Amft, P. M. Oppeneer, J. A. Mydosh,Nat. Mater. 8, 337 (2009).20. P. Chandra, P. Coleman, J. A. Mydosh, V. Tripathi, Nature417, 831 (2002).21. C. M. Varma, L. Zhu, Phys.Rev.Lett.96, 036405 (2006).22. See supporting material on Science Online.23. A. del Moral, J. I. Arnaudas, J. S. Cockaday, E. W. Lee,J. Magn. Magn. Mater. 40, 331 (1984).24. O. O. Bernal et al., Phys. Rev. Lett. 87, 196402 (2001).25. O. O. Bernal et al., Physica B 378–380, 574 (2006).26. R. Okazaki et al., J. Phys. Soc. Jpn. 79, 084705(2010).27. E. Fradkin, S. A. Kivelson, M. J. Lawler, J. P. Eisenstein,A. P. Mackenzie, Annu. Rev. Condens. Matter Phys. 1, 153(2010).28. R. A. Borzi et al., Science 315, 214 (2007);10.1126/science.1134796.29. M. Vojta, Adv. Phys. 58, 699 (2009).30. N. Shannon, T. Momoi, P. Sindzingre, Phys. Rev. Lett. 96,027213 (2006).31. We thank D. F. Agterberg, A. V. Balatsky, A. Buzdin,H. Harima, R. Ikeda, K. Ishida, Y. Kohori, T. Sakakibara,A. Schofield, H. Shishido, M. Sigrist, T. Takimoto,A. Tanaka, P. Thalmeier, and Y. Yanase for helpfuldiscussions. Supported by Grant-in-Aid for the Global COEprogram “The Next Generation of Physics, Spun fromUniversality and Emergence,” Grant-in-Aid for ScientificResearch on Innovative Areas “Heavy Electrons”(20102002, 20102006) from the Ministry of Education,Culture, Sports, Science & Technology, and Grant-in-Aidfor Scientific Research from the Japan Society for thePromotion of Science.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/439/DC1Materials and MethodsFigs. S1 to S3References3 September 2010; accepted 21 December 201010.1126/science.1197358High-Gain Backward Lasing in AirArthur Dogariu, 1 * James B. Michael, 1 Marlan O. Scully, 1,2 Richard B. Miles 1The compelling need for standoff detection of hazardous gases and vapor indicators ofexplosives has motivated the development of a remotely pumped, high-gain air laser thatproduces lasing in the backward direction and can sample the air as the beam returns. Wedemonstrate that high gain can be achieved in the near-infrared region by pumping with afocused ultraviolet laser. The pumping mechanism is simultaneous resonant two-photondissociation of molecular oxygen and resonant two-photon pumping of the atomic oxygenfragments. The high gain from the millimeter-length focal zone leads to equally strong lasingin the forward and backward directions. Further backward amplification is achieved with the useof earlier laser spark dissociation. Low-divergence backward air lasing provides possibilities forremote detection.1 Mechanical and Aerospace Engineering Department, PrincetonUniversity, Princeton, NJ 08544, USA. 2 Institute of QuantumScience and Engineering, Texas A&M University, CollegeStation, TX 77843, USA.*To whom correspondence should be addressed. E-mail:adogariu@princeton.eduOptical techniques for the remote detectionof atoms and molecules rely on theuse of lasers to selectively identify andquantify species of interest. To enable singlesideddetection, collection of light must beaccomplished in the backward direction. Collectionof incoherent light emission from moleculesof interest is limited by the nondirectional natureof spontaneous emission. More sensitive detectiontechniques, aided by the coherent nature andwell-defined direction of emission, are restrictedin the direction of emission by the phase-matchingrelation. For commonly employed nonlinear techniquessuch as coherent anti–Stokes Raman spectroscopy(1) and stimulated Raman scattering (2),phase-matching results in a coherent beam propagatingin the direction of the pumping laser,away from the source.These limitations have motivated the explorationof backward air lasing and stimulated gainconcepts, which can produce coherent scatteringthat returns to the pump-laser location (3). To date,the only approach that has shown promise is basedon the electron recombination of ionized molecularnitrogen from a femtosecond-produced filament(4, 5). This scheme leads to gain at 337 nm,the same wavelength as the molecular nitrogenlaser. Amplified spontaneous emission gain onthe order of 0.3 cm −1 has been observed (5).We demonstrate the generation of high-gainlasing in air with the use of a 100-ps remotepump laser, which simultaneously drives a twophotondissociation of molecular oxygen and atwo-photon excitation of one of the resulting oxygenatom fragments. Both processes are resonantlyenhanced at the 226-nm wavelength of the pumplaser. The excitation is followed by lasing from theexcited atomic oxygen (Fig. 1A). The pump laseris focused such that there is no laser-inducedbreakdown of the air, and excitation followed bystimulated emission is achieved throughout the1-mm-long focal region. The result is the formationof well-collimated backward and forwardpropagating laser beams at 845 nm with parameterscorresponding to the ultraviolet (UV) pumpbeamfocusing.Two-photon laser-induced fluorescence fromatomic oxygen has been developed for quantitativediagnostics of combusting gases whereatomic oxygen is an important radical species(6–10). The two-photon excitation transition isfrom the 2p 3 P ground state to the 3p 3 Pexcitedstate with 226-nm laser radiation. That excitationis followed by spontaneous relaxation fromthe 3p 3 P state to the 3s 3 S state, producing fluorescenceemission at 845 nm (Fig. 1A). Theuse of the two-photon excitation to produce stimulatedemission at 845 nm in atomic oxygen hasbeen observed in flames at subatmospheric pressures(11).The same two-photon transition can be usedas the initial step in a 2+1 resonance enhancedmultiphoton ionization (REMPI) (12). This processcan be remotely monitored by microwavescattering from the free electrons [radar REMPI(13)]. Figure 1B shows the radar REMPI excitationspectra from a flame containing oxygenatoms (squares) and from ambient air (circles)with a 100-ps laser tuned in wavelength throughthe two-photon oxygen atom transition at 226 nm.44228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSA226 nm226 nm845 nm3p 3 P226 nm226 nm2p 3 P845 nm3s 3 SO atom excitationFig. 1. (A) Two-photon dissociation of the oxygen molecule and subsequent twophotonresonant excitation of the ground-state oxygen atom fragment result inemission at 845 nm. (B) Atomic oxygen 2+1 radar REMPI signals for preexistingoxygen atoms in a flame (red squares) and produced by photodissociation in air(blue circles).BRadar Rempi signal (V)543210225.4 225.5 225.6 225.7 225.8Wavelength (nm)Intensity (a.u.)150100500AIntensity at 845nm (V)0.0750.0500.025843 844 845 846 847 -1.0 -0.5 0 0.5 1.0Wavelength (nm)Time (ns)Fig. 2. Backscattered atomic oxygen emission (A) measured spectrum and (B) pulse in time. a.u.,arbitrary units.0BThe 100-ps laser pulse is short enough to suppressavalanche breakdown (14), so multiphotonionization of atomic oxygen dominates, and themeasured radar REMPI signal reflects the densityof atomic oxygen in the focal region of thelaser. Note that in room air, even in the absenceof the flame, the radar REMPI signal peaks atthe same wavelength as the resonant atomic oxygentransition. The fact that this is not coincidentwith the resonant peak of the molecularoxygen ionization (15) indicates that the laserhas caused the dissociation of molecular oxygen,and the observed peak is due to atomic oxygen,broadened by the high velocity of the oxygenatom fragments and their collisions with surroundingmolecules (16). Similar molecular oxygendissociation and atomic oxygen excitationhave previously been observed with nanosecondpulses using two-photon laser-induced fluorescence(8, 11). The high velocity of the fragmentssuggests that they are formed by two-photon,rather than single-photon, dissociation. This observationis consistent with the measurements ofBuijsse et al.,whouseda2+1REMPItime-offlightion-imaging scheme to measure molecularoxygen photodissociation mechanisms and atomicoxygen fragment velocities (17). Their results indicatethat three dissociation pathways are presentat 226 nm: (i) a quite weak single-photon dissociativepath via the Herzberg continuum formingO( 3 P 2 )+O( 3 P j ) with 0.38-eV excess energy, (ii)a dominant two-photon dissociative path via aresonance at 225.66 nm with the predissociative3dd 3 P 0,1 (v =2,wherev is the vibrational quantumnumber) Rydberg state leading to O( 3 P 2 )+O( 1 D 2 ) with 3.91-eV excess energy, and (iii) aweak two-photon resonant pathway throughthe same Rydberg state leading to O( 3 P 2 )+O( 3 P j )with 5.88-eV excess energy. The dominantresonant two-photon pathway produces O( 3 P 2 )atoms with velocities on the order of 4800 m/s.The observation that the 100-ps laser pulseat 226 nm can simultaneously dissociate themolecular oxygen to ground-state atoms and excitethe fast-moving oxygen atoms led to therealization that lasing might be possible in roomtemperature,atmospheric-pressure air. The experimentalsetup is arranged to measure the gainand lasing properties in the backward direction,toward the pump source. The backward lasingis separated from the oppositely propagatingpump by a dichroic filter and monitored by a detector,a fast-gated charge-coupled device (CCD)camera, or a spectrometer. The 10-Hz tunable226-nm pump-laser pulses are obtained by mixingthe output of a picosecond amplified tunableTi:Sapphire laser with Nd:yttrium-aluminumgarnet(YAG) pulses and further frequency doubling.The 100-mJ pulses are focused into air witha30-cm–focal-length lens, leading to a two-photonexcitation volume of ~10-mm diameter and 1-mmlength. The estimated focused peak intensity of300 GW/cm 2 is low enough to ensure that thephotodissociation process takes place withoutmeasurable ionization or spark formation withinthe 100 ps of excitation, so the pump laser isnot scattered by plasma effects, and the gainpath length of the excited atomic oxygen reflectsthe focal depth of the pump laser.Figure 2 shows the measured spectrum andtemporal pulse shape of the backscattered emissionfrom atomic oxygen at 845 nm. The atomicoxygen spectrum shown in Fig. 2A is limited bythe 0.1-nm resolution of the spectrometer. The3-GHz oscilloscope used to measure the backscatteredpulse duration in Fig. 2B indicates anupper limit of 300 ps for the atomic oxygen emission.The fact that the backward emission pulsewidth is more than two orders of magnitudeshorter than the reported atomic oxygen fluorescencelifetime of 36 ns (11) is one indicationof stimulated emission.To establish the presence of lasing, the signaldetected along the axis of the volume is comparedto the signal detected when observing fromthe side. Because both approaches observe thesame excited volume and collect the same solidangle, if the signal is incoherent fluorescence, thetwo measurements should give similar results.On the other hand, if the signal is due to stimulatedemission, then there is exponential gainand the path length is important. In that case, theemission from the volume along the axis will besubstantially stronger than the off-axis emission.Using a photomultiplier tube, we measure a ratioof more than 8000 between the emissionwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 443

REPORTSBack emissionwith OD3 filter (V)845nm side emission (V)1.000.750.500.2500.080.04ABFig. 4. Far-field, single-pulse backscattered oxygen laser beam at 845 nm measured 30 cm away fromthe 226-nm laser focused in (A) air,(B) a methane-air flame, and (C) air with a 532-nm pre-pulse–generated spark on the far side of the gain region.collected within the same 0.78-steradian (sr) solidangle along the axis in the backscattered directionand that collected from the side (Fig. 3, A andB), indicating that the backscattered emission isstrongly gain-dominated.Furthermore, if the emission were primarilyincoherent, the scaling of the detected emissionwith the input UV pump-laser energy would beindependent of the direction of observation.Figure 3 shows that the scaling is dramaticallydifferent. The axial signal (A) scales as the 4.2power of the pump laser and has a thresholdvalue, whereas the side-view signal (B) scales asthe 2.3 power of the pump laser and has nothreshold. From experiments at lower intensitywith nanosecond laser pumping (11), we wouldexpect the two-photon fluorescence scaling tohave a power dependence of 3 or 4, dependingon the route of dissociation of O 2 (one or twophotons in addition to the two for excitation).The 2.3 power scaling shown in Fig. 3B impliesstrong saturation of either the dissociation orBackscattered (0.006 steradians) emission (V)00 20 40 60 80 100 0.1 0.2 0.5 1Pump pulse energy (µJ)100010010CSpherical (4π steradians)spontaneous emission (V)Fig. 3. The backward (A) andside(B) 845-nm emission in air as functions of the 226-nm pump-laserenergy show the scaling difference between backward lasing (4th power) and incoherent emission (2ndpower). Error bars indicate one SD. The backward coherent emission versus the total nondirectionalincoherent emission (C) shows the stimulated emission gain.excitation process. In a methane/air flame, whereatomic oxygen is naturally present, we observe asubquadratic power dependence, indicating strongsaturation of the two-photon excitation process.This suggests that in air, where the atomic oxygendensity is smaller than in the flame, the twophotonexcitation of atomic oxygen is saturated,whereas the dissociation of molecular oxygen retainsa higher-order dependence.Furthermore, in Fig. 3C, a direct comparisonbetween the backward emission and side emission(while varying the pump power) shows anonlinear dependence with the power of 2. Asthe voltage measurements of the two photomultiplierscan be directly related to the respectivesignal amplitudes, we can also estimate theoptical gain seen by the axial beam due to thestimulated emission process (Fig. 3C): Thebackscattered beam (which can be all collectedin a 0.006-sr solid angle) is about 500 timesstronger than the overall incoherent emission (obtainedby extending the 0.78-sr side-collectionefficiency over the whole 4p solid angle). Thiscorresponds to an optical-gain coefficient of62 cm −1 over the 1-mm path length.An overall signal gain in the backscattereddirection can be estimated by using the ratio ofthe laser energy to the energy of the fluorescencecollected into the same solid angle as the laserbeam (6 × 10 −3 sr). The solid angle ratio, multipliedwith the optical gain of 500 times theenergy increase of the axial beam when comparedto the total fluorescence, leads to an increaseon the order of 10 6 along the 1-mm axialpath length of the excitation volume. Similar gainis measured in the forward direction.As a final confirmation of lasing, Fig. 4 showsthe single-shot, far-field images recorded 30 cmaway in the backward direction by a nonamplifiedPrinceton Instruments (Trenton, New Jersey) CCDcamera. Figure 4A shows the image of the backwardpropagatinglasing in air. The beam is well localized,and the divergence of 40 mrad is consistentwith diffraction-limited lasing from the crosssectionalsize of the pump volume. The measuredenergy of the atomic oxygen laser pulse obtainedin air for this focusing geometry is ~20 nJ, whichprovides an overall 2 × 10 −4 efficiency for thebackward-generated air laser.The observed donut-shaped mode arises fromthe interaction of the pump laser with molecularoxygen. The presence of preexisting atomic oxygengenerated by other means leads to a muchstronger signal and produces a mode with a centralmaximum. For example, Fig. 4B shows theimage of backward lasing from naturally occurringatomic oxygen in an atmospheric-pressure, stoichiometricmethane/air flame, whereas Fig. 4Cshows the atomic oxygen lasing from air when aspark is induced a few microseconds earlier by a532-nm pulse from a frequency-doubled Nd:YAGlaser. Using this pre-pulse to create atomic oxygenenhances the coherent backscattered emission by afactor of ~50. For that enhancement to be effectivefor standoff applications, the spark must be locatedjust on the far side of the lasing volumeto minimize the distortions of the backwardpropagatinglaser beam induced by the hightemperaturespark region.The optical gain and directional emission ofthe coherent backward radiation from theremote air laser open new opportunities for thestandoff detection of atmospheric contaminantsand trace species through stimulated and nonlinearbackward scattering processes and alsoprovide a much greater range for the detectionof optical molecular and atomic features from adistant target.References and Notes1. P. J. Wrzesinski et al., J. Raman Spectrosc.; publishedonline 16 June 2010 (10.1002/jrs.2709).2. A. Owyoung, IEEE J. Quan. Elec. 14, 192 (1978).3. V. Kocharovsky et al., Proc. Natl. Acad. Sci. U.S.A. 102,7806 (2005).4. Q. Luo, A. Hosseini, W. Liu, S. L. Chin, Opt. Phot. News15, 44 (2004).5. Q. Luo, W. Liu, S. L. Chin, Appl. Phys. B 76, 337(2003).44428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

6. W. K. Bischel, B. E. Perry, D. R. Crosley, Chem. Phys. Lett.82, 85 (1981).7. M. Aldén, H. Edner, P. Grafström, S. Svanberg, Opt.Commun. 42, 244 (1982).8. J. E. M. Goldsmith, Appl. Opt. 26, 3566 (1987).9. K. C. Smyth, P. Tjossem, Proc. Combust. Inst. 23, 1829 (1991).10. J. H. Frank, T. B. Settersten, Proc. Combust. Inst. 30,1527 (2005).11. M. Aldén, U. Westblom, J. E. M. Goldsmith, Opt. Lett. 14,305 (1989).12. T. A. Cool, Appl. Opt. 23, 1559 (1984).13. R. B. Miles, Z. Zhang, S. H. Zaidi, M. N. Shneider, Am.Inst. Aeronaut. Astronaut. J. 45, 513 (2007).14. Y. P. Raizer, Laser-Induced Discharge Phenomena(Consultants Bureau, New York, 1977).15. R. J. Yokelson, R. J. Lipert, W. A. Chupka, J. Chem. Phys.97, 6153 (1992).16. A.Dogariu,J.Michael,E.Stockman,R.B.Miles,“Atomic oxygen detection using radar REMPI” inConference on Lasers and Electro-Optics/InternationalQuantum Electronics Conference, OpticalSocietyofAmerica (OSA) Technical Digest (OSA, Washington,DC, 2009), paper CFU4.17. B. Buijsse et al., J. Chem. Phys. 108, 7229 (1998).18. We gratefully acknowledge support from the U.S. Officeof Naval Research under the Science AddressingAsymmetric Explosive Threats Program.25 October 2010; accepted 16 December 201010.1126/science.1199492REPORTSMagnetic Bistability of Moleculesin Homogeneous Solution atRoom TemperatureS. Venkataramani, 1 U. Jana, 1 * M. Dommaschk, 1 F. D. Sönnichsen, 1 F. Tuczek, 2 R. Herges 1 †Magnetic bistability, as manifested in the magnetization of ferromagnetic materials or spincrossover in transition metal complexes, has essentially been restricted to either bulk materialsor to very low temperatures. We now present a molecular spin switch that is bistable atroom temperature in homogeneous solution. Irradiation of a carefully designed nickel complexwith blue-green light (500 nanometers) induces coordination of a tethered pyridine ligandand concomitant electronic rearrangement from a diamagnetic to a paramagnetic state in up to75% of the ensemble. The process is fully reversible on irradiation with violet-blue light(435 nanometers). No fatigue or degradation is observed after several thousand cycles at roomtemperature under air. Preliminary data show promise for applications in magnetic resonanceimaging.1 Otto-Diels-Institute for Organic Chemistry, University of Kiel,24118-Kiel, Germany. 2 Institute for Inorganic Chemistry, Universityof Kiel, 24118-Kiel, Germany.*Present address: Department of Chemistry, Jadavpur University,Kolkata-700032, India.†To whom correspondence should be addressed. E-mail:rherges@oc.uni-kiel.de.Bulk materials exhibit magnetic propertiesfundamentally different from those of isolatedatoms or molecules, as magnetismon the macroscopic scale is a collective phenomenoninvolving the mutual cooperation of largenumbers of microscopic magnetic moments. Aparticularly important macroscopic property, generatedby cooperative effects in the solid state, ismagnetic bistability, or hysteresis, as found in themagnetization of ferromagnetic materials and spincrossover (SCO) complexes. Unfortunately, themaximum temperature at which the magneticstates are stable decreases with decreasing numberof interacting atoms or molecules. This limitationhas generally restricted technical applicationsat room temperature to particles of 7 to 10 nmand larger.In contrast to ferro- and ferrimagnetic systems,where the magnetically interacting spinsare located in neighboring atoms, the SCO effectprovides control over the interaction ofelectron spins within a single atom (1). The phenomenonis observed in first-row transition metalcomplexes (d 4 to d 7 orbital occupancy). Dependingon the nature and field strength of the surroundingligands, the metal ion may exist in twodifferent electronic configurations, low spin andhigh spin, which respectively minimize and maximizethe number of unpaired d-electrons. If (forexample, for intermediate ligand field strengths),both configurations are of similar energy, externalperturbations such as temperature, pressure, orlight can be used to switch between the twomagnetic states (2). However, single molecules insolution undergo a gradual spin crossover followinga Boltzmann distribution. Upon reversalof the external stimulus, the system immediatelyreturns to the original state. In solids, cooperativeeffects (lattice effects) (3) may give rise to a hysteresisand thus to bistable behavior. For technicalapplications, the low-spin and high-spin statesshould be stable for a reasonable period of timewithin a large temperature range close to roomtemperature. It has been estimated that crystalswith cooperative interaction of a minimum of afew thousand molecules are needed to meet thesecriteria (4). Temperature- and light-induced excitedspin-state trapping (LIESST) in bulk crystallinematerials has recently been demonstrated at ambienttemperatures (5–8). However, these SCOsystems exhibit bistability only in the solid state.We herein demonstrate that, based on visiblelight–mediated reversible ligand coordination, theswitching of magnetic properties can be performedat room temperature in magnetically noninteractingmolecules in homogeneous solution.Hysteresis is achieved without cooperative effects,such as magnetic coupling (e.g., in ferromagneticmaterials) or lattice interactions (e.g., inspin crossover systems).Our technique is based on the idea that thespin state of a transition metal ion can be controlledby switchable ligands. In principle, thereare three approaches toward this end: varyingthe ligand field strength, switching the coordinationnumber, or switching the degree of antiferromagneticcoupling between two high-spinmetal ions (Fig. 1). The first approach, coinedlight-driven ligand-induced spin change (LD-LISC) has been investigated by multiple researchgroups using iron(II) (9–16) and iron(III) (17)complexes. A disadvantage of this concept is thatthe change in ligand field strength induced byexternal stimuli tends to be small and a completeswitching between magnetic states over a widetemperature range is difficult to achieve. In contrastto the ligand field strength, the coordinationnumber is a discrete, rather than smoothly variable,property, and the resulting magnetic statesshould be more robust to changes in the environment,such as temperature variations. We thereforepursued this approach for the design andsynthesis of a prototype system. The tuning ofantiferromagnetic coupling (Fig. 1C) is probablymore difficult to achieve in transition metalcomplexes.Nickel(II) is known to form complexes in twodifferent spin states depending on the coordinationnumber (n = 4, 5, and 6) (18). In square planarcomplexes (n =4),Ni 2+ tends to be low spin(diamagnetic, S = 0). Octahedral complexes (n =6) are nearly always high spin (paramagnetic,S = 1). Square pyramidal complexes (n =5)areeither high or low spin depending on the natureof the ligands. Controlled coordination of twoaxial ligands to a square planar Ni 2+ complex(i.e., modulation of the coordination numberfrom 4 to 6) thus leads to a spin change from low(S =0)tohigh(S = 1) spin. A simultaneous additionand removal of two ligands is difficult toFig. 1. (A to C) Three principles for ligand-drivenspin-state change.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 445

REPORTSrealize in a robust molecular framework. Wetherefore performed preliminary investigationsto find suitable organonickel systems that switchthe spin state on changing the coordination numberbetween n = 4 and 5 (19). Square planarNi-tetrakis(pentafluorophenyl)porphyrin combinedwith pyridine as the axial ligand meets the criteria.To drive the coordination/decoordination processesof the axial ligand, we took advantage ofthe photochromic properties of phenylazopyridine(PAP). As with azobenzene, ultraviolet (UV) lightirradiation (typically 365 nm) of trans-PAP inducesisomerization to the cis configuration,whereas the back reaction is induced by visiblelight (435 nm) (20). This reversible switching processis accompanied by a large geometry changethat in a well-designed molecular framework canbe used to control coordination (21–23)andspinstate [light-driven coordination-induced spin-stateswitching (LD-CISSS)]. Hereby, the PAP unitserves both as a movable arm and as the axialligand (24), thus keeping the system structurallysimple. As in a macroscopic record player, thepickup arm (PAP) with the needle (lone pair ofpyridine nitrogen) has to be connected to theturntable housing (Ni-porphyrin) in such a waythat upon operation the needle is precisely manipulatedto a well-defined position on the disk.In our case, the lone pair of the pyridine shouldbe placed exactly perpendicular to the porphyrinplane and at a distance of 2.1 Å above the Ni 2+ion. There are three possible regioisomers ofPAP (o, m, and p), two points of attachment tothe porphyrin (b or meso) and numerous tethersfor connecting the PAP arm with the porphyrin.We checked a large number of potentially suitablemolecular structures and optimized their geometryby density functional theory (DFT) calculations.The target parameter for finding the mostpromising candidate was the distance between thepyridine nitrogen and Ni 2+ ion. Ni-azoporphyrin1 (Fig. 2) was selected for synthesis. DFT calculations(Perdew-Burke-Ernzerhof/Triple Zeta ValencePolarization) (25) predicted that the 1-cisisomer would be paramagnetic (triplet groundstate) with an axial N-Ni bond of 2.086 Å, and1-trans diamagnetic (singlet ground state) with anN-Ni distance of 6.350 Å (or larger depending onthe conformation) (figs. S1 and S2).The azo-functionalized porphyrin 1 was synthesizedin a six-step sequence (26) (fig. S3).The key step was cyclization of phenyl dipyrromethane,PAP-functionalized benzaldehyde andpentafluorobenzaldehyde to form the metal-freemonosubstituted tetraphenyl porphyrin. Ni(II) wasinserted in the final step, and the resulting nickelporphyrin 1 was obtained in the thermodynamicallymore stable trans configuration. Uponirradiation of a dilute solution (~1 mM) of 1-transin acetonitrile with bluish-green light (500 nm) atroom temperature, the paramagnetic cis isomer isformed in ~65% yield (photostationary equilibrium).Thus, the wavelength inducing the trans-cisisomerization in 1 is red-shifted by about 135 nmas compared with most azobenzenes and azopyridines(~365 nm). We attribute this bathochromicshifttoanenergytransferfromthep,p* excitedstate of the porphyrin (Q band) to the p,p* state ofthe azopyridine. The isomerization is accompaniedby a striking downfield shift and broadening of thenuclear magnetic resonance (NMR) signals of theprotons in the vicinity of the paramagnetic Ni 2+center (Fig. 3, table S1, and figs. S4 to S7). Weattribute these spectral changes to a combinationof Fermi contact (contact shift) and dipolar couplings(pseudocontact shift) interactions betweenvarious nuclei in the molecule with the unpairedelectrons in the nickel (27, 28). The pyrroleprotons shift from 9.0 and 9.1 parts per million(ppm) in 1-trans to 42.3, 43.2, and 43.7 ppm in1-cis. The largest downfield shift is observed forthe o-pyridine protons, which give rise to broadsignals at 100 and 108 ppm.Because of the paramagnetism of 1-cis wewere not able to use nuclear Overhauser experiments(NOE) to obtain more detailed structuralinformation. We therefore synthesized theanalogous Zn-porphyrin from the metal-free precursor.The Zn complex exhibits an analogouscis-trans switching behavior to the Ni-porphyrin1; however, it remains in the diamagnetic spinstate. We isolated the cis and the trans configurationin pure form and unambiguously eluci-Fig. 2. Reversible light-induced magnetic switching of azopyridine functionalized Ni-porphyrin 1.Thermal conversion of the paramagnetic cis compound to the thermodynamically more stable transisomer at room temperature in the dark is very slow. The half-life of 1-cis at 54°C in acetonitrile is 27hours. The thermal stability in DMSO is even higher.Fig. 3. 1 H-NMR spectra of the trans (A) andcis(B) configurations of porphyrin 1. The pyrrol protons(13 to 20) and, in even more pronounced fashion, the pyridine protons (1 and 2) in 1-cis exhibit a verylarge downfield shift and line broadening caused by magnetic coupling with the paramagnetic nickelcenter. The latter protons resonate at 100 and 108 ppm.44628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSFig. 4. UV spectra of the trans(green) and cis (violet-blue) configurationof Ni porphyrin 1. The l maxfor each absorption band is marked,and the extinction coefficients of thel max of Soret bands of 1-cis and1-trans are given in brackets.Achieva 3.0 Tesla). A solution of nickel-porphyrin1 (3.14 mL, 1.8 mM in DMSO) was irradiatedwith bluish-green light (500 nm, 75% conversionto 1-cis), and with white light (70% conversionto 1-trans). We observed a contrast difference(due to the change in relaxation time) of 43%,which corresponds well to the change in concentrationof the paramagnetic 1-cis (45%) (fig. S22).dated the structure by quantitative NOE experimentsand DFT calculations (for details, see tables S4toS6andfigs.S16to21).In the UV-visible absorption spectrum duringisomerization of 1-trans to 1-cis, theSoretband of the square planar Ni porphyrin (l max =406 nm) and the Q bands at 524 and 557 nmdecrease in intensity, whereas new bands appearat l max = 422 and 544 nm (Fig. 4). The cis isomeris extremely light sensitive and converts backto the diamagnetic trans species by irradiationwith violet-blue light (435 nm) or by exposure todiffuse ambient light (photostationary states atdifferent wavelengths are shown in fig. S11).The unusually high photoreactivity of 1-cis mightbe due to the fact that paramagnetic nickel(II)porphyrins release their axial ligands upon irradiationinto the Soret band (29–31). The switchingefficiency is even higher in dimethyl sulfoxide(DMSO). Upon irradiation at 500 nm, a photostationaryequilibrium of 75% cis isomer is obtained.Irradiation at 435 nm switches the ensembleback to 97% of the trans form (3% cis).The light-driven switching of magnetic propertiesin 1 is fully reversible, and the compoundis completely stable under ambient conditions.Two samples of Ni porphyrin 1 (0.5 ml, ~50 mmin CD 3 CN in an NMR tube, and 3 ml, ~5 mminCH 3 CN in a UV cuvette) were alternately irradiatedwith light-emitting diodes emitting light of500 nm for 90 s (conversion to 45 and 50% 1-cis)and 435 nm for 30 s (conversion to 80 and 90%1-trans) (fig. S12). No degradation or fatigue wasdetected after more than 10,000 cycles, eventhough the experiment was performed at roomtemperature under air (fig. S12). Particularly remarkableis the thermal stability of the cis isomer.Whereas cis-azobenzenes and cis-azopyridinestypically isomerize to the more stable trans configurationwith a half-life of several hours, 1-cisis stable for weeks (25% isomerization within10 weeks at 21°C; half-life at 54°C, 27.2 hours)(figs. S13 and S14). The thermal stability of 1-cisis even higher in DMSO (

REPORTS37. F. Matino et al., Chem. Commun. (Camb.) 46,6780 (2010).38. R. B. Lauffer, Chem. Rev. 87, 901 (1987).39. L. Knutsson, F. Ståhlberg, R. Wirestam, Magn. Reson.Mater. Phys. 23, 1 (2010).40. H. Sugimoto, K. Kuramoto, S. Inoue, Perkin Trans. 1,1826 (2002).41. This work was supported by the DeutscheForschungsgemeinschaft through SFB 677. We thank theAlexander von Humboldt Foundation for a researchfellowship for S.V. We are also grateful for supportfrom R. Siewertsen and F. Temps (Institut für PhysikalischeChemie, Universität Kiel), G. Peters (Institut fürAnorganische Chemie, Universität Kiel), and O. Jansen,C. Riedel and M. Helle (Institut für Neuroradiologie,Universitätsklinikum Schleswig-Holstein). TheChristian-Albrechts-Universität zu Kiel andUniversitätsklinikum Schleswig-Holstein have filed apatent application for the compound described herein asan MRI contrast agent.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/445/DC1Materials and MethodsSOM TextFigs. S1 to S22Tables S1 to S6References2 December 2010; accepted 30 December 201010.1126/science.1201180Kinetic Isotope Effects for theReactions of Muonic Heliumand Muonium with H 2Donald G. Fleming, 1 * Donald J. Arseneau, 1 Oleksandr Sukhorukov, 1,2 Jess H. Brewer, 3Steven L. Mielke, 4 George C. Schatz, 5 Bruce C. Garrett, 6 Kirk A. Peterson, 7 Donald G. Truhlar 4 *The neutral muonic helium atom may be regarded as the heaviest isotope of the hydrogen atom, with amass of ~4.1 atomic mass units ( 4.1 H), because the negative muon almost perfectly screens one protoncharge. We report the reaction rate of 4.1 Hwith 1 H 2 to produce 4.1 H 1 H+ 1 H at 295 to 500 kelvin.The experimental rate constants are compared with the predictions of accurate quantum-mechanicaldynamics calculations carried out on an accurate Born-Huang potential energy surface and withpreviously measured rate constants of 0.11 H(where 0.11 H is shorthand for muonium). Kinetic isotopeeffects can be compared for the unprecedentedly large mass ratio of 36. The agreement with accuratequantum dynamics is quantitative at 500 kelvin, and variational transition-state theory is used tointerpret the extremely low (large inverse) kinetic isotope effects in the 10 −4 to 10 −2 range.Isotopic substitution provides a widely usedtool for mechanistic analysis of chemicalreaction rates and is a test bed for fundamentaltheories of chemical kinetics (1). A key simplifyingassumption in chemical dynamics isthe Born-Oppenheimer (BO) approximation (2),which allows the potential energy surface (PES)that governs nuclear motion to be calculated independentlyof the motion of the nuclear masses.The BO approximation facilitates the calculationand interpretation of kinetic isotope effects(KIEs), defined as the ratio of rate constants fortwo reactions differing only in isotopic masses,which provide mechanistic information even forcomplex reactions. The largest conventional KIEsare associated with substitution of 2 Hor 3 Hfor1 H. However, by using muons, one can accessKIEs corresponding to isotopic mass ratios muchgreater than even 3 (3–5), and here we present anexample with the unprecedentedly large ratio of 36.4.The lightest H atom isotope is muonium (Mu),inwhichanelectron(e) orbits a positive muon (m + )1 TRIUMF and Department of Chemistry, University of BritishColumbia, Vancouver, British Columbia, Canada V6T 1Z1.2 Department of Chemistry, University of Alberta, Edmonton,Alberta, Canada T6G 2G2. 3 Department of Physics, University ofBritish Columbia, Vancouver, British Columbia, Canada V6T 1Z1.4 Department of Chemistry and Supercomputing Institute,University of Minnesota, Minneapolis, MN 55455–0431, USA.5 Department of Chemistry, Northwestern University, Evanston, IL60208–3113, USA. 6 Chemical and Material Sciences Division,Pacific Northwest National Laboratory, Richland, WA 99352,USA. 7 Department of Chemistry, Washington State University,Pullman, WA 99164–4630, USA.*To whom correspondence should be addressed. E-mail:flem@triumf.ca (D.G.F.); truhlar@umn.edu (D.G.T.)“nucleus,” which is 206 times heavier than anelectron, so that Mu [ 0.11 H, 0.113 atomic massunits (amu)] behaves chemically like a very lightH atom (unlike positronium, for instance). Herewe label Mu as 0.11 H to emphasize its importancefor kinetics as a pseudo-isotope of H. Because ofits light mass, 0.11 H has proven to be an unusuallysensitive probe of quantum effects on nuclearmotion in reaction rates, both in terms of zero–point-energy (ZPE) shifts and in its propensity tomanifest quantum tunneling (5, 6).In contrast to Mu, the heaviest usable isotopeof H can be created by using a negative muon(m – ) to replace an electron in He to form Hem,with an atomic mass of 4.116 amu (7). Becausethe muon is much heavier than an electron, its 1sorbital is very close to the nucleus (mean radiusof 0.2 pm), effectively screening one protoncharge, so that Hem may be considered a virtualisotope of H, and so we will label it 4.1 Hasshorthand. [The decimal character of 4.1 is areminder that this is not a conventional H isotopesuch as quadrium, 4 H, or other isotopes up to 7 H(8), with lifetimes, when known, in the 0.1 to 1 zsrange.] Here we report a measurement of the rateconstant k 4.1 for reaction of this heaviest isotopeof the H atom with diprotium, and we comparethe results with theory.It has been said (9)that“Explaining the H + H 2reaction is without question the most importantproblem in chemical dynamics. . . . A successfultheoretical description is crucial because the credibilityof every other theoretical advance in thefield rests with the solution to the H + H 2problem.” For 2 H+ 1 H 2 and 1 H+ 2 H 2 , theory andexperiment converged in 2003 (10). Here wereport the application of the same theoreticalmethods to the reactions of 4.1 H and 0.11 H with1 H 2 , along with results for the new experimentson the former. Combining these with previous(11) experimental results for the rate constant k 0.11for the reaction 0.11 H+ 1 H 2 allows a comparisonof experiment and theory for this fundamentalreaction over a mass ratio of 36.4. The theoreticalresults are based on an accurate PES and accuratequantum dynamics calculations so that—in contrastto the usual situation—the theoretical resultsfor this system can be used as benchmarks to validatethe experimental approach. We also reportand compare calculations based on variationaltransition-state theory (12) (VTST); this methodis approximate, but testing it is important becauseof its applicability to complex reactions for whichaccurate quantum dynamics are impractical (1, 13).Muon beams can be produced with essentially100% longitudinal spin polarization at a nuclearFig. 1. Relaxation rates versus concentrationof H 2 at 405 K (blue datapoints and fitted line) and 500 K(red data points and fitted line) at500 bar of He.44828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

accelerator such as TRIUMF, where the presentexperiments with negative muons (m – ) were carriedout. In the decay of the muon (m − → e − n e n m ), theresulting electron, which is detected in the experiments,is emitted preferentially opposite to themuon spin direction [see the supporting onlinematerial (SOM) (14)]. In a low transverse magneticfield, a fixed counter in the plane of precessionwill detect an oscillating signal with a characteristicLarmor frequency as the muon spin sweeps past itssolid angle (15); this is the basis of the transversefieldmuon spin rotation technique (16)weused.Although the kinetic energies of muons areseveral million electron volts upon entering thegas target, most of this energy is lost to ionizationand inelastic scattering processes down toenergies of ~100 keV (16);thereisnolossinmuon spin polarization because the muon spin isunaffected by these Coulomb interaction processes.At lower energies, the m – is captured into a muonicorbit by He, which ejects both of its electrons,leaving the 4.1 H + ion, which is neutralized in a chargeexchangecollision (15) with ammonia dopant.The experimental data are in histograms ofthe time differences between the detection of anincoming muon and the detection of its electrondecay product. Superimposed on the signal is afunction that reflects the interaction of the spinpolarizedmuon with its environment. Any processthat causes a spin flip or spin dephasing causes arelaxation of the signal, with relaxation rate l;themost important process here is the pseudo–firstorderreaction of 4.1 H with H 2 . The procedurefor analyzing the data is presented elsewhere(7), and further experimental details are givenin the SOM (14). Examples of the measuredrelaxation rates are in Fig. 1.For high accuracy, the present calculationsreplace the usual BO PES with the Born-Huang(17) PES. The latter is obtained by adding the BOdiagonal correction, which depends on nuclearmasses. The BO surface (18) is a fit to essentiallycomplete configuration interaction calculations. Thediagonal correction (19) raises the BO barrier of9.60 kcal/mol to 9.73 for 4.1 Heandto9.97for 0.11 H.The quantum-dynamics calculations used theoutgoing wave variational principle (20), numericallyconverged to better than 1%. The VTSTcalculations were carried out by improved canonicalvariational theory (ICVT) (21), with least-actionground-state (LAG) multidimensional tunnelingcontributions (22), and with the bound stretchingvibration treated with the Wentzel-Kramers-Brillouin approximation (23) and bends treatedwith the centrifugal oscillator approximation (24).The rate constants, k 4.1 , determined from theslopes of plots such as those in Fig. 1, are shownin the Arrhenius plot of Fig. 2, where they arecompared to theory. The VTST results are shownboth with (ICVT/LAG) and without (ICVT)tunneling (both VTST results include ZPE); comparisonof these results shows that tunneling makesa large contribution to the rate. The accuratequantum-mechanical (QM) calculations are inexcellent agreement with the experimental data atFig. 2. Comparison of experimental and theoreticalthermal rate constants of the 4.1 H+ 1 H 2 reaction(in units of cubic centimeters per molecule per second).REPORTS500 and 405 K, where they are well withinexperimental error, but the calculated rates arebelow experimental values at 295 K by 30%, asurprisingly large disagreement. Possible systematicerrors that could be more important at thelowest temperature (where the rate is smallest)are impurities in the H 2 or additional relaxationcaused by nonreactive collisions. The ICVT/LAG calculations of k 4.1 are consistently belowboth accurate QM theory and experiment at alltemperatures, although they agree with the QMresults within 23% over the temperature rangeplotted. The largest uncertainties in the ICVT/LAGcalculations are in the tunneling contributions.The present experimental results for the reactionrate of the heaviest H-atom isotope, muonic He,with the protium molecule can now be comparedwith that for the lightest H-atom isotope, muonium,corresponding to an unprecedented factor of 36.4in atomic mass. The ratio of rate constants is givenin Fig. 3, which contains only one experimentalpoint because the 0.11 H+ 1 H 2 experiments (11)extend down only to 473 K. The usual conventionis to place the rate constant for the lighter isotopein the numerator, because that yields a normal KIEgreater than 1. The present KIE defined this way,k 0:11 =k 4:1 , however, is inverse (less than 1) andstrikingly small, with an experimental value of0.0108 and an accurate quantum value of 0.0104at 500 K. The finding that accurate quantumdynamics agrees essentially perfectly with experiment,even though the KIE is a very strongfunction of temperature (Fig. 3), argues convincinglythat 4.1 H is indeed a H isotope, and nothing(such as geometric phases or nonadiabaticbehavior) beyond a nonrelativistic treatment ofsingle–potential-energy-surface scattering on theBorn-Huang surface needs to be considered. Thisresult is a dramatic confirmation of the success ofquantum scattering calculations and the experimentsin a difficult arena. When the temperatureis lowered, the accurate quantum scattering resultbecomes 2.46 × 10 −3 at 405 K and 1.74 × 10 −4 at295 K, with the latter being (to our knowledge)the smallest KIE ever reported.Conventional transition-state theory withouttunneling, which is widely used to interpret reactionmechanisms (25), predicts a KIE of 1/19at 500 K, in contrast to 1/90 by ICVT/LAG andthe accurate value of 1/96. The good agreementof VTST calculations with accurate quantum resultsallowsthemtobeusedtoassessvibrationalcontributions at the transition state and the importanceof quantum tunneling, neither of whichis explicit in the fully quantum calculations. TheVTST calculations show that the large inverse KIEmay be attributed primarily to the difference instretching-vibration ZPE at the two transition states(the two reactions have identical reactant ZPEs).At their respective variational transition states at500 K, the 4.1 He reaction has a ZPE of 5.3 kcal/mol(2.9 in stretch and 2.4 in bend), whereas the0.11 H reaction has a ZPE of 13.8 kcal/mol (9.8in stretch and 4.0 in bend). The good agreementFig. 3. Comparison of experimental and theoreticalkinetic isotope effects (k 0.11 /k 4.1 ). The experimentalpoint was obtained using the publishedexperimental fit for the 0.11 Hreactionandtheexperimental datum at 500 K for the 4.1 Hreaction.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 449

REPORTSFig. 4. Comparison of Arrhenius activation energies(E a ≡ E Arrhenius ) calculated via QM (red solidlines), ICVT (dashed black lines), and ICVT/LAG (solidblack lines) for the reactions of 4.1 Hand 0.11 HwithH 2 as a function of temperature (T ).(within 23%) of the VTST calculations with experimentlends credence to the existence of quantizedtransition states (26, 27) with these largeamounts of vibrational energy.The tunneling transmission coefficient, whichis the ratio of the ICVT/LAG rate constant to theICVT rate constant, is 4.3, 2.1, and 1.6 at 295,405, and 500 K, respectively. Reactions of 0.11 Hmay exhibit greater tunneling effects than heavierisotopes because of the small mass (6), and thiswould be true in any one-dimensional treatmentwith an isotope-independent effective barrier. However,the LAG approximation is multidimensional,including ZPE effects (28, 29) in the effectivetunneling barriers and isotope-dependent tunnelingpaths, and the effective barrier to tunneling is muchbroader for the 0.11 H case because of the large ZPEof the 0.11 H 1 H product (30). It is very encouragingthat the results based on the LAG tunnelingtreatment, which is affordable for complex systems(31), correctly accounts for the KIE despite thequite different effective potentials for the twoisotopes, confirming the physicality of the isotopedependentbarriers.The Arrhenius (32) activation energy is definedasE a ¼ −R d ln kð1Þd(1/T)which is proportional to the negative slope of anArrhenius plot. It is well known (33, 34) thatE acan exhibit substantial temperature dependence,and the present reactions illustrate this in Fig. 4;this kind of detail cannot be revealed yet byexperiment. Figure 4 shows that E a would bequite different in magnitude and temperaturedependence without tunneling, but with tunneling,the ICVT/LAG calculation agrees very wellwith the dramatic temperature dependence andisotope dependence shown by the accurate quantumdynamicalresults. These results confirm the usefulnessof VTST for interpreting large quantum effects.References and Notes1. A Kohen, H.-H. Limbach, Eds., Isotope Effects in Chemistryand Biology (Taylor & Francis, Boca Raton, FL, 2006).2. M. Born, J. R. Oppenheimer, Ann. Phys. 84, 457 (1927).3. J. Espinosa-Garcia, Phys. Chem. Chem. Phys. 10, 1277(2008).4. D. C. Walker, J. Phys. Chem. 85, 3960 (1981).5. S. Baer et al., ACS Symp. Ser. 502, 111 (1992).6. T. Tanaka, T. Takayanagi, Chem. Phys. Lett. 496, 248(2010).7. D. J. Arseneau et al., Physica B 404, 946 (2009).8. G. Audi et al., Nucl. Phys. A. 729, 3 (2003).9. J. V. Michael, K. P. Lim, Annu. Rev. Phys. Chem. 44, 429(1993).10. S. L. Mielke et al., Phys. Rev. Lett. 91, 063201 (2003).11. I. D. Reid et al., J. Chem. Phys. 86, 5578 (1987).12. D. G. Truhlar, B. C. Garrett, Acc. Chem. Res. 13, 440 (1980).13. A. Dybala-Defratyka, P. Paneth, R. Banerjee, D. G. Truhlar,Proc. Natl. Acad. Sci. U.S.A. 104, 10774 (2007).14. Experimental details, including scheme S1, are in theSOM on Science Online.15. P. A. Souder et al., Phys. Rev. A 22, 33 (1980).16. M. Senba et al., Phys.Rev.A74, 042708 (2006).17. M. Born, K. Huang, The Dynamical Theory of CrystalLattices (Oxford Univ. Press, London, 1954).18. S. L. Mielke, B. C. Garrett, K. A. Peterson, J. Chem. Phys.116, 4142 (2002).19. S. L. Mielke, D. W. Schwenke, G. C. Schatz, B. C. Garrett,K. A. Peterson, J. Phys. Chem. A 113, 4479 (2009).20. Y. Sun et al., Phys.Rev.A41, 4857 (1990).21. B. C. Garrett, D. G. Truhlar, R. S. Grev, A. W. Magnusson,J. Phys. Chem. 84, 1730 (1980).22. B.C.Garrett,D.G.Truhlar,J. Phys. Chem. 79, 4931(1983).23. B. C. Garrett, D. G. Truhlar, J. Chem. Phys. 81, 309(1984).24. B. C. Garrett, D. G. Truhlar, J. Phys. Chem. 95, 10374(1991).25. C. F. Bernasconi, Ed., Investigation of Rates andMechanisms of Reactions (Wiley, New York, ed. 4, 1986).26. D. C. Chatfield, R. S. Friedman, D. G. Truhlar, D. W. Schwenke,Faraday Discuss. Chem. Soc. 91, 289 (1991).27. D. C. Chatfield, R. S. Friedman, D. W. Schwenke,D. G. Truhlar, J. Phys. Chem. 96, 2414 (1992).28. R. A. Marcus, J. Chem. Phys. 41, 610 (1964).29. A. Kuppermann, D. G. Truhlar, J. Am. Chem. Soc. 93,1840 (1971).30. B. C. Garrett et al., Hyperfine Interact. 32, 779 (1986).31. R.Meana-Pañeda,D.G.Truhlar,A.Fernández-Ramos,J. Chem. Theory Comput. 6, 6 (2010).32. S. Arrhenius, Z. Physikal. Chem 4, 226 (1889).33. N. C. Blais et al., J. Phys. Chem. 85, 1094 (1981).34. B. C. Garrett et al., J. Am. Chem. Soc. 108, 3515 (1986).35. We thank the Natural Sciences and Engineering ResearchCouncil of Canada, the Office of Basic Energy Sciences of theU.S. Department of Energy (DOE), and the Air Force Office ofScientific Research for their support of this work. Battelleoperates the Pacific Northwest National Laboratory for DOE.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/448/DC1MethodsScheme S1References22 October 2010; accepted 7 December 201010.1126/science.1199421Enhanced Modern Heat Transfer to theArctic by Warm Atlantic WaterRobert F. Spielhagen, 1,2 * Kirstin Werner, 2 Steffen Aagaard Sørensen, 3 Katarzyna Zamelczyk, 3Evguenia Kandiano, 2 Gereon Budeus, 4 Katrine Husum, 3 Thomas M. Marchitto, 5 Morten Hald 3The Arctic is responding more rapidly to global warming than most other areas on our planet. NorthwardflowingAtlantic Water is the major means of heat advection toward the Arctic and strongly affects the sea icedistribution. Records of its natural variability are critical for the understanding of feedback mechanisms andthe future of the Arctic climate system, but continuous historical records reach back only ~150 years. Here,we present a multidecadal-scale record of ocean temperature variations during the past 2000 years, derivedfrom marine sediments off Western Svalbard (79°N). We find that early–21st-century temperatures ofAtlantic Water entering the Arctic Ocean are unprecedented over the past 2000 years and are presumablylinked to the Arctic amplification of global warming.Rising air temperatures (1–3) andadeclineof the sea ice cover (4, 5) evidencea rapid warming in the Arctic that hasreversed a long-term cooling trend (6). Relativelywarm Atlantic Water (AW) in the FramStrait Branch (FSB) of the North Atlantic Currentis the major carrier of oceanic heat to theArctic Ocean (Fig. 1). It maintains perenniallyice-free conditions in the eastern Fram Straittoday and supplies salt to the intermediate andbottom waters of the Arctic Ocean, thereby stabilizingthe stratification (7, 8). In the easternFram Strait, AW with temperatures of 2° to 6°Cand salinities of >35.0 is found at 50- to 600-mwater depth (Fig. 2). Most of the year it is overlainby a mixed layer of lower salinity, seasonallyvariable temperatures, and ice coverage in ex-1 Academy of Sciences, Humanities, and Literature, 53151Mainz, Germany. 2 Leibniz Institute of Marine Sciences (IFM-GEOMAR), 24148 Kiel, Germany. 3 Department of Geology,University of Tromsø, 9037 Tromsø, Norway. 4 Alfred WegenerInstitute of Polar and Marine Research, 27515 Bremerhaven,Germany. 5 Department of Geological Sciences and Institute ofArctic and Alpine Research, University of Colorado, Boulder,CO 80309, USA.*E-mail: rspielhagen@ifm-geomar.de45028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSFig. 1. Bathymetric map of theFram Strait area and the easternArctic Ocean (inset; source: www.ibcao.org). Average sea ice coveragefor April [1989 to 1995;stippled line: 1963 to 1969 (31)]and September (inset; 1979 to2000; source: http://nsidc.org) isindicated by white shading. Whitearrows indicate ice flow directionin Fram Strait area. Red arrows indicateflow direction of Atlantic Water.Atlantic water flow is below haloclinewaters in the Arctic Ocean proper.Yellow spot marks station MSM5/5-712 at 78°54.94'N, 6°46.04'E, 1491-mwaterdepth.BS,BarentsSea;LS,Laptev Sea.Water depth (m)200400-2 00Atlantic Water20°WBStreme winters. These parameters also varied inthe past 2000 years, as shown by sea ice observationsand sediment core studies (5, 9).Highly variable AW advection to the Arcticin the Late Quaternary has been recorded byproxies like microfossil abundances, flux rates,and species ratios (subpolar versus polar species)in sediment cores. Advection maxima occurredmostly during relatively warm periods (interglacialsand interstadials) (10–12). The Holocene[past 12 ky (ky, thousand years)] was characterizedbyathermalmaximumat10to9kyanda cooling thereafter (13). Previous studies, however,were unable to clearly resolve variationswithin a few hundred years or less. Such variationsare well known from historical and proxydata of European climate (14) and subdivide thelast2kyintotheRomanWarmPeriod(RWP,until ~600 CE), the Dark Ages Cold Period(DACP, ~600 to ~900 CE), the Medieval ClimateAnomaly (MCA, ~900 to ~1500 CE), theLittle Ice Age (LIA, ~1500 to ~1900 CE), and theModern (Industrial) Period. Ages of boundariesbetween individual periods may vary regionally.To reconstruct the temperature variability ofAW in the FSB within the past ~2000 years, weinvestigated planktic foraminifers in a sedimentcore obtained in August 2007 from the westernSvalbard continental margin at site MSM5/5-712 (Fig. 1). This site is strategically situated inLS0° 20°EArctic OceanMSM5/5 -712FramStraitTemperature ( C)2 4 6 8Mixed LayerMain habitat depth ofplanktic foraminifers60034.65 34.75 34.85 34.95 35.05Salinity35.15SvalbardFig. 2. Water-mass structure of the upper 600 m at station MSM5/5-712. Seasonal variability of temperature(red lines) and salinity (blue lines) in the Atlantic Water (pink) is revealed by our measurementsin summer (4 August 2007; bold lines) and early winter (11 October 2006; thin lines). The main habitat ofplanktic foraminifers in the eastern Fram Strait (16, 18) is marked by a dashed pattern.80°75°Nthe path of AW inflow to the Arctic Ocean. Radiocarbondates in core MSM5/5-712-1 (table S1)revealed sedimentation rates of 18 to 20 cm ky −1for the period 120 BCE to 1475 CE and 27.7 cmky −1 thereafter, resulting in a resolution of 28 and18 years per 0.5-cm sample, respectively (Fig. 3).The core top had a modern age. Bioturbationby bottom-living organisms is expected to somewhatsmooth paleoenvironmental signals, but aquantification of the effect is difficult. Our temperaturereconstruction of AW inflow to the ArcticOcean is based on two independent methods:(i) The SIMMAX modern analog technique(15) applied on planktic foraminifer speciescounts to calculate temperatures at 50-m waterdepth and (ii) Mg/Ca measurements on the speciesNeogloboquadrina pachyderma (sinistral).Method details are given in the Supporting OnlineMaterial (SOM). Habitat and calcificationdepth of planktic foraminifers in the eastern FramStrait are below 50-m water depth and reachdown to ~300 m, but the distribution maximumis above 150 m (16). In this area, plankton bloomsusually occur in August (16). Thus, our temperaturereconstructions reflect mid-summer conditionsin the uppermost part of the AW layer,averaged over a few decades.Planktic foraminifer associations are usefulindicators of water mass distributions and oceanicfronts in the northernmost North Atlantic (17).The planktic foraminiferal data from core MSM5/5-712-1 reveal multicentennial changes underlyinga high-frequency (multidecadal) variability (Fig.3A; for details of species distributions, see fig.S1). High flux rates of both polar and subpolarspecies characterize the RWP, the MCA-LIAtransition, and the Modern Period, whereas lowfluxes are recorded for the DACP and the lateLIA (Fig. 3A). These flux variations likely reflectsea ice cover, because lowest planktic foraminiferalfluxes are today found in the ice-coveredareas of the Fram Strait, and highest fluxes atthe ice margins (18). In sediments from before~1900 CE, 10 to 40% of all planktic foraminifersbelong to subpolar species. In contrast, the youngestsediments reflecting the past ~100 years showa steep increase in subpolar foraminifer fluxes andan unprecedented inversion of the subpolar/polarspecies ratio, reaching 66% subpolar specimensin the surface sample (Fig. 3B). Because the percentageof subpolar specimens is closely relatedto the water mass distribution, with low valuesin Arctic waters and high values in AW (16–18),highest subpolar foraminifer fluxes and percentagesin samples from the Modern Period indicatea strongly increased influence of warm AW advectedfrom the Norwegian Sea.Results from SIMMAX and Mg/Ca measurementsallowed us to quantify the temperatureincrease associated with the stronger influenceof AW off Western Svalbard during the ModernPeriod. Until ~1850 CE, average summer temperaturesvaried between 2.8° and 4.4°C (mean TSE: 3.4° T 0.06°C) according to SIMMAX, withlow values during the well-known cold periodswww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 451

AW Temperature ( C)BS ice margin anomaly (km)Temperature anomaly ( C)AWCT anomaly (SD)AW Temperature ( C)% subpolar foraminifersTemperature anomaly ( C)REPORTS(early DACP, LIA) and maxima during the earlyMCA at 800 to 1100 CE (Fig. 3C). The recordshows strong similarities to terrestrial records ofNorthern Hemispheric and circum-Arctic temperaturevariability (6, 19) (Fig.3E).AWtemperaturessince 1890 CE were 4.1° to 6.0°C (mean TSE: 5.2° T 0.22°C) and thus ~2°C higher thanduring the previous 2 ky. The rapid increase toan unprecedented maximum of 6°C calculatedfor the surface sample apparently started alreadyaround 1850 CE. However, the gradual transitionmay be an artifact caused by bioturbationmixing of foraminifer-poor sediments from thelate LIA and foraminifer-rich sediments fromthe Modern Period. Temperature reconstructionsfrom Mg/Ca measurements give results very similarto those of SIMMAX (Fig. 3D). Variabilitybefore the Modern Period is high (0.7° to 5.7°C),probably because the Mg/Ca method could notprecisely reproduce temperatures below 3°C [(20);see discussion in the SOM], but the temperaturemean of 3.6°C (T0.3°C, SE) confirms theSIMMAX results. The same agreement appliesfor the Modern Period: Temperatures since 1890,as calculated from Mg/Ca data, range from 4.4to 7.1 with a mean value of 5.8°C (T0.5°C, SE).For both methods, the temperature mean of theModern Period exceeds all individual values fromthe preceding 2000 years. These results reveala rapid warming by ~2°C of uppermost AW inthe FSB in the Arctic Gateway during the past~120 years, consistent with the documented seaice retreat in the Barents Sea (5), terrestrial paleoclimatereference records (6, 19) (Fig.3,CtoE),and atmospheric measurements (3). Notably,modern summer temperatures of uppermostAW in the eastern Fram Strait are >1.5°C higherthan multidecadal mean temperature maxima(averaged by bioturbation and sampling) duringthe MCA.Our reconstructed warming of ~2°C since theLIA matches the reported temperature increaseof the Arctic Atlantic Water Layer (AAWL), obtainedfrom observational data of the past ~120years (21) (Fig. 3C). At present, there are nosubcentennial-scale open ocean proxy data seriesavailable to document the temperature evolutionof AAWL in the Arctic Ocean proper in thepreceding two millennia. Further upstream, in theeastern Norwegian Sea, high-resolution recordsof summer sea surface temperature (SSST) variabilityfrom diatoms (22) and stable isotopes ofplanktic foraminifera (23) indicate a warming of~1.5°C off western Norway since the LIA. However,eastern Norwegian Sea SSSTs in the late1990s do not clearly deviate from those occurringperiodically during the MCA (22, 23). Our findingof unequaled warm modern AW temperaturesin the eastern Fram Strait with respect to theprevious 2000 years (including the warm periodsin Roman and Medieval times) may thusexpress another facet of the Arctic amplification(1, 24) of global warming. Recent model results(25, 26) reveal the important role of sea ice andatmospheric pressure fields in the Barents Sea asSpecimens * 10 3 * cm-2 * ky-1121086420654302004000.50-0.5-1.0ABCDE3.4 C3.6 CPolar speciesSubpolar species5.2 C5.8 C-10 400 800 1200 1600 2000Year CEFig. 3. Planktic foraminiferal data and temperature reconstructions of upper Atlantic Water in the easternFram Strait over the past ~2100 years from sediment core MSM5/5-712-1. Thin lines are raw data, bold linesare three-point running means. Black triangles on the age scale mark calibrated accelerator mass spectrometry14 Cages.(A) Fluxes of polar and subpolar planktic foraminifers (100- to 250-mmfraction).(B)Percentageofsubpolar planktic foraminifers in the 100- to 250-mmfraction.(C) Summer temperatures at 50-m water depth(red) calculated by the SIMMAX Modern Analog Technique. Gray bars mark averages until 1835 CE and 1890to 2007 CE. Blue line is the normalized Atlantic Water core temperature (AWCT) record (standard deviations)from the Arctic Ocean (1895 to 2002; 6-year averages) obtained from (21). (D) Summer temperatures (purple)calculated from Mg/Ca ratios in planktic foraminifers N. pachyderma (sinistral). Gray bars mark averages until1835 CE and 1890 to 2007 CE. Blue line is the sea ice margin anomaly (11-year means, less ice is up) in theBarents Sea (BS) obtained from (5). Dashed lines mark less reliable data before 1850 CE. (E) Terrestrial Arctic[green, from (6)] and Northern Hemisphere [black, 25-year means, from (19)] temperature anomaly recordswith reference to the 980 to 1800 CE and 1961 to 1990 CE averages, respectively.706050403010-1654321045228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

a possible amplifier, which may, at least in part,be responsible for the exceptionally warm AWadvection.Instrumental air and AW temperatures in theArctic during the 20th century and beyond displayquasi-synchronous multidecadal oscillationsthat make isolation of the industrial warmingtrend difficult (3, 21). Basinwide observationssince the 1980s detected multiyear events ofAW spreading in the Arctic Ocean that featuredboth a strong warming and an increased inflowto the Arctic (7, 27, 28). Although we cannotquantify from our data the variability of previousAW inflow to the Arctic by volume, our temperaturedata series and the above observationallink suggest that the modern warm AW inflow(averaged over two to three decades) is anomalousand unique in the past 2000 years and notjust the latest in a series of natural multidecadaloscillations. Both effects—a temperature rise aswell as a volume transport increase—introduce alarger heat input into the Arctic Ocean. Althoughthere is no direct contact of the AAWL withthe ocean surface in the Arctic, such an increasedheat input has far-reaching consequences. Thestrong AW warming event in the Arctic Oceanin the 1990s caused a shoaling of the AW coreand an enhanced heat flux to the surface (29),concurrent with decreasing sea ice (4). Recentoceanographic data from the Laptev Seacontinental margin indicate the impact of warmAW-related water masses on the shallow (80-year time series. The data also provide evidencefor a significant heat flux to the overlying shelfwaters (30). Even without any modification ofthe vertical heat transfer processes, the enhancedtemperature contrast between the AW and thesurface sea water freezing point (increased from~5 to 7 K as identified here) leads to an increasein the vertical heat flux of ~40%. Any positivefeedbackmechanism will magnify the effect ofthis flux increase on the ice cover. Complementingthe strong feedback between ice and atmospherictemperatures (1), warming of the AWlayer, unprecedented in the past 2000 years, ismost likely another key element in the transitiontoward a future ice-free Arctic Ocean.References and Notes1. J. A. Screen, I. Simmonds, Nature 464, 1334 (2010).2. R. G. Graversen, T. Mauritsen, M. Tjernström, E. Källén,G. Svensson, Nature 451, 53 (2008).3. P. Chylek, C. K. Folland, G. Lesins, M. K. Dubey,M. Muyin Wang, Geophys. Res. Lett. 36, L14801 (2009).4. J. C. Comiso, C. L. Parkinson, R. Gersten, L. Stock,Geophys. Res. Lett. 35, L01703 (2008).5. D. V. Divine, C. Dick, J. Geophys. Res. 111, C01001 (2006).6. D. S. Kaufman et al., Science 325, 1236 (2009).7. B. Rudels, E. P. Jones, L. G. Anderson, G. Kattner, in ThePolar Oceans and Their Role in Shaping the GlobalEnvironment, O. M. Johannessen, R. D. Muench,J. E. Overland, Eds. (American Geophysical Union,Washington, DC, 1994), pp. 33–46.8. U. Schauer, E. Fahrbach, S. Osterhus, G. Rohardt,J. Geophys. Res. 109, C06026 (2004).9. S. Bonnet, A. de Vernal, C. Hillaire-Marcel, T. Radi,K. Husum, Mar. Micropaleontol. 74, 59 (2010).10. D. Hebbeln, T. Dokken, E. S. Andersen, M. Hald,A. Elverhøi, Nature 370, 357 (1994).11. T. M. Dokken, M. Hald, Geology 24, 599 (1996).12. R. F. Spielhagen et al., Quat. Sci. Rev. 23, 1455 (2004).13. M. Hald et al., Quat. Sci. Rev. 26, 3423 (2007).14. H. H. Lamb, Climate: Present, Past and Future(Methuen, London, 1977).15. U. Pflaumann, J. Duprat, C. Pujol, L. D. Labeyrie,Paleoceanography 11, 15 (1996).16. R. Volkmann, J. Foraminiferal Res. 30, 157 (2000).17. T. Johannessen, E. Jansen, A. Flatøy, A. C. Ravelo,in Carbon Cycling in the Glacial Ocean: Constraintson the Ocean’s Role in Global Change, R. Zahn,T. F. Pedersen, M. A. Kaminski, L. Labeyrie, Eds.(Springer-Verlag, Berlin, 1994), pp. 61–85.18. J. Carstens, D. Hebbeln, G. Wefer, Mar. Micropaleontol.29, 257 (1997).19. A. Moberg, D. M. Sonechkin, K. Holmgren, N. M. Datsenko,W. Karlén, Nature 433, 613 (2005).20. R. Kozdon, A. Eisenhauer, M. Weinelt, M. Y. Meland,D. Nürnberg, Geochem. Geophys. Geosyst. 10, Q03005(2009).21. I. V. Polyakov et al., J. Clim. 17, 4485 (2004).22. N. Koç, E. Jansen, in Climate Development and Historyof the North Atlantic Realm, G. Wefer, W. H. Berger,K.-E. Behre, E. Jansen, Eds. (Springer-Verlag, Berlin,2002), pp. 165–173.23. D. Klitgaard Kristensen, H. P. Sejrup, H. Haflidason,I. M. Berstad, G. Mikalsen, Paleoceanography 19,PA2007 (2004).24. M. Serreze, J. A. Francis, Clim. Change 76, 241 (2006).25. H. Goosse, M. M. Holland, J. Clim. 18, 3552 (2005).26. V. A. Semenov, W. Park, M. Latif, Geophys. Res. Lett. 36,L14709 (2009).27. M. J. Karcher, R. Gerdes, F. Kauker, C. Köberle,J. Geophys. Res. 108, 3034 (2003).28. I. A. Dmitrenko et al., J. Geophys. Res. 113, C05023(2008).29. M. Steele, T. Boyd, J. Geophys. Res. 103, 10419 (1998).30. I. A. Dmitrenko et al., J. Geophys. Res. 115, C08010 (2010).31. R. R. Dickson et al., J. Clim. 13, 2671 (2000).32. We thank the captains, crews, and shipboard scientificparties of research vessels Jan Mayen and Maria S.Merian for assistance during cruises JM06-WP andMSM5/5. R.F.S. and K.W. acknowledge support fromthe German Research Foundation (DeutscheForschungsgemeinschaft Priority Core Program 1266INTERDYNAMIK, project HOVAG); R.F.S. received supportfrom the Academy of Sciences, Humanities, andLiterature, Mainz; and S.A.S., K.Z., K.H., and M.H.acknowledge support from the Norwegian ResearchCouncil (projects SciencePub and WARMPAST). All dataare available at www.pangaea.de (doi:10.1594/PANGAEA.755114).Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/450/DC1MethodsFig. S1Table S1References6 September 2010; accepted 20 December 201010.1126/science.1197397REPORTSThe Southern Route “Out of Africa”:Evidence for an Early Expansionof Modern Humans into ArabiaSimon J. Armitage, 1 Sabah A. Jasim, 2 Anthony E. Marks, 3 Adrian G. Parker, 4Vitaly I. Usik, 5 Hans-Peter Uerpmann 6 *The timing of the dispersal of anatomically modern humans (AMH) out of Africa is a fundamentalquestion in human evolutionary studies. Existing data suggest a rapid coastal exodus via the IndianOcean rim around 60,000 years ago. We present evidence from Jebel Faya, United Arab Emirates,demonstrating human presence in eastern Arabia during the last interglacial. The tool kit foundat Jebel Faya has affinities to the late Middle Stone Age in northeast Africa, indicating thattechnological innovation was not necessary to facilitate migration into Arabia. Instead, we proposethat low eustatic sea level and increased rainfall during the transition between marine isotopestages 6 and 5 allowed humans to populate Arabia. This evidence implies that AMH may have beenpresent in South Asia before the Toba eruption (1).The deserts of the Arabian Peninsula havebeen thought to represent a major obstaclefor human dispersal out of Africa. AMHwere present in East Africa by about 200 thousandyears ago (ka) (2). It is likely that the firstmigration of AMH out of Africa occurred immediatelybefore or during the last interglacial[marine isotope stage (MIS) 5e] (3). During MIS6, the Afro-Asiatic arid belt was hyperarid, restrictingmovements of human populations out ofAfrica. Finds from Qafzeh and Skhul in the NearEast, dated between 119 T 18 and 81 T 13 thousandyears ago (ka) (4, 5), suggest that AMH firstmigrated along the “Nile Corridor” andintotheLevant. A later pulse from

REPORTSand Asia (6–8). An early southern dispersal fromEast Africa into south Arabia has been suggestedas an alternative route out of Africa, drawing uponfaunal and floral evidence (9), human genetics(10–12), and several Paleolithic surface occurrences(3), some of which were first described inthe 1930s (13).Here,wedescribeevidenceforthe presence of AMH by about 125 ka at JebelFaya (25.119°N 55.847°E, Fig. 1) (14). Jebel Fayais a 10-km-long, north-south–oriented limestonemountain outlier rising to ~350 m above sea level(m asl). It is located ~55 km from both the Gulf ofOman and the Persian Gulf and directly south ofthe Straits of Hormuz. At its closest, the presentcoast of Iran is 165 km away.The archaeological site, named FAY-NE1, is arock-shelter (180 m asl) at the northeastern end ofJebel Faya. Excavations at the site were undertakenbetween 2003 and 2010. A 24-m-long section,up to 5 m in depth, was cut from the backwall of the shelter outward (fig. S3). Severalother trenches were opened within and in front ofthe shelter. An area exceeding 150 m 2 has beenexcavated. The site contains archaeological levelsand artifacts that include the Iron and BronzeAges, the Neolithic, and the Paleolithic. We usedsingle-grain optically stimulated luminescence(OSL) dating (15) to determine the age of Paleolithiclayers. The Holocene levels are separatedfrom the Paleolithic layers by archaeologicallysterile sediments. We deal here with the three Paleolithicassemblages designated as A, B, and C(16) because they do not conform to any knownnamed Paleolithic industries. Assemblage A isthe uppermost and is separated from the lowerassemblages by sterile sediments. AssemblagesB and C are not always well separated but areclearly superimposed upon one another.Three OSL samples associated with assemblageC yielded ages of 127 T 16 (1 SE uncertainties),123 T 10, and 95 T 13 ka (table S8).Lithics from assemblage C (Fig. 2) exhibit a numberof different reduction strategies; blank productionwas by the Levallois, volumetric blade, andsimple parallel methods. Typologically the toolsinclude small hand axes, foliates, foliate preforms,end scrapers, sidescrapers, and denticulates (16).Assemblage B blanks, mainly flakes and fewblades, were largely produced from flat flakingsurfaces with parallel, converging, and crossed removals.Tools include sidescrapers, end scrapers,denticulates, retouched pieces, burins, and perforators.Unlike assemblage C, there was no evidencefor the Levallois technique, although there weresome small volumetric blade cores. The numberand range of tools suggests that the occupationwas multipurpose.Assemblage A contains mainly flakes struckfrom multiple platform cores with parallel removalson each face. Blades are rare. Tools includeburins, retouched pieces, end scrapers, sidescrapers,and denticulates. Both assemblages A and B lackLevallois flaking and bifacial reduction, and,typologically, the use of backing was unknown.Assemblage A is overlain by ~40 cm of sterilesand. Two OSL samples from within assemblageA yielded ages of 38.6 T 3.1 and 40.2 T 3.0 ka,and two samples from the overlying sterile layeryielded ages of 38.6 T 3.2 and 34.1 T 2.8 ka.Scattered early Holocene lithics, including Fasadpoints, were found immediately above the sterilesand layer. Two marine shells found in associationwith Fasad points yielded radiocarbon agesof 10,405 to 9711 and 10,380 to 10,078 calendaryears before the present (cal. yr B.P.), indicatingearly Holocene human occupation of the southeasternArabian Peninsula (17).Assemblage C ages are consistent with thoseof the Levantine Middle Palaeolithic, but theassemblage is neither technologically nor typologicallyrelated to that of the Levant. NotablyFig. 1. The location of Jebel Faya, United Arab Emirates, along with key sites mentioned in the text. Thedashed line represents the –120-m paleoshoreline, indicating the maximum exposure of land duringmarine lowstands.Fig. 2. Assemblage C: 1,bifacial foliate; 2, Levalloisflake; 3, bifacial preform;4, radial core; and 5, handax preform.45428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSABCDEFGHδ 18 O (‰)δ 18 Ο (‰)-9-8-7-6-5-4-3-2Age (ka)different is the reduction by façonnage for theproduction of small hand axes and foliate toolforms. Technological patterns at FAY-NE1 showgreater similarities with East and northeast Africa(18) than with other sites known in Arabia. On thebasis of these affinities and the contemporaneouspresence of AMH in East and northeastAfrica, we suggest that assemblage C occupationis attributed to AMH expanding out of Africaduring early MIS 5. An autochthonous developmentfor assemblages A and B is suggested becausethey bear no affinities with Middle StoneAge/Late Stone Age assemblages known fromEast Africa, the Upper Palaeolithic from theLevant, or the Zagros. Assemblage A occupationmay have been terminated by the onset of hyperaridconditions, as indicated by the sterile sandlayer. Certainly, there is no evidence for humanpresence at Jebel Faya between 38 and 10 ka.This hiatus coincides with a period of hyperaridityand the emplacement of major aeoliandune networks across the Arabian Peninsula (19).0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160-2.5-2-1.5-1-0.500.511.522.50 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160Age (ka)Fig. 3. Compilation of the Jebel Faya chronology with regional and global proxy records. Vertical graybars represent periods of humidity (19). (A) OSL ages from Jebel Faya. Solid squares represent ages fromwithin occupation phases C and A, whereas open squares represent ages from the sterile sand layersuperimposed on top of assemblage A (this study). Error bars indicate 1 SE. (B) Phases of aeolian duneaccumulation in the Wahiba Sands, Oman (28). (C) TimingofwetphasesinOmanfromHotiandQunfCave speleothem records (29, 30). (D) Phasesoflacustrinedevelopment,Mudawwara,Jordan(31). (E)Indian Ocean Monsoon Index (32). (F) Oxygen isotope (d 18 O) record from Soreq Cave (33). ‰ indicatesper mil. (G) Red Sea eustatic changes (20). (H) SPECMAP stacked oxygen isotope record (34).240230220210200190180200-20-40-60-80-100-120Monsoon index (W/m 2 )sea-level(m asl)Within the Arabian section of the southernroute, the Nejd Plateau, along with the Straits ofBab al-Mandab and Hormuz, are the three majorobstacles to human expansions. The Nejd Plateauis flat and today has poor vegetation cover andlacks surface water. Contrasting climatic conditionsare required to allow human migration throughthese bottlenecks. The straits were easiest to crossduring periods of low sea level, associated withglacial epochs, during which the interior of Arabiaand the Nejd Plateau was hyperarid. Conversely,during interglacials, when sea level was high, theNejd Plateau had increased vegetation densityand water was more available. Consequently, thephasing of changes in eustatic sea level and terrestrialmoisture availability is the key to understandingthe timing of human expansion alongthe southern route.We suggest that the initial expansion of AMHfrom East Africa to southern Arabia occurred atthe transition from MIS 6 to MIS 5e, when thewidth of the Bab al-Mandab was at a minimum(20, 21). Once humans had crossed into southernArabia, this migrant population would have experienceddecreased predation and competitionfor resources. Southern Arabia may have becomea secondary center for human population growth.During pluvial episodes, southern Arabia wouldthus have facilitated further hunter-gatherer rangeexpansions onto the peninsula. There is evidencefor wetter phases in southern Arabia at 135 to120ka(MIS6-5e)and82to78ka(MIS5a)(19)(Fig. 3). It is likely that populations expanded andmoved through the interior of Arabia, as well asvia the coastline, and used adaptive strategies incorporatingterrestrial resources. The presence ofhumans at Jebel Faya early in MIS 5 indicatesthat a significant range expansion occurred duringMIS 5e (22). The African affinity of assemblageC implies that a wetter southern Arabian climaterather than technological innovation was responsiblefor this range expansion (9). With more aridconditions during MIS 5d and 5b, this populationprobably became disconnected from that in southArabia, as the corridor across the Nejd Plateauwas lost. The continued deposition of Paleolithicartifacts at FAY-NE1, which have no known affinitiesto other lithic industries in the surroundingareas (16), suggests that AMH presence continuedin southeast Arabia throughout MIS 5.Because the Persian Gulf averages just 40 min depth (23), lowered eustatic sea levels duringMIS 5d and 5b, and between 75 ka and 14 ka,would have provided a land connection betweensoutheastern Arabia and Iran (24). The formerTigris-Euphrates river system flowed through theexposed Persian Gulf and into the Indian Ocean.This landscape periodically formed a corridor foranimals including Mesopotamian fallow deer,aurox, and water buffalo (25), as well as Indianelephants, which existed in northwestern Mesopotamiauntil early historic times (26).Consequently, the extended Persian Gulf regionis likely to have formed another populationcenter from which early modern humans couldwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 455

REPORTSradiate during favorable times (22). This probablyincludes the reoccupation of FAY-NE1 duringMIS 3 by the population that produced assemblageA. Access from southeast Arabia to thePersian Gulf and vice versa is likely to have beenvia the numerous wadi channels that extend fromthe Hajar Mountains and into the Persian Gulfbasin, passing Jebel Faya to the north and thesouth. These channels would also have facilitatedhuman migration by providing access to freshwater along the shores of the proto-Gulf (22, 27).References and Notes1. M. D. Petraglia et al., Science 317, 114 (2007).2. I. McDougall, F. H. Brown, J. G. Fleagle, Nature 433,733 (2005).3. M. D. Petraglia, J. I. Rose, Eds. Evolution of HumanPopulations in Arabia: Paleoenvironments, Prehistoryand Genetics (Springer, Dordrecht, Netherlands, 2009),pp. 1–14.4. R. Derricourt, J. World Prehist. 19, 119 (2005).5. J. J. Shea, Quat. Sci. Rev. 27, 2253 (2008).6. R. Klein, Evol. Anthropol. 9, 17 (2000).7. P. Mellars, Proc. Natl. Acad. Sci. U.S.A. 103, 9381 (2006).8. C. B. Stringer, Nature 405, 24 (2000).9. D. E. Wildman et al., Mol. Phylogenet. Evol. 32, 287(2004).10. L. O. Quintana-Murci et al., Nat. Genet. 23, 437 (1999).11. T. Kivisild et al., Am. J. Hum. Genet. 75, 752 (2004).12. P. Forster, S. Matsumura, Science 308, 965 (2005).13. H. Field, Am. J. Archaeol. 36, 426 (1932).14. Materials and methods are available as supportingmaterial on Science Online.15. Z. Jacobs, R. R. Roberts, Evol. Anthropol. 16, 210(2007).16. A. Marks, in Evolution of Human Populations in Arabia:Paleoenvironments, Prehistory and Genetics,M. D. Petraglia, J. I. Rose, Eds. (Springer, Dordrecht,Netherlands, 2009), pp. 295–308.17. H.-P. Uerpmann, D. T. Potts, M. Uerpmann, in Evolutionof Human Populations in Arabia: Paleoenvironments,Prehistory and Genetics, M. D. Petraglia, J. I. Rose,Eds. (Springer, Dordrecht, Netherlands, 2009),pp. 205–214.18. S. McBrearty, World Archaeol. 19, 388 (1988).19. A. G. Parker, in Evolution of Human Populations inArabia: Paleoenvironments, Prehistory and Genetics,M. D. Petraglia, J. I. Rose, Eds. (Springer, Dordrecht,Netherlands, 2009), pp. 38–49.20. G. N. Bailey et al., Coast. Arch 2, 127 (2007).21. M. Siddall et al., Nature 423, 853 (2003).22. J. I. Rose, Proc. Semin. Arabian Stud. 37, 219 (2007).23. M. Sarnthein, Mar. Geol. 12, 245 (1972).24. K. Lambeck, Earth Planet. Sci. Lett. 142, 43 (1996).25. H.-P. Uerpmann, The Ancient Distribution of UngulateMammals in the Middle East - Fauna and ArchaeologicalSites in Southwest Asia and Northeast Africa, Beiheftezum Tübinger Atlas des Vorderen Orients, Reihe A(Naturwissenschaften), Band 27 (Ludwig Reichert Verlag,Wiesbaden, Germany, 1987).26. C. Becker, Munibe 57, 445 (2005).27. H. Faure et al., Global Planet. Change 33, 47 (2002).28. F. Preusser, C. R. Geosci. 341, 621 (2009).29. S. J. Burns, D. Fleitmann, A. Matter, U. Neff, A. Mangini,Geology 29, 623 (2001).30. D. Fleitmann, A. Matter, C. R. Geosci. 341, 633 (2009).31. N. Petit-Maire et al., Global Planet. Change 72, 368(2010).32. D. C. Leuschner, F. Sirocko, Palaeogeogr. Palaeoclimatol.Palaeoecol. 197, 83 (2003).33. M. Bar-Matthews, A. Ayalon, M. Gilmour, A. Matthews,C. J. Hawkesworth, Geochim. Cosmochim. Acta 67,3181 (2003).34. D. G. Martinson et al., Quat. Res. 27, 1 (1987).35. Funding information: We thank the government ofSharjah for funding the excavations and elements ofthe lab work at Tübingen. Funding was also providedby the Role of Culture in Early Expansions of Humansproject (Heidelberg Academy of Sciences), HumboldtFoundation, Oxford Brookes University, and the GermanScience Foundation (Deutsche Forschungsgemeinschaft).Author responsibilities: S.J.A., OSL dating; S.A.J.,excavation codirector; A.E.M. and V.I.U., lithic analysis;A.G.P., paleoenvironments; H.-P.U., codirector ofexcavations, director of laboratory analyses at TübingenUniversity. Special thanks are due to the third codirector,M. Uerpmann, for handling the logistics of the project.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/453/DC1Materials and MethodsSOM TextFigs. S1 to S10Tables S1 to S8References and Notes15 October 2010; accepted 28 December 201010.1126/science.1199113Phosphorylation of ULK1 (hATG1) byAMP-Activated Protein Kinase ConnectsEnergy Sensing to MitophagyDaniel F. Egan, 1 David B. Shackelford, 1 Maria M. Mihaylova, 1,2 Sara Gelino, 4Rebecca A. Kohnz, 1 William Mair, 1 Debbie S. Vasquez, 1 Aashish Joshi, 5 Dana M. Gwinn, 1Rebecca Taylor, 1 John M. Asara, 6 James Fitzpatrick, 3 Andrew Dillin, 1,2 Benoit Viollet, 7Mondira Kundu, 5 Malene Hansen, 4 Reuben J. Shaw 1,2 *Adenosine monophosphate–activated protein kinase (AMPK) is a conserved sensor of intracellularenergy activated in response to low nutrient availability and environmental stress. In a screenfor conserved substrates of AMPK, we identified ULK1 and ULK2, mammalian orthologs of theyeast protein kinase Atg1, which is required for autophagy. Genetic analysis of AMPK or ULK1in mammalian liver and Caenorhabditis elegans revealed a requirement for these kinases inautophagy. In mammals, loss of AMPK or ULK1 resulted in aberrant accumulation of the autophagyadaptor p62 and defective mitophagy. Reconstitution of ULK1-deficient cells with a mutantULK1 that cannot be phosphorylated by AMPK revealed that such phosphorylation is requiredfor mitochondrial homeostasis and cell survival during starvation. These findings uncover aconserved biochemical mechanism coupling nutrient status with autophagy and cell survival.Ahighly conserved sensor of cellular nutrientstatus found in all eukaryotes isthe adenosine monophosphate (AMP)–activated protein kinase (AMPK). In response todecreases in intracellular ATP, AMPK is activatedand serves as a metabolic checkpoint, restoringATP levels through acute regulation of metabolicenzymes and inhibition of pro-growth anabolicpathways (1). Inactivation of LKB1, the upstreamkinase necessary for activation of AMPK underlow-energy conditions, is a frequent event in severalforms of human cancer (2). In addition, LKB1signaling is required in the liver for the therapeuticeffect of metformin, the most prevalent type 2diabetes drug worldwide, and LKB1 inactivationin mouse liver results in a type 2 diabetes–likemetabolic disease (3). Thus the LKB1-AMPKpathway provides a direct link between tumorsuppression and control of cellular and organismalmetabolism.Similar to AMPK activation, the cellular processof autophagy is initiated under nutrient-poorand low-energy conditions as a survival mechanismto ensure availability of critical metabolicintermediates and to eliminate damaged organelles,including mitochondria (4). Autophagy isthought to be initiated under nutrient-limited conditionsby a conserved kinase complex containingthe serine-threonine kinase Atg1 and its associatedsubunits, Atg13 and Atg17 (5). In mammals,this complex is encoded by two Atg1 homologs,ULK1 and ULK2, and the subunits Atg13 andFIP200, which signal to downstream autophagyregulators through still poorly understood mechanisms.In yeast and mammalian cells, Atg1 orULK1 activity is suppressed under nutrient-richconditions by the TOR (target of rapamycin) complex1 (TORC1) (6). However, biochemical eventsthat activate Atg1 or ULK1 have not yet beenidentified.We used a two-part screen to identify substratesof AMPK that mediate its effects on cell1 Molecular and Cell Biology Laboratory, Dulbecco Center forCancer Research, Salk Institute for Biological Studies, La Jolla,CA 92037, USA.2 Howard Hughes Medical Institute, SalkInstitute for Biological Studies, La Jolla, CA 92037, USA. 3 WaittAdvanced Biophotonics Center, Salk Institute for BiologicalStudies, La Jolla, CA 92037, USA. 4 Graduate School of BiomedicalSciences, Del E. Webb Neuroscience, Aging and Stem CellResearch Center, Sanford-Burnham Medical Research Institute,LaJolla,CA92037,USA. 5 Department of Pathology, St. JudeChildren’s Research Hospital, Memphis, TN 38105, USA.6 Division of Signal Transduction, Beth Israel Deaconess MedicalCenter, and Department of Medicine, Harvard Medical School,Boston, MA 02115, USA.7 Inserm, U1016, Institut Cochin,Université Paris Descartes, CNRS (UMR 8104), Paris, France.*To whom correspondence should be addressed. E-mail:shaw@salk.edu45628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

growth and metabolism (7). First, we used an optimalAMPK substrate motif (8)tosearcheukaryoticdatabases for proteins containing conservedcandidate target sites. Many in vivo substrates ofAMPK not only conform to this motif but alsobind to the phospho-binding protein 14-3-3 induciblyupon phosphorylation by AMPK. Wetherefore screened for proteins that bound torecombinant 14-3-3 in wild-type (WT) but notAMPK-deficient cells, and only under conditionsof energy stress, when AMPK would be active.One protein we identified that contained multipleconserved candidate AMPK phosphorylation sitesandassociatedwith14-3-3inanAMPK-dependentmanner was the mammalian Atg1 homolog ULK1(Fig. 1, A and B). ULK1 contains four sites (Ser 467 ,REPORTSSer 555 ,Thr 574 ,andSer 637 ) matching the optimalAMPK substrate motif, all of which are conservedin higher eukaryotes. Two of the sites areconserved back to Caenorhabditis elegans (Ser 555and Ser 574 ) and in the mammalian family memberULK2, though not the more distant familymembers ULK3 and ULK4, which unlike ULK1and ULK2 are not thought to function in autophagy.Indeed, endogenous AMPK subunitsco-immunoprecipitated with ULK1 and ULK2but not ULK3 (fig. S1), and AMPK subunits werefound in unbiased identifications of proteins coimmunoprecipitatingwith overexpressed ULK2(fig. S2), which is consistent with recent proteomicanalyses (9). To examine ULK1 in vivo phosphorylationsites, we used tandem mass spectrometryon epitope-tagged ULK1 isolated fromcells treated with or without the mitochondrialcomplex I inhibitor phenformin (10). We detectedpeptides spanning three of the four candidateAMPKsitesinULK1(Ser 555 ,Thr 574 ,andSer 637 ),and all three were phosphorylated only afterphenformin treatment (figs. S3 and S4). To examinewhether ULK1 could serve as a directsubstrate for AMPK in vitro, we created a kinaseinactiveallele (K46I) (11) to remove its autophosphorylation.AMPK phosphorylated ULK1to a greater extent than it did an established substrate,Raptor (Fig. 1C and fig. S5), which mayreflect the presence of at least four potentialAMPK sites in ULK1, as compared with Raptor,which has two reported AMPK sites (8). We gen-Fig. 1. ULK1 is a conservedsubstrate of AMPK.AB(A) Clustalalignmentoffour conserved sites inULK1 and two sites inULK2 matching the optimalAMPK substrate motif.(B) ULK1 and glutathioneS-transferase (GST) or GST-HEK293T14-3-3 expression vectorswere transfected into humanembryonic kidney (HEK)C293T cells, and placed inmedia containing 20 mMSTO-609 (STO), vehicle(veh), or 5 mM phenformin(Phen) for 1 hour. Cell lysatesand GST pulldownswere immunoblotted as indicated.In vitroKinaseAssay(C) Invitrokinaseassays with myc-taggedcatalytically inactive (KI:K46I) ULK1 or myc-taggedWT raptor that were immunoprecipitatedfromHEK293TcellsandusedasDFsubstrates for purified activeAMPK in the presence ofP-(C)-ATP. (D) HEK293Tcells transfected with myctaggedWT ULK1 or indicatedserine-to-alanineULK1 mutants were treatedwith either vehicle or 1 mMphenformin for 1 hour orwere cotransfected with aconstitutively active AMPKa1HEK293T(aa1-312) mammalian expressionvector (12). Proteinsfrom lysates were immunoblottedEwith phospho-specificantibodies as indicated. (E)In vitro kinase assays usingmyc-ULK1 and purified AMPKas above. PhosphorylationIn vitroKinaseAssayof myc-ULK1 detectedthrough immunoblottingwith indicated phosphospecificantibodies. (F)PrimaryMEFs were treated with 2 mM AICAR or vehicle for 1 hour. Lysates were immunoblotted as indicated, including detection of endogenous ULK1 P-Ser 555 .www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 457

REPORTSFig. 2. Genetic deficiency for AMPK or ULK1 inmouse liver or primary mouse hepatocytes results inautophagy defects. (A) Liver lysates from littermatematchedmice were immunoblotted for the indicatedantibodies. p62-to-actin ratio calculated from densitometryperformed on immunoblots. Data shownas mean T SEM. *P < 0.01. (B) Primary hepatocytesderived from ULK1 +/+ or ULK1 −/− mice or AMPKa1 +/− a2 L/+or AMPKa1 −/− a2 L/L as described in (7) wereplacedinmedia containing 2 mM metformin (met) or vehicle(veh) for 2 hours. Lysates were immunoblotted with theindicated antibodies. (C)TEMwasperformedonprimarymouse hepatocytes of the indicated genotypes, revealingaccumulation of mitochondria in both AMPK- and ULK1-deficient cells. Red, mitochondria; blue, cytoplasm;green, nuclei; yellow, lipid droplets. (D) Primary mousehepatocytes of the indicated genotypes stained bymeans of immunocytochemistry for the mitochondrialmarker TOM20 (red) and nuclei (blue). Scale bar, 10 mm.ACAMPK+/+AMPK-/-Murine liver lysatesp62 /actin*+/+ -/-BDULK1+/+ULK1-/-Primary HepatocytesPrimary HepatocytesFig. 3. AMPK is necessary and sufficient for autophagyinduction in C. elegans. (A) Insulin receptordaf-2(e1370) mutant worms expressing GFP::LGG-1 (equivalentto GFP-LC3) were treated with control RNAi orRNAi against bec-1 (Beclin), aak-2 (AMPKa2), or unc-51 (ULK1), and the number of LGG-1/LC3–positivepuncta per hypodermal seam cell were quantified. (B)aak-2(ok524) mutants or WT N2 (WT) animalsexpressing GFP::LGG-1 were treated with control ordaf-2 RNAiandscoredforLGG-1positivepunctaperseam cell. (C) Transgenic worms expressing constitutivelyactive AAK-2 (12) (amino acids 1-321)::TOMATO (CA-AAK-2::TOMATO) fusion or controls (WT) were analyzedfor LGG-1–positive puncta per seam cell. (D) Animalsexpressing both CA-AAK-2 (1-321) ::TOMATO and GFP::LGG-1 were treated with control or unc-51 RNAi andscored for LGG-1/LC3–positive puncta per seam cell. Allpanels show relative counts; see fig. S10 for details. Datashown as mean T SEM. *P < 0.0001.AC* * **BD** **45828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

erated phosphospecific antibodies against Ser 467and Ser 555 of ULK1. Phosphorylation of bothsites was induced by means of phenformin treatmentor expression of ULK1 with a constitutivelyactive AMPKa1 allele (12) in the absence ofREPORTSenergy stress (Fig. 1D). Purified AMPK also inducedphosphorylation at these sites in an in vitrokinase assay, which is consistent with their directphosphorylation (Fig. 1E). Using AMPK- andULK1-deficient primary mouse embryonic fibroblasts(MEFs) or matched control WT MEFs, weobserved phosphorylation of endogenous ULK1on Ser 555 in an AMPK-dependent manner aftertreatment of cells with the AMP-mimetic aminoimidazolecarboxamide ribonucleotide (AICAR)Fig. 4. AMPK phosphorylationof ULK1 is requiredfor mitophagy and cellsurvival upon nutrient deprivation.(A) U2OS cellsstably expressing mouseWT, KI, or AMPK 4SAULK1 cDNA or the emptyretroviral vector (v) alongwith a shRNA against endogenoushuman ULK1andULK2wereplacedinmedia containing 5 mMphenformin (Phen) or vehiclefor 1 hour. Lysateswere immunoblotted asindicated. (B) ULK1 −/−MEFs stably expressingWT, KI, or 4SA ULK1 cDNAor the empty retroviralvector (v) along with ashRNA against endogenousULK2 were placed inEarle’s balanced salts solutionstarvation media (starv)or control media (ctl) for6 hours in the presence orabsence of BafilomycinA(BafA) and immunoblottedas indicated. (C) Cellsfrom (B) analyzed withTEM and Inform morphometricsoftware. Red,mitochondria; blue, cytoplasm;green, nuclei. (D)Flourescence-activatedcell sorting (FACS) analysison cells from (B), whichwere stained with JC-1under basal conditions orwith the mitochondrial uncouplercarbonyl cyanidem-chloro phenyl hydrazone(CCCP) as a control in orderto measure mitochondrialmembrane potential. Compromisedmitochondrialmembrane potential isshifted to the left, as observedin cells treated withCCCP. (E) WTMEFstransfectedwith 20 nM siRNApools to a universal control(ctl), mouse Atg5, ormouse ULK1 and ULK2 for72 hours were then placedin starvation medium (starv)or standard media (ctl) forABEFULK1-/- MEFs with stable ULK2 shRNA reconstituted with:12 hours, and cell death was scored by means of AnnexinV-FACS. (F) Cells from (B) were placed in starvation medium (starv) or standard media (ctl) for 12 hours, andcell death was scored by means of AnnexinV-FACS. (G) Model for AMPK activation of ULK1 in a two-pronged mechanism via direct phosphorylation of ULK1 andinhibition of mTORC1 suppression of ULK1.CWTULK1KIULK14SAULK1GDULK1-/- MEFs with stable ULK2 shRNA reconstituted with:www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 459

REPORTS(Fig. 1F). The phosphorylation of ULK1 in thesecells paralleled that of two bona fide AMPK substrates,acetyl-CoA carboxylase (ACC) and Raptor(Fig. 1F and fig. S6).We examined the phenotypic consequencesof AMPK or ULK1 deficiency on markers ofautophagy in mouse liver and primary hepatocytes.Immunoblot and immunohistochemicalanalysis of AMPK-deficient livers (13) showedaccumulation of the p62 protein (Fig. 2A and fig.S7), whose selective degradation by autophagyhas established it as a widely used marker of thisprocess (14). p62 contains a ubiquitin-associated(UBA) ubiquitin-binding domain, which mediatesbinding to ubiquitinated cargo targeted forautophagy-mediated degradation (14). Consistentwith this function, p62 aggregates colocalizedwith ubiquitin aggregates in AMPK-deficient livers(fig. S7). p62 is recruited to mitochondria targetedfor mitophagy and is involved in mitochondrialaggregation and clearance (15, 16). ULK1-deficientmice exhibit accumulation of defective mitochondriain mature red blood cells, which are normallydevoid of mitochondria (17). Given the aberrantaccumulation of p62 in the absence of AMPK inmouse liver and the fact that rodent hepatocytesundergo substantial mitophagy upon culturing (18),we examined whether AMPK or ULK1 deficiencyin primary hepatocytes might exhibit mitochondrialdefects. Protein levels of p62 and the mitochondrialmarker protein CoxIV were similarlyelevated in lysates from AMPK- or ULK1-deficienthepatocytes cells but not from WTcontrols (Fig. 2Band fig. S8). Increased phosphorylation of endogenousULK1 Ser 555 was observed in WT but notAMPK-deficient hepatocytes after AMPK activationby metformin treatment (Fig. 2B). Furtheranalysis of the ULK1 and AMPK hepatocytesby use of transmission electron microscopy (TEM)revealed elevated levels of abnormal mitochondria,which was analyzed quantitatively with morphometricsoftware (Fig. 2C, right). Similar tofindings in other autophagy-mutant hepatocytes(19), the number of mitochondria per cell wassignificantly increased in AMPK- and ULK1-deficient hepatocytes as compared with that ofWT controls (fig. S9), which is also seen withimmunocytochemical staining for the mitochondrialmembrane protein TOM20 (Fig. 2D).Given the conservation of AMPK sites inULK1, we examined whether these two proteinsfunction together to play conserved roles in autophagyin the nematode C. elegans. In a reporterassay based on the C. elegans LC3 homologLGG-1 (20, 21), we observed that loss of insulinsignaling through genetic mutation [daf-2(e1370)] orRNAinterference(RNAi)againstthe insulin receptor daf-2 resulted in increasednumbers of green fluorescent protein (GFP)::LGG-1–positive foci in hypodermal seam cells,which is indicative of increased autophagy andconsistent with the established role for insulinsignaling in the suppression of autophagy in C.elegans (20–24). daf-2 mutant worms treatedwith RNAi to aak-2 or unc-51, theAMPKandULK1 orthologs, respectively, resulted in adecrease in abundance of LGG-1–containingpuncta (Fig. 3A). daf-2 RNAifailedtoincreasethe number of LGG-1–positive foci in AMPKdeficientworms (Fig. 3B). These data indicatethat both AMPK and ULK1 have critical roles inautophagy induced by means of reduced insulinsignaling in C. elegans. Transgenic worms expressingconstitutively active AMPK exhibitedan approximately threefold increase in the numberofLGG-1–positivefoci in seam cells as comparedwith the number of foci in controls (Fig.3C). The number of LGG-1–positive foci wassignificantly reduced when these animals werefed unc-51 RNAi (Fig. 3D) (all data can be foundin fig. S10). These observations indicate thatAMPK activation is sufficient to induce autophagyinworms,andULK1isrequiredforthisinduction.To test whether AMPK phosphorylation ofULK1 is required for ULK1 function, we stablyintroduced WT, catalytically inactive (KI), or theAMPK nonphosphorylatable (4SA) ULK1 cDNAinto human osteosarcoma U2OS cells in whichwe subsequently reduced endogenous ULK1 andULK2 with lentiviral short hairpin RNAs (shRNAs)against each (fig. S11). U2OS cells stably expressingULK1 and ULK2 shRNA exhibited increasedamounts of p62 indicative of defectiveautophagy as compared with that of parentalU2OS cells infected with an empty lentiviral vector(Fig. 4A; compare lane 1 and 2). Stable retroviralreconstitution of a myc-tagged WT ULK1cDNA, but not the 4SA or KI mutant, restoredp62 degradation (Fig. 4A, lanes 3 to 5, and fig.S12). Furthermore, we reconstituted ULK1 −/−MEFs that were also knocked down for endogenousULK2 (fig. S13) with WT, KI, or 4SAULK1 cDNAs and examined the extent of autophagyafter placement of these cell lines into starvationmedia. MEFs deficient for ULK1 and ULK2 containedelevated levels of p62 upon starvation.Cells reconstituted with WT ULK1 had reducedp62 levels, unlike the KI- or 4SA-expressing cells,which behaved like the ULK-deficient state (Fig.4B and fig. S14). To test whether the 4SA mutantexhibited effects on mitochondrial homeostasis,we used TEM and mitochondrial-selective dyeson the WT, KI, and 4SA ULK1 stably reconstitutedULK-deficient MEFs. TEM and Mitotracker Redstaining revealed that the KI- and 4SA-ULK1–expressing cells had altered mitochondrial homeostasisas compared with that of WT ULK1cells, denoted by increases in the overall numberand aberrant morphology of mitochondria (Fig.4C and figs. S18 and S19). The altered cristae andaberrant morphology of the mitochondria in theKI- and 4SA-ULK1–reconstituted cells was enhancedupon starvation (fig. S19). To test whetherthese mitochondria were functionally impaired,we analyzed the mitochondrial membrane potentialwith the activity-dependent JC-1 dye, whichrevealed defects in KI- and 4SA-reconstitutedMEFs (Fig. 4D).A hallmark of cells defective for autophagy isa predisposition to undergo apoptosis after stressstimuli that normally would activate autophagyto promote cell survival (25). We examined howULK1/2 deficiency would compare with loss ofcentral downstream autophagic regulator such asAtg5 in terms of requirement for cell survivalafter starvation. WT MEFs were treated with control,Atg5, or combined ULK1 and ULK2 smallinterfering RNA (siRNA) and analyzed for effectson cell viability after being placed into starvationconditions. Simultaneous depletion of ULK1 andULK2 mirrored the magnitude and kinetics of celldeath observed with Atg5 loss upon starvation(Fig. 4E and fig. S20). We next investigated whethermutation of the AMPK sites in ULK1 mightalso mimic ULK1/2 loss of function in this cellsurvival assay. ULK-deficient MEFs reconstitutedwith WT, but not KI or 4SA ULK1, restoredcell survival upon starvation (Fig. 4F).ULK1-deficient cells expressing the KI or 4SAmutant ULK1 showed rates of cell death similarto those of WT MEFs treated with Ulk1 andUlk2 siRNA. Thus, loss of the AMPK sites inULK1 mimics complete loss of ULK1 and ULK2in control of cell survival after nutrient deprivation.Our findings reveal a direct connection betweenenergy sensing and core conserved autophagyproteins. In mammals, phosphorylation of ULK1by AMPK is required for ULK1 function in theresponse to nutrient deprivation. Because AMPKsuppresses mammalian TOR (mTOR) activity andmTOR inhibits ULK1 (26–30), AMPK controlsULK1 via a two-pronged mechanism, ensuringactivation only under the appropriate cellularconditions (Fig. 4G and fig. S21). There are anumber of physiological and pathological contextsinwhichthispathwayislikelytoplayacritical role (31). Beyond the conserved natureof these signaling events and the role of someautophagy genes as tumor suppressors (25, 32),AMPK is defective in a variety of human cancersbearing inactivating mutations in its upstreamkinase LKB1. Thus, ULK1 may have a centralrole in the beneficial effects of the LKB1/AMPKpathway on tumor suppression or in treatment ofmetabolic disease, as observed here with metforminstimulation of ULK1 phosphorylation inliver and the profound defect in autophagy inAMPK-deficient livers. ULK1-dependent effectson mitochondrial homeostasis and cell survivalmay represent additional beneficial effects ofmetformin and other AMPK activators in overallorganismal health and life span (33).References and Notes1. D. G. Hardie, Nat. Rev. Mol. Cell Biol. 8, 774 (2007).2. D. B. Shackelford, R. J. Shaw, Nat. Rev. Cancer 9, 563 (2009).3. R. J. Shaw et al., Science 310, 1642 (2005).4. C. He, D. J. Klionsky, Annu. Rev. Genet. 43, 67 (2009).5. N. Mizushima, Curr. Opin. Cell Biol. 22, 132 (2010).6. E. Y. Chan, Sci. Signal. 2, pe51 (2009).7. Materials and methods are available as supportingmaterial on Science Online.8. D. M. Gwinn et al., Mol. Cell 30, 214 (2008).9. C. Behrends, M. E. Sowa, S. P. Gygi, J. W. Harper, Nature466, 68 (2010).10. D. G. Hardie, Gastroenterology 131, 973 (2006).11. E. Y. Chan, A. Longatti, N. C. McKnight, S. A. Tooze,Mol. Cell. Biol. 29, 157 (2009).46028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

12. B. E. Crute, K. Seefeld, J. Gamble, B. E. Kemp,L. A. Witters, J. Biol. Chem. 273, 35347 (1998).13. F. Andreelli et al., Endocrinology 147, 2432 (2006).14. V. Kirkin, D. G. McEwan, I. Novak, I. Dikic, Mol. Cell 34,259 (2009).15. S. Geisler et al., Nat. Cell Biol. 12, 119 (2010).16. D. P. Narenda, L. A. Kane, D. N. Hauser, I. M. Fearnley,R. J. Youle, Autophagy 6, 1 (2010).17. M. Kundu et al., Blood 112, 1493 (2008).18. S. Rodriguez-Enriquez, Y. Kai, E. Maldonado, R. T. Currin,J. J. Lemasters, Autophagy 5, 1099 (2009).19. M. Martinez-Vicente et al., Nat. Neurosci. 13, 567 (2010).20. A. Melendez et al., Science 301, 1387 (2003).21. C. Kang, Y. J. You, L. Avery, Genes Dev. 21, 2161 (2007).22. M. Hansen et al., PLoS Genet. 4, e24 (2008).23. K. Jia et al., Proc. Natl. Acad. Sci. U.S.A. 106, 17534(2009).24. E. S. Hars et al., Autophagy 3, 93 (2007).25. B. Levine, G. Kroemer, Cell 132, 27 (2008).26. C. H. Jung et al., Mol. Biol. Cell 20, 1992 (2009).27. Y. Y. Chang, T. P. Neufeld, Mol. Biol. Cell 20, 2004(2009).28. N. Hosokawa et al., Mol. Biol. Cell 20, 1981 (2009).29. Y. Kamada et al., Mol. Cell. Biol. 30, 1049 (2010).30. I. G. Ganley et al., J. Biol. Chem. 284, 12297 (2009).31. D. Nakada, T. L. Saunders, S. J. Morrison, Nature 468,653 (2010).32. R. Mathew, V. Karantza-Wadsworth, E. White, Nat. Rev.Cancer 7, 961 (2007).33. S. Fogarty, D. G. Hardie, Biochim. Biophys. Acta 1804,581 (2010).34. We thank L. Gerken for mouse colony assistance,F. Esterman for confocal assistance, C. Chu for assistancein scoring LGG-1 assays, M. Wood at The ScrippsResearch Institute EM facility, X. Yang in the Beth IsraelDeaconess Medical Center mass spectometry facility,R. Chitta from St. Jude’s proteomic core, and K. Lamiafor comments on the manuscript. D.F.E., D.M.G., andM.M.M. were supported through the T32 CMG traininggrant to the UCSD-Salk Biological Sciences GraduateProgram. D.B.S. was funded by T32 CA009370 to theSalk Institute Center for Cancer Research. R.J.S. is fundedby the NIH grants R01 DK080425 and 1P01CA120964,National Cancer Institute grant P30CA014195, anAmerican Cancer Society Research Scholar Award, theAmerican Diabetes Association Junior Faculty Award1-08-JF-47, and a Howard Hughes Medical Institute EarlyCareer Scientist Award. J.M.A. is supported by grants5P30CA006516-43 and 1P01CA120964-01A. M.K. isfunded by grant NHBLI K08, Burroughs Welcome Fund,and the American Lebanese Syrian Association. B.V. issupported by Agence Nationale de la Recherche (ANR).M.H. is an Ellison Medical Foundation New Scholar inAging. We also thank the Leona M. and Harry B.Helmsley Charitable Trust for their generous support.Several of the authors (D.F.E., D.B.S., M.M.M, and R.J.S.)have filed a patent related to this work (U.S. patentapplication 61/325361).Supporting Online Materialwww.sciencemag.org/cgi/content/full/science.1196371/DC1Materials and MethodsFigs. S1 to S21References11 August 2010; accepted 8 December 2010Published online 23 December 2010;10.1126/science.1196371REPORTSEffects of Experimental SeaweedDeposition on Lizard and AntPredation in an Island Food WebJonah Piovia-Scott,* David A. Spiller, Thomas W. SchoenerThe effect of environmental change on ecosystems is mediated by species interactions. Environmentalchange may remove or add species and shift life-history events, altering which species interact at agiven time. However, environmental change may also reconfigure multispecies interactions when bothspecies composition and phenology remain intact. In a Caribbean island system, a major manifestationof environmental change is seaweed deposition, which has been linked to eutrophication, overfishing,and hurricanes. Here, we show in a whole-island field experiment that without seaweed twopredators—lizards and ants—had a substantially greater-than-additive effect on herbivory. Whenseaweed was added to mimic deposition by hurricanes, no interactive predator effect occurred. Thusenvironmental change can substantially restructure food-web interactions, complicating efforts topredict anthropogenic changes in ecosystem processes.Global environmental change is expectedto have a profound impact on the structureand function of ecological communitiesby changing interactions between theircomponent species. Range shifts, extinctions, andspecies introductions change community composition,deleting some interactions and adding others(1, 2). In addition, changes in the seasonal timingof migration and other life-history events can producephenological mismatches, which can affect communitieseven when species composition is unchanged(1–3). The absence of alterations in speciescomposition and phenology, however, does notnecessarily mean that ecosystems will remainunaltered (4): Environmental change can alsoinfluence ecosystem processes by reconfiguringinteractions in communities whose species listsremain intact (5–9).Section of Evolution and Ecology and Center for PopulationBiology, One Shields Avenue, University of California Davis,Davis, CA 95616–8755, USA.*To whom correspondence should be addressed. E-mail:jpioviascott@ucdavis.eduAn important aspect of global environmentalchange is the mobilization and transport of resourcesbetween ecosystems. Seaweed deposition in particularis likely to become a more common featurein shoreline ecosystems as anthropogenic effects(such as overfishing and eutrophication) facilitate ashift toward algae-dominated marine ecosystems(10). Furthermore, intense storms, which are associatedwith the deposition of large amounts of seaweed(11, 12), have increased in frequency—atrendthatisexpectedtocontinuewithincreasingglobal warming (13). Such pulsed inputs of externalresources can increase prey availability, “subsidizing”in situ predators (14–17) and altering predatoreffects on lower trophic levels (12, 16, 18–21).We used a whole-island field experiment inthe Bahamas to probe how major seasonal pulsesof seaweed deposition—mimicking what occursin an active storm year (22)—influence theeffects of multiple predators on herbivory in aterrestrial food web. Twelve small islands (one isshown in Fig. 1A), six with lizards and sixnaturally lizard-free, were used in the experiment.Seaweed was added or removed from islands in acrossed design; each combination of seaweedand lizard presence or absence was representedby three islands. Seaweed was manipulated inOctober and December of 2008, coinciding withthat season when large storms are most likely tocause natural deposition events. On seaweedadditionislands, 0.4 to 1.4 kg/m 2 of seaweed wasdistributed patchily throughout each island,mimicking what occurs after a large storm (22).The removal treatment maintained seaweed at alevel near zero, which is consistent with naturallevels of about half of the islands before manipulation.Ant exclusions were established on branchesof four (three in one case) Conocarpus erectusplants on each island; ants were excluded with asticky resin, which lizards were able to bypass bycrossing a narrow gap between wire mesh cones(Fig. 1B). Each of the four predator treatments inthe experiment—(i) ants and lizards absent; (ii)ants present and lizards absent; (iii) ants absent andlizards present; and (iv) ants and lizards present—was represented by 12 branches in the absence ofseaweedsubsidyand11or12branchesinthepresence of seaweed subsidy.Anolis sagrei was the only lizard species onthe lizards-present islands. Previous experimentsin this ecosystem demonstrated that A. sagrei reducedherbivory on C. erectus (23, 24), one of the mostcommon plants in shoreline habitats. Extrafloralnectaries on C. erectus foliage attract mutualisticants, which can also decrease herbivory (25). Thepossibility that these two predators interactivelyaffect lower trophic levels (26) had not beenexplored in previous studies.Figure 1C shows a working model of the mainfood web components in our system. Seaweeddeposits support an abundance of detritivores, whichattract both ants and lizards (12). The most commonherbivorous arthropods include Coleoptera,Lepidoptera, and Hemiptera (12). Previous studiesin this system indicate that a shift in predatorforaging behavior after two closely spaced pulsesof seaweed is associated with increased herbivoryon C. erectus (12), suggesting that the subsidieswww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 461

REPORTSinterfere with herbivore suppression by predators(27).The present study shows that seaweed subsidiesdiminish the effect of predators on C. erectusby eliminating the greater-than-additive effect ofants and lizards on herbivory found in the absenceof seaweed subsidies. Specifically, without seaweedthe combined effect of ants and lizards onherbivory was more than three times greater thanexpected on the basis of the sum of their individualFig. 1. Study system. (A)Oneofthetwelveexperimentalislands.(B) An ant-exclusion treatment on oneof the experimental plants. The wire mesh cones on either side allow lizards to pass over the sticky resinbarrier. (C) Working hypothesis of food-web relationships on the small islands used in this experiment.Red text denotes the components that were manipulated in this study.Fig. 2. Effects of seaweed,lizards, and ants on herbivoryin C. erectus. (A) Leafdamage on C. erectus infour predator treatmentson seaweed-removal andseaweed-addition islands.Data are mean T SE; n =11 branches for the seaweedaddition–lizardstreatments and 12branchesforallothertreatments.The interactive effectof lizards and ants onC. erectus leaf damagediffered between seaweedtreatments [mixed modelanalysis of covariance(ANCOVA); seaweed*lizard*ant interaction:F 1,39 =6.0,P =0.02].Onseaweed-removal islands,there was a greater-thanadditiveeffect of ants andlizards on herbivory(mixed model ANCOVA;lizard*ant interaction onseaweed-removal islands:F 1,20 = 11.5, P = 0.003), but there was no such interactive effect on seaweed-addition islands (mixedmodel ANCOVA; lizard*ant interaction on seaweed-addition islands: F 1,19 =0.16,P =0.7).(B) Predatoreffects on leaf damage. Effect sizes were calculated as ln (leaf damage with predators/leaf damage withoutpredators) by using least-squares means from the mixed model ANCOVA described in (22).effects (Fig. 2, left). In other words, the differencebetween the treatment with both predators presentand the no-predators treatment was threetimes greater than the difference between theants/no-lizards treatment and the no-predatorstreatment plus the difference between the lizards/no-ants treatment and the no-predators treatment.In contrast, in the presence of seaweed subsidiesthere was no interactive effect of ants and lizardson herbivory (Fig. 2, right). Ant abundance increasedon islands with seaweed subsidies—the(least-squares) mean ant abundance on seaweedadditionislands was higher than that on seaweedremovalislands—but was not affected by lizards(Fig. 3A). In contrast to ants, lizards did not increasein overall abundance on subsidized islands(Fig. 3B).We suggest that the synergistic effect of lizardsandantsonherbivoryintheabsenceofseaweedsubsidies derives from the fact that these two predatorsare active at different times of day. The onlylizard species present, A. sagrei,isdiurnal.Incontrast,the dominant ant species (comprising 88% ofall pan-trapped ants), Camponotus tortuganus,isnocturnal. This temporal partitioning of foragingactivity may create a dilemma for herbivores:They can avoid A. sagrei by feeding at night, andthey can avoid C. tortuganus by feeding duringthe day, but they cannot simultaneously avoidboth types of predators. This hypothesis is analo-Fig. 3. Effects of seaweed on ant and lizardabundance. (A) Ant abundance in pan traps onexperimental islands. Ant abundance was higher onseaweed-addition islands than seaweed-removalislands (ANCOVA; F 1,7 =7.7,P =0.03);therewasno difference between lizard and no-lizard islands[analysis of variance (ANOVA); F 1,7 = 0.15, P =0.71]. (B) Lizard abundance on experimental islands.There was no effect of seaweed on lizard density(ANOVA; F 1,4 =0.38,P = 0.60). Data are mean T SEof log-transformed abundances; n = 3 islands foreach treatment combination.46228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSgous to what has been observed when predatorsexhibit spatial complementarity in foraging: Preyseeking to escape one predator enter habitats inwhich they are more vulnerable to the other predator(26). In addition to eliminating prey refuges,temporal partitioning may reduce the frequencyof antagonistic interactions between predators,which probably explains why lizards did not affectant abundance in our study.When seaweed subsidies were added to thesystem, the greater-than-additive effect of antsand lizards on herbivory was not present. We suggestthat the seaweed caused both ants and lizardsto spend more time on the ground foraging fordetritivores associated with seaweed deposits, reducingtheir combined impact on herbivory. This hypothesisis consistent with the results of an earlierstudy, in which seaweed was added or removedfrom shoreline plots on large islands (12). Despitehigher abundances of both ants and lizards inseaweed-addition plots, C. erectus sustained higherlevels of herbivory. An analysis of carbon-stableisotopesignatures indicated that lizard diets containedmore marine-derived prey in seaweed-addition plots,and ants were observed foraging for detritivores inthe seaweed deposits, suggesting that shifts inpredator foraging behavior were responsible for theobserved increases in herbivory (12). In the currentexperiment, the increase in ant density in pan trapsmay have been caused by a shift in foraging patterns,although an increase in the overall abundanceof ants resulting from increased food supplymay have also been contributory. Although lizarddensity increased in response to seaweed subsidiesin mainland plots (12), we observed no such increasein the current study. We suggest that theabsence of a numerical response in lizards on smallislands was caused by (i) the inability of lizards toimmigrate to the experimental sites from surroundingareas (as they could in the previous, mainlandplotexperiment) and (ii) the lack of reproductiveactivity during much of the study period (12). Becauseemigration and reproductive lags influencethe timing of predator responses to subsidy, weexpect the long-term impact of pulsed seaweedsubsidies on predator effects to depend on thefrequency of pulses and degree of habitat isolation(15, 16). Seaweed did not appear to cause a reductionin the effects of either ants or lizards by themselves(Fig. 2B), suggesting that the interactiveeffect of these two predators is more sensitive tosubsidy than their individual effects.Predicting the effects of environmental changeon ecosystems is an important challenge. There isincreasing recognition that species interactions stronglyinfluence how environmental change affectsecosystem processes, complicating efforts to makereliable forecasts (28, 29). Our results show thatlarge seaweed-deposition events affect the structureand function of an ecological community byreconfiguring the effects of multiple predators onlower trophic levels. This suggests that predictionsthat are based on single-species responses or pairwiseinteractions may not adequately represent communityresponses to environmental perturbations.Experiments such as the one we report here, conductedat a spatial scale large enough to capturecommunity-wide dynamics, are particularly relevantfor conservation and management decisions in theface of ever-increasing anthropogenic disturbances.References and Notes1. C. Parmesan, Annu. Rev. Ecol. Evol. Syst. 37, 637 (2006).2. G. R. Walther et al., Nature 416, 389 (2002).3. L. H. Yang, V. H. W. Rudolf, Ecol. Lett. 13, 1 (2010).4. G. R. Walther, Philos. Trans. R. Soc. B 365, 2019 (2010).5. B. T. Barton, A. P. Beckerman, O. J. Schmitz, Ecology 90,2346 (2009).6. B. T. Barton, O. J. Schmitz, Ecol. Lett. 12, 1317 (2009).Metagenomic Discovery ofBiomass-Degrading Genes andGenomes from Cow RumenMatthias Hess, 1,2 * Alexander Sczyrba, 1,2 * Rob Egan, 1,2 Tae-Wan Kim, 3 Harshal Chokhawala, 3Gary Schroth, 4 Shujun Luo, 4 Douglas S. Clark, 3,5 Feng Chen, 1,2 Tao Zhang, 1,2Roderick I. Mackie, 6 Len A. Pennacchio, 1,2 Susannah G. Tringe, 1,2 Axel Visel, 1,2 Tanja Woyke, 1,2Zhong Wang, 1,2 Edward M. Rubin 1,2 †The paucity of enzymes that efficiently deconstruct plant polysaccharides represents a major bottleneck forindustrial-scale conversion of cellulosic biomass into biofuels. Cow rumen microbes specialize in degradationof cellulosic plant material, but most members of this complex community resist cultivation. To characterizebiomass-degrading genes and genomes, we sequenced and analyzed 268 gigabases of metagenomic DNAfrom microbes adherent to plant fiber incubated in cow rumen. From these data, we identified 27,755 putativecarbohydrate-active genes and expressed 90 candidate proteins, of which 57% were enzymatically activeagainst cellulosic substrates. We also assembled 15 uncultured microbial genomes, which were validated bycomplementary methods including single-cell genome sequencing. These data sets provide a substantiallyexpanded catalog of genes and genomes participating in the deconstruction of cellulosic biomass.Biofuels derived from lignocellulosic plantmaterial represent an important renewableenergy alternative to transportation fossilfuels (1, 2). A major obstacle to industrial-scaleproduction of fuel from lignocellulose lies in theinefficient deconstruction of plant material, owing7. J.P.Harmon,N.A.Moran,A.R.Ives,Science 323, 1347 (2009).8. E. Post, R. O. Peterson, N. C. Stenseth, B. E. McLaren,Nature 401, 905 (1999).9. C. C. Wilmers, W. M. Getz, PLoS Biol. 3, 571 (2005).10. J. B. C. Jackson, Proc. Natl. Acad. Sci. U.S.A. 105 (suppl. 1),11458 (2008).11. H. L. Blomquist, J. H. Pyron, Am. J. Bot. 30, 28 (1943).12. D. A. Spiller et al., Ecology 91, 1424 (2010).13. M. A. Bender et al., Science 327, 454 (2010).14. L. B. Marczak, R. M. Thompson, J. S. Richardson, Ecology88, 140 (2007).15. L. H. Yang et al., Ecol. Monogr. 80, 125 (2010).16. R. D. Holt, Ecology 89, 671 (2008).17. G. A. Polis, W. B. Anderson, R. D. Holt, Annu. Rev. Ecol. Syst.28, 289 (1997).18. C. V. Baxter, K. D. Fausch, M. Murakami, P. L. Chapman,Ecology 85, 2656 (2004).19. S. J. Leroux, M. Loreau, Ecol. Lett. 11, 1147 (2008).20. S. Nakano, H. Miyasaka, N. Kuhara, Ecology 80, 2435 (1999).21. J. L. Sabo, M. E. Power, Ecology 83, 1860 (2002).22. Materials and methods are available as supportingmaterial on Science Online.23. D. A. Spiller, T. W. Schoener, Ecology 75, 182 (1994).24. T. W. Schoener, D. A. Spiller, Am. Nat. 153, 347 (1999).25. J. Piovia-Scott, J. Ecol. 99, 327 (2011).26. A. Sih, G. Englund, D. Wooster, Trends Ecol. Evol. 13, 350 (1998).27. G.Takimoto,T.Iwata,M.Murakami,Am. Nat. 173, 200 (2009).28. K.B.Suttle,M.A.Thomsen,M.E.Power,Science 315,640 (2007).29. S. E. Gilman, M. C. Urban, J. Tewksbury, G. W. Gilchrist,R. D. Holt, Trends Ecol. Evol. 25, 325 (2010).30. WethankL.H.Yang,J.J.Stachowicz,M.L.Stanton,andthreeanonymous reviewers for comments on the manuscript;A. N. Wright and S. S. Porter for help in the field; L. Wongfor help in the lab; and the Bahamas Ministry of Agricultureand Marine Resources for permission to conduct thisresearch. This project was supported by grants from NSF andthe University of California Davis Center for PopulationBiology to the authors.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/461/DC1Materials and MethodsSOM TextFigs. S1 to S3References10 November 2010; accepted 23 December 201010.1126/science.1200282to the recalcitrant nature of the substrate towardenzymatic breakdown and the relatively low activityof currently available hydrolytic enzymes.Although the success of protein engineering toimprove the performance of existing lignocellulosedegradingenzymes has been limited (3), retrievingenzymes from naturally evolved biomass-degradingmicrobial communities offers a promising strategyfor the identification of new lignocellulolyticenzymes with potentially improved activities (4).Metagenomics, the direct analysis of DNAfrom environmental samples, represents a strategyfor discovering diverse enzymes encoded in nature(5, 6). Although metagenomics has been used1 Department of Energy, Joint Genome Institute, Walnut Creek,CA 94598, USA. 2 Genomics Division, Lawrence Berkeley NationalLaboratory, Berkeley, CA 94720, USA. 3 Energy BiosciencesInstitute, University of California, Berkeley, CA 94720,USA. 4 Illumina Inc., Hayward, CA 94545, USA. 5 Department ofChemical and Biomolecular Engineering, University of California,Berkeley, CA 94720, USA. 6 Department of Animal Sciences,Institute for Genomic Biology and Energy Biosciences Institute,University of Illinois, Urbana, IL 61801, USA.*These authors contributed equally to this work.†To whom correspondence should be addressed. E-mail:emrubin@lbl.govwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 463

REPORTSsuccessfully to identify enzymes with desiredactivities (7), it has relied primarily on relativelylow-throughput function-based screening of environmentalDNA clone libraries (8, 9). Sequencebasedmetagenomic discovery of complete genesfrom environmental samples has been limited bythe microbial species complexity of most environmentsand the consequent rarity of full-lengthgenes in low-coverage metagenomic assemblies(8, 10, 11).In this study, we generated 268 gigabases ofmetagenomic sequence data from the microbiotain cow rumen to identify genes and genomesparticipating in biomass deconstruction. To isolaterumen microbes associated with definedplant substrates for subsequent genomic assessment,we incubated biomass-containing nylonbags in two fistulated cows (12) (Fig. 1 and tableS1). We isolated organisms that had become adherentto the plant fiber material during incubationin an attempt to target microbes specificallyinvolved in biomass degradation.We used switchgrass (Panicum virgatum), apromising cellulosic energy crop (13), as the plantsubstrate for our studies. To determine the cow’sability to degrade this substrate, we compared thechemical composition of the switchgrass beforeand after rumen incubation. Switchgrass degradationwas substantial (37% dry mass reduction after72 hours of incubation). Further analysis confirmedthat the decrease in mass of the switchgrass fiberwas largely due to degradation of both celluloseand hemicellulose, which together accounted for72% of the reduction in dry mass during incubation(table S2). The remaining reduction in drymass is likely in large part due to the degradationof pectin, protein, and other components of plantbiomass (14). These results indicate that the cowrumen microbiota is able to degrade this fibersource and support previous observations that therumen environment contains some of the mostcellulolytic mesophilic microbes described fromany habitat (15).To examine whether a unique fiber-degradingmicrobial community was enriched on switchgrassincubated within the nylon bags, we comparedthe community composition of microbes adherentto rumen-incubated switchgrass to the microbialpopulation from bulk rumen fluid. We usedpyrotag sequencing of small subunit ribosomalRNA genes (16) to identify operational taxonomicunits (OTUs) in two fistulated cows for each ofthe two samples. Rarefaction analysis indicatedthat the pyrotag sequencing depth was sufficientto capture the vast majority of OTUs in eachsample and suggests that about 1000 differentOTUs were present in each of these samples (fig.S1), which is consistent with previous estimatesof microbial complexity in the rumen (17). Comparisonof the OTUs identified in the switchgrassfiber-adherent community and the communitypresent in rumen fluid revealed overlaps betweencows and substrates, as well as reproducibleenrichment of specific bacterial phylotypes inthe switchgrass-adherent fraction (18).We targeted a single sample of switchgrassadherentrumen microbes for deep metagenomicsequencing with the goal of maximizing the likelihoodof obtaining large contiguous stretches ofoverlapping sequence reads (contigs) containingfull-length lignocellulolytic genes. We generatedseveral sequencing libraries from this sample withpaired-end read separations (equivalent to insertsizes in clone-based libraries) of 200 base pairs(bp), 300 bp, 3 kbp, and 5 kbp. Massively parallelsequencing (19) from all libraries yielded 1.5 billionread pairs, ranging in length from 2 × 36 bpto 2 × 125 bp and amounting to a total of 268 Gbpof sequence information. A summary of the libraryand sequencing technologies used is providedin table S3.To identify candidate carbohydrate-activegenes from this metagenomic sequence data set,AFig. 1. (A) A surgically created fistula (arrow) sealed with a flexiblecannula was used to study the degradation of switchgrasswithin the rumen. (B) Switchgrass before rumen incubation. (C)Nylon bags filled with switchgrass before insertion into the rumen.(D) Switchgrass after 72 hours of rumen incubation.Fig. 2. (A) Sequenceidentity of 90 candidatesequences assembled fromthe switchgrass-associatedrumen microbiome andtested for carbohydratedegradingactivity to knowncarbohydrate-active enzymes.Sequence identityto known enzymes isshown for tested candidates(blue) and candidatesfound to be active(red) toward at least oneof the substrates used inthe activity assays. (B)ANumber of genes2520151050Candidate CAZymestested for activity (n=90)with cellulolyticactivity (n=51)we performed de novo assembly and predicted2,547,270 open reading frames (ORFs). The averageORF length was 542 bp, and 55% of theORFs were predicted to represent full-length genes.All predicted genes were screened for candidateproteins with potential enzymatic activity towardplant cell wall polysaccharides. To minimize thedependence on overall sequence similarity ofcandidate genes to known carbohydrate-activeenzymes, we searched candidate genes for thepresence of individual predicted functional domains,rather than global sequence similarity toknown carbohydrate-active enzymes. We identified27,755 candidate genes with a significantmatch to at least one relevant catalytic domain orcarbohydrate-binding module (18) (table S4). Thesequence domains identified in our sample werelargely consistent with a prokaryotic origin of can-BCDpre-incubationpost-incubation30 40 50 60 70 80 90 100Sequence identity to best BLAST hit in CAZy (%)Similarity distribution of CAZyme candidates (n = 27,755) containing a catalytic domain (CD) associatedwith carbohydrolytic activity or a carbohydrate-binding module (CBM). Sequences were comparedto the CAZy (blue, 25,947 hits), NCBI-nr (black, 26,679 hits), and NCBI-env (green, 26,030 hits)databases (best BLAST hit, E-value ≤ 1e-5); 482 genes contained both a CD and CBM, whereas 23,804and 3469 genes contained only a CD or CBM, respectively.BNumber of geneswith BLAST hit40003000200010000BLASTP hits to CAZyBLASTP hits to NCBI_NRTBLASTN hits to NCBI_ENV20 30 40 50 60 70 80 90 100Sequence identity to best BLAST hit (%)46428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

didate gene sequences, with isolated examples ofdomains that are characteristic for eukaryotes.Comparison of the 268 Gbp obtained by sequencingof the switchgrass-adherent microbiometo data from previously published lower-depthmetagenomic studies of other plant-feedinganimals (8, 10, 11) revealed that the number ofcandidate carbohydrate-active genes identified inthe present study was larger by a factor of 5 thanthe combined number of candidate carbohydrateactivegenes from all previous studies (table S5).The total amount of sequence analyzed in theseearlier studies (combined: 0.21 Gbp) was threeorders of magnitude less than the present data setand thus resulted predominantly in identification ofpartial genes. In contrast, genes in the present studyare derived from assemblies with an average of 56-fold sequence coverage, and more than 15,000 ofthe candidate carbohydrate-active enzymes reportedhere were predicted to represent full-length genes.Rarefaction analysis indicates that even at theconsiderable sequencing depth of this study, onlya subset of genes present in the cow rumen microbiotawas assembled (figs. S4 and S5).Although the present study focuses on thevalidation of a subset of carbohydrate-active enzymefamilies, we expect the full repertoire ofgenes involved in biomass deconstruction to bepresent in the fiber-adherent rumen metagenomicdata set. To test this hypothesis, we searched ourdata set for cohesins and dockerins, proteins commonlyinvolved in the formation of lignocellulolyticmulti-enzyme complexes (cellulosomes) (20),and cellobiose phosphorylases, proteins belongingREPORTSto the family of glycosyltransferases. We wereable to identify 80 and 188 ORFs containing thecohesin- and dockerin-specific PFAM domains,respectively. We also identified 811 genes from theswitchgrass-adherent rumen microbiome that hadsignificant similarity to cellobiose phosphorylasesdeposited in NCBI-nr (BLAST search, E ≤ 1e-5).These results indicate that a wide spectrum ofbiomass-degrading genes can be identified throughanalysis of the sequence data generated in this study.Focusing on our set of 27,755 predictedcarbohydrate-active enzymes, we compared theirsequences to entries in the Carbohydrate ActiveEnzyme (CAZy) database, which contains bothexperimentally verified and inferred carbohydrateactiveenzymes (21). In the CAZy database, 1075,1199, and 251 entries are annotated as b-(1,4)Fig. 3. Carbohydrolytic potential of candidate carbohydrate-active enzymeson glycosidic substrates of different complexity. (A) Summary of carbohydrolyticactivities of 90 tested candidates on ten substrates. Candidates thatwere not active on any of the tested substrates and the four substrates thatwere recalcitrant to all tested candidates are not shown. (B and C) Cellulolyticactivities of candidate carbohydrate-active enzymes on five substrates.Carbohydrate-active gene candidates were expressed in a cell-free system (B)or using Escherichia coli as expression host (C). Samples were only considered“active”▪ ( ) if the measured mean glucose equivalent quantitatively exceededthe activity of negative controls plus one standard deviation by at least 50%(indicated by shaded horizontal line in each panel) and was significantlyhigher (*P < 0.05, Student’s t test) than the negative controls. Samples notmeeting both criteria were considered as “not active” (□; n.s. = not significant).All measurements were performed in duplicate. IL, ionic liquid.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 465

REPORTSendoglucanases, b-glucosidases, and cellobiohydrolases,respectively (63%, 86%, and 87% ofthese entries lack an Enzyme Commission (EC)number, indicating that their assigned activity hasnot been verified biochemically). In our rumenderiveddata set, we identified 1086, 1477, and 153sequences whose most significant matches (BLASTsearch, E ≤ 1e-5) were to a b-(1,4) endoglucanase,b-glucosidase, or cellobiohydrolase, respectively,within the CAZy database. Only 1% of these 2716sequences were highly similar (>95% sequenceidentity) to any CAZy database entry, indicating thatnearly all of these enzymes had not been previouslydeposited in CAZy. The overall lower efficiency atwhich new candidate cellobiohydrolases were identifiedmay be due to their underrepresentation inthe reference database, but even in this category weobserved a 56% increase of candidate sequenceswith 95% sequence identity,28 of 29 putative genes). These results suggestthat a substantial proportion of the genespredicted on the basis of short-read assembliesextracted from the metagenomic data representauthentic genes present in rumen microbes.To evaluate the biochemical activity of theputative carbohydrate-active genes identifiedby metagenomic sequencing of the switchgrassassociatedmicrobiome, we chose 90 candidategenes predicted to contain a glycoside hydrolasefamily 3, 5, 8, 9, 10, 26, or 48 domain or acarbohydrate-binding module. The selected candidategenes were expressed using two complementaryexpression systems, and the obtainedproteins were subjected to biochemical activityassays. The genes selected for expression rangedfrom 29 to 96% amino acid sequence identity toknown carbohydrate-active proteins, with an averageof less than 55% identity (Fig. 2A). Wetested all 90 proteins for enzymatic activity on apanel of 10 different substrates. This panel includedeight model substrates—carboxymethylcellulose (CMC), p-nitrophenyl b-glucoside, gumguar, lichenan, laminarin, mannopentose, Avicel,and xylan—along with two potential biofuel feedstocks,miscanthus and switchgrass (22), to providean initial understanding of the substratespecificity for each of the tested candidates. Thebiofuel crop substrates and Avicel were subjectedto ionic liquid pretreatment before conducting theactivity assays. In total, 51 of 90 (57%) testedproteins showed enzymatic activity against atleast one of the substrates, suggesting that thecandidate genes predicted by our metagenomicstrategy are highly enriched in enzymes withrelevant activities (Fig. 3 and table S6). Therewas no evidence that proteins with high sequenceidentity to known enzymes were more likely tobe active than proteins with low sequence similarity(P = 0.66, Kolmogorov-Smirnov test; Fig.2A). Inactivity of the remaining carbohydrateactivecandidates in these assays could be due toa number of reasons, including false-positive predictionof carbohydrate-active enzyme domains,minimal expression and/or misfolding of candidateproteins, or suboptimal reaction conditions.The overall high validation rate observed in theseassays suggests that the number and sequencediversity of known genes encoding hydrolyticenzymes from these and possibly other enzymeFig. 4. (A) Draft genomes AGenomeEstimated Bassembled from switchgrassadherentrumen microbes.(Mb)nessGenomePhylogeneticSizeComplete-BinOrderCompleteness was estimatedas fraction of the number ofAFa 2.87 Spirochaetales 92.98%identified and the number ofAMa 2.21 Spirochaetales 91.23%expected core genes withinAIa 2.53 Clostridiales 90.10%the phylogenetic order. Formore information on the assembledAGa 3.08 Bacteroidales 89.77%genomes, see tableAN 2.02 Clostridiales 78.50%S10. (B) Circular representationof the draft genome (ge-AJ 2.24 Bacteroidales 75.96%Genome Bin APbAC2a 2.07 Bacteroidales 75.96%55 Scaffoldsnome bin APb) validatedbyAWa 2.02 Clostridiales 75.77%single amplified genome analysis.2.41 MbFrom outside toward theAH 2.52 Bacteroidales 75.45%center: outermost circle, scaffoldswithin the draft genomeAQ 1.91 Bacteroidales 71.36%(random order, low-quality regionsAS1a 1.75 Clostridiales 70.99%removed); circle 2, IlluminaAPb 2.41 Clostridiales 64.85%read coverage in metagenomicdata (each horizontal red lineBOa 1.67 Clostridiales 64.16%indicates 25-fold coverage); circle3(bluetickmarks),regionsADa 2.99 Myxococcales 62.13%of the draft genome simultaneouslycovered by 454 readsATa 1.87 Clostridiales 60.41%derived from a single amplified genome; innermost circle (green tick marks), location of glycoside hydrolase genes on draft genome.46628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

families was substantially increased through thismetagenomic data set.A considerable fraction (20%) of the testedcarbohydrate-active enzyme candidates showedactivity toward biofuel crops pretreated with ionicliquids, one of the most promising initial stepsin the deconstruction of biomass (23). Becauseionic liquids can inhibit enzymatic biomass degradation(24), the retention of enzymatic activityin their presence makes these proteins promisingcandidates for more detailed physicochemicalanalyses. In addition, the tested set of targetcandidates also included two enzymes (MH-9and MH-10) that showed activity on CMC agarplates and contained only a carbohydrate-bindingmodule, but no known catalytic domain specificfor carbohydrate-active enzymes (table S6). Itis possible that these two enzymes contain catalyticdomains that share little similarity with thecatalytic domains of currently known carbohydrateactiveenzymes.To enable genomic studies of these microbes,we developed a strategy for producing draft genomesfrom deep metagenomic data. An initialassembly of the 268 Gbp of metagenomic sequenceresulted in 179,092 scaffolds, of whichthe 65 largest ranged in size from 0.5 to 1.5 Mbp(tables S8 and S9). Only 47 (0.03%) of the assembledscaffolds showed high levels of similarity(≥90% identity over ≥1000 bp) to previouslysequenced genomes available in GenBank. Mostof these alignable scaffolds were small (medianlength: 1626 bp) and the aligned regions typicallycovered nearly the entire length of the scaffold(median: 91% of scaffold length). Theseresults suggest that the vast majority of theassembled scaffolds represent segments of hithertouncharacterized microbial genomes. We furthervalidated these assemblies via two independentindicators of scaffold integrity: (i) level and uniformityof read depth in subregions, and (ii) matepairsupport. We identified 26,042 scaffolds greaterthan 10 kbp that satisfied these criteria for scaffoldintegrity (18), totaling 568 Mbp (N50: 24 kbp;longest scaffold: 541 kbp). To generate draft genomes,we binned the validated scaffolds by meansof two complementary properties expected tobe present in scaffolds derived from the samegenome: (i) tetranucleotide frequencies (TNFs)and (ii) read coverage. TNF signatures are generallyan effective approach for distinguishingsequences derived from different genomes butcan be similar for closely related species (25, 26).In contrast, read coverage is directly correlated tothe relative abundance of each organism in thesample and can thus be used to distinguish scaffoldsthat are likely derived from different closelyrelated organisms. In total, 446 genome bins withconsistent TNF and read coverage were formed.To estimate the completeness of the largest potentialmicrobial draft genomes identified throughthis approach, we first determined the most likelyphylogenetic order from which each of these binswas derived. For each of these orders, we used allavailable sequenced reference genomes (tableS11) to identify a minimal set of core genes thatare present in all members of this order (27, 28).Comparison of each draft genome to the pangenomeof the respective phylogenetic orderdemonstrated that between 60% and 93% of thecore genes were included in the 15 draft genomesfound to be most complete by this measure,similar to the fraction found in each referencegenome used for comparison (Fig. 4A and tablesS10 and S12). These observations suggest thatnear-complete draft genomes were successfullyassembled. To address the possibility that thecompleteness of individual draft genomes wasoverestimated as a result of binning of scaffoldsderived from multiple organisms, we further validatedtheir authenticity by copy number analysisof genes that were present only in single copy inall reference genomes of the respective phylogeneticorder (18) (tables S10 and S12).To test experimentally the validity and completenessof draft genomes derived from metagenomicscaffold bins, we obtained genomesequence data from individual uncultured microbialcells isolated directly from the same complexrumen community. Single cells loosely adherentto switchgrass were isolated using fluorescenceactivatedcell sorting (29) followed by wholegenome amplification (30). Screening of 16S sequencessuggested that one of the single cellsanalyzed was related to the fibrolytic Butyrivibriofibrisolvens and matched bin APb, one of thelargest bins assembled from metagenomic data(Fig. 4A and table S10). From this particular singlecell, we generated 65,272 reads (22.5 Mbpafter appropriate filtering) of which 55% mappedto genome bin APb. The remaining mappablesingle-cell reads matched either unbinned scaffoldsor assembly regions with poor scaffold integrity.Each of the 55 scaffolds in bin APb wassupported by substantial numbers of mapped singlecell–derived reads, suggesting that all scaffoldsin bin APb represent segments of the genome ofthe same single organism (Fig. 4B). These resultsdirectly support the assumption that individualgenome bins derived from our assembly representauthentic draft genomes and suggest that substantialproportions of the respective genomesare covered by these bins.Discovery of full-length genes with definedfunctions from complex microbial communitieshas previously been severely limited by the lowthroughput of the required cellular and molecularmanipulations (8, 11, 31). Our study demonstratesthe potential of deep sequencing of a complexcommunity to accurately reveal genes of interestat a massive scale, and to generate draft genomesof uncultured novel organisms involved inbiomass deconstruction. Although this work focusedon the identification and validation of newcarbohydrate-active enzymes, these data sets providean extensive resource for the discovery of amultitude of other classes of enzymes known toexist in the rumen, and the general approachpresented here will be applicable to other environmentalmicrobial communities.References and Notes1. H. W. Blanch et al., ACS Chem. Biol. 3, 17 (2008).2. K. Sanderson, Nature 444, 673 (2006).3. F. Wen, N. U. Nair, H. Zhao, Curr. Opin. Biotechnol. 20,412 (2009).4. E. M. Rubin, Nature 454, 841 (2008).5. M. Ferrer et al., Environ. Microbiol. 7, 1996 (2005).6. F. Wang, F. Li, G. Chen, W. Liu, Microbiol. Res. 164, 650(2009).7. L. L. Li, S. R. McCorkle, S. Monchy, S. Taghavi,D. van der Lelie, Biotechnol. Biofuels 2, 10 (2009).8. P. B. Pope et al., Proc. Natl. Acad. Sci. U.S.A. 107,14793 (2010).9. T. Uchiyama, K. Miyazaki, Curr. Opin. Biotechnol. 20,616 (2009).10. J. M. Brulc et al., Proc. Natl. Acad. Sci. U.S.A. 106, 1948(2009).11. F. Warnecke et al., Nature 450, 560 (2007).12. J. H. Meyer, R. I. Mackie, Appl. Environ. Microbiol. 51,622 (1986).13. D.J.Parrish,J.H.Fike,Methods Mol. Biol. 581, 27 (2009).14. W. J. Kelly et al., PLoS ONE 5, e11942 (2010).15. P. J. Weimer, J. Dairy Sci. 79, 1496 (1996).16. Z. Liu, C. Lozupone, M. Hamady, F. D. Bushman,R. Knight, Nucleic Acids Res. 35, e120 (2007).17. D. O. Krause et al., FEMS Microbiol. Rev. 27, 663 (2003).18. See supporting material on Science Online.19. D. R. Bentley et al., Nature 456, 53 (2008).20. M. T. Rincon et al., PLoS ONE 5, e12476 (2010).21. B. L. Cantarel et al., Nucleic Acids Res. 37 (databaseissue), D233 (2009).22. T. Demura, Z. H. Ye, Curr. Opin. Plant Biol. 13, 299 (2010).23. H. Zhao, G. A. Baker, J. V. Cowins, Biotechnol. Prog. 26,127 (2010).24. H. Zhao et al., J. Biotechnol. 139, 47 (2009).25. H. Teeling, A. Meyerdierks, M. Bauer, R. Amann,F. O. Glöckner, Environ. Microbiol. 6, 938 (2004).26. T. Woyke et al., Nature 443, 950 (2006).27. P. Lapierre, J. P. Gogarten, Trends Genet. 25, 107 (2009).28. D. Medini, C. Donati, H. Tettelin, V. Masignani,R. Rappuoli, Curr. Opin. Genet. Dev. 15, 589 (2005).29. R. Stepanauskas, M. E. Sieracki, Proc. Natl. Acad.Sci. U.S.A. 104, 9052 (2007).30. T. Woyke et al., PLoS ONE 4, e5299 (2009).31. H. García Martín et al., Nat. Biotechnol. 24, 1263 (2006).32. We thank J. Bristow, P. Hugenholtz, F. Warnecke, andK. Mavrommatis for critical discussions and reading themanuscript. We acknowledge technical support by theJGI production team, L. M. Sczyrba, M. Harmon-Smith,J. Froula, J. Martin, C. Wright, A. Lipzen, J. Zhao,S. Malfatti and Stefan Bauer. We thank P. D’Haeseleerfor sequences extracted from the CAZy database, JonasLøvaas Gjerstad for the picture of the fistulated cow,T. Shinkei, T. Yannarell, J. Kim and staff at the Dairy Farm,Department of Animal Sciences for assistance with themaintenanceofthefistulatedcows,nylonbagexperimentsand lab procedures carried out at the University of Illinois.The work conducted by the U.S. Department of Energy JointGenome Institute was supported in part by the Office ofScience of the U.S. Department of Energy under contractDE-AC02-05CH112 and U.S. Department of Energy undercontract DE-AC02-05CH11231 (cow rumen metagenomicsdata analysis and informatics). Supported by a researchgrant from the Energy Biosciences Institute at theUniversity of California, Berkeley (M.H.). Data areavailable at the NCBI Short Read Archive under accessionnumber SRA023560 and GenBank accession numbersHQ706005-HQ706094. Complete data can also beaccessed through the Web site of the DOE Joint GenomeInstitute (www.jgi.doe.gov).Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/463/DC1Materials and MethodsSOM TextTables S1 to S13Figs. S1 to S8References12 November 2010; accepted 29 December 201010.1126/science.1200387REPORTSwww.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 467

REPORTSCleavage of NIK by the API2-MALT1Fusion Oncoprotein Leads toNoncanonical NF-kB ActivationShaun Rosebeck, 1 ‡ Lisa Madden, 1 *‡ Xiaohong Jin, 2 Shufang Gu, 1 Ingrid J. Apel, 2 Alex Appert, 3Rifat A. Hamoudi, 3 Heidi Noels, 4,5 † Xavier Sagaert, 6 Peter Van Loo, 4,5 Mathijs Baens, 4,5Ming-Qing Du, 3 Peter C. Lucas, 2 § Linda M. McAllister-Lucas 1 §Proper regulation of nuclear factor kB (NF-kB) transcriptional activity is required for normallymphocyte function, and deregulated NF-kB signaling can facilitate lymphomagenesis. Wedemonstrate that the API2-MALT1 fusion oncoprotein created by the recurrent t(11;18)(q21;q21)in mucosa-associated lymphoid tissue (MALT) lymphoma induces proteolytic cleavage ofNF-kB–inducing kinase (NIK) at arginine 325. NIK cleavage requires the concerted actions ofboth fusion partners and generates a C-terminal NIK fragment that retains kinase activity and isresistant to proteasomal degradation. The resulting deregulated NIK activity is associated withconstitutive noncanonical NF-kB signaling, enhanced B cell adhesion, and apoptosis resistance.Our study reveals the gain-of-function proteolytic activity of a fusion oncoprotein and highlightsthe importance of the noncanonical NF-kB pathway in B lymphoproliferative disease.Mucosa-associated lymphoid tissue (MALT)lymphoma, the most common extranodalB cell tumor, accounts for 8% ofnon-Hodgkin’s lymphomas (1). The API2-MALT1fusion oncoprotein present in t(11;18)-positiveMALT lymphomas is composed of the N terminusof API2 [also termed cellular inhibitor ofapoptosis 2 (cIAP2)] linked to the C terminus ofMALT1 (1). Wild-type MALT1 mediates antigeninducednuclear factor kB (NF-kB) stimulation,which leads to lymphocyte survival and proliferation(2). MALT1 activates canonical NF-kBsignalingafter autooligomerization induced by upstream1 Department of Pediatrics and Communicable Diseases, Universityof Michigan, 1150 West Medical Center Drive, AnnArbor, MI, 48109, USA. 2 Department of Pathology, Universityof Michigan, 1150 West Medical Center Drive, Ann Arbor, MI,48109, USA. 3 Division of Molecular Histopathology, Departmentof Pathology, University of Cambridge, Laboratory Block,Addenbrooke’s Hospital, Hills Road, Cambridge CB2 0QQ, UK.4 Human Genome Laboratory, Molecular Genetics, Center forHuman Genetics, Catholic University Leuven, B-3000 Leuven,Belgium. 5 Human Genome Laboratory, Department of Molecularand Developmental Genetics, Flanders Institute for Biotechnology(VIB), B-3000 Leuven, Belgium. 6 Section of Morphologyand Molecular Pathology, Department of Pathology, CatholicUniversity Leuven, B-3000 Leuven, Belgium.*Present address: Department of Pediatrics, Washington UniversitySchool of Medicine, 660 South Euclid Avenue, St. Louis,MO 63110, USA.†Present address: Institute for Molecular CardiovascularResearch, RWTH Aachen University, 52074 Aachen, Germany.‡These authors contributed equally to the work.§To whom correspondence should be addressed. E-mail:plucas@umich.edu (P.C.L.); lindaluc@umich.edu (L.M.M-L.)ANIK: + - -- - - - -Flag-p100: + + + + + +API2-MALT1:p100p52GAPDHWB:α-FlagWB:α-GAPDHBFlag-API2-MALT1:MG132:p100p52API2-MALT1- + +- - + - +WB:α-p100/52WB:α-FlagCFlag-API2-MALT1: - +p52RelBHDAC1API2-MALT1WB:α-p52WB:α-RelBWB:α-HDAC1WB:α-FlagNuclearextractCytoplasmicextractDp100SSK41 controlSSK41 API2-MALT1WB:α-p100/52EFlag-API2-MALT1: - + - + - +HA-NIK 624-947 : - + + -HA-NIK KK429/430AA : - - -- - - + +p100p52WB:α-p100/52FFlag-API2-MALT1:HA-NIK 624-947 :p52HDAC1- + +- - + - +WB:α-p100/52WB:α-HDAC1Nuclearextractp52API2-MALT1GAPDHWB:α-FlagWB:α-GAPDHAPI2-MALT1NIK 624-947*NIK KK429/430AAWB:α-FlagWB:α-HAAPI2-MALT1NIK 624-947WB:α-FlagWB:α-HACytoplasmicextractFig. 1. API2-MALT1 induces noncanonical NF-kB signaling through NIK.(A and B) HEK293T cells were transfected as indicated, and p100processing to p52 was assessed by Western blot (WB) with an antibodyagainst Flag (a-Flag) (A) or a-p100/52 (to detect endogenous p100/52)(B). Where indicated, cells were treated with proteasome inhibitor,MG132. (C) After transfection of HEK293T cells, nuclear extracts wereprepared and analyzed for p52 and RelB by WB. (D) p100 processing inlysates from control SSK41 cells or SSK41 cells stably expressing API2-MALT1 was analyzed by WB. (E and F) HEK293T cells were transfectedwith API2-MALT1 in the absence or presence of NIK mutants. Endogenousp100 processing was analyzed by WB (E), or nuclear extracts wereanalyzed for the presence of p52 (F). *Nonspecific band. HDAC1, histonedeacetylase 1 (loading control for nuclear extract). Data are representativeof at least three separate experiments.46828 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

factors CARMA1 and Bcl10 (3, 4). It is thoughtthat because the API2 moiety mediates autooligomerization,API2-MALT1 can stimulate NF-kBindependent of upstream signals (5, 6). This mayexplain why t(11;18)-positive MALT lymphomasare not dependent on antigenic stimulation forprogression, whereas t(11;18)-negative tumorsrequire ongoing chronic inflammation for survival.The phenomenon is best exemplified by gastricMALT lymphomas, the majority of which arise inthe setting of chronic Helicobacter pylori gastritisand are cured by eradication of H. pylori withantibiotics. In contrast, t(11;18)-positive gastrictumors are resistant to this treatment and areassociated with advanced-stage disease (1).We discovered that besides activating canonicalNF-kB, expression of API2-MALT1 in humanembryonic kidney 293T (HEK293T) or the humanB lymphoma (SSK41) cells induced proteasomedependentprocessing of the NF-kB precursor,p100, to its mature form, p52, and stimulatednuclear translocation of p52/RelB (Fig. 1, A toD, and fig. S1) (7). These findings indicate thatAPI2-MALT1 activates the “noncanonical” NF-kBpathway, which requires NF-kB–inducing kinase(NIK)–dependent phosphorylation and activationof inhibitor of NF-kB (IkB) kinase-a (IKKa).REPORTSThis, in turn, triggers proteasome-mediated partialdegradationofp100top52andgenerationoftranscriptionallyactive p52/RelB NF-kB dimers(8).Consistent with this notion, dominant-negative NIKmutants (9, 10) blocked API2-MALT1–dependentp100 processing and p52 nuclear translocation(Fig.1,EandF).cIAP1 and cIAP2 (API2) associate with NIKand promote NIK degradation via RING domainubiquitin ligase activity (11–13). We hypothesizedthat API2-MALT1, which lacks the cIAP2RING domain, stimulates noncanonical signalingthrough competitive inhibition of cIAP-mediatedNIK degradation. In testing this, we discoveredthat expression of API2-MALT1 instead inducedproteolytic cleavage of NIK, generating ~37-kDN-terminal and ~70-kD C-terminal NIK fragments(Fig.2,AandB,andfig.S2).API2-MALT1 fusion transcripts invariably containthree intact baculoviral IAP repeat (BIR)domains from API2 and an intact “caspaselike”domain from MALT1, which suggests that thesedomains are critical for lymphomagenesis (1).The caspaselike domain of wild-type MALT1 hasproteolytic activity, and Bcl10 and the NF-kBinhibitor, A20, are the only known substrates(14, 15). We therefore investigated whether the caspaselikedomain within API2-MALT1 is also ableto cleave NIK. Deletion mutants of API2-MALT1lacking portions of the caspaselike domain andAPI2-MALT1-C678A (16), in which the catalyticcysteine within the MALT1 proteolytic domainis replaced with alanine, were unable to induceNIK cleavage (Fig. 2C and fig. S3, A to C). Furthermore,treatment with z-Val-Arg-Pro-Argfluoromethylketone(z-VRPR-fmk), a MALT1protease inhibitor (15), blocked API2-MALT1–induced NIK cleavage, whereas z-IETD-fmk, acaspase-8 inhibitor, had no effect (fig. S3, Dand E). Finally, an in vitro cleavage reaction usingpurified recombinant proteins showed that NIK isa direct substrate of the MALT1 protease domain(Fig. 2D and fig. S4).Cellular expression of the MALT1 moietyalone was unable to induce NIK cleavage, whichsuggests that the API2 moiety also contributes insome way (Fig. 2E). Indeed, analyses revealedthat NIK physically associates with API2-MALT1via the API2 moiety (Fig. 2F) and that the regionwithin the API2 moiety that mediates autooligomerizationof API2-MALT1 (amino acids 49to 98) (5) is required for efficient API2-MALT1–dependent NIK cleavage and p100 processing(fig. S5). The collaborative relationship of theADFlag-API2-MALT1:HA-NIK: +F.L. NIK (110 kD)Cleaved NIK (70 kD)(C-terminal)Cleaved NIK (37 kD)(N-terminal)API2-MALT1*+ +STII-MALT1: - + +V5-NIK: + + +AC-LSSR-CHO: - - +STII-MALT1F.L. NIK (110 kD)Cleaved NIK (70 kD)WB:α-NIKWB:α-HAWB:α-FlagWB: α-FlagWB: α-V5BDoxycycline (hrs):FL NIK (110 kD)Cleaved NIK (70 kD)EAPI2-MALT1GAPDHFlag-API2-MALT1:Flag-API2(1-441):Flag-MALT1(324-813):HA-NIK:F.L. NIK (110 kD)Cleaved NIK (37 kD)*BJAB B cells0 4 8 16 24BIR BIR BIR Ig Casp-like+ - -- - + - +- + --- + + +WB:α-NIKWB:α-FlagAPI2-MALT11 441 324 813WB:α-GAPDHWB: α-HAWB: α-FlagCFFlag-API2-MALT1construct:HA-NIK:F.L. NIKF.L. NIKAPI2-MALT1Flag-API2-MALT1:HA-NIK: +F.L. NIK (110 kD)Cleaved NIK (70 kD)Cleaved NIK (37 kD)API2-MALT1-Full length+ +- -- + + + + + + + +Full lengthE47A/R48A∆1-49∆1-98∆1-166C678AWB:α-NIKWB:α-HAWB:α-FlagAPI2 moiety onlyMALT1 moiety onlyIP: α-FlagWB: α-HAWB:α-HAWB:α-FlagTotallysateFig. 2. API2-MALT1 induces NIK cleavage, a phenomenon requiring both theAPI2 moiety and MALT1 protease activity. (A) HEK293T cells were transfectedas indicated, and NIK cleavage fragments were detected by WB. (B)BJABcellsexpressing API2-MALT1 from a tetracycline-inducible promoter were treatedwith doxycycline. WB with an antibody raised against a C-terminal NIK sequencerevealed time-dependent generation of an endogenous 70-kD NIKcleavage fragment. To enhance detection of full-length (FL) NIK, cells wereincubated with 25 mM MG132.(C) HEK293T cells were transfected as indicated,and the presence of the N-terminal 37-kD and the C-terminal 70-kDNIK cleavage fragments was analyzed by WB. (D) RecombinantpurifiedNIK-V5-bioC and StrepII-Flag–tagged MALT1 were incubated in kosmotropic saltbuffer for 6 hours at 37°C, with or without 100 mM MALT1 protease inhibitorAc-LSSR-CHO, and analyzed by WB. (E) HEK293T cells were transfectedas indicated, and the 37-kD NIK cleavage fragment was detected by WB.(F) HEK293T cells were transfected, and immunoprecipitations were carriedout using a-Flag–agarose. For a detailed description of API2-MALT1 mutants,see the legend to fig. S5. *Nonspecific band. Data are representative of at leastthree separate experiments.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 469

REPORTSAPI2 and MALT1 moieties in achieving NIKcleavage was further underscored in several differentcontexts. First, unlike API2-MALT1, inducedexpression of wild-type MALT1 in BJABB cells did not result in NIK cleavage (fig. S6A).Second, NIK cleavage was not observed in SSK41B cells, which are characterized by MALT1 geneamplification, overexpression of MALT1, andconstitutive MALT1 protease activity (14, 17, 18).In contrast to A20, NIK was cleaved only if SSK41cells were engineered to express API2-MALT1 (fig.S6B). Third, coexpression of wild-type MALT1with Bcl10, which triggers MALT1 oligomerizationand activation (3), did lead to cleavage of A20but not of NIK (fig. S6, C and D). Fourth, ligandinducedB cell receptor stimulation, which similarlyactivates the MALT1 protease (14), did nottrigger NIK cleavage (fig. S6E). Furthermore, althoughMALT1 oligomerization and activationrequire Bcl10 (3), API2-MALT1–mediated NIKcleavage occurred in the absence of Bcl10 (fig.S6F). Together, these findings suggest that NIK isa substrate for the MALT1 protease domain, butonly when this domain is present within the contextof API2-MALT1.Structural analyses predict that the MALT1protease should show specificity for substrates witha basic or uncharged residue at P1 (N-terminal tothe cleavage site) (19). Furthermore, the MALT1cleavage sites of Bcl10 and A20 both contain aP2-serine preceding a P1-arginine (14, 15, 20). Thus,we identified candidate P2-Ser/P1-Arg MALT1cleavage sites within NIK that would generate fragmentsof ~37 and 70 kD (fig. S7A), and we individuallychanged each candidate P1-Arg to Ala.The R366A and R368A NIK mutants werereadily cleaved by API2-MALT1; however, theR325A mutant was resistant (Fig. 3A), whichsuggests that API2-MALT1–dependent cleavageof NIK occurs at R325 (fig. S7B). Expressionof the resulting C-terminal NIK cleavagefragment, NIK(326–947), which retains the kinasedomain, induced robust p100 processing (fig.S7C), p52 nuclear translocation (Fig. 3B), andnoncanonical NF-kB target gene expression (fig.S7D). NIK(326–947) also induced nuclear translocationof the p65 NF-kB subunit, which indicatedactivation of the canonical NF-kBpathwayFig. 3. API2-MALT1–dependentcleavageofNIKat R325 generates an activeC-terminal fragment. (A)HEK293T cells were transfectedas indicated, and the37-kD N-terminal NIK cleavagefragment was detectedby WB. (B) HEK293T cellswere transfected as indicated,and nuclear translocationofp52andp65NF-kBsubunitswas assessed. (C andD) HEK293T cells were transfectedas indicated, and endogenousp100 processing(C) and nuclear translocationof NF-kB subunits (D) wereassessed. (E) HEK293T cellswere transfected as indicated,and the ability of endogenousTRAF3 or IKKa toimmunoprecipitate with eachNIK protein was assessed.(F) HEK293T cells were transfectedas indicated and thenincubated in the absence orpresence of 25 mM MG132.ThepresenceofNIKwasdetectedby WB. (G) Cell lysateswere prepared in the absenceof MG132 and analyzed byWB to detect full-length (F.L.)NIK and the 70-kD C-terminalNIK cleavage fragment. Dataare representative of at leastthree separate experiments.AHA-NIK:Flag-API2-MALT1:F.L. NIK (110 kD)Cleaved NIK (37 kD)CFFlag-API2-MALT1p100p52API2-MALT1NIKGAPDHFlag-TRAF3:F.L. NIK(110 kD)C-terminal NIK(70 kD)TRAF3WT R325A R366A R368A- + - + - + - +NIK-API2-MALT1API2-MALT1(C678A)Dp52p65API2-MALT1MG132HA-NIK: + - + - + -HA-NIK 326-947 : - + - + - +- - + + + +HDAC1-WB:α-HAWB:α-FlagWB:α-HAWB:α-FlagBAPI2-MALT1GHA-NIKconstruct:p52HDAC1F.L. NIK (110 kD)C-terminal NIK (70 kD)N-terminal NIK (37 kD)API2-MALT1(C678A)NuclearextractF.L. NIK (110 kD)CytoplasmicextractCleaved NIK (70 kD)API2-MALT1MALT1GAPDHp65E-HA-NIKconstruct:TRAF3IKKαTRAF3IKKαF.L. NIK(110 kD)C-terminal NIK(70 kD)SSK41 cellsCont.API2-MALT1WB:α-NIK*WB:α-MALT1WB:α-GAPDHWT (F.L. NIK)R325A1-325 (N-term frag)326-947 (C-term frag)-NuclearextractCytoplasmicextractWT (F.L. NIK)326-947 (C-term frag)IP:α-HATotallysate47028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

as well (Fig. 3B). Conversely, API2-MALT1-C678A,the catalytically inactive mutant that cannot induceNIK cleavage to produce NIK(326–947),failed to stimulate p100 processing (Fig. 3C) andp52 nuclear translocation (Fig. 3D and fig. S8).NIK associates with the adaptor protein, TRAF3,via an N-terminal NIK domain (amino acids 78to 84), and this interaction targets NIK for proteasomaldegradation (21–23). Because cleavageof NIK at R325 separates this TRAF3-bindingsite from the NIK kinase domain, we hypothesizedthat the active C-terminal NIK cleavageproduct would be resistant to TRAF3-directeddegradation. Indeed, NIK(326–947) retained bindingto IKKa but not to TRAF3 (Fig. 3E) and,unlike full-length NIK, was resistant to TRAF3-dependent proteasomal degradation (Fig. 3F).We also demonstrated the unique stability of theAPI2-MALT1–generated C-terminal NIKcleavage fragment in SSK41 B lymphoma cells.In the absence of MG132 (a proteosome inhibitor),expression of full-length NIK was verylow, regardless of whether API2-MALT1 waspresent, consistent with the fact that NIK issubject to constitutive proteasomal degradation(21, 24) (Fig. 3G). In contrast, high levels ofREPORTSendogenous C-terminal NIK cleavage fragmentwere detected in SSK41 cells expressing API2-MALT1 (Fig. 3G).API2-MALT1–dependent generation of theC-terminal NIK cleavage fragment in SSK41cells was associated with enhanced transcriptionof noncanonical NF-kB target genes, includingPim-2, an oncogenic kinase that blocks apoptosisby phosphorylating the proapoptotic Bcl-2 familymember BAD (fig. S9, A and B) (25–29).RNA interference–mediated knockdown of NIKin API2-MALT1–expressing SSK41 cells led toloss of the 70-kD NIK fragment, loss of p100ASSK41 API2-MALT1siRNA: Cont.p100p52Cleaved NIK(70 kD)API2-MALT1EGAPDHNIKWB:α-p100/p52WB:α-NIKWB:α-MALT1WB:α-GAPDHBCase #: 1 2 3 4 5F.L. NIK(110 kD) WB:Cleaved NIKα-NIK(70 kD)API2-MALT1MALT1p52/p100 ratiop100p52Actin0.80.60.40.2MALT lymphomat(11;18)positivet(11;18)negative0.0Case #: 1 2 3 4 5FollicularlymphomaSSK41 API2-MALT1shRNA: Cont. NIKCleaved NIK(70 kD)Pim2p-BADGAPDHWB:α-p100/p52WB:α-MALT1WB:α-ActinFRunning Enrichment Score (RES)-0.8 -0.6 -0.4 -0.2 0.0 0.2CCND1WB:α-NIKWB:α-Pim2WB:α-p-BADWB:α-GAPDHP = 0.0021FDR = 0.0021API2-MALT1NegativeCPeak at 2051Percent survival (vs untreated)BAFFPNAD100806040200***1 10 100Dexamethasone (nM)CXCL12MADCAM1CXCL9RANKLCXCL13Pim2Zero crossingat 5171Peak at 85400 2000 4000 6000 8000 10000Gene List IndexNumber of genes: 9794 (in list), 185 (in gene set)API2-MALT1API2-MALT1+ cont shRNAAPI2-MALT1+ NIK shRNAControlBcl2A1IRF3CCL21CCL19CXCR4API2-MALT1PositiveDGPercent Adhesion5040302010BSA coatedVCAM-1 coated0siRNA: Cont. NIK- -BJABControlHigher inAPI2-MALT1-casesMADCAM1*CXCR4CCL19CCL21IRF3CXCL12CXCL9RANKLCXCL13Pim2BAFFPNADAPI2-MALT1Cleaved NIK(70 kD)BJABAPI2-MALT1Higher inAPI2-MALT1+casessiRNA:200 100 0 100 200 300 700**##**#CCND1*##Percent Change(API2-MALT1 Positive versus Negative)**Cont.NIKFig. 4. API2-MALT1–dependent NIK cleavage is associated with up-regulationof noncanonical NF-kB target genes, results in an altered B cell phenotype, andoccurs in t(11;18)-positive MALT lymphoma. (A)API2-MALT1–expressing SSK41cells were transiently transfected with control or NIK small interfering RNA(siRNA), and p100 processing was assessed by WB. (B and C) API2-MALT1–expressing SSK41 cells were stably infected with control or NIK small hairpinRNA (shRNA) lentiviral particles. The 70-kD NIK cleavage fragment, Pim-2, andphospho-Ser 112 –BAD levels were compared by WB (B). Cells were treated for48 hours with dexamethasone, and percent viability was compared (C). Data areexpressed as average TSEM for four separate experiments. *P

REPORTSprocessing, and loss of API2-MALT1–dependentinduction of Pim-2 kinase and BAD phosphorylation(Fig. 4, A and B). In accordance with itseffect on antiapoptotic signal transduction, API2-MALT1 expression protected SSK41 cells fromdexamethasone-induced cell death, which wasreversed by NIK knockdown (Fig. 4, B and C,and fig. S10). Knockdown of IKKa impairedAPI2-MALT1–dependent protection, whichsupports a role for the noncanonical NF-kBpathway in mediating this effect of API2-MALT1–induced NIK cleavage (fig. S11). IKKbknockdown impaired API2-MALT1–dependentprotection as well, which implies that the canonicalpathway may also contribute (fig. S11).We next investigated the impact of API2-MALT1–dependent NIK cleavage on B celladhesion because we had observed that API2-MALT1 induced the expression of B cell integrins(fig. S9A), known noncanonical NF-kB genetargets (27, 28). API2-MALT1 expression wasassociated with increased B cell adhesion toplates coated with the endothelial protein vascularcell adhesion molecule VCAM-1, and thisproadhesive phenotype was fully dependent onNIK (Fig. 4D). Lymphocyte adhesion is thoughtto play a role in lymphoma dissemination, thus NIKcleavage–dependentAPI2-MALT1–induced adhesionmay contribute to the higher rate of tumorspread among t(11;18)-positive MALT lymphomas(1). Again, knockdown of IKKa or IKKbimpaired API2-MALT1–dependent adhesion,which suggests that both noncanonical and canonicalpathways contribute to the proadhesivephenotype after API2-MALT1–dependent NIKcleavage (fig. S12).The striking pattern of NIK cleavage and stabilityobserved in API2-MALT1–expressing Blymphoma cell lines was recapitulated in MALTlymphoma patient specimens. Full-length NIKlevels were relatively low in all lymphomas, whereasan endogenous 70-kD C-terminal NIK fragmentwas detected only in t(11;18)-positive MALTlymphomas expressing API2-MALT1 (Fig. 4E).The presence of the NIK cleavage product wasassociated with an elevated p52/p100 ratio in thet(11;18)-positive tumors, which indicated enhancednoncanonical NF-kB activation (Fig. 4E). Weperformed absolute gene set enrichment analysis(GSEA) comparing the expression of NF-kBtarget genes between t(11;18)-positive MALTlymphomas (n = 9) and those with no translocation(n = 8). Analysis revealed a significantdifference in the pattern of expression, with mostknown noncanonical NF-kB target genes overrepresentedin the t(11;18)-positive cases (Fig.4F, fig. S13, and tables S1 and S2). One exceptionis cyclin D1, although its categorizationas a noncanonical NF-kB target gene is controversial(30). We then compared the expression ofthese noncanonical target genes in a completelyseparate group of six t(11;18)-positive and eightt(11;18)-negative MALT lymphomas that werecollected at a different institution and, again, foundthat many noncanonical NF-kB target genes weremore highly expressed in the t(11;18)-positive tumors(Fig. 4G). CXCR4, which encodes a chemokinereceptor whose expression is associatedwith widespread lymph node involvement in Blymphomas (31–33), is one intriguing exampleof a noncanonical gene target that is up-regulatedin t(11;18)-positive tumors in both tumor collections(Fig.4,FandG)(34).Deregulated NIK activity has been increasinglyimplicated in the pathogenesis of B cellneoplasms. For example, an EFTUD2-NIK fusiononcoprotein that retains the NIK kinase domain—but lacks the TRAF3-binding site and is resistantto proteasomal degradation—was recently identifiedin a case of multiple myeloma (24). Anotherrecent report described a murine model of B lymphoproliferativedisease in which a NIK mutantlacking the TRAF3-binding domain (NIKDT3)wasexpressedinBcells(35). Compared with controltransgenic mice expressing full-length NIK,the NIKDT3 mice demonstrated increased NIKlevels with enhanced p100 processing in B cells.They also showed expanded MALT and profoundsplenic marginal zone B cell hyperplasia, aphenotype that bears similarity to the Em-API2-MALT1 transgenic mouse (36). Together withour results, these findings suggest that separatingthe TRAF3-binding site on NIK from the kinasedomain, either through aberrations of the NIKgene or through proteolytic cleavage of NIK protein,may represent a common mechanism forderegulating NIK activity in B cell neoplasms.Our findings suggest that in API2-MALT1–expressing MALT lymphomas, the API2 moietymediates autooligomerization of API2-MALT1and recruitment of NIK, and the MALT1 proteasedomain cleaves NIK, which leads to degradationresistantNIK kinase and deregulated noncanonicalNF-kB signaling (see model, fig. S14). Data suggestthat NIK cleavage protects API2-MALT1–expressing B cells from apoptosis and promotesB cell adhesion, both of which could contributeto the more aggressive phenotype of t(11;18)-positive MALT lymphomas (1). Disrupting theAPI2-NIK interaction and/or blocking MALT1protease or NIK kinase activity could representnew treatment approaches for refractory t(11;18)-positive MALT lymphoma.References and Notes1. P.G.Isaacson,M.Q.Du,Nat. Rev. Cancer 4, 644(2004).2. M. Thome, Nat. Rev. Immunol. 8, 495 (2008).3. P. C. Lucas et al., J. Biol. Chem. 276, 19012 (2001).4. L. Sun, L. Deng, C. K. Ea, Z. P. Xia, Z. J. Chen,Mol. Cell 14, 289 (2004).5. P. C. Lucas et al., Oncogene 26, 5643 (2007).6. H. Zhou, M. Q. Du, V. M. Dixit, Cancer Cell 7, 425(2005).7. Materials and methods are available as supportingmaterial on Science Online.8. T. D. Gilmore, Oncogene 25, 6680 (2006).9. N. L. Malinin, M. P. Boldin, A. V. Kovalenko, D. Wallach,Nature 385, 540 (1997).10. H. Y. Song, C. H. Régnier, C. J. Kirschning, D. V. Goeddel,M. Rothe, Proc.Natl.Acad.Sci.U.S.A.94, 9792(1997).11. S. L. Petersen et al., Cancer Cell 12, 445 (2007).12. E. Varfolomeev et al., Cell 131, 669 (2007).13. J. E. Vince et al., Cell 131, 682 (2007).14. B. Coornaert et al., Nat. Immunol. 9, 263 (2008).15. F. Rebeaud et al., Nat. Immunol. 9, 272 (2008).16. Single-letter abbreviations for the amino acidresiduesareasfollows: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.17. H. Noels et al., J. Biol. Chem. 282, 10180 (2007).18. D. Sanchez-Izquierdo et al., Blood 101, 4539(2003).19. A. G. Uren et al., Mol. Cell 6, 961 (2000).20. L. M. McAllister-Lucas, P. C. Lucas, Nat. Immunol. 9,231 (2008).21. G. Liao, M. Zhang, E. W. Harhaj, S.-C. Sun, J. Biol. Chem.279, 26243 (2004).22. S. Vallabhapurapu et al., Nat. Immunol. 9, 1364(2008).23. B. J. Zarnegar et al., Nat. Immunol. 9, 1371 (2008).24. C. M. Annunziata et al., Cancer Cell 12, 115 (2007).25. G. Bonizzi et al., EMBO J. 23, 4202 (2004).26. J. L. Chen, A. Limnander, P. B. Rothman, Blood 111,1677 (2008).27. T. Enzler et al., Immunity 25, 403 (2006).28. F. Guo, D. Weih, E. Meier, F. Weih, Blood 110,2381 (2007).29. R. T. Woodland et al., Blood 111, 750 (2008).30. I. I. Witzel, L. F. Koh, N. D. Perkins, Biochem. Soc. Trans.38, 217 (2010).31. S. López-Giral et al., J. Leukoc. Biol. 76, 462 (2004).32. Y. Nie et al., J. Exp. Med. 200, 1145 (2004).33. M. Luftig et al., Proc. Natl. Acad. Sci. U.S.A. 101,141 (2004).34. See Materials and Methods in Supporting OnlineMaterials for a description of the tumor collectionsnos. 1 and 2.35. Y. Sasaki et al., Proc. Natl. Acad. Sci. U.S.A. 105,10883 (2008).36. M. Baens et al., Cancer Res. 66, 5270 (2006).37. We thank M. Dyer for the SSK41 cells, R. Rennefor the BJAB Tet-On cells, and G. Nunez for severalexpression plasmids. A Material Transfer Agreement isrequired for use of the SSK41 cells expressingAPI2-MALT1. Microarray data from MALT lymphomatumor collections are Minimum Information About aMicroarray Experiment (MIAME) compliant and areavailable online at the Gene Expression Omnibus (GEO)with accession number GSE25527 for collectionno. 1 and accession number GSE25550 for collectionno. 2 (www.ncbi.nlm.nih.gov/projects/geo/). L.M. is arecipient of the Nancy Newton Loeb Pediatric CancerResearch and Helen L. Kay Pediatric Cancer ResearchAwards and received support from National Institute ofChild Health and Human Development, NIH,T32-HD07513. L.M. and S.R. were both supportedby National Heart, Lung, and Blood Institute, NIH,T32-HL007622-21A2. M.B. is supported by grantsfrom the Research Foundation—Flanders (FWO) andBelgian Foundation against Cancer. H.N. was anaspirant of the FWO-Vlaanderen. X.S. is a SeniorClinical Investigator of FWO-Vlaanderen, and P.V.L.is a postdoctoral researcher of the FWO. This workwas supported by the Shirley K. Schlafer Foundation,the Elizabeth Caroline Crosby Fund, and grantsfrom the University of Michigan ComprehensiveCancer Center (G007839), Leukemia and LymphomaResearch UK, and National Cancer Institute NIH(R01CA124540).Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/468/DC1Materials and MethodsFigs. S1 to S14Tables S1 and S2References12 October 2010; accepted 22 December 201010.1126/science.119894647228 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

Intramembrane Cleavage of AMA1Triggers Toxoplasma to Switch froman Invasive to a Replicative ModeJoana M. Santos, 1,3 David J. P. Ferguson, 2 Michael J. Blackman, 3 Dominique Soldati-Favre 1 *Apicomplexan parasites invade host cells and immediately initiate cell division. The extracellularparasite discharges transmembrane proteins onto its surface to mediate motility and invasion.These are shed by intramembrane cleavage, a process associated with invasion but otherwise poorlyunderstood. Functional analysis of Toxoplasma rhomboid 4, a surface intramembrane protease,by conditional overexpression of a catalytically inactive form produced a profound block inreplication. This was completely rescued by expression of the cleaved cytoplasmic tail ofToxoplasma or Plasmodium apical membrane antigen 1 (AMA1). These results reveal an unexpectedfunction for AMA1 in parasite replication and suggest that invasion proteins help to promoteparasite switch from an invasive to a replicative mode.Host cell invasion by Toxoplasma andPlasmodium involves discharge of secretoryorganelles called micronemes(1) and rhoptries. Apical membrane antigen1 (AMA1), a microneme protein, is crucial forinvasion (2) and is part of the moving junctioncomplex formed during invasion (3). In Toxoplasmagondii, AMA1 and microneme protein–2(MIC2), MIC6, MIC8, and MIC12 are cleaved duringinvasion within their transmembrane domain(TMD) by a rhomboid activity (4–7), releasingthem from the parasite surface. In Plasmodium,the majority of AMA1 shedding is mediated bya subtilisin protease (6, 8, 9), but cleavage by arhomboid protease also occurs (6).TheroleofAMA1 shedding is unknown. Toxoplasma encodessix rhomboids (10), two of which have been functionallydissected: ROM1 is expressed in themicronemes but does not play a crucial role ininvasion (11, 12); ROM4 localizes to the plasmamembrane (13, 14) and cleaves MIC2, AMA1,andMIC8(15). Plasmodium falciparum ROM4sheds the micronemal erythrocyte-binding proteinEBA175 (16) and possibly other adhesins(17). After invasion, the parasite initiates replicationwithin a parasitophorous vacuole. Toxoplasmatachyzoites divide by endodyogeny, in which repeatedcycles of replication produce numerousnew parasites equipped for egress and invasion.In contrast, Plasmodium replicates by schizogony,in which a multinucleated syncytium (schizont) isformed that undergoes cytokinesis only after completionof nuclear division. The signals governinginitiation of replication are unknown.Rhomboids are broad–substrate-specificityserine proteases that recognize helix-destabilizing1 Department of Microbiology, Faculty of Medicine, Universityof Geneva, 1 rue-Michel Servet, 1211 Geneva 4, Switzerland.2 Nuffield Department of Clinical Laboratory Science, Universityof Oxford, John Radcliffe Hospital, Oxford OX3 9DU,UK. 3 Division of Parasitology, Medical Research Council NationalInstitute for Medical Research, Mill Hill, London NW71AA, UK.*To whom correspondence should be addressed. E-mail:Dominique.Soldati-Favre@unige.chresidues within the substrate’s TMD(18). Wereasoned that expression of a T. gondii ROM4mutant, able to bind the substrate but unable tocleave it, would sequester the substrate from theendogenous protease and behave as dominantnegative. To regulate expression of the protease,we used the FK506 binding protein destabilizationdomain (dd) system in which fusion proteinsare degraded unless a ligand, Shld-1, is added(19). We expressed either a control wild-type (WT)construct (ddROM4) or a mutant form, in whichthe catalytic Ser 409 was substituted with an Alaresidue (ddROM4 S-A )(20). Similar mutations ablaterhomboid activity (13, 14, 17). To validate thesystem, a ROM1 mutant (ddROM1 S-A ) was alsogenerated. The transgenic proteins were correctlylocalized and expressed in a Shld-1–dependentmanner (Fig. 1, A and B). Expression of ddROM4was detectable 5 min after Shld-1 treatment, butlevels similar to those of the endogenous proteinand correct trafficking were only reached by 480or 180 min, respectively (fig. S1, A and B).Plaque assays reflect multiple lytic cycles,including invasion, replication, egress, and invasionof neighboring cells. Parasites expressingddROM4 or yellow fluorescent protein (RH-2YFP)formed plaques of similar size with or withoutShld-1, indicating that overexpression of ROM4was not detrimental. The ddROM1 S-A parasitesformed slightly smaller plaques in the presence ofShld-1, suggesting a modest growth defect. TheddROM4 S-A parasites produced no plaques in thepresence of Shld-1 (Fig. 1C), indicating that itsexpression exerts a dominant negative effect.Closer examination of the ddROM4 S-A phenotyperevealed an unanticipated defect in replication.Quantification of the number of parasites pervacuole 24 hours postinvasion indicated that theRH-2YFP and ddROM1 S-A parasites replicated ata similar rate regardless of the presence of Shld-1,whereas the ddROM4 parasites grew slightly better.In contrast, the ddROM4 S-A line grew normallywithoutShld-1butwasseverelyimpairedinthepresence of Shld-1 (Fig. 1D and fig. S1C). Modificationof the extreme C terminus of rhomboidsREPORTSinterferes with activity (13, 14, 16), and inclusionof a C-terminal Ty-1 epitope tag abrogated thedeleterious effect of ddROM4 S-A (fig. S2), supportingthe notion that stabilization of ddROM4 S-Aproduces a dominant-negative effect.Delivery of organelles into daughter cells duringreplication occurs in a highly coordinatedfashion, starting with the centriole and Golgi,followed by the apicoplast, the nucleus, andendoplasmic reticulum, and finishing with themitochondrion and de novo synthesis of themicronemes and rhoptries (21). To define the pointof cell-cycle arrest after ddROM4 S-A stabilization,we scrutinized the integrity and inheritance ofvarious organelles. Though the inner membranecomplex and apicoplast appeared normal, the mitochondrion,micronemes, rhoptries, and nuclei weredefective in arrested parasites (Fig. 2, fig. S3, andtables S2 and S3) in a phenotype characteristic ofan arrest late in the cell cycle at the S phase (22).ddROM4 S-A stabilization also reduced the amountof vacuoles containing developing daughter cells(Fig. 2B and table S2).Because ROM4 plays an important thoughnonessential role during invasion (15), we investigatedthe effect of ddROM4 S-A stabilization onmotility and invasion. As prolonged Shld-1 treatmentof intracellular parasites prevented egress asa direct result of the block in replication, we stabilizedddROM4 S-A in nondividing extracellularparasites. Consistent with (15), there was a modestdefect in invasion and gliding (fig. S4).To determine whether the impairment in celldivision after ddROM4 S-A stabilization was dependenton an invasion-related event, we performedpulse-chase assays. Although ddROM4 S-Aparasites treated with Shld-1 for 12 hours beforeegress and for 6 hours postinvasion recovered andunderwent normal cell division (Fig. 3A), parasitestreated at 6 hours postinvasion were impaired(Fig. 3B). Thus, the defect is reversible andindependent of invasion, resulting from the noncleavageof one or more substrates at the intracellularparasite surface, whose processing isrequired to trigger or maintain replication.All T. gondii rhomboid substrates characterizedto date are micronemal proteins. Amongthese, AMA1 is unique in that it functions exclusivelyduring invasion (2), it is detected oninvading parasites (3), and its C-terminal tail isdetectable in the newly invaded (6, 7, 9) parasites.We hypothesized that the AMA1 cytosolictailgeneratedbyROM4cleavageduringinvasiontriggers parasite replication and that stimulationof each replicative cycle is engendered byfurther cleavage of AMA1 or another substrate.To test this hypothesis, we determined whetherthe block in division could be reverted by expressingthe T. gondii (ddAMA1) or P. falciparum(ddPfAMA1) AMA1 tail in either the WT formor mutated in residues important for the AMA1invasion-related function (23–25). We substitutedwith Ala either the conserved Phe and Trp (23)(FW) (26)residuesinTgAMA1(ddAMA1 FW-AA )or the Ser residue phosphorylated in PfAMA1 (25)www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 473

REPORTS(ddPfAMA1 S-A ) (Fig. 4A). Expression of alltransgenes was Shld-1–regulated (fig. S5) and didnot affect the growth of WT or ddROM4 parasites(fig. S6), but it was able to rescue thereplication phenotype of ddROM4 S-A parasites(Fig. 4, B and C, and fig. S7), indicating independent,dual functions for AMA1 in replicationand invasion. To identify residues involvedin the replication function of AMA1, we testedddAMA1 constructs carrying an Ala replacementat the conserved C-terminal Tyr residue(ddAMA1 DY-AA ) or with the most N-terminal region(ddAMA1 535-570 ) or the 20 most C-terminalresidues (ddAMA1 504-549 ) deleted (Fig. 4A).None of the mutants impaired function (Fig. 4,B and C, and fig. S7), suggesting that the conservedcentral region PSDLMQEAEPS is importantfor function. To assess the specificity of theseresults, we expressed the MIC2 cleavage product(ddMIC2) and verified that it did not affect parasitegrowth (fig. S6) or complement the ddROM4 S-A -mediated arrest (Fig. 4, B and C).Two previous studies found no evidence for arole of ROM4 or AMA1 in replication (2, 15).Reexamination of the phenotype of parasites conditionallydepleted for AMA1 (2) found that theywere modestly affected in division (fig. S8), con-Fig. 1. Expression of ddROM4 S-A severely impairs intracellular growth.Stabilization of ddROM4, ddROM4 S-A , and ddROM1 S-A expression on parasitestreated 12 hours T Shld-1, as shown by indirect immunofluorescence assay [(A),green] or Western-blot [(B), top panels] with a-myc antibodies. The additionallower form of ddROM1 S-A (B) probably corresponds to a degradation product or otherwisemodified form (13). Gliding-associated protein 45 (a-GAP45) labels the innermembranecomplex [(A), red]. Scale bars in (A): 5 mm. ROM4 N-terminal (a-ROM4)antibodies were used to compare the expression level of ddROM4/ddROM4 S-A toendogenous ROM4 [(B), bottom panel]. The surface marker surface antigen 1 (aSAG1)served as loading control [(B), top panels]. (C) Plaque assays of RH-2YFP, ddROM4,ddROM4 S-A ,andddROM1 S-A parasites grown 7 days T Shld-1. The assays wereperformedsimultaneouslyforthesameparasitestrainT Shld-1. (D) Replicationprofile of RH-2FP, ddROM1 S-A , ddROM4,andddROM4 S-A , parasites pretreated12 hours T Shld-1 before host cell egress and for the time of the assay in a total of36 hours. We counted the number of parasites per vacuole 24 hours after host cellinoculation. Asterisks indicate statistically significant results (**P = 0.004; ***P =0.0005), as determined by the Student’s t test. Data are represented as mean T SD(error bars) of four independent experiments.47428 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTSFig. 2. ddROM4 S-A parasites arrest late in the cell cycle.Indirect immunofluorescence assay of ddROM4 S-A parasitesafter 24 hours T Shld-1 showing the mitochondrion (a-F1adenosine triphosphatase b subunit, green) broken down(A); replication arrested after a single round of division,as determined by staining of the nascent apical cones ofthe mother and daughter parasites (a-ISP1, green) onlyin nontreated parasites [(B), arrowheads]; and defectivekaryocytokinesis with the nuclei enlarged and uncondensed[(C), arrowheads]. In red, parasites are labeledwith a-GAP45. Longitudinal section through two daughterscells after 18 hours plus (D)orminus(E) Shld-1 and after36 hours plus Shld-1 (F). (G) Section through a nontreatedvacuole at 36 hours showing three cycles of endodyogeny.(H) Detail of a vacuole 18 hours plus Shld-1, showing themitochondrion running between the posterior pole andthe residual body, which is left after the budding ofdaughter cells from the mother parasite. (I) Longitudinalsection through one of two daughters at 18 hours plusShld-1, showing the elongated and lobed appearance ofthe nucleus. N, nucleus; DG, dense granule; R, rhoptry;C,conoid;MN,micronemes;RB,residualbody;Mi,mitochondrion;PP, posterior pole. Scale bars: (A) to (C), 5 mm;(D) to (G), 1 mm; (H) and (I), 0.5 mm.Fig. 3. The dominant negative effect of ddROM4 S-A is reversible andindependent of invasion. Intracellular ddROM4 and ddROM4 S-A parasiteswere treated 12 hours with Shld-1 before host cell egress andShld-1 was then removed 6 hours postinvasion [(A), blue bars], or parasiteswere allowed to invade and Shld-1 was added 6 hours postinvasionand was maintained for the duration of the assay in a total of 18 hours[(B), blue bars]. Replication was compared to either nontreated parasites(gray bars) or parasites pretreated 12 hours with Shld-1 beforehost cell egress and for the time of the assay in a total of 36 hours (blackbars). We counted the number of parasites per vacuole 24 hours afterinvasion. Asterisks indicate statistically significant results (*P =0.01;**P =0.004), as determined with the Student’s t test. Data are represented asmean T SD (error bars) of four independent experiments.www.sciencemag.org SCIENCE VOL 331 28 JANUARY 2011 475

REPORTSFig. 4. Expression ofddAMA1, ddAMA1 FW-AA ,and ddPfAMA1 transcomplementsthe negativeeffect of ddROM4 S-A on replication.(A) Schematic ofthe T. gondii (ddAMA1,ddAMA1 FW-AA ,ddAMA1 DY-AA ,ddAMA1 535-570 , ddAMA1 504-549 ,and ddMIC2) and P. falciparum(ddPfAMA1 and ddPfAMA1 S-A )fusion proteins used in thisstudy. The sequence of thecloned tails is shown. Mutatedresidues are boxed in gray;residues conserved in the T.gondii and P. falciparum AMA1tail are boxed in blue. (B)Plaque assays of ddROM4 S-Aparasites expressing ddAMA1,ddAMA1 FW-AA , ddAMA1 504-549 ,ddPfAMA1, or ddMIC2. Parasiteswere incubated 7 days TShld-1 before fixation andGiemsa staining. The assayswere only performed simultaneouslyfor the same parasitestrain T Shld-1. (C) Intracellularreplication assays of parasitesexpressing ddROM4 S-Aalone or in combination withddAMA1, ddAMA1 FW-AA ,ddMIC2, ddAMA1 504-549 ,orddPfAMA1. Parasites weretreated 24hours T Shld-1beforefixation. Data are representedas mean T SD (errorbars) of three independentexperiments.firming a role for AMA1 in replication. Failure toobserve a replication defect in (15) may havebeen the result of an inherent limitation of thesystem [see supporting online material (SOM)].The apicomplexan life cycle consists ofconsecutive transmissive and replicative phases.Premature differentiation into replicative formswould be potentially lethal, so commitment tocell division needs to be tightly regulated in timeand space. We show here that a mechanism ofregulated intramembrane proteolysis (27) actingon AMA1 is implicated in a signaling pathwayleading to replication (see SOM). Our study highlightsa role in this process for one of the mostconserved apicomplexan proteins and shows thatthis group of parasites has opted to use invasionmolecules to couple invasion with replicativegrowth.References and Notes1. D. Soldati-Favre, Parasite 15, 197 (2008).2. J. Mital, M. Meissner, D. Soldati, G. E. Ward, Mol. Biol.Cell 16, 4341 (2005).3. V. B. Carruthers, O. K. Giddings, L. D. Sibley,Cell. Microbiol. 1, 225 (1999).4. C. Opitz et al., EMBO J. 21, 1577 (2002).5. V. B. Carruthers, G. D. Sherman, L. D. Sibley,J. Biol. Chem. 275, 14346 (2000).6. S. A. Howell et al., Mol. Microbiol. 57, 1342 (2005).7. C. G. Donahue, V. B. Carruthers, S. D. Gilk, G. E. Ward,Mol. Biochem. Parasitol. 111, 15 (2000).8. P. K. Harris et al., PLoS Pathog. 1, e29 (2005).9. S. A. Howell et al., J. Biol. Chem. 278, 23890 (2003).10. T. J. Dowse, D. Soldati, Trends Parasitol. 21, 254(2005).11. F. Brossier, G. L. Starnes, W. L. Beatty, L. D. Sibley,Eukaryot. Cell 7, 664 (2008).12. P. Srinivasan, I. Coppens, M. Jacobs-Lorena, PLoS Pathog.5, e1000262 (2009).13. T. J. Dowse, J. C. Pascall, K. D. Brown, D. Soldati, Int. J.Parasitol. 35, 747 (2005).14. F. Brossier, T. J. Jewett, L. D. Sibley, S. Urban, Proc. Natl.Acad. Sci. U.S.A. 102, 4146 (2005).15. J. S. Buguliskis, F. Brossier, J. Shuman, L. D. Sibley,PLoS Pathog. 6, e1000858 (2010).16. R. A. O’Donnell et al., J. Cell Biol. 174, 1023 (2006).17. R. P. Baker, R. Wijetilaka, S. Urban, PLoS Pathog. 2,e113 (2006).18. S. Urban, M. Freeman, Mol. Cell 11, 1425 (2003).19. A. Herm-Götz et al., Nat. Methods 4, 1003 (2007).20. Material and methods are available as supportingmaterial on Science Online.21. M. Nishi, K. Hu, J. M. Murray, D. S. Roos, J. Cell Sci. 121,1559 (2008).22. M. J. Gubbels et al., PLoS Pathog. 4, e36 (2008).47628 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

REPORTStimes. As predicted, infants attended longer tothe Unexpected Outcome, in which a big agentbowed and yielded to a small one [mean (M) =20.0 T 4.2 (SEM)], than vice versa [M = 12.0 T 3.2(SEM); F (1,14) = 9.4, P = 0.008; partial ratio ofvariance accounted for (h 2 ) = 0.40; table S1] (31).A second experiment explored the developmentalcourse of coming to expect that a small,novel agent will yield to a larger one in theirfirst right-of-way conflict, using stimuli identicalto those in experiment 1. Participants were8-month-old (n = 16), 9-month-old (n =16),10-month-old (n = 16), and 12- to 13-monthold(n =16)infants(31).A repeated-measures ANOVA with test trials(expected or unexpected) varied within subjectsand presentation order (expected first or unexpectedfirst) and age category (8, 9, 10, and 12to 13 months) varied between subjects again revealedthat infants’ attention was drawn especiallyto those trials in which the larger agentmade way for the smaller one [M =17.2 T 1.6(SEM)], rather than the reverse [M =11.8T 1.2(SEM); F (1, 56) = 11.40, P=0.001, partial h 2 =0.17]. This effect interacted with participant age(F (3, 56) = 3.14, P=0.032, partial h 2 = 0.14).Planned follow-up analyses (with one-tailed testsof the unidirectional hypothesis that the main effectsof experiments 1 and 2 should replicate withineach age group of the present experiment) demonstratedthat 8-month-olds failed [M BigYields =6.0 T 1.0 (SEM), M SmallYields =8.2T 2.5 (SEM);paired-samples t (15) = –1.15, P (one-tailed) =0.13],and9-month-olds marginally succeeded [M BigYields =11.9 T 3.0 (SEM), M SmallYields =6.7T 1.9 (SEM),paired-samples t (15) =1.54,P (one-tailed) =0.073],indifferentiating between unexpected and expectedtest trials, whereas 10-month-olds [M BigYields =21.7 T 4.1 (SEM), M SmallYields =14.4T 2.6 (SEM);paired-samples t (15) =1.78,P (one-tailed) = 0.048] and12- to 13-month-olds [M BigYields = 29.0 T 3.9(SEM), M SmallYields = 18.0 T 2.9 (SEM), pairedsamplest (15) = 3.49, P (one-tailed) = 0.0015] wereable to do so (Fig. 1 and table S1) (31).Replicating the results of experiment 1,which used children 11 months and older, experiment2 thus demonstrates that the differenceof attention when a large agent bows andyields to a smaller agent rather than vice versadevelops between 8 and 10 months.It is possible that the results of experiments1 and 2 reflect merely that there is more mass inmotion, and thus perhaps more salient motioncatching the infants’ attention, when the largerrather than the smaller agent bows and scoots awayfor the other. As an initial control for this possibility,a third Isolated Motion Control experiment (fig.S2) exposed 11- to 16-month-olds (n = 16) to thesame familiarization events used in experiments1 and 2 (movies S1 and S2), but without conflictinggoals in intertrials and test trials. Instead, eitherthe large or small agent was alone on stage, performingthe identical motions as in the ConflictingGoal trials. Two intertrials (movies S6 and S7)showed infants the actions of each agent immediatelybefore it bowed and scooted back on subsequenttest trials (movies S8 and S9), while notimplying conflicting goals, because each agent wasalone on the stage (31). Hence, if the results of theConflicting Goals experiments had been driven bydifferences in the total mass in motion across theunexpected and expected test trials, then infantsshould also differentiate the Isolated Motion testtrials. Conversely, if the results so far reflect infants’use of relative size as a cue for dominancerank predicting the outcome of a conflict of goals,then infants should not differentiate the currentIsolated Motion trials.Here, a repeated-measures ANOVA with thetest factor (expected or unexpected) varied withinsubjects and presentation order varied betweensubjects revealed that 11- to 16-month-olds nolonger differentiated between a large [M =6.9T2.2 (SEM)] and small [M =6.3T 2.0 (SEM)]agent bowing and scooting away [F (1,14) = 0.089,ns (not significant); table S1] (31). Pooling thesedata with those of experiment 1, the 11- to 16-month-olds in these two experiments looked15.0010.005.000.00-5.0012.009.006.003.000.008 9 10 12 & 13Age in MonthsConflicting Goals Isolated Motion ControlExperimentequally long at the identical familiarization events,ruling out preexisting group differences in attentiveness.Crucially, an interaction confirmed thatinfants’ differentiation of unexpected and expectedtest trials varied significantly across experimentsin such a way that the size of the bowing agentmattered in the Conflicting Goals experiment butnot in the Isolated Motion Control experiment(Test X Experiment: F (1,28) = 4.75, P=0.038, partialh 2 = 0.15) (Fig. 2) (31).Still, it is possible that infants are only sensitiveto the greater mass in motion when the largeblock prostrates itself if cues to relative size arealso present. This was not the case in the IsolatedMotion Control experiment, because infants neverwatched the two agents together. Thus, low-levelperceptual factors might still have driven the resultspresented so far. Experiment 4 (Motion BehindControl, fig. S3) addressed this possibilityand explored another alternative interpretation ofthe results of experiments 1 and 2: that young infantsmerely expect that smaller agents are morelikely to fall over than are bigger ones.Fig. 1. Developmental onsetoftheConflictingGoalseffect (experiment 2). Continuedmean differential lookingtimes (TSEM) at unexpectedover expected test trials by categoricalage, once the animationshad frozen to stills (after19.1 s), are shown. N=64infantparticipants (8-, 9-, 10-, and 12-to 13-month-olds). #P (one-tailed) =0.073, *P (one-tailed) = 0.048, **P=0.032, ***P (one-tailed) = 0.0015.Fig. 2. Conflicting Goals (experiment1) versus IsolatedMotion Control (experiment 3).Continued mean differentiallooking times (TSEM) at unexpectedover expected testtrials by experimental condition,once the animations hadfrozen to stills (after 19.1 and13.2 s, respectively), are shown.N = 32 infant participants (11-to 16-month-olds). *P

REPORTScontext (Fig. 4) (31). This rules out the possibilitythat the greater complexity of the occlusion eventsin the unexpected Conflicting Goals test trialsdrove the infants’ interest to them.Together, the five experiments presented hereshow that preverbal infants use relative size topredict which of two novel agents has the right ofway. This effect is not driven by increased saliencyof the greater area in motion when thelarge agent bows; nor by expectations that small,rather than large, agents or objects fall over; norby differential attention to partial versus full occlusionwhen the agents pass each other. In additionto the results from the three quite differentcontrol conditions, the fact that 8-month-olds fail,9-month-olds marginally succeed, and 10- and12- to 13-month-olds robustly succeed at differentiatingthe Conflicting Goals outcomes speaksfurther against low-level perceptual saliency accountsthat would predict no such age differences.One final detail of our data provides additionalsupport for the interpretation that our findingsin experiments 1 and 2 depend on the representationsof dominance ranks for predicting theoutcomes of conflicting goals. What the qualitativelydifferent control experiments have in commonis that they do not establish conflictinggoals, and infants were far less interested in all ofthem than in the Conflicting Goals test trials (31).Other lean interpretations that make no referenceto goal-directed agents and social dominancecannot account for our data. The animationswere designed so that they would not invokemechanistic physics. One agent did not knockover the other one; instead the two agents movedback from each other and paused, before one ofthem lay down forwardly, opposite to whatbilliard-ball causality would entail. Finally, a leanaccount might argue that infants’ attention couldsimply be drawn to relative (nonsocial) size difference.However, relative size difference is maximizedin the expected Conflicting Goals testtrials (fig. S1 and movies S4 and S5), whereas wefound that infants attend more to the unexpectedtrials, and infants did not differentiate the exactsame relative size differences in the Motion Behindtest trials (fig. S3 and movies S16 and S17).Social relations are irreducible to features ofindividual agents. Although a social interactiongives evidence that intentional agents arepresent, the presence of several intentional agentsalone does not predict the kind of social relationshipthey have with one another. Previous researchhas documented that young infants representinteractions among others in terms of affiliativeand cooperative relations (13–16). The presentstudies show that preverbal infants also representdominance between competing agents. Moreover,consistent with existing research on infants’representations of intentional agency (7–12), ourevidence suggests that this conceptual understandingdevelops between 8 and 10 monthsof age.Nine- and 10-month-old infants are too youngto have actively participated in dominancefights, and American infants are unlikely to havewatched small agents bow and prostrate in subordinationto others of more formidable physicalsize, such as their parents. They may have experiencedolder siblings taking their toys orobserved older siblings struggle with each other,and learned that the bigger one often gets his orher way over the little one. If infants use theseexperiences to predict what will happen in ouranimations, they must see the similarity betweenprevious experiences and the present animations,even though the spatiotemporal parameters of theagents and their physical movements do notexactly match those they have experienced. Hence,representations of conflicting goals and socialdominance, rather than spatiotemporal primitives,must underpin the relevant similarity metricthat allows infants to apply their experiences tothe situations depicted in our animations.We cannot know, at present, what aspects ofour stimuli led the children to encode the familiarizationand intertrial events in experiments1 and 2 in terms of conflicting goals between twoagents. It is likely that the presence of face-likefeatures contributed to infants’ categorizationof the blue and green blocks as agents (13, 14).But equifinality of motion is also sufficient toestablish goals for novel agents with no facialfeatures (9). It is also uncertain whether infantswould have extrapolated that the agents’ goalsconflict had the agents not bumped into eachother when physically blocking each other’sway (indeed, the infants might have thought thatthe agents could pass or jump over each other sothat both could complete their paths to the endof the stage). These results also leave open justwhat about the unexpected conflicting goals eventis unexpected. Infants’ attention may be drawnwhen the smaller agent prevails, or when thelarger agent displays features of subordination(by prostrating itself), or both. Importantly, ourresults do not yet address which other conflictinggoals and features of dominance infants canrepresent, nor whether infants expect the dominancerelation between two agents to be stableacross different contexts of conflicting goals, evenwhen the physical dominance cue of relative sizeis absent.A crucial task for the developing child is tolearn the social structure of his or her world, inorder to interact appropriately with kin and nonkin,friend and foe, superiors, inferiors, and peers.Constraints on infants’ mental representationsof social relations may direct infants’ attentionto the relevant features among the myriad stimulipresent in any social scenario and assist them ininterpreting social interactions (3–6). Our findingthat preverbal infants mentally represent conflictinggoals and social dominance between twoagents suggests that just as infants possess earlydevelopingmechanisms for learning about thephysical world and the world of individual intentionalagents (3), they also have early-developingrepresentational resources tailored to understandingthe social world, allowing infants to understandand learn the dominance structures thatsurround them.References and Notes1. P. J. Richerson, R. Boyd, Not By Genes Alone(Univ. of Chicago Press, Chicago, 2006).2. J. B. Silk, Science 317, 1347 (2007).3. S. Carey, The Origin of Concepts (Oxford Univ. Press,Oxford, 2009).4. A. P. Fiske, Structures of Social Life (Free Press, New York,1991).5. A. P. Fiske, Pers. Soc. Psychol. Rev. 4, 76 (2000).6. A. P. Fiske, in Relational Models Theory, N. Haslam,Ed. (Erlbaum, Mahwah, NJ, 2004), pp 61–146.7. G. Gergely, Z. Nadasdy, G. Csibra, S. Biro, Cognition 56,165 (1995).8. G. Gergely, H. Bekkering, I. Kiraly, Nature 415, 755(2002).9. G. Csibra, S. Biro, O. Koos, G. Gergely, Cogn. Sci. 27,111 (2003).10. A. L. Woodward, Cognition 69, 1 (1998).11. R. Saxe, J. B. Tenenbaum, S. Carey, Psychol. Sci. 16, 995(2005).12. S. C. Johnson, Philos. Trans. R. Soc. London Ser. B 358,549 (2003).13. V. Kuhlmeier, K. Wynn, P. Bloom, Psychol. Sci. 14, 402(2003).14. K. Hamlin, K. Wynn, P. Bloom, Nature 450, 557 (2007).15. F. Warneken, M. Tomasello, Science 311, 1301 (2006).16. H. Over, M. Carpenter, Psychol. Sci. 20, 1189 (2009).17. J. Silk, Philos. Trans. R. Soc. London Ser. B 364, 3243(2009).18. J. Call, B. Hare, M. Carpenter, M. Tomasello, Dev. Sci. 7,488 (2004).19. W. Phillips, J. L. Barnes, N. Mahajan, M. Yamaguchi,L. R. Santos, Dev. Sci. 12, 938 (2009).20. A. P. Melis, B. Hare, M. Tomasello, Science 311, 1297(2006).21. F. Aureli, F. de Waal, Natural Conflict Resolution(Univ. of California Press, Berkeley, CA, 2000).22. C. Darwin, The Origin of Species [Penguin Classics,Berkeley, CA, 1982 (1859)].23. J. Sidanius, F. Pratto, Social Dominance (CambridgeUniv. Press, Cambridge, 1999).24. W. C. McGrew, An Ethological Study of Children’sBehavior (Academic Press, New York, 1972).25. A. M. Sluckin, P. K. Smith, Child Dev. 48, 917 (1977).26. J. H. Brown, B. A. Maurer, Nature 324, 248 (1986).27. P. Buston, Nature 424, 145 (2003).28. T. H. Clutton-Brock et al., Nature 444, 1065 (2006).29. T. Schubert, S. Waldzus, B. Seibt, in EmbodiedGrounding, E. Smith, Ed. (Cambridge Univ. Press,New York, 2008), pp. 160–185.30. G. Lakoff, M. Johnson, Metaphors We Live By(Univ. of Chicago Press, Chicago, 2003).31. See supporting material on Science Online.32. This research formed part of L.T.’s doctoral dissertation.It was supported by Harvard University and WinklerFoundation dissertation fellowships, together with apostdoctoral fellowship and an early-career award fromthe National Danish Science Foundation to L.T., and aNIH/NICHD grant (2 ROI HD038338) to S.C. We thankL. Schultz and the MIT Play Lab in the Boston Children’sMuseum for lending us their lab space; J. Cotton,M. Renno, and B. Walker-Meade for help in datacollection; S. Grum-Thomsen for help with graphics;and A. Fiske, J. Sidanius, T. Teasdale, and two anonymousreviewers for helpful feedback on an earlier draft.Supporting Online Materialwww.sciencemag.org/cgi/content/full/331/6016/477/DC1Materials and MethodsSOM TextFigs. S1 to S4Table S1ReferenceMovies S1 to S2518 October 2010; accepted 20 December 201010.1126/science.119919848028 JANUARY 2011 VOL 331 SCIENCE www.sciencemag.org

online @sciencecareers.orgPOSITIONS OPENTENURE-TRACK FACULTY POSITION inTeaching BiochemistryThe Department of Biochemistry and Molecular Biologyat The George Washington University Medical Centerinvites applications for a tenure-track ASSISTANT/ASSOCIATE PROFESSOR beginning in September2011.Basic qualifications: Applicants must have a Ph.D.degree in the life sciences (preferably in biochemistry)and have had extensive experience teaching biochemistryto medical students as well as to undergraduateand graduate (M.S.) students.Preferred qualifications: Preference will be given tocandidates with experience in teaching biochemistryand genetics to medical students involving didacticlectures, problem-based learning, and small group discussions.Candidates with this experience as well astraining and interests in a medically relevant field arepreferred. Strong communication and collaborativeskills and a commitment to innovative approaches toeducation are also desirable.Application process: Interested applicants mustsend complete curriculum vitae; a statement of teachingexperience (including evidence of teaching excellence),philosophy, and interests; and three formal letters of recommendationto: Ms. Kelly D. Crist, Executive Coordinator,Department of Biochemistry and MolecularBiology, Faculty Search, The George WashingtonUniversity Medical Center, Suite 530, 2300 EyeSt NW, Washington, DC 20037. Or, electronically toe-mail: bcmkdc@gwumc.edu.Review of applications by the Search Committee willbegin on March 15, 2011 and will continue until theposition has been filled. Only complete applications willbe considered.The George Washington University is an Affirmative Action/Equal Opportunity Employer.COURSE ANNOUNCEMENTEnvironmental Biomechanics, Physiology, andGenomics of Marine SpeciesMarine species are widely used as a model systemin community ecology, physiology, and genetics. Manyof the factors responsible for structuring these communitiesare abiotic variables such as wave exposure,temperature, wind speed, and light. The physical andbiological environment also sets the geographic scalefor dispersal, adaptation, and gene flow. The interactionbetween the physical environment and individualfitness or performance is especially topical in lightof climate change. This four-week summer course isdesigned to offer experimental ecologists and physiologiststheoretical and hands-on instruction in cuttingedgemethods in biomechanics, physiological molecularbiology, and genome-wide investigations of populationdifferentiation and adaptation. The relationships betweencomplex environmental mosaics and genome responsesinthecontextoffutureclimatechangewillbeamajortheme.Instructors: Drs. Mark Denny, Steve Palumbi andGeorge Somero. Dates: June 13 through July 8, 2011.Independent research following the course is possible.Location: Hopkins Marine Station, StanfordUniversity, Pacific Grove, CA 93950-3094. Foradditional information, including a course prospectusand instructions for application, see website: http://hms.stanford.edu/HMSweb/mech.htm or contactthe instructors (e-mails: mwdenny@stanford.edu,spalumbi@stanford.edu, or somero@stanford.edu).Deadline for receipt of all application materials isApril 1, 2011.HARVARD MEDICAL SCHOOLPOSTDOCTORAL POSITION available to studythe molecular and cellular mechanisms of Alzheimer_sand Parkinson_sdiseases(Nature 460:632, 2009; PNAS107:9879, 2010; J. Neurosci. 30:13066, 2010; Science330:1055, 2010). Recent Ph.D.s with strong backgroundin molecular biology, biochemistry, or electrophysiologyare encouraged to apply. Prior experience inslice physiology, calcium imaging, or mitochondrial analysisis preferred. Send curriculum vitae to Dr. Jie Shen(website: http://www.shenlab.net)ate-mail: jshen@rics.bwh.harvard.edu.POSITIONS OPENRADIOCHEMISTRY FACULTY POSITIONThe University of Nevada, Las Vegas (UNLV),Department of Chemistry, invites applications for atenure-track ASSISTANT, ASSOCIATE, or FULLPROFESSOR with emphasis on Radiochemistry. TheUNLV Radiochemistry program (website: http://radchem.nevada.edu) has a research focus on thechemistry of technetium and the actinides in solids,solutions, the environment, and chemical syntheses.Applicants must hold a Doctorate in chemistry, radiochemistry,or a closely related field. Applicants shouldhave demonstrated records of scholarly productivityand a track record in external funding, commensuratewith the position level. The position requires teachingundergraduate and graduate-level courses in bothchemistry and radiochemistry. In addition, mentoringgraduate and undergraduate students in an internationallycompetitive research environment is required. Thesuccessful applicant is expected to provide leadership inthe UNLV Radiochemistry program and the UNLVChemistry Department, commensurate with the positionlevel (Assistant, Associate, or Full Professor), utilizeand enhance existing facilities and laboratory equipmentin the UNLV Radiochemistry Program, interact closelywith other research and academic faculty in the program,and maintain a rigorous, externally funded researchprogram. For a complete position description and applicationdetails, please visit website: http://jobs.unlv.edu or call telephone: 702-895-2894 for assistance.Equal Employment Opportunity/Affirmative Action Employer.POSTDOCTORAL FELLOWSHIPSCedars-Sinai Medical CenterApplications are solicited for federally funded basic,clinical, and translational medicine research fellowshipsin Endocrinology, Diabetes, and Metabolism at Cedars-Sinai Medical Center.We are an academic medical center and major teachinghospital of the David Geffen School of Medicine atUCLA. We seek qualified applicants with M.D., Ph.D.,D.V.M., or D.D.S. degrees who seek postdoctoral trainingfrom a strong faculty of research scientists.Applicants must be U.S. citizens or possess permanent residencestatus.Underrepresented minorities and women are particularly encouragedto apply, specifically African Americans, Hispanic Americans,Native Americans, Alaskan Natives, and Pacific Islanders.Submit inquires to:Office of Academic AffairsCedars-Sinai Medical Center8700 Beverly BoulevardLos Angeles, CA 90048E-mail: billy.gellepis@cshs.orgWebsite: http://www.csmc.eduHelp employersfind you. Postyour resume/cv.www.ScienceCareers.org48228 JANUARY 2011 VOL 331 SCIENCE www.sciencecareers.org

online @sciencecareers.orgThe Medical Faculty at the Goethe-University Frankfurt in cooperation with theFoundation “Institute for Biomedical Research - Georg-Speyer-Haus - (GSH)”invites applications for the following position:Full Professorship (W3) in Cancer Biology(Succession Prof. Bernd Groner, PhD)The holder of the position will be Director of the Georg-Speyer-Haus in Frankfurtam Main. The Georg-Speyer-Haus is an academic research institute with a greattradition in biomedical research, located adjacent to the campus of the UniversityClinics of the Goethe-University in Frankfurt am Main. The Institute is seekingan exceptional candidate to lead the scientific activities of the Georg-Speyer-Haus,to supervise its organization and to serve as a member of the Medical Faculty ofthe Goethe-University, Frankfurt.The Georg-Speyer-Haus is an independent research foundation and employs astaff of about 100 members, currently organized into 9 research groups. Theresearch program is focussed on cancer biology and molecular and cellular therapy,and comprises basic and translational research in the areas of solid tumors as wellas hematological malignancies and inherited diseases. The individual researchgroups are working on aspects of stem cell biology and stem cell therapy, cells ofthe immune system in cancer therapy and immunodeficiency, and signal transductionin cancer cells. The Georg-Speyer-Haus is supported by the State of Hessen andthe Federal Ministry of Health with a stable significant budget complemented bycompetitive research grants from public and private funding organizations.The position provides a unique opportunity to lead an institute dedicated touncovering new knowledge in the area of cancer biology and the exploitation ofthese insights for clinical applications through innovative technologies in closecooperation with the Medical Faculty of the Goethe-University. The candidatemust be committed to scientific excellence, be able to integrate collaborativeapproaches across basic and clinical disciplines, design and implement successfulresearch strategies and lead a research group within the institute. A high degreeof dedication and experience in academic teaching are explicitly expected,especially in the planned course in Molecular Medicine.Candidates for this position must have an MD or a PhD degree, an internationallyrecognized reputation in basic and applied cancer research, senior level researchleadership experience and the ability to manage complex research programs andbudgets. In addition, superb communicative and organizational skills are importantprerequisites.The designated salary for the position is based on “W3” on the German universityscale or equivalent. Frankfurt University is an equal opportunity employer.Applications from or nominations of female candidates are encouraged. For furtherinformation regarding the general conditions for professorial appointments, pleasesee: http://www.uni-frankfurt.de/aktuelles/ausschreibung/professuren/index.htmlTo be considered for this position, please submit a curriculum vitae, a bibliography,a short description of the major achievements, a list of current funding and ascientific concept for future research in the institute. Applications should be sentwithin four weeks after publication of this advertisement to: The Dean, Facultyof Medicine of the Johann Wolfgang Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, E-Mail: dekan@kgu.de and (in electronicform) to: Dr. Rolf E. Breuer, Chairman of the Board of the Foundation“Georg-Speyer-Haus”, c/o Deutsche Bank, Taunusanlage 17, D-60325Frankfurt am Main, E-Mail: claudia.rode@db.comNontraditionalCareers:OpportunitiesAway Fromthe BenchWebinarWant to learn more about excitingand rewarding careers outside ofacademic/industrial research?View a roundtable discussion thatlooks at the various career optionsopen to scientists and strategiesyou can use to pursue anonresearch career.Now AvailableOn Demandwww.sciencecareers.org/webinarProduced by theScience/AAAS Business Office.

NORTHWEST A&F UNIVERSITY FACULTY POSITIONS in MULTIPLE DISCIPLINESonline @sciencecareers.orgThe Northwest A&F University (http://www.nwsuaf.edu.cn) invites applications for full-time faculty positionsin the following areas: biological sciences either cellular biology, molecular biology, ecology, microbiology,physiology, entomology or plant pathology; agricultural sciences either soil sciences, crop sciences, food sciences,animal science or veterinary medicine; engineering either mechanical engineering or computer engineering;and social sciences either agricultural and resource economics or sociology, etc. The successful candidates willhave a doctoral degree or equivalent degree and experience in academic research and teaching. Outstandingapplicants are encouraged to apply.The successful candidates will be provided with internationally competitive startup support, salary, benefits andhousing (apartments) according to qualifications and experience.Applicants should submit: (1) letter describing teaching and research experiences and career goals, (2) curriculumvitae, and (3) three letters of reference directly from the referees to: Talent Introduction Office, PersonnelDivision, Northwest A&F University, 3 Taicheng Rd., Yangling, Shaanxi 712100, P.R.China. Formal inquiries canbe directed to Mr Zhang Pengfei (or Ms Chen Xiaoyan): E-mail: rencaike@nwsuaf.edu.cn or rencaiban@gmail.com; Telephone: + 86-29-87082855. Interested persons can also visit our website: http://rcb.nwsuaf.edu.cn/.Consideration of applications will start immediately and the positions will remain open until suitable candidatesare selected.www.mssm.eduThe Mount Sinai School of Medicine is internationally acclaimed forexcellence in clinical care, education, and scientific research.Flow CytometryCore DirectorThe University of Connecticut Health Centerseeks candidates at the Ph.D. level with extensiveexperience with flow cytometryinstrumentation and applications to direct aheavily utilized and well equipped flow cytometrycore (http://flowcytometry.uchc.edu/). Thefacility currently houses a B-D FACSVantageSE, a FACSAria, three LSRII and twoFACSCaliburs. The director will be an expert inoperation of the equipment and will be expectedto direct the daily operations of the facility.He/she will independently design and implementnew cytometry protocols and assist userswith specialized applications. A faculty appointmentis associated with this position.Salary will be commensurate with experience.UCHC offers an excellent benefits package.Interested applicants should apply at88473,79--/.9, search number2011-622 and upload a curriculum vitae andthe names of three references through thewebsite. Questions regarding this searchshould be addressed to Leo Lefrancois atllefranc@neuron.uchc.eduUCHC is an Equal Opportunity EmployerM/F/V/PwDCHAIR, DEPARTMENT OF GENETICSAND GENOMIC SCIENCESMount Sinai School of Medicine, New York, NYWe seek a national and international leader to serve as Chair of this outstanding academicdepartment. Candidates must have an MD and/or PhD, a superior record of scientificachievement in genetics or genomics research, demonstrated commitment to education, andproven leadership ability. The Chair will have a unique opportunity to build upon an alreadyleading department.Mount Sinai School of Medicine and Mount Sinai Hospital are among the worlds leadingbiomedical institutions. The Medical Center is in the middle of a $1 billion capital campaignin support of our $2.25 billion strategic plan whose primary focus is on basic and translationalresearch leading to therapeutic discoveries and on the delivery of outstanding clinical care. Amajor new research building, the Center for Science and Medicine, focusing on translationalresearch is under construction with an expected occupancy mid-2012. Our leadership iscommitted to maintaining and growing this outstanding Genetics and Genomic Sciencesdepartment. The Medical Center will provide the exceptional resources necessary to recruit thehighest quality scientists.We offer a competitive salary and an excellent benefits package. Please send a letter ofapplication or nomination, with curriculum vitae, to: Dr.EricJ.Nestler,Director,TheFriedman Brain Institute, Mount Sinai School of Medicine, One Gustave Levy Place,Box 1065, New York, NY 10029.To receive full consideration, applications should arrive by March 11, 2011.Mount Sinai Medical Center is an equal opportunity/affirmative action employer. We recognize the power and importanceof a diverse employee population and strongly encourage applicants with various experiences and backgrounds.Mount Sinai Medical Center--An EEO/AA-D/V Employer

Postdoctoral Positions AvailableProgram in Genomics of DifferentiationNATIONAL INSTITUTE OF CHILD HEALTH AND HUMAN DEVELOPMENTFully funded postdoctoral positions are available in the NICHD Program in Genomics of Differentiation (PGD). The PGD is a diverse and highly interactive program in basic cellular, molecular,and developmental biology research at the National Institutes of Health in Bethesda, Maryland, just outside of Washington, DC. The 19 laboratories making up the PGD encompass a wide varietyof research areas, with strong emphasis on developmental patterning and differentiation, chromatin dynamics and epigenetics, the immune system, the viral life cycle, DNA replication, generegulation, and RNA metabolism. Program investigators perform research using a wide variety of models including viruses, bacteria, mammalian cell culture, yeast, fruit flies, zebrafish, frogs,and mice. Environment, resources, and stipend support are excellent- for additional information on the PGD and its researchers see http://science.nichd.nih.gov/confluence/display/pgd/Home.Interested applicants with a Ph.D. or M.D. and less than 5 years’ postdoctoral experience should send a CV, bibliography, cover letter with a brief description of research experience and interests,and the names of 3 references (with phone numbers) to individual investigators listed below:- Sohyun Ahn – Neurogenesis in the developing and mature nervous system of the mouse- Harold Burgess – Analysis of neural circuits controlling behavior in zebrafish- Mike Cashel – Molecular regulation of the prokaryotic stress response- Ajay Chitnis - Early neural and lateral line development, Notch signaling- David Clark – The role of chromatin structure in gene activation in yeast- Robert Crouch – RNase H function in health and disease- Igor Dawid – Early embryonic development and patterning in the frog and zebrafish- Melvin DePamphilis - Regulation of DNA replication in normal and in cancer cells in the mammal- Bruce Howard – Experimental and bioinformatic analysis of epigenome structure- Judith Kassis – Gene silencing by Polycomb group genes (PcG) in Drosophila- James Kennison – Drosophila developmental genetics- Judith Levin – HIV-1 replication; reverse transcription, host restriction, and viral assembly- Paul Love – Genes regulating mammalian hematopoiesis- Richard Maraia – RNA metabolism, the La protein and RNA Polymerase III- Keiko Ozato – Gene regulation in innate immunity- Karl Pfeifer – Epigenetic regulation of gene expression in the mouse; mouse cardiac development and function- Thomas Sargent – Epidermal and neural crest development in frog and fish- Brant Weinstein – Vascular development in the zebrafish- Heiner Westphal – Murine embryogenesis and induced pluripotent stem cellsDepartment of Health and Human ServicesNational Institutes of HealthMedical Informatics FellowshipsThe Lister Hill National Center for Biomedical Communication (LHNCBC) at the National Library of Medicine seeks postdoctoral fellows as well as graduate and medical students,who are interested in collaborative research within a variety of biomedical informatics areas, including- Capture, processing and analysis of clinical data for care and research applications- Biomedical and document image analysis- Development of health informatics resources.Successful applicants are matched with LHNCBC staff and participate directly in ongoing research. LHNCBC’s research activities include basic and applied research in fields suchas:- Tools and standards development for electronic medical records- Natural language processing for understanding medical text and improving information retrieval- Medical knowledge representation- Text mining- Multimedia database design- Interactive publications- Machine learning techniques- Image processing researchLHNCBC has a tradition of advancing health information systems and its world class research staff is involved in activities that define and support the research infrastructure fornext generation medical information systems.Postdoctoral candidates should have a Ph.D., MD/OD/DDS or equivalent degree in medical informatics, information science, computer science, engineering, applied mathematics,or related disciplines. Candidates should have research experience in these areas. Medical student rotation programs are available as well as programs for graduate students. Postdoctoral fellowships are in residence at LHNCBC in Bethesda, MD for one year with the possibility of renewal. Time in residence is variable for other awards, including visitingscholars, visiting faculty and graduate student candidates.Stipends are commensurate with research experience and education. The annual application deadlines are: January 15, March 15 and October 15. However, applications also areconsidered year round under special circumstances. For additional information and instructions to submit an application, please see our website: http://lhncbc.nlm.nih.gov. TheHHS and NIH are equal opportunity employers.

online @sciencecareers.orgWEBINARNow availableon demand.FACTS&FICTIONCareers in Industry and AcademiaTrying to figure out the next step in your career? Join us for a roundtablediscussion that will look at facts and fiction surrounding academic andindustry career options for PhD-level scientists. Get some nuts and boltsadvice on how to research career options, what questions to ask, andhow to best prepare for various careers.• Do industry and academic careers require different skill sets?• Do industry jobs have better compensation? Less autonomy?• Do academic scientists have less work/life balance?For answers view our roundtable discussion for free at:ScienceCareers.org/webinarProduced by the Science/AAAS Business Office.

ASSISTANT/ASSOCIATE PROFESSOR – CHEMISTRY – theDepartment of Chemistry at the University of Alabama at Birminghamseeks candidates for a tenure-track faculty position at the assistant orassociate professor level in CHEMISTRY with research expertise in areasof chemical or biophysical applications of NMR and/or mass spectroscopy.Candidates whose research is complementary to existing strengths withinthe Department and School will be given preference. The University ofAlabama at Birmingham (UAB) is a comprehensive research university andmedical center with over 2,000 full-time faculty and over 17,000 studentsfor Fall, 2010. UAB is ranked among the top tier research universities interms of federal grant support. The Department of Chemistry offers B.S.(ACS-Certified), M.S., and Ph.D. degrees and has major research thrustareas in drug discovery, structural biochemistry, biophysical chemistry,computational chemistry, and polymer/advanced materials. Applicationswill be considered beginning March 6, 2011. Applications past that datewill be considered until this position is filled. Candidates must have a Ph.D.degree in chemistry or biochemistry, postdoctoral or equivalent experience,and a commitment to teaching excellence at undergraduate and graduatelevels.Qualified applicants should send a letter indicating their interest, detailedcurriculum vitae, description of research plans, a statement on theirteaching experience and philosophy, and the names and contact informationof a minimum of four references. At least one reference should be ableto address your teaching potential, experience, and ability. Electronicsubmissions are encouraged and should be sent to Ms. Laura Knighten(knighten@uab.edu), or mailed to the Department of Chemistry,Faculty Search, University of Alabama at Birmingham, CHEM Suite201, 1530 3 rd Avenue South, Birmingham, AL 35294-1240.The Department of Chemistry and the University of Alabama at Birminghamare committed to building a culturally diverse workforce and stronglyencourage applications from women and individuals from underrepresentedgroups. UAB has a Dual Career Assistance Program to support and offerresources to help spouses and partners of newly recruited UAB faculty. UABis an Affirmative Action/Equal Employment Opportunity Employer.The Salk Institute for Biological Studies, a world classscientific environment located in La Jolla, CA, seeks candidatesfor faculty positions in the following areas:JUNIOR OR SENIOR FACULTY POSITION IN THE WAITTADVANCED BIOPHOTONICS CENTERWe are seeking a creative leader in biological imaging toestablish an independent research program in our WaittAdvanced Biophotonics Center. The goal of this position is thedevelopment and application of new tools for transforming thestudy of biological processes.TWO JUNIOR FACULTY POSITIONS IN NEUROSCIENCEThe successful candidates will be individuals with specificinterests and expertise in studying 1) neural circuits andbehavior using genetic approaches coupled with imagingand/or electrophysiology, and 2) synaptic physiology anddynamics. The Salk Institute maintains a world-renownprogram in the neurosciences with particular strengths inneuroendocrinology, vision, computational, developmental,behavioral and systems neuroscience.online @sciencecareers.orgLos Alamos National Laboratory, a premier national security researchinstitution, has a chemistry leadership opportunity for a chemical orphysical sciences research manager.RESEARCH & DEVELOPMENT MANAGER 6With your signicant record of advocacy and innovation, you will lead morethan 300 researchers and staff to technical excellence and operationaleffectiveness as division leader for LANLs long-standing Chemistry Division.In addition to senior-level management experience, the best candidatewill have an outstanding record in chemistry or physical science researchdemonstrated through publication, grantsmanship and large projects; theabilitytoleadinterdisciplinaryactivitiesinsuchareasasnanoscience,nontraditionalchemical processing, radiological and nuclear science, as well asadeptness at cost controls, ESH&Q, security and other operational matters.The selected applicant must have ability to obtain a DOE Q clearance,which typically requires U.S. Citizenship.For more information or to apply, go towww.lanl.gov/jobs and reference jobnumber 220659.Los Alamos supports a drug-freeworkplace and is an Equal OpportunityEmployer.www.lanl.gov/jobsThe Salk Institute is a collaborative research environment thatcovers a broad range of topics including molecular biology,cellular biology, plant biology, developmental biology,neurobiology, systems biology, structural biology, and biologicalchemistry. We seek interactive individuals who are passionatein their pursuit of important problems in basic science.The Salk Institute offers an excellent start-up package alongwith a competitive salary and benefits.Qualified candidates are invited to submit their curriculum vitae,description of present and future scientific endeavors, and alsoarrange to have three letters of recommendation sent directlyfrom the referees to: biophotonics_search@salk.edu (pleasereference job code: #A033) for the position in the WaittAdvanced Biophotonics Center ORneuroscience_search@salk.edu (please reference jobcode: #A034) for positions in Neuroscience.Applications with all required materials will be reviewedbeginning March 1, 2011 and will be accepted until positionsare filled. Faculty positions available at the Salk Institute maybe found at:http://www.salk.edu/careers/faculty_positions.html.The Salk Institute for Biological Studies is anEqual Opportunity Employer.

AAAS is here – connecting government to the scientific community.As a part of its efforts to introduce fully open government, the White House is reaching out to the scientific community for aconversation around America’s national scientific and technological priorities.To enable the White House’s dialogue with scientists, AAAS launched Expert Labs, under the direction of blogger and tech guruAnil Dash. Expert Labs is building online tools that allow government agencies to ask questions of the scientific community andthen sort and rank the answers they receive.On April 12, 2010, AAAS asked scientists everywhere to submit their ideas to the Obama administration and at the same timelaunched the first of Expert Labs tools, Think Tank, to help policy makers collect the subsequent responses. The result wasthousands of responses to the White House’s request, many of which are already under consideration by the Office of Scienceand Technology Policy.As a AAAS member, your dues support our efforts to help government base policy on direct feedback from the scientificcommunity. If you are not already a member, join us. Together we can make a difference.To learn more, visit aaas.org/plusyou/expertlabs

AAAS is here – promoting universal science literacy.In 1985, AAAS founded Project 2061 with the goal of helping all Americans become literate in science, mathematics, andtechnology. With its landmark publications Science for All Americans and Benchmarks for Science Literacy, Project 2061 set outrecommendations for what all students should know and be able to do in science, mathematics, and technology by the time theygraduate from high school. Today, many of the state standards in the United States have drawn their content from Project 2061.Every day Project 2061 staff use their expertise as teachers, researchers, and scientists to evaluate textbooks and assessments,create conceptual strand maps for educators, produce groundbreaking research and innovative books, CD-ROMs, and professionaldevelopment workshops for educators, all in the service of achieving our goal of universal science literacy.As a AAAS member, your dues help support Project 2061 as it works to improve science education. If you are not yet a AAASmember, join us. Together we can make a difference.To learn more, visit aaas.org/plusyou/project2061

online @sciencecareers.orgPOSITIONS OPENDARWIN FELLOWUniversity of Massachusetts AmherstThe Graduate Program in Organismic and EvolutionaryBiology (OEB) at University of MassachusettsAmherst announces a two-year postdoctoral fellowship/lectureship. OEB draws together more than 80 facultyfrom the Five Colleges (University of MassachusettsAmherst and Smith, Hampshire, Mount Holyoke, andAmherst Colleges), offering unique training and researchopportunities in the fields of ecology, organismic, andevolutionary biology. Our research/lecture position providesrecent Ph.D._s withanopportunityforindependentresearch with an OEB faculty sponsor as well asexperience developing and teaching a one-semesterundergraduate biology course. Proven teaching skillsare required. Position subject to availability of funds.First-year salary: $35,000. Second-year salary: $37,000.To apply, send curriculum vitae, three letters of reference,statements of research and teaching interests,and arrange for a letter of support from your proposedOEB faculty sponsor. A list of faculty and additionalinformation is available at website: http://www.bio.umass.edu/oeb.OEB Darwin Fellowship319 Morrill Science Center611 N. Pleasant StreetUniversity of Massachusetts AmherstAmherst, MA 01003Telephone: 413-545-0928E-mail: darwin@bio.umass.eduApplication review begins: March 4, 2011Start date: August 14, 2011The University of Massachusetts Amherst is an AffirmativeAction/Equal Opportunity Employer. Women and members ofminority groups are encouraged to apply.Findyour futurehere.↓www.ScienceCareers.org√Morescientists agree—we are the mostuseful website.www.ScienceCareers.orgWidelyRecognizedOriginal &GuaranteedUS Pat #5,436,149Call: Ab PeptidesFax: 314•968•8988MARKETPLACEKlenTaq18¢/uTruncatedTaq DNAPolymeraseWithstand 99 o Ce-mail: abpeps@msn.com1•800•383•3362www.abpeps.com49428 JANUARY 2011 VOL 331 SCIENCE www.sciencecareers.org

Antibodies and Related Reagents for Signal Transduction ResearchXP Monoclonal Antibodies,eXceptional Performance Unparalleled product quality, validation, and technical support.XP monoclonal antibodies are a line of high qualityrabbit monoclonal antibodies exclusively availablefrom Cell Signaling Technology. Any product labeledwith XP has been carefully selected based on superiorperformance in all approved applications.XP monoclonal antibodies are generated using XMTTechnology, a proprietary monoclonal method developedat Cell Signaling Technology. This technology providesaccess to a broad range of antibody-producing B cellsunattainable with traditional monoclonal technologies,allowing more comprehensive screening and theidentification of XP monoclonal antibodies.For additional information and a complete listof available XP Monoclonal Antibodies visit…www.cellsignal.comeXceptional specificityAs with all of our antibodies, the antibody is specific to your target ofinterest, saving you valuable time and resources.+eXceptional sensitivityThe antibody will provide a stronger signal for your target protein incells and tissues, allowing you to monitor expression of low levels ofendogenous proteins, saving you valuable materials.+eXceptional stability and reproducibilityXMT Technology combined with our stringent quality control ensuresmaximum lot-to-lot consistency and the most reproducible results.=eXceptional PerformanceXMT Technology coupled with our extensive antibody validation andstringent quality control delivers XP monoclonal antibodies witheXceptional Performance in the widest range of applications.Above: Confocal IF analysis of rat cerebellum using β3-Tubulin (D71G9) XP Rabbit mAb#5568 (green) and Neurofilament-L (DA2) Mouse mAb #2835 (red). Blue pseudocolor =DRAQ5 ® #4084 (fluorescent DNA dye).© 2011 Cell Signaling Technology, Inc. XMT, XP , eXceptional Performance, CST, and Cell Signaling Technology ®are trademarks of Cell Signaling Technology, Inc. / DRAQ5 ® is a registered trademark of Biostatus LimitedOrders (toll-free) 1-877-616-2355 | Technical support (toll-free) 1-877-678-8324 support@cellsignal.com | Inquiries info@cellsignal.com | Environmental Commitment eco.cellsignal.com