8 MARCH 2024, VOL 383 ISSUE 6687 Science

Citation preview

A French science colossus faces a reckoning p. 1046

Sundance film picks for scientists p. 1052

Using ultrasound to monitor tissue health pp. 1058 & 1096

$15 8 MARCH 2024 science.org

TROUBLE BELOW Deepwater sharks threatened by harvest p. 1135

Publish your research in the Science family of journals The Science family of journals (Science, Science Advances, Science Immunology, Science Robotics, Science Signaling, and Science Translational Medicine) are among the most highlyregarded journals in the world for quality and selectivity. Our peer-reviewed journals are committed to publishing cutting-edge research, incisive scientific commentary, and insights on what’s important to the scientific world at the highest standards.

Submit your research today! Learn more at Science.org/journals

QUALITY CONTENT FOR THE GLOBAL SCIENTIFIC COMMUNITY Multiple ways to stay informed on issues related to your research Sponsored Collection Booklets

Podcasts

Advertorials

Webinars

Posters

Scan the code and start exploring the latest advances in science and technology innovation! Science.org/custom-publishing

Brought to you by the Science/AAAS Custom Publishing Office.

Posters

Podcasts

Sponsored Collection Booklets

Advertorials

Webinars

CONTENTS 1052 8 M A R C H 2 0 24 • VO LU M E 3 8 3 • I S S U E 6 6 87

IN BRIEF

1036 News at a glance IN DEPTH

1038 U.S. giant telescopes imperiled by funding limit NSF faces choice between multibillion-dollar projects after board sets cost cap By D. Clery

1041 Gars truly are ‘living fossils,’ massive DNA data set shows The fish’s genomes change so slowly that species separated since the dinosaurs can produce fertile hybrids today By A. Heidt

1042 Brazil is hoping and waiting for a new vaccine as dengue rages A locally produced vaccine did well in a phase 3 clinical trial but won’t be available until at least 2025 By M. Triunfol

1043 Final spending bills offer gloomy outlook for science

1039 Surprise RNA paints colorful patterns on butterfly wings

Congress makes sizable cuts at key funding agencies By Science News Staff

Understudied means of regulating genes is likely widespread in butterflies— and perhaps other animals

1044 Skin side effects stymie advance of HIV vaccine

By E. Pennisi

Strategy of using multiple mRNA shots to hone powerful antibodies hits a pothole By J. Cohen

1040 Smithsonian urged to speed repatriation of human remains Task force says museum should return many of its 30,000 remains and seek descendants’ consent for research By R. Pérez Ortega

INSIGHTS BOOKS ET AL.

1052 Review roundup Science at Sundance 2024 PERSPECTIVES

1057 Two rings to rule them all A single photonic device accommodates three different modes of operation By A. Rolland and B. M. Heffernan RESEARCH ARTICLE p. 1080

1058 Monitoring homeostasis with ultrasound An implant could allow at-home monitoring of deep-tissue changes after surgery By S. N. Sharma and Y. Lee RESEARCH ARTICLE p. 1096

FEATURES

1059 Breathing control of vocalization

1046 The reckoning

A crucial brainstem circuit for vocalrespiratory coordination of the larynx is revealed By S. R. Hage

Didier Raoult and his institute found fame during the pandemic. Then, a group of dogged critics exposed major ethical failings By C. O’Grady b At the height of his fame, French microbiologist Didier Raoult inspired a nativity figurine.

RESEARCH ARTICLE p. 1074

1060 Amphibian hatchlings find mother’s milk Egg-laying amphibian females produce lipid-rich “milk” to feed offspring after hatching By M. H. Wake RESEARCH ARTICLE p. 1092

POLICY FORUM

1062 Accounting for the increasing benefits from scarce ecosystems As people get richer, and ecosystem services scarcer, policy-relevant estimates of ecosystem value must rise By M. A. Drupp et al. 1032

science.org SCIENCE

CREDITS: (ILLUSTRATION) HEDOF; (PHOTO) NICOLAS TUCAT/AFP VIA GETTY IMAGES

NEWS

LETTERS

1066 Reform wildlife trade in the European Union By P. Cardoso et al. 1066 Incorporate ethics into US public health plans By R. Anthony et al. 1067 Mangrove forest decline on Iran’s Gulf coast By H. Yarahmadi and Z. Khorsandi

RESEARCH IN BRIEF

1068 From Science and other journals REVIEW

1071 Neuroscience Structure, biophysics, and circuit function of a “giant” cortical presynaptic terminal D. Vandael and P. Jonas

1111 The bacterium Wolbachia blocks sperm development in the primary spermatocytes of its insect host by targeting a long noncoding RNA (shown in cyan in this fluorescence confocal image; nuclei are yellow).

REVIEW SUMMARY; FOR FULL TEXT: DOI.ORG/10.1126/SCIENCE.ADG6757

1092 Life history

1135 Conservation

Milk provisioning in oviparous caecilian amphibians P. L. Mailho-Fontana et al.

Fishing for oil and meat drives irreversible defaunation of deepwater sharks and rays

Vitamin A resolves lineage plasticity to orchestrate stem cell lineage choices

PERSPECTIVE p. 1060

B. Finucci et al.

M. T. Tierney et al.

1096 Biomedicine

1142 Quantum imaging

RESEARCH ARTICLE SUMMARY; FOR FULL TEXT: DOI.ORG/10.1126/SCIENCE.ADI7342

Bioresorbable shape-adaptive structures for ultrasonic monitoring of deep-tissue homeostasis J. Liu et al.

Adaptive optical imaging with entangled photons P. Cameron et al.

RESEARCH ARTICLES

1072 Adult stem cells

PODCAST

1073 Plant science Enhancing rice panicle branching and grain yield through tissue-specific brassinosteroid inhibition X. Zhang et al. RESEARCH ARTICLE SUMMARY; FOR FULL TEXT: DOI.ORG/10.1126/SCIENCE.ADK8838

1074 Neuroscience Brainstem control of vocalization and its coordination with respiration J. Park et al. RESEARCH ARTICLE SUMMARY; FOR FULL TEXT: DOI.ORG/10.1126/SCIENCE.ADI8081 PERSPECTIVE p. 1059

1075 Geology

PHOTO: RUPINDER KAUR/PENNSYLVANIA STATE UNIVERSITY

CO2 drawdown from weathering is maximized at moderate erosion rates A. Bufe et al.

PERSPECTIVE p. 1058

DEPARTMENTS

1104 HIV Induction of durable remission by dual immunotherapy in SHIV-infected ART-suppressed macaques S.-Y. Lim et al.

1111 Symbiosis Prophage proteins alter long noncoding RNA and DNA of developing sperm to induce a paternal-effect lethality R. Kaur et al.

1118 Attosecond science Attosecond-pump attosecond-probe x-ray spectroscopy of liquid water S. Li et al.

1080 Photonics

1122 Cell biology

Multimodality integrated microresonators using the Moiré speedup effect Q.-X. Ji et al.

Sister chromatid cohesion is mediated by individual cohesin complexes F. Ochs et al.

PERSPECTIVE p. 1057

1130 Paleoecology 1084 Neuroscience Axonal self-sorting without target guidance in Drosophila visual map formation E. Agi et al.

Climate change is an important predictor of extinction risk on macroevolutionary timescales C. M. Malanoski et al.

1035 Editorial Collections are truly priceless By C. C. Davis

1150 Working Life Writing my ticket By V. J. Rodriguez ON THE COVER Rough sharks (Oxynotidae) are a small family of deepwater sharks consisting of five species. Three species are threatened with extinction from overfishing. Their slow growth and few young, combined with an unusual diet of shark eggs, make this group of deepwater sharks susceptible to overfishing, which highlights the need to provide refuge from human activities. See page 1135. Photo: Jordi Chias/ NPL/Minden Pictures Science Staff ............................................1034 Science Careers ........................................1149

SCIENCE (ISSN 0036-8075) is published weekly on Friday, except last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue, NW, Washington, DC 20005. Periodicals mail postage (publication No. 484460) paid at Washington, DC, and additional mailing offices. Copyright © 2024 by the American Association for the Advancement of Science. The title SCIENCE is a registered trademark of the AAAS. Domestic individual membership, including subscription (12 months): $165 ($74 allocated to subscription). Domestic institutional subscription (51 issues): $2627; Foreign postage extra: Air assist delivery: $107. First class, airmail, student, and emeritus rates on request. Canadian rates with GST available upon request, GST #125488122. Publications Mail Agreement Number 1069624. Printed in the U.S.A. Change of address: Allow 4 weeks, giving old and new addresses and 8-digit account number. Postmaster: Send change of address to AAAS, P.O. Box 96178, Washington, DC 20090–6178. Single-copy sales: $15 each plus shipping and handling available from backissues.science.org; bulk rate on request. Authorization to reproduce material for internal or personal use under circumstances not falling within the fair use provisions of the Copyright Act can be obtained through the Copyright Clearance Center (CCC), www.copyright.com. The identification code for Science is 0036-8075. Science is indexed in the Reader’s Guide to Periodical Literature and in several specialized indexes.

SCIENCE science.org

8 MARCH 2024 • VOL 383 ISSUE 6687

1033

Editor-in-Chief Holden Thorp, [email protected]

BOARD OF REVIEWING EDITORS (Statistics board members indicated with S)

Executive Editor Valda Vinson Editor, Research Jake S. Yeston Editor, Insights Lisa D. Chong Managing Editor Lauren Kmec DEPUTY EDITORS Gemma Alderton (UK),Stella M. Hurtley (UK), Phillip D. Szuromi, Sacha Vignieri SR. EDITORS Caroline Ash (UK),

Michael A. Funk, Brent Grocholski, Di Jiang, Priscilla N. Kelly, Marc S. Lavine (Canada), Sarah Lempriere (UK), Mattia Maroso, Yevgeniya Nusinovich, Ian S. Osborne (UK), L. Bryan Ray, H. Jesse Smith, Keith T. Smith (UK), Jelena Stajic, Peter Stern (UK), Valerie B. Thompson, Brad Wible ASSOCIATE EDITORS Bianca Lopez, Sarah Ross (UK), Madeleine Seale (UK), Corinne Simonti, Yury V. Suleymanov, Ekeoma Uzogara SENIOR LETTERS EDITOR Jennifer Sills NEWSLETTER EDITOR Christie Wilcox RESEARCH & DATA ANALYST Jessica L. Slater LEAD CONTENT PRODUCTION EDITORS Chris Filiatreau, Harry Jach SR. CONTENT PRODUCTION EDITOR Amelia Beyna CONTENT PRODUCTION EDITORS Anne Abraham, Robert French, Julia Haber-Katris, Nida Masiulis, Abigail Shashikanth, Suzanne M. White SR. EDITORIAL MANAGERS Carolyn Kyle, Beverly Shields SR. PROGRAM ASSOCIATE Maryrose Madrid EDITORIAL ASSOCIATES Aneera Dobbins, Joi S. Granger, Lisa Johnson, Anita Wynn SR. EDITORIAL COORDINATORS Jeffrey Hearn, Alexander Kief, Ronmel Navas, Jerry Richardson, Alice Whaley (UK) EDITORIAL COORDINATORS Maura Byrne, Clair Goodhead (UK), Isabel Schnaidt, Qiyam Stewart, Brian White ADMINISTRATIVE COORDINATOR Karalee P. Rogers ASI DIRECTOR, OPERATIONS Janet Clements (UK) ASI OFFICE MANAGER Victoria Smith ASI SR. OFFICE ADMINISTRATORS Dawn Titheridge (UK), Jessica Waldock (UK) COMMUNICATIONS DIRECTOR Meagan Phelan DEPUTY DIRECTOR Matthew Wright SENIOR WRITER Walter Beckwith WRITERS Joseph Cariz, Abigail Eisenstadt, Nyla Husain SENIOR COMMUNICATIONS ASSOCIATE Zachary Graber COMMUNICATIONS ASSOCIATES Kiara Brooks, Haley Riley, Mackenzie Williams, Sarah Woods

News Editor Tim Appenzeller NEWS MANAGING EDITOR John Travis INTERNATIONAL EDITOR David Malakoff DEPUTY NEWS EDITORS Rachel Bernstein, Shraddha Chakradhar, Elizabeth Culotta, Martin Enserink, David Grimm, Eric Hand, Kelly Servick, Matt Warren (Europe) SR. CORRESPONDENTS Daniel Clery (UK), Jon Cohen, Jeffrey Mervis, Elizabeth Pennisi ASSOCIATE EDITORS Jeffrey Brainard, Michael Price NEWS REPORTERS Adrian Cho, Jennifer Couzin-Frankel, Phie Jacobs, Jocelyn Kaiser, Rodrigo Pérez Ortega (Mexico City), Robert F. Service, Erik Stokstad, Paul Voosen, Meredith Wadman CONTRIBUTING CORRESPONDENTS Warren Cornwall, Andrew Curry (Berlin), Ann Gibbons, Sam Kean, Kai Kupferschmidt (Berlin), Andrew Lawler, Mitch Leslie, Virginia Morell, Dennis Normile (Tokyo), Elisabeth Pain (Careers), Charles Piller, Gabriel Popkin, Joshua Sokol, Richard Stone, Emily Underwood, Gretchen Vogel (Berlin), Lizzie Wade (Mexico City) CAREERS Katie Langin (Associate Editor) COPY EDITORS Julia Cole (Senior Copy Editor), Hannah Knighton, Cyra Master (Copy Chief) ADMINISTRATIVE SUPPORT Meagan Weiland

Creative Director Beth Rakouskas DESIGN MANAGING EDITOR Chrystal Smith GRAPHICS MANAGING EDITOR Chris Bickel PHOTOGRAPHY MANAGING EDITOR Emily Petersen MULTIMEDIA MANAGING PRODUCER Kevin McLean WEB CONTENT STRATEGY MANAGER Kara Estelle-Powers DESIGN EDITOR Marcy Atarod DESIGNER Noelle Jessup SENIOR SCIENTIFIC ILLUSTRATOR Noelle Burgess SCIENTIFIC ILLUSTRATORS Austin Fisher, Kellie Holoski, Ashley Mastin SENIOR GRAPHICS EDITOR Monica Hersher GRAPHICS EDITOR Drew An-Pham SENIOR GRAPHICS SPECIALISTS Holly Bishop, Nathalie Cary SENIOR PHOTO EDITOR Charles Borst PHOTO EDITOR Elizabeth Billman SENIOR PODCAST PRODUCER Sarah Crespi SENIOR VIDEO PRODUCER Meagan Cantwell SOCIAL MEDIA STRATEGIST Jessica Hubbard SOCIAL MEDIA PRODUCER Sabrina Jenkins WEB DESIGNER Jennie Pajerowski

Chief Executive Officer and Executive Publisher Sudip Parikh Publisher, Science Family of Journals Bill Moran DIRECTOR, BUSINESS SYSTEMS AND FINANCIAL ANALYSIS Randy Yi DIRECTOR, BUSINESS OPERATIONS & ANALYSIS Eric Knott MANAGER, BUSINESS OPERATIONS Jessica Tierney MANAGER, BUSINESS ANALYSIS Cory Lipman MANAGER, WEB ANALYTICS Samantha Cressman BUSINESS ANALYSTS Kurt Ennis, Maggie Clark FINANCIAL ANALYST Isacco Fusi BUSINESS OPERATIONS ADMINISTRATOR Taylor Fisher SENIOR PRODUCTION MANAGER Jason Hillman SENIOR MANAGER, PUBLISHING AND CONTENT SYSTEMS Marcus Spiegler CONTENT OPERATIONS MANAGER Rebecca Doshi SENIOR CONTENT & PUBLISHING SYSTEMS SPECIALIST Jacob Hedrick SENIOR PRODUCTION SPECIALIST Kristin Wowk PRODUCTION SPECIALISTS Kelsey Cartelli, Audrey Diggs DIGITAL PRODUCTION MANAGER Lisa Stanford SENIOR DIGITAL ADVERTISING SPECIALIST Kimberley Oster ADVERTISING PRODUCTION OPERATIONS MANAGER Deborah Tompkins DESIGNER, CUSTOM PUBLISHING Jeremy Huntsinger SR. TRAFFIC ASSOCIATE Christine Hall SPECIAL PROJECTS ASSOCIATE Shantel Agnew ASSOCIATE DIRECTOR, BUSINESS DEVELOPMENT Justin Sawyers GLOBAL MARKETING MANAGER Allison Pritchard DIGITAL MARKETING MANAGER Aimee Aponte JOURNALS MARKETING MANAGER Shawana Arnold MARKETING ASSOCIATES Ashley Hylton, Mike Romano, Lorena Chirinos Rodriguez, Jenna Voris SENIOR DESIGNER Kim Huynh DIRECTOR AND SENIOR EDITOR, CUSTOM PUBLISHING Erika Gebel Berg ASSISTANT EDITOR, CUSTOM PUBLISHING Jackie Oberst PROJECT MANAGER Melissa Collins DIRECTOR, PRODUCT & PUBLISHING DEVELOPMENT Chris Reid DIRECTOR, BUSINESS STRATEGY AND PORTFOLIO MANAGEMENT Sarah Whalen DIRECTOR, PRODUCT MANAGEMENT Kris Bishop PRODUCT DEVELOPMENT MANAGER Scott Chernoff PUBLISHING PLATFORM MANAGER Jessica Loayza SR. PRODUCT ASSOCIATE Robert Koepke PRODUCT ASSOCIATES Caroline Breul, Anne Mason ASSOCIATE DIRECTOR, INSTITUTIONAL LICENSING MARKETING Kess Knight BUSINESS DEVELOPMENT MANAGER Rasmus Andersen ASSOCIATE DIRECTOR, INSTITUTIONAL LICENSING SALES Ryan Rexroth INSTITUTIONAL LICENSING MANAGER Marco Castellan, Claudia Paulsen-Young SENIOR MANAGER, INSTITUTIONAL LICENSING OPERATIONS Judy Lillibridge SENIOR OPERATIONS ANALYST Lana Guz SYSTEMS & OPERATIONS ANALYST Ben Teincuff FULFILLMENT ANALYST Aminta Reyes ASSOCIATE DIRECTOR, US ADVERTISING Stephanie O'Connor US MID WEST, MID ATLANTIC AND SOUTH EAST SALES Chris Hoag US WEST COAST SALES Lynne Stickrod ASSOCIATE DIRECTOR, ROW Roger Goncalves SALES REP, ROW Sarah Lelarge SALES ADMIN ASSISTANT, ROW Victoria Glasbey DIRECTOR OF GLOBAL COLLABORATION AND ACADEMIC PUBLISHING RELATIONS, ASIA Xiaoying Chu ASSOCIATE DIRECTOR, INTERNATIONAL COLLABORATION Grace Yao SALES MANAGER Danny Zhao MARKETING MANAGER Kilo Lan ASCA CORPORATION, JAPAN

Rie Rambelli (Tokyo), Miyuki Tani (Osaka) DIRECTOR, COPYRIGHT, LICENSING AND SPECIAL PROJECTS Emilie David RIGHTS AND PERMISSIONS ASSOCIATE Elizabeth Sandler LICENSING ASSOCIATE Virginia Warren RIGHTS AND LICENSING COORDINATOR Dana James CONTRACT SUPPORT SPECIALIST Michael Wheeler EDITORIAL

MEDIA CONTACTS

[email protected]

[email protected]

NEWS

PRODUCT ADVERTISING & CUSTOM PUBLISHING

[email protected] INFORMATION FOR AUTHORS

science.org/authors/ science-information-authors

science.org/subscriptions MEMBER BENEFITS

advertising.science.org/ aaas.org/membership/ products-services benefits [email protected] CLASSIFIED ADVERTISING

REPRINTS AND PERMISSIONS

MEMBERSHIP AND INDIVIDUAL SUBSCRIPTIONS

INSTITUTIONAL SALES AND SITE LICENSES

science.org/help/ reprints-and-permissions

advertising.science.org/ science.org/librarian science-careers [email protected] AAAS BOARD OF DIRECTORS

MULTIMEDIA CONTACTS

JOB POSTING CUSTOMER SERVICE

[email protected] [email protected]

employers.sciencecareers.org [email protected]

CHAIR Gilda A. Barabino PRESIDENT Keith R. Yamamoto PRESIDENT-ELECT Willie E. May

TREASURER Carolyn N. Ainslie CHIEF EXECUTIVE OFFICER

Sudip Parikh BOARD Cynthia M. Beall Janine Austin Clayton Kaye Husbands Fealing Kathleen Hall Jamieson Jane Maienschein Robert B. Millard Babak Parviz William D. Provine Juan S. Ramírez Lugo Susan M. Rosenberg Vassiliki Betty Smocovitis

Science serves as a forum for discussion of important issues related to the advancement of science by publishing material on which a consensus has been reached as well as including the presentation of minority or conflicting points of view. Accordingly, all articles published in Science—including editorials, news and comment, and book reviews—are signed and reflect the individual views of the authors and not official points of view adopted by AAAS or the institutions with which the authors are affiliated.

1034

8 MARCH 2024 • VOL 383 ISSUE 6687

Erin Adams, U. of Chicago Takuzo Aida, U. of Tokyo Leslie Aiello, Wenner-Gren Fdn. Deji Akinwande, UT Austin James Analytis, UC Berkeley Paola Arlotta, Harvard U. Delia Baldassarri, NYU Nenad Ban, ETH Zürich Christopher Barratt, U. of Dundee Franz Bauer, Pontificia U. Católica de Chile Ray H. Baughman, UT Dallas Carlo Beenakker, Leiden U. Yasmine Belkaid, NIAID, NIH Kiros T. Berhane, Columbia U. Joseph J. Berry, NREL Chris Bowler, École Normale Supérieure Ian Boyd, U. of St. Andrews Malcolm Brenner, Baylor Coll. of Med. Emily Brodsky, UC Santa Cruz Ron Brookmeyer, UCLA (S) Christian Büchel, UKE Hamburg Johannes Buchner, TUM Dennis Burton, Scripps Res. Carter Tribley Butts, UC Irvine György Buzsáki, NYU School of Med. Mariana Byndloss, Vanderbilt U. Med. Ctr. Annmarie Carlton, UC Irvine Simon Cauchemez, Inst. Pasteur Ling-Ling Chen, SIBCB, CAS Wendy Cho, UIUC Ib Chorkendorff, Denmark TU Chunaram Choudhary, Københavns U. Karlene Cimprich, Stanford U. Laura Colgin, UT Austin James J. Collins, MIT Robert Cook-Deegan, Arizona State U. Virginia Cornish, Columbia U. Carolyn Coyne, Duke U. Roberta Croce, VU Amsterdam Molly Crocket, Princeton U. Christina Curtis, Stanford U. Ismaila Dabo, Penn State U. Jeff L. Dangl, UNC Nicolas Dauphas, U. of Chicago Frans de Waal, Emory U. Claude Desplan, NYU Sandra DÍaz, U. Nacional de CÓrdoba Samuel Díaz-Muñoz, UC Davis Ulrike Diebold, TU Wien Stefanie Dimmeler, Goethe-U. Frankfurt Hong Ding, Inst. of Physics, CAS Dennis Discher, UPenn Jennifer A. Doudna, UC Berkeley Ruth Drdla-Schutting, Med. U. Vienna Raissa M. D'Souza, UC Davis Bruce Dunn, UCLA William Dunphy, Caltech Scott Edwards, Harvard U. Todd A. Ehlers, U. of Glasgow Nader Engheta, UPenn Tobias Erb, MPS, MPI Terrestrial Microbiology Karen Ersche, U. of Cambridge Beate Escher, UFZ & U. of Tübingen Barry Everitt, U. of Cambridge Vanessa Ezenwa, U. of Georgia Toren Finkel, U. of Pitt. Med. Ctr. Natascha Förster Schreiber, MPI Extraterrestrial Phys. Peter Fratzl, MPI Potsdam Elaine Fuchs, Rockefeller U. Caixia Gao, Inst. of Genetics and Developmental Bio., CAS Daniel Geschwind, UCLA Lindsey Gillson, U. of Cape Town Gillian Griffiths, U. of Cambridge Simon Greenhill, U. of Auckland Nicolas Gruber, ETH Zürich Hua Guo, U. of New Mexico Taekjip Ha, Johns Hopkins U.

Daniel Haber, Mass. General Hos. Sharon Hammes-Schiffer, Yale U. Wolf-Dietrich Hardt, ETH Zürich Kelley Harris, U. of Wash Carl-Philipp Heisenberg, IST Austria Christoph Hess, U. of Basel & U. of Cambridge Heather Hickman, NIAID, NIH Hans Hilgenkamp, U. of Twente Janneke Hille Ris Lambers, ETH Zürich Kai-Uwe Hinrichs, U. of Bremen Deirdre Hollingsworth, U. of Oxford Christina Hulbe, U. of Otago, New Zealand Randall Hulet, Rice U. Auke Ijspeert, EPFL Gwyneth Ingram, ENS Lyon Darrell Irvine, MIT Akiko Iwasaki, Yale U. Erich Jarvis, Rockefeller U. Peter Jonas, IST Austria Sheena Josselyn, U. of Toronto Matt Kaeberlein, U. of Wash. Daniel Kammen, UC Berkeley Kisuk Kang, Seoul Nat. U. V. Narry Kim, Seoul Nat. U. Nancy Knowlton, Smithsonian Etienne Koechlin, École Normale Supérieure Alex L. Kolodkin, Johns Hopkins U. LaShanda Korley, U. of Delaware Paul Kubes, U. of Calgary Chris Kuzawa, Northwestern U. Laura Lackner, Northwestern U. Gabriel Lander, Scripps Res. (S) Mitchell A. Lazar, UPenn Hedwig Lee, Duke U. Fei Li, Xi'an Jiaotong U. Ryan Lively, Georgia Tech Luis Liz-Marzán, CIC biomaGUNE Omar Lizardo, UCLA Jonathan Losos, WUSTL Ke Lu, Inst. of Metal Res., CAS Christian Lüscher, U. of Geneva Jean Lynch-Stieglitz, Georgia Tech David Lyons, U. of Edinburgh Fabienne Mackay, QIMR Berghofer Zeynep Madak-Erdogan, UIUC Vidya Madhavan, UIUC Anne Magurran, U. of St. Andrews Ari Pekka Mähönen, U. of Helsinki Asifa Majid, U. of Oxford Oscar Marín, King’s Coll. London Charles Marshall, UC Berkeley Christopher Marx, U. of Idaho David Masopust, U. of Minnesota Geraldine Masson, CNRS Jennifer McElwain, Trinity College Dublin Scott McIntosh, NCAR Rodrigo Medellín, U. Nacional Autónoma de México Mayank Mehta, UCLA C. Jessica Metcalf, Princeton U. Tom Misteli, NCI, NIH Jeffery Molkentin, Cincinnati Children's Hospital Medical Center Alison Motsinger-Reif, NIEHS, NIH (S) Danielle Navarro, U. of New South Wales Daniel Neumark, UC Berkeley Thi Hoang Duong Nguyen, MRC LMB Beatriz Noheda, U. of Groningen Helga Nowotny, Vienna Sci. & Tech. Fund Pilar Ossorio, U. of Wisconsin Andrew Oswald, U. of Warwick Isabella Pagano, Istituto Nazionale di Astrofisica Giovanni Parmigiani, Dana-Farber (S) Sergiu Pasca, Standford U. Ana Pêgo, U. do Porto Julie Pfeiffer, UT Southwestern Med. Ctr.

Philip Phillips, UIUC Matthieu Piel, Inst. Curie Kathrin Plath, UCLA Martin Plenio, Ulm U. Katherine Pollard, UCSF Elvira Poloczanska, Alfred-Wegener-Inst. Julia Pongratz, Ludwig Maximilians U. Philippe Poulin, CNRS Suzie Pun, U. of Wash Lei Stanley Qi, Stanford U. Simona Radutoiu, Aarhus U. Trevor Robbins, U. of Cambridge Joeri Rogelj, Imperial Coll. London John Rubenstein, SickKids Mike Ryan, UT Austin Miquel Salmeron, Lawrence Berkeley Nat. Lab Nitin Samarth, Penn State U. Erica Ollmann Saphire, La Jolla Inst. Joachim Saur, U. zu Köln Alexander Schier, Harvard U. Wolfram Schlenker, Columbia U. Susannah Scott, UC Santa Barbara Anuj Shah, U. of Chicago Vladimir Shalaev, Purdue U. Jie Shan, Cornell U. Beth Shapiro, UC Santa Cruz Jay Shendure, U. of Wash. Steve Sherwood, U. of New South Wales Brian Shoichet, UCSF Robert Siliciano, JHU School of Med. Lucia Sivilotti, UCL Emma Slack, ETH Zürich & U. of Oxford Richard Smith, UNC (S) John Speakman, U. of Aberdeen Allan C. Spradling, Carnegie Institution for Sci. V. S. Subrahmanian, Northwestern U. Sandip Sukhtankar, U. of Virginia Naomi Tague, UC Santa Barbara Eriko Takano, U. of Manchester A. Alec Talin, Sandia Natl. Labs Patrick Tan, Duke-NUS Med. School Sarah Teichmann, Wellcome Sanger Inst. Rocio Titiunik, Princeton U. Shubha Tole, Tata Inst. of Fundamental Res. Maria-Elena Torres Padilla, Helmholtz Zentrum München Kimani Toussaint, Brown U. Barbara Treutlein, ETH Zürich Li-Huei Tsai, MIT Jason Tylianakis, U. of Canterbury Matthew Vander Heiden, MIT Wim van der Putten, Netherlands Inst. of Ecology Ivo Vankelecom, KU Leuven Henrique Veiga-Fernandes, Champalimaud Fdn. Reinhilde Veugelers, KU Leuven Bert Vogelstein, Johns Hopkins U. Julia Von Blume, Yale School of Med. David Wallach, Weizmann Inst. Jane-Ling Wang, UC Davis (S) Jessica Ware, Amer. Mus. of Natural Hist. David Waxman, Fudan U. Alex Webb, U. of Cambridge Chris Wikle, U. of Missouri (S) Terrie Williams, UC Santa Cruz Ian A. Wilson, Scripps Res. (S) Hao Wu, Harvard U. Li Wu, Tsinghua U. Amir Yacoby, Harvard U. Benjamin Youngblood, St. Jude Yu Xie, Princeton U. Jan Zaanen, Leiden U. Kenneth Zaret, UPenn School of Med. Lidong Zhao, Beihang U. Bing Zhu, Inst. of Biophysics, CAS Xiaowei Zhuang, Harvard U. Maria Zuber, MIT

science.org SCIENCE

EDITO RIAL

Collections are truly priceless

L

ast month, Duke University in North Carolina announced that it was shuttering its herbarium. The collection consists of nearly 1 million specimens representing the most comprehensive and historic set of plants from the southeastern United States. It also includes extensive holdings from other regions of the world, especially Mexico, Central America, and the West Indies. Duke plans to disperse these samples to other institutions for use or storage over the next 2 to 3 years, but this decision reflects a lack of awareness by academia that such collections are being leveraged as never before. With modern technologies spanning multiple fields of study, the holdings in herbaria and other natural history collections are not only facilitating a deeper and broader understanding of the past and present world but are also providing tools to meet both known and unforeseen challenges facing humanity. Science and society can hardly risk the loss of such an important resource. Sadly, Duke is not the first worldclass institution to withdraw support from, and cease the operation of, its natural history collections. In the late 1970s and early 1980s, Princeton and Stanford Universities did the same. Ostensibly, the decisions to close those collections were made to shift priority to research programs in molecular biology and biochemistry, which were considered closer to science’s cutting edge of discovery and able to attract more external funding. Ironically, nearly half a century on, biological sciences departments at these institutions and comparable ones in China, Brazil, some regions in Africa, and in most of Western Europe are filled with world-class scholars who—knowingly or unknowingly—use herbaria, zoological collections, and their derivatives every day for transformative research published in the highestimpact journals. Herbaria have long been a critical resource for ecological and evolutionary research but have recently become relevant to many more fields, including climate science, anthropology, genetics, computer science, chemistry, and medicine. Specimens are being mobilized to investigate plant–animal and plant–pathogen interactions, crop domestication, compounds with potential applications in agriculture and pharmaceutics, and human migration over time and space. Advances in genome sequencing and machine learning are guiding

biodiversity monitoring efforts and revealing knowledge gaps where specimen sampling is needed. The decision by Duke comes at a time when widespread awareness of and access to herbaria are growing in tandem. This is principally a result of the large-scale digitization of natural history collections, an endeavor that has been extensively supported by governmental agencies and philanthropic organizations worldwide. This innovation is arguably one of the greatest transformations in biodiversity science since DNA sequencing. In short, creation of the Global Metaherbarium—an open-access, global interlinked virtual resource—makes physical herbaria discoverable and is attracting new interest in the utility of these collections for sophisticated multiomic investigations (genomics, transcriptomics, metabolomics, proteomics, and microbiomics) and for research that connects science with the broader society. Closure of the Duke Herbarium also points to changes needed in formally recognizing herbaria and other natural history collections in research initiatives and agendas. Collections increasingly have become the first line of genetic and genomic sampling for investigators who otherwise eschew conventional field work. Requests to destructively sample specimens are often central to rapidly expanding big data initiatives. These requests place enormous demands on the institutions and staff who support collections but who largely go unrecognized for their crucial work. In turn, users of these collections, many of whom are not based at these institutions, benefit from grants and high-profile papers in which herbaria are only briefly acknowledged, if they are mentioned at all. Scientists who oversee collections should be fully funded partners in research initiatives. Institutions, herbarium curators, and support staff should be coauthors of studies, with contributions indicated through the Contributor Roles Taxonomy (CRediT) system, for example. Such recognition could help more directly measure the impact and influence of natural history collections on scholarly research. Universities should support the priceless resources and heritage represented in natural history collections. They also should have the vision to provide for, and commit to, the long-term stewardship and robust intellectual environment for open inquiry and deep research that these collections provide across generations. –Charles C. Davis

PHOTO: KRIS SNIBBE/HARVARD UNIVERSITY

“…society can hardly risk the loss of such an important resource.”

Charles C. Davis is a professor in the Department of Organismic and Evolutionary Biology, and Curator of Vascular Plants, Harvard University Herbaria, Cambridge, MA, USA. cdavis@oeb. harvard.edu

10.1126/science.ado9732

SCIENCE science.org

8 MARCH 2024 • VOL 383 ISSUE 6687

1035

NEWS IN BRIEF

650%

Increase since 2009 in the share price of scientific publishing giant Elsevier’s parent company RELX and its predecessor. The stock is the top performer on the U.K.’s FTSE 100 index in its 40-year history. In 2023, RELX’s scientific division reaped a profit margin of 38%. (Financial Times, RELX annual report)

Edited by Jeffrey Brainard

U.S. deports Chinese students | An unusual town hall last week at Yale University highlighted a recent spate of incidents in which immigration authorities blocked Chinese graduate students from returning to U.S. universities after visiting family in China. More than a dozen students in Ph.D. science programs at Yale, John Hopkins University, and other major U.S. research institutions had their visas revoked and were immediately sent back home. U.S. Customs and Border Protection (CBP) declined to discuss specific cases. Immigration lawyers suspect the influence of a 2020 presidential directive that gives CBP agents the authority to deny entry to Chinese graduate students and postdocs who have received support from entities suspected of stealing U.S. technology. Yale’s graduate school of arts and sciences hosted the 26 February event for its international students, who make up nearly half the school’s enrollment.

SECURITY

A 1953 nuclear test in Nevada was among the human activities that could have marked the Anthropocene.

Methane satellite begins work | The Environmental Defense Fund (EDF) this week became the first nonprofit group to launch a satellite to track methane emission sources. MethaneSAT, funded by EDF donors, is designed to detect methane emissions in high resolution above known oil-and-gas facilities, filling a gap in coverage. Its data will support efforts to regulate and reduce leaks and other sources of the potent greenhouse gas. The group plans to provide the data for free, in nearly real time, at www.MethaneSAT.org.

C L I M AT E S C I E N C E

Anthropocene epoch gets voted down

A

group of two dozen geologists has turned down a proposal to classify the Anthropocene as an “epoch” that would mark humanity’s overwhelming influence on the planet, a tally released this week indicates. For 15 years, researchers had considered designating this formal unit of geologic time, and in 2023 they chose a marker of when it started, a layered sediment core from Canada’s Crawford Lake that shows a global acceleration in carbon dioxide emissions and atmospheric nuclear weapons testing during the 1950s. But over the past month, the proposal failed to win a supermajority of votes from a panel of the International Commission on Stratigraphy, with some members stating that the proposed start date failed to account for earlier human influences. Barring an unexpected reversal, the formal classification cannot be reconsidered for another decade. But even opponents of the proposal acknowledge humanity’s potent, transformative effects on Earth and the power of the term Anthropocene, and some suggest considering it, like some other great changes in the planet’s history, a geologic “event”—a usage that requires no formal ratification or exact start date. 1036

8 MARCH 2024 • VOL 383 ISSUE 6687

U.K. funder clears diversity panel | The United Kingdom’s national funding agency has reinstated its advisory panel on diversity, equity, and inclusion, which was suspended in October 2023 after science minister Michelle Donelan said members of the newly created panel had posted “extremist” views on social media about the Israel-Hamas conflict. This week, UK Research and Innovation (UKRI) reported the results of its investigation into the matter, concluding that the panel members had not violated a code

POLITICS

science.org SCIENCE

PHOTO: NNSA/NEVADA FIELD OFFICE/SCIENCE SOURCE

STRATIGRAPHY

NATURAL HISTORY

Scanning project creates huge digital menagerie

B

iologists have completed a free, online repository containing x-ray scans of vertebrate specimens from 16 museum collections across the United States. The openVertebrate collection, one of the largest of its kind, covers more than 13,000 specimens, including more than half the genera of amphibians, reptiles, fishes, and mammals. Led by the Florida Museum of Natural History, researchers spent 5 years making computer tomography scans and creating 3D reconstructions; most show only the animals’ skeletons, but some samples were

of conduct for public servants or posted problematic views. Although Donelan had asked UKRI to shut down the diversity panel, UKRI’s statement said the investigation concluded the panel’s work is necessary, and it will reconvene. Separately, a lawyer for a panel member, Heriot-Watt University gender studies professor Kate Sang, announced on 5 March that Donelan had agreed to pay Sang an undisclosed settlement and retract her “false” statement about Sang’s social media post.

the institute is funded only by nonstate sources, including its own endowment and the university’s foundation. A proposal floated earlier would have created a new nonprofit organization to fund and manage some of the institute’s administrative functions while allowing its faculty and collections to remain within the university. But some researchers worried the split would expose the institute to future legislative crackdowns, The Guardian reported.

Trustees protect Kinsey Institute

Finding new uses for drugs

| The Kinsey Institute, the famed research center on human sexuality, will remain part of Indiana University (IU), despite a 2023 state law that blocks the institute from receiving taxpayer dollars. Conservative lawmakers targeted the institute after one claimed its research promotes sexual abuse, an allegation Kinsey’s defenders call baseless. Last week, IU’s board of trustees voted unanimously to develop a plan ensuring

C L I N I CA L R E S E A R C H | A nonprofit that seeks to repurpose approved drugs for new indications will receive more than $48 million from the U.S. Advanced Research Projects Agency for Health to supercharge its work, the agency said on 28 February. Every Cure plans to use artificial intelligence to predict the power of more than 3000 approved drugs against more than 10,000 rare diseases, most without effective treatments. The

POLITICS

IMAGE: OPENVERTEBRATE

stained before being scanned to reveal internal organs. As of December 2023, the database had received more than 1 million views and nearly 100,000 downloads. The digital collection has already led to new research findings, including unusual bones in African spiny mice (pictured, with tail colored red) and evidence that frogs have lost and regained teeth more than 20 times during their evolution. Project organizers also trained secondary school teachers to use the images for science education. The project’s impact is described in the 6 March issue of BioScience.

SCIENCE science.org

Philadelphia-based nonprofit was cofounded by University of Pennsylvania immunologist David Fajgenbaum, who a decade ago identified a treatment— sirolimus, which prevents organ rejection—for his own rare, life-threatening immune condition, Castleman disease.

Pesticide database restored | The U.S. Geological Survey (USGS) has backtracked on cuts to a widely used database of approximately 400 agricultural pesticides after pleas from scientists. The agency had reduced the number of compounds tracked in 2019 by the Pesticide National Synthesis Project, which documents estimated annual application rates, from 400 to 72, citing budget constraints. Then last year, USGS halted the annual release of preliminary data, opting instead to publish final data every 5 years. Last week, the agency said it will restore the database’s pre-2019 scope, and data for 2018 to 2022 will be published in 2025.

AG R I C U LT U R E

8 MARCH 2024 • VOL 383 ISSUE 6687

1037

NE WS

The Thirty Meter Telescope (artist’s conception) in Hawaii is one of two projects seeking public funding.

IN DEP TH ASTRONOMY

U.S. giant telescopes imperiled by funding limit By Daniel Clery

F

or several years, U.S. astronomers have hoped the government would help build a pair of giant ground-based telescopes. But the National Science Board (NSB), the panel of scientists that oversees the National Science Foundation (NSF), says the field can only afford one. At a meeting on 22 February, NSB capped the budget of the U.S. Extremely Large Telescope Program (US-ELTP) at $1.6 billion and gave the agency until May to come up with a process to choose one of the two 30-meter class telescopes. With a rival European telescope rapidly taking shape on a mountaintop in Chile, the NSB decision is a relief to those who want U.S. astronomy to unite behind a realistic plan and catch up. “I think the decision was long overdue,” says John Monnier of the University of Michigan. But for Richard Ellis of University College London, “It’s a tragedy, given the investment made in both telescopes.” He adds, “There were many opportunities to merge or down select. Now, the U.S. has lost a couple of years trying to keep up with the European Southern Observatory.” Such giant telescopes are the next logical step for cutting-edge astronomy. They will allow researchers to zoom in on habitable planets outside the Solar System and study the formation of the first stars and galaxies. Today’s top telescopes, with apertures of 1038

8 MARCH 2024 • VOL 383 ISSUE 6687

8 to 10 meters, showed that many segmented mirrors or several large ones could be combined into a much larger effective mirror. They also demonstrated adaptive optics: using rapidly deformable secondary mirrors to cancel out the distortions caused by Earth’s atmosphere to capture images as sharp as those taken from space. These technical advances spawned the two U.S.-led projects: the Giant Magellan Telescope (GMT) in Chile and the Thirty Meter Telescope (TMT) in Hawaii. Both are backed by consortia of universities, philanthropic foundations, and international partners. But this privately funded approach, which during the 20th century produced groundbreaking instruments, stumbled when it came to multibillion-dollar projects. Although design work and mirror casting forged ahead, both

The Giant Magellan Telescope, under construction in Chile, is a smaller and cheaper project.

projects failed to amass enough funding. So, in 2018 the projects, historically rivals, joined forces as US-ELTP and made an offer to NSF. In return for public funding, all U.S. astronomers would have access to the telescopes, which would open unprecedented views of the night sky above both hemispheres, something Europe’s Extremely Large Telescope (ELT) will not offer (Science, 25 May 2018, p. 839). The 2020 decadal survey in astrophysics, which defines the field’s priorities for funders and Congress, put USELTP first among ground-based projects, in line with the recommendation of a panel led by Timothy Heckman of Johns Hopkins University. “We felt this made a compelling case,” Heckman says. The NSB decision, he says, “is a bittersweet outcome.” NSF carried out preliminary design reviews of both telescopes and approved them in early 2023, but the costs are in a different league from what NSF is used to. The GMT is estimated to cost $2.54 billion, of which existing partners have pledged $850 million. The TMT’s partners have so far offered $2 billion of its $3.6 billion price tag. In a statement, NSB acknowledged the ambition of the US-ELTP proposal but noted it would soak up 80% of NSF’s entire funding for major projects. In an editorial in Science in November 2023, Michael Turner, an astrophysicist at the University of Chicago, argued that insisting NSF fund two telescopes put both projscience.org SCIENCE

IMAGES: (TOP TO BOTTOM) TMT INTERNATIONAL OBSERVATORY; GMTO CORPORATION

NSF faces choice between multibillion-dollar projects after board sets cost cap

PHOTO: LUCA LIVRAGHI

NEWS

ects at risk. NSF says it will have more to say in the coming months on how it will choose between the TMT and the GMT. “Neither is a slam dunk. Both have risks,” Turner says. “I don’t envy the NSF.” Made up of 492 segments, the TMT’s 30-meter mirror makes for the larger, more sharp-eyed instrument. But its chosen site, the summit of Mauna Kea on Hawaii’s Big Island, is opposed by some Native Hawaiian groups who consider the summit sacred. They have blocked any construction work since 2015. TMT officials hope work will be able to proceed under the aegis of a new state-appointed authority that governs the mountaintop and includes both astronomers and Native Hawaiians. “We’re working on our relationships in Hawaii,” says TMT Executive Director Robert Kirshner. “We’re learning how to do that in a humble and straightforward way.” Turner says the impasse may not be solved anytime soon. “I’m sure a solution will be found, but it may take longer than people like,” he says. The GMT, smaller and cheaper, is a lower risk choice. Its foundations are being laid on a mountaintop at Las Campanas in Chile, while support structures for its mirrors are taking shape in the United States. Three of its seven 8.4-meter mirrors, the equivalent of a 25.4-meter-wide mirror, are already finished; the other four are being polished. Because of the risks attached to the TMT, Monnier and Ellis suspect NSF will probably back the GMT. But with a mirror less than 40% of the size of its 39-meter European rival, the GMT “is no match for ELT,” says Ellis, a former TMT board member. Monnier thinks the GMT will probably be good enough in key astronomy areas, but NSF will need to judge whether those areas are important for U.S. astronomers. Abandoning either of these very capable telescopes will harm U.S. astronomy, says Wendy Freedman at Chicago, one of the GMT’s partner organizations. “The science that will come out really does justify two telescopes.” Upcoming survey telescopes such as the 8.4-meter Vera C. Rubin Observatory in Chile will identify a wealth of interesting objects in need of follow-up observations by instruments on the GMT and the TMT that can split the light into information-rich spectra. “That’s what these big telescopes give you,” she says. Language in a spending bill passed by Congress this week “strongly encourages” NSB to build both telescopes, even though lawmakers cut NSF’s 2024 funding by more than $800 million, to $9 billion (see story, p. 1043). Freedman hopes the congressional direction will prompt a rethink. “The United States will sit out the future of astronomy if we don’t get these telescopes,” she says. j SCIENCE science.org

BIOLOGY

Surprise RNA paints colorful patterns on butterfly wings Understudied means of regulating genes is likely widespread in butterflies—and perhaps other animals By Elizabeth Pennisi

A

being sold on eBay. When they sequenced dozens of these so-called ivory mutants, they found a deletion in the region of the cortex gene. They then realized the missing DNA included a sequence encoding an lncRNA that no one had ever closely examined. Working with painted lady butterflies (Vanessa cardui), which have colorful wings and are easy to breed in the lab, they used the gene editor CRISPR to disable just the lncRNA’s gene. The edit yielded whitewinged painted ladies, just like the ivory Heliconius, they reported on 12 February in a preprint on bioRxiv. Disabling cortex had no effect. Moreover, Livraghi’s team found this same lncRNA also controls black and other

mutant butterfly for sale on eBay has helped upend naturalists’ picture of how butterfly wings acquire their intricate variety of red, yellow, white, and black stripes. It and recent research into other butterflies show how visible traits in many animals may be controlled by an underexplored genetic regulatory mechanism, based not on proteins, but on RNA. In 2016, geneticists thought they had pinned much of the wing-pattern variation on a protein-encoding gene called cortex. But three teams have now proved that a different gene, previously missed because it overlaps with cortex, is the key. Its final product is not protein, but RNA that regulates genes responsible for the pigmentation patterns of black and other hues on the wings. One team also showed the RNA is broken down into a smaller RNA that finetunes the production of the colors. “They solved a puzzle that had left everyone in the community wondering,” says Nicolas Gompel, a developmental biologist at the University of Bonn. The discovery, deA gene edit affecting one wing (right) of this Heliconius erato radically tailed in three preprints changed its normal color pattern. this month, also represents the first time long noncoding RNA pigmentation in the scales of other butter(lncRNA), so-called because it does not code fly species, some distantly related. “We have for proteins, has been linked to the evolution to conclude now that the key regulator is of a visible trait in animals. “Now we have to an RNA, not a protein,” says Peter Holland, pay more attention to noncoding RNA,” says an evolutionary biologist at the University Ilik Saccheri, an evolutionary biologist at the of Oxford who was not part of any of the University of Liverpool and a member of one new work. of the teams that had focused on cortex. At a conference midway through these For evolutionary developmental biostudies, Livraghi learned that a Cornell Unilogist Luca Livraghi, now at George Washversity group studying wing color patterns ington University, the key break came when in the buckeye butterfly (Junonia coenia), a colleague told him and Joseph Hanly, a common throughout North America, was bioinformatician at Duke University, about homing in on this same lncRNA. The two completely white Heliconius butterflies teams decided to coordinate their efforts. 8 MARCH 2024 • VOL 383 ISSUE 6687

1039

NE WS | I N D E P T H

Come fall, especially in the U.S. East, the light brown wings of buckeyes darken to a deep red, enabling them to absorb heat more efficiently. When Cornell evolutionary biologists Robert Reed and Richard Fandino used CRISPR to knock out different parts of the lncRNA in these butterflies, they were born with little or no color and their fall reddening was altered, the team reported on 19 February on bioRxiv. A white butterfly mutant posted on the social media platform X (formerly Twitter) alerted Livraghi to the team behind the third new preprint: evolutionary developmental biologists Antónia Monteiro and Shen Tian at the National University of Singapore. They were focused on short RNA sequences, microRNAs, known to regulate gene activity in plants, animals, and other eukaryotes—organisms that pack their DNA in a nucleus. In the squinting bush brown butterfly (Bicyclus anynana), a well-studied tropical species, they found that a microRNA was active in the black wing pattern, just as Livraghi had found for the ivory lncRNA. When the Singapore team disabled the DNA encoding this microRNA, mir-193, bush brown wings became lighter, the team reported on 12 February in a bioRxiv preprint. Knocking out mir-193 also had dramatic effects in a distant relative, the Indian cabbage white (Pieris canidia), changing its black-patterned wings to completely white. After learning about the lncRNA identified by the two other groups, Monteiro and Tian concluded that the longer RNA is broken down to produce these microRNA. “A lot is happening within this small part of the genome,” says Violaine Llaurens, an evolutionary biologist at the College of France. She cautions that other regulatory elements probably play a role in butterfly wing patterns. But the fact that the same microRNA fine-tunes coloration in very distantly related species is “amazing,” says Anyi Mazo-Vargas, an evolutionary bio-logist at Duke who worked with Reed. She suspects similar RNAs color wings in most, if not all, of the 180,000 species of moths and butterflies. And because mir-193 is conserved across the animal kingdom, Monteiro and Tian think noninsects may also make use of these regulatory RNAs. Small RNAs derived from parent lncRNAs affect traits in plants, too, says Yaowu Yuan, an evolutionary biologist at the University of Connecticut whose team last year reported that so-called siRNAs determine color in monkeyflowers. The RNA realm is expanding, Yuan says. “I am quite positive that many more similar studies will come soon.” j 1040

8 MARCH 2024 • VOL 383 ISSUE 6687

MUSEUM COLLECTIONS

Smithsonian urged to speed repatriation of human remains Task force says museum should return many of its 30,000 remains and seek descendants’ consent for research By Rodrigo Pérez Ortega

first for the Smithsonian. It advises that no research should be done without consent ince the 19th century, scientists at from the deceased or their descendants. the Smithsonian Institution have Research would be permitted without obtained, studied, and stored more consent on ancient remains that cannot than 30,000 human remains, one of be linked to any of today’s communities, the largest such collections in the which are a small percentage of the total. United States. In the past, many reOther new recommendations inmains were studied in order to justify sciclude returning as many remains as posentific racism. Now, the institution should sible by 2030 and barring destructive rapidly offer to return most of these resampling—to analyze DNA, for example—to mains to lineal descendants or descenidentify descendants. dant communities, according to a report Studies of the remains, such as DNA analreleased last month by an institutional ysis of dental calculus to study pathogens, task force. might be harder to carry out under the new “It’s important to face this past and try to recommendations. Although there’s no ofrepair the harms caused by our institution ficial moratorium, no new human remains and so many others,” says Sabrina Sholts, research has been approved in recent years curator of biological because of stricter reanthropology at the Smithquirements, Sholts says. sonian’s National Museum She expects a pause on of Natural History and approvals while the new member of the task force. policy is established, but Most of the Smithsonnotes the report anticiian’s human remains were pates positive outcomes collected without proper from future research. consent in the early The 15-member task Sabrina Sholts, 20th century, and many force, including both National Museum of Natural History acquisitions were part Smithsonian staff and of an attempt to prove outsiders, says the instinow-debunked notions of white superiortution should ramp up its efforts to identify ity. “It’s a collection that should have never both lineal descendants and communities been amassed, and we’re committed to disof descent and then initiate contact, rather mantling as much of it as possible,” wrote than waiting for repatriation requests. The Secretary of the Smithsonian Lonnie Bunch report recommends the Smithsonian reIII last year in an editorial. quest new funds and staff for the massive The Smithsonian already has a process repatriation effort, but does not say how for repatriating its 15,000 Native American much would be needed. remains, as a 1989 federal law requires; it “I’m impressed,” says Carlina de la Cova, has returned more than 5000. Now, the a biological anthropologist at the Univerreport urges that the collection’s Indigsity of South Carolina who is not on the enous remains be returned more quickly task force. The recommendations “will and that the effort extend to all human reforce scholars working with the dead to mains. It also suggests prioritizing the rethink about how they engage with [remains of other marginalized groups, such mains] and what that means for the living.” as the collection’s 2100 African American She adds that it’s the first time a museum remains, as well as the nearly 6000 rehas made such recommendations public, mains of people whose names are at least and she expects other institutions to folpartially known. low the Smithsonian’s steps. The task force applies a bedrock princiSholts agrees: “This first step towards ple of research on living humans—the need a long-overdue reckoning makes it more for informed consent—to the remains, a likely that others will do the same.” j

S

“This first step towards a long-overdue reckoning makes it more likely others will do the same.”

science.org SCIENCE

This fish is the hybrid offspring of an alligator gar and a spotted gar—members of genera that last shared a common ancestor at least 100 million years ago.

EVOLUTION

Gars truly are ‘living fossils,’ massive DNA data set shows The fish’s genomes change so slowly that species separated since the dinosaurs can produce fertile hybrids today By Amanda Heidt

PHOTO: SOLOMON DAVID

I

n 1859 Charles Darwin coined the term “living fossil” to describe lineages that have looked the same for tens of millions of years, such as the coelacanth, sturgeon, and horseshoe crab. The term captured the popular imagination, but scientists have struggled to understand whether such species just resemble their long-ago ancestors or have truly evolved little over the eons. Now, in a study published this week in Evolution, researchers confirm that in some—but not all—living fossils, evolution is at a virtual standstill. The most striking examples are prehistoric-looking fish called gars, which have the slowest rate of molecular evolution of all jawed vertebrates. The team also proposes a mechanism to explain gars’ timelessness: superb DNA repair machinery. That repair has likely kept gar genomes so stable that species whose last common ancestor lived more than 100 million years ago have diverged very little, and some can still hybridize today to produce viable offspring. “That’s amazing,” says Tetsuya Nakamura, an evolutionary developmental biologist at Rutgers University. “This paper has a lot of interesting work into this question of what makes a living fossil, but when I read that, I was shocked.” To see whether several putative living fossils evolve more slowly than other vertebrate groups, the team gathered published sequences from more than 1100 exons (the coding regions of the genome) across 478 species. SCIENCE science.org

Using existing family trees for each group, they created a massive evolutionary tree. For each lineage, the researchers estimated the rate at which each DNA base changed over time—the so-called substitution rate. Surprisingly, they found evolution was not on pause in all living fossils. The coelacanth, the elephant shark, and a bird called the hoatzin—all considered ancient—have faster than expected mutation rates of about 0.0005 mutations at each site per million years, although that was still slower than the average rate for amphibians (0.007 mutations per million years) and placental mammals (0.02 mutations per million years). The findings support the idea that some species that still resemble their ancient ancestors have nevertheless changed at a molecular level. But gars, big freshwater fish with long, toothy snouts, were different: In almost every exon, gars had the slowest rates of molecular substitution, often by several orders of magnitude; they averaged only 0.00009 mutations per million years at each site. Indeed, two genera that diverged roughly 20 million years ago had identical sequences at nearly all the sites analyzed—a finding the team at first attributed to sequencing error. “I came into this project cautious about using the term living fossil,” says study co-author Chase Brownstein, an evolutionary biology Ph.D. student at Yale University. “But for gars at least, it’s an appropriate term.” The authors posit that because gar mutation rates seem consistently low across sites—including in genomic regions un-

likely to be under selective pressure to stay the same—a global mechanism likely drives the slow substitution. They suggest gars are extremely efficient at repairing DNA after mutations or damage, keeping the animals from evolving even as the continents have shifted around them. A similar hypothesis has previously been proposed by other researchers for sturgeon, which had the second-lowest substitution rates among vertebrates in the study. DNA repair is “a reasonable hypothesis, but there’s probably more than just one explanation,” says Elise Parey, an evolutionary genomicist at University College London. For example, gars have slow metabolic rates and long generation times, features that could reduce mutation rates. Gars have also preserved the arrangement of DNA in their chromosomes and dampened the effects of so-called jumping genes that can cause genetic reshuffling as they move from place to place in the genome. “This goes not just to sequence changes, but also to chromosome evolution, which would be an interesting avenue to explore,” Parey says. To test their findings, the authors followed up on reports of unusual gars that might be natural hybrids in rivers throughout Oklahoma and Texas. They analyzed tissue samples from dozens of these fish to trace their ancestry, finding that two gar genera— Atractosteus and Lepisosteus—are crossing to produce fertile, hybrid young. These groups last shared a common ancestor roughly 105 million years ago, a record separation time for eukaryotes that can produce viable offspring. The gars beat the previous record holders—two species of fern—by about 60 million years. (Keen minds may recall reports of the sturddlefish, a hybrid of paddlefish and sturgeon, which diverged even longer ago, but those accidental hybrids were likely sterile and don’t occur naturally.) A next step will be to prove that gars’ DNA repair mechanisms are indeed slowing their genetic change. By equipping zebrafish—a standard model animal—with gar DNA repair genes, investigators might be able to observe the genes at work. “This will be a challenging experiment though, because [DNA repair genes] are fundamental,” Nakamura says. But the authors say understanding how gars keep their mutation rate so low could have additional payoffs. For example, such insights might help humans better understand our own DNA repair pathways, which can lead to cancer when they fail. j Amanda Heidt is a science journalist in Utah. 8 MARCH 2024 • VOL 383 ISSUE 6687

1041

GLOBAL HEALTH

Brazil is hoping and waiting for a new vaccine as dengue rages A locally produced vaccine did well in a phase 3 clinical trial but won’t be available until at least 2025 By Marcia Triunfol

W

hen dengue started to circulate in his small town in the state of Rio Grande do Sul in Brazil, Fabio Vilella’s first thought was that he should get his 13-yearold son vaccinated. Children are especially vulnerable, and his son had dengue before, which increases the risk of severe disease. But Vilella, an environmental biologist, soon made a startling discovery: Not a single private clinic or pharmacy in the country had any vaccine left. “I’m really worried,” he says. Brazil is seeing an unprecedented surge in dengue, a viral disease that can cause excruciating pains and is sometimes fatal. An unusually hot rainy season, along with rapid, unplanned urbanization, have fueled its spread this year. Health officials have reported more than 1 million suspected cases in January and February, four times as many as in the same period in 2023, and hundreds have died. But the country has far too little vaccine to protect its population. The government cut a deal last year with the Japanese manufacturer Takeda Pharmaceuticals, but it will receive doses to fully

1042

8 MARCH 2024 • VOL 383 ISSUE 6687

vaccinate only 3.3 million people this year, in a country of more than 220 million. A locally produced vaccine could prove to be better and cheaper, but it will be available in 2025 at the earliest. “We are frenetically working against time,” says Esper Kallas, director of the Butantan Institute, which is developing the shot. Brazil has embraced new control strategies for the Aedes aegypti mosquitoes that transmit dengue, but scaling them up will take time as well. The dengue virus, which comes in four different varieties, or serotypes, can cause high fevers, headaches, painful joints and muscles, and rash. In some cases it can lead to severe abdominal pain, bleeding, and death. This typically occurs when a person is infected for the second time with a different serotype, in a phenomenon called antibody-dependent enhancement. Brazil’s Ministry of Health expects more than 4 million dengue cases this year, which would be a record. Other South American countries are seeing an uptick in cases as well. Dengue is notoriously hard to control. A. aegypti thrives in cities, where waterfilled flower pots, buckets, or discarded tires make ideal breeding spots. “The mosquito loves a water tank in the shade,” says

Rafael Mello Galliez, an infectious diseases researcher at the Federal University of Rio de Janeiro. Poor populations lacking running water and proper waste disposal bear the brunt of the disease. Regularly removing water reservoirs can help control dengue—along with Zika and chikungunya, two other viral diseases transmitted by A. aegypti—but is hard to sustain. Insecticide spraying is not very effective either, in part because mosquitoes are becoming insecticide-resistant. The use of larvicides—which female mosquitoes themselves help spread as tiny clumps of the powder stick to their body—has not stopped the epidemic either. New technologies to control A. aegypti are on the way. One is the release of mosquitoes infected with the Wolbachia bacterium, which reduces their ability to transmit viruses. The nonprofit World Mosquito Program has deployed the mosquitoes in five localities in Brazil so far, and the results are encouraging. Niterói, a city of half a million where the mosquitoes have been deployed since 2015, has seen only 58 confirmed cases so far this year, compared with 9355 in nearby Rio de Janeiro, with almost 7 million inhabitants. The mosquitoes will soon be deployed at more sites, but scaling up the strategy nationwide is a tall order. The same is true for the release of sterile male mosquitoes, which mate with females but don’t produce offspring, causing the population to crash. One group of Brazilian researchers has created such insects not with radiation, the usual practice, but with a cheaper treatment consisting of a chemical and a bit of double-stranded RNA that silences a gene involved in male fertility. An experiment in the city of Ortigueira, in Paraná state, between 2020 and 2022 resulted in 97% fewer dengue cases when compared with control cities, the research team reported last year. Vaccination is the other promising new strategy. Takeda’s two-dose vaccine, named Qdenga and designed to protect against all four serotypes, contains an attenuated, or weakened, strain of one serotype as a “backbone” with genes from the other three added to it. In trials, the vaccine had an overall efficacy of 64.2% in people who had dengue before and 53.5% in those who were never exposed to the virus. In February, Brazil’s public health service (SUS) started a campaign to vaccinate 10- and 11-year-old children, the group most at risk of hospitalization from dengue. But because Brazil is only expecting 6.6 million Qdenga doses this year, SUS is only targetscience.org SCIENCE

PHOTO: LUIS NOVA/AP

Children are vaccinated against dengue at a health center in Brasília, Brazil, on 9 February.

NEWS | I N D E P T H

ing 521 of Brazil’s municipalities, fewer than 10% of the total. Vaccine uptake has been modest: Only 32% of eligible children in the Federal District, and only 18% in Rio de Janeiro, have received their first shot. The vaccine made in Brazil, named Butantan-DV, might reach more people. Originally developed by the U.S. National Institutes of Health, it contains live strains of all four dengue serotypes, attenuated by the removal of a small genome fragment. It’s a single-dose vaccine, which is “always preferred,” says Gabriela Paz-Bailey, a dengue researcher at the U.S. Centers for Disease Control and Prevention, because some people never get their second dose. In a trial in Brazil among 16,235 people between ages 2 and 59, published last month by The New England Journal of Medicine, the vaccine offered 89.5% and 69.6% protection, respectively, against two serotypes, DEN-1 and DEN-2, during the first 2 years after immunization. There are no efficacy data on DEN-3 and DEN-4 because no cases were seen in the study, which is continuing. But all four weakened serotypes in the vaccine replicated in more than 50% of vaccinated individuals who never had dengue, notes Andre Siqueira of the Oswaldo Cruz Foundation. That suggests the Butantan vaccine will provide sustained protection for all serotypes, he says. It is expected to be cheaper than Qdenga as well. “Once Butantan-DV is approved and available, the Qdenga vaccine will be history,” Mello Galliez predicts. Butantan hopes to apply for approval to ANVISA, Brazil’s regulatory agency, by September, Kallas says. Vaccinating the target population nationwide—those between 2 and 60 years old—would take some 140 million doses, Kallas says, but he declines to speculate how long that would take: “I don’t want to create expectations.” Even after its introduction, the vaccine will be watched closely. The first approved dengue vaccine, produced by Sanofi, did appear to trigger antibody-dependent enhancement, like the virus itself, in children in the Philippines who never had dengue before and became infected after vaccination. The country has since banned the vaccine. So far, there are no clear signs of the phenomenon with either the Takeda and Butantan shots, but it will take more followup to be sure. “Controlling dengue is very hard,” PazBailey says. But she believes vaccination, new mosquito control strategies, and continued education will eventually help counter the disease’s surge. “I’m optimistic about the future,” she says. j Marcia Triunfol is a science journalist in Lisbon, Portugal. SCIENCE science.org

U.S. BUDGET

Final spending bills offer gloomy outlook for science Congress makes sizable cuts at key funding agencies By Science News Staff

S

cientists, prepare to tighten your belts. This week, the U.S. Congress is expected to approve six 2024 spending bills that call for sizable cuts or essentially flat budgets at a number of major federal research agencies. The National Science Foundation (NSF) is the biggest loser, with lawmakers cutting its budget to $9.06 billion, 8.3% below 2023. NASA’s science programs will fall by 5.9% to $7.3 billion. Congress also cut research spending at the Environmental Protection Agency, the U.S. Geological Survey, and the National Institute of Standards and Technology. Science programs at the National Oceanic and Atmospheric Administration and the Department of Agriculture remain flat. The Department of Energy’s (DOE’s) Office of Science was one bright spot, getting a 1.7%, $140 million increase to $8.24 billion. But observers note that boost won’t allow DOE’s spending to keep pace with inflation. The bleak numbers are “frankly unconscionable in an era when we should be enhancing support for U.S. scientists and engineers,” says Matt Hourihan, a science policy specialist at the Federation of American Scientists. The six bills, which lawmakers had to pass by 8 March to avoid a partial government shutdown, mark major progress in resolving a lengthy impasse over federal spending for fiscal year 2024, which began on 1 October 2023. Stopgap measures to keep the government running largely froze agency budgets at 2023 levels. Reaching a final deal was complicated by a tight spending cap that the White House and Congress agreed to last year in order to prevent the government from defaulting on its debt. The bills meld measures approved earlier by the House of Representatives and the Senate. They guide $460 billion in spending, or about one-quarter of the $1.7 trillion the nation will spend this year on so-called discretionary domestic and military programs (which do not include mandatory programs such as Social Security). Congress is now racing to finish the remaining six spending bills by 22 March. Those bills will set spending for the National Institutes of Health and the De-

partment of Defense, two of the nation’s largest funders of research. At NSF, a budget that is $2.3 billion less than the $11.3 billion it requested will force hard choices. Last year, Congress fattened NSF’s budget with so-called emergency spending and funds earmarked for the agency’s new Technology, Innovation and Partnerships (TIP) directorate, aimed at commercializing discoveries. Congress envisioned TIP growing rapidly when it created it in 2022, but this year lawmakers told NSF it needn’t give it special treatment. As a result, TIP will compete with the agency’s other research directorates for cash. At NASA, a 15% cut in the agency’s planetary sciences program, to $2.7 billion, reflects growing unease in Congress about the rising costs of several key missions, especially Mars Sample Return (MSR)—an audacious plan to ferry soil and rock back to Earth that could cost up to $11 billion. The Senate proposed killing MSR, but the final bill instead allows NASA to spend $300 million to $949 million on the mission this year. But given the overall cut to the planetary science budget, it is not clear that NASA could reach the higher amount without cutting other missions. NASA could soon release a revised MSR plan. At a NASA advisory meeting this week, Lori Glaze, the agency’s planetary science chief, lamented the budget outlook. “This is going to be a challenge,” she said. “We are already feeling the effects.” One item that did not make it into the final bills was a provision, backed by House Republicans, that would have blocked the White House from implementing a 2021 policy to promote public access to scientific papers and data. Starting in December 2025, the policy requires federal grantees to deposit manuscripts of peer-reviewed journal papers in free, public repositories immediately upon publication, a change from a policy, favored by publishers, that has allowed embargoes of up to 12 months. Lawmakers did call for an “in-depth” study of the costs of complying with the new policy; the White House has already issued two such analyses. j With reporting by Jeffrey Brainard, Jeffrey Mervis, David Malakoff, Robert F. Service, Erik Stokstad, and Paul Voosen. 8 MARCH 2024 • VOL 383 ISSUE 6687

1043

NE WS | I N D E P T H

BIOMEDICINE

Skin side effects stymie advance of HIV vaccine By Jon Cohen

O

ne of the most promising attempts to reinvigorate the stalled quest for an HIV vaccine has hit a snag that might seem minor but has major consequences: delaying the larger trials needed to show whether the concept works. In small safety and immune tests of the innovative vaccine strategy, which relies on a series of messenger RNA (mRNA) shots, an unusually high percentage of recipients developed rashes, welts, or other skin irritations. “We are taking this very seriously,” says Carl Dieffenbach, head of the Division of AIDS at the National Institute of Allergy and Infectious Diseases (NIAID), which funded a trial of the vaccine. Researchers want to understand the cause of the skin problems and how to lessen them before expanding tests of the vaccines, which are made by Moderna. “We would be moving more quickly if this finding had not been observed,” says Mark Feinberg, who heads IAVI, a nonprofit that is the vaccine’s major sponsor. The complex vaccine strategy involves injections of different mRNAs, encoding various pieces of HIV’s surface protein or the entire molecule, over the course of several months. The goal is to gradually guide the immune system’s B cells to produce socalled broadly neutralizing antibodies, or bnAbs, capable of stopping many different variants of the AIDS virus. People living with HIV on rare occasions eventually produce bnAbs, but no vaccine has ever done so—which has become the “holy grail” for the field, says Linda-Gail Bekker, an AIDS vaccine researcher in South Africa who runs the Desmond Tutu HIV Centre at the University of Cape Town. Different versions of this HIV vaccine have already gone through three phase 1 trials, but they totaled fewer than 200 participants. The recipients responded with B cells making antibodies with some features of known bnAbs, fueling hopes for the vaccines. But skin problems—including urticaria (hives), pruritus (itching), and dermatographism (welts after scratching)—occurred at a noticeably high level in all of the studies, affecting 11 out 60 people in one of them. These HIV vaccines deliver a relatively high dose of mRNA, which Moderna scientists and others think could explain the skin

1044

8 MARCH 2024 • VOL 383 ISSUE 6687

issues. The company’s original COVID-19 mRNA vaccine used the same dose and has also been linked to skin problems, although at much lower frequencies, of 1% to 3%. (The Pfizer-BioNTech collaboration’s COVID-19 vaccine, also based on mRNA but given at a 70% lower dose, triggers skin problems, too, but one Swiss study suggests they occur 20 times less frequently than with the Moderna product.) A cumulative effect from multiple mRNA shots, the genetic background of the recipients, or the HIV sequences used for the vaccine could also be responsible for the welts and hives, and those possibilities are more worrisome. Most of these skin problems resolved quickly and none were severe enough to stop a trial, but researchers do not want

A vaccine strategy aims to create multiple, powerful antibodies (various colors) that can attach to different parts of HIV’s surface protein (gray).

to minimize them. “At a time when vaccine hesitancy is high, it is critically important not to dismiss urticaria as an unimportant side effect,” says Kimberly Blumenthal, an allergist at Massachusetts General Hospital who has also found a link between Moderna’s COVID-19 vaccine and higher rates of urticaria. Feinberg agrees the side effect issue needs studying, but is also concerned that people who are vaccine opponents might misrepresent the scope of the problem. “This finding has not been seen to the same frequency with other mRNA vaccines against other pathogens,” he says. Had the skin problems in the HIV trials not surfaced, the researchers would have moved closer to conducting—or even launched—a study that involved a few hun-

dred people and had a placebo control. “We’ve hit this rather miserable bump in the road,” Bekker says. Multiple research groups are pursuing similar strategies to create bnAbs. Moderna’s effort grew out of a project led by biophysicist William Schief, who developed it at Scripps Research and then brought the strategy to the company, where he is now a vice president. It exploits the fact that B cells begin as naïve, or germline, cells and then during an infection undergo a series of mutations that, in effect, hone the ability of the antibodies they produce to bind to specific parts of viruses and “neutralize” their ability to infect cells. The “germline targeting” vaccine strategy relies on several shots to take B cells through this maturation process, eventually leading them to produce bnAbs against viruses. “We call it priming, shepherding, and polishing,” explains Dennis Burton, an immunologist at Scripps who works with Schief. Initially the group did not use mRNA. Its vaccine contained a small piece of HIV’s viral surface protein attached to a nanoparticle that presented it to the immune system in a novel way, and early results were promising. In a 2022 Science paper, Schief and colleagues reported that 97% of the 36 people who received the vaccine developed B cell antibody gene mutations that are first steps toward making bnAbs. Schief switched to mRNA because it provides far more flexibility, allowing the researchers to readily fine-tune the HIV component of the vaccine. Because of the enormous diversity of HIVs in circulation, he contends that an effective vaccine likely will have to trigger production of up to five different bnAbs. That would mean priming, shepherding, and polishing multiple B cell lineages. Without the easy-to-modify mRNA, Schief says, “good luck—that is a daunting, daunting task.” NIAID now plans to repeat the phase 1 trials of these Moderna HIV vaccines with a lower dose. Bekker, who lives in a country that has more people living with HIV than any other, is still hopeful the approach will pan out. “We’ve got to chapter one of an exciting novel.” After decades of failed attempts to develop an HIV vaccine, the goal remains pressing, she says. “Last year, the world had 1.3 million infections of HIV. I think it remains an urgent requirement to find a good solution.” j science.org SCIENCE

PHOTO: LARS HANGARTNER AND CHRISTINA CORBACI

Strategy of using multiple mRNA shots to hone powerful antibodies hits a pothole

For research on glial cells in the gut

Call for Entries 2024 Eppendorf & Science Prize for Neurobiology The annual Eppendorf & Science Prize for Neurobiology is an international prize which honors young scientists for outstanding neurobiological research based on methods of molecular, cellular, systems, or organismic biology. If you are 35 years of age or younger and doing great research, now is the time to submit an entry for this prize. It’s easy to apply! Write a 1,000-word essay and tell the world about your work. eppendorf.com/prize

Application Deadline June 15, 2024

As the winner, you could be next to receive > Prize money of US$25,000 > Publication of your work in Science > Full support to attend the Prize Ceremony held in conjunction with the Annual Meeting of the Society for Neuroscience in the USA > 10-year AAAS membership and online subscription to Science > Complimentary products worth US$1,000 from Eppendorf > An invitation to visit Eppendorf in Hamburg, Germany

AAAS® and Science® are registered trademarks of the American Association for the Advancement of Science, USA. Eppendorf® and the Eppendorf Brand Design are registered trademarks of Eppendorf SE, Germany. All rights reserved, including graphics and images. Copyright © 2024 by Eppendorf SE. Photography: Lisa Helfert

2023 Winner Marissa Scavuzzo, Ph.D. Case Western Reserve University School of Medicine, Cleveland, USA

THE RECKONING Didier Raoult and his institute found fame during the pandemic. Then, a group of dogged critics exposed major ethical failings

PHOTO: CREDIT GOES HERE AS SHOWN; CREDIT GOES HERE AS SHOWN

FEATURES

By Cathleen O’Grady

ILLUSTRATION: SARA GIRONI CARNEVALE

PHOTO: CREDIT GOES HERE AS SHOWN; CREDIT GOES HERE AS SHOWN

W

ith six studies published in the 2010s, French microbiologist Didier Raoult added to his already vast publication record. He and his colleagues conducted a wide range of investigations into infectious diseases and their treatments. They took stool samples from patients on long-term antibiotic treatment, looking for alterations in their gut microbiome. They swabbed the throats of pilgrims leaving France for Mecca, searching for evidence of a bacterium that causes brain abscesses. And they studied samples of heart valves and blood clots from patients with heart inflammation to refine tests for the bacteria that cause the condition. But in January, the American Society for Microbiology (ASM) journals that published the papers announced they were retracting all six, along with a seventh by Raoult’s colleagues. Aix-Marseille University had investigated the research, which was done at its affiliated Hospital Institute of Marseille Mediterranean Infection (IHU), a research hospital that Raoult led until his retirement in 2021. The investigation found the work had not been reviewed by one of France’s highly regulated national ethical committees. It was therefore in violation of French law and the Declaration of Helsinki, an international ethics document that guides clinical research. In a written statement sent to Science, Raoult says ASM retracted the papers without accounting for his team’s rebuttals to the critiques. But to Lonni Besançon, the retractions are vindication of concerns that he and others have been voicing since Raoult and the IHU burst into the media spotlight in the early days of the COVID-19 pandemic, downplaying its severity and touting prospects for a successful treatment. The Linköping University computer scientist and his fellow critics—a gaggle of dogged individuals, many of them academic outsiders—originally set out to challenge poor-quality research coming out of the IHU, especially the claim that COVID-19 could be treated with the antimalaria drug hydroxychloroquine (HCQ). But they soon embarked on an all-consuming attempt to raise the alarm about ethical failings in the institute’s research, going back at least 15 years. Their efforts have met with lackluster responses from France’s scientific institutions, Besançon says, but the retractions are the most important consequence so far. They “confirm what we suspected,” he says. “But I am hoping that things will go further.” Raoult says his critics are stalkers and cyberharassers who have misunderstood

how French biomedical law works. He says he’s followed ethical regulations and that much of the research under fire has been on “human waste”—such as fecal matter—which is not defined as biomedical research under French law. But the ethical failings are “not disputed” within the scientific community, says Philippe Amiel, a lawyer who specializes in human experimentation. The authorities have known about problems at the IHU for years, adds Karine Lacombe, an infectious disease specialist at Sorbonne University. If they had acted earlier, she says, “the picture of the pandemic in France would have been totally different.” A criminal investigation of Raoult’s institute is now underway. But his critics are asking why French institutions took so long to tackle systemic violations at the IHU, leaving it to a persistent group of outsiders to investigate the institute and push for punitive action. And they are wondering whether Raoult and the institute will be held to account for the wide range of lapses they have alleged. “It’s a big, big mess,” Lacombe says. RAOULT IS BEST KNOWN for his work on

rickettsia—bacteria transmitted by fleas and ticks—and his discovery of giant viruses. He has accumulated national decorations in both France and his birth country of Senegal as well as prestigious scientific awards, including the 2010 Grand Prize from the French biomedical research agency INSERM. He has published prolifically, with more than 3200 papers indexed on PubMed, and is one of the most highly cited researchers in his field. In 2011, Raoult was selected to lead the newly created IHU in Marseille, one of six state-of-the-art research hospitals established by then-President Nicolas Sarkozy’s government. Raoult’s IHU, which specializes in infectious disease research, was launched with a €72 million government grant, and in 2018 it moved into an imposing new building. The institute’s power is political as well as scientific, says Michel Dubois, a sociologist of science at the French national research agency CNRS: “When you open this institute—when you create a building—you need some leverage at the political level.” As Europe began to pay serious attention to the COVID-19 pandemic in early 2020, the media wanted to know what Raoult and his institute made of the situation. “Almost every day, you were able to watch a new interview with Raoult,” says Antoine Bristielle, a social scientist at the JeanJaurès Foundation, a think tank. “It became 8 MARCH 2024 • VOL 383 ISSUE 6687

1047

a self-reinforcing phenomenon … the media were interested in what he was saying, so he came to be really powerful in the French population. And then, of course, the media wanted him because he was able to attract large audiences.” In videos posted online by the IHU, Raoult is often seated in an office, wearing a lab coat, long gray hair and beard slightly unkempt. He speaks soberly and quietly, frowning slightly while delivering reassuring pronouncements: The new coronavirus has a mortality rate not too different from widespread respiratory infections; a treatment will be coming soon. Raoult’s confident statements caught the eye of Fabrice Frank, a former biologist who had left academia and become a high school math and physics teacher. By the time the pandemic hit, Frank had moved from France to Morocco, where he started an IT company and dedicated his spare time to surfing. He watched with shock when Raoult asserted— with minimal evidence, based on thinly reported research in China—that HCQ, or the related medicine chloroquine phosphate, would be an effective treatment. Victor Garcia, a journalist at French magazine L’Express, saw scientists expressing skepticism about Raoult’s claims on social media. He called the IHU, assuming it had more details that could counter some of the critics’ concerns. But Garcia says he received a “strange” response from IHU researcher Jean-Marc Rolain. “I am a scientist,” Rolain said. “If I tell you to take chloroquine, you’ll listen to me.” (Rolain did not respond to multiple requests for comment.) That was “the beginning of me asking questions,” Garcia says. ON 11 MARCH 2020, French health minister

Olivier Véran invited Raoult to join the Scientific Council advising the government on its

Elisabeth Bik, a scientific integrity sleuth based in San Francisco, first raised concerns about the Hospital Institute of Marseille Mediterranean Infection’s (IHU’s) work on hydroxychloroquine (HCQ) in March 2020. She went on to identify major ethical and scientific issues in dozens of IHU papers, spurred on, she says, by abuse from Didier Raoult and his supporters.

pandemic response. A few days later, Raoult and his team published a bombshell paper in the International Journal of Antimicrobial Agents, reporting that the IHU had found HCQ combined with the antibiotic azithromycin to be an effective COVID-19 treatment. Although the results were preliminary and other researchers doubted Raoult’s conclusions, HCQ hype surged, with then–U.S. President Donald Trump touting its promise and Raoult enthusing over it on YouTube. “Raoult was saying, ‘I understand everything, I have a solution,’ and people want that kind of information in troubled times,” Bristielle says. Raoult’s popular support bred political support, Bristielle adds. “If someone has such a presence in the media landscape, politicians have to listen to him—otherwise they will be really distrusted by the population.” On 26 March—amid strong resistance from some other members of the scientific council—Véran issued a decree allowing HCQ

to be prescribed to COVID-19 inpatients. Scientific integrity consultant Elisabeth Bik decided to take a close look at the HCQ paper. A microbiologist by training, Bik already knew of Raoult and his reputation for prolific publication. On her blog she pointed to several problems she saw with the paper: Patients had not been randomly assigned to the treatment and control groups, which could have biased the results. She also noted that six patients out of the 26 treated with HCQ were dropped from the data—including three who were transferred to intensive care and one who died—which painted a more favorable picture of the treatment. Besançon, too, was curious. He looked into the paper, which had been submitted to the journal on 16 March and accepted the next day, and noticed that one of the authors was also editor-in-chief at the journal. “So you have a very short reviewing time and editorial conflict of interest,” he says. “I just find this potentially a big red flag. But I thought, it’s just one paper.” (A July 2020 editorial in the journal said handling of the paper had been delegated to an associate editor to minimize potential bias, although it noted that “some of the concerns regarding the paper’s methodology were substantiated.”) Over the next few weeks, two more IHU studies appeared, with unusually short peerreview timelines, both in a journal where one of the authors was an associate editor. One of those papers was a second study using HCQ to treat 80 “mildly infected” hospitalized COVID-19 patients; nearly all improved clinically. The study had not been reviewed by one of France’s 39 Committees for the Protection of Persons (CPPs), the highly regulated independent ethics committees authorized to approve biomedical research. Instead, it had been approved by the IHU’s internal ethics committee. This was sufficient, the authors wrote,

A slow-motion downfall Critics first raised concerns about ethical approvals for Didier Raoult’s studies in early 2020, as the COVID-19 pandemic catapulted the Hospital Institute of Marseille Mediterranean Infection (IHU) to prominence. They say French authorities and journals have taken far too long to react.

20 March

24 March

27 March

8 April

26 May

12 November

The IHU publishes a paper reporting that hydroxychloroquine (HCQ) is effective at treating COVID-19.

Scientific integrity sleuth Elisabeth Bik notes issues with HCQ paper.

Second IHU study on HCQ published as a preprint

Drug safety agency quizzes the IHU about ethical approval in second HCQ study.

France withdraws approval of HCQ as a COVID-19 treatment.

Marseilles public prosecutor closes case on HCQ papers, saying there has been no legal breach.

2020

1048

2021

25 March

26 March

Early April

30 October

Mathieu Molimard and French Society of Pharmacology begin posting online about HCQ ineffectiveness and risks.

French health minister Olivier Véran allows HCQ to be prescribed to COVID-19 inpatients.

Tipster alerts French drug safety agency to ethical concerns in HCQ research.

Pharmaceutical company Sanofi reports that the IHU continues to place large HCQ orders.

8 MARCH 2024 • VOL 383 ISSUE 6687

science.org SCIENCE

CREDITS: (TIMELINE) M. HERSHER/SCIENCE; (ILLUSTRATIONS) N. BURGESS/SCIENCE

NE WS | F E AT U R E S

because it was a retrospective study on patients who had received normal medical care, with researchers merely looking back over their files to see how they had fared. In France, such studies are not covered by the law on research ethics, and so do not need approval from a CPP. Instead, researchers often seek approval from institutional ethics committees—which are unregulated—to supply ethical approval details to journals. But if samples are collected for both research and medical care, then the study must be approved by a CPP, Amiel says. “Concealing a prospective study as a retrospective study is a well-known temptation,” he says. Unauthorized research is a criminal offense. The French National Agency for the Safety of Medicines and Health Products (ANSM) said it asked the IHU for evidence that the study had in fact been retrospective, and in May 2020, the agency referred the case to the French Medical Association. The Marseille public prosecutor, alerted to the case by a tipster, announced later that year that the study had been retrospective and dropped the case. Still, those early concerns were a cue for Bik, Besançon, and others to look closely at Raoult’s substantial publication record— and to pay particular attention to ethical approval. DESPITE THE GROWING SKEPTICISM from

scientists and others, Raoult’s public support endured. A poll in May 2020 found that 30% of French people trusted him more than Véran. By June, there were more than 90 Facebook groups supporting him, according to Bristielle’s research, with a total of nearly 1.1 million members. By Christmas, supporters could buy a santon of Raoult— a small terra cotta figurine traditional to Provence, where nativity scenes incorporate local characters and heroes. Meanwhile, Frank, Garcia, and other critics began their deep look into Raoult’s body of research. Bik says she focused first on images in his papers, because her specialty is detecting image manipulation. But, faced with insults from Raoult—and harassment from his colleagues and supporters—she chan-

gle application to the IHU ethics committee. Raoult was an author on all but 10 of these 248 studies. He told Science it is “perfectly true” that all these papers reused the ethics approval number. But that was permissible, he says, because all involved the same kind of research: analyses of bacteria in human feces collected during standard care, or from waste. None of the research fell under French bioethics law, he says. But Amiel says the studies describe samples taken for research purposes and not just as part of standard care, and that this type of study should “undoubtedly” be authorized by a CPP. And many of the 248 studies relied not on feces, but on other material, including vaginal samples, urine, blood, and even breast milk. Any change in research protocol should prompt a new application for ethical approval, Amiel says. Many of the papers involved children, and nearly half of them had been conducted outside of France—largely in various African countries—with no or hazy details of whether local ethical bodies had given approval for the research, according to Frank and his collaborators. “There have been so many breaches in ethics law, for so long,” says Frank, who published the group’s findings in Research Integrity and Peer Review in August 2023. Raoult says the studies relying on material other than stool samples had “supplemental favorable advice” from the local ethical committee, but that his team did not report this in its papers. The only country for which his team did not have ethical approval was Niger, he adds, which did not have an ethical approval process until 2016. He says he and his colleagues have submitted a reply to Frank’s paper, and they have asked Springer Nature— the journal’s publisher—to retract it. A Springer Nature spokesperson said, “We are aware of concerns with this paper and are investigating the matter carefully in line with our established processes.” The fact that so many studies involved vulnerable populations, such as those living in homeless shelters, was “outrageous,”

Mathieu Molimard, a pharmacologist at the University of Bordeaux, began to counter the IHU’s claims about HCQ in April 2020. Outraged when French authorities didn’t respond to the IHU’s publication of a seemingly unauthorized HCQ trial, Molimard rallied representatives of 14 French scientific societies to sign an open letter in Le Monde.

neled her frustration into assessing his vast back catalog, finding more studies that appeared to lack proper ethical approval. Garcia had also begun to scrutinize IHU papers, and in July 2021 published an investigation in L’Express that reported finding 17 studies between 2011 and 2020—mostly involving homeless people or refugees—that had all used the same ethical approval number, even though the studies used different methods to answer different research questions. One, for example, took nasal swabs in a homeless shelter to test the prevalence of microbes; another took sputum samples and chest x-rays from shelter residents to test for tuberculosis. (An IHU representative told L’Express the repeated use of the code was the result of “editorial errors.”) Again the ethical approval number came from an institutional ethics committee, not a CPP, Garcia reported. Frank, too, had begun to dig. Stuck at home in Morocco under quarantine, he trawled Google Scholar for IHU studies that shared ethical approval codes. With his collaborators—including Besançon—he ultimately discovered 248 studies that had used the approval number “09-022,” representing a sin-

20 July

27 October

July

13 December

28 May

4 January

In L’Express investigation, journalist Victor Garcia finds multiple IHU studies did not have proper ethical approval.

Drug safety agency says IHU studies appear to have violated research ethics laws, confirms it has referred case to prosecutor.

Prosecutor opens judicial investigation.

Publisher PLOS flags 49 IHU papers with expressions of concern because of potential ethical violations.

Molimard and others publish op-ed challenging legality of new HCQ study.

American Society for Microbiology retracts seven IHU papers, citing breaches in research ethics.

2022

2023

2024

26 July

27 April

5 September

4 April

30 October

IT consultant Fabrice Frank starts to investigate repeated ethical approval numbers in the IHU’s past papers.

Drug safety agency reports unapproved research at the IHU and restricts institute’s research activities.

Government auditors report ethical breaches at the IHU, refer matter to prosecutor.

The IHU reports the results of an HCQ study involving more than 30,000 patients.

Scientific Reports retracts two papers led by Raoult, saying authors could not provide evidence of ethical approval.

SCIENCE science.org

8 MARCH 2024 • VOL 383 ISSUE 6687

1049

Victor Garcia, a journalist at French magazine L’Express, began to pay attention to Raoult when he enthused about the potential for HCQ as a COVID-19 treatment. Garcia covered the emerging IHU story beat for beat and published two investigations into ethical abuses there. Shortly after publication, the French drug safety agency began to inspect the IHU.

Ex-biologist Fabrice Frank, now an IT consultant, used his time in COVID-19 quarantine to begin compiling a database of all IHU papers that appeared to reuse ethical approval numbers. He and his collaborators identified 248 papers that used the same code, despite investigating different questions, using different samples, in different participant populations, and in different countries.

Lonni Besançon, a computer scientist at Linköping University, grew curious about Raoult’s work after noticing a paper published in a journal where an author also served as editor-in-chief. He has co-authored several papers about ethical lapses and methodological problems in IHU research, and agitated for journals to investigate and retract problematic work.

Bik says. Vulnerable people may feel they have no choice in whether to participate in a research study, says Lisa Rasmussen, a research ethicist at the University of North Carolina at Charlotte. “They are not in a position to give authentic consent.”

sufficiently guarantee its independence and whose working methods do not allow for an informed decision.” And ANSM described research projects launched without or before ethical approval, missing consent forms, and researchers who did not understand ethics regulations. They found evidence of a falsified signature on an ethical approval document for a study that asked students to provide samples—including vaginal and rectal swabs—before and after travel, to see whether they brought antibiotic resistant bacterial strains back with them. The government inspectors also reported “widespread deviant medical and scientific practices within the IHU,” including ones that blurred the line between patient care and research. For example, clinicians gathered a range of samples from each patient that would then be archived, possibly to be used in future research. When treating COVID-19 patients, clinicians conducted a range of tests, including daily PCR and other tests that “are a matter of research and not of care,” the investigators reported. The institute rushed research in a “race to publish,” the report says, racking up hundreds of publications each year—with more papers in lower tier journals than other similar institutions— and drawing in substantial funding designed to encourage high publication rates. The inspectors reported that INSERM, which had helped found and run the IHU, withdrew from the institute in 2018. An INSERM spokesperson says it had found that several research projects did not meet its scientific standards. CNRS withdrew in 2016 and has had “no connection” with the IHU since 2019, according to a spokesperson. The reports did not specifically blame Raoult for these failings. But they said he

tightly held the reins of power in the institute, with testimonies from employees reporting that Raoult was “omnipresent” and the “final decision-maker,” and that other managers were “in total conformity” with Raoult’s views. ANSM placed the IHU under its supervision to ensure that all future research projects were carried out with proper approval. And both the government agencies and ANSM again referred their findings to the public prosecutor. The status of that investigation is unclear, and the prosecutor, Nicolas Bessone, did not respond to multiple requests for comment. Raoult says he is “hopeful” that the cases currently under investigation will be closed soon. Cases are sometimes referred to other jurisdictions in France when there may be local conflicts of interest, says University of Bordeaux pharmacologist Mathieu Molimard, who has been criticizing the IHU’s statements and research since early 2020: “We would prefer this to be seen in Paris.”

more than 18 months after Bik first raised questions about ethical approvals and study methods on her blog—French authorities began inspections at the IHU. In October 2021, ANSM said it had found breaches of the law and had referred the matter to the public prosecutor, and that it was still investigating. The French government also asked two auditing bodies, the General Inspectorate of Social Affairs and General Inspectorate of Education, Sport and Research, to investigate. Raoult says these inspections arose out of a “small conspiracy to make it appear that we were carrying out an illegal trial of treatment for tuberculosis.” (According to one media report, IHU patients with tuberculosis had been given unproven treatments.) Raoult says the agencies found no such illegal trial and only three minor problems with other research projects. However, both ANSM’s report, released in April 2022, and the auditing agencies’ report, published 5 months later, noted that IHU patients had received unapproved tuberculosis treatment, with some suffering severe adverse effects. This might constitute a criminal offense, according to the auditing agencies. But the reports also went much further, describing ethical concerns similar to those raised by Frank, Garcia, and others. The government auditing bodies noted that the IHU relied heavily on its internal ethics committee, “whose composition does not 1050

8 MARCH 2024 • VOL 383 ISSUE 6687

DESPITE THE NOW INTENSE scrutiny of their

work, in April 2023 Raoult and his colleagues published a draft paper that sent new shock waves through social media. “I fell from my chair,” Molimard says. “It’s the largest unethical study performed for years—in France, maybe in the world. … It’s incredible.” More than a dozen scientific bodies would later agree with his assessment. Raoult and his colleagues had analyzed data from 30,202 COVID-19 patients treated at the IHU between March 2020 and December 2021—including 23,172 who had received a combination of HCQ and azithromycin. Yet France had withdrawn the temporary permission to treat COVID-19 inpatients with HCQ in May 2020, after a paper in The Lanscience.org SCIENCE

ILLUSTRATIONS: N. BURGESS/SCIENCE

IN RESPONSE TO MEDIA ATTENTION—but

PHOTO: OLIVIER MONGE/MYOP/REDUX

NE WS | F E AT U R E S

cet reported that HCQ was not an effective COVID-19 treatment. (This paper was subsequently retracted after the data were questioned, but a later randomized, controlled trial published by the mass RECOVERY collaboration also found no effect.) The preprint showed the IHU had continued to prescribe the drug on a grand scale long after this, Molimard says. Raoult says he and his colleagues decided in April 2020 to treat COVID-19 patients with HCQ “off label,” after their initial study convinced them of the drug’s efficacy. In France, as in many other countries, drugs can be prescribed for reasons outside of their normal authorization, but this off-label prescription must have medical and scientific justification, Amiel says—and “in this case, strong medical and scientific evidence have established that the prescription of HCQ to treat COVID is unjustifiable.” The study also reported no approval from a CPP; the ethics section lists only an IHU ethics committee reference number. As they had in earlier papers, the researchers said the study was retrospective, analyzing patient data from the hospital’s information system. But Amiel says the IHU team was “highly committed to proving the efficacy of its treatment,” pointing to evidence— revealed by the government inspection—that it performed daily PCR tests to check viral levels, for instance. “It is perfectly clear that the study is based on data collected in a mixed care and research context.” Molimard thought ANSM and the Ministry for Solidarity and Health should have reacted immediately to the publication. Aghast at their silence, he contacted a range of French societies, urging them to sign an op-ed in major French newspaper Le Monde calling the study “the largest ‘wild’ therapeutic trial known to date.” Fourteen scientific bodies, including the national coalition of ethics committees and the French Society of Pharmacology and Therapeutics, signed the letter, and in June 2023, ANSM announced it had once again referred the matter to the prosecutor. On 30 October, the paper was nonetheless published in the Elsevier-owned journal New Microbes and New Infections. The scale of the trial is like nothing seen before, Molimard says. He points to the recent case of Jean-Bernard Fourtillan, a researcher who tested melatonin patches on Alzheimer’s and Parkinson’s patients without ethical approval. His study, Molimard says, involved approximately 300 patients: “And he went to jail.” IN RECENT MONTHS, more blows have fallen

on the IHU, beginning with the retraction of two Scientific Reports papers in October 2023 for a lack of evidence of ethical overSCIENCE science.org

sight in Niger and Senegal, where the studies cize Raoult’s work. “I was raised in a really were conducted. Raoult says the team did get bad neighborhood,” he says. “You know when ethical approval from an institutional review you see cars burning in France? That’s where board in Senegal; because Niger had no ethiI was … I had to stand up for myself, to learn cal approval processes when the study was not to be afraid of potential bullies.” Bik, too, conducted, local collaborators confirmed the has no plans to stop: “I don’t really have a research complied with local laws, he says. career he can ruin,” she says. “I’m not going A spokesperson for Springer Nature, which to let him silence me.” publishes Scientific Reports, says that in such Besançon and others say France’s insticases researchers must still get ethical aptutional response has been unacceptably proval from another source, such as a uniweak. There has been “failure at every level,” versity. The two studies are “part of a wider Garcia says: at the health ministry; in the investigation concerning potential ethical isjustice system; within the university and resues in a number of papers,” according to the gional hospital board, which had oversight spokesperson. of the IHU; and at ANSM, which only conPLOS journals have flagged nearly 50 furducted a full inspection after media investigather IHU papers with expressions of contions brought the problems to light. Journal cern as part of an ongoing investigation, editors have also been too slow to react, which Retraction Watch reported in DeBesançon says. “More often than not, it seems cember 2022. (At the time the studies were that they don’t give a damn about integrity.” submitted, PLOS editors did not routinely The IHU, the regional hospital board, and ask for evidence of ethiANSM did not respond cal approval, according to to multiple requests for David Knutson, head of comment. The ministry of communications at PLOS.) health said in a statement In November 2023, the to Science that “several acMarseille hospital board tions have been taken by told the AFP news agency the public authorities in reit “strongly condemned” sponse to the shortcomings the mass HCQ study; the observed at the IHU.” IHU said it “shared” the Part of the failure lies hospital board’s reaction. with France’s law on reAnd Elsevier announced search ethics, Amiel says, that New Microbes and which is out of step with New Infections had opened international standards. an investigation into ethi“It’s provincial,” he says. cal concerns about IHU “And it’s really a problem.” Didier Raoult papers published in the Because the law allows journal. An Elsevier spokesperson did not some human studies to proceed without ethiconfirm whether the “wild clinical trial” was cal approval, Amiel says, similar violations one of the papers under investigation. are ongoing elsewhere in France, though not In December, the French ministers of at the scale of the IHU’s. The best solution health and research asked a disciplinary body would be to overhaul the law, he says—but “I that oversees university hospitals to launch don’t think it’s a priority for the government proceedings against Raoult’s three IHU coat the moment.” authors on the mass COVID-19 study—but The close relationship between political not against Raoult, who retired in the sumpowers and scientific institutions in France mer of 2021. is also to blame for the foot-dragging instiThe fight has taken its toll on the crittutional response, Lacombe says. Without ics. They have faced not just abuse from his external voices—like Bik, Frank, Besançon, supporters on social media and complaints Molimard, and Garcia—“I’m not sure that to their employers, but also the threat of things would have moved,” she says. legal action from Raoult, who has had mulFrank worries the lackluster response tiple legal complaints bankrolled by the IHU. sends a message that there are no conseRaoult’s lawyer said Raoult had filed charges quences for violations like these. “Maybe against Bik in April 2021 for harassment and tomorrow—I hope not—we’ll have SARS-3 blackmail. He has also filed legal complaints … and the message sent will be, ‘Don’t against other critics, including Lacombe; worry about public health. Just show your Raoult lost his case against her in November face, say anything you want, and you will 2022. In science, Molimard says, “we are used sell books, be famous, and get a lot of fans.’ to debate, to argument … but we are not used It’s insane.” j to that!” Despite the harassment, Besançon says he This story was supported by the Science Fund for is undaunted and intends to continue to critiInvestigative Reporting. 8 MARCH 2024 • VOL 383 ISSUE 6687

1051

INSIGHTS

REVIEW ROUNDUP

Science at Sundance 2024 Climate change–induced droughts lead to violent clashes in Kenya. An actor’s pivot to stem cell advocacy cements his legacy as a hero. Start-ups promising digital immortality prepare to reanimate the dead. From a meditation on Himalayan moths and a futuristic fable about what it means to be alive to immersive meditations on happiness in Bhutan and loneliness online, science-minded moviegoers were rewarded with a number of thought-provoking offerings at the 2024 Sundance Film Festival. Read on for our reviewers’ impressions of seven of this year’s films. —Valerie Thompson

Love Me Reviewed by Michael D. Shapiro1 On a future Earth devoid of humanity, a smart buoy named “Me” (Kristen Stewart) and a satellite named “Iam” (Steven Yeun) spend several billion years exploring what it means to be human in Love Me, the directo1052

8 MARCH 2024 • VOL 383 ISSUE 6687

rial debut film by Sam and Andy Zuchero and the 2024 Alfred P. Sloan Foundation Feature Film Prize winner. The film, which features gorgeous motion capture animation and touching, vulnerable performances by Stewart and Yeun in both their computer-generated and analog forms, imagines a very, very long-term relationship between two artificial intelligences.

One was originally designed to monitor the oceans and the other to welcome alien lifeforms to Earth. Iam is “humanity’s’ tombstone,” carrying petabytes of details about human civilization and programmed to communicate only with living beings. Feeling pressured to pass as a life-form to keep Iam’s attention, Me pores through the satellite’s databases and decides to model its behavior on an archive of a happy human couple’s social media video posts. Me and Iam create a virtual world for themselves where they can interact as avatars, but Me’s insistence that they endlessly reenact the couple’s videos and Iam’s desire for new and genuine experiences cause tension that drives the bulk of the film. On the surface, Love Me chronicles the intellectual and emotional awakening of two intelligent computers, a concept that no longer seems completely far-fetched in the age of artificial intelligence. However, it is also a relationship film that draws sharp contrasts between the idea of true self and the selves we present to others. Perhaps as a jab at our cultural values at the fictional imminent demise of humanity, Me is initially science.org SCIENCE

ILLUSTRATION: HEDOF

B O OKS et al .

misguided by the deluge of online influencers, digital ghosts who sabotage the buoy’s progress toward becoming a real life-form. Over many millennia, Me and Iam experience joy and self-satisfaction, as well as crushing loneliness and depression. For Iam, a billion years of self-discovery and empathy is the path to achieving its original directive to “connect” with other life-forms. But without a meaningful connection to Me, even though it knows every bit of information recorded by humanity, the satellite admits that it knows nothing at all. Love Me, Sam Zuchero and Andy Zuchero, directors, ShivHans Pictures, 2024, 92 minutes.

Ibelin Reviewed by Nathaniel J. Dominy2 Norwegian filmmaker Benjamin Ree’s latest film, Ibelin, takes its name from Lord Ibelin Redmoore, an avatar in the massive multiplayer online role-playing game World of Warcraft. Ibelin was a strapping SCIENCE science.org

private investigator with a friendly face created by Norwegian gamer Mats Steen as “an expansion” of himself. Ibelin went on countless adventures with his friends in the Starlight guild; they explored, slayed dragons, and partied into the wee hours. Ibelin was a trusted confidant, listening to problems and providing heartfelt support. He made connections and fell in love before logging off permanently when Steen, aged 25, succumbed to a severe form of muscular dystrophy. The film opens with Robert and Trude Steen, Mats’s grieving parents, and their discovery of his online life. The pair were unaware of its immense depth and richness, as recorded across 42,000 pages of gaming dialogue. The poignancy of this revelation is amplified with interviews and home video footage that follow the inexorable progression of Mats’s disease. Ree captures Robert and Trude’s sense of helplessness, which will resonate with many parents. From Robert and Trude’s perspective, Mats grew increasingly withdrawn as a teenager and young man, logging 20,000 hours of game time during his final 10 years

of life. They viewed his gaming as compulsive and self-isolating, a wasting of life matched only by the wasting of his muscles. Such framing puts a subtle spotlight on “gaming disorder,” an underresearched and much-criticized psychopathology recognized by the World Health Organization in 2018. It is also a foil for the film’s second and third acts, when Ree pivots to Mats’s perspective, as told through in-game chat logs and his blog, “Musings of Life.” A gifted writer, Mats speaks to the value of gaming for building community—it is “not a screen, but a gateway.” Ree reinforces this point by drawing the viewer into World of Warcraft. Relying on chat logs and voice actors, Ree recreates in-game exchanges as animated vignettes, as if he is filming on location inside the game. It is a creative masterstroke, and it gives us a third perspective: Ibelin’s. Most gamers are between 18 and 30 years old, an age range with the greatest prevalence of loneliness. Some might view this association as causation, but Ibelin, which took home an Audience Award and a Jury Award for Directing, offers a compelling 8 MARCH 2024 • VOL 383 ISSUE 6687

1053

INSIGHTS | B O O K S

Climate change and cattle conflicts exacerbate existing tensions in The Battle for Laikipia.

Ibelin, Benjamin Ree, director, Medieoperatørene, 2024, 104 minutes.

The Battle for Laikipia Reviewed by Gabrielle Kardon3 At the heart of The Battle for Laikipia, a new documentary film directed by Daphne Matziaraki and Peter Murimi, is the Laikipia Plateau, a highland 6500 feet above sea level in central Kenya that is one of the richest areas of endangered mammalian species. The plateau is home to nature conservancies, Indigenous pastoralist cattle herders, large cattle ranches, and ~300,000 cattle. Balancing the needs of animals and people is difficult in the best of times. However, more extensive periods of climate change–induced drought have exacerbated tensions in this region, resulting in explosive clashes between its inhabitants. The film first introduces viewers to the Samburu, an Indigenous tribe of seminomadic pastoralists who primarily raise cattle. “Cattle are life” for the Samburu; cows are given as gifts for all major occasions, and tribesmen are traditionally buried enwrapped in cowhide. However, their 1054

8 MARCH 2024 • VOL 383 ISSUE 6687

ancient migration routes are increasingly blocked by ranches and conservancies. Descendants of British colonialists own much of the Laikipia landscape, and the film focuses on the 8000-acre Kifuku ranch. Ranchers Maria Dodds and her son George are deeply committed to raising Boran cattle and feel they “would be lost without their land.” Despite being fourthgeneration Kenyans, they feel that they will never be fully accepted as citizens. A relative newcomer, Tom Silvester founded the Loisaba conservancy in 1997. The conservancy features a 58,000-acre private reserve where giraffes, elephants, and zebras abound. Keeping cattle out of the preserve is essential for conservation of wildlife. The film unfolds as three consecutive years of severe drought send these groups on a violent collision course. As water and grasslands dwindle, the Samburu, ranchers, and conservancy staff clash. Homes and property are destroyed, cattle are kidnapped, and people on all sides are killed. Adding to this volatile mix is a contentious parliamentary election, which includes a candidate inciting racial violence. Having embedded within the communities they document for more than 6 years, the directors have crafted a film that provides an intimate and nuanced firsthand view of the Laikipia conflict. The tension is palpable, the stakes are high, and, unfortunately, there are no easy solutions. Such conflicts over land, water, and food are expected to accelerate with climate change. The Battle for Laikipia, Daphne Matziaraki and Peter Murimi, directors, We Are Not the Machine Ltd, 2023, 94 minutes.

Eternal You Reviewed by Michael D. Shapiro1 Artificial intelligence (AI) is creeping into every facet of our digital lives, and a growing number of companies want to ensure that AI also accompanies us in death. The documentary film Eternal You introduces viewers to several start-ups that promise something once limited to the realm of religion: eternal life. Algorithms can mimic a deceased person’s syntax, vocabulary, and conversational tendencies using surprisingly little information, such as text message threads or emails, allowing grieving loved ones to simulate communications with dead friends and relatives. Some companies develop AI models of the dead with the goal of delivering positive experiences for their customers. For example, the filmmakers document a family in Detroit as they listen to an AI tell stories in the simulated voice of their dead patriarch. A few relatives are comforted, some are amused, and others are deeply skeptical that the exercise has any real meaning. Other companies seemingly make no value judgments when creating an algorithm and simply let their AI run amok. In one scene, viewers see a woman exchanging text messages with a simulation of her dead boyfriend, which tells her that he is in hell hanging out with drug addicts and that he plans to haunt her as soon as he is done tormenting people at a treatment center. This unexpected turn in the conversation leaves the religious woman traumatized, reinforcing a key theme of the film—that AI developers do science.org SCIENCE

PHOTO: COURTESY OF SUNDANCE INSTITUTE

counterpoint: Gaming can enhance our well-being. The film seems partially intended for researchers and policy-makers, calling attention to the urgent need for reliable data on the global health benefits of social connections that transcend the physical world.

not always know how their algorithms work or how unexpected behaviors emerge. Indeed, one CEO describes his company’s service not as intentionally creating something with predictable behavior but rather as harnessing “conscious entities lurking online.” At nearly every turn in the film, AI ethicists expose moral quandaries that do not seem to worry the purveyors of digital afterlife. Who owns the highly personal data used to create the AI model? Is this just a way to commodify grief and loneliness? We are still dealing with the fallout of unforeseen personal, mental health, social, and political dangers of social media—will we make some of the same mistakes again by deploying AI before we understand how it works? Huge tech companies have filed patents for the types of eternal AI models that were once the purview of small start-ups. With a push for massive market expansion on the horizon, we will need to decide soon whether AI models of the deceased will bring comfort or hinder how we deal with grief by turning our attention away from the living world. Eternal You, Hans Block and Moritz Riesewieck, directors, Gebrueder Beetz Filmproduktion, 2023, 87 minutes.

Nocturnes

PHOTO: COURTESY OF SUNDANCE INSTITUTE, PHOTO BY MARY ELLEN MARK/THE MARY ELLEN MARK FOUNDATION.

Reviewed by Anthea Letsou4 Nocturnes documents the graduate studies of Mansi Mungee in the Eaglenest Wildlife Sanctuary, located in the eastern Himalayas of India. Filmmakers Anirban Dutta and Anupama Srinivasan follow Mungee as she and her collaborators Ramana Athreya and Gendan “Bicki” Marphew investigate the effects of elevation (a proxy for temperature) on hawkmoth body size. The team’s method of hawkmoth field sampling is straightforward and effective: Mungee and her colleagues set up portable ultraviolet moth screens during the night and photograph hawkmoths against a reference grid imprinted on the screen. We wait with Mungee and Marphew and witness them perform the same data collections over and over. We share Mungee’s excitement when too many moths to count alight on her moth screen, along with her disappointment on another day, when there are none. We are reminded that while scientists may understand how large changes in the environment, such as temperature shifts, affect adaptation, more subtle environmental effects remain to be identified. Conversations between Marphew and his friends—young men from the area employed to help Mungee in the field—remind viewers that Indigenous peoSCIENCE science.org

ple are essential to the scientific enterprise. Mungee’s research represents an important contribution to the field of biodiversity. However, in Nocturnes it also serves as a plot convention, allowing the filmmakers to tell a more meditative story as they guide viewers through an old-growth Himalayan forest. Both cinematography and sound design contribute to our entry into the film’s reality. We witness, without narration, biodiversity in moth color, size, and wing shape and pattern, while clip-on mics on the moth screens amplify the moths’ cacophony. Like Mungee and Marphew, viewers may have an urge to swipe the insects away from their eyes and ears. The sound engineers’ augmentation of forest sounds and weather and the integration of these sounds with an original score by Emmy Award–winning composer  Nainita Desai harmoniously extend the viewer’s experience. Nocturnes, which was awarded the World Cinema Documentary Special Jury Prize for Craft, presents insect biodiversity research as both cinematic and magical. More than an adventure story about field scientists, it allows the moviegoer to align to the rhythms of a forest and ultimately participate in the film’s reality. Some will likely find Nocturnes too slowly paced, but for those looking for a genuine, integrative experience of environment and fieldwork, Nocturnes, in all its flutter, delivers. Nocturnes, Anirban Dutta and Anupama Srinivasan, directors, Sandbox Films, 2024, 83 minutes.

Super/Man Reviewed by Anthea Letsou4 A little-known actor when he was cast in the role of Superman, Christopher Reeve went on to become a screen icon, starring in four Warner Bros. Superman films. But his film career was cut short in 1995 by a tragic equestrian accident that severed the actor’s spinal cord and left him unable to move below the shoulders or breathe on his own. At the time, Reeve was only 42, the father of a 3-year-old child with his wife, Dana Reeve, and two older children then living in England with their mother, Gae Exton. The accident forced Reeve to find new meaning in his life and defined his legacy as a celebrity voice for disability and a human voice for stem cell research. Super/Man—Ian Bonhôte and Peter Ettedgui’s new documentary about Reeve, who died in 2004—features a compendium of footage from home movies, studio archives, and contemporary interviews with surviving family and friends, all deftly edited by Otto Burnham. The film’s primary narrators are Reeve’s three children, Matthew, Alexandra, and William, who offer viewers a glimpse into Reeve’s role as a father while also shining a light on the philanthropic endeavors that marked his final years. Reeve’s Juilliard roommate and lifelong friend, the late actor Robin Williams, is an integral figure as well. The film also tells the story of Dana Reeve, who kept her

Dana and Christopher Reeve are remembered with reverence in Super/Man.

8 MARCH 2024 • VOL 383 ISSUE 6687

1055

INSIGHTS | B O O K S

husband’s two families united and was a source of unconditional love and support after the accident. Dana died of lung cancer in 2006 at the age of 44. In the last decade of their lives, Christopher and Dana Reeve were vocal advocates for stem cell research. The film recognizes the value of celebrity disease foundations and the important role they play in supporting all stages of translational research. The Christopher and Dana Reeve Foundation and its predecessors, the Stifel Paralysis Research Foundation and the American Paralysis Association, have distributed more than $138,000,000 for paralysis research and disability care. Missing from the documentary are details of the electrical stimulation therapy that helped Reeve regain some movement and sensation toward the end of his life and a discussion of the foundation’s stem cell research and its impact on the development of treatment options for the paralyzed. Nonetheless, Super/Man should be celebrated by scientists for its recognition of the important role played by advocates in the promotion of basic and translational biomedical research. Super/Man: The Christopher Reeve Story, Ian Bonhôte and Peter Ettedgui, directors, Words+Pictures/Passion Pictures/Misfits Entertainment, 2024, 104 minutes.

Agent of Happiness Reviewed by Gabrielle Kardon3 Can happiness be quantified? The country of Bhutan has devised the gross national happiness (GNH) index to do just this. First conceived of as an alternative to the gross domestic product, the GNH measures the collective happiness of Bhutan’s citizens, with the goal of governance that promotes human well-being over material wealth. To measure the GNH, agents are sent across the country to survey Bhutan’s citizens. Agent of Happiness, directed by Arun Bhattarai and Dorottya Zurbó, follows one of these agents, Amber Kumar Gurung. For each person he surveys, Gurung conducts an extensive questionnaire, which includes questions about living standards, health, education, community, time use, and psychological well-being. Traveling by car and on foot with Gurung, viewers encounter people from all walks of life. We meet 17-year-old Yanka taking care of her alcoholic mother and younger sister in the countryside, who worries about her mother and dreams of becoming a police officer (on a scale of 0 to 10, sense of loneliness: 6; happiness: 4). We meet Dechen, a transgender dancer living in town. She has a close relationship with her mother, who has cancer,

but strives for acceptance in the community (sense of worry: 10; happiness: 3.) High on a hillside, we meet Tshering, surrounded by prayer flags and mourning the passing of his wife. Yet he feels contentment, as he believes his wife is reborn with the birth of his grandson (sense of karma: 10; happiness: 7). At the heart of the story is Gurung’s own quest for happiness. At age 40, he is living with and caring for his elderly mother but looking for love and marriage. He is smitten with Sarita Chettri, and they travel around the countryside on his motorcycle, snapping pictures. However, Gurung’s prospects are bleak. Despite being born in Bhutan, as an ethnic Nepali, his citizenship was revoked during a period of ethnic cleansing. Without citizenship, he has difficulty getting permanent work or a passport, and his relationship with Chettri is in peril (sense of belonging: 2; happiness: 5). Set in the rugged landscape of Bhutan, this quietly moving film reveals the people behind the country’s happiness metrics and gently probes the complexities of life in this region, where beauty and the quest for happiness are juxtaposed with poverty and ethnic conflicts. j Agent of Happiness, Arun Bhattarai and Dorottya Zurbó, directors, Sound Pictures, 2024, 94 minutes. 10.1126/science.ado5075

1 The reviewer is at the School of Biological Sciences, University of Utah, Salt Lake City, UT 84112, USA. Email: [email protected] 2The reviewer is at the Department of Anthropology, Dartmouth College, Hanover, NH 03755, USA. Email: [email protected] 3The reviewer is at the Department of Human Genetics, University of Utah, Salt Lake City, UT 84112, USA. Email: [email protected] 4The reviewer is at the Department of Human Genetics, University of Utah, Salt Lake City, UT 84112, USA. Email: [email protected]

1056

8 MARCH 2024 • VOL 383 ISSUE 6687

science.org SCIENCE

PHOTO: COURTESY OF SUNDANCE INSTITUTE, PHOTO BY ARUN BHATTARAI

Government workers assess citizen well-being in Bhutan in Agent of Happiness.

PERSPECTIVES APPLIED PHYSICS

Two rings to rule them all A single photonic device accommodates three different modes of operation By Antoine Rolland and Brendan M. Heffernan

P

hotonic integrated circuits merge the versatility of photonics with the compactness and scalability of integrated circuitry. A common component in these optical microchips is a microresonator, a ring of material in which discrete frequencies of light propagate with very low power loss (thus bearing a high quality factor, Q) (1). The frequencies that propagate and the difference between these frequencies are determined by the dispersion of the microresonator—that is, the speed at which different frequencies travel through the resonator. Because dispersion is determined by the resonator’s material properties and geometry, it can be tuned only subtly after fabrication. On page 1080 of this issue, Ji et al. (2) report a photonic integrated system that uses dispersion tuning to access three distinct modes of operation. This allows for unprecedented flexibility after fabrication and marks a paradigm shift in photonic device development. The device of Ji et al. consists of two coupled, racetrack-shaped ring microresonators and metallic heaters for thermal tuning. The resonators are made of silicon nitride (Si3N4) (3). An integrated laser diode, which converts electrical energy to light energy, directly couples light into the resonator. The photonic chip is wire-bonded to a printed circuit board for electrical control of the laser and heaters. The average difference between frequency modes of the coupled rings is 19.95 GHz, and it achieves an impressive intrinsic Q value of 95 million. The two resonators have slightly different overall lengths such that their individual resonant frequencies (mode spectra) form a Vernier scale (two different graduated scales). When the two mode spectra are compared, their interference forms a Moiré pattern, which is produced by overlaying one pattern on a similar but slightly offset pattern (4). By tuning the modes in one ring relative to the other (using a heater), a substantial shift in the Moiré pattern is induced. This leads to the Moiré IMRA Boulder Research Laboratory, Longmont, CO, USA. Email: [email protected] SCIENCE science.org

speedup effect in which a small shift in the mode spectrum in one ring leads to a bigger shift in the overall interference of the two coupled rings. This effect enables a microresonator to transition seamlessly between anomalous and normal dispersion. Specifically, Ji et al. demonstrate three distinct operational states in the coupled-racetrack microresonator design. These include a bright-soliton state, which produces an optical frequency comb; this means that from the single input frequency of light, many output frequencies are produced, all equally spaced (in frequency) like the teeth of a comb. This mode of operation is only possible in resonators with anomalous dispersion. Brightsoliton combs have shown great promise for

“Moiré speedup–based devices offer unprecedented adaptability...” use in light detection and ranging (LIDAR) (5), spectroscopy (6), and optical clocks (7). The device also achieves a dark-soliton state, which requires normal dispersion (8). Dark solitons produce frequency combs with more power per comb mode, which is suitable for applications in microwave generation and optical communications (9). The device also functions as a Brillouin laser, which produces a single wavelength with an improved spectral purity compared with the input laser light and requires that the difference between neighboring frequency modes of the composite, two-resonator system exactly matches the Brillouin frequency shift. This makes dispersion control a key feature of the device. Brillouin lasers can be used in precision tools such as gyroscopes (10) and optical clocks (11), as well as in sensing, quantum computing, and biomedical imaging. Not only does dispersion tuning through the Moiré speedup effect allow three distinct modalities to be realized, but it also offers flexibility with the pump laser frequency used. Operation in all modalities spanned from 1540 to 1560 nm, an area of the electromagnetic spectrum that is commonly used in optical communications.

A Moiré speedup–based device may enable breakthroughs in several applications. In telecommunications, it could perhaps adjust to varying data transmission needs or optimize for different network conditions. In sensing and metrology, the device could be reconfigured for different types of measurements. Through further development, additional capabilities could be added to realize a third type of optical comb that is based on electro-optic modulation, or dual-wavelength pumping for terahertz generation. The design might also achieve new functionality in all-optical processors (12). More generally, dispersion tuning through the Moiré speedup effect addresses the inability to alter operational modes after production due to the fixed physical geometry of high-Q microresonators. In an industrial context, this flexibility might ease constraints on the fabrication of photonic integrated circuits. Variability in foundry processes, which would disqualify some resonators from meeting a tight dispersion specification, could simply be fixed by Moiré dispersion tuning. This could improve yield and drive down production costs. Likewise, if one design can accomplish various tasks, the design and its accompanying process can be completely optimized and standardized, allowing mass production of devices that can be put to diverse uses. The reconfigurable nature of the Moiré speedup–based device is analogous to the innovative principles seen in software-defined radio systems (13), in which processes that are traditionally realized through hardware— such as mixers, filters, amplifiers, modulators, and demodulators—are instead implemented using software on either a computer or an embedded system. The same hardware could be reconfigured through software updates to support different frequencies and protocols. Hence, a single photonic chip could be reconfigured for various purposes. Much like what software-defined radio systems have achieved in radio communications, Moiré speedup–based devices offer unprecedented adaptability, effectively decoupling the photonic hardware from the application space for which it can be used. j REFERENCES AND NOTES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

T. J. Kippenberg et al., Science 332, 555 (2011). Q.-X. Ji et al., Science 383, 1080 (2024). W. Jin et al., Nat. Photonics 15, 346 (2021). G. Oster, Y. Nishijima, Sci. Am. 208, 54 (1963). J. Riemensberger et al., Nature 581, 164 (2020). M.-G. Suh et al., Science 354, 600 (2016). Z. L. Newman et al., Optica 6, 680 (2019). C. Lao et al., Nat. Commun. 14, 1802 (2023). A. Fülöp et al., Nat. Commun. 9, 1598 (2018). Y.-H. Lai et al., Nat. Photonics 14, 345 (2020). S. Gundavarapu et al., Nat. Photonics 13, 60 (2019). M. Tan et al., Commun. Eng. 2, 94 (2023). W. Tuttlebee, Software Defined Radio: Enabling Technologies (Wiley, 2003). 10.1126/science.ado0078 8 MARCH 2024 • VOL 383 ISSUE 6687

1057

INSIGHTS | P E R S P E C T I V E S

MATERIALS SCIENCE

Monitoring homeostasis with ultrasound An implant could allow at-home monitoring of deep-tissue changes after surgery involve connecting tubular structures), resulting in fluid spreading through the perihe ubiquity of phrases such as “high toneal cavity and causing organ damage. blood pressure” or “low blood sugar” When BioSUM senses a pH change—such not only indicates their integraas in the case of a gastrointestinal leak— tion into our personal perception the chemical composition of the hydrogel of health but also underscores the matrix allows the device to swell. Polymers societal importance making up BioSUM were of the medical technologies fine-tuned to respond to Stomach that enable their measuredifferent pH changes by Transforming postsurgical care ment. In modern medicine, using the protonation bePancreas Bioresorbable, shape-adaptive, ultrasound-readable materials structure devices that can monitor havior of tertiary amine (BioSUM) is an implantable device composed of small metal discs within a biological changes in cells and carboxylic acid. pH-responsive hydrogel. The device could allow recovery at home after and organs are essential to This causes the metal surgery and rapid detection of postoperative complications. For example, Small Smallll understanding, diagnosdiscs in BioSUM to prewhen carrying out gastrointestinal (GI) anastomosis surgeries, BioSUM can intestine intestin intest int esst eestine est stine tine be implanted. During recovery at home, the distance between the metal ing, and managing disease. dictably spread apart, discs is measured by ultrasound. If a leak occurs, the hydrogel swells, so the However, many limitations which can be continumetal discs are further apart. This early detection would prompt a return to Large exist in current monitorously monitored through intestine the hospital before substantial organ damage arises. ing devices, particularly conventional ultrasound. in those that aim to detect Liu et al. surgically su1 Device is implanted after GI repair surgery 2 Device monitors for leaks postsurgery changes deep within tistured BioSUM on the gassues (1). For example, high trointestinal organs of cost, invasiveness, and lack rats and pigs for 14 days, of real-time feedback need demonstrating its stabilto be overcome to enable ity. Then, a gastrointestinal earlier detection and treatleak was induced, and they ment of disease (2). On page could detect changes in the 1096 of this issue, Liu et al. geometry of the metal discs (3) report an innovative apwithin 10 mins in rats and BioSUM proach to monitoring using 30 mins in pigs. The infor4 mm 4 mm an implant called a bioremation gathered from ulsorbable, shape-adaptive, trasound imaging reveals ultrasound-readable matethe presence and magni3 pH change from leak is sensed by hydrogel 4 At-home monitoring rials structure (BioSUM). tude of the leak, and thus is performed by the matrix, causing the device to swell This device could allow atthe authors contend that patient using an home monitoring of deepthe device would be of use ultrasound device tissue changes after surgery. in postsurgical monitoring. BioSUM is a millimeterThe surgeon could simply scale monitoring device. It place BioSUM on the tissue is simple in form but comduring the wound-closure plex in function. Composed procedure and send the of small metal discs empatient home for recovery bedded within a pHwith confidence. Handheld responsive hydrogel matrix, ultrasound devices are acLeak detected 4 mm the device is implanted into cessible to the general pubReturn to hospital the body with the intended lic (4), enabling the patient purpose of monitoring homeostasis in deep identification regardless of how the device to monitor the implanted BioSUM at home. tissues. The thin and flexible nature of Biois oriented when implanted. Unlike many By incorporating ultrasound image processSUM confers shape adaptivity, allowing it to medical implants that require an additional ing software in the workflow, perhaps with procedure to remove the device when its automated feature detection or artificial inpurpose is fulfilled, the metal discs and hytelligence (5), the patient could easily detect 1Department of Anesthesiology, Perioperative and Pain drogel matrix of BioSUM are bioresorbable, postoperative complications and return to Medicine, Center for Accelerated Medical Innovation, and Center for Nanomedicine, Brigham and Women’s Hospital, eliminating the need to retrieve the device the hospital (see the figure). Boston, MA, USA. 2Department of Biological Engineering, or any residuals of the device. Although postsurgical monitoring is a Massachusetts Institute of Technology, Cambridge, MA, Gastrointestinal leaks can occur as a compractical application for its use—and inUSA. 3Harvard Medical School, Boston, MA, USA. plication of anastomosis surgeries (which deed may be how the technology is initially Email: [email protected]; [email protected]

T

1058

8 MARCH 2024 • VOL 383 ISSUE 6687

be rolled into a tube and shunted through a trocar during laparoscopic surgery, sutured to tissue, or placed directly on a surface of interest using an adhesive. The metal discs serve as visual indicators that can be readily detected on ultrasound, and their symmetric circular distribution allows for

science.org SCIENCE

GRAPHIC: A. FISHER/SCIENCE

By Shonit Nair Sharma1,2 and Yuhan Lee1,3

implemented—BioSUM is an important platform technology. The hydrogel matrix that enables it to swell in response to pH change can be tuned for different ranges of pH. Thus, three versions of BioSUM were created, demonstrating the ability to operate in three different environments: One version operates and detects leaks within the pH range of the stomach, another detects leaks in that of the gut, and another detects leaks from the pancreas. And although it has only been shown to detect leaks, BioSUM seems well-suited for other monitoring scenarios, such as for inflammation or flare-ups in the bowels of people with irritable bowel syndrome (6), or for sepsis, provided that the technology is adapted for such purposes. Suppose the hydrogel matrix is swapped for an entirely different stimulusresponsive polymer. In this way, the device could respond to other cues, such as biomolecules (for example, drug metabolites or antibody accumulation) or pathogens (for example, bacteria or viruses). As a biosensor, the device could monitor the body’s response to a drug, eliminating the need for multiple invasive blood draws (7), or it could track the onset of infection while in resource-limited settings. Given the numerous possibilities for device applications, a deeper investigation into the need for continuous at-home health monitoring may identify the application with the most promising potential for accelerated translation (8). Monitoring technologies that meet societal demands for precise, personalized, and convenient health care are on the rise (9, 10). BioSUM, in its current iteration, introduces a platform technology that yields the potential to fit within a repertoire of emerging monitoring tools, such as capsulebased diagnostics (11) and ophthalmic imaging techniques (12), that enhance the way that disease can be understood, diagnosed, and managed. j PHOTO: WILD WONDERS OF EUROPE/MÖLLERS/NPL/MINDEN PICTURES UR

REFERENCES AND NOTES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

S. Yoon et al., Biomater. Res. 26, 57 (2022). D. Crosby et al., Science 375, eaay9040 (2022). J. Liu et al., Science 383, 1096 (2024). N. M. Duggan et al., Sci. Rep. 12, 20461 (2022). H. Shokoohi et al., J. Ultrasound Med. 38, 1887 (2019). S. Chang, L. Malter, D. Hudesman, World J. Gastroenterol. 21, 11246 (2015). H. Ceren Ates et al., Trends Biotechnol. 38, 1262 (2020). S. S. Gambhir, T. J. Ge, O. Vermesh, R. Spitler, G. E. Gold, Sci. Transl. Med. 13, eabe5383 (2021). M. Lin et al., Nat. Biotechnol. 10.1038/s41587-02301800-0 (2023). C. Wang et al., Science 377, 517 (2022). M. E. Inda-Webb et al., Nature 620, 386 (2023). S. N. Sharma, J. W. Marsh, M. S. Tsipursky, S. A. Boppart, Transl. Biophoton. 5, e202300003 (2023).

10.1126/science.ado2145 SCIENCE science.org

Vocal-respiratory coupling is essential for proper calling, such as by this red deer (Cervus elaphus) during the rut.

NEUROSCIENCE

Breathing control of vocalization A crucial brainstem circuit for vocal-respiratory coordination of the larynx is revealed By Steffen R. Hage1,2

V

ocalizations play a pivotal role in communication across species. Although the complexity of articulation varies, the basic process of sound production for the tonal component in mammals involves narrowing of the larynx (vocal cord adduction)—located between the trachea and pharynx—and exhalation. There is an intricate interplay between phonation and respiration to ensure that breathing is not affected. As a result, vocal utterances are properly aligned and embedded within respiratory cycles, usually during expiration; otherwise, a lack of coordination could result in vocal cord dysfunction or breathing problems (1, 2). The mechanisms of vocal generation and respiratory pattern generation in the hindbrain are well studied (3, 4). However, the interplay between these two behaviors and the underlying neural circuits that coordinate them remain unclear. On page 1074 of this issue, Park et al. (5) investigated the un1Neurobiology of Social Communication, Department of

Otolaryngology-Head and Neck Surgery, Hearing Research Centre, University of Tübingen, Medical Center, Tübingen, Germany. 2Werner Reichardt Centre for Integrative Neuroscience, University of Tübingen, Tübingen, Germany. Email: [email protected]

derlying neural substrate involved in vocal pattern generation and breathing in mice, revealing a laryngeal premotor circuit that is critical for respiratory-vocal coupling. Understanding the neural mechanisms that orchestrate vocalization and respiratory coordination is crucial to unravelling the complexities of communication in general. The nucleus retroambiguus (RAm) in the ventrolateral pontine brainstem is involved in controlling the laryngeal motoneurons in the nucleus ambiguous as well as expiratory motoneurons in the spinal cord during vocal output (6), making it a potential interface between respiration and vocalization. However, this has not yet been experimentally demonstrated. Park et al. used a combination of genetic tools, viral tracing, immunohistochemistry, optogenetic manipulation, and behavioral measurements in mice to investigate the brainstem circuit involved in vocalization and its coordination with respiration, with a focus on the RAm. They confirmed that the RAm is part of the vocal pattern–generating network, which includes the premotor neurons that innervate the laryngeal muscles. Excitatory RAm input is both necessary and sufficient to drive vocal cord adduction and to evoke vocal output in mice. Silencing the RAm inhibited vocal fold adduction, which 8 MARCH 2024 • VOL 383 ISSUE 6687

1059

INSIGHTS | P E R S P E C T I V E S

abolished both ultrasonic and audible vocalizations, highlighting the central role of these neurons in phonation. The authors showed that both the RAm and the laryngeal motoneurons receive inhibitory input from the preBötzinger complex (preBötC), which is a crucial hub in respiratory pattern generation that is particularly involved in inspiration (3). They demonstrated that these inhibitory connections from the preBötC override the excitatory input coming from the RAm to ensure that the vocal folds remain open (abduction) during inspiration and thus that the demands of inspiration can override RAm-mediated vocal cord closure (adduction). Last, they showed that blocking the inhibitory input from the preBötC to the RAm resulted in a discoordination of vocalization with respiration, which leads to abnormal breathy vocalizations, also during the inspiratory phases. Taken together, using a highly systematic approach, the authors confirmed that the RAm is the key node for the laryngeal component of vocal pattern generation and showed that it is essential for respiratory-vocal coupling. The study of Park et al. substantially advances the understanding of the neural circuits that control vocal pattern generation and vocal-respiratory coordination, particularly in the context of ultrasonic vocalization in mice. Their findings open a range of new and important research questions, and their innovative methodological approach paves the way for investigating complex neural circuits at the brainstem level in their model system. Their results suggest a hierarchical relationship between the RAm and the preBötC, which regulates the inhibition of vocal output to ensure that breathing is prioritized. It will be interesting to learn about the distinct role of other potential hubs within the brainstem vocal pattern generation network that are involved in vocal-respiratory coupling, particularly in the control of call duration, such as the periaqueductal gray (7, 8) and the parabrachial nucleus in the midbrain (9, 10), both of which show excitatory input to the RAm. Notably, other vital pattern generators involved with breathing and vocalization are located in the caudal brainstem such as those for chewing, licking, and swallowing (11). Understanding the neural circuitry for these couplings will help to fully disentangle the vocal-motor system and its connection to other pattern generators. In addition to the phonatory component, which involves laryngeal and respiratory movements, complex mammalian vocalizations are primarily characterized by articulation. This encompasses movements of the tongue, lips, and jaw that modulate the shape of the vocal tract, which leads to 1060

8 MARCH 2024 • VOL 383 ISSUE 6687

changes in the spectral characteristics of the vocal output. As the authors point out, mouse vocalization is largely produced with phonation and lacks articulatory movements, making mice an ideal model organism to study the laryngeal component. Therefore, the mouse model established by Park et al. could allow investigation of the neural control of other phonatory aspects of vocalization, such as call frequency, duration, and amplitude. On the basis of their findings, it appears that mouse vocal production requires more than RAm activity because this does not control call frequency. The ventrolateral reticular formation in the pontine brainstem, another crucial part of the vocal pattern–generating network that also projects to the laryngeal motoneuron pool (4), has been shown to be involved in call frequency control in several mammals, including mice (4, 12). It will be intriguing to learn about the precise role of this structure in vocal pattern generation by using the approach established by Park et al. together with other recently developed rodent model systems (8, 13) in the near future. Insights into the neural mechanisms that control vocalization and breathing have potential applications in understanding speech disorders and respiratory dysfunction. In particular, human speech—one of the most complex mammalian vocal behaviors—involves laryngeal-articulatory coupling, which is crucial for speech production (14). Such coupling has recently been shown to be present in nonhuman primates (15). These findings suggest that future studies should investigate how complex temporal patterns as well as articulation are orchestrated with the larynx during complex vocalizations in other mammals, including primates. j REFERENCES AND NOTES

1. K. L. Christopher et al., N. Engl. J. Med. 308, 1566 (1983). 2. J. M. Hintze, C. L. Ludlow, S. F. Bansberg, C. H. Adler, D. G. Lott, Otolaryngol. Head Neck Surg. 157, 558 (2017). 3. J.-M. Ramirez, N. A. Baertsch, Annu. Rev. Neurosci. 41, 475 (2018). 4. S. R. Hage, A. Nieder, Trends Neurosci. 39, 813 (2016). 5. J. Park et al., Science 383, 1074 (2024). 6. U. Jürgens, Neurosci. Biobehav. Rev. 26, 235 (2002). 7. O. K. Faull, H. H. Subramanian, M. Ezra, K. T. S. Pattinson, Neurosci. Biobehav. Rev. 98, 135 (2019). 8. K. Tschida et al., Neuron 103, 459 (2019). 9. M. Smotherman, K. Kobayasi, J. Ma, S. Zhang, W. Metzner, J. Neurosci. 26, 4860 (2006). 10. J. W. Arthurs, A. J. Bowen, R. D. Palmiter, N. A. Baertsch, Nat. Commun. 14, 963 (2023). 11. J. D. Moore, D. Kleinfeld, F. Wang, Trends Neurosci. 37, 370 (2014). 12. K. Hartmann, M. Brecht, iScience 23, 101804 (2020). 13. M. Concha-Miranda, W. Tang, K. Hartmann, M. Brecht, J. Neurosci. 42, 8252 (2022). 14. P. F. MacNeilage, Behav. Brain Sci. 21, 499, discussion 511 (1998). 15. C. Risueno-Segovia, S. R. Hage, Curr. Biol. 30, 4276 (2020). 10.1126/science.ado2114

EVOLUTION

Amphibian hatchlings find mother’s milk Egg-laying amphibian females produce lipid-rich “milk” to feed offspring after hatching By Marvalee H. Wake

D

uring their evolution, amphibians have performed diverse and sometimes bizarre “natural experiments” in reproductive biology. Most members of the class, which includes frogs, salamanders, and caecilians, retain certain ancestral features, including courtship, external fertilization, egg laying (oviparity), hatching of larvae, metamorphosis in water, and emerging terrestrially. However, many frogs, salamanders, and especially caecilians have developed distinct and complex modes of egg protection and nutrition. Brooding developing embryos in pouches or pits on a maternal female’s back, in her stomach, in her oviducts, or in a paternal male’s vocal sacs, and eating of a mother’s skin are but a few examples. On page 1092 of this issue, Mailho-Fontana et al. (1) report that females of a species of oviparous Brazilian caecilian, Siphonops annulatus, provide “milk” to their offspring during parental care after hatching. This previously unobserved form of maternal provisioning challenges existing understanding of the evolution of parental care modes. Caecilians, members of the order Gymnophiona, are less well studied than frogs and salamanders, although interest in these enigmatic amphibians is rapidly growing, with increasing field work, observation, and research into multiple dimensions of their biology. Basal amphibian taxa share most elements of the ancestral reproductive traits, yet all caecilians carry out internal fertilization, which facilitates interesting natural experiments. For example, retention of fertilized ova in the maternal oviducts and provision of secreted nutrition before giving birth to fully metamorphosed juveniles occur in one species of frog and one species of salamander, but are found in several different families of caecilians. Why science.org SCIENCE

PHOTO: CARLOS JARED/COPYRIGHT UR

this mode of maternal provisioning has The study by Mailho-Fontana et al. opens seemingly remarkable species such as arisen in only 2 of 8500 species of frogs and new areas of research for caecilians and for S. annulatus provide the “raw material” salamanders, but more than 20 of the 222 amphibian biology in general. It also proof evolutionary biology. They contribute species of caecilians, remains unclear (2). vides an expanded approach to investigate to broadening and bettering understandThe mechanism of providing nutrients the evolution of derived modes of reproing of the scope and role of evolution. Rein the form of oviductal “milk” evolved induction in the broadest sense, and to betsearch that combines inquiries at various dependently several times in amphibians. ter understand key aspects of evolutionary levels of the hierarchy of biology, whether It has long been thought that the embryos biology. However, several parts of the story molecular, cellular, organismal, or ecosyshatch in the oviducts and consume the milk are still missing. For example, the genetic tem scale, within a framework of robust until they are born fully metamorphosed, and endocrinological factors that support phylogenetic hypotheses can provide new after months or even years of gestation. both dermatophagy and milk production and powerful ways to study evolutionary The species of frog that uses this approach in S. annulatus, or the relevant ecological biology and generate insights that cut has a 9-month gestation period, which is parameters that influence milk production across kingdoms of life. For example, it strongly correlated with enis not yet known how, why, vironmental seasonality (3). when, or where “amphibMeanwhile, the salamander ian milk” came into being, that produces oviductal milk or which evolutionary forces has a flexible gestation peinfluenced caecilian biochemriod that can vary between istry, morphology, endocrinol2 and 5 years, depending on ogy, or behavior to make this environmental factors (4). adaptation possible. FurtherThe gestation periods of caemore, why the composition cilian species are also flexible, of the milk is so like that of changing according to their mammalian species remains environment (5). unclear. However, answering Several oviparous caecilian these questions will provide species are known to exhibit new ways to consider what dermatophagy, in which the evolutionary constraints apply maternal female lays her ferto parental care mechanisms, tilized eggs on land and stays and how new innovations dewith her clutch until they hatch velop across the tree of life. and eat patches of her proliferA research mindset that ated, lipid-rich skin (6). Widely emphasizes exploring the muldistributed in Brazil, S. annutiple facets of the biology of latus is one of these species. organisms like caecilians will However, as Mailho-Fontana et enable better understanding Hatchlings of the Brazilian caecilian, Siphonops annulatus, seek their al. have now discovered, S. anof the patterns and processes mother’s “milk” in an unusual innovation in parental care. nulatus is exceptional among of evolution and the adapamphibians in being the first tive capacities that maintain known species in which oviparous mothers by this burrowing (fossorial) species of amlineages, which give them flexibility and also provide oviductal “milk,” which is prephibian, are poorly understood. Similarly, resilience and that influence potential sumably generated just as in live-bearing the neurobiology underpinning mother– extinction. Observations of natural ex(viviparous) species. The authors present hatchling communication and milk release, periments reveal what can happen, but extensive evidence that hatchlings gather to as well as behaviors such as mate recognirarely how it happens. Using the sort of receive milk at the mother’s vent (the outtion, deserves investigation. Furthermore, integrative approach employed by Mailholet through which intestinal and urogenital the evolutionary forces that drove the Fontana et al. to study such phenomena output is extruded) and that through both physiological, biochemical, ecological, and will be increasingly vital to understanding physical contact and by making noises, they behavioral changes that have resulted in evolution and its bearing on the survival of stimulate the mother to release her milk, this innovation in parental care remain to biodiversity on this planet. j which is rich in lipids and carbohydrates, be elucidated. REFERENCES AND NOTES similar to mammalian milk. By coupling At face value, the research by Mailho1. P. L. Maihlo-Fontana et al., Science 383, 1092 (2024). video and audio recordings with measurFontana et al., and studies like it, could be 2. AmphibiaWeb, University of California, Berkeley, (2024); https://amphibiaweb.org. able physical characterization, as well as viewed simply as descriptions of curious 3. F. Xavier, “An exceptional reproductive strategy in Anura: biochemical and histological analyses, the phenomena. However, there are numerNectophrynoides occidentalis Angel (Bufonidae), an authors provide compelling support for this ous ways in which they have a far greater example of adaptation in terrestrial life by viviparity” in newly discovered behavior. They speculate context and relevance. For example, they Major Patterns of Vertebrate Evolution, M. K. Hecht, P. C. Goody, B. M. Hecht, Eds. (Plenum, 1977), pp. 545. that this is a mode of parent–offspring comdemonstrate the value in adopting an in4. G. D. Guex, H. Greven, Mertensiella 4, 161 (1994). munication to solicit feeding, a mechanism tegrative approach, bringing diverse ex5. M. H. Wake, “How do homoplasies arise? Mechanisms of not known for any other amphibians and pertise and subject matter knowledge to evolution of derived modes of reproduction in amphibians” in Great Transformations in Vertebrate Evolution, an important advance in understanding the investigate an interesting biological quesK. Dial, N. Shubin, B. Brainerd, Eds. (Univ. of California evolution of parental care modes. tion or observation, which will invariably Press, 2015), pp. 373. have multiple dimensions and depths of 6. M. Wilkinson et al., Biol. Lett. 4, 356 (2008). Department of Integrative Biology and Museum of 7. M. H. Wake, Integr. Comp. Biol. 43, 239 (2003). complexity (7). Furthermore, studies of Vertebrate Zoology, University of California, Berkeley, Berkeley, CA, USA. Email: [email protected] 10.1126/science.ado2094 the natural experiments undertaken by SCIENCE science.org

8 MARCH 2024 • VOL 383 ISSUE 6687

1061

INSIGHTS

P OLICY FORUM ENVIRONMENTAL ECONOMICS

Accounting for the increasing benefits from scarce ecosystems As people get richer, and ecosystem services scarcer, policy-relevant estimates of ecosystem value must rise By M. A. Drupp1, M. C. Hänsel2,3, E. P. Fenichel4, M. Freeman5, C. Gollier6, B. Groom7,8, G. M. Heal9, P. H. Howard10, A. Millner11, F. C. Moore12, F. Nesje13, M. F. Quaas2,14, S. Smulders15, T. Sterner16, C. Traeger17, F. Venmans8

G

overnments are catching up with economic theory and practice by increasingly integrating ecosystem service values into national planning processes, including benefitcost analyses of public policies. Such analyses require information not only about today’s benefits from ecosystem services but also on how benefits change over time. We address a key limitation of existing policy guidance, which assumes that benefits from ecosystem services remain unchanged. We provide a practical rule that is grounded in economic theory and evidence-based as a guideline for how benefits change over time: They rise as societies get richer and even more so when ecosystem services are declining. Our proposal will correct a substantial downward bias in currently used estimates of future ecosystem service values. This will help governments to reflect the importance of ecosystems more accurately in benefit-cost analyses and policy decisions they inform. Besides nature’s intrinsic value, ecosystems provide diverse benefits to humans (1, 2). People regularly exchange goods derived from ecosystem services, such as fruits, fish, and timber, in market economies and can see the values of those benefits in the market prices they pay. Ecosystems also provide nonmarket goods or services that have real value to humans without involving market transactions. Examples include water and air purification by forests, soil nutrient cycling by earthworms, the enjoyment of natural areas through recreation or aesthetic appreciation,

and the importance people attach to the existence of diverse species (1, 2). Although there are many philosophical and practical challenges involved in putting dollar values to ecosystem services, the main motivation for doing so is that policy processes require an analysis of trade-offs—for instance, using benefit-cost analyses. Here, the absence of a monetary value is often equated to having no value at all, which leads society to underinvest in healthy ecosystems. The benefits of these nonmarket goods can be assessed in monetary terms using “shadow” prices (2). One can estimate current shadow prices from information on current marginal “willingness to pay” (WTP) for changes in ecosystem services. WTP for ecosystem services can be estimated with nonmarket valuation techniques using revealed consumer behavior (e.g., in housing markets, travel behavior, or donations) or surveys (3, 4). Governments are making progress integrating the value of ecosystem services in policy planning frameworks as they implement the Global Biodiversity Framework under the United Nations (UN) Convention of Biological Diversity and work towards the UN Sustainable Development Goals. Policy guidance on benefit-cost analysis already recognizes the principle of relative scarcity. For example, as real income grows, the benefits that people derive from their health or from reduced travel time grows and policy guidelines account for this (5). Yet, with few notable exceptions (6), the changing benefits from scarce ecosystem services have so far been overlooked in policy guidance.

One of the barriers to including ecosystem services in benefit-cost analysis is the lack of a straightforward approach for adjusting future WTPs in response to growing real incomes and changing scarcities. Several recent initiatives have put the issue on the policy agenda. The UK Treasury recently convened an expert Working Group to develop guidance on this matter (7) (authors M.A.D., M.F., C.G., B.G., A.M., and T.S. were members). The US Office of Management and Budget (OMB) proposed guidance on “Assessing Changes in Environmental and Ecosystem Services in Benefit-Cost-Analysis” (authors E.P.F. and F.C.M. contributed while seconded to the government). The US National Science and Technology Council has established a Subcommittee on Frontiers of Benefit-Cost Analysis, which has flagged ecosystem services effects as a priority. These movements reflect opportunities to rectify how governments account for ecosystem services in regulatory guidance and policy decisions. We propose a simple and transparent rule for estimating future WTPs that can be applied independently of how current WTP is estimated. INCREASING RELATIVE SCARCITY

Although real incomes, and thus the consumption of market goods, continue to grow—reflected in real per capita gross domestic product (GDP) growth of around 2% per year (8)—the supply of ecosystem services is far from keeping pace. Many ecosystem services are in decline because of habitat destruction, over-harvesting, and climate change (2). Global forest areas and populations of threatened species are on a downward trend. Even if nature is preserved in current conditions (denoted as “Environmental Stagnation”), ecosystem services would become scarcer relative to real income or market goods, both of which continue to grow (see the figure, first panel). Rising real incomes coupled with a stagnation or decline of ecosystem services means that the benefits society derives from scarce ecosystem services, measured in WTP, increase over time. This is conceptually similar to how WTP for ecosystem services increases with income across individuals, from poor to rich, at any point in time (4, 9, 10): If people get richer, they want to spend more on all types of goods and services, such as additional Netflix shows and trips to natural parks. The market responds by supplying more TV

1Department of Economics and Center for Earth System Research and Sustainability (CEN), University of Hamburg, Hamburg, Germany. 2Faculty of Economics and Management Science, Leipzig

University, Leipzig, Germany. 3Potsdam Institute for Climate Impact Research (PIK), Leibniz Association, Potsdam, Germany. 4School of the Environment, Yale University, New Haven, CT, USA. for Business and Society, University of York, York, UK. 6Toulouse School of Economics, Toulouse, France. 7Dragon Capital Chair of Biodiversity Economics, LEEP Institute, Department of Economics, University of Exeter, Exeter, UK. 8Grantham Research Institute on Climate Change and the Environment, London School of Economics and Political Science, London, UK. 9Columbia Business School, New York, NY, USA. 10Institute for Policy Integrity, New York University School of Law, New York, NY, USA.11Department of Economics, University of California (UC)–Santa Barbara, Santa Barbara, CA, USA. 12Department of Environmental Science and Policy, UC-Davis, Davis, CA, USA. 13Department of Economics, University of Copenhagen, Copenhagen, Denmark. 14German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Leipzig, Germany. 15Department of Economics, Tilburg University, Tilburg, Netherlands. 16Department of Economics, University of Gothenburg, Gotheenburg, Sweden. 17Department of Economics, University of Oslo, Oslo, Norway. Email: [email protected]

5School

1062

8 MARCH 2024 • VOL 383 ISSUE 6687

science.org SCIENCE

Scarce ecosystems and increasing ecosystem-service values Several estimates of ecosystem services trends are used, reflected by global forest area, species conservation status according to the International Union for the Conservation of Nature (IUCN) Red List Index for threatened species, and biodiversity as reflected by the Living Planet Index. See supplementary materials for details and further analyses. Environmental stagnation

Threatened species

Biodiversity

Relative to growth in market goods [or real income, reflected by gross domestic product (GDP) per capita], there is increasing scarcity of many ecosystem services, all of which are on a declining trend that is projected forward (10). An “Environmental stagnation” scenario reflects ecosystem services remaining unchanged.

The relative price change (RPC) rule maps growth rates into yearly relative price adjustments against the rate at which “willingness to pay” (WTP) for ecosystem services changes with income, i.e., the income elasticity of WTP (j). We contrast the current, old, default (j = 0), and the proposed new default (j = 1). 10

700 500

GDP per capita Growth rate (gc) = 2%

Ecosystem services growth rates (gE) 0.0% –0.1% –0.4% –2.8%

300 100 0

20

40 60 Years (t)

80

8 RPC (%)

Normalized growth (varied units)

2 Yearly relative price change

6

RPC | j = 1

j=1 RPC = gC − gE

RPC = j·(gC − gE) j=0 RPC = 0

2.00% 2.11% 2.42% 4.84%

4 2 0

100

0.0

0.5 1.0 1.5 Income elasticity of WTP (j)

2.0

The RPC adjustment increases the present value of ecosystem services over a century compared to current government guidance on benefit-cost analysis. The increase of 131% in the “Environmental stagnation” scenario captures the real income effect that is common to all ecosystem service values.

2100

WTP(t) = WTP(0)·eRPC·t

WTP calculated with RPC | j = 1

1600 1100 600

WTP calculated with RPC | j = 0

100 20

40 60 Years (t)

80

shows. Nature, however, does not respond to people’s demand. If natural parks do not expand, people may value them more in the future, expressing a higher WTP to extend or preserve these parks. This increase in WTP for the same ecosystem service is thus due to ecosystem services becoming scarcer relative to market goods (or real income). WTPs will increase even more if ecosystem services are declining, such as coral reefs or threatened species, and thus become absolutely scarcer. Estimates of future WTPs that do not reflect the increasing relative scarcity of ecosystem services due to growing incomes, or the changing real scarcity due to ecosystems loss, will systematically undervalue the ongoing contribution of ecosystem services to society. As a result, the increasing importance of the natural environment for future generations will be overlooked and society will underinvest in measures to safeguard nature (9–11). Economic theory provides a path for governments to reflect the changing relative and absolute scarcity of ecosystem services in benefit-cost analysis (9–14). To derive a simple rule for estimating future WTPs using relative price change (RPC) adjustments, we follow the standard constant elasticity frameSCIENCE science.org

100

Present value increase (%)

4 Increase in present value of ecosystem services

WTPs increase over time when applying the RPC adjustments using the new default (j = 1) from panel 2. Future WTPs for stagnating ecosystem services would rise in proportion with real income. For declining ecosystem services, future WTPs would rise faster. Normalized WTP ($)

3 Evolution of willingness to pay

0

GRAPHIC: M. HERSHER/SCIENCE

Forest area

1 Projected real growth

2000

j=1

Increase in value | j = 1

1500

131% 145% 188% 1213%

1000 500 0 0.0

work that typically underpins guidance on benefit-cost analysis [see supplementary materials (SM)]. The RPC adjustment of future WTPs depends on the rate at which WTP for ecosystem services changes with income (the income elasticity of WTP, denoted by j). In the standard framework (9–15), the income elasticity of WTP is directly related to the degree to which people consider market and nonmarket goods as complementary, rather than as substitutes for one another (see SM): If people feel that market goods provide a good substitute for, say, a walk in a national park, then the elasticity j is low. Conversely, a high degree of complementarity implies a high j (9–11, 15). The more ecosystem services serve as complements to, rather than substitutes for, market goods, the faster WTP for ecosystem services rises as income grows. The RPC adjustment depends on two ingredients that interact with the income elasticity. First, it depends on the growth rate of market consumption goods, gC, measured as GDP per capita. As real incomes grow, the larger is the budget that people can spend on any good, market and nonmarket. This describes the “real income effect” (j 3 gC). Second, the adjustment depends on the

0.5 1.0 1.5 Income elasticity of WTP (j)

2.0

growth rate of ecosystem services, gE. WTPs for ecosystem services rise more when ecosystems are in decline, and thus do not only become scarcer relative to market goods but also scarcer in absolute terms. This describes the “real scarcity effect” (−j 3 gE). Combining both effects yields the RPC rule: RPC = j 3 [gC − gE] (see the figure, second panel). When growth rates remain constant, WTPs increase exponentially with the RPC (see the figure, third panel). Thus, as incomes grow and ecosystems decline, the benefits from ecosystem services reflected in policy analysis must rise. To this end, policy guidance should incorporate the RPC rule to adjust estimates of future WTPs for scarce ecosystem services. The first step for integrating the RPC rule into policy guidance is to account for the real income effect. This is already routine practice for other nonmarket goods, such as health or travel time. In a functioning market, when the demand for a good increases because of greater wealth in the economy, firms have an incentive to produce more of the good. Such an increase in supply counteracts the price increase. But ecosystems do not respond to (shadow) prices. It is the job of policy to account for the real income effect. The second 8 MARCH 2024 • VOL 383 ISSUE 6687

1063

INSIGHTS | P O L I C Y F O RU M

step is to account for the real scarcity effect. Our proposal relates closely to two standard concepts in benefit-cost analysis: discounting (a method to make future monetary benefits or costs comparable with today’s) and benefit transfer. First, an alternative to estimating future WTPs adjusted for relative price changes is to instead use different discount rates for ecosystem services and market goods (9–14). This, however, would also require changing the standard discount rate. The alternative that we propose here, which is mathematically equivalent (see SM), is to adjust future WTPs and use a single discount rate schedule. This proposal is simpler, more transparent, and often more compatible with how guidelines deal with other nonmarket goods (5–7) (see SM). Second, benefit-cost analysis routinely draws on benefit transfer to estimate missing WTPs, using WTP estimates from a study site to transfer or scale it to another geographical setting. Benefit transfer “in space” commonly adjusts for differences in average incomes across locations (15). The RPC rule can be thought of as a dynamic extension to perform benefit transfer “in time,” adjusting past or current WTP estimates to future dates where real incomes and real scarcities have changed. A NEW DEFAULT FOR POLICY AND ACTION Most current policy guidance implicitly assumes that WTP for ecosystem services does not increase with income over time (j = 0). This ignores both income and scarcity effects—in contrast to empirical evidence (4, 9, 10). It also produces an inconsistency in the treatment of nonmarket goods when adjustments for real income effects are considered for time and health effects but not ecosystem services. We propose to shift policy guidance to a new default, in which benefits from ecosystem services are considered to increase proportionally with real income or the consumption of market goods (j = 1). This strikes a balance between indirect evidence from nonmarket valuation studies, yielding elasticity estimates of around 0.4 to 0.8 (4, 9, 10), and expert judgments that employ values of up to 2 (9, 11), and accords with what governmental bodies use for valuing reductions in mortality risk (5) or travel time. Under the new default, WTPs for stagnating ecosystem services would rise in proportion with real income (see the figure, third panel). For declining ecosystem services, WTPs would rise faster, accounting for larger absolute scarcity of ecosystems (see the figure, third panel). The fourth panel of the figure illustrates how shifting from current valuation practice to our proposed RPC rule affects today’s value of ecosystem changes over a century. We compare the present value (i.e., the discounted sum) of future WTPs using the new default 1064

8 MARCH 2024 • VOL 383 ISSUE 6687

from the third panel of the figure for RPC adjustments to the present value of unadjusted WTPs as in current policy guidance (see SM). Against the backdrop of expected increases in real incomes, we first consider the case of “Environmental Stagnation.” Here, a proportional increase of WTP with real income—the new proposed default—results in WTP for ecosystem services increasing by 2% per year (see the figure, second panel). Considering adjustments to future WTPs over a century (see the figure, third panel), at a discount rate of 2% as in the US OMB Circular A-4, the RPC rule adjustment yields an increase in the present value of ecosystem services of 131% (see the figure, fourth panel). Projecting forward the decline rate of global forest areas, species conservation status according to the International Union for the Conservation of Nature (IUCN) Red List Index for threatened species, or biodiversity according to the Living Planet Index, the increase in present values would be more than 140, 180, and 1200%, respectively (see the figure, fourth panel). Accounting for the effects of growing real income and increasing real scarcities of ecosystems thus clearly matters and will make projects that have long-term positive effects on ecosystem services more attractive. In a benefit-cost analysis of climate change, for instance, neglecting such relative price changes underestimates the social cost of carbon (an estimate of the cost of damage resulting from each additional ton of carbon emissions) by more than 50% (9). To put this shift in guidance into action, we recommend that governments in their policy analyses immediately start accounting for the real income effect with a proportional increase of WTPs as real incomes grow. Focusing on the real income effect is a pragmatic starting point, as it is common for all ecosystem service benefits and closely aligned with how guidelines commonly value benefits of travel time reductions and of health (4). Forecasts for GDP growth are also available (8), whereas forecasts for ecosystem services require further research (10). Real scarcity effects should be integrated whenever forecasts for ecosystem services are available. Policy guidance should be periodically revised as more evidence becomes available. Governments may consider creating advisory groups (7) to distil evidence on income and scarcity effects, including growth rates of various ecosystem services, and to inform setting income or substitution elasticities, which may vary across ecosystems and geographies. Elasticities are likely heterogeneous, and estimates of elasticities and growth rates are also inherently uncertain (13). Furthermore, ecosystem services are affected by expanding economies, yet they also provide inputs to producing market goods. Their increasing

scarcity may thus also change the growth rate of GDP (14). Future refinements should seek to reflect these complexities.  Our proposal helps level the playing field so that ecosystem services are treated more consistently with other goods, whose (shadow) prices, or WTP estimates, change over time. As governmental guidelines in Germany, the UK, and the US are undergoing major updates, our proposal would help governments operationalize guidance on assessing the changing values of ecosystem services. Applying a simple relative price change rule, as we propose here, would ensure that the importance of scarce ecosystems for future generations is appropriately reflected when deliberating over public investments, evaluating regulatory change, and meeting sustainability requirements. j REFERENCES AND NOTES

1. IPBES, “Global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services,” E. S. Brondizio, J. Settele, S. Díaz, H. T. Ngo, Eds. (IPBES secretariat, 2019). 2. P. Dasgupta, The Economics of Biodiversity: The Dasgupta Review (HM Treasury, 2021). 3. R. C. Bishop et al., Science 356, 253 (2017). 4. J. B. Jacobsen, N. Hanley, Environ. Resour. Econ. 43, 137 (2009). 5. US Department of Transportation, Departmental Guidance: Treatment of the Value of Preventing Fatalities and Injuries in Preparing Economic Analyses (2021); https://www.transportation.gov/sites/dot.gov/ files/2021-03/DOT%20VSL%20Guidance%20-%20 2021%20Update.pdf. 6. B. Groom, M. A. Drupp, M. C. Freeman, F. Nesje, Annu. Rev. Resour. Econ. 14, 467 (2022). 7. H. M. Treasury, Environmental Discount Rate Review: Conclusion (HM Treasury, 2021). 8. U. K. Müller, J. H. Stock, M. W. Watson, Rev. Econ. Stat. 104, 857 (2022). 9. M. A. Drupp, M. C. Hänsel, Am. Econ. J. Econ. Policy 13, 168 (2021). 10. M. A. Drupp, Z. Turk, B. Groom, J. Heckenhahn, arXiv:2308.04400 [econ.GN] (2023). 11. T. Sterner, U. M. Persson, Rev. Environ. Econ. Policy 2, 61 (2008). 12. G. M. Heal, Valuing the Future: Economic Theory and Sustainability (Columbia Univ, Press, 1998). 13. C. Gollier, J. Environ. Econ. Manage. 94, 54 (2019). 14. X. Zhu, S. Smulders, A. de Zeeuw, J. Environ. Econ. Manage. 98, 102272 (2019). 15. S. Baumgärtner, M. A. Drupp, J. N. Meya, J. M. Munz, M. F. Quaas, J. Environ. Econ. Manage. 85, 35 (2017). ACKNOWLEDGMENTS

M.A.D. acknowledges support by the German Federal Ministry of Education and Research (BMBF) under grant no. 01UT2103B, and from the DFG Excellence Strategy (EXC 2037 and CLICCS) project no. 390683824, contribution to the Center for Earth System Research and Sustainability (CEN) of Universität Hamburg. B.G. is funded by Dragon Capital. E.P.F. is funded by the Knobloch Family Foundation. F.N. acknowledges financial support from DFF (ref. 3126-00011B). T.S. acknowledges support from the Mistra program Biopath and the Kamprad Stiftelse. F.V. is funded by Grantham Institute and CCCEP (ref. ES/R009708/1). C.G. acknowledges funding from the Agence Nationale de la Recherche (ref. ANR-17-EURE-0010). F.V. and B.G. acknowledge financial support from the BIOADD NERC grant (ref. NE/X002292/1). The contributions by E.P.F. and F.C.M. were made after leaving the White House. SUPPLEMENTARY MATERIALS

science.org/doi/10.1126/science.adk2086

10.1126/science.adk2086

science.org SCIENCE

The NOMIS & Science Young Explorer Award recognizes and rewards early-career M.D., Ph.D., or M.D./Ph.D. scientists that perform research at the intersection of the social and life sciences. Essays written by these bold researchers on their recent work are judged for clarity, scientific quality, creativity, and demonstration of cross-disciplinary approaches to address fundamental questions. A cash prize of up to USD 15,000 will be awarded to essay winners, and their engaging essays will be published in Science. Winners will also be invited to share their work and forward-looking perspective with leading scientists in their respective fields at an award ceremony.

Apply by May 15, 2024

The European Union must ensure that all traded wildlife, such as these tarantulas, is documented and legally and sustainably sourced.

Edited by Jennifer Sills

Reform wildlife trade in the European Union Wildlife trade affects a wide variety of species (1), and trade numbers are growing globally (2). The European Union (EU) is a major hub for wildlife trade (3), but its wildlife trade regulations are primarily based on the adoption and adaptation of provisions written by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES). In 2022, the EU adopted a welcome new action plan against wildlife trafficking (4), but its measures fall short of ensuring that wildlife traded within the EU is thoroughly documented and sourced legally and sustainably. The EU Trade in Wildlife Information Exchange (EU-TWIX) database (5) provides information about illegal trade, but the data are limited to CITES-listed species, excluding many traded species (6, 7). Building on the 2022 plan, which focuses on the enforcement of existing legislation, the EU should expand the EU-TWIX database to cover all wildlife trade, including all species, legal and illegal, and regulated and unregulated. The data collected should adhere to FAIR (findable, accessible, interoperable, reusable) principles (8). Other countries, such as the US (9), already track all wildlife trade, and the EU should follow that example. The EU Action Plan also overlooks the possibility of illegally sourced and trafficked species legally entering EU borders. Such species are subsequently traded within the EU and often later exported. To address this issue, the EU should implement a regulation comparable to the US Lacey Act, which prohibits imports that violate the laws of a specimen’s country of origin (10). This change could fill the gaps 1066

8 MARCH 2024 • VOL 383 ISSUE 6687

in current regulatory frameworks and also aid in the preservation of biodiversity in nations that, owing to resource limitations, lack effective enforcement. Finally, many legally traded species face overlooked risks, with little or no evidence of sustainable trade practices. Unsustainable trade poses potential threats to wild populations and human health (11). EU authorities should require evidence of the sustainable trade of all imported species and populations. Although identifying sustainability is often more challenging than determining legality, information about how trade affects legally traded species (which are not listed by CITES) can be obtained by consulting the International Union for Conservation of Nature Red List (12) or species experts, many of whom are already in the EU. The EU is a global hub for unregistered, illegal, and unsustainable activities. The recent action plan is a step forward, but tracking all species, observing the laws of the species’ home countries, and ensuring sustainability of legal traffic would transform the region into a model of environmental protection in relation to wildlife trade. Pedro Cardoso1,2*, Caroline S. Fukushima2,3, Armela Maxhelaku4, Peter Poczai5,6, Miguel Porto7,8,9, Andrius Puksas10, Luís Reino7,8,9, Indrek Saar11,12, Oliver Stringham13, Adam Toomes14, Thiago Vargas15, Diogo Veríssimo16 1Centre

for Ecology, Evolution and Environmental Changes & CHANGE–Institute for Global Changes and Sustainability, Faculdade de Ciências, Universidade de Lisboa, Lisboa, Portugal. 2Laboratory for Integrative Biodiversity Research, Finnish Museum of Natural History Luomus, University of Helsinki, Helsinki, Finland. 3Biodiversity and Sustainability Solutions Lab, University of Turku, Turku, Finland. 4Faculty of Law, University of Tirana, Tirana, Albania. 5Botany and Mycology Unit, Finnish Museum of Natural History Luomus, University of Helsinki, Helsinki, Finland. 6Museomics Research Group, Helsinki Institute of Life Science, Helsinki, Finland. 7Centro de Investigação em Biodiversidade e Recursos Genéticos (CIBIO), Research Network in Biodiversity and Evolutionary Biology (InBIO)

Laboratório Associado, Instituto Superior de Agronomia, Universidade de Lisboa, Lisboa, Portugal. 8CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, InBIO Laboratório Associado, Campus de Vairão, Universidade do Porto, Vairão, Portugal. 9Program in Genomics, Biodiversity, and Land Planning, CIBIO, Campus de Vairão, Vairão, Portugal. 10Vytautas Magnus University, Kaunas, Lithuania. 11Financial College, Estonian Academy of Security Sciences, Tallinn, Estonia. 12School of Governance, Law and Society, Tallinn University, Tallinn, Estonia. 13Rutgers Climate and Energy Institute, Rutgers, The State University of New Jersey, New Brunswick, NJ, USA. 14Invasion Science and Wildlife Ecology Group, School of Biological Sciences, The University of Adelaide, Adelaide, SA, Australia. 15Brazilian Network to Fight Wildlife Trafficking (RENCTAS), Brasilia, Brazil. 16Department of Biology, University of Oxford, Oxford, UK. *Corresponding author. Email: [email protected] REFERENCES AND NOTES

1. C. S. Fukushima, S. Mammola, P. Cardoso, Biol. Conserv. 247, 108503 (2020). 2. M. Harfoot et al., Biol. Conserv. 223, 47 (2018). 3. M. Halbwax, Biol. Conserv. 251, 108798 (2020). 4. European Commission, “Revision of the EU action plan against wildlife trafficking” (2022); https://eur-lex. europa.eu/legal-content/EN/TXT/ ?uri=COM%3A2022%3A581%3AFIN& qid=1667989438184. 5. EU Trade in Wildlife Information Exchange (2024); www. eu-twix.org. 6. B. M. Marshall, C. T. Strine, A. C. Hughes, Nat. Comm. 11, 4738 (2020). 7. B. M. Marshall et al., Commun. Biol. 5, 448 (2022). 8. M. Wilkinson et al., Sci. Data 3, 160018 (2016). 9. E. A. Eskew et al., Sci. Data 7, 22 (2020). 10. L. Slobodian, A. Chatziantoniou, Forum Crime Soc. 9, 43 (2018). 11. A. Hughes et al., J. Environ. Manag. 341, 117987 (2023). 12. D. W. S. Challender et al., Nat. Ecol. Evol. 7, 1211 (2023). COMPETING INTERESTS

D.V. is director of Oxford Biodiversity Research, a research consulting firm. 10.1126/science.ado1142

Incorporate ethics into US public health plans In November 2023, the Centers for Disease Control and Prevention (CDC) concluded its public comment period on a new framework (1) to coordinate the US response to public health and zoonotic disease challenges. The proposed plan will establish an interagency One Health program that brings practitioners and academics together to improve biodefense and pandemic preparation (2, 3). The initiative rightly recognizes the importance of broad scientific expertise, but it does not sufficiently incorporate public health ethics. Zoonotic disease management is characterized by uncertainty, knowledge deficits, and conflicting interests. Public science.org SCIENCE

PHOTO: IBAMA ARCHIVE

LET TERS

INSIGHTS

health–based ethical analysis can clarify the available courses of action and their impacts on humans, animals, and the environment. Each course of action favors one set of interests at the expense of others, and the resulting inequities may undermine societal support for zoonotic disease control. Public skepticism and internecine conflicts are particularly concerning during disease outbreaks, when time constraints, resource limitations, and uncertainty complicate high-stakes decision-making. Incorporating ethical analysis into US One Health zoonotic disease preparedness is vital. Other countries have applied ethicsoriented approaches that could serve as models. In Austria, the Federal Ministry of Health commissioned the development of an ethics-oriented decision-making tool to prepare veterinarians for managing zoonoses, including the consideration of animal welfare during emergency killing (4). The tool is now used in official training of state veterinarians in Austria (5). In the Netherlands, the Council on Animal Affairs, a strategic body advising the government on One Health policy assessment, includes members with ethics expertise who help reveal the value-based aspects of policy decisions that affect humans as well as animals (6). The UK Biological Security Strategy emphasizes a One Health approach, recognizing the need for assessing how ethical values apply in the governance of institutional and legal frameworks. The UK analysis process aims to identify the conflicting and commensurate values that underlie policies at the intersection of climate change, zoonotic disease preparedness, and animal welfare (7). The Austrian, Netherlands, and UK policies show how the incorporation of ethics into public health plans can benefit animal welfare and support collaborative, long-term multiagency and interdisciplinary disease management efforts among state and nongovernmental entities. Increased attention to ethics analysis in decision-making approaches for disease management also has the potential to rekindle waning public trust in science and in public health officials. Raymond Anthony1*, David S. Miller2, Donald E. Hoenig3, Kate M. Millar4, Jean Goodwin5, Wesley R. Dean6, Herwig Grimm7, Franck L. B. Meijboom8, Julia Murphy9, Elyse Persico Murphy10, H. Morgan Scott11, Andreia De Paula Vieira12 1Department

of Philosophy, University of Alaska Anchorage, Anchorage, AK, USA. 2Miller Veterinary Services, Loveland, CO, USA. 3Made in Maine Veterinary Consulting, Belfast, ME, USA. 4Centre for Applied Bioethics, School of Biosciences and School of Veterinary Medicine and Science, University of Nottingham, Nottingham, UK. 5Department of Communication, North Carolina State University, Raleigh, NC, USA. 6Department of Food and Resource Economics, University of Copenhagen, Copenhagen, Denmark. 7Messerli

SCIENCE science.org

Research Institute, Department of Interdisciplinary Life Science, University of Veterinary Medicine Vienna, Medical University of Vienna, University of Vienna, Vienna, Austria. 8Centre for Sustainable Animal Stewardship, Department of Population Health Sciences, Faculty of Veterinary Medicine, Utrecht University, Utrecht, Netherlands. 9Virginia Department of Health, Richmond, VA, USA. 10Webbing the Wild, Virginia Beach, VA, USA. 11Department of Veterinary Pathobiology, Texas A&M University, College Station, TX, USA. 12Curitiba, Brazil. *Corresponding author. Email: [email protected] REFERENCES AND NOTES

1. CDC, “National One Health Framework to address zoonotic diseases and advance public health preparedness in the United States: A framework for One Health coordination and collaboration across federal agencies,” Federal Register (2023); https://www.federalregister. gov/documents/2023/09/20/2023-20338/nationalone-health-framework-to-address-zoonotic-diseasesand-advance-public-health-preparedness-in. 2. “National biodefense strategy and implementation plan” (The White House, 2022). 3. “PREVENT Pandemics Act (P.L. 117-328, Division FF, Title II)” (Congressional Research Service, 2022); https:// crsreports.congress.gov/product/pdf/R/R47649. 4. K. Weich, C. Dürnberger, H. Grimm, Ethik in der amtstierärztlichen praxis: Ein Wegweiser (Harald Fischer Verlag, 2016) [in German]. 5. C. Thöne-Reineke, S. Hartnack, P. Kunzmann, H. Grimm, K. Weich, Berl. Munch. Tierarztl. Wochenschr. 133, 207 (2020) [in German]. 6. Netherlands Council on Animal Affairs, “One Health a policy assessment framework” (Raad voor Dierenaangelegenheden, 2015); https://english.rda.nl/ publications/publications/2016/02/08/one-health. 7. H. Wolmuth-Gordon, N. Mutebi, “Public health and climate change: A One Health approach” (UK Parliament POSTnote 701, 2023). 10.1126/science.adn7640

Mangrove forest decline on Iran’s Gulf coast Iran’s mangrove forests, which cover about 9370 ha along the Persian Gulf and the Sea of Oman, are home to distinct species (1–3), store carbon (1, 2), purify water (4), and support coastal communities (1). However, these forests face threats from human activities and climate change (5). Iran must protect its mangrove forests (1). Mangrove forests harbor distinctive habitats, containing plants that have adapted to oxygen-poor environments (1). The diverse plant and animal species that mangroves support sustain the local food chain (1). The roots of trees such as Avicennia marina and Rhizophora mucronata stabilize coastal sediments, offering shelter to wildlife and fostering fertile conditions (1). As a transitional zone between land and sea, mangrove forests provide optimal conditions for marine life such as plants, algae, bioluminescent organisms, and juvenile fish (1, 3). Mangroves, which store 30% of Earth’s carbon (6), also provide valuable ecosystem

services. By reducing wave action and water currents, they increase carbon sequestration and reduce resuspension of microplastics in the Persian Gulf’s surface waters (4, 7). Mangrove forests also provide economic, social, and recreational advantages to Iran’s coastal communities (5). Human activities have resulted in a decline in mangrove forests. Factory construction, wastewater discharge, oil pollutants, and improper land management have polluted the mangroves on the Persian Gulf shores (1). Mangroves also suffer from reduced freshwater flow, imbalances between freshwater and saline water, dam construction, agricultural development, shrimp farms, and discharge from desalination plants (1, 8). Despite the deterioration of this vital resource, Iran has failed to take coordinated protective action. Insufficient funding has been allocated to preserving mangrove forests, and unsustainable policies and limited regulations have exacerbated environmental degradation (1). Iran’s socioeconomic challenges, including poverty, unemployment, and inadequate rights and resources, have led to overexploitation and unsustainable use of mangrove forests and other natural resources (1, 9). To protect the Persian Gulf coast’s mangrove forests, Iran should strengthen and rigorously enforce the relevant environmental laws and improve socioeconomic conditions to reduce local residents’ reliance on them. The government should also partner with and support nongovernmental organizations that educate the public about the importance of mangroves and that work toward achieving mangroverelated conservation goals. Iran should base its policies on those of developed nations, including the European Union, that have implemented plans to completely prevent deforestation by 2030 (10). Hossein Yarahmadi1 and Zahra Khorsandi2* 1Department

of Chemical Engineering, Sirjan University of Technology, Sirjan, Iran. 2Department of Chemistry, Isfahan University of Technology, Isfahan, Iran. *Corresponding author. Email: [email protected] REFERENCES AND NOTES

1. M. Savari, H. E. Damaneh, H. E. Damaneh, J. Nat. Conserv. 66, 126153 (2022). 2. A. Goudarzi, M. Moslehi, Crop Prot. 128, 104987 (2020). 3. P. Hajializadeh et al., Front. Mar. Sci. 7, 575480 (2020). 4. C. Martin et al., Sci. Adv. 6, eaaz5593 (2020). 5. A. Asadi et al., J. Iran. Nat. Res. 61, 849 (2009) [in Farsi]. 6. S. Sjögersten et al., Geoderma 403, 115173 (2021). 7. C. M. Duarte et al., Nat. Clim. Chang. 3, 961 (2013). 8. H. Etemadi, J. M. Smoak, E. Abbasi, Oceanologia 63, 99 (2021). 9. H. Azadi, D. Samari, K. Zarafshani, G. Hosseininia, F. Witlox, Sustain. Sci. 8, 543 (2013). 10. M. Karimi, Political Spatial Planning 3, 205 (2021) [in Farsi]. 10.1126/science.ado0376 8 MARCH 2024 • VOL 383 ISSUE 6687

1067

RESEARCH

Mountainous regions with high erosion rates, such as the coastal mountains of eastern Taiwan pictured here, are likely a carbon dioxide source rather than sink.

IN S CIENCE JOURNAL S Edited by Michael Funk

GEOLOGY

Quantifying chemical weathering

T

he role of erosion in the modulation of atmospheric carbon dioxide has been challenging to untangle because increased weathering fluxes do not always result in the drawdown of this greenhouse gas. Bufe et al. found that carbon dioxide drawdown peaks at a range of erosion rates that are much lower than the most active mountain belts. Peak drawdown occurs in various places on the planet, but very actively eroding landscapes are likely contributing carbon dioxide to the atmosphere. This observation helps to reconcile conflicting data on the role of weathering and atmospheric carbon dioxide. —BG Science p. 1075, 10.1126/science.adk0957

Guide star–free quantum adaptive optics Adaptive optics provides a way to remove aberrations from imaging systems and has revolutionized imaging applications ranging from astronomy to microscopy. Typically, a “guide star” is required, and a deformable mirror or spatial light modulator corrects the distortions in the propagating wavefront, resulting in a sharp image. However, not all samples or imaging systems are amenable to a guide star. Cameron et al. show that entangled photons can be used in adaptive optics, demonstrating the ability to remove 1068

aberrations in biological samples without the need for a guide star. Such quantum adaptive optics enables guide star-free imaging of biological samples in the presence of aberrations. —ISO Science p. 1142, 10.1126/science.adk7825

ATTOSECOND SCIENCE

Attosecond x-ray imaging of liquid water The low intensity of attosecond pulses remains one of the main challenges in conducting true attosecond-pump/attosecondprobe experiments that could directly observe electron dynamics in real time when all

8 MARCH 2024 • VOL 383 ISSUE 6687

nuclear motions are frozen. Using a recently upgraded x-ray freeelectron laser that enables the production and synchronization of two-color attosecond x-ray pulse pairs, Li et al. performed all x-ray attosecond transient absorption spectroscopy measurements on an important condensed phase process: the initial electronic response to valence ionization in liquid water. This work experimentally resolves a long-standing debate around the interpretation of the 1b1 x-ray emission doublet. The presented technique could resolve many complex condensed phase issues for which nuclear motions need to be frozen to capture the

information content of the probe signal. —YS Science p. 1118, 10.1126/science.adn6059

CONSERVATION

Still in danger Over the past decade, the plight of the world’s sharks has received increasing attention, leading to increased regulation and finning bans. However, whether this increased attention has translated into improved outcomes for sharks is unclear. Finucci et al. found a need for increased regulations in their study of deep sea sharks and rays, which are experiencing declines due to increased fishing mortality, science.org SCIENCE

PHOTO: CHON KIT LEONG/ALAMY STOCK PHOTO

QUANTUM IMAGING

particularly when targeted for oil and meat. Increased regulations are urgent because the potential for most sharks to recover from such declines is limited due to their slow development and reproduction rates. —SNV

rings. Understanding the mechanistic basis of sister chromatid cohesion is especially important for understanding the etiology of age-related infertility and aneuploidy-driven diseases. —DJ

Edited by Caroline Ash and Jesse Smith

IN OTHER JOURNALS

Science p. 1122, 10.1126/science.adl4606

Science p. 1135, 10.1126/science.ade9121

INNATE IMMUNE SYSTEM NEUROSCIENCE

A pain in the bladder

Self-determined visual map

Recurrent urinary tract infections (rUTIs) are associated with urinary frequency and chronic pelvic pain even in the absence of bacteria. Hayes et al. examined the mechanisms of rUTIs and observed that rUTI patient bladder biopsies had an increase in the neuropeptide substance P in the lamina propria, a layer of the bladder made of connective tissue and containing blood vessels, nerves, and glands. This finding indicates an increase in pain-sensing nerves. In a mouse model of rUTIs, bladders showed striking sensory nerve sprouting that was linked to elevated nerve growth factor in monocytes and mast cells, which was associated with increased pain and urinary frequency. —CNF

Determining the mechanisms guiding axon pathfinding during development is critical for understanding how neuronal circuits develop in a dynamic milieu without fixed targets. Agi et al. studied the developing visual system in fruit flies to determine how thousands of axons form the intricate wiring pattern of the visual map. In contrast to the accepted idea of target-dependent guidance of axonal growth cones, the authors showed that depleting the target neurons does not prevent correct axonal growth. Axonal growth thus follows a self-organizing principle in the fly visual map. These results pave the way for further work in other systems to determine whether the self-organizing principle of axonal growth is a general phenomenon during brain development in the absence of fixed targets. —MMa Science p. 1084, 10.1126/science.adk3043

CELL BIOLOGY

IMAGE: JULIUS T CSOTONYI & THE CHINESE UNIVERSITY OF HONG KONG

One ring binds sister chromatids Cohesin is a ring-shaped multiprotein complex that not only organizes chromatin fibers by DNA loop extrusion, but also mediates the sister chromatid cohesion necessary for chromosome segregation. Ochs et al. determined the stoichiometry of cohesin by imaging single molecules bound to the chromatin of intact human cells. Although many complexes appeared dimeric, those engaged in sister chromatid cohesion were exclusively monomeric, as predicted by the notion that cohesion arises from the co-entrapment of sister DNAs inside individual cohesin SCIENCE science.org

Sci. Immunol. (2024) 10.1126/sciimmunol.adi5578

CANCER

eEF1A2 and PTEN against breast cancer Isoforms of the protein synthesis factor eEF1A can either promote or suppress tumor growth. Treekitkarnmongkol et al. found that eEF1A2 teams up with the tumor suppressor PTEN to repress cell cycling. The interaction between eEF1A2 and PTEN resulted in the assembly of a multiprotein complex in the premitotic phases of the cell cycle and promoted the degradation of the mitotic kinase Aurora A. The abundance of Aurora A and cell proliferation markers was increased in PTEN-deficient breast cancers. A two-drug combination that derepressed PTEN and inhibited Aurora A suppressed tumor growth in mice. —LKF Sci. Signal. (2024) 10.1126/scisignal.adh4475

EVOLUTION

Skull, tooth, and claw

A

bout 100 million years ago during the Mesozoic, Enanthiornithines were the dominant and most diverse bird type. One of the groups, the Bohaiornithidae, are known to have been large and colorful with prominent teeth and claws, but what they ate was unknown. Miller et al. measured a large range of fossils and, by comparison with modern birds, attempted to discern the dietary preferences of some of these creatures. For diet proxies, the authors measured body mass, mechanical advantage, and related functional aspects of the jaws and bite combined with traditional morphometric analysis of claw shape and size. The data showed that these birds exhibited a surprising range of trophic levels, from being plant and fish eaters to ambush predators and raptors. It is likely that the capacity to fly allowed access to more niches and fueled diversification. —CA eLife (2024) 10.7554/eLife.89871.2

Artist’s reconstruction of enantiornithine birds (left: Longipteryx, center: Bohaiornis, and right: Pengornis) feeding in the Early Cretaceous forests of northeastern China roughly 120 million years ago.

CANCER

How lung cancer evades therapy Mutations in the KRAS gene are the most frequently

occurring mutation in nonsmall-cell lung cancer (NSCLC). Pharmacological drugs that inhibit KRAS are used to treat NSCLC, but tumors often develop resistance to therapy. 8 MARCH 2024 • VOL 383 ISSUE 6687

1069

RESEARCH | I N O T H E R J O U R NA L S

METABOLISM

Why you like dessert

S

ome of the reward sensation in the brain from eating comes from taste, but other signals are transmitted back to the brain as the food reaches the intestine. McDougle et al. analyzed how such signals from ingested sugars and fat influence the brain. Imaging of neurons communicating from the gut to the brain showed that separate neurons respond to fats and sugars. Each also connect to dopamine-mediated reward circuits in the brain. The combined signals from fat and sugar produce a synergistically greater reward signal and promote overeating. The authors note that if such reward circuits could be controlled, they might offer a strategy more effective than conscious dieting efforts for the treatment of obesity. —LBR Cell Metab. (2024) 10.1016/j.cmet.2023.12.014

al. performed a kinetic analysis of the effects of exon deletions on the aggregation of aSyn and investigated the effects of some of the shorter aSyn isoforms in their monomeric and fibrillar forms on the aggregation of the main isoform, aSyn-140. One of these, aSyn-112, even at low levels, enhanced the aggregation of aSyn-140. These results indicate that we should consider proteoforms, i.e., distinct versions of a protein, when investigating protein behavior and implications for disease pathology. —PRS Proc. Natl. Acad. Sci. U.S.A. (2024) 10.1073/pnas.2313465121

Cancer Discov. (2024) 10.1158/2159-8290.CD-23-0289

GEOENGINEERING NEUROSCIENCE

Alternative splicing of a-synuclein The aggregation of a-synuclein (aSyn) is central to the pathogenesis of Parkinson’s disease and other synucleinopathies. In addition to the main isoform of aSyn, several other protein variants are produced by alternative splicing. Röntgen et 1070

Learning what we don’t know Various types of geoengineering interventions have been proposed to help slow the drastic rise of temperature that humans have caused over the past century, but we know so little about the potential side effects of such schemes that it is not clear if the cure might be worse than the disease. Wunderlin

8 MARCH 2024 • VOL 383 ISSUE 6687

et al. explored the possible consequences of low-latitude sulfur-based stratospheric aerosol injections, finding that they could cause heating of the tropical lower stratosphere. The authors calculate that some of the atmospheric effects and regional climate changes that this approach risks are comparable in their negative impacts to the carbon dioxide–driven warming that they are trying to counteract. —HJS Geophys. Res. Lett. (2024) 10.1029/2023GL107285

PHYSICS

Laser-cooled positronium Positronium, a bound state of an electron and its antiparticle, a positron, can be used to test the limits of theories such as quantum electrodynamics and the equivalence principle. Typically, positronium ensembles are created with broad velocity distributions, and reducing their width would benefit these fundamental tests. However, coming up with an effective cooling protocol for positronium has proven tricky. Glöggler et al. used a custom-built laser

to reduce the temperature of a positronium cloud from 380 to 170 K. Starting from a colder source and adding a second cooling stage may lead to even lower temperatures. —JS Phys. Rev. Lett. (2024) 10.1103/PhysRevLett.132.083402

CHEMISTRY

Sulfate-binding cryptands Sulfate is a biologically and environmentally relevant anion, but it has a very high solvation enthalpy so there are few highaffinity sulfate-binding motifs, either natural or synthetic. Jing et al. developed a class of cryptand ligands with six central-facing urea moieties that can donate hydrogen bonds to a small central cavity. This environment is well suited to coordinating a central sulfate anion with millimolar affinity in water. The authors observed much higher affinities for sulfate by this ligand in a detergent solution and quantitative, selective extraction in a mixture with an organic solvent and a lipophilic cation. —MAF Nat. Chem. (2024) 10.1038/s41557-024-01457-5 science.org SCIENCE

PHOTO: ADSHOOTER/ISTOCK PHOTO

Li et al. uncover a mechanism that helps to explain how lung adenocarcinoma (LUAD), a subtype of NSCLC, evades therapy. Using patient samples and experimental model systems, KRAS inhibitors were found to promote the transition of LUAD cells into alveolar type 1-like (AT1) cells, a differentiated and quiescent cell subset that does not respond to therapy. Ablation of the AT1-like cell state improved drug response, indicating that strategies to target alveolar differentiation could be used in combination with KRAS inhibitors to improve lung cancer treatment. —PNK

RESEAR CH

REVIEW SUMMARY

◥

step closer to answering a fundamental question in neuroscience: How basic synaptic properties shape higher-network computations.

NEUROSCIENCE OUTLOOK: The hippocampal mossy fiber synapse

Structure, biophysics, and circuit function of a “giant” cortical presynaptic terminal David Vandael and Peter Jonas*

BACKGROUND: The hippocampus plays a key role in learning and memory. Exactly in the center of this circuit is found a glutamatergic synapse with remarkable properties—the mossy fiber synapse, which provides the connection between the granule cell layer and the CA3 subfield. This synapse has fascinated neuroscientists for decades. As already noted by Ramón y Cajal, this synapse is characterized by the large size of the presynaptic terminals and the proximal location on the dendrites of the postsynaptic CA3 pyramidal neurons, where it impinges on large spines, termed “thorny excrescences.” These properties have led to the idea that the mossy fiber synapse is a very powerful synapse that may efficiently activate its postsynaptic target cells. Thus, the mossy fiber synapse in the hippocampus is often depicted as a “teacher” or “detonator,” somewhat analogous to the climbing fiber synapse in the cerebellum. However, this hypothesis has not been fully tested, and the underlying synaptic mechanisms remain unclear. ADVANCES: Because of its comparatively large size, the hippocampal mossy fiber synapse is accessible to direct presynaptic patch-clamp

A

recording. Therefore, the mossy fiber synapse allows a rigorous investigation of the biophysical mechanisms of synaptic transmission and plasticity, being comparable to the squid giant synapse in the invertebrate nervous system and the calyx of Held in the auditory brainstem of mammals. Recent work based on new technologies in both nanoanatomy and nanophysiology, including presynaptic patch-clamp recording, paired recording, superresolution light-microscopy, and freeze-fracture and “flash-and-freeze” electron microscopy, has provided new insights into structure (see figure, panel A), biophysics (see figure, panel B), and network function (see figure, panel C) of this intriguing synapse. At the synaptic level, its distinctive facilitation has been linked to the coupling between Ca2+ channels and release sensors, and the potentiation has been related to the size and organization of synaptic vesicle pools. At the network level, the mossy fiber synapse not only contributes to storage of information, as expected for a “teacher” synapse, but is also involved in recall and participates in higher-order computations (pattern separation and pattern completion). Structural changes associated with plasticity may represent “synaptic engrams.” This brings us one

B

Structure

not only plays a key role in higher-network functions but also seems to be impaired in several brain diseases. In epilepsy, mossy fiber sprouting leads to the formation of mossy fiber synapses with distinct functional properties. In genetic mouse models of neuropsychiatric diseases, including schizophrenia, mossy fiber synaptic transmission is affected, suggesting a new entity of “mossy fiber synaptopathy.” Furthermore, in neurodegenerative diseases, the entorhinalhippocampal circuit appears to be the first affected circuit, and the hippocampal mossy fiber synapse is one of the first impaired synapses. This suggests the possibility that the mossy fiber synapse may play a causal role in the memory impairments in these diseases. Conversely, the mossy fiber synapse could be used for therapeutically targeting complex neuropsychiatric or neurodegenerative diseases with unprecedented selectivity. Thus, the hippocampal mossy fiber synapse could become the first synapse in the history of neuroscience for which we reach complete understanding of synaptic biophysics, network function, and contribution to disease, and exploit this information for therapeutic purposes.

▪

Institute of Science and Technology Austria (ISTA), A-3400 Klosterneuburg, Austria. *Corresponding author. Email: [email protected]. Cite this article as D. Vandael and P. Jonas, Science 383, eadg6757 (2024). DOI: 10.1126/science.adg6757

READ THE FULL ARTICLE AT https://doi.org/10.1126/science.adg6757

C

Biophysics

Network function Recurrent collateral CA3-CA3 synapses

Mossy fiber input

2 µm

Structure, biophysics, and network function of a cortical presynaptic terminal. (A) Structural analysis of hippocampal mossy fiber terminals on the apical dendrite of a single CA3 pyramidal neuron (golden yellow) by superresolution microscopy. Different terminals represented in different colors. Modified from Michalska et al. (106). (B) Functional analysis of synaptic transmission at hippocampal mossy fiber synapses. Infrared videomicrograph; small pipette on top right attached to presynaptic terminal (red); large SCIENCE science.org

4 µm

CA3 pyramidal neurons

pipette on bottom left connected to postsynaptic neuron (blue). Reprinted (modified) from (29), with permission from Elsevier. (C) Network function of the hippocampal mossy fiber synapse in a model of pattern completion of the CA3 pyramidal neuron population. Proximal synapses, mossy fiber synapses; distal synapses, recurrent collateral CA3-CA3 synapses. Red, active; black, inactive inputs or cells. Reprinted (modified) from (11), with permission from Elsevier. 8 MARCH 2024 • VOL 383 ISSUE 6687

107 1

RES EARCH

RESEARCH ARTICLE SUMMARY

◥

ADULT STEM CELLS

Vitamin A resolves lineage plasticity to orchestrate stem cell lineage choices Matthew T. Tierney*, Lisa Polak, Yihao Yang, Merve Deniz Abdusselamoglu, Inwha Baek, Katherine S. Stewart, Elaine Fuchs*

INTRODUCTION: Adult stem cells reside within

specialized niches that tightly regulate the type and timing of tissue regeneration. And yet, barrier tissues, such as the skin, are exposed to a barrage of insults that can disrupt niche architecture. Relieved from their normal constraints, these stem cells can be mobilized to those wounded areas most in need. To obtain the flexibility required to cross these homeostatic boundaries, stem cells enter a transient state of lineage plasticity in which both prior and new cell fates are coexpressed. Proper execution of this fate switch is critical to restoring tissue health because failure to do so can lead to defective lineage selection and chronic wound repair. RATIONALE: Despite a growing appreciation of

the importance of lineage plasticity, the signaling pathways involved remain poorly underHomeostasis Niche-restricted lineages

stood. Skin epithelium is ideal for exploring this problem because hair follicle stem cells (HFSCs) act as the primary responders to reepithelialization after scratches and abrasions. To contribute to repair, the stem cells must enter a lineage plastic state before taking up new residences in epidermal niches. Similarly to stem cell behavior in wounds, cultured mouse HFSCs grown in serum and growth factor–rich medium also exhibit analogous molecular signatures. These cultures provide a platform to pursue high-throughput functional approaches to elucidate the biology that underlies stem cell lineage plasticity. RESULTS: By screening a library of small mol-

ecules for their ability to regulate lineage plasticity in mouse HFSCs, we identified all-trans retinoic acid (atRA) as a vitamin A metabolite essential to restoring their physiological identity Tissue regrowth Lineage plasticity resolved

Wounding Stem cell lineage plasticity

Epidermal differentiation Stem cell fate switching

EpSCs

HFSCs (atRA+)

Mobilized HFSCs

Hair regrowth HFSC origin

Lineage plasticity

EpSCs

Lineage plasticity

atRA-dependent nuclear RAR-RXR receptor activity in HFSCs

HFSCs atRA+BMP+

TACs atRA+WNT+

HS/IRS WNT+BMP+

Retinoic acid orchestrates stem cell lineage plasticity during wound healing. Local retinoic acid signaling is transiently reduced in HFSCs when niche architecture is disrupted by injury, promoting lineage plasticity and enabling cell fate switching to replace the lost epidermis. Once the skin barrier is repaired, retinoic acid signaling must then be restored to rebalance stem cell contribution to hair regrowth. EpSCs, epidermal stem cells; TACs, transit-amplifying cells; HS, hair shaft; IRS, inner root sheath; RAR, retinoic acid receptor; RXR, retinoid X receptor. 1072

8 MARCH 2024 • VOL 383 ISSUE 6687

in culture. These same effects were recapitulated in vivo in mouse skin, where atRA was produced locally as a stem cell niche component and was required to maintain lineage specification. Layering transcriptomic and chromatin landscaping data with retinoid-activated nuclear retinoic acid receptor–DNA interactions, we identified targets that affect this outcome. Included in this downstream regulatory network was pioneer factor SOX9, which is known to directly up-regulate key hair follicle genes while silencing epidermal fates. By resolving lineage plasticity alongside comprehensive media optimization, atRA enabled cultured HFSCs to collaborate with WNT (Wingless-related integration site) and BMP (bone morphogenetic protein) signals to effectively orchestrate the distinct lineages required to make hair follicles. A host of behaviors specific to stem cells, including the transition from quiescence to active self-renewal and their direction along orderly routes to differentiated fates, were accurately modeled in this culture platform. Studying the influence of atRA during wound healing using the laboratory mouse as a model system, we found that genetic, dietary, and topical interventions all confirmed atRA’s power in balancing stem cell contributions to epidermal repair and hair regrowth. The availability of atRA was inversely correlated with lineage plasticity, transiently reduced early in repair to allow for stem cell fate switching in the wound bed, and then later restored to fuel hair regeneration once the barrier had reformed. CONCLUSION: Lineage plasticity has emerged

as a crucial feature of the stem cell stress response. In this work, we found that local vitamin A metabolism acts as a potent upstream regulator of this state during skin re-epithelialization. In applying this knowledge to different settings, we learned that stem cells rely on atRA availability as an essential niche component to balance their contributions to distinct fate choices and lineages during homeostasis and repair. We also defined the minimal requirements needed to produce a suite of skin and hair cell types outside the body, offering potential opportunities to study their biology uncomplicated by the many variables still unaccounted for in the in vivo microenvironment. Given the widespread occurrence of lineage plasticity in normal tissues as well as nonhealing wounds and cancer, these findings may have substantial implications for a variety of applications in regenerative medicine.

▪

The list of author affiliations is available in the full article online. *Corresponding author. Email: [email protected] (E.F.); [email protected] (M.T.T.) Cite this article as M. T. Tierney et al., Science 383, eadi7342 (2024). DOI: 10.1126/science.adi7342

READ THE FULL ARTICLE AT https://doi.org/10.1126/science.adi7342 science.org SCIENCE

RESEAR CH

RESEARCH ARTICLE SUMMARY

◥

function and its role in controlling similar growth patterns in various species. Cloning the CL gene is a crucial first step in exploring these possibilities.

PLANT SCIENCE

Enhancing rice panicle branching and grain yield through tissue-specific brassinosteroid inhibition Xiaoxing Zhang†, Wenjing Meng†, Dapu Liu†, Dezhuo Pan†, Yanzhao Yang, Zhuo Chen, Xiaoding Ma, Wenchao Yin, Mei Niu, Nana Dong, Jihong Liu, Weifeng Shen, Yuqin Liu, Zefu Lu, Chengcai Chu, Qian Qian*, Mingfu Zhao*, Hongning Tong*

INTRODUCTION: Crop breeding is essentially the

RATIONALE: Clustered-spikelet rice (CL) is a

art of balancing various traits to achieve the best performance in plants. However, achieving the optimal balance is challenging owing to complex relationships among traits. One major hurdle is the trade-offs among various traits such as between grain size and grain number. Brassinosteroids (BRs), a class of steroid hormones, promote grain size and have been demonstrated to enhance crop yield. Despite their potential, the pleiotropic effects of BRs have hindered their use in crop improvement, and their role in regulating grain number remains unclear.

distinctive germplasm with historical importance, producing multiple grains from a single position without affecting grain size. The causal gene for CL has remained elusive for almost a century. We hypothesized that dissecting the molecular genetics of CL could provide insights into panicle branching and potentially unlock the trade-off between grain size and number. Observations of weakly clustered growth in BR-deficient plants suggested a possible role of BRs in controlling this trait. Understanding the mechanism behind CL occurrence could offer new insights into BR

RESULTS: We successfully cloned the causal

gene for CL through extensive screening of CL suppressor mutants. Genomic analysis revealed that CL is associated with complex structural variations in chromosomes, activating the BR catabolic gene BRASSINOSTEROID-DEFICIENT DWARF3 (BRD3) in secondary branch meristems. Spatial-specific activation of BRD3 promotes grain number without the common negative effects of BR deficiency on grain size. We uncovered a tissue-specific BR pathway that supports this function, involving the BR signaling inhibitor kinase GSK3/SHAGGYLIKE KINASE2 (GSK2), which phosphorylates and stabilizes the OsMADS1 transcriptional factor. OsMADS1, in turn, targets and promotes the TERMINAL FLOWER1-like gene RICE CENTRORADIALIS2 (RCN2) to suppress meristem identity. Introduction of CL into different backgrounds substantially improved rice yield. Additionally, we found consistent alterations in BR content in pepper and rose with clustered growth patterns, suggesting a broader role of BRs in controlling this phenotype in nature. CONCLUSION: The identification of complex

Typical BR synthesis or signaling defective rice mutants

Reduced grain size and other defects

Unique clustered-spikelet rice germplasm

Prolonged meristem transition

NCL

SM

SBM Structural variations

BRD3 activation

BR degradation

GSK2 accumulation

CL Enhanced grain number with unaltered grain size and no other defects

OsMADS1 stabilization

up-regulation

BR-associated clustered growth in nature BR alteration

NCL

BR alteration

CL

NCL

▪

CL

Tissue-specific BR inhibition promotes rice panicle branching. Structural variations in CL activate BRD3 in specific tissues (red) such as secondary branch meristems (SBMs), where it causes BR degradation, GSK2 accumulation, OsMADS1 stabilization, and RCN2 up-regulation. This affects spikelet meristem (SM) identity, prolonging the SBM to SM transition, increasing panicle branching and grain number without affecting grain size. BR change may be associated with the clustered growth in various plants. SCIENCE science.org

structural variations preceding BRD3 explains previous unsuccessful attempts to clone the gene. Our research unveils a groundbreaking role of BRs in coordinating panicle branching and grain number through precise meristem transitions. We describe a tissue-specific BR pathway (BRD3-BR-GSK2-OsMADS1-RCN2) that underpins this previously unknown function. The positive and negative roles of BRs in regulating grain size and number, respectively, represent a crucial trade-off mechanism. Notably, this spatially targeted mechanism within the BR pathway enhances grain number without compromising size. Therefore, manipulating BR distribution provides effective breeding strategies for finely tuning crop traits, ultimately boosting crop yield. Our findings showcase the efficacy of tissue-specific hormonal manipulation in overcoming trade-offs among various traits and unlocking crop yield potential. Changes in BR levels may represent a fundamental mechanism governing natural inflorescence architecture. The list of author affiliations is available in the full article online. *Corresponding author. Email: [email protected] (H.T.); [email protected] (M.Z.); [email protected] (Q.Q.) †These authors contributed equally to this work. Cite this article as X. Zhang et al., Science 383, eadk8838 (2024). DOI: 10.1126/science.adk8838

READ THE FULL ARTICLE AT https://doi.org/10.1126/science.adk8838 8 MARCH 2024 • VOL 383 ISSUE 6687

1073

RES EARCH

RESEARCH ARTICLE SUMMARY

◥

NEUROSCIENCE

Brainstem control of vocalization and its coordination with respiration Jaehong Park*, Seonmi Choi, Jun Takatoh, Shengli Zhao, Andrew Harrahill, Bao-Xia Han, Fan Wang*

INTRODUCTION: Phonation, the pivotal process

governing vocalization and speech, requires two simultaneous actions of narrowing the larynx (vocal cord adduction) and exhaling air from the lungs. Speech cannot occur during inhalation, because inspiration dominantly inhibits vocalization. This breathing primacy is crucial for survival. Although prior studies have identified neurons in the midbrain periaqueductal gray (PAG) as a permissive gate for eliciting vocalizations, the alternating patterns of inspiration and vocalization are not changed by PAG stimulation in experimental animals. This prompted our investigation to identify a neural population directly driving phonation and to elucidate its interactions with the breathing circuit that ensure vocal-respiratory coordination and prioritize breathing. To this end, we used mouse ultrasonic vocalizations (USVs) as a model, in which vocal cord adduction is required for USVs and USV syllables are periodically interrupted by inspirations. RATIONALE: Our hypothesis centers on laryngeal premotor neurons in the brainstem as being key controllers of vocal cord adduction and its coordination with breathing. Whereas

Normal

Phonation

past literature has identified the nucleus of retroambiguus (RAm) in the caudal hindbrain as a critical node for vocalization, its heterogeneity, including neurons modulating respirations and other orofacial movements, necessitates precise targeting of vocalizationspecific premotor neurons within the RAm to unravel the mechanistic intricacies of vocal cord control. Using monosynaptic rabies virus– mediated transsynaptic tracing, we labeled a population of excitatory laryngeal premotor neurons in the RAm in adult mice. Furthermore, courtship USVs induced robust expression of the immediate early gene Fos in these rabies-traced RAm neurons (RAmVOC), leading us to use a Fos-based targeting method (CANE) to label and manipulate RAmVOC neurons and examine their role in phonation and the vocal-respiration interaction. RESULTS: Silencing RAmVOC neurons using

tetanus toxin light chain abolished courtship USVs and pain-elicited audible squeaks in adult mice, along with a lack of phonationrelated abdominal muscle activity, indicating that RAmVOC neurons are necessary for phonation. Optogenetic activation of RAmVOC

was sufficient to induce vocal cord closure and to elicit USVs, with the duration of activation influencing USV syllable lengths and concurrent expiration periods. Inspiration needs could override RAmVOC-mediated vocal cord closure. Both laryngeal motoneurons and RAmVOC neurons receive inhibitory inputs from the preBötzinger complex (preBötC), which is known for containing inspiration rhythm-generating neurons. Ablating inhibitory synapses in RAmVOC neurons compromised the inspiration gating of vocal cord adduction, resulting in abnormal hoarse vocalizations during inspiration periods upon PAG stimulation. Additionally, disinhibited RAmVOC led to spontaneous USVs in the absence of a social context. CONCLUSION: Our study unveils the circuits

and mechanisms underlying phonation and vocal-respiration interaction (see the figure). RAmVOC forms the critical premotor node downstream of PAG necessary for all phonations by driving vocal cord adduction and coordinating expiratory muscle activity. Furthermore, inhibitory inputs from the preBötC to both RAmVOC and laryngeal motoneurons enable rhythmic inspiration to gate and pace vocalization, thereby ensuring breathing primacy.

▪

The list of author affiliations is available in the full article online. *Corresponding author. Email: [email protected] (F.W.); [email protected] (J.P.) Cite this article as J. Park et al., Science 383, eadi8081 (2024). DOI: 10.1126/science.eadi8081

READ THE FULL ARTICLE AT https://doi.org/10.1126/science.adi8081

PAG

Vocal cords

RAmVOC Vocal drive (Tonic excitation) +

Inspiration

preBötC

+ preBötC

RAmVOC -

RAmVOC VOC

Block synaptic inhibition to RAm

+

-

Inspiratory rhythm

Inspiration

preBötC

RAmVOC Vocal motoneurons

Neurons and circuit mechanisms for phonation and vocalization-respiration coordination. RAmVOC represents vocal premotor neurons downstream of the PAG that drive vocal cord closure and phonations (ultrasonic vocalizations in mice). During inspiration, inhibitory neurons in the inspiration rhythm generator preBötC suppress activities of RAmVOC and vocal motoneurons to ensure breathing. Blocking inhibitory inputs to RAmVOC results in abnormal vocalization during inspiration. 1074

8 MARCH 2024 • VOL 383 ISSUE 6687

science.org SCIENCE

RESEAR CH

RESEARCH ARTICLES

◥

GEOLOGY

CO2 drawdown from weathering is maximized at moderate erosion rates Aaron Bufe1*, Jeremy K. C. Rugenstein2, Niels Hovius3,4 Uplift and erosion modulate the carbon cycle over geologic timescales by exposing minerals to chemical weathering. However, the erosion sensitivity of mineral weathering remains difficult to quantify. Solutechemistry datasets from mountain streams in different orogens isolate the impact of erosion on silicate weathering—a carbon dioxide (CO2) sink—and coupled sulfide and carbonate weathering—a CO2 source. Contrasting erosion sensitivities of these reactions produce a CO2-drawdown maximum at erosion rates of ~0.07 millimeters per year. Thus, landscapes with moderate uplift rates bolster Earth’s inorganic CO2 sink, whereas more rapid uplift decreases or even reverses CO2 sequestration. This concept of an “erosion optimum” for CO2 drawdown reconciles conflicting views on the impact of mountain building on the carbon cycle and permits estimates of geologic CO2 fluxes dependent upon tectonic changes.

he chemical dissolution of rocks at Earth’s surface modulates the concentration of CO2 in the atmosphere and, thereby, Earth’s climate (1–3). Understanding how weathering rates are affected by tectonic and climatic boundary conditions is critical to modeling Earth’s evolution over geologic time and to assessing Earth’s systemic response to natural and anthropogenic perturbations. Rates of chemical weathering are controlled, to first order, by the supply of fresh minerals to the weathering zone, the availability of reactive fluids, and the kinetics of mineral dissolution (3, 4). In turn, the supply of fresh minerals is directly linked to the exhumation of rocks at Earth’s surface (5–7). Hence, mountain building and attendant erosion are thought to play a role in modulating Earth’s inorganic carbon cycle (8–11). Quantifying the link between exhumation, chemical weathering, and the inorganic carbon cycle is complicated by a large range of observed sensitivities of weathering fluxes to erosion. Compilations of weathering fluxes from global streams, soils, and sedimentary archives have shown an apparent increase of silicate weathering fluxes with increasing erosion rates (5, 7, 12–15). By contrast, other studies have suggested that silicate weathering fluxes can be insensitive to erosion across wide erosionrate gradients (5, 16–19). Existing weathering models attribute these varying responses to a trade-off between erosion, subsurface water flow,

T

1

Department of Earth and Environmental Sciences, LudwigMaximilians-Universität München, Munich, 80333, Germany. Department of Geosciences, Colorado State University, Fort Collins, CO 80523, USA. 3GFZ German Research Center for Geosciences, Potsdam, 14473, Germany. 4Department of Earth Sciences, University of Potsdam, Potsdam, 14476, Germany.

2

*Corresponding author. Email: [email protected]

SCIENCE science.org

and temperature in modulating weathering rates on the landscape scale (5, 6, 20–25) (Fig. 1). Under slow erosion, weathering fluxes are limited by the supply of fresh minerals (mineralsupply limit). In this case, weathering fluxes scale linearly with erosion rate and are insensitive to climatic parameters (Fig. 1). By contrast, in uplands with high erosion rates, weathering can become limited either by rates of reactive fluid flow through the weathering zone or by the rate of mineral dissolution (weathering limit). In these environments, weathering fluxes become insensitive or even negatively sensitive to erosion rate (5, 6, 26) (Fig. 1). These concepts have been incorporated into long-term carbon cycle models (10, 27, 28), but the validation of their predictive power and the quantification of the

resulting erosion sensitivity of inorganic carbon fluxes remains limited for at least three reasons. First, runoff (commonly used as a proxy for subsurface water flow) and erosion covary in available datasets (5, 7, 13, 15, 23, 24). This covariation poses a challenge to distinguishing the relative impact of runoff and erosion on weathering fluxes because that relative impact depends on the limit to weathering (Fig. 1). Second, we lack systematic studies of the erosion sensitivity of weathering fluxes from nonsilicate mineral phases with respect to the erosion sensitivity of silicate weathering fluxes. A growing canon of studies suggests that carbonates and sulfides dominate weathering fluxes from active mountain ranges, even if they constitute only minor proportions of the exhumed rock mass (15, 16, 18, 23, 29–33). Moreover, carbonate and sulfide weathering fluxes may be sensitive to erosion beyond the point at which silicate weathering fluxes plateau (16, 18, 19, 30, 31, 34). Considering these weathering reactions within the context of the inorganic carbon cycle is important because combined carbonate and sulfide weathering acts as a transient carbon source to the atmosphere over timescales shorter than the residence time of sulfate in the oceans [~107 years (35)] and could offset the carbon sink from silicate weathering for millions of years (27, 30, 33, 36). Moreover, sulfide and carbonate weathering may modulate solution pH and directly affect silicate weathering rates (18, 37). Third, apparent nonlinearities exist in the sensitivity between erosion and weathering fluxes over multiple orders of magnitude (13, 15, 18, 23) and may strongly differ between mineral phases. To constrain the sensitivity of the inorganic carbon cycle Fig. 1. Sensitivity of weathering to erosion and climate. All lines are calculated with the weather model from West (20) with model parameters set arbitrarily and plotted in log-log space with the x and y axes each spanning four orders of magnitude. (A) Sensitivity of weathering flux to erosion calculated for three different runoff (qw) values: qw, 10qw, and 100qw. Sensitivity of weathering flux to climate and erosion trade off between the mineral-supply limit and the weathering limits. However, a nonlinear erosion sensitivity of weathering fluxes is possible even at the weathering limit [dotted line; s in equation S6 in (42)]. (B and C) Sensitivity of concentration (B) and weathering flux (C) to runoff calculated for three different erosion rate (E) values: E, 10E, and 100E. Note that weathering fluxes (concentration multiplied by runoff) have different runoff sensitivities compared with concentrations. The runoff sensitivity factor can vary between b = 0 (concentrations are insensitive to runoff; weathering fluxes are positivelinearly sensitive to runoff) and b = −1 (concentrations are negative-linearly sensitive to runoff; weathering fluxes are insensitive to runoff).

8 MARCH 2024 • VOL 383 ISSUE 6687

1075

RES EARCH | R E S E A R C H A R T I C L E S

to erosion, therefore, the sensitivity of weathering fluxes of all relevant mineral phases to erosion must be considered in isolation from any covariation with climate. We compiled and analyzed four datasets that report solute concentrations in mountain streams underlain by relatively uniform siliceous metasedimentary rocks, across well-constrained order-of-magnitude erosion-rate gradients, spanning from ~0.02 mm/year to ~7 mm/year (5, 15, 16, 18). Three datasets comprise local erosion-rate gradients in southern Taiwan (18), the eastern Tibetan Plateau in Sichuan (16), and the Southern Alps of New Zealand (15). In addition, we included a global compilation of mountain streams (5) (hereafter referred to as West data), which has been extensively discussed in the literature and serves as a point of reference, even if the sampling differs from the three regional datasets and includes catchments underlain by felsic igneous rocks. We did not consider global compilations of rivers where substantial topographic, lithologic, climatic, and erosional contrasts within single catchments complicate the attribution of weathering to a single driver (13, 23). For each dataset,

we estimated silicate, carbonate, and sulfide weathering fluxes from solute concentrations and runoff. To investigate the sensitivity of these fluxes to erosion, we fit a widely used model (20) that describes chemical weathering fluxes from eroding landscapes as a function of erosion flux and climatic parameters (runoff and temperature). A function of runoff (water discharge per catchment area) is used in the model to approximate the dependence of weathering fluxes on fluid flow through the subsurface (22). The model captures the transition from a mineral-supply limit to a weathering limit and allows for nonlinearity between the weathering flux and erosion at the weathering limit (Fig. 1A). The functional form of this model and the predicted limits to weathering are similar to those of alternative models (6, 22). In addition to the model fits, we estimated an average erosion sensitivity, a, of weathering fluxes to erosion across all erosion-rate gradients by fitting a power law with exponent a. We use the word sensitivity to describe the relative response of a dependent variable (e.g., weathering flux) to a change in an explanatory variable (e.g., erosion rate). We note that the existing

Fig. 2. Data and model fits calculated with bsil = 0 and bsulf,carb = −0.75. (A to K) Accepted fits (gray lines) are equivalent to a 95% confidence intervals, and quartiles are the 25th and 75th percentiles. White points and associated red dash-dotted best-fit lines are calculated for bsil = −1 (equivalent to no runoff correction); these are only plotted for New Zealand 1076

8 MARCH 2024 • VOL 383 ISSUE 6687

weathering models and our compiled data cannot account for weathering of sediment deposits downslope of major mountain ranges, which could play a major part in global weathering fluxes (38–40) and could be affected by changes in uplift and erosion rates in the upstream source area. The multiparameter model assumes that erosion, runoff, and temperature are independent variables. An order-of-magnitude range in runoff values and a strong covariation between runoff and erosion in the New Zealand dataset— and to a lesser extent in the West dataset— violate this assumption (fig. S1). To isolate the direct influence of erosion on chemical weathering flux in these data, we normalized weathering fluxes in all four datasets to a reference runoff with a prescribed runoff sensitivity factor b (Fig. 1). The runoff sensitivity depends on the limit to weathering (Fig. 1) and can, therefore, differ with erosion rate, mineral phases, and study area. Without a priori knowledge of the runoff sensitivity, we assumed that silicate weathering fluxes are weathering limited and linearly sensitive to runoff across the sampled erosion rates (bsil = 0) and that sulfide

and the West data where the difference between the bsil = 0 and bsil = −1 fits is substantial. For all fits, see figs. S2 and S3. Data points in (D) and (H) with inserted black circles are low–erosion rate samples from cratonic regions not included in the analysis (42). These data are plotted without runoff correction (b = −1). science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

Fig. 3. Degrees of nonlinearity from the power law fit. Colored points are calculated with bsil = 0 and bsulf,carb = −0.75. White points are calculated with b = −1 (equivalent to no runoff correction) and gray points for b = 0 (linear runoff sensitivity of weathering fluxes).

and carbonate weathering fluxes have an intermediate sensitivity to runoff (bcarb,sulf = −0.75). This assumption is based on weathering fluxes in the locations that are unaffected by runoff gradients (Taiwan and Sichuan), on existing theory and observations (5, 19, 20, 22), and on a global compilation of runoff sensitivities (41) [see section 2 in (42)]. We emphasize that the choice of b substantially affects the results for the New Zealand dataset only, and we describe the impact of b on these data below. The erosion sensitivity of silicate, carbonate, and sulfide weathering fluxes

Our data compilation and model fits allow us to explore the erosion sensitivity of silicate, carbonate, and sulfide weathering fluxes in parallel across three regional and one global erosionrate gradient [Fig. 2, Fig. 3, and section 2 in (42)]. Despite differences in climatic, lithologic, and tectonic boundary conditions, our data reveal telling consistencies in the sensitivity of solute fluxes to erosion (Fig. 2 and Fig. 3) that suggest that common mechanisms link erosion and weathering across all locations. Silicate weathering fluxes in all study areas have a low erosion sensitivity, ranging from weakly positive to weakly negative (mean and standard deviation of all four datasets sil ¼ 0:00T0:15 ) (Fig. 3). The model fits a require a transition from a supply to a weathering limit at erosion rates of 35% in 30 min (Fig. 2H). BioSUM2 can respond to 1 ml of simulated intestinal fluid (SIF; pH ~6.8) applied directly onto the device by swelling >10% within 10 min and >15% in 30 min (Fig. 2K). Figure 2, L and M, presents the response of BioSUM3 to alkaline conditions. BioSUM3 swells by 35% in pH 9 within 30 min (Fig. 2L), to an equilibrium swelling ratio of ~40% (Fig. 2M). BioSUM3 swells by 10% within 1 min after contact with 1 ml of simulated pancreatic juice (SPJ; pH ~8.2), and it reaches 30% in 30 min (Fig. 2N). FEA modeling accurately predicts the swelling kinetics for all cases (Fig. 2, F, I, and L). Multiple devices at strategic anatomical locations relative to a surgical site provided the basis for spatiotemporal monitoring of the convective spread of SGF (Fig. 2, O and P, and figs. S22 and S23). Benchtop demonstrations involved the introduction of SGF on one end of a slab of agarose gel. Seven sensors distributed along a straight line in a thin layer of PBS solution on this gel allowed measurements of the swelling ratios at corresponding locations, to a distance of 30 cm from the origin at a spacing of 5 cm. Figure 2P shows the timeresolved responses of BioSUM1 at different locations, indicating expected performance and a spatial resolution comparable to the sizes and spacings of the devices. Sensing by ultrasound imaging in deep tissues

Ultrasound B-mode imaging serves as a mechanism for quantitatively evaluating the pHdependent geometry of these responsive hydrogels when implanted at deep tissue locations. The disks of Zn, an established bioresorbable metal that reacts with water to yield the benign end

product Zn(OH)2 (30), act as ultrasonic indicators. The acoustic impedance mismatch between these Zn structures and the surrounding hydrogel matrix and adjacent soft tissues enhances their visibility in B-mode images. The reflection coefficient is approximately proportional to (Z1 − Z2)2/(Z1 + Z2)2 for structure thickness greater than ~20 mm for normal incidence specular reflections (39), where Z1 and Z2 are the impedances for the Zn and the hydrogel, respectively (see supplementary text and figs. S11 and S24 for details about reflections for thicknesses of 80%. By contrast, coefficients between typical surrounding soft tissues and the hydrogel are 200 alleles analyzed per cosmid, averaged from two independent biological repeats. (C) RASER-FISH outcomes of two probes. Marking sites of sister chromatid cohesion are shown, either in wild-type conditions or conditions of cohesin loss induced by PROTAC3-mediated degradation of RAD21 (8-hour treatment). G2 cells were identified by sororin mean intensity (wild type) or Cyclin A mean intensity (no cohesin). (Left) Example images of G2 cells. (Right) Quantification. Experiments were carried out in biological duplicates. n = 57 HS443D9 wild type, n = 43 HSD443D9 no cohesin, n = 28 HS306A4 wild type, and n = 31 HS306A4 no cohesin. Kruskal-Wallis test yielded P = 0.0095, mean and SD are plotted. (D) Colocalization of sororin to (top) cohesed and (bottom) noncohesed sites and absolute quantification of sororin occupancy of the respective sites. Samples were prepared under preextraction conditions to visualize sororin, which maintained cohesion. Experiments were carried out in biological duplicates. n = 48 HS443D9 (top), n = 28 HSD443D9 (bottom), n = 31 HS306A4 (top), and n = 20 HS306A4 (bottom). Kruskal-Wallis test yielded (top) P = 0.0075 and (bottom) P = 0.0086; mean and SD are plotted. (E) Distance analysis between sister chromatids (distance between intensity maxima) at sites occupied by or void of sororin. Experiments were carried out in biological duplicates, n = 49 for sororin positive sites and n = 55 for sororin negative sites. Mann-Whitney U test yielded P < 0.000001; mean and SD are plotted. Graphs are artificially jittered in x to show distributions.

SCIENCE science.org

8 MARCH 2024 • VOL 383 ISSUE 6687

1125

RES EARCH | R E S E A R C H A R T I C L E S

Fig. 4. Sororin is a monomer and associated with less than one-third of chromatin-bound cohesin. (A) Schematic depiction of 3D-SIM imaging approach of endogenously tagged RAD21 together with sororin in G2 phase and on chromatin. (B) Representative widefield and 3D-SIM images of chromatin-bound sororin. Single-nucleus z-slice or cropped region are shown. False coloring by means of sororin signal intensity shows uniform population. (C) Frequency distribution of chromatin-bound sororin signal intensities with fitted nonlinear regression (Gaussian distribution). SumInt, sum intensity; A.U., arbitrary units; n > 140,000 sororin signals from two biologically independent experiments. (D) Frequency distribution of signal intensities of chromatin-bound sororin in (C) is overlaid with nonspecific 1126

8 MARCH 2024 • VOL 383 ISSUE 6687

antibody conjugate intensities observed on imaging slides outside nuclei as shown in (E). n > 3700 antibody conjugates, and data are from two biologically independent experiments. (E) Example image showing analysis of single sororin intensities. On imaging slides, primary and secondary antibody conjugates, which occur through nonspecific binding, were segmented, and their frequency distribution was analyzed. Arrows indicate nonspecific antibody conjugates on slide (blue) and specific sororin signals (green). Antibody specificity in cells has been controlled for in fig. S6, A to C. Quantification of absolute distributions of chromatin-bound sororin from Fig. 4C yielded the following frequencies: 74.4% (monomer), 15.75% (dimer), and 9.85% (multimer). Quantification is based on science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

nonlinear regression analysis. (F) Example image of RAD21-Halo and sororin colocalization in G2 phase cells. Maximum projections of five z-slices are shown, as an entire nucleus and a cropped region. Overlaps are indicated in white circles, and numbered lines indicate the normalized intensity profile measurements. (G) Colocalization analysis of RAD21 at sites of sister chromatid cohesion marked

for trans-sister contacts. Cis contact maps showed abundant interactions extending away from SMC3 sites containing sororin as well as those lacking sororin, indicating that the loopforming cohesin is present at both types of loci (Fig. 2, D and E). By contrast, trans contact maps were associated with increased contacts around SMC3 sites containing sororin but not around those lacking it (Fig. 2, D and E). Profiling trans contact densities at individual genomic regions confirmed that the concentration of trans contacts around SMC3 sites containing sororin occurred throughout the genome (Fig. 2, F and G). In other words, trans contacts were found in the vicinity of cohesin associated with sororin but not with cohesin lacking it. Cohesin complexes associated with sororin specifically mark sites of sister chromatid cohesion

Although consistent with the notion that sororin is exclusively present at cohesion sites, data from populations of cells cannot per se demonstrate that sororin is actually present at sites where sisters are held together by cohesin. To do this, it would be necessary to measure sister cohesion of particular genomic sequences as well as their association with sororin within individual cells (Fig. 3A). To identify a site of cohesion by using fluorescence in situ hybridization (FISH), we used a cosmid DNA probe library covering 2 Mb of telomeric chromosome band 16p13.3 (34). Thirty cosmid probes were labeled, spaced on average every 75 kb midpoint to midpoint to cover coordinates 266 to 1922 kb from the telomere, and four more cosmids were labeled to extend the analysis by a further 1.5 Mb, to ~3.5 Mb from the telomere. FISH was carried out in a human lymphoblastoid cell line established from a normal individual (TA) and a mouse erythroleukemia hybrid cell line containing a normal human chromosome 16 (MEL JY5.4). Initially, with conventional microscopy, we analyzed whether sister DNAs appeared as “split dots,” distinct fluorescent signals for each sister, or as single signals, indicating spatial proximity of sisters. This revealed, in both cell lines, a 320-kb interval within which >60% of sister DNAs were closely juxtaposed (Fig. 3B). To address whether juxtaposition was cohesin dependent and associated with sororin, we combined super-resolution imaging of sororin in U2OS G2 cells with resolution after singlestrand exonuclease resection (RASER)–FISH, which allows visualization of DNA sequences SCIENCE science.org

by sororin. Bar charts consist of dots, each indicating an analyzed RAD21 signal. Colocalization was analyzed in two biologically independent experiments with >38,000 RAD21 molecules analyzed per repeat. Absolute numbers for RAD21 colocalization with sororin are 28.47% (repeat 1) and 28.56% (repeat 2). Twotailed Student’s t test determined P < 0.0001; means are shown.

in structurally preserved cells through a method that avoids heat denaturation or acid treatment (Fig. 3A) (35). Analysis of two probes HS306A4 (probe midpoint 898 kb) and HS443D9 (probe midpoint 955 kb) within the interval by using 3D-SIM showed that sister DNAs were present either as “sausage”- or “dumbbell”shaped single territories in ~70% of analyzed alleles (HS443D9, 69.13 ± 9.38%; HS306A4, 71.28 ± 2.9%) but segregated into distinct territories in the remainder (Fig. 3C and fig. S5A). The presence of both sisters within a single territory in most cases was consistent with the data from using conventional FISH from TA and MELY J5.4 cells (Fig. 3B). The greater resolution achieved by using RASER-FISH combined with 3D-SIM revealed that even when sister DNAs are closely juxtaposed, the signals were not uniform within each territory but instead concentrated at opposite poles of elongated territories, suggesting that sister sequences occupy distinct positions even in those cells where they are closely juxtaposed. Sister DNA juxtaposition was greatly reduced by degradation of the Halo-tagged RAD21 after an 8-hour PROTAC 3 (proteolysis targeting chimera 3) treatment (fig. S5B), which caused the appearance of split dots in >90% of cases (HS443D9, 97.68 ± 3.29%; HS306A4, 87.61 ± 7.43%) (Fig. 3C). It also resulted in a spatial separation of sister DNAs from 95% of cases but rarely with either sister when (in different G2 cells) they were clearly separate (Fig. 3D). Likewise, distance measurements of sister signals revealed a strong correlation between sororin presence and the spatial coupling of sister DNAs. Thus, DNA loci with and without sororin have average distances of 116 ± 68 and 233 ± 103 nm, respectively (Fig. 3E). The colocalization of sororin with this genomic site was also detected with ChIP-seq analysis, which showed discrete accumulation of both cohesin, marked by SMC3, and sororin at this locus (fig. S5D). These findings suggest that sororin is not merely required for holding sisters together but is exclusively associated with cohesive cohesin and therefore a suitable marker to study this subpopulation. Sororin is a monomer, associated with less than one-third of chromatin-bound cohesin

To assess the oligomerization state of cohesive cohesin, we used 3D-SIM to image chromatin-

bound cohesin (RAD21-Halo) and sororin (antibody) in G2 U2OS cells (Fig. 4A). We first analyzed the status of sororin, whose foci were sparse and had a uniform Gaussian intensity distribution (Fig. 4, B and C). To measure their stoichiometry, we used fluorescent signals from primary-secondary antibody conjugates binding nonspecifically to the glass surface outside cells to calibrate those associated with chromatin within cells. The distribution of single primary-secondary antibody conjugates fell exactly into the chromatin-bound sororin distribution (Fig. 4, D and E), implying that the majority of chromatin-bound sororin is monomeric (74.4%). The remaining 25.6% were initially classified as multimeric. To confirm the specificity of our antibody, we analyzed the cell cycle distribution of chromatin-bound sororin by using high-content widefield microscopy [quantitative imagebased cytometry (QIBC)] (36, 37). Specifically, we classified cell cycle stages according to EdU pulse-labeling, DNA content, and MCM2 immunofluorescence, and costaining of chromatin-bound sororin showed its absence on G1 chromatin and accumulation from early S phase onward, increasing to G2, which is consistent with previous reports (fig. S6, A and B) (26). We also confirmed that RNA interference–mediated depletion of sororin abrogated the Western blot signal (fig. S6C). To determine the stoichiometry of chromatinbound sororin by use of an independent method, we added a C-terminal SNAP-tag homozygously to the endogenous locus in the previously established RAD21-Halo, SMC3-E602-Spot U2OS cell line (fig. S6, D and E). The distribution of the dye-labeled sororin-SNAP resembled antibody-detected sororin regarding chromatin recruitment during cell cycle (fig. S6F). Step-wise photobleaching of SNAP-tagged sororin revealed one-step bleaching in 92.5 ± 4.7% of cases (fig. S6H), confirming that the majority of chromatin-bound sororin is present as a monomer. Having demonstrated that chromosomal sororin is localized exclusively at cohesion sites and that it exists predominantly as individual molecules, we addressed its association with cohesin. We used SIMinspector to analyze the association in G2 cells of sororin with chromatin-bound cohesin marked by RAD21Halo. This showed that 28.5 ± 0.064% of chromatin-bound cohesin colocalizes with sororin (Fig. 4, F and G), a proportion that matches the fraction of cohesin stably associated with chromatin in G2 phase HeLa cells, 8 MARCH 2024 • VOL 383 ISSUE 6687

1127

RES EARCH | R E S E A R C H A R T I C L E S

Fig. 5. Sister chromatid cohesion is maintained by individual cohesin rings. (A) Schematic depiction of labeling approach. U2OS RAD21-Halo cells were incubated with two different HaloTag ligands, each labeling ~50% of RAD21-Halo (controls are provided in fig. S7, A to C) The table shows expected frequencies of observing one Halo dye at a sororin site, or the other, or both at the same time for monomeric, dimeric, and multimeric cohesin. (B) Example 3D-SIM images of a G2 cell after incubation with JFX554 and JFX650 against RAD21-Halo (each labeling ~50%) and immunostaining with sororin antibody. A single z-slice of a nucleus is shown, and quantification is in (D) and (H). (C) Examples of classes I to III (sororin-containing signals) detected in G2 cells. Single z-slices of 3D-SIM example crops are shown. (D) Quantification of (C). Three-way colocalization was measured in two biologically independent experiments, with 1200 sororin sites analyzed per repeat. Means are 8.21% (sororin without RAD21-Halo), 45.00% (sororin with RAD21-Halo JFX554), 44.46% (sororin with RAD21-Halo JFX650), and 2.33% (sororin with RAD21-Halo JFX554 and RAD21-Halo JFX650). Dots indicate individual cells. Statistical testing by means of two-tailed Student’s t test yielded P < 0.000001. (E) Examples of classes IV to VI (cohesin at sororin-absent sites) detected in G2 cells. Single z-slices of 3D-SIM example crops are shown. (F) Cohesin at sites without sororin was analyzed in two biologically independent experiments, with 1200 cohesin signals analyzed per repeat. Means are 26.67% (2 × RAD21-Halo JFX554), 24.79% (2 × RAD21-Halo JFX650) and 48.54% (RAD21-Halo JFX554 and RAD21-Halo JFX650). Dots indicate individual cells. Statistical testing by two-tailed Student’s t test yielded P < 0.000001. (G) (Left) RASER-FISH example of probe HS443D9 colocalizing with RAD21-Halo in a dye-mixing experiment. (Right) Quantification. n = 34 probe signals from two biologically independent experiments. Mann-Whitney U test yielded P = 0.33. 1128

8 MARCH 2024 • VOL 383 ISSUE 6687

science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

estimated by using fluorescence recovery after photobleaching (FRAP) (22, 28). Sister chromatid cohesion is maintained by individual cohesin rings

To ascertain whether the cohesin complexes associated with sororin correspond to monomers, dimers, or multimers, we developed a labeling strategy to differentiate monomeric and multimeric proteins by means of 3D-SIM, making use of each HaloTag binding only a single dye ligand. Using our RAD21-Halo cell line, we determined conditions in which about half the RAD21-Halo molecules were labeled with one dye and the other half with a second dye, and only very few were unlabeled. If under these conditions sororin were associated with cohesin monomers, then we should never observe colocalization of the two dyes, whereas if it were associated with dimers, then we should observe colocalization in 50% of cases (Fig. 5A). To implement this, we determined the concentrations of the HaloTag dyes JFX554 and JFX650 necessary to saturate labeling when on their own and subsequently the concentrations necessary to obtain equal labeling when combined (fig. S7A). The extent of RAD21-Halo molecules that were unlabeled was then measured by use of a chase with a third dye. Using these conditions (fig. S7, B and C), we co-stained for sororin and measured the frequency of colocalization with neither, one or the other, or both Halo dyes. Given that we had previously identified monomeric, dimeric, and multimeric fractions of chromatin-bound cohesin (Fig. 1C), there were four potential outcomes of this experiment, with either monomers, dimers, multimers, or a mix of fractions enriched at cohesion sites labeled by sororin (Fig. 5A) and an equivalent set at sites lacking it. Sororin signals were detected with neither dye (8.21%), with dye 1 alone (JFX554, 45.00%), or with dye 2 alone (JFX650, 44.46%) (Fig. 5, B to F). They were observed to colocalize with both Halo dyes in only 2.33% of cases (Fig. 5, C and D), and inspection revealed that these instances consisted of a pair of close but clearly separate sororin molecules, each associated with a RAD21Halo molecule labeled with only a single dye (fig. S7D). The observed distribution implies that sites of chromatid cohesion marked by single molecules of sororin are associated with a monomeric form of cohesin. It also shows that most if not all chromatin-bound sororin colocalizes with cohesin in G2 because 90% reduction in regional populations over time spans equivalent to three generation lengths (table S3). Only one species’ population increased: the rosette skate (Leucoraja garmani) from the Northwest Atlantic. Global extinction risk

Fig. 1. Intrinsic population growth rate and population trends of deepwater sharks and rays. (A and B) Comparison of maximum intrinsic rate of population increase (rmax) and body size (where body size is the maximum linear dimension measured as maximum length or maximum disk width, in log scale) for 86 shark and ray species and 22 marine mammal species. Marine mammals are indicated by red, coastal sharks and rays by teal, pelagic sharks and rays by blue, and deepwater sharks and rays by dark blue. Barrels in (A) and the red line in (B) indicate deepwater species identified in the liver-oil trade. In (B), black lines indicate the median, and shaded areas indicate the kernel density estimate. (C) Regional population trends for deepwater sharks and rays where data were available from before 1980. Numbers represent changes in the abundance index over the observed time period. Lines denote the mean, and shaded regions indicate the 95% credible intervals. Species are ordered by rate of population change, and species color indicates global Red List status: CR, Critically Endangered; EN, Endangered; VU, Vulnerable; NT, Near Threatened; LC, Least Concern; and DD, Data Deficient. 1136

8 MARCH 2024 • VOL 383 ISSUE 6687

Sharks and rays are the most comprehensively assessed and threatened major radiation of the eight major deepwater lineages, totaling 2746 species assessed to date (Fig. 2). We estimate that one in seven (14.1%) deepwater sharks and rays are threatened with an elevated risk of extinction worldwide, on the basis of the observed number of threatened species combined with the estimated number of IUCN Data Deficient species that are likely to be threatened (10). Of a total of 521 species of deepwater sharks and rays, there are presently 60 (11.5%) threatened species: nine (1.7%) Critically Endangered, 20 (3.8%) Endangered, and 31 (6.0%) Vulnerable (Fig. 3, A and B). This level of extinction risk is more than twice the number reported from the first global assessment in 2014 [25 of 479 species (5.2%) (15)]. Deepwater sharks (15.2%, n = 43 of 283 species) are twice as threatened as deepwater rays (7.1%, n = 17 of 238 species). No species were flagged as Possibly Extinct nor assessed as globally Extinct or Extinct in the Wild; however, gulper sharks often become “commercially extinct” where intensive targeted fishing has occurred (16). Almost two-thirds of species (62.0%, n = 323) are assessed as Least Concern and 43 (8.3%) as Near Threatened. Trends in extinction risk

Many status changes since 2005 arose from the acquisition of new information combined science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

Fig. 2. Percentage of IUCN Red List categories for other deepwater taxa. Red lines indicate the best estimate of threatened species if all Data Deficient species faced a similar level of threat to data-sufficient species in the taxa. The shaded box shows groups where not all taxa have been assessed, as indicated by the number of species on the IUCN Red List (Red List), species diversity (Diversity), and percentage of species assessed (% Assessed).

with a more precautionary approach to recent reassessment (11). Nevertheless, these changes likely masked considerable, genuine change. We used a back-casting approach to retrospectively infer “genuine” changes that were most likely to have occurred (10) and to calculate a Red List Index spanning a half century (1970– 2020), where a value of 1 indicates that all species are Least Concern and a value of 0 indicates that species are Extinct. There was very little threat in the deep ocean before 1970, and the Red List Index declined by 7% from a retrospective estimate of 0.98 in 1970 to 0.91 in 2020 (Fig. 3D). The Red List Index for deepwater sharks and rays is comparable to estimates for all fish (sampled RLI2010 = 0.91) but worse than for marine fish (sampled RLI2010 = 0.97), and deepwater sharks and rays are generally less at risk than members of the class Chondrichthyes (RLI2020 = 0.77, n = 1199 species). The Red List Index trends were similar for deepwater sharks and rays, with a faster rate of decline between 1980 and 2005, which coincided with the advent and expansion of most deepwater fishing (17). Between 1980 and 2005, the inferred number of threatened species more than doubled from 22 to 57 (i.e., 4.2 to 10.9% of total deepwater diversity; Fig. 3, A and B). Overfishing is the main threat to the deep ocean

Overfishing is the primary threat to deepwater sharks and rays. Where threat was assessed, nearly every species (99.3%, n = 435 of 438) is threatened by overfishing (fig. S4). Most (87.7%, SCIENCE science.org

n = 384 of 438) deepwater species are taken as incidental catch in trawl, longline, and gillnet fisheries that target groups such as grenadiers (Macrouridae) and hakes (Merlucciidae). One-tenth of deepwater sharks and rays are targeted in fisheries (11.6%, n = 51 of 438 species). Nevertheless, one-third (35%, n = 21 of 60) of threatened species are targeted, primarily from three families: gulper sharks (Centrophoridae; 33.3%, n = 7 of 21), dogfishes (Squalidae; 19.0%, n = 4 of 21), and hardnose skates (Rajidae; 14.3%, n = 3 of 21). Other threats were minor and were identified for only 2.5% of species (n = 11 of 438), including pollution, climate change, and ecosystem modification. Shark liver-oil trade as a driver of defaunation and rising extinction risk

The use of sharks for their liver oil dates to ancient civilizations (e.g., for wound healing, heating fuel, or waterproofing boats) (16, 18, 19), but the globalized expansion of the trade and diversification of use is a relatively new phenomenon (Fig. 3E). International liver-oil trade is now a major driver of targeted fisheries and retention of incidental catch for many deepwater sharks around the world (table S6). Although both coastal and deepwater sharks are used for their liver oil, deepwater shark livers are preferred for their high squalene content (13, 16). Nearly two-thirds of threatened deepwater sharks (58.1%, n = 25 of 43 species) have been used for their liver oil (table S6). Of the 53 species (18.7% of deepwater shark diversity) taken for their liver oil, half (47.1%, n = 25 of

53) are threatened, with 7.5% Critically Endangered (n = 4), 22.6% Endangered (n = 12), and 17.0% Vulnerable (n = 9) (Fig. 3C, fig. S5, and table S6). Specifically, gulper shark liver oil is prized for its very high squalene content, and this family accounts for more than one-fourth of species taken for liver oil (26.4%). Most (93.3%, n = 14 of 15) gulper shark species have been identified in trade or inferred to be traded on the basis of information from regional fisheries that report catch under generic categories (e.g., Centrophorus spp.). The mixed end product makes it notably difficult to identify and quantify species composition. Other species identified in the trade include dogfishes (Squalidae; 26.5%, n = 9 of 34 species), sleeper sharks (Somniosidae; 50.0%, n = 8 of 16 species), and cow sharks (Hexanchidae; 100%, n = 4 of 4 species) (fig. S5). We show that liver-oil fisheries are not sustainable and result in steep population reductions and elevated extinction risk (Figs. 1C and 4). Indeed, these boom-and-bust fisheries are better thought of as nonrenewable mining extractions (20), peaking within 2 to 3 years of commencement and collapsing soon after (4738 km2 at a depth between 200 and 650 m (36). The design of such closures also needs to consider the management of cooccurring commercial fish species to minimize the trade-off of fishing displaced from new closures having a greater impact on sharks outside the closures. In addition to reducing fishing mortality, habitat protection is needed science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

for recovery and maintaining healthy population levels. Habitat dependence is poorly understood, but some documented examples exist, including deepwater skates using hydrothermal vents to incubate their eggs (37). Hydrothermal vents have received global attention for prospective mineral mining operations (3), and this provides additional motivation for their protection. Such spatial protection measures have positive impacts, not only on deepwater sharks and rays but on all demersal biodiversity biomass and richness (38). Trade and fishing regulations specific to deepwater sharks and rays are needed to ensure legal, traceable, and sustainable trade and prevent further endangerment. For example, international liver-oil trade regulations could be established to ensure that the trade is not driving extinction risk. At present, there are limited means of determining what species make up internationally traded squalene oil; either the oil may be a by-product of sustainable fisheries or the present lack of regulations could be masking the trade of threatened species (39). In the first instance, highly threatened species (e.g., gulper sharks and look-alikes; fig. S5) could be listed in Appendix II of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) to regulate international trade. There are >145 shark and ray species listed in CITES Appendix II (40), but to date, none are deepwater species. National catch regulations, such as landing the shark whole (much like a finsnaturally-attached policy), would improve species identification and fisheries characterization and monitoring and would guide appropriate national-level management, such as catch limits, catch prohibitions, or closed areas. There is also an urgent need to assess the risk to human health and determine the appropriateness of sourcing and using shark squalene for medical applications. The extinction risk of deepwater sharks and rays is presently much less than that of shallowwater species, but their potential for recovery is low. We have the evidence to act more proactively for the deep ocean and learn from the mistakes that have driven more than half of coastal and pelagic species to be threatened (11). Achieving sustainable fisheries for most deepwater sharks and rays would be challenging and require high management capacity, ecological knowledge, and implementation of routine rigorous monitoring. Effective precautionary actions are needed to ensure that the largest ecosystem on the planet maintains its biodiversity and that half of the world’s shark and ray species have refuge from the global extinction crisis. RE FE RENCES AND N OT ES

1. E. Ramirez-Llodra et al., Biogeosciences 7, 2851–2899 (2010).

SCIENCE science.org

2. M. J. Costello, C. Chaudhary, Curr. Biol. 27, R511–R527 (2017). 3. Z. Da Ros et al., Mar. Policy 108, 103642 (2019). 4. United Nations General Assembly, “Further revised draft text of an agreement under the United Nations Convention on the Law of the Sea on the conservation and sustainable use of marine biological diversity of areas beyond national jurisdiction” (A/CONF.232/2022/5, United Nations, 2022); https://www.un.org/bbnj/. 5. Convention on Biological Diversity (CBD), “15/4. KunmingMontreal Global Biodiversity Framework” (CBD/COP/ DEC/15/4, United Nations Environment Programme, 2022); https://www.cbd.int/doc/decisions/cop-15/ cop-15-dec-04-en.pdf. 6. E. Cortés, Methods Ecol. Evol. 7, 1136–1145 (2016). 7. N. Pacoureau et al., Nature 589, 567–571 (2021). 8. S. H. Butchart et al., Science 328, 1164–1168 (2010). 9. S. H. Butchart et al., PLOS ONE 2, e140 (2007). 10. R. H. L. Walls, N. K. Dulvy, Biol. Conserv. 246, 108459 (2020). 11. N. K. Dulvy et al., Curr. Biol. 31, 4773–4787.e8 (2021). 12. J. Nielsen et al., Science 353, 702–704 (2016). 13. P. M. Kyne, C. A. Simpfendorfer, Marine Conservation Biology Institute, IUCN Shark Specialist Group, A Collation and Summarization of Available Data on Deepwater Chondrichthyans: Biodiversity, Life History and Fisheries (Florida Museum of Natural History, 2007). 14. C. G. Mull et al., Sci. Data 9, 559 (2022). 15. N. K. Dulvy et al., eLife 3, e00590 (2014). 16. R. C. Anderson, H. Ahmed, “The shark fisheries in the Maldives” (Ministry of Fisheries and Agriculture, Republic of Maldives, Food and Agriculture Organization of the United Nations, 1993). 17. T. Morato, R. Watson, T. J. Pitcher, D. Pauly, Fish Fish. 7, 24–34 (2006). 18. S. F. de Borhegyi, Southwest. J. Anthropol. 17, 273–296 (1961). 19. M. A. MacNeil et al., J. Fish Biol. 80, 991–1018 (2012). 20. E. A. Norse et al., Mar. Policy 36, 307–320 (2012). 21. K. V. Akhilesh, C. Anulekshmi, K. K. Bineesh, U. Ganga, N. G. K. Pillai, Indian J. Fish. 67, 8–15 (2020). 22. F. Dent, S. Clarke, “State of the global market for shark products” (FAO Fisheries and Aquaculture Technical Paper no. 590, Food and Agriculture Organization of the United Nations, 2015). 23. K. J. Fisher et al., NPJ Vaccines 8, 14 (2023). 24. S. Corsolini, K. Pozo, J. S. Christiansen, Rend. Lincei Sci. Fis. Nat. 27, 201–206 (2016). 25. A. Arkhipkin et al., ICES J. Mar. Sci. 80, 578–590 (2023). 26. M. Estalles, N. M. Coller, M. R. Perier, E. E. Di Giácomo, Aquat. Living Resour. 24, 193–199 (2011). 27. C. A. Simpfendorfer, N. K. Dulvy, Curr. Biol. 27, R97–R98 (2017). 28. L. Fauconnet et al., Fish. Res. 209, 230–241 (2019). 29. B. Talwar, E. J. Brooks, J. W. Mandelman, R. D. Grubbs, Mar. Ecol. Prog. Ser. 582, 147–161 (2017). 30. International Council for the Exploration of the Sea (ICES), Working Group on Elasmobranch Fishes (WGEF), ICES Scientific Reports series, vol. 2, issue 77, J. Batsleer, P. Lorance, Eds. (ICES, 2020); https://doi.org/10.17895/ ices.pub.7470. 31. N. K. Dulvy, H. K. Kindsvater, in Conservation for the Anthropocene Ocean, P. S. Levin, M. R. Poe, Eds. (Academic Press, 2017), pp. 39–64. 32. L. Georgeson et al., ICES J. Mar. Sci. 77, 1711–1727 (2020). 33. European Union, “Establishing specific conditions for fishing for deep-sea stocks in the North-East Atlantic and provisions for fishing in international waters of the North-East Atlantic and repealing council regulation (EC)” (Regulations 2347/2002 and 2016/2336, European Union, 2016). 34. General Fisheries Commission for the Mediterranean (GFCM), “On the management of certain fisheries exploiting demersal and deep-water species and the establishment of a fisheries restricted area below 1000 m” (REC.CM-GFCM/ 29/2005/1, Food and Agriculture Organization of the United Nations, 2005). 35. C. Hyde et al., Front. Mar. Sci. 9, 968853 (2022). 36. Australian Fisheries Management Authority (AFMA), “Upperslope dogfish management strategy” (AFMA, 2012). 37. P. Salinas-de-León et al., Sci. Rep. 8, 1788 (2018).

38. J. G. Hiddink, T. Hutton, S. Jennings, M. J. Kaiser, ICES J. Mar. Sci. 63, 822–830 (2006). 39. D. Cardeñosa et al., Conserv. Lett. 15, e12910 (2022). 40. Secretariat of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), “Notifications to the Parties” (No. 2023/005, CITES, 2023); https://cites.org/eng/notif/index.php. 41. IUCN, The IUCN Red List of Threatened Species, version 2022-1 (2022); https://www.iucnredlist.org. AC KNOWLED GME NTS

We thank all members of the IUCN Species Survival Commission Shark Specialist Group and invited national, regional, and international experts who attended Shark Specialist Group Red List workshops, as well as all the experts who contributed data and their expertise in person or by correspondence. The scientific results and conclusions, as well as any views or opinions expressed herein, are those of the author(s) and do not necessarily reflect those of institutions or data providers. We thank E. Digel for the collation of life history data for the rmax estimation, as well as J. Lawson, J. Bigman, Z. Crysler, A. Hood, G. Ralph, and H. Yan for their contributions to workshop organization where IUCN Red List assessments for deepwater species were undertaken. The sawfish vector was created by C. Camilo Julián-Caballero (with no additional modifications; https://creativecommons.org/ licenses/by/3.0/). Funding: This project was funded by the Shark Conservation Fund, a philanthropic collaborative that pools expertise and resources to meet the threats facing the world’s sharks and rays. The Shark Conservation Fund is a project of Rockefeller Philanthropy Advisors. This work was funded by the Shark Conservation Fund as part of the Global Shark Trends Project to N.K.D. and C.A.S. N.K.D. was supported by Natural Science and Engineering Research Council Discovery and Accelerator Awards and the Canada Research Chairs Program. D.A.E. was supported by Save Our Seas Foundation and the South African Shark and Ray Protection Project, implemented by WILDTRUST, funded by the Shark Conservation Fund. This work was also funded by Natural Science and Engineering Research Council Discovery Award 5013566 (N.K.D.), Natural Science and Engineering Research Council Discovery Accelerator Award 462291 (N.K.D.), and Canada Research Chairs Program 1228186 (N.K.D.). Author contributions: Conceptualization: B.F., N.P., C.L.R., J.H.M., C.S.S., W.J.V., C.A.S., N.K.D.; Methodology: N.P., J.H.M., N.F.-B.; Validation: B.F., N.P., C.L.R., J.H.M., N.F.-B., C.S.S., W.J.V., C.A.S., N.K.D.; Formal analysis: B.F., N.P., C.L.R., J.H.M., N.F.-B., R.W.J., C.A.S., N.K.D.; Resources: B.F., N.P., C.L.R., R.W.J., P.C., P.A.M.-F., A.F.N., B.K., C.F.C., J.-M.C., R.K.D., D., D.A.E., D.F., S.M.C.F., M.P.F., C.H., H.I., D.W.K., R.W.L., F.N., A.M.O., G.R., G.J.S., I.V.V., T.I.W.; Data curation: B.F., N.P., C.L.R., P.M.K., R.A.P., R.H.L.W., K.B.H., C.A.S., N.K.D.; Writing – original draft: B.F., N.P., C.L.R., J.H.M., N.F.-B., C.S.S., W.J.V., C.A.S., N.K.D.; Writing – review and editing: B.F., N.P., C.L.R., J.H.M., N.F.-B., C.S.S., W.J.V., R.W.J., P.C., P.A.M.-F., A.F.N., P.M.K., R.A.P., R.H.L.W., D.H.D., K.B.H., B.K., C.F.C., J.-M.C., R.K.D., D., D.A.E., D.F., S.M.C.F., M.P.F., C.H., H.I., D.W.K., R.W.L., F.N., A.M.O., G.R., G.J.S., I.V.V., T.I.W., C.A.S., N.K.D.; Visualization: B.F., N.P., C.L.R., J.H.M., N.F.-B., C.S.S., W.J.V., D.H.D., C.A.S., N.K.D.; Supervision: B.F., C.L.R., C.A.S., N.K.D.; Project administration: B.F., C.L.R., R.W.J., P.C., P.A.M.-F., A.F.N., P.M.K., C.A.S., N.K.D.; Funding acquisition: R.W.J., P.C., P.A.M.-F., A.F.N., C.A.S., N.K.D. Competing interests: The authors declare that they have no competing interests. Data and materials availability: IUCN Red List of Threatened Species assessments are publicly available on the IUCN Red List (https://www.iucnredlist.org/), and a permanent summary is documented here (41). Time-series and trait data are publicly available on https://www.sharkipedia.org/. License information: Copyright © 2024 the authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original US government works. https://www.science.org/ about/science-licenses-journal-article-reuse SUPPLEMENTARY MATERIALS

science.org/doi/10.1126/science.ade9121 Materials and Methods Figs. S1 to S6 Tables S1 to S8 References (42–75) MDAR Reproducibility Checklist Submitted 17 September 2022; resubmitted 28 April 2023 Accepted 2 November 2023 10.1126/science.ade9121

8 MARCH 2024 • VOL 383 ISSUE 6687

1141

RES EARCH | R E S E A R C H A R T I C L E S

QUANTUM IMAGING

Adaptive optical imaging with entangled photons Patrick Cameron1*, Baptiste Courme2,3, Chloé Vernière2, Raj Pandya2,3,4, Daniele Faccio1, Hugo Defienne1,2* Adaptive optics (AO) has revolutionized imaging in fields from astronomy to microscopy by correcting optical aberrations. In label-free microscopes, however, conventional AO faces limitations because of the absence of a guide star and the need to select an optimization metric specific to the sample and imaging process. Here, we propose an AO approach leveraging correlations between entangled photons to directly correct the point spread function. This guide star–free method is independent of the specimen and imaging modality. We demonstrate the imaging of biological samples in the presence of aberrations using a bright-field imaging setup operating with a source of spatially entangled photon pairs. Our approach performs better than conventional AO in correcting specific aberrations, particularly those involving substantial defocus. Our work improves AO for label-free microscopy and could play a major role in the development of quantum microscopes.

abel-free microscopes are essential for studying biological systems in their most native states, and in recent years, their performance has been enhanced by the use of nonclassical light sources. In particular, sources of entangled photon pairs, which illuminate an object and are detected in coincidence to form an image, are at the basis of numerous protocols (1). For example, they are used in bright-field imaging configurations to enhance the spatial resolution (2–5), achieve sub–shot-noise imaging (6), and improve the contrast in the presence of noise and losses (7, 8). In phase imaging, they can be used to augment the contrast in both confocal (9) and wide-field (10, 11) differential interference contrast (DIC) systems and are the basis of new modalities, including quantum holography (12, 13), reconfigurable phase-contrast microscopy (14), and three-dimensional (3D) imaging (15). Finally, they can also improve time-gated imaging protocols such as optical coherence tomography (OCT) by reducing dispersion (16, 17) and enhancing depth sensitivity (18). However, whether in their classical or quantum version, all of these methods are sensitive to optical aberrations created by the specimens being imaged or by the imaging system itself. If left uncorrected, these effects negate the benefits gained by these techniques and compromise their practical use. Adaptive optics (AO) can be used to mitigate these aberrations. To operate, a light-emitting source or a point-like structure in the sample is identified as a guide star. The wavefront accumulates aberrations while propagating

L

1

School of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK. 2Sorbonne Université, CNRS, Institut des NanoSciences de Paris, INSP, F-75005 Paris, France. 3 Laboratoire Kastler Brossel, ENS-Universite PSL, CNRS, Sorbonne Universite, College de France, 75005 Paris, France. 4 Cavendish Laboratory, University of Cambridge, Cambridge CB3 0HE, UK. *Corresponding author. Email: [email protected] (P.C.); [email protected] (H.D.)

1142

8 MARCH 2024 • VOL 383 ISSUE 6687

out of the sample, and these aberrations are then measured by a Shack-Hartmann sensor (direct AO) or a focus-forming process (indirect AO). Wavefront correction is then applied to cancel out the aberrations using a deformable mirror or a spatial light modulator (SLM). Over the past decades, AO has played a major role in the development of advanced imaging systems, particularly fluorescence microscopes (19, 20). In the absence of a guide star, however, the point spread function (PSF) and thus the aberration information are not directly accessible. This is especially the case in most label-free and linear microscopy systems. To circumvent this issue, wavefront sensorless, image-based AO methods have been developed (20–22). They are based on the principle that the image resulting from the convolution between the specimen and the PSF has optimum quality only when the aberrations have been fully

compensated. In practice, an image metric is first defined and then optimized by acting with the wavefront-shaping device. The appropriate choice of the metric depends on the image formation process of the microscope used and the nature of the sample. The most commonly used metrics include total output intensity (23), image contrast (24), lowfrequency content (25), and sharpness (26, 27). In recent years, AO has enabled aberration correction in several label-free microscope modalities, including bright-field (25), quantitative phase-contrast (28), DIC (29), and OCT (30). One of the primary hurdles in achieving effective image-based AO lies in the requirement to define distinct metrics for each microscope modality and for varying specimen types. Furthermore, certain metrics may introduce systematic errors. For instance, when capturing volumetric samples, the utilization of an image sharpness metric to correct defocus aberration typically yields multiple solutions corresponding to different imaging planes within the sample. In this work, we present a quantum-assisted AO (QAO) method that harnesses the entanglement between photon pairs to directly access the imaging system PSF, and thus the aberration information, in the absence of a guide star. This approach also eliminates the need to define a specific image-based metric and is thus independent of the imaging modality and specimen under study. We demonstrate the effectiveness of this approach by imaging biological samples using classical and quantum bright-field transmission imaging systems in the presence of aberrations. In particular, we present experimental situations in which our technique leads unambiguously to

Fig. 1. Concept of QAO. (A) An object is illuminated by spatially entangled photon pairs and imaged onto a singlephoton–sensitive camera. The imaging system between the object and the camera is not represented for clarity. Photon pairs are strongly correlated in the object plane. (B) Without optical aberrations, a sharp intensity image of the object is acquired, and photon pairs are still correlated at the camera plane. (C) Photon pair correlations are visualized by measuring the spatial second-order correlation function G(2) and projecting it onto specific coordinates. Such a G(2) projection is proportional to the system’s PSF and shows a narrow peak at its center. (D to F) With aberrations present, the system is not limited by diffraction and the pairs are no longer correlated at the camera plane (D), resulting in a blurred intensity image (E) and a distorted G(2) projection (F). In QAO, aberrations are corrected using an SLM to maximize the central value of the G(2) projection. science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

the optimal correction while classical imagebased AO methods fail. Concept

In the QAO scheme, spatially entangled photon pairs are incident on an object (t), which is then imaged onto a single-photon–sensitive camera (Fig. 1A). As in classical incoherent illumination, the intensity image (I) produced at the output results from a convolution between the absolute value-squared PSF (h) and the object as I = |h|2  |t|2 (Fig. 1B). In addition, the photons forming the image are also pairwise correlated in space, which arises from their entanglement (31). The second-order spatial correlation function G(2) can be written as G ð2Þ ðr10 ; r20 Þ ¼ jfðr1 ; r2 Þt ðr1 Þt ðr2 Þhðr1 Þhðr2 Þj2

ð1Þ

where f(r1,r2) is the spatial two-photon wave function of the photon pair in the object plane with transverse coordinates r1 and r2 (32). In general, G(2) is a complicated function that depends on the PSF, the object, and the spatial correlations between photon pairs. Under specific experimental conditions, however, one can simplify Eq. 1 and average G(2) along specific spatial axes to extract information only linked to the system’s PSF. In particular, if the object is positioned in the Fourier plane of the source, then the two-photon wave function can be approximated by f(r1,r2) ≈ d(r1 + r2), which describes near-perfect anticorrelations between photon pairs originating from spatial entanglement. Using this configuration, we can measure the sum coordinate projection of G(2), defined as C+(dr+) = ∫G(2)(r, dr+ – r)dr, with dr+ = r1 + r2 being the sum coordinate. Assuming weak optical aberrations in the imaging system, C+ can be approximated as C þ ðdrþ Þ ≈ K j½h  hŠðdrþ Þj2

ð2Þ

where K = ∫|t(r)t(–r)|2d(r) is a constant independent of h, and K represents the photonpair transmission rate through the sample. For example, Fig. 1C shows a sum-coordinate projection simulated in the case of a diffractionlimited imaging system. It has a very specific shape, with a narrow peak at its center, just like the corresponding PSF. In the presence of optical aberrations, however, the PSF is distorted, as is the sum-coordinate projection (Fig. 1F), with a central correlation peak that decreases and spreads. The value of the central peak is therefore maximal when the imaging system is limited by diffraction. In QAO, we use this value as a feedback signal to compensate for optical aberrations in the imaging system using a modal-based AO algorithm. Simulations and additional experimental data supporting SCIENCE science.org

Fig. 2. Experimental setup. Spatially entangled photon pairs centered at 810 nm are produced with type I SPDC using a 405-nm collimated continuous-wave laser and a 0.5-mm-thick BBO nonlinear crystal (NLC). Blue photons are then filtered out by a low-pass filter (LP) at 650 nm. The sample is illuminated by the photon pairs while being positioned in the Fourier plane of the crystal (f1 = 100 mm). It is subsequently imaged (with a magnification of 1) onto the EMCCD camera using two 4f imaging systems, f2-f3 and f4-f5. The SLM used to correct aberrations is positioned in a Fourier plane of the sample between f4 and f5. For clarity, it is depicted in transmission, but in practice, it operates in reflection. Optical aberrations can be introduced at either the optical planes A1 (near the sample plane) or A2 (near the Fourier plane). Note that plane A1 is deliberately placed at a small distance from the object plane to introduce sufficient aberrations. To detect only near-degenerate photon pairs, a band-pass filter (BP) at 810 ± nm is positioned in front of the camera.

this result are provided in section III of the supplementary materials. Results

Figure 2 shows the experimental setup. Spatially entangled photon pairs are generated through spontaneous parametric down conversion (SPDC) in a thin b-barium-borate (BBO) crystal cut for type I phase matching. Using lens f1, the output surface of the crystal is Fourier imaged onto the sample. Subsequently, the sample is imaged onto the camera using two 4f imaging systems, f2-f3 and f4-f4. Specimen- and system-induced aberrations can be introduced in the imaging system in planes A1 and A2, respectively. An SLM, used to correct for aberrations, is placed in a Fourier plane of the sample. Photon pairs transmitted through the system are detected at the output using an electron-multiplying charge-coupled device (EMCCD) camera for measuring both conventional intensity images and photon correlations following the technique described in (33) (see also sections I and II of the supplementary materials). To illustrate our approach, we placed a biological sample, a honeybee mouthpiece on a microscope slide, in the sample plane and captured its intensity image in transmission (Fig. 3A). In the absence of aberrations, the sum-coordinate projection exhibits a distinct and sharp peak, as shown in Fig. 3D. However, when aberrations are present, the image becomes blurred, and the correlation peak is spread and distorted, as depicted in Fig. 3, B and E, respectively. In this demonstration, we induced aberrations by introducing a second SLM at plane A2 that displays a low-frequency random phase pattern (see section V of the supplementary materials). To correct aberrations, we used a modalbased AO algorithm that includes C0þ as a feedback parameter, where C0þ ¼ C þ ðdrþ ¼ 0Þ. This algorithm involves introducing predetermined aberrations on the SLM using Zernike

polynomial modes. In our study, we consider all modes with radial numbers n ≤ 5 and azimuthal numbers |m| ≤ n, excluding piston, tip, and tilt. For each Zernike mode (Znm), we recorded five sum-coordinate projections with distinct, known bias amplitudes (anm). In each measurement, the SLM phase qnm is thus modulated according to qnm ¼ qnm

1

þ amn Znm

ð3Þ

where qnm–1 represents the optimal phase correction obtained for the previous mode. Such a phase modulation approach is commonly used in classical modal AO (22). For example, the values of C0þ obtained from the sum-coordinate projections for the modes Z3 3 and Z31 are shown in Fig. 3H. The positions of the maxcorr ima, denoted as acorr 33 and a13 , representing the optimal corrections for their respective mode, are determined using a Gaussian fitting model (see section IV of the supplementary materials). After several optimization steps, a narrow peak is recovered in the sumcoordinate projection (Fig. 3F), and a sharp image appears in the intensity (Fig. 3C). Visual comparison with the aberration-free images shows a clear improvement after correction. Quantitatively, one can use the structural similarity index measure (SSIM) as a metric to quantify image similarity. Using the aberrationfree image as a reference (Fig. 3A), we found that SSIM = 77.89% for the uncorrected image (Fig. 3C) and SSIM = 98.41% for the corrected image (Fig. 3B). Note that here, although the object is illuminated by a source of entangled photon pairs, the quantum properties of which are crucial for measuring C+ and thus correcting aberrations, the imaging process itself is purely “classical” because the output image is obtained through a simple intensity measurement. QAO offers several advantages compared with classical AO. First, as demonstrated in Fig. 3, it does not require a guide star. All 8 MARCH 2024 • VOL 383 ISSUE 6687

1143

RES EARCH | R E S E A R C H A R T I C L E S

Fig. 3. Results of QAO correction. (A to C) Intensity images of a biological sample (bee head) acquired in transmission without aberrations (A), with aberrations before correction (B), and after correction (C). Using the aberration-free intensity image as a reference, we found structural similarity values of SSIM = 77.89% and SSIM = 98.41% for the uncorrected and corrected images, respectively. (D to F) Correlation images C+ (dr+ = r1 + r2) measured without aberrations (D), with aberrations before correction (E), and with aberrations after correction (F).

photon pairs forming the image have information about the system aberrations at every point because these are encoded in their spatial correlations. Additionally, QAO perfor1144

8 MARCH 2024 • VOL 383 ISSUE 6687

(G) Optimal phase pattern obtained after correction and displayed on the SLM. (H) Values of the sum-coordinate projection peaks Cþ 0 in function of the coefficient amn for two Zernike modes, Z3 3 and Z31 (crosses). acorr 33 ¼ 0:2253 and acorr ¼ 0:6881 are the two optimal correction values for each mode returned 13 by the fit (solid lines). Each intensity and sum-coordinate projection was obtained from 105 frames approximately equivalent to a 2-min acquisition. White scale bar, 400 mm.

mance does not depend on the sample properties or the imaging modality. The spatial correlation structure is a property of the illumination itself and is only affected by the system

aberrations. This implies that QAO will converge irrespective of the observed sample type, ranging from nearly transparent samples (e.g., cells) to denser ones (e.g., layered minerals), science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

A

Metric (normalised)

1 0.8 0.6 0.4 0.2 0

-5

-4

-3

-1

-2

0

1

Defocus Coefficient ( 1

x (pixel)

B

100 1

1

x (pixel)

C

100

x (pixel)

100 1

E

0

1

1

y (pixel)

100

x (pixel)

100 1

F

100

0

1

y (pixel)

and regardless of their complexity or smoothness of structure. In fig. S9 of the supplementary materials, we provide additional experimental results obtained using various sample types that demonstrate this. QAO thus surpasses all image-based AO techniques for which the chosen metrics and optimization performances depend on the properties of the sample. We show in the next section that in certain imaging situations, image-based approaches can

1

1

y (pixel)

100

x (pixel)

100 1

100

100

1

0

1

y (pixel)

100

Correlations (AU)

Intensity (AU) 2000

0

G

Norm. Intensity

Norm. Intensity

Norm. Intensity y (pixel)

100

1

100

1

100 1

y (pixel)

100

1

x (pixel)

y (pixel)

y (pixel)

0

1

D

Norm. Intensity

Norm. Intensity

1

y (pixel)

100 1

100

1 Norm. Intensity

1

5

4

y (pixel)

0

3

2

02)

y (pixel)

100

1

1

SCIENCE science.org

Quantum

Low freq.

Contrast

PIB

y (pixel)

Fig. 4. Comparison between QAO and classical image-based AO. (A) Values of three image quality metrics, PIB, image contrast, and low frequencies, and Cþ 0 in function of the defocus correction coefficient a02. Data are shown by the crosses, and the fits used to find optimal values (acorr 02 ) are shown by solid lines. In this experiment, the object is 3D (three thin copper wires). (B to G) Intensity images (grayscale), central regions of C+ (inset), and intensity profile for a single column (line plot) for various defocus corrections on SLM without correction (acorr 02 ¼ 0 and SSIM = 76.39%) (B); optimal correction found using the PIB metric (acorr 02 ¼ 3:1427 and SSIM = 50.56%) (C); optimal correction found using an image contrast metric (acorr 02 ¼ 3:1427 and SSIM = 52.29%) (D); optimal correction found using the low spatial frequencies metric (acorr 02 ¼ 0:2677 and SSIM = 72.61%) (E); optimal correction found using QAO (acorr 02 ¼ 1:6622 and SSIM = 96.83%) (F); and no aberration (G). Vertical red lines show selected column for profile plots. Each intensity image and sum-coordinate projection was obtained from 105 frames approximately equivalent to a 2-min acquisition. White scale bar, 400 mm.

4500

-2

lead to systematic error in aberration correction, whereas QAO converges to the correct solution. We consider a situation where the sample has a 3D structure, which is very common in microscopy. In such a case, it is known that it is not possible to correct for defocus aberrations properly. Indeed, when using an image quality metric, it may optimize for the wrong focal plane within the sample. Because the

12

sample structure has no effect in QAO, defocus correction is possible. In our demonstration, we chose an object consisting of three copper wires, each with an approximate thickness of 0.15 mm and spaced ~5 mm apart along the optical axis. We then induced defocus aberration with strength aaber 02 ¼ 2 by placing a second SLM in plane A2. Sum-coordinate projection and intensity images are acquired for a wide range of defocus corrections (a02 ∈ [–5,5]) 8 MARCH 2024 • VOL 383 ISSUE 6687

1145

RES EARCH | R E S E A R C H A R T I C L E S

Fig. 5. Application to quantum imaging. (A) Intensity image formed by one photon of a pair used as the reference photon. (B) Intensity image formed by the other photon used to illuminate the sample, here a bee’s leg. (C) One-centimeterthick piece of PDMS inserted in plane A1 in the setup in Fig. 2 to induce aberrations. (D) Optimal phase pattern obtained after correction and displayed on the SLM. (E to G) Anticorrelation images R(r) ≈ G(2)(r, –r) obtained without induced aberration (SNR = 29) (E), with aberration (SNR ≈ 3) (F), and after aberration correction

programmed on the correction SLM. At each step, values of three standard AO image quality metrics are calculated from the intensity image: power in bucket (PIB) (23), image contrast (24), and low-frequency content (25). C0þ is also retrieved from the sumcoordinate projection. Figure 4A shows the four corresponding optimization curves. First, we observed that the various classical AO 1146

8 MARCH 2024 • VOL 383 ISSUE 6687

(SNR = 15) (G). Insets show the sum-coordinate projection in each case. Each sumcoordinate projection to achieve QAO was obtained from 2.2 × 104 frames and ~3 min of acquisition, and each anticorrelation image was obtained from 107 frames and ~24 hours of acquisition. Note that the EMCCD camera used here is different from that used in Figs. 2 and 3 and has a frame rate of just 130 frames/s. The total intensity (i.e., total number of photons) measured on the camera was the same in the presence of aberrations before and after correction. White scale bar, 400 mm.

metrics return different optimization coefficients, highlighting their dependency on the object’s structure. Then, by examining the intensity images captured while programming each optimal correction phase pattern (Fig. 4, B to E), it becomes evident that none of these metrics properly corrected the aberrations. Indeed, the aberration-free image in Fig. 4G clearly shows that only the bottom wire is in

the focal plane, which is not the case in any of the intensity images shown in Fig. 4, B to E. Conversely, QAO converges to the correct solution, as seen in the intensity image shown in Fig. 4F (SSIM = 96.83%). The optimum value found with QAO is acorr 02 ¼ 1:622, which differs slightly from the value of 2 (opposite of aaber 02 ¼ 2) that we would expect to find. This is because QAO corrects not only for the science.org SCIENCE

RESE ARCH | R E S E A R C H A R T I C L E S

intentionally introduced defocus aberrations in the A2 plane but also for those inherent in the imaging system. This is also shown by the fact that the correlation peak in Fig. 4F (inset) is slightly narrower than the one in Fig. 4G (inset). This demonstration uses a very simple 3D sample: three spaced wires. However, QAO can in principle be used with more complex 3D samples as long as they remain within the regime of weak aberrations, i.e., no strong scattering and absorption. Such samples are typically studied with optical tomography methods, where QAO can therefore be used after adapting the mathematical formalism to account for the thickness of these objects (15). Finally, to showcase its potential for quantum imaging, QAO was applied to a “quantum” variant of the bright-field imaging setup depicted in Fig. 2. In such a scheme, only one photon of a pair interacts with the object, and its twin serves as a reference. For that, the sample is placed on only one half of the object plane (x > 0), as observed in the intensity images shown in Fig. 5, A and B. To interpret this specific arrangement in Eq. 1, we theoretically define the object such that t(x < 0) = 1 and t(x > 0) describes the object. Then, the final image (R) is obtained by measuring photon correlations between all symmetric pixel pairs of the two halves, i.e., R(r) ≈ G(2)(r,–r) (see section II of the supplementary materials). This image is called an anticorrelation image and is shown in Fig. 5E. As demonstrated in previous studies (5, 8, 12), such a quantum scheme offers some advantages over its classical counterpart, including an enhanced transverse spatial resolution and increased resilience against noise and stray light. In the presence of aberrations, however, we show that this imaging technique becomes highly impractical and thereby loses all its purported advantages. For example, Fig. 5F shows an anticorrelation image acquired after inserting a 1-cm-thick layer of polydimethylsiloxane (PDMS) (Fig. 5C) on both photon paths in plane A1 to induce optical aberrations. Not only is the resulting image blurred, leading to a complete loss of the expected resolution advantage, its signal-to-noise ratio (SNR) is greatly reduced, rendering the sample almost indiscernible (SNR ≈ 3). After applying QAO, we retrieve an anticorrelation image shown in Fig. 5G that has a spatial resolution closer to that without aberrations and is of much better quality (SNR = 15). The inset of Fig. 5G shows the corresponding sum-coordinate projection, exhibiting a much narrower and more intense peak, and Fig. 5D shows the optimal SLM phase pattern. In addition, when comparing carefully the sum-coordinate projections without aberrations (Fig. 5E, inset) and after correction (Fig. 5G, inset), we observe that QAO also corrected for a small PSF asymSCIENCE science.org

metry present in the initial system. Compensating for this asymmetry results in a more uniform output image (Fig. 5G) than that obtained in the aberration-free case (Fig. 5E). By using QAO, we then show a substantial improvement of the output image quality in terms of resolution, SNR, and uniformity, effectively restoring the operational capability of this quantum imaging technique. Discussion

We have introduced a QAO method that eliminates the need for a guide star. By optimizing the spatial correlations of entangled photon pairs, we can directly optimize the system PSF and compensate for optical aberrations. QAO circumvents certain limitations linked to conventional image-based AO and is particularly well suited for classical and quantum full-field, label-free, and linear microscopy systems. In our study, we demonstrate QAO in the regime of weak optical aberrations. We used artificial layers to simulate aberrations commonly encountered in real-world microscopy systems, including system-induced (e.g., astigmatism, defocus, comatic aberrations due to objectives, and misalignment) and weak specimen-induced aberrations (e.g., translucent tissues surrounding the sample, immersion liquid, and sample support). At this stage, QAO is not demonstrated in the scattering regime, although preliminary results obtained with more complex aberrations show promise (figs. S14 to S16). Within this regime of weak aberrations, there are no fundamental barriers preventing the use of QAO in other, more advanced label-free imaging systems. For instance, QAO could: (i) improve current image-based approaches used in optical coherence tomography (34); (ii) be combined with 3D imaging techniques, some of which already used entangled photon pair sources (15); (iii) be used in phase imaging and high-numerical aperture imaging schemes (figs. S19 and S20); and (iv) be adapted to reflection geometries by using multiple SLMs (simulation in fig. S17). As with classical AO, the effectiveness of the correction found with QAO will always depend on the imaging modality and the nature of aberrations present. For instance, spatially variant aberrations will restrict the field of view in the corrected image, although this limitation might be circumvented by using alternative AO designs such as conjugate and multiconjugate AO (35, 36). Finally, it is important to note that QAO is not yet adaptable in fluorescence microscopy, but this could change in the future with the emergence of biomarkers that emit photon pairs (37). In practice, the main limitation of QAO is its long operating time. Using an EMCCD camera, acquisition times of ~1 min are required to

measure one sum-coordinate projection. This means that correcting for multiple orders of aberration can take up to several hours. However, this technical limitation can be overcome by using alternative camera technologies, some of which are already available commercially. For example, single-photon avalanche diode (SPAD) cameras have been used to capture sum-coordinate projections at speeds up to 100 times faster than EMCCD cameras using similar photon pair sources (38, 39). Another promising technology is the intensified Tpx3cam camera, which has recently been used for similar correlation measurements (40–42). As technology improves, we expect acquisition times soon to be on the order of seconds, which would result in correction times on the order of minutes. In addition, here, we chose Zernike polynomials as the basis set for aberration representation even though they may not be optimal (43). In particular, if the aberrations are more complex, then wavefront-shaping approaches using Hadamard or random bases should be considered (44–48). In our demonstration, QAO uses entanglement between photons. Indeed, replacing our source by classically anticorrelated photons would yield a formally output mea different surement, i.e., Cclþ ¼ hj2  hj2 (see section XIV of the supplementary materials). Such a metric could still be used for AO but is genuinely less sensitive compared with entangled photons (fig. S18) and thus is not suitable for phase imaging. In addition, producing such nearperfect classical anticorrelations is challenging in practice. One potential approach could use thermal light that is naturally position correlated and adapting the output measurement by using the minus-coordinate projection of G(2). This measurement will have lower contrast and sensitivity than entangled photons and will face issues with camera cross-talk but could benefit from a higher brightness. Finally, it should also be noted that prior studies (49–53) have explored the use of entangled photon pairs to correct specific types of optical aberrations without using AO. In summary, we have demonstrated that QAO works for bright-field imaging (classical and quantum) and that it can also extend to more complex label-free modalities, such as phase imaging and reflection configurations. Another crucial point is that QAO can be used in all of the quantum versions of these systems (10, 11, 14, 16–18). This could prove very useful because, as shown in the bright-field case in Fig. 5 and in a quantum-enhanced phase scheme shown in fig. S20, such quantum schemes are extremely sensitive to optical aberrations to the point of preventing them from working. QAO thus has the potential to optimize the operation of any imaging system based on photon pairs and could therefore 8 MARCH 2024 • VOL 383 ISSUE 6687

1147

RES EARCH | R E S E A R C H A R T I C L E S

play a major role in the development of future quantum optical microscopes. RE FE RENCES AND N OT ES

1. P.-A. Moreau, E. Toninelli, T. Gregory, M. J. Padgett, Nat. Rev. Phys. 1, 367–380 (2019). 2. A. N. Boto et al., Phys. Rev. Lett. 85, 2733–2736 (2000). 3. E. Toninelli et al., Optica 6, 347 (2019). 4. H. Defienne et al., Nat. Commun. 13, 3566 (2022). 5. Z. He, Y. Zhang, X. Tong, L. Li, L. V. Wang, Nat. Commun. 14, 2441 (2023). 6. G. Brida, M. Genovese, I. Ruo Berchera, Nat. Photonics 4, 227–230 (2010). 7. H. Defienne, M. Reichert, J. W. Fleischer, D. Faccio, Sci. Adv. 5, eaax0307 (2019). 8. T. Gregory, P.-A. Moreau, E. Toninelli, M. J. Padgett, Sci. Adv. 6, eaay2652 (2020). 9. T. Ono, R. Okamoto, S. Takeuchi, Nat. Commun. 4, 2426 (2013). 10. R. Camphausen et al., Sci. Adv. 7, eabj2155 (2021). 11. A. N. Black et al., Optica 10, 952 (2023). 12. H. Defienne, B. Ndagano, A. Lyons, D. Faccio, Nat. Phys. 17, 591–597 (2021). 13. S. Töpfer et al., Sci. Adv. 8, eabl4301 (2022). 14. H. Hodgson, Y. Zhang, D. England, B. Sussman, Appl. Phys. Lett. 122, 034001 (2023). 15. Y. Zhang, A. Orth, D. England, B. Sussman, Phys. Rev. A 105, L011701 (2022). 16. M. B. Nasr, B. E. A. Saleh, A. V. Sergienko, M. C. Teich, Phys. Rev. Lett. 91, 083601 (2003). 17. A. F. Abouraddy, M. B. Nasr, B. E. A. Saleh, A. V. Sergienko, M. C. Teich, Phys. Rev. A 65, 053817 (2002). 18. B. Ndagano et al., Nat. Photonics 16, 384–389 (2022). 19. N. Ji, Nat. Methods 14, 374–380 (2017). 20. Q. Zhang et al., Biomed. Opt. Express 14, 1732–1756 (2023). 21. K. M. Hampson, J. Antonello, R. Lane, M. J. Booth, “Sensorless adaptive optics” (Zenodo, 2020); https://zenodo. org/records/4271425. 22. M. Booth, Opt. Express 14, 1339–1352 (2006). 23. D. Débarre et al., Opt. Lett. 34, 2495–2497 (2009). 24. X. Zhou et al., Biomed. Opt. Express 6, 3577–3595 (2015). 25. D. Debarre, M. J. Booth, T. Wilson, Opt. Express 15, 8176–8190 (2007).

1148

8 MARCH 2024 • VOL 383 ISSUE 6687

26. J. R. Fienup, J. J. Miller, J. Opt. Soc. Am. A Opt. Image Sci. Vis. 20, 609–620 (2003). 27. L. P. Murray, J. C. Dainty, E. Daly, in SPIE Proceedings Volume 5823: Opto-Ireland 2005: Imaging and Vision, F. D. Murtagh, Ed. (SPIE, 2005). 28. Y. Shu et al., PhotoniX 3, 24 (2022). 29. Z. Kam, B. Hanser, M. G. L. Gustafsson, D. A. Agard, J. W. Sedat, Proc. Natl. Acad. Sci. U.S.A. 98, 3790–3795 (2001). 30. Y. Jian et al., Biomed. Opt. Express 5, 547–559 (2014). 31. S. Walborn, C. Monken, S. Pádua, P. Souto Ribeiro, Phys. Rep. 495, 87–139 (2010). 32. A. F. Abouraddy, B. E. A. Saleh, A. V. Sergienko, M. C. Teich, J. Opt. Soc. Am. B 19, 1174 (2002). 33. H. Defienne, M. Reichert, J. W. Fleischer, Phys. Rev. Lett. 120, 203604 (2018). 34. P. Xiao, M. Fink, A. C. Boccara, J. Biomed. Opt. 21, 121505 (2016). 35. J. Mertz, H. Paudel, T. G. Bifano, Appl. Opt. 54, 3498–3506 (2015). 36. R. D. Simmonds, M. J. Booth, J. Opt. 15, 094010 (2013). 37. N. Frenkel et al., ACS Nano 17, 14990–15000 (2023). 38. B. Ndagano et al., NPJ Quantum Inf. 6, 94 (2020). 39. R. Camphausen et al., Opt. Express 31, 6039–6050 (2023). 40. A. Nomerotski et al., J. Instrum. 18, C01023 (2023). 41. V. Vidyapin, Y. Zhang, D. England, B. Sussman, Sci. Rep. 13, 1009 (2023). 42. B. Courme et al., Opt. Lett. 48, 3439–3442 (2023). 43. Q. Hu et al., APL Photonics 5, 100801 (2020). 44. T. Yeminy, O. Katz, Sci. Adv. 7, eabf5364 (2021). 45. I. M. Vellekoop, A. P. Mosk, Opt. Lett. 32, 2309–2311 (2007). 46. H. Defienne, M. Reichert, J. W. Fleischer, Phys. Rev. Lett. 121, 233601 (2018). 47. O. Lib, G. Hasson, Y. Bromberg, Sci. Adv. 6, eabb6298 (2020). 48. B. Courme, P. Cameron, D. Faccio, S. Gigan, H. Defienne, PRX Quantum 4, 010308 (2023). 49. C. Bonato, A. V. Sergienko, B. E. Saleh, S. Bonora, P. Villoresi, Phys. Rev. Lett. 101, 233603 (2008). 50. L. A. Filpi, M. V. da Cunha Pereira, C. H. Monken, Opt. Express 23, 3841–3850 (2015).

51. D. Simon, A. Sergienko, Phys. Rev. A 80, 053813 (2009). 52. A. N. Black et al., Phys. Rev. Lett. 123, 143603 (2019). 53. D. Simon, A. Sergienko, J. Opt. Soc. Am. B 28, 247 (2011). 54. P. Cameron, B. Courme, D. Faccio, H. Defienne, Data for: Adaptive optical imaging with entangled photons, Dryad (2024). AC KNOWLED GME NTS

H.D. and P.C. thank K. Kassem, T. Chaigne, and A. Badon for experimental help and for discussions. Funding: This work was supported by the Royal Academy of Engineering Chairs in Emerging Technologies Scheme and funding from the United Kingdom Engineering and Physical Sciences Research Council (grant nos. EP/M01326X/1, EP/R030081/1, and EP/Y029097/1 to D.F.); the European Union Horizon 2020 Research and Innovation Program (grant 801060 to D.F.); and the European Research Council (starting grant SQIMIC-101039375 to H.D.). P.C. and H.D. were supported by the SPIE Early Career Researcher Accelerator fund in Quantum Photonics. R.P. received funding from Clare College, Cambridge, through a junior research fellowship. Author contributions: P.C. analyzed the results and performed the experiments. P.C., B.C., C.V., and R.P. designed the experiments. P.C. and H.D. conceived the original idea. H.D. and D.F. supervised the project. All authors discussed the data and contributed to the manuscript. Competing interests: The authors declare no competing interests. Data and material availability: All data are available in the main manuscript or the supplementary materials or have been deposited at Dryad (54). License information: Copyright © 2024 the authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original US government works. https://www.science.org/ about/science-licenses-journal-article-reuse SUPPLEMENTARY MATERIALS

science.org/doi/10.1126/science.adk7825 Materials and Methods Supplementary Text Figs. S1 to S20 References (55–63) Submitted 13 September 2023; accepted 23 January 2024 10.1126/science.adk7825

science.org SCIENCE

PROFESSOR OF BIOMEDICAL DATA SCIENCE (ASSISTANT, ASSOCIATE, AND/OR PROFESSOR LEVEL)

For your career in science, there’s only one

Features in myIDP include:



Exercises to help you examine your skills, interests, and values.



A list of 20 scientific career paths with a prediction of which ones best fit your skills and interests.

Visit the website and start planning today! myIDP.sciencecareers.org

In partnership with:

The Center for Biomedical Data Science (CBDS) at the Knight Cancer Institute (KCI) at Oregon Health and Science University (OHSU) is searching for multiple tenured or tenure-track faculty positions at all ranks (Assistant Professor, Associate Professor, and/or Professor) in the area of Biomedical Data Science. Faculty will be members of CBDS with primary tenure home in the Division of Oncological Sciences (DOS) at KCI. A secondary appointment in a different unit at OHSU is possible depending on the candidate’s discipline and research focus. The newly founded CBDS, as a nexus for research in data science at OHSU, is dedicated to advancing precision health in general, and precision early cancer detection and therapies in particular, by analyzing large, complex, multi-modal multi-scale data through the development and implementation of novel computational analytical tools and methods. The data science goals include identifying biomarkers associated with response and resistance to therapy; modeling etiologic causes of cancer; modeling the development and progression of cancer; organizing and processing large data resources to create harmonized datasets; and mining datasets for novel subtypes of diseases. The field analyzes all kinds of cancer data, including omics, anatomical imaging to molecular imaging to nanoscale imaging, cellular and pathway interactions, electronic medical records, and published literature. A key area of focus is integrating datasets, such as combining public and private datasets or merging data across modalities, and doing novel analyses on integrated datasets. To serve these goals, CBDS seeks interdisciplinary scholars whose work spans diverse areas, including but not limited to: machine learning, computer vision, natural language processing, AI-enabled health, computational biology, systems biology, bioinformatics, biostatistics, dynamic modeling, responsible and interpretable AI, multi-scale multi-modal data integration, and data analytics and visualization.

Science Virtual Career Fair: Accelerating Your Career Path March 20, 2024 ScienceCareers.org ScienceCareers org

Join Science Careers for a chance to meet with top scientific organizations and gain insightful career advice — all at once. Benefits of Attending • Connect with employers virtually one-on-one. • Access four career exploration videos from an industry expert in short, digestible segments. • Login from anywhere — no travel required. • Manage your experience with queues that tell you how many people are in line ahead of you.

Scan code for more details and to register or visit bit.ly/science-vcf

ScienceCareers.org

myIDP: A career plan customized for you, by you.

WORKING LIFE By Violeta J. Rodriguez

Writing my ticket

“H

ow did you publish so much as an immigrant?” Realizing I wasn’t a native English speaker, people asked variations of this question countless times as I interviewed for faculty positions last year. I knew that disclosing my struggles was unlikely to land me a job, so I would laugh and say, “I enjoy writing!” But the truth is writing did not come easily. I arrived in the United States at age 9 knowing only Spanish. As an undergrad, I spent months laboring over my senior thesis. I finally thought it was good enough—only to get it back from my mentor covered in bright red changes, showing how far I had yet to go. But I knew I needed to develop my publication record to stand out in a world where backgrounds like mine are scarce.

From early on I wanted to pursue an academic career. But my first applications for Ph.D. programs were rejected. After learning I would be a stronger applicant if I had some publications, I decided to take an industry job to pay the bills while also pursuing a postbaccalaureate project in my undergraduate mentor’s lab, hoping for publishable findings. When I had results to write up, I agonized over finding the right words and telling a coherent, wellstructured research story. I was gratified that my mentor’s edits were less extensive than they had been on my thesis, highlighting my growth. Still, the paper was rejected from three journals. At the same time, I was getting rejected by more Ph.D. programs. Maybe an immigrant like me, who had started in a community college and graduated thanks to a Pell Grant, did not belong in academia. But I was determined to change that. I tried to focus on what I was learning: Revisions and rejections are a normal part of the scientific process, and what mattered was how I adapted and grew from them. Realizing I might need an alternate path to a Ph.D., I decided to pivot to a master’s degree in quantitative methods to expand my technical expertise while also working as a research assistant at an academic medical center. This combination opened the door to analyzing large data sets for several federally funded global health studies—and opportunities to publish. When it came time to write up the results, I sat beside my supervisors as they revised my work, absorbing every revision and comment. I began to see writing not just as a task, but as an art form that I could slowly master. Learning how they approached writing helped me find my voice. I went on to publish several first-author

manuscripts and grew more confident. Yet, I knew my learning journey was far from over. After completing my master’s I finally dared to apply to Ph.D. programs again, 4 years out of college, and was thrilled to get in. I even managed to secure competitive national fellowships. I’m sure being able to list multiple publications on my CV played a role in both successes; it helped me stand out from other candidates who may have gone to prestigious universities and had more “traditional” backgrounds but had less tangible research output to their names. I went on to spend my Ph.D. refining my research and prioritizing writing, including beginning a weekly writing group with other trainees, saying “yes” to collaborations, leveraging my quantitative training to contribute to various projects, and regularly applying for fellowships and other funding opportunities. As my Ph.D. program neared its end, my publication record and funding success distinguished me in a competitive academic job market and boosted my confidence during faculty interviews. Writing had become a testament to my journey from an uncertain immigrant feeling like an outsider in academia to a confident researcher whose background is a source of pride and strength. Now, as I embark on a new career journey in pursuit of tenure, I see the challenges and rejections from my early days not as setbacks, but as vital opportunities to grow and improve as a writer— opportunities to embrace rather than dread. j

1150

8 MARCH 2024 • VOL 383 ISSUE 6687

Violeta J. Rodriguez is an assistant professor at the University of Illinois Urbana-Champaign. Send your career story to [email protected].

science.org SCIENCE

ILLUSTRATION: ROBERT NEUBECKER

“Revisions and rejections are a normal part of the scientific process.”

Hear the stories behind the latest news and research from Science.

LISTEN & SUBSCRIBE NEW EPISODES EVERY THURSDAY

YOUR RESEARCH HAS ITS REWARDS.

Submit an essay by 3/15/24 to enter to win $25K. The prestigious international Science & PINS Prize is awarded annually to outstanding early-career researchers in the emerging field of neuromodulation. Findings already show improved outcomes in several neurological disorders. To enter, just write a 1000-word essay about your research performed in the last three years. The Grand Prize winner will be presented a US$25,000 award, as well as have their essay published in Science magazine and on Science Online. A runner-up will have their essay published online. For additional inquiries, email [email protected].

SCIENCE.ORG/PINS