Volume 144 Number 5 March 4, 2011
www.cell.com
Hallmarks of Cancer
INTRODUCING OUR NEW
20TH ANNIVERSARY
2011 RESEARCH PRODUCT CATALOG FROM THE LEADING SUPPLIER OF RESEARCH PRODUCTS:
SANTA CR U Z B IO TECHN OL OGY, Y I NC.
䡲
over 16,720 monoclonal and 34,860 polyclonal antibodies
䡲
over 135,300 siRNA and shRNA (plasmid and lentivirus) products
䡲
UltraCruz ™, ExactaCruz ™, and CrystalCruz ™ brand labware
䡲
over 141,000 ChemCruz ™ specialty biochemicals
䡲
broad range of support products and secondary antibodies
䡲
website updated daily with new antibodies, biochemicals, labware, product citations and support data
C ELEBR ATI NG 20 YEA R S 19 91 – 2 011
anti an t bo ti bodi dies es s ExactaCruz™
lab la bwar ware arre
ChemCruz™
siRN si RNA/ A/shRN NA
CrystalCruz™
chem ch emic ical als s
UltraCruz™
ImmunoCruz™
WWW. SC B T .COM
HEADQUARTERS
EUROPEAN SUPPORT
ASIAN SUPPORT
Santa Cruz Biotechnology, Inc. 2145 Delaware Avenue Santa Cruz, California 95060
Santa Cruz Biotechnology, Inc. Bergheimer Str. 89-2 69115 Heidelberg, Germany
Santa Cruz Biotechnology (Shanghai) Co., Ltd. Building No. 2, Lane 315, No. 1-6, Jianye Road Pudong New District, Shanghai, 201201
TOLL FREE: 800.457.3801 or
TOLL FREE: +00800.4573.8000 or
N. CHINA TOLL FREE: 10.800.711.0752
PHONE: 831.457.3800
PHONE: +49.6221.4503.0
S. CHINA TOLL FREE: 10.800.110.0694 JAPAN TOLL FREE: (010) 800.40402026 S. KOREA TOLL FREE: 00798.1.1.002.0297
© 2011 Santa Cruz Biotechnology, Inc., the Santa Cruz Biotechnology, Inc. logo, UltraCruz™, ExactaCruz ™, ChemCruz ™, ImmunoCruz ™ and CrystalCruz ™ are registered trademarks of Santa Cruz Biotechnology, Inc.
! W E N
Intelligent design
inCuSaFe™ copper enriched stainless steel interior
+
+
+
+
Single-beam, dual capture infrared CO2 sensor
+
Zirconia O2 Sensor for multigas model
SafeCell UV protection in situ
Hydrogen peroxide vapor sterilization in situ
1
Integrated Cooling Coil (optional)
The industry’s first in situ H2O2 sterilization with the fastest turnaround. For maximum productivity in clinical, general purpose or the most highly compliant GMP applications, the new SANYO Sterisonic™ GxP CO2 incubator offers an impressive return on investment. With multiple contamination control safeguards, exclusive on-board H2O2 sterilization, FDA-21CFR data capture and graphical LCD display, the Sterisonic™ GxP rewards good laboratory technique with performance you can trust. Learn more, visit www.sterisonic.com or call 800-858-8442. pictured above: Sterisonic™ GxP MCO-19AICUVH with rapid H2O2 vapor sterilization system.
FREEE R OFF [
!
E S O N L IN D E TA IL
]
FREE! FREE!
H2O2 sterilization system accessories ($571 Value), plus BD labware consumables ($250 Value). Limited time upgrade offer with purchase of the Sterisonic™ GxP MCO-19AIC(UVH). BD stem cell starter kit with Sterisonic™ GxP quote. No purchase necessary! Act now. Supplies are limited. ($150 Value)
www.sterisonic.com
©2011 Sanyo Biomedical
Editor Emilie Marcus Senior Deputy Editor Elena Porro Deputy Editor Robert Kruger Scientific Editors Karen Carniol Kara Cerveny Michaeleen Doucleff Fabiola Rivas Niki Scaplehorn Lara Szewczak Senior Managing Editor Meredith Adinolfi Deputy Managing Editor Andy Smith Lead Illustrator Andrew A. Tang Illustrators Yvonne Blanco Kate Mahan Production Staff Reyna Clancy Editorial Assistant Mary Beth O’Leary
Editorial Board Abul Abbas C. David Allis Genevie`ve Almouzni Uri Alon Angelika Amon Johan Auwerx Richard Axel Cori Bargmann Konrad Basler Bonnie Bassler David Baulcombe Jeffrey Benovic Carolyn Bertozzi Wendy Bickmore Elizabeth Blackburn Joan Brugge Lewis Cantley Joanne Chory David Clapham Andrew Clark Hans Clevers Stephen Cohen Pascale Cossart Jeff Dangl Ted Dawson Pier Paolo di Fiore Marileen Dogterom Julian Downward Bruce Edgar Steve Elledge Anne Ephrussi Ronald Evans Witold Filipowicz Marco Foiani Elaine Fuchs Yukiko Goda Stephen Goff Joe Goldstein
Douglas Green Leonard Guarente Taekjip Ha Daniel Haber Ulrike Heberlein Nobutaka Hirokawa Mark Hochstrasser Arthur Horwich Tony Hunter James Hurley Richard Hynes Thomas Jessell Narry Kim Mary-Claire King David Kingsley Frank Kirchhoff Richard Kolodner John Kuriyan Robert Lamb Mark Lemmon Beth Levine Wendell Lim Jennifer Lippincott-Schwartz Dan Littman Richard Losick Scott Lowe Tom Maniatis Matthias Mann Kelsey Martin Joan Massague´ Iain Mattaj Satyajit Mayor Ruslan Medzhitov Craig Mello Tom Misteli Tim Mitchison Alex Mogilner Paul Nurse Roy Parker
Dana Pe’er Kathrin Plath Carol Prives Klaus Rajewsky Venki Ramakrishnan Rama Ranganathan Anne Ridley James Roberts Alexander Rudensky Helen Saibil Joshua Sanes Randy Schekman Ueli Schibler Joseph Schlessinger Hans Scho¨ler Trina Schroer Geraldine Seydoux Kevan Shokat Pamela Sklar Nahum Sonenberg James Spudich Paul Sternberg Bruce Stillman Azim Surani Keiji Tanaka Craig Thompson Robert Tjian Ju¨rg Tschopp Ulrich von Andrian Gerhard Wagner Detlef Weigel Alan Weiner Jonathan Weissman Matthew Welch Tian Xu Shinya Yamanaka Marino Zerial Xiaowei Zhuang Huda Zoghbi
Cell Office Cell, Cell Press, 600 Technology Square, 5th Floor, Cambridge, Massachusetts 02139 Phone: (+1) 617 661 7057, Fax: (+1) 617 661 7061, E-mail:
[email protected] Online Publication: http://www.cell.com Cell (ISSN 0092-8674) is published biweekly by Cell Press, 600 Technology Square, 5th Floor, Cambridge, Massachusetts 02139. The institutional subscription rate for 2011 is $1,605 (US and Canada) or $1,847 (elsewhere). The individual subscription rate is $320 (US and Canada) or $363 (elsewhere). The individual copy price is $50. Periodicals postage paid at Boston, Massachusetts and additional mailing offices. Postmaster: send address changes to Elsevier Customer Service Americas, Cell Press Journals, 11830 Westline Industrial Drive, St. Louis, MO 63146, USA. The paper used in this publication meets the requirments of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed by Dartmouth Printing Company, Hanover, NH.
BE THE FIRST
to read the latest issue of any Cell Press journal.
Register for Cell Press Email Alerts and get the complete table of contents as soon as the issue publishes online — FREE! Cell Press Email Alerts deliver the news, research, and commentaries featured in each journal’s latest issue, including the full title of every article, direct links to the articles, and the complete author list. Plus, to save you time, each research article has a brief summary highlighting its significant findings. You don’t have to be a subscriber to sign up for Cell Press Email Alerts. While subscribers have instant access to the full text of all articles listed in the Email Alerts, non-subscribers can read the abstracts of all articles as well as the full text of the issue’s Featured Article.
www.cellpress.com
Cell Press President & CEO Lynne Herndon Editor in Chief, Vice President of Content Development Emilie Marcus Vice President of Business Development Joanne Tracy Vice President of Web Development and Operations Keith Wollman Senior Product Manager Mark Van Hussen Director of Marketing Jonathan Atkinson Production Manager Meredith Adinolfi
Display Advertising Northeast/Mid-Atlantic: Victoria Macomber, ph: 508 928 1255; fax: 508 928 1256; e-mail:
[email protected] Midwest/Southeast/Eastern Canada: Inez Herrero-Redman, ph: 585 678 4395; fax: 585 678 4722; e-mail: i.herrero@elsevier. com Northwest/Southwest/Western Canada: Lori Young, ph: 646 370 6312; fax: 212 462 1915; e-mail:
[email protected] California: Elizabeth Loennborn, ph: 714 655 1877; fax: 214 452 9627; e-mail:
[email protected] UK/Europe: James Kenney, ph: +44 20 7424 4216; fax: +44 18 6585 3136; e-mail:
[email protected] Asia: Wendy Xie, ph: +86 10 8520 8827; e-mail: w.xie@ elsevier.com Classified Advertising United States and Canada: Gordon Sheffield, Key Account Manager, ph: 617 386 2189; fax: 617 397 2805; e-mail: g.sheffi
[email protected]
Press Officer Cathleen Genova
UK, Europe, and Asia: Sabrina Dodge, Key Account Manager, ph: +44 20 7424 4997; fax: +44 18 6585 3136; e-mail:
[email protected]
ª2011 Elsevier Inc. All rights reserved. This journal and the individual contributions contained in it are protected under copyright by Elsevier Inc., and the following terms and conditions apply to their use:
advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. Although all advertising material is expected to conform to ethical (medical) standards, inclusion in this publication does not constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer.
Photocopying: Single photocopies of single articles may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee are required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for nonprofit educational classroom use. For information on how to seek permission, visit www.elsevier. com/permissions or call (+44) 1865 843830 (UK) / (+1) 215 239 3804 (US). Permissions: For information on how to seek permission, visit www.elsevier.com/ permissions or call (+44) 1865 843830 (UK) / (+1) 215 239 3804 (US). Derivative Works: Subscribers may reproduce tables of contents or prepare lists of articles including summaries for internal circulation within their institutions. Permission of the Publisher is required for resale or distribution outside the institution. Permission of the Publisher is required for all other derivative works, including compilations and translations (please consult www.elsevier.com/permissions). Electronic Storage or Usage: Permission of the Publisher is required to store or use electronically any material contained in this journal, including any article or part of an article (please consult www.elsevier.com/permissions). Except as outlined above, no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the Publisher. Notice: No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Because of rapid
Reprints: Article reprints are available through Cell’s reprint service; for information, contact Nicholas Pavlow (e-mail:
[email protected]; ph: (+1) 212 633 3960). Subscription Orders and Inquiries: Mail, fax, or e-mail address changes to Elsevier Customer Service Americas, allowing 4–6 weeks for processing. Lost or damaged issues will be replaced, subject to availability, if Cell Press is notified within the claim period (US and airmail delivery: 3 months from issue date; surface delivery: 4 months from issue date). Periodical delivery in the US can take up to 3 weeks. Airmail delivery can take 2–4 weeks. The price of a single copy of Cell is $50 (excluding special issues). All orders must be prepaid and in writing. Please include the volume and issue number, payment (check or credit card, MasterCard, Visa, or American Express only), and a delivery address. Allow 4–6 weeks for delivery. Mailing address: Elsevier Customer Service Americas, Cell Press Journals, 11830 Westline Industrial Drive, St. Louis, MO 63146, USA. Toll-free phone within USA/Canada: 866 314 2355; phone for outside US/Canada: (+1) 314 453 7038; fax: (+1) 314 523 5170; e-mail:
[email protected]; internet: www.cellpress.com or <www.cell.com>. Funding Body Agreements and Policies: Elsevier has established agreements and developed policies to allow authors whose articles appear in journals published by Elsevier to comply with potential manuscript archiving requirements as specified as conditions of their grant awards. To learn more about existing agreements and policies, visit http://www.cell.com/cellpress/FundingBodyAgreements. Guide for Authors: For a full and complete guide for authors, please go to www.cell.com/ authors.
THE ONE YOU’VE BEEN WAITING FOR
AT-rich kb
OneTaq DNA Polymerase ™
The ONE polymerase for your endpoint PCR needs s Robust yield with minimal optimization
M
29
37
Standard 47
55
GC-rich 65
66
High GC 73
79
%GC
10.0 8.0 6.0 5.0 4.0 3.0 2.0 1.5 1.0
s Ideal for routine, AT- or GC-rich templates s Hot start and master mix versions available
0.5
Standard Reaction Buffer
Request a sample at
www.neb.com/OneTaq
GC Reaction Buffer Plus High GC Enhancer
Amplification of a selection of sequences with varying AT and GC content from human and C. elegans genomic DNA, using OneTaq DNA Polymerase.
Leading Edge Cell Volume 144 Number 5, March 4, 2011 IN THIS ISSUE SELECT 631
Cell Death and Destruction
ANALYSIS 635
Exome Sequencing Deciphers Rare Diseases
A. Maxmen
PREVIEWS 638
Uncovering a Tumor Suppressor for Triple-Negative Breast Cancers
J.G. Albeck and J.S. Brugge
640
The Ribosome and TORC2: Collaborators for Cell Growth
X. Xie and K.-L. Guan
642
The Skeleton: The New Controller of Male Fertility?
L.B. Smith and P.T.K. Saunders
644
Astrocytes: Powering Memory
P. Bezzi and A. Volterra
REVIEW 646
Hallmarks of Cancer: The Next Generation
D. Hanahan and R.A. Weinberg
SNAPSHOT 826
Neuromuscular Junction
S.J. Burden
Antibodies and Related Reagents for Signal Transduction Research
™
XP
Monoclonal Antibodies, eXceptional Performance
™
Unparalleled product quality, validation, and technical support.
XP monoclonal antibodies are generated using XMT™ technology, a proprietary monoclonal method developed at Cell Signaling Technology. This technology provides access to a broad range of antibody-producing B cells unattainable with traditional monoclonal technologies, allowing more comprehensive screening and the identification of XP monoclonal antibodies.
eXceptional specificity As with all of our antibodies, the antibody is specific to your target of interest, saving you valuable time and resources.
+ eXceptional sensitivity The antibody will provide a stronger signal for your target protein in cells and tissues, allowing you to monitor expression of low levels of endogenous proteins, saving you valuable materials.
+ eXceptional stability and reproducibility XMT technology combined with our stringent quality control ensures maximum lot-to-lot consistency and the most reproducible results.
= eXceptional Performance™ XMT Technology coupled with our extensive antibody validation and stringent quality control delivers XP monoclonal antibodies with eXceptional Performance in the widest range of applications. Above: Confocal IF analysis of rat cerebellum using β3-Tubulin (D71G9) XP™ Rabbit mAb #5568 (green) and Neurofilament-L (DA2) Mouse mAb #2835 (red). Blue pseudocolor = DRAQ5® #4084 (fluorescent DNA dye).
For additional information and a complete list of available XP™ Monoclonal Antibodies visit…
www.cellsignal.com Orders (toll-free) 1-877-616-2355
| Technical support (toll-free) 1-877-678-8324
[email protected] | Inquiries
[email protected] | Environmental Commitment eco.cellsignal.com
© 2011 Cell Signaling Technology, Inc. XMT™, XP™ , eXceptional Performance™, CST™, and Cell Signaling Technology® are trademarks of Cell Signaling Technology, Inc. / DRAQ5® is a registered trademark of Biostatus Limited
XP™ monoclonal antibodies are a line of high quality rabbit monoclonal antibodies exclusively available from Cell Signaling Technology. Any product labeled with XP has been carefully selected based on superior performance in all approved applications.
Articles Cell Volume 144 Number 5, March 4, 2011 675
TLR Signaling Is Required for Salmonella typhimurium Virulence
N. Arpaia, J. Godec, L. Lau, K.E. Sivick, L.M. McLaughlin, M.B. Jones, T. Dracheva, S.N. Peterson, D.M. Monack, and G.M. Barton
689
PARIS (ZNF746) Repression of PGC-1a Contributes to Neurodegeneration in Parkinson’s Disease
J.-H. Shin, H.S. Ko, H. Kang, Y. Lee, Y.-I. Lee, O. Pletinkova, J.C. Troconso, V.L. Dawson, and T.M. Dawson
703
Activation of Multiple Proto-oncogenic Tyrosine Kinases in Breast Cancer via Loss of the PTPN12 Phosphatase
T. Sun, N. Aceto, K.L. Meerbrey, J.D. Kessler, C. Zhou, I. Migliaccio, D.X. Nguyen, N.N. Pavlova, M. Botero, J. Huang, R.J. Bernardi, E. Schmitt, G. Hu, M.Z. Li, N. Dephoure, S.P. Gygi, M. Rao, C.J. Creighton, S.G. Hilsenbeck, C.A. Shaw, D. Muzny, R.A. Gibbs, D.A. Wheeler, C.K. Osborne, R. Schiff, M. Bentires-Alj, S.J. Elledge, and T.F. Westbrook
719
A Hierarchical Combination of Factors Shapes the Genome-wide Topography of Yeast Meiotic Recombination Initiation
J. Pan, M. Sasaki, R. Kniewel, H. Murakami, H.G. Blitzblau, S.E. Tischfield, X. Zhu, M.J. Neale, M. Jasin, N.D. Socci, A. Hochwagen, and S. Keeney
732
Double-Strand Breaks in Heterochromatin Move Outside of a Dynamic HP1a Domain to Complete Recombinational Repair
I. Chiolo, A. Minoda, S.U. Colmenares, A. Polyzos, S.V. Costes, and G.H. Karpen
745
Mpk1 MAPK Association with the Paf1 Complex Blocks Sen1-Mediated Premature Transcription Termination
K.-Y. Kim and D.E. Levin
757
Activation of mTORC2 by Association with the Ribosome
V. Zinzalla, D. Stracka, W. Oppliger, and M.N. Hall
769
The Mechanism of Linkage-Specific Ubiquitin Chain Elongation by a Single-Subunit E2
K.E. Wickliffe, S. Lorenz, D.E. Wemmer, J. Kuriyan, and M. Rape
782
Yap1 Acts Downstream of a-Catenin to Control Epidermal Proliferation
K. Schlegelmilch, M. Mohseni, O. Kirak, J. Pruszak, J.R. Rodriguez, D. Zhou, B.T. Kreger, V. Vasioukhin, J. Avruch, T.R. Brummelkamp, and F.D. Camargo
796
Endocrine Regulation of Male Fertility by the Skeleton
F. Oury, G. Sumara, O. Sumara, M. Ferron, H. Chang, C.E. Smith, L. Hermo, S. Suarez, B.L. Roth, P. Ducy, and G. Karsenty
810
Astrocyte-Neuron Lactate Transport Is Required for Long-Term Memory Formation
A. Suzuki, S.A. Stern, O. Bozdagi, G.W. Huntley, R.H. Walker, P.J. Magistretti, and C.M. Alberini
(continued)
'O 8TREME
5SE THESE NEW POWERFUL TRANSFECTION REAGENTS AND OBTAIN AN IDEAL BALANCE BETWEEN HIGH EFlCIENCY AND LOW CYTOTOXICITY
4RANSFECTION %FFICIENCY
4OXICITY
,$( ACTIVITY
,UCIFERASE ACTIVITY 2,5 X
2OCHES .EW 8 TREME'%.% 4RANSFECTION 2EAGENTS
%FFICIENTLY TRANSFECT A BROAD SPECTRUM OF CELLS INCLUDING DIFFICULT TO TRANSFECT CELL TYPES 'ENERATE PHYSIOLOGICALLY RELEVANT RESULTS USING THESE LOW CYTOTOXIC REAGENTS
&OR MORE INFORMATION VISIT WWWX TREMEGENEROCHECOM
8 TREME'%.%
#OMPETITOR ,4+
8 TREME'%.% (0
#OMPETITOR ,480
4HE 8 TREME'%.% AND (0 $.! 4RANSFECTION 2EAGENTS SHOW AN EXCELLENT BALANCE OF HIGH TRANSFECTION EFFICIENCY AND LOW CYTOTOXICITY IN (E,A CELLS 4RANSFECTION EFFICIENCY WAS MEASURED IN RELATIVE LUCIFERASE LIGHT UNITS AND CYTOTOXICITY USING LACTATE DEHYDRO GENASE ACTIVITY
&OR LIFE SCIENCE RESEARCH ONLY .OT FOR USE IN DIAGNOSTIC PROCEDURES 8 42%-%'%.% IS A TRADEMARK OF 2OCHE ¹ 2OCHE $IAGNOSTICS !LL RIGHTS RESERVED
2OCHE $IAGNOSTICS 'MB( 3ANDHOFER 3TRAE -ANNHEIM 'ERMANY
CORRECTION 824
Rallying the Exocyst as an Autophagy Scaffold
J.-C. Farre and S. Subramani
ERRATA 825
Functional and Mechanistic Diversity of Distal Transcription Enhancers
M. Bulger and M. Groudine
825
Role for Dpy-30 in ES Cell-Fate Specification by Regulation of H3K4 Methylation within Bivalent Domains
H. Jiang, A. Shukla, X. Wang, W.-y. Chen, B.E. Bernstein, and R.G. Roeder
POSITIONS AVAILABLE
On the cover: In their classic Review, Douglas Hanahan and Robert A. Weinberg formulated the six hallmarks of cancer: self-sufficiency in growth signals, insensitivity to antigrowth signals, tissue invasion and metastasis, limitless replicative potential, sustained blood vessel growth, and resistance to cell death (Hallmarks of Cancer, Cell 100, 57–70, January 7, 2000). In this issue, Hanahan and Weinberg (pp. 646–674) revisit these original hallmarks in light of the past decade of progress and add four new hallmarks of cancer: genomic instability, inflammation, evasion of the immune system, and reprogrammed energy metabolism. The icons on the cover depict the ten hallmarks of cancer.
017.A1.0115.A © 2010 Eppendorf AG
l
Joystick provides intuitive control
l
Patented axial injection movement of the capillary
l
Semi-Automatic microinjection into adherent cells
l
Pre-pull capillaries for reproducible injection of adherent cells
microinjection
search level
A B C
injection level
carrier
InjectMan NI 2
Microinjection simplified! Microinjection is one of the core methods to introduce foreign DNA and other non-permeable molecules into cells. Nuclear injection of plasmid DNA enables rapid expression of proteins in specific cells within a population. The menu-controlled, programmable micromanipulator InjectMan NI 2 is ideally suited for microinjection of adherent cells. Connection with the FemtoJet and the axial mounting allows injections at 45˚ angle reducing cell damage during injection and increases cell viability. This guarantees a very rapid, safe and reproducible microinjection process.
Eppendorf InjectMan NI 2 microinjector has it all: Motorized X-Y-Z movements provide precise movement l Pre-setting and storage of up to 2 locations in X-Y-Z, saves time in returning to pre-set work locations l Automated Home function for rapid capillary exchange l Joystick-controlled provides overall ergonomic manipulator l Fine adjustment of work speed made easy with positioning wheel l Can be adapted to all common microscopes l
For more information visit www.eppendorf.com
www.eppendorf.com Þ Email:
[email protected] In the U.S.: Eppendorf North America, Inc. 800-645-3050 Þ In Canada: Eppendorf Canada Ltd. 800-263-8715
Leading Edge
In This Issue TLRs Signal Safety for Salmonella PAGE 675
Toll-like receptors (TLRs) are critical for detecting pathogens and then triggering an immune response. Arpaia et al. now find that TLRs are surprisingly required for the survival and replication of Salmonella typhimurium in mice. They show that, in the absence of TLRs, bacteria fail to induce virulence genes or to build replicative compartments within the cell. This study suggests that immune signals serve as intracellular contextual cues for certain pathogens, ensuring that their virulence genes are expressed at the right time and place.
Defining Meiotic Breakpoints PAGE 719
Meiotic recombination relies on DNA double-strand breaks (DSBs) that impact inheritance and genome evolution. Pan et al. now provide a global map of meiotic recombination events with single-nucleotide resolution. This map provides a comprehensive picture of where and why meiotic recombination initiates in yeast, including how chromosomal features shape the recombination landscape. Surprisingly, many recombination events occur outside meiotic hot spots, suggesting that Spo11, the enzyme responsible for generating DSBs, is an opportunistic cutter.
Parkin, PARIS, and PGC-1a PAGE 689
Parkinson ’s disease is often caused by the inactivation of the E3 ubiquitin ligase parkin. Now Shin et al. uncover a new substrate for parkin, named PARIS, which provides a molecular explanation for how parkin inactivation leads to neurodegeneration. PARIS is a transcriptional repressor that accumulates in the brains of Parkinson’s disease patients and is required for degeneration of dopamine neurons in mouse models. PARIS blocks expression of PGC-1a and NRF-1, and manipulation of this pathway protects dopamine neurons from degeneration in Parkinson’s disease.
New Target for Refractory Breast Cancer PAGE 703
‘‘Triple-negative’’ breast cancers are common malignancies with poor prognosis for which no targeted therapeutics exists. Using a genetic screen, Sun et al. now identify the tyrosine phosphatase PTPN12 as a tumor suppressor in triple-negative breast cancers. PTPN12 activity is frequently lost in human tumors, and inhibiting two kinase pathways regulated by PTPN12 (i.e., EGFR/HER2 and PDGFR-b) impairs tumorigenic and metastatic potential of PTPN12-deficient triple breast cancer cells. These findings provide therapeutic inroads to triple-negative breast cancers and reveal a mechanism for activation of proto-oncogenic tyrosine kinases that could be relevant to other cancers.
Stalling Repair until Breaks Move to Safety PAGE 732
Double-strand breaks in heterochromatin pose a serious threat to genome integrity because their repetitive sequences have increased potential for erroneous exchanges during recombinational repair. Chiolo et al. now discover how heterochromatin overcomes this challenge: heterochromatin proteins postpone strand invasion and final stages of repair until the damaged DNA is moved outside the heterochromatin domain. This spatial and temporal separation of the repair steps prevents aberrant recombination and represents a new mechanism for safeguarding genome integrity.
Elongation Runs on Stress PAGE 745
Transcriptional attenuation is a regulatory mechanism that prematurely terminates elongation. Now Kim and Levin explain how the Mpk1 MAP kinase in yeast alleviates attenuation to activate the expression of stress-induced genes. Mpk1 blocks recruitment of a termination complex by directly interacting with the elongation factor Paf1C. Moreover, the human homologs ERK5 MAPK and Paf1 can substitute for the yeast factors during this process, suggesting that this MAPK-driven mechanism of transcriptional attenuation is conserved. Cell 144, March 4, 2011 ª2011 Elsevier Inc. 627
3TRUGGLINGåTOåKEEPåUPåWITHåå THEåLATESTåLIFEåSCIENCEåNEWS
#ELLå$AILY¬.EWS¬!GGREGATORåHASåTHEåSOLUTION 3UBSCRIBEåTOå&2%%å$AILYå.EWSå!LERTSåATåNEWSCELLCOMåANDå GETåTHEåLATESTåLIFEåSCIENCEåHEADLINESåDELIVEREDåTOåYOURåINåBOX
NEWSCELLCOM
mTORC2 Activation by Ribosome Association PAGE 757
The mammalian target of rapamycin complex 2 (mTORC2) is a highly conserved kinase deregulated in cancer and diabetes, but what activates mTORC2 is unknown. Zinzalla et al. now demonstrate that mTORC2 activates upon association with the ribosome. Insulin stimulates this interaction through PI3K signaling, and the mTORC2-ribosome complex promotes Akt signaling in cancer cells. Together, these results suggest that activation by ribosome association is a conserved mechanism for mTORC2 in cancer and diabetes.
Ubiquitin Only Has Eyes for Lysine 11, Thanks to E2 PAGE 769
Ubiquitin chains linked together at different lysine residues trigger distinct functional consequences in the cell. Wickliffe et al. now determine how the monomeric E2 enzyme Ube2S specifically generates lysine 11-linked ubiquitin chains. A combination of NMR and biochemistry reveals how the Ube2B orients the donor ubiquitin such that the complex transiently recognizes the surface around lysine 11 on the acceptor ubiquitin, but not around ubiquitin’s six other lysines. This fleeting interaction creates a competent active site, indicating that that specific chain topology is achieved through substrate-assisted catalysis.
Yap-ing about Crowd Control for Stem Cells PAGE 782
Proliferation of tissue-specific stem cells is tightly controlled to produce organs of a predetermined size. Schlegelmilch et al. now demonstrate that Yap1, the transcriptional effector of the Hippo-signaling pathway, is a critical modulator of epidermal stem cell proliferation. They find that a-catenin, which has been implicated in sensing cell density, negatively regulates Yap1’s nuclear translocation by modulating its interactions with 14-3-3 and the PP2A phosphatase. These findings provide insights into the molecular circuitry mediating ‘‘crowd control’’ in mammalian skin.
Boning Up on Male Fertility PAGE 796
Not just a simple scaffold, bone is now recognized as an endocrine organ regulating energy homeostasis. Here Oury et al. show that bone also plays a surprising role in regulating male fertility. They find that bone osteoblasts secrete the hormone osteocalcin, which stimulates testosterone production in testicular Leydig cells by binding the receptor Gprc6a.
Astrocytes Feed Memories PAGE 810
Astrocytes are known primarily for their supportive functions in the central nervous system. Now Suzuki et al. show that astrocytes play an active role in long-term plasticity and long-term memory in rats. Astrocytes in the rat hippocampus generate extracellular lactate during a memory task, and disrupting the lactate transporters on either neurons or astrocytes causes amnesia. These findings suggest that a metabolic coupling between astrocytes and neurons is critical for long-term memory formation.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 629
R&D Systems Tools for Cell Biology Research™ Cancer Development Endocrinology Immunology Neuroscience
R&D Systems MultiAnalyte Profiling Kits for
Cancer Research
Proteases Stem Cells
Cancer is a disease of context-dependent cellular dysfunction that can involve a range of biological activities. Because of its complexity, quality research tools designed to simultaneously study multiple cancer-related factors are crucial for rapidly advancing our understanding of cellular transformation. R&D Systems Proteome Profiler™ Antibody Arrays allow for the concurrent detection of multiple intracellular or secreted proteins without the need for individual Western blot experiments or specialized laboratory equipment. Streptavidin-HRP Biotin
Proteome Profiler
Light
Target Protein
Detection Antibody Capture Antibody
Array Membrane
Prostate Cancer MMP-8
Relative Protein Expression was Measured in Two Cancer Cell Types Using the Proteome Profiler Human Angiogenesis Array (Catalog # ARY007)
TIMP-1 PDGF-AA FGF Acidic Angiogenin PC
50
20
MMP-9
IGFBP-1 IGFBP-2
uPA
VEGF IL-8
For more information visit our website at www.RnDSystems.com/go/ProteomeProfiler For research use only. Not for use in diagnostic procedures.
R&D Systems, Inc. www.RnDSystems.com R&D Systems Europe, Ltd. www.RnDSystems.co.uk R&D Systems China Co., Ltd. www.RnDSystemsChina.com.cn
uPA
Serpin B5
VEGF
0
IL-8
10
PDGF-AA
PC
30
IGFBP-1
Breast Cancer TIMP-1 PDGF-AA MMP-8 Angiogenin FGF Acidic
Breast Cancer
Serpin B5
VEGF IL-8
FGF acidic
uPA
TIMP-1
IGFBP-1 MMP-9 IGFBP-2
MMP-9
Serpin B5
Pixel Density (x 103)
Prostate Cancer
40
Leading Edge
Select Cell Death and Destruction Programmed cell death and autophagy, a limited form of cellular destruction, are implicated in an ever-expanding range of biological processes. In this issue’s Select, we highlight recent insights into how these events impinge on cancer progression, Crohn’s disease, innate immune responses, and development.
Neutrophils Make the Ultimate Sacrifice Upon contact with their bacterial targets, neutrophils fire a barrage of antimicrobial peptides, sometimes releasing structures called neutrophil-associated extracellular traps, or NETs. In releasing their NETs, they commit suicide—unleashing their entire antibacterial arsenal via a specialized form of programmed cell death called netosis. Hakkim et al. (2011) now provide new insight into the molecular triggers for netosis. The authors first employ a high-throughput approach to rapidly and quantitatively examine the cellular changes that accompany NET formation in isolated human neutrophils. They then screen for small molecules that inhibit or delay netosis, isolating two classes of compounds that block NET formation. One includes chemicals that block the production of reactive oxygen species (ROS), which have been implicated in netosis, and one contains molecules that are known to inhibit transduction through the Raf-MEK-ERK signaling cascade. The authors go on to dissect how this pathway influences NET formation by challenging neutrophils with either proinflammatory lipids or Helicobacter pylori, the stomach bacterium that causes ulcers. Despite some differences, all stimuli cause netosis by elevating production of ROS by stimulation of NADPH oxidase via the Raf-MEK-ERK pathway. Though NET formation can be beneficial, excessive netosis is associated with a variety of autoimmune diseases, and this new understanding of the molecular player that is responsible for netosis could provide a foothold for an assault on excessive innate immune responses. Hakkim, A., et al. (2011). Nat. Chem. Biol. 7, 75–77. Web-like neutrophil extracellular traps (NETs) catch and neutralize the dysenteryinducing rod-shaped bacteria, Shigella flexneri. Image courtesty of V. Brinkmann.
Death Silences Developmental Noise Apoptosis is well known for its role in removing excess cells during normal developmental processes, including the destruction of the skin cells that form webs between our fingers and toes. Recent findings by Koto et al. (2011) show that apoptosis is also deployed to remove mis-specified cells that result from inborn errors of development. The authors make movies of developing sensory organs in living fly pupae and then ask a few simple questions. When, where, and how frequently are aberrant sensory cells produced? Though much is known about how Notch-Delta signaling regulates the patterns that arise during maturation of sensory organ precursors (SOPs), this study provides new insight into how fluctuations in Notch signaling synergize with neural specification factors to promote either cell differentiation or cell death. The authors suggest that transcriptional noise during cell fate determination causes nearly 20% of epithelial cells to initiate neuronal differentiation in the wrong place at the wrong time. These mis-specified SOP-like cells appear confused, following neither the neuronal nor epithelial differentiation program. Further examination of the SOP-like cells indicates that a caspase-dependent mechanism kills them, ensuring that sensory bristles on the fly notum occur at regularly spaced intervals. The data provide several hints that activation of Notch, in the presence of neural determinants, favors apoptosis. Though this study focuses on the development of sensory organs on a specific patch of the fly, it is tempting to speculate that similar failsafe systems operate in many places and in many organisms, guaranteeing robust control over the number, type, and patterns of cells during development. Koto, A., et al. (2011). Curr. Biol. 21, 1–10.
Aberrant SOP-like cells (white arrows) express both neurogenic genes (magenta) and high Notch activity (green) before they die. Image courtesy of A. Koto and M. Miura.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 631
Expand your stem cell library and save today on the latest books on stem cells and regenerative medicine Stem Cells
Stem Cell Anthology
Scientific Facts and Fiction
From Stem Cell Biology, Tissue Engineering, Regenerative Medicine, Cloning and Stem Cell Methods
Christine Mummery, Ian Wilmut, Anja Van de Stolpe and Bernard Roelen November 2010 | 400 pages | Paperback | $79.95 | €57.95 | £48.99 | ISBN: 9780123815354
Principles of Regenerative Medicine, 2nd Edition
Bruce M. Carlson October 2009 | 450 pp. | Hardback | $150.00 | €100.00 | £95.00 |AU$222.00 | ISBN: 9780123756824
Essential Stem Cell Methods
Anthony Atala and Robert Lanza November 2010 | 1400 pages | Hardback | $199.95 | €143.00 | £125.00 | ISBN: 9780123814227
A Volume in the Reliable Lab Solutions Series
Heart Development and Regeneration, 2-Volume Set
Tissue Engineering
Robert Lanza and Irina Klimanskaya April 2009 | 628 pp. | Paperback | $75.00 | €50.95 | £45.99 |AU$111.00 | ISBN: 9780123750617
Nadia Rosenthal and Richard P. Harvey June 2010 | 1072 pp. | Hardback | $199.95 | €143.00 | £125.00 | AU$296.00 | ISBN: 9780123813329
Clemens van Blitterswijk, Peter Thomsen, Jeffrey Hubbell, Ranieri Cancedda, Anders Lindahl Sahlgrenska, Jerome Sohier and David F. Williams March 2008 | 760 pp. | Hardback | $115.00 | €76.95 | £69.99 |AU$170.00 | ISBN: 9780123708694
Essentials of Stem Cell Biology, 2nd Edition
Human Stem Cell Manual
Robert Lanza, Roger Pedersen, John Gearhart, E. Donnall Thomas, Brigid Hogan, James Thomson, Douglas Melton and Sir Ian Wilmut June 2009 | 600 pp. | Hardback | $199.95 | €134.00 | £125.00 | AU$302.00 | ISBN: 9780123747297
Jeanne F. Loring, Robin L. Wesselschmidt and Philip H. Schwartz June 2007 | 488 pp. | Spiral bound | $88.95 | €59.95 | £53.99 |AU$132.00 | ISBN: 9780123704658
Foundations of Regenerative Medicine Clinical and Therapeutic Applications Anthony Atala, Robert Lanza, James Thomson and Robert Nerem September 2009 | 750 pp. | Hardback | $99.95 | €66.95 | £60.99|AU$148.00 | ISBN: 9780123750853
A Laboratory Guide
Handbook of Stem Cells 2-Volume Set with CD-ROM Vol. 1–2 Vol. 1 – Embryonic Stem Cells Vol. 2 – Adult & Fetal Stem Cells Robert Lanza, Roger Pedersen, Helen Blau, E. Donnall Thomas, John Gearhart, James Thomson, Brigid Hogan, Catherine Verfaillie, Douglas Melton, Irving Weissman, Malcolm Moore and Michael West September 2004 | 1,760 pp. | Hardback | $566.00 | €380.00 | £345.00 | AU$817.00 | ISBN: 9780124366435
Cell Stem Cell subscribers save 25% on their book order Secure ordering online at elsevierdirect.com Enter promo code 28024 at check out Prices and publication dates subject to change without notice.
One Way or Another, Ras Is Gonna Getcha Although activating mutations in oncogenic proteins, such as Rasv12, promote hyperproliferation in a wide variety of cancers, including squamous cell carcinomas and neuroblastomas, these same mutated proteins can also trigger senescence and cell death. How an activated oncoprotein promotes programmed cell death has been controversial. Elgendy et al. (2011) now supply evidence that oncogenic Ras signals through the kinase ERK, stimulating excessive autophagy that culminates in caspase-independent cell death. The authors conditionally induce Rasv12 in a variety of epithelial cell lines and then study how these cells die. They show that induction of the constitutively active Rasv12, but not the dominant-negative RasN17, promotes proliferative arrest followed by cell death. Prior to death, these cells upregulate the BH3-only protein Noxa, which has been linked to initiation of apoptosis. Surprisingly, the dying cells display none of the typical hallmarks of apoptosis, instead exhibiting excessive autophagy. Generally, BH3-only High levels of Ras-induced autophproteins promote apoptosis by binding to antiapoptotic members of the Bcl-2 family. Recently, agy trigger the death of human several Bcl-2-related proteins have been found to bind and inhibit a proautophagy protein, ovarian epithelial cells. Image Beclin-1. The authors tested whether Rasv12-induced upregulation of Noxa promoted autophagy courtesy of S. Martin. by antagonizing the inhibition of Beclin by Bcl-2. Consistent with this notion, Elgendy and colleagues find that Noxa can bind Bcl-2 family members and sequester them from Beclin-1. In addition, they show that downregulation of Beclin-1 or the autophagosome protein Atg5 prevents Rasv12-induced cell death. The authors speculate that this same mechanism of induced autophagic cell death may be responsible for the spontaneous regression of some human neuroblastoma tumors, raising the possibility that understanding the innate potential of oncogenic cells to switch from unrestrained proliferation to cell death may provide a new avenue for cancer therapies. M. Elgendy et al. (2011). Mol. Cell, in press. Published online February 24, 2011. 10.1016/j.molcel.2011.02.009.
miRs Link Autophagy and Crohn’s Disease Increased risk of Crohn’s disease has been linked to more than 70 genetic polymorphisms in loci-encoding proteins that are required for myriad cellular processes, including cell fate determination, inflammation, and autophagy. However, the pathobiology of this chronic disorder remains poorly defined. In a recent study, Brest et al. (2011) connect deregulation of a specialized form of autophagy that degrades phagocytosed bacteria, xenophagy, with the pathology of Crohn’s disease. The authors find that a synonymous single-nucleotide polymorphism (SNP) in the coding region of IRGM, a gene that is required for xenophagy, has profound implications for the regulation of IRGM expression. They show that, rather than changing protein function, this nucleotide change makes IRGM transcripts less susceptible to regulation by the mircoRNA, miR-196. Because the disease-linked SNP is located within the miRNA seed region of IRGM, it alters the binding affinity between the mRNA-miRNA duplex and the RNA-induced silencing complex, RISC. Brest and coworkers find that miR-196 binds more tightly to transcripts containing the protective (nondisease associated) allele and suggest that altered control of IRGM levels by miR-196 explains, in part, the risk of developing Crohn’s disease. They go on to investigate how deregulation of IRGM could explain the massive inflammation observed in the intestinal mucosa of Crohn’s patients. Under normal circumstances, gut inflammation in response to E. coli infection dampens autophagic flux and eventually downregulates the immune response. However, in the inflamed intestinal epithelium of Crohn’s patients, levels of autophagy and xenophagy remain high, prolonging the inflammatory immune response that leads to bowel dysfunction. Brest, P., et al. (2011). Nat. Genet., in press. Published online January 30, 2011. 10.1038/ng.762.
The Shape of Death’s Dark Shadow The multimeric ring-shaped apoptosome binds specific procaspases, initiating the caspase cascade that is responsible for the intrinsic programmed cell death pathway. With their most recent study, Yuan et al. (2011) offer new insights into apoptosome assembly, function, and evolution. The authors use cryo-electron microscopy and homology-based modeling to determine the structure of the biologically active apoptosome formed by the Drosophila Apaf-1related killer protein, Dark. They then compare their latest structure with those from other species. In all organisms, apoptosomes are homo-oligomers, comprised of a variety of domains that bind nucleotides, facilitate inter- and intramolecular interactions, and recruit procaspases. The authors highlight differences in caspase recruitment domain (CARD) organization in different organisms’ apoptosomes, suggesting that the complex evolutionary history of the apoptotic pathways is reflected in these variations. Additional structural insights from this work include variability in the necessity of cytochrome c and ATP for apoptosome formation, as well as the presence of distinct nucleotide binding motifs. Despite many differences among these divergent procaspase activating platforms, including distinct cylindrical symmetries, the authors find that the pattern of connections between apoptosome subunits in flies, worms, and humans is remarkably similar. Yuan, S., et al. (2011). Structure 19, 128–140.
The eight subunits of the Dark apoptosome form a ring-like platform with a central hub, topped by an octagonal crown of CARDs (shown in green). Image courtesy of C. Akey.
Kara L. Cerveny Cell 144, March 4, 2011 ª2011 Elsevier Inc. 633
I M AG I N G S Y S T E M S // G E L D O C E Z
Quality images at the push of a button.
That’s im mag gef fo or tllessly. Now you can save time, space, and hassle with the Gel Doc™ EZ imager. All you do is push the green button and publication-ready images are generated in just seconds. Q
Superior image resolution produces high-quality images at any zoom level.
Q
Filters, lenses, and light sources are controlled automatically, saving time and eliminating error.
Q
The modular design lets you purchase only the applications you need.
See how easy it is to imageffortlessly with Gel Doc EZ. Go to www.bio-rad.com/ad/imageffortlessly or contact your Bio-Rad sales representative.
Rese earch.. Tog geth her.
To find your local sales office, visit www.bio-rad.com/contact/ In the U.S., call toll free at 1-800-4BIORAD (1-800-424-6723)
Visit us at www.bio-rad.com
Leading Edge
Analysis Exome Sequencing Deciphers Rare Diseases Two years ago, NIH’s Undiagnosed Diseases Program began delivering genomics to the clinic on an unprecedented scale. Now, with 128 exomes sequenced and 39 rare diseases diagnosed, the program’s success is paving the way for widespread personal genomics while pioneering new techniques for reigning in the ‘‘tsunami’’ of genomics data. In 2009, a healthy Colombian couple watched their two sons suffer from a mysterious neurological illness, featuring seizures, tremors, and other complications. When the younger son succumbed to the illness at age 13, the family decided to try a new approach. They enrolled their elder son, who will be referred to here as Carlos, in NIH’s Undiagnosed Diseases Program, a transinstitute initiative dedicated to deciphering the cause of rare, mysterious health conditions. The Undiagnosed Diseases Program at NIH started in May 2008 as a pilot program with initial funding of $280,000. Since then, it has grown to include 75 physicians and scientists from almost every institution at NIH, including endocrinologists, immunologists, oncologists, and cardiologists. The team now works together on 300 cases with $3.5 million in funding each year. What sets the program apart from other large clinical projects is their state-of-theart genetic analyses. The team’s Illumina platforms can sequence the entire exome—that is, the 180,000 exons in the
human genome—of a patient and his family in only 9 weeks. In addition, high-resolution microarrays can genotype a million single-nucleotide polymorphisms (SNPs) for each family member, providing information about the remaining 99% of the genome—the introns and noncoding regions. In this way, the Undiagnosed Diseases Program is leading the charge in bringing genomics to the clinic. So why is this type of clinical genomics not routinely available outside the NIH? Although the price of sequencing has dropped to roughly $15,000 per genome or about $3000 for an exome, data storage and the ensuing analysis remains too costly, too laborious, and frankly too inefficient to put into common practice in the clinic. It’s cheap and fast to sequence, but to glean diagnostic information from the data is still quite laborious and costly. But recent successes by the Undiagnosed Diseases Program are shifting this trend. First, the program is learning how to identify patients who might benefit most from genomic analysis. Second, the
team has developed tools for quickly narrowing down the flood of genomics data to a handful of candidate genes for functional tests. With these new strategies, the program has made 39 diagnoses, including the team’s first reported discovery of a new disease, which was published last month in The New England Journal of Medicine. Through the program, ‘‘scientists and clinicians will get an understanding of the difficulties involved in next-gen sequencing, as well as the benefits; of what cases are appropriate for genomic study and of what filters should be applied to the data, ‘‘ says William Gahl, the program’s director and clinical director of the National Human Genome Research Institute. Family Matters First Before Carlos and his family arrived at the NIH Clinical Center, their first major hurdle was acceptance into the program. The NIH has received 4600 inquiries for the program and 1700 applications, but it has enrolled only 380 cases. In general, the program selects patients whose maladies appear genetic in origin, based on their family history or a sign that their symptoms have a shared underlying cause. Typically, the NIH team first attempts to diagnose illnesses by looking for known genetic markers and using standard molecular and biochemical tests that are commercially available. However, if they have no good leads on the disease and therefore no candidate genes to search for, they’ll look to a patient’s whole exome. An entire exome yields an overload of information, however, and must be carefully edited. Exome sequences typically vary at about
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 635
20,000 places, explains Thomas Markello, a clinician involved with the program at NHGRI. This vast variation makes it nearly impossible to find 1–2 rare alleles underlying a mysterious disease. In other words, the nucleotide haystack is just too big. After some trial-and-error, the team has developed a strategy to quickly shrink the haystack: they eliminate many variants by comparing the patient’s DNA with genomic information from the family. This added information enhances the signal-to-noise ratio and allows them to reduce the number of candidate genes. ‘‘Data reduction is pretty much everything,’’ Markello says. ‘‘It’s simply not worth doing whole exome sequencing on a single individual,’’ says Gahl, ‘‘when the clinical manifestations don’t point to a specific group of genes.’’ With family information, the team can apply classical pedigree tools reminiscent of population genetics in the 1970s to filter out mutations that don’t follow predicted modes of Mendelian inheritance. And DNA from healthy family members allows them to eliminate harmless variations that run in the family but don’t lead to the disease. ‘‘Family data makes a fundamental difference,’’ says David Adams, an NHGRI clinician who worked on Carlos’s case. ‘‘You don’t just need family history; we’ve learned you need family DNA to succeed.’’ So for Carlos’s case, the geneticists sequenced his exome, as well as that of his parents and his deceased brother. In general, the team gets 88% of the complete exome with 99.9% confidence. The exome data from Carlos and his family initially generated 120,000 variants for program geneticists to sift through. Remarkably, the team then narrowed this list to three candidate genes using a series of software programs developed internally at the NIH. The geneticists first align the patient’s exome with a reference sequence (typically the one generated in the Human Genome Project) and exclude unlikely candidate mutations according to a killlist of variants present in more than 1% of exomes and genomes stored in databases from two separate projects, NIH’s ClinSeq and the 1000 Genomes Project. Mutations are then ranked by how
severely they alter the coding sequence. For example, a mutation encoding a stop codon is considered more severe than one with no effect on the resulting amino acid. Another in-house software program then eliminates mutations if their inheritance pattern doesn’t match that of the disease. For example, Carlos’s disease appeared to follow a recessive mode of inheritance, and thus, his team kept only the variants that behaved accordingly. Finally, the team exports the results to their program VarSifter, which lists the candidate variants in order from least to most likely. Currently, some of this software is available to the scientific community, and according to NHGRI’s Nancy Hansen, the entire suite may be released once it’s optimized so that clinicians and scientists might follow the lead of the Undiagnosed Diseases Program. Success Stories Accumulating Although the program’s strategy is still evolving, this ‘‘whittling down’’ approach is producing results. At the end of 2010, the program identified the root of Carlos’s neurological symptoms, and a report on the teen’s diagnosis is pending publication. And data from the Undiagnosed Diseases Program team also helped Manfred Boehm at the National Heart, Lung and Blood Institute quickly locate the mutation behind a mysterious vascular calcification disorder. The team handed him a list of 100 candidate genes, and within 1 month he had pinpointed mutations in the NT5E gene as the causal culprit. NT5E encodes a membrane-bound nucleotidase involved in extracellular ATP metabolism. Boehm’s study identified in three families numerous mutations, which destroy the activity of the nucleotidase. By targeting this enzyme, clinicians may soon be able to treat a disease that didn’t even have a name a few months ago. The study was published last month in The New England Journal of Medicine. In contrast, Boehm says, a couple of years ago, NIH geneticists handed him a list of 2000 genes potentially underlying a known disorder, and a year passed before he could identify the causal mutation.
636 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
Thus far, the program has diagnosed 39 cases—3 of which are neurological and muscular diseases discovered through exome analysis. Another 3 were discovered with SNP analysis, whereas the remaining cases were diagnosed with commercial tests, including the identification of the rare disease congenital disorder of glycosylation type 2B. The Undiagnosed Diseases team also counts concise lists of candidate mutations as partial successes. These ‘‘short lists’’ are handed off to bench scientists for functional studies on how the mutations contribute to disease. Right now, these collaborations are currently all within the NIH. But program leaders say they’ll work with basic researchers outside of the NIH once they’ve established a portal for collaboration. Marjan Huizing, a metabolic disorders researcher at NHGRI, is not involved with the program but says she’s ‘‘piggybacking’’ off their exomic techniques. She studies endocytic trafficking defects that lead to albinism, bleeding, and infections. Traditional candidate gene approaches have frustrated her team. They spent 12 years screening patients for candidate genes and never found the right one. So she decided to follow in the program’s lead and analyze the exomes of three patients. Right off the bat, she identified the causal mutation underlying two of the cases. For the third, more difficult case, she’s planning to sequence the patient’s parents. In the meantime, geneticists are keeping an eye on the program for clues about the nature of rare diseases and whether sequencing alone can identify ‘‘medically actionable alleles,’’ says Harvard genomicist George Church. ‘‘If that turns out to be true, it’s a paradigm shift.’’ Although the clinicians at the Undiagnosed Diseases Program echo Church’s curiosity, they keep the focus tightly on their patients. The thrill of a diagnosis through genomics is a triumph for scientists and clinicians, but it means even more to patients and their loved ones, who have sought explanations for years. Still, it is a bittersweet success. Carlos now has a name to call his malady, but there remains no cure or powerful drug
to combat it. ‘‘Even when we get lucky, the best we can do is offer hope to the patient that down the road, someone might use
the diagnosis to develop a treatment,’’ says John Gallin, director of the NIH Clinical Center. ‘‘Our patients have referred
to this as the House of Hope, but from the perspective of a care provider, I wish I could do even more.’’
Amy Maxmen New York, NY, USA DOI 10.1016/j.cell.2011.02.033
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 637
Leading Edge
Previews Uncovering a Tumor Suppressor for Triple-Negative Breast Cancers John G. Albeck1 and Joan S. Brugge1,* 1Department of Cell Biology, Harvard Medical School, Boston, MA 02115, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.030
‘‘Triple-negative’’ breast cancers are aggressive malignancies that respond poorly to treatments. Now Sun et al. (2011) find that the activity of the protein tyrosine phosphatase PTPN12 is lost in a large percentage of this breast cancer subtype, offering molecular drivers and possible therapeutic targets for this heterogeneous and intractable cancer. The molecular characterization of breast cancer subtypes has lead to significant progress in treating this common disease. Tumors that express the estrogen receptor (ER) or progesterone receptor (PR) respond well to ER antagonists or aromatase inhibitors (which block estrogen synthesis), whereas HER2-positive tumors are effectively treated with a HER2-blocking antibody (Trastuzumab) or a HER2 kinase inhibitor (Trikerb). However, 15%–20% of breast cancers fall through the cracks of this classification scheme. Called ‘‘triple-negative’’ breast cancers, these cells lack detectable expression of the three key molecular signatures: ER, PR, and overexpressed HER2. Triple-negative breast cancers have the worst prognosis among all breast cancers and the fewest treatment options. Because these cancers are defined by what they lack, no common molecular ‘‘drivers’’ are available for developing targeted therapies. Moreover, triplenegative breast cancers are a highly heterogeneous group, and it has been difficult to identify common mutations underlying the disease. In this issue of Cell, Sun et al. (2011) uncover a new molecular feature common to a large fraction of triple-negative breast cancers: the loss of the protein tyrosine phosphatase PTPN12 (also known as PTP-PEST). Sun et al. begin by performing a lossof-function screen for human kinases and phosphatases that regulate anchorage-independent growth in human breast epithelial cells. The top hit was PTPN12. They then show that loss of PTPN12 expression enables these cells to grow in the absence of extracellular
matrix and also leads to their hyperproliferation in a three-dimensional culture model that recapitulates structures of the mammary gland. These two assays are widely used to assess oncogenic transformation, and PTNP12’s behavior in these assays establishes it as a candidate tumor suppressor gene in mammary epithelial cells. Importantly, the ability of PTPN12 to block growth in these assays depends on its tyrosine phosphatase activity. Sun and colleagues then characterize the signaling events altered by disruption of PTPN12. The phosphorylation of specific tyrosines on both the EGFR and HER2 receptor tyrosine kinases increases, as does the activity of the downstream ERK/RSK (extracellular signal-regulated kinase/ribosomal S6 kinase) pathway. Together, these results indicate that PTPN12 acts as a growth suppressor by antagonizing key receptor tyrosine kinase pathways. Further, an analysis of PTPN12 gene sequences reveals a series of point mutations found only in triple-negative breast cancer cells. Using in vitro functional tests, Sun and colleagues confirm that these mutations represent loss-of-function alleles. Additionally, 20% of all breast cancers contain deletions encompassing PTPN12, although other genes within the deleted region may also be involved in tumorigenesis. Perhaps most strikingly, Sun and colleagues find that 60% of triple-negative breast cancer cells are histologically negative for PTPN12 in a panel of 185 breast cancers, whereas only 9% of HER2-positive cancers are PTPN12 negative.
638 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
Sun and colleagues also note that the PTPN12 mRNA contains three binding sites for miR-124, a microRNA repressed by the tumor suppressor REST (RE1-silencing transcription factor). The expression of REST correlates with that of PTPN12, which is consistent with miR-124 suppressing PTPN12. In addition, overexpressing miR-124 promotes anchorageindependent growth, as was observed when PTPN12 was disrupted. The miR124 locus is amplified in 20% of breast cancers, suggesting yet another mechanism by which PTPN12 expression may be lost in cancer. Together, these findings strongly suggest that PTPN12 loss plays a role in triple-negative breast cancers. Moreover, the near mutual exclusivity of HER2 overexpression and loss of PTPN12 suggests that disrupting PTPN12 represents an alternate route for activating survival and proliferation pathways known to drive cancer development. Another intriguing finding by Sun and colleagues is that a common singlenucleotide polymorphism in the PTPN12 locus (Thr573Ala) represents a loss-offunction variant. Although the correlation of this polymorphism with breast cancer falls below a strict threshold of statistical significance (p = 0.2), the trend hints that this variant may predispose an individual to breast cancer. Indeed, a previous study sequenced this polymorphic locus in 19 breast cancer cell lines (Streit et al., 2006), with the loss-of-function allele occurring more frequently in triple-negative breast cancer cell lines (4 out of 9) than in other types of breast cancer (2 out of 9). Interestingly, this previous study also found that the alanine substitution,
which resides outside the catajunctions (e.g., caveolin, deslytic domain, enhanced phosmoplakin, cingulin, and clauphatase activity in vitro. Given din-1). Together, these data that Sun et al. found loss of are consistent both with Sun tumor suppressor function in and colleagues’ conclusion cultured cell transformation that loss of PTPN12 in tripleassays, these results raise the negative breast cancer cells possibility that the alanine stimulates a wide range of substitution alters PTPN12 signaling programs and with localization or its access to other studies linking PTPN12 substrates inside the cell. to the regulation of cell adheUnfortunately, despite the sion and motility (Angersprevalence of PTPN12 loss in Loustau et al., 1999; Garton triple-negative breast cancer and Tonks, 1999). cells, there is currently no feaThe promiscuity of tyrosine sible way to directly restore kinase signaling networks the function of this phosphahas also been observed in tase therapeutically. Morelung cancer, in which signaling over, although PTPN12 apbehaves as a highly interconnected web of substrates pears to regulate the EGFR/ that collapses dramatically HER2 pathway, inhibitors of EGFR and HER2 have failed when key ‘‘hub’’ tyrosine kinases are inhibited (Guo to show clinical efficacy for et al., 2008; Rikova et al., triple-negative breast can2007). These webs appear to cers, even though many of be highly flexible, resulting in these breast cancers display cells with variable sensitivity high levels of EGFR. Thereto tyrosine kinase inhibitors, fore, reversing the effects of depending on the cell’s profile PTPN12 loss may require inhiof kinases and phosphatase bition of multiple tyrosine levels. Given that a number of kinases. This prompted Sun tyrosine kinase inhibitors are and colleagues to identify addicurrently approved for treattional targets of PTPN12 using ment of other cancers, each a bimolecular fluorescence Figure 1. Triple-Negative Breast Cancer Cell Lines Are Enriched of which targets three or complementation assay. with Tyrosine Phosphorylation Sites more different tyrosine These experiments pinpoint ‘‘Triple-negative’’ breast cancers, which do not express the three major kinases, it will be of great the platelet-derived growth molecular signatures of breast cancer cells (estrogen receptor, progesterone interest to determine whether factor receptor-b (PDGFR-b) receptor, and HER2), are heterogeneous and difficult to treat. Sun et al. (2011) now find that the activity of the protein tyrosine phosphatase PTPN12 is lost in combinations of these drugs as another tyrosine kinase that many of these cancers. Here a Wilcoxon rank sum test was used to identify are effective against tripleinteracts with PTPN12. Indeed, tyrosine sites preferentially phosphorylated in triple-negative breast cancer inhibiting both EGFR/HER2 negative breast cancer, and cell lines versus other types of breast cancer cells (all sites shown are signifand PDGFR-b with two inhibiwhether PTPN12 status can icant at the p < 0.05 level and are ranked from top to bottom in order of increasing p value). Asterisks indicate sites that may also be phosphorylated tors, lapatinib and sunitinib, predict sensitivity. on closely related proteins. Data provided by Ting-lei Gu of Cell Signaling slows the growth of xenoAlthough the work of Sun Technology, and the breast cancer dataset is downloadable from Phosphografted triple-negative breast and colleagues opens a new SitePlus at http://www.phosphosite.org/downloads/breastcancerdataset.gz. cancer tumors. window on a poorly underImportantly, the central stood and aggressive maligconcept advanced by Sun and colleagues tion are enriched in triple-negative breast nancy, it also underscores the inherent is not that PDGFR-b is a key target in triple- cancer cell lines (Figure 1), based on difficulties in targeting complex cancers negative breast cancer—indeed, datasets mass spectrometry analysis of immunoaf- like triple-negative breast cancers. The from Oncomine (http://oncomine.org) finity-purified phosphopeptides (Ting- signaling changes that accompany the indicate that PDGF-a/b are not preferen- lei Gu, Cell Signaling Technology; loss of PTPN12 or overexpression are tially expressed in triple-negative breast http://www.phosphosite.org/downloads/ modest, in the range of 2- to 3-fold. This cancer—but rather that the treatment of breastcancerdataset.gz). Many of these suggests that PTPN12 exerts its potent this obstinate cancer may require combi- sites are located on receptor tyrosine phenotypic effects not through a single natorial inhibition of multiple tyrosine kinases (e.g., EGFR, MET, Axl, and main effector but rather by many small kinases. In support of this idea, we note EphAs), Src family kinases (SFKs), and changes distributed throughout a signalthat many sites of tyrosine phosphoryla- proteins involved in cell adhesion and tight ing network. Cell 144, March 4, 2011 ª2011 Elsevier Inc. 639
Lastly, it is important to note that a sizable fraction of the general population (2.5%) harbors single-nucleotide polymorphisms that inactivate the tumor suppressor activity of PTPN12, but the majority of these individuals do not develop cancer. Therefore, loss of PTPN12 alone is not sufficient to drive tumor formation. Interestingly, the effect of PTPN12 appears confined to the ERK/RSK pathway, without affecting the important PI3K/Akt pathway. This observation suggests that PTPN12 loss cooperates with mutations in other pathways, including the PI3K/Akt pathway, which is hyperactivated in nearly 50% of
breast cancers. These considerations together lead us to conclude that, although PTPN12 may be an important piece of the triple-negative breast cancer puzzle, it is still a member of an even larger constellation of alterations that together drive this recalcitrant malignancy. REFERENCES Angers-Loustau, A., Cote, J.F., Charest, A., Dowbenko, D., Spencer, S., Lasky, L.A., and Tremblay, M.L. (1999). J. Cell Biol. 144, 1019–1031. Garton, A.J., and Tonks, N.K. (1999). J. Biol. Chem. 274, 3811–3818.
Guo, A., Villen, J., Kornhauser, J., Lee, K.A., Stokes, M.P., Rikova, K., Possemato, A., Nardone, J., Innocenti, G., Wetzel, R., et al. (2008). Proc. Natl. Acad. Sci. USA 105, 692–697. Rikova, K., Guo, A., Zeng, Q., Possemato, A., Yu, J., Haack, H., Nardone, J., Lee, K., Reeves, C., Li, Y., et al. (2007). Cell 131, 190–203. Streit, S., Ruhe, J.E., Knyazev, P., Knyazeva, T., Iacobelli, S., Peter, S., Hoefler, H., and Ullrich, A. (2006). Cancer Genet. Cytogenet. 170, 48–53. Sun, T., Aceto, N., Meerbrey, K.L., Kessler, J.D., Zhao, C., Migliaccio, I., Nguyen, D.X., Pavlova, N.N., Botero, M., Huang, J., et al. (2011). Cell 144, this issue, 703–718.
The Ribosome and TORC2: Collaborators for Cell Growth Xiaoduo Xie1 and Kun-Liang Guan1,* 1Department of Pharmacology and Moores Cancer Center, University of California, San Diego, La Jolla, CA 92093-0815, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.029
The target of rapamycin complex 2 (TORC2) is a key regulator of cell growth. Zinzalla et al. (2011) now provide evidence that TORC2 is activated by direct association with the ribosome, which may ensure that TORC2 activity is calibrated to match the cell’s intrinsic growth capacity. TOR is an atypical serine/threonine protein kinase conserved from yeast to mammals that forms two distinct physical and functional complexes termed TORC1 and TORC2 (Figure 1) (Wullschleger et al., 2006). TORC1, a rapamycin-sensitive complex, regulates translation, autophagy, cell growth, and cell size. In contrast, TORC2 is not directly inhibited by rapamycin and controls cell survival and morphology. The frequent dysregulation of mammalian TOR (mTOR) signaling observed in human cancer is thought to contribute to tumorigenesis (Zoncu et al., 2011). Extensive studies have revealed the molecular mechanisms of mTORC1 regulation in response to signals, such as growth factors, cellular energy status, nutrient availability, and stress. Although compelling evidence
has placed TORC2 downstream of phosphatidylinositol 3-kinase (PI3K) and upstream of the serine/threonine kinase AKT, it has been unknown how TORC2 is regulated (Wullschleger et al., 2006). Using a combination of elegant genetic screening and sophisticated biochemical studies, Zinzalla et al. (2011) significantly advance our understanding of TOR biology by identifying the ribosome as a missing link between PI3K and mTORC2 (Figure 1). In yeast, TORC2 phosphorylates and activates YPK2, the ortholog of mammalian kinase SGK1, a known substrate of mammalian TORC2. Loss of TORC2 function is lethal in yeast; however, overexpression of a constitutively active YPK2 suppresses the lethality caused by a loss-of-function TORC2 mutation
640 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
(Kamada et al., 2005). Zinzalla et al. designed a clever reverse suppressor screen in search of yeast mutants that require the expression of constitutively active YPK2 for survival. This strategy was aimed at uncovering mutations in TORC2 upstream activators. Perhaps not surprisingly, many mutations isolated were found in genes encoding components of TORC2. Interestingly, the only non-TORC2 component isolated was NIP7, which encodes a protein involved in the maturation of rRNA and ribosome biogenesis. Survival of NIP7 mutant yeast requires overexpression of the constitutively active YPK2. Many scientists would abstain from studying ribosomal proteins because their inactivation may disrupt protein translation and lead to pleiotropic effects. Undeterred, Hall and colleagues
investigated further, eventuthe ribosome regulated in ally discovering an exciting yeast? And if so, by what biochemical mechanism linksignals? The TORC2-riboing the ribosome to TORC2 some interaction is conserved activation. from yeast to mammals; The genetic studies not however, there is no evidence only confirm a role for NIP7 that yeast TORC2 is activated in TORC2 activation, but also in a PI3K-dependent manner show that ribosomal proteins as in mammalian cells. are important for TORC2 How could mTORC2 be function in yeast. Mutation of responsible for phosphoryNIP7 mimics the TORC2 lating two different sites on the same protein with drastiloss-of-function phenotypes, cally different manners of indicating a strong functional regulation? mTORC2 phosrelationship between these phorylates both the hydrotwo genes. The authors phobic motif and the turn further extended their study motif of AKT. Notably, phosto mammalian cells and show that mammalian NIP7 phorylation of the hydrophobic motif is posttransla(mNIP7) and ribosome Figure 1. Regulation of TORC2 by PI3K and Ribosome Association tional and induced in a assembly are important for In response to upstream stimulation, increased phosphatidylinositol 3-kinase PI3K-dependent manner, mTORC2 activation. Knock(PI3K) signaling activates mTORC2 (consisting of mTOR, Rictor, Sin1, and mLST8) by promoting its association with ribosomes. mTORC2 phosphorywhereas phosphorylation of down of mNIP7 reduces lates the C-terminal turn motif (T450) and hydrophobic motif (S473) in AKT. the turn motif is cotranslamTORC2 activity, as indiAKT activation also requires phosphorylation of the activation loop (T308) by tional and constitutive (Oh cated by a decrease in the PDK1. Notably, phosphorylation of the turn motif of AKT is constitutive and et al., 2010). Ribosomal phosphorylation of mTORC2 cotranslational, whereas phosphorylation of the hydrophobic motif is highly dependent on PI3K signaling and posttranslational. TORC2 similarly phosassociation may explain the substrates. Moreover, knockphorylates the turn motif and hydrophobic motif in SGK and conventional constitutive and cotransladown of either Rpl7 (a subunit protein kinase C (Garcı´a-Martı´nez and Alessi, 2008; Jacinto et al., 2004; Sartional phosphorylation of the of the 60S ribosome) or bassov et al., 2005). turn motif in AKT (Figure 1). Rps16 (a subunit of the Membrane association has 40S ribosome) inactivates mTORC2, demonstrating the importance the interaction between mTORC2 and been suggested to be important for of ribosome in mTORC2 activation. Inhibi- ribosome, demonstrating the necessary AKT hydrophobic motif phosphorylation. tion of protein translation had no effect on and sufficient role of PI3K in regulating Does this imply that there are two mTORC2 activation, supporting the the mTORC2-ribosomal association and different pools of mTORC2? What is the exact role of the ribosome in notion that mTORC2 is activated by the thus mTORC2 activation. The role of PI3K ribosome, but not translation. Addition- in promoting the interaction of mTORC2 mTORC2 activation? Does ribosome ally, extensive biochemical studies and ribosome was further validated in binding regulate mTORC2 function by demonstrate that mTORC2 can associate multiple cancer cell lines with high PI3K affecting localization, contributing to actiwith the ribosome, and the ribosomal- signaling. Collectively, the study by Zin- vation, or maintaining an active mTORC2 associated mTORC2 displays kinase zalla et al. has convincingly demonstrated kinase? activity toward AKT in vitro. It appears that the PI3K-dependent mTORC2Lastly, why the ribosome? As pointed that the mTORC2 components, rictor ribosome interaction plays a major role in out by Zinzalla et al., ribosome content and/or sin1, which are not found in TORC2 activation. Interestingly, Oh et al. reflects the cell’s potential for growth. TORC1, interact with the 60S subunit of (2010) have also recently reported the Therefore, a mechanism that involves ribosome. association of mTORC2 with the ribosome the ribosome may ensure that TORC2 Importantly, the authors link this associ- and propose that the ribosomal associa- activation is tightly linked with the cell’s ation with known upstream regulators of tion is important for the cotranslational ability to respond to growth signals. In TORC2, demonstrating that the interaction phosphorylation of the AKT turn motif. addition, given that TORC1 regulates between ribosome and TORC2 is strongly The findings of Zinzalla et al. raise many ribosome biogenesis, this study suggests enhanced by insulin stimulation. Inhibition interesting questions. How is the interac- a possible interplay between the two of PI3K activity blocks the interaction tion between mTORC2 and the ribosome TORC complexes, in which TORC2 between the ribosome and mTORC2, as regulated? There may be intermediate modulates TORC1 by activating AKT well as inhibits mTORC2 activation in players between PI3K and the TORC2- and TORC1 affects TORC2 by increasing response to insulin. Complementary acti- ribosome interaction. For example, phos- ribosomal biogenesis and S6K-mediated vation of PI3K by knockdown of PTEN— phorylation of rictor may play a role feedback inhibition. Future studies are a tumor suppressor lipid phosphatase (Dibble et al., 2009; Treins et al., 2010). needed to address these important questhat counters PI3K signaling—increases Is the interaction between TORC2 and tions and possible new roles of the Cell 144, March 4, 2011 ª2011 Elsevier Inc. 641
Treins, C., Warne, P.H., Magnuson, M.A., Pende, M., and Downward, J. (2010). Oncogene 29, 1003–1016.
Dibble, C.C., Asara, J.M., and Manning, B.D. (2009). Mol. Cell. Biol. 29, 5657–5670.
Jacinto, E., Loewith, R., Schmidt, A., Lin, S., Ru¨egg, M.A., Hall, A., and Hall, M.N. (2004). Nat. Cell Biol. 6, 1122–1128. Kamada, Y., Fujioka, Y., Suzuki, N.N., Inagaki, F., Wullschleger, S., Loewith, R., Hall, M.N., and Ohsumi, Y. (2005). Mol. Cell. Biol. 25, 7239–7248. Oh, W.J., Wu, C.C., Kim, S.J., Facchinetti, V., Julien, L.A., Finlan, M., Roux, P.P., Su, B., and Jacinto, E. (2010). EMBO J. 29, 3939–3951.
Garcı´a-Martı´nez, J.M., and Alessi, D.R. (2008). Biochem. J. 416, 375–385.
Sarbassov, D.D., Guertin, D.A., Ali, S.M., and Sabatini, D.M. (2005). Science 307, 1098–1101.
Zoncu, R., Efeyan, A., and Sabatini, D.M. (2011). Nat. Rev. Mol. Cell Biol. 12, 21–35.
ribosome in cellular signaling, especially in the regulation of cell growth.
REFERENCES
Wullschleger, S., Loewith, R., and Hall, M.N. (2006). Cell 124, 471–484. Zinzalla, V., Stracka, D., Oppliger, W., and Hall, M. (2011). Cell 144, this issue, 757–768.
The Skeleton: The New Controller of Male Fertility? Lee B. Smith1 and Philippa T.K. Saunders1,* 1MRC Centre for Reproductive Health, The University of Edinburgh, The Queen’s Medical Research Institute, 47 Little France Crescent, Edinburgh EH16 4TJ, UK *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.028
Sex steroids, including testosterone, regulate the development and function of the male skeleton. Oury et al. (2011) identify a surprising new connection between the skeleton and the testis, which has implications for male fertility. They show that testosterone production in the testis is directly influenced by the bone-derived hormone osteocalcin. The sex steroids testosterone and estrogen are essential regulators of fertility. Acting through nuclear hormone receptors expressed in somatic cells, testosterone regulates the development and maintenance of the male reproductive system and, by modulating the activity of support cells, promotes germ cell survival. Testosterone synthesis occurs within Leydig cells in the testes and is regulated by luteinizing hormone (LH) produced by the pituitary gland. In addition to LH and follicle-stimulating hormone (which is also secreted from the pituitary), there is emerging evidence that metabolic hormones such as leptin (Tena-Sempere and Barreiro, 2002) and insulin (Ma and Wittert, 2010) have impacts on testis function. Now, Oury et al. (2011) demonstrate that the skeleton may play a previously unforeseen role in control of male fertility, showing that osteocalcin (Ocn), a peptide hormone that is secreted by bone-form-
ing osteoblast cells, directly controls testosterone production by Leydig cells in the testis (Figure 1). Using ex vivo cell assays, the authors demonstrate that a factor that is secreted by osteoblasts, but not other skeletal cell-types, markedly increases testosterone production both by testis explants and by isolated Leydig cells. Speculating that this effect is likely to be mediated by osteocalcin, the authors show that this response is lost when Ocndeficient osteoblasts are used but is rescued in a dose-dependent manner by introduction of exogenous osteocalcin. Injection of osteocalcin also increases serum testosterone in mice. Together, these data strongly suggest that osteocalcin is indeed one of the osteoblast-secreted factors that are responsible for enhanced testosterone production. To support their in vitro data, the authors exploit two mouse models in which osteocalcin function is either ablated (Ocn
642 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
knockout mice) or enhanced (Esp knockout mice, which lack an osteotesticular phosphatase that negatively regulates osteocalcin maturation). Compared to controls, testis weight, sperm counts, and levels of circulating testosterone are all significantly reduced in Ocn knockout mice and conversely increased in Esp knockout males. Consistent with these physiological changes, Ocn-deficient males sire fewer offspring than controls, although no significant increase is observed using Esp-deficient males, possibly because the availability of oocytes acts as a limiting factor. Tissuespecific ablation of these genes in bone or testis confirms that osteocalcin derived from the skeleton acts as an endocrine factor to regulate male fertility. How does osteocalcin influence testosterone synthesis? To identify the receptor through which osteocalcin signals, the authors postulate that, because the
changes in the T:E ratio have hormone stimulates cyclic been implicated in the develadenosine monophosphate (cAMP) production in Leydig opment of benign and maligcells, it is most likely a nant prostate disease. InterG protein-coupled receptor estingly, a recent paper has (GPCR). By performing a identified GPRC6A as comprehensive analysis of a susceptibility locus for prosthe expression pattern of 103 tate cancer (Takata et al., GPCRs, they identify four that 2010). Disturbances in funcare enriched in Leydig cells. tional activity of bone cells Notably, a previous study reassociated with aging may ported that deletion of one of therefore have ramifications not only for male fertility, but these, Gprc6a, results in also for susceptibility to prosa testicular phenotype, with tate cancer. increased circulating levels of estradiol and reduced levels The identification of osteoof testosterone (Pi et al., calcin as a regulator of male 2008). Oury and colleagues fertility represents a step now demonstrate that Leydig change in our understanding of the homeostatic mechacell-specific impairment of nisms that are responsible Gprc6a expression generates mice with an identical phenofor regulation of androgen biosynthesis in the testis. G type to Ocn-deficient animals. protein-coupled receptors Their experiments suggest Figure 1. Hormonal Integration of the Skeleton and Testis Testosterone (T) biosynthesis takes place within Leydig cells (identified here by such as Gprc6A are popular that osteocalcin signals immunostaining for 3b-HSD, an enzyme that is involved in androgen biosyntargets for drug development, through Gprc6a to regulate thesis), located within the interstitial compartment of the testis. Testosterone and these studies may therecAMP and the transcription also acts as a pro-hormone for biosynthesis of estrogens (E); both hormones fore pave the way for treatfactor CREB, which in turn have been implicated in regulation of bone turnover. Leydig cells express G protein-coupled receptors (GPCRs) that bind the endocrine hormones luteiments for male infertility or promotes the expression of nizing hormone (Lhcgr) and osteocalcin (Gprc6a). Luteinizing hormone (LH) is other hormone-dependent key steroidogenic enzymes. synthesized and secreted by the pituitary gland and the latter produced by disorders. One outstanding question bone osteoblasts. Stimulation of both GPCRs results in increased concentrations of intracellular cAMP and increased expression of key regulators of the that remains to be addressed steroidogenic pathway. Secretion of LH is subject to negative feedback is the relationship between regulation by testosterone and estrogens. Testosterone, acting via androgen REFERENCES LH- and osteocalcin-depenreceptors expressed in somatic cells, but not germ cells, plays a critical role in the regulation of male fertility. dent signaling in regulation of Ma, P.H., and Wittert, G.A. (2010). Leydig cell function. Both of Mol. Cell. Endocrinol. 316, 180–186. these hormones exert endocrine control over testosterone production, Androgens such as testosterone not Oury, F., Sumara, G., Sumara, O., Ferron, M., and although they act through different only regulate testis function, but also Chang, H., Smith, C.E., Hermo, L., Suarez, S., Roth, B.L., Ducy, P., and Karsenty, G. (2011). Cell GPCRs, both are coupled to activation of play a key role in the function of the rest 144, this issue, 796–809. cAMP-dependent signaling cascades. In of the male urogenital system. Oury et al. this study however, even though animals show that mice in which Leydig cell func- Pi, M., Chen, L., Huang, M.Z., Zhu, W., Ringhofer, lacking Ocn have significantly increased tion is not stimulated by osteocalcin, due B., Luo, J., Christenson, L., Li, B., Zhang, J., Jackson, P.D., et al. (2008). PLoS ONE 3, e3858. levels of serum LH, the increase in LH is to the absence of either the ligand or its apparently insufficient to compensate for receptor, have significantly smaller Takata, R., Akamatsu, S., Kubo, M., Takahashi, A., loss of function of the osteocalcin pathway. epididymides and seminal vesicles. Ocn- Hosono, N., Kawaguchi, T., Tsunoda, T., Inazawa, This suggests that LH and osteocalcin may deficient males also have significantly J., Kamatani, N., Ogawa, O., et al. (2010). Nat. Genet. 42, 751–754. act through different effector molecules. It increased levels of serum estradiol, would therefore be interesting to establish leading to a disturbance in the testostero- Tena-Sempere, M., and Barreiro, M.L. (2002). Mol. whether exogenous osteocalcin could ne:estradiol (T:E) ratio. Estrogens are Cell. Endocrinol. 188, 9–13. rescue the phenotype of LH-beta knockout known to regulate both male fertility and Zhang, F.P., Poutanen, M., Wilbertz, J., and Huhtamice (Zhang et al., 2001). prostate function, and age-dependent niemi, I. (2001). Mol. Endocrinol. 15, 172–183.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 643
Leading Edge
Previews Astrocytes: Powering Memory Paola Bezzi1 and Andrea Volterra1,* 1Department of Cell Biology and Morphology, University of Lausanne, Rue du Bugnon, 9 1005 Lausanne, Switzerland *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.027
Creating long-term memory requires a cellular program in neurons involving gene expression, protein synthesis, and formation of new synaptic connections. Suzuki et al. (2011) show that astrocytes, glial cells of the brain, play a necessary role in this program by converting glycogen to lactate and transporting it to neurons. Memory is a mechanism for retaining newly learned information. External events are represented in the brain as specific spatiotemporal patterns of neural activity, and this activity can affect the way neurons communicate with each other at specialized junctions called synapses. The physical location of memory storage thus occurs at synapses that undergo activity-dependent changes in synaptic efficiency. A key brain region for storage of explicit memories—that is, memories involving conscious recall of people, places, objects, or events—is the hippocampus, where synaptic connections exhibit activity-dependent plasticity changes such as long-term potentiation (LTP) (Kandel, 2001). A memory role for astrocytes, glial cells strategically intercalated between blood vessels and synapses, has not been considered until very recently. Astrocytes are the main energy reservoirs of the brain. They accumulate glycogen and help fulfill the high-energy demands associated with neuronal activity (Attwell and Gibb, 2005; Rouach et al., 2008). However, their precise roles and mechanism of action remain unclear. Suzuki and colleagues (2011) now demonstrate a direct role for astrocyte metabolism in the memory process. Using an elegant combination of in vivo approaches in rats, the authors show that glycogenolysis, the breakdown of glycogen into glucose, occurs immediately after animals undergo a learning task and demonstrate that this metabolic event is required for consolidation of the learned paradigm (Figure 1). Consolidation is the process by which memories, initially labile and highly sensitive to disruption, are stabilized and
stored for a long time. At the cellular level, the early phase of memory (minutes to hours) relies on posttranslational modification of pre-existing proteins, whereas consolidation involves gene expression, new protein synthesis, and growth of new synaptic connections (Kandel, 2001) (Figure 1). Suzuki and colleagues show that pharmacological inhibition of glycogen metabolism during the initial phase of memory storage, but not at later stages, prevents retention of the learned task. This result suggests that glycogenolysis in astrocytes is necessary for triggering the conversion of memory from a transient short-term form to a more stable long-term one. Why is glycogenolysis, and not another energy source, required for memory consolidation? Does the process require an initial extra-energy burst? Intriguingly, the authors find that the relevant metabolic product is not glucose but lactate, a downstream product of glycogenolysis. They show that extracellular lactate levels increase in the hippocampus immediately after rats engage in a learning task. Furthermore, when they block the endogenous release of lactate, administration to the animals of exogenous lactate prevents amnesia. Glucose administration, by contrast, even at doses that are higher calorie-wise, produces only a transient effect on memory formation. Why do lactate and glucose have different effects? Perhaps different timing or location of production and delivery determine the outcome. Lactate transport among cells, for example, depends on the action of monocarboxylate transporters (MCTs), which do not transport glucose. Indeed, bilateral injection into the hippocampus of antisense oligonucleotides
644 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
directed against specific MCT isoforms (MCT1, 2, and 4) affects memory consolidation in a way that resembles the effects of a blockade of glycogenolysis. The same effect is observed when targeting either astrocytic transporters (MCT4) or neuronal ones (MCT2). However, the memory defects caused by knockout of MCT4 (the astrocytic transporter) but not those caused by knockout of MCT2 (the neuronal transporter) can be rescued by addition of exogenous lactate. These results suggest that lactate is transported from astrocytes into neurons. It is therefore likely that a specific shuttle system delivers lactate from astrocytes to neurons, as has been postulated long ago (Pellerin and Magistretti, 1994) but never directly proven. The shuttle system may be located strategically, allowing it to deliver lactate directly where it is needed, possibly in the neuronal dendritic spines that will undergo plastic transformation during memory consolidation. This hypothesis is consistent with the reported localization of MCT2 in dendritic spines at excitatory hippocampal synapses (Bergersen, 2007) (Figure 1). However, the authors find that knocking out MCT1, a transporter widely distributed in astrocytes, blood cells, and oligodendrocytes, also affects memory consolidation, which may argue against the specific shuttle hypothesis. Further studies will be needed to delineate the lactate transport route and the exact localization and specific function of each transporter in the process. Another potential difference between lactate and glucose may be that lactate has not only purely energetic functions but also (or instead) signaling functions, in analogy to other neuroactive agents
released by astrocytes (gliobetween astrocytes and neurons that involves glycogenoltransmitters; Volterra and Meldolesi, 2005). This idea is ysis. Other real-time methods consistent with recent studies may need to be developed. implicating a role for lactate in These unresolved questions the control of vessels by underline our present limited astrocytes. In this process, understanding of the modes astrocytes release vasodilatof synapse-astrocyte interacing prostaglandins and tion and the methodological lactate acts as modulator of challenges associated with prostaglandin transport, not their study. However, with as an energy source (Gordon more and more studies impliet al., 2008). cating astrocyte signaling in In the context of synaptic synaptic plasticity and now, plasticity, unexpected roles directly, in memory processes, are emerging for additional the Pandora’s Box has molecules released from asbeen opened. Understanding trocytes. For example, astroneuron-glia interactions may cyte-released TNF-a particiwell reveal new critical and pates in synaptic scaling, unexpected aspects of the a form of homeostatic plasneurobiology of memory. ticity that modulates the ACKNOWLEDGMENTS strength of an entire synaptic network depending on its Research in A.V.’s lab is supported activity history (Stellwagen by grants from the Swiss National and Malenka, 2006). On the Science Foundation and from other hand, astrocyte-reSynapsis Foundation (Zurich, Switzerland); in P.B.’s lab, from the leased D-serine plays a role in University of Lausanne and from the induction of long-term Figure 1. Astrocytes in Long-Term Memory Formation Novartis Foundation (Basel, Suzuki et al. (2011) show that glycogenolysis (the breakdown of glycogen) and potentiation at hippocampal Switzerland). the release of its downstream product lactate from astrocytes are required for synapses by acting as an memory consolidation in response to a learning paradigm. endogenous NMDA receptor (A) Rats participate in a learning task, creating a stable memory in the CA1 REFERENCES region of the hippocampus. coagonist (Henneberger (B) Some of the key molecular and structural changes at CA1 synapses (brown) et al., 2010). Interestingly, the Attwell, D., and Gibb, A. (2005). Nat. and neighboring astrocytes (light blue) that lead to the establishment of longD-serine study also suggested Rev. Neurosci. 6, 841–849. term memory. that LTP formation requires The figure is based on the available literature and the new results by Suzuki Bergersen, L.H. (2007). Neurosciet al. The learning task stimulates memory formation that, in the early phase, ence 145, 11–19. intact metabolic activity in requires both glutamatergic synaptic activity in neurons and glycogen Gordon, G.R., Choi, H.B., Rungta, astrocytes. metabolism and lactate release in astrocytes. The memory program in the R.L., Ellis-Davies, G.C., and MacViBy showing that glycopostsynaptic neuron triggers activation of protein kinases with subsequent car, B.A. (2008). Nature 456, phosphorylation of the nuclear transcription factor CREB and of cofilin (an genolysis and its product, 745–749. actin-binding protein), as well as rapid expression of the activity-dependent lactate, are required for megene Arc, all necessary steps for the establishment of long-term memory. In Henneberger, C., Papouin, T., Oliet, mory consolidation in vivo, the late phase, new protein synthesis leads to structural changes and formaS.H., and Rusakov, D.A. (2010). tion of new synapses. The authors’ data suggest that lactate is exported from Suzuki et al. enhance our Nature 463, 232–236. astrocytes via the monocarboxylate transporters MCT1 or MCT4 and imported understanding of the links Kandel, E.R. (2001). Science 294, into neurons via MCT2. Lactate transport is necessary for the expression of Arc between metabolic and and for the phosphorylation of CREB and cofilin. Question marks in the figure 1030–1038. memory processes. Neverhighlight the several issues that remain to be resolved: (1) the neuronal signal Pellerin, L., and Magistretti, P.J. responsible for triggering glycogen metabolism in astrocytes; (2) the specific theless, several aspects of (1994). Proc. Natl. Acad. Sci. USA location of the different lactate transporters in astrocytes and neurons; (3) the this system still await clarifi91, 10625–10629. mechanism by which lactate acts in the cascade of events leading to memory cation. For example, we consolidation. Rouach, N., Koulakoff, A., Abudara, don’t know what triggers V., Willecke, K., and Giaume, C. (2008). Science 322, 1551–1555. glycogenolysis in astrocytes, the time of the signal, and whether it Another, more general question concerns Stellwagen, D., and Malenka, R.C. (2006). Nature originates from presynaptic or postsyn- the experimental approaches required for 440, 1054–1059. aptic neuronal activity. Likewise, the clarifying the above points. For instance, Suzuki, A., Stern, S.A., Bozdagi, O., Huntley, G.W., mechanism by which lactate acts in the it is unclear whether Ca2+ imaging, the Walker, R.H., Magistretti, P.J., and Alberini, C.M. neuronal cellular cascade leading to main current approach for real-time study (2011). Cell 144, this issue, 810–823. memory consolidation and the exact of synapse-astrocyte communications, Volterra, A., and Meldolesi, J. (2005). Nat. Rev. timing of its action remain unknown. can be used to study the interplay Neurosci. 6, 626–640. Cell 144, March 4, 2011 ª2011 Elsevier Inc. 645
Leading Edge
Review Hallmarks of Cancer: The Next Generation Douglas Hanahan1,2,* and Robert A. Weinberg3,* 1The
Swiss Institute for Experimental Cancer Research (ISREC), School of Life Sciences, EPFL, Lausanne CH-1015, Switzerland Department of Biochemistry & Biophysics, UCSF, San Francisco, CA 94158, USA 3Whitehead Institute for Biomedical Research, Ludwig/MIT Center for Molecular Oncology, and MIT Department of Biology, Cambridge, MA 02142, USA *Correspondence: dh@epfl.ch (D.H.),
[email protected] (R.A.W.) DOI 10.1016/j.cell.2011.02.013 2The
The hallmarks of cancer comprise six biological capabilities acquired during the multistep development of human tumors. The hallmarks constitute an organizing principle for rationalizing the complexities of neoplastic disease. They include sustaining proliferative signaling, evading growth suppressors, resisting cell death, enabling replicative immortality, inducing angiogenesis, and activating invasion and metastasis. Underlying these hallmarks are genome instability, which generates the genetic diversity that expedites their acquisition, and inflammation, which fosters multiple hallmark functions. Conceptual progress in the last decade has added two emerging hallmarks of potential generality to this list—reprogramming of energy metabolism and evading immune destruction. In addition to cancer cells, tumors exhibit another dimension of complexity: they contain a repertoire of recruited, ostensibly normal cells that contribute to the acquisition of hallmark traits by creating the ‘‘tumor microenvironment.’’ Recognition of the widespread applicability of these concepts will increasingly affect the development of new means to treat human cancer. INTRODUCTION We have proposed that six hallmarks of cancer together constitute an organizing principle that provides a logical framework for understanding the remarkable diversity of neoplastic diseases (Hanahan and Weinberg, 2000). Implicit in our discussion was the notion that as normal cells evolve progressively to a neoplastic state, they acquire a succession of these hallmark capabilities, and that the multistep process of human tumor pathogenesis could be rationalized by the need of incipient cancer cells to acquire the traits that enable them to become tumorigenic and ultimately malignant. We noted as an ancillary proposition that tumors are more than insular masses of proliferating cancer cells. Instead, they are complex tissues composed of multiple distinct cell types that participate in heterotypic interactions with one another. We depicted the recruited normal cells, which form tumor-associated stroma, as active participants in tumorigenesis rather than passive bystanders; as such, these stromal cells contribute to the development and expression of certain hallmark capabilities. During the ensuing decade this notion has been solidified and extended, revealing that the biology of tumors can no longer be understood simply by enumerating the traits of the cancer cells but instead must encompass the contributions of the ‘‘tumor microenvironment’’ to tumorigenesis. In the course of remarkable progress in cancer research subsequent to this publication, new observations have served both to clarify and to modify the original formulation of the hallmark capabilities. In addition, yet other observations have raised questions and highlighted mechanistic concepts that were not integral to our original elaboration of the hallmark traits. Moti646 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
vated by these developments, we now revisit the original hallmarks, consider new ones that might be included in this roster, and expand upon the functional roles and contributions made by recruited stromal cells to tumor biology. HALLMARK CAPABILITIES—CONCEPTUAL PROGRESS The six hallmarks of cancer—distinctive and complementary capabilities that enable tumor growth and metastatic dissemination—continue to provide a solid foundation for understanding the biology of cancer (Figure 1; see the Supplemental Information for downloadable versions of the figures for presentations). In the first section of this Review, we summarize the essence of each hallmark as described in the original presentation in 2000, followed by selected illustrations (demarcated by subheadings in italics) of the conceptual progress made over the past decade in understanding their mechanistic underpinnings. In subsequent sections we address new developments that broaden the scope of the conceptualization, describing in turn two enabling characteristics crucial to the acquisition of the six hallmark capabilities, two new emerging hallmark capabilities, the constitution and signaling interactions of the tumor microenvironment crucial to cancer phenotypes, and we finally discuss the new frontier of therapeutic application of these concepts. Sustaining Proliferative Signaling Arguably the most fundamental trait of cancer cells involves their ability to sustain chronic proliferation. Normal tissues carefully control the production and release of growth-promoting signals that instruct entry into and progression through the cell growthand-division cycle, thereby ensuring a homeostasis of cell
Figure 1. The Hallmarks of Cancer This illustration encompasses the six hallmark capabilities originally proposed in our 2000 perspective. The past decade has witnessed remarkable progress toward understanding the mechanistic underpinnings of each hallmark.
number and thus maintenance of normal tissue architecture and function. Cancer cells, by deregulating these signals, become masters of their own destinies. The enabling signals are conveyed in large part by growth factors that bind cell-surface receptors, typically containing intracellular tyrosine kinase domains. The latter proceed to emit signals via branched intracellular signaling pathways that regulate progression through the cell cycle as well as cell growth (that is, increases in cell size); often these signals influence yet other cell-biological properties, such as cell survival and energy metabolism. Remarkably, the precise identities and sources of the proliferative signals operating within normal tissues were poorly understood a decade ago and in general remain so. Moreover, we still know relatively little about the mechanisms controlling the release of these mitogenic signals. In part, the understanding of these mechanisms is complicated by the fact that the growth factor signals controlling cell number and position within tissues are thought to be transmitted in a temporally and spatially regulated fashion from one cell to its neighbors; such paracrine signaling is difficult to access experimentally. In addition, the bioavailability of growth factors is regulated by sequestration in the pericellular space and extracellular matrix, and by the actions of a complex network of proteases, sulfatases, and possibly other enzymes that liberate and activate them, apparently in a highly specific and localized fashion. The mitogenic signaling in cancer cells is, in contrast, better understood (Lemmon and Schlessinger, 2010; Witsch et al., 2010; Hynes and MacDonald, 2009; Perona, 2006). Cancer cells can acquire the capability to sustain proliferative signaling in a number of alternative ways: They may produce growth factor ligands themselves, to which they can respond via the expression of cognate receptors, resulting in autocrine proliferative stimulation. Alternatively, cancer cells may send signals to stimulate normal cells within the supporting tumor-associated stroma, which reciprocate by supplying the cancer cells with various growth factors (Cheng et al., 2008; Bhowmick et al., 2004). Receptor signaling can also be deregulated by elevating the levels of receptor proteins displayed at the cancer cell
surface, rendering such cells hyperresponsive to otherwise-limiting amounts of growth factor ligand; the same outcome can result from structural alterations in the receptor molecules that facilitate ligand-independent firing. Growth factor independence may also derive from the constitutive activation of components of signaling pathways operating downstream of these receptors, obviating the need to stimulate these pathways by ligand-mediated receptor activation. Given that a number of distinct downstream signaling pathways radiate from a ligand-stimulated receptor, the activation of one or another of these downstream pathways, for example, the one responding to the Ras signal transducer, may only recapitulate a subset of the regulatory instructions transmitted by an activated receptor. Somatic Mutations Activate Additional Downstream Pathways High-throughput DNA sequencing analyses of cancer cell genomes have revealed somatic mutations in certain human tumors that predict constitutive activation of signaling circuits usually triggered by activated growth factor receptors. Thus, we now know that 40% of human melanomas contain activating mutations affecting the structure of the B-Raf protein, resulting in constitutive signaling through the Raf to mitogenactivated protein (MAP)-kinase pathway (Davies and Samuels 2010). Similarly, mutations in the catalytic subunit of phosphoinositide 3-kinase (PI3-kinase) isoforms are being detected in an array of tumor types, which serve to hyperactivate the PI3kinase signaling circuitry, including its key Akt/PKB signal transducer (Jiang and Liu, 2009; Yuan and Cantley, 2008). The advantages to tumor cells of activating upstream (receptor) versus downstream (transducer) signaling remain obscure, as does the functional impact of crosstalk between the multiple pathways radiating from growth factor receptors. Disruptions of Negative-Feedback Mechanisms that Attenuate Proliferative Signaling Recent results have highlighted the importance of negativefeedback loops that normally operate to dampen various types of signaling and thereby ensure homeostatic regulation of the flux of signals coursing through the intracellular circuitry (Wertz and Dixit, 2010; Cabrita and Christofori, 2008; Amit et al., 2007; Mosesson et al., 2008). Defects in these feedback mechanisms are capable of enhancing proliferative signaling. The prototype of this type of regulation involves the Ras oncoprotein: the oncogenic effects of Ras do not result from a hyperactivation of its signaling powers; instead, the oncogenic mutations affecting ras genes compromise Ras GTPase activity, which Cell 144, March 4, 2011 ª2011 Elsevier Inc. 647
operates as an intrinsic negative-feedback mechanism that normally ensures that active signal transmission is transitory. Analogous negative-feedback mechanisms operate at multiple nodes within the proliferative signaling circuitry. A prominent example involves the PTEN phosphatase, which counteracts PI3-kinase by degrading its product, phosphatidylinositol (3,4,5) trisphosphate (PIP3). Loss-of-function mutations in PTEN amplify PI3K signaling and promote tumorigenesis in a variety of experimental models of cancer; in human tumors, PTEN expression is often lost by promoter methylation (Jiang and Liu, 2009; Yuan and Cantley, 2008). Yet another example involves the mTOR kinase, a coordinator of cell growth and metabolism that lies both upstream and downstream of the PI3K pathway. In the circuitry of some cancer cells, mTOR activation results, via negative feedback, in the inhibition of PI3K signaling. Thus, when mTOR is pharmacologically inhibited in such cancer cells (such as by the drug rapamycin), the associated loss of negative feedback results in increased activity of PI3K and its effector Akt/PKB, thereby blunting the antiproliferative effects of mTOR inhibition (Sudarsanam and Johnson, 2010; O’Reilly et al., 2006). It is likely that compromised negative-feedback loops in this and other signaling pathways will prove to be widespread among human cancer cells and serve as an important means by which these cells can achieve proliferative independence. Moreover, disruption of such selfattenuating signaling may contribute to the development of adaptive resistance toward drugs targeting mitogenic signaling. Excessive Proliferative Signaling Can Trigger Cell Senescence Early studies of oncogene action encouraged the notion that ever-increasing expression of such genes and the signals manifested in their protein products would result in correspondingly increased cancer cell proliferation and thus tumor growth. More recent research has undermined this notion, in that excessively elevated signaling by oncoproteins such as RAS, MYC, and RAF can provoke counteracting responses from cells, specifically induction of cell senescence and/or apoptosis (Collado and Serrano, 2010; Evan and d’Adda di Fagagna, 2009; Lowe et al., 2004). For example, cultured cells expressing high levels of the Ras oncoprotein may enter into the nonproliferative but viable state called senescence; in contrast, cells expressing lower levels of this protein may avoid senescence and proliferate. Cells with morphological features of senescence, including enlarged cytoplasm, the absence of proliferation markers, and expression of the senescence-induced b-galactosidase enzyme, are abundant in the tissues of mice engineered to overexpress certain oncogenes (Collado and Serrano, 2010; Evan and d’Adda di Fagagna, 2009) and are prevalent in some cases of human melanoma (Mooi and Peeper, 2006). These ostensibly paradoxical responses seem to reflect intrinsic cellular defense mechanisms designed to eliminate cells experiencing excessive levels of certain types of signaling. Accordingly, the relative intensity of oncogenic signaling in cancer cells may represent compromises between maximal mitogenic stimulation and avoidance of these antiproliferative defenses. Alternatively, some cancer cells may adapt to high levels of oncogenic signaling by disabling their senescence- or apoptosis-inducing circuitry. 648 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
Evading Growth Suppressors In addition to the hallmark capability of inducing and sustaining positively acting growth-stimulatory signals, cancer cells must also circumvent powerful programs that negatively regulate cell proliferation; many of these programs depend on the actions of tumor suppressor genes. Dozens of tumor suppressors that operate in various ways to limit cell growth and proliferation have been discovered through their characteristic inactivation in one or another form of animal or human cancer; many of these genes have been validated as bona fide tumor suppressors through gain- or loss-of-function experiments in mice. The two prototypical tumor suppressors encode the RB (retinoblastoma-associated) and TP53 proteins; they operate as central control nodes within two key complementary cellular regulatory circuits that govern the decisions of cells to proliferate or, alternatively, activate senescence and apoptotic programs. The RB protein integrates signals from diverse extracellular and intracellular sources and, in response, decides whether or not a cell should proceed through its growth-and-division cycle (Burkhart and Sage, 2008; Deshpande et al., 2005; Sherr and McCormick, 2002). Cancer cells with defects in RB pathway function are thus missing the services of a critical gatekeeper of cell-cycle progression whose absence permits persistent cell proliferation. Whereas RB transduces growth-inhibitory signals that originate largely outside of the cell, TP53 receives inputs from stress and abnormality sensors that function within the cell’s intracellular operating systems: if the degree of damage to the genome is excessive, or if the levels of nucleotide pools, growth-promoting signals, glucose, or oxygenation are suboptimal, TP53 can call a halt to further cell-cycle progression until these conditions have been normalized. Alternatively, in the face of alarm signals indicating overwhelming or irreparable damage to such cellular subsystems, TP53 can trigger apoptosis. Notably, the various effects of activated TP53 are complex and highly context dependent, varying by cell type as well as by the severity and persistence of conditions of cell stress and genomic damage. Although the two canonical suppressors of proliferation— TP53 and RB—have preeminent importance in regulating cell proliferation, various lines of evidence indicate that each operates as part of a larger network that is wired for functional redundancy. For example, chimeric mice populated throughout their bodies with individual cells lacking a functional Rb gene are surprisingly free of proliferative abnormalities, despite the expectation that loss of RB function would allow continuous firing of the cell division cycle in these cells and their lineal descendants; some of the resulting clusters of Rb null cells should, by all rights, progress to neoplasia. Instead, the Rb null cells in such chimeric mice have been found to participate in relatively normal tissue morphogenesis throughout the body; the only neoplasia observed was in the development of pituitary tumors late in life (Lipinski and Jacks, 1999). Similarly, TP53 null mice develop normally, show largely proper cell and tissue homeostasis, and again develop abnormalities later in life, in the form of leukemias and sarcomas (Ghebranious and Donehower, 1998). Both examples must reflect the operations of redundantly acting mechanisms that serve to constrain inappropriate replication of cells lacking these key proliferation suppressors.
Mechanisms of Contact Inhibition and Its Evasion Four decades of research have demonstrated that the cell-tocell contacts formed by dense populations of normal cells propagated in two-dimensional culture operate to suppress further cell proliferation, yielding confluent cell monolayers. Importantly, such ‘‘contact inhibition’’ is abolished in various types of cancer cells in culture, suggesting that contact inhibition is an in vitro surrogate of a mechanism that operates in vivo to ensure normal tissue homeostasis, one that is abrogated during the course of tumorigenesis. Until recently, the mechanistic basis for this mode of growth control remained obscure. Now, however, mechanisms of contact inhibition are beginning to emerge. One mechanism involves the product of the NF2 gene, long implicated as a tumor suppressor because its loss triggers a form of human neurofibromatosis. Merlin, the cytoplasmic NF2 gene product, orchestrates contact inhibition via coupling cell-surface adhesion molecules (e.g., E-cadherin) to transmembrane receptor tyrosine kinases (e.g., the EGF receptor). In so doing, Merlin strengthens the adhesivity of cadherin-mediated cell-to-cell attachments. Additionally, by sequestering growth factor receptors, Merlin limits their ability to efficiently emit mitogenic signals (Curto et al., 2007; Okada et al., 2005). A second mechanism of contact inhibition involves the LKB1 epithelial polarity protein, which organizes epithelial structure and helps maintain tissue integrity. LKB1 can, for example, overrule the mitogenic effects of the powerful Myc oncogene when the latter is upregulated in organized, quiescent epithelial structures; in contrast, when LKB1 expression is suppressed, epithelial integrity is destabilized, and epithelial cells become susceptible to Myc-induced transformation (Partanen et al., 2009; Hezel and Bardeesy, 2008). LKB1 has also been identified as a tumor suppressor gene that is lost in certain human malignancies (Shaw, 2009), possibly reflecting its normal function as a suppressor of inappropriate proliferation. It remains to be seen how frequently these two mechanisms of contact-mediated growth suppression are compromised in human cancers; no doubt yet other contact-induced proliferative barriers are yet to be discovered. Clearly mechanisms like these that enable cells to construct and maintain architecturally complex tissues represent important means of suppressing and counterbalancing inappropriate proliferative signals. Corruption of the TGF-b Pathway Promotes Malignancy TGF-b is best known for its antiproliferative effects, and evasion by cancer cells of these effects is now appreciated to be far more elaborate than simple shutdown of its signaling circuitry (Ikushima and Miyazono, 2010; Massague´, 2008; Bierie and Moses, 2006). In many late-stage tumors, TGF-b signaling is redirected away from suppressing cell proliferation and is found instead to activate a cellular program, termed the epithelial-to-mesenchymal transition (EMT), that confers on cancer cells traits associated with high-grade malignancy, as discussed in further detail below. Resisting Cell Death The concept that programmed cell death by apoptosis serves as a natural barrier to cancer development has been established by compelling functional studies conducted over the last two decades (Adams and Cory, 2007; Lowe et al., 2004: Evan and
Littlewood, 1998). Elucidation of the signaling circuitry governing the apoptotic program has revealed how apoptosis is triggered in response to various physiologic stresses that cancer cells experience during the course of tumorigenesis or as a result of anticancer therapy. Notable among the apoptosis-inducing stresses are signaling imbalances resulting from elevated levels of oncogene signaling, as mentioned earlier, and DNA damage associated with hyperproliferation. Yet other research has revealed how apoptosis is attenuated in those tumors that succeed in progressing to states of high-grade malignancy and resistance to therapy (Adams and Cory, 2007; Lowe et al., 2004). The apoptotic machinery is composed of both upstream regulators and downstream effector components (Adams and Cory, 2007). The regulators, in turn, are divided into two major circuits, one receiving and processing extracellular death-inducing signals (the extrinsic apoptotic program, involving for example the Fas ligand/Fas receptor), and the other sensing and integrating a variety of signals of intracellular origin (the intrinsic program). Each culminates in activation of a normally latent protease (caspases 8 and 9, respectively), which proceeds to initiate a cascade of proteolysis involving effector caspases responsible for the execution phase of apoptosis, in which the cell is progressively disassembled and then consumed, both by its neighbors and by professional phagocytic cells. Currently, the intrinsic apoptotic program is more widely implicated as a barrier to cancer pathogenesis. The ‘‘apoptotic trigger’’ that conveys signals between the regulators and effectors is controlled by counterbalancing pro- and antiapoptotic members of the Bcl-2 family of regulatory proteins (Adams and Cory, 2007). The archetype, Bcl-2, along with its closest relatives (Bcl-xL, Bcl-w, Mcl-1, A1) are inhibitors of apoptosis, acting in large part by binding to and thereby suppressing two proapoptotic triggering proteins (Bax and Bak); the latter are embedded in the mitochondrial outer membrane. When relieved of inhibition by their antiapoptotic relatives, Bax and Bak disrupt the integrity of the outer mitochondrial membrane, causing the release of proapoptotic signaling proteins, the most important of which is cytochrome c. The released cytochrome c activates, in turn, a cascade of caspases that act via their proteolytic activities to induce the multiple cellular changes associated with the apoptotic program. Bax and Bak share protein-protein interaction domains, termed BH3 motifs, with the antiapoptotic Bcl-2-like proteins that mediate their various physical interactions. The activities of a subfamily of related proteins, each of which contains a single such BH3 motif, are coupled to a variety of sensors of cellular abnormality; these ‘‘BH3-only’’ proteins act either by interfering with antiapoptotic Bcl-2 proteins or by directly stimulating the proapoptotic members of this family (Adams and Cory, 2007; Willis and Adams, 2005). Although the cellular conditions that trigger apoptosis remain to be fully enumerated, several abnormality sensors that play key roles in tumor development have been identified (Adams and Cory, 2007; Lowe et al., 2004). Most notable is a DNAdamage sensor that functions via the TP53 tumor suppressor (Junttila and Evan, 2009); TP53 induces apoptosis by upregulating expression of the Noxa and Puma BH3-only proteins, doing so in response to substantial levels of DNA breaks and other chromosomal abnormalities. Alternatively, insufficient survival Cell 144, March 4, 2011 ª2011 Elsevier Inc. 649
factor signaling (for instance inadequate levels of interleukin-3 in lymphocytes or of insulin-like growth factor 1/2 [Igf1/2] in epithelial cells) can elicit apoptosis through a BH3-only protein called Bim. Yet another condition leading to cell death involves hyperactive signaling by certain oncoproteins, such as Myc, which triggers apoptosis (in part via Bim and other BH3-only proteins) unless counterbalanced by antiapoptotic factors (Junttila and Evan, 2009; Lowe et al., 2004). Tumor cells evolve a variety of strategies to limit or circumvent apoptosis. Most common is the loss of TP53 tumor suppressor function, which eliminates this critical damage sensor from the apoptosis-inducing circuitry. Alternatively, tumors may achieve similar ends by increasing expression of antiapoptotic regulators (Bcl-2, Bcl-xL) or of survival signals (Igf1/2), by downregulating proapoptotic factors (Bax, Bim, Puma), or by short-circuiting the extrinsic ligand-induced death pathway. The multiplicity of apoptosis-avoiding mechanisms presumably reflects the diversity of apoptosis-inducing signals that cancer cell populations encounter during their evolution to the malignant state. The structure of the apoptotic machinery and program, and the strategies used by cancer cells to evade its actions, were widely appreciated by the beginning of the last decade. The most notable conceptual advances since then have involved other forms of cell death that broaden the scope of ‘‘programmed cell death’’ as a barrier to cancer. Autophagy Mediates Both Tumor Cell Survival and Death Autophagy represents an important cell-physiologic response that, like apoptosis, normally operates at low, basal levels in cells but can be strongly induced in certain states of cellular stress, the most obvious of which is nutrient deficiency (Levine and Kroemer, 2008; Mizushima, 2007). The autophagic program enables cells to break down cellular organelles, such as ribosomes and mitochondria, allowing the resulting catabolites to be recycled and thus used for biosynthesis and energy metabolism. As part of this program, intracellular vesicles termed autophagosomes envelope intracellular organelles and then fuse with lysosomes wherein degradation occurs. In this fashion, lowmolecular-weight metabolites are generated that support survival in the stressed, nutrient-limited environments experienced by many cancer cells. Like apoptosis, the autophagy machinery has both regulatory and effector components (Levine and Kroemer, 2008; Mizushima, 2007). Among the latter are proteins that mediate autophagosome formation and delivery to lysosomes. Of note, recent research has revealed intersections between the regulatory circuits governing autophagy, apoptosis, and cellular homeostasis. For example, the signaling pathway involving the PI3kinase, AKT, and mTOR kinases, which is stimulated by survival signals to block apoptosis, similarly inhibits autophagy; when survival signals are insufficient, the PI3K signaling pathway is downregulated, with the result that autophagy and/or apoptosis may be induced (Levine and Kroemer, 2008; Sinha and Levine, 2008; Mathew et al., 2007). Another interconnection between these two programs resides in the Beclin-1 protein, which has been shown by genetic studies to be necessary for induction of autophagy (Levine and Kroemer, 2008; Sinha and Levine, 2008; Mizushima, 2007). Beclin-1 is a member of the BH3-only subfamily of apoptotic regulatory 650 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
proteins, and its BH3 domain allows it to bind the Bcl-2/Bcl-xL proteins. Stress-sensor-coupled BH3 proteins can displace Beclin-1 from its association with Bcl-2/Bcl-xL, enabling the liberated Beclin-1 to trigger autophagy, much as they can release proapoptotic Bax and Bak to trigger apoptosis. Hence, stresstransducing BH3 proteins (e.g., Bid, Bad, Puma, et al.) can induce apoptosis and/or autophagy depending on the physiologic state of the cell. Mice bearing inactivated alleles of the Beclin-1 gene or of certain other components of the autophagy machinery exhibit increased susceptibility to cancer (White and DiPaola, 2009: Levine and Kroemer, 2008). These results suggest that induction of autophagy can serve as a barrier to tumorigenesis that may operate independently of or in concert with apoptosis. Accordingly, autophagy appears to represent yet another barrier that needs to be circumvented during tumor development (White and DiPaola, 2009). Perhaps paradoxically, nutrient starvation, radiotherapy, and certain cytotoxic drugs can induce elevated levels of autophagy that are apparently cytoprotective for cancer cells, impairing rather than accentuating the killing actions of these stressinducing situations (White and DiPaola, 2009; Apel et al., 2009; Amaravadi and Thompson, 2007; Mathew et al., 2007). Moreover, severely stressed cancer cells have been shown to shrink via autophagy to a state of reversible dormancy (White and DiPaola, 2009; Lu et al., 2008). This survival response may enable the persistence and eventual regrowth of some latestage tumors following treatment with potent anticancer agents. Thus, in analogy to TGF-b signaling, which can be tumor suppressing at early stages of tumorigenesis and tumor promoting later on, autophagy seems to have conflicting effects on tumor cells and thus tumor progression (Apel et al., 2009; White and DiPaola, 2009). An important agenda for future research will involve clarifying the genetic and cell-physiologic conditions that dictate when and how autophagy enables cancer cells to survive or causes them to die. Necrosis Has Proinflammatory and Tumor-Promoting Potential In contrast to apoptosis, in which a dying cell contracts into an almost-invisible corpse that is soon consumed by neighbors, necrotic cells become bloated and explode, releasing their contents into the local tissue microenvironment. Although necrosis has historically been viewed much like organismic death, as a form of system-wide exhaustion and breakdown, the conceptual landscape is changing: cell death by necrosis is clearly under genetic control in some circumstances, rather than being a random and undirected process (Galluzzi and Kroemer, 2008; Zong and Thompson, 2006). Perhaps more important, necrotic cell death releases proinflammatory signals into the surrounding tissue microenvironment, in contrast to apoptosis and autophagy, which do not. As a consequence, necrotic cells can recruit inflammatory cells of the immune system (Grivennikov et al., 2010; White et al., 2010; Galluzzi and Kroemer, 2008), whose dedicated function is to survey the extent of tissue damage and remove associated necrotic debris. In the context of neoplasia, however, multiple lines of evidence indicate that immune inflammatory cells can be actively tumor promoting, given that such cells are capable
of fostering angiogenesis, cancer cell proliferation, and invasiveness (see below). Additionally, necrotic cells can release bioactive regulatory factors, such as IL-1a, which can directly stimulate neighboring viable cells to proliferate, with the potential, once again, to facilitate neoplastic progression (Grivennikov et al., 2010). Consequently, necrotic cell death, while seemingly beneficial in counterbalancing cancer-associated hyperproliferation, may ultimately do more damage than good. Accordingly, incipient neoplasias and potentially invasive and metastatic tumors may gain an advantage by tolerating some degree of necrotic cell death, doing so in order to recruit tumor-promoting inflammatory cells that bring growth-stimulating factors to the surviving cells within these growths. Enabling Replicative Immortality By 2000, it was widely accepted that cancer cells require unlimited replicative potential in order to generate macroscopic tumors. This capability stands in marked contrast to the behavior of the cells in most normal cell lineages in the body, which are able to pass through only a limited number of successive cell growth-and-division cycles. This limitation has been associated with two distinct barriers to proliferation: senescence, a typically irreversible entrance into a nonproliferative but viable state, and crisis, which involves cell death. Accordingly, when cells are propagated in culture, repeated cycles of cell division lead first to induction of senescence and then, for those cells that succeed in circumventing this barrier, to a crisis phase, in which the great majority of cells in the population die. On rare occasion, cells emerge from a population in crisis and exhibit unlimited replicative potential. This transition has been termed immortalization, a trait that most established cell lines possess by virtue of their ability to proliferate in culture without evidence of either senescence or crisis. Multiple lines of evidence indicate that telomeres protecting the ends of chromosomes are centrally involved in the capability for unlimited proliferation (Blasco, 2005; Shay and Wright, 2000). The telomeres, composed of multiple tandem hexanucleotide repeats, shorten progressively in nonimmortalized cells propagated in culture, eventually losing the ability to protect the ends of chromosomal DNAs from end-to-end fusions; such fusions generate unstable dicentric chromosomes whose resolution results in a scrambling of karyotype that threatens cell viability. Accordingly, the length of telomeric DNA in a cell dictates how many successive cell generations its progeny can pass through before telomeres are largely eroded and have consequently lost their protective functions, triggering entrance into crisis. Telomerase, the specialized DNA polymerase that adds telomere repeat segments to the ends of telomeric DNA, is almost absent in nonimmortalized cells but expressed at functionally significant levels in the vast majority (90%) of spontaneously immortalized cells, including human cancer cells. By extending telomeric DNA, telomerase is able to counter the progressive telomere erosion that would otherwise occur in its absence. The presence of telomerase activity, either in spontaneously immortalized cells or in the context of cells engineered to express the enzyme, is correlated with a resistance to induction of both senescence and crisis/apoptosis; conversely, suppres-
sion of telomerase activity leads to telomere shortening and to activation of one or the other of these proliferative barriers. The two barriers to proliferation—senescence and crisis/ apoptosis—have been rationalized as crucial anticancer defenses that are hard-wired into our cells, being deployed to impede the outgrowth of clones of preneoplastic and frankly neoplastic cells. According to this thinking, most incipient neoplasias exhaust their endowment of replicative doublings and are stopped in their tracks by one or the other of these barriers. The eventual immortalization of rare variant cells that proceed to form tumors has been attributed to their ability to maintain telomeric DNA at lengths sufficient to avoid triggering senescence or apoptosis, achieved most commonly by upregulating expression of telomerase or, less frequently, via an alternative recombination-based telomere maintenance mechanism. Hence, telomere shortening has come to be viewed as a clocking device that determines the limited replicative potential of normal cells and thus one that must be overcome by cancer cells. Reassessing Replicative Senescence Whereas telomere maintenance has been increasingly substantiated as a condition critical to the neoplastic state, the concept of replication-induced senescence as a general barrier requires refinement and reformulation. (Differences in telomere structure and function in mouse versus human cells have also complicated investigation of the roles of telomeres and telomerase in replicative senescence.) Recent experiments have revealed that the induction of senescence in certain cultured cells can be delayed and possibly eliminated by the use of improved cell culture conditions, suggesting that recently explanted primary cells may be able to proliferate unimpeded in culture up the point of crisis and the associated induction of apoptosis triggered by critically shortened telomeres (Ince et al., 2007; Passos et al., 2007; Zhang et al., 2004; Sherr and DePinho, 2000). In contrast, experiments in mice engineered to lack telomerase indicate that the consequently shortened telomeres can shunt premalignant cells into a senescent state that contributes (along with apoptosis) to attenuated tumorigenesis in mice genetically destined to develop particular forms of cancer (Artandi and DePinho, 2010). Such telomerase null mice with highly eroded telomeres exhibit multiorgan dysfunction and abnormalities that include evidence for both senescence and apoptosis, perhaps analogous to the senescence and apoptosis observed in cell culture (Artandi and DePinho, 2010; Feldser and Greider, 2007). Of note, and as discussed earlier, a morphologically similar form of cell senescence induced by excessive or unbalanced oncogene signaling is now well documented as a protective mechanism against neoplasia; the possible interconnections of this form of senescence with telomerase and telomeres remain to be ascertained. Thus, cell senescence is emerging conceptually as a protective barrier to neoplastic expansion that can be triggered by various proliferation-associated abnormalities, including high levels of oncogenic signaling and, apparently, subcritical shortening of telomeres. Delayed Activation of Telomerase May Both Limit and Foster Neoplastic Progression There is now evidence that clones of incipient cancer cells often experience telomere loss-induced crisis relatively early during Cell 144, March 4, 2011 ª2011 Elsevier Inc. 651
the course of multistep tumor progression due to their inability to express significant levels of telomerase. Thus, extensively eroded telomeres have been documented in premalignant growths through the use of fluorescence in situ hybridization (FISH), which has also revealed the end-to-end chromosomal fusions that signal telomere failure and crisis (Kawai et al., 2007; Hansel et al., 2006). These results also suggest that such cells have passed through a substantial number of successive telomere-shortening cell divisions during their evolution from fully normal cells-of-origin. Accordingly, the development of some human neoplasias may be aborted by telomere-induced crisis long before they succeed in becoming macroscopic, frankly neoplastic growths. In contrast, the absence of TP53-mediated surveillance of genomic integrity may permit other incipient neoplasias to survive initial telomere erosion and attendant chromosomal breakage-fusion-bridge (BFB) cycles. The genomic alterations resulting from these BFB cycles, including deletions and amplifications of chromosomal segments, evidently serve to increase the mutability of the genome, thereby accelerating the acquisition of mutant oncogenes and tumor suppressor genes. The realization that impaired telomere function can actually foster tumor progression has come from the study of mutant mice that lack both p53 and telomerase function (Artandi and DePinho, 2010, 2000). The proposition that these two defects can cooperatively enhance human tumorigenesis has not yet been directly documented. Circumstantial support for the importance of transient telomere deficiency in facilitating malignant progression has come, in addition, from comparative analyses of premalignant and malignant lesions in the human breast (Raynaud et al., 2010; Chin et al., 2004). The premalignant lesions did not express significant levels of telomerase and were marked by telomere shortening and nonclonal chromosomal aberrations. In contrast, overt carcinomas exhibited telomerase expression concordantly with the reconstruction of longer telomeres and the fixation (via clonal outgrowth) of the aberrant karyotypes that would seem to have been acquired after telomere failure but before the acquisition of telomerase activity. When portrayed in this way, the delayed acquisition of telomerase function serves to generate tumor-promoting mutations, whereas its subsequent activation stabilizes the mutant genome and confers the unlimited replicative capacity that cancer cells require in order to generate clinically apparent tumors. New Functions of Telomerase Telomerase was discovered because of its ability to elongate and maintain telomeric DNA, and almost all telomerase research has been posited on the notion that its functions are confined to this crucial function. However, in recent years it has become apparent that telomerase exerts functions that are relevant to cell proliferation but unrelated to telomere maintenance. The noncanonical roles of telomerase, and in particular its protein subunit TERT, have been revealed by functional studies in mice and cultured cells; in some cases novel functions have been demonstrated in conditions where the telomerase enzymatic activity has been eliminated (Cong and Shay, 2008). Among the growing list of telomere-independent functions of TERT/telomerase is the ability of TERT to amplify signaling by 652 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
the Wnt pathway, by serving as a cofactor of the b-catenin/LEF transcription factor complex (Park et al., 2009). Other ascribed telomere-independent effects include demonstrable enhancement of cell proliferation and/or resistance to apoptosis (Kang et al., 2004), involvement in DNA-damage repair (Masutomi et al., 2005), and RNA-dependent RNA polymerase function (Maida et al., 2009). Consistent with these broader roles, TERT can be found associated with chromatin at multiple sites along the chromosomes, not just at the telomeres (Park et al., 2009; Masutomi et al., 2005). Hence, telomere maintenance is proving to be the most prominent of a diverse series of functions to which TERT contributes. The contributions of these additional functions of telomerase to tumorigenesis remain to be fully elucidated. Inducing Angiogenesis Like normal tissues, tumors require sustenance in the form of nutrients and oxygen as well as an ability to evacuate metabolic wastes and carbon dioxide. The tumor-associated neovasculature, generated by the process of angiogenesis, addresses these needs. During embryogenesis, the development of the vasculature involves the birth of new endothelial cells and their assembly into tubes (vasculogenesis) in addition to the sprouting (angiogenesis) of new vessels from existing ones. Following this morphogenesis, the normal vasculature becomes largely quiescent. In the adult, as part of physiologic processes such as wound healing and female reproductive cycling, angiogenesis is turned on, but only transiently. In contrast, during tumor progression, an ‘‘angiogenic switch’’ is almost always activated and remains on, causing normally quiescent vasculature to continually sprout new vessels that help sustain expanding neoplastic growths (Hanahan and Folkman, 1996). A compelling body of evidence indicates that the angiogenic switch is governed by countervailing factors that either induce or oppose angiogenesis (Baeriswyl and Christofori, 2009; Bergers and Benjamin, 2003). Some of these angiogenic regulators are signaling proteins that bind to stimulatory or inhibitory cellsurface receptors displayed by vascular endothelial cells. The well-known prototypes of angiogenesis inducers and inhibitors are vascular endothelial growth factor-A (VEGF-A) and thrombospondin-1 (TSP-1), respectively. The VEGF-A gene encodes ligands that are involved in orchestrating new blood vessel growth during embryonic and postnatal development, and then in homeostatic survival of endothelial cells, as well as in physiological and pathological situations in the adult. VEGF signaling via three receptor tyrosine kinases (VEGFR-1–3) is regulated at multiple levels, reflecting this complexity of purpose. Thus, VEGF gene expression can by upregulated both by hypoxia and by oncogene signaling (Ferrara, 2009; Mac Gabhann and Popel, 2008; Carmeliet, 2005). Additionally, VEGF ligands can be sequestered in the extracellular matrix in latent forms that are subject to release and activation by extracellular matrix-degrading proteases (e.g., MMP-9; Kessenbrock et al., 2010). In addition, other proangiogenic signals, such as members of the fibroblast growth factor (FGF) family, have been implicated in sustaining tumor angiogenesis when their expression is chronically upregulated (Baeriswyl and Christofori, 2009). TSP-1, a key counterbalance in the
angiogenic switch, also binds transmembrane receptors displayed by endothelial cells and thereby evokes suppressive signals that can counteract proangiogenic stimuli (Kazerounian et al., 2008). The blood vessels produced within tumors by chronically activated angiogenesis and an unbalanced mix of proangiogenic signals are typically aberrant: tumor neovasculature is marked by precocious capillary sprouting, convoluted and excessive vessel branching, distorted and enlarged vessels, erratic blood flow, microhemorrhaging, leakiness, and abnormal levels of endothelial cell proliferation and apoptosis (Nagy et al., 2010; Baluk et al., 2005). Angiogenesis is induced surprisingly early during the multistage development of invasive cancers both in animal models and in humans. Histological analyses of premalignant, noninvasive lesions, including dysplasias and in situ carcinomas arising in a variety of organs, have revealed the early tripping of the angiogenic switch (Raica et al., 2009; Hanahan and Folkman, 1996). Historically, angiogenesis was envisioned to be important only when rapidly growing macroscopic tumors had formed, but more recent data indicate that angiogenesis also contributes to the microscopic premalignant phase of neoplastic progression, further cementing its status as an integral hallmark of cancer. The past decade has witnessed an astonishing outpouring of research on angiogenesis. Amid this wealth of new knowledge, we highlight several advances of particular relevance to tumor physiology. Gradations of the Angiogenic Switch Once angiogenesis has been activated, tumors exhibit diverse patterns of neovascularization. Some tumors, including such highly aggressive types as pancreatic ductal adenocarcinomas, are hypovascularized and replete with stromal ‘‘deserts’’ that are largely avascular and indeed may even be actively antiangiogenic (Olive et al., 2009). Many other tumors, including human renal and pancreatic neuroendocrine carcinomas, are highly angiogenic and consequently densely vascularized (Zee et al., 2010; Turner et al., 2003). Collectively, such observations suggest an initial tripping of the angiogenic switch during tumor development that is followed by a variable intensity of ongoing neovascularization, the latter being controlled by a complex biological rheostat that involves both the cancer cells and the associated stromal microenvironment (Baeriswyl and Christofori, 2009; Bergers and Benjamin, 2003). Of note, the switching mechanism can vary in its form, even though the net result is a common inductive signal (e.g., VEGF). In some tumors, dominant oncogenes operating within tumor cells, such as Ras and Myc, can upregulate expression of angiogenic factors, whereas in others, such inductive signals are produced indirectly by immune inflammatory cells, as discussed below. The direct induction of angiogenesis by oncogenes that also drive proliferative signaling illustrates the important principle that distinct hallmark capabilities can be coregulated by the same transforming agents. Endogenous Angiogenesis Inhibitors Present Natural Barriers to Tumor Angiogenesis Research in the 1990s revealed that TSP-1 as well as fragments of plasmin (angiostatin) and type 18 collagen (endostatin) can act as endogenous inhibitors of angiogenesis (Ribatti, 2009;
Kazerounian, et al., 2008; Folkman, 2006, 2002; Nyberg et al., 2005). The last decade has seen reports of another dozen such agents (Ribatti, 2009; Folkman, 2006; Nyberg et al., 2005). Most are proteins, and many are derived by proteolytic cleavage of structural proteins that are not themselves angiogenic regulators. A number of these endogenous inhibitors of angiogenesis can be detected in the circulation of normal mice and humans. The genes encoding several endogenous angiogenesis inhibitors have been deleted from the mouse germline without untoward physiological effects; the growth of autochthonous and implanted tumors, however, is enhanced as a consequence (Ribatti, 2009; Nyberg et al., 2005). By contrast, if the circulating levels of an endogenous inhibitor are genetically increased (e.g., via overexpression in transgenic mice or in xenotransplanted tumors), tumor growth is impaired (Ribatti, 2009; Nyberg et al., 2005); interestingly, wound healing and fat deposition are impaired or accelerated by elevated or ablated expression of such genes (Cao, 2010; Seppinen et al., 2008). The data suggest that such endogenous angiogenesis inhibitors serve under normal circumstances as physiologic regulators that modulate transitory angiogenesis during tissue remodeling and wound healing; they may also act as intrinsic barriers to induction and/or persistence of angiogenesis by incipient neoplasias. Pericytes Are Important Components of the Tumor Neovasculature Pericytes have long been known as supporting cells that are closely apposed to the outer surfaces of the endothelial tubes in normal tissue vasculature, where they provide important mechanical and physiologic support to the endothelial cells. Tumor-associated vasculature, in contrast, was portrayed as lacking appreciable coverage by these auxiliary cells. However, careful microscopic studies conducted in recent years have revealed that pericytes are associated, albeit loosely, with the neovasculature of most if not all tumors (Raza et al., 2010; Bergers and Song, 2005). More importantly, mechanistic studies discussed below have revealed that pericyte coverage is important for the maintenance of a functional tumor neovasculature. A Variety of Bone Marrow-Derived Cells Contribute to Tumor Angiogenesis It is now clear that a repertoire of cell types originating in the bone marrow play crucial roles in pathological angiogenesis (Qian and Pollard, 2010; Zumsteg and Christofori, 2009; Murdoch et al., 2008; De Palma et al., 2007). These include cells of the innate immune system—notably macrophages, neutrophils, mast cells, and myeloid progenitors—that infiltrate premalignant lesions and progressed tumors and assemble at the margins of such lesions; the peri-tumoral inflammatory cells help to trip the angiogenic switch in previously quiescent tissue and to sustain ongoing angiogenesis associated with tumor growth, in addition to facilitating local invasion, as noted below. In addition, they can help protect the vasculature from the effects of drugs targeting endothelial cell signaling (Ferrara, 2010). Additionally, several types of bone marrow-derived ‘‘vascular progenitor cells’’ have been observed in certain cases to have migrated into neoplastic lesions and become intercalated into the neovasculature as pericytes or endothelial cells (Patenaude et al., 2010; Kovacic and Boehm, 2009; Lamagna and Bergers, 2006). Cell 144, March 4, 2011 ª2011 Elsevier Inc. 653
Activating Invasion and Metastasis In 2000, the mechanisms underlying invasion and metastasis were largely an enigma. It was clear that as carcinomas arising from epithelial tissues progressed to higher pathological grades of malignancy, reflected in local invasion and distant metastasis, the associated cancer cells typically developed alterations in their shape as well as in their attachment to other cells and to the extracellular matrix (ECM). The best characterized alteration involved the loss by carcinoma cells of E-cadherin, a key cell-tocell adhesion molecule. By forming adherens junctions with adjacent epithelial cells, E-cadherin helps to assemble epithelial cell sheets and maintain the quiescence of the cells within these sheets. Increased expression of E-cadherin was well established as an antagonist of invasion and metastasis, whereas reduction of its expression was known to potentiate these phenotypes. The frequently observed downregulation and occasional mutational inactivation of E-cadherin in human carcinomas provided strong support for its role as a key suppressor of this hallmark capability (Berx and van Roy, 2009; Cavallaro and Christofori, 2004). Additionally, expression of genes encoding other cell-to-cell and cell-to-ECM adhesion molecules is demonstrably altered in some highly aggressive carcinomas, with those favoring cytostasis typically being downregulated. Conversely, adhesion molecules normally associated with the cell migrations that occur during embryogenesis and inflammation are often upregulated. For example, N-cadherin, which is normally expressed in migrating neurons and mesenchymal cells during organogenesis, is upregulated in many invasive carcinoma cells. Beyond the gain and loss of such cell-cell/matrix attachment proteins, the master regulators of invasion and metastasis were largely unknown or, when suspected, lacking in functional validation (Cavallaro and Christofori, 2004). The multistep process of invasion and metastasis has been schematized as a sequence of discrete steps, often termed the invasion-metastasis cascade (Talmadge and Fidler, 2010; Fidler, 2003). This depiction envisions a succession of cell-biologic changes, beginning with local invasion, then intravasation by cancer cells into nearby blood and lymphatic vessels, transit of cancer cells through the lymphatic and hematogenous systems, followed by escape of cancer cells from the lumina of such vessels into the parenchyma of distant tissues (extravasation), the formation of small nodules of cancer cells (micrometastases), and finally the growth of micrometastatic lesions into macroscopic tumors, this last step being termed ‘‘colonization.’’ Research into the capability for invasion and metastasis has accelerated dramatically over the past decade as powerful new research tools and refined experimental models have become available, and as critical regulatory genes were identified. While still an emerging field replete with major unanswered questions, significant progress has been made in delineating important features of this complex hallmark capability. An admittedly incomplete representation of these advances is highlighted below. The EMT Program Broadly Regulates Invasion and Metastasis A developmental regulatory program, referred to as the ‘‘epithelial-mesenchymal transition’’ (EMT), has become prominently implicated as a means by which transformed epithelial cells 654 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
can acquire the abilities to invade, to resist apoptosis, and to disseminate (Klymkowsky and Savagner, 2009; Polyak and Weinberg, 2009; Thiery et al., 2009; Yilmaz and Christofori, 2009; Barrallo-Gimeno and Nieto, 2005). By co-opting a process involved in various steps of embryonic morphogenesis and wound healing, carcinoma cells can concomitantly acquire multiple attributes that enable invasion and metastasis. This multifaceted EMT program can be activated transiently or stably, and to differing degrees, by carcinoma cells during the course of invasion and metastasis. A set of pleiotropically acting transcriptional factors, including Snail, Slug, Twist, and Zeb1/2, orchestrate the EMT and related migratory processes during embryogenesis; most were initially identified by developmental genetics. These transcriptional regulators are expressed in various combinations in a number of malignant tumor types and have been shown in experimental models of carcinoma formation to be causally important for programming invasion; some have been found to elicit metastasis when ectopically overexpressed (Micalizzi et al., 2010; Taube et al., 2010; Schmalhofer et al., 2009; Yang and Weinberg, 2008). Included among the cell-biological traits evoked by such transcription factors are loss of adherens junctions and associated conversion from a polygonal/epithelial to a spindly/fibroblastic morphology, expression of matrix-degrading enzymes, increased motility, and heightened resistance to apoptosis—all traits implicated in the processes of invasion and metastasis. Several of these transcription factors can directly repress E-cadherin gene expression, thereby depriving neoplastic epithelial cells of this key suppressor of motility and invasiveness (Peinado et al., 2004). The available evidence suggests that these transcription factors regulate one another as well as overlapping sets of target genes. No rules have yet been established to describe their interactions and the conditions that govern their expression. Evidence from developmental genetics indicates that contextual signals received from neighboring cells in the embryo are involved in triggering expression of these transcription factors in those cells destined to pass through an EMT (Micalizzi et al., 2010); in an analogous fashion, increasing evidence suggests that heterotypic interactions of cancer cells with adjacent tumor-associated stromal cells can induce expression of the malignant cell phenotypes that are known to be choreographed by one or more of these transcriptional regulators (Karnoub and Weinberg, 2006–2007; Brabletz et al., 2001). Moreover, cancer cells at the invasive margins of certain carcinomas can be seen to have undergone an EMT, suggesting that these cancer cells are subject to microenvironmental stimuli distinct from those received by cancer cells located in the cores of these lesions (Hlubek et al., 2007). Although the evidence is still incomplete, it would appear that EMT-inducing transcription factors are able to orchestrate most steps of the invasion-metastasis cascade save the final step of colonization. We still know rather little about the various manifestations and temporal stability of the mesenchymal state produced by an EMT. Although expression of EMT-inducing transcription factors has been observed in certain nonepithelial tumor types, such as sarcomas and neuroectodermal tumors, their roles in programming malignant traits in these tumors are
presently poorly documented. Additionally, it remains to be determined whether invasive carcinoma cells necessarily acquire their capability through activation of parts of the EMT program, or whether alternative regulatory programs can also enable this capability. Heterotypic Contributions of Stromal Cells to Invasion and Metastasis It is increasingly apparent that crosstalk between cancer cells and cells of the neoplastic stroma is involved in the acquired capability for invasive growth and metastasis (Egeblad et al., 2010; Qian and Pollard, 2010; Joyce and Pollard, 2009; Kalluri and Zeisberg, 2006). Such signaling may impinge on carcinoma cells and act to alter their hallmark capabilities as suggested above. For example, mesenchymal stem cells (MSCs) present in the tumor stroma have been found to secrete CCL5/RANTES in response to signals released by cancer cells; CCL5 then acts reciprocally on the cancer cells to stimulate invasive behavior (Karnoub et al., 2007). Macrophages at the tumor periphery can foster local invasion by supplying matrix-degrading enzymes such as metalloproteinases and cysteine cathepsin proteases (Kessenbrock et al., 2010; Joyce and Pollard, 2009; Palermo and Joyce, 2008; Mohamed and Sloane, 2006); in one model system, the invasionpromoting macrophages are activated by IL-4 produced by the cancer cells (Gocheva et al., 2010). And in an experimental model of metastatic breast cancer, tumor-associated macrophages (TAMs) supply epidermal growth factor (EGF) to breast cancer cells, while the cancer cells reciprocally stimulate the macrophages with CSF-1; their concerted interactions facilitate intravasation into the circulatory system and metastatic dissemination of the cancer cells (Qian and Pollard, 2010; Wyckoff et al., 2007). Observations like these indicate that the phenotypes of highgrade malignancy do not arise in a strictly cell-autonomous manner, and that their manifestation cannot be understood solely through analyses of tumor cell genomes. One important implication, still untested, is that the ability to negotiate most of the steps of the invasion-metastasis cascade may be acquired in certain tumors without the requirement that the associated cancer cells undergo additional mutations beyond those that were needed for primary tumor formation. Plasticity in the Invasive Growth Program The role of contextual signals in inducing an invasive growth capability (often via an EMT) implies the possibility of reversibility, in that cancer cells that have disseminated from a primary tumor to a more distant tissue site may no longer benefit from the activated stroma and invasion/EMT-inducing signals that they experienced while residing in the primary tumor; in the absence of ongoing exposure to these signals, carcinoma cells may revert in their new homes to a noninvasive state. Thus, carcinoma cells that have undergone an EMT during initial invasion and metastatic dissemination may pass through the reverse process, termed the mesenchymal-epithelial transition (MET). This plasticity may result in the formation of new tumor colonies of carcinoma cells exhibiting a histopathology similar to those of carcinoma cells in the primary tumor that never underwent an EMT (Hugo et al., 2007). Moreover, the notion that cancer cells routinely pass through a complete EMT program is likely to be
simplistic; instead, in many cases, cancer cells may enter into an EMT program only partially, thereby acquiring new mesenchymal traits while continuing to express residual epithelial traits. Distinct Forms of Invasion May Underlie Different Cancer Types The EMT program regulates a particular type of invasiveness that has been termed ‘‘mesenchymal.’’ In addition, two other distinct modes of invasion have been identified and implicated in cancer cell invasion (Friedl and Wolf, 2008, 2010). ‘‘Collective invasion’’ involves nodules of cancer cells advancing en masse into adjacent tissues and is characteristic of, for example, squamous cell carcinomas; interestingly, such cancers are rarely metastatic, suggesting that this form of invasion lacks certain functional attributes that facilitate metastasis. Less clear is the prevalence of an ‘‘amoeboid’’ form of invasion (Madsen and Sahai, 2010; Sabeh et al., 2009), in which individual cancer cells show morphological plasticity, enabling them to slither through existing interstices in the extracellular matrix rather than clearing a path for themselves, as occurs in both the mesenchymal and collective forms of invasion. It is presently unresolved whether cancer cells participating in the collective and amoeboid forms of invasion employ components of the EMT program, or whether entirely different cell-biological programs are responsible for choreographing these alternative invasion programs. Another emerging concept, noted above, involves the facilitation of cancer cell invasion by inflammatory cells that assemble at the boundaries of tumors, producing the extracellular matrix-degrading enzymes and other factors that enable invasive growth (Kessenbrock et al., 2010; Qian and Pollard, 2010; Joyce and Pollard, 2009); these functions may obviate the need of cancer cells to produce these proteins through activation of EMT programs. Thus, cancer cells may secrete the chemoattractants that recruit the proinvasive inflammatory cells rather than producing the matrix-degrading enzymes themselves. The Daunting Complexity of Metastatic Colonization Metastasis can be broken down into two major phases: the physical dissemination of cancer cells from the primary tumor to distant tissues, and the adaptation of these cells to foreign tissue microenvironments that results in successful colonization, i.e., the growth of micrometastases into macroscopic tumors. The multiple steps of dissemination would seem to be in the purview of the EMT and similarly acting migratory programs. Colonization, however, is not strictly coupled with physical dissemination, as evidenced by the presence in many patients of myriad micrometastases that have successfully disseminated but never progress to macroscopic metastatic tumors (Talmadge and Fidler, 2010; McGowan et al., 2009; Aguirre-Ghiso, 2007; Townson and Chambers, 2006; Fidler, 2003). In some types of cancer, the primary tumor may release systemic suppressor factors that render such micrometastases dormant, as revealed clinically by explosive metastatic growth soon after resection of the primary growth (Demicheli et al., 2008; Folkman, 2002). In others, however, such as breast cancer and melanoma, macroscopic metastases may erupt decades after a primary tumor has been surgically removed or pharmacologically destroyed; these metastatic tumor growths evidently Cell 144, March 4, 2011 ª2011 Elsevier Inc. 655
reflect dormant micrometastases that have solved, after much trial and error, the complex problem of tissue colonization (Barkan, et al., 2010; Aguirre-Ghiso, 2007; Townson and Chambers, 2006). One can infer from such natural histories that micrometastases may lack other hallmark capabilities necessary for vigorous growth, such as the ability to activate angiogenesis; indeed the inability of certain experimentally generated dormant micrometastases to form macroscopic tumors has been ascribed to their failure to activate tumor angiogenesis (Naumov et al., 2008; Aguirre-Ghiso, 2007). Additionally, recent experiments have shown that nutrient starvation can induce intense autophagy that causes cancer cells to shrink and adopt a state of reversible dormancy; such cells may exit this state and resume active growth and proliferation when changes in tissue microenvironment, such as access to more nutrients, permit (Kenific et al., 2010; Lu et al., 2008). Other mechanisms of micrometastatic dormancy may involve anti-growth signals embedded in normal tissue extracellular matrix (Barkan et al., 2010) and tumor-suppressing actions of the immune system (Teng et al., 2008; Aguirre-Ghiso, 2007). Most disseminated cancer cells are likely to be poorly adapted, at least initially, to the microenvironment of the tissue in which they have landed. Accordingly, each type of disseminated cancer cell may need to develop its own set of ad hoc solutions to the problem of thriving in the microenvironment of one or another foreign tissue (Gupta et al., 2005). These adaptations might require hundreds of distinct colonization programs, each dictated by the type of disseminating cancer cell and the nature of the tissue microenvironment in which colonization is proceeding. As further discussed below, however, certain tissue microenviroments may be preordained to be intrinsically hospitable to disseminated cancer cells (Peinado et al., 2011; Talmadge and Fidler, 2010). Metastatic dissemination has long been depicted as the last step in multistep primary tumor progression, and indeed for many tumors that is likely the case, as illustrated by recent genome sequencing studies that present genetic evidence for clonal evolution of pancreatic ductal adenocarcinoma to metastasis (Campbell et al., 2010; Luebeck, 2010; Yachida et al., 2010). On the other hand, evidence has recently emerged indicating that cells can disseminate remarkably early, dispersing from ostensibly noninvasive premalignant lesions in both mice and humans (Coghlin and Murray, 2010; Klein, 2009). Additionally, micrometastases can be spawned from primary tumors that are not obviously invasive but possess a neovasculature lacking in lumenal integrity (Gerhardt and Semb, 2008). Although cancer cells can clearly disseminate from such pre-neoplastic lesions and seed the bone marrow and other tissues, their capability to colonize these sites and develop into pathologically significant macrometastases remains unproven. At present, we view this early metastatic dissemination as a demonstrable phenomenon in mice and humans whose clinical significance is yet to be established. Beyond the timing of their dissemination, it also remains unclear when and where cancer cells develop the ability to colonize foreign tissues as macroscopic tumors. This capability may arise during primary tumor formation as a result of a tumor’s particular developmental path prior to any dissemination, such 656 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
that primary tumor cells entering the circulation are fortuitously endowed with the ability to colonize certain distant tissue sites (Talmadge and Fidler, 2010). Alternatively, the ability to colonize specific tissues may only develop in response to the selective pressure on already disseminated cancer cells to adapt to growth in foreign tissue microenvironments. Having developed such tissue-specific colonizing ability, the cells in metastatic colonies may proceed to disseminate further, not only to new sites in the body but also back to the primary tumors in which their ancestors arose. Accordingly, tissuespecific colonization programs that are evident among cells within a primary tumor may originate not from classical tumor progression occurring within the primary lesion but instead from emigrants that have returned home (Kim et al., 2009). Such reseeding is consistent with the aforementioned studies of human pancreatic cancer metastasis (Campbell et al., 2010; Luebeck, 2010; Yachida et al., 2010). Stated differently, the phenotypes and underlying gene expression programs of the populations of cancer cells (and of the cancer stem cells discussed below) within primary tumors may be significantly modified by reverse migration of their distant metastatic progeny. Implicit in this self-seeding process is another notion: the supportive stroma that arises in a primary tumor and contributes to its acquisition of malignant traits may intrinsically provide a hospitable site for reseeding and colonization by circulating cancer cells emanating from metastatic lesions. Clarifying the regulatory programs that enable metastatic colonization represents an important agenda for future research. Substantial progress is being made, for example, in defining sets of genes (‘‘metastatic signatures’’) that correlate with and appear to facilitate the establishment of macroscopic metastases in specific tissues (Coghlin and Murray, 2010; Bos et al., 2009; Olson et al., 2009; Nguyen et al., 2009; Gupta et al., 2005). The challenge is considerable, given the apparent multitude of distinct colonization programs cited above. Moreover, colonization is unlikely to depend exclusively on cell-autonomous processes. Instead, it almost certainly requires the establishment of a permissive tumor microenvironment composed of critical stromal support cells. For these reasons, the process of colonization is likely to encompass a large number of cellbiological programs that are, in aggregate, considerably more complex and diverse than the preceding steps of metastatic dissemination. Programming of Hallmark Capabilities by Intracellular Circuitry In 2000, we presented a metaphor, in which the numerous signaling molecules affecting cancer cells operate as nodes and branches of elaborate integrated circuits that are reprogrammed derivatives of the circuits operating in normal cells. The ensuing decade has both solidified the original depiction of these circuits and expanded the catalog of signals and the interconnections of their signaling pathways. It is difficult if not impossible to graphically portray this circuit comprehensively and coherently, as was already the case in 2000. We now suggest a portrayal of this circuitry that is aligned with individual hallmarks of cancer. Thus, the intracellular integrated
Figure 2. Intracellular Signaling Networks Regulate the Operations of the Cancer Cell An elaborate integrated circuit operates within normal cells and is reprogrammed to regulate hallmark capabilities within cancer cells. Separate subcircuits, depicted here in differently colored fields, are specialized to orchestrate the various capabilities. At one level, this depiction is simplistic, as there is considerable crosstalk between such subcircuits. In addition, because each cancer cell is exposed to a complex mixture of signals from its microenvironment, each of these subcircuits is connected with signals originating from other cells in the tumor microenvironment, as outlined in Figure 5.
circuit can be segmented into distinct subcircuits, each of which is specialized to support a discrete cell-biological property in normal cells and is reprogrammed in order to implement a hallmark capability in cancer cells (Figure 2). Only a subset of hallmark capabilities are addressed in this figure, either because their underlying control circuits remain poorly understood or because they overlap extensively with those portrayed here. An additional dimension of complexity involves considerable interconnections and thus crosstalk between the individual subcircuits. For example, certain oncogenic events can affect multiple capabilities, as illustrated by the diverse effects that prominent oncogenes, such as mutant RAS and upregulated MYC, have on multiple hallmark capabilities (e.g., proliferative signaling, energy metabolism, angiogenesis, invasion, and survival). We anticipate that future renditions of this integrated circuit will encompass subcircuits and associated hallmark capabilities that are still not addressed here.
ENABLING CHARACTERISTICS AND EMERGING HALLMARKS We have defined the hallmarks of cancer as acquired functional capabilities that allow cancer cells to survive, proliferate, and disseminate; these functions are acquired in different tumor types via distinct mechanisms and at various times during the course of multistep tumorigenesis. Their acquisition is made possible by two enabling characteristics. Most prominent is the development of genomic instability in cancer cells, which generates random mutations including chromosomal rearrangements; among these are the rare genetic changes that can orchestrate hallmark capabilities. A second enabling characteristic involves the inflammatory state of premalignant and frankly malignant lesions that is driven by cells of the immune system, some of which serve to promote tumor progression through various means. Cell 144, March 4, 2011 ª2011 Elsevier Inc. 657
Figure 3. Emerging Hallmarks and Enabling Characteristics An increasing body of research suggests that two additional hallmarks of cancer are involved in the pathogenesis of some and perhaps all cancers. One involves the capability to modify, or reprogram, cellular metabolism in order to most effectively support neoplastic proliferation. The second allows cancer cells to evade immunological destruction, in particular by T and B lymphocytes, macrophages, and natural killer cells. Because neither capability is yet generalized and fully validated, they are labeled as emerging hallmarks. Additionally, two consequential characteristics of neoplasia facilitate acquisition of both core and emerging hallmarks. Genomic instability and thus mutability endow cancer cells with genetic alterations that drive tumor progression. Inflammation by innate immune cells designed to fight infections and heal wounds can instead result in their inadvertent support of multiple hallmark capabilities, thereby manifesting the now widely appreciated tumor-promoting consequences of inflammatory responses.
Yet other distinct attributes of cancer cells have been proposed to be functionally important for the development of cancer and might therefore be added to the list of core hallmarks (Negrini et al., 2010; Luo et al., 2009; Colotta et al., 2009). Two such attributes are particularly compelling. The first involves major reprogramming of cellular energy metabolism in order to support continuous cell growth and proliferation, replacing the metabolic program that operates in most normal tissues and fuels the physiological operations of the associated cells. The second involves active evasion by cancer cells from attack and elimination by immune cells; this capability highlights the dichotomous roles of an immune system that both antagonizes and enhances tumor development and progression. Both of these capabilities may well prove to facilitate the development and progression of many forms of human cancer and therefore can be considered to be emerging hallmarks of cancer. These enabling characteristics and emerging hallmarks, depicted in Figure 3, are discussed individually below. An Enabling Characteristic: Genome Instability and Mutation Acquisition of the multiple hallmarks enumerated above depends in large part on a succession of alterations in the genomes of neoplastic cells. Simply depicted, certain mutant genotypes confer selective advantage on subclones of cells, enabling their outgrowth and eventual dominance in a local tissue environment. Accordingly, multistep tumor progression can be portrayed as a succession of clonal expansions, each of which is triggered by the chance acquisition of an enabling mutant genotype. Because heritable phenotypes, e.g., inactivation of tumor suppressor genes, can also be acquired through epigenetic mechanisms such as DNA methylation and histone modifications (Berdasco and Esteller, 2010; Esteller, 2007; Jones and Baylin, 2007), some clonal expansions may well be triggered by nonmutational changes affecting the regulation of gene expression. 658 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
The extraordinary ability of genome maintenance systems to detect and resolve defects in the DNA ensures that rates of spontaneous mutation are usually very low during each cell generation. In the course of acquiring the roster of mutant genes needed to orchestrate tumorigenesis, cancer cells often increase the rates of mutation (Negrini et al., 2010; Salk et al., 2010). This mutability is achieved through increased sensitivity to mutagenic agents, through a breakdown in one or several components of the genomic maintenance machinery, or both. In addition, the accumulation of mutations can be accelerated by compromising the surveillance systems that normally monitor genomic integrity and force genetically damaged cells into either senescence or apoptosis (Jackson and Bartek, 2009; Kastan, 2008; Sigal and Rotter, 2000). The role of TP53 is central here, leading to its being called the ‘‘guardian of the genome’’ (Lane, 1992). A diverse array of defects affecting various components of the DNA-maintenance machinery—often referred to as the ‘‘caretakers’’ of the genome (Kinzler and Vogelstein, 1997)—have been documented. The catalog of defects in these caretaker genes includes those whose products are involved in (1) detecting DNA damage and activating the repair machinery, (2) directly repairing damaged DNA, and (3) inactivating or intercepting mutagenic molecules before they have damaged the DNA (Negrini et al., 2010; Ciccia and Elledge, 2010; Jackson and Bartek, 2009; Kastan, 2008; Harper and Elledge, 2007; Friedberg et al., 2006). From a genetic perspective, these caretaker genes behave much like tumor suppressor genes, in that their functions can be lost during the course of tumor progression, with such losses being achieved either through inactivating mutations or via epigenetic repression. Mutant copies of many of these caretaker genes have been introduced into the mouse germline and result, predictably, in increased cancer incidence, supporting their potential involvement in human cancer development (Barnes and Lindahl, 2004).
In the decade since we first enumerated the cancer hallmarks, another major source of tumor-associated genomic instability has been uncovered: as described earlier, the loss of telomeric DNA in many tumors generates karyotypic instability and associated amplification and deletion of chromosomal segments (Artandi and DePinho, 2010). When viewed in this light, telomerase is more than an enabler of the hallmark capability for unlimited replicative potential and must also be added to the list of critical caretakers responsible for maintaining genome integrity. Advances in the molecular-genetic analysis of cancer cell genomes have provided the most compelling demonstrations of function-altering mutations and of ongoing genomic instability during tumor progression. One type of analysis—comparative genomic hybridization (CGH)—documents the gains and losses of gene copy number across the cell genome; in many tumors, the pervasive genomic aberrations revealed by CGH provide clear evidence for loss of control of genome integrity. Importantly, the recurrence of specific aberrations (both amplifications and deletions) at particular sites in the genome indicates that such sites are likely to harbor genes whose alteration favors neoplastic progression (Korkola and Gray, 2010). More recently, with the advent of efficient and economical DNA-sequencing technologies, higher-resolution analyses have become possible. Early studies are revealing distinctive patterns of DNA mutations in different tumor types (see http:// cancergenome.nih.gov/). In the not-too-distant future, the sequencing of entire cancer cell genomes promises to clarify the prevalence of ostensibly random mutations scattered across cancer cell genomes. Thus, recurring genetic alterations may point to a causal role of particular mutations in tumor pathogenesis. Although the specifics of genome alteration vary dramatically between different tumor types, the large number of genome maintenance and repair defects that have already been documented in human tumors, together with abundant evidence of widespread destabilization of gene copy number and nucleotide sequence, persuade us that instability of the genome is inherent to the great majority of human cancer cells. This leads, in turn, to the conclusion that the defects in genome maintenance and repair are selectively advantageous and therefore instrumental for tumor progression, if only because they accelerate the rate at which evolving premalignant cells can accumulate favorable genotypes. As such, genome instability is clearly an enabling characteristic that is causally associated with the acquisition of hallmark capabilities. An Enabling Characteristic: Tumor-Promoting Inflammation Pathologists have long recognized that some tumors are densely infiltrated by cells of both the innate and adaptive arms of the immune system and thereby mirror inflammatory conditions arising in non-neoplastic tissues (Dvorak, 1986). With the advent of better markers for accurately identifying the distinct cell types of the immune system, it is now clear that virtually every neoplastic lesion contains immune cells present at densities ranging from subtle infiltrations detectable only with cell typespecific antibodies to gross inflammations that are apparent
even by standard histochemical staining techniques (Page`s et al., 2010). Historically, such immune responses were largely thought to reflect an attempt by the immune system to eradicate tumors, and indeed, there is increasing evidence for antitumoral responses to many tumor types with an attendant pressure on the tumor to evade immune destruction, as discussed below. By 2000, there were already clues that the tumor-associated inflammatory response had the unanticipated, paradoxical effect of enhancing tumorigenesis and progression, in effect helping incipient neoplasias to acquire hallmark capabilities. In the ensuing decade, research on the intersections between inflammation and cancer pathogenesis has blossomed, producing abundant and compelling demonstrations of the functionally important tumor-promoting effects that immune cells—largely of the innate immune system—have on neoplastic progression (DeNardo et al., 2010; Grivennikov et al., 2010; Qian and Pollard, 2010; Colotta et al., 2009). Inflammation can contribute to multiple hallmark capabilities by supplying bioactive molecules to the tumor microenvironment, including growth factors that sustain proliferative signaling, survival factors that limit cell death, proangiogenic factors, extracellular matrix-modifying enzymes that facilitate angiogenesis, invasion, and metastasis, and inductive signals that lead to activation of EMT and other hallmark-facilitating programs (DeNardo et al., 2010; Grivennikov et al., 2010; Qian and Pollard, 2010; Karnoub and Weinberg, 2006–2007). Importantly, inflammation is in some cases evident at the earliest stages of neoplastic progression and is demonstrably capable of fostering the development of incipient neoplasias into full-blown cancers (Qian and Pollard, 2010; de Visser et al., 2006). Additionally, inflammatory cells can release chemicals, notably reactive oxygen species, that are actively mutagenic for nearby cancer cells, accelerating their genetic evolution toward states of heightened malignancy (Grivennikov et al., 2010). As such, inflammation can be considered an enabling characteristic for its contributions to the acquisition of core hallmark capabilities. The cells responsible for this enabling characteristic are described in the section below on the tumor microenvironment. An Emerging Hallmark: Reprogramming Energy Metabolism The chronic and often uncontrolled cell proliferation that represents the essence of neoplastic disease involves not only deregulated control of cell proliferation but also corresponding adjustments of energy metabolism in order to fuel cell growth and division. Under aerobic conditions, normal cells process glucose, first to pyruvate via glycolysis in the cytosol and thereafter to carbon dioxide in the mitochondria; under anaerobic conditions, glycolysis is favored and relatively little pyruvate is dispatched to the oxygen-consuming mitochondria. Otto Warburg first observed an anomalous characteristic of cancer cell energy metabolism (Warburg, 1930, 1956a, 1956b): even in the presence of oxygen, cancer cells can reprogram their glucose metabolism, and thus their energy production, by limiting their energy metabolism largely to glycolysis, leading to a state that has been termed ‘‘aerobic glycolysis.’’ The existence of this metabolic switch in cancer cells has been substantiated in the ensuing decades. Such reprogramming of Cell 144, March 4, 2011 ª2011 Elsevier Inc. 659
energy metabolism is seemingly counterintuitive, in that cancer cells must compensate for the 18-fold lower efficiency of ATP production afforded by glycolysis relative to mitochondrial oxidative phosphorylation. They do so in part by upregulating glucose transporters, notably GLUT1, which substantially increases glucose import into the cytoplasm (Jones and Thompson, 2009; DeBerardinis et al., 2008; Hsu and Sabatini, 2008). Indeed, markedly increased uptake and utilization of glucose have been documented in many human tumor types, most readily by noninvasively visualizing glucose uptake using positron emission tomography (PET) with a radiolabeled analog of glucose (18F-fluorodeoxyglucose, FDG) as a reporter. Glycolytic fueling has been shown to be associated with activated oncogenes (e.g., RAS, MYC) and mutant tumor suppressors (e.g., TP53) (DeBerardinis et al., 2008; Jones and Thompson, 2009), whose alterations in tumor cells have been selected primarily for their benefits in conferring the hallmark capabilities of cell proliferation, avoidance of cytostatic controls, and attenuation of apoptosis. This reliance on glycolysis can be further accentuated under the hypoxic conditions that operate within many tumors: the hypoxia response system acts pleiotropically to upregulate glucose transporters and multiple enzymes of the glycolytic pathway (Semenza, 2010a; Jones and Thompson, 2009; DeBerardinis et al., 2008). Thus, both the Ras oncoprotein and hypoxia can independently increase the levels of the HIF1a and HIF2a transcription factors, which in turn upregulate glycolysis (Semenza, 2010a, 2010b; Kroemer and Pouyssegur, 2008). A functional rationale for the glycolytic switch in cancer cells has been elusive, given the relatively poor efficiency of generating ATP by glycolysis relative to mitochondrial oxidative phosphorylation. According to one long-forgotten (Potter, 1958) and recently revived and refined hypothesis (Vander Heiden et al., 2009), increased glycolysis allows the diversion of glycolytic intermediates into various biosynthetic pathways, including those generating nucleosides and amino acids; this facilitates, in turn, the biosynthesis of the macromolecules and organelles required for assembling new cells. Moreover, Warburg-like metabolism seems to be present in many rapidly dividing embryonic tissues, once again suggesting a role in supporting the large-scale biosynthetic programs that are required for active cell proliferation. Interestingly, some tumors have been found to contain two subpopulations of cancer cells that differ in their energy-generating pathways. One subpopulation consists of glucose-dependent (‘‘Warburg-effect’’) cells that secrete lactate, whereas cells of the second subpopulation preferentially import and utilize the lactate produced by their neighbors as their main energy source, employing part of the citric acid cycle to do so (Kennedy and Dewhirst, 2010; Feron, 2009; Semenza, 2008). These two populations evidently function symbiotically: the hypoxic cancer cells depend on glucose for fuel and secrete lactate as waste, which is imported and preferentially used as fuel by their better-oxygenated brethren. Although this provocative mode of intratumoral symbiosis has yet to be generalized, the cooperation between lactatesecreting and lactate-utilizing cells to fuel tumor growth is in fact not an invention of tumors but rather again reflects cooption of a normal physiological mechanism, in this case one operating in 660 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
muscle (Kennedy and Dewhirst, 2010; Feron, 2009; Semenza, 2008). Additionally, it is becoming apparent that oxygenation, ranging from normoxia to hypoxia, is not necessarily static in tumors but instead fluctuates temporally and regionally (Hardee et al., 2009), likely as a result of the instability and chaotic organization of the tumor-associated neovasculature. Altered energy metabolism is proving to be as widespread in cancer cells as many of the other cancer-associated traits that have been accepted as hallmarks of cancer. This realization raises the question of whether deregulating cellular energy metabolism is therefore a core hallmark capability of cancer cells that is as fundamental as the six well-established core hallmarks. In fact, the redirection of energy metabolism is largely orchestrated by proteins that are involved in one way or another in programming the core hallmarks of cancer. When viewed in this way, aerobic glycolysis is simply another phenotype that is programmed by proliferation-inducing oncogenes. Interestingly, activating (gain-of-function) mutations in the isocitrate dehydrogenase 1/2 (IDH) enzymes have been reported in glioma and other human tumors (Yen et al., 2010). Although these mutations may prove to have been clonally selected for their ability to alter energy metabolism, there is confounding data associating their activity with elevated oxidation and stability of the HIF-1 transcription factors (Reitman and Yan, 2010), which could in turn affect genome stability and angiogenesis/invasion, respectively, thus blurring the lines of phenotypic demarcation. Currently, therefore, the designation of reprogrammed energy metabolism as an emerging hallmark seems most appropriate, to highlight both its evident importance as well as the unresolved issues surrounding its functional independence from the core hallmarks. An Emerging Hallmark: Evading Immune Destruction A second, still-unresolved issue surrounding tumor formation involves the role that the immune system plays in resisting or eradicating formation and progression of incipient neoplasias, late-stage tumors, and micrometastases. The long-standing theory of immune surveillance proposes that cells and tissues are constantly monitored by an ever-alert immune system, and that such immune surveillance is responsible for recognizing and eliminating the vast majority of incipient cancer cells and thus nascent tumors. According to this logic, solid tumors that do appear have somehow managed to avoid detection by the various arms of the immune system or have been able to limit the extent of immunological killing, thereby evading eradication. The role of defective immunological monitoring of tumors would seem to be validated by the striking increases of certain cancers in immunocompromised individuals (Vajdic and van Leeuwen, 2009). However, the great majority of these are virus-induced cancers, suggesting that much of the control of this class of cancers normally depends on reducing viral burden in infected individuals, in part through eliminating virus-infected cells. These observations, therefore, seem to shed little light on the possible role of the immune system in limiting formation of the >80% of tumors of nonviral etiology. In recent years, however, an increasing body of evidence, both from genetically engineered mice and from clinical epidemiology, suggests that
the immune system operates as a significant barrier to tumor formation and progression, at least in some forms of non-virusinduced cancer. When mice genetically engineered to be deficient for various components of the immune system were assessed for the development of carcinogen-induced tumors, it was observed that tumors arose more frequently and/or grew more rapidly in the immunodeficient mice relative to immunocompetent controls. In particular, deficiencies in the development or function of CD8+ cytotoxic T lymphocytes (CTLs), CD4+ Th1 helper T cells, or natural killer (NK) cells each led to demonstrable increases in tumor incidence; moreover, mice with combined immunodeficiencies in both T cells and NK cells were even more susceptible to cancer development. The results indicated that, at least in certain experimental models, both the innate and adaptive cellular arms of the immune system are able to contribute significantly to immune surveillance and thus tumor eradication (Teng et al., 2008; Kim et al., 2007). In addition, transplantation experiments have shown that cancer cells that originally arose in immunodeficient mice are often inefficient at initiating secondary tumors in syngeneic immunocompetent hosts, whereas cancer cells from tumors arising in immunocompetent mice are equally efficient at initiating transplanted tumors in both types of hosts (Teng et al., 2008; Kim et al., 2007). Such behavior has been interpreted as follows: Highly immunogenic cancer cell clones are routinely eliminated in immunocompetent hosts—a process that has been referred to as ‘‘immunoediting’’—leaving behind only weakly immunogenic variants to grow and generate solid tumors; such weakly immunogenic cells can thereafter colonize both immunodeficient and immunocompetent hosts. Conversely, when arising in immunodeficient hosts, the immunogenic cancer cells are not selectively depleted and can, instead, prosper along with their weakly immunogenic counterparts. When cells from such nonedited tumors are serially transplanted into syngeneic recipients, the immunogenic cancer cells are rejected when they confront, for the first time, the competent immune systems of their secondary hosts (Smyth et al., 2006). (Unanswered in these particular experiments is the question of whether the chemical carcinogens used to induce such tumors are prone to generate cancer cells that are especially immunogenic.) Clinical epidemiology also increasingly supports the existence of antitumoral immune responses in some forms of human cancer (Bindea et al., 2010; Ferrone and Dranoff, 2010; Nelson, 2008). For example, patients with colon and ovarian tumors that are heavily infiltrated with CTLs and NK cells have a better prognosis than those that lack such abundant killer lymphocytes (Page`s et al., 2010; Nelson, 2008); the case for other cancers is suggestive but less compelling and is the subject of ongoing investigation. Additionally, some immunosuppressed organ transplant recipients have been observed to develop donorderived cancers, suggesting that in the ostensibly tumor-free donors, the cancer cells were held in check, in a dormant state, by a fully functional immune system (Strauss and Thomas, 2010). Still, the epidemiology of chronically immunosuppressed patients does not indicate significantly increased incidences of the major forms of nonviral human cancer, as noted above.
This might be taken as an argument against the importance of immune surveillance as an effective barrier to tumorigenesis and tumor progression. We note, however, that HIV and pharmacologically immunosuppressed patients are predominantly immunodeficient in the T and B cell compartments and thus do not present with the multicomponent immunological deficiencies that have been produced in the genetically engineered mutant mice lacking both NK cells and CTLs; this leaves open the possibility that such patients still have residual capability for an immunological defense against cancer that is mounted by NK and other innate immune cells. In truth, the above discussions of cancer immunology simplify tumor-host immunological interactions, as highly immunogenic cancer cells may well evade immune destruction by disabling components of the immune system that have been dispatched to eliminate them. For example, cancer cells may paralyze infiltrating CTLs and NK cells, by secreting TGF-b or other immunosuppressive factors (Yang et al., 2010; Shields et al., 2010). More subtle mechanisms operate through the recruitment of inflammatory cells that are actively immunosuppressive, including regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs). Both can suppress the actions of cytotoxic lymphocytes (Mougiakakos et al., 2010; Ostrand-Rosenberg and Sinha, 2009). In light of these considerations and the still-rudimentary demonstrations of antitumor immunity as a significant barrier to tumor formation and progression in humans, we present immunoevasion as another emerging hallmark, whose generality as a core hallmark capability remains to be firmly established. THE TUMOR MICROENVIRONMENT Over the past decade, tumors have increasingly been recognized as organs whose complexity approaches and may even exceed that of normal healthy tissues. When viewed from this perspective, the biology of a tumor can only be understood by studying the individual specialized cell types within it (Figure 4, upper) as well as the ‘‘tumor microenvironment’’ that they construct during the course of multistep tumorigenesis (Figure 4, lower). This depiction contrasts starkly with the earlier, reductionist view of a tumor as nothing more than a collection of relatively homogeneous cancer cells, whose entire biology could be understood by elucidating the cellautonomous properties of these cells. We enumerate here a set of cell types known to contribute in important ways to the biology of many tumors and discuss the regulatory signaling that controls their individual and collective functions. Most of these observations stem from the study of carcinomas, in which the neoplastic epithelial cells constitute a compartment (the parenchyma) that is clearly distinct from the mesenchymal cells forming the tumor-associated stroma. Cancer Cells and Cancer Stem Cells Cancer cells are the foundation of the disease; they initiate tumors and drive tumor progression forward, carrying the oncogenic and tumor suppressor mutations that define cancer as a genetic disease. Traditionally, the cancer cells within tumors Cell 144, March 4, 2011 ª2011 Elsevier Inc. 661
Figure 4. The Cells of the Tumor Microenvironment (Upper) An assemblage of distinct cell types constitutes most solid tumors. Both the parenchyma and stroma of tumors contain distinct cell types and subtypes that collectively enable tumor growth and progression. Notably, the immune inflammatory cells present in tumors can include both tumor-promoting as well as tumor-killing subclasses. (Lower) The distinctive microenvironments of tumors. The multiple stromal cell types create a succession of tumor microenvironments that change as tumors invade normal tissue and thereafter seed and colonize distant tissues. The abundance, histologic organization, and phenotypic characteristics of the stromal cell types, as well as of the extracellular matrix (hatched background), evolve during progression, thereby enabling primary, invasive, and then metastatic growth. The surrounding normal cells of the primary and metastatic sites, shown only schematically, likely also affect the character of the various neoplastic microenvironments. (Not shown are the premalignant stages in tumorigenesis, which also have distinctive microenvironments that are created by the abundance and characteristics of the assembled cells.)
have been portrayed as reasonably homogeneous cell populations until relatively late in the course of tumor progression, when hyperproliferation combined with increased genetic instability spawn distinct clonal subpopulations. Reflecting such clonal heterogeneity, many human tumors are histopathologically diverse, containing regions demarcated by various degrees of differentiation, proliferation, vascularity, inflammation, and/or invasiveness. In recent years, however, evidence has accumulated pointing to the existence of a new dimension of intratumor heterogeneity and a hitherto-unappreciated subclass of neoplastic cells within tumors, termed cancer stem cells (CSCs). Although the evidence is still fragmentary, CSCs may prove to be a common constituent of many if not most tumors, albeit being present with widely varying abundance. CSCs are defined operationally through their ability to efficiently seed new tumors upon inoculation into recipient host mice (Cho and Clarke, 2008; Lobo et al., 2007). This functional definition is often complemented by including the expression in CSCs of markers that are also expressed by the normal stem cells in the tissue-oforigin (Al-Hajj et al., 2003). CSCs were initially implicated in the pathogenesis of hematopoietic malignancies (Reya et al., 2001; Bonnet and Dick, 1997) and then years later were identified in solid tumors, in particular breast carcinomas and neuroectodermal tumors (Gilbertson and Rich, 2007; Al-Hajj et al., 2003). Fractionation of cancer cells on the basis of displayed cell-surface markers has yielded subpopulations of neoplastic cells with a greatly enhanced ability, relative to the corresponding majority populations, to seed new tumors upon implantation in immunodeficient mice. These 662 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
often-rare tumor-initiating cells proved to share transcriptional profiles with certain normal tissue stem cell populations, motivating their designation as stem-like. The origins of CSCs within a solid tumor have not been clarified and indeed may well vary from one tumor type to another. In some tumors, normal tissue stem cells may serve as the cells-of-origin that undergo oncogenic transformation to yield CSCs; in others, partially differentiated transit-amplifying cells, also termed progenitor cells, may suffer the initial oncogenic transformation thereafter assuming more stem-like character. Once primary tumors have formed, the CSCs, like their normal counterparts, may self-renew as well as spawn more differentiated derivatives; in the case of neoplastic CSCs, these descendant cells form the great bulk of many tumors. It remains to be established whether multiple distinct classes of increasingly neoplastic stem cells form during inception and subsequent multistep progression of tumors, ultimately yielding the CSCs that have been described in fully developed cancers. Recent research has interrelated the acquisition of CSC traits with the EMT transdifferentiation program discussed above (Singh and Settleman, 2010; Mani et al., 2008; Morel et al., 2008). Induction of this program in certain model systems can induce many of the defining features of stem cells, including self-renewal ability and the antigenic phenotypes associated with both normal and cancer stem cells. This concordance suggests that the EMT program not only may enable cancer cells to physically disseminate from primary tumors but also can confer on such cells the self-renewal capability that is crucial to their subsequent clonal expansion at sites of dissemination (Brabletz et al., 2005). If generalized, this connection raises an important corollary hypothesis: the heterotypic signals that trigger an EMT, such as those released by an activated, inflammatory stroma, may also be important in creating and maintaining CSCs.
An increasing number of human tumors are reported to contain subpopulations with the properties of CSCs, as defined operationally through their efficient tumor-initiating capabilities upon xenotransplantation into mice. Nevertheless, the importance of CSCs as a distinct phenotypic subclass of neoplastic cells remains a matter of debate, as does their oftcited rarity within tumors (Boiko et al., 2010; Gupta et al., 2009; Quintana et al., 2008). Indeed, it is plausible that the phenotypic plasticity operating within tumors may produce bidirectional interconversion between CSCs and non-CSCs, resulting in dynamic variation in the relative abundance of CSCs. Such plasticity could complicate definitive measurement of their prevalence. Analogous plasticity is already implicated in the EMT program, which can be engaged reversibly (Thiery and Sleeman, 2006). These complexities notwithstanding, it is evident that this new dimension of tumor heterogeneity holds important implications for successful cancer therapies. Increasing evidence in a variety of tumor types suggests that cells with properties of CSCs are more resistant to various commonly used chemotherapeutic treatments (Singh and Settleman, 2010; Creighton et al., 2009; Buck et al., 2007). Their persistence may help to explain the almost-inevitable disease recurrence following apparently successful debulking of human solid tumors by radiation and various forms of chemotherapy. Indeed, CSCs may well prove to underlie certain forms of tumor dormancy, whereby latent cancer cells persist for years or even decades after surgical resection or radio/chemotherapy, only to suddenly erupt and generate life-threatening disease. Hence, CSCs may represent a double-threat, in that they are more resistant to therapeutic killing and, at the same time, endowed with the ability to regenerate a tumor once therapy has been halted. This phenotypic plasticity implicit in CSC state may also enable the formation of functionally distinct subpopulations within a tumor that support overall tumor growth in various ways. For example, an EMT can convert epithelial carcinoma cells into mesenchymal, fibroblast-like cancer cells that may well assume the duties of cancer-associated fibroblasts (CAFs) in some tumors. Remarkably, several recent reports have documented the ability of glioblastoma cells (or possibly their associated CSC subpopulations) to transdifferentiate into endothelial-like cells that can substitute for bona fide host-derived endothelial cells in forming a tumor-associated neovasculature (Soda et al., 2011; El Hallani et al., 2010; Ricci-Vitiani et al., 2010; Wang et al., 2010). Observations like these indicate that certain tumors may acquire stromal support by inducing some of their own cancer cells to undergo various types of metamorphosis to produce stromal cell types rather than relying on recruited host cells to provide their functions. The discovery of CSCs and biological plasticity in tumors indicates that a single, genetically homogeneous population of cells within a tumor may nevertheless be phenotypically heterogeneous due to the presence of cells in distinct states of differentiation. However, an equally important source of phenotypic variability may derive from the genetic heterogeneity within a tumor that accumulates as cancer progression proceeds. Thus, elevated genetic instability operating in later stages of
tumor progression may drive rampant genetic diversification that outpaces the process of Darwinian selection, generating genetically distinct subpopulations far more rapidly than they can be eliminated. Such thinking is increasingly supported by in-depth sequence analysis of tumor cell genomes, which has become practical due to recent major advances in DNA (and RNA) sequencing technology. Thus the sequencing of the genomes of cancer cells microdissected from different sectors of the same tumor (Yachida et al., 2010) has revealed striking intratumoral genetic heterogeneity. Some of this genetic diversity may be reflected in the long-recognized histological heterogeneity within individual human tumors. Alternatively, this genetic diversification may enable functional specialization, producing subpopulations of cancer cells that contribute distinct, complementary capabilities, which then accrue to the common benefit of overall tumor growth as described above. Endothelial Cells Much of the cellular heterogeneity within tumors is found in their stromal compartments. Prominent among the stromal constituents are the cells forming the tumor-associated vasculature. Mechanisms of development, differentiation, and homeostasis of endothelial cells composing the arteries, veins, and capillaries were already well understood in 2000. So too was the concept of the ‘‘angiogenic switch,’’ which activates quiescent endothelial cells, causing them to enter into a cellbiological program that allows them to construct new blood vessels (see above). Over the last decade, a network of interconnected signaling pathways involving ligands of signal-transducing receptors displayed by endothelial cells (e.g., Notch, Neuropilin, Robo, and Eph-A/B) has been added to the already-prominent VEGF, angiopoietin, and FGF signals. These newly characterized pathways have been functionally implicated in developmental and tumor-associated angiogenesis and illustrate the complex regulation of endothelial cell phenotypes (Pasquale, 2010; Ahmed and Bicknell, 2009; Dejana et al., 2009; Carmeliet and Jain, 2000). Other avenues of research are revealing distinctive gene expression profiles of tumor-associated endothelial cells and identifying cell-surface markers displayed on the lumenal surfaces of normal versus tumor endothelial cells (Nagy et al., 2010; Ruoslahti et al., 2010; Ruoslahti, 2002). Differences in signaling, in transcriptome profiles, and in vascular ‘‘ZIP codes’’ will likely prove to be important for understanding the conversion of normal endothelial cells into tumor-associated endothelial cells. Such knowledge may lead, in turn, to opportunities to develop novel therapies that exploit these differences in order to selectively target tumor-associated endothelial cells. Closely related to the endothelial cells of the general circulation are those forming lymphatic vessels (Tammela and Alitalo, 2010). Their role in the tumor-associated stroma, specifically in supporting tumor growth, is poorly understood. Indeed, because of high interstitial pressure within solid tumors, intratumoral lymphatic vessels are typically collapsed and nonfunctional; in contrast, however, there are often functional, actively growing (‘‘lymphangiogenic’’) lymphatic vessels at the peripheries of tumors and in the adjacent normal tissues that cancer cells Cell 144, March 4, 2011 ª2011 Elsevier Inc. 663
invade. These associated lymphatics likely serve as channels for the seeding of metastases in the draining lymph nodes that are commonly observed in a number of cancer types. Pericytes As noted earlier, pericytes represent a specialized mesenchymal cell type (related to smooth muscle cells) with finger-like projections that wrap around the endothelial tubing of blood vessels. In normal tissues, pericytes are known to provide paracrine support signals to the normally quiescent endothelium. For example, Ang-1 secreted by pericytes conveys antiproliferative stabilizing signals that are received by the Tie2 receptors expressed on the surface of endothelial cells; some pericytes also produce low levels of VEGF that serve a trophic function in endothelial homeostasis (Gaengel et al., 2009; Bergers and Song, 2005). Pericytes also collaborate with the endothelial cells to synthesize the vascular basement membrane that anchors both pericytes and endothelial cells and helps vessel walls to withstand the hydrostatic pressure of blood flow. Genetic and pharmacological perturbation of the recruitment and association of pericytes has demonstrated the functional importance of these cells in supporting the tumor endothelium (Pietras and Ostman, 2010; Gaengel et al., 2009; Bergers and Song, 2005). For example, pharmacological inhibition of signaling through the PDGF receptor expressed by tumor pericytes and bone marrow-derived pericyte progenitors results in reduced pericyte coverage of tumor vessels, which in turn destabilizes vascular integrity and function (Pietras and Ostman, 2010; Raza et al., 2010; Gaengel et al., 2009); interestingly, and in contrast, the pericytes of normal vessels are not prone to such pharmacological disruption, providing another example of the differences in regulation of normal quiescent and tumor vasculature. An intriguing hypothesis, still to be fully substantiated, is that tumors with poor pericyte coverage of their vasculature may be more prone to permit cancer cell intravasation into the circulatory system, enabling subsequent hematogenous dissemination (Raza et al., 2010; Gerhardt and Semb, 2008). Immune Inflammatory Cells As also discussed above, infiltrating cells of the immune system are increasingly accepted to be generic constituents of tumors. These inflammatory cells operate in conflicting ways: both tumor-antagonizing and tumor-promoting leukocytes can be found, in various proportions, in most if not all neoplastic lesions. Although the presence of tumor-antagonizing CTLs and NK cells is not surprising, the prevalence of immune cells that functionally enhance hallmark capabilities was largely unanticipated. Evidence began to accumulate in the late 1990s that the infiltration of neoplastic tissues by cells of the immune system serves, perhaps counterintuitively, to promote tumor progression. Such work traced its conceptual roots back to the association of sites of chronic inflammation with tumor formation, and to the observation that tumors could be portrayed as wounds that never heal (Scha¨fer and Werner, 2008: Dvorak, 1986). In the course of normal wound healing and fighting infections, immune inflammatory cells appear transiently and then disappear, in contrast to their persistence in sites of chronic inflammation, where their presence has been associated with various tissue pathologies, 664 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
including fibrosis, aberrant angiogenesis, and neoplasia (Grivennikov et al., 2010; Karin et al., 2006). Over the past decade, the manipulation of genes involved in the determination or effector functions of various immune cell types, together with pharmacological inhibitors of such cells or their functions, has shown them to play diverse and critical roles in fostering tumorigenesis. The roster of tumor-promoting inflammatory cells now includes macrophage subtypes, mast cells, and neutrophils, as well as T and B lymphocytes (Coffelt et al., 2010; DeNardo et al., 2010; Egeblad et al., 2010; Johansson et al., 2008; Murdoch et al., 2008; DePalma et al., 2007). Such studies are yielding a growing list of signaling molecules released by inflammatory cells that serve as effectors of their tumor-promoting actions. These include the tumor growth factor EGF, the angiogenic growth factor VEGF, other proangiogenic factors such as FGF2, chemokines, and cytokines that amplify the inflammatory state; in addition, these cells may produce proangiogenic and/or proinvasive matrix-degrading enzymes, including MMP-9 and other matrix metalloproteinases, cysteine cathepsin proteases, and heparanase (Qian and Pollard, 2010; Murdoch et al., 2008). Consistent with their expression of these diverse effectors, tumor-infiltrating inflammatory cells have been shown to induce and help sustain tumor angiogenesis, to stimulate cancer cell proliferation, to facilitate, via their presence at the margins of tumors, tissue invasion, and to support the metastatic dissemination and seeding of cancer cells (Coffelt et al., 2010; Egeblad et al., 2010; Qian and Pollard, 2010; Mantovani, 2010; Joyce and Pollard, 2009; Mantovani et al., 2008; Murdoch et al., 2008; DePalma et al., 2007). In addition to fully differentiated immune cells present in tumor stroma, a variety of partially differentiated myeloid progenitors have been identified in tumors (Murdoch et al., 2008). Such cells represent intermediaries between circulating cells of bone marrow origin and the differentiated immune cells typically found in normal and inflamed tissues. Importantly, these progenitors, like their more differentiated derivatives, have demonstrable tumor-promoting activity. Of particular interest, a class of tumor-infiltrating myeloid cells (defined as coexpressing the macrophage marker CD11b and the neutrophil marker Gr1) has been shown to suppress CTL and NK cell activity, having been independently identified as MDSCs (Qian and Pollard, 2010; Ostrand-Rosenberg and Sinha, 2009). This attribute raises the possibility that recruitment of certain myeloid cells may be doubly beneficial for the developing tumor, by directly promoting angiogenesis and tumor progression while at the same time affording a means to evade immune destruction. The counterintuitive existence of both tumor-promoting and tumor-antagonizing immune cells can be rationalized by invoking the diverse roles of the immune system: On the one hand, the immune system specifically detects and targets infectious agents with the adaptive immune response, which is supported by cells of the innate immune system. On the other, the innate immune system is involved in wound healing and clearing dead cells and cellular debris. These specialized tasks are accomplished by distinct subclasses of inflammatory cells, namely a class of conventional macrophages and neutrophils (engaged in supporting adaptive immunity), and subclasses of ‘‘alternatively activated’’ macrophages, neutrophils, and
myeloid progenitors that are engaged in wound healing and tissue housecleaning (Egeblad et al., 2010; Mantovani, 2010; Qian and Pollard, 2010; Johansson et al., 2008). The latter subtypes of immune cells are one of the major sources of the angiogenic, epithelial, and stromal growth factors and matrix-remodeling enzymes that are needed for wound healing, and it is these cells that are recruited and subverted to support neoplastic progression. Similarly, subclasses of B and T lymphocytes may facilitate the recruitment, activation, and persistence of such wound-healing and tumor-promoting macrophages and neutrophils (DeNardo et al., 2010; Egeblad et al., 2010; Biswas and Mantovani, 2010). Of course, other subclasses of B and T lymphocytes and innate immune cell types can mount demonstrable tumor-killing responses. The balance between the conflicting inflammatory responses in tumors is likely to prove instrumental in prognosis and, quite possibly, in therapies designed to redirect these cells toward tumor destruction. Cancer-Associated Fibroblasts Fibroblasts are found in various proportions across the spectrum of carcinomas, constituting in many cases the preponderant cell population of the tumor stroma. The term ‘‘cancer-associated fibroblast’’ subsumes at least two distinct cell types: (1) cells with similarities to the fibroblasts that create the structural foundation supporting most normal epithelial tissues and (2) myofibroblasts, whose biological roles and properties differ markedly from those of tissue-derived fibroblasts. Myofibroblasts are identifiable by their expression of a-smooth muscle actin (SMA). They are rare in most healthy epithelial tissues, although certain tissues, such as the liver and pancreas, contain appreciable numbers of a-SMA-expressing cells. Myofibroblasts transiently increase in abundance in wounds and are also found in sites of chronic inflammation. Although beneficial to tissue repair, myofibroblasts are problematic in chronic inflammation, contributing to the pathological fibrosis observed in tissues such as lung, kidney, and liver. Recruited myofibroblasts and reprogrammed variants of normal tissue-derived fibroblastic cells have been demonstrated to enhance tumor phenotypes, notably cancer cell proliferation, angiogenesis, and invasion and metastasis; their tumorpromoting activities have largely been defined by transplantation of cancer-associated fibroblasts admixed with cancer cells into mice, and more recently by genetic and pharmacologic perturbation of their functions in tumor-prone mice (Dirat et al., 2010; Pietras and Ostman, 2010; Ra¨sa¨nen and Vaheri, 2010; Shimoda et al., 2010; Kalluri and Zeisberg, 2006; Bhowmick et al., 2004). Because they secrete a variety of extracellular matrix components, cancer-associated fibroblasts are implicated in the formation of the desmoplastic stroma that characterizes many advanced carcinomas. The full spectrum of functions contributed by both subtypes of cancer-associated fibroblasts to tumor pathogenesis remains to be elucidated. Stem and Progenitor Cells of the Tumor Stroma The various stromal cell types that constitute the tumor microenvironment may be recruited from adjacent normal tissue—the most obvious reservoir of such cell types. However, in recent
years, the bone marrow has increasingly been implicated as a key source of tumor-associated stromal cells (Bergfeld and DeClerck, 2010; Fang and Salven, 2011; Giaccia and Schipani, 2010; Patenaude et al., 2010; Lamagna and Bergers, 2006). Mesenchymal stem and progenitor cells have been found to transit into tumors from the marrow, where they may differentiate into the various well-characterized stromal cell types. Some of these recent arrivals may also persist in an undifferentiated or partially differentiated state, exhibiting functions that their more differentiated progeny lack. The bone marrow origins of stromal cell types have been demonstrated using tumor-bearing mice in which the bone marrow cells and thus their disseminated progeny have been selectively labeled with reporters such as green fluorescent protein (GFP). While immune inflammatory cells have been long known to derive from the bone marrow, more recently the progenitors of pericytes and of various subtypes of cancer-associated fibroblasts originating from the bone marrow have been described in various mouse models of cancer (Bergfeld and DeClerck, 2010; Fang and Salven, 2011; Giaccia and Schipani, 2010; Lamagna and Bergers, 2006); the prevalence and functional importance of endothelial progenitors for tumor angiogenesis is currently unresolved (Fang and Salven, 2011; Patenaude et al., 2010). Taken together, these various lines of evidence indicate that tumor-associated stromal cells may be supplied to growing tumors by proliferation of preexisting stromal cells, by differentiation in situ of local stem/progenitor cells originating in the neighboring normal tissue, or via recruitment of bone marrow-derived stem/progenitor cells. Heterotypic Signaling Orchestrates the Cells of the Tumor Microenvironment Depictions of the intracellular circuitry governing cancer cell biology (e.g., Figure 2) will need to be complemented by similar diagrams charting the complex interactions between the neoplastic and stromal cells within a tumor and the dynamic extracellular matrix that they collectively erect and remodel (Egeblad et al., 2010; Kessenbrock et al., 2010; Pietras and Ostman, 2010; Polyak et al., 2009). A reasonably complete, graphic depiction of the network of microenvironmental signaling interactions is still far beyond our reach, as the great majority of signaling molecules and pathways remain to be identified. We provide instead a hint of such interactions in Figure 5, upper. These few well-established examples are intended to exemplify a signaling network of remarkable complexity that is of critical importance to tumor pathogenesis. Another dimension of complexity is not represented in this simple schematic: both neoplastic cells and the stromal cells around them change progressively during the multistep transformation of normal tissues into high-grade malignancies. This histopathological progression must reflect underlying changes in heterotypic signaling between tumor parenchyma and stroma. Such stepwise progression is likely to depend on back-andforth reciprocal interactions between the neoplastic cells and the supporting stromal cells, as depicted in Figure 5, lower. Thus, incipient neoplasias begin the interplay by recruiting and activating stromal cell types that assemble into an initial preneoplastic stroma, which in turn responds reciprocally by enhancing Cell 144, March 4, 2011 ª2011 Elsevier Inc. 665
Figure 5. Signaling Interactions in the Tumor Microenvironment during Malignant Progression (Upper) The assembly and collective contributions of the assorted cell types constituting the tumor microenvironment are orchestrated and maintained by reciprocal heterotypic signaling interactions, of which only a few are illustrated. (Lower) The intracellular signaling depicted in the upper panel within the tumor microenvironment is not static but instead changes during tumor progression as a result of reciprocal signaling interactions between cancer cells of the parenchyma and stromal cells that convey the increasingly aggressive phenotypes that underlie growth, invasion, and metastatic dissemination. Importantly, the predisposition to spawn metastatic lesions can begin early, being influenced by the differentiation program of the normal cell-of-origin or by initiating oncogenic lesions. Certain organ sites (sometimes referred to as ‘‘fertile soil’’ or ‘‘metastatic niches’’) can be especially permissive for metastatic seeding and colonization by certain types of cancer cells, as a consequence of local properties that are either intrinsic to the normal tissue or induced at a distance by systemic actions of primary tumors. Cancer stem cells may be variably involved in some or all of the different stages of primary tumorigenesis and metastasis.
666 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
the neoplastic phenotypes of the nearby cancer cells. The cancer cells, which may further evolve genetically, again feed signals back to the stroma, continuing the reprogramming of normal stromal cells to serve the budding neoplasm; ultimately signals originating in the tumor stroma enable cancer cells to invade normal adjacent tissues and disseminate. This model of reciprocal heterotypic signaling must be extended to encompass the final stage of multistep tumor progression—metastasis (Figure 5, lower right). The circulating cancer cells that are released from primary tumors leave a microenvironment created by the supportive stroma of such tumors. However, upon landing in a distant organ, these cancer cells encounter a naive, fully normal, tissue microenvironment. Consequently, many of the heterotypic signals that shaped their phenotype while they resided within primary tumors may be absent in sites of dissemination, constituting a barrier to growth of the seeded cancer cells. Thus, the succession of reciprocal cancer cell to stromal cell interactions that defined multistep progression in the primary tumor now must be repeated anew in distant tissues as disseminated cancer cells proceed to colonize their newfound organ sites. Although this logic applies in some cases of metastasis, in others, as mentioned earlier, certain tissue microenvironments may, for various reasons, already be supportive of freshly seeded cancer cells; such permissive sites have been referred to as ‘‘metastatic niches’’ (Peinado et al., 2011; Coghlin and Murray, 2010). Implicit in this term is the notion that cancer cells seeded in such sites may not need to begin by inducing a supportive stroma because it already preexists, at least in part. Such permissivity may be intrinsic to the tissue site (Talmadge and Fidler, 2010) or preinduced by circulating factors released by the primary tumor (Peinado et al., 2011). The most well-documented components of induced premetastatic niches are tumor-promoting inflammatory cells, although other cell types and the ECM may well prove to play important roles in different metastatic contexts. The likelihood that signaling interactions between cancer cells and their supporting stroma evolve during the course of multistage tumor development clearly complicates the goal of fully elucidating the mechanisms of cancer pathogenesis. For example, this reality poses challenges to systems biologists seeking to chart the crucial regulatory networks than orchestrate malignant progression. Moreover, it seems likely that understanding these dynamic variations will become crucial to the development of novel therapies designed to successfully target both primary and metastatic tumors. THERAPEUTIC TARGETING The introduction of mechanism-based targeted therapies to treat human cancers has been heralded as one of the fruits of three decades of remarkable progress of research into the mechanisms of cancer pathogenesis. We do not attempt here to enumerate the myriad therapies that are under development or have been introduced of late into the clinic. Instead, we consider how the description of hallmark principles is beginning to inform therapeutic development at present and may increasingly do so in the future.
The rapidly growing armamentarium of targeted therapeutics can be categorized according to their respective effects on one or more hallmark capabilities, as illustrated in the examples presented in Figure 6. Indeed, the observed efficacy of these drugs represents, in each case, a validation of a particular capability: if a capability is truly important for the biology of tumors, then its inhibition should impair tumor growth and progression. We note that most of the hallmark-targeting cancer drugs developed to date have been deliberately directed toward specific molecular targets that are involved in one way or another in enabling particular capabilities. Such specificity of action has been considered a virtue, as it presents inhibitory activity against a target while having, in principle, relatively fewer off-target effects and thus less nonspecific toxicity. In fact, resulting clinical responses have generally been transitory, being followed by almost-inevitable relapses. One interpretation of this history, supported by growing experimental evidence, is that each of the core hallmark capabilities is regulated by partially redundant signaling pathways. Consequently, a targeted therapeutic agent inhibiting one key pathway in a tumor may not completely shut off a hallmark capability, allowing some cancer cells to survive with residual function until they or their progeny eventually adapt to the selective pressure imposed by the therapy being applied. Such adaptation, which can be accomplished by mutation, epigenetic reprogramming, or remodeling of the stromal microenvironment, can reestablish the functional capability, permitting renewed tumor growth and clinical relapse. Given that the number of parallel signaling pathways supporting a given hallmark must be limited, it may become possible to target all of these supporting pathways therapeutically, thereby preventing the development of adaptive resistance. In response to therapy, cancer cells may also reduce their dependence on a particular hallmark capability, becoming more dependent on another; this represents a quite different form of acquired drug resistance. This concept is exemplified by recent discoveries of unexpected responses to antiangiogenic therapies. Some have anticipated that effective inhibition of angiogenesis would render tumors dormant and might even lead to their dissolution (Folkman and Kalluri, 2004). Instead, the clinical responses to antiangiogenic therapies have been found to be transitory (Azam et al., 2010; Ebos et al., 2009; Bergers and Hanahan, 2008). In certain preclinical models, where potent angiogenesis inhibitors succeed in suppressing this hallmark capability, tumors adapt and shift from a dependence upon continuing angiogenesis to heightening the activity of another instead—invasiveness and metastasis (Azam et al., 2010: Ebos et al., 2009; Bergers and Hanahan, 2008). By invading nearby tissues, initially hypoxic cancer cells evidently gain access to normal, preexisting tissue vasculature. Initial clinical validation of this adaptive/evasive resistance is apparent in the increased invasion and local metastasis seen when human glioblastomas are treated with antiangiogenic therapies (Ellis and Reardon, 2009; Norden et al., 2009; Verhoeff et al., 2009). The applicability of this lesson to other human cancers has yet to be established. Analogous adaptive shifts in dependence on other hallmark traits may also limit efficacy of analogous hallmark-targeting Cell 144, March 4, 2011 ª2011 Elsevier Inc. 667
Figure 6. Therapeutic Targeting of the Hallmarks of Cancer Drugs that interfere with each of the acquired capabilities necessary for tumor growth and progression have been developed and are in clinical trials or in some cases approved for clinical use in treating certain forms of human cancer. Additionally, the investigational drugs are being developed to target each of the enabling characteristics and emerging hallmarks depicted in Figure 3, which also hold promise as cancer therapeutics. The drugs listed are but illustrative examples; there is a deep pipeline of candidate drugs with different molecular targets and modes of action in development for most of these hallmarks.
therapies. For example, the deployment of apoptosis-inducing drugs may induce cancer cells to hyperactivate mitogenic signaling, enabling them to compensate for the initial attrition triggered by such treatments. Such considerations suggest that drug development and the design of treatment protocols will benefit from incorporating the concepts of functionally discrete hallmark capabilities and of the multiple biochemical pathways involved in supporting each of them. Thus, in particular, we can envisage that selective cotargeting of multiple core and emerging hallmark capabilities and enabling characteristics (Figure 6) in mechanism-guided combinations will result in more effective and durable therapies for human cancer. CONCLUSION AND FUTURE VISION We have sought here to revisit, refine, and extend the concept of cancer hallmarks, which has provided a useful conceptual framework for understanding the complex biology of cancer. 668 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
The six acquired capabilities—the hallmarks of cancer—have stood the test of time as being integral components of most forms of cancer. Further refinement of these organizing principles will surely come in the foreseeable future, continuing the remarkable conceptual progress of the last decade. Looking ahead, we envision significant advances during the coming decade in our understanding of invasion and metastasis. Similarly, the role of aerobic glycolysis in malignant growth will be elucidated, including a resolution of whether this metabolic reprogramming is a discrete capability separable from the core hallmark of chronically sustained proliferation. We remain perplexed as to whether immune surveillance is a barrier that virtually all tumors must circumvent, or only an idiosyncrasy of an especially immunogenic subset of them; this issue too will be resolved in one way or another. Yet other areas are currently in rapid flux. In recent years, elaborate molecular mechanisms controlling transcription through chromatin modifications have been uncovered, and there are
clues that specific shifts in chromatin configuration occur during the acquisition of certain hallmark capabilities (Berdasco and Esteller, 2010). Functionally significant epigenetic alterations seem likely to be factors not only in the cancer cells but also in the altered cells of the tumor-associated stroma. It is unclear at present whether an elucidation of these epigenetic mechanisms will materially change our overall understanding of the means by which hallmark capabilities are acquired or simply add additional detail to the regulatory circuitry that is already known to govern them. Similarly, the discovery of hundreds of distinct regulatory microRNAs has already led to profound changes in our understanding of the genetic control mechanisms that operate in health and disease. By now dozens of microRNAs have been implicated in various tumor phenotypes (Garzon et al., 2010), and yet these only scratch the surface of the real complexity, as the functions of hundreds of microRNAs known to be present in our cells and altered in expression in different forms of cancer remain total mysteries. Here again, we are unclear as to whether future progress will cause fundamental shifts in our understanding of the pathogenetic mechanisms of cancer or only add detail to the elaborate regulatory circuits that have already been mapped out. Finally, the circuit diagrams of heterotypic interactions between the multiple distinct cell types that assemble and collaborate to produce different forms and progressively malignant stages of cancer are currently rudimentary. In another decade, we anticipate that the signaling circuitry describing the intercommunication between these various cells within tumors will be charted in far greater detail and clarity, eclipsing our current knowledge. And, as before (Hanahan and Weinberg, 2000), we continue to foresee cancer research as an increasingly logical science, in which myriad phenotypic complexities are manifestations of a small set of underlying organizing principles.
Al-Hajj, M., Wicha, M.S., Benito-Hernandez, A., Morrison, S.J., and Clarke, M.F. (2003). Prospective identification of tumorigenic breast cancer cells. Proc. Natl. Acad. Sci. USA 100, 3983–3988. Amaravadi, R.K., and Thompson, C.B. (2007). The roles of therapy-induced autophagy and necrosis in cancer treatment. Clin. Cancer Res. 13, 7271–7279. Amit, I., Citri, A., Shay, T., Lu, Y., Katz, M., Zhang, F., Tarcic, G., Siwak, D., Lahad, J., Jacob-Hirsch, J., et al. (2007). A module of negative feedback regulators defines growth factor signaling. Nat. Genet. 39, 503–512. Apel, A., Zentgraf, H., Bu¨chler, M.W., and Herr, I. (2009). Autophagy-A doubleedged sword in oncology. Int. J. Cancer 125, 991–995. Artandi, S.E., and DePinho, R.A. (2000). Mice without telomerase: what can they teach us about human cancer? Nat. Med. 6, 852–855. Artandi, S.E., and DePinho, R.A. (2010). Telomeres and telomerase in cancer. Carcinogenesis 31, 9–18. Azam, F., Mehta, S., and Harris, A.L. (2010). Mechanisms of resistance to antiangiogenesis therapy. Eur. J. Cancer 46, 1323–1332. Baeriswyl, V., and Christofori, G. (2009). The angiogenic switch in carcinogenesis. Semin. Cancer Biol. 19, 329–337. Baluk, P., Hashizume, H., and McDonald, D.M. (2005). Cellular abnormalities of blood vessels as targets in cancer. Curr. Opin. Genet. Dev. 15, 102–111. Barkan, D., Green, J.E., and Chambers, A.F. (2010). Extracellular matrix: a gatekeeper in the transition from dormancy to metastatic growth. Eur. J. Cancer 46, 1181–1188. Barnes, D.E., and Lindahl, T. (2004). Repair and genetic consequences of endogenous DNA base damage in mammalian cells. Annu. Rev. Genet. 38, 445–476. Barrallo-Gimeno, A., and Nieto, M.A. (2005). The Snail genes as inducers of cell movement and survival: implications in development and cancer. Development 132, 3151–3161. Berdasco, M., and Esteller, M. (2010). Aberrant epigenetic landscape in cancer: How cellular identity goes awry. Dev. Cell 19, 698–711. Bergers, G., and Benjamin, L.E. (2003). Tumorigenesis and the angiogenic switch. Nat. Rev. Cancer 3, 401–410. Bergers, G., and Hanahan, D. (2008). Modes of resistance to anti-angiogenic therapy. Nat. Rev. Cancer 8, 592–603. Bergers, G., and Song, S. (2005). The role of pericytes in blood-vessel formation and maintenance. Neuro-oncol. 7, 452–464.
SUPPLEMENTAL INFORMATION
Bergfeld, S.A., and DeClerck, Y.A. (2010). Bone marrow-derived mesenchymal stem cells and the tumor microenvironment. Cancer Metastasis Rev. 29, 249–261.
Supplemental Information includes six figures that are downloadable for presentations and can be found with this article online at doi:10.1016/j.cell. 2011.02.013.
Berx, G., and van Roy, F. (2009). Involvement of members of the cadherin superfamily in cancer. Cold Spring Harb. Perspect. Biol. 1, a003129.
ACKNOWLEDGMENTS We thank Terry Schoop (OFC Graphics, Kensington, CA, USA) for exceptional efforts in preparing the figures. And we thank Gerard Evan (Cambridge, UK), Erwin Wagner (Madrid, ESP), and Zena Werb (San Francisco, USA) for valuable comments and suggestions on the manuscript. D.H. and R.A.W. are American Cancer Society Research Professors. Research in the authors’ laboratories has been largely supported by the U.S. National Cancer Institute. Due to space limitations, many primary and historical publications have not been cited, particularly in cases where topical reviews are available. REFERENCES Adams, J.M., and Cory, S. (2007). The Bcl-2 apoptotic switch in cancer development and therapy. Oncogene 26, 1324–1337. Aguirre-Ghiso, J.A. (2007). Models, mechanisms and clinical evidence for cancer dormancy. Nat. Rev. Cancer 7, 834–846. Ahmed, Z., and Bicknell, R. (2009). Angiogenic signalling pathways. Methods Mol. Biol. 467, 3–24.
Bhowmick, N.A., Neilson, E.G., and Moses, H.L. (2004). Stromal fibroblasts in cancer initiation and progression. Nature 432, 332–337. Bierie, B., and Moses, H.L. (2006). Tumour microenvironment: TGFbeta: the molecular Jekyll and Hyde of cancer. Nat. Rev. Cancer 6, 506–520. Bindea, G., Mlecnik, B., Fridman, W.H., Page`s, F., and Galon, J. (2010). Natural immunity to cancer in humans. Curr. Opin. Immunol. 22, 215–222. Biswas, S.K., and Mantovani, A. (2010). Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nat. Immunol. 11, 889–896. Blasco, M.A. (2005). Telomeres and human disease: ageing, cancer and beyond. Nat. Rev. Genet. 6, 611–622. Boiko, A.D., Razorenova, O.V., van de Rijn, M., Swetter, S.M., Johnson, D.L., Ly, D.P., Butler, P.D., Yang, G.P., Joshua, B., Kaplan, M.J., et al. (2010). Human melanoma-initiating cells express neural crest nerve growth factor receptor CD271. Nature 466, 133–137. Bonnet, D., and Dick, J.E. (1997). Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat. Med. 3, 730–737. Bos, P.D., Zhang, X.H., Nadal, C., Shu, W., Gomis, R.R., Nguyen, D.X., Minn, A.J., van de Vijver, M.J., Gerald, W.L., Foekens, J.A., and Massague´, J. (2009). Genes that mediate breast cancer metastasis to the brain. Nature 459, 1005–1009.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 669
Brabletz, T., Jung, A., Reu, S., Porzner, M., Hlubek, F., Kunz-Schughart, L.A., Knuechel, R., and Kirchner, T. (2001). Variable beta-catenin expression in colorectal cancers indicates tumor progression driven by the tumor environment. Proc. Natl. Acad. Sci. USA 98, 10356–10361. Brabletz, T., Jung, A., Spaderna, S., Hlubek, F., and Kirchner, T. (2005). Opinion: migrating cancer stem cells - an integrated concept of malignant tumour progression. Nat. Rev. Cancer 5, 744–749. Buck, E., Eyzaguirre, A., Barr, S., Thompson, S., Sennello, R., Young, D., Iwata, K.K., Gibson, N.W., Cagnoni, P., and Haley, J.D. (2007). Loss of homotypic cell adhesion by epithelial-mesenchymal transition or mutation limits sensitivity to epidermal growth factor receptor inhibition. Mol. Cancer Ther. 6, 532–541. Burkhart, D.L., and Sage, J. (2008). Cellular mechanisms of tumour suppression by the retinoblastoma gene. Nat. Rev. Cancer 8, 671–682. Cabrita, M.A., and Christofori, G. (2008). Sprouty proteins, masterminds of receptor tyrosine kinase signaling. Angiogenesis 11, 53–62. Campbell, P.J., Yachida, S., Mudie, L.J., Stephens, P.J., Pleasance, E.D., Stebbings, L.A., Morsberger, L.A., Latimer, C., McLaren, S., Lin, M.L., et al. (2010). The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 467, 1109–1113.
DeBerardinis, R.J., Lum, J.J., Hatzivassiliou, G., and Thompson, C.B. (2008). The biology of cancer: Metabolic reprogramming fuels cell growth and proliferation. Cell Metab. 7, 11–20. Dejana, E., Orsenigo, F., Molendini, C., Baluk, P., and McDonald, D.M. (2009). Organization and signaling of endothelial cell-to-cell junctions in various regions of the blood and lymphatic vascular trees. Cell Tissue Res. 335, 17–25. Demicheli, R., Retsky, M.W., Hrushesky, W.J., Baum, M., and Gukas, I.D. (2008). The effects of surgery on tumor growth: a century of investigations. Ann. Oncol. 19, 1821–1828. DeNardo, D.G., Andreu, P., and Coussens, L.M. (2010). Interactions between lymphocytes and myeloid cells regulate pro- versus anti-tumor immunity. Cancer Metastasis Rev. 29, 309–316. De Palma, M., Murdoch, C., Venneri, M.A., Naldini, L., and Lewis, C.E. (2007). Tie2-expressing monocytes: regulation of tumor angiogenesis and therapeutic implications. Trends Immunol. 28, 519–524. Deshpande, A., Sicinski, P., and Hinds, P.W. (2005). Cyclins and cdks in development and cancer: a perspective. Oncogene 24, 2909–2915. de Visser, K.E., Eichten, A., and Coussens, L.M. (2006). Paradoxical roles of the immune system during cancer development. Nat. Rev. Cancer 6, 24–37.
Cao, Y. (2010). Adipose tissue angiogenesis as a therapeutic target for obesity and metabolic diseases. Nat. Rev. Drug Discov. 9, 107–115.
Dirat, B., Bochet, L., Escourrou, G., Valet, P., and Muller, C. (2010). Unraveling the obesity and breast cancer links: a role for cancer-associated adipocytes? Endocr. Dev. 19, 45–52.
Carmeliet, P. (2005). VEGF as a key mediator of angiogenesis in cancer. Oncology 69 (Suppl 3), 4–10.
Dvorak, H.F. (1986). Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing. N. Engl. J. Med. 315, 1650–1659.
Carmeliet, P., and Jain, R.K. (2000). Angiogenesis in cancer and other diseases. Nature 407, 249–257.
Ebos, J.M., Lee, C.R., and Kerbel, R.S. (2009). Tumor and host-mediated pathways of resistance and disease progression in response to antiangiogenic therapy. Clin. Cancer Res. 15, 5020–5025.
Cavallaro, U., and Christofori, G. (2004). Cell adhesion and signalling by cadherins and Ig-CAMs in cancer. Nat. Rev. Cancer 4, 118–132. Cheng, N., Chytil, A., Shyr, Y., Joly, A., and Moses, H.L. (2008). Transforming growth factor-beta signaling-deficient fibroblasts enhance hepatocyte growth factor signaling in mammary carcinoma cells to promote scattering and invasion. Mol. Cancer Res. 6, 1521–1533. Chin, K., de Solorzano, C.O., Knowles, D., Jones, A., Chou, W., Rodriguez, E.G., Kuo, W.L., Ljung, B.M., Chew, K., Myambo, K., et al. (2004). In situ analyses of genome instability in breast cancer. Nat. Genet. 36, 984–988. Cho, R.W., and Clarke, M.F. (2008). Recent advances in cancer stem cells. Curr. Opin. Genet. Dev. 18, 1–6. Ciccia, A., and Elledge, S.J. (2010). The DNA damage response: making it safe to play with knives. Mol. Cell 40, 179–204. Coffelt, S.B., Lewis, C.E., Naldini, L., Brown, J.M., Ferrara, N., and De Palma, M. (2010). Elusive identities and overlapping phenotypes of proangiogenic myeloid cells in tumors. Am. J. Pathol. 176, 1564–1576. Coghlin, C., and Murray, G.I. (2010). Current and emerging concepts in tumour metastasis. J. Pathol. 222, 1–15. Collado, M., and Serrano, M. (2010). Senescence in tumours: evidence from mice and humans. Nat. Rev. Cancer 10, 51–57. Colotta, F., Allavena, P., Sica, A., Garlanda, C., and Mantovani, A. (2009). Cancer-related inflammation, the seventh hallmark of cancer: links to genetic instability. Carcinogenesis 30, 1073–1081. Cong, Y., and Shay, J.W. (2008). Actions of human telomerase beyond telomeres. Cell Res. 18, 725–732. Creighton, C.J., Li, X., Landis, M., Dixon, J.M., Neumeister, V.M., Sjolund, A., Rimm, D.L., Wong, H., Rodriguez, A., Herschkowitz, J.I., et al. (2009). Residual breast cancers after conventional therapy display mesenchymal as well as tumor-initiating features. Proc. Natl. Acad. Sci. USA 106, 13820–13825. Curto, M., Cole, B.K., Lallemand, D., Liu, C.H., and McClatchey, A.I. (2007). Contact-dependent inhibition of EGFR signaling by Nf2/Merlin. J. Cell Biol. 177, 893–903. Davies, M.A., and Samuels, Y. (2010). Analysis of the genome to personalize therapy for melanoma. Oncogene 29, 5545–5555.
670 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
Egeblad, M., Nakasone, E.S., and Werb, Z. (2010). Tumors as organs: complex tissues that interface with the entire organism. Dev. Cell 18, 884–901. El Hallani, S., Boisselier, B., Peglion, F., Rousseau, A., Colin, C., Idbaih, A., Marie, Y., Mokhtari, K., Thomas, J.L., Eichmann, A., et al. (2010). A new alternative mechanism in glioblastoma vascularization: tubular vasculogenic mimicry. Brain 133, 973–982. Ellis, L.M., and Reardon, D.A. (2009). Cancer: The nuances of therapy. Nature 458, 290–292. Esteller, M. (2007). Cancer epigenomics: DNA methylomes and histone-modification maps. Nat. Rev. Genet. 8, 286–298. Evan, G.I., and d’Adda di Fagagna, F. (2009). Cellular senescence: hot or what? Curr. Opin. Genet. Dev. 19, 25–31. Evan, G., and Littlewood, T. (1998). A matter of life and cell death. Science 281, 1317–1322. Fang, S., and Salven, P. (2011). Stem cells in tumor angiogenesis. J. Mol. Cell. Cardiol. 50, 290–295. Feron, O. (2009). Pyruvate into lactate and back: from the Warburg effect to symbiotic energy fuel exchange in cancer cells. Radiother. Oncol. 92, 329– 333. Feldser, D.M., and Greider, C.W. (2007). Short telomeres limit tumor progression in vivo by inducing senescence. Cancer Cell 11, 461–469. Ferrara, N. (2009). Vascular endothelial growth factor. Arterioscler. Thromb. Vasc. Biol. 29, 789–791. Ferrara, N. (2010). Pathways mediating VEGF-independent tumor angiogenesis. Cytokine Growth Factor Rev. 21, 21–26. Ferrone, C., and Dranoff, G. (2010). Dual roles for immunity in gastrointestinal cancers. J. Clin. Oncol. 28, 4045–4051. Fidler, I.J. (2003). The pathogenesis of cancer metastasis: the ‘seed and soil’ hypothesis revisited. Nat. Rev. Cancer 3, 453–458. Folkman, J. (2002). Role of angiogenesis in tumor growth and metastasis. Semin. Oncol. 29(6, Suppl 16), 15–18. Folkman, J. (2006). Angiogenesis. Annu. Rev. Med. 57, 1–18. Folkman, J., and Kalluri, R. (2004). Cancer without disease. Nature 427, 787.
Friedberg, E.C., Aguilera, A., Gellert, M., Hanawalt, P.C., Hays, J.B., Lehmann, A.R., Lindahl, T., Lowndes, N., Sarasin, A., and Wood, R.D. (2006). DNA repair: from molecular mechanism to human disease. DNA Repair (Amst.) 5, 986–996. Friedl, P., and Wolf, K. (2008). Tube travel: the role of proteases in individual and collective cancer cell invasion. Cancer Res. 68, 7247–7249. Friedl, P., and Wolf, K. (2010). Plasticity of cell migration: a multiscale tuning model. J. Cell Biol. 188, 11–19. Gaengel, K., Genove´, G., Armulik, A., and Betsholtz, C. (2009). Endothelialmural cell signaling in vascular development and angiogenesis. Arterioscler. Thromb. Vasc. Biol. 29, 630–638. Galluzzi, L., and Kroemer, G. (2008). Necroptosis: a specialized pathway of programmed necrosis. Cell 135, 1161–1163. Garzon, R., Marcucci, G., and Croce, C.M. (2010). Targeting microRNAs in cancer: rationale, strategies and challenges. Nat. Rev. Drug Discov. 9, 775– 789. Gerhardt, H., and Semb, H. (2008). Pericytes: gatekeepers in tumour cell metastasis? J. Mol. Med. 86, 135–144. Ghebranious, N., and Donehower, L.A. (1998). Mouse models in tumor suppression. Oncogene 17, 3385–3400. Giaccia, A.J., and Schipani, E. (2010). Role of carcinoma-associated fibroblasts and hypoxia in tumor progression. Curr. Top. Microbiol. Immunol. 345, 31–45. Gilbertson, R.J., and Rich, J.N. (2007). Making a tumour’s bed: glioblastoma stem cells and the vascular niche. Nat. Rev. Cancer 7, 733–736. Gocheva, V., Wang, H.W., Gadea, B.B., Shree, T., Hunter, K.E., Garfall, A.L., Berman, T., and Joyce, J.A. (2010). IL-4 induces cathepsin protease activity in tumor-associated macrophages to promote cancer growth and invasion. Genes Dev. 24, 241–255. Grivennikov, S.I., Greten, F.R., and Karin, M. (2010). Immunity, inflammation, and cancer. Cell 140, 883–899. Gupta, G.P., Minn, A.J., Kang, Y., Siegel, P.M., Serganova, I., Cordo´n-Cardo, C., Olshen, A.B., Gerald, W.L., and Massague´, J. (2005). Identifying sitespecific metastasis genes and functions. Cold Spring Harb. Symp. Quant. Biol. 70, 149–158. Gupta, P.B., Chaffer, C.L., and Weinberg, R.A. (2009). Cancer stem cells: mirage or reality? Nat. Med. 15, 1010–1012. Hanahan, D., and Folkman, J. (1996). Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86, 353–364.
Ikushima, H., and Miyazono, K. (2010). TGFbeta signalling: a complex web in cancer progression. Nat. Rev. Cancer 10, 415–424. Ince, T.A., Richardson, A.L., Bell, G.W., Saitoh, M., Godar, S., Karnoub, A.E., Iglehart, J.D., and Weinberg, R.A. (2007). Transformation of different human breast epithelial cell types leads to distinct tumor phenotypes. Cancer Cell 12, 160–170. Jackson, S.P., and Bartek, J. (2009). The DNA-damage response in human biology and disease. Nature 461, 1071–1078. Jiang, B.H., and Liu, L.Z. (2009). PI3K/PTEN signaling in angiogenesis and tumorigenesis. Adv. Cancer Res. 102, 19–65. Johansson, M., Denardo, D.G., and Coussens, L.M. (2008). Polarized immune responses differentially regulate cancer development. Immunol. Rev. 222, 145–154. Jones, P.A., and Baylin, S.B. (2007). The epigenomics of cancer. Cell 128, 683– 692. Jones, R.G., and Thompson, C.B. (2009). Tumor suppressors and cell metabolism: a recipe for cancer growth. Genes Dev. 23, 537–548. Joyce, J.A., and Pollard, J.W. (2009). Microenvironmental regulation of metastasis. Nat. Rev. Cancer 9, 239–252. Junttila, M.R., and Evan, G.I. (2009). p53—a Jack of all trades but master of none. Nat. Rev. Cancer 9, 821–829. Kalluri, R., and Zeisberg, M. (2006). Fibroblasts in cancer. Nat. Rev. Cancer 6, 392–401. Kang, H.J., Choi, Y.S., Hong, S.B., Kim, K.W., Woo, R.S., Won, S.J., Kim, E.J., Jeon, H.K., Jo, S.Y., Kim, T.K., et al. (2004). Ectopic expression of the catalytic subunit of telomerase protects against brain injury resulting from ischemia and NMDA-induced neurotoxicity. J. Neurosci. 24, 1280–1287. Karin, M., Lawrence, T., and Nizet, V. (2006). Innate immunity gone awry: linking microbial infections to chronic inflammation and cancer. Cell 124, 823–835. Karnoub, A.E., and Weinberg, R.A. (2006–2007). Chemokine networks and breast cancer metastasis. Breast Dis. 26, 75–85. Karnoub, A.E., Dash, A.B., Vo, A.P., Sullivan, A., Brooks, M.W., Bell, G.W., Richardson, A.L., Polyak, K., Tubo, R., and Weinberg, R.A. (2007). Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 449, 557–563. Kastan, M.B. (2008). DNA damage responses: mechanisms and roles in human disease: 2007 G.H.A. Clowes Memorial Award Lecture. Mol. Cancer Res. 6, 517–524.
Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cell 100, 57–70.
Kawai, T., Hiroi, S., Nakanishi, K., and Meeker, A.K. (2007). Telomere length and telomerase expression in atypical adenomatous hyperplasia and small bronchioloalveolar carcinoma of the lung. Am. J. Clin. Pathol. 127, 254–262.
Hansel, D.E., Meeker, A.K., Hicks, J., De Marzo, A.M., Lillemoe, K.D., Schulick, R., Hruban, R.H., Maitra, A., and Argani, P. (2006). Telomere length variation in biliary tract metaplasia, dysplasia, and carcinoma. Mod. Pathol. 19, 772–779.
Kazerounian, S., Yee, K.O., and Lawler, J. (2008). Thrombospondins in cancer. Cell. Mol. Life Sci. 65, 700–712.
Hardee, M.E., Dewhirst, M.W., Agarwal, N., and Sorg, B.S. (2009). Novel imaging provides new insights into mechanisms of oxygen transport in tumors. Curr. Mol. Med. 9, 435–441. Harper, J.W., and Elledge, S.J. (2007). The DNA damage response: Ten years after. Mol. Cell 28, 739–745. Hezel, A.F., and Bardeesy, N. (2008). LKB1; linking cell structure and tumor suppression. Oncogene 27, 6908–6919. Hlubek, F., Brabletz, T., Budczies, J., Pfeiffer, S., Jung, A., and Kirchner, T. (2007). Heterogeneous expression of Wnt/beta-catenin target genes within colorectal cancer. Int. J. Cancer 121, 1941–1948. Hugo, H., Ackland, M.L., Blick, T., Lawrence, M.G., Clements, J.A., Williams, E.D., and Thompson, E.W. (2007). Epithelial—mesenchymal and mesenchymal—epithelial transitions in carcinoma progression. J. Cell. Physiol. 213, 374–383. Hsu, P.P., and Sabatini, D.M. (2008). Cancer cell metabolism: Warburg and beyond. Cell 134, 703–707. Hynes, N.E., and MacDonald, G. (2009). ErbB receptors and signaling pathways in cancer. Curr. Opin. Cell Biol. 21, 177–184.
Kenific, C.M., Thorburn, A., and Debnath, J. (2010). Autophagy and metastasis: another double-edged sword. Curr. Opin. Cell Biol. 22, 241–245. Kennedy, K.M., and Dewhirst, M.W. (2010). Tumor metabolism of lactate: the influence and therapeutic potential for MCT and CD147 regulation. Future Oncol. 6, 127–148. Kessenbrock, K., Plaks, V., and Werb, Z. (2010). Matrix metalloproteinases: Regulators of the tumor microenvironment. Cell 141, 52–67. Kim, M.Y., Oskarsson, T., Acharyya, S., Nguyen, D.X., Zhang, X.H., Norton, L., and Massague´, J. (2009). Tumor self-seeding by circulating cancer cells. Cell 139, 1315–1326. Kim, R., Emi, M., and Tanabe, K. (2007). Cancer immunoediting from immune surveillance to immune escape. Immunology 121, 1–14. Kinzler, K.W., and Vogelstein, B. (1997). Cancer-susceptibility genes. Gatekeepers and caretakers. Nature 386, 761–763. Klein, C.A. (2009). Parallel progression of primary tumours and metastases. Nat. Rev. Cancer 9, 302–312. Klymkowsky, M.W., and Savagner, P. (2009). Epithelial-mesenchymal transition: a cancer researcher’s conceptual friend and foe. Am. J. Pathol. 174, 1588–1593.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 671
Korkola, J., and Gray, J.W. (2010). Breast cancer genomes—form and function. Curr. Opin. Genet. Dev. 20, 4–14.
Mooi, W.J., and Peeper, D.S. (2006). Oncogene-induced cell senescence— halting on the road to cancer. N. Engl. J. Med. 355, 1037–1046.
Kovacic, J.C., and Boehm, M. (2009). Resident vascular progenitor cells: an emerging role for non-terminally differentiated vessel-resident cells in vascular biology. Stem Cell Res. (Amst.) 2, 2–15.
Morel, A.-P., Lie`vre, M., Thomas, C., Hinkal, G., Ansieau, S., and Puisieux, A. (2008). Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS ONE 3, e2888.
Kroemer, G., and Pouyssegur, J. (2008). Tumor cell metabolism: Cancer’s Achilles’ heel. Cancer Cell 13, 472–482.
Mosesson, Y., Mills, G.B., and Yarden, Y. (2008). Derailed endocytosis: an emerging feature of cancer. Nat. Rev. Cancer 8, 835–850.
Lamagna, C., and Bergers, G. (2006). The bone marrow constitutes a reservoir of pericyte progenitors. J. Leukoc. Biol. 80, 677–681.
Mougiakakos, D., Choudhury, A., Lladser, A., Kiessling, R., and Johansson, C.C. (2010). Regulatory T cells in cancer. Adv. Cancer Res. 107, 57–117.
Lane, D.P. (1992). Cancer. p53, guardian of the genome. Nature 358, 15–16.
Murdoch, C., Muthana, M., Coffelt, S.B., and Lewis, C.E. (2008). The role of myeloid cells in the promotion of tumour angiogenesis. Nat. Rev. Cancer 8, 618–631.
Lemmon, M.A., and Schlessinger, J. (2010). Cell signaling by receptor tyrosine kinases. Cell 141, 1117–1134. Levine, B., and Kroemer, G. (2008). Autophagy in the pathogenesis of disease. Cell 132, 27–42. Lipinski, M.M., and Jacks, T. (1999). The retinoblastoma gene family in differentiation and development. Oncogene 18, 7873–7882. Lobo, N.A., Shimono, Y., Qian, D., and Clarke, M.F. (2007). The biology of cancer stem cells. Annu. Rev. Cell Dev. Biol. 23, 675–699. Lowe, S.W., Cepero, E., and Evan, G. (2004). Intrinsic tumour suppression. Nature 432, 307–315. Luebeck, E.G. (2010). Cancer: Genomic evolution of metastasis. Nature 467, 1053–1055. Lu, Z., Luo, R.Z., Lu, Y., Zhang, X., Yu, Q., Khare, S., Kondo, S., Kondo, Y., Yu, Y., Mills, G.B., et al. (2008). The tumor suppressor gene ARHI regulates autophagy and tumor dormancy in human ovarian cancer cells. J. Clin. Invest. 118, 3917–3929. Luo, J., Solimini, N.L., and Elledge, S.J. (2009). Principles of cancer therapy: Oncogene and non-oncogene addiction. Cell 136, 823–837. Mac Gabhann, F., and Popel, A.S. (2008). Systems biology of vascular endothelial growth factors. Microcirculation 15, 715–738. Madsen, C.D., and Sahai, E. (2010). Cancer dissemination—Lessons from leukocytes. Dev. Cell 19, 13–26. Maida, Y., Yasukawa, M., Furuuchi, M., Lassmann, T., Possemato, R., Okamoto, N., Kasim, V., Hayashizaki, Y., Hahn, W.C., and Masutomi, K. (2009). An RNA-dependent RNA polymerase formed by TERT and the RMRP RNA. Nature 461, 230–235. Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks, M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704–715. Mantovani, A. (2010). Molecular pathways linking inflammation and cancer. Curr. Mol. Med. 10, 369–373. Mantovani, A., Allavena, P., Sica, A., and Balkwill, F. (2008). Cancer-related inflammation. Nature 454, 436–444. Massague´, J. (2008). TGFb in cancer. Cell 134, 215–230. Masutomi, K., Possemato, R., Wong, J.M., Currier, J.L., Tothova, Z., Manola, J.B., Ganesan, S., Lansdorp, P.M., Collins, K., and Hahn, W.C. (2005). The telomerase reverse transcriptase regulates chromatin state and DNA damage responses. Proc. Natl. Acad. Sci. USA 102, 8222–8227.
Nagy, J.A., Chang, S.H., Shih, S.C., Dvorak, A.M., and Dvorak, H.F. (2010). Heterogeneity of the tumor vasculature. Semin. Thromb. Hemost. 36, 321–331. Naumov, G.N., Folkman, J., Straume, O., and Akslen, L.A. (2008). Tumorvascular interactions and tumor dormancy. APMIS 116, 569–585. Negrini, S., Gorgoulis, V.G., and Halazonetis, T.D. (2010). Genomic instability—an evolving hallmark of cancer. Nat. Rev. Mol. Cell Biol. 11, 220–228. Nelson, B.H. (2008). The impact of T-cell immunity on ovarian cancer outcomes. Immunol. Rev. 222, 101–116. Nguyen, D.X., Bos, P.D., and Massague´, J. (2009). Metastasis: from dissemination to organ-specific colonization. Nat. Rev. Cancer 9, 274–284. Norden, A.D., Drappatz, J., and Wen, P.Y. (2009). Antiangiogenic therapies for high-grade glioma. Nat. Rev. Neurol. 5, 610–620. Nyberg, P., Xie, L., and Kalluri, R. (2005). Endogenous inhibitors of angiogenesis. Cancer Res. 65, 3967–3979. Okada, T., Lopez-Lago, M., and Giancotti, F.G. (2005). Merlin/NF-2 mediates contact inhibition of growth by suppressing recruitment of Rac to the plasma membrane. J. Cell Biol. 171, 361–371. Olive, K.P., Jacobetz, M.A., Davidson, C.J., Gopinathan, A., McIntyre, D., Honess, D., Madhu, B., Goldgraben, M.A., Caldwell, M.E., Allard, D., et al. (2009). Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 324, 1457–1461. Olson, P., Lu, J., Zhang, H., Shai, A., Chun, M.G., Wang, Y., Libutti, S.K., Nakakura, E.K., Golub, T.R., and Hanahan, D. (2009). MicroRNA dynamics in the stages of tumorigenesis correlate with hallmark capabilities of cancer. Genes Dev. 23, 2152–2165. O’Reilly, K.E., Rojo, F., She, Q.B., Solit, D., Mills, G.B., Smith, D., Lane, H., Hofmann, F., Hicklin, D.J., Ludwig, D.L., et al. (2006). mTOR inhibition induces upstream receptor tyrosine kinase signaling and activates Akt. Cancer Res. 66, 1500–1508. Ostrand-Rosenberg, S., and Sinha, P. (2009). Myeloid-derived suppressor cells: linking inflammation and cancer. J. Immunol. 182, 4499–4506. Page`s, F., Galon, J., Dieu-Nosjean, M.C., Tartour, E., Saute`s-Fridman, C., and Fridman, W.H. (2010). Immune infiltration in human tumors: a prognostic factor that should not be ignored. Oncogene 29, 1093–1102. Palermo, C., and Joyce, J.A. (2008). Cysteine cathepsin proteases as pharmacological targets in cancer. Trends Pharmacol. Sci. 29, 22–28.
Mathew, R., Karantza-Wadsworth, V., and White, E. (2007). Role of autophagy in cancer. Nat. Rev. Cancer 7, 961–967.
Park, J.I., Venteicher, A.S., Hong, J.Y., Choi, J., Jun, S., Shkreli, M., Chang, W., Meng, Z., Cheung, P., Ji, H., et al. (2009). Telomerase modulates Wnt signalling by association with target gene chromatin. Nature 460, 66–72.
McGowan, P.M., Kirstein, J.M., and Chambers, A.F. (2009). Micrometastatic disease and metastatic outgrowth: clinical issues and experimental approaches. Future Oncol. 5, 1083–1098.
Partanen, J.I., Nieminen, A.I., and Klefstrom, J. (2009). 3D view to tumor suppression: Lkb1, polarity and the arrest of oncogenic c-Myc. Cell Cycle 8, 716–724.
Micalizzi, D.S., Farabaugh, S.M., and Ford, H.L. (2010). Epithelial-mesenchymal transition in cancer: parallels between normal development and tumor progression. J. Mammary Gland Biol. Neoplasia 15, 117–134.
Pasquale, E.B. (2010). Eph receptors and ephrins in cancer: bidirectional signalling and beyond. Nat. Rev. Cancer 10, 165–180.
Mizushima, N. (2007). Autophagy: process and function. Genes Dev. 21, 2861– 2873.
Passos, J.F., Saretzki, G., and von Zglinicki, T. (2007). DNA damage in telomeres and mitochondria during cellular senescence: is there a connection? Nucleic Acids Res. 35, 7505–7513.
Mohamed, M.M., and Sloane, B.F. (2006). Cysteine cathepsins: multifunctional enzymes in cancer. Nat. Rev. Cancer 6, 764–775.
Patenaude, A., Parker, J., and Karsan, A. (2010). Involvement of endothelial progenitor cells in tumor vascularization. Microvasc. Res. 79, 217–223.
672 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
Peinado, H., Lavothskin, S., and Lyden, D. (2011). The secreted factors responsible for pre-metastatic niche formation: Old sayings and new thoughts. Semin. Cancer Biol. Published online January 18, 2011. 10.1016/j.semcancer. 2011.01.002. Peinado, H., Marin, F., Cubillo, E., Stark, H.J., Fusenig, N., Nieto, M.A., and Cano, A. (2004). Snail and E47 repressors of E-cadherin induce distinct invasive and angiogenic properties in vivo. J. Cell Sci. 117, 2827–2839. Perona, R. (2006). Cell signalling: growth factors and tyrosine kinase receptors. Clin. Transl. Oncol. 8, 77–82. Pietras, K., and Ostman, A. (2010). Hallmarks of cancer: interactions with the tumor stroma. Exp. Cell Res. 316, 1324–1331. Polyak, K., and Weinberg, R.A. (2009). Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat. Rev. Cancer 9, 265–273. Polyak, K., Haviv, I., and Campbell, I.G. (2009). Co-evolution of tumor cells and their microenvironment. Trends Genet. 25, 30–38. Potter, V.R. (1958). The biochemical approach to the cancer problem. Fed. Proc. 17, 691–697. Qian, B.Z., and Pollard, J.W. (2010). Macrophage diversity enhances tumor progression and metastasis. Cell 141, 39–51. Quintana, E., Shackleton, M., Sabel, M.S., Fullen, D.R., Johnson, T.M., and Morrison, S.J. (2008). Efficient tumour formation by single human melanoma cells. Nature 456, 593–598. Raica, M., Cimpean, A.M., and Ribatti, D. (2009). Angiogenesis in pre-malignant conditions. Eur. J. Cancer 45, 1924–1934. Ra¨sa¨nen, K., and Vaheri, A. (2010). Activation of fibroblasts in cancer stroma. Exp. Cell Res. 316, 2713–2722. Raynaud, C.M., Hernandez, J., Llorca, F.P., Nuciforo, P., Mathieu, M.C., Commo, F., Delaloge, S., Sabatier, L., Andre´, F., and Soria, J.C. (2010). DNA damage repair and telomere length in normal breast, preneoplastic lesions, and invasive cancer. Am. J. Clin. Oncol. 33, 341–345. Raza, A., Franklin, M.J., and Dudek, A.Z. (2010). Pericytes and vessel maturation during tumor angiogenesis and metastasis. Am. J. Hematol. 85, 593–598. Reitman, Z.J., and Yan, H. (2010). Isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism. J. Natl. Cancer Inst. 102, 932–941. Reya, T., Morrison, S.J., Clarke, M.F., and Weissman, I.L. (2001). Stem cells, cancer, and cancer stem cells. Nature 414, 105–111. Ribatti, D. (2009). Endogenous inhibitors of angiogenesis: a historical review. Leuk. Res. 33, 638–644. Ricci-Vitiani, L., Pallini, R., Biffoni, M., Todaro, M., Invernici, G., Cenci, T., Maira, G., Parati, E.A., Stassi, G., Larocca, L.M., and De Maria, R. (2010). Tumour vascularization via endothelial differentiation of glioblastoma stemlike cells. Nature 468, 824–828. Ruoslahti, E. (2002). Specialization of tumour vasculature. Nat. Rev. Cancer 2, 83–90. Ruoslahti, E., Bhatia, S.N., and Sailor, M.J. (2010). Targeting of drugs and nanoparticles to tumors. J. Cell Biol. 188, 759–768. Sabeh, F., Shimizu-Hirota, R., and Weiss, S.J. (2009). Protease-dependent versus -independent cancer cell invasion programs: three-dimensional amoeboid movement revisited. J. Cell Biol. 185, 11–19. Salk, J.J., Fox, E.J., and Loeb, L.A. (2010). Mutational heterogeneity in human cancers: origin and consequences. Ann. Rev. Pathol. 5, 51–75. Scha¨fer, M., and Werner, S. (2008). Cancer as an overhealing wound: an old hypothesis revisited. Nat. Rev. Mol. Cell Biol. 9, 628–638. Schmalhofer, O., Brabletz, S., and Brabletz, T. (2009). E-cadherin, beta-catenin, and ZEB1 in malignant progression of cancer. Cancer Metastasis Rev. 28, 151–166.
Semenza, G.L. (2010b). Defining the role of hypoxia-inducible factor 1 in cancer biology and therapeutics. Oncogene 29, 625–634. Seppinen, L., Sormunen, R., Soini, Y., Elamaa, H., Heljasvaara, R., and Pihlajaniemi, T. (2008). Lack of collagen XVIII accelerates cutaneous wound healing, while overexpression of its endostatin domain leads to delayed healing. Matrix Biol. 27, 535–546. Shaw, R.J. (2009). Tumor suppression by LKB1: SIK-ness prevents metastasis. Sci. Signal. 2, pe55. Shay, J.W., and Wright, W.E. (2000). Hayflick, his limit, and cellular ageing. Nat. Rev. Mol. Cell Biol. 1, 72–76. Sherr, C.J., and DePinho, R.A. (2000). Cellular senescence: Mitotic clock or culture shock? Cell 102, 407–410. Sherr, C.J., and McCormick, F. (2002). The RB and p53 pathways in cancer. Cancer Cell 2, 103–112. Shields, J.D., Kourtis, I.C., Tomei, A.A., Roberts, J.M., and Swartz, M.A. (2010). Induction of lymphoidlike stroma and immune escape by tumors that express the chemokine CCL21. Science 328, 749–752. Shimoda, M., Mellody, K.T., and Orimo, A. (2010). Carcinoma-associated fibroblasts are a rate-limiting determinant for tumour progression. Semin. Cell Dev. Biol. 21, 19–25. Sigal, A., and Rotter, V. (2000). Oncogenic mutations of the p53 tumor suppressor: the demons of the guardian of the genome. Cancer Res. 60, 6788–6793. Singh, A., and Settleman, J. (2010). EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 29, 4741–4751. Sinha, S., and Levine, B. (2008). The autophagy effector Beclin 1: a novel BH3only protein. Oncogene 27 (Suppl 1), S137–S148. Smyth, M.J., Dunn, G.P., and Schreiber, R.D. (2006). Cancer immunosurveillance and immunoediting: the roles of immunity in suppressing tumor development and shaping tumor immunogenicity. Adv. Immunol. 90, 1–50. Soda, Y., Marumoto, T., Friedmann-Morvinski, D., Soda, M., Liu, F., Michiue, H., Pastorino, S., Yang, M., Hoffman, R.M., Kesari, S., and Verma, I.M. (2011). Feature Article: Transdifferentiation of glioblastoma cells into vascular endothelial cells. Proc. Natl. Acad. Sci. USA. Published online January 24, 2011. Sudarsanam, S., and Johnson, D.E. (2010). Functional consequences of mTOR inhibition. Curr. Opin. Drug Discov. Devel. 13, 31–40. Strauss, D.C., and Thomas, J.M. (2010). Transmission of donor melanoma by organ transplantation. Lancet Oncol. 11, 790–796. Talmadge, J.E., and Fidler, I.J. (2010). AACR centennial series: the biology of cancer metastasis: historical perspective. Cancer Res. 70, 5649–5669. Tammela, T., and Alitalo, K. (2010). Lymphangiogenesis: Molecular mechanisms and future promise. Cell 140, 460–476. Taube, J.H., Herschkowitz, J.I., Komurov, K., Zhou, A.Y., Gupta, S., Yang, J., Hartwell, K., Onder, T.T., Gupta, P.B., Evans, K.W., et al. (2010). Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc. Natl. Acad. Sci. USA 107, 15449–15454. Teng, M.W.L., Swann, J.B., Koebel, C.M., Schreiber, R.D., and Smyth, M.J. (2008). Immune-mediated dormancy: an equilibrium with cancer. J. Leukoc. Biol. 84, 988–993. Thiery, J.P., and Sleeman, J.P. (2006). Complex networks orchestrate epithelial-mesenchymal transitions. Nat. Rev. Mol. Cell Biol. 7, 131–142. Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009). Epithelialmesenchymal transitions in development and disease. Cell 139, 871–890. Townson, J.L., and Chambers, A.F. (2006). Dormancy of solitary metastatic cells. Cell Cycle 5, 1744–1750.
Semenza, G.L. (2008). Tumor metabolism: cancer cells give and take lactate. J. Clin. Invest. 118, 3835–3837.
Turner, H.E., Harris, A.L., Melmed, S., and Wass, J.A. (2003). Angiogenesis in endocrine tumors. Endocr. Rev. 24, 600–632.
Semenza, G.L. (2010a). HIF-1: upstream and downstream of cancer metabolism. Curr. Opin. Genet. Dev. 20, 51–56.
Vajdic, C.M., and van Leeuwen, M.T. (2009). Cancer incidence and risk factors after solid organ transplantation. Int. J. Cancer 125, 1747–1754.
Cell 144, March 4, 2011 ª2011 Elsevier Inc. 673
Vander Heiden, M.G., Cantley, L.C., and Thompson, C.B. (2009). Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033. Verhoeff, J.J., van Tellingen, O., Claes, A., Stalpers, L.J., van Linde, M.E., Richel, D.J., Leenders, W.P., and van Furth, W.R. (2009). Concerns about anti-angiogenic treatment in patients with glioblastoma multiforme. BMC Cancer 9, 444. Wang, R., Chadalavada, K., Wilshire, J., Kowalik, U., Hovinga, K.E., Geber, A., Fligelman, B., Leversha, M., Brennan, C., and Tabar, V. (2010). Glioblastoma stem-like cells give rise to tumour endothelium. Nature 468, 829–833. Warburg, O. (1956a). On the origin of cancer cells. Science 123, 309–314. Warburg, O. (1956b). On respiratory impairment in cancer cells. Science 124, 269–270. Warburg, O.H. (1930). The Metabolism of Tumours: Investigations from the Kaiser Wilhelm Institute for Biology, Berlin-Dahlem (London, UK: Arnold Constable). Wertz, I.E., and Dixit, V.M. (2010). Regulation of death receptor signaling by the ubiquitin system. Cell Death Differ. 17, 14–24. White, E., Karp, C., Strohecker, A.M., Guo, Y., and Mathew, R. (2010). Role of autophagy in suppression of inflammation and cancer. Curr. Opin. Cell Biol. 22, 212–217. White, E., and DiPaola, R.S. (2009). The double-edged sword of autophagy modulation in cancer. Clin. Cancer Res. 15, 5308–5316.
phage-assisted tumor cell intravasation in mammary tumors. Cancer Res. 67, 2649–2656. Yachida, S., Jones, S., Bozic, I., Antal, T., Leary, R., Fu, B., Kamiyama, M., Hruban, R.H., Eshleman, J.R., Nowak, M.A., et al. (2010). Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 467, 1114–1117. Yang, J., and Weinberg, R.A. (2008). Epithelial-mesenchymal transition: At the crossroads of development and tumor metastasis. Dev. Cell 14, 818–829. Yang, L., Pang, Y., and Moses, H.L. (2010). TGF-beta and immune cells: an important regulatory axis in the tumor microenvironment and progression. Trends Immunol. 31, 220–227. Yen, K.E., Bittinger, M.A., Su, S.M., and Fantin, V.R. (2010). Cancer-associated IDH mutations: biomarker and therapeutic opportunities. Oncogene 29, 6409–6417. Yilmaz, M., and Christofori, G. (2009). EMT, the cytoskeleton, and cancer cell invasion. Cancer Metastasis Rev. 28, 15–33. Yuan, T.L., and Cantley, L.C. (2008). PI3K pathway alterations in cancer: variations on a theme. Oncogene 27, 5497–5510. Zee, Y.K., O’Connor, J.P., Parker, G.J., Jackson, A., Clamp, A.R., Taylor, M.B., Clarke, N.W., and Jayson, G.C. (2010). Imaging angiogenesis of genitourinary tumors. Nat. Rev. Urol. 7, 69–82.
Willis, S.N., and Adams, J.M. (2005). Life in the balance: how BH3-only proteins induce apoptosis. Curr. Opin. Cell Biol. 17, 617–625.
Zhang, H., Herbert, B.S., Pan, K.H., Shay, J.W., and Cohen, S.N. (2004). Disparate effects of telomere attrition on gene expression during replicative senescence of human mammary epithelial cells cultured under different conditions. Oncogene 23, 6193–6198.
Witsch, E., Sela, M., and Yarden, Y. (2010). Roles for growth factors in cancer progression. Physiology (Bethesda) 25, 85–101.
Zong, W.X., and Thompson, C.B. (2006). Necrotic death as a cell fate. Genes Dev. 20, 1–15.
Wyckoff, J.B., Wang, Y., Lin, E.Y., Li, J.F., Goswami, S., Stanley, E.R., Segall, J.E., Pollard, J.W., and Condeelis, J. (2007). Direct visualization of macro-
Zumsteg, A., and Christofori, G. (2009). Corrupt policemen: inflammatory cells promote tumor angiogenesis. Curr. Opin. Oncol. 21, 60–70.
674 Cell 144, March 4, 2011 ª2011 Elsevier Inc.
TLR Signaling Is Required for Salmonella typhimurium Virulence Nicholas Arpaia,1 Jernej Godec,1 Laura Lau,1 Kelsey E. Sivick,1 Laura M. McLaughlin,2 Marcus B. Jones,3 Tatiana Dracheva,3 Scott N. Peterson,3 Denise M. Monack,2 and Gregory M. Barton1,* 1Division of Immunology & Pathogenesis, Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720-3200, USA 2Department of Microbiology and Immunology, Stanford University, Stanford, CA 94305, USA 3Pathogen Functional Genomics Resource Center, J. Craig Venter Institute, Rockville, MD 20850, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.01.031
SUMMARY
Toll-like receptors (TLRs) contribute to host resistance to microbial pathogens and can drive the evolution of virulence mechanisms. We have examined the relationship between host resistance and pathogen virulence using mice with a functional allele of the nramp-1 gene and lacking combinations of TLRs. Mice deficient in both TLR2 and TLR4 were highly susceptible to the intracellular bacterial pathogen Salmonella typhimurium, consistent with reduced innate immune function. However, mice lacking additional TLRs involved in S. typhimurium recognition were less susceptible to infection. In these TLR-deficient cells, bacteria failed to upregulate Salmonella pathogenicity island 2 (SPI-2) genes and did not form a replicative compartment. We demonstrate that TLR signaling enhances the rate of acidification of the Salmonella-containing phagosome, and inhibition of this acidification prevents SPI-2 induction. Our results indicate that S. typhimurium requires cues from the innate immune system to regulate virulence genes necessary for intracellular survival, growth, and systemic infection. INTRODUCTION During early stages of infection the innate immune system is essential for limiting microbial replication and spread before an adaptive response is mounted. Accordingly, pathogens have evolved virulence strategies to antagonize innate immune function (Hedrick, 2004; Rausher, 2001; Woolhouse et al., 2002). The interplay between host innate immune function and pathogen virulence mechanisms largely determines the outcome of most infections. Despite the logic of this conceptual framework, our understanding of the molecular interactions driving the emergence of virulence mechanisms remains relatively poor. Innate immune receptors detect infection by recognizing conserved microbial features common to broad classes of
microbes (Janeway, 1989; Medzhitov, 2007). The Toll-like receptors (TLRs) target a range of microbial ligands, including lipopolysaccharide (TLR4), lipoproteins (TLR2), flagellin (TLR5), unmethylated CpG motifs in DNA (TLR9), doublestranded RNA (TLR3), and single-stranded RNA (TLR7 and TLR8) (Akira et al., 2001; Kawai and Akira, 2005). Expression of TLRs on innate immune cells links microbial recognition to induction of antimicrobial mechanisms, such as production of reactive oxygen species (ROS) and reactive nitrogen species (RNS) and expression of antimicrobial peptides (AMPs). In addition, TLR activation can promote adaptive immunity through control of dendritic cell (DC) maturation (Iwasaki and Medzhitov, 2004). To study the evolution of pathogen virulence and its relationship to innate immunity, we have focused on TLR-mediated recognition of Salmonella enterica serovar typhimurium. S. typhimurium is a gram-negative bacterium that can survive and replicate within host macrophages (Coburn et al., 2007). Survival within macrophages requires a set of genes, many of which are encoded within Salmonella pathogenicity island 2 (SPI-2) (Galan, 2001; Shea et al., 1996; Waterman and Holden, 2003). SPI-2 encodes a type 3 secretion system (T3SS) that is expressed after the bacterium is phagocytosed (Cirillo et al., 1998; Pfeifer et al., 1999; Valdivia and Falkow, 1997). Translocation of SPI-2 effectors into the host cell transforms the phagosome into a compartment that supports bacterial replication, the Salmonella-containing vacuole (SCV) (Marcus et al., 2000). Multiple signals have been implicated in the transcriptional induction of SPI-2, including cation deprivation, phosphate starvation, and low pH (Chakravortty et al., 2005; Cirillo et al., 1998; Deiwick et al., 1999; Kim and Falkow, 2004; Rappl et al., 2003). Most of the studies implicating these signals have been performed on bacteria grown in vitro; whether the same signals are responsible for induction of SPI-2 genes within the phagosome remains unclear. Recognition of S. typhimurium is largely mediated by TLR2, TLR4, and TLR5 (Feuillet et al., 2006; Hapfelmeier et al., 2005; O’Brien et al., 1980; Royle et al., 2003; Smith et al., 2003; Uematsu et al., 2006; Vazquez-Torres et al., 2004). Consistent with a central role for these receptors, S. typhimurium has evolved mechanisms to subvert this recognition or to avoid the consequences of TLR activation. For example, modification of Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 675
A
HEK293 (endogenous TLR5)
Fold NF-κB Induction
30
30
20
20
10
10 0
0
BLP Fla
Fla
Fold NF-κB Induction
40
MOI
MOI
20
20
10
10 0
LPS Fla
MOI
MOI
C
40 30
MOI
CpG Fla
100
20
S.t. S.t. MOI 5 MOI 10
LPS
CpG
Side Scatter
Unstim.
Uninfected
100
0.38
Wildtype
80
DMSO
60
40
40
20
20
20
0
0
100
101
102
TLR2x4 KO
103
100
16.9
80
100
101
14.5
80
Unstim. CpG
676 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
100
60
40
40
40
20
20
20
0
100
101
102
103
0
100
101
R848
S.t. MOI 5
102
103
100
101
72.5
102
0
103
MyD88xTRIF KO
0.7
80
60
Bafilomycin A1
S.t. MOI 5
0
60
5
R848
103
TLR2x4x9 KO
WT TLR 2x4 KO TLR 2x4x9 KO MyD88xTRIF KO
Unstim. CpG
102
TLR2 KO
80
60
100 75 50 25
0
100
40
TNF
D
81.2
60
100
0
genomic DNA
80
WT TLR 2x4 KO TLR 2x4x9 KO MyD88xTRIF KO
10
% TNF-positive cells
Purified Ligand S.t. S.t.ΔfljB/fliC
30
B 50
MOI
HEK293-TLR9N4C
HEK293-TLR4/MD-2/CD14
30
0
Nitrite (uM)
HEK293-TLR2/CD14
100
101
102
103
lipid A by pagP reduces recognition by TLR4, although this modification is probably most relevant for resistance to AMPs (Bader et al., 2005; Detweiler et al., 2003; Guo et al., 1997, 1998). Mice lacking TLRs, especially TLR4, are more susceptible to S. typhimurium (Weiss et al., 2004). To circumvent the problem of redundancy, many studies have used mice lacking the common TLR adaptor MyD88 or lacking both MyD88 and another adaptor, TRIF (Hapfelmeier et al., 2005; Weiss et al., 2004). Although these mice are very susceptible to S. typhimurium, these studies suffer from the caveat that MyD88 is also required for signaling by members of the IL-1 receptor (IL-1R) family. Because mice deficient in IL-1R are more susceptible to infection, the phenotype of MyD88 knockout (KO) mice cannot be unequivocally attributed to TLRs (Mayer-Barber et al., 2010; Raupach et al., 2006). In the studies described here, we sought to eliminate TLRbased recognition of S. typhimurium and examine the effect on pathogen virulence, while avoiding the caveats associated with MyD88-KO mice. In addition, we were concerned that the extreme susceptibility of C57Bl/6 mice (the genetic background on which most studies with TLR-KO mice have been performed) to S. typhimurium infection might mask any relationships between TLRs and bacterial virulence strategies. Many inbred mouse strains, including C57Bl/6, possess a nonfunctional allele of the nramp-1 gene. nramp-1 encodes a multipass transmembrane protein that localizes to lysosomes and functions as a transporter of divalent cations, and mice with the nonfunctional allele are extremely susceptible to a number of intracellular pathogens (Forbes and Gros, 2001; Govoni et al., 1996; Vidal et al., 1993, 1995, 1996). To avoid the caveats associated with nonfunctional Nramp-1 and TLR-independent functions of MyD88, we generated mice with a functional allele of nramp1 that lack individual or multiple TLRs. Studies in these mice led to a striking finding. Whereas mice lacking a subset of the TLRs involved in S. typhimurium recognition showed increased susceptibility to infection, a lack of additional TLRs resulted in reduced susceptibility. The loss of virulence correlated with an inability of bacteria to survive and replicate within macrophages. We show that TLR signaling leads to rapid acidification of the SCV, and this signal is required for regulation of virulence gene expression. In the absence of this contextual cue, S. typhimurium is unable to survive and replicate intracellularly. Altogether, this work describes the molecular interactions underlying a bacterial pathogen’s dependence on the innate immune system for virulence.
RESULTS Multiple TLRs Are Involved in Recognition of S. typhimurium To identify which TLRs are relevant for innate recognition of S. typhimurium, we utilized HEK293 reporter cell lines expressing an NF-kB-luciferase reporter construct. Stimulation of these cells with heat-killed bacteria resulted in robust induction of NF-kB, which we attributed to endogenous TLR5 expressed by these cells (Figure 1A). This response was abrogated when cells were stimulated with bacteria lacking flagellin. To measure activation of other TLR family members, HEK293 reporter cells stably expressing individual TLRs were stimulated with bacteria lacking flagellin (to eliminate the contribution of endogenous TLR5). Using this approach, we observed activation of TLR2 and TLR4 by S. typhimurium (Figure 1A). Furthermore, S. typhimurium genomic DNA was capable of activating a surface-localized version of TLR9 (Figure 1A). Although these results confirmed that TLR2, TLR4, TLR5, and TLR9 may play a role in recognition of S. typhimurium, they did not address the relative importance of each TLR during infection. To this end, we infected bone marrow-derived macrophages (BMMs) lacking combinations of TLRs and measured production of nitric oxide (NO). In agreement with previously published studies, BMMs lacking both TLR2 and TLR4 (TLR234-KO) produced much less NO than wild-type BMMs (Figure 1B). The remaining response was partially dependent on TLR9, as BMMs lacking TLR2, TLR4, and TLR9 (TLR23439-KO) produced even less NO. Similar results were observed when tumor necrosis factor alpha (TNF) production was measured (Figure 1C). Importantly, all genotypes of BMMs responded equivalently to the TLR7 ligand R848, indicating that the cells were otherwise equivalent (Figure 1D and Figure S1 available online). The small amount of TNF and NO produced in TLR23439-KO BMMs was dependent on other TLRs, as BMMs lacking both MyD88 and TRIF (and therefore all TLRdependent signaling) did not respond to S. typhimurium (Figures 1B and 1C). As TLR5 is not expressed in murine BMMs, we reasoned that the residual TNF and NO produced by TLR23439-KO BMMs was most likely due to TLR7 or TLR3 signaling. To address this possibility directly, we pretreated TLR23439-KO BMMs with bafilomycinA1, an inhibitor of the vacuolar ATPase (V-ATPase) that prevents activation of endosomal TLRs. BafilomycinA1 treatment inhibited TNF production in TLR234-KO and TLR23439-KO BMMs to almost background
Figure 1. Multiple TLRs Recognize Products of S. typhimurium (A) HEK293 cells expressing the indicated TLRs or TLR accessory proteins together with an NF-kB luciferase reporter were treated with heat-killed wild-type LT2 S. typhimurium (S.t.), flagellin-deficient LT2 (S.t.DfljB/fliC), or genomic DNA isolated from flagellin-deficient LT2. Luciferase activity was measured after 8 hr. Relative moi range: 100 to 3.125, DNA concentration: 375 ng/ml to 12.5 ng/ml. Data are representative of two independent experiments and shown as mean ± standard error of the mean (SEM). LPS, lipopolysaccharide; Fla, flagellin; BLP, bacterial lipopeptide. (B) BMMs differentiated from the indicated mice were treated overnight with 100 U/ml recombinant IFN-g and infected the next morning with wild-type S. typhimurium (SL1344) at the indicated moi. Nitrite production was measured 36 hr post-infection by Griess assay. Data are representative of three independent experiments and presented as mean ± standard deviation (SD). (C) BMMs were infected as in (B) at an moi of 5 for 8 hr, followed by intracellular cytokine staining for TNF. Percent TNF-positive cells are indicated in each panel. Data are representative of three independent experiments. (D) BMMs pretreated for 2 hr with bafilomycinA1 or DMSO vehicle were infected with S. typhimurium (SL1344) at an moi of 5 for 6 hr. Cells were processed and stained for intracellular cytokine staining as in (C).
Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 677
levels, suggesting that TLR7 and/or TLR3 are responsible for the remaining TNF production in response to S. typhimurium (Figure 1D). Collectively, these data indicate that TLR2, TLR4, TLR9, and TLR7 (and/or TLR3) each contribute to the recognition of S. typhimurium in infected BMMs. TLR Signaling Is Required for S. typhimurium Virulence Having established which TLRs respond to ligands derived from S. typhiumurium in BMMs, we sought to test the effect of TLR deficiency on bacterial virulence in vivo. We crossed a functional allele of the nramp1 gene onto the C57BL/6 background and generated TLR-deficient or TLR-adaptor-deficient mice with functional Nramp1 (see Extended Experimental Procedures). We expected that reduced TLR function would lead to greater susceptibility to infection. Indeed, all TLR234-KO mice died within 16 days when challenged orally with S. typhimurium, whereas 75% of the wild-type mice survived for the duration of the experiment (Figure 2A). By contrast, TLR23439-KO mice were less susceptible to infection than TLR234-KO mice, despite a greater impairment in TLR function (Figure 2A). This increased survival was not a consequence of reduced immunopathology due to reduced TLR function. In fact, TLR23439-KO mice had lower numbers of bacteria 4 days post-infection in spleens, livers, ceca, and mesenteric lymph nodes (MLNs) relative to TLR234-KO mice (Figure 2B). Thus, despite less robust innate immune function, S. typhimurium was less virulent in TLR23439-KO mice. The difference in susceptibility between TLR234-KO and TLR23439-KO mice could indicate that TLR9 plays a negative role in immunity to S. typhimurium. To test this possibility, we challenged mice lacking TLR4 and TLR9 (TLR439-KO). We reasoned that if TLR9 were playing a negative role in immunity, then any genotype lacking TLR9 would be resistant to infection. Instead, TLR439-KO mice were as susceptible to infection as TLR234-KO mice, indicating that lack of TLR9 by itself does not confer increased resistance to infection (Figure 2A). Thus, the data presented suggest that overall TLR signaling is in some way required for S. typhimurium virulence. Despite this apparent requirement, MyD88-KO and MyD883TRIF-KO mice (with wild-type Nramp1) were highly susceptible to S. typhimurium infection (Figure S2). As discussed earlier, the extreme sensitivity of these mice relative to TLR23439-KO mice is likely due to the role of MyD88 downstream of the IL-1, IL-18, and IL-33 receptors (Mayer-Barber et al., 2010; Raupach et al., 2006). Thus, to examine the role for TLR signaling in S. typhimurium virulence, we must use TLR-deficient mice, not mice lacking common signaling adaptors. One potential caveat of these in vivo studies is that the commensal flora may be different between TLR234-KO and TLR23439-KO mice. Recent studies have reported alterations in commensal communities in mice lacking certain TLRs or TLR-signaling adaptors (Vijay-Kumar et al., 2008; Wen et al., 2008). To address this possibility, we challenged mice with a different gram-negative enteric pathogen, Yersinia enterocolitica (Y. enterocolitica), which shares a similar route of intestinal colonization but remains extracellular after crossing the intestinal epithelia. In contrast to our experiments with S. typhimurium, TLR23439-KO mice were equally, if not more, susceptible rela678 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
tive to TLR234-KO mice (Figure 2C). The differential sensitivity of TLR23439-KO mice to these two enteric bacteria argues that alterations in commensal flora are not contributing to the phenotypes of TLR234-KO and TLR23439-KO mice. Instead, the reduction in TLR signaling in TLR23439-KO mice appears to specifically impact the virulence of S. typhimurium. TLR Signaling Is Required for Intracellular Growth of Bacteria Because survival within macrophages is required for systemic infection (Fields et al., 1986; Leung and Finlay, 1991), we next used a gentamicin protection assay to examine survival and replication in BMMs lacking various TLRs. Consistent with our in vivo experiments, S. typhimurium was able to replicate in TLR234-KO but not TLR23439-KO BMMs (Figure 3A). When we counted bacteria in individual BMMs by immunofluorescence microscopy (IF), the number of bacteria per cell in TLR234-KO BMMs accumulated over time, whereas the number of bacteria per cell in TLR23439-KO BMMs remained constant, indicating that bacterial replication was responsible for the differences in colony-forming units (CFU) between genotypes (Figures 3B and 3C). We observed a similar lack of bacterial replication in MyD883TRIF-KO BMMs (Figure 3A). Unlike our in vivo experiments, the phenotype of MyD883TRIF-KO BMMs is most likely due to a deficiency in TLR signaling, as the IL-1 receptor family is not involved in the initial recognition of S. typhimurium within BMMs in vitro. Furthermore, TLR439-KO BMMs supported bacterial replication similarly to TLR234-KO BMMs, corroborating the conclusions from our in vivo experiments (Figures 3B and 3C). TLR23439-KO and MyD883TRIF-KO BMMs did support replication of Listeria monocytogenes and Legionella pneumophila (Figure S3). In addition, S. typhimurium replicated well in MyD883TRIF-KO BMMs lacking functional Nramp1 (Figure S3). These results indicate that phagosomes of TLR-deficient cells are formally capable of supporting bacterial growth, but the combination of functional Nramp1 and lack of TLR signaling prevents S. typhimurium replication. Collectively, these data suggest that S. typhimurium requires TLR signaling for replication in macrophages. However, the lack of replication in wild-type BMMs would appear to contradict this conclusion, as TLR function is normal in these cells. When we counted the number of bacteria per cell by IF, though, we observed a similar increase in bacteria per cell over time as in TLR234-KO BMMs (Figure 3C). This contradiction was resolved when we measured cell death of BMMs after infection. Wild-type BMMs exhibited greater cell death relative to each of the other genotypes (Figures 3D and 3E). Because only wildtype BMMs express functional TLR4, the increased death of these cells seems likely to be due to a previously described TLR4-dependent cell death that occurs in S. typhimurium infected cells (Cook et al., 2007; Hsu et al., 2004; Park et al., 2002). Thus, the apparent lack of replication as measured by CFU in wild-type BMMs is the result of macrophage death followed by gentamicin-mediated killing of the bacteria. In contrast, the inability of S. typhimurium to replicate in TLR23439-KO or MyD883TRIF-KO BMMs is due to a different mechanism.
0
5
Spleen p<0.05
O
CFU/ Organ (Log)
TL
R
p<0.05
TL
R
x9
TL
R
4 2x
KO
CFU/ Organ (Log)
CFU/ Organ (Log)
Spleen p<0.05
KO
2x
R
2
x4
KO T
LR
2x
4
x9
KO
Liver p<0.05
6 5 4 3 2 1 0
OO
T W
4
R
4 2x
4
R
TL
R
4 2x
x9
KO
KO
2x
p<0.05
3 2 1 0
T W
x4
KO
2
R
KO
9 4x
R TL
T
MLN
2x
KO
MLN
Cecum
TL
OO
TL
9 8 7 6 5 4 3 2 1 T W
30
TL
O
C
4
R TL
9 4x
KO
p<0.05
TL
2x
4 2x
KO
Cecum 7 6 5 4 3 2 1 0
T W
T W
OOO
O
T W
9 8 7 6 5 4 3 2 1
10 15 20 25 Days Post-infection
LR
2x
4
x9
KO
Peyer’s Patches
p<0.05
p<0.05
9 8 7 6 5 4 3 2 1 T W
TL
R
2x
4
KO
x9
4 2x
R
TL
KO
CFU/ Organ (Log)
6 5 4 3 2 1 0
*
CFU/ Organ (Log)
CFU/ Organ (Log)
B
Wildtype, n=28 TLR4x9 KO, n=17 TLR2x4 KO, n=19 TLR2x4x9 KO, n=23
CFU/ Organ (Log)
100 90 80 70 60 50 40 30 20 10 0
CFU/mg Organ (Log)
A
p<0.05
9 8 7 6 5 4 3 2 1 T W TL
R
2x
4
KO
x9
KO
4 2x
R
TL
Figure 2. TLR23439-KO Mice Are Less Susceptible to S. typhimurium than TLR234-KO Mice (A) Survival plots of mice orally inoculated with 1.6 3 108 CFU S. typhimurium (SL1344) are shown. *p < 0.005 by log-rank curve comparison test. Data are representative of at least two independent experiments. (B) Groups of 8- to 10-week-old mice were orally inoculated with 1 3 109 CFU S. typhimurium (SL1344). Four days post-infection organs were harvested and homogenized for colony enumeration. Data are representative of at least three independent experiments. (C) Groups of mice of the indicated genotype were orally inoculated with 2 3 108 Y. enterocolitica and CFU were measured in the indicated organs 3 days post-infection. Data presented are the combined results from two independent experiments. For (B) and (C), bars represent mean CFU of all mice, with data significance determined by Mann-Whitney U test. Open circles indicate mice for which no colonies were detected. MLN, mesenteric lymph node. See also Figure S2.
Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 679
A
C
* *
*
* *
B
*
D WT TLR2x4 KO TLR2x4x9 KO MyD88xTRIF KO Uninfected
E
*
Figure 3. S. typhimurium Is Unable to Replicate in TLR23439-KO and MyD883TRIF-KO BMMs (A) BMMs derived from mice of the indicated genotypes were infected with S. typhimurium (SL1344) at an moi of 1, and intracellular CFU were measured by a gentamicin protection assay at the indicated time points. Data are presented as the average of three independent experiments ± SEM. *p < 0.05 by student’s t test comparing TLR234-KO to MyD883TRIF-KO and TLR23439-KO. (B) BMMs infected as described in (A) were fixed and permeablized at the indicated times post-infection followed by staining with anti-Salmonella LPS antibody (green) and wheat germ agglutinin (red). (C) Intracellular bacteria per cell were counted in random fields at the 2 and 24 hr time points from Z-stacked images as shown in (B). Data are representative of two independent experiments, p value determined by student’s t test (*p < 0.05). (D) BMMs of the indicated genotypes were infected with S. typhimurium (SL1344) at an moi of 5. Eight hours post-infection, cells were harvested and stained with Annexin V. Data are representative of two independent experiments. (E) BMMs were infected at an moi of 10 with S. typhimurium (SL1344), and release of lactate dehydrogenase (LDH) was measured in supernatants at the indicated time points. Data are presented as mean ± SD and are representative of at least two independent experiments; p value determined by student’s t test (*p < 0.05) comparing wild-type to all other genotypes. See also Figure S3.
680 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
Figure 4. S. typhimurium Fails to Form an SCV in TLR23439-KO and MyD883TRIF-KO BMMs
A TLR 2x4 KO 1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
1μm
0.5μm
0.5μm
TLR 2x4x9 KO
MyD88 x TRIF KO
B TLR 2x4 KO
TLR 2x4x9 KO
MyD88 x TRIF KO 0.5μm
1μm
1μm
2μm
1μm
0.5μm
0.5μm
0.5μm
1μm
0.5μm
0.5μm
0.5μm
C TLR 2x4 KO
TLR 2x4x9 KO
MyD88 x TRIF KO
TLR234-KO, TLR23439-KO, or MyD883TRIF-KO BMMs were infected at an moi of 10 with S. typhimurium (SL1344), and cells were fixed and processed for electron microscopy at 2 hr (A), 8 hr (B), and 22 hr (C) post-infection. Bacteria in intact vacuoles are shown with filled black arrowheads, cytosolic bacteria with filled white arrowheads, and bacteria that are degraded or have fused with lytic compartments are indicated with open black arrowheads. Micron bars are in the lower left corner of each panel. Three representative images are shown from different sections and independent infections.
TLR Signaling Is Required for Establishment of the SCV To better understand why S. typhimurium is unable to replicate in TLR23439-KO and MyD883TRIF-KO BMMs, we used transmission electron microscopy (EM) to investigate the fate of bacteria in infected BMMs. At 2 hr post-infection, bacteria were clearly visible in well-defined vacuoles in BMMs of all three genotypes (Figure 4A, black triangles). By 8 hr and 22 hr postinfection, bacteria in TLR234-KO BMMs remained largely unchanged, although evidence of replication was evident, especially at 22 hr (Figures 4B and 4C). In contrast, phagosomes containing bacteria in TLR23439-KO and MyD883TRIF-KO BMMs were quite distinct. The bacteria often appeared mottled or irregular in shape, and in many cases bacteria were surrounded by electron-dense staining material consistent with lysosomal fusion (Figure 4, open triangles). In some instances, bacteria were no longer surrounded by membrane, suggesting that they entered the cytosol (Figure 4, white triangles). Cytosolic bacteria have been described when bacteria fail to secrete certain SPI-2 effectors (Beuzon et al., 2000). In total, the images clearly demonstrate a defect in the ability of S. typhimurium to establish a replicative compartment in TLR23439-KO and MyD883TRIFKO BMMs. Induction of SPI-2 Genes by TLR Signaling Our studies thus far indicate that intracellular growth of S. typhimurium is impaired in TLR23439-KO and MyD883TRIF-KO BMMs and suggest that this defect may be related to inefficient SCV formation. We next sought to define the underlying basis for impaired growth in BMMs lacking TLR function by profiling gene expression of bacteria isolated from BMMs of each genotype. To overcome the lack of sensitivity of microarray-based approaches, we performed quantitative RT-PCR to measure expression of all genes in the S. typhimurium genome (Figure 5A). Using K-means clustering analysis, we identified subsets of genes with differential expression profiles between the BMM genotypes. Genes within the SPI-2 locus were upregulated in bacteria in wild-type and TLR234-KO BMMs but not in bacteria in TLR23439-KO or MyD883TRIF-KO BMMs. For validation, we reanalyzed expression of each gene within the SPI-2 locus and adjacent to the locus (as controls), using independent RNA samples from infected BMMs of all genotypes. As shown in Figure 5B, 13 genes within the SPI-2 locus were upregulated in wildtype and TLR234-KO BMMs but not in TLR23439-KO and MyD883TRIF-KO BMMs. These 13 genes most likely underestimate the extent to which the entire SPI-2 locus is differentially expressed between BMM genotypes, as many genes were statistically excluded due to extremely low levels of message Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 681
A
B
-4
+4
C
D
PipB2 DnaK tubulin Figure 5. S. typhimurium Fails to Upregulate and Secrete SPI-2 Effectors in TLR23439-KO and MyD883TRIF-KO BMMs (A) Schematic of quantitative expression analyses for S. typhimurium genes. Total RNA was isolated from infected BMMs of the indicated genotypes followed by processing for quantitative RT-PCR (see Experimental Procedures). (B) Heat map of normalized expression data for SPI-2 genes in bacteria within BMMs of the indicated genotypes. For each BMM genotype, data are shown relative to the 2 hr time point. The gene designation is to the right of each row. (C) Relative induction of individual SPI-2 genes in bacteria isolated from BMMs of the indicated genotypes. Data are normalized to the average expression values of a set of control genes. The data presented in (B) and (C) represent the mean of two independent experiments. (D) BMMs of the indicated genotypes were infected with an S. typhimurium strain (12032) expressing an HA-tagged allele of pipB2 expressed from the endogenous pipB2 locus. The presence of PipB2 in BMM lysates was detected by immunoprecipitation and immunoblot with anti-HA antibodies. Controls for numbers of BMM (tubulin) and bacteria (DnaK) are also shown. Data are representative of three independent experiments.
682 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
A
Figure 6. TLR Signaling Is Necessary for SPI-2 Induction and Intracellular Growth
B
2h 1μm
1μm
1μm
8h 0.5μm
0.5μm
0.5μm
22h 0.5μm
C
in TLR23439-KO or MyD883TRIF-KO samples. For most SPI-2 genes, induction was higher in wild-type BMMs relative to TLR234-KO BMMs (Figure 5C), suggesting that induction correlates with the strength of TLR signaling. Thus, the lack of intracellular replication in TLR-deficient cells may be due to a failure to upregulate SPI-2 genes. These expression-profiling studies indicated that transcription of SPI-2 genes within BMMs depends on signals downstream of TLR activation. To view SPI-2 induction at the protein level, we utilized a strain of S. typhimurium (12023) with an HA-tagged allele of pipB2, a SPI-2 effector. 12023 displays the same dependence on TLR signaling for intracellular growth as SL1344 (Figure S3D). PipB2 was strongly induced and secreted in infected TLR234-KO BMMs (Figure 5D). In contrast, the levels of PipB2 were significantly reduced in TLR23439-KO BMMs and barely detectable in MyD883TRIF-KO BMMs, despite equivalent numbers of bacteria in all samples (indicated by DnaK levels). These data are consistent with our transcriptional analyses and indicate that TLR signaling is required for the induction of SPI-2 genes. SPI-2 Genes Are Required for Intracellular Growth We hypothesized that the impaired induction of SPI-2 genes in bacteria isolated from TLR23439-KO and MyD883TRIF-KO BMMs was responsible for the defect in SCV formation and
0.5μm
0.5μm
(A) BMMs of the indicated genotypes were infected (moi of 1) with SPI-2 mutant S. typhimurium (SL1344 ssaV::Kan). Intracellular CFU were measured by a gentamicin protection assay. (B) TLR234-KO BMMs were infected as in (A) followed by fixation and processing for electron microscopy. Bacteria in intact vacuoles are indicated by filled black triangles, cytosolic bacteria by filled white triangles, and bacteria that are degraded or have fused with lytic compartments by open black triangles. Three representative images from different sections and from independent infections are shown. (C) BMMs of the indicated genotypes were infected (moi of 1) with wild-type S. typhimurium or a strain with constitutive SPI-2 expression (SL1344 hha::Cm). Intracellular CFU were measured via a gentamicin protection assay at the indicated time points post-infection. Data are presented as mean fold over the first time point (to control for minor inoculum differences between strains) ± SEM and are representative of three independent experiments. *p < 0.05 by student’s t test comparing TLR234-KO to MyD883TRIF-KO and TLR23439-KO at the indicated time point.
intracellular replication in these cells. To test this hypothesis, we compared the fates of bacteria lacking a functional SPI-2 secretion system (ssaV::Kan) in BMMs of each genotype. As expected, SPI-2 mutant bacteria were unable to replicate in BMMs of any genotype (Figure 6A). Moreover, EM analysis of SPI-2 mutant bacteria in TLR234-KO BMMs revealed the same lack of SCV formation observed for wild-type bacteria in TLR23439-KO and MyD883TRIF-KO BMMs (Figure 6B). If the lack of intracellular growth in TLR23439-KO and MyD883TRIF-KO BMMs is due to failure to induce SPI-2 genes, then an S. typhimurium strain with constitutive expression of SPI-2 genes should regain the ability to grow in these cells. To test this possibility directly, we constructed a strain lacking hha (Dhha, hha::Cm), a negative regulator of SPI-2 genes. Previous work has demonstrated that the Dhha mutant strain expresses SPI-2 genes constitutively (Silphaduang et al., 2007). Remarkably, Dhha mutant bacteria replicated equivalently in BMMs of all genotypes, except wild-type cells where the lack of growth is due to TLR4-dependent cell death (Figure 6C). Although Hha probably negatively regulates additional S. typhimurium virulence genes, restoration of growth in TLR-deficient BMMs is consistent with the conclusion that constitutive expression of SPI-2 genes can bypass the requirement for TLR signaling. Induction of SPI-2 Genes Requires TLR-Dependent Acidification of the SCV Our results thus far indicate that TLR signaling provides a cue used by S. typhimurium to regulate SPI-2 expression. TLR activation induces host transcription as well as more proximal Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 683
in pu t
A Time p.i.
EtOH 2
4
BAF 6
8
2
4
APDC
6
8
2
4
Apocynin
6
8
2
4
6
8
Figure 7. TLR-Dependent Phagosomal Acidification Is Required for SPI-2 Expression
L-NIL 2
4
6
8 PipB2 DnaK
Wildtype
tubulin EtOH Time p.i.
2
4
BAF 6
8
2
4
Apocynin
APDC 6
8
2
4
6
8
2
4
6
L-NIL 8
2
4
6
CHX 8
2
4
6
8 PipB2 DnaK
TLR2x4 KO
tubulin
B
effects, such as production of ROS and RNS and phagosome maturation and acidification, although this last aspect remains controversial. Using pharmacological inhibitors to block each of these potential signals we measured the effect on PipB2 induction and secretion. Treatment of TLR234-KO BMMs with cyclohexamide (CHX) had no effect on PipB2 induction, indicating that host translation was not required for generation of the signal sensed by S. typhimurium (Figure 7A, bottom panel). Similarly, blocking ROS or RNS production did not prevent PipB2 induction. However, inhibition of the V-ATPase with bafilomycinA1 blocked S. typhimurium induction of PipB2 in both TLR234-KO and wild-type BMMs. The block in TLR234KO cells could be due to an inhibition of TLR signaling, as bafilomycinA1 almost completely inhibits the residual response to S. typhimurium (Figure 1D). In wild-type cells, though, TLR2 and TLR4 signaling is largely unaffected by bafilomycinA1, suggesting that TLR-dependent acidification of the SCV may be the signal required by S. typhimurium for SPI-2 gene induction (Figure 7A, top panel). Experiments analyzing the induction of SPI-2 genes at the transcriptional level also indicated a requirement for phagosome acidification (data not shown). Based on these data, we hypothesized that the lack of SPI-2 induction in TLR23439-KO and MyD883TRIF-KO BMMs is due to failure of SCVs to acidify. The issue of whether TLR signaling influences the kinetics of phagosomal maturation remains controversial (Blander and Medzhitov, 2004, 2006a, 2006b; Russell and Yates, 2007; Yates and Russell, 2005). To investigate this issue, we used ratiometric imaging to measure the pH of Salmonella-containing phagosomes in BMMs of each genotype. Whereas the mean pH of SCVs in wild-type and TLR234-KO BMMs dropped below 6 within 60 min post684 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
(A) BMMs pretreated with each of the indicated inhibitors were infected with an S. typhimurium strain (12032) expressing an HA-tagged allele of pipB2. PipB2 levels in BMM lysates were detected by immunoprecipitation and immunoblot with antiHA antibodies at 2, 4, 6, and 8 hr post-infection. Immunoblots for tubulin and DnaK serve as loading controls for BMMs and bacteria, respectively. Data are representative of two independent experiments for wild-type cells and three independent experiments for TLR234-KO BMMs. (B) BMMs were infected with FITC-labeled S. typhimurium (SL1344), and the fluorescence intensities of individual bacteria excited at 490 nm or 440 nm were measured over time by live-cell imaging. The fluorescence intensity ratio (490/440) reflects the pH within the phagosome (see Experimental Procedures). The plots presented represent the mean pH calculated from at least 35 independent bacteria in multiple imaging fields. The right panel shows the percent of bacteria at or below pH 6 at 30 min post-infection. See Experimental Procedures and Figure S4.
infection, SCVs in TLR23439-KO and MyD883TRIF-KO BMMs failed to acidify to the same extent and exhibited slower acidification kinetics (Figure 7B). By 30 min post-infection, over 70% of SCVs in wild-type and TLR234-KO BMMs had reached pH 6, whereas less than 35% of SCVs in TLR23439-KO and MyD883TRIF-KO BMMs had similarly acidified (Figure 7B). Consistent with the lower transcriptional induction of SPI-2 genes in TLR234-KO BMMs (relative to wild-type), the rate of acidification in TLR234-KO cells was slower than in wild-type cells, despite ultimately reaching pH 6 by 60 min. Collectively, these data support a model in which TLR signaling accelerates phagosomal acidification, which is used by S. typhimurium as a cue for SPI-2 gene induction. DISCUSSION Biological interactions are strong drivers of evolution, and the dynamics of host-pathogen interactions provide some of the clearest examples of this principle. Hosts have evolved resistance mechanisms, such as TLRs, that work by reducing pathogen fitness and drive the evolution of pathogen virulence (Hedrick, 2004; Rausher, 2001; Woolhouse et al., 2002). Although virulence genes provide a fitness advantage, they can be energetically costly and often serve as targets of host sensors (Miao et al., 2010; Vance et al., 2009). Therefore, the ability to regulate expression of virulence genes based on changing environments is a key feature of microbial pathogenesis. In this study, we report the requirement of TLR signaling for S. typhimurium to establish a successful infection and cause disease. We demonstrate that this requirement stems, at least in part, from the need for TLR-dependent phagosome acidification
to induce SPI-2 genes, resulting in replication and virulence of the microbe. These data demonstrate that a pathogen can evolve to require innate immune signaling for full virulence. Previous studies have demonstrated that host genetic variation can result in prolonged survival upon infection (Raberg et al., 2007). However, these phenotypes are generally attributable to reduced inflammation and immunopathology, suggesting that the host is more tolerant to an increased pathogen burden (Gowen et al., 2006; Wang et al., 2004). By contrast, our work demonstrates that mice lacking sufficient TLR signaling are less susceptible to an S. typhimurium infection due to reduced bacterial growth. A similar relationship has been described in Drosophila, where mutations in the melanization arm of the innate immune response render flies less susceptible to Streptococcus pneumoniae with reduced levels of bacteria, but the mechanism behind this observation remains unclear (Ayres and Schneider, 2008). Our work clearly shows that S. typhimurium fails to induce virulence genes when deprived of innate immune signals. Crosstalk between Nramp-1 and TLR Signaling Two aspects of our approach were crucial for our ability to observe the requirement for TLR-dependent signals in Salmonella virulence. First, by using mice that are deficient in multiple TLRs, as opposed to mice lacking MyD88 and TRIF, we were able to circumvent the susceptibility associated with lack of IL-1R family function. Indeed, the difference that we observe in susceptibility between MyD883TRIF-KO and TLR23439-KO mice underscores the importance of the IL-1R family in defense against infection. We were somewhat surprised by the role for nucleic acid-sensing TLRs in innate recognition of Salmonella, although TLR9 and TLR7 have been implicated in recognition of bacterial nucleic acid (Bafica et al., 2005; Mancuso et al., 2009). Although the simplest explanation for this observation is that some bacteria are degraded, it is also possible that a nucleic acid ligand is secreted by Salmonella or present on the bacterial surface (Whitchurch et al., 2002; Woodward et al., 2010). A second critical aspect of our study is that we used mice with a functional allele of nramp1. Why the lack of this protein renders mice so susceptible to intracellular pathogens remains unclear, but this heightened sensitivity may simply obviate any requirement for TLR-dependent SPI-2 induction. The presence of functional Nramp1 has been shown to enhance SPI-2 expression as well as TLR-dependent responses (Fritsche et al., 2003; Valdez et al., 2008; Zaharik et al., 2002). Regardless of the precise mechanism responsible for the strong TLR dependence when Nramp1 is functional, it is important to recognize that infection of cells with functional Nramp1 represents the ‘‘wild-type’’ scenario. Indeed, mutations in the human Nramp1 gene are associated with increased susceptibility to several intracellular pathogens (Bellamy et al., 1998; Malik et al., 2005). Therefore, examining virulence in the presence of functional Nramp1 most accurately reflects the host-pathogen interactions between S. typhimurium and the mammalian immune system. TLR Signaling Alters the pH of the Salmonella-Containing Vacuole Our studies indicate that the difference in susceptibility of TLR-deficient mice is due to lack (or substantial delay) of SPI-2
induction. We show that SPI-2 induction requires phagosome acidification, and our measurements of phagosomal pH indicate that acidification is impaired and/or delayed in TLR-deficient cells. The extent to which TLR signaling influences phagosomal maturation (including increasing phagolysosomal fusion, acidification, and proteolytic activity) has remained a contentious issue (Blander and Medzhitov, 2004, 2006a, 2006b; Russell and Yates, 2007; Yates and Russell, 2005). Although our studies were not designed to address this controversy, we clearly show that TLR signaling is required for rapid acidification of the SCV and has profound implications for the fate of intracellular bacteria and disease outcome. The mechanism is likely similar to the acidification of lysosomes during DC maturation, when TLR signaling leads to recruitment of the V1 subunit of the vacuolar ATPase to the lysosomal membrane (Trombetta et al., 2003). The precise signaling pathways that lead to assembly of this machinery are unknown. Moreover, whether bacteria sense pH directly or utilize other phagosomal features that require acidic pH remains unclear. Importantly, we are not suggesting that phagosome maturation cannot occur without TLR signaling. Indeed, our EM images of infected TLR23439-KO and MyD883TRIF-KO BMMs at late time points show bacteria within electron-dense compartments, suggestive of phagolysosomal fusion. Due to technical limitations, we have not extended our pH measurements beyond 60 min post-infection, but our images suggest that phagosomes in TLR-deficient cells eventually mature. In fact, we do observe a small percentage of SCV in TLR23439-KO and MyD883 TRIF-KO cells with significant reductions in pH within 1 hr (Figure S4). The lack of bacterial replication in TLR-deficient cells suggests that the eventual maturation of phagosomes is not sufficient to induce SPI-2 genes or that the induction occurs too late to prevent bacterial killing by lysosomal contents. It is also possible that the phagosome breaks down in the absence of SPI-2 function, and bacteria enter the cytosol where they are unable to replicate (Beuzon et al., 2000). Our EM analyses suggest that both of these possibilities may contribute to the lack of bacterial replication. Innate Immune Signaling as an Environmental Cue for Virulence Gene Regulation These findings have important implications for our understanding of the evolution of host-pathogen interactions and virulence mechanisms. Many pathogens trigger these mechanisms by utilizing signals downstream of innate receptors, most likely as a reliable mechanism to properly induce genes necessary for survival in the presence of antimicrobial mechanisms. For example, the PhoP/PhoQ two-component system, when activated by AMPs, induces expression of genes that modify lipidA and render the bacterial membrane more resistant to AMPs (Bader et al., 2005; Guo et al., 1998). By a potentially similar mechanism, prior activation of cells with TLR ligands can increase replication of S. typhimurium (Wong et al., 2009). Our finding that S. typhimurium has evolved to require host-resistance signals for proper expression of virulence genes is conceptually distinct from these previously described antagonistic strategies. Notably, S. typhimurium is unable to replicate in TLR-deficient cells, despite the absence of the antimicrobial Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 685
mechanisms normally induced by TLRs. One implication of this remaining dependence is that the virulence genes induced by TLR signaling are required for purposes other than simply evading TLR-induced antimicrobial mechanisms to promote S. typhimurium fitness. Why would Salmonella use signals downstream of TLRs to broadly coordinate expression of virulence genes required for intracellular growth? These signals may be the most reliable contextual cues that Salmonella can use to sense its presence within a macrophage phagosome. In general, a fundamental problem faced by Salmonella is the need to interact with multiple cell types through the course of an infection. Unique sets of virulence genes are required to survive each of these stages, and Salmonella must recognize its environment and induce the appropriate genes. For example, Salmonella must detect when it has encountered a macrophage and induce SPI-2 genes, which are necessary for formation of the SCV and maintenance of the integrity of the phagosome. Precise regulation of such virulence genes is clearly essential for optimal growth, as mutant bacteria with constitutive expression of SPI-2 genes (e.g., Dhha mutants) are attenuated in vivo (Coombes et al., 2005; Silphaduang et al., 2007). Inappropriate expression of certain virulence genes could result in decreased fitness due to recognition by innate sensors or may disrupt proper regulation of other virulence genes required at specific stages of infection. Therefore, Salmonella utilizes TLR-dependent signals within the phagosome to detect its presence within a macrophage. Linking the induction of virulence genes (including SPI-2) to phagosomal signals downstream of TLRs may be an efficient way of coordinating multiple virulence mechanisms in response to a unifying contextual cue. EXPERIMENTAL PROCEDURES Cell Culture BMMs were differentiated from bone marrow for 5 days using macrophage colony-stimulating factor (M-CSF) as previously described (Ewald et al., 2008). See Extended Experimental Procedures for details. Bacterial Strains and Infections Overnight cultures of S. typhimurium were opsonized for infection. BMMs were spin-infected, incubated at 37 C, then washed with PBS before the addition of 10 mg/ml gentamicin media. For intracellular CFU determination, cells were washed with PBS and lysed in 1% Triton X-100 in PBS. Lysates were plated on LB agar plates containing 200 mg/ml streptomycin (Life Technologies). See Extended Experimental Procedures for strain details and descriptions of assays with L. monocytogenes and L. pneumophila. Measurement of Cell Death Lactate dehydrogenase (LDH) release was quantified using the CytoTox 96 Nonradioactive cytotoxicity kit (Promega) according to manufacturer’s instructions. Annexin V staining was performed using Annexin V-FITC (BD Pharmingen) in Annexin V staining buffer according to manufacturer’s instructions.
presence of 10 mg/ml gentamicin. Cells were analyzed on an FC500 flow cytometer (Beckman Coulter). S. typhimurium Effector Secretion BMMs were infected (multiplicity of infection [moi] of 25) with S. typhimurium pipB2-2xHA (12032). At the indicated time points, cells were washed with PBS and lysed in 1% NP-40 in PBS with protease-inhibtor cocktail (Roche) and EDTA (Fisher), and lysates were subjected to immunoprecipitation with rat anti-HA agarose beads (Roche). Cells were pretreated with inhibitors for 1 hr before infection. See Extended Experimental Procedures for a more detailed explanation of sample processing and detection of effector secretion. Mice and In Vivo Infections All animal experiments were performed in accordance with University of California Animal Care and Use Committee guidelines. See Extended Experimental Procedures for descriptions of strains and backcrossing analyses. For survival and CFU enumeration experiments, age-matched mice were fasted for 14 hr followed by oral gavage with 100 ml S. typhimurium (SL1344) or Yersinia enterocolitica (8081) in PBS (see figure legends for CFU). For CFU enumeration, organs were harvested, homogenized in PBS using a Polytron PT2100 homogenizer (Kinematica), diluted, and plated on streptomycin (for Salmonella) or irgasan (1 mg/ml) (for Yersinia) LB-agar plates. Gene Expression Analyses RNA from infected BMMs (moi of 5) was extracted with Trizol RNA reagent, purified using PureLink Micro-to-Midi total RNA purification system, DNasetreated, reverse transcribed using random hexamers (all reagents from Life Technologies), and processed for quantitative PCR. Sample processing, primer sequences/design, and description of data analysis can be found in the Extended Experimental Procedures and Table S1. Microscopy For IF, BMMs on coverslips were stained with FITC-conjugated mouse anti-Salmonella antibody (clone 1E6, Santa Cruz Biotechnology) and Cy3conjugated wheat germ agglutinin (Life Technologies) at the indicated time points. Cells were imaged on a Nikon E800 fluorescent microscope and bacteria per cell were counted in random, blinded, Z-stacked images. For pH determination and video microscopy, BMMs plated on 4-chamber slides (Nunc) were infected with FITC-labeled bacteria. Following infection, chambers were incubated at 37 C for 5 min, washed extensively with PBS, incubated with phenol-free DMEM containing 10 mg/ml gentamicin, then visualized on a Nikon TE2000 inverted fluorescent microscope with environmental control. Images were collected with excitation at both 440 nm and 490 nm and analyzed using Imaris Scientific 3D/4D image processing and analysis software (Bitplane) to track individual intracellular bacteria. Background-subtracted fluorescence intensity values were used to determine the 490/440 ratios for each bacterium at each time point. Absolute pH values were determined by generating a standard curve using buffered pH solutions (see Extended Experimental Procedures). For EM studies, cells were infected (moi of 10) with wild-type or ssaV::Kan SL1344 (as described above). At each time point, cells were fixed, embedded, sectioned, and stained for EM (see Extended Experimental Procedures). SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, four figures, and one table and can be found with this article online at doi:10.1016/j.cell.2011.01.031. ACKNOWLEDGMENTS
Measurement of BMM Activation NO was quantified in supernatants from BMMs treated overnight with 100 U/ml recombinant IFN-g (R&D Systems) using the Griess assay (all reagents from Sigma Aldrich). TNF-a production was measured by intracellular cytokine staining using anti-TNF-a antibody (eBioscience) according to manufacturer’s instructions (eBioscience). All steps prior to fixation were performed in the
686 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
We thank T. Machen, P. Herzmark, J. Ross, and E. Peled for assistance with ratiometric imaging; members of the Barton lab, J. Ayres, and R. Vance for critical reading of this manuscript; C. Rae and N. Meyer-Morse for assistance with L. monocytogenes; M. Fontana and K. Monroe for assistance with L. pneumophila; R. Zalpuri and K. McDonald for assistance
with EM; and H. Nolla for assistance with flow cytometry. This work was supported in part by grants from the NIH (P01-AI063302 to D.M.M. and G.M.B.; Y1-AI-8401 to S.N.P.), the Burroughs Wellcome Fund (D.M.M.), and the University of California Cancer Research Coordinating Committee (G.M.B.) and an NIH NRSA Trainee appointment on grant T32-GM007232 (N.A.). Received: April 21, 2010 Revised: December 6, 2010 Accepted: January 10, 2011 Published: March 3, 2011 REFERENCES Akira, S., Takeda, K., and Kaisho, T. (2001). Toll-like receptors: critical proteins linking innate and acquired immunity. Nat. Immunol. 2, 675–680. Ayres, J.S., and Schneider, D.S. (2008). A signaling protease required for melanization in Drosophila affects resistance and tolerance of infections. PLoS Biol. 6, 2764–2773.
Ewald, S.E., Lee, B.L., Lau, L., Wickliffe, K.E., Shi, G.P., Chapman, H.A., and Barton, G.M. (2008). The ectodomain of Toll-like receptor 9 is cleaved to generate a functional receptor. Nature 456, 658–662. Feuillet, V., Medjane, S., Mondor, I., Demaria, O., Pagni, P.P., Gala´n, J.E., Flavell, R.A., and Alexopoulou, L. (2006). Involvement of Toll-like receptor 5 in the recognition of flagellated bacteria. Proc. Natl. Acad. Sci. USA 103, 12487–12492. Fields, P.I., Swanson, R.V., Haidaris, C.G., and Heffron, F. (1986). Mutants of Salmonella typhimurium that cannot survive within the macrophage are avirulent. Proc. Natl. Acad. Sci. USA 83, 5189–5193. Forbes, J.R., and Gros, P. (2001). Divalent-metal transport by NRAMP proteins at the interface of host-pathogen interactions. Trends Microbiol. 9, 397–403. Fritsche, G., Dlaska, M., Barton, H., Theurl, I., Garimorth, K., and Weiss, G. (2003). Nramp1 functionality increases inducible nitric oxide synthase transcription via stimulation of IFN regulatory factor 1 expression. J. Immunol. 171, 1994–1998. Galan, J.E. (2001). Salmonella interactions with host cells: type III secretion at work. Annu. Rev. Cell Dev. Biol. 17, 53–86.
Bader, M.W., Sanowar, S., Daley, M.E., Schneider, A.R., Cho, U., Xu, W., Klevit, R.E., Le Moual, H., and Miller, S.I. (2005). Recognition of antimicrobial peptides by a bacterial sensor kinase. Cell 122, 461–472.
Govoni, G., Vidal, S., Gauthier, S., Skamene, E., Malo, D., and Gros, P. (1996). The Bcg/Ity/Lsh locus: genetic transfer of resistance to infections in C57BL/6J mice transgenic for the Nramp1 Gly169 allele. Infect. Immun. 64, 2923–2929.
Bafica, A., Scanga, C.A., Feng, C.G., Leifer, C., Cheever, A., and Sher, A. (2005). TLR9 regulates Th1 responses and cooperates with TLR2 in mediating optimal resistance to Mycobacterium tuberculosis. J. Exp. Med. 202, 1715– 1724.
Gowen, B.B., Hoopes, J.D., Wong, M.H., Jung, K.H., Isakson, K.C., Alexopoulou, L., Flavell, R.A., and Sidwell, R.W. (2006). TLR3 deletion limits mortality and disease severity due to Phlebovirus infection. J. Immunol. 177, 6301–6307.
Bellamy, R., Ruwende, C., Corrah, T., McAdam, K.P., Whittle, H.C., and Hill, A.V. (1998). Variations in the NRAMP1 gene and susceptibility to tuberculosis in West Africans. N. Engl. J. Med. 338, 640–644.
Guo, L., Lim, K.B., Gunn, J.S., Bainbridge, B., Darveau, R.P., Hackett, M., and Miller, S.I. (1997). Regulation of lipid A modifications by Salmonella typhimurium virulence genes phoP-phoQ. Science 276, 250–253.
Beuzon, C.R., Meresse, S., Unsworth, K.E., Ruiz-Albert, J., Garvis, S., Waterman, S.R., Ryder, T.A., Boucrot, E., and Holden, D.W. (2000). Salmonella maintains the integrity of its intracellular vacuole through the action of SifA. EMBO J. 19, 3235–3249.
Guo, L., Lim, K.B., Poduje, C.M., Daniel, M., Gunn, J.S., Hackett, M., and Miller, S.I. (1998). Lipid A acylation and bacterial resistance against vertebrate antimicrobial peptides. Cell 95, 189–198.
Blander, J.M., and Medzhitov, R. (2004). Regulation of phagosome maturation by signals from toll-like receptors. Science 304, 1014–1018. Blander, J.M., and Medzhitov, R. (2006a). On regulation of phagosome maturation and antigen presentation. Nat. Immunol. 7, 1029–1035. Blander, J.M., and Medzhitov, R. (2006b). Toll-dependent selection of microbial antigens for presentation by dendritic cells. Nature 440, 808–812. Chakravortty, D., Rohde, M., Ja¨ger, L., Deiwick, J., and Hensel, M. (2005). Formation of a novel surface structure encoded by Salmonella Pathogenicity Island 2. EMBO J. 24, 2043–2052. Cirillo, D.M., Valdivia, R.H., Monack, D.M., and Falkow, S. (1998). Macrophage-dependent induction of the Salmonella pathogenicity island 2 type III secretion system and its role in intracellular survival. Mol. Microbiol. 30, 175–188.
Hapfelmeier, S., Stecher, B., Barthel, M., Kremer, M., Mu¨ller, A.J., Heikenwalder, M., Stallmach, T., Hensel, M., Pfeffer, K., Akira, S., et al. (2005). The Salmonella pathogenicity island (SPI)-2 and SPI-1 type III secretion systems allow Salmonella serovar typhimurium to trigger colitis via MyD88dependent and MyD88-independent mechanisms. J. Immunol. 174, 1675– 1685. Hedrick, S. (2004). The acquired immune system: A vantage from beneath. Immunity 21, 607–615. Hsu, L.C., Park, J.M., Zhang, K., Luo, J.L., Maeda, S., Kaufman, R.J., Eckmann, L., Guiney, D.G., and Karin, M. (2004). The protein kinase PKR is required for macrophage apoptosis after activation of Toll-like receptor 4. Nature 428, 341–345. Iwasaki, A., and Medzhitov, R. (2004). Toll-like receptor control of the adaptive immune responses. Nat. Immunol. 5, 987–995.
Coburn, B., Grassl, G.A., and Finlay, B.B. (2007). Salmonella, the host and disease: a brief review. Immunol. Cell Biol. 85, 112–118.
Janeway, C.A., Jr. (1989). Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb. Symp. Quant. Biol. 54, 1–13.
Cook, P., To¨temeyer, S., Stevenson, C., Fitzgerald, K.A., Yamamoto, M., Akira, S., Maskell, D.J., and Bryant, C.E. (2007). Salmonella-induced SipB-independent cell death requires Toll-like receptor-4 signalling via the adapter proteins Tram and Trif. Immunology 122, 222–229.
Kawai, T., and Akira, S. (2005). Pathogen recognition with Toll-like receptors. Curr. Opin. Immunol. 17, 338–344.
Coombes, B.K., Wickham, M.E., Lowden, M.J., Brown, N.F., and Finlay, B.B. (2005). Negative regulation of Salmonella pathogenicity island 2 is required for contextual control of virulence during typhoid. Proc. Natl. Acad. Sci. USA 102, 17460–17465. Deiwick, J., Nikolaus, T., Erdogan, S., and Hensel, M. (1999). Environmental regulation of Salmonella pathogenicity island 2 gene expression. Mol. Microbiol. 31, 1759–1773. Detweiler, C.S., Monack, D.M., Brodsky, I.E., Mathew, H., and Falkow, S. (2003). virK, somA and rcsC are important for systemic Salmonella enterica serovar Typhimurium infection and cationic peptide resistance. Mol. Microbiol. 48, 385–400.
Kim, C.C., and Falkow, S. (2004). Delineation of upstream signaling events in the Salmonella pathogenicity island 2 transcriptional activation pathway. J. Bacteriol. 186, 4694–4704. Leung, K.Y., and Finlay, B.B. (1991). Intracellular replication is essential for the virulence of Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 88, 11470– 11474. Malik, S., Abel, L., Tooker, H., Poon, A., Simkin, L., Girard, M., Adams, G.J., Starke, J.R., Smith, K.C., Graviss, E.A., et al. (2005). Alleles of the NRAMP1 gene are risk factors for pediatric tuberculosis disease. Proc. Natl. Acad. Sci. USA 102, 12183–12188. Mancuso, G., Gambuzza, M., Midiri, A., Biondo, C., Papasergi, S., Akira, S., Teti, G., and Beninati, C. (2009). Bacterial recognition by TLR7 in the lysosomes of conventional dendritic cells. Nat. Immunol. 10, 587–594.
Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc. 687
Marcus, S.L., Brumell, J.H., Pfeifer, C.G., and Finlay, B.B. (2000). Salmonella pathogenicity islands: big virulence in small packages. Microbes Infect. 2, 145–156. Mayer-Barber, K.D., Barber, D.L., Shenderov, K., White, S.D., Wilson, M.S., Cheever, A., Kugler, D., Hieny, S., Caspar, P., Nunez, G., et al. (2010). Caspase-1 independent IL-1beta production is critical for host resistance to mycobacterium tuberculosis and does not require TLR signaling in vivo. J. Immunol. 184, 3326–3330. Medzhitov, R. (2007). Recognition of microorganisms and activation of the immune response. Nature 449, 819–826. Miao, E.A., Mao, D.P., Yudkovsky, N., Bonneau, R., Lorang, C.G., Warren, S.E., Leaf, I.A., and Aderem, A. (2010). Innate immune detection of the type III secretion apparatus through the NLRC4 inflammasome. Proc. Natl. Acad. Sci. USA 107, 3076–3080. O’Brien, A.D., Rosenstreich, D.L., Scher, I., Campbell, G.H., MacDermott, R.P., and Formal, S.B. (1980). Genetic control of susceptibility to Salmonella typhimurium in mice: role of the LPS gene. J. Immunol. 124, 20–24.
Valdez, Y., Diehl, G.E., Vallance, B.A., Grassl, G.A., Guttman, J.A., Brown, N.F., Rosenberger, C.M., Littman, D.R., Gros, P., and Finlay, B.B. (2008). Nramp1 expression by dendritic cells modulates inflammatory responses during Salmonella Typhimurium infection. Cell. Microbiol. 10, 1646–1661. Valdivia, R.H., and Falkow, S. (1997). Fluorescence-based isolation of bacterial genes expressed within host cells. Science 277, 2007–2011. Vance, R.E., Isberg, R.R., and Portnoy, D.A. (2009). Patterns of pathogenesis: discrimination of pathogenic and nonpathogenic microbes by the innate immune system. Cell Host Microbe 6, 10–21. Vazquez-Torres, A., Vallance, B.A., Bergman, M.A., Finlay, B.B., Cookson, B.T., Jones-Carson, J., and Fang, F.C. (2004). Toll-like receptor 4 dependence of innate and adaptive immunity to Salmonella: importance of the Kupffer cell network. J. Immunol. 172, 6202–6208. Vidal, S.M., Malo, D., Vogan, K., Skamene, E., and Gros, P. (1993). Natural resistance to infection with intracellular parasites: Isolation of a candidate for Bcg. Cell 73, 469–485.
Park, J.M., Greten, F.R., Li, Z.W., and Karin, M. (2002). Macrophage apoptosis by anthrax lethal factor through p38 MAP kinase inhibition. Science 297, 2048– 2051.
Vidal, S., Tremblay, M.L., Govoni, G., Gauthier, S., Sebastiani, G., Malo, D., Skamene, E., Olivier, M., Jothy, S., and Gros, P. (1995). The Ity/Lsh/Bcg locus: natural resistance to infection with intracellular parasites is abrogated by disruption of the Nramp1 gene. J. Exp. Med. 182, 655–666.
Pfeifer, C.G., Marcus, S.L., Steele-Mortimer, O., Knodler, L.A., and Finlay, B.B. (1999). Salmonella typhimurium virulence genes are induced upon bacterial invasion into phagocytic and nonphagocytic cells. Infect. Immun. 67, 5690– 5698.
Vidal, S.M., Pinner, E., Lepage, P., Gauthier, S., and Gros, P. (1996). Natural resistance to intracellular infections: Nramp1 encodes a membrane phosphoglycoprotein absent in macrophages from susceptible (Nramp1 D169) mouse strains. J. Immunol. 157, 3559–3568.
Raberg, L., Sim, D., and Read, A.F. (2007). Disentangling genetic variation for resistance and tolerance to infectious diseases in animals. Science 318, 812–814. Rappl, C., Deiwick, J., and Hensel, M. (2003). Acidic pH is required for the functional assembly of the type III secretion system encoded by Salmonella pathogenicity island 2. FEMS Microbiol. Lett. 226, 363–372. Raupach, B., Peuschel, S.-K., Monack, D.M., and Zychlinsky, A. (2006). Caspase-1-mediated activation of interleukin-1beta (IL-1beta) and IL-18 contributes to innate immune defenses against Salmonella enterica serovar Typhimurium infection. Infect. Immun. 74, 4922–4926. Rausher, M.D. (2001). Co-evolution and plant resistance to natural enemies. Nature 411, 857–864. Royle, M.C., To¨temeyer, S., Alldridge, L.C., Maskell, D.J., and Bryant, C.E. (2003). Stimulation of Toll-like receptor 4 by lipopolysaccharide during cellular invasion by live Salmonella typhimurium is a critical but not exclusive event leading to macrophage responses. J. Immunol. 170, 5445–5454. Russell, D.G., and Yates, R.M. (2007). Toll-like receptors and phagosome maturation. Nat. Immunol. 8, 217, author reply 217–218. Shea, J.E., Hensel, M., Gleeson, C., and Holden, D.W. (1996). Identification of a virulence locus encoding a second type III secretion system in Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 93, 2593–2597. Silphaduang, U., Mascarenhas, M., Karmali, M., and Coombes, B.K. (2007). Repression of intracellular virulence factors in Salmonella by the Hha and YdgT nucleoid-associated proteins. J. Bacteriol. 189, 3669–3673. Smith, K.D., Andersen-Nissen, E., Hayashi, F., Strobe, K., Bergman, M.A., Barrett, S.L., Cookson, B.T., and Aderem, A. (2003). Toll-like receptor 5 recognizes a conserved site on flagellin required for protofilament formation and bacterial motility. Nat. Immunol. 4, 1247–1253.
Vijay-Kumar, M., Aitken, J.D., Kumar, A., Neish, A.S., Uematsu, S., Akira, S., and Gewirtz, A.T. (2008). Toll-like receptor 5-deficient mice have dysregulated intestinal gene expression and nonspecific resistance to Salmonella-induced typhoid-like disease. Infect. Immun. 76, 1276–1281. Wang, T., Town, T., Alexopoulou, L., Anderson, J.F., Fikrig, E., and Flavell, R.A. (2004). Toll-like receptor 3 mediates West Nile virus entry into the brain causing lethal encephalitis. Nat. Med. 10, 1366–1373. Waterman, S.R., and Holden, D.W. (2003). Functions and effectors of the Salmonella pathogenicity island 2 type III secretion system. Cell. Microbiol. 5, 501–511. Weiss, D.S., Raupach, B., Takeda, K., Akira, S., and Zychlinsky, A. (2004). Toll-like receptors are temporally involved in host defense. J. Immunol. 172, 4463–4469. Wen, L., Ley, R.E., Volchkov, P.Y., Stranges, P.B., Avanesyan, L., Stonebraker, A.C., Hu, C., Wong, F.S., Szot, G.L., Bluestone, J.A., et al. (2008). Innate immunity and intestinal microbiota in the development of Type 1 diabetes. Nature 455, 1109–1113. Whitchurch, C.B., Tolker-Nielsen, T., Ragas, P.C., and Mattick, J.S. (2002). Extracellular DNA required for bacterial biofilm formation. Science 295, 1487. Wong, C.E., Sad, S., and Coombes, B.K. (2009). Salmonella enterica serovar typhimurium exploits toll-like receptor signaling during the host-pathogen interaction. Infect. Immun. 77, 4750–4760. Woodward, J.J., Iavarone, A.T., and Portnoy, D.A. (2010). c-di-AMP secreted by intracellular Listeria monocytogenes activates a host type I interferon response. Science 328, 1703–1705. Woolhouse, M.E., Webster, J.P., Domingo, E., Charlesworth, B., and Levin, B.R. (2002). Biological and biomedical implications of the co-evolution of pathogens and their hosts. Nat. Genet. 32, 569–577.
Trombetta, E.S., Ebersold, M., Garrett, W., Pypaert, M., and Mellman, I. (2003). Activation of lysosomal function during dendritic cell maturation. Science 299, 1400–1403.
Yates, R.M., and Russell, D.G. (2005). Phagosome maturation proceeds independently of stimulation of toll-like receptors 2 and 4. Immunity 23, 409–417.
Uematsu, S., Jang, M.H., Chevrier, N., Guo, Z., Kumagai, Y., Yamamoto, M., Kato, H., Sougawa, N., Matsui, H., Kuwata, H., et al. (2006). Detection of pathogenic intestinal bacteria by Toll-like receptor 5 on intestinal CD11c+ lamina propria cells. Nat. Immunol. 7, 868–874.
Zaharik, M.L., Vallance, B.A., Puente, J.L., Gros, P., and Finlay, B.B. (2002). Host-pathogen interactions: Host resistance factor Nramp1 up-regulates the expression of Salmonella pathogenicity island-2 virulence genes. Proc. Natl. Acad. Sci. USA 99, 15705–15710.
688 Cell 144, 675–688, March 4, 2011 ª2011 Elsevier Inc.
PARIS (ZNF746) Repression of PGC-1a Contributes to Neurodegeneration in Parkinson’s Disease Joo-Ho Shin,1,2 Han Seok Ko,1,2 Hochul Kang,1,2 Yunjong Lee,1,3 Yun-Il Lee,1,2 Olga Pletinkova,4 Juan C. Troconso,2,4 Valina L. Dawson,1,2,3,5 and Ted M. Dawson1,2,5,* 1NeuroRegeneration
and Stem Cell Programs, Institute for Cell Engineering of Neurology 3Department of Physiology 4Department of Pathology, Division of Neuropathology 5Solomon H. Snyder Department of Neuroscience Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.010 2Department
SUMMARY
A hallmark of Parkinson’s disease (PD) is the preferential loss of substantia nigra dopamine neurons. Here, we identify a new parkin interacting substrate, PARIS (ZNF746), whose levels are regulated by the ubiquitin proteasome system via binding to and ubiquitination by the E3 ubiquitin ligase, parkin. PARIS is a KRAB and zinc finger protein that accumulates in models of parkin inactivation and in human PD brain. PARIS represses the expression of the transcriptional coactivator, PGC-1a and the PGC-1a target gene, NRF-1 by binding to insulin response sequences in the PGC-1a promoter. Conditional knockout of parkin in adult animals leads to progressive loss of dopamine (DA) neurons in a PARIS-dependent manner. Moreover, overexpression of PARIS leads to the selective loss of DA neurons in the substantia nigra, and this is reversed by either parkin or PGC-1a coexpression. The identification of PARIS provides a molecular mechanism for neurodegeneration due to parkin inactivation. INTRODUCTION Parkinson’s disease (PD) is a progressive neurodegenerative disorder that is characterized phenotypically by bradykinesia, rigidity, tremor, and neuropsychiatric disturbances (Savitt et al., 2006). Although the cause of PD in the majority of cases is unknown, there are rare familial cases for which the genes have been identified. There are at least sixteen PD associated loci (Gasser, 2007). Mutations in a-synuclein and leucine rich repeat kinase 2 (LRRK2) cause autosomal dominant PD. Four genes have been linked to autosomal recessive PD (AR-PD) and include mutations in parkin, DJ-1, PINK1, and ATP13A2. Investigating the biology of these genes and their mutant protein
has provided tremendous insight into the pathogenesis of both familial and sporadic PD (Gasser, 2007; Savitt et al., 2006). Parkin is an ubiquitin E3 ligase (Shimura et al., 2000; Zhang et al., 2000). In general, PD-associated mutations in parkin lead to loss of its E3 ligase function (Tanaka et al., 2004). Moreover, oxidative, nitrosative, and dopaminergic stress, which play important pathogenic roles in PD inactivate parkin, suggesting that parkin inactivation may play a role in sporadic PD (Chung et al., 2004; LaVoie et al., 2005; Winklhofer et al., 2003). Thus, substrates of parkin that are subject to proteasomal degradation should accumulate in animal and cellular models of parkin inactivation and AR-PD due to parkin mutations, and also in sporadic PD. There are a diverse array of parkin substrates that has hindered the generation of a consensus in the field on parkin’s physiologic function and pathologic role in PD. Moreover, parkin’s ability to mono- and polyubiquitinate as well as ubiquitinate proteins with both lysine-48 and lysine-63 chains has made it difficult to reconcile a common biochemical pathway for parkin’s role in PD (Dawson and Dawson, 2010). Here, we identify a new parkin interacting substrate, PARIS, which provides a molecular mechanism for neurodegeneration due to parkin inactivation in PD. Our results show that parkin regulates the levels of PARIS via the ubiquitin proteasome system (UPS). We show that PARIS is a major transcriptional repressor of peroxisome proliferator-activated receptor gamma (PPARg) coactivator-1a (PGC-1a) expression and that conditional knockout (KO) of parkin in adult mice leads to progressive loss of dopamine (DA) neurons through PARIS overexpression and transcriptional repression of PGC-1a. RESULTS PARIS Is a KRAB and C2H2 Zinc Finger Protein PARIS was identified by yeast two-hybrid screening using parkin as bait (Zhang et al., 2000). Human PARIS (ZNF746) is a 644 amino acid protein that contains a Kruppel-associated box (KRAB) at its N-terminus and a C2HC/C2H2 type zinc finger at its C terminus (Figure 1A). There is a high degree of homology Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 689
Figure 1. Identification of a Parkin Interacting Substrate, PARIS (A) Schematic representation of PARIS. The conserved Kruppel associated Box (KRAB) and Zinc Finger motifs and their location are indicated. (B) Regional analysis and levels of PARIS protein expression via immunoblot in various brain regions. Data = mean ± SEM, n = 3. (C) Immunohistochemical distribution of PARIS in a sagittal and coronal sections of 6-week-old male C57BL mouse brain. Right upper panel shows an antigen preabsorption control. Ctx, cerebral cortex; Hip, hippocampus; CPu, caudate putamen; SNpr, SN pars reticulata; SNpc, SN pars compacta; Th, thalamus; Cb, cerebellum; (Cb); PC, Purkinje cells; ML, molecular cell layer; GL, granule cell layer; DG, dentate gyrus. Dashed (-) line outlines SN. High power view of SNpc (rectangles in third row) is shown in the third row middle and right panels and lower middle and right panel. Scale bars represent 200 mm unless indicated, n = 3. (D) Confocal microscopy demonstrates that endogenous PARIS and parkin are colocalized mainly in the cytoplasm of rat cortical neurons. Top Panel, Parkin-red; PARIS-green. Inset – high power view of an individual neuron. Bottom panel, Parkin-green; PARIS-red, Nucleus-DAPI-blue, merge-yellow, n = 4. See also Figure S1.
among human, mouse and rat PARIS proteins (Figure S1A). Northern blot and immunoblot analysis reveals that PARIS is expressed in all organs examined (Figures S1B and S1C). PARIS is differentially expressed in the brain with low levels in cerebellum and midbrain (Figure 1B). Immunohistochemisty reveals that PARIS is heterogeneously distributed throughout the brain and that it is localized to neurons, including substantia nigra (SN) pars compacta DA containing neurons (Figure 1C). Confocal imaging indicates that PARIS is colocalized with parkin in primary cortical neuron cultures (Figure 1D) and in SH-SY5Y dopaminergic-like cells (Figure S1D). Parkin Interacts with PARIS MYC-tagged parkin and FLAG-tagged PARIS coimmunoprecipitate in SH-SY5Y cells, whereas XIAP, a RING finger ubiquitin E3 690 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
ligase and ZNF398, a highly conserved homolog of PARIS do not interact with PARIS and parkin, respectively (Figure 2A). Recombinant GST-PARIS pulls down recombinant His-Parkin indicating that PARIS and parkin directly interact (Figure 2B). The familial PD-associated mutations in parkin, C431F, R275W, and Q311X bind to PARIS less avidly than WT parkin, whereas G430D mutant parkin binds to PARIS in a manner similar to WT parkin (Figure 2C). PARIS coimmunoprecipitates with parkin and parkin coimmunoprecipitates with PARIS from whole human striatum (Figure 2D) or mouse brain (Figure S2A). This endogenous interaction between parkin and PARIS is not observed in parkin exon 7 KO brain (Figure 2E) confirming the specificity of the interaction. Domain mapping indicates that parkin binds to the C-terminal domain, but not the N-terminal domain of PARIS (Figure S2B).
Figure 2. Parkin Interacts with PARIS (A) Immunoprecipitated (IP) FLAG-PARIS interacts with MYC-Parkin, but not MYC-XIAP or the PARIS homolog, V5-ZNF398 (lane 6), n = 3. (B) GST-pull down assay between parkin and PARIS indicates a robust interaction between parkin and PARIS, n = 4. (C) Immunoprecipitated FLAG-PARIS interacts with WT Parkin and Parkin mutants (C431F, G430D, R275W, Q311X) in SH-SY5Y cells. Lower panel, relative binding, data = mean ± SEM, n = 3, *p < 0.05, Student’s t test. (D) Immunoblot (IB) shows that Parkin and PARIS coimmunoprecipitate in human striatum, n = 3. (E) Parkin and PARIS coimmunoprecipitate from WT mouse ventral midbrain, but not parkin KO ventral midbrain, n = 3. See also Figure S2.
To ascertain which domain of parkin binds to PARIS, various deletion constructs of MYC-tagged parkin were utilized. Coimmunoprecipitation experiments show that both the RING1 or RING2 domains are required for parkin binding to PARIS (Figures S2C and S2D). Taken together these results indicate that parkin interacts with the zinc finger domain of PARIS and PARIS binds to either the RING1 or RING2 domain of parkin. Parkin Ubiquitinates PARIS and Regulates PARIS Levels FLAG-tagged PARIS is ubiquitinated by MYC-tagged parkin in SH-SY5Y cells as shown by the substantial HA immunoreactivity in the form of a smear, which is characteristic of polyubiquitinated proteins (lane 3, Figure 3A). Familial linked parkin mutations C431F, G430D, and Q311X have substantially reduced ubiquitination activity against PARIS, whereas R275W has modest activity (Figure 3A). The reduction in ubiquitination of PARIS by these mutants is due, in part, to their reduced binding (see Figure 2C) as well as reduced E3 ligase activity of these parkin mutants (Sriram et al., 2005). FLAG-PARIS is ubiquitinated in the absence of exogenous parkin (lane 2, Figures 3A and 3B). This endogenous ubiquitination is enhanced with coexpression of WT parkin (lane 3, Figures
3A and 3B). ShRNA-Parkin eliminates the endogenous ubiquitination (lane 4, Figure 3B). To control for potential off-targets effects of shRNA-parkin, a shRNAresistant WT parkin, but not a shRNAresistant Q311X parkin mutant restores the ubiquitination of PARIS in the presence of shRNA-parkin (Figure 3B). In vitro ubiquitin assays shows that parkin ubiquitinates PARIS in the presence of various E2 enzymes including UbcH5c (Figure S3A). There is no ubiquitination signal in the absence of PARIS (Figure S3B) and in the absence of parkin (Figure S3C) indicating that the ubiquitination signal is specific for PARIS ubiqutination. CHIP, which acts as a E4 for parkin (Imai et al., 2002), enhances the ubiquitination of PARIS by parkin, but it has no affect in the absence of parkin indicating that CHIP does not ubiquitinate PARIS directly (Figure S3D). OTU1, a K48-linkage specific deubiquitinating enzyme (Messick et al., 2008), successfully hydrolyzed the polyubiquitin chain on PARIS (Figures S3E and S3F) and a K48-specific anti-ubiquitin antibody, but not a K63 antibody detects the polyubiquitin chain on PARIS (Figure S3E), suggesting that endogenous and exogenous parkin ubiquitinates PARIS via K48 linkages. The steady-state level of PARIS is regulated by the ubiquitin proteasome system since the level of PARIS increases approximately two to three fold when SH-SY5Y cells are treated with the proteasome inhibitor, MG132 (Figure 3C). The levels of PARIS and parkin are tightly and significantly correlated (R2 = 0.9985) in coexpression experiments in SH-SY5Y cells in which increased levels of parkin lead to decreased levels of PARIS (Figure 3D). Cyclohexamide (100 mg/ml) chase experiments in SH-SY5Y cells transiently expressing GFP-tagged PARIS with or without WT parkin or Q311X parkin demonstrate that the decrease in the steady-state levels of PARIS is accelerated in the presence of WT parkin compared with cells transfected with GFP-vector or familial linked parkin Q311X mutant Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 691
Figure 3. Parkin Ubiquitinates and Regulates the Ubiquitin Proteasomal Degradation of PARIS (A) WT MYC-Parkin ubiquitinates FLAG-PARIS (lane 3). Parkin mutants (C431F, G430D, Q311X) are unable to efficiently ubiquitinate FLAG-PARIS (lanes 4-7). Ubiquitination (Ub(n)) is indicated on right with brackets, n = 3. (B) Endogenous ubiquitination of PARIS (lane 2) is enhanced with exogenous WT Parkin (lane 3) and it is eliminated with shRNA-Parkin (lane 4). In the presence of shRNA-Parkin, robust ubiquitination of PARIS is observed via coexpression of shRNA-resistant WT parkin (WTR) but not shRNA-resistant Q311X mutant parkin (Q311XR), n = 3. (C) 10 mM MG-132 increases PARIS steady-state levels compared to DMSO control. Bottom panel, relative PARIS levels normalized to b-actin, n = 3. (D) Increasing ratio (1:1 to 4:1) of MYC-Parkin results in decreased steady-state levels of FLAG-PARIS (lanes 1-4). Bottom panel, relative PARIS and parkin levels normalized to b-actin, n = 3; regression analysis, R2 = 0.9985, p < 0.05). (E) WT Parkin decreases the steady-state levels of PARIS compared to mutant Q311X parkin or GFP transfected control cells in cyclohexamide (CHX)-chase experiments in SH-SY5Y cells transiently expressing FLAG-PARIS. Bottom panel, relative PARIS levels normalized to b-actin, n = 3. (F) MYC-parkin leads to degradation of FLAG-PARIS at a 4 to 1 ratio, respectively. 10 mM MG-132 prevents the degradation of FLAG-PARIS and MYC-Q311X parkin has no effect, n = 3). (G) PARIS accumulates after shRNA-Parkin and coexpression of shRNA resistant parkin (MYC-ParkinR) leads to robust degradation of PARIS, n = 3. Data = mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001, ANOVA with Student-Newman-Keuls post hoc analysis (E, F, G) or Student’s t test (C). See also Figure S3.
(Figure 3E). Parkin at a 4 to 1 ratio with PARIS leads to a statistically significant reduction in PARIS that is blocked by MG132, whereas the catalytically inactive Q311X parkin fails to reduce PARIS levels (Figure 3F). ShRNA-Parkin leads to a significant increase in the level of PARIS, which is reduced by expression of a shRNA-resistant WT parkin (Figure 3G). These results taken 692 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
together suggest that parkin controls the levels of PARIS via the ubiquitin proteasome system. PARIS Accumulates in the Absence of Parkin Activity PARIS levels are increased two fold in the cingulate cortex of ARPD patients, which lack functional parkin, versus age-matched
Figure 4. PARIS accumulates in AR-PD, sporadic PD and in animal models of parkin inactivation (A) Immunoblot analysis of PARIS and b-actin in cingulate cortex from age-matched controls and AR-PD patient brains with parkin mutations. (B) Quantitation of the immunoblots in panel A normalized to b-actin, n = 4. (C) PARIS levels in cerebellum (CBM), frontal cortex (FC), striatum (STR) and SN of sporadic PD patient brains compared to age-matched controls. (D) Relative PARIS levels normalized to b-actin in panel B, Controls n = 4; PD n = 5. (E) Immunoblot analysis of PARIS in cortex (CTX), STR and ventral midbrain (VM) from WT and parkin exon 7 KO 18-24 month old mice. (F) Relative protein levels of PARIS normalized to b-actin for panel E, WT n = 9; parkin KO n = 10. (G) PARIS mRNA levels in indicated brain regions from WT and parkin exon 7 KO 18-24 month old mice. (H) Top panel, experimental illustration of stereotaxic intranigral virus injection. Bottom panels, immunofluorescent images of TH (red), GFP (green) and merged (yellow) in exon 7 floxed parkin mice (parkinFlx/Flx) after stereotactic delivery of Lenti-GFP or Lenti-GFPCre into the SNpc. 84.9 ± 1.9% and 78.1 ± 2.6% of TH neurons express GFP and GFPCre, respectively, n = 3 per group. Enlarged images in the right bottom panels were taken from the white rectangle region from the merged images of Lenti-GFPCre and Lenti-GFP, bar = 100 mm. (I) Immunoblot analysis of parkin, PARIS, actin and GFP 4 weeks after intranigral Lenti-GFPCre or Lenti-GFP injection into parkinFlx/Flx mice. (J) Relative protein levels of PARIS normalized to b-actin for panel I. Data = mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001, unpaired two-tailed Student’s t test (B, J) and ANOVA test with Student-Newman-Keuls post hoc analysis (D, F, G). See also Table S1 and Table S2.
controls (Figures 4A and 4B). Due to the lack of tissue availability, we were not able to assess the levels of PARIS in other brain regions of AR-PD patients. PARIS levels are also increased by more than two fold in both the striatum and the SN of sporadic PD compared to controls (Figures 4C and 4D). We find no alteration in the levels of PARIS mRNA in PD striatum or PD SN versus control striatum and SN indicating that
the increased level of PARIS is not due to a transcriptional effect (see Figure 6A and Figure S5A). PARIS levels are not increased in regions of the brain that are relatively unaffected in PD including the cerebellum and the frontal cortex of sporadic PD patients compared to controls (Figures 4C and 4D) suggesting that the upregulation is primarily within the nigrostriatal pathway. Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 693
We next evaluated whether PARIS is increased in germline parkin exon 7 KO mice (Von Coelln et al., 2004). PARIS is modestly upregulated by approximately 48% in the striatum and by 63% in ventral midbrain of germline parkin exon 7 KO mice compared to age matched WT controls, whereas the levels of PARIS in the cortex is not changed (Figures 4E and 4F). The upregulation of PARIS is not due to a transcriptional effect since we find no difference in the level of PARIS mRNA in germline parkin KO ventral midbrain and striatum versus WT ventral midbrain and striatum(Figure 4G). Since germline parkin exon 7 KO mice lack parkin from the point of conception, it is possible that compensatory mechanisms account for the lack of a more substantial upregulation of PARIS (for review see (Dawson et al., 2010)). To avoid potential developmental compensation, exon 7 was deleted in 6-8 week old conditional parkin KO mice in which exon 7 was flanked by loxP sites (parkinFlx/Flx) by SN stereotactic injection of a GFP-fused Cre recombinase lentivirus (Lenti-GFPCre) and compared to control SN injections of lentivirus expressing GFP (Lenti-GFP) in parkinFlx/Flx mice (Figures 4H and 4I). Four weeks after injection of the lentiviruses, Lenti-GFP and Lenti-GFPCre effectively transduce neurons in the SN including DA neurons (Figure 4H). Lenti-GFPCre leads to almost a complete loss of parkin from the ventral midbrain of parkinFlx/Flx mice compared to Lenti-GFP mice (Figures 4I and 4J). Accompanying the loss of parkin is greater than a two-fold upregulation of PARIS. The upregulation of PARIS in the conditional parkin exon 7 KO model is not due to a transcriptional effect since there is no alteration in PARIS mRNA in the conditional parkin KO ventral midbrain versus WT ventral midbrain (Figure S5E). Thus, the levels of PARIS are increased in conjunction with parkin inactivation and impairment of ubiquitin-mediated proteasomal degradation in sporadic PD, AR-PD and in an animal model of parkin inactivation. PARIS Is a Transcriptional Repressor of PGC-1a Proteins with KRAB domains can function as transcriptional repressors (Witzgall et al., 1994). GAL4-BD fused PARIS leads to decreased luciferase activity from a 5 X GAL4-luciferase reporter construct, which is restored by coexpression of WT parkin, but not Q311X mutant parkin (Figure S4A). To identify the PARIS DNA-binding consensus sequence, a chimeric protein containing the zinc finger domain (ZF) of PARIS (amino acids, 453–589) fused in frame to glutathione-S-tranferase (GST), GST-ZF-PARIS, was used in a cyclic amplification and selection of targets (CAST) assay (Figure 5A) (Funk and Wright, 1992; Shields and Yang, 1998). Immobilized GST-ZF-PARIS was incubated with a pool of oligonucleotides containing 26 random nucleotides. The final pool of oligonucleotides remaining after four rounds of CASTing followed by three rounds of electrophoretic mobility shift assays (EMSA) were cloned, sequenced and analyzed. Alignment of all the sequences using the program MACAW (National Center for Biotechnology Information) reveals a consensus sequence with a core sequence composed of TATTTT (T/G) (Figure 5A). Nineteen out of twenty-four sequences contained the core sequence and 3 out of 19 sequence tags were identified as duplicates indicating this is the primary DNA-binding sequence for PARIS. The TATTTT (T/G) consensus 694 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
sequence is an insulin response sequence (IRS) designated the phosphoenolpyruvate carboxykinase (PEPCK)-like motif (PLM) (Daitoku et al., 2003; O’Brien et al., 2001), which is involved in the regulation of transcripts involved in energy metabolism and insulin responsiveness (Mounier and Posner, 2006). A NCBI survey of IRS/PLM responsive transcripts reveals that members of the PPARg coactivator-1 (PGC-1) family of transcriptional coactivators are regulated by IRS sites. PGC-1a contains three IRS/PLM elements within its 50 -promoter region (Figure 5B). The activity of the 1kb human PGC-1a promoter is decreased approximately 40% in the presence of PARIS, which is rescued by parkin overexpression (Figure 5B). Similar results are observed with the 2 Kb mouse PGC-1a promoter (Figure S4B). The familial parkin mutant, Q311X, has minimal effects on the PGC-1a promoter (Figure 5B). PARIS represses PGC-1a promoter activity by binding to the IRS sites, since it fails to inhibit the reporter in which the three IRS sites are deleted in the context of the PGC-1a promoter-reporter construct (Figure 5B). Individual IRS site mutants (IRS1-M, IRS2-M, IRS3-M) within the PGC-1a reporter that disrupt PARIS binding to the PGC-1a promoter as determined by EMSA (see Figure 5C) and a promoter construct, IRS123-M, containing all three mutations were evaluated in the promoter reporter assay (Figure S4C). IRS1-M and IRS3-M substantially reduce PGC-1a reporter promoter activity, whereas IRS2-M increases the activity (Figure S4C). PARIS overexpression inhibits IRS1-M, IRS2-M, IRS3-M PGC-1a reporter promoter activity suggesting that PARIS binds to and regulates all three sites (Figure S4C) as indicated from the EMSA assays (Figure 5C). EMSA shows that PARIS binds to the PGC-1a IRS elements since GST-PARIS (full length) elicits a shift of [32P]-labeled oligonucleotides containing the IRS1, IRS2, and IRS3 sequences of the PGC-1a promoter, whereas it fails to cause a shift of IRS1, IRS2 and IRS3 containing a single base substitution within the IRS sequence (Figure 5C). Additionally eight zinc finger mutants of PARIS were assessed for their ability to repress the PGC-1a reporter promoter construct (Figure S4D and S4E). M1, M2, M8 PARIS mutants are able to repress PGC-1a reporter promoter activity, similar to WT PARIS. M3, M4, M5, M6 PARIS mutants partially repress PGC-1a reporter promoter activity, whereas the M7 (C571A) PARIS mutant has no affect on PGC-1a reporter promoter activity. The M7 GST-C571A-PARIS mutant has substantially reduced IRS-binding capacity as determined by EMSA (Figure S4F). Chromatin immunoprecipitation (ChIP) indicates that PARIS binds to the endogenous PGC-1a promoter in SH-SY5Y cells (Figure 5D) and mouse brain (Figure 5E). PARIS also binds to the endogenous PGC-1a promoter in human brain (Figure 5F) and consistent with its upregulation in PD striatum (see Figures 4C and 4D) there is enhanced PARIS occupancy of endogenous PGC-1a in PD striatum compared to control (Figure 5G). We performed luciferase reporter assay in SH-SY5Y cells and ChIP assays in PD versus control striatum and SH-SY5Y cells with phosphoenolpyruvate carboxykinase (PEPCK) (Figure S4G) and glucose-6-phosphatase (G6Pase) (Figure S4H). The luciferase reporter assay shows that overexpression of PARIS enhances the promoter activity of rat PEPCK, but not mouse
Figure 5. PARIS Acts as Transcriptional Repressor of PGC-1 a (A) Identification and MACAW alignment of the PARIS DNA-binding sequence as determined by CASTing. Darker colors represent a greater degree of overlap of the segment pairs (bottom right - % overlap). Asterisk, duplicate sequence tags. (B) Relative luciferase activity of the 1-kilobase human PGC-1a (992 to +90) compared to Renilla luciferase ± PARIS or ± parkin or ± familial mutant Q311X parkin, n = 3. Location of IRS, CRE motifs and oligos for human ChIP (arrows) in the PGC-1a promoter construct (top of panel). Immunoblot analysis confirms the expression of FLAG-PARIS, MYC-Parkin and MYC-Q311X parkin (right panel). (C) EMSA of GST-PARIS of 32P-labeled WT (WT-32P) IRS oligonucleotides (IRS1-WT, IRS2-WT, IRS3-WT) of the human PGC-1a promoter and 32P-labeled mutant (T/g) (MT-32P) IRS oligonucleotides (IRS1-MT, IRS2-MT, IRS3-MT). Unlabeled WT (WT cold) IRS oligonucleotides disrupt the GST-PARIS-DNA protein complexes with the WT-32P IRS oligonucleotides, n = 3. Unlabeled mutant probes (MT cold) has no effect on mutant (MT-32P). Arrow indicates specific PARISshifted probe; NS, nonspecific; FS, fragmented PARIS-shifted probe; FP, free probe. (D) PARIS occupies the endogenous PGC-1 a promoter as determined by ChIP assay with anti-PARIS polyclonal antibodies in SH-SY5Y cells, n = 3. (E) PARIS occupies the endogenous mouse PGC-1 a promoter as determined by ChIP in mouse whole brain, n = 3. Mouse specific IRS primers are indicated in Figure S4B. (F) ChIP assay of endogenous PARIS binding to the IRS region of the human PGC-1a promoter in human PD and aged-matched control (CTL) striatum, control n = 3; PD n = 4. (G) Quantitation of ChIP in panel F. Human specific IRS primers for D and F are indicated in panel B. (H) Real-time qRT-PCR of PGC-1a, GFP and b-actin following transient transfection of GFP, GFP-PARIS or GFP-C571A PARIS mutant, n = 4. (I) Immunoblot analysis of PGC-1a, GFP and b-actin following transient transfection of GFP, GFP-PARIS or GFP-C571A PARIS mutant, n = 4. (J) Quantitation of the immunoblots in panel I normalized to b-actin, n = 4. (K) Immunoblot analysis of parkin, PARIS, PGC-1a and b-actin in double-knockdown experiments via lentiviral transduction of shRNA-parkin and/or shRNAPARIS in SH-SY5Y cells, n = 3. (L) Quantitation of the immunoblots in panel K normalized to b-actin, n = 3, sh = shRNA. (M) Relative mRNA levels of PGC-1a normalized to GAPDH, n = 3. Quantitative data = mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired two-tailed Student’s t test (G), ANOVA with Student-Newman-Keuls post hoc analysis (B, H, J, L, M). See also Figure S4 and Table S2, Table S3, and Table S5.
Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 695
G6Pase promoter activity (Figure S4I). The ChIP assay demonstrates that PARIS binds to the endogenous promoter of human PEPCK and G6Pase in SH-SY5Y cells and in control and PD postmortem striatum (Figure S5J). In contrast to PGC-1a, there is not enhanced occupancy of the PEPCK and G6Pase promoter by PARIS. These data suggest that PARIS can bind to the promoter of PEPCK and G6Pase, but in contrast to PGC-1a it positively regulates PEPCK and it has no appreciable effect on G6Pase. Thus, the transcriptional repressive effects of PARIS are relatively specific to PGC-1a. GFP-WT-PARIS overexpression leads to approximately a 75% reduction in PGC-1a mRNA (Figure 5H) and approximately a 60% reduction in protein levels of PGC-1a (Figures 5I and 5J), whereas the GFP-C571A-PARIS mutant has no effect on PGC1-a protein or message levels (Figures 5H–5J). Lentiviral shRNA-parkin leads to a two fold increase in the level of PARIS followed by a 66% reduction of PGC-1a (Figures 5K and 5L). To determine whether the reduction in PGC-1a levels induced by the absence of parkin is dependent on the presence of PARIS, a double-knockdown experiment was performed by lentiviral shRNA-parkin and/or shRNA-PARIS in SH-SY5Y cells. Knockdown of PARIS prevents the downregulation of PGC-1a levels induced by parkin knockdown (Figures 5K and 5L). Knockdown of PARIS results in a 3 fold increase in PGC-1a protein levels (Figures 5K and 5L) and a 3.5 fold increase in PGC-1a mRNA (Figure 5M). Knockdown of PARIS in the setting of parkin knockdown also prevents the downregulation of PGC-1a mRNA (Figure 5M). These results taken together indicate that PARIS is a transcriptional repressor that negatively regulates the levels of endogenous PGC-1a and that the downregulation of PGC-1a in the absence of parkin is due to the upregulation of PARIS. Identification of NRF-1 as a Potential In Vivo PGC-1a Target Gene in PD PGC-1a is a transcriptional coactivator that regulates a variety of genes (Lin et al., 2005; St-Pierre et al., 2006). To determine which PGC-1a target genes are coregulated by PARIS in PD, real-time quantitative RT-PCR (qRT-PCR) was performed on a variety of PGC-1a target genes in PD SN and striatum. We measured the levels PGC-1a coregulated genes that play important roles in mitochondrial function and oxidant metabolism including nuclear respiratory factor-1 (NRF-1), copper/zinc superoxide dismutase (SOD1), manganese SOD (SOD2), glutathione peroxidase (GPx1), catalase (CAT), mitochondrial uncoupling proteins (UCP2 and UCP3), mitochondrial transcription factor A (Tfam) and the oxidative phosphorylation regulators, ATP5b, cytochrome C (CytC) and cytochrome C oxidase (COX II and IV) (Figure 6A and Figure S5A). In addition, the levels of other genes containing IRS/PLM in their promoter including PEPCK, G6Pase, insulin-like-growth-factor-binding protein 1 (IGFBP-1), tyrosine aminotransferase (TAT), and apolipoprotein C III (APOC3) were monitored along with PGC-1a to assess whether PGC-1a is selectively affected in PD (Figure 6A and Figure S5A). The levels of PARIS and parkin were also assessed as controls. We find that PGC-1a and NRF-1 mRNA are downregulated in PD SN and striatum compared to control SN and striatum (Figure 6A and Figure S5A). In PD SN ATP5B is also significantly downregulated and CAT is significantly upregulated (Figure 6A). All other 696 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
PGC-1a dependent genes are not significantly altered (Figure 6A and Figure S5A). In addition there is no significant change in the levels of the IRS/PLM responsive transcripts PEPCK, G6Pase and IGFBP-1 (Figure 6A and Figure S5A). TAT and APOC3 are not detectable. No significant alteration in the mRNA level of PARIS and parkin is observed between PD and control SN and striatum (Figure 6A and Figure S5A) indicating that the upregulation in PARIS protein levels (see Figures 4C and 4D) are most likely due to impairment of parkin E3 ubiquitin ligase activity. Moreover, the absence of an alteration in the mRNA levels of PARIS and parkin suggest that the changes in the mRNA levels of PGC-1a and NRF-1 are specific and not due to the degenerative process that occurs in PD. As shown above (see Figures 4C and 4D) PARIS protein is upregulated almost 3-fold in PD SN (Figures 6B and 6C) and greater than two-fold in PD striatum (Figure S5B and S5C) compared to control SN and striatum respectively. Accompanying the upregulation of PARIS is the downregulation of PGC1a and NRF-1 in SN (Figures 6B and 6C) and striatum (Figures S5B and S5C). There is a trend toward redistribution of parkin from the soluble to insoluble fraction in SN (Figures 6B and 6C) and parkin shifts from the soluble to insoluble fraction in PD striatum (Figures S5B and S5C). There is a strong negative correlation between the protein levels of PARIS and PGC-1a (R2 = 0.5195, p < 0.05) and NRF-1 (R2 = 0.8015, p < 0.01) in the striatum and between PARIS and PGC-1a (R2 = 0.6955, p < 0.05) and NRF-1 (R2 = 0.5979, p < 0.05) in the SN and a positive correlation between PGC-1a and NRF-1 striatum (R2 = 0.6827, p < 0.05) and SN (R2 = 0.6488, p < 0.05) (Figure S5D). These results taken together indicate that PARIS accumulates in the nigrostriatal pathway in PD leading to the downregulation of PGC-1a and the PGC-1a dependent gene, NRF-1. Downregulation of PGC-1a and NRF-1 in Conditional Parkin KO Mice Requires PARIS In adult conditional parkin KO mice (see Figures 4H–4J) four weeks after Lenti-GFPCre mediated parkin deletion there is a greater than two fold upregulation of PARIS (Figures 6D and 6E and Figures 4H–4J), comparable to that which occurs in sporadic PD SN (see Figures 4C and 4D) and a concomitant downregulation of PGC-1a and NRF-1 (Figures 6D and 6E). The alteration of PGC-1a and NRF1 results from the reduction of their mRNA levels (Figure 6F). Moreover, we analyzed the mRNA levels of PGC-1a target genes in the ventral midbrain of the Lenti-GFPCre-mediated conditional parkin KO and show a significant reduction of PGC-1a, SOD2, and NRF-1 and no significant alteration in other sampled PGC-1a regulated transcripts (Figure S5E) similar to what occurs in sporadic PD brain. In addition, the levels of other genes containing IRS/PLM in their promoter including PEPCK, G6Pase, IGFBP-1, TAT, and APOC were monitored. Of the genes containing IRS/PLM in their promoters, only PGC-1a is significantly downregulated (Figure S5E). The upregulation of PARIS and subsequent downregulation of PGC-1a and NRF-1 occurs prior to the loss of DA neurons since there is no appreciable loss of DA neurons at 3 months after the Lenti-GFPCre injection (see Figure 6H). Moreover, laser capture microdissection (LCM) was performed prior
Figure 6. Identification of PGC-1a and NRF-1 as Pathological In Vivo Targets of Accumulated PARIS in PD Brain and Conditional Parkin KO Mice (A) Real-time qRT-PCR of IRS (PEPCK-like motif)-containing genes and PGC-1a dependent genes in PD SN compared to age-matched CTL-SN normalized to GAPDH, n = 3–4 per group. (B) Immunoblots of PARIS, PGC-1a, parkin and NRF-1 in soluble and insoluble fractions of PD SN compared to age-matched CTL-SN. (C) Quantitation of the immunoblots in panel B normalized to b-actin, PD, n = 4; Control, n = 3. (D) Top panel, experimental illustration of stereotaxic intranigral virus injection. Below are immunoblots of parkin, PARIS, PGC-1a, NRF-1, b-actin and GFP, 4 weeks after stereotactic delivery of Lenti-GFP, Lenti-GFPCre, Lenti-GFPCre + shRNA-dsRed, or Lenti-GFPCre + shRNA-PARIS into exon 7 floxed parkin mice (parkinFlx/Flx), n = 3 per group. *nonspecific band. (E) Quantitation of the immunoblots in panel D normalized to b-actin, n = 3 per group. (F) The alteration of PGC-1a and NRF1 shown in panel D and E results from transcriptional changes as determined by real-time qRT-PCR, n = 3 per group. (G) Top panel, experimental illustration of stereotaxic intranigral virus injection. Below is TH immunostaining of representative mice midbrain sections from SN of parkinFlx/Flx mice injected with Lenti-GFP, and Lenti-GFPCre ± Lenti-shRNA-PARIS 10 months post-injection of virus. (H) Stereological assessment of TH and Nissl positive neurons in the SN of parkinFlx/Flx mice injected with Lenti-GFP, and Lenti-GFPCre ± Lenti-shRNA-PARIS 3 (n = 3 per group) and 10 months (n = 7 per group) after injection of virus. Data = mean ± SEM, *p < 0.05, **p < 0.01 and ***p < 0.001, unpaired two-tailed Student’s t test (A and C) and ANOVA with Student-Newman-Keuls post hoc analysis (E,F, and H). See also Figure S5, Table S4, and Table S5.
Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 697
to the loss of DA neurons to obtain mRNA from TH positive neurons transduced with GFP-Cre from conditional parkin KO mice 4 weeks after the Lenti-GFPCre injection to ascertain whether the reduction of PGC-1a is cell autonomous in DA neurons (Figures S5F and S5G). We find a robust reduction of PGC-1a mRNA in TH-positive DA neurons of conditional parkin KO mice, whereas the levels of PARIS mRNA are unchanged (Figure S5F and S5G). PARIS is only modestly upregulated in germline parkin exon 7 KO mice (also see Figures 4E and 4F). PGC-1a and NRF-1 protein levels in germline parkin exon 7 KO mice are comparable to those of WT mice (Figures S5H and S5I). Thus, germline deletion of parkin apparently leads to compensatory changes that prevent substantial alterations in the levels of PGC-1a and NRF-1. We further developed the Cre-flox conditional parkin exon 7 KO model by introducing lentiviral shRNA-PARIS along with lenti-GFPCre into parkinFlx/Flx mice to address whether the changes in PGC-1a and NRF-1 are due to PARIS (Figures 6D–6F). Coadministration of lentiviral shRNA-PARIS along with Lenti-GFPCre prevents the changes in PGC-1a and NRF-1 protein and mRNA as compared to control lentiviral shRNA-dsRed plus Lenti-GFPCre (Figures 6D–6F). These results taken together indicate that PARIS accumulates in the nigrostriatal pathway in PD and in models of parkin inactivation leading to the downregulation of PGC-1a and the PGC-1a dependent gene, NRF-1. These changes in PGC-1a and NRF-1 due to the loss of parkin are cell autonomous and precede the loss of DA neurons and are due to the upregulation of PARIS, since knockdown of PARIS prevents these changes. Moreover, since deletion of parkin from adult mice leads to similar events that occur in sporadic PD, it is likely that the absence and/or inactivation of parkin in PD leads to PARIS upregulation and impairment of PGC-1a signaling. Neurodegeneration in Conditional Parkin KO Mice Requires PARIS Conditional KO of parkin leads to a significant reduction in tyrosine hydroxylase (TH) positive and Nissl stained DA neurons 10 months after stereotaxic injection of Lenti-GPFCre into the SN of 6- to 8-week-old parkinFlx/Flx mice compared to parkinFlx/Flx mice injected with control Lenti-GFP (Figures 6G and 6H). The loss of DA neurons is progressive, since there is no substantial loss of DA neurons after 3 months (Figure 6H). PARIS is required for the loss of DA neurons in conditional parkin KO mice, since coadministration of lentiviral shRNA-PARIS along with Lenti-GFPCre significantly reduces the loss of DA neurons due to conditional KO of parkin (Figures 6G and 6H). Taken together these results indicate that conditional KO of parkin in adult mice leads to degeneration of DA neurons and the upregulation of PARIS is necessary to contribute to the demise of DA neurons. Overexpression of PARIS Kills Dopamine Neurons In Vivo: Restoration by Parkin and PGC-1a A PARIS overexpression model was developed in which AAV1-PARIS was stereotactically injected into the SN of C57Bl/6 mice and compared to mice injected with control AAV1-GFP virus (Figure 7). Stereotactic intranigral injection of 698 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
AAV1 effectively transduces the entire SN (Figure 7A). One month after stereotactic injection of the viruses, AAV1 mediated overexpression of PARIS leads to a greater than two-fold upregulation of PARIS levels in the SN of mice (Figures 7B and 7C) and it has no affect on parkin levels (Figures 7B and 7C). Accompanying the increase in PARIS levels is a concomitant downregulation of PGC-1a and NRF-1 protein levels (Figures 7B and 7C). PARIS overexpression leads to a greater than 40% reduction in TH positive and Nissl stained DA neurons positive neurons (Figure 7D and Figure 7E). We do not observe any substantial decrement in GABAergic neurons as assessed by GAD65/67 immunoreactivity via immunohistochemistry (Figure S6A) or via immunoblot (Figures S6B and S6C) in AAV1-PARIS versus AAV1-GFP transduced SN, indicating that PARIS overexpression is selectively detrimental to DA neurons. Coadministration of AAV1-Parkin with AAV1-PARIS prevents the loss of dopamine neurons induced by PARIS overexpression (Figures 7D and 7E). Lentiviral PGC-1a also prevents the loss of DA neurons induced by PARIS overexpression (Figures 7D and 7E). AAV1-mediated overexpression of PARIS was confirmed by immunoblot analysis and reduces PGC-1a and NRF-1 levels by 52% and 60%, respectively. These reductions are restored by co-overexpression of parkin or PGC-1a (Figures7F and 7G). These results indicate that PARIS overexpression is sufficient to downregulate PGC-1a, NRF-1 and selectively kill DA kills neurons through a PGC-1a dependent mechanism. DISCUSSION PARIS (ZNF746) Is a Parkin Substrate PARIS fulfills several criteria for a parkin substrate and in the absence of parkin activity in PD it accumulates making it an attractive pathogenic substrate. In sporadic PD, PARIS only accumulates in the striatum and SN. The selective inactivation of parkin in the striatum and SN in sporadic PD through nitrosative and dopaminergic stress (Chung et al., 2004; LaVoie et al., 2005) and c-Abl phosphorylation (Ko et al., 2010) most likely accounts for the accumulation of PARIS. Similarly, in parkin exon 7 KO mice, PARIS is only upregulated in the striatum and SN. However, PARIS accumulates in the cingulate cortex of AR-PD brains, suggesting that parkin can regulate the expression of PARIS outside the striatum and SN. Whether there is a differential regional or temporal upregulation of PARIS in AR-PD brains is not known as tissue from these areas was unavailable for analysis. The lack of an upregulation in the cortex of parkin exon 7 mice, suggests that there may differential and selective regulatory mechanisms in the cortex versus the striatum and SN. PARIS Is a Transcriptional Repressor of PGC-1a PARIS is a member of the family of KRAB zinc-finger proteins (KRAB-ZFPs) transcriptional repressors (Looman et al., 2002). Homologs of PARIS exist in simpler organisms suggesting that PARIS function may be evolutionarily conserved. PARIS seems to bind exclusively to IRS/PLM motifs, which provide an important site of regulation of a variety of IRS/PLM responsive proteins (Daitoku et al., 2003; Mounier and Posner, 2006; O’Brien
Figure 7. Introduction of AAV1-Parkin or Lenti-PGC-1a in Mice SN Protects from AAV1-PARIS-Mediated Selective Dopaminergic Neuronal Toxicity (A) Schematic illustration of intranigral viral injection and transduced brain regions. (B) Immunoblot analysis of PARIS, PGC-1a, parkin and NRF-1 four-weeks post intranigral injection of AAV1-PARIS, n = 5 per group. (C) Quantitation of the immunoblots in panel B normalized to b-actin. (D) TH staining of a representative section of mice injected with AAV1-GFP, AAV1-PARIS ± AAV1-parkin or AAV1-PARIS ± Lenti-PGC-1a. Each panel shows the noninjected side (Non) and contralateral injected side (Inj) and white pentagonal box indicates the SNpc. Enlarged images containing SNpc and SNpr are shown on the right panels. AAV1 encoding GFP was used as transduction control in all injection procedures. Broad regions including SNpc and SNpr were successfully transduced (left top panel). In right bottom panel, yellow rectangle indicates the region that PARIS and lenti- PGC-1a cotransduced. Approximately 30% of the SNpc was transduced with lenti-PGC-1a and this is the region, which is protected from PARIS toxicity, n = 6 per group. (E) Stereological TH, Nissl-positive neuronal counting, n = 6 per group. (F) Immunoblot analysis of PARIS, PGC-1a, parkin and NRF-1, n = 3. (G) Quantitation of the immunoblots in panel F normalized to b-actin. (H) Parkin-PARIS-PGC-1a pathway as a model in PD. Endogenous PARIS acts to maintain the balance of PGC-1a levels. In PD, parkin is inactivated by diverse insults such as familial mutations, reactive oxygen species (ROS), nitrosative (NO) and dopamine (DA) stress and PARIS accumulates. Accumulated PARIS continuously inhibits PGC-1a transcription leading to reduction in PGC-1a dependent genes. Ultimately this situation results in neurodegeneration in PD. Data = mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001; ANOVA with the Student-Newman-Keuls post hoc test. See also Figure S6.
et al., 2001). Our studies reveal that PGC-1a levels are potently and selectively regulated by PARIS, consistent with the observation that PGC-1a transcription is controlled by IRS/PLM motifs (Finck and Kelly, 2006). In PD SN and striatum and conditional parkin KO midbrain, PGC-1a is the only IRS/PLM regulated gene that is downregulated suggesting that PARIS selectively represses PGC-1a expression in PD. Notably PARIS is a physiological transcriptional repressor of PGC-1a, which directly and endogenously occupies the cis-regulatory elements of PGC-1a. PGC-1a is a transcriptional coactivator that controls the transcription of many genes involved in cellular metabolism including mitochondrial biogen-
esis and respiration and ROS metabolism (Finck and Kelly, 2006; St-Pierre et al., 2006). The levels of PGC-1a dependent genes are controlled in large part by the nature and composition of the PGC-1a transcriptional coactivator complex (Finck and Kelly, 2006). The PGC-1a dependent gene, NRF-1, appears to be particularly susceptible to the inhibitory effects of PARIS. It is likely that there is not only a PARIS transcriptional repressor protein complex that plays important regulatory roles in PARIS-mediated transcriptional repression and specificity but the repression of PARIS relies on the genomic context rather than simple IRS motif. In addition, under certain contexts it appears as though PARIS can act a transcriptional activator. Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 699
Consistent with this notion are our observations that PARIS can bind the PEPCK or G6Pase endogenous promoters via ChIP and it has no effect on G6Pase promoter-reporter activity, but enhances PEPCK promoter-reporter activity. DNA response elements can function as allosteric effectors that determine the transcriptional activity of regulators, explaining that regulators may activate transcription in the context of one gene, yet repress transcription in another (Lefstin and Yamamoto, 1998). Indeed, the IRS motifs in the PGC-1a promoter are organized differently then the IRS motifs of the PEPCK and G6Pase promoters, which may, in part account for the ability of PARIS to act as both a repressor and an activator. Other IRS/PLM responsive genes are likely to be regulated by PARIS and the identification of these genes, their potential regulation by PARIS and their potential role in PD requires additional study. Dopaminergic Degeneration in Conditional Parkin KO Mice Germline deletion of parkin using a variety of approaches created parkin KO mice with minimal phenotypes and no loss of DA neurons (Goldberg et al., 2003; Von Coelln et al., 2004). A number of reasons were recently put forward regarding the lack of overt degeneration of DA neurons in genetic animal models of PD including compensatory mechanisms (Dawson et al., 2010). Similar compensatory mechanisms are likely to occur in PD (Bezard et al., 2003; Palop et al., 2006), and compensation accounts, in part, for the age-dependence of PD. The ultimate failure of these compensatory mechanisms contributes to neurodegeneration. Since, our germline parkin exon 7 KO mice (Von Coelln et al., 2004) do not have degeneration of DA neurons, we generated conditional parkin exon 7 KO mice and deleted parkin from adult mice to avoid developmental compensation. Similar to germline deletion of parkin, embryonic deletion of glial-cell-line-derived neurotrophic factor (GDNF) had no deleterious effects on DA neurons (Pascual et al., 2008). However, conditional KO of GDNF in adult animals using a tamoxifen sensitive Cre-Lox recombination unmasked the ‘‘true physiologic action of GDNF’’ and led to mice with profound degeneration of catecholaminergic neurons (Pascual et al., 2008), indicating that developmental compensation can be overcome by deleting genes in the DA system in adulthood (for review see (Dawson et al., 2010)). Similar to the adult GDNF deleted mice, we observed progressive loss of DA neurons when parkin is deleted in adult mice. PARIS levels increase in parkin conditional KO mice similar to levels in AR-PD brain and in sporadic PD SN. Accompanying the upregulation of PARIS is downregulation of PGC-1a and NRF in conditional KO mice similar to the downregulation in sporadic PD striatum and SN. Consistent with the notion that compensation occurs in the germline parkin KO mice is the observation that PARIS is only modestly elevated and there is no alteration in the levels of PGC-1a and NRF and there is no loss of DA neurons. Parkin, PARIS, PGC-1a, and Neurodegeneration We cannot be certain that PARIS is the sole substrate or mechanism that contributes to DA neuron degeneration following parkin inactivation. Recent studies suggest that PINK1 in a mitochondrial membrane potential-dependent manner signals 700 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
and recruits parkin through as of yet unclear mechanisms from the cytoplasm to the mitochondria to initiate degradation of damaged mitochondria through autophagy (mitophagy) (VivesBauza and Przedborski, 2010). Since PGC-1a and NRF-1 are major transcriptional regulators of mitochondrial biogenesis (Finck and Kelly, 2006; St-Pierre et al., 2006; Ventura-Clapier et al., 2008; Wu et al., 1999), it is conceivable that when parkin decreases the number of mitochondria through mitophagy in response to mitochondrial damage, it is counter balanced by downregulation of PARIS levels as a homeostatic mechanism to increase mitochondrial size and number through regulation of PGC-1a and NRF-1 levels. The contribution of other parkin substrates that are regulated by the UPS and accumulate in AR-PD, sporadic PD and models of parkin inactivation, such as the aminoacyl-tRNA synthetase interacting multifunctional protein type 2 (AIMP2) also known as p38/JTV-1 (Corti et al., 2003; Ko et al., 2005) and far upstream element-binding protein 1 (FBP-1) (Ko et al., 2006), as well as others are not known. Additional substrates and mechanisms will need to be evaluated in future studies to determine their relative contributions to DA neuron degeneration due to parkin inactivation. However, we can be confident that PARIS upregulation is necessary and sufficient to cause DA neuron degeneration in models of parkin inactivation and that maintaining PGC-1a function is beneficial. Indeed PGC-1a has been implicated in other models of PD (St-Pierre et al., 2006). Moreover, PGC-1a-responsive genes are underexpressed in microdissected dopaminergic neurons of PD suggesting that the alteration of PGC-1a might be a cause of PD pathogenesis, not the consequence (Zheng et al., 2010). Herein, we hypothesize a model that accumulated PARIS in the setting of parkin’s inactivation represses PGC-1a expression leading to neurodegeneration. Consistent with this hypothesis are the observations that the reduction in PGC-1a levels and neurodegeneration are substantially reduced by knocking down PARIS levels in the setting of parkin inactivation. Since PARIS is likely to regulate other genes, the Parkin-PARISPGC-1a pathway is one potential contributory mechanism to PD pathogenesis. This parkin-PARIS-PGC-1a neurodegenerative pathway ultimately results in the selective vulnerability of dopamine neurons and accounts, in part, for the neurodegeneration in PD (Figure 7H). Further study is required to determine the full implications of the parkin-PARIS-PGC-1a neurodegenerative pathway in PD. Nonetheless, our results suggest that parkin inactivation acting through PARIS and downregulation of PGC-1a contributes to the pathogenesis of PD. EXPERIMENTAL PROCEDURES Yeast Two-Hybrid Screening Saccharomyces cerevisiae MaV203 was transformed with pDBLeu-R1-parkin, and 3 3 106 stable transformants were further transformed with 15 mg of pPC86 human brain cDNA library (Life Tech/GIBCO). Transformants were selected and confirmed according to the manufacturer’s instructions as previously described (Zhang et al., 2000). Antibodies Polyclonal PARIS antibodies were generated as described in supplemental information. Primary antibodies used include the following: goat antiPGC-1a (K-15, Santa Cruz Biotechnology), mouse anti-PGC-1a (4C1.3,
Calbiochem), goat anti-NRF1 (A-19, Santa Cruz Biotechnology), rabbit anti-NRF1 (ab34682, Abcam), mouse anti-parkin (Park8, Cell Signaling), rabbit anti-TH (Novus Biologicals), rabbit anti-glutamate decarboxylase (GAD) 65&67 (Millipore), rabbit anti-GFP (ab661, Abcam), mouse anti-GFP (ab1218, Abcam); Secondary antibodies used include Biotin-SP-conjugated goat anti-rabbit (Jackson ImmunoResearch lab), donkey anti-goat-Cy3, donkey anti-rabbit-Cy2/Cy3, donkey anti-mouse-Cy2 for immunostaining.
SUPPLEMENTAL INFORMATION
Plasmid Constructions Full-length parkin, and deletion mutants, Q311X, R42P, R275W, G430D and C431F parkin, HA-ubiquitin, PARIS, PARIS deletion and point mutations, GST-PARIS, ZNF398 vector were constructed as described in supplemental information. Construct integrity was verified by sequencing. Lentiviral pLV-PGC-1a plasmid was kindly provided by Dr. Dimitri Krainc (Massachusetts General Hospital, Harvard Medical School, Charlestown, USA), MYC-tagged XIAP was generously given from Dr. Kenny K. K. Chung (Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong).
This work was supported by NS38377, NS048206, NS051764, and the Bachmann Strauss Dystonia and Parkinson’s Disease Foundation. Y.L. is supported by Samsung Scholarship Foundation. T.M.D. is the Leonard and Madlyn Abramson Professor in Neurodegenerative Diseases. The authors thank Drs. Xin Guo, Sathya Siram, and members of the Dawson laboratory for assistance and contributions to the manuscript. We also thank Drs. Alex Kolodkin, David Ginty, and Solomon H. Snyder for helpful suggestions.
Cell culture and Transfection Human neuroblastoma SH-SY5Y cells (ATCC, Manassas, VA) were grown in DMEM containing 10% FBS and antibiotics in a humidified 5% CO2/95% air atmosphere at 37 C. For transient transfection, cells were transfected with indicated amounts of target vector using Lipofectamine Plus (Invitrogen), according to manufacturer’s instructions. For coimmunoprecipitation from cell cultures, SH-SY5Y cells were transfected with 2 mg of each plasmid, unless otherwise indicated in supplemental information. For the ubiquitination assay, SH-SY5Y cells were transiently transfected with 2 mg of pRK5Myc-tagged parkin, Myc-tagged parkin (C431F, G430D, R275W, Q311X) pCMV-FLAG-PARIS, and 2 mg of pMT123-HA-ubiquitin plasmids for 48 hr. For the luciferase assay SH-SY5Y cells were transiently transfected with pCMV-empty vector or pCMV-FLAG-PARIS with either wild-type or Q311X parkin, or pGL3-Basic, pGL3-PGC-1a promoter-Luciferase, or pGL3-PGC1a promoter deletion mutant (a gift from Akyoshi Fukamizu, University of Tsukuba, Japan) (Daitoku et al., 2003) for firefly Luciferase assay and 0.1 mg pRL-TK vector (Promega) for Renilla luciferase control. Immunocytochemistry and Immunblot Analysis Immunocytochemistry and immunoblot analysis was performed as described in the supplemental information. In vitro Interaction and Ubiquitination Assays GST-PARIS and His-Parkin were used in in vitro interaction and ubiquitination assays as described in the supplemental information. CAST, EMSA, ChIP, qRT-PCR Assays We followed a previous published protocol with modification for CAST ((Cyclic Amplification and Selection of Targets) (Voz et al., 2000) as described in the supplemental information. GST, GST-PARIS, GST-C571A-PARIS were used for electrophoretic mobility shift assays (EMSA) as described in the supplemental information. Chromatin immunoprecipitation was carried out according to the manufacturer’s instruction as described in the supplemental information. Primers used for real-time qRT-PCR are listed in Table S4. Conditional Parkin Knockout To generate Cre-flox conditional model of parkin knock out, a lentiviral vector expressing GFP fused Cre recombinase (Lenti-GFPCre) was stereotaxically introduced into exon 7 floxed parkin mice (parkinFlx/Flx) using the coordinates in Supplemental Information. Furthermore lentiviral shRNA-PARIS was coadministrated along with Lenti-GFPCre to demonstrate whether the changes in PGC-1a and NRF-1 are due to PARIS. Statistics Quantitative data are presented as the mean ± SEM. Statistical significance was either assessed via an unpaired two-tailed Student’s t test or an ANOVA test with Student-Newman-Keuls post hoc analysis. Assessments were considered significant with a p < 0.05.
Supplemental Information includes Extended Experiment Procedures, six figures, and five tables and can be found with this article online at doi:10. 1016/j.cell.2011.02.010. ACKNOWLEDGMENTS
Received: August 4, 2010 Revised: November 22, 2010 Accepted: February 4, 2011 Published: March 3, 2011 REFERENCES Bezard, E., Gross, C.E., and Brotchie, J.M. (2003). Presymptomatic compensation in Parkinson’s disease is not dopamine-mediated. Trends Neurosci. 26, 215–221. Chung, K.K., Thomas, B., Li, X., Pletnikova, O., Troncoso, J.C., Marsh, L., Dawson, V.L., and Dawson, T.M. (2004). S-nitrosylation of parkin regulates ubiquitination and compromises parkin’s protective function. Science 304, 1328–1331. Corti, O., Hampe, C., Koutnikova, H., Darios, F., Jacquier, S., Prigent, A., Robinson, J.C., Pradier, L., Ruberg, M., Mirande, M., et al. (2003). The p38 subunit of the aminoacyl-tRNA synthetase complex is a Parkin substrate: linking protein biosynthesis and neurodegeneration. Hum. Mol. Genet. 12, 1427– 1437. Daitoku, H., Yamagata, K., Matsuzaki, H., Hatta, M., and Fukamizu, A. (2003). Regulation of PGC-1 promoter activity by protein kinase B and the forkhead transcription factor FKHR. Diabetes 52, 642–649. Dawson, T.M., and Dawson, V.L. (2010). The role of parkin in familial and sporadic Parkinson’s disease. Mov. Disord. 25 (Suppl 1), S32–S39. Dawson, T.M., Ko, H.S., and Dawson, V.L. (2010). Genetic animal models of Parkinson’s disease. Neuron 66, 646–661. Finck, B.N., and Kelly, D.P. (2006). PGC-1 coactivators: inducible regulators of energy metabolism in health and disease. J. Clin. Invest. 116, 615–622. Funk, W.D., and Wright, W.E. (1992). Cyclic amplification and selection of targets for multicomponent complexes: myogenin interacts with factors recognizing binding sites for basic helix-loop-helix, nuclear factor 1, myocyte-specific enhancer-binding factor 2, and COMP1 factor. Proc. Natl. Acad. Sci. USA 89, 9484–9488. Gasser, T. (2007). Update on the genetics of Parkinson’s disease. Mov. Disord. 22, S343–S350. Goldberg, M.S., Fleming, S.M., Palacino, J.J., Cepeda, C., Lam, H.A., Bhatnagar, A., Meloni, E.G., Wu, N., Ackerson, L.C., Klapstein, G.J., et al. (2003). Parkin-deficient mice exhibit nigrostriatal deficits but not loss of dopaminergic neurons. J. Biol. Chem. 278, 43628–43635. Imai, Y., Soda, M., Hatakeyama, S., Akagi, T., Hashikawa, T., Nakayama, K.I., and Takahashi, R. (2002). CHIP is associated with Parkin, a gene responsible for familial Parkinson’s disease, and enhances its ubiquitin ligase activity. Mol. Cell 10, 55–67. Ko, H.S., Kim, S.W., Sriram, S.R., Dawson, V.L., and Dawson, T.M. (2006). Identification of far upstream element-binding protein-1 as an authentic Parkin substrate. J. Biol. Chem. 281, 16193–16196. Ko, H.S., Lee, Y., Shin, J.H., Karuppagounder, S.S., Gadad, B.S., Koleske, A.J., Pletnikova, O., Troncoso, J.C., Dawson, V.L., and Dawson, T.M. (2010).
Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc. 701
Phosphorylation by the c-Abl protein tyrosine kinase inhibits parkin’s ubiquitination and protective function. Proc. Natl. Acad. Sci. USA 107, 16691–16696. Ko, H.S., von Coelln, R., Sriram, S.R., Kim, S.W., Chung, K.K., Pletnikova, O., Troncoso, J., Johnson, B., Saffary, R., Goh, E.L., et al. (2005). Accumulation of the authentic parkin substrate aminoacyl-tRNA synthetase cofactor, p38/JTV-1, leads to catecholaminergic cell death. J. Neurosci. 25, 7968–7978. LaVoie, M.J., Ostaszewski, B.L., Weihofen, A., Schlossmacher, M.G., and Selkoe, D.J. (2005). Dopamine covalently modifies and functionally inactivates parkin. Nat. Med. 11, 1214–1221. Lefstin, J.A., and Yamamoto, K.R. (1998). Allosteric effects of DNA on transcriptional regulators. Nature 392, 885–888. Lin, J., Handschin, C., and Spiegelman, B.M. (2005). Metabolic control through the PGC-1 family of transcription coactivators. Cell Metab. 1, 361–370. Looman, C., Abrink, M., Mark, C., and Hellman, L. (2002). KRAB zinc finger proteins: an analysis of the molecular mechanisms governing their increase in numbers and complexity during evolution. Mol. Biol. Evol. 19, 2118–2130. Messick, T.E., Russell, N.S., Iwata, A.J., Sarachan, K.L., Shiekhattar, R., Shanks, J.R., Reyes-Turcu, F.E., Wilkinson, K.D., and Marmorstein, R. (2008). Structural basis for ubiquitin recognition by the Otu1 ovarian tumor domain protein. J. Biol. Chem. 283, 11038–11049. Mounier, C., and Posner, B.I. (2006). Transcriptional regulation by insulin: from the receptor to the gene. Can. J. Physiol. Pharmacol. 84, 713–724. O’Brien, R.M., Streeper, R.S., Ayala, J.E., Stadelmaier, B.T., and Hornbuckle, L.A. (2001). Insulin-regulated gene expression. Biochem. Soc. Trans. 29, 552–558. Palop, J.J., Chin, J., and Mucke, L. (2006). A network dysfunction perspective on neurodegenerative diseases. Nature 443, 768–773. Pascual, A., Hidalgo-Figueroa, M., Piruat, J.I., Pintado, C.O., Gomez-Diaz, R., and Lopez-Barneo, J. (2008). Absolute requirement of GDNF for adult catecholaminergic neuron survival. Nat. Neurosci. 11, 755–761. Savitt, J.M., Dawson, V.L., and Dawson, T.M. (2006). Diagnosis and treatment of Parkinson disease: molecules to medicine. J. Clin. Invest. 116, 1744–1754.
Sriram, S.R., Li, X., Ko, H.S., Chung, K.K., Wong, E., Lim, K.L., Dawson, V.L., and Dawson, T.M. (2005). Familial-associated mutations differentially disrupt the solubility, localization, binding and ubiquitination properties of parkin. Hum. Mol. Genet. 14, 2571–2586. St-Pierre, J., Drori, S., Uldry, M., Silvaggi, J.M., Rhee, J., Jager, S., Handschin, C., Zheng, K., Lin, J., Yang, W., et al. (2006). Suppression of reactive oxygen species and neurodegeneration by the PGC-1 transcriptional coactivators. Cell 127, 397–408. Tanaka, K., Suzuki, T., Hattori, N., and Mizuno, Y. (2004). Ubiquitin, proteasome and parkin. Biochim. Biophys. Acta 1695, 235–247. Ventura-Clapier, R., Garnier, A., and Veksler, V. (2008). Transcriptional control of mitochondrial biogenesis: the central role of PGC-1alpha. Cardiovasc. Res. 79, 208–217. Vives-Bauza, C., and Przedborski, S. (2010). PINK1/Parkin direct mitochondria to autophagy. Autophagy 6, 315–316. Von Coelln, R., Thomas, B., Savitt, J.M., Lim, K.L., Sasaki, M., Hess, E.J., Dawson, V.L., and Dawson, T.M. (2004). Loss of locus coeruleus neurons and reduced startle in parkin null mice. Proc. Natl. Acad. Sci. USA 101, 10744–10749. Voz, M.L., Agten, N.S., Van de Ven, W.J., and Kas, K. (2000). PLAG1, the main translocation target in pleomorphic adenoma of the salivary glands, is a positive regulator of IGF-II. Cancer Res. 60, 106–113. Winklhofer, K.F., Henn, I.H., Kay-Jackson, P.C., Heller, U., and Tatzelt, J. (2003). Inactivation of parkin by oxidative stress and C-terminal truncations: a protective role of molecular chaperones. J. Biol. Chem. 278, 47199–47208. Witzgall, R., O’Leary, E., Leaf, A., Onaldi, D., and Bonventre, J.V. (1994). The Kruppel-associated box-A (KRAB-A) domain of zinc finger proteins mediates transcriptional repression. Proc. Natl. Acad. Sci. USA 91, 4514–4518. Wu, Z., Puigserver, P., Andersson, U., Zhang, C., Adelmant, G., Mootha, V., Troy, A., Cinti, S., Lowell, B., Scarpulla, R.C., et al. (1999). Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1. Cell 98, 115–124.
Shields, J.M., and Yang, V.W. (1998). Identification of the DNA sequence that interacts with the gut-enriched Kruppel-like factor. Nucleic Acids Res. 26, 796–802.
Zhang, Y., Gao, J., Chung, K.K., Huang, H., Dawson, V.L., and Dawson, T.M. (2000). Parkin functions as an E2-dependent ubiquitin- protein ligase and promotes the degradation of the synaptic vesicle-associated protein, CDCrel-1. Proc. Natl. Acad. Sci. USA 97, 13354–13359.
Shimura, H., Hattori, N., Kubo, S., Mizuno, Y., Asakawa, S., Minoshima, S., Shimizu, N., Iwai, K., Chiba, T., Tanaka, K., et al. (2000). Familial Parkinson disease gene product, parkin, is a ubiquitin-protein ligase. Nat. Genet. 25, 302–305.
Zheng, B., Liao, Z., Locascio, J.J., Lesniak, K.A., Roderick, S.S., Watt, M.L., Eklund, A.C., Zhang-James, Y., Kim, P.D., Hauser, M.A., et al. (2010). PGC-1alpha, a potential therapeutic target for early intervention in Parkinson’s disease. Sci. Transl. Med. 2, 52ra73.
702 Cell 144, 689–702, March 4, 2011 ª2011 Elsevier Inc.
Activation of Multiple Proto-oncogenic Tyrosine Kinases in Breast Cancer via Loss of the PTPN12 Phosphatase Tingting Sun,1 Nicola Aceto,10,14 Kristen L. Meerbrey,1,2,14 Jessica D. Kessler,1,2 Chunshui Zhou,11 Ilenia Migliaccio,6 Don X. Nguyen,13 Natalya N. Pavlova,11 Maria Botero,6 Jian Huang,6 Ronald J. Bernardi,3 Earlene Schmitt,1,2 Guang Hu,11 Mamie Z. Li,11 Noah Dephoure,12 Steven P. Gygi,12 Mitchell Rao,2 Chad J. Creighton,4,5 Susan G. Hilsenbeck,4,5 Chad A. Shaw,2 Donna Muzny,9 Richard A. Gibbs,2,5,9 David A. Wheeler,5,9 C. Kent Osborne,5,6,7,8 Rachel Schiff,5,6,7,8 Mohamed Bentires-Alj,10,15 Stephen J. Elledge,11,15 and Thomas F. Westbrook1,2,3,5,* 1Verna
& Marrs McLean Department of Biochemistry & Molecular Biology of Molecular & Human Genetics 3Department of Pediatrics 4Division of Biostatistics 5Dan L. Duncan Cancer Center 6The Lester & Sue Smith Breast Center 7Department of Medicine 8Department of Molecular & Cellular Biology 9The Human Genome Sequencing Center Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA 10Friedrich Miescher Institute for Biomedical Research, Basel CH-4002, Switzerland 11Howard Hughes Medical Institute, Department of Genetics, Harvard Medical School, Division of Genetics, Brigham & Women’s Hospital, Boston, MA 02115, USA 12Department of Cell Biology, Harvard Medical School, Boston, MA 02115, USA 13Department of Pathology, Yale University School of Medicine, New Haven, CT 06520, USA 14These authors contributed equally to this work 15These authors contributed equally to this work *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.003 2Department
SUMMARY
Among breast cancers, triple-negative breast cancer (TNBC) is the most poorly understood and is refractory to current targeted therapies. Using a genetic screen, we identify the PTPN12 tyrosine phosphatase as a tumor suppressor in TNBC. PTPN12 potently suppresses mammary epithelial cell proliferation and transformation. PTPN12 is frequently compromised in human TNBCs, and we identify an upstream tumor-suppressor network that posttranscriptionally controls PTPN12. PTPN12 suppresses transformation by interacting with and inhibiting multiple oncogenic tyrosine kinases, including HER2 and EGFR. The tumorigenic and metastatic potential of PTPN12deficient TNBC cells is severely impaired upon restoration of PTPN12 function or combined inhibition of PTPN12-regulated tyrosine kinases, suggesting that TNBCs are dependent on the proto-oncogenic tyrosine kinases constrained by PTPN12. Collectively, these data identify PTPN12 as a commonly inactivated tumor suppressor and provide a rationale for combinatorially targeting proto-oncogenic tyrosine
kinases in TNBC and other cancers based on their profile of tyrosine-phosphatase activity. INTRODUCTION Breast cancer is the most common malignancy among women and is comprised of a heterogeneous group of diseases stratified into three major subtypes (Di Cosimo and Baselga, 2010; Jemal et al., 2008). Two of these are defined by expression of steroid hormone receptors (estrogen receptor [ER] and progesterone receptor [PR]) or amplification/overexpression of the receptor tyrosine kinase (RTK) HER2. Agents targeting these proteins have led to significant increases in patient survival (Osborne, 1998; Slamon et al., 1989). In contrast, the triple-negative breast cancer (TNBC) subtype is defined only by the absence of ER and PR expression or HER2 amplification, underscoring our lack of understanding of key pathways driving TNBC. TNBC comprises approximately 20% of breast cancer, and the prognosis for patients with TNBC is poor because of its propensity for recurrence and metastasis and a lack of effective targeted therapeutics (Hurvitz and Finn, 2009). Consequently, a major challenge remaining in breast cancer treatment is to identify aberrant signaling networks underlying this aggressive subtype of breast cancer. Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 703
Protein tyrosine phosphorylation plays a central role in cellular physiology and in cancer (Hunter, 2009). For instance, aberrant activation of the human epidermal growth factor receptor (HER) family of RTKs occurs frequently in many malignancies, including breast, lung, and brain cancer (Yarden and Sliwkowski, 2001). Although HER2 signaling is required for tumor maintenance in HER2-amplified disease (Slamon et al., 1989), the role of HER2 and other HER family tyrosine kinases (TKs) in breast cancer subtypes lacking HER2 amplification is unclear. Notably, in the absence of TK mutational activation or amplification, other pathways that regulate HER2 and other TKs may also be involved in tumorigenesis but remain to be elucidated. Protein tyrosine phosphatases (PTPs) also regulate the equilibrium of tyrosine phosphorylation and, in principle, can serve as antagonists to TK signaling to play a prominent role in tumor suppression (Tonks, 2006). However, much less is known about the role of PTPs in suppressing tumorigenesis. In this study we have employed an unbiased functional screen for tumor suppressors and identified a role for the tyrosine phosphatase PTPN12 in TNBC. Loss of PTPN12 leads to malignant transformation of human mammary epithelial cells (HMECs) through multi-TK activation. PTPN12 function is frequently compromised in TNBC by deletions, defective sequence variants, or loss of expression, suggesting that HER2/EGFR and other RTK signaling is aberrantly activated in non-HER2-amplified breast cancers. Restoring PTPN12 expression in PTPN12-deficient TNBC cells inhibits their proliferation, tumorigenicity, and metastatic potential in vivo. These results identify PTPN12 as a tumor suppressor and suggest that combinatorial TK signaling is a key dependency in TNBC and, therefore, a target for cancer therapies. RESULTS PTPN12 Suppresses Transformation of HMECs Signal transduction networks play key roles in the malignant behavior of cancer cells. To identify new networks that regulate cellular transformation in human breast cancer, we performed a genetic screen for kinases and phosphatases that suppress cellular transformation in genetically engineered HMECs (Figure 1A). HMECs isolated from healthy human breast tissue were transduced with lentiviruses expressing hTERT and SV40Large T. These cells (herein termed TLM-HMECs) are immortal but do not proliferate in the absence of extracellular matrix (ECM) (Westbrook et al., 2005). For the screen we generated a shRNA library targeting all human kinases and phosphatases (six shRNAs/gene). TLM-HMECs were transduced with the shRNA library and assessed for cellular transformation by culturing in the absence of ECM. We isolated 530 anchorage-independent colonies from two independent screens and identified proviral shRNAs by sequencing. Genes identified in both replicate screens were considered candidate suppressors of transformation. Several genes were targeted by multiple independent shRNAs, including the well-documented tumor suppressors PTEN and LKB1 (Hemminki et al., 1998; Li et al., 1997). The top-scoring candidate from this genetic screen was the PTP PTPN12 (aka PTP-PEST) (Yang et al., 1993). Many TKs have been shown to be important drivers of human cancer, but PTPs that antagonize proto-oncogenic TKs have not received 704 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
equal attention. Notably, PTPN12 has not been previously implicated in tumor suppression. Three independent PTPN12 shRNAs exhibited robust cellular transformation in TLM-HMECs (Figure 1C), and the degree of depletion correlated with the severity of the phenotype (Figures 1B and 1C). In addition, restoring PTPN12 expression with an exogenous PTPN12 cDNA completely suppressed transformation (Figure 1D), ruling out RNAi off-target effects. Collectively, these data indicate that PTPN12 is a potent suppressor of transformation in mammary epithelial cells. Loss of PTPN12 Disrupts Proper 3D Acinar Formation of Mammary Epithelial Cells Proper control of cell proliferation, survival, and polarity in the mammary epithelium is critical for normal mammary gland function, and dysregulation of these processes is considered to be a driver in breast cancer initiation and progression (Bissell et al., 2002). To determine whether PTPN12 regulates these processes during acini formation, we tested the effects of PTPN12 loss of function in 3D culture of MCF10A cells, a nontumorigenic mammary epithelial cell line. When cultured in semisolid ECM, these cells form a polarized acinar structure that resembles mammary acini in vivo (Petersen et al., 1992). MCF10A cells transduced with control- or PTPN12-shRNAs were analyzed for PTPN12 expression and formation of 3D acini (Figure 1E, and see Figure S1E available online). PTPN12 depletion significantly disrupted normal acini formation, with >85% forming aberrant structures (Figure 1E). Confocal microscopy revealed that PTPN12 loss of function significantly increased ectopic proliferation (Figures S1B and S1C) but did not lead to a compensatory increase in apoptosis (data not shown) like other oncogenic insults (Debnath et al., 2002), consistent with the significantly expanded cellularity and filled lumen in acini upon PTPN12 depletion. These observations suggest that PTPN12 is required to establish proper proliferative arrest during acinar formation. To determine whether PTPN12 is required to maintain proliferative arrest in preestablished acini, we utilized an inducible shRNA vector that encodes a shRNA and tRFP on the same inducible transcript (Figure S1D, top panel) (Meerbrey et al., 2011). Addition of doxycycline (dox) results in tRFP fluorescence and PTPN12 depletion within 72 hr (Figure S1E). MCF10A cells encoding an inducible control- or PTPN12-shRNA were seeded in 3D culture, and after acini formation (day 9), dox was added to the culture medium, and acini morphology was assessed at day 15. Addition of dox had no effect in cells expressing control shRNA, but cells expressing PTPN12 shRNA exhibited a significant increase in aberrant acini (Figure S1F). Thus, PTPN12 is required to establish and maintain proliferative arrest in mammary epithelial acini. PTPN12 Phosphatase Activity Is Required to Suppress Cellular Transformation To determine the role of the PTPN12 phosphatase activity in transformation, we mutated amino acid C231 to S, which is known to ablate phosphatase activity (Garton et al., 1996). A shRNA-resistant PTPN12-C231S cDNA was transduced into TLM-HMECs expressing a PTPN12 shRNA. Unlike wild-type
A
shRNA viruses targeting all kinases & phosphatases
B PTPN12 Load control
PTPN12 expression (% control)
PCR / sequence proviral shRNA
Culture cells in absence of extracellular matrix
Control shRNA
PTPN12 shRNA
75 50 25 0
Human Mammary Epithelial Cells (TLM-HMECs)
C
100
shRNA:
D
AnchorageIndependent colonies
AnchorageIndependent colonies
Transformation Assay
Transformation Assay 300 200
400 300 200 100
0 PTPN12 cDNA:
100 0
-
-
+
shRNA:
shRNA:
Control shRNA
PTPN12 shRNA
F
E
%PTPN12 100
7 107
100
*
Load control P<2.8x10-8
*
50
0
AnchorageIndep. colonies
% Aberrant Acini
PTPN12
150 100 50 0
PTPN12-shRNA:
shRNA: PTPN12-cDNA:
-
+ + - C231S
Figure 1. A Genetic Screen for Tumor Suppressors Identifies PTPN12 (A) Schematic of genetic screen for suppressors of HMEC transformation. A pool of retroviral shRNAs was transduced into TLM-HMECs in duplicate, and assessed for anchorage-independent proliferation. shRNAs were PCR amplified and sequenced from macroscopic colonies. (B) Depletion of PTPN12. PTPN12 protein expression in TLM-HMECs transduced with vectors expressing the indicated shRNAs with quantification below. (C) PTPN12 loss of function transforms TLM-HMECs. Anchorage-independent proliferation in TLM-HMECs transduced with the indicated shRNAs. (D) Restoring PTPN12 expression suppresses transformation by PTPN12 shRNA. Anchorage-independent proliferation in TLM-HMECs transduced with control or PTPN12-shRNA in combination with PTPN12 cDNA as indicated. (E) PTPN12 regulates acinar morphogenesis. MCF10A cells expressing the indicated shRNAs were analyzed for 3D acinar morphogenesis in vitro (day 15 after seeding) and quantified for the number of aberrant mammary acini. (F) The enzymatic activity of PTPN12 is required for transformation suppression. TLM-HMECs expressing a PTPN12 shRNA were transduced with lentivirusencoding control or shRNA-resistant PTPN12-C231S mutant cDNA and assessed for PTPN12 expression by western (top) and anchorage-independent proliferation (bottom). Error bars represent standard error.
Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 705
B
A
HMECs 20
PTPN12
YFP
C-YFP
N-YFP
D
RTK-X
PTPN12
RTK-X
E
% YFP+ cells
*
15 10
*
5 0
Fold increase 2.3x
PTPN12 HER2-pY1248
p=0.04
% YFP+ cellls
C
F
Fold decrease 2.0x
PTPN12 HER2-pY1248
HER2
0.7x
HER2
1.1x
EGFR-pY1148
2.4x
EGFR-pY1148
4.4x
EGFR-pY992
1.1x
EGFR-pY992
1.2x
15
EGFR-pY845
1.1x
EGFR-pY845
1.3x
0
EGFR-pY1045
1.5x
EGFR-pY1045
1.2x
EGFR
1.0x
EGFR
1.1x
45 30
WT
C231S
Akt pS473 Akt-pS473
1.2x
Akt-pS473 Akt pS473
1.1x
Akt
1.2x
Akt
1.0x
S6K-pT421/S424
1.5x
S6K-pT421/S424
0.6x
MAPK-pT202/Y204
2.8x
MAPK-pT202/Y204
2.8x
MAPK
0.9x
MAPK
1.2x
RSK-pT380
2.1x
RSK pT380 RSK-pT380
1.8x
RSK
1.0x
RSK
0.9x
S6-S235/S236
5.2x
S6-S235/S236
6.5x
S6
0.8x
S6
1.2x
Load control (dox) PTPN12-shRNA:
706 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
Load control
- +
(dox) PTPN12-cDNA:
-
+
PTPN12 (Figure 1D), the C231S mutant had no effect on HMEC transformation (Figure 1F), suggesting that the tyrosine-phosphatase activity of PTPN12 is required for suppressing transformation. PTPN12 Regulates an EGFR/HER2-Centered RTK Network in HMECs To elucidate the role of the phosphatase activity of PTPN12 in suppressing transformation, it is critical to identify the phosphorylation events altered in response to PTPN12 loss. Thus, we used a quantitative proteome-wide method that combines anti-phosphotyrosine peptide immunoprecipitation, differential peptide labeling, and LC-MS/MS-based phosphopeptide identification and quantitation (Hsu et al., 2003) to search for phosphotyrosine-peptides whose abundance increases when PTPN12 is depleted. TLM-HMECs engineered with an inducible PTPN12shRNA were profiled with and without PTPN12 depletion. We identified 99 phosphotyrosine peptides corresponding to 69 proteins whose tyrosine phosphorylation increased by greater than 1.5-fold in the absence of PTPN12 (see Table S1). Many of these proteins have been previously described to interact with PTPN12, including the known substrate p130CAS (or BCAR1). However, whereas p130CAS was highly phosphorylated, depletion experiments indicate it plays only a minor role in cellular transformation due to PTPN12 loss (Figure S2), suggesting that other signaling proteins may play more important roles in the tumor-suppressive function of PTPN12. Analysis of these PTPN12-regulated proteins using Ingenuity and the Human Protein Reference Database (HPRD) resources revealed two highly connected protein-protein interaction networks. The number of interactions within this PTPN12-regulated network was highly enriched (p < 0.001, Monte Carlo procedure), consisting of 46% of all PTPN12-regulated proteins identified (Figures 2A and 2B). The first network consists of proteins and signaling pathways known to govern proliferation and survival in human cancer, with the RTK EGFR being a central component of both the literature-based Ingenuity network and the protein interaction-based HPRD network (p < 0.001, Monte Carlo procedure) (Figures 2A and 2B). The second network is comprised of proteins controlling the actin cytoskeleton, consistent with the role of PTPN12 in regulating cell motility and possibly metastasis (Angers-Loustau et al., 1999) (Figures S2A and S2B). Because EGFR and its related receptors play critical roles in breast cancer initiation and progression (Brandt et al., 2000; Muller et al., 1988; Slamon et al., 1989), we focused our
attention on the potential regulatory interaction between PTPN12 and the EGFR family of receptor TKs. EGFR is one of four RTKs in the HER family, with HER2 playing the most prominent role in human breast cancer. These RTKs are known to promote cell survival and proliferation and signal via autophosphorylation and recruitment of additional substrates through recognition of these autophosphorylation sites (Yarden and Sliwkowski, 2001). In principle, tyrosine phosphatases can counter the activity of these RTKs by dephosphorylation of RTK substrates or the RTKs themselves. To determine if PTPN12 interacts with EGFR and other RTKs in HMECs, we employed a bimolecular fluorescence complementation (BiFC) system (Giepmans et al., 2006; Kerppola, 2006). Each of the HER family RTKs (EGFR, HER2, HER3, and HER4) was fused on its C termini with the C-terminal half of YFP. These fusion cDNAs were transduced into TLM-HMECs expressing PTPN12 fused to the N terminus of YFP. If an interaction occurs between PTPN12 and the candidate RTK, this enables folding of the N- and C-terminal fragments of YFP to produce a fluorescent YFP protein (Figure 2C, left panel). EGFR and HER2 exhibited strong interaction with PTPN12, as determined by cellular YFP fluorescence (Figure 2C, right panel). In addition, HER2 interaction was enhanced with the substrate-trapping C231S mutant of PTPN12 (Figure 2D). These data suggest that PTPN12 may directly interact with and inhibit EGFR/HER2 signaling to suppress transformation. PTPN12 Inhibits an EGFR/HER2-MAPK Signaling Axis to Suppress Cellular Transformation HER2 and EGFR signal via homo- and heterodimerization and subsequent phosphorylation of their C-terminal tails on sites that serve as hubs for recruitment and activation of signaling complexes (Yarden and Sliwkowski, 2001). To determine whether PTPN12 controls EGFR and HER2, we assessed the phosphorylation status of EGFR and HER2 in cells expressing an inducible PTPN12 shRNA. Depletion of PTPN12 led to an increase in HER2 (Y1248) and EGFR (Y1148) phosphorylation. Consistent with our phosphoproteomic data, the Y1148 residue of EGFR showed the strongest differential phosphorylation (>2-fold) in response to PTPN12 depletion (Figure 2E). Reciprocally, inducible expression of a PTPN12 cDNA in TLM-HMECs decreased HER2-pY1248 and EGFR-pY1148, but not other phosphotyrosine residues on EGFR (Figure 2F; data not shown). Thus, PTPN12 selectively regulates phosphorylation of a subset of tyrosine residues on EGFR and HER2.
Figure 2. PTPN12 Interacts with and Inhibits the HER2/EGFR Signaling Axis (A and B) Tyrosine phosphoproteins regulated by PTPN12. HMECs expressing an inducible PTPN12-shRNA were quantified in the presence and absence of PTPN12 for tyrosine-phosphorylated peptides using a quantitative proteomic approach (described in Experimental Procedures). Interactions between the 69 PTPN12-regulated phosphoproteins were analyzed via (A) Ingenuity and (B) the HPRD. (C) HER2 and EGFR RTKs interact with PTPN12 in HMECs. The left panel shows the experimental design of the BiFC system. PTPN12 was fused with the N terminus of YFP, and RTKs were fused with the C terminus of YFP. HMECs were transduced with retroviruses expressing PTPN12-N-YFP and individual RTKC-YFP cDNAs. In the right panel the interaction between PTPN12 and HER family RTKs was assessed by cellular fluorescence. Asterisk indicates p < 0.01. (D) Substrate-trapping PTPN12 mutant displays increased interaction with HER2. Breast cancer cells expressing PTPN12-WT-N-YFP or mutant PTPN12-C231SN-YFP in combination with HER2-C-YFP were analyzed for cellular fluorescence. (E) PTPN12 loss of function elicits hyperactivation of HER2, EGFR, and a MAPK-signaling cascade. HMECs engineered with an inducible PTPN12-shRNA were cultured ± dox for 3 days, starved of growth factors, and analyzed for levels of the indicated total and phosphorylated proteins by western. (F) PTPN12 suppresses HER2, EGFR, and MAPK signaling. HMECs engineered with an inducible PTPN12-cDNA were cultured and analyzed as in (E). Error bars represent standard error.
Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 707
We next evaluated the effects of PTPN12 suppression on known EGFR/HER2 effector pathways (RAS/MAPK and PI3K/ AKT). PTPN12 depletion led to hyperactivation of ERK/RSK/S6 signaling but had no effect on phosphorylation of PI3K-effectors AKT and S6K1 (Figure 2E). Likewise, ectopic expression of PTPN12 decreased the phosphorylation of ERK/RSK/S6 but not PI3K signaling (Figure 2F). This is consistent with previous observations that pY1148 of EGFR serves as a binding site for the adaptor protein SHC and mediates activation of RASMAPK signaling (Batzer et al., 1995; Songyang et al., 1995). Furthermore, SHC is phosphorylated upon EGFR recruitment, and we found SHC1 phosphorylation on Y317 to be regulated by PTPN12 (Table S1). Taken together, these results indicate that loss of PTPN12 function leads to hyperphosphorylation of HER2/EGFR and activation of downstream RAS/MAPK signaling in HMECs. To determine whether HER2/EGFR is required for transformation, TLM-HMECs expressing a dox-inducible PTPN12-shRNA were depleted of EGFR or HER2 (Figure 3A). Transformation by the PTPN12 shRNA was significantly impaired upon depletion of EGFR or HER2 (Figure 3B). Depletion of the adaptor protein SHC also reduced cellular transformation in response to a PTPN12 shRNA (Figures 3C and 3D), consistent with the hypothesis that PTPN12 suppresses transformation by inhibiting a HER2/EGFR/SHC/MAPK signaling axis. Furthermore, pharmacologic inhibition of HER2 and EGFR or MEK strongly suppressed transformation in cells depleted of PTPN12 (Figures 3E and 3F, respectively). Collectively, these data demonstrate that PTPN12 suppresses cellular transformation in HMECs by antagonizing HER2/EGFR phosphorylation and downstream MAPK signaling. PTPN12 Is Inactivated by Mutation in Human TNBC The role of PTPN12 in transformation and control of protooncogenic pathways led us to ask whether PTPN12 is inactivated in breast cancer. Initially, we examined whether the PTPN12 locus is deleted frequently in breast cancer by analyzing a publicly available data set of 243 human primary breast tumors and tumor-derived cell lines for which genomic copy number has been assessed (Beroukhim et al., 2010). Indeed, 22.6% of breast cancers exhibit evidence of deletion (one homozygous deletion of 15.0 Mb) at the PTPN12 locus, though the deletions exhibit a median size of 22.9 Mb, suggesting that multiple driver mutations may exist in this region. PTPN12 deletion in lung cancer was also frequent (13.8%) and typically encompassed large chromosomal regions (median deletion size of 22.2 Mb; two homozygous deletions of 1.3 and 0.7 Mb). An analysis of focal deletions (<2.0 Mb) encompassing PTPN12 in these tumors revealed a minimum common region (MCR) of only 0.61 Mb that spans five genes (Figure S3). These data suggest that PTPN12 is inactivated, in part, via deletion in a wide range of cancers and support the hypothesis that PTPN12 is a frequently inactivated tumor suppressor. Because PTPN12 depletion leads to hyperactivation of HER2/ EGFR, PTPN12 inactivation may occur more frequently in nonHER2-amplified human breast cancers. Therefore, to search for potential tumorigenic sequence variants, we sequenced the 708 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
coding exons of PTPN12 in 83 TNBCs and cell lines (75 primary tumors and eight cell lines). Primary sequence analysis revealed nonsynonymous sequence alterations in PTPN12 in 4.8% of TNBCs (three primary tumors; one tumor-derived cell line) (Figure 4A). Three variants were heterozygous, one was homozygous, and none was present in reference genomes or singlenucleotide polymorphism (SNP) databases. These amino acid changes (shown in Figure 4B) occurred in the highly conserved catalytic cleft of PTPN12 (H230Y) and within close proximity to a known protein-interaction domain (E690G and W699G) (Pao et al., 2007). In contrast we found no evidence of sequence alterations in 202 primary breast cancers from the other predominant breast cancer subtypes (ER+ and HER2-amplified), suggesting that PTPN12 may be inactivated more frequently in the TNBC subtype (p < 0.001, Fisher’s exact test). To determine whether these sequence variants affected PTPN12 function, wild-type (WT) or mutant PTPN12 cDNAs lacking the 30 UTR were introduced into TLM-HMECs expressing PTPN12 shRNA targeting the PTPN12 30 UTR. All cDNAs restored PTPN12 protein to endogenous levels (Figures 4C and 4D, top panels). Similar to WT PTPN12, the R30Q mutant suppressed PTPN12-shRNA induced transformation. In contrast, the three remaining tumor-derived mutants (H230Y, E690G, and W699G) did not suppress transformation. In fact, TLM-HMECs expressing these mutants formed 25%–50% more colonies than cells with PTPN12 shRNA alone (Figures 4C and 4D, bottom panels), suggesting that these loss-of-function PTPN12 mutations are potentially dominant negative. We further examined whether the defective H230Y variant perturbed PTPN12 function in an independent system (MCF10A acinar formation). WT and H230Y mutant PTPN12 cDNAs were transduced into MCF10A cells expressing control- or PTPN12shRNAs and were expressed at comparable levels (Figure S4). In contrast to WT PTPN12, the H230Y mutant did not suppress aberrant acini formation in the presence of PTPN12-shRNA (Figures 4F and 4G). In addition the H230Y mutant elicited aberrant acini in the presence of endogenous PTPN12, again suggesting dominant-negative activity. Although we were unable to determine if these defective variants were somatic tumor mutations due to the lack of patient-matched nontumor DNA, our functional data indicate that these defective PTPN12 variants are likely causal in TNBCs. During sequencing of the PTPN12 locus, we observed a SNP occurring more frequently in tumors (7.3%, n = 274) than in nondisease control patients (2.5%, n = 1142) (p = 0.004) (Easton et al., 2007). Notably, this SNP results in a threonine to alanine change in PTPN12 (T573A). We have previously shown that this threonine is phosphorylated in cells (Dephoure et al., 2008), suggesting that this residue may have a regulatory function for PTPN12. Based on these observations, we tested the hypothesis that the PTPN12-T573A allele may have reduced ability to suppress transformation. Indeed, a PTPN12-T573A cDNA did not completely suppress transformation in PTPN12depleted TLM-HMECs (Figure 4E) and conferred aberrant acinar morphogenesis in 3D culture (Figure 4H). These results suggest that the PTPN12 T573A SNP is a partial loss-of-function allele with dominant-negative properties. Intriguingly, our analysis of primary data from a genome-wide association study (GWAS) of
A
EGFR
HER2
Vinculin
Vinculin
shRNA:
shRNA:
AnchorageIndependent colonies
B
Transformation Assay +Dox
200
-Dox 150 100 50
0 shRNA: Control
D
HER2
Transformation Assay AnchorageIndependent colonies
C
EGFR
SHC
Vinculin shRNA:
+Dox -Dox
300 200 100 0
shRNA: Control
E
Shc
F Transformation Assay # of colonies
# of colonies
Transformation Assay
200
100
0 PTPN12-shRNA: Lapatinib:
-
+ -
+ +
200
100
0 PTPN12-shRNA: U0126:
-
+ -
+ +
Figure 3. PTPN12 Suppresses Transformation by Inhibiting HER2/EGFR Signaling (A) EGFR and HER2 depletion in TLM-HMECs. TLM-HMECs expressing control, EGFR, or HER2-targeting shRNAs were analyzed by western blotting for EGFR, HER2, and vinculin (loading control) as indicated. (B) EGFR and HER2 RTKs are required for cellular transformation upon PTPN12 depletion. TLM-HMECs encoding an inducible PTPN12-shRNA were transduced with the indicated shRNAs and assessed for anchorage-independent proliferation ± dox. (C) SHC depletion in HMECs. TLM-HMECs expressing control or SHC-targeting shRNAs were analyzed by western for SHC and vinculin as indicated. (D) SHC is required for cellular transformation upon PTPN12 depletion. TLM-HMECs encoding an inducible PTPN12-shRNA were transduced with the indicated shRNAs and assessed for anchorage-independent proliferation ± dox. (E) HER2/EGFR inhibitors block PTPN12 depletion-induced transformation. TLM-HMECs expressing the indicated shRNAs were assessed for anchorageindependent growth ± a HER2/EGFR inhibitor (lapatinib). (F) Transformation by PTPN12 inactivation requires MAPK signaling. TLM-HMECs expressing the indicated shRNAs were assessed for anchorage-independent proliferation ± a MEK inhibitor (U0126). Error bars represent standard error.
SNPs contributing to breast cancer (Easton et al., 2007) revealed that the homozygous PTPN12-573A genotype occurred more frequently in the germline of patients with breast cancer. Although this observation did not reach statistical significance (p = 0.2), the trend (combined with our functional studies) is consistent with the allele conferring enhanced susceptibility to breast cancer. This raises the possibility that a relatively frequent allele of PTPN12 may confer a predisposition to breast cancer,
a hypothesis that will need to be rigorously tested in further experiments. Loss of PTPN12 Expression Occurs More Frequently in TNBC Our results suggest that PTPN12 functions as a suppressor of malignant transformation and may be inactivated in TNBC. Given the deletions and sequence alterations in PTPN12, we wished Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 709
A
B Phos Domain Pro1
ER- / HER2-
4/83 (4.8%)
ER+ or HER2+
0/202 (0.0%)
% PTPN12 100 22 84 20 88 PTPN12
100 50 0
PTPN12-shRNA:
- + + + + - - + - + WT
Pro5
PTPN1 PTPN6 PTPRO
E
% PTPN12
100 9 105 214
PTPN12 Load control
200 100
0 PTPN12-shRNA:
- + + + + + PTPN12-cDNA: -
p=10-5 200 150 100 50
0 PTPN12-shRNA: PTPN12-cDNA:
- + + + -
H
*P<0.001
H230Y
G vector
Pro4
PTPN12
% PTPN12 100 20 117 92 77 101 PTPN12 Load control
F cDNA:
HCSAGCGR HCSAGCGR HCSAGCGR HCSAGCGR HCSAG I GR HCSAG I GR HCSAGVGR
Pro3
PTPN12-WT PTPN12-H230Y
Control shRNA
PTPN12 shRNA
*P<0.015 **P<0.001 100
**
0
**
*
50
PTPN12 shRNA:
30
% Aberrant Acini
Dox:
D
AnchorageIndep. colonies
AnchorageIndep. colonies
Load control
p<0.001
H. sapiens M. musculus X. tropicalis D. rerio H. sapiens H. sapiens H. sapiens
Pro2
AnchorageIndep. colonies
PTPN12-mutant
% Aberrant Acini
C
Tumor-type
- - - + + +
*
20
10
0 PTPN12 cDNA:
PTPN12 cDNA:
p<0.0001
HER2+ Invasive Cancer
ER- / HER2Invasive Cancer
% of tumors PTPN12-negative
J
I
100 PTPN12 status
0 HER2 status: ER status:
710 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
negative
50
positive
+ -
+
-
to test whether PTPN12 function may also be frequently compromised by loss of expression. Unfortunately, PTPN12 mRNA expression is high in stromal compartments, thus precluding the use of RNA profiling to evaluate PTPN12 levels in public breast cancer data sets. To circumvent this problem we developed a specific immunohistochemical assay for PTPN12 protein (Figure S5A) and evaluated expression of PTPN12 in an independent cohort of 185 breast cancers. PTPN12 protein was consistently expressed in normal breast tissue (Figure S5B). In contrast, PTPN12 was undetectable in 37% of invasive breast cancers (example images shown in Figure 4I and Figure S5C). Strikingly, loss of PTPN12 expression occurred most frequently in TNBC (60.4% of TNBCs exhibit no detectable PTPN12 protein) (Figure 4J). In contrast, HER2-amplified tumors only rarely exhibited loss of PTPN12 expression (9.1% of HER2amplified tumors). The near-mutual exclusivity of HER2 amplification and PTPN12 loss (p < 0.0001 by Fisher’s exact test) suggests functional redundancy between these two events in tumorigenesis, and is consistent with the model that PTPN12 and HER2/EGFR RTKs operate in the same genetic pathway. Collectively, these results indicate that PTPN12 is frequently inactivated by deletion, sequence variation, or loss of protein expression, and PTPN12 loss of function may be a major determinant in aggressive TNBC. PTPN12 Is Posttranscriptionally Regulated by a RESTmiR-124 Network in Human Breast Cancer Our observation that PTPN12 protein levels were frequently undetectable in primary human breast cancers led to us to examine the mechanism(s) by which PTPN12 expression is lost. By examining other suppressors of HMEC transformation, we observed that the tumor suppressor REST was also lost in primary breast cancers (example images in Figure 5A), and the expression of REST and PTPN12 was highly correlated (p < 0.0001; Figure 5B), suggesting that PTPN12 and REST may be
coordinately regulated. REST is a transcription factor that represses neuronal genes in non-neural tissues, and plays a prominent tumor suppressor role in epithelial tissues (Westbrook et al., 2005), though the mechanism by which REST suppresses tumorigenesis is poorly understood. Based on the coordinate expression of REST and PTPN12 in human breast cancers and their similar phenotypes, we tested whether REST functions in a genetic pathway with PTPN12 by regulating its expression. Consistent with this hypothesis, transgenic REST expression led to a substantial increase in endogenous PTPN12 protein level in HCC70 (Figure 5C) and other TNBC cell lines (data not shown). Based on the established role of REST as a transcriptional repressor, we hypothesized that REST positively regulates PTPN12 expression via an indirect mechanism, perhaps by repressing an inhibitor of PTPN12 protein levels. microRNAs are an emerging class of negative regulators, and we observed that the PTPN12 30 UTR has three evolutionarily conserved binding sites for miR-124 (Figure 5D), a microRNA suggested to play a role in neural development (Lim et al., 2005). Importantly, miR-124 is transcriptionally repressed by REST during organismal development, and REST inactivation leads to elevated miR-124 expression in cells and tissues (Conaco et al., 2006; Yoo et al., 2009). To determine whether miR-124 regulates PTPN12 protein levels, we ectopically expressed miR-124-3 in HMECs. Transgenic miR-124 expression led to a significant decrease in endogenous PTPN12 protein levels (Figure 5E), suggesting that miR-124 may in part mediate REST’s ability to inhibit PTPN12 protein expression. The ability of miR-124 to inhibit the PTPN12 tumor suppressor makes a strong prediction that miR-124 may function as an oncogene. Indeed, ectopic expression of miR-124 led to robust transformation of TLM-HMECs (Figure 5F), thus phenocopying PTPN12 loss of function. Consistent with the role of miR-124 as a putative human oncogene, we observed frequent and focal
Figure 4. PTPN12 Is Functionally Inactivated in Human TNBC via Multiple Mechanisms (A) PTPN12 mutations occur more frequently in human TNBC. Frequency of mutations observed in TNBCs (75 primary tumors and eight cell lines) and 202 primary breast cancers positive for ER and/or HER2. (B) Schematic of PTPN12 mutations in TNBCs. Red stars indicate altered amino acids. The sequence surrounding the catalytic cysteine residue is expanded. The preceding histidine, H230, is conserved among all tyrosine phosphatases. The phosphatase domain (Phos Domain) and Proline rich regions (Pro1-5) are shown. (C) H230Y mutant PTPN12 fails to suppress transformation. TLM-HMECs expressing the PTPN12-shRNA were engineered with the indicated dox-inducible cDNAs. Cells were assessed for anchorage-independent growth. (D) E690 and W699 PTPN12 mutants fail to suppress transformation. TLM-HMECs expressing the PTPN12-shRNA were transduced with lentiviral PTPN12 cDNAs (as indicated) and assessed for anchorage-independent growth. (E) The PTPN12-T573A SNP is a partial loss-of-function allele for suppressing transformation. TLM-HMECs were transduced with PTPN12-shRNA in combination with lentiviral cDNAs encoding PTPN12-WT (threonine at residue 573), or PTPN12-T573A (alanine at residue 573). Cells were measured for anchorage-independent growth. (F) PTPN12-H230Y mutation disrupts proper acinar formation. MCF10A cells expressing control or PTPN12 shRNA in combination with wild-type or H230Y mutant PTPN12 were analyzed for 3D acinar formation (day 15 after seeding). (G) PTPN12-H230Y mutation disrupts proper acinar formation. Quantification of aberrant mammary acini from (F). (H) The PTPN12-T573A SNP allele disrupts acinar formation. MCF10A cells transduced with lentiviral cDNAs encoding control, PTPN12-WT (threonine at residue 573), or PTPN12-T573A (alanine at residue 573) as indicated were analyzed for 3D acinar morphogenesis in vitro (day 15 after seeding). (I) Loss of PTPN12 expression occurs more frequently in human TNBC. Primary human breast cancers (n = 185) were analyzed by immunohistochemistry for PTPN12 expression. Representative panels exhibiting positive PTPN12 expression in HER2-amplified breast cancer (left panel) and lack of expression in TNBC (right panel). (J) Loss of PTPN12 expression occurs predominantly in TNBC. Primary human breast cancers (n = 185) were analyzed by immunohistochemistry for PTPN12 expression. The number of samples showing no detectable PTPN12 expression (red area of bars) was quantified in HER2-positive, ER-positive, and triplenegative subtypes. Association between PTPN12 expression and breast cancer subtypes was tested by Fisher’s exact test. Error bars represent standard error.
Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 711
A
p < 0.00001
REST-negative invasive cancer
8 6 4 2 0
PTPN12 protein score
B
REST-positive invasive cancer
0
0.1 – 2.0
> 2.0
REST protein score
D
C
%PTPN12
100
262
miR-124
REST
PTPN12 REST
cDNA: Hsa: Eca: Bta: Mdo: Aca: Gga:
E
%PTPN12
1
PTPN12
Vinculin
100
42 Hsa: Eca: Bta: Mdo: Aca: Gga:
PTPN12 Vinculin miRNA:
Hsa: Eca: Bta: Mdo: Aca: g Gga:
3
2
1
133
162
GACCAUCUAC-CUGCCUUA---UACUA---CACUUAG GACCAUCUAC-CUGCCUUA---UGCUA---CACUU-G GACCAUCUAC-CUGCCUUA---UAUAACACUU-A GACCAUCUACUCUGCCUUAUUAUUGUA---CACUUAG GGCCAUCAAU-CUGCCUUA---U--UA---CACUUAG GGCCAUCUAC-CUGCCUUA---U--UA---CACUUAG
2
200
230
GUAUUAUUGCCUUAAUGUCUCUUAACCCUGU GUAUUAUUGCCUUAAUGUCUCUUAACCCUGU GUAUUAUUGCCUUAAUGUCUCUUAACCCUGU GUAUUAUUGCCUUAAUGUCUCUUAACCCUGU GUAUUAUUGCCUUAAUGUCUUUUAAACUGGU GUAUUAUUGCCUUAAUGUCUUUUAACCCUGU
640
3
670
AGUAAGUGCCUUGGAACAAUAUUGAAUUCUC AGUAAGUGCCUUGGAACAAUGUUGAAUUCUC AGUAAGUGCCUUGGAACAAUAUUGAAUUCUC AGUAAGUGCCUUGGAACAAUAUUGAAUUCUC CAUAAGUGCCUUUGA--AAUAUCAGAAU-UC AGUAAGUGCCUUGGAACAAUAUUGAAUUCUC
Transformation Assay # of colonies
F 200
100
0 miRNA:
Vector
miR-124
Figure 5. PTPN12 Is Regulated by the REST Tumor Suppressor via miR-124 (A) Loss of REST expression in human breast cancer. Primary human breast cancers (n = 185) were analyzed by immunohistochemistry for REST expression. Representative panels exhibiting negative and positive REST expression in invasive breast cancers. (B) Loss of REST expression strongly correlates with loss of PTPN12 expression. Primary human breast cancers (n = 185) were analyzed by immunohistochemistry for REST and PTPN12 expression. The level of PTPN12 (y axis) is plotted for tumors with absent, intermediate, or high REST levels (x axis). The median and mean PTPN12 values for each group are represented by a solid red line and plus symbol, respectively. The boxes represent the 25th to 75th percentiles. Association between PTPN12 and REST expression was tested by Fisher’s exact test. Error bars represent maximum and minimum observations within inner fences. (C) Ectopic REST expression increases PTPN12 protein levels in REST-deficient TNBC cells. HCC70 TNBC cells were transduced with control or REST cDNA, cultured for 9 days, and analyzed for expression of REST and PTPN12 by western. (D) Model for REST regulation of PTPN12 expression. REST regulates transcription of the neuronal microRNA miR-124. The PTPN12 30 UTR contains three conserved binding sites for miR-124. The sequences surrounding the three miR-124 binding sites are shown for human and five other vertebrate species. (E) Ectopic miR-124 expression decreases PTPN12 protein levels in HMECs. HMECs were transduced with control or miR-124-containing plasmid, cultured for 7 days, and analyzed for PTPN12 expression by western. (F) Ectopic miR-124 expression transforms TLM-HMECs. Cells from (E) were assessed for anchorage-independent proliferation. Error bars represent standard error.
712 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
amplifications of the miR-124-3 locus in human breast cancers (Figures S6A and S6C). In a cohort of 243 breast cancers (Beroukhim et al., 2010), 20.1% of tumors harbored amplifications in the miR-124-3 locus, defining a MCR of amplification of 0.1 Mb and encompassing only two NCBI-annotated genes (including miR-124-3). miR-124-3 was similarly amplified (12.1%) in human lung cancers (Figures S6B and S6C), suggesting that miR-124-3 and possibly other oncogenes exist on these amplicons. Collectively, these data suggest that miR-124 may be a novel human oncogene in epithelial cancers that transforms cells, in part, by suppressing PTPN12 function. PTPN12 Suppresses Growth and Metastasis of PTPN12-Deficient Breast Cancer Cells The observations described above implicate PTPN12 as a tumor suppressor in TNBC. We next explored whether breast cancer cells are functionally dependent on PTPN12 inactivation. We first determined levels of PTPN12 in a panel of established breast cancer cell lines. PTPN12 levels were similar in normal HMECs and in HER2-amplified breast cancer cells (Figure 6A). In contrast, several TNBC cell lines exhibited low PTPN12 protein (Figure 6A), consistent with our observation that a significant fraction of TNBC tumors have low or undetectable PTPN12 protein (Figure 4J). To determine whether TNBC cells with low PTPN12 are sensitive to restoring PTPN12 levels, we transduced TNBC cells with control- or PTPN12-expressing retrovirus (Figure S7). As shown in Figures 6B and 6C, ectopic expression of PTPN12 decreased colony formation by TNBC cells, suggesting that these cells are sensitive to reconstituting PTPN12 function. The ability of cancer cells to grow in microenvironments with altered or absent ECM is a hallmark of metastasis and is mimicked, in part, by in vitro culture in the absence of ECM support. Because our data indicate that PTPN12 is a potent suppressor of anchorage-independent proliferation (Figure 1), we tested the hypothesis that PTPN12 suppresses the metastatic propensity of TNBC cells. Notably, in our analysis of PTPN12 protein levels, we observed that the TNBC cell line MDA-MB231 exhibited high PTPN12, whereas a highly tumorigenic and metastatic subpopulation of MDA-MB231 (‘‘MDAMB231-LM2’’ cells; Minn et al. [2005]) exhibited 7-fold less PTPN12 (Figure 6D). We hypothesized that suppression of PTPN12 expression in this subpopulation of TNBC cells contributes to their aggressive tumorigenic and metastatic behavior. To test this hypothesis we engineered MDA-MB231-LM2 cells with a dox-inducible PTPN12 cDNA (Meerbrey et al., 2011) (termed ‘‘LM2-IP’’ cells) and tested whether PTPN12 restoration suppresses the tumorigenic and metastatic potential of LM2-IP cells. Addition of dox to LM2-IP cells resulted in PTPN12 protein levels comparable to parental MDA-MB231 cells (Figure 6D). To assay their tumorigenic potential, LM2-IP or parental MDAMB231 cells were transplanted orthotopically into the mammary gland in the presence or absence of dox. As shown in Figure 6E, LM2-IP cells formed tumors rapidly. However, dox-induced PTPN12 expression significantly reduced the tumorigenicity of LM2-IP cells (Figure 6E; p < 0.0001) to the levels of parental MDA-MB231 cells. Lungs collected from these animals revealed significantly fewer metastases than dox-free animals (data not shown). However, interpretation of these results was partially
confounded by the significantly larger primary tumor burden of dox-free animals. To circumvent the issue of tumor burden and directly assess the effects of PTPN12 on lung metastatic colonization and growth, we utilized an experimental model of metastasis (tail vein injection) that measures colonization and expansion in the lung. We injected LM2-IP cells into the tail vein, maintained animals in the presence or absence of dox, and monitored lung metastatic growth using luciferase luminescence-based imaging (examples shown in Figure 6F, top panels). LM2-IP cells injected into dox-free animals showed rapid expansion in the lung (Figure 6F, bottom panel), consistent with the previously reported behavior of MDA-MD231-LM2 cells (Minn et al., 2005). In contrast, LM2-IP cells exhibited a significantly reduced rate of expansion in dox-administered animals and showed no increase in lung metastatic growth after day 21. To confirm the luminescence-based readout, lungs were extracted from doxpositive and dox-negative animals at the experimental endpoint (day 35) and assessed for lung metastases. As shown in Figure 6G, dox-free animals showed significantly higher lung metastatic burden relative to dox-administered animals. Taken together, these data indicate that restoring PTPN12 function constrains the tumorigenic and metastatic behavior of aggressive TNBC cells. PTPN12 Suppresses Proliferation and Tumorigenicity of TNBCs by Inhibiting Multiple RTKs Together, our observations strongly suggest that restoring PTPN12 function impairs the tumorigenesis and proliferation of PTPN12-deficient TNBCs, suggesting that these cancers are dependent on the TK signaling constrained by PTPN12. Consequently, we sought to identify TKs regulated by PTPN12 in TNBC cells. Based on our observations that PTPN12 physically associates with and inhibits EGFR and HER2 RTKs in HMECs, we tested whether PTPN12 interacts with and regulates HER2 in PTPN12-deficient TNBC cells. BiFC analysis revealed a strong interaction between PTPN12 and HER2 in TNBC cells (Figure 7A, left panel). In addition, restoring PTPN12 expression in these cells decreased HER2 activity, as measured by HER2 tyrosine phosphorylation (Figure 7B). However, inhibition of HER family RTKs (HER2 and EGFR) with a pharmacologic inhibitor did not significantly affect the proliferation of these TNBC cells (Figure 7C), indicating that inhibition of HER2 and EGFR activity does not phenocopy PTPN12 restoration. Consequently, we hypothesized that PTPN12 regulates other TKs that, alone or in combination with HER2, are required for proliferation of TNBC cells. To identify which additional TKs are regulated by PTPN12 in TNBC cells, we tested several candidate TKs from our proteomic analysis and previous reports (Cong et al., 2000; Markova et al., 2003) for interaction with PTPN12 in TNBC cells via BiFC. Notably, we observed a strong interaction between PTPN12 and PDGFR-b (Figure 7A, right panel), and ectopic PTPN12 expression reduced tyrosine phosphorylation of endogenous PDGFR-b (Figure 7B), suggesting that PTPN12 regulates this RTK in TNBCs. Notably, whereas inhibition of PDGFR-b alone had only a modest effect on TNBC proliferation, combined inhibition of HER2 family and PDGFR-b RTKs significantly impaired TNBC proliferation (Figure 7C). Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 713
A
F
B
C
D G
E
Figure 6. PTPN12 Suppresses Growth and Metastasis of TNBC Cells (A) PTPN12 expression is reduced in TNBC cell lines. PTPN12 protein levels were quantified by western in HMECs, TNBC cells, and HER2-amplified breast cancer cells as indicated. (B) Reconstituting PTPN12 expression suppresses proliferation in PTPN12-deficient breast cancer cells. TNBC cells expressing low endogenous PTPN12 were transduced with equivalent multiplicity of infection (moi) of retrovirus encoding eGFP (control) or PTPN12 cDNAs and analyzed for macroscopic colony formation in vitro. (C) Reconstituting PTPN12 expression suppresses proliferation in PTPN12-deficient breast cancer cells. Quantification of colony number from cells in (B). (D) PTPN12 expression is reduced in aggressive lung metastatic subpopulation of TNBC MDA-MB231 cells. PTPN12 protein expression was assessed in MDA-MB231 breast cancer cells (231-parent) and in MDA-MB231-LM2 subpopulation that exhibits enhanced primary and lung metastatic tumor growth. MDA-MB231-LM2 cells were engineered with an inducible PTPN12-cDNA (LM2-IP cells) that expresses similar PTPN12 levels as parental MDA-MB231 cells upon addition of dox. (E) Restoring PTPN12 expression suppresses primary tumor growth in aggressive TNBC cells. Cells from (D) were transplanted in the mouse mammary gland and monitored for primary tumor growth in the presence or absence of dox (n = 12 for each group). (F) Restoring PTPN12 expression suppresses lung metastatic growth in aggressive TNBC cells. Cells from (D) were tail vein injected and monitored for lung metastatic growth (via luminescence detection) in the presence or absence of dox (n = 6 for each group). Representative +dox and dox images and quantification are shown in upper and lower panels, respectively. (G) Restoring PTPN12 expression suppresses lung metastatic growth in aggressive TNBC cells. Cells from (D) were tail vein injected (as in F) in the presence or absence of dox (n = 7) and analyzed for lung metastatic lesions via standard H&E. Error bars represent standard error.
714 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
A
Figure 7. PTPN12 Inhibits Proliferation and Survival of TNBCs by Inhibiting Multiple RTKs
20
% YFP+ cells
% YFP+ cells
B
10
0
30
PTPN12 HER2-pY1248
15
PDGFRβ-pY1021 0
Vinculin (load control)
Vector PDGFRβ
Vector HER2
-
PTPN12 DNA PTPN12-cDNA:
C
8 4
+
-
Lapatinib:
E
Tumor Volume (mm3)
Cell number (x1 10-3)
p=10-8 p
0 Sunitinib:
+ -
Lapatinib+Sunitinib
400
0 5
12
19
26
33
Days post-injection
p=0.003 800
400
-
40
+
+ -
+ +
F Percent E Event-free Survival
Lapatinib Sunitinib
800
1200
Lapatinib:
Tumor growth in mammary gland
1200
p=0.001
0 Sunitinib:
+ +
Vehicle
Tumor Volume V (mm3)
Tumor growth in mammary gland
D p=10-10 12
+
100 80 60
p=0.002 p
40 20 0 10
20
30
40
Days post-injection
50
(A) HER2 and PDGFR-b RTKs interact with PTPN12 in TNBC cells. HCC1937 cells expressing PTPN12-N-YFP and individual RTK-C-YFP cDNAs (as indicated) were analyzed for cellular fluorescence using flow cytometry. (B) Ectopic PTPN12 expression inhibits HER2 and PDGFR-b RTK signaling in TNBC cells. HCC1937 cells engineered with control or PTPN12-cDNA were assessed for PTPN12 expression and levels of phosphorylated HER2 and PDGFR-b by western. (C) Combined HER family and PDGFR inhibitors suppress proliferation of PTPN12-deficient TNBC cells. HCC1937 cells were cultured ± HER2/ EGFR inhibitor lapatinib (1 mM) ± PDGFR inhibitor sunitinib (5 mM) for 8 days. Cell numbers were determined by DAPI cell counting. (D and E) Combined HER family and PDGFR inhibitors suppress tumorigenicity of PTPN12deficient TNBC cells. MDA-MB231-LM2 cells were transplanted in the mouse mammary gland and monitored for primary tumor growth in the presence or absence of HER2/EGFR inhibitor (lapatinib) and PDGFR inhibitor (sunitinib) as indicated (n = 10 for each group). Tumor volumes on day 26 postinjection are shown in (D). Tumor growth curves are shown in (E). (F) Combined HER family and PDGFR inhibitors extend event-free survival of animals harboring PTPN12-deficient TNBC tumors. Animals transplanted with MDA-MB231-LM2 cells (as above) were treated with the indicated inhibitor and monitored for tumor volume. Events are denoted as animals with tumors greater than 1000 mm3. Error bars represent standard error.
Vehicle Lapatinib Sunitinib Lapatinib+Sunitinib
To determine whether PTPN12-deficient TNBCs are combinatorially dependent on the HER family and additional RTKs, such as PDGFR-b in vivo, we tested the effects of the pharmacologic inhibitors lapatinib and sunitinib alone or in combination on the tumorigenicity of aggressive PTPN12-deficient TNBCs. Lapatinib is a dual inhibitor of HER2 and EGFR, and sunitinib is an inhibitor of PDGFR-b and other TKs. MDA-MB231-LM2 cells were transplanted orthotopically, and mice were treated for 40 days with the indicated agents. Both agents were well tolerated by all animals at the administered doses (alone and in combination). Treatment with sunitinib alone resulted in a decrease in tumor growth rate (Figures 7D and 7E). Strikingly, whereas lapatinib alone did not affect tumor growth, lapatinib treatment significantly increased the efficacy of sunitinib in reducing tumor growth rate and burden (p = 0.003; Figures 7D and 7E). In addition, combination therapy resulted in a significant extension in event-free animal survival (p = 0.002; Figure 7F). These data suggest that PTPN12-deficient TNBC tumors may be combinatorially dependent on HER2 family and other RTKs.
Although our interaction and signaling studies suggest that PDGFR-b is one of these additional RTKs, and this is supported by the efficacy of sunitinib, given the broad specificity of sunitinib, it is possible that RTKs in addition to PDGFR-b may be involved. Thus, additional genetic and pharmacologic studies are necessary to determine the precise identities of the relevant RTKs driving tumorigenesis in tumors lacking PTPN12. Collectively, these studies identify PTPN12 as a significant tumor suppressor and illustrate a dependency on PTPN12-regulated TK signaling in TNBC, suggesting that these tumors may be successfully treated with the appropriate combination of TK inhibitors. DISCUSSION The Tyrosine Phosphatase PTPN12 Is a Tumor Suppressor in Human Breast Cancer In this study, we demonstrate that the tyrosine phosphatase PTPN12 is a potent tumor suppressor in human breast cancer. Loss of PTPN12 phosphatase activity leads to aberrant acinar Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 715
morphogenesis and cellular transformation in mammary epithelial cells. PTPN12 is frequently compromised in breast cancer by deletion, inactivating sequence variants, or loss of expression. Importantly, in breast cancer cells exhibiting PTPN12 deficiency, restoring PTPN12 expression to levels observed in normal mammary epithelial cells suppresses proliferation, tumorigenesis, and metastasis. Together, these observations strongly support the conclusion that PTPN12 is a suppressor of human breast cancer.
a significant subset of HER2-negative tumors may harbor aberrant HER2/EGFR signaling. Notably, whereas inhibitors targeting HER2 or EGFR individually have not been effective in HER2negative breast cancers, dual EGFR/HER2 inhibitors have not been rigorously tested in HER2-negative cancers such as TNBC. Further studies will be necessary to determine whether dual EGFR/HER2 inhibitors in combination with other TK inhibitors are effective in HER2-negative cancers and whether PTPN12 represents a biomarker for sensitivity to such agents.
A Network Governing Cellular Transformation and Tumor Suppression Our study shows that PTPN12 functions in concert with a collection of other oncogenes and tumor suppressors in a serial inhibitory network culminating in the regulation of proto-oncogenic RTKs. We previously discovered that the oncogenic F box protein b-TRCP acts to negatively regulate the REST tumor suppressor (Westbrook et al., 2008). Here, we find that REST positively regulates PTPN12 levels, in part, by negatively regulating miR-124 that represses PTPN12 protein levels. Ectopic miR-124 expression can also transform HMECs, and miR-124 is focally and frequently amplified in epithelial cancers of the breast and lung, and is likely to act as an oncogene in these contexts. Thus, we have discovered an extensive network of serial negative regulation with alternating oncogenes and tumor suppressors consisting of b-TRCP, REST, miR-124, and PTPN12, which inhibits proto-oncogenic RTKs such as EGFRHER2 to control cell proliferation, survival, and tumorigenesis.
A New Rationale for Combinatorially Targeting TK Signaling in TNBC and Other Cancers TKs have been shown to be critical pathogenetic drivers in some cancers, and the identification of oncogenic mutations and amplifications in TKs have provided important biomarkers for selecting cancers that are dependent on TK signaling and amenable to targeted therapies (Druker, 2004; Slamon et al., 1989). In contrast, TNBC is a particularly aggressive subtype of breast cancer for which no single, dominantly acting TK has been shown to drive the disease. Our discovery that PTPN12 is a tumor suppressor frequently inactivated in TNBC that acts as a negative regulator of HER2/EGFR and other TKs, such as PDGFR-b and ABL (Markova et al., 2003), raises the possibility that inhibitors of these proto-oncogenic TKs may be therapeutic in TNBC when used in the appropriate combination. In support of this prediction, our data indicate that lapatinib combined with another TK inhibitor (sunitinib) significantly reduces the proliferation and tumorigenicity of TNBCs. Given the prevalence of PTPN12 inactivation in TNBC and the addiction of these cancers to PTPN12 dysfunction, it will be important to define the full spectrum of TKs that PTPN12 regulates in TNBC and other cancers in order to rationally combine agents targeting these kinases. In summary, these studies establish an important role for tyrosine phosphatases in antagonizing tumorigenesis. Our observations raise the important prediction that many malignancies considered to be non-TK driven because of the absence of a dominant TK mutation may indeed be dependent on TK signaling. It is likely that in different cell types, different PTPs may play roles similar to PTPN12 in suppressing tumorigenesis, possibly by antagonizing different combinations of TKs. These studies provide a rationale for identifying and employing TK inhibitors to treat cancers not previously thought to be driven by TKs due to an absence of biomarkers for TK dependency. As noted above for TNBC, it is likely that these tumors would not be fully dependent on a single TK but, rather, a combination of TKs within the cell. Thus, a combinatorial inhibitor approach would be required to treat these cancers, most of which are likely to be untreatable today. Our studies warrant a closer examination of the status of tyrosine phosphorylation in what has been previously considered non-TK driven disease and the identification of TKs whose basal activity might combinatorially contribute to cancers harboring defects in PTPs.
The Role of PTPN12 and HER Family Receptors in ‘‘HER2-Negative’’ Breast Cancer Approximately 20% of all breast cancers exhibit amplification and overexpression of HER2, and many of these malignancies are sensitive to HER2 inhibitors (Di Cosimo and Baselga, 2010). However, there is significant controversy as to whether HER2 and other HER family RTKs play a role in breast cancers that do not exhibit amplification/overexpression of the HER2 locus (termed ‘‘HER2-negative’’). Our data support a model in which PTPN12 inactivation leads to HER2/EGFR hyperactivity and cellular transformation in HER2-negative breast cancers. This is supported by several observations. First, PTPN12 interacts with the HER2 and EGFR receptors in HMECs, and loss of PTPN12 function leads to hyperactivation of HER-receptor signaling in these cells. Conversely, transgenic expression of PTPN12 suppresses HER2 signaling in HMECs and PTPN12deficient breast cancer cells. Second, PTPN12 inactivation leads to mammary epithelial cell transformation that is at least partially dependent on HER2 and EGFR. Third, HER2 activity is diminished in PTPN12-deficient breast cancers when PTPN12 expression is restored. Fourth, inhibitors of HER2 and EGFR, when combined with other TK inhibitors, reduce proliferation and tumorigenicity of HER2-negative breast cancers lacking PTPN12. Finally, we observe frequent inactivation of PTPN12 specifically in HER2-negative breast cancers (96% of PTPN12deficient breast cancers were HER2 negative). This mutually exclusive relationship is consistent with PTPN12 and HER2 functioning in a common genetic pathway. Moreover, the frequent inactivation of PTPN12 in HER2-negative cancers suggests that 716 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
EXPERIMENTAL PROCEDURES Vectors The shRNA library targeting human kinases and phosphatases (six shRNAs/ gene) was synthesized using published methods (Cleary et al., 2004) and
cloned into a MSCV-based retroviral vector. Individual GIPZ lentiviral shRNAs were from the Hannon-Elledge shRNA collection (Open Biosystems). shRNAs targeting the 30 UTR region of PTPN12 were designed using the BiopredSI and RNAi Codex algorithms. For inducible RNAi or overexpression experiments, shRNAs or cDNAs were subcloned into a dox-inducible lentiviral expression system (Meerbrey et al., 2011).
Cell Culture HMECs expressing hTERT and SV40 LT (TLM-HMECs) (Westbrook et al., 2005) were cultured in mammary epithelial growth medium (MEGM, Lonza). Lung metastatic MDA-MB231-LM2 cells (Minn et al., 2005) were cultured in DMEM (GIBCO) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. HCC1954, HCC38, and HCC1937 cells were cultured in RPMI-1640 (ATCC) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. MCF10A cells (from J. Brugge, Harvard Medical School) were grown as previously described (Petersen et al., 1992). All cell lines were incubated at 37 C and 5% CO2. Stable cell lines expressing indicated shRNAs or cDNAs were generated by retroviral/lentiviral infection in the presence of 8 mg/ml polybrene, followed by selection with appropriate antibiotic-resistance markers.
Tumorigenicity and Experimental Metastasis Assays Orthotopic tumorigenicity assays were performed as previously described (Minn et al., 2005). In drug treatment experiments, animals were gavaged with lapatinib (100 mg/kg) and/or sunitinib (40 mg/kg) once daily and monitored for tumor growth. For event-free survival analysis, events were denoted as animals with tumors greater than 1000 mm3. Comparison between groups was performed using Wilcoxon analysis. For experimental metastasis assays, NOD/SCID female mice (NCI) aged-matched between 5 and 7 weeks were treated with PBS or dox at 2 mg/kg by intraperitoneal (IP) injection. The 2 3 105 cells were resuspended in 0.1 ml PBS and injected into the lateral tail vein. Lung metastatic progression was monitored and quantified using noninvasive bioluminescence as previously described (Minn et al., 2005). Linear model analyses (F test) were performed to test the effect of PTPN12 induction on tumorigenicity (tumor growth) or metastatic expansion (luminescence) in both assays.
SUPPLEMENTAL INFORMATION
REFERENCES Angers-Loustau, A., Cote, J.F., Charest, A., Dowbenko, D., Spencer, S., Lasky, L.A., and Tremblay, M.L. (1999). Protein tyrosine phosphatase-PEST regulates focal adhesion disassembly, migration, and cytokinesis in fibroblasts. J. Cell Biol. 144, 1019–1031. Bakalarski, C.E., Elias, J.E., Villen, J., Haas, W., Gerber, S.A., Everley, P.A., and Gygi, S.P. (2008). The impact of peptide abundance and dynamic range on stable-isotope-based quantitative proteomic analyses. J. Proteome Res. 7, 4756–4765. Batzer, A.G., Blaikie, P., Nelson, K., Schlessinger, J., and Margolis, B. (1995). The phosphotyrosine interaction domain of Shc binds an LXNPXY motif on the epidermal growth factor receptor. Mol. Cell. Biol. 15, 4403–4409. Beroukhim, R., Mermel, C.H., Porter, D., Wei, G., Raychaudhuri, S., Donovan, J., Barretina, J., Boehm, J.S., Dobson, J., Urashima, M., et al. (2010). The landscape of somatic copy-number alteration across human cancers. Nature 463, 899–905. Bissell, M.J., Radisky, D.C., Rizki, A., Weaver, V.M., and Petersen, O.W. (2002). The organizing principle: microenvironmental influences in the normal and malignant breast. Differentiation 70, 537–546. Brandt, R., Eisenbrandt, R., Leenders, F., Zschiesche, W., Binas, B., Juergensen, C., and Theuring, F. (2000). Mammary gland specific hEGF receptor transgene expression induces neoplasia and inhibits differentiation. Oncogene 19, 2129–2137. Cleary, M.A., Kilian, K., Wang, Y., Bradshaw, J., Cavet, G., Ge, W., Kulkarni, A., Paddison, P.J., Chang, K., Sheth, N., et al. (2004). Production of complex nucleic acid libraries using highly parallel in situ oligonucleotide synthesis. Nat. Methods 1, 241–248. Conaco, C., Otto, S., Han, J.J., and Mandel, G. (2006). Reciprocal actions of REST and a microRNA promote neuronal identity. Proc. Natl. Acad. Sci. USA 103, 2422–2427. Cong, F., Spencer, S., Cote, J.F., Wu, Y., Tremblay, M.L., Lasky, L.A., and Goff, S.P. (2000). Cytoskeletal protein PSTPIP1 directs the PEST-type protein tyrosine phosphatase to the c-Abl kinase to mediate Abl dephosphorylation. Mol. Cell 6, 1413–1423. Debnath, J., Muthuswamy, S.K., and Brugge, J.S. (2003). Morphogenesis and oncogenesis of MCF-10A mammary epithelial acini grown in three-dimensional basement membrane cultures. Methods 30, 256–268.
Supplemental Information includes Extended Experimental Procedures, seven figures, and one table and can be found with this article online at doi:10.1016/ j.cell.2011.02.003.
Debnath, J., Mills, K.R., Collins, N.L., Reginato, M.J., Muthuswamy, S.K., and Brugge, J.S. (2002). The role of apoptosis in creating and maintaining luminal space within normal and oncogene-expressing mammary acini. Cell 111, 29–40.
ACKNOWLEDGMENTS
Dephoure, N., Zhou, C., Villen, J., Beausoleil, S.A., Bakalarski, C.E., Elledge, S.J., and Gygi, S.P. (2008). A quantitative atlas of mitotic phosphorylation. Proc. Natl. Acad. Sci. USA 105, 10762–10767.
We thank J. Luo, J. Rosen, and A. Lee for suggestions and critical reading of the manuscript. We are grateful to A. Lee, J. LaBaer, O. Lee, Z. Songyang, and J. Qin for providing reagents, and A. Elia for help with phosphopeptide analysis. This work was supported by a Susan G. Komen for the Cure Career Catalyst Award (KG090355) to T.F.W., S.P.O.R.E developmental grant (P50 CA058183) to T.F.W., S.P.O.R.E. grant (P50 CA058183) to R.S. and C.K.O., grant HG3456 to S.P.G., and a U.S. Army Innovator Award (W81XWH0410197) to S.J.E. Research in the laboratory of M.B.-A. is supported by the Novartis Research Foundation, the European Research Council (ERC starting grant), the Swiss Cancer League, and the Krebsliga Beider Basel. S.J.E. is an Investigator with the Howard Hughes Medical Institute. T.F.W. is a scholar of The V Foundation and Mary Kay Ash Foundation for Cancer Research. Received: April 15, 2010 Revised: December 22, 2010 Accepted: February 1, 2011 Published: March 3, 2011
Di Cosimo, S., and Baselga, J. (2010). Management of breast cancer with targeted agents: importance of heterogeneity. Nat Rev Clin Oncol 7, 139–147. Druker, B.J. (2004). Imatinib as a paradigm of targeted therapies. Adv. Cancer Res. 91, 1–30. Easton, D.F., Pooley, K.A., Dunning, A.M., Pharoah, P.D., Thompson, D., Ballinger, D.G., Struewing, J.P., Morrison, J., Field, H., Luben, R., et al. (2007). Genome-wide association study identifies novel breast cancer susceptibility loci. Nature 447, 1087–1093. Garton, A.J., Flint, A.J., and Tonks, N.K. (1996). Identification of p130(cas) as a substrate for the cytosolic protein tyrosine phosphatase PTP-PEST. Mol. Cell. Biol. 16, 6408–6418. Garton, A.J., Burnham, M.R., Bouton, A.H., and Tonks, N.K. (1997). Association of PTP-PEST with the SH3 domain of p130cas; a novel mechanism of protein tyrosine phosphatase substrate recognition. Oncogene 15, 877–885. Giepmans, B.N., Adams, S.R., Ellisman, M.H., and Tsien, R.Y. (2006). The fluorescent toolbox for assessing protein location and function. Science 312, 217–224.
Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc. 717
Hemminki, A., Markie, D., Tomlinson, I., Avizienyte, E., Roth, S., Loukola, A., Bignell, G., Warren, W., Aminoff, M., Hoglund, P., et al. (1998). A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature 391, 184–187.
Muller, W.J., Sinn, E., Pattengale, P.K., Wallace, R., and Leder, P. (1988). Single-step induction of mammary adenocarcinoma in transgenic mice bearing the activated c-neu oncogene. Cell 54, 105–115.
Hsu, J.L., Huang, S.Y., Chow, N.H., and Chen, S.H. (2003). Stable-isotope dimethyl labeling for quantitative proteomics. Anal. Chem. 75, 6843–6852.
Osborne, C.K. (1998). Tamoxifen in the treatment of breast cancer. N. Engl. J. Med. 339, 1609–1618.
Hunter, T. (2009). Tyrosine phosphorylation: thirty years and counting. Curr. Opin. Cell Biol. 21, 140–146.
Pao, L.I., Badour, K., Siminovitch, K.A., and Neel, B.G. (2007). Nonreceptor protein-tyrosine phosphatases in immune cell signaling. Annu. Rev. Immunol. 25, 473–523.
Hurvitz, S.A., and Finn, R.S. (2009). What’s positive about ‘triple-negative’ breast cancer? Future Oncol. 5, 1015–1025. Jemal, A., Siegel, R., Ward, E., Hao, Y., Xu, J., Murray, T., and Thun, M.J. (2008). Cancer statistics, 2008. CA Cancer J. Clin. 58, 71–96. Kerppola, T.K. (2006). Visualization of molecular interactions by fluorescence complementation. Nat. Rev. Mol. Cell Biol. 7, 449–456. Li, J., Yen, C., Liaw, D., Podsypanina, K., Bose, S., Wang, S.I., Puc, J., Miliaresis, C., Rodgers, L., McCombie, R., et al. (1997). PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science 275, 1943–1947.
Petersen, O.W., Ronnov-Jessen, L., Howlett, A.R., and Bissell, M.J. (1992). Interaction with basement membrane serves to rapidly distinguish growth and differentiation pattern of normal and malignant human breast epithelial cells. Proc. Natl. Acad. Sci. USA 89, 9064–9068. Slamon, D.J., Godolphin, W., Jones, L.A., Holt, J.A., Wong, S.G., Keith, D.E., Levin, W.J., Stuart, S.G., Udove, J., Ullrich, A., et al. (1989). Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 244, 707–712. Songyang, Z., Margolis, B., Chaudhuri, M., Shoelson, S.E., and Cantley, L.C. (1995). The phosphotyrosine interaction domain of SHC recognizes tyrosinephosphorylated NPXY motif. J. Biol. Chem. 270, 14863–14866.
Lim, L.P., Lau, N.C., Garrett-Engele, P., Grimson, A., Schelter, J.M., Castle, J., Bartel, D.P., Linsley, P.S., and Johnson, J.M. (2005). Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs. Nature 433, 769–773.
Tonks, N.K. (2006). Protein tyrosine phosphatases: from genes, to function, to disease. Nat. Rev. Mol. Cell Biol. 7, 833–846.
Markova, B., Herrlich, P., Ronnstrand, L., and Bohmer, F.D. (2003). Identification of protein tyrosine phosphatases associating with the PDGF receptor. Biochemistry 42, 2691–2699.
Westbrook, T.F., Martin, E.S., Schlabach, M.R., Leng, Y., Liang, A.C., Feng, B., Zhao, J.J., Roberts, T.M., Mandel, G., Hannon, G.J., et al. (2005). A genetic screen for candidate tumor suppressors identifies REST. Cell 121, 837–848.
Matsuoka, S., Ballif, B.A., Smogorzewska, A., McDonald, E.R., 3rd, Hurov, K.E., Luo, J., Bakalarski, C.E., Zhao, Z., Solimini, N., Lerenthal, Y., et al. (2007). ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage. Science 316, 1160–1166.
Westbrook, T.F., Hu, G., Ang, X.L., Mulligan, P., Pavlova, N.N., Liang, A., Leng, Y., Maehr, R., Shi, Y., Harper, J.W., and Elledge, S.J. (2008). SCFbeta-TRCP controls oncogenic transformation and neural differentiation through REST degradation. Nature 452, 370–374.
Meerbrey, K.L., Hu, G., Kessler, J.D., Roarty, K., Li, M.Z., Fang, J.E., Herschkowitz, J.I., Burrows, A.E., Ciccia, A., Sun, T., et al. (2011). The pINDUCER lentiviral toolkit for inducible RNA interference in vitro and in vivo. Proc. Natl. Acad. Sci. USA. Published online February 9, 2011. 10.1073/pnas. 1019736108.
Yang, Q., Co, D., Sommercorn, J., and Tonks, N.K. (1993). Cloning and expression of PTP-PEST. A novel, human, nontransmembrane protein tyrosine phosphatase. J. Biol. Chem. 268, 6622–6628. Yarden, Y., and Sliwkowski, M.X. (2001). Untangling the ErbB signalling network. Nat. Rev. Mol. Cell Biol. 2, 127–137.
Minn, A.J., Gupta, G.P., Siegel, P.M., Bos, P.D., Shu, W., Giri, D.D., Viale, A., Olshen, A.B., Gerald, W.L., and Massague, J. (2005). Genes that mediate breast cancer metastasis to lung. Nature 436, 518–524.
Yoo, A.S., Staahl, B.T., Chen, L., and Crabtree, G.R. (2009). MicroRNA-mediated switching of chromatin-remodelling complexes in neural development. Nature 460, 642–646.
718 Cell 144, 703–718, March 4, 2011 ª2011 Elsevier Inc.
A Hierarchical Combination of Factors Shapes the Genome-wide Topography of Yeast Meiotic Recombination Initiation Jing Pan,1,9,10 Mariko Sasaki,1,5,9 Ryan Kniewel,1,5 Hajime Murakami,1 Hannah G. Blitzblau,6 Sam E. Tischfield,1,7 Xuan Zhu,1,5 Matthew J. Neale,1,8 Maria Jasin,2 Nicholas D. Socci,3 Andreas Hochwagen,6 and Scott Keeney1,4,* 1Molecular
Biology Program Biology Program 3Computational Biology Center 4Howard Hughes Medical Institute Memorial Sloan-Kettering Cancer Center, New York, NY 10065, USA 5Weill Graduate School of Medical Sciences of Cornell University, New York, NY 10065, USA 6Whitehead Institute for Biomedical Research, Cambridge, MA 02142, USA 7Tri-Institutional Training Program in Computational Biology and Medicine, Cornell University, New York, NY 10065, USA 8Genome Damage and Stability Centre, University of Sussex, Brighton BN1 9RQ, UK 9These authors contributed equally to this work 10Present address: Cell Biology Department, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.009 2Developmental
SUMMARY
The nonrandom distribution of meiotic recombination influences patterns of inheritance and genome evolution, but chromosomal features governing this distribution are poorly understood. Formation of the DNA double-strand breaks (DSBs) that initiate recombination results in the accumulation of Spo11 protein covalently bound to small DNA fragments. By sequencing these fragments, we uncover a genome-wide DSB map of unprecedented resolution and sensitivity. We use this map to explore how DSB distribution is influenced by large-scale chromosome structures, chromatin, transcription factors, and local sequence composition. Our analysis offers mechanistic insight into DSB formation and early processing steps, supporting the view that the recombination terrain is molded by combinatorial and hierarchical interaction of factors that work on widely different size scales. This map illuminates the occurrence of DSBs in repetitive DNA elements, repair of which can lead to chromosomal rearrangements. We also discuss implications for evolutionary dynamics of recombination hot spots. INTRODUCTION Most sexual species induce homologous recombination in meiosis via a developmentally programmed pathway that forms numerous DNA double-strand breaks (DSBs) (Keeney, 2007). Recombination helps homologous chromosomes pair and become physically connected by crossovers, which promote
accurate chromosome segregation at Meiosis I. Recombination also alters genome structure by disrupting linkage of sequence polymorphisms on the same DNA molecule (Kauppi et al., 2004). Thus, meiotic recombination is a powerful determinant of genome diversity and evolution. Recombination is more likely to occur in some genomic regions than others, largely because of nonrandom DSB distributions (Petes, 2001; Kauppi et al., 2004). DSBs in S. cerevisiae show many levels of spatial organization. There are large (tens of kb) DSB hot and cold domains, within which are short regions, called hot spots, where DSBs form preferentially. Important determinants of this organization include open chromatin structure, presence of certain histone modifications, and, at some loci, binding of sequence-specific transcription factors (TFs) (Petes, 2001; Lichten, 2008). However, detailed understanding is lacking of how these and other factors influence DSB locations. Meiotic DSBs are formed by the conserved topoisomeraserelated Spo11 protein via a reaction in which a tyrosine severs the DNA backbone and attaches covalently to the 50 end of the cleaved strand (Keeney, 2007) (Figure 1A). Two Spo11 molecules work in concert to cut both strands of a duplex. Endonucleolytic cleavage adjacent to the covalent proteinDNA complex liberates Spo11 bound to a short oligonucleotide (oligo) (Neale et al., 2005). In S. cerevisiae there are two major oligo subpopulations differing in length. The longer (mostly 21–37 nt) and shorter oligos (<12 nt) are equally abundant and may reflect asymmetry of DSB processing. Further resection of 50 DSB termini yields 30 -single stranded DNA (ssDNA) that is a substrate for strand exchange proteins. Prior studies of genome-wide DSB distributions used either covalent Spo11-DSB complexes that accumulate in rad50Slike mutants or ssDNA generated by DSB resection as microarray hybridization probes (e.g., Gerton et al., 2000; Blitzblau Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 719
Spo11
C Spo11 oligos (hpM/bp)
3'
nick nick
F 5'–3' resection & repair
bp 350 300 250 200 150
k
c mo
Sp
P 1I o1
100 75
40
Data set 2
30
Data set 1C
20
G
non-repetitive (98.2%)
Data set 1B
10 Data set 1A
0
repetitive (1.8%)
220
260
300 rDNA (0.15%)
Spo11 oligos ssDNA
15
others Telomeres (0.43%)
10 5
tRNA Ty/LTR (0.28%) (0.30%)
0 190
210
P
Expected size range
Spo11 oligo distribution:
230
E
*
spo11yf sae2
DSBs (% of DNA)
B
D
50
180 Spo11 oligos (hpM/bp) ssDNA hyb. signal / 2
3'
Spo11 oligos (hpM/bp)
A
15 10 5 0
ORFs: 42c
43c
206
44c
45c
46c 47c
48w
210 Position on Chr III (kb)
51w 52w
214
10
dmc1, r2 = 0.861 sae2, r2 = 0.873
1
0.1 102 103 104 Spo11 oligos (hpM)
Figure 1. Mapping DSBs by Sequencing Spo11 Oligos (A) Formation and processing of meiotic DSBs. (B) Amplification of adaptor-ligated Spo11 oligos and mock immunoprecipitation processed in parallel. (C) Reproducibility of Spo11 oligo maps. Oligo counts are in hpM. For the complete Spo11 oligo maps for each of the 16 chromosomes, see Figure S2. (D) Proportions of reads mapped to unique regions and repetitive elements. (E) Quantitative agreement of Spo11 oligos and DSBs. Oligo counts in 19 hot spots were compared with DSBs (mean ± SD, three cultures) assayed by Southern blot of DNA from dmc1 and sae2 mutants. (F) The Spo11 oligo map agrees with dmc1 ssDNA microarray analysis (Buhler et al., 2007) but has higher resolution. (G) Agreement of the Spo11 oligo map with direct DSB detection in sae2 genomic DNA (spo11yf = DSB-deficient mutant spo11-Y135F. P, parental band; asterisk, cross-hybridization). Oligo counts in (C), (F), and (G) were smoothed with a 201 bp sliding Hann window. See also Figure S1 and Table S1.
et al., 2007; Buhler et al., 2007). These studies provided considerable insight but had limited quantitative and spatial resolution due to microarray design, dynamic range of hybridization signal, and the large size of DSB-associated DNA used as probes. We overcame these limitations by using each Spo11 oligo as a tag that records precisely where a break was made. Sequencing these oligos allowed us to quantitatively map DSBs across the genome at nucleotide resolution with high sensitivity. This map elucidated chromosome features that govern DSB distributions, allowed us to test long-standing hypotheses concerning the influence of TFs, chromatin, and other factors, and uncovered mechanistic details of the formation and early nucleolytic processing of DSBs. RESULTS AND DISCUSSION A Nucleotide Resolution Map of Meiotic DSBs Spo11 oligos were purified from meiotic cultures, and adaptors were added (see Figure S1A available online). Because shorter oligos are difficult to map uniquely, longer ones were enriched 720 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
by size fractionation. PCR yielded products of the anticipated size that were absent in controls from mock precipitation of the meiotic extracts (Figure 1B). We deep sequenced three replicates from one culture and one from an independent culture, obtaining 2.19 million reads that were mapped to the genome of strain S288C and to a draft genome of SK1 (Liti et al., 2009), the source of Spo11 oligos (Table S1). More than 95% mapped to one or both genomes, mostly to unique sites. The maps agreed well: <0.8% of oligos mapped to different positions in the two strains (Table S1 and data not shown). The SK1 genome assembly is incomplete, so the S288C map was used for most analyses. Mapped reads matched sizes expected for longer oligos (Figure S1B). Replicates were highly reproducible (Pearson’s r = 0.95–0.99) (Figure 1C and Figure S1C), so data were pooled. Sequenced DNA was highly specific for bona fide Spo11 oligos. The rDNA cluster, 100–200 copies of a 9.1 kb repeat on Chr XII, is strongly repressed for meiotic recombination (Petes and Botstein, 1977). Only 0.15% of mappable reads were from rDNA (Figure 1D; other repeats are discussed below). Supposing
r = –0.89 p < 0.0001
0.5 0.4 0.3 0
0.5
1.0
1.5
Chrom. length (Mbp)
B
E
~20 kb
8
D8
6
6
4
4
2
2
0
50 100 150 Distance from telomere (kb)
200 Smoothing window
0.2
Figure 2. Large-Scale DSB Patterns
<5 kb
0 50 Distance from centromere (kb)
F
50 kb 0.12 r = –0.81 p = 0.0001
0.10 0.08 0.06
0
0.5
1.0
1.5
Chrom. length (Mbp)
0
cen
0.2
30 kb
0 0.3 15 kb 0 0.8 5 kb
GC content
Correlation with Spo11 oligos
Spo11 oligos (hpM/bp)
Mean Spo11 oligos (hpM/kb, log2)
C
Spo11 oligos (hpM/bp)
Crossover density (cM/kb)
A
0.4
0.2 Condensin 0
Smc6 Cohesin (Scc1)
-0.2 Rec8 -0.4 0
5
10
15
20
Window size (kb)
(A and B) Smaller chromosomes are hotter for crossing over (A, centimorgans [cM] per kb; data of Mancera et al., 2008) and DSBs (B). (C and D) Telomere-proximal and pericentric DSB suppression. Points are oligo densities in 500 bp segments averaged across all 32 chromosome arms. Dashed line shows genome average, magenta line indicates smoothed fit (Lowess), and shading illustrates DSB suppression zones. (E) Multiple levels of spatial organization of the DSB landscape. Oligo distributions on Chr V were smoothed with different-sized Hann windows. Dashed lines, genome average. (F) Correlation (Pearson’s r) between Spo11 oligo density and GC content or binding of chromosome structure proteins, analyzed in variable-sized windows. Rec8 data were from 2 hr in meiosis (Kugou et al., 2009); others were from vegetative cells (Lindroos et al., 2006; D’Ambrosio et al., 2008). See also Extended Experimental Procedures and Figure S3.
0 0
100
200
300
400
500
Position on Chr V (kb)
that none of the rDNA reads is a true Spo11 oligo, then the Spo11-independent background is 0.0011 hits per million mapped reads (hpM) per bp (assuming 150 rDNA repeats). This is likely an overestimate because meiotic DSBs probably do form in the rDNA. Even so, this value is 75-fold below genome average (0.083 hpM/bp) and is 146- to 6646-fold below oligo densities in hot spots. The Spo11 oligo map showed spatial and quantitative agreement with direct assays of DSBs in genomic DNA (Figures 1E and 1G), and matched or exceeded sensitivity of DSB detection from rad50S-like mutants (e.g., note weak signals in the YCR048w ORF, Figure 1G). This agreement allows us to convert oligo counts to percentage of DNA broken (Figure 1E), from which we estimate that 160 DSBs form in nonrepetitive sequences per meiotic cell in wild-type (see Extended Experimental Procedures). This value agrees with prior estimates (Buhler et al., 2007) and can account for detectable crossovers and noncrossovers (mean = 136.7 recombination events per meiosis) (Mancera et al., 2008). As expected from prior studies (Petes, 2001; Lichten, 2008), most Spo11 oligos were from intergenic regions containing promoters, but a significant number mapped within ORFs (Figure 1G and Figure S1D). Our oligo map agrees with microarray hybridization of ssDNA from dmc1 mutants (Blitzblau et al., 2007; Buhler et al., 2007; Borde et al., 2009) but provides much higher resolution (Figure 1F and Figure S1E). Thus, sequencing Spo11 oligos provides a genome-wide DSB map with unprecedented spatial and quantitative accuracy in recombination-proficient strains (Figure S2). Below, we explore this map at increasingly finer scale, from whole chromosome to single nucleotide. This analysis defines factors that interact in a hierarchical and combinatorial manner to shape DSB distributions.
Chromosome Size-Correlated Variation in DSB Frequencies We exploited the quantitative nature of our data to address the mechanisms behind chromosome size-associated variation in recombination. Small chromosomes cross over more often per kb than longer chromosomes (Kaback et al., 1992) (Figure 2A). Previously proposed mechanisms include smaller chromosomes having higher hot spot density, having more DSBs, favoring a crossover instead of noncrossover recombination outcome, and/or having less crossover interference (Kaback et al., 1992; Gerton et al., 2000; Martini et al., 2006; Blitzblau et al., 2007). Similar to crossovers, more Spo11 oligos per kb were recovered from smaller chromosomes (Figure 2B and Figure S3A), so crossover density correlated strongly with oligo density (r = 0.79) (Figure S3B). In contrast, there appeared to be little difference between large and small chromosomes for either the crossover versus noncrossover decision (Figure S3C), or the choice of homolog versus sister chromatid as partner for recombination (Figure S3D). We infer that smaller chromosomes tend to experience more DSBs per kb, accounting for much of the crossover density variation. Spo11 oligo hot spots (described below) occurred at similar density on all chromosomes (Figure S3E), so the greater DSB density on smaller chromosomes is not simply because of a higher density of favorable DSB sites. Subchromosomal Domains of Suppressed or Enhanced DSB Formation Telomere-Proximal Regions Spo11 oligos were less frequent in the 20 kb closest to each telomere (Figure 2C), matching DSB suppression zones seen by ssDNA mapping (Blitzblau et al., 2007; Buhler et al., 2007). Telomere structures in SK1 are not well defined, but inferring Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 721
from the S288C map, oligo counts were 3.5-fold lower than genome average in the telomere-proximal 20 kb (p <1015 compared to a random sample, Mann-Whitney). Although most oligos from subtelomeric repeats do not map uniquely, their aggregate contribution can be estimated. If repeats were omitted, oligo counts appeared even more reduced (6.5-fold; data not shown). Notwithstanding this suppression, 1.5% of oligos mapped within 20 kb of a telomere, suggesting that meiotic cells experience two to three such DSBs on average, consistent with crossover rates near chromosome ends (Barton et al., 2008). Pericentric Regions DSBs are suppressed near centromeres, but as resolution of whole-genome methods has increased, size estimates for suppressed zones have decreased from 20 kb (Gerton et al., 2000) to 8–10 kb (Buhler et al., 2007). In our study, strong reduction extended only a short distance compared to telomeres: Spo11 oligo density was 7-fold lower in the 3 kb surrounding centromeres compared with a randomized sample (p < 104, Mann-Whitney), whereas segments further away were 2- to 3-fold lower than random but within genome-wide variation (Figure 2D and Figure S3F). We observed hot spots near centromeres in agreement with Blitzblau et al. (2007) (Table S2), but hot spot density was lower than expected within 10 kb of centromeres (60%; p < 0.02), and hot spot strength within 5 kb tended to be weaker (mean = 3.3-fold; p < 0.01) (Figures S3G and S3H). We infer that DSBs are rare within 1–3 kb of centromeres and that 5–10 kb on either side is below average. In total, 0.4% of oligos mapped within 5 kb of centromeres, equivalent to 0.6 DSB per meiosis. Interestingly, pericentric oligo density varied 6.5-fold between chromosomes (Figure S3I), suggesting that different chromosomes may have different propensity toward missegregation caused by recombination disrupting pericentric cohesion (Rockmill et al., 2006; Chen et al., 2008). Interstitial Regions Chromosomes show alternating domains of inherently higher or lower DSB frequency (Borde et al., 1999; Petes, 2001; Blat et al., 2002), which can be visualized by smoothing Spo11 oligo distributions with windows of increasing size (Figure 2E). Analyzed this way, peak spacing and peak-to-valley ratios varied substantially between regions (Figure 2E and data not shown), so the domains do not alternate in a highly regular fashion. To explore mechanisms underlying these domains, we compared oligo distributions to several higher-order chromosome structural features. Consistent with prior studies (Gerton et al., 2000; Blat et al., 2002), Spo11 oligos correlated positively with GC content. However, additional patterns emerged when correlations were evaluated using data binned in windows of varying sizes, such that the correlation with GC content was weak over short distances (1 kb) but was uniformly strong at longer ranges (Figure 2F). This pattern reflects superposition of at least two levels of spatial organization: DSBs occur more often in relatively GC-rich domains but at finer scale are mostly in intergenic regions (Figure S1D), which tend to be more AT rich than their surroundings. Spo11 oligos correlated negatively with presence of meiosisspecific cohesin subunit Rec8, as expected (Kugou et al., 2009), but anticorrelation was strongest at short range (<5 kb) and was weaker at larger scales (Figure 2F). This pattern also likely reflects superposition of different levels of spatial organiza722 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
tion. Anticorrelation in larger windows is consistent with the hypothesis that a fundamental organizing principle of DSB distributions is the arrangement of chromosomes as 10–20 kb chromatin loops emanating from a cohesin-enriched axis, with DSBs forming preferentially in cohesin-poor loops (Blat et al., 2002; Kleckner, 2006). Why anticorrelation is even stronger at short range is unknown but may reflect a tendency for Rec8 to be especially depleted in promoters. We also compared our data with mitotic distributions of other chromosome structure proteins (Lindroos et al., 2006; D’Ambrosio et al., 2008). Spo11 oligos showed only a weak correlation at short distances with mitotic condensin, but there was a strong anticorrelation with the G2/M distribution of Smc6 and, as previously noted, the mitotic cohesin subunit Scc1/Mcd1 (Blat and Kleckner [1999]; Figure 2F). This is consistent with the known correlation between Scc1 and the Smc5/6 complex (Lindroos et al., 2006), but the anticorrelation of Spo11 oligos with the two proteins had different size dependence. Taken together, these patterns point to existence of multiple levels of spatial organization of the DSB terrain, supporting the view that DSB distributions are shaped by numerous high-order chromosome structures that vary over different size scales and that intersect in complex combinations (Petes, 2001; Kleckner, 2006; Keeney, 2007). Comprehensive Identification of DSB Hot Spots We defined 3604 DSB hot spots as clusters of Spo11 oligos (Figure 3A, Table S2, and Extended Experimental Procedures). These hot spots agreed well with direct DSB detection both spatially (e.g., Figure 1G) and quantitatively (Figure 1E), and included 94 hot spots previously documented in SK1 by Southern blot (Extended Experimental Procedures). Spo11 oligo hot spots account for nearly all hot spots identified by ssDNA mapping (Blitzblau et al., 2007; Buhler et al., 2007; Borde et al., 2009), if allowance is made for spatial ambiguity from DSB hyperresection in dmc1 mutants (Figure 3A and Figure S4A). However, increased spatial precision of Spo11 oligo hot spot determination (Figure 3A and Figure S4A) resolved hot spots that were merged in microarray data by overlapping ssDNA resection tracts (Figure 3A). Thus, Spo11 oligos provide the highest resolution and most complete compilation of DSB hot spots available to date in a recombination-proficient organism. Canonical and Noncanonical Characteristics of DSB Hot Spots Apparent hot spot traits emerged from prior studies, such as a narrow width (50–250 bp) and a tendency to overlap promoters (Petes, 2001; Lichten, 2008). However, because few hot spots have been studied in detail and previous wholegenome data do not resolve individual hot spots, the full range of variability was unknown. Spo11 oligos address this issue and reveal additional features. Oligo hot spots had a median width of 189 bp, and 73.4% were 50–300 bp wide (Figure S4B). Most (88.2%) overlapped with promoters (Table S2), agreeing with studies of Chr III (Baudat and Nicolas, 1997). Thus, most hot spots conform to stereotypical patterns inferred from direct mapping of a small subset. Nonetheless, there were many exceptions. For example,
Figure 3. Identification of DSB Hot Spots
A ssDNA hotspots
150 100
Spo11 oligo hotspots
50 0 NTG1
127 P
Spo11 oligos (hpM/bp) Raw Smoothed (x15)
B
SYN8
128
CYS3
129 130 131 Position on Chr I (kb)
DSBs
spo11yf dmc1 sae2
100 80 60 40 20 0
DEP1
*
C
Cumulative frxn of hotspots
Spo11 oligos (hpM/bp) Raw Smoothed (x15)
200
YAL011w
132
133
and
Properties
(A) Hot spots near CYS3. Green bars are hot spots from this study. Hot spots from ssDNA mapping are also indicated (red and teal circles, peaks only [Buhler et al., 2007; Borde et al., 2009]; orange circle and line, peak and region with signal over threshold [Blitzblau et al., 2007]). (B) Verification of an unusually broad hot spot by Southern blot of genomic DNA (labels as in Figure 1G). DSB signal in dmc1 spreads out because of hyperresection. (C) Hot spot intensities vary over a smooth continuum. Inset shows cumulative fraction of hot spots versus cumulative fraction of Spo11 oligos. Red lines mark the fraction of hot spot oligos in the weakest 67% of hot spots. For more comprehensive data for DSB hot spots, see also Figure S4 and Table S2.
1.0 0.8 0.8
2001; Lichten, 2008). Thus, an open chromatin structure is inferred to be necessary 0.4 0.4 for Spo11 to access its DNA substrate, but the genome-wide relationship 0 0.2 between DSBs and nucleosome occu0 0.4 0.8 Cum. frxn of total hits in hotspots pancy remains unexplored. Therefore, 0 MST28 YAT1 SWH1 we generated high-resolution maps of 4 6 8 10 12 Hotspot intensity: Spo11 oligos (hpM, log2) micrococcal nuclease (MNase)-resistant 189 190 191 192 193 Position on Chr I (kb) mononucleosomes during meiosis and compared them to Spo11 oligos. Nucleosome occupancy was determined as previously described (Kaplan 10.4% of hot spots were R500 bp wide, and nine were >1.5 kb et al., 2009) (Figure S5A and Extended Experimental Procewide (Figure S4B). Where tested, direct assays verified anoma- dures). Chromatin digestion and deep sequencing were carried lously wide hot spots, some of which overlapped ORFs (Fig- out in two laboratories at 0, 1, 2, and 3 hr (data set N1) ,and ure 3B and Figure S4C; see also below). Moreover, nonpromoter 0 and 4 hr in meiosis (data set N2), respectively. Data set N2 hot spots accounted for 4.8% of uniquely mapped Spo11 oligos. samples were more heavily MNase digested, but the two data When hot spots were rank ordered, their oligo counts followed sets nonetheless agreed well for patterns described below. a smooth continuum over a 410-fold range (Figure 3C). This (See Extended Experimental Procedures and Figures S5B–S5D pattern has several implications. First, lack of an obvious break for additional discussion.) Relatively few differences were in the continuum indicates that the cutoff is arbitrary between observed when comparing premeiotic (0 hr) with meiotic sites that are hot spots and those that are not. Second, most samples, indicating that steady-state nucleosome occupancy hot spots were very similar to neighbors on the continuum. changes little during early meiosis. Moreover, patterns agreed Thus, small changes in measured DSB activity can cause large well with prior studies in vegetatively growing haploids of changes in hot spot rank, which may contribute to variability of different strains (Jiang and Pugh, 2009), attesting to the hot spot compilations that implicitly rely on rank ordering. Third, conserved structure of yeast chromatin (Radman-Livaja and the strongest 33% of hot spots contained >75% of all hot spot- Rando, 2010; Tsankov et al., 2010). Most S. cerevisiae promoters exhibit a short (130 bp) NDR associated oligos (Figure 3C, inset). Thus, most DSBs occur in a small subset of hot spots, whereas many different sites flanked by well-positioned nucleosomes, with the transcription account for a smaller (but still substantial) fraction of DSBs. Strik- start site (TSS) in the first (+1) nucleosome (Radman-Livaja and ingly, 11% of uniquely mapped Spo11 oligos fell outside of clear Rando, 2010) (Figure 4A). Substantial overlap of DSBs with hot spots. Thus, a purely hot spot-centric view misses a consid- promoters is largely explained by DNA accessibility in this stereotypical structure (Lichten, 2008). As expected, Spo11 erable fraction of total recombination events. oligos mapped preferentially in promoter NDRs (Figures 4B and 4C). Nucleosome-Depleted Regions as Windows Most Pol II promoters (80%) lack a TATA sequence; this of Opportunity for DSB Formation Of the few hot spots tested to date, most are nuclease-hypersen- class is enriched for constitutively expressed genes, whereas sitive sites in chromatin of meiotic and vegetative cells (Petes, TATA-containing promoters are more common for inducible 0.6
Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 723
TSS
C
1.4
NDR 1.0 -1
+1 +2
0.4
+3
0.6 100
5
80
4
60
3 40
2 1
20
0
0 DEP1
129
130 Position on Chr I (kb)
1
0
0
-400 0 400 Position relative to hotspot center (bp) 0 hr Data set N1 3 hr
Mean nucleosome occupancy
1
0 TATA 0.6
1.0
0.4 0.6
0.2
0.2
0 TATA-less 0.6
1.0
0.4 0.2
131 0.2
E Mean Spo11 oligos (hpM/bp)
Mean nucleosome occupancy
D
0.2
0.2
0.6
CYS3
F 1
0
1 2 3 4 5
colder hotspot quintile
Mean nucleosome occupancy
SYN8
Mean nucleosome occupancy
6
Spo11 oligos (hpM/bp)
Norm. nucleosome occupancy (genome avg. = 1)
B
Mean Spo11 oligos (hpM/bp)
-2
Figure 4. DSB Formation Chromatin Structure
All TSS 0.8 0.6
hotter
-400 0 400 Position relative to hotspot center (bp)
0 -400 0 400 Position relative to TSS (bp)
1.4 1.0
3' ends
0.8 0.6 0.4
0.6 0.2
0.2
0 -400 0 400 Position relative to stop codon of convergent genes (bp)
Mean Spo11 oligos (hpM/bp)
A
and
Local
(A) Stereotypical chromatin structure of most yeast promoters. (B) Spo11 oligos cluster in promoter NDRs. Normalized nucleosome occupancy is the number of sequence reads covering a base pair divided by genome average coverage (Kaplan et al., 2009). (C) Genome average nucleosome occupancy and Spo11 oligo density around all annotated TSS and separately for TATA-containing and TATA-less genes. (D) Hot spots are nucleosome depleted (n = 3600, omitting exogenous hot spots). (E) DSB activity is not correlated with degree of nucleosome depletion at hot spot centers. Hot spots were divided into quintiles by oligo count. (F) Low nucleosome occupancy is not sufficient for strong DSB activity. Mean profiles are shown for 30 ends of convergent genes. Spo11 oligos in (C), (D), (F) were smoothed with a 75 bp Hann window. For more details regarding the relationship between chromatin structure and DSB formation, see also Figure S5.
0 hr Data set N2 4 hr
genes (Basehoar et al., 2004). These classes differed in average chromatin structure: TATA-less promoters had a narrower average NDR and well-positioned +1 and 1 nucleosomes; whereas TATA-containing promoters had a wider average NDR (Figure 4C) (Mavrich et al., 2008). Spo11 oligo distributions matched this difference (Figure 4C). TATA-containing promoters also had a 1.5-fold higher mean oligo count (Figure 4C and Figure S5E), which may account for amino acid biosynthetic genes being enriched in hot spots in a prior study (Gerton et al., 2000). TATA-containing promoters have a higher level of histone turnover, potentially providing opportunities for increased access by Spo11 (Tirosh and Barkai, 2008). The conclusion that DSBs occur nearly exclusively on nonnucleosomal DNA is reinforced by the fact that essentially all Spo11 oligo hot spots had low nucleosome occupancy (Figure 4D and Figure S5F). However, hot spot oligo counts did not correlate with quantitative scores for nucleosome occupancy (Figure 4E). Moreover, low nucleosome occupancy is not sufficient for robust DSB formation. For example, NDRs are also prominent at 30 ends of genes (Kaplan et al., 2009) (Figures S5D and S5G), but these are not strong DSB sites unless they coincide with the promoter NDR of a downstream gene (Figure 4F). The hottest fifth of hot spots showed a wider average zone of low nucleosome occupancy (red line, Figure 4E). Because stronger hot spots tended to be wider on average (Figure S4B), we examined chromatin separately for ‘‘normal’’ width hot spots and unusually wide ones. Indeed, wider hot spots tended to have wider regions of nucleosome depletion (Figure S5H), suggesting that chromatin structure is a primary determinant of hot spot width. This conclusion is further supported by the 724 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
exceptionally wide hot spots at YAT1, NAR1, and WHI5, where overall nucleosome occupancy was low, and nucleosomes appeared relatively disordered (Figure S4D), suggesting that stably bound nucleosomes are sparse and variably positioned among cells. Our findings support the view that stable nucleosomes occlude Spo11 access to DNA in vivo, in turn suggesting that variability of nucleosome occupancy contributes to variation in the DSB landscape between individual cells or between strains. However, although lack of a nucleosome is a prerequisite for DSB formation, other factors play a more dominant role in determining the probability of DNA cleavage. Influence of TFs on DSB Spatial Patterns The effect on DSB formation of a few TFs—Bas1, Bas2, and Rap1—has been explored (reviewed in Petes, 2001). It was hypothesized that TF binding (but not transcription) promotes DSBs nearby by influencing chromatin structure and/or interacting with the DSB machinery (Petes, 2001). It was also hypothesized that TFs compete with Spo11 for DNA access, occluding DSB formation at their binding sites (Xu and Petes, 1996; Petes, 2001). Spo11 oligos allowed us to test these hypotheses genome wide. Spo11 oligos mapped frequently near 4233 binding sites of 77 TFs annotated based on chromatin immunoprecipitation and conservation (MacIsaac et al., 2006) (Figures 5A and 5B), which is not surprising because TF sites are enriched in promoters. We examined fine-scale patterns by grouping TFs based on local oligo distributions (Figure 5C and Table S3). We discuss three of these groups below. Other TFs are not considered further because they showed little spatial correlation with Spo11 oligos,
0.8
0.6
0.4
Mean nucleosome occupancy
0 1.2 1.0 0.8 0.6 0.4 0.2 0
Class 1
Class 2
1.0 0.8 0.6 0.4 0.2
1.2 1.0 0.8 0.6 0.4 0.2 0
0 Class 3
-400 0 400 Position relative to TF site (bp)
Mean normalized Spo11 oligos (arbitrary units; smoothed)
0.4 0.2
D
Bas1
0 0.8
Sum1
Abf1
0.4 0
-100 0 100 -100 0 100 Position relative to TF site (bp)
C
E Class 5
Class 4 Class 3 Class 2 Class 1 -200 0 200 Position relative to TF site (bp) Spo11 oligo count Low
High
Mean normalized Spo11 oligos
Rap1
0.8
Protection from: DNase I Spo11
2 1.6 1.2 0.8 0.4 0
-50 0 50 Position rel. to Reb1 sites (bp)
Hottest
B
1.0
,QR 3KR /HX +DS ,QR
Middle
All TF sites
)NK 6ZL +DS 6RN 0HW
Coldest
A
5PH
0
1 Fraction of TF sites Hotspot association None
1 2 3 4 Cold Hot Hotspot quartiles
Figure 5. DSB Patterns around TF Binding Sites (A and B) Mean normalized Spo11 oligo profiles (see Extended Experimental Procedures), smoothed with a 75 bp window, and nucleosome occupancy (3-hr sample) around TF sites. (C) TFs clustered according to local spatial pattern of Spo11 oligos. Each horizontal line on the heat map shows the mean profile for binding sites of a single TF, grouped by k-means clustering. Because locally normalized oligo counts were used, color coding reflects DSB spatial pattern, not total DSB intensity. (D) Zones of protection from Spo11 or DNase I (Hesselberth et al., 2009) around Reb1-binding sites (n = 156). (E) TFs that are most or least associated with frequent DSB formation nearby. TFs were rank ordered by mean oligo count ±500 bp from their binding sites. A subset of TFs is shown; others are in Figure S6B. Bars depict the fraction of sites not in hot spots (black) and divide the remainder according to hot spot quartile. For further information about the relationship between TFs and DSB formation, see Figure S6 and Table S3.
having either local oligo enrichment offset from the TF sites (Class 4 in Figure 5C) or evenly distributed oligos (Class 5). For 12 TFs, there was strong evidence for DSB occlusion at their binding sites (Class 1, Figures 5A and 5C). The two most striking examples were Abf1 and Reb1, whose binding sites were often located in Spo11 oligo hot spots (Table S3). Both proteins showed strong oligo enrichment adjacent to their sites but depletion in the central 40 bp (Figures 5B and 5D and Figure S6A). Abf1 or Reb1 binding promotes nucleosome exclusion nearby (Badis et al., 2008; Kaplan et al., 2009), and both bind chromatin in meiosis (Schlecht et al., 2008), so it is likely that they influence hot spot activity at least indirectly by providing favorable chromatin structure. Class 1 also includes Rap1 (Figure 5B and Figure S6A), whose binding sites occluded DSB formation in an altered HIS4 hot spot (Xu and Petes, 1996). Our results show that Spo11 tends to be prevented from cutting in natural Rap1-binding sites genome wide. Abf1, Reb1, and Rap1 footprints of protection against Spo11 cleavage (40–42 bp) were larger than for protection from DNase I cleavage in chromatin (19–24 bp) (Hesselberth et al., 2009) (Figure 5D and Figure S6A). We infer that Spo11 (and associated proteins) has a larger effective size than DNase I for cleaving DNA and that steric constraints place the Spo11 active site
R10 bp (30–40 A˚) away from surfaces of competing DNAbinding proteins. The findings also suggest that it is unlikely that Spo11-associated proteins must form an extensive DNAbinding surface prior to DNA binding by Spo11 itself. Seven TFs showed only weak DSB occlusion at their binding sites (Class 2, Figures 5A and 5C). A good example is Bas1. Consistent with prior studies (Mieczkowski et al., 2006), 32/37 (86.5%) analyzed Bas1 sites were in 18 Spo11 oligo hot spots (Table S3). However, Bas1 sites showed only modest depression of oligo counts in their immediate vicinity (Figure 5B). Thus, not all TFs that affect DSBs can block Spo11 access to DNA. Class 2 TFs may have low steady-state occupancy of their binding sites when DSBs form (e.g., because of short dwell time on DNA, or because they act earlier), or Spo11 and/or accessory factors can displace them. Binding sites for another 17 TFs showed local oligo enrichment, but, unlike Classes 1 or 2, no detectable DSB occlusion (Class 3, Figures 5A and 5C). An example is Sum1 (Figure 5B), which represses meiotic genes in vegetative cells and is displaced during meiosis (Ahmed et al., 2009). Thus, some Class 3 TFs may not block Spo11 simply because they are not chromatin bound at the relevant time. Nevertheless, some may influence DSBs by prior hit-and-run action on nucleosome occupancy or histone modifications. Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 725
Correlating TF-Binding Sites with DSB Frequency We also examined whether TFs can be linked, positively or negatively, to DSB hot spot activity. TFs differed widely when we compared the average total oligo counts near their binding sites (Figure 5E and Figure S6B). TF classes defined above did not correlate with oligo counts (Figure S6C); thus, DSB frequency and DSB spatial distribution are separate features of the interplay between TFs and DSBs. The five TFs with the highest mean oligo counts (the Ino2/Ino4 complex, Pho4, Leu3, and Hap1; Figure 5E) are not known to influence meiotic recombination, but enrichment of Pho4 sites in hot spots was noted before (Gerton et al., 2000). It is not yet clear whether these TFs are active players or innocent bystanders in hot spot activity, but their known properties are consistent with their being bound to target promoters during sporulation (Figure S6 legend). On the other end of the scale, several TFs had low oligo counts near their binding sites (Figure 5E). Most have not been characterized in meiosis, but it may be that they do not exist in SK1 (Figure S6 legend), are not expressed or not bound to targets during meiosis, or are linked to formation of closed chromatin that inhibits DSBs. Thus, these findings reveal TFs whose binding sites are predictive of hot spot activity or lack thereof. However, for most TFs, oligo counts varied widely between individual binding sites. For example, Fkh2 and Swi4 sites were about equally likely to be in a hot spot as not, and the hot spots they were associated with ran the gamut from weak to strong (Figure 5E). Most TFs had similar characteristics (Figure 5E and Figure S6B), so presence of these binding sites is a poor predictor of DSB frequency. This pattern reinforces the view that promoters provide windows of opportunity for Spo11, but DSB frequency is more strongly dictated by other factors. Fine-Scale Analysis of DSB Sites Spo11 displays biases for which phosphodiester bonds are cleaved, but it has been difficult to discern the patterns behind these preferences (Keeney, 2007; Murakami and Nicolas, 2009). Our data now provide a large library of individual cleavage sites, with the fine-scale Spo11 oligo distribution agreeing with direct DSB mapping (r = 0.77; Figure S7A) (Murakami and Nicolas, 2009), after accounting for spatial ambiguity of oligos whose 50 ends map next to C residues (Figure S1A). To explore Spo11 preferences, we aligned DNA sequences around each uniquely mapped oligo, using the SK1 genome sequence (Figure 6A and Figure S7B). All mapped oligos were included, but conclusions discussed below were also obtained if we used only oligos without 50 -C ambiguity (data not shown). No consensus was apparent, supporting the idea that Spo11 is flexible in terms of DNA sequences it can cleave (Murakami and Nicolas, 2009). Nonetheless, base composition was highly nonrandom from 16 to +30 relative to the predicted dyad axis of cleavage (Figure S7B). Figure 6B summarizes this pattern. The strongest bias encompassed 10–12 bp centered on the dyad axis (segment ‘‘a,’’ Figure 6B), a region predicted to contact Spo11 based on docking DNA against Top6A, the archaeal Spo11 homolog (Nichols et al., 1999) (Figure 6C). This biased composition likely reflects DNA properties promoting Spo11 binding and/or catalysis, so we examined this region in detail. Overall, it is AT enriched (64.6% versus 60.3% local average), 726 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
and the dinucleotide composition is consistent with a preference for relatively narrow, deep grooves on the side of the DNA facing Spo11 (Figure S7D). G was enriched, and C was depleted at the third base in Spo11 oligos (Figure S7B), which is the complement of the base 50 of the scissile phosphate on the opposite strand (Figure 6D). Thus, Spo11 cleavage is favored 30 of C, and as previously shown (Murakami and Nicolas, 2009), cleavage 30 of G is disfavored. Dinucleotide frequencies further refined Spo11 preference at the scissile phosphate: 50 -C[A/C/T] and TA were favored, whereas G[A/C/T] and AA were disfavored (Figure 6D). Bias was also observed at 11–16 bp symmetrically to the right and left of the dyad axis, outside the predicted Spo11 footprint (‘‘b’’ segments, Figures 6B and 6C). These zones, which are modestly GC enriched (41.8% versus 39.7% local average), likely reflect preference of a Spo11-associated protein or a Spo11 domain not modeled by the Top6A structure. Another region that was asymmetric relative to the dyad axis (segment ‘‘c,’’ Figures 6B and 6C) probably reflects bias for oligo 30 -end formation, discussed below. For the central 32 bp, dinucleotide composition on the right correlated with the reverse complementary composition on the left (Figures S7E and S7F), as predicted for 2-fold rotational symmetry around the theoretical dyad axis. This symmetry does not imply that individual Spo11 cleavage sites are palindromic. Instead, it appears that left and right half-sites contribute separately because sites with a favored base composition on one side were less likely to show favored composition on the other side (data not shown). Importantly, DNA 50 of each oligo was engaged by Spo11 and accessory factors in vivo but was never encountered by enzymes used in vitro to manipulate the oligos. Thus, left:right symmetry demonstrates that observed biases are inherent to DSB formation and cannot be artifacts of methods to sequence Spo11 oligos. Formation of Spo11 Oligo 30 Ends Because reads on the 454 platform are relatively long, we could define the 30 end of each oligo, which is likely formed by nuclease activity of Mre11, or possibly Sae2 (Figure 6A) (Keeney, 2007). Base composition around 30 ends was nonrandom in a pattern distinct from 50 ends (Figure 6E and Figure S7C). Strong bias was limited to 3–4 bp centered on the 30 ends, most notably a small but significant enrichment for T. Thus, Mre11 (or Sae2) activity appears to be only modestly affected by DNA composition around the scissile phosphate, with a slight preference for homopolymeric T runs. Unlike 50 ends, 30 ends frequently mapped within boundaries of positioned nucleosomes and in TF-binding sites where 50 ends were rare (Figures 6F and 6G and Figure S6A). Thus, neither nucleosomes nor the TFs that block Spo11 appear to be a barrier to endonucleolytic DSB processing. Possibly, Mre11-dependent cleavage can occur on DNA still bound by histones or TFs, but it seems more likely that these protein-DNA interactions are disrupted before cleavage, either through normal dynamics of histone- or TF-DNA interactions or via active displacement by Spo11 and/or accessory proteins. In principle, disruption of these protein-DNA interactions could occur either prior to or as a consequence of DSB formation. However, Mre11 associates with DSB sites independent of DSB formation (Borde
3' end of Spo11 oligo
Spo11 oligo
5'-NNNNNNNNNNNNNNNNNNNN 3'-NNNNNNNNNNNNNNNNNNNN
E Deviation of dinucleotide frequency from local avg.
Deviation of dinucleotide frequency from local avg.
B
a
20
b c
10
0 -100
-50
0
50
100
Position relative to dyad axis (bp) a
C
b
5'-NNNNNNNNNNNNNNNNNNNN 3'-NNNNNNNNNNNNNNNNNNNN
20
10
0 -100
-50
0
50
100
Position relative to scissile phosphate at 3' end (bp) c
b
F YAL043c
Abf1
YAL042w
40 5' ends
-20
-10 0 10 Position relative to dyad axis (bp)
High
20
Spo11 oligos (hpM/bp)
Low
30
G 5'-NNNNNN 3'-NNNNNN
1
+1
0
-1
AT
TA
TG CG AG TC AC TT AA GT GC GA
GG
CA
CC
CT
-2
Dinucleotide at Spo11 cleavage site
Spo11 oligos (hpM/bp)
Fold enrichment over local average (log2)
D
0.6
-1 Watson
40 Crick 40 3' ends
-150 0 150 Position relative to nucleosome center (bp)
-2 2 1 0
0
-1 -2
40
0.2 0
0
0
5' ends 3' ends
0.4
2 1
Biased dinucleotide composition
Normalized nucleosome occupancy (log2)
dyad axis
A
61.1 61.2 61.3 Position on Chr I (kb)
Figure 6. High-Resolution View of DSB Sites (A) Schematic of a Spo11 DSB. Staggered cuts (blue arrows) by a Spo11 dimer generate a 2 nt 50 overhang, the middle of which is a 2-fold rotational symmetry axis. Nucleolytic cleavage (black arrow) forms the oligo 30 end. (B) Nonrandom dinucleotide composition around Spo11 cleavage sites. At each position, deviation of dinucleotide frequencies from local average was summed (Extended Experimental Procedures). Gray line indicates deviation for a randomized sample of 50-mers. (C) Biased DNA composition relative to predicted Spo11 binding. Dimeric structure is shown of a fragment of a Spo11 homolog, archaeal Top6A (Nichols et al., 1999), with monomers in green and blue, and catalytic tyrosines in magenta. White bars indicate scissile phosphates in DNA docked on the dimer and color coded by composition bias in (B). (D) Spo11 preference at the scissile phosphate (dinucleotide indicated by the red circle). (E) Nonrandom dinucleotide composition around 30 ends, as in (B). (F and G) Nucleosomes and TFs are not a barrier to 30 -end formation. (F) Spo11 oligos in an intergenic region. Cartoon depicts annotated nucleosomes (ellipses), TSS (arrows), and Abf1-binding site. Graphs show 50 and 30 ends of oligos from top (red) and bottom (blue) strands. Nucleosome occupancy (3 hr) is in gray. (G) Oligos around +1 nucleosomes (n = 5036). Genome averages were smoothed with 5 bp sliding window. Figure S7 provides nucleotide resolution analysis of many DSB sites.
et al., 2004), and is required for DSBs (Keeney, 2007). Thus, we propose that assembly of Spo11-containing pre-DSB complexes on DNA competitively replaces or actively displaces other proteins. This scenario can explain increased MNase sensitivity observed in hot spots prior to DSBs (Ohta et al., 1994), much of which is in NDRs themselves and is not accompanied by loss of positioned nucleosomes (Lichten, 2008). Evidence for Asymmetry in an Early DSB Processing Step We proposed three models to explain the 1:1 ratio of prominent oligo subpopulations (Neale et al., 2005): each DSB could be
processed asymmetrically to yield one short and one long oligo (Figure 7Aa); each DSB could be processed symmetrically to yield two long or two short oligos, with the two outcomes equally likely genome wide (Figure 7Ab); or nucleolytic cleavage could occur either near or far from each DSB, with independent positions on the left and right, and an 50% chance for near versus far (Figure 7Ac). The latter model predicts a mix of DSBs with oligos that are asymmetric, symmetric long, or symmetric short. Because we recovered mostly the longer oligos, symmetry can be evaluated by asking whether oligos were recovered equally from the top (‘‘Watson’’) and bottom (‘‘Crick’’) strands Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 727
A
b. Symmetric
a. Asymmetric
or or
C
B Total=44
Total=1,615 Crick
400 Spo11 oligos
Total=2,008 200
Watson
0 Crick
Total=1,546
200
E
Spo11 oligos on Crick (log2)
Spo11 oligos
Watson
1,449
D
Individual cleavage sites
10
0 5'- & & $ 7 & $ $ $ 7 $ & & & * * & & * 7 * * 0
200
Spo11 oligos on Crick (log2)
or
nick
200
Figure 7. Spo11 Oligo Asymmetry
c. Mixed
nick
8 6 4 2 0 0 2 4 6 8 10 Spo11 oligos on Watson (log2)
12 10
Within hotspots
(A) Models for origin of long and short oligos. (a). Nicks (red arrows) positioned asymmetrically relative to Spo11. (b). Two DSB classes with symmetric nicks. (c) Variable nick placement, independent on left and right. (B) Asymmetry at DSB sites in the YCR048w hot spot (Chr III, 211,744–211,764 bp). Arrows show two known DSB sites that cannot be resolved due to ambiguity of oligos with 50 -C. Gray brackets indicate regions pooled to tally oligos for each strand. (C) Substantial asymmetry at Spo11 cleavage sites genome wide. Strand-specific counts were tallied for sites with eight or more oligos total (n = 49,115). Points in red are significantly asymmetric (Poisson test, p % 0.05 after correction for false discovery). Green lines show 95% confidence intervals for Poisson sampling with equal frequencies on Watson and Crick. (D) Net asymmetry in the CYS3 hot spot. (E) Net asymmetry of hot spots genome wide (n = 3604, colors as in C).
8 6 4
(data not shown). It remains to be determined whether hot spot asymmetry is simply the 2 4 6 8 10 12 130.4 130.7 131.0 aggregate of independent DSB sites, or if direcSpo11 oligos on Watson (log ) Position on Chr I (kb) 2 tionality is influenced by as yet unknown local chromosomal features. Nonetheless, the findings provide candidate sites to test the proposal that DSB asymmetry might influence later at individual DSB positions. Figure 7B shows two strong recombination steps, such as which end is first to invade the cleavage sites in the YCR048w hot spot (Murakami and Nicolas, unbroken homolog (Neale et al., 2005). 2009). These sites are not resolved in our study because of 50 -C ambiguity, but total oligos from these sites can be tallied for the DSBs at Risk for Genome Rearrangements two strands (gray brackets, Figure 7B). We recovered 1615 DSBs in repetitive DNA can lead to genome rearrangement if Crick oligos, but only 44 Watson oligos, revealing strong asym- nonallelic homologous sequences are used as recombination metry (p < 2.2 3 1016, Poisson test). We infer that equal oligo templates (reviewed in Sasaki et al., 2010). Our data allowed numbers were formed on both strands in vivo but that Watson us to examine these ‘‘at-risk’’ DSBs. Only 1.53% of Spo11 oligos oligos often escaped detection, perhaps because they were mapped to two or more positions in the genome, and an additional 0.23% mapped to unique positions within boundaries of too short. Many DSB sites analyzed showed significant asymmetry repetitive elements (Figure 1D). High copy number repeats (39.3%, Figure 7C), which is incompatible with the obligate (rDNA, telomeres, retrotransposons, and tRNA genes) acsymmetry model (Figure 7Ab) and with versions of the mixed counted for only 1.16% of total oligos, despite these repeats model (Figure 7Ac) in which 30 -end positions are random every occupying 14% of the genome. Remaining oligos that mapped time a DSB is made. Instead, our findings are compatible with to multiple locations were from low copy repeats such as multithe obligate asymmetry model (Figure 7Aa). Oligo counts are gene families or from regions with low sequence complexity. a population average; thus, the fact that not all DSB sites showed As noted above, few oligos were from rDNA; these mapped fairly asymmetry could mean that every DSB is processed asymmet- uniformly across the repeat unit (Figure S1F). Subtelomeric X and Y0 elements accounted for 0.43% of rically, with different sites showing greater or lesser propensity for the direction to be the same in different cells. The results mappable reads, or 0.7 DSBs per meiosis assuming all recovare also compatible with the mixed cleavage model (Figure 7Ac) ered sequences were from bona fide Spo11 oligos (Figure 1D). if degree of asymmetry is particular to individual DSB sites rather Most of these oligos were from Y0 elements (Figure S1F), consistent with frequent rearrangement of chromosome ends through than being the same genome wide. We also evaluated hot spot asymmetry. On a population Y0 recombination in meiosis (Horowitz et al., 1984). S288C has 50 full Ty retrotransposons and many more solo basis, asymmetric DSBs in a hot spot can either reinforce or cancel one another. For example the CYS3 hot spot had net long terminal repeats (LTRs) or LTR fragments, totaling 3% of 1.3-fold asymmetry in favor of Watson (p < 1014) (Figure 7D). its genome, but SK1 has only approximately half as many fullOf the 3604 hot spots identified, 1754 (48.7%) showed net length Tys, and insertion sites are not conserved (Gabriel et al., asymmetry (Figure 7E). Direction or presence of asymmetry 2006; Liti et al., 2009) (unpublished data). Meiotic recombination does not correlate with orientation of adjacent transcription units was rare in artificial Ty constructs, and meiotic DSBs were not DEP1
CYS3
2
728 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
detected in a full Ty element, likely reflecting its relatively closed chromatin structure (Ben-Aroya et al., 2004). Correspondingly, only 0.28% of Spo11 oligos were from full Tys or LTRs (Figure 1D and Figure S1F), indicating that meiotic DSBs tend to be somewhat suppressed in natural Ty elements. These findings demonstrate that DSB formation is suppressed within repetitive DNA genome wide, albeit to various degrees for different repeat families. Nonetheless, the total burden of such DSBs is substantial: from the number of Spo11 oligos recovered, we estimate an average of approximately two to three DSBs per meiosis. Not all types of repeats have the same potential to generate lethal chromosome rearrangements through nonallelic homologous recombination, but all have potential to contribute to genome plasticity and evolution, and all have potential to adversely affect meiotic chromosome pairing and disjunction. Thus, as yet poorly understood mechanisms that control nonallelic recombination are clearly critical in nearly every meiosis to maintain genome integrity (Sasaki et al., 2010). Conclusions Our findings reinforce the view that the DSB landscape in S. cerevisiae is shaped by combinatorial action of many factors (Petes, 2001; Kleckner, 2006; Keeney, 2007; Lichten, 2008). These factors operate over many size scales and include whole chromosome variation, large subchromosomal domains, chromosome structure proteins, chromatin structure (including nucleosomes and sequence-specific DNA-binding proteins), and local DNA composition. Moreover, these factors work hierarchically, with the general trend that level in hierarchy is proportional to scale. Thus, for example, two DNA segments that are equally free of nucleosomes may have substantially different DSB probability depending on their locations relative to loop-axis chromosome organization or telomeres. These studies also lead to reexamination of the definition of DSB hot spots. From the earliest proposals that recombination initiates preferentially at defined sites (Holliday, 1964), the concept of hot spots has been useful for describing how recombination is distributed and for cataloging preferred sites. However, we suggest that this concept also tends to lead, intentionally or not, to the view that hot spots are discrete functional entities. We show here that many base pairs—and maybe all— are potential DSB sites, with a continuous distribution of cleavage probability. We also show that hot spot definition is arbitrary, i.e., that it is difficult to draw a biologically meaningful boundary between what is and what is not a hot spot. Moreover, whereas most hot spots comprise a narrow cluster of DSB sites within a promoter, many do not fit this stereotype, and many DSBs occur outside of hot spots entirely. These findings argue against a granular, hot spot-centric view of DSB distributions. Instead, the DSB landscape is better seen as a quantitative, probabilistic distribution across the entire genome. A hot spot is a group of phosphodiester bonds that share a high local likelihood of cleavage, i.e., simply one spatial organization level among many. In this view, Spo11 is an opportunistic cutter, and hot spots are its windows of opportunity (Lichten, 2008). As shown here and in prior studies, chromatin structure is a key determinant of these windows. Low nucleosome occupancy in promoters is conserved in S. cerevisiae strains and
related species (this work and Tsankov et al. [2010]), likely because of selection to maintain appropriate chromatin structure for gene expression. In fact, chromosome structures at all size scales are probably under evolutionary constraints because of functions important for DNA replication, compaction, and segregation. Hence, many factors that govern Spo11 activity are under selective pressures independent of meiosis or recombination per se (Nicolas et al., 1989). This predicts that some features of DSB landscapes will show significant similarity between related yeast species: quantitative values are likely to differ substantially, but many hot spots in one strain or species are likely to be hot spots in another. In this regard, S. cerevisiae and its relatives may contrast with mammals, where a meiosisspecific factor, PRDM9, appears to specifically target SPO11 to sites that may not have another intrinsic function (Neale, 2010), and which might thus be easier to evolve toward hot spot extinction because of different selective constraints. It will be interesting to compare Spo11 oligo maps between species, such as mammals and budding and fission yeasts, whose chromosomes have distinct architectures and evolutionary dynamics. EXPERIMENTAL PROCEDURES Detailed methods are in Extended Experimental Procedures. Strains are diploid derivatives of SK1 (Table S4). The Spo11-HA construct used here gives reduced DSBs (80% of wild-type) but does not appear to grossly alter DSB distributions (Figure S7). Spo11 oligos were sequenced as follows (Figure S1A). Cultures were harvested at 4 hr of synchronous meiosis, then denaturing extracts were prepared from nuclei of hypotonically lysed spheroplasts. Spo11-HA-oligo complexes were immunoprecipitated, size fractionated by SDS-PAGE, then protease digested. The 30 ends of resulting Spo11 oligos were extended with GTP and terminal deoxynucleotidyl transferase, then ligated to a double-stranded DNA adaptor using T4 RNA ligase 2. The complementary strands of Spo11 oligos were synthesized, then gel purified on denaturing PAGE, and subjected to another round of 30 -GTP tailing, adaptor ligation, and strand synthesis. Sequencing primers were added by low-cycle numbers of PCR, products were sequenced on the 454 platform (Roche) according to manufacturer’s instructions, and reads were mapped to S288C and SK1 genomes. Nucleosome mapping was carried out by deep sequencing of MNase-digested chromatin according to an established method (Kaplan et al., 2009). Data analysis utilized R (http://www.r-project.org/) and Bioconductor (http://www.bioconductor.org/). Map files in several formats are available for download (see Table S5 and Data S1). ACCESSION NUMBERS Raw sequence data are available from the GEO repository (accession numbers GSE26449 and GSE26452). SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, seven figures, five tables, and one data file and can be found with this article online at doi:10.1016/j.cell.2011.02.009. ACKNOWLEDGMENTS We thank A. Viale (MSKCC Genomics Core Laboratory) for sequencing; S. Shuman and J. Nandakumar (MSKCC) for generous gifts of RNA ligase; and M. Lichten (NCI), I. Whitehouse (MSKCC), and G. Smith (Fred Hutchinson Cancer Research Center) for comments on the manuscript. This work was supported in part by NIH grants GM58673 (S.K.) and HD53855 (S.K. and
Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 729
M.J.). M.J.N. is supported by a Royal Society University Research Fellowship, an MRC New Investigator Grant, and an HFSP Career Development Award. J.P. was supported by a Leukemia and Lymphoma Society Fellowship. R.K. was supported by NIH training grant T32 GM008539. S.K. is an Investigator of the Howard Hughes Medical Institute.
Hesselberth, J.R., Chen, X., Zhang, Z., Sabo, P.J., Sandstrom, R., Reynolds, A.P., Thurman, R.E., Neph, S., Kuehn, M.S., Noble, W.S., et al. (2009). Global mapping of protein-DNA interactions in vivo by digital genomic footprinting. Nat. Methods 6, 283–289. Holliday, R. (1964). A mechanism for gene conversion in fungi. Genet. Res. 5, 282–304.
Received: December 3, 2010 Revised: January 13, 2011 Accepted: February 3, 2011 Published: March 3, 2011
Horowitz, H., Thorburn, P., and Haber, J.E. (1984). Rearrangements of highly polymorphic regions near telomeres of Saccharomyces cerevisiae. Mol. Cell. Biol. 4, 2509–2517.
REFERENCES
Jiang, C., and Pugh, B.F. (2009). A compiled and systematic reference map of nucleosome positions across the Saccharomyces cerevisiae genome. Genome Biol. 10, R109.
Ahmed, N.T., Bungard, D., Shin, M.E., Moore, M., and Winter, E. (2009). The Ime2 protein kinase enhances the disassociation of the Sum1 repressor from middle meiotic promoters. Mol. Cell. Biol. 29, 4352–4362.
Kaback, D.B., Guacci, V., Barber, D., and Mahon, J.W. (1992). Chromosome size-dependent control of meiotic recombination. Science 256, 228–232.
Badis, G., Chan, E.T., van Bakel, H., Pena-Castillo, L., Tillo, D., Tsui, K., Carlson, C.D., Gossett, A.J., Hasinoff, M.J., Warren, C.L., et al. (2008). A library of yeast transcription factor motifs reveals a widespread function for Rsc3 in targeting nucleosome exclusion at promoters. Mol. Cell 32, 878–887. Barton, A.B., Pekosz, M.R., Kurvathi, R.S., and Kaback, D.B. (2008). Meiotic recombination at the ends of chromosomes in Saccharomyces cerevisiae. Genetics 179, 1221–1235. Basehoar, A.D., Zanton, S.J., and Pugh, B.F. (2004). Identification and distinct regulation of yeast TATA box-containing genes. Cell 116, 699–709.
Kaplan, N., Moore, I.K., Fondufe-Mittendorf, Y., Gossett, A.J., Tillo, D., Field, Y., LeProust, E.M., Hughes, T.R., Lieb, J.D., Widom, J., et al. (2009). The DNA-encoded nucleosome organization of a eukaryotic genome. Nature 458, 362–366. Kauppi, L., Jeffreys, A.J., and Keeney, S. (2004). Where the crossovers are: recombination distributions in mammals. Nat. Rev. Genet. 5, 413–424. Keeney, S. (2007). Spo11 and the formation of DNA double-strand breaks in meiosis. In Recombination and Meiosis, D.H. Lankenau, ed. (Heidelberg, Germany: Springer-Verlag), pp. 81–123.
Baudat, F., and Nicolas, A. (1997). Clustering of meiotic double-strand breaks on yeast Chromosome III. Proc. Natl. Acad. Sci. USA 94, 5213–5218.
Kleckner, N. (2006). Chiasma formation: chromatin/axis interplay and the role (s) of the synaptonemal complex. Chromosoma 115, 175–194.
Ben-Aroya, S., Mieczkowski, P.A., Petes, T.D., and Kupiec, M. (2004). The compact chromatin structure of a Ty repeated sequence suppresses recombination hot spot activity in Saccharomyces cerevisiae. Mol. Cell 15, 221–231.
Kugou, K., Fukuda, T., Yamada, S., Ito, M., Sasanuma, H., Mori, S., Katou, Y., Itoh, T., Matsumoto, K., Shibata, T., et al. (2009). Rec8 guides canonical Spo11 distribution along yeast meiotic chromosomes. Mol. Biol. Cell 20, 3064–3076.
Blat, Y., and Kleckner, N. (1999). Cohesins bind to preferential sites along yeast chromosome III, with differential regulation along arms versus the centric region. Cell 98, 249–259.
Lichten, M. (2008). Meiotic chromatin: the substrate for recombination initiation. In Recombination and Meiosis: Models, Means, and Evolution, R. Egel and D.H. Lankenau, eds. (Berlin: Springer-Verlag), pp. 165–193.
Blat, Y., Protacio, R.U., Hunter, N., and Kleckner, N. (2002). Physical and functional interactions among basic chromosome organizational features govern early steps of meiotic chiasma formation. Cell 111, 791–802.
Lindroos, H.B., Strom, L., Itoh, T., Katou, Y., Shirahige, K., and Sjogren, C. (2006). Chromosomal association of the Smc5/6 complex reveals that it functions in differently regulated pathways. Mol. Cell 22, 755–767.
Blitzblau, H.G., Bell, G.W., Rodriguez, J., Bell, S.P., and Hochwagen, A. (2007). Mapping of meiotic single-stranded DNA reveals double-strand-break hot spots near centromeres and telomeres. Curr. Biol. 17, 2003–2012.
Liti, G., Carter, D.M., Moses, A.M., Warringer, J., Parts, L., James, S.A., Davey, R.P., Roberts, I.N., Burt, A., Koufopanou, V., et al. (2009). Population genomics of domestic and wild yeasts. Nature 458, 337–341.
Borde, V., Wu, T.-C., and Lichten, M. (1999). Use of a recombination reporter insert to define meiotic recombination domains on Chromosome III of Saccharomyces cerevisiae. Mol. Cell. Biol. 19, 4832–4842.
MacIsaac, K.D., Wang, T., Gordon, D.B., Gifford, D.K., Stormo, G.D., and Fraenkel, E. (2006). An improved map of conserved regulatory sites for Saccharomyces cerevisiae. BMC Bioinformatics 7, 113.
Borde, V., Lin, W., Novikov, E., Petrini, J.H., Lichten, M., and Nicolas, A. (2004). Association of Mre11p with double-strand break sites during yeast meiosis. Mol. Cell 13, 389–401.
Mancera, E., Bourgon, R., Brozzi, A., Huber, W., and Steinmetz, L.M. (2008). High-resolution mapping of meiotic crossovers and non-crossovers in yeast. Nature 454, 479–485.
Borde, V., Robine, N., Lin, W., Bonfils, S., Geli, V., and Nicolas, A. (2009). Histone H3 lysine 4 trimethylation marks meiotic recombination initiation sites. EMBO J. 28, 99–111.
Martini, E., Diaz, R.L., Hunter, N., and Keeney, S. (2006). Crossover homeostasis in yeast meiosis. Cell 126, 285–295.
Buhler, C., Borde, V., and Lichten, M. (2007). Mapping meiotic single-strand DNA reveals a new landscape of DNA double-strand breaks in Saccharomyces cerevisiae. PLoS Biol. 5, e324.
Mavrich, T.N., Ioshikhes, I.P., Venters, B.J., Jiang, C., Tomsho, L.P., Qi, J., Schuster, S.C., Albert, I., and Pugh, B.F. (2008). A barrier nucleosome model for statistical positioning of nucleosomes throughout the yeast genome. Genome Res. 18, 1073–1083.
Chen, S.Y., Tsubouchi, T., Rockmill, B., Sandler, J.S., Richards, D.R., Vader, G., Hochwagen, A., Roeder, G.S., and Fung, J.C. (2008). Global analysis of the meiotic crossover landscape. Dev. Cell 15, 401–415. D’Ambrosio, C., Schmidt, C.K., Katou, Y., Kelly, G., Itoh, T., Shirahige, K., and Uhlmann, F. (2008). Identification of cis-acting sites for condensin loading onto budding yeast chromosomes. Genes Dev. 22, 2215–2227. Gabriel, A., Dapprich, J., Kunkel, M., Gresham, D., Pratt, S.C., and Dunham, M.J. (2006). Global mapping of transposon location. PLoS Genet. 2, e212. Gerton, J.L., DeRisi, J., Shroff, R., Lichten, M., Brown, P.O., and Petes, T.D. (2000). Inaugural article: global mapping of meiotic recombination hot spots and coldspots in the yeast Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 97, 11383–11390.
730 Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc.
Mieczkowski, P.A., Dominska, M., Buck, M.J., Gerton, J.L., Lieb, J.D., and Petes, T.D. (2006). Global analysis of the relationship between the binding of the Bas1p transcription factor and meiosis-specific double-strand DNA breaks in Saccharomyces cerevisiae. Mol. Cell. Biol. 26, 1014–1027. Murakami, H., and Nicolas, A. (2009). Locally, meiotic double-strand breaks targeted by Gal4BD-Spo11 occur at discrete sites with a sequence preference. Mol. Cell. Biol. 29, 3500–3516. Neale, M.J. (2010). PRDM9 points the zinc finger at meiotic recombination hot spots. Genome Biol. 11, 104. Neale, M.J., Pan, J., and Keeney, S. (2005). Endonucleolytic processing of covalent protein-linked DNA double-strand breaks. Nature 436, 1053–1057.
Nichols, M.D., DeAngelis, K., Keck, J.L., and Berger, J.M. (1999). Structure and function of an archaeal topoisomerase VI subunit with homology to the meiotic recombination factor Spo11. EMBO J. 18, 6177–6188.
Sasaki, M., Lange, J., and Keeney, S. (2010). Genome destabilization by homologous recombination in the germ line. Nat. Rev. Mol. Cell Biol. 11, 182–195.
Nicolas, A., Treco, D., Schultes, N.P., and Szostak, J.W. (1989). An initiation site for meiotic gene conversion in the yeast Saccharomyces cerevisiae. Nature 338, 35–39. Ohta, K., Shibata, T., and Nicolas, A. (1994). Changes in chromatin structure at recombination initiation sites during yeast meiosis. EMBO J. 13, 5754–5763.
Schlecht, U., Erb, I., Demougin, P., Robine, N., Borde, V., Nimwegen, E., Nicolas, A., and Primig, M. (2008). Genome-wide expression profiling, in vivo DNA binding analysis, and probabilistic motif prediction reveal novel Abf1 target genes during fermentation, respiration, and sporulation in yeast. Mol. Biol. Cell 19, 2193–2207.
Petes, T.D. (2001). Meiotic recombination hot spots and cold spots. Nat. Rev. Genet. 2, 360–369.
Tirosh, I., and Barkai, N. (2008). Two strategies for gene regulation by promoter nucleosomes. Genome Res. 18, 1084–1091.
Petes, T.D., and Botstein, D. (1977). Simple Mendelian inheritance of the reiterated ribosomal DNA of yeast. Proc. Natl. Acad. Sci. USA 74, 5091–5095.
Tsankov, A.M., Thompson, D.A., Socha, A., Regev, A., and Rando, O.J. (2010). The role of nucleosome positioning in the evolution of gene regulation. PLoS Biol. 8, e1000414.
Radman-Livaja, M., and Rando, O.J. (2010). Nucleosome positioning: how is it established, and why does it matter? Dev. Biol. 339, 258–266. Rockmill, B., Voelkel-Meiman, K., and Roeder, G.S. (2006). Centromere-proximal crossovers are associated with precocious separation of sister chromatids during meiosis in Saccharomyces cerevisiae. Genetics 174, 1745–1754.
Xu, F., and Petes, T.D. (1996). Fine-structure mapping of meiosis-specific double-strand DNA breaks at a recombination hot spot associated with an insertion of telomeric sequences upstream of the HIS4 locus in yeast. Genetics 143, 1115–1125.
Cell 144, 719–731, March 4, 2011 ª2011 Elsevier Inc. 731
Double-Strand Breaks in Heterochromatin Move Outside of a Dynamic HP1a Domain to Complete Recombinational Repair Irene Chiolo,1,* Aki Minoda,1 Serafin U. Colmenares,1 Aris Polyzos,1 Sylvain V. Costes,1 and Gary H. Karpen1,2,* 1Department
of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720, USA *Correspondence:
[email protected] (I.C.), karpen@fruitfly.org (G.H.K.) DOI 10.1016/j.cell.2011.02.012 2Department
SUMMARY
Double-strand breaks (DSBs) in heterochromatic repetitive DNAs pose significant threats to genome integrity, but information about how such lesions are processed and repaired is sparse. We observe dramatic expansion and dynamic protrusions of the heterochromatin domain in response to ionizing radiation (IR) in Drosophila cells. We also find that heterochromatic DSBs are repaired by homologous recombination (HR) but with striking differences from euchromatin. Proteins involved in early HR events (resection) are rapidly recruited to DSBs within heterochromatin. In contrast, Rad51, which mediates strand invasion, only associates with DSBs that relocalize outside of the domain. Heterochromatin expansion and relocalization of foci require checkpoint and resection proteins. Finally, the Smc5/6 complex is enriched in heterochromatin and is required to exclude Rad51 from the domain and prevent abnormal recombination. We propose that the spatial and temporal control of DSB repair in heterochromatin safeguards genome stability by preventing aberrant exchanges between repeats. INTRODUCTION Double-strand breaks (DSBs) are efficiently repaired by two major pathways, HR (homologous recombination) and NHEJ (nonhomologous end-joining). In HR repair, resection of DSBs generates single-stranded DNA (ssDNA) ends that invade homologous sequences, which serve as templates for DNA synthesis and repair. Repair of single-copy sequences through HR is normally error free because a unique homologous sequence is present on the donor sister chromatid or homolog. NHEJ, by contrast, is intrinsically mutagenic because it simply joins broken ends without restoring any missing sequence. Heterochromatin, a specialized domain enriched for highly repetitive sequences, represents a specific challenge for DSB repair. This cytologically distinct region of the nucleus comprises about 30% of fly and human genomes (Hoskins et al., 2007; 732 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
Lander et al., 2001). The large number of repeated sequences in heterochromatin and their close proximity within nuclei exacerbate the risk for genome rearrangements in the presence of DSBs, particularly during HR repair. Recombination among repetitive sequences results in loss or duplication of information (Peng and Karpen, 2008). Recombination between identical repeats on nonhomologous chromosomes produces dicentric and acentric chromosomes that are known to contribute to human diseases such as cancer and infertility (Pearson et al., 2005). NHEJ repair of a DSB in repetitive DNA is potentially less problematic because small deletions or mutations do not affect the function of tandem repeats as severely as genes. Thus, two unresolved issues are whether DSBs in heterochromatin are repaired by NHEJ or HR and how repair occurs without threatening the stability of the genome. Heterochromatin in S. pombe, flies, and mammals is characterized by enrichment of specific histone modifications (e.g., H3K9me2 and H3K9me3) and associated proteins, such as heterochromatin protein 1a (HP1a) and the histone methyltransferase (HMTase) Su(var)3–9 (Grewal and Jia, 2007). Chromatin composition and regulation are integral to many aspects of the DSB response (Downs et al., 2007), and heterochromatin components and structure help to maintain repeat stability (Peng and Karpen, 2008). For example, the absence of Su(var) 3–9 results in the accumulation of DSBs in heterochromatin, formation of extrachromosomal circular repeated DNAs, translocations, and loss of heterozygosity (Peng and Karpen, 2007, 2009). How heterochromatin components promote genome stability and whether they regulate DSB repair have not been determined, but current evidence suggests three possible mechanisms. First, compaction or chromatin composition could make heterochromatin intrinsically less responsive to DSB formation or processing (Kim et al., 2007). DSBs produced by ionizing radiation (IR) in mammalian cells are detected at lower frequencies and are resolved with slower kinetics in heterochromatin (Cowell et al., 2007; Goodarzi et al., 2008), and delocalization of HP1b and Kap1 (binding partner for HP1) seem to facilitate repair (Ayoub et al., 2008; Goodarzi et al., 2008). Second, heterochromatin components could repress HR and promote less-damaging repair processes, similar to suppression of reciprocal recombination in pericentric heterochromatin during meiosis, which requires heterochromatin components (Westphal and Reuter,
2002). Third, HR repair could occur in heterochromatin, but aberrant recombination is prevented by specific mechanisms, as suggested by studies in S. cerevisiae (Torres-Rosell et al., 2007). Here, we show that IR-induced DSBs are efficiently formed and processed in Drosophila heterochromatin and that their repair is, surprisingly, dependent on HR. We demonstrate that heterochromatin responds dynamically to IR: proteins involved in early stages of HR are quickly recruited to DSBs and promote expansion of the domain, and repair sites display a dramatic relocalization to outside heterochromatin, where they first recruit a protein (Rad51) required for strand invasion and completion of HR repair. In addition, we identify the Smc5/6 complex as a new heterochromatin component that serves as a key regulator of HR repair in time and space. These results provide new insights into the dynamics of heterochromatic DSB processing and repair and suggest a mechanism for limiting the risk associated with HR repair of repetitive sequences. RESULTS DSBs Rapidly Disappear from Heterochromatin and Are Not Associated with Rad51 Foci In Drosophila, the DAPI-bright heterochromatin is organized as a distinct nuclear domain (Figure S1 available online) that is enriched for the canonical heterochromatin markers histone H3K9me2, H3K9me3, and HP1a (Eissenberg and Reuter, 2009). In contrast, marks associated with gene expression (e.g., H3K4me2 and me3) are excluded from the heterochromatin domain. This striking three-dimensional (3D) separation is maintained throughout the cell cycle (Figures S1D and S1E). An early marker for DSB formation is a phosphoepitope that appears on histone variant H2Av (H2AX in mammals), which is catalyzed by the ATM and ATR checkpoint kinases and contributes to checkpoint signaling and repair (Downs et al., 2007). Previous studies in mouse cells showed that gH2AX foci are mostly excluded from the heterochromatic chromocenters (DAPI-bright regions) at > 30 min after IR (Cowell et al., 2007), suggesting that DSBs are not formed or processed in heterochromatin. However, it is also possible that repair foci form in heterochromatin at earlier time points and are either quickly repaired or moved outside of the domain. Therefore, we performed a kinetic analysis of the DSB response in heterochromatin, using Kc tissue culture cells fixed at different time points after IR. TUNEL analysis reveals DNA damage in DAPI-bright in most cells at 10 min after IR (Figure S2A), demonstrating that heterochromatin is not refractory to IR-induced damage. However, at 30 min after IR, only 10% of the cells display TUNEL signals in DAPI-bright, when they are still abundant in DAPI-weak regions. Immunofluorescence (IF) analysis and quantitation show that the frequencies of gH2Av and Rad51 (a marker of HR repair) foci peak at 30 min (gH2Av) or 60 min (Rad51) after IR in DAPI-weak euchromatin (Figures 1A and 1B), as observed in mammalian cells (Costes et al., 2009). At later time points, foci frequencies steadily decrease (Figure 1B), which likely reflects ongoing repair. In contrast, the frequencies of gH2Av foci in DAPI-bright heterochromatin peak at 10 min after IR and then drop sharply (Figure 1B). Notably, Rad51 foci are rarely observed in DAPI-bright,
whereas in DAPI-weak, about half of the gH2Av foci are associated with Rad51 foci (Figure 1B). Similar kinetics are observed in S2 cells (data not shown). We conclude that: (1) DSB formation and processing do occur in heterochromatin soon after damage is induced, (2) DSBs and gH2Av foci rapidly disappear from heterochromatin, suggesting that heterochromatic DSBs are either repaired quickly or relocate elsewhere for repair, and (3) Rad51 foci are rarely observed in heterochromatin throughout the time course. Heterochromatic DSB Repair Requires the HR Pathway and Is Associated with Early HR Repair Components Repair by NHEJ would be consistent with the absence of Rad51 foci and the faster kinetics of DSB removal in the heterochromatin domain. To determine which repair pathway is involved, we analyzed the effect of NHEJ or HR inactivation on foci kinetics. Importantly, Kc cells are predominantly in S and G2 phases of the cell cycle (Figures S1A–S1D), when both pathways are functional (Huertas, 2010). In DAPI-weak regions, depleting proteins required for NHEJ (Ku70 and Ku80) or HR (Rad51 or Rad54) results in persistent gH2Av foci after IR (Figure 1C and Figure S2D), confirming the involvement of both pathways in repairing euchromatic DSBs in Kc cells. However, Ku70/80 depletion has no effect on gH2Av or Rad51 foci kinetics in DAPI-bright (Figure S2D), suggesting that NHEJ is not required for DSB repair in heterochromatin. In contrast, depletion of either Rad51 or Rad54 results in persistent gH2Av foci in DAPI-bright (Figure 1C and Figure S2E), reflecting unrepaired DSBs (gH2Av foci overlapping with TUNEL signals) (Figure S2E). We conclude that the fast disappearance of DSBs from Drosophila heterochromatin predominantly depends on the HR pathway. The requirement for HR is surprising, given the lack of Rad51 foci in heterochromatin. To determine whether HR steps that precede Rad51 loading occur in heterochromatin, we analyzed the kinetics of foci formation of ATRIP, which is recruited to RPA-covered resected DNA (Zou and Elledge, 2003). ATRIP and Rad51 kinetics are similar in DAPI-weak regions (compare Figure 1D to Figure 1B). In sharp contrast, ATRIP foci in DAPIbright peak at 10 min after IR and then quickly disappear, similar to gH2Av foci (Figure 1D). We conclude that early steps in DSB processing for HR repair (resection and ATRIP loading) occur within heterochromatin, whereas later steps (assembly of Rad51 foci) are excluded from this domain. This separation of early and late HR events is not observed in euchromatin, in which ATRIP and Rad51 foci display similar kinetics. Heterochromatin Rapidly Expands and Becomes Dynamic after IR One reason why HR in heterochromatin is potentially dangerous is because of the close proximity of homologous sequences on different chromosomes. Therefore, we explored whether the organization of this domain changes in response to IR. Timelapse studies of cells expressing mCherry-tagged HP1a (mChHP1a) show that the HP1a domain is compact and contiguous prior to IR treatment. However, soon after IR, this domain expands and becomes fragmented, with rapid extension and retraction of HP1a ‘‘fingers’’ (Figure 2A and Movie S1). HP1a Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 733
domain volume increases within minutes after IR (Figure S3A) and peaks at 20–40 min (1.5-fold compared to before IR) (Figure 2B and Figure S3A). This is followed by a partial contraction, which is maintained until at least 100 min after IR (Figure 2B and Figure S3B). Fixed cells display a similar increase in HP1a, H3K9me2, and DAPI-bright volumes after IR, with no increase in total nuclear volume (Figures S3C and S3D and data not shown). HP1a expansion was also observed after low dose radiation (1.67 Gy, data not shown; this dose is equivalent to 10cGy treatment of human cells). The increased volumes of HP1a and H3K9 methylation could result from spreading of heterochromatin into euchromatic regions in response to IR, as observed for chromosome rearrangements that juxtapose euchromatic and heterochromatic sequences (Eissenberg and Reuter, 2009). Alternatively, the heterochromatic domain could physically expand, in which case HP1a would still be restricted to heterochromatic sequences. Genome-wide HP1a ChIP-array analysis shows that HP1a does not become enriched in the euchromatin after IR (Figure 2C and Figure S3E), demonstrating that the volume of the heterochromatin domain expands without spreading into euchromatic sequences. In addition, we used fluorescence in situ hybridization (FISH) analysis to monitor the locations of the AACAC satellite sequences with respect to DAPI-bright before and after IR. Quantitative analysis demonstrates that these repeats localize predominantly within or close to DAPI-bright in untreated controls (Figure 2D and Figure S1F). However, the average distance between satellite signals and DAPI-bright increases significantly between 10 and 60 min after IR and then decreases by 3 hr (Figure 2D). We conclude that changes in the spatial distribution of heterochromatic repeated sequences occur in concert with the expansion and partial contraction of the HP1a domain.
A
v
60 min
10 min
v
B H
C
D
v v
Heterochromatic Repair Foci Relocate Outside of the Heterochromatin Domain Relocalization of DSBs to outside of heterochromatin to complete HR repair provides one explanation for the fast disappearance of gH2Av and ATRIP foci, the exclusion of Rad51 foci from heterochromatin, and the requirement for HR repair. To test this hypothesis, we performed live studies of cells expressing fluorescent-tagged versions of HP1a and components of the DSB response. Mu2/Mdc1 recruitment to DSBs relies on direct
v
Figure 1. Processing of DSBs in Heterochromatin Occurs with Different Kinetics Than Euchromatin and Requires HR (A) gH2Av, but not Rad51, foci form in DAPI-bright (dashed circles) after IR. IF with gH2Av or Rad51 antibodies shows gH2Av foci in DAPI-bright at 10 min and not at 60 min after IR. Rad51 foci are rare in DAPI-bright at both time points (see one example in Figure S2B). (B) IF was performed at different time points after IR as in (A), and quantitative analysis of foci was performed. In DAPI-weak, frequencies of gH2Av and Rad51 foci peak at 30 and 60 min after IR, respectively, and then show a slow decay. In DAPI-bright, rapid reduction of gH2Av foci is observed between
734 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
10 and 30 min after IR (p < 0.001; n > 60). No significant differences in the low numbers of Rad51 foci were observed in DAPI-bright throughout the time course (n > 100). (C) HR proteins are required for removal of gH2Av foci from heterochromatin. Quantitative kinetic analysis shows persistent gH2Av foci in DAPI-bright and DAPI-weak after Rad51 or Rad54 RNAi at 60 and 240 min after IR (p < 0.01; n > 100). Examples of cells are shown in Figure S2E. (D) ATRIP foci form in DAPI-bright in response to IR. Cells expressing GFPATRIP were fixed 10 min after IR and stained with gH2Av antibodies (top). Quantitation (bottom) reveals rapid reduction of ATRIP and gH2Av foci between 10 and 30 min after IR (p = 0.001; n > 60). In DAPI-weak, the kinetics of ATRIP foci is similar to Rad51 (see B). Graphs show mean ± SEM. Only Z-stacks including the DAPI-bright are shown as projections in (A) and (D). Scale bars, 1 mm. See also Figure S1 and Figure S2.
A
HP1a -34’
-31’
-28’
-25’
-22’
-19’
-16’
23’
26’
29’
32’
35’
38’
41’
-IR
+IR
fold increase
B
29’
32’
HP1a volume
35’
38’ C
chromosome arm 3L
2 Expansion 1.8 Contraction UNT 1.6 1.4 30 1.2 Unt min 1 60 0.8 40 80 120 min -40 min 0 before IR
after IR
euchromatin
20
21
heterochromatin
22
23 Mb
D
AACAC-DAPI bright
distance / nuclear size
26’
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 10 30 60 180
time after IR (min.)
Figure 2. Heterochromatin Expands and Becomes Dynamic in Response to IR (A) HP1a expands after IR and forms dynamic protrusions. Stills from Movie S1 are shown, in which a single cell expressing GFP-HP1a was imaged before IR (IR) and starting 20 min after IR (+IR). Arrows and zoomed detail of the outlined region show HP1a protrusions. Scale bars, 5 mm. (B) Automated quantitation shows HP1a expansion (p < 0.0001) followed by partial contraction (p = 0.0011) (see examples in Figure S3B). After contraction, the volume of HP1a is still significantly larger than the untreated control (p < 0.0001) (two-tailed unpaired t test; n = 11; graph shows mean ± SEM). (C) ChIP-array analysis of HP1a localization was performed before (UNT), 30 min, and 60 min after IR. The browser view of the proximal half of chromosome 3L shows the position of the euchromatin-heterochromatin border (red line); x axis = arm coordinates in Mbs. y axis = enrichment p values, normalized to input. No major increases in HP1a enrichments are observed in the euchromatin after IR (other chromosome arms are shown in Figure S3). (D) IR induces expansion of heterochromatin sequences. Quantitation of the distance between AACAC signals, detected by FISH, and DAPI-bright shows higher values between 10 and 60 min after IR and partial decrease at 180 min (p < 0.001) (n = 150 signals). Red bar, mean; black bars, ± SEM. See also Figure S3 and Movie S1.
interactions with gH2Av and is an excellent marker for the early response to DSBs (Dronamraju and Mason, 2009). To detect later steps in HR repair, we tracked proteins recruited to RPAcovered resected DNA (ATRIP and TopBP1) and Rad51, which promotes strand invasion (Su, 2006). Time-lapse studies and 3D reconstructions of nuclei revealed that Mu2, ATRIP, and TopBP1 foci appear within the HP1a domain by 3 min after IR and then move to the HP1a periphery; by 30–40 min after IR, most foci are located at the HP1a periphery or outside of the domain (Figures 3A and 3C, Movie S2, Movie S3, Figures S4A–S4E, and Figure S5A). Interestingly, ATRIP and TopBP1 foci are more abundant and much brighter within the HP1a domain at early time points, compared to the euchromatin (Figure 3A and Figure S4B). ATRIP foci dynamically join and split within the HP1a domain (Figure S5B and Movie S4) and often split into multiple foci when relocated outside of the
HP1a domain (Figure 3C and Figure S5A). This suggests that the bright ATRIP and TopBP1 foci could result from early clustering of DSBs in heterochromatin, which separate after relocalization. Interestingly, assembly of ATRIP and TopBP1 foci occurs with kinetics similar to Mu2 foci in heterochromatin but is delayed outside of the HP1a domain (Figure 3A and Figures S4A and S4B), consistent with our observations in fixed cells. Thus, DSB recognition (Mu2 and gH2Av foci) occurs with the same kinetics in euchromatin and heterochromatin, whereas subsequent steps (resection or ATRIP/TopBP1 recruitment) are accelerated or enhanced in heterochromatin. In contrast, Rad51 foci are rarely observed within the HP1a domain at all time points, consistent with the analysis of fixed cells (Figure 3B). Instead, they appear at the periphery of the HP1a domain starting 30 min after IR (Figure 3B) and move with the ends of the dynamic HP1a fingers (Figure 3D, Figures Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 735
A 60’
10’
Rad51 HP1a
HP1a
HP1a
ATRIP
Rad51 ATRIP
Position of foci vs HP1a domain:
avg. # foci / cell
6 4
IN Periphery (in) Periphery (out) OUT
2 0
0
100
200
300
3’
10’
60’
Rad51
time after IR (min.) HP1a
30’
avg. # foci / cell
30’
ATRIP HP1a
C
Figure 3. Heterochromatic DSBs Relocalize Outside of the HP1a Domain, Where They Associate with Rad51
B 10’
3 2 1 0 0
100
200
300
time after IR (min.) 13’
31’
43’
ATRIP
E
HP1a Rad51 54’
DAPI
γH2Av AACAC -IR
+IR 46’
54’
48’
56’
50’
58’
52’
60’
AACAC-DAPI bright distance / nuclear size
D
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
γH2Av Rad51
0 60 60 0 60 60 + + γH2Av Rad51
time after IR (min.)
S4E and S4F, and Movie S5). The assembly of Rad51 foci correlates in space and time with relocalization of ATRIP and TopBP1 foci outside of the HP1a domain, suggesting that relocalization precedes assembly of Rad51 foci at heterochromatic DSBs (compare kinetics in Figure 3A with Figure S4A and images in Figure S4C with Figure S4E). Indeed, colocalization of TopBP1 with Rad51 foci progressively increases at the periphery of HP1a between 10 and 60 min after IR (when TopBP1 foci relocalize) (Figure S4G). The intensity reduction for TopBP1 and ATRIP foci during relocalization and Rad51 foci formation (Figures S4C, S4E, and S4G) likely results from the local removal 736 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
(A) ATRIP foci form within the HP1a domain and relocate outside of the domain. Stills from Movie S2 are shown, in which cells expressing GFPATRIP and mCh-HP1a were imaged before and after IR. (Top) Maximum intensity projections of one cell at 10, 30, and 60 min after IR show the relocalization of ATRIP foci to the periphery and outside of the HP1a domain. More time points are shown in Figure S4C. (Bottom) Quantitation shows increased ATRIP foci within the HP1a domain at 4 and 10 min and reductions after 10 min. ATRIP foci accumulate at the periphery of the HP1a domain from 10 to 60 min (p < 0.001; n = 30 cells). (B) Rad51 foci form at the periphery and outside of the HP1a domain after IR. Cells expressing GFPRad51 and mCh-HP1a were analyzed as in (A). (Top) Images of the same cell at 10, 30, and 60 min after IR; more time points are shown in Figure S4E. (Bottom) Quantitation shows that Rad51 foci only appear near the HP1a periphery or outside of the domain (n = 40 cells). (C) 3D modeling of stills from Movie S3 (Figure S5A), in which cells expressing GFP-ATRIP and mCh-HP1a were imaged before and after IR, shows ATRIP foci formed within the HP1a domain that translocate outside (arrows) and frequently split (see red, cyan, yellow foci). (D) Cells expressing GFP-Rad51 and mCh-HP1a, imaged before and after IR, show Rad51 foci localized at the ‘‘tips’’ of dynamic HP1a protrusions. Stills from Movie S5 are shown. Representative time points of the region outlined are shown below. (E) Damaged heterochromatic repeats are more distant from DAPI-bright compared to undamaged repeats. AACAC repeats and gH2Av or Rad51 foci were detected by FISH-IF of cells fixed before () and 60 min after (+) IR. Images and quantitation of the distance from DAPI-bright (dashed blue circle) show that AACAC signals that colocalize with gH2Av (60 + gH2Av) or Rad51 (60 + Rad51) are more distant than all AACAC signals at the same time point (60) (p < 0.001 and p < 0.02, respectively, by two-tailed Mann-Whitney test; n > 100 signals). The distances in untreated cells are also shown for each experiment (0). Images are projections of the Z-stacks, including the DAPI-bright. Graphs show mean ± SEM. Scale bars, 1 mm. See also Figure S4, Figure S5, Movie S2, Movie S3, Movie S4, and Movie S5.
of RPA during assembly of Rad51 nucleofilaments (Sugiyama and Kowalczykowski, 2002). We also confirmed that the relocalization of repair foci reflects the behavior of damaged repeated sequences associated with heterochromatic proteins. First, IF for gH2Av or Rad51 combined with DNA FISH (AACAC satellite repeats) shows that satellite sequences associated with gH2Av or Rad51 repair foci are more distant from DAPI-bright compared to all satellite signals at 60 min after IR (Figure 3E). This result directly links repeated heterochromatic sequences to HR repair foci relocalized outside of the heterochromatin domain. Second, coimmunoprecipitation
experiments show that ATRIP is strongly associated with HP1a and H3K9me3 at 6 and 40 min after IR (Figure S5C), when ATRIP foci are concentrated in heterochromatin. Reduced association between ATRIP and HP1a at 60 min, after relocalization of foci is complete, could result from the disassembly of ATRIP foci or local heterochromatin changes at the sites of damage. Interestingly, we observe some association of ATRIP with heterochromatin components prior to IR and focus formation. We conclude that repair foci are highly mobile in Drosophila cultured cells and that early steps in HR repair (e.g., resection) occur within the heterochromatic domain, whereas late HR steps (e.g., strand invasion) do not occur until DSBs relocate to the HP1a periphery. Heterochromatin Expansion and Relocalization of Foci Require Resection and Checkpoint Proteins We investigated whether proteins involved in the initial processing of DSBs, specifically checkpoint kinases and resection components, mediate heterochromatin expansion or relocalization of foci. Treatment with caffeine, a potent inhibitor of ATM and ATR kinase activity (Sarkaria et al., 1999), results in defective HP1a expansion after IR (3.6-fold reduction) (Figures 4A and 4B and Movie S6). Caffeine also delays formation of ATRIP foci in euchromatin but prevents formation of ATRIP foci within the HP1a domain (Figures 4A and 4C), suggesting a role for checkpoint kinases in the assembly of ATRIP foci specifically in heterochromatin. We directly addressed the roles of ATM and ATR in heterochromatin expansion by RNAi depletion (Figures S6A and S6B). HP1a expansion is partially defective after ATM RNAi (1.7-fold reduction) and severely impaired after ATR RNAi (3-fold reduction) (Figure 4D and Figure S6E). Simultaneous depletion of ATM and ATR results in defective expansion, but the magnitude of the defect is not significantly different from the removal of ATR alone (Figure 4D and Figure S6E). We conclude that checkpoint kinases are required to promote heterochromatin expansion after IR and that ATR plays a major role. Although regulators of resection have not previously been identified in Drosophila, there are homologs of three proteins required for resection in yeast and mammals (CtIP, Exo1 [Tosca in Drosophila], and Blm) (Digilio et al., 1996; Kusano et al., 1999; Uanschou et al., 2007). Simultaneous depletion of all three proteins (Figure S6A) severely impaired the assembly of ATRIP foci and not gH2Av and Mu2 foci (Figures S6C and S6D), indicating that resection is indeed inhibited. Importantly, HP1a expansion is almost completely abolished after CtIP/Tosca/ Blm RNAi (15-fold reduction) (Figure 4D and Figure S6E). Interestingly, loss of resection seems to have a stronger impact than checkpoint inactivation on HP1a expansion (Figure 4D). Moreover, gH2Av and Mu2 foci are weaker and less numerous after ATR RNAi (Figure S6B), but not after depletion of resection components (Figure S6D). Thus, ATR is still partially active after CtIP+Tosca+Blm RNAi, yet expansion is blocked; this suggests that resection could contribute to expansion independent of its role in promoting ATR recruitment. To address the impact of checkpoint proteins and resection on DSB relocalization, we quantified the frequencies of gH2Av foci
within and outside of the DAPI-bright region at 60 min after IR. We observe that gH2Av foci persist in heterochromatin when checkpoint or resection proteins are depleted (Figure 4E and Figure S6F). As for HP1a expansion, ATR, ATR+ATM, and CtIP+ Tosca+Blm depletions display the strongest effects. Time-lapse analysis of GFP-Mu2 confirms that relocalization of heterochromatic DSBs to the periphery and outside of the HP1a domain is defective when resection is inhibited (Figure 4F). We conclude that IR-induced expansion of the heterochromatin domain and relocalization of repair foci require checkpoint kinases (mainly ATR) and resection. Heterochromatin Proteins Regulate DSB Repair by Blocking Formation of Rad51 Foci Careful analysis of fixed and living cells revealed that Rad51 foci assemble in HP1a ‘‘holes’’ or ‘‘pockets’’ at the heterochromatin periphery after IR or where HP1a is not visible after NIR-laser targeting to the heterochromatin (Figure S4F and Figure S7B). This mutual exclusivity suggests that assembly of Rad51 foci could be blocked by heterochromatin proteins. We therefore tested the effect of heterochromatic protein removal on the distributions of Rad51 foci. Depletion of the Su(var)3–9 HMTase results in severe reduction of H3K9me2 and H3K9me3 levels and disassembly of HP1a, and these effects are exacerbated by simultaneous depletion of the SetDB1 HMTase (Figure S7C). The absence of H3K9 HMTases results in a 3- to 4-fold increased frequency of Rad51 foci after IR, specifically in DAPI-bright (Figure 5A). Similarly, HP1a depletion results in the abnormal presence of Rad51 foci in DAPIbright (Figure 5A), and the kinetics of gH2Av and Rad51 foci become similar to those observed normally in DAPI-weak euchromatin (Figure 5B, compare to Figure 1). Because H3K9 methylation levels are almost normal after HP1a RNAi (Figure S7D), we conclude that HP1a acts downstream of H3K9 methylation in regulating exclusion of Rad51 foci. Enrichment of the Drosophila Smc5/6 Complex in Heterochromatin Requires HP1a HP1a could counteract the formation of Rad51 foci directly or by recruiting other proteins that provide antirecombinase functions. One good candidate for this role is the Smc5/6 complex, which prevents the formation of DSB-induced Rad52 foci in the nucleolus and suppresses spontaneous recombination of rDNA repeats in S. cerevisiae (Torres-Rosell et al., 2007). In S. pombe, HU treatment promotes Smc5/6 recruitment to pericentric heterochromatin, suggesting a role in responding to the presence of stalled forks (Pebernard et al., 2008). We first determined whether subunits of the Drosophila Smc5/ 6 complex are localized to heterochromatin. Live imaging shows that GFP-tagged Smc5, Smc6, and Nse2 colocalize with mChHP1a through all interphase stages of the cell cycle (Figure 6A, Figure S8A, and data not shown). Similar to HP1a (Kellum et al., 1995), Smc5/6 components disassemble from heterochromatin during mitosis and reassemble in G1 (Figure S8A and data not shown). Finally, Smc5 and Smc6 coimmunoprecipitate with FLAG-HP1a (Figure 6B). These results identify the Smc5/6 complex as a new heterochromatin component in Drosophila that is physically associated with HP1a. Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 737
-20 min 5 min
C
B
20 min 180 min
Ctrl + Caffeine
HP1a
fold increase in HP1a volume
+ Caffeine
HP1a ATRIP HP1a
1.4 1.2 1.0 0.8 -50
0
50
100 150 200
time after IR (min)
C tr
+ a
C
C tIP
RNAi
RNAi
tIP
B
l lm
0
avg. # MU2 foci / cell
0.4
+
To s
ca Ct + rl B lm
C tr l A TM A TM AT + R A TR
1.0
*
0.8
To sc
1.1
50
100
150
+ Caffeine
6
Position of foci vs HP1a domain: IN Periphery (in) Periphery (out) OUT
4 2 0
50
100
150
Ctrl RNAi 6
***
l TM A TM AT + R A TR
1.2
1.2
C tr
**
0
F
DAPI-bright *** **
A
1.3
***
0
time after IR (min)
avg. # γH2Av foci / cell
1.4
***
2
0
ATRIP E
fold increase in HP1a volume
4
1.6
ATRIP
D
Ctrl
6
avg. # ATRIP foci / cell
Ctrl
HP1a ATRIP
+
A
4 2 0
Position of foci vs HP1a domain:
0
50
100
150
CtIP+Tosca+Blm RNAi 6
IN Periphery (in) Periphery (out) OUT
4 2 0
0
50
100
150
time after IR (min) Figure 4. Checkpoint Kinases and Resection Are Required for Heterochromatin Expansion and DSB Relocalization (A) Caffeine treatment suppresses HP1a dynamics and prevents ATRIP foci assembly within the HP1a domain. Stills from Movie S6 are shown, in which cells expressing GFP-ATRIP and mCh-HP1a were imaged before and after IR, in the presence (+ caffeine) or absence (Ctrl) of caffeine. Scale bar, 1 mm. (B) Quantitation of the experiment described in (A) shows defective expansion of the HP1a volume after caffeine treatment compared to untreated cells (p < 0.0001; n > 15). (C) Quantitation of the experiment described in (A) highlights defective formation of ATRIP foci in the HP1a domain after caffeine treatment (p < 0.001; n > 30). ATM is not required for formation of ATRIP foci in heterochromatin (Figure S6B), suggesting that this defect results from ATR inactivation. (D) Quantitation of HP1a volume shows defective expansion after depletion of checkpoint or resection proteins. Values = mean fold increase in HP1a volume for time points between 10 and 40 min after IR, compared to control RNAi (***p < 0.0001, **p < 0.01, two-tailed unpaired t test with Welch correction; n > 12). All time points are shown in Figure S6E. (E) Depletion of checkpoint or resection proteins results in persistent gH2Av foci in DAPI-bright at 60 min after IR (***p < 0.0001, **p < 0.01, *p < 0.05, two-tailed Mann-Whitney test; n > 100). Kinetics are shown in Figure S6F. (F) RNAi of CtIP+Tosca+Blm results in persistent Mu2 foci within the HP1a domain after IR and defective relocalization to the HP1a periphery (p < 0.0001; n > 20). This analysis could not be performed after ATR depletion because Mu2 foci do not form (Figure S6B). Similarly, ATRIP foci could not be monitored after RNAi of ATR or resection components (Figures S6B–S6D). Graphs show mean ± SEM. See also Figure S6 and Movie S6.
RNAi depletions were used to determine the interdependency of Smc5/6 and HP1a recruitment to heterochromatin. GFP-Nse2 colocalizes with intense HOECHST staining (which identifies the equivalent of DAPI-bright in living cells) in control cells, but not after HP1a RNAi (Figure 6C). Conversely, depletion of Smc5/6 has no impact on HP1a localization to heterochromatin or on the levels and nuclear distributions of H3K9me2 and H3K9me3 (Figure S8C). Similarly, Smc5/6 RNAi did not increase the DAPI-bright volume, a phenotype observed after depletion of Su(var)3–9, Su(var)3–9 + SetDB1, or HP1a (Figure S8D). 738 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
We conclude that the Drosophila Smc5/6 complex is a heterochromatin component, and its recruitment to heterochromatin is HP1a dependent. Smc5 and Smc6 Are Required to Prevent Formation of Rad51 Foci and Aberrant Recombination Events in Heterochromatin We next determined whether loss of the Smc5/6 complex affects the behavior of DSBs in heterochromatin. RNAi depletion of Smc5, Smc6, or Smc5+Smc6 significantly increased Rad51
HP1a
DAPI Rad51 Rad51
Ctrl
RNAi SetDB1 Su(var)3-9 Su(var)3-9
DAPI-weak DAPI-bright
avg. # Rad51 foci / cell
A
4 3
avg. # foci / cell
DAPI-weak
0.3
2
0.2
1
0.1
0 RNAi
B
0.4
0 Ctrl Su(var)3-9 Su(var)3-9 SetDB1
DAPI-bright
(A) IF analysis 60 min after IR and quantitation shows that Su(var)3–9, Su(var)3–9 + SetDB1, or HP1a depletion (quantitation is in B) results in increased Rad51 foci in DAPI-bright (dashed blue circles) and not in DAPI-weak (3- to 4-fold increase in DAPI-bright after Su(var)3–9 and SetDB1 RNAi [right]; p < 0.001; n > 100). (B) Quantitation of gH2Av and Rad51 foci at different time points after IR shows that depletion of HP1a results in increased foci in DAPI-bright (*p < 0.01; n > 100) and not in DAPI-weak. Graphs show mean ± SEM. Scale bars, 1 mm. See also Figure S7.
0.4 3
0.3
2 1 0
Ctrl RNAi HP1a RNAi
0.2 0.1
Rad51 0
100
200
0
0
100
200
time after IR (min.) avg. # foci / cell
Figure 5. HP1a or H3K9 HMTase Depletion Results in Formation of Rad51 Foci in Heterochromatin
10
0.8
8
0.6
6 2 0
Ctrl RNAi HP1a RNAi
0.4
4
γH2Av 0
100
200
0.2 0
0
100
200
time after IR (min.)
foci formation only in DAPI-bright (Figure 6D and Figure S8E), similar to the phenotypes observed after Su(var)3–9 or HP1a depletion (Figure 5). The number of Rad51 foci that were observed in DAPI-bright after simultaneous depletion of Smc5, Smc6, and HP1a by RNAi is comparable to HP1a depletion alone (Figure 6D), suggesting that they control HR repair in heterochromatin through the same pathway. Similar to HP1a RNAi, the kinetics of heterochromatic gH2Av and Rad51 foci become equivalent to euchromatin after Smc5/6 depletion (compare Figures S8F and S8G with Figure 5B). Because Smc5/6 RNAi does not affect levels and distributions of H3K9me2, me3, and HP1a, as well as compaction of the DAPI-bright domain, we conclude that Smc5/6 acts downstream of HP1a in controlling Rad51 exclusion from the heterochromatin domain. Finally, RNAi depletion of Smc5/6 complex components results in the appearance of extended DNA filaments between 50% of nuclei (cells do not complete cytokinesis) (Figure 6E and Figure S8E). Similar phenotypes were observed after HP1a or Su(var)3–9 depletion (data not shown). These DNA filaments are enriched in H3K9me2 (Figure 6E and Figure S8E), indicating that they contain heterochromatic DNA. Importantly, the filament phenotype is rescued by blocking HR (simultaneous depletion of
Smc5/6, Rad51, and/or Rad54) (Figure 6F), suggesting that they arise from aberrant recombination events. We conclude that the presence of the Smc5/6 in heterochromatin precludes the assembly of Rad51 foci at early time points after damage. Removal of the Smc5/6 complex or of any of the heterochromatin components responsible for its recruitment results in Rad51 foci formation inside of heterochromatin and failure to properly resolve HR events. DISCUSSION
Our findings reveal that DSBs are formed in Drosophila heterochromatin after IR and are rapidly recognized and processed for repair by HR. Furthermore, both the heterochromatin domain and heterochromatic repair foci display dramatic, dynamic behaviors in response to IR. Within minutes, gH2Av, Mu2, TopBP1, and ATRIP foci are assembled in heterochromatin, and the HP1a domain globally expands and forms dynamic protrusions. Later, repair foci relocate to the HP1a periphery or outside of the domain. Importantly, we directly show that damaged repeated sequences relocalize, not just DSBs within the HP1a domain. We also demonstrate that resection and checkpoint kinases, predominantly ATR, play crucial roles in both heterochromatin expansion and foci relocalization. Despite the HR requirement for repairing heterochromatin, Rad51 foci only form at the HP1a periphery or outside of the domain, where they associate with the HP1a dynamic protrusions and regions where HP1a is locally absent. These results demonstrate that the initial recognition and resection of heterochromatic DSBs are spatially and temporally separated from Rad51 foci assembly. Heterochromatin components are crucial for this regulation; the Smc5/6 complex, recruited to heterochromatin by HP1a, is the key component that is required to prevent formation of IR-induced Rad51 foci in heterochromatin. Smc5/6 depletion results in recombination-dependent heterochromatic DNA Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 739
Smc6
Nse2
B
WCE
GFP Flag-HP1a HP1a HP1a Smc6 GFP HP1a DAPI-bright 0.4
4 3
0.3
2
0.2
1
0.1
0
60
240
0
60
240
time after IR (min.) Ctrl HP1a
Smc5/6RNAi
E
DAPI H3K9me2 H3K9me2
avg. # Rad51 foci / cell
DAPI-weak
RNAi Smc5/6 HP1a + Smc5/6
HOECHST Nse2
- + - + Ctrl RNAi HP1a RNAi
Smc5
DAPI
D
C
IP
F
60 48.6
45 30 15 0
0
0
0
2.7 1.8 2.1
Ct Sm rl c5 Ra /6 d5 Sm R 1 a c5 d5 Sm Sm /6 R 4 c5 c5 ad /6 /6 51 Ra R a d5 d5 1R 4 ad 54
Smc5
% of nuclei connected by DNA threads
A
RNAi
Figure 6. The Smc5/6 Complex Is Required to Prevent IR-Induced Rad51 Foci and Aberrant Recombination in Heterochromatin (A) Images of cells expressing GFP-tagged Smc5, Smc6, or Nse2 and mCh-HP1a show that different Smc5/6 subunits colocalize with HP1a. (B) The Smc5/6 complex interacts with HP1a. FLAG-HP1a (+) expressing S2 cells were used to immunoprecipitate HP1a. Western blot with anti-FLAG (HP1a), Smc5, and Smc6 antibodies shows enrichment of Smc5 and Smc6 compared to control cells (). Extracts were treated with Benzonase prior to IP, suggesting that Smc5/6-HP1a interaction is not mediated by DNA. (C) HP1a is required for localization of the Smc5/6 complex to heterochromatin. HP1a RNAi results in delocalization of GFP-Nse2 from heterochromatin (see Figure S8B for quantitation). (D) Smc5/6 and HP1a are part of the same pathway that excludes Rad51 foci from heterochromatin. Bw (control), HP1a, Smc5/6, or HP1a + Smc5/6 were depleted for 5 days by RNAi, and then cells were fixed after IR. Quantitation shows similar increases in DAPI-bright Rad51 foci after HP1a or Smc5/6 + HP1a RNAi (n > 150). Graphs show mean ± SEM. (E) Smc5/6 depletion alone (without IR) results in heterochromatic DNA filaments connecting dividing cells. Smc5 and Smc6 were depleted for 6.5 days by RNAi, spun down, mixed, and settled to adhere to the slide. Then, cells were fixed and stained with DAPI and H3K9me2 antibodies. Images are maximum intensity projections. The brightness of the DAPI-only image was uniformly increased to visualize the filaments. Similar results were observed after separate depletion of Smc5 or Smc6 (Figure S8E). (F) DNA filaments formed in the absence of Smc5/6 are dependent on recombination. Bw (control), Smc5/6, Rad51, and/or Rad54 were depleted and cells were processed as in (E). DNA filaments resulting from Smc5/6 RNAi are suppressed by simultaneous depletion of Rad51 and/or Rad54 (p < 0.0001; n > 500). Scale bars, 5 mm. See also Figure S8.
filaments between nuclei, revealing the importance of the Smc5/ 6 complex in suppressing abnormal HR in heterochromatin. A Model for How HR Repair Is Controlled in Heterochromatin to Prevent Abnormal Recombination of Repeats Our results suggest a model for how HR repair of damaged heterochromatic repeats could occur without generating genome instability (Figure 7 and Movie S7). Heterochromatic DSBs are quickly detected by early components of the DNA damage response, resulting in resection and checkpoint signaling, whereas recruitment of Rad51 is initially suppressed by HP1a through its association with the Smc5/6 complex. Expansion of the HP1a domain, triggered by resection and ATR activity, facilitates the relocalization of DSBs to the heterochromatin periphery. Next, the local absence or removal of HP1a 740 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
at the periphery of the heterochromatin domain allows recruitment of Rad51 and other proteins required for recombination, which may also participate in tethering DSBs outside of the heterochromatin domain until repair is complete. We propose that the dynamic spatial and temporal regulation of the DSB response in heterochromatin helps to ensure repeat and genome stability. Separation of the repeats with DSBs from the rest of the (undamaged) heterochromatin before strand invasion would preclude the use of sequences on nonhomologous chromosomes as templates for repair, thus preventing chromosome rearrangements. Pairing of sister chromatids and homologous chromosomes is maintained throughout interphase in Drosophila cells (McKee, 2004). Thus, genome stability could be ensured by completing recombination repair using homologous sequences on sister chromatids or homologs that relocalize in concert with the damaged DNA.
0-20 min
20-40 min
checkpoint activation resection
> 40 min strand invasion
IR Heterochromatin expansion DSB relocalization
HP1a
γH2Av-Mu2
TopBP1-ATRIP
Rad51
HP1a - Smc5/6
ectopic recombination
HP1a absent at DSBs HR repair
Figure 7. A Model for the Spatiotemporal Regulation of DSB Repair Events in Heterochromatin DSB detection, checkpoint activation, and resection occur quickly in heterochromatin, where ATR and HP1a facilitate the assembly of TopBP1/ATRIP foci. Meanwhile, Rad51 recruitment and strand invasion are suppressed by the Smc5/6 complex, which is recruited to heterochromatin by HP1a (yellow background). Heterochromatin expansion facilitates DSB relocalization to the HP1a periphery; both dynamic behaviors require checkpoint kinases (mainly ATR) and resection proteins. The euchromatic environment or local disassembly of HP1a at DSBs allows Rad51 loading and completion of HR repair (blue background). Once initiated, strand invasion facilitates the retention of DSBs at the HP1a periphery or the euchromatic domain during HP1a contraction. This mechanism prevents ectopic recombination by isolating the damaged site from the undamaged heterochromatic repeats before strand invasion. Homolog or chromatid pairing, maintained during relocalization, likely provide the templates for completing HR repair. Only processes related to DSB repair in heterochromatin are shown. See Movie S7 for animated version.
Aspects of this model share striking similarities with mechanisms described in S. cerevisiae to preserve rDNA stability, where the Smc5/6 complex plays a role in suppressing formation of Rad51 foci until DSBs relocalize to outside nucleoli (Torres-Rosell et al., 2007). This similarity is surprising because S. cerevisiae lacks H3K9 methylation and HP1 proteins, which in Drosophila are critical components of DSB repair in heterochromatin and are not enriched in nucleoli. In addition, nucleolar rDNAs are the most actively expressed genes in the genome, whereas Drosophila heterochromatin is mostly transcriptionally silent and contains a wider spectrum of repeats (simple, short satellites and transposable elements). Our work extends the insightful yeast results by demonstrating that the spatiotemporal separation of HR repair events also occurs in the heterochromatin domain and that the response to IR includes physical expansion of HP1a. Moreover, we demonstrate that relocalization is coordinated with heterochromatin expansion and is orchestrated by checkpoint and resection activities, as well as heterochromatin components. DSB Processing Is Highly Efficient in Heterochromatin Contrary to previous conclusions (Cowell et al., 2007; Goodarzi et al., 2008; Kim et al., 2007), we show that heterochromatin does not represent an obstacle to DSB recognition and processing. In fact, Mu2 and gH2Av foci are quickly assembled in heterochromatin, with kinetics similar to euchromatin. Unexpectedly, TopBP1 and ATRIP foci appear even faster in heterochromatin than in euchromatin and are unusually bright compared to euchromatic foci. These observations suggest that heterochromatin features may amplify some aspects of the DSB response, such as resection, or formation or clustering of TopBP1/ATRIP foci. TopBP1/ATRIP foci formation could be enhanced by the interaction that we identified between HP1a and ATRIP, which is increased after IR. Checkpoint kinases
could also play a role, as caffeine treatment abrogates ATRIP foci specifically in heterochromatin. Because resection is required for ATRIP-ATR loading and, together with ATR, for heterochromatin expansion and DSB relocalization, the efficiency of resection and/or ATRIP assembly is likely to play a pivotal role in promoting heterochromatin stability in the presence of DSBs. One advantage of efficient resection in heterochromatin might be to suppress NHEJ and promote HR repair (Huertas, 2010). Because NHEJ can generate translocations from DSBs in close proximity (Soutoglou and Misteli, 2008), safe repair of DSBs in repeats could require both immediate repression of NHEJ plus spatial and temporal separation of early and later steps in HR repair. What Is the Significance of Heterochromatin Expansion? We discovered that IR rapidly induces a global change in heterochromatin organization in Drosophila (even at low dose), visualized as dynamic HP1a protrusions, and expansion of both HP1a and DAPI-bright domains. This response may be conserved, as increased size of the HP1b domain was observed after laser microirradiation in mouse cells. However, it is unclear whether the HP1b response reflects heterochromatin relaxation (Ayoub et al., 2008; Kruhlak et al., 2006) or HP1b recruitment to euchromatic DSBs next to heterochromatin (Luijsterburg et al., 2009). In contrast, we never observe HP1a recruitment to gH2Av foci in Drosophila euchromatin, and multiple cell biological and epigenomic analyses demonstrate that heterochromatin volume expands in response to IR in Drosophila cells. Previous studies suggested that heterochromatin relaxation is required to permit accessibility of repair components (Ayoub et al., 2008; Goodarzi et al., 2008). However, we observe that: (1) the peak of gH2Av, Mu2, ATRIP, and TopBP1 foci formation in heterochromatin occurs before the peak of heterochromatin expansion, (2) depletion of proteins required for expansion Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 741
(e.g., resection components) does not interfere with the formation of Mu2 and gH2Av foci in heterochromatin, and (3) removal of Smc5/6, which does not interfere with heterochromatin compaction or the presence of heterochromatin marks (H3K9me2/3 and HP1a), permits Rad51 foci formation in heterochromatin. Nevertheless, the fact that heterochromatin expansion requires resection and checkpoint kinases, indicating that it is specifically induced by DSBs, suggests that expansion does play a role in the DSB response. Importantly, the peak of heterochromatin expansion coincides with the timing of DSB relocalization, and conditions that suppress expansion also prevent DSB relocalization. Thus, we propose that heterochromatin expansion may be required for mobilizing DSBs, rather than accessibility of repair components (see below). What Regulates Relocalization of Foci and the Progression of HR Repair? One of the most striking observations reported here is the movement of DSBs and repair foci to outside of the heterochromatin domain. Such directional, long-range DSB movements were previously considered a peculiarity of S. cerevisiae (Lisby et al., 2003), in contrast to the positional stability of repair foci in mammalian cells (Kruhlak et al., 2006; Soutoglou et al., 2007). An exception is that unprotected telomeres, which resemble DSBs, become mobilized in mouse cells (Dimitrova et al., 2008). Telomeres are enriched for repeats and HP1, suggesting the intriguing possibility that relocalization of DSBs is conserved in mammals, but only for regions of the genome where proximity of repeated sequences presents a dangerous environment for repair. Relocalization of DSBs to the heterochromatin periphery would also explain the absence of gH2Ax foci in chromocenters of mouse cells at > 30 min after IR (Cowell et al., 2007). Detailed studies of the kinetics of the DSB response and mobility of foci in mammalian heterochromatin are required to determine whether the dynamics that we observe in Drosophila are conserved in vertebrates. Importantly, we show that relocalization of foci requires the presence of checkpoint and resection components. It is currently unknown whether they affect relocalization directly or indirectly through regulation of expansion. Further investigations are also required to determine the role of interactions between heterochromatin proteins and resection/checkpoint signaling (e.g., HP1a and ATRIP) in promoting DSB relocalization. In terms of mechanisms, relocalization within nuclei could be an ‘‘active’’ directed process that is mediated by cytoskeletal elements and motors, which regulate other types of chromosome dynamics (Chuang et al., 2006; Sato et al., 2009). The movement of Rad51 foci at the tips of the HP1a protrusions suggest that ‘‘pulling’’ forces could be present that help to isolate DSBs from the main heterochromatin domain. However, inhibition of actin or tubulin polymerization (by latrunculin A or colchicine treatment, respectively) did not affect relocalization of heterochromatic foci (data not shown); additional studies are required to definitively determine whether cytoskeletal elements or motors are required. Alternatively, relocalization of foci could be regulated by ‘‘passive’’ mechanisms. The initial compaction of the domain could limit the motion of DSBs, with expansion allowing increased mobility. One possibility, based on the persis742 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
tence of DSBs in heterochromatin after Rad51 or Rad54 depletion, is that relocalization is ensured by restricting strand invasion to outside of the heterochromatin domain, where active Rad51 and associated proteins ‘‘capture’’ resected DSBs. Interestingly, the human Ku70/80 complex is required to constrain the local motion of DSBs (Soutoglou et al., 2007). GFP-Ku80 is mostly absent in the HP1a compartment in Drosophila (data not shown), suggesting that suppression of NHEJ repair may contribute to the mobility of DSBs. These models are not mutually exclusive; for example, mobility could be triggered by active processes even if relocation involves a later capture mechanism. Heterochromatin proteins, and particularly the Smc5/6 complex, could play different roles in relocalization and the progression of HR repair. Studies in yeast demonstrate that the Smc5/6 complex not only suppresses illegitimate HR, but also facilitates sister chromatid exchanges (SCEs) (De Piccoli et al., 2009). Thus, Drosophila Smc5/6 could promote SCEs at heterochromatic DSBs during completion of HR repair, in addition to its earlier role in exclusion of Rad51 foci. How the Smc5/6 complex directly or indirectly controls HR is unknown. The Smc5/6 complex has an SMC structure, like cohesins and condensins, and the Nse2 subunit is an E3 SUMO-ligase (De Piccoli et al., 2009). An important challenge will be to determine whether the Smc5/6 complex controls the stability of Drosophila heterochromatic sequences by SUMOylating relevant targets of the recombination pathway or by recruiting other proteins, and whether relocalization of DSBs relies on the unique proprieties of SMC complexes, such as the ATP-dependent ability to link DNA molecules and interact with subnuclear structures. Consequences of Misregulation of HR Repair in Heterochromatin We previously showed that loss of HP1a or Su(var)3–9 results in excision of repeats and chromosome translocations in flies (Peng and Karpen, 2007, 2009). The results presented here suggest that heterochromatin components ensure genome stability by preventing aberrant recombination at DSBs after IR and that the Smc5/6 complex mediates this function. In addition, Smc5/6 removal without IR results in dramatic accumulation of abnormal, recombination-dependent heterochromatic filaments between dividing nuclei. These connections likely arise from deregulation of repair followed by chromosome segregation in the presence of unresolved recombination intermediates. Thus, heterochromatin components could also help to maintain genome stability during normal replication, when spontaneous damage can result from passage of replication forks through highly repeated DNA (Branzei and Foiani, 2010). Because the Smc5/6 complex is a constitutive heterochromatin component in Drosophila, we suggest that it is a key contributor to ensuring the stability of repeats after both spontaneous and induced damage. We propose that other examples of heterochromatic ‘‘threads’’ in human cells (Baumann et al., 2007) and Drosophila (Hughes et al., 2009; Royou et al., 2010) could arise from failures in control of heterochromatin HR repair or local, directed deregulation of the mechanisms described here.
Despite the challenges that they pose to genome stability, repeats are likely maintained in complex eukaryotes because they encode essential functions such as centromeres, meiotic pairing, and telomeres (Bernard et al., 2001; de Lange, 2005; Dernburg et al., 1996; Karpen et al., 1996). Our results suggest that heterochromatin components play an additional, major role in preventing genome and repeat instability through temporal and spatial regulation of HR repair. It will be important to determine whether the dynamic behaviors and regulatory roles of heterochromatin described here are conserved in vertebrates and whether defects in these processes contribute to human diseases associated with chromosome instability and aberrant heterochromatin, such as cancer (Dialynas et al., 2008) and aging (Shin et al., 2010).
ACKNOWLEDGMENTS This work was supported by postdoctoral fellowships from FIRC and Unesco to I.C., by Ruth Kirchstein NIH Postdoctoral Fellowship to S.U.C. (1F32GM086111), and by the Low Dose Radiation Research Program U.S. Department of Energy (DOE) (DE-AC02-05CH11231) to S.V.C. and G.H.K. We are grateful to P.-O. Mari for assistance with the NIR experiments and to S. Elgin, J. Kadonaga, B. Mellone, and D. Rio for reagents. We thank H.V. Le, R. Kunitake, and C. Pham for their help; A. Dernburg, E. Dunleavy, J. Swenson, and W. Zhang for comments on the manuscript; S. Langley for advice on ChIP-array data analysis; C-Y. Chen for software development; and the Karpen lab for sharing reagents and helpful discussions. Received: June 3, 2010 Revised: December 15, 2010 Accepted: February 7, 2011 Published online: February 24, 2011
EXPERIMENTAL PROCEDURES REFERENCES Cell Culture, IR Treatments Kc167 (Kc) cells were used for all experiments unless indicated otherwise and were maintained as logarithmically growing cultures. In IR experiments, the culture was exposed to 5Gy using a 320 kV x-ray source. In kinetics with fixed cells, time 0 (Unt) corresponds to cells fixed without exposing them to IR. In time-lapse experiments, time 0 (Unt) corresponds to cells imaged 5–10 min before IR treatment, unless otherwise indicated. Cytological Methods, Time-Lapse Experiments, and Image Processing IF, FISH, and TUNEL assays were described in Peng and Karpen (2007, 2009). Classification of foci inside or outside of the heterochromatin domain was done by analyzing their positions with respect to DAPI-bright or the HP1a domain in each of the Z stacks and in 3D image reconstructions with softWoRx. In time-lapse experiments, the same field of cells was imaged before and after IR, as described in Extended Experimental Procedures. Deltavision movie files were processed with Imaris (Bitplane AG, Zurich, Switzerland) for 3D rendering and compensation of movement, when necessary. Movies were created in Imaris, and time points were added using Photoshop (Adobe). 3D modeling (Figure 3C, Figure S5A, Movie S2, and Movie S3) was done with Imaris. Quantitation of HP1a Expansion 3D volumes occupied by HP1a were analyzed for consecutive time points in time-lapse images of nuclei. Quantitation was done with custom Matlab scripts, as described in Extended Experimental Procedures. In the graphs, the volume quantified for each cell was normalized to the value (or mean value) of the same cell before IR. ChIP-Chip ChIP samples were prepared as described in Extended Experimental Procedures, and data were analyzed using the Affymetrix Tiling Analysis Software. Browser shots were generated with the Affymetrix Integrated Genome Browser. Statistical Analysis Unless otherwise indicated, n = number of cells/sample, and p values were calculated using the Kruskal-Wallis test with multiple-comparison Dunnet’s post test. Analyses were done with InStat (Graphpad).
SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, eight figures, and seven movies and can be found with this article online at doi:10.1016/j.cell.2011.02.012.
Ayoub, N., Jeyasekharan, A.D., Bernal, J.A., and Venkitaraman, A.R. (2008). HP1-beta mobilization promotes chromatin changes that initiate the DNA damage response. Nature 453, 682–686. Baumann, C., Ko¨rner, R., Hofmann, K., and Nigg, E.A. (2007). PICH, a centromere-associated SNF2 family ATPase, is regulated by Plk1 and required for the spindle checkpoint. Cell 128, 101–114. Bernard, P., Maure, J.F., Partridge, J.F., Genier, S., Javerzat, J.P., and Allshire, R.C. (2001). Requirement of heterochromatin for cohesion at centromeres. Science 294, 2539–2542. Branzei, D., and Foiani, M. (2010). Maintaining genome stability at the replication fork. Nat. Rev. Mol. Cell Biol. 11, 208–219. Chuang, C.-H., Carpenter, A.E., Fuchsova, B., Johnson, T., de Lanerolle, P., and Belmont, A.S. (2006). Long-range directional movement of an interphase chromosome site. Curr. Biol. 16, 825–831. Costes, S.V., Chiolo, I., Pluth, J.M., Barcellos-Hoff, M.H., and Jakob, B. (2009). Spatiotemporal characterization of ionizing radiation induced DNA damage foci and their relation to chromatin organization. Mutat. Res. 704, 78–87. Cowell, I.G., Sunter, N.J., Singh, P.B., Austin, C.A., Durkacz, B.W., and Tilby, M.J. (2007). gammaH2AX foci form preferentially in euchromatin after ionisingradiation. PLoS ONE 2, e1057. de Lange, T. (2005). Shelterin: the protein complex that shapes and safeguards human telomeres. Genes Dev. 19, 2100–2110. De Piccoli, G., Torres-Rosell, J., and Arago´n, L. (2009). The unnamed complex: what do we know about Smc5-Smc6? Chromosome Res. 17, 251–263. Dernburg, A.F., Sedat, J.W., and Hawley, R.S. (1996). Direct evidence of a role for heterochromatin in meiotic chromosome segregation. Cell 86, 135–146. Dialynas, G.K., Vitalini, M.W., and Wallrath, L.L. (2008). Linking Heterochromatin Protein 1 (HP1) to cancer progression. Mutat. Res. 647, 13–20. Digilio, F.A., Pannuti, A., Lucchesi, J.C., Furia, M., and Polito, L.C. (1996). Tosca: a Drosophila gene encoding a nuclease specifically expressed in the female germline. Dev. Biol. 178, 90–100. Dimitrova, N., Chen, Y.C., Spector, D.L., and de Lange, T. (2008). 53BP1 promotes non-homologous end joining of telomeres by increasing chromatin mobility. Nature 456, 524–528. Downs, J.A., Nussenzweig, M.C., and Nussenzweig, A. (2007). Chromatin dynamics and the preservation of genetic information. Nature 447, 951–958. Dronamraju, R., and Mason, J.M. (2009). Recognition of double strand breaks by a mutator protein (MU2) in Drosophila melanogaster. PLoS Genet. 5, e1000473. Eissenberg, J.C., and Reuter, G. (2009). Cellular mechanism for targeting heterochromatin formation in Drosophila. Int. Rev. Cell Mol. Biol. 273, 1–47.
Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc. 743
Goodarzi, A.A., Noon, A.T., Deckbar, D., Ziv, Y., Shiloh, Y., Lo¨brich, M., and Jeggo, P.A. (2008). ATM signaling facilitates repair of DNA double-strand breaks associated with heterochromatin. Mol. Cell 31, 167–177.
Pebernard, S., Schaffer, L., Campbell, D., Head, S.R., and Boddy, M.N. (2008). Localization of Smc5/6 to centromeres and telomeres requires heterochromatin and SUMO, respectively. EMBO J. 27, 3011–3023.
Grewal, S.I., and Jia, S. (2007). Heterochromatin revisited. Nat. Rev. Genet. 8, 35–46.
Peng, J.C., and Karpen, G.H. (2007). H3K9 methylation and RNA interference regulate nucleolar organization and repeated DNA stability. Nat. Cell Biol. 9, 25–35.
Hoskins, R.A., Carlson, J.W., Kennedy, C., Acevedo, D., Evans-Holm, M., Frise, E., Wan, K.H., Park, S., Mendez-Lago, M., Rossi, F., et al. (2007). Sequence finishing and mapping of Drosophila melanogaster heterochromatin. Science 316, 1625–1628. Huertas, P. (2010). DNA resection in eukaryotes: deciding how to fix the break. Nat. Struct. Mol. Biol. 17, 11–16. Hughes, S.E., Gilliland, W.D., Cotitta, J.L., Takeo, S., Collins, K.A., and Hawley, R.S. (2009). Heterochromatic threads connect oscillating chromosomes during prometaphase I in Drosophila oocytes. PLoS Genet. 5, e1000348. Karpen, G.H., Le, M.H., and Le, H. (1996). Centric heterochromatin and the efficiency of achiasmate disjunction in Drosophila female meiosis. Science 273, 118–122. Kellum, R., Raff, J.W., and Alberts, B.M. (1995). Heterochromatin protein 1 distribution during development and during the cell cycle in Drosophila embryos. J. Cell Sci. 108, 1407–1418. Kim, J.-A., Kruhlak, M., Dotiwala, F., Nussenzweig, A., and Haber, J.E. (2007). Heterochromatin is refractory to gamma-H2AX modification in yeast and mammals. J. Cell Biol. 178, 209–218. Kruhlak, M.J., Celeste, A., Dellaire, G., Fernandez-Capetillo, O., Mu¨ller, W.G., McNally, J.G., Bazett-Jones, D.P., and Nussenzweig, A. (2006). Changes in chromatin structure and mobility in living cells at sites of DNA double-strand breaks. J. Cell Biol. 172, 823–834. Kusano, K., Berres, M.E., and Engels, W.R. (1999). Evolution of the RECQ family of helicases: A drosophila homolog, Dmblm, is similar to the human bloom syndrome gene. Genetics 151, 1027–1039. Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al; International Human Genome Sequencing Consortium. (2001). Initial sequencing and analysis of the human genome. Nature 409, 860–921.
Peng, J.C., and Karpen, G.H. (2008). Epigenetic regulation of heterochromatic DNA stability. Curr. Opin. Genet. Dev. 18, 204–211. Peng, J.C., and Karpen, G.H. (2009). Heterochromatic genome stability requires regulators of histone H3 K9 methylation. PLoS Genet. 5, e1000435. Royou, A., Gagou, M.E., Karess, R., and Sullivan, W. (2010). BubR1- and Polo-coated DNA tethers facilitate poleward segregation of acentric chromatids. Cell 140, 235–245. Sarkaria, J.N., Busby, E.C., Tibbetts, R.S., Roos, P., Taya, Y., Karnitz, L.M., and Abraham, R.T. (1999). Inhibition of ATM and ATR kinase activities by the radiosensitizing agent, caffeine. Cancer Res. 59, 4375–4382. Sato, A., Isaac, B., Phillips, C.M., Rillo, R., Carlton, P.M., Wynne, D.J., Kasad, R.A., and Dernburg, A.F. (2009). Cytoskeletal forces span the nuclear envelope to coordinate meiotic chromosome pairing and synapsis. Cell 139, 907–919. Shin, D.M., Kucia, M., and Ratajczak, M.Z. (2010). Nuclear and chromatin reorganization during cell senescence and aging - a mini-review. Gerontology 57, 76–84. Soutoglou, E., and Misteli, T. (2008). On the contribution of spatial genome organization to cancerous chromosome translocations. J. Natl. Cancer Inst. Monogr. 2008, 16–19. Soutoglou, E., Dorn, J.F., Sengupta, K., Jasin, M., Nussenzweig, A., Ried, T., Danuser, G., and Misteli, T. (2007). Positional stability of single double-strand breaks in mammalian cells. Nat. Cell Biol. 9, 675–682. Su, T.T. (2006). Cellular responses to DNA damage: one signal, multiple choices. Annu. Rev. Genet. 40, 187–208. Sugiyama, T., and Kowalczykowski, S.C. (2002). Rad52 protein associates with replication protein A (RPA)-single-stranded DNA to accelerate Rad51mediated displacement of RPA and presynaptic complex formation. J. Biol. Chem. 277, 31663–31672.
Lisby, M., Mortensen, U.H., and Rothstein, R. (2003). Colocalization of multiple DNA double-strand breaks at a single Rad52 repair centre. Nat. Cell Biol. 5, 572–577.
Torres-Rosell, J., Sunjevaric, I., De Piccoli, G., Sacher, M., Eckert-Boulet, N., Reid, R., Jentsch, S., Rothstein, R., Arago´n, L., and Lisby, M. (2007). The Smc5-Smc6 complex and SUMO modification of Rad52 regulates recombinational repair at the ribosomal gene locus. Nat. Cell Biol. 9, 923–931.
Luijsterburg, M.S., Dinant, C., Lans, H., Stap, J., Wiernasz, E., Lagerwerf, S., Warmerdam, D.O., Lindh, M., Brink, M.C., Dobrucki, J.W., et al. (2009). Heterochromatin protein 1 is recruited to various types of DNA damage. J. Cell Biol. 185, 577–586.
Uanschou, C., Siwiec, T., Pedrosa-Harand, A., Kerzendorfer, C., SanchezMoran, E., Novatchkova, M., Akimcheva, S., Woglar, A., Klein, F., and Schlo¨gelhofer, P. (2007). A novel plant gene essential for meiosis is related to the human CtIP and the yeast COM1/SAE2 gene. EMBO J. 26, 5061–5070.
McKee, B.D. (2004). Homologous pairing and chromosome dynamics in meiosis and mitosis. Biochim. Biophys. Acta 1677, 165–180.
Westphal, T., and Reuter, G. (2002). Recombinogenic effects of suppressors of position-effect variegation in Drosophila. Genetics 160, 609–621.
Pearson, C.E., Nichol Edamura, K., and Cleary, J.D. (2005). Repeat instability: mechanisms of dynamic mutations. Nat. Rev. Genet. 6, 729–742.
Zou, L., and Elledge, S.J. (2003). Sensing DNA damage through ATRIP recognition of RPA-ssDNA complexes. Science 300, 1542–1548.
744 Cell 144, 732–744, March 4, 2011 ª2011 Elsevier Inc.
Mpk1 MAPK Association with the Paf1 Complex Blocks Sen1-Mediated Premature Transcription Termination Ki-Young Kim1 and David E. Levin1,2,* 1Department
of Molecular and Cell Biology, Boston University Goldman School of Dental Medicine of Microbiology, Boston University School of Medicine Boston University, Boston, MA, 02118, USA *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.01.034 2Department
SUMMARY
The Mpk1 MAPK of the yeast cell wall integrity pathway uses a noncatalytic mechanism to activate transcription of stress-induced genes by recruitment of initiation factors to target promoters. We show here that Mpk1 additionally serves a function in transcription elongation that is also independent of its catalytic activity. This function is mediated by an interaction between Mpk1 and the Paf1 subunit of the Paf1C elongation complex. A mutation in Paf1 that blocks this interaction causes a specific defect in transcription elongation of an Mpk1-induced gene, which results from Sen1-dependent premature termination through a Nab3-binding site within the promoterproximal region of the gene. Our findings reveal a regulatory mechanism in which Mpk1 overcomes transcriptional attenuation by blocking recruitment of the Sen1-Nrd1-Nab3 termination complex to the elongating polymerase. Finally, we demonstrate that this mechanism is conserved in an interaction between the human ERK5 MAPK and human Paf1. INTRODUCTION The cell wall of the budding yeast S. cerevisiae is required to maintain cell shape and integrity (Klis, 1994). The cell must remodel this rigid structure during vegetative growth and during pheromone-induced morphogenesis. Wall remodeling is monitored and regulated by the cell wall integrity (CWI) signaling pathway, which activates a MAP kinase (MAPK) cascade (reviewed in Levin, 2005). The MAPK cascade is a linear pathway comprised of Pkc1, a MEKK (Bck1), a pair of redundant MEKs (Mkk1 and Mkk2), and a MAPK (Mpk1/Slt2). Mpk1 is a functional homolog of human ERK5 (Truman et al., 2006), a MAPK activated in response to growth factors as well as physical and chemical stresses (Abe et al., 1996; Yan et al., 2001). CWI signaling is induced in response to cell wall stressors, including elevated growth temperature, pheromone-induced morphogenesis, hypo-osmotic shock, and chemical cell wall
antagonists (Levin, 2005). This pathway activates two known transcription factors. One of these is Rlm1 (Dodou and Treisman, 1997; Watanabe et al., 1997), which is activated through phosphorylation by Mpk1 (Jung et al., 2002). The second is SBF (Madden et al., 1997; Baetz et al., 2001), a dimeric transcriptional regulator comprised of Swi4 and Swi6, which is essential for normal regulation of G1-specific transcription (reviewed in Breeden, 2003). Swi4 is the sequence-specific DNA-binding subunit, but Swi6 is required for binding to cell cycle-regulated promoters (Andrews and Moore, 1992; Sidorova and Breeden, 1993). Swi6 allows Swi4 to bind DNA at cell cycle-regulated promoters by relieving an autoinhibitory intramolecular association of the Swi4 C terminus with its DNA-binding domain. Additionally, Swi6 is the transcriptional activation component of SBF (Sedgwick et al., 1998). SBF drives gene expression in response to cell wall stress in a manner that is independent of its role in G1-specific transcription (Kim et al., 2008; Truman et al., 2009). Mpk1 and its pseudokinase paralog, Mlp1, which is also activated by the MEKs of the CWI signaling pathway, use a noncatalytic mechanism to activate SBF. Activated (phosphorylated) Mpk1, or Mlp1, forms a complex with Swi4 that associates with SBF-binding sites in the promoters of cell wall stress target genes independent of Swi6 (Kim et al., 2008; Kim and Levin, 2010). In this context, Mpk1 (or Mlp1) relieves the autoinhibitory Swi4 interaction by binding to a MAPK docking site on Swi4 (a D motif) adjacent to the C-terminal Swi6-binding site (Truman et al., 2009). Although Swi6 is not required for Swi4 to bind to the promoters of cell wall stress genes, it must be recruited to the Mpk1-Swi4 complex to activate transcription. Here, we show that Mpk1 also serves a noncatalytic function in transcription elongation through an association with the RNA polymerase II (Pol II)-associated complex (Paf1C). This complex, originally identified in yeast, is comprised of five subunits (Paf1, Cdc73, Rtf1, Ctr9, and Leo1; Mueller et al., 2004). It has been implicated in transcription start site selection (Stolinski et al., 1997) and elongation (Costa and Arndt, 2000; Betz et al., 2002; Kim et al., 2010b) and as a platform for the recruitment of histone methyltransferases (Krogan et al., 2003; Wood et al., 2003) and 30 end processing factors to the elongation complex (Mueller et al., 2004; Penheiter et al., 2005; Sheldon et al., 2005). The importance of this complex is underscored by the identification Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 745
P -579 to -233
ATG
N
M 3121 to 3370
1477 to 1770 Vector
Heat Stress
+
Mpk1-TAYF Mpk1-K54R
Mpk1
-
TGA
C 5459 to 5655
-
+
+
-
+
Input
Mpk1-HA ChIP
B
Vector Heat Stress
+
Mlp1
-
+
Input
+
Input
DYN1 P N M FKS2 C
+
Input
DYN1 P N FKS2 M C
Mlp1-HA ChIP
C
Vector Heat Stress
+
Vector Heat Stress
Swi6-HA ChIP
+
DYN1 P N M FKS2 C
Swi4
-
Swi4-HA ChIP
D
Figure 1. Mpk1 and Mlp1 Associate with the FKS2 Promoter and Coding Region, but Swi4 and Swi6 Bind Only to the Promoter
FKS2
A
Swi6
-
of various subunits of the human Paf1C as candidate oncogenes and tumor suppressor genes (Chaudhary et al., 2007). Despite this, little is known about the regulation of the Paf1C or of mechanisms governing gene expression through this complex. The Paf1C associates with the open reading frames of all Pol II-transcribed genes examined, consistent with its binding to Pol II (Pokholok et al., 2002; Krogan et al., 2002; Jaehning, 2010). However, global gene expression analyses suggested that the Paf1C is important for the expression of fewer than 5% of yeast genes (Penheiter et al., 2005), including many involved in progression of the cell cycle (Koch et al., 1999; Porter et al., 2002) and some for cell wall biosynthesis (Chang et al., 1999). Consistent with the idea that the Paf1C is important for the expression of a subset of Pol II-transcribed genes, the yeast Paf1C components are not essential, but their loss results in hypersensitivity to a variety of stresses (Betz et al., 2002), implicating this complex in the expression of stress-responsive genes. Among the conditions to which mutants in the Paf1C are hypersensitive is cell wall stress, prompting the suggestion that the CWI pathway plays a regulatory role in Paf1C-mediated transcription (Chang et al., 1999). We demonstrate that the Mpk1-Paf1C association allows transcription elongation of the FKS2 gene by preventing recruitment of and premature termination by the Sen1-Nrd1-Nab3 complex. This complex is known to function specifically in the termination of short, nonpolyadenylated Pol II transcripts, including small nucleolar RNAs (snoRNAs), cryptic unstable transcripts (CUTs), and at a few known mRNA attenuation sites (Ursic et al., 1997; Rasmussen and Culbertson, 1998; Steinmetz et al., 2001, 2006; Wyers et al., 2005; Arigo et al., 2006). Thus, our findings reveal a MAPK-driven mechanism for suppressing transcriptional attenuation of a stress-induced gene. 746 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
DYN1 P N FKS2 M C
(A) (Top) Schematic of the FKS2 gene showing regions amplified for ChIP analyses along the length of the gene. P, promoter; N, N-terminal coding region; M, middle coding region; C, C-terminal coding region. (Bottom) Mpk1 ChIP on FKS2. HA-tagged forms of Mpk1, an activationdefective phosphorylation site mutant (Mpk1-TAYF), or a catalytically inactive form (Mpk1-K54R) were expressed from plasmids in an mpk1D mlp1D strain (DL3327). Cells were exposed to heat stress at 39 C for 2 hr or unstressed at RT. A region from the middle of the DYN1-coding region was used as a negative control. (B) Mlp1 ChIP on FKS2. HA-tagged Mlp1 was expressed from a plasmid in an mpk1D mlp1D strain (DL3327). (C) Swi4 ChIP on FKS2. HA-tagged Swi4 was expressed from a plasmid in a swi4D strain (DL3145). (D) Swi6 ChIP on FKS2. HA-tagged Swi6 was expressed from a plasmid in a swi6D strain (DL3148).
RESULTS We showed previously by chromatin immunoprecipitation (ChIP) analysis that activated Mpk1 MAPK binds to the FKS2 promoter independent of its protein kinase activity through association with the Swi4 transcription factor (Kim et al., 2008). Because other MAPKs, most notably Hog1, have been shown to associate not only with the promoters of target genes, but with the coding regions as well (Pokholok et al., 2006; Proft et al., 2006), we extended our analysis of Mpk1 to ask whether this MAPK binds to the FKS2 coding region. We found that Mpk1 associates with genomic regions along the length of the FKS2 gene in response to activation by cell wall stress induced by elevated growth temperature (Figure 1A). Moreover, the catalytically inactive Mpk1-K54R mutant form of Mpk1 behaved like wild-type in this assay. However, a mutant form of Mpk1 that cannot be phosphorylated by its upstream activating kinases (Mpk1TAYF) failed to associate with any part of the FKS2 gene, revealing the requirement for activating signal. We demonstrated additionally that the pseudokinase paralog of Mpk1, Mlp1, also associates with the FKS2 promoter (Kim et al., 2008). In fact, it was this finding that led us to the understanding that Mpk1driven FKS2 transcription does not require its protein kinase activity. Like Mpk1, activated Mlp1 also bound to the FKS2 coding region (Figure 1B). In contrast to the presence of Mpk1 and Mlp1 across the FKS2 gene, we found that the Swi4 and Swi6 transcription factors were present only at the promoter (Figures 1C and 1D), consistent with their established role in transcription initiation. This suggested a model in which Mpk1 and Mlp1 move from the transcription initiation complex to the transcription elongation complex, leaving Swi4-Swi6 behind. Mpk1 Associates with the Paf1C at the FKS2 Promoter Our starting assumption was that Mpk1 association with the elongation complex is integral to the mechanism by which this MAPK controls cell wall stress-induced transcription. Therefore,
A
U
HS
CR
Figure 2. Requirements for Association of Mpk1 with Paf1
CW
Mpk1-HA
IP: Flag
Paf1-Flag Mpk1-HA
Input
Paf1-Flag
B
Mpk1 Heat Stress Mpk1-HA
-
+
Mpk1-TAYF Mpk1-K54R
-
+
-
Vector
-
+
+ IP: Flag
Paf1-Flag Mpk1-HA
Input
Paf1-Flag
C
Wild-type Heat Stress
-
+
swi4Δ
swi6Δ
rlm1Δ
-
-
-
+
+
+
Mpk1-HA IP: Flag
(A) The Mpk1-Paf1 interaction is stimulated by cell wall stress. Paf1-Flag was tested for coimmunoprecipitation (IP) of Mpk1-HA activated by various cell wall stresses for 2 hr—U, unstressed (RT); HS, heat stress at 39 C; CR, 50 mg/ml Congo red; CW, 40 mg/ml calcofluor white—in an mpk1D mlp1D (DL3183) strain expressing differentially tagged Mpk1 and Paf1. (B) The Paf1-Mpk1 interaction requires activating signal to Mpk1, but not catalytic activity. Mpk1-HA, an activationdefective phosphorylation site mutant (Mpk1-TAYF-HA), or its catalytically inactive form (Mpk1-K54R-HA) was coexpressed with Paf1-Flag in an mpk1D mlp1D strain (DL3183). (C) The Paf1-Mpk1 association requires Swi4 and Swi6, but not Rlm1. Yeast strains were: wild-type (DL3187), swi4D (DL3405), swi6D (DL3233), and rlm1D (DL3586). (D) Neither Pol II nor Paf1 is recruited to the FKS2 gene in the absence of Swi6. Rpb3-HA and Paf1-HA ChIP on FKS2 in wild-type cells (DL3187) and a swi6D strain (DL3233). The ADH1 promoter was used as a reference control.
Paf1-Flag Mpk1-HA Paf1-Flag
Input
dispensable for Paf1 binding, as judged by the behavior of the catalytically inactive Mpk1Paf1-HA ChIP Rpb3-HA ChIP D K54R protein. SWI6 swi6Δ SWI6 swi6Δ In addition to an activating signal, formation of Input Heat Stress + + + + the Mpk1-Paf1 complex required the presence ADH1 P of Swi4 (Figure 2C). Thus, the association N occurs under conditions in which activated FKS2 M Mpk1 forms a complex with Swi4 on the proC moters of target genes. We interpret this to mean that the Mpk1-Paf1 interaction is restricted to genes regulated by Mpk1-Swi4. we wished to identify the factors required for Mpk1 to bind the However, Mpk1-Swi4 promoter binding was not sufficient to FKS2 coding region. As a first step, we tested individual deletion allow the Mpk1-Paf1 interaction, which also required the presmutants in nonessential genes that encode established or puta- ence of Swi6 (Figure 2C). Although Swi6 is not needed for tive transcription elongation factors for their effects on FKS2- Mpk1-Swi4 promoter occupancy, it is required for transcriptional lacZ reporter expression. From among 71 deletion mutants, we activation of FKS2 (Kim et al., 2008). Therefore, the Mpk1-Paf1 identified 25 that displayed impaired induction of FKS2-lacZ in association is dependent on the same factors required for response to heat stress (Table S1 available online). Tests of FKS2 transcription initiation in response to cell wall stress. two-hybrid interactions between Mpk1 and this subset of factors Consistent with this, the Mpk1-Paf1 interaction was independent led us to focus on the Paf1C. All five components of this complex of the Rlm1 transcription factor (Figure 2C), which controls cell (Paf1, Cdc73, Rtf1, Ctr9, and Leo1; Mueller et al., 2004) inter- wall stress transcription of other genes in response to Mpk1 actiacted with Mpk1 in a cell wall stress-dependent manner (Fig- vation (Jung and Levin, 1999). We noted additionally that, in the ure S1). In addition to the Paf1C subunits, we also detected absence of Swi6, neither Paf1 nor the Rpb3 Pol II subunit were stress-induced association of Mpk1 with the Gcn5 and Ada2 recruited stably to the FKS2 gene (Figure 2D), even though the subunits of the SAGA and ADA histone acetyltransferase (HAT) Mpk1-Swi4 complex resides at the promoter under these conditions (Kim et al., 2008). These results suggest that the complexes (see below). Because we were specifically interested in the mechanism by Pol II-Paf1C complex must also be recruited to Mpk1-Swi4which Mpk1 associates with the transcription elongation com- Swi6-regulated promoters for the MAPK to bind the Paf1C. In contrast to the above results, the cell wall stress-induced plex, we next examined by coimmunoprecipitation the requirements for association of Mpk1 with Paf1. A variety of cell wall association of Mpk1 with either Gcn5 or Ada2 was only modestly stress treatments stimulated this association (Figure 2A), indi- reduced in a swi6D mutant (Figure S2), indicating that these cating a requirement for activating signal to Mpk1. In support interactions occur largely independent of Mpk1-Swi4-Swi6of this conclusion, we found that formation of the Mpk1-Paf1 driven transcription. However, both interactions were greatly complex was blocked in a nonphosphorylatable form of Mpk1 reduced in an rlm1D mutant, suggesting a more general role (Mpk1-TAYF; Figure 2B). However, Mpk1 catalytic activity was for HAT complexes in cell wall stress transcription. For this Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 747
A
Paf1-Gal4AD fusions
1
85
131
205
β-gal. Spec. 445 Act. (U)
319
Vector
C Cdc73 Heat Stress - + Mpk1-Flag
PAF1 Ctr9 Leo1 Rtf1 - + - + - +
10.2 ± 0.5 24.4 ± 4.5 20.4 ± 1.3 31.1 ± 4.9 31.1 ± 5.3 10.7 ± 2.1 18.0 ± 2.1 9.5 ± 1.5 10.8 ± 2.3 12.9 ± 1.0
B
Paf1-Flag Paf1-4AFlag
Heat Stress
-
+
-
+
Paf1
-
+
Mpk1-HA IP: Flag
Paf1-Flag Mpk1-HA
Input Paf1-Flag
D
LiCl
YPD
HS
Form.
CR
Caff.
paf1-4A Cdc73 Ctr9 - + - +
Leo1 - +
TF-ZZ-HA
-
Rtf1 +
IP: ZZ
CFW
MgCl2
TBZ
Van.
CHX
Hyg.
Mpk1-Flag TF-ZZ-HA
Input
Top: PAF1 Middle: paf1-4A Bottom: paf1Δ Figure 3. A Paf1 Mutant that Is Defective for Interaction with Mpk1 (A) Yeast two-hybrid interaction of Mpk1 with Paf1 fragments. An Mpk1-Gal4DBD fusion was tested for interaction with the indicated Paf1-Gal4AD fusion under conditions of cell wall stress. Endpoints of Paf1 fusions are indicated at top. Fusions that contained residues 85–131 (black) displayed b-galactosidase activities (U, units) higher than empty vector. (B) A Paf1 mutant that fails to bind Mpk1. A paf1D strain (DL3548) expressing the indicated forms of Mpk1 and Paf1 was treated as in Figure 2. (C) Mpk1 fails to associate with any Paf1C subunit in the paf1-4A mutant. A paf1D strain (DL3548) expressing Mpk1-Flag, the indicated TAP-tagged Paf1C subunit (TF-ZZ-HA), and either Paf1 or Paf1-4A were processed for coimmunoprecipitation of Mpk1 with the various Paf1C subunits. (D) The paf1-4A mutant is specifically hypersensitive to cell wall stress. A paf1D strain (DL3792), expressing PAF1 (top of each panel), paf1-4A (middle), or bearing empty vector (bottom) was subjected to the indicated stresses on YPD plates or unstressed (YPD). Stresses were: HS, heat stress (39 C); CR, Congo red (50 mg/ml); CFW, calcofluor white (20 mg/ml); TBZ, thiabendazole (50 mg/ml); CHX, cycloheximide (50 ng/ml); LiCl (50 mM); Form., formamide (3%); Caff., caffeine (8 mM); MgCl2 (750 mM); Van., sodium orthovanadate (3 mM); and Hyg., hygromycin (20 mg/ml). Equivalent cell numbers of each strain were spotted as serial 10-fold dilutions onto plates and incubated for 3 days at 30 C (or 39 C).
reason and because the Paf1C had been implicated previously in the maintenance of cell wall integrity (Chang et al., 1999), we focused our attention on this complex. Mpk1 Associates with the Paf1C through a Docking Motif in the Paf1 Subunit Both Mpk1 and Paf1 are part of larger complexes on the DNA, so their interactions might be indirect. However, we speculated that Mpk1 associates directly with Paf1 for two reasons. First, Swi4 and Swi6 do not associate with the FKS2 coding region, indicating that Mpk1 does not interact with the Paf1C indirectly through either of these factors. Second, among the two-hybrid associations detected between Mpk1 and various Paf1C subunits, the strongest was with Paf1 itself, which appears to be the major scaffolding subunit of the complex (Jaehning, 2010; Kim et al., 2010b). Therefore, we first tested a set of N-terminal and C-terminal truncations of Paf1 for two-hybrid 748 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
association with Mpk1. This analysis narrowed the region that is important for interaction to residues 86–131 (Figure 3A). MAPKs associate with their substrates and regulators through a canonical D motif, (Arg/Lys)1-2-(X)2-6-FA-x-FB (wherein F are hydrophobic residues Leu, Ile, or Val), recognized by a common docking site in the MAPK (Zhang et al., 2003). Because the Mpk1-Paf1 association did not require its catalytic activity, we reasoned that it might involve the Mpk1 docking site in the same manner that this MAPK associates with Swi4 (Truman et al., 2009). To explore this possibility, we next identified a single potential D motif in Paf1 within the stretch of sequence identified by the two-hybrid analysis (93-KDDRILLR-100) and constructed a mutant in which this site is ablated (residues 97–100 mutated to Ala; Paf1-4A). Consistent with the notion that this is a functional D motif, we found that the Paf1-4A protein failed to coimmunoprecipitate with Mpk1 in response to cell wall stress (Figure 3B).
To determine whether the observed associations between Mpk1 and the other Paf1C subunits were indirect through Paf1, we tested by coimmunoprecipitation the effect of the paf1-4A mutation on pairwise associations of Mpk1 with Paf1C subunits. Although Mpk1 associated in vivo with each of the Paf1C subunits, all of those associations were lost in the paf1-4A mutant (Figure 3C), suggesting that Mpk1 engages in a stable interaction with the Paf1C at a single point, which is ablated in the paf1-4A mutant. The paf1-4A Allele Is Defective for Transcription Elongation of FKS2 The Paf1-4A protein is generally functional, as judged by several criteria. First, the mutant subunit was detected in complex with each of the other Paf1C subunits and with the Rpb3 subunit of Pol II in vivo (Figure S3), indicating that the structural integrity of the Paf1C as well as its interaction with Pol II, which recruits Paf1C to the promoter (Jaehning, 2010), remain intact. Second, the phenotypes of the paf1-4A mutant revealed specific functional defects. Betz et al. (2002) demonstrated that a paf1D mutant is hypersensitive to growth inhibition by a wide variety of stress agents. Therefore, we compared the paf1-4A mutant to an isogenic paf1D mutant for growth inhibition by a panel of stressors (Figure 3D). Remarkably, the paf1-4A mutant behaved like the paf1D mutant only in response to cell wall stresses (i.e., heat stress, Congo red, and calcofluor white). By contrast, this mutant exhibited the wild-type level of resistance to other stresses to which the paf1D mutant was sensitive. The only exception to this rule was caffeine, to which the paf1-4A mutant was resistant despite the fact that caffeine is a cell wall stress agent (Martı´n et al., 2000; Jung et al., 2002). However, caffeine has many cellular targets, and we showed recently that, although it activates Mpk1, it also provokes additional Ser phosphorylation on this MAPK through the DNA damage checkpoint pathway (Truman et al., 2009). The consequence of this phosphorylation of Mpk1 (on Ser423) is that it blocks its association with Swi4, thereby preventing transcription through Swi4/Swi6. Therefore, the caffeine sensitivity of the paf1D mutant cannot be the consequence of loss of the noncatalytic Mpk1 transcription pathway, consistent with the finding that the paf1-4A mutant is not sensitive to this agent. The above results suggested that the paf1-4A mutant is specifically defective in gene expression that requires the Mpk1-Paf1 interaction. To test this directly, we compared FKS2 to CLN2 with respect to several parameters of gene expression in wild-type cells and in the paf1-4A and paf1D mutants. We chose CLN2 because its expression, which is periodic through the cell cycle, is strongly dependent on both the Paf1C (Koch et al., 1999) and Swi4/Swi6 (Breeden, 2003) but independent of Mpk1/Mlp1 (Igual et al., 1996; Kim et al., 2008). First, transcription of CLN2 in asynchronous cultures was blocked in a paf1D mutant, but CLN2 mRNA levels were unaffected by the paf1-4A mutation, as measured by RT-PCR (Figure 4A). Second, the Paf1-4A protein retained the ability to associate with the CLN2 promoter and the CLN2 coding region (Figure 4B). Third, the Pol II subunit, Rpb3, associated with the CLN2 promoter and the CLN2 coding region in the paf1-4A mutant but was not detected bound to the CLN2 gene at all in
the paf1D mutant (Figure 4C). By all of these measures, the paf1-4A mutant is functional for CLN2 transcription. In contrast to the above, heat stress-induced transcription of FKS2 was blocked in both the paf1-4A and paf1D mutants (Figure 4A), consistent with the observed sensitivity of both mutants to cell wall stress agents. Although heat stress induced recruitment of Pol II (Rpb3) to the FKS2 promoter in the paf1-4A mutant, the polymerase did not progress to the coding region (Figure 4D), indicating a defect in transcription elongation of this gene. This conclusion was supported by the further finding that the Paf1-4A mutant protein was recruited to the FKS2 promoter but did not progress to the FKS2 coding region (Figure 4E). Consistent with the absence of an elongation complex along the length of the FKS2 gene, Mpk1 that was recruited to the promoter also failed to progress to the coding region in either the paf1D or paf1-4A mutants (Figure 4F). Such a defect in elongation could be caused either by stalling of Pol II shortly after initiation or by premature termination of transcription, a point to which we will return below. As was observed for the CLN2 promoter, Pol II failed even to be recruited stably to the FKS2 promoter in the paf1D mutant, supporting a previously suggested role for the Paf1C in promoter site selection (Stolinski et al., 1997). This result contrasted with that observed for the ADH1 promoter, to which Pol II was recruited independent of Paf1 (Figures 4C and 4D), despite the presence of Paf1 on the ADH1 promoter in wild-type cells (Figures 4B and 4E; Krogan et al., 2002). Human Paf1 Complements the Mpk1-Dependent Function of Yeast Paf1 When expressed in yeast, the human ERK5 MAPK (hERK5) is activated in response to cell wall stress and suppresses the phenotypic defects of an mpk1D mutant (Truman et al., 2006). hERK5 also carries out the noncatalytic transcriptional function of Mpk1/Mlp1 (Kim et al., 2008), strongly suggesting that it interacts with Paf1 in the same manner as does Mpk1. This finding further suggests that the human Paf1C may be subject to similar regulation. Therefore, we examined the human PD2/ PAF1 (hPAF1) sequence for potential MAPK-binding D motifs analogous to that present in yeast Paf1. Although yeast Paf1 shares only 24% amino acid sequence identity with the hPAF1 protein, we identified a potential D motif (93-RIDPNVLL-100; italicized residues defined by the D motif) within the same region of hPAF1 as the D motif in yeast Paf1 and constructed a mutant predicted to ablate this site (residues Val98 and Leu100 mutated to Ala; hPAF1-AA). hPAF1 and its point mutant were expressed in a yeast paf1D mutant and tested for the ability to complement its deficiencies in FKS2 and CLN2 transcription. The wild-type hPAF1 was able to restore CLN2 transcription and to restore partially cell wall stress-induced FKS2 transcription (Figure 5A), indicating functional complementation of the paf1D mutation. Significantly, the hPAF1-AA mutant was specifically blocked for FKS2 induction, suggesting that MAPK regulation of the Paf1C is conserved in humans. To determine whether hERK5 can function in combination with hPAF1 to drive FKS2 expression in a yeast cell, we expressed both heterologous proteins in an mpk1D mlp1D paf1D mutant. FKS2 expression was induced in response to cell wall stress in Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 749
A
PAF1 Heat Stress
-
paf1-4A
-
+
Figure 4. The paf1-4A Mutant Is Blocked for Transcription Elongation of FKS2, but Not CLN2
paf1Δ
-
+
WCE
+
FKS2 CLN2
RT-PCR
18S
B P -743 to -470
CLN2
ATG
704 to 921 Paf1-HA
Heat Stress
TGA
M
-
C 1473 to 1631
Paf1-4A-HA
-
+
Paf1
-
+
+
Input ADH1 P M C
Paf1-HA ChIP
C PAF1 Heat Stress
-
+
Rpb3-HA paf1-4A
-
+
paf1Δ
-
+
CLN2
(A) RT-PCR of FKS2 and CLN2. A paf1D strain (DL3792) expressing PAF1, paf1-4A, or bearing empty vector was subjected to heat stress. Total RNA was processed for RT-PCR detection of FKS2 and CLN2 mRNA and 18S RNA. WCE, whole-cell extract. (B–F) ChIP analyses in a paf1D strain (DL3548) expressing the indicated forms of Paf1. (B) (Top) Schematic of the CLN2 gene showing regions amplified for ChIP along the length of the gene. P, promoter; M, middle coding region; C, C-terminal coding region. (Bottom) Paf1-HA ChIP on CLN2. (C) Rpb3-HA ChIP on CLN2. (D) Rpb3-HA ChIP on FKS2. (E) Paf1-HA ChIP on FKS2. (F) Mpk1-HA ChIP on FKS2.
Vector PAF1
-
+
Input ADH1
Rpb3-HA ChIP
reports, we detected attenuated transcripts across the FKS2 promoter-proximal region under noninducing conditions in both wild-type and paf1-4A cells (Figure 6A). Strikingly, Rpb3-HA Vector D although the paf1-4A mutant did not express PAF1 paf1-4A PAF1 paf1Δ Heat Stress extended FKS2 mRNA under stress conditions, + + + + Input ADH1 it did induce attenuated transcripts, confirming P Rpb3-HA ChIP N that this mutant is defective in transcription M FKS2 elongation and suggesting that the Mpk1-Paf1 C interaction serves an antitermination function. E Paf1-HA Paf1-4A-HA Paf1 The Sen1-Nrd1-Nab3 transcription terminaInput Heat Stress + + + ADH1 tion complex associates with Pol II near proP moters and appears to be most important for Paf1-HA ChIP N M FKS2 termination within 500 bp of transcription start C sites (Steinmetz et al., 2006; Kim et al., 2010a). PAF1 paf1-4A paf1Δ F Significantly, the Paf1C has been shown to be Input Heat Stress + + + DYN1 important for recruitment of the Sen1-Nrd1P Nab3 complex to snoRNA genes, and mutations Mpk1-HA ChIP N FKS2 in genes encoding Paf1C subunits result in tranM C scriptional readthrough products of these genes (Sheldon et al., 2005). To test the involvement of the Sen1-Nrd1-Nab3 termination complex in this strain, but not in a strain expressing the hPAF1-AA mutant FKS2 expression, we asked whether a temperature-sensitive (Figure 5B), supporting the conclusion that hERK5 can interact allele of SEN1 (sen1-1; Ursic et al., 1997) would suppress the FKS2 transcription defect in the paf1-4A mutant. In this experiwith the D motif in hPaf1 to drive transcription elongation. ment, raising the growth temperature to 39 C had two effects. First, it induced cell wall stress as in other experiments. Second, Sen1-Nrd1-Nab3-Mediated Transcription Attenuation it inactivated Sen1 function. Figure 6B shows that FKS2 expresof FKS2 A large number of short (200–400 nt) sense transcripts that termi- sion was restored to the paf1-4A mutant when Sen1 function nate between 243 and 213 of FKS2 (relative to the ATG) have was disabled, indicating that the FKS2 elongation defect that been identified through global transcription analyses conducted results from ablation of the Mpk1-Paf1 interaction is the conseunder noninducing conditions (Miura et al., 2006, Neil et al., quence of Sen1-dependent premature termination. ChIP anal2009). This suggests that attenuation (premature termination) ysis also revealed that Pol II occupancy of the entire FKS2 shortly after initiation may be used to minimize the basal level gene was restored to the paf1-4A mutant in the absence of of FKS2 transcription. Therefore, we explored the possibility Sen1 function (Figure 6C). To determine whether the Mpk1-Paf1 interaction blocks that the critical function of the Mpk1-Paf1 association in transcription elongation was to prevent attenuation under inducing recruitment or function of the Sen1-Nrd1-Nab3 termination conditions. The paf1-4A mutant was tested for the production complex in this setting, we next conducted Sen1 ChIP on of prematurely terminated FKS2 transcripts using RT-PCR to FKS2. In wild-type cells, Sen1 was detected weakly in associadetect the 50 regions of FKS2 RNAs. Consistent with previous tion with the FKS2 promoter under noninducing conditions, but 750 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
P M C
CLN2
hPAF1
A
-
Heat Stress
paf1Δ Δ
hPAF1-AA
-
+
-
+
Figure 5. hERK5 and hPAF1 Interact to Drive FKS2 Expression
WCE
+
(A and B) RT-PCR of FKS2. (A) Expression of either hPAF1 or hPAF1-AA in a paf1D mutant (DL3792). (B) Coexpression of hERK5 and hPAF1 in a paf1D mpk1D mlp1D mutant (DL3980). WCE, whole-cell extract.
FKS2 CLN2
RT-PCR
18S
hERK5 hPAF1
B Heat Stress
-
+
hERK5 hPAF1-AA
-
+
hERK5 Vector
Vector hPAF1
-
-
+
+
WCE
terminator sequence inserted into the ACT1 intron (Figure 6E) disrupts expression of the full-length message, resulting 18S in copper sensitivity of a yeast strain lacking the endogenous CUP1 gene. The region of FKS2 from 480 to 213 (relative to the ATG) behaved as a Sen1dependent terminator (Figure 6E). Within this sequence, we found a single potential Nab3 recognition site at position 476. A point mutation at this site relieved dependence on the Mpk1-Paf1 interaction for transcription in the context of an FKS2-lacZ reporter (Figure 6F), revealing that the Sen1 termination complex acts through this solitary Nab3 site. Additionally, mutation of the Nab3 site caused a 3-fold elevation of basal FKS2-lacZ expression, supporting the conclusion that transcriptional attenuation by the Sen1 complex is used under noninducing conditions to minimize FKS2 expression. Finally, returning to our initial question concerning how Mpk1 associates with the FKS2 coding region during transcription, we tested the requirement for the Mpk1-Paf1 interaction for this association. Figure 6G shows that Mpk1 failed to associate with the FKS2 coding region in the paf1-4A mutant whose elongation defect was suppressed by the sen1-1 mutation. This result indicates that Mpk1 is tethered to the elongation complex through its interaction with Paf1. FKS2
RT-PCR mpk1Δ mlp1Δ paf1Δ
not with other parts of the FKS2 gene (Figure 6D). However, this weak association was not evident under inducing conditions, lending support to the suggestion that Sen1-dependent attenuation plays a role in repressing FKS2 expression under noninducing conditions. Sen1 was also detected on the FKS2 promoter in the paf1-4A mutant under noninducing conditions. However, in contrast to wild-type cells, its association increased under inducing conditions (Figure 6D), suggesting that the Mpk1-Paf1 interaction normally interferes with recruitment of the termination complex. Recruitment of the Sen1-Nrd1-Nab3 complex to Pol II is influenced by the phosphorylation status of the C-terminal heptapeptide repeat domain (CTD) of the Pol II largest subunit (Rpb1). Both Nrd1 and Sen1 possess CTD interaction domains (Steinmetz and Brow, 1998; Finkel et al., 2010) that contribute to binding of the Sen1 complex to CTD Ser5-phosphorylated Pol II (Vasiljeva et al., 2008). In contrast to this, CTD Ser2 phosphorylation appears to inhibit Sen1 complex binding (Gudipati et al., 2008). Additionally, a positional correlation between CTD Ser7 phosphorylation and recruitment of Nrd1 was observed recently (Kim et al., 2010a). Therefore, to determine whether the Mpk1-Paf1 interaction affects recruitment of the Sen1 complex indirectly by altering the phosphorylation status of Pol II during FKS2 transcription, we carried out FKS2 ChIP analyses using antibodies specific for the Ser5-, Ser2-, or Ser7-phosphorylated forms of the CTD. For this experiment, we compared PAF1 to the paf1-4A mutant whose elongation defect was suppressed by the sen1-1 mutation. The Mpk1Paf1 interaction did not affect the CTD phosphorylation status on FKS2 (Figure S4), thus ruling out this potential mechanism for control of Sen1 recruitment. Sen1-mediated termination is dictated by the recognition of specific sequences in the nascent RNA by Nrd1 (GTA[A/G]) and/or Nab3 (TCTT) (Steinmetz et al., 2001; Morlando et al., 2002; Carroll et al., 2004). As noted above, a large number of short transcripts are initiated from numerous start sites across the FKS2 promoter (Miura et al., 2006, Neil et al., 2009). Therefore, we tested several regions of the FKS2 promoter for the presence of a Sen1-dependent terminator using an ACT1CUP1 fusion described by Steinmetz et al. (2001). In brief, a
DISCUSSION MAPKs regulate transcription not only by phosphorylation of target transcription factors, but also through the formation of stable interactions on the DNA. This has been studied most thoroughly in the yeast Hog1 MAPK, which is activated by osmotic stress and uses several distinct mechanisms for inducing gene expression including recruitment of transcription factors, chromatin remodelers, and Pol II to responsive promoters (reviewed in de Nadal and Posas, 2010). All of these mechanisms require Hog1 protein kinase activity. By contrast, the yeast Mpk1 MAPK and its pseudokinase paralog, Mlp1, drive a portion of the cell wall stress-induced transcriptional program through a pathway that does not require protein kinase catalytic activity (Kim et al., 2008; Truman et al., 2009; Kim and Levin, 2010). Mpk1 and Mlp1 recruit the Swi4/Swi6 (SBF) transcription factor to target promoters through direct binding to Swi4. In this study, we investigated the role of Mpk1 association with the transcription elongation complex on a transcriptional target whose expression is driven through this noncatalytic pathway. Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 751
- 564
A
- 243
E
Unstressed
FKS2
Heat Stress Prom.
ATG
PAF1
-
Heat Stress
-
CUP1 ACT1 intron
Inserted sequence
paf1-4A
+
Inserted terminator ACT1
+
SNR13 +125 to 232
W CE - 564 to + 48
FKS2 -480 to -213 forward
- 564 to - 243
RT-PCR
FKS2 -480 to -213 reverse
18S
-4 A
SEN1 SNR13 +125 to 232
pa f1 Δ
PA
pa f1
F1
sen1-1 pa f1 Δ
F1 PA
SEN1 pa f1 -4 A
No insert
B
WCE
FKS2 -480 to -213 forward
FKS2 FKS2 -480 to -213 reverse
RT-PCR
18S
SEN1 Rpb3-HA
V
sen1-E1597K
F
FKS2 (-700 to +6)-lacZ
Input ADH1 P N M FKS2 C
Rpb3-HA ChIP
Sen1-ZZ
D PAF1 Heat Stress
No insert
sen1-1 Rpb3-HA
PA F1 PA F1 pa f1 -4 A pa f1 Δ
PA F1
PA
F1
V
pa f1 -4 A pa f1 Δ
C
-
Sen1
paf1-4A
-
+
PAF1
paf1Δ
-
+
-
+
+
Input
Sen1-ZZ ChIP
paf1Δ
FKS2 (-700 to +6)-lacZ with Nab3-binding site mutation
A
Δ
-4
f1
f1 pa
pa
F1 PA
f pa
pa
P
1 AF
paf1-4A
sen1-1
4A 1-
f1 Δ
SEN1
G
PAF1
ADH1 P N M FKS2 C
Mpk1-HA ChIP
Input
ADH1 P N M FKS2 C
PAF1
paf1-4A
paf1Δ
Figure 6. The Mpk1-Paf1 Interaction Prevents Sen1-Mediated Premature Termination (A) RT-PCR of the FKS2 50 region. A paf1D strain (DL3977) expressing PAF1 or paf1-4A was processed for RT-PCR detection of FKS2 transcripts. One primer pair detects only extended mRNA (564 to +48). A second pair detects both attenuated transcripts and extended mRNA (564 to 243). (B) RT-PCR of FKS2. A SEN1 paf1D strain (DL3977) and a sen1-1 paf1D strain (DL3978) expressing PAF1, paf1-4A, or bearing empty vector were processed for RT-PCR detection of FKS2 expression after heat stress. (C) Rpb3-HA ChIP on FKS2. The same strains as in (B), but expressing Rpb3-ZZ-HA, were processed for ChIP after heat stress. V, vector. (D) Sen1-ZZ ChIP on FKS2 in a paf1D strain expressing Sen1-CBD-ZZ (DL3979) and the indicated forms of Paf1. (E) A region of the FKS2 promoter functions as a Sen1-dependent terminator. The indicated region of FKS2 sequence was inserted in both orientations into the ACT1 intron of an ACT1-CUP1 fusion plasmid, which was transformed into SEN1 cup1D (DL3990) and sen1-E1597K cup1D (DL3995) strains. The SNR13 terminator was used as a positive control. Serial dilutions were spotted onto plates with 0.4 mM CuSO4 and incubated for 3 days at the semipermissive temperature for sen1-E1597K of 30 C. (F) The Nab3-binding site at position 476 in FKS2 is responsible for transcriptional dependence on the Mpk1-Paf1 interaction. Expression of the indicated FKS2-lacZ fusion was measured before and after heat stress in a paf1D strain (DL3548) expressing PAF1, paf1-4A, or bearing empty vector. Each value represents the mean and standard deviation from three independent transformants. (G) Mpk1-HA ChIP on FKS2. The same strains as in (B), but expressing Mpk1-HA, were processed for ChIP analysis on FKS2 after heat stress.
752 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
Figure 7. Model for Mpk1-Driven FKS2 Transcription
A
Paf1
Sen1 Paf1 Nab3
FKS2 B
Pol II SCB
Mpk1
FKS2
Swi6
PolNrd1 II
(A) Under noninducing conditions, weak transcription initiation is attenuated by association of the Sen1-Nrd1-Nab3 complex to the elongation complex. (B) Under inducing conditions, activated Mpk1 recruits Swi4-Swi6 to the FKS2 promoter. (C) Pol II and the Paf1C are recruited to the promoter in a Swi6-dependent manner. (D) Mpk1 associates with Paf1, possibly by handoff from Swi4. Mpk1 functions as an antitermination factor by blocking recruitment of the termination complex.
Swi4
SCB
C
Paf1
the local context of such genes and is the consequence of a handoff of Mpk1 from Swi4 to Paf1 on the assembled initiation complex (Figure 7).
The Mpk1-Paf1 Interaction Is Required for Transcription Elongation of FKS2 A mutation that ablated the D motif in Paf1 SCB (paf1-4A) had the specific effect of blocking Mpk1-dependent functions of the Paf1C. Strikingly, D the paf1-4A mutation blocked FKS2 transcription, Mpk1 Mpk1 Paf1 Paf1 though not the transcription of another gene (CLN2) whose expression is dependent on both Swi6 SBF and the Paf1C, but not dependent on Mpk1. Pol II Pol II Swi4 Consequently, this mutation caused hypersensitivity to cell wall stress agents, but not to other SCB agents to which a paf1D mutant is hypersensitive. However, the most interesting behavior of the paf1-4A mutation is that it blocked transcription elongation of FKS2. Although in this mutant, Mpk1, Pol II, and Mpk1 Moves from the Initiation Complex Paf1-4A were all recruited to the FKS2 promoter in response to to the Elongation Complex on the Paf1C Scaffold We found that active (phosphorylated) Mpk1 and Mlp1 associate cell wall stress, none of these proteins was detected along the with the coding region of the FKS2 gene during transcription by FKS2 coding region. binding to the Paf1C elongation complex. MAPKs possess a docking site (called the D motif-binding site) for binding targets Mpk1 Drives FKS2 Transcription Elongation and regulators, which resides on the opposite side of the protein by Blocking Recruitment of the Sen1-Nrd1-Nab3 from the catalytic site (Zhang et al., 2003). The docking site on Termination Complex active Mpk1 (Truman et al., 2009) engages the Paf1C through In principle, the observed failure of the elongation complex to contact with a D motif near the N terminus of the Paf1 subunit. progress beyond the FKS2 promoter in the paf1-4A mutant could It appears that this is the only stable association made between result either from stalling of Pol II early in the elongation cycle or Mpk1 and the Paf1C because all of the other in vivo associations from premature termination of transcription at a point close to the detected between Mpk1 and Paf1C subunits were lost upon promoter. We found that, under inducing conditions, the paf1-4A mutant accumulated short attenuated transcripts across the mutational ablation of the D motif in Paf1. An interesting aspect of the Mpk1-Paf1 interaction is that it FKS2 promoter-proximal region, indicating premature terminarequires both activating signal to Mpk1 and the presence of tion. Moreover, the paf1-4A transcription defect was suppressed both Swi4 and Swi6. Mpk1 recruits the latter two factors to the by disabling the Sen1-Nrd1-Nab3 termination complex. This led FKS2 promoter through association of its docking site with a D to our identification of a Sen1-dependent terminator within the motif in Swi4 (Kim et al., 2008; Truman et al., 2009). Although FKS2 promoter-proximal region and the conclusion that expresan Mpk1-Swi4 complex binds to the FKS2 promoter in the sion of this gene is regulated, in part, by transcriptional attenuaabsence of the Swi6 transcription activation subunit (Kim et al., tion. Under inducing conditions, we detected Sen1 on the 2008), we found that Swi6 was required for stable recruitment FKS2 promoter in the paf1-4A mutant, but not in wild-type cells, of Pol II and the Paf1C to this promoter. We conclude from these suggesting that the Mpk1-Paf1 interaction interferes with recruitresults that the Mpk1-Paf1 association occurs only under condi- ment of the Sen1 complex to Pol II, rather than preventing its tions in which Pol II and the Paf1C have also been recruited to the activity after recruitment. Therefore, in this setting, active Mpk1 promoters of Mpk1-Swi4-Swi6-regulated genes. It seems likely, functions as an antitermination factor that allows transcription therefore, that the Mpk1-Paf1 interaction occurs specifically in of FKS2 by blocking recruitment of the Sen1 complex to the Mpk1
FKS2
Swi6 Swi4
Pol II
FKS2
Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 753
elongation complex early in the transcription cycle (Figure 7). It was reported previously that the Paf1C facilitates recruitment of the Sen1 complex to snoRNA genes (Sheldon et al., 2005), but the mechanism remains obscure. Our data suggest that, whatever the nature of the interaction between the Paf1C elongation complex and the Sen1 termination complex, it extends beyond snoRNA genes. It is noteworthy that active Mpk1 is required for two steps in the FKS2 transcription cycle: initiation and antitermination. We propose that, because FKS2 is expressed only under emergent conditions, Sen1-mediated termination is used to minimize FKS2 basal transcription during periods of unstressed growth (Figure 7). Attenuation would, therefore, have to be suppressed under inducing conditions. This hypothesis is supported by three findings. First, in the absence of cell wall stress, a large number of attenuated sense transcripts are produced across the FKS2 promoter-proximal region (Miura et al., 2006; Neil et al., 2009; Figure 6A). Second, under noninducing conditions, we detected weak recruitment of Sen1 specifically to the FKS2 promoter (Figure 6D). Third, mutation of the Sen1-dependent terminator in the FKS2 promoter-proximal region caused an elevation of basal FKS2 expression (Figure 6F). Considering that the Mpk1-Paf1 interaction serves its antitermination function shortly after initiation, it is not clear why Mpk1 remains associated with the elongation complex across most of the FKS2 gene. This MAPK may serve additional as yet undiscovered functions later in the FKS2 transcription cycle. The Sen1 complex is required for termination and 30 end formation of Pol II transcripts that are not polyadenylated (Ursic et al., 1997; Rasmussen and Culbertson, 1998). However, it has also been implicated in the transcriptional attenuation of several protein-coding genes, including NRD1, HRP1, and IMD2 (Steinmetz et al., 2001, 2006; Arigo et al., 2006). Both NRD1 and HRP1 encode proteins involved in Sen1-mediated termination. Thus, their autoregulation represents a specialized feedback mechanism. The IMD2 gene represents a similarly special case (Steinmetz et al., 2006). Our finding of the requirement for Mpk1-Paf1 association in blocking premature termination of FKS2 transcription reveals both an expanded repertoire for the Sen1-Nrd1-Nab3 termination complex on protein coding genes and a regulatory mechanism for its action. Indeed, two recent findings suggest a wider role for Sen1-dependent termination in transcription regulation than previously thought. First, Nrd1 is recruited widely to Pol II transcription units of all types (Kim et al., 2010a). Second, the genome-wide distribution of Pol II across both noncoding and protein coding genes is strongly influenced by disabling the Sen1 helicase (Steinmetz et al., 2006). We note additionally that short sense transcripts are produced across the promoter regions of a significant fraction of protein coding genes in yeast (10%) (Neil et al., 2009), suggesting that transcriptional attenuation may be a widespread phenomenon. Paf1 Is Required for Stable Recruitment of Pol II to Some Promoters We found that Paf1 is required for stable recruitment of Pol II to the CLN2 and FKS2 promoters, both genes whose expression requires Paf1. However, Paf1 is not universally required for Pol 754 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
II recruitment. Indeed, that a paf1D mutant is not lethal indicates that a requirement for the Paf1C in Pol II recruitment is restricted to a small number of genes, despite its ubiquitous association with Pol II transcription units (Jaehning, 2010). Two lines of evidence previously implicated the Paf1C in transcription initiation. First, the gene encoding the Rtf1 (restores TBP function) subunit of the Paf1C was isolated initially through the ability of a mutant allele to alter transcription start site selection (Stolinski et al., 1997). Second, the gene encoding the Ctr9 (CLN three requiring) subunit was isolated in two independent genetic screens through the defect of ctr9 mutants in Swi4/ Swi6 (SBF)-driven gene expression (Di Como et al., 1995; Koch et al., 1999). Notably, the results of Koch et al. (1999) implicate the Paf1C in transcriptional activation of SBF-dependent promoters. Our results reveal that the Paf1C is required for stable recruitment of Pol II to at least a subset of SBF-dependent promoters. Because SBF is a key cell cycle-regulated transcription factor and cell cycle-regulated genes represent a large fraction of those whose expression is at least partially dependent on Paf1 (Porter et al., 2002), this may reflect a general role for the Paf1C in recruitment of Pol II to SBF-dependent promoters. MAPK Regulation of Paf1C Function Is Conserved in Human PAF1 We found that hPAF1 complemented the transcription defects of a paf1D mutant, at least with regard to FKS2 and CLN2 expression. Remarkably, mutation of a predicted D motif in hPAF1 within the same region as that found in yeast Paf1 specifically blocked FKS2 induction by cell wall stress, strongly suggesting that MAPK regulation of the Paf1C is conserved in humans. Moreover, because human ERK5 complements the defects associated with loss of yeast MPK1 and MLP1 function, including FKS2 induction (Truman et al., 2006; Kim et al., 2008), it seems likely that there is a role for an ERK5-PAF1 pairing in the regulation of human gene expression. Our finding that an mpk1D, mlp1D, paf1D yeast strain was complemented for FKS2 induction with hERK5 and hPAF1 supports this conclusion. It is important to note that our findings are necessarily restricted to a small subset of stress-induced genes. Even among those activated in response to cell wall stress, most are not under the control of the noncatalytic Mpk1/Mlp1 pathway described here. It remains to be seen whether the D motif in Paf1 is recognized by additional yeast MAPKs in response to other signals. In any case, that a paf1D mutant displays hypersensitivity to such a wide array of stress agents (Betz et al., 2002) suggests that the Paf1C serves a poorly understood function in stress-induced gene expression. We were able to separate its function in cell wall stress transcription from its other functions by a point mutation in Paf1. Perhaps expression of other Paf1C-dependent genes is subject to a similar mode of regulation by factors yet to be identified. EXPERIMENTAL PROCEDURES Plasmids and Strains The S. cerevisiae strains and plasmids used in this study are listed in Table S2 and Table S3, respectively. Their construction and growth conditions are described in Extended Experimental Procedures.
Chromatin Immunoprecipitation Assays Chromatin immunoprecipitation (ChIP) assays were conducted as described by Kim et al., (2008). Primers used to amplify specific regions of FKS2, CLN2, DYN1, and ADH1 are listed in Table S4. For each experiment, representative gels were chosen from at least three independent sets of extracts. See Extended Experimental Procedures for additional details. Coimmunoprecipitation Assays Protein extractions and coimmunoprecipitations were conducted as described previously (Kamada et al., 1995) with modification as described in Extended Experimental Procedures. RT-PCR Assays Total RNA was prepared as described previously (Schmitt et al., 1990), and cDNAs were made using Superscript III reverse transcriptase (Invitrogen), as per the manufacturer’s instructions, and amplified by PCR with Taq polymerase. Primers used to amplify FKS2, CLN2, and 18S cDNA sequences are listed in Table S4. See Extended Experimental Procedures for additional details. SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, three figures, and four tables and can be found online with this article online at doi:10.1016/j.cell.2011.01.034. ACKNOWLEDGMENTS We thank Judith Jaehning for comments on the manuscript; Steve Buratowski for helpful discussions; David Brow for the ACT1-CUP1 reporter system; Michael Culbertson for the sen1-1 mutant; Peter Piper for two-hybrid fusions; Elizabeth Yranski for assistance with strain construction; and two anonymous reviewers for insightful comments. This work was supported by a grant from the NIH (GM48533) to D.E.L. Received: August 10, 2010 Revised: November 16, 2010 Accepted: January 26, 2011 Published: March 3, 2011
and Ccr4p plays a role in protein kinase C signaling. Mol. Cell. Biol. 19, 1056–1067. Chaudhary, K., Deb, S., Moniaux, N., Ponnusamy, M.P., and Batra, S.K. (2007). Human RNA polymerase II-associated factor complex: dysregulation in cancer. Oncogene 26, 7499–7507. Costa, P.J., and Arndt, K.M. (2000). Synthetic lethal interactions suggest a role for the Saccharomyces cerevisiae Rtf1 protein in transcription elongation. Genetics 156, 535–547. de Nadal, E., and Posas, F. (2010). Multilayered control of gene expression by stress-activated protein kinases. EMBO J. 29, 4–13. Di Como, C.J., Chang, H., and Arndt, K.T. (1995). Activation of CLN1 and CLN2 G1 cyclin gene expression by BCK2. Mol. Cell. Biol. 15, 1835–1846. Dodou, E., and Treisman, R. (1997). The Saccharomyces cerevisiae MADSbox transcription factor Rlm1 is a target for the Mpk1 mitogen-activated protein kinase pathway. Mol. Cell. Biol. 17, 1848–1859. Finkel, J.S., Chinchilla, K., Ursic, D., and Culbertson, M.R. (2010). Sen1p performs two genetically separable functions in transcription and processing of U5 small nuclear RNA in Saccharomyces cerevisiae. Genetics 184, 107–118. Gudipati, R.K., Villa, T., Boulay, J., and Libri, D. (2008). Phosphorylation of the RNA polymerase II C-terminal domain dictates transcription termination choice. Nat. Struct. Mol. Biol. 15, 786–794. Igual, J.C., Johnson, A.L., and Johnston, L.H. (1996). Coordinated regulation of gene expression by the cell cycle transcription factor Swi4 and the protein kinase C MAP kinase pathway for yeast cell integrity. EMBO J. 15, 5001–5013. Jaehning, J.A. (2010). The Paf1 complex: platform or player in RNA polymerase II transcription? Biochim. Biophys. Acta 1799, 379–388. Jung, U.S., and Levin, D.E. (1999). Genome-wide analysis of gene expression regulated by the yeast cell wall integrity signalling pathway. Mol. Microbiol. 34, 1049–1057. Jung, U.S., Sobering, A.K., Romeo, M.J., and Levin, D.E. (2002). Regulation of the yeast Rlm1 transcription factor by the Mpk1 cell wall integrity MAP kinase. Mol. Microbiol. 46, 781–789. Kamada, Y., Jung, U.S., Piotrowski, J., and Levin, D.E. (1995). The protein kinase C-activated MAP kinase pathway of Saccharomyces cerevisiae mediates a novel aspect of the heat shock response. Genes Dev. 9, 1559–1571.
REFERENCES
Kim, K.-Y., and Levin, D.E. (2010). Transcriptional reporters for genes activated by cell wall stress through a non-catalytic mechanism involving Mpk1 and SBF. Yeast 27, 541–548.
Abe, J., Kusuhara, M., Ulevitch, R.J., Berk, B.C., and Lee, J.D. (1996). Big mitogen-activated protein kinase 1 (BMK1) is a redox-sensitive kinase. J. Biol. Chem. 271, 16586–16590.
Kim, K.-Y., Truman, A.W., and Levin, D.E. (2008). Yeast Mpk1 mitogen-activated protein kinase activates transcription through Swi4/Swi6 by a noncatalytic mechanism that requires upstream signal. Mol. Cell. Biol. 28, 2579–2589.
Andrews, B.J., and Moore, L.A. (1992). Interaction of the yeast Swi4 and Swi6 cell cycle regulatory proteins in vitro. Proc. Natl. Acad. Sci. USA 89, 11852– 11856.
Kim, H., Erickson, B., Luo, W., Seward, D., Graber, J.H., Pollock, D.D., Megee, P.C., and Bentley, D.L. (2010a). Gene-specific RNA polymerase II phosphorylation and the CTD code. Nat. Struct. Mol. Biol. 17, 1279–1286.
Arigo, J.T., Carroll, K.L., Ames, J.M., and Corden, J.L. (2006). Regulation of yeast NRD1 expression by premature transcription termination. Mol. Cell 21, 641–651.
Kim, J., Guermah, M., and Roeder, R.G. (2010b). The human PAF1 complex acts in chromatin transcription elongation both independently and cooperatively with SII/TFIIS. Cell 140, 491–503.
Baetz, K., Moffat, J., Haynes, J., Chang, M., and Andrews, B. (2001). Transcriptional coregulation by the cell integrity mitogen-activated protein kinase Slt2 and the cell cycle regulator Swi4. Mol. Cell. Biol. 21, 6515–6528. Betz, J.L., Chang, M., Washburn, T.M., Porter, S.E., Mueller, C.L., and Jaehning, J.A. (2002). Phenotypic analysis of Paf1/RNA polymerase II complex mutations reveals connections to cell cycle regulation, protein synthesis, and lipid and nucleic acid metabolism. Mol. Genet. Genomics 268, 272–285. Breeden, L.L. (2003). Periodic transcription: a cycle within a cycle. Curr. Biol. 13, R31–R38.
Klis, F.M. (1994). Review: cell wall assembly in yeast. Yeast 10, 851–869. Koch, C., Wollmann, P., Dahl, M., and Lottspeich, F. (1999). A role for Ctr9p and Paf1p in the regulation G1 cyclin expression in yeast. Nucleic Acids Res. 27, 2126–2134. Krogan, N.J., Kim, M., Ahn, S.H., Zhong, G., Kobor, M.S., Cagney, G., Emili, A., Shilatifard, A., Buratowski, S., and Greenblatt, J.F. (2002). RNA polymerase II elongation factors of Saccharomyces cerevisiae: a targeted proteomics approach. Mol. Cell. Biol. 22, 6979–6992.
Carroll, K.L., Pradhan, D.A., Granek, J.A., Clarke, N.D., and Corden, J.L. (2004). Identification of cis elements directing termination of yeast nonpolyadenylated snoRNA transcripts. Mol. Cell. Biol. 24, 6241–6252.
Krogan, N.J., Dover, J., Wood, A., Schneider, J., Heidt, J., Boateng, M.A., Dean, K., Ryan, O.W., Golshani, A., Johnston, M., et al. (2003). The Paf1 complex is required for histone H3 methylation by COMPASS and Dot1p: linking transcriptional elongation to histone methylation. Mol. Cell 11, 721–729.
Chang, M., French-Cornay, D., Fan, H.Y., Klein, H., Denis, C.L., and Jaehning, J.A. (1999). A complex containing RNA polymerase II, Paf1p, Cdc73p, Hpr1p,
Levin, D.E. (2005). Cell wall integrity signaling in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 69, 262–291.
Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc. 755
Madden, K., Sheu, Y.J., Baetz, K., Andrews, B., and Snyder, M. (1997). SBF cell cycle regulator as a target of the yeast PKC-MAP kinase pathway. Science 275, 1781–1784.
Steinmetz, E.J., and Brow, D.A. (1998). Control of pre-mRNA accumulation by the essential yeast protein Nrd1 requires high-affinity transcript binding and a domain implicated in RNA polymerase II association. Proc. Natl. Acad. Sci. USA 95, 6699–6704.
Martı´n, H., Rodrı´guez-Pacho´n, J.M., Ruiz, C., Nombela, C., and Molina, M. (2000). Regulatory mechanisms for modulation of signaling through the cell integrity Slt2-mediated pathway in Saccharomyces cerevisiae. J. Biol. Chem. 275, 1511–1519.
Steinmetz, E.J., Conrad, N.K., Brow, D.A., and Corden, J.L. (2001). RNAbinding protein Nrd1 directs poly(A)-independent 30 -end formation of RNA polymerase II transcripts. Nature 413, 327–331.
Miura, F., Kawaguchi, N., Sese, J., Toyoda, A., Hattori, M., Morishita, S., and Ito, T. (2006). A large-scale full-length cDNA analysis to explore the budding yeast transcriptome. Proc. Natl. Acad. Sci. USA 103, 17846–17851.
Steinmetz, E.J., Warren, C.L., Kuehner, J.N., Panbehi, B., Ansari, A.Z., and Brow, D.A. (2006). Genome-wide distribution of yeast RNA polymerase II and its control by Sen1 helicase. Mol. Cell 24, 735–746.
Morlando, M., Greco, P., Dichtl, B., Fatica, A., Keller, W., and Bozzoni, I. (2002). Functional analysis of yeast snoRNA and snRNA 30 -end formation mediated by uncoupling of cleavage and polyadenylation. Mol. Cell. Biol. 22, 1379–1389.
Stolinski, L.A., Eisenmann, D.M., and Arndt, K.M. (1997). Identification of RTF1, a novel gene important for TATA site selection by TATA box-binding protein in Saccharomyces cerevisiae. Mol. Cell. Biol. 17, 4490–4500.
Mueller, C.L., Porter, S.E., Hoffman, M.G., and Jaehning, J.A. (2004). The Paf1 complex has functions independent of actively transcribing RNA polymerase II. Mol. Cell 14, 447–456. Neil, H., Malabat, C., d’Aubenton-Carafa, Y., Xu, Z., Steinmetz, L.M., and Jacquier, A. (2009). Widespread bidirectional promoters are the major source of cryptic transcripts in yeast. Nature 457, 1038–1042. Penheiter, K.L., Washburn, T.M., Porter, S.E., Hoffman, M.G., and Jaehning, J.A. (2005). A posttranscriptional role for the yeast Paf1-RNA polymerase II complex is revealed by identification of primary targets. Mol. Cell 20, 213–223. Pokholok, D.K., Hannett, N.M., and Young, R.A. (2002). Exchange of RNA polymerase II initiation and elongation factors during gene expression in vivo. Mol. Cell 9, 799–809. Pokholok, D.K., Zeitlinger, J., Hannett, N.M., Reynolds, D.B., and Young, R.A. (2006). Activated signal transduction kinases frequently occupy target genes. Science 313, 533–536. Porter, S.E., Washburn, T.M., Chang, M., and Jaehning, J.A. (2002). The yeast pafl-rNA polymerase II complex is required for full expression of a subset of cell cycle-regulated genes. Eukaryot. Cell 1, 830–842. Proft, M., Mas, G., de Nadal, E., Vendrell, A., Noriega, N., Struhl, K., and Posas, F. (2006). The stress-activated Hog1 kinase is a selective transcriptional elongation factor for genes responding to osmotic stress. Mol. Cell 23, 241–250. Rasmussen, T.P., and Culbertson, M.R. (1998). The putative nucleic acid helicase Sen1p is required for formation and stability of termini and for maximal rates of synthesis and levels of accumulation of small nucleolar RNAs in Saccharomyces cerevisiae. Mol. Cell. Biol. 18, 6885–6896. Schmitt, M.E., Brown, T.A., and Trumpower, B.L. (1990). A rapid and simple method for preparation of RNA from Saccharomyces cerevisiae. Nucleic Acids Res. 18, 3091–3092. Sedgwick, S.G., Taylor, I.A., Adam, A.C., Spanos, A., Howell, S., Morgan, B.A., Treiber, M.K., Kanuga, N., Banks, G.R., Foord, R., and Smerdon, S.J. (1998). Structural and functional architecture of the yeast cell-cycle transcription factor swi6. J. Mol. Biol. 281, 763–775.
Truman, A.W., Millson, S.H., Nuttall, J.M., King, V., Mollapour, M., Prodromou, C., Pearl, L.H., and Piper, P.W. (2006). Expressed in the yeast S. cerevisiae, human ERK5 is a client of the Hsp90 chaperone that complements loss of the Slt2 (Mpk1) cell integrity stress-activated protein kinase. Eukaryot. Cell 5, 1914–1924. Truman, A.W., Kim, K.Y., and Levin, D.E. (2009). Mechanism of Mpk1 MAPK binding to the Swi4 transcription factor and its regulation by a novel caffeine-induced phosphorylation. Mol. Cell. Biol. 29, 6449–6461. Ursic, D., Himmel, K.L., Gurley, K.A., Webb, F., and Culbertson, M.R. (1997). The yeast SEN1 gene is required for the processing of diverse RNA classes. Nucleic Acids Res. 25, 4778–4785. Vasiljeva, L., Kim, M., Mutschler, H., Buratowski, S., and Meinhart, A. (2008). The Nrd1-Nab3-Sen1 termination complex interacts with the Ser5-phosphorylated RNA polymerase II C-terminal domain. Nat. Struct. Mol. Biol. 15, 795–804. Watanabe, Y., Takaesu, G., Hagiwara, M., Irie, K., and Matsumoto, K. (1997). Characterization of a serum response factor-like protein in Saccharomyces cerevisiae, Rlm1, which has transcriptional activity regulated by the Mpk1 (Slt2) mitogen-activated protein kinase pathway. Mol. Cell. Biol. 17, 2615– 2623. Wood, A., Schneider, J., Dover, J., Johnston, M., and Shilatifard, A. (2003). The Paf1 complex is essential for histone monoubiquitination by the Rad6Bre1 complex, which signals for histone methylation by COMPASS and Dot1p. J. Biol. Chem. 278, 34739–34742. Wyers, F., Rougemaille, M., Badis, G., Rousselle, J.C., Dufour, M.E., Boulay, J., Re´gnault, B., Devaux, F., Namane, A., Se´raphin, B., et al. (2005). Cryptic pol II transcripts are degraded by a nuclear quality control pathway involving a new poly(A) polymerase. Cell 121, 725–737. Yan, C., Luo, H., Lee, J.D., Abe, J., and Berk, B.C. (2001). Molecular cloning of mouse ERK5/BMK1 splice variants and characterization of ERK5 functional domains. J. Biol. Chem. 276, 10870–10878.
Sheldon, K.E., Mauger, D.M., and Arndt, K.M. (2005). A Requirement for the Saccharomyces cerevisiae Paf1 complex in snoRNA 30 end formation. Mol. Cell 20, 225–236.
Zhang, J., Zhou, B., Zheng, C.-F., and Zhang, Z.-Y. (2003). A bipartite mechanism for ERK2 recognition by its cognate regulators and substrates. J. Biol. Chem. 278, 29901–29912.
Sidorova, J., and Breeden, L. (1993). Analysis of the SWI4/SWI6 protein complex, which directs G1/S-specific transcription in Saccharomyces cerevisiae. Mol. Cell. Biol. 13, 1069–1077.
Zhao, C., Jung, U.S., Garrett-Engele, P., Roe, T., Cyert, M.S., and Levin, D.E. (1998). Temperature-induced expression of yeast FKS2 is under the dual control of protein kinase C and calcineurin. Mol. Cell. Biol. 18, 1013–1022.
756 Cell 144, 745–756, March 4, 2011 ª2011 Elsevier Inc.
Activation of mTORC2 by Association with the Ribosome Vittoria Zinzalla,1 Daniele Stracka,1 Wolfgang Oppliger,1 and Michael N. Hall1,* 1Biozentrum, University of Basel, CH-4056 Basel, Switzerland *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.014
SUMMARY
The target of rapamycin (TOR) is a highly conserved protein kinase and a central controller of growth. Mammalian TOR complex 2 (mTORC2) regulates AGC kinase family members and is implicated in various disorders, including cancer and diabetes. Here, we investigated the upstream regulation of mTORC2. A genetic screen in yeast and subsequent studies in mammalian cells revealed that ribosomes, but not protein synthesis, are required for mTORC2 signaling. Active mTORC2 was physically associated with the ribosome, and insulin-stimulated PI3K signaling promoted mTORC2-ribosome binding, suggesting that ribosomes activate mTORC2 directly. Findings with melanoma and colon cancer cells suggest that mTORC2-ribosome association is important in oncogenic PI3K signaling. Thus, TORC2-ribosome interaction is a likely conserved mechanism of TORC2 activation that is physiologically relevant in both normal and cancer cells. As ribosome content determines growth capacity of a cell, this mechanism of TORC2 regulation ensures that TORC2 is active only in growing cells. INTRODUCTION Target of rapamycin (TOR) is a central controller of cell growth and metabolism in response to nutrients, growth factors, and energy status. TOR is found in two structurally and functionally distinct multiprotein complexes termed TOR complex 1 (TORC1) and TORC2 (Wullschleger et al., 2006). The TOR complexes, originally described in yeast (Loewith et al., 2002), are conserved across all eukaryotes and regulate a wide spectrum of cellular processes that mediate cell growth (Laplante and Sabatini, 2009; Soulard et al., 2009; Wullschleger et al., 2006; Yang and Guan, 2007). In mammalian cells, mammalian TORC1 (mTORC1) contains mTOR, raptor, and mLST8 and is sensitive to the immunosuppressant and anticancer drug rapamycin (Hara et al., 2002; Kim et al., 2002; Loewith et al., 2002). mTORC1 controls transcription, ribosome biogenesis, protein synthesis, lipid synthesis, nutrient transport, autophagy, and other growth-related processes. The best-characterized substrates of mTORC1 are S6K and 4E-BP via which mTORC1 controls
translation (Sonenberg and Hinnebusch, 2009). mTORC2 consists of mTOR, rictor, mSIN1, mLST8, and PRR5/PRR5L (also known as protor1 and 2) and is insensitive to rapamycin, although long-term rapamycin treatment can indirectly inhibit mTORC2 in some cell types (Cybulski and Hall, 2009; Jacinto et al., 2004; Sarbassov et al., 2004; Sarbassov et al., 2006; Sparks and Guertin, 2010). mTORC2 directly phosphorylates and activates the AGC kinases Akt (also known as PKB), SGK1, and likely PKC (Facchinetti et al., 2008; Garcı´a-Martı´nez and Alessi, 2008; Hresko and Mueckler, 2005; Ikenoue et al., 2008; Jacinto and Lorberg, 2008; Sarbassov et al., 2004; Sarbassov et al., 2005). mTORC2 promotes cell survival via Akt and mediates organization of the actin cytoskeleton (Cybulski and Hall, 2009; Sparks and Guertin, 2010). Both mTORC1 and mTORC2 are activated by growth factors, including insulin, IGF-1, and others. Growth factors activate mTORC1 via phosphatidylinositol 3-kinase (PI3K), PDK1, Akt, the TSC1-TSC2 complex, and Rheb, a small guanosine triphosphate (GTP)-binding protein that binds and activates mTORC1 directly (Avruch et al., 2009; Manning and Cantley, 2003). In contrast, the mechanism via which growth factors activate mTORC2 has been elusive. Growth factors activate mTORC2 via PI3K (Frias et al., 2006; Garcı´a-Martı´nez and Alessi, 2008; Yang et al., 2006), but signaling steps beyond PI3K are distinct from those upstream of mTORC1 and unknown (Cybulski and Hall, 2009; Sparks and Guertin, 2010). The nature of the upstream regulators of TORC2 in unicellular model organisms such as yeasts, which lack a growth factor signaling pathway, is completely unknown (Soulard et al., 2009). Here, we describe a genetic screen in yeast and subsequent studies in mammalian cells that identify the ribosome as an activator of TORC2. We demonstrate that the ribosome, independent of protein synthesis, is required for mTORC2 signaling in vivo and mTORC2 kinase activity in vitro. Active mTORC2 is associated with the ribosome. Insulin stimulates the association of mTORC2 with the ribosome via PI3K signaling. Findings with cancer cells suggest that ribosome-dependent mTORC2 activation is physiologically relevant in tumors with hyperactive PI3K signaling. RESULTS A Genetic Screen Reveals that NIP7 Is Required for TORC2 Signaling in Yeast To identify upstream activators of TORC2, we performed a genetic screen in the budding yeast S. cerevisiae. Furthermore, Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 757
we assumed that whatever activates TORC2 in yeast would also be upstream of mTORC2 in mammals. In other words, we assumed that growth factor signaling was grafted onto a heretofore unknown ancestral input controlling TORC2 in unicellular yeast. This reasoning was supported by the fact that growth factor signaling was grafted onto the ancestral nutrient input in the case of TORC1. mTORC2 phosphorylates the hydrophobic motif in the AGC kinase SGK1 and thereby activates SGK1. In yeast, TORC2 similarly phosphorylates and activates the SGK1 ortholog YPK2. Our genetic screen was based on the observation that overexpression of constitutively active YPK2 (YPK2D239A, hereafter referred to as YPK2*) suppresses the lethality of a TORC2 defect (Aronova et al., 2008; Kamada et al., 2005). Assuming that YPK2* would also suppress lethality caused by a defect in an upstream activator of TORC2, we performed a so-called reverse suppressor screen to isolate yeast mutants that depend on YPK2* for viability (Figure 1A and Experimental Procedures). This is referred to as a reverse suppressor screen because it starts with a suppressor mutation (YPK2*) to identify unknown ‘‘suppressee’’ mutations—the reverse order of a normal (forward) suppressor screen. The screen was predicted to identify loss-of-function mutations in an essential upstream activator of TORC2 or in the essential components of TORC2, including TOR2, AVO1 (mSIN1 ortholog), and AVO3 (rictor ortholog). The screen yielded a total of 44 independent mutants defective in TORC2, thereby validating the screen. The 44 mutants consisted of 25, 13, and 6 tor2, avo1, and avo3 mutants, respectively. In addition, we obtained a temperature-sensitive nip7 mutant that we hereafter refer to as nip7-1 (Figure 1A and Figures S1A and S1B). Sequence analysis of the nip7-1 allele identified a point mutation that converts glycine 71 to aspartic acid. Western blot analysis of extracts from nip7-1 cells showed that the point mutation mildly and strongly decreased NIP7 protein levels at permissive temperature (25 C) and nonpermissive temperature (37 C), respectively (Figure S1A). YPK2*, but not wild-type YPK2, suppressed the growth defect of the nip7-1 mutant at semipermissive temperature (30 C and 34 C) (Figure S1B). NIP7 is an essential protein required for maturation of rRNA of the 60S ribosomal subunit (Zanchin et al., 1997). Confirming this role of NIP7 in ribosome biogenesis, we observed a reduction in the amounts of the 60S subunit, the 80S ribosome, and polysomes, with a concomitant appearance of halfmer polysomes, in extracts of the nip7-1 mutant grown at semipermissive temperature (30 C) (Figure S1C). To examine further whether NIP7 is required for TORC2 signaling, we investigated whether the nip7-1 mutant phenocopied TORC2 mutants. The nip7-1 mutant indeed exhibited several defects similar to those observed in temperature-sensitive TORC2 mutants (avo3-1 and tor2-21) (Aronova et al., 2008; Beeler et al., 1998; Facchinetti et al., 2008; Helliwell et al., 1998; Kamada et al., 2005; Schmidt et al., 1997). First, the nip7-1 mutant exhibited reduced signaling through the cell wall integrity pathway, as evidenced by decreased MPK1 phosphorylation and PKC1 protein levels, depolarization of the actin cytoskeleton, and restoration of growth in the presence of the osmotic stabilizer sorbitol (Figure 1B and Figure S1D). Second, the nip7-1 mutant showed impaired sphingolipid biosynthesis, 758 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
including hypersensitivity to myriocin, an inhibitor of the first step in the sphingolipid biosynthetic pathway, and restoration of growth in the presence of Ca2+ in a csg2D background (Figure S1E). Third, the nip7-1 mutant showed decreased YPK2 kinase activity, as measured by an in vitro kinase assay with immunopurified YPK2 and, as a control, kinase-dead YPK2K373A (Figure 1C and data not shown for kinase dead). Finally and most importantly, TORC2 kinase activity was reduced in the nip7-1 mutant, as measured by an in vitro kinase assay with TORC2 immunopurified from wild-type cells and nip7-1 mutant cells grown at 30 C (Figure 1D). The above results strongly suggest that NIP7 is required, directly or indirectly, for TORC2 kinase activity and signaling. mNIP7 Is Required for mTORC2 Activity and Ribosome Maturation in Mammals Yeast NIP7 shares 75% identity with an uncharacterized mammalian protein also termed NIP7 (Figure S2A). We examined whether mammalian NIP7 (mNIP7) is required for mTORC2 signaling. mTORC2 directly phosphorylates Ser473 in the hydrophobic motif of Akt and Ser422 in the hydrophobic motif of SGK1 and thereby activates Akt and SGK1 toward substrates such as FoxO3a (Thr32) and NDRG1 (Thr346), respectively (Garcı´aMartı´nez and Alessi, 2008; Sarbassov et al., 2005). mTORC2 also autophosphorylates sites in rictor (Sarbassov et al., 2004; Jacinto et al., 2004) and is required for phosphorylation of Thr450 in the turn motif of Akt (Facchinetti et al., 2008; Ikenoue et al., 2008). Knockdown of mNIP7 in HeLa and HEK293 cells strongly decreased basal and insulin-stimulated phosphorylation of Ser473 and Thr450 in Akt, Thr32 in FoxO3a, Thr346 in NDRG1, and rictor in mTORC2 (Figure 2A, Figure S2B, and data not shown). mNIP7 knockdown had no effect on Erk phosphorylation or mTORC1-dependent S6K phosphorylation (Figure 2A). These findings suggest that mNIP7 is specifically required for mTORC2 signaling. Finally, we observed that mNIP7 knockdown reduced (by 58%) mTORC2 kinase activity in vitro, with no effect on mTORC2 amount or integrity (Figure 2B). These findings indicate that mNIP7 is required for mTORC2 activity and signaling, and not for mTORC2 synthesis, assembly, or stability. Furthermore, NIP7 is required for TORC2 signaling in both yeast and mammalian cells, suggesting a conserved mechanism of TORC2 activation. Is mNIP7 a 60S maturation factor like NIP7 in yeast? Knockdown of mNIP7 reduced the amounts of the 60S ribosomal subunit and the 80S ribosome (80S), with no effect on the amount of the 40S subunit (Figure S2C). Knockdown of mTORC2 component mSIN1 had no effect on the ribosome profile (Figure S2C), indicating that the effect of mNIP7 on ribosome maturation was not due to a defect in signaling downstream of mTORC2. Thus, NIP7 is conserved from yeast to human as a 60S ribosome maturation factor and a TORC2 activator. Ribosomes, but Not Protein Synthesis, Are Required for mTORC2 Signaling Does mNIP7 control mTORC2 via its role in ribosome maturation? To address this question, we examined whether ribosome content affects mTORC2 signaling. Knockdown of ribosomal protein Rpl7 (60S subunit) reduced the amounts of the 60S
Figure 1. NIP7, Identified by a Reverse Suppressor Screen, Is Required for TORC2 Activity In Vivo and In Vitro (A) Schematic representation of the reverse suppressor screen. A wild-type strain (JK9-3da) overexpressing constitutively active YPK2* (YPK2D239A) on an URA3-based plasmid was mutagenized with ethyl-methanesulfonate (EMS). Mutants that could not grow on plates containing 5-FOA were chosen for further analysis. A plasmid-borne NIP7 gene (pNIP7) was isolated by complementation of the nip7-1 mutation. (B) nip7-1 mutant exhibits depolarization of actin cytoskeleton. Wild-type (JK9-3da), avo3-1 (BAS65-2a), and nip7-1 (DS1) cells grown in YPD at 25 C were shifted to 37 C for 6 hr. The actin cytoskeleton was stained with rhodamine-coupled phalloidin. A representative figure of three independent experiments for each strain is shown. (C) nip7-1 mutant shows decreased YPK2 activity. Wild-type cells (JK9-3da) (mock) and wild-type, avo3-1, or nip7-1 cells expressing plasmid-borne YPK2-HA (pYPK2-HA) were grown in YPD at 30 C. YPK2-HA was immunoprecipitated and subjected to in vitro kinase assay using cross-tide as a substrate. Substrate phosphorylation was quantified, and the average ± standard deviation from the mean based on three independent experiments is shown. Immunoprecipitated YPK2-HA was detected by western blotting (bottom). (D) nip7-1 mutant shows decreased TORC2 kinase activity in vitro. Wild-type cells (JK9-3da; mock) and wild-type (DS2; WT) or nip7-1 (DS3; nip7-1) cells expressing AVO3-HA were grown in YPD at 30 C. AVO3-HA was immunoprecipitated and subjected to a TORC2 kinase assay in vitro using recombinant YPK2 protein as a substrate. Also shown are Coomassie blue-stained total YPK2 protein and immunoblot of immunoprecipitated AVO3-HA. See also Figure S1.
subunit and assembled ribosomes (80S monosomes and polysomes), whereas knockdown of ribosomal protein Rps16 (40S subunit) reduced the amounts of the 40S subunit and assembled ribosomes (Figure S2C). We also observed that knockdown of Rpl7 or Rps16 reduced the amounts of other
proteins in the corresponding ribosomal subunit (Figure 3A and Figure S3A), consistent with the published finding that ribosomal subunit assembly determines the level of ribosomal proteins (Idol et al., 2007). Like knockdown of mNIP7, knockdown of either Rpl7 or Rps16 decreased basal and insulin-stimulated Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 759
Figure 2. mNIP7 Is Required for mTORC2 Activity In Vivo and In Vitro (A) siRNA-mediated knockdown of mNIP7 inhibits mTORC2 signaling. HeLa cells, 48 hr after transfection with the indicated siRNA, were harvested (left) or serum starved for 3 hr and then restimulated with insulin for the indicated times before harvesting (right). Phosphorylation and protein levels were determined by immunoblotting with the appropriate antibodies, as indicated. (B) Knockdown of mNIP7 inhibits mTORC2 kinase activity toward Akt with no effect on mTORC2 integrity. rictor immunoprecipitates were immunoblotted with the indicated antibodies to determine mTORC2 integrity. mTORC2 in vitro kinase assay was performed using immunopurified mTORC2 (rictor) and recombinant, kinase-dead Akt as a substrate. See also Figure S2.
phosphorylation of Akt (Ser473 and Thr450), FoxO3a (Thr32), NDRG1 (Thr346), and rictor, with no effect on Erk or S6K phosphorylation (Figure 3A, Figure S2B, and Figure S3A). Finally, we observed that Rpl7 knockdown reduced mTORC2 kinase activity, as determined by an in vitro kinase assay with immunopurified mTORC2 and recombinant, kinase-dead Akt as a substrate. Rpl7 knockdown had no effect on mTORC2 amount or integrity, as determined by unaffected coimmunoprecipitation of mTOR, mSIN1, or mLST8 with rictor (Figure 3B). Thus, assembled ribosomes (not 60S or 40S ribosomal subunits alone) are required for mTORC2 kinase activity and signaling. Furthermore, the above findings suggest that mNIP7 controls mTORC2 via its role in ribosome maturation. mTORC2 promotes cell survival via phosphorylation and activation of Akt, which in turn phosphorylates and inhibits proapoptotic Bad (Brazil et al., 2004; Datta et al., 1997; Jacinto et al., 2006; Yang et al., 2006). To investigate further the physiological relevance of ribosome-mediated mTORC2 regulation, we examined the effect of mNIP7, Rpl7, or Rps16 knockdown on induction of apoptosis by etoposide or hydrogen peroxide (H2O2). Both treatments induce cell death in an Akt-sensitive manner (He et al., 2010; Kim et al., 2001; Wang et al., 2000). Knockdown of mNIP7, Rpl7, or Rps16 enhanced the induction of apoptosis by etoposide or hydrogen peroxide, as indicated by an increase in caspase 3 and PARP cleavage and a decrease in cell viability (Figure 4A and Figure S4). Knockdown of mSIN1 (mTORC2) similarly enhanced apoptosis. Knockdown of Bad 760 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
blocked the proapoptotic effect of mSIN1, mNIP7, Rpl7, or Rps16 knockdown (Figure 4B). Thus, ribosomes appear to regulate mTORC2 signaling in a physiologically relevant manner. Although ribosomal knockdown reduces mTORC2 kinase activity and signaling without affecting mTORC2 synthesis, is the process of protein synthesis per se required for mTORC2 activation? To address this question, we examined whether the protein synthesis inhibitors salubrinal, cycloheximide, anisomycin, or puromycin acutely inhibit mTORC2 signaling. Salubrinal blocks translation initiation by selective inhibition of eIF2a dephosphorylation (Boyce et al., 2005; Cnop et al., 2007) (Figure S3B). Cycloheximide, anisomycin, and puromycin inhibit translation elongation. Unlike knockdown of mNIP7, Rpl7, or Rps16, salubrinal, cycloheximide, anisomycin, and puromycin had no effect on basal or insulin-stimulated Akt Ser473 phosphorylation (Figures S3B–S3D and Figure S5C). All drugs were used under conditions known to affect protein synthesis in the expected manner, as determined by polysome gradient analysis (Figures S3B–S3D and data not shown). Thus, ribosomes mediate mTORC2 signaling independently of protein synthesis. Active mTORC2 Is Associated with the Ribosome To investigate the molecular mechanism by which ribosomes activate mTORC2 signaling, we examined by four unrelated methods whether ribosomes and mTORC2 physically interact. First, we determined whether mTORC2 coimmunoprecipitates with ribosomal protein Rpl26. Endogenous Rpl26 coimmunoprecipitated with endogenous mTOR, rictor, mSIN1, and ribosomal protein Rpl7 (Figure 5A), but not with raptor. This suggests that mTORC2, but not mTORC1, associates with the ribosome. Second, we determined whether mTORC2 copurifies with total ribosomes isolated by sedimentation through a sucrose cushion (see Experimental Procedures). mTOR and rictor, but not raptor, cosedimented with Rpl26, Rpl7, and Rps16 (Figure 5B). Third, although ribosomes activate mTORC2 independently of protein synthesis, we also determined whether mTORC2 cosediments
Figure 3. Ribosomes Are Required for mTORC2 Activity and Signaling (A) siRNA-mediated knockdown of Rpl7 inhibits mTORC2 signaling. HeLa cells, 24 hr after transfection with the indicated siRNA, were harvested (left, basal activity) or serum starved for 3 hr and then restimulated with insulin for the indicated times before harvesting (right, insulinstimulated activity). Phosphorylation and protein levels were determined by immunoblotting with the appropriate antibodies, as indicated. (B) Knockdown of Rpl7 inhibits mTORC2 kinase activity toward Akt with no effect on mTORC2 integrity. rictor immunoprecipitates were immunoblotted with the indicated antibodies. mTORC2 in vitro kinase assay was performed using immunopurified mTORC2 (rictor) and recombinant, kinase-dead Akt as a substrate. See also Figure S3.
with polysomes in a sucrose gradient. Lysates were fractionated in a sucrose gradient to separate polysomes from 80S, 60S, and 40S ribosomes. mTOR, rictor, and mSIN1 were found in both the polysomal and ribosomal fractions, to the same extent as Rpl7, Rpl26, and Rps16 (Figure S5A). Finally, we determined whether mTORC2 copurifies with mRNA-bound ribosomes. mRNAbound ribosomes were purified by pull-down of poly(A) mRNA with oligo(dT) cellulose (Figure S5B) (Ceci et al., 2003). In this experiment, mTOR and rictor copurified with Rpl26, Rpl7, and the 40S ribosomal protein RACK1. RNase A treatment of a lysate before pull-down prevented isolation of any of the above proteins (Figure S5B), confirming interaction of mTORC2 with mRNA-bound ribosomes. The above data taken together suggest that mTORC2 physically interacts with translating (mRNA-bound) and nontranslating 80S ribosomes. An interaction between mTORC2 and the 80S ribosome is also supported by published mass spectrometry studies that identified several large and small subunit ribosomal proteins in mTORC2 immunoprecipitates (Pearce et al., 2007; Thedieck et al., 2007). We also note that nonionic detergent was required during any of our four above ribosome purifications to detect copurification of mTORC2 and that protein synthesis inhibitors had no effect on the interaction between mTORC2 and Rpl26 (Figure S5C). To obtain more insight into the mTORC2-ribosome interaction, we performed coimmunoprecipitation experiments after knockdown of mTORC2 subunits or ribosomal proteins. First, the amount of mTOR in Rpl26 immunoprecipitates from rictor or mSIN1 knockdown cells was significantly reduced compared to control cells (Figure S6A), suggesting that mTORC2 interacts
with the ribosome via rictor and/or mSIN1. We could not distinguish a requirement specifically for rictor or mSIN1 because knockdown of either rictor or mSIN1 alone also results in loss of the other. Second, knockdown of Rpl7 (60S subunit) abolished the interaction between mTORC2 and Rpl26, whereas knockdown of Rps16 (40S subunit) only moderately decreased the interaction between mTORC2 and Rpl26 (60S subunit) (Figure S6B). These results suggest that mTORC2 associates with the ribosome via rictor and/or mSIN1 binding to the 60S subunit. The above findings suggest that ribosomes bind and activate mTORC2. To test this model further, we investigated whether ribosome-bound mTORC2 is indeed active. We performed an mTORC2 kinase assay with ribosomes purified either by Rpl26 immunoprecipitation or by sedimentation through a sucrose cushion, as described above. In both cases, a ribosome-associated kinase phosphorylated Ser473 in recombinant, kinase-dead Akt (Figure 5C). The mTOR inhibitor PP242 (Feldman et al., 2009) abolished the in vitro phosphorylation of Ser473 (Figure 5C), confirming that the ribosome-associated Ser473 kinase was mTORC2. Thus, ribosome-associated mTORC2 is active. Furthermore, we analyzed whether mTORC2 kinase activity is required for mTORC2-ribosome interaction. Treatment of HeLa cells with PP242 abolished Ser473 phosphorylation, as expected, but had no effect on mTORC2Rpl26 interaction (Figure S6D), suggesting that mTORC2 activity is not required for mTORC2-ribosome association. Insulin-PI3K Signaling Stimulates mTORC2-Ribosome Association We next investigated whether mTORC2 association with the ribosome is regulated by insulin. Insulin treatment of serumstarved cells significantly increased the amount of mTOR and rictor that coimmunoprecipitated with Rpl26 (Figure 6A). Importantly, this stimulation of the interaction between mTORC2 and Rpl26 correlated with phosphorylation of the mTORC2 target site Ser473 in endogenous Akt. Thus, insulin stimulates mTORC2-ribosome association. Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 761
Figure 4. Ribosome or mTORC2 Knockdown Enhances Stress-Induced, Bad-Dependent Apoptosis (A) siRNA-mediated knockdown of mSIN1, mNIP7, Rpl7, or Rps16 increases etoposide-induced apoptosis. HeLa cells were transfected with the various siRNAs and then treated with 25 mM etoposide for the indicated times (top) or for 24 hr (bottom). Extracts were analyzed by immunoblotting to assess efficiency of siRNA knockdown and induction of apoptosis with the indicated antibodies (top). Apoptosis was assessed by blotting for cleaved PARP (cPARP) and cleaved caspase 3 (cCaspase 3). Cells were fixed and stained with DAPI and cleaved PARP antibody to detect apoptotic cells (bottom). (B) Bad is required for enhanced apoptosis in cells with knockdown of mSIN1, mNIP7, Rpl7, or Rps16. HeLa cells were transfected with the indicated siRNA and then treated with 25 mM etoposide for 24 hr. Extracts were analyzed by immunoblotting to assess the induction of apoptosis with cleaved PARP antibody. The efficiency of Bad knockdown is shown. See also Figure S4.
Insulin activates mTORC2 via PI3K (Frias et al., 2006; Garcı´aMartı´nez and Alessi, 2008; Yang et al., 2006). To investigate whether PI3K signaling regulates mTORC2 association with the ribosome, we examined whether the mTORC2-Rpl26 interaction was affected upon inhibition or hyperactivation of PI3K signaling. Inhibition of PI3K strongly decreased mTORC2Rpl26 interaction, as determined by a reduction in the amount of mTOR and rictor that coimmunoprecipitated with Rpl26 in 762 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
lysates from cells treated with the PI3K inhibitor LY294002 (Figure 6B and Figure S6D). In contrast, hyperactivation of PI3K signaling by knockdown of PTEN, a negative regulator of PI3K signaling, increased the amount of mTOR and rictor in Rpl26 immunoprecipitates, compared to control cells (Figure 6C). Again, the change in mTORC2-Rpl26 interaction upon PI3K inhibition or hyperactivation was paralleled by a corresponding change in phosphorylation of the mTORC2 target site Ser473
Figure 5. Active mTORC2 Is Associated with the Ribosome (A) Endogenous Rpl26 coimmunoprecipitates with endogenous mTOR, rictor, and mSIN1. Rpl26 and mock immunoprecipitations were performed with HeLa cell extracts and analyzed by immunoblotting with the indicated antibodies. (B) mTORC2 associates with ribosomes. Ribosomes were purified from HeLa cells by sedimentation through a sucrose cushion. Ribosomes were probed with the indicated antibodies (left) or were resedimented through a sucrose gradient to monitor 40S, 60S, and 80S ribosomes and polysomes. The absorbance profile of the sucrose gradient was determined at 254nm (right). (C) Ribosome-associated mTORC2 is active. mTORC2 kinase assays were performed with Rpl26 or mock immunoprecipitates using recombinant, kinase-dead Akt as a substrate and in the presence or absence of the mTOR inhibitor PP242 (left). mTORC2 kinase assay was performed with ribosomes purified as described in (B), with recombinant, kinase-dead Akt as a substrate and in the presence or absence of PP242 (right). See also Figure S5.
in endogenous Akt (Figures 6B and 6C). Taken together, the above data suggest that insulin stimulates mTORC2-ribosome association in a physiologically relevant manner via PI3K. mTORC2-Ribosome Interaction Promotes Akt Signaling in Cancer Cells PI3K-dependent association of mTORC2 with the ribosome, ribosome-mediated mTORC2 activation, and the physiological
relevance of mTORC2-ribosome interaction in promoting cell survival prompted us to analyze mTORC2-ribosome association in cancer cells with hyperactive PI3K signaling. First, we examined whether PTEN-deficient metastatic melanoma cells have elevated mTORC2 activity. Approximately 60% of metastatic melanomas have reduced PTEN expression and elevated Akt phosphorylation (Robertson, 2005; Stahl et al., 2004). We analyzed PTEN expression and Akt Ser473 phosphorylation in Figure 6. Insulin-PI3K Signaling mTORC2-Ribosome Association
Stimulates
(A) Insulin stimulates mTORC2-ribosome association. HeLa cells were serum starved and then restimulated with insulin for the indicated time. Rpl26 immunoprecipitates (IP: Rpl26) and cell extracts (lysate) were immunoblotted with the indicated antibodies. (B) PI3K inhibition decreases mTORC2-ribosome association. HeLa cells were treated with LY294002 (50 mM final concentration) for 30 min before harvesting. Rpl26 immunoprecipitates and cells extracts (lysate) were immunoblotted with the indicated antibodies. (C) siRNA-mediated knockdown of PTEN increases mTORC2-ribosome association. HeLa cells were transfected with the indicated siRNA and then harvested after 48 hr. Rpl26 immunoprecipitates and cell extracts (lysate) were analyzed by immunoblotting with the indicated antibodies. See also Figure S6.
Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 763
Figure 7. mTORC2-Ribosome Interaction Promotes Akt Signaling in Cancer Cells (A) Akt Ser473 phosphorylation inversely correlates with PTEN expression in human melanoma cells. Cell extracts from six human melanoma cell lines were analyzed by immunoblotting with indicated antibodies. (B) PTEN-deficient cell lines exhibit increased mTORC2ribosome association. Ribosomes were purified from PTEN-deficient (NA8 and ME67) and PTEN-positive melanoma cell lines (Juso) by sedimentation through a sucrose cushion. Ribosomes were probed with the indicated antibodies. (C) Knockdown of mSIN1 or Rpl7 induces apoptosis in PTEN-deficient (NA8 and ME67) melanoma cell lines. Cells were transfected with the indicated siRNA and harvested after 72 hr. Extracts were analyzed by immunoblotting to check the efficiency of the knockdown and the induction of apoptosis (PARP and caspase 3 cleavage) with the indicated antibodies. See also Figure S7.
six human metastatic melanoma cell lines generated from different patients (Certa et al., 2001; Deschodt-Lanckman et al., 1990; Gervois et al., 1996). PTEN expression inversely correlated with Akt Ser473 phosphorylation (Figure 7A). For further study, we chose the two cell lines NA8 and ME67 with low PTEN expression and high Akt Ser473 phosphorylation (mTORC2 activity) and, as a control, the cell line Juso with the opposite signaling profile. To examine mTORC2-ribosome association, ribosomes were isolated from the three cell lines and probed for mTOR, rictor, and mSIN1. In all cases, the total yield of ribosomes obtained by sedimentation through a sucrose cushion was similar. However, the PTEN-deficient NA8 and ME67 cells exhibited significantly increased levels of mTORC2 associated with the ribosomal fraction as compared to the PTEN-positive Juso cells (Figure 7B), consistent with our PTEN knockdown studies in HeLa cells described above (Figure 6C). We also examined mTORC2-ribosome interaction in colon cancer cells harboring an activating mutation in the PI3K gene PIK3CA. The PIK3CA mutant cell lines HT29 and HCT116 exhibited both Akt Ser473 hyperphosphorylation (mTORC2 activity) and increased mTORC2-ribosome association as compared to SW60 colon cells harboring an unaltered PIK3CA gene (Figure S7A). Thus, mTORC2-ribosome interaction correlates with mTORC2 activity in both melanoma and colon cancer cells. To investigate further the physiological relevance of mTORC2 activation via ribosome association, we examined the effect of mSIN1 and Rpl7 knockdown on mTORC2 signaling and cell survival in NA8, ME67, and Juso cells. Knockdown of either mSIN1 or Rpl7 decreased Akt Ser473 phosphorylation 764 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
(Figure 7C), supporting a role for ribosomedependent activation of mTORC2 in cancer. Furthermore, knockdown of either mSIN1 or Rpl7 enhanced apoptosis both in the presence and absence of etoposide. Interestingly, the increase in apoptosis was more pronounced in the PTEN-deficient NA8 and ME67 cells than in Juso cells, suggesting that NA8 and ME67 cells are more ‘‘addicted’’ to mTORC2 (Figure 7C and Figure S7B). The above results taken together suggest that mTORC2-ribosome association, mediating PI3K-mTORC2-Akt signaling and cell survival, is functionally important in cancer cells. DISCUSSION We investigated the upstream regulation of TORC2. A genetic screen in yeast and subsequent studies in mammalian cells revealed that ribosomes are upstream of TORC2. In particular, we found the following. First, the genetic screen in yeast revealed that knocking down ribosome biogenesis inhibits TORC2 kinase activity in vitro (Figure 1D) and TORC2 signaling in vivo (Figures 1B and 1C and Figure S1). Second, the ribosome is required for mTORC2 activity in vitro (Figure 2B and Figure 3B) and mTORC2 signaling in vivo in mammalian cells (Figure 2A, Figure 3A, and Figure S3A). Knockdown of ribosome maturation factor mNIP7 or ribosomal proteins (Rpl7 or Rps16) in mammalian cells decreased mTORC2 kinase activity and mTORC2 signaling. Third, the ribosome (translating or nontranslating) interacts directly with mTORC2 (Figures 5 and Figure S5). mTORC2 copurified with ribosomes isolated by four independent methods. Furthermore, mTORC2 copurified with ribosomes isolated from growing cells or cells treated with protein synthesis inhibitors. Fourth, ribosome-bound mTORC2 is active (Figure 5C), and ribosome-free mTORC2 is inactive in vitro (Figure 2B, Figure 3B, and Figure S6B). Fifth, PI3K-dependent insulin signaling stimulates binding of the ribosome to mTORC2 (Figure 6 and Figure S6D) with the same kinetics that it stimulates mTORC2
activation (Figure 6A). Finally, the mTORC2-ribosome interaction correlates with mTORC2 activity in both melanoma and colon cancer cells (Figure 7 and Figure S7). Melanoma and colon cancer cells with high PI3K activity (due to loss of PTEN or an activating mutation in the PI3K gene) exhibited both enhanced mTORC2-ribosome interaction and increased mTORC2 activity. Our findings suggest that the translating or nontranslating 80S ribosome binds and activates mTORC2 in response to growth factor-stimulated PI3K signaling (Figure S7C). TORC2-ribosome association is a mechanism of TORC2 activation that is likely conserved from unicellular yeast to human. The ribosome is presumably a primordial activator of TORC2 onto which growth factor signaling was grafted during the evolution of multicellularity. In a parallel and complementary study, Oh et al. also showed that mTORC2 associates with the ribosome (Oh et al., 2010). Furthermore, they showed that mTORC2 phosphorylates the Akt turn motif cotranslationally and the Akt hydrophobic motif posttranslationally. Thus, Oh et al. investigated the role of the mTORC2-ribosome interaction in downstream signaling by mTORC2. Our study addresses the separate issue of upstream regulation of mTORC2. We show that an mTORC2-ribosome interaction activates mTORC2, and this activation is independent of translation. In other words, an mTORC2-ribosome interaction activates mTORC2 regardless of whether mTORC2 is phosphorylating a substrate co- or posttranslationally. We note that Oh et al. did not examine a requirement for the ribosome in posttranslational phosphorylation. A connection between ribosomes and TOR signaling is well established. TORC1 activates ribosome biogenesis and protein synthesis and inhibits autophagy as key readouts in the control of cell growth. Why should ribosomes control TORC2? Ribosome content determines growth capacity of the cell and TORC2 regulates growth-related processes. Thus, regulation of TORC2 by ribosomes ensures that TORC2 is not inappropriately activated in cells that are unable to grow. The above also implies that TORC1, via activation of ribosome biogenesis and inhibition of autophagy-mediated ribosome turnover, indirectly controls TORC2. Indeed, Sarbassov et al. (Sarbassov et al., 2006) have shown that inhibition of mTORC1 by long-term rapamycin treatment indirectly inhibits mTORC2. Our findings suggest that the effect of rapamycin on mTORC2 is due, at least in part, to a reduction in ribosome content. Interestingly, the literature also indicates that ribosomal defects induce apoptotic cell death, although the underlying mechanism is not understood (Warner and McIntosh, 2009). We find that ribosomal defects inhibit mTORC2 and its downstream effector Akt, which in turn leads to Bad-dependent apoptosis. Thus, our findings also provide a mechanism for the induction of apoptosis by a ribosomal defect. Our findings suggest that ribosomes bind and activate mTORC2 directly. The fraction of total mTORC2 that associates with ribosomes varies depending on the cell type and the growth conditions. For example, under normal growth conditions, 20% of total rictor (mTORC2) was associated with ribosomes in HeLa cells, whereas 30%–40% of rictor was associated with ribosomes in PTEN-deficient cells such as melanoma and PTEN knockdown cells. Thus, ribosome association appears to be
a major if not the sole mechanism of TORC2 activation. Consideration of the fraction of total ribosomes that associate with mTORC2 is also potentially informative. Given that ribosomes are 100- to 1000-fold more abundant than signaling kinases such as mTORC2, only a small fraction of total ribosomes bind mTORC2. This excess of ribosomes would require strong regulation of mTORC2-ribosome binding by upstream PI3K signaling to achieve physiologically relevant regulation of mTORC2. Alternatively, TORC2 could be regulated by a specific subpopulation of ribosomes. Previous studies have demonstrated that TORC2 is associated with membranes, including the endoplasmic reticulum (ER) and Golgi apparatus, and that mTORC2 isolated from ER microsomes phosphorylates Akt Ser473 in vitro (Drenan et al., 2004; Hresko and Mueckler, 2005; Liu and Zheng, 2007; Schroder et al., 2007; Sturgill et al., 2008). These findings suggest that TORC2 might associate specifically with membrane-bound ribosomes. In support of this notion, we observed that mTORC2 copurifies with ribosomes only when the ribosomes are isolated in the presence of detergent. Membrane-bound ribosomes constitute 10% of total ribosomes and include both translating and nontranslating ribosomes (Seiser and Nicchitta, 2000). It is also interesting to note that Komili et al. (Komili et al., 2007) have proposed a ribosome code in which there is a specialization of ribosomes for specific cellular process. Our findings are consistent with other studies proposing the ribosome as a kinase platform. The two kinases PKCbII and Pim-1 are associated with the ribosome via RACK1 and Rps19, respectively (Ceci et al., 2003; Chiocchetti et al., 2005; Grosso et al., 2008b). Overall, the ribosome appears to be a signaling platform for mTORC2 and other kinases. Furthermore, ribosomal proteins have been shown to modulate the activity of NF-kB, p53, and c-Myc (Lindstro¨m, 2009). PI3K-Akt signaling is upregulated and contributes to tumorigenesis in 60% of advanced-stage melanomas (Stahl et al., 2003, 2004). PTEN expression or Akt inhibition increases sensitivity of melanoma cells to apoptosis-inducing agents and prevents tumor development (Madhunapantula and Robertson, 2009; Stahl et al., 2004). Our data, with melanoma, cancer colon, and HeLa cells, suggest that PI3K signaling promotes Akt phosphorylation via stimulation of mTORC2-ribosome binding. Furthermore, disruption of the mTORC2-ribosome supercomplex selectively induces apoptosis in PTEN-deficient melanoma cells (Figure 7C). The extent to which cells of other cancers require ribosome-dependent mTORC2 activation is unclear, although we expect that other cancers driven by mutations promoting PI3K signaling may also depend on mTORC2-ribosome association. Disrupting the mTORC2-ribosome interaction may be a useful strategy in the treatment of melanomas, colon carcinomas, and possibly other cancers. Several findings suggest that upregulation of the protein synthesis machinery contributes to the development of cancer and other diseases (Ruggero and Pandolfi, 2003). Consistent with our findings, a ribosomal protein deficiency inhibits Aktdriven tumorigenesis (Hsieh et al., 2010). Furthermore, the myc oncogene enhances ribosome biogenesis, and myc oncogenicity in mice can be blocked by mutations in ribosomal protein genes (Barna et al., 2008; Ruggero, 2009). Our findings and the observation that mTORC2 is required for tumor progression in Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 765
at least some cancers (Guertin et al., 2009; Gulhati et al., 2009; Masri et al., 2007) suggest that myc and increased ribosomal content may promote tumorigenicity via stimulation of mTORC2 and its downstream effector Akt. Curiously, our genetic screen in yeast yielded 44 TORC2 mutants but only a single mutant (nip7-1) that was defective in ribosome biogenesis. Why did we not obtain more mutants that were defective in ribosomal maturation factors or ribosomal proteins? First, ribosomal genes are duplicated in yeast, thereby precluding identification of recessive, loss-of-function mutations in these genes. Second, YPK2 is downstream of the ribosome in activation of TORC2 but is not downstream of the ribosome in mediating protein synthesis, precluding full suppression of a ribosome biogenesis defect by YPK2*. YPK2* only partly suppresses the nip7-1 mutation and only at semipermissive temperature. EXPERIMENTAL PROCEDURES Detailed protocols for apoptosis assays and statistical analyses can be found in the Extended Experimental Procedures. Yeast Strains, Media, Kinase Assays, Actin Staining, and Antibodies Yeast strains used in this study are listed in the Table S1. All strains are isogenic derivatives of JK9-3da. Plasmids used in this study are described in Table S2. Standard techniques and media were used for yeast manipulation (Kamada et al., 2005; Loewith et al., 2002). Unless indicated otherwise, cells were grown in rich YPD medium. YPK2 and TORC2 kinase assays were performed as described previously (Casamayor et al., 1999; Kamada et al., 2005; Loewith et al., 2002; Wullschleger et al., 2005). Rhodamine phalloidin staining of polymerized actin was performed as described (Loewith et al., 2002; Mulet et al., 2006). Immunoprecipitations were performed as described previously (Loewith et al., 2002; Wullschleger et al., 2005). Reverse Suppressor Screen A wild-type strain (JK9-3da) was transformed with an URA3-based plasmid overexpressing YPK2* (pYPK2*) (Figure 1A). Cells were randomly mutagenized with 100 mM ethyl-methanesulfonate (EMS) (Sigma) for 15 min in SD-Ura medium, washed, and then allowed to recover in SD-Ura medium without mutagen for 2–4 hr. Cells were plated on solid SD-Ura or SD supplemented with 5-FOA and incubated at 30 C. The SD medium supplemented with 5-FOA counterselected against URA3 such that only those cells that had spontaneously lost the URA3-based pYPK2* plasmid were able to form a colony. Mutants that were unable to grow on SD 5-FOA (and hence in the absence of YPK2*) were isolated from the master SD-Ura plate. 45,000 colonies from mutagenized cells were screened, from which 45 mutants were isolated and the corresponding mutations were identified. Mutated genes were isolated by complementation with a LEU2 centromeric plasmid-based yeast genomic library. Complementing members of the genomic library were selected by growth on SD-Leu medium containing 5-FOA. Genomic inserts of library-derived plasmids were identified by sequencing. Complementation with subclones of isolated inserts identified the complementing ORF within a given insert. Sequencing of the genomic copy of NIP7 and meiotic segregation studies confirmed that NIP7 was indeed the relevant mutant gene. nip7-1 was found to be temperature sensitive upon subsequent characterization. For experiments at semipermissive temperature (30 C), nip7-1 cells were grown in YPD at permissive temperature (25 C), diluted to OD600 = 0.1, and grown at 30 C to approximately OD600 = 0.6–0.8. For experiments at the nonpermissive temperature (37 C), nip7-1 mutant was grown in YPD at 25 C and then shifted to 37 C for 6 hr.
Thedieck et al., 2007). In brief, cells were seeded and grown for 48 hr in DMEM supplemented with 10% serum (basal conditions). Cells were starved of serum for 3 hr before restimulation with 100 nM insulin (Sigma). For mSIN1, mNIP7, Rpl7, Rps16, PTEN, or Bad knockdown, a pool of four different synthetic siRNA or of the appropriate control siRNA (Dharmacon) was used as described (Thedieck et al., 2007). All transfections were done according to the manufacturer’s instructions (Lipofectamine, Invitrogen transfection). Protein extracts were prepared as previously described (Jacinto et al., 2004; Thedieck et al., 2007), resolved on SDS-PAGE, and transferred to nitrocellulose membranes (Protran, Whatman). Immunoprecipitation, immunoblotting, and mTORC2 kinase assays were performed as previously described (Jacinto et al., 2004; Thedieck et al., 2007).
Polysome Profiles, Ribosome Purification, and Poly(A) mRNA Pull-Down Polysome analysis using sucrose gradients was performed as described previously (Grosso et al., 2008b; Idol et al., 2007). For ribosome purification by sedimentation through a sucrose cushion, HeLa or melanoma cells were washed in PBS, trypsinized, and lysed in buffer A (50 mM Tris-HCl [pH 7.4], 100 mM NaCl, 30 mM MgCl2, 0.3% CHAPS, 100 ug/ml cycloheximide, 40 U/ml RNase inhibitor, protease inhibitor cocktail, and 100 ug/ml cycloheximide). Whole-cell extracts were clarified at 4 C, 10 min at 15,000 3 g. Extracts were loaded on a 30% sucrose cushion in 50 mM Tris-acetate (pH 7.5), 50 mM NH4Cl, 12 mM MgCl2, and 1 mM DTT and ultracentrifuged for 17 hr in a SW41Ti Beckman rotor at 39,000 rpm. For the mTORC2 kinase assay, the ribosomal pellet was resuspended in mTORC2 kinase buffer (Jacinto et al., 2004; Thedieck et al., 2007). Ribosome-sucrose gradient fractionation was performed as previously described (Grosso et al., 2008a). For poly(A) pull-down, HeLa cells were washed with PBS, trypsinized, and lysed in buffer A. Whole-cell extracts were clarified at 4 C, 10 min at 8000 3 g. Lysates corresponding to 5 3 107 cells were incubated with oligo(dT) cellulose (Invitrogen) for 1 hr at room temperature. After incubation, the oligo(dT) cellulose was pelleted and washed five times with buffer A. The bound fraction was eluted with elution buffer (100 mM Tris [pH 7.4], 500 mM NaCl, 10 mM EDTA, 1% sodium dodecyl sulfate (SDS), and 5 mM DTT). Purified ribosome fractions and the bound and unbound fractions after poly(A) pull-down were concentrated with Vivaspin 500 (Sartorius Stedim) and analyzed by immunoblotting with the indicated antibodies.
SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, seven figures, and two tables and can be found with this article online at doi:10.1016/j.cell.2011.02.014.
ACKNOWLEDGMENTS We thank Rachel Idol and Stefano Grosso for valuable discussions and technical assistance and Giulio Spagnoli for the melanoma cell lines. We acknowledge support from the Swiss National Science Foundation, SystemsX.ch, the Swiss Cancer League, the Louis Jeantet Foundation, and the Canton of Basel. We declare that no competing interests exist. Received: July 14, 2010 Revised: October 20, 2010 Accepted: February 7, 2011 Published: March 3, 2011
REFERENCES Cell Culture, Immunoprecipitations, Immunoblotting, and mTORC2 Kinase Assay HeLa, melanoma cells, and colon cancer cells were cultured, transfected, stimulated, and harvested as described previously (Jacinto et al., 2004;
766 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
Aronova, S., Wedaman, K., Aronov, P.A., Fontes, K., Ramos, K., Hammock, B.D., and Powers, T. (2008). Regulation of ceramide biosynthesis by TOR complex 2. Cell Metab. 7, 148–158.
Avruch, J., Long, X., Ortiz-Vega, S., Rapley, J., Papageorgiou, A., and Dai, N. (2009). Amino acid regulation of TOR complex 1. Am. J. Physiol. Endocrinol. Metab. 296, E592–E602.
Gervois, N., Guilloux, Y., Diez, E., and Jotereau, F. (1996). Suboptimal activation of melanoma infiltrating lymphocytes (TIL) due to low avidity of TCR/MHCtumor peptide interactions. J. Exp. Med. 183, 2403–2407.
Barna, M., Pusic, A., Zollo, O., Costa, M., Kondrashov, N., Rego, E., Rao, P.H., and Ruggero, D. (2008). Suppression of Myc oncogenic activity by ribosomal protein haploinsufficiency. Nature 456, 971–975.
Grosso, S., Volta, V., Sala, L.A., Vietri, M., Marchisio, P.C., Ron, D., and Biffo, S. (2008a). PKCbetaII modulates translation independently from mTOR and through RACK1. Biochem. J. 415, 77–85.
Beeler, T., Bacikova, D., Gable, K., Hopkins, L., Johnson, C., Slife, H., and Dunn, T. (1998). The Saccharomyces cerevisiae TSC10/YBR265w gene encoding 3-ketosphinganine reductase is identified in a screen for temperature-sensitive suppressors of the Ca2+-sensitive csg2Delta mutant. J. Biol. Chem. 273, 30688–30694.
Grosso, S., Volta, V., Vietri, M., Gorrini, C., Marchisio, P.C., and Biffo, S. (2008b). Eukaryotic ribosomes host PKC activity. Biochem. Biophys. Res. Commun. 376, 65–69.
Boyce, M., Bryant, K.F., Jousse, C., Long, K., Harding, H.P., Scheuner, D., Kaufman, R.J., Ma, D., Coen, D.M., Ron, D., and Yuan, J. (2005). A selective inhibitor of eIF2alpha dephosphorylation protects cells from ER stress. Science 307, 935–939. Brazil, D.P., Yang, Z.Z., and Hemmings, B.A. (2004). Advances in protein kinase B signalling: AKTion on multiple fronts. Trends Biochem. Sci. 29, 233–242. Casamayor, A., Torrance, P.D., Kobayashi, T., Thorner, J., and Alessi, D.R. (1999). Functional counterparts of mammalian protein kinases PDK1 and SGK in budding yeast. Curr. Biol. 9, 186–197.
Guertin, D.A., Stevens, D.M., Saitoh, M., Kinkel, S., Crosby, K., Sheen, J.H., Mullholland, D.J., Magnuson, M.A., Wu, H., and Sabatini, D.M. (2009). mTOR complex 2 is required for the development of prostate cancer induced by Pten loss in mice. Cancer Cell 15, 148–159. Gulhati, P., Cai, Q., Li, J., Liu, J., Rychahou, P.G., Qiu, S., Lee, E.Y., Silva, S.R., Bowen, K.A., Gao, T., and Evers, B.M. (2009). Targeted inhibition of mammalian target of rapamycin signaling inhibits tumorigenesis of colorectal cancer. Clin. Cancer Res. 15, 7207–7216. Hara, K., Maruki, Y., Long, X., Yoshino, K., Oshiro, N., Hidayat, S., Tokunaga, C., Avruch, J., and Yonezawa, K. (2002). Raptor, a binding partner of target of rapamycin (TOR), mediates TOR action. Cell 110, 177–189.
Ceci, M., Gaviraghi, C., Gorrini, C., Sala, L.A., Offenha¨user, N., Marchisio, P.C., and Biffo, S. (2003). Release of eIF6 (p27BBP) from the 60S subunit allows 80S ribosome assembly. Nature 426, 579–584.
He, H.N., Wang, X., Zheng, X.L., Sun, H., Shi, X.W., Zhong, Y.J., Huang, B., Yang, L., Li, J.K., Liao, L.C., et al. (2010). Concurrent blockade of the NF-kappaB and Akt pathways potently sensitizes cancer cells to chemotherapeutic-induced cytotoxicity. Cancer Lett. 295, 38–43.
Certa, U., Seiler, M., Padovan, E., and Spagnoli, G.C. (2001). High density oligonucleotide array analysis of interferon- alpha2a sensitivity and transcriptional response in melanoma cells. Br. J. Cancer 85, 107–114.
Helliwell, S.B., Howald, I., Barbet, N., and Hall, M.N. (1998). TOR2 is part of two related signaling pathways coordinating cell growth in Saccharomyces cerevisiae. Genetics 148, 99–112.
Chiocchetti, A., Gibello, L., Carando, A., Aspesi, A., Secco, P., Garelli, E., Loreni, F., Angelini, M., Biava, A., Dahl, N., et al. (2005). Interactions between RPS19, mutated in Diamond-Blackfan anemia, and the PIM-1 oncoprotein. Haematologica 90, 1453–1462.
Hresko, R.C., and Mueckler, M. (2005). mTOR.RICTOR is the Ser473 kinase for Akt/protein kinase B in 3T3-L1 adipocytes. J. Biol. Chem. 280, 40406–40416.
Cnop, M., Ladriere, L., Hekerman, P., Ortis, F., Cardozo, A.K., Dogusan, Z., Flamez, D., Boyce, M., Yuan, J., and Eizirik, D.L. (2007). Selective inhibition of eukaryotic translation initiation factor 2 alpha dephosphorylation potentiates fatty acid-induced endoplasmic reticulum stress and causes pancreatic beta-cell dysfunction and apoptosis. J. Biol. Chem. 282, 3989–3997.
Hsieh, A.C., Costa, M., Zollo, O., Davis, C., Feldman, M.E., Testa, J.R., Meyuhas, O., Shokat, K.M., and Ruggero, D. (2010). Genetic dissection of the oncogenic mTOR pathway reveals druggable addiction to translational control via 4EBP-eIF4E. Cancer Cell 17, 249–261.
Cybulski, N., and Hall, M.N. (2009). TOR complex 2: a signaling pathway of its own. Trends Biochem. Sci. 34, 620–627.
Idol, R.A., Robledo, S., Du, H.Y., Crimmins, D.L., Wilson, D.B., Ladenson, J.H., Bessler, M., and Mason, P.J. (2007). Cells depleted for RPS19, a protein associated with Diamond Blackfan Anemia, show defects in 18S ribosomal RNA synthesis and small ribosomal subunit production. Blood Cells Mol. Dis. 39, 35–43.
Datta, S.R., Dudek, H., Tao, X., Masters, S., Fu, H., Gotoh, Y., and Greenberg, M.E. (1997). Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91, 231–241.
Ikenoue, T., Inoki, K., Yang, Q., Zhou, X., and Guan, K.L. (2008). Essential function of TORC2 in PKC and Akt turn motif phosphorylation, maturation and signalling. EMBO J. 27, 1919–1931.
Deschodt-Lanckman, M., Vanneste, Y., Loir, B., Michel, A., Libert, A., Ghanem, G., and Lejeune, F. (1990). Degradation of alpha-melanocyte stimulating hormone (alpha-MSH) by CALLA/endopeptidase 24.11 expressed by human melanoma cells in culture. Int. J. Cancer 46, 1124–1130.
Jacinto, E., Facchinetti, V., Liu, D., Soto, N., Wei, S., Jung, S.Y., Huang, Q., Qin, J., and Su, B. (2006). SIN1/MIP1 maintains rictor-mTOR complex integrity and regulates Akt phosphorylation and substrate specificity. Cell 127, 125–137.
Drenan, R.M., Liu, X., Bertram, P.G., and Zheng, X.F. (2004). FKBP12-rapamycin-associated protein or mammalian target of rapamycin (FRAP/mTOR) localization in the endoplasmic reticulum and the Golgi apparatus. J. Biol. Chem. 279, 772–778.
Jacinto, E., Loewith, R., Schmidt, A., Lin, S., Ru¨egg, M.A., Hall, A., and Hall, M.N. (2004). Mammalian TOR complex 2 controls the actin cytoskeleton and is rapamycin insensitive. Nat. Cell Biol. 6, 1122–1128.
Facchinetti, V., Ouyang, W., Wei, H., Soto, N., Lazorchak, A., Gould, C., Lowry, C., Newton, A.C., Mao, Y., Miao, R.Q., et al. (2008). The mammalian target of rapamycin complex 2 controls folding and stability of Akt and protein kinase C. EMBO J. 27, 1932–1943.
Jacinto, E., and Lorberg, A. (2008). TOR regulation of AGC kinases in yeast and mammals. Biochem. J. 410, 19–37. Kamada, Y., Fujioka, Y., Suzuki, N.N., Inagaki, F., Wullschleger, S., Loewith, R., Hall, M.N., and Ohsumi, Y. (2005). Tor2 directly phosphorylates the AGC kinase Ypk2 to regulate actin polarization. Mol. Cell. Biol. 25, 7239–7248.
Feldman, M.E., Apsel, B., Uotila, A., Loewith, R., Knight, Z.A., Ruggero, D., and Shokat, K.M. (2009). Active-site inhibitors of mTOR target rapamycin-resistant outputs of mTORC1 and mTORC2. PLoS Biol. 7, e38.
Kim, A.H., Khursigara, G., Sun, X., Franke, T.F., and Chao, M.V. (2001). Akt phosphorylates and negatively regulates apoptosis signal-regulating kinase 1. Mol. Cell. Biol. 21, 893–901.
Frias, M.A., Thoreen, C.C., Jaffe, J.D., Schroder, W., Sculley, T., Carr, S.A., and Sabatini, D.M. (2006). mSin1 is necessary for Akt/PKB phosphorylation, and its isoforms define three distinct mTORC2s. Curr. Biol. 16, 1865–1870.
Kim, D.H., Sarbassov, D.D., Ali, S.M., King, J.E., Latek, R.R., ErdjumentBromage, H., Tempst, P., and Sabatini, D.M. (2002). mTOR interacts with raptor to form a nutrient-sensitive complex that signals to the cell growth machinery. Cell 110, 163–175.
Garcı´a-Martı´nez, J.M., and Alessi, D.R. (2008). mTOR complex 2 (mTORC2) controls hydrophobic motif phosphorylation and activation of serum- and glucocorticoid-induced protein kinase 1 (SGK1). Biochem. J. 416, 375–385.
Komili, S., Farny, N.G., Roth, F.P., and Silver, P.A. (2007). Functional specificity among ribosomal proteins regulates gene expression. Cell 131, 557–571.
Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc. 767
Laplante, M., and Sabatini, D.M. (2009). mTOR signaling at a glance. J. Cell Sci. 122, 3589–3594. Lindstro¨m, M.S. (2009). Emerging functions of ribosomal proteins in genespecific transcription and translation. Biochem. Biophys. Res. Commun. 379, 167–170. Liu, X., and Zheng, X.F. (2007). Endoplasmic reticulum and Golgi localization sequences for mammalian target of rapamycin. Mol. Biol. Cell 18, 1073–1082. Loewith, R., Jacinto, E., Wullschleger, S., Lorberg, A., Crespo, J.L., Bonenfant, D., Oppliger, W., Jenoe, P., and Hall, M.N. (2002). Two TOR complexes, only one of which is rapamycin sensitive, have distinct roles in cell growth control. Mol. Cell 10, 457–468. Madhunapantula, S.V., and Robertson, G.P. (2009). The PTEN-AKT3 signaling cascade as a therapeutic target in melanoma. Pigment Cell Melanoma Res. 22, 400–419. Manning, B.D., and Cantley, L.C. (2003). Rheb fills a GAP between TSC and TOR. Trends Biochem. Sci. 28, 573–576. Masri, J., Bernath, A., Martin, J., Jo, O.D., Vartanian, R., Funk, A., and Gera, J. (2007). mTORC2 activity is elevated in gliomas and promotes growth and cell motility via overexpression of rictor. Cancer Res. 67, 11712–11720.
Schmidt, A., Bickle, M., Beck, T., and Hall, M.N. (1997). The yeast phosphatidylinositol kinase homolog TOR2 activates RHO1 and RHO2 via the exchange factor ROM2. Cell 88, 531–542. Schroder, W.A., Buck, M., Cloonan, N., Hancock, J.F., Suhrbier, A., Sculley, T., and Bushell, G. (2007). Human Sin1 contains Ras-binding and pleckstrin homology domains and suppresses Ras signalling. Cell. Signal. 19, 1279– 1289. Seiser, R.M., and Nicchitta, C.V. (2000). The fate of membrane-bound ribosomes following the termination of protein synthesis. J. Biol. Chem. 275, 33820–33827. Sonenberg, N., and Hinnebusch, A.G. (2009). Regulation of translation initiation in eukaryotes: mechanisms and biological targets. Cell 136, 731–745. Soulard, A., Cohen, A., and Hall, M.N. (2009). TOR signaling in invertebrates. Curr. Opin. Cell Biol. 21, 825–836. Sparks, C.A., and Guertin, D.A. (2010). Targeting mTOR: prospects for mTOR complex 2 inhibitors in cancer therapy. Oncogene 29, 3733–3744. Stahl, J.M., Cheung, M., Sharma, A., Trivedi, N.R., Shanmugam, S., and Robertson, G.P. (2003). Loss of PTEN promotes tumor development in malignant melanoma. Cancer Res. 63, 2881–2890.
Mulet, J.M., Martin, D.E., Loewith, R., and Hall, M.N. (2006). Mutual antagonism of target of rapamycin and calcineurin signaling. J. Biol. Chem. 281, 33000–33007.
Stahl, J.M., Sharma, A., Cheung, M., Zimmerman, M., Cheng, J.Q., Bosenberg, M.W., Kester, M., Sandirasegarane, L., and Robertson, G.P. (2004). Deregulated Akt3 activity promotes development of malignant melanoma. Cancer Res. 64, 7002–7010.
Oh, W.J., Wu, C.C., Kim, S.J., Facchinetti, V., Julien, L.A., Finlan, M., Roux, P.P., Su, B., and Jacinto, E. (2010). mTORC2 can associate with ribosomes to promote cotranslational phosphorylation and stability of nascent Akt polypeptide. EMBO J. 29, 3939–3951.
Sturgill, T.W., Cohen, A., Diefenbacher, M., Trautwein, M., Martin, D.E., and Hall, M.N. (2008). TOR1 and TOR2 have distinct locations in live cells. Eukaryot. Cell 7, 1819–1830.
Pearce, L.R., Huang, X., Boudeau, J., Paw1owski, R., Wullschleger, S., Deak, M., Ibrahim, A.F., Gourlay, R., Magnuson, M.A., and Alessi, D.R. (2007). Identification of Protor as a novel Rictor-binding component of mTOR complex-2. Biochem. J. 405, 513–522. Robertson, G.P. (2005). Functional and therapeutic significance of Akt deregulation in malignant melanoma. Cancer Metastasis Rev. 24, 273–285. Ruggero, D. (2009). The role of Myc-induced protein synthesis in cancer. Cancer Res. 69, 8839–8843. Ruggero, D., and Pandolfi, P.P. (2003). Does the ribosome translate cancer? Nat. Rev. Cancer 3, 179–192.
Thedieck, K., Polak, P., Kim, M.L., Molle, K.D., Cohen, A., Jeno¨, P., Arrieumerlou, C., and Hall, M.N. (2007). PRAS40 and PRR5-like protein are new mTOR interactors that regulate apoptosis. PLoS ONE 2, e1217. Wang, X., McCullough, K.D., Franke, T.F., and Holbrook, N.J. (2000). Epidermal growth factor receptor-dependent Akt activation by oxidative stress enhances cell survival. J. Biol. Chem. 275, 14624–14631. Warner, J.R., and McIntosh, K.B. (2009). How common are extraribosomal functions of ribosomal proteins? Mol. Cell 34, 3–11. Wullschleger, S., Loewith, R., Oppliger, W., and Hall, M.N. (2005). Molecular organization of target of rapamycin complex 2. J. Biol. Chem. 280, 30697– 30704.
Sarbassov, D.D., Ali, S.M., Kim, D.H., Guertin, D.A., Latek, R.R., ErdjumentBromage, H., Tempst, P., and Sabatini, D.M. (2004). Rictor, a novel binding partner of mTOR, defines a rapamycin-insensitive and raptor-independent pathway that regulates the cytoskeleton. Curr. Biol. 14, 1296–1302.
Wullschleger, S., Loewith, R., and Hall, M.N. (2006). TOR signaling in growth and metabolism. Cell 124, 471–484.
Sarbassov, D.D., Guertin, D.A., Ali, S.M., and Sabatini, D.M. (2005). Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098–1101.
Yang, Q., Inoki, K., Ikenoue, T., and Guan, K.L. (2006). Identification of Sin1 as an essential TORC2 component required for complex formation and kinase activity. Genes Dev. 20, 2820–2832.
Sarbassov, D.D., Ali, S.M., Sengupta, S., Sheen, J.H., Hsu, P.P., Bagley, A.F., Markhard, A.L., and Sabatini, D.M. (2006). Prolonged rapamycin treatment inhibits mTORC2 assembly and Akt/PKB. Mol. Cell 22, 159–168.
Zanchin, N.I., Roberts, P., DeSilva, A., Sherman, F., and Goldfarb, D.S. (1997). Saccharomyces cerevisiae Nip7p is required for efficient 60S ribosome subunit biogenesis. Mol. Cell. Biol. 17, 5001–5015.
768 Cell 144, 757–768, March 4, 2011 ª2011 Elsevier Inc.
Yang, Q., and Guan, K.L. (2007). Expanding mTOR signaling. Cell Res. 17, 666–681.
The Mechanism of Linkage-Specific Ubiquitin Chain Elongation by a Single-Subunit E2 Katherine E. Wickliffe,1,5 Sonja Lorenz,1,5,* David E. Wemmer,1,2,4 John Kuriyan,1,2,3,4 and Michael Rape1,* 1Department
of Molecular and Cell Biology of Chemistry 3Howard Hughes Medical Institute University of California, Berkeley, CA 94720, USA 4Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA 5These authors contributed equally to this work *Correspondence:
[email protected] (S.L.),
[email protected] (M.R.) DOI 10.1016/j.cell.2011.01.035 2Department
SUMMARY
Ubiquitin chains of different topologies trigger distinct functional consequences, including protein degradation and reorganization of complexes. The assembly of most ubiquitin chains is promoted by E2s, yet how these enzymes achieve linkage specificity is poorly understood. We have discovered that the K11-specific Ube2S orients the donor ubiquitin through an essential noncovalent interaction that occurs in addition to the thioester bond at the E2 active site. The E2-donor ubiquitin complex transiently recognizes the acceptor ubiquitin, primarily through electrostatic interactions. The recognition of the acceptor ubiquitin surface around Lys11, but not around other lysines, generates a catalytically competent active site, which is composed of residues of both Ube2S and ubiquitin. Our studies suggest that monomeric E2s promote linkage-specific ubiquitin chain formation through substrate-assisted catalysis. INTRODUCTION By regulating protein stability, activity, or localization, ubiquitination exerts control over almost every cellular process. As this includes pathways responsible for the duplication and separation of genetic material, aberrant ubiquitination often results in tumorigenesis. Despite the importance for cellular regulation, the mechanisms determining the specificity and efficiency of ubiquitination reactions are still incompletely understood. Ubiquitination requires at least three enzymatic activities. An E1 enzyme forms a thioester between a cysteine at its active site and the C terminus of ubiquitin (Schulman and Harper, 2009). The activated ubiquitin is transferred to a cysteine of an E2 (Ye and Rape, 2009). In the third step, the charged E2 cooperates with E3s to catalyze formation of an isopeptide bond
between the C terminus of ubiquitin and the 3-amino group of a substrate lysine (Deshaies and Joazeiro, 2009). The 600 human RING-E3s interact with E2s and substrates at the same time, allowing them to promote the transfer of ubiquitin directly from the E2 to the substrate. The modification of a substrate with a single ubiquitin usually leads to changes in protein interactions (Dikic et al., 2009). In many cases, additional ubiquitin molecules are attached to a substrate-linked ubiquitin, giving rise to polymeric ubiquitin chains. Such chains can be connected through the N terminus of ubiquitin or through one of its seven Lys residues, and all linkages have been detected in cells (Ye and Rape, 2009; Xu et al., 2009). Ubiquitin chains of different topologies can have distinct structures and functions (Dikic et al., 2009; Matsumoto et al., 2010). K48-linked chains, for example, drive protein degradation, whereas K63-linked chains regulate the assembly of protein complexes (Ye and Rape, 2009). Thus, the efficiency and specificity of chain formation have profound consequences for the modified protein. We recently identified K11-linked ubiquitin chains as critical cell-cycle regulators in human cells (Jin et al., 2008). Most K11-linked chains are synthesized during mitosis by the E3 anaphase-promoting complex (APC/C) and its E2s Ube2C/ UbcH10 and Ube2S (Williamson et al., 2009; Matsumoto et al., 2010; Wu et al., 2010). Together, these enzymes modify mitotic regulators, such as cyclin B, securin, or HURP, to trigger their degradation (Jin et al., 2008; Song and Rape, 2010). As a result, inhibiting the formation of K11-linked chains blocks mitotic progression in Xenopus, Drosophila, and humans (Jin et al., 2008; Williamson et al., 2009; Garnett et al., 2009), whereas their untimely assembly causes inaccurate cell division and tumorigenesis (Wagner et al., 2004; Jung et al., 2006). How the APC/C and its E2s assemble K11-linked chains, however, is poorly understood. Linkage between two ubiquitin moieties involves the covalent connection of one ubiquitin, the donor, to the active site cysteine of the E2, followed by nucleophilic attack by a lysine of an acceptor ubiquitin (Figure 1A). Much of our knowledge about the basis of linkage specificity is limited to the E2 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 769
B
C WT-ubi/UBC 0.4
acceptor
K11
Ube2S
0.3 0.2
ubi~ubi ubi
0.1 0
1
Ube2S 10
20
Coomassie
30
40
50
60
Ube2S APC/C
70 A
E
INPUT WT L8 A I44 V7 A 0 I44 A A/V
11 4
+E m +s i1 ec +U urin be 2C C
INPUT
S
ubiquitin residue number
D
ubiquitin mutant:
G
WT
donor/ acceptor
WT
Ubn-L-cycA
Ubn-L-cycA
Ub2-L-cycA Ub-L-cycA
Ub2-L-cycA Ub-L-cycA
76
F
ubiquitin mutant:
Ube2S APC/C
70
K11A
only
His ΔGG acceptor WT K1 1 I44 R A
C95
Δδ(1H15N)/ppm
K11 donor
WT-ubi/Ube2S ubi-I44A/UBC
WT L8A I44A V70A I44A/V70A
A
I44A/ V70A
0 60 60 0 60 60 0 60 60 time (min)
60 60 60 60 time (min)
/
/
autoradiography
Western: anti-ubiquitin
Western: anti-ubiquitin
autoradiography
Figure 1. Ube2S Recognizes the Hydrophobic Patch of Donor Ubiquitin (A) Overview of K11-specific linkage formation. Lys11 of acceptor ubiquitin attacks the thioester bond between Cys95 of Ube2S and the C terminus of the donor ubiquitin. (B) Ube2S interacts with ubiquitin noncovalently. Weighted combined chemical shift perturbations, Dd(1H15N), are plotted over residue number. The asterisk indicates the disappearance of the resonance for His68 of ubiquitin in the presence of Ube2S due to intermediate exchange. (C) Mutation of the hydrophobic patch in ubiquitin interferes with formation of K11-linked ubiquitin dimers (ubiubi) by Ube2S, as monitored by Coomassie staining. (D) Ube2S and APC/C extend ubiquitin chains on a fusion between ubiquitin and cyclin A (Ub-L-cycA), as analyzed by autoradiography. (E) The hydrophobic patch of ubiquitin is required for chain elongation by APC/CCdh1 and Ube2S, as analyzed by autoradiography. (F) The hydrophobic patch is not required on acceptor ubiquitin. Ube2S was mixed with acceptor His6ubiquitinDGG mutants (ubiDGG; purple) and WT-ubiquitin (blue) and analyzed by a-ubiquitin-western. (G) The hydrophobic patch is required on the donor ubiquitin. Ube2S was mixed with WT-ubiDGG and ubiquitin mutants and analyzed by a-ubiquitin-western. See also Figure S1.
Ube2N-Uev1A (Ubc13-Mms2; VanDemark et al., 2001). In this system, the catalytically inactive Uev1A orients the acceptor ubiquitin, such that Lys63 of the acceptor is at the active site of Ube2N charged with the donor (Eddins et al., 2006). In contrast, K11- and K48-linkage-specific E2s promote chain elongation in reconstituted systems lacking UEVs (Li et al., 2009; Pierce et al., 2009; Williamson et al., 2009). Although kinetic analyses suggest that these E2s also engage acceptor ubiquitin residues (Petroski and Deshaies, 2005; Rodrigo-Brenni et al., 2010), the molecular details of acceptor recognition by monomeric E2s and its importance for linkage-specific chain formation have not been established. Here, we have combined functional studies with nuclear magnetic resonance (NMR) and computational docking to dissect the mechanism of linkage-specific chain formation by single-subunit E2s. We show that the K11-specific Ube2S orients the donor ubiquitin by a noncovalent interaction that is in addition 770 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
to the flexible covalent linkage between these molecules at the E2 active site. We find that a similar tethering mechanism is used by other E2s independently of linkage specificity. The Ube2S-donor ubiquitin complex transiently engages the acceptor ubiquitin through electrostatic interactions. As indicated by our analysis, only binding of the acceptor surface around Lys11, but not around other lysines, leads to formation of a catalytically competent active site composed of residues of both Ube2S and ubiquitin. Hence, linkage-specific ubiquitin chain formation by Ube2S is the result of substrate-assisted catalysis. RESULTS A Noncovalent Interaction with Ubiquitin Is Required for Ube2S Activity In the absence of the APC/C, Ube2S generates K11-linked ubiquitin dimers (ubi2) and chains attached to Lys residues of its UBC
A
UBC/ubiquitin
UBC/ubi-K6E
UBCI121A/ubiquitin
UBC/ubi-I44A
B WT C118A I121A
Δδ(1H15N)/ppm
0.4
0.3
Ube2S mutant:
ubi~ubi
0.2
ubi
0.1
Ube2S Coomassie
0
Ube2S residue number
Ube2S mutant: Ube2S APC/C
E
F HeLa
WT C118A I121A
INPUT WT C1 I1218A 1A
D
cyclin A
HeLa+ Ube2SI121A
siUBE2S 0 2 0 2 0 2 0 2 0 2 0 2 time (h)
Ube2S mutant:
Ubn-L-cycA
HeLa+ Ube2S
WT C118A I121A C118A+C95A I121A+C95A
C
K11-linked chains
Ubi
Ube2S mutant:
ubi~ubi ubi Ube2S
Ube2S
Ub2-L-cycA Ub-L-cycA autoradiography
cyclin A
Coomassie
β-actin
autoradiography Western
Figure 2. Noncovalent Donor Ubiquitin Binding Is Required for Ube2S Activity (A) Identification of Ube2S residues involved in noncovalent binding of ubiquitin. Weighted combined chemical shift perturbations, Dd(1H15N), are plotted over the residue number. (B) Donor binding is required for formation of ubi2 dimers (ubiubi) by Ube2S mutants, as analyzed by Coomassie staining. (C) Donor binding by Ube2S is required for chain elongation on Ub-L-cycA with APC/C, as analyzed by autoradiography. (D) Donor binding by Ube2S is required for chain formation in a full APC/C assay. Ubiquitination of cyclin A by APC/C, Ube2C, and Ube2S mutants was analyzed by autoradiography. (E) Donor binding is required for Ube2S activity in vivo. HeLa cell lines expressing Ube2S or Ube2SI121A were treated with siRNAs against the 30 UTR of Ube2S, which specifically depletes endogenous Ube2S. Cells were synchronized in prometaphase (t = 0 hr) or late mitosis (t = 2 hr), and K11-linked ubiquitin chains were detected by aK11-western. (F) Donor binding occurs in cis. Ube2SC118A or Ube2SI121A (lack the noncovalent ubiquitin-binding site) and Ube2SC95A (no active site) were mixed, and ubi2 formation was monitored by Coomassie staining. See also Figure S2.
domain and its C-terminal tail (Figure S1A available online). The UBC domain of Ube2S (UBCUbe2S) promotes ubi2 formation with similar kinetics and specificity as Ube2S (Figure S1A). Like Ube2S, UBCUbe2S is monomeric in ubiquitination buffers, as suggested by gel filtration, small-angle x-ray scattering (SAXS), and other biophysical techniques (Figures S1B and S1C; data not shown). Thus, UBCUbe2S contains all elements required for the synthesis of K11 linkages, making it an appropriate system for analyzing the mechanism of linkage-specific chain formation. The prevailing model of linkage-specific chain formation posits that an elongating E2, charged with the donor ubiquitin, binds the acceptor in such a way that a preferred acceptor lysine is at the E2 active site (Eddins et al., 2006). To test for such a noncovalent interaction between Ube2S and ubiquitin, we performed titrations of 15N-enriched ubiquitin with Ube2S and UBCUbe2S,
respectively, and monitored 1H-15N HSQC spectra. The presence of either E2 caused significant resonance-specific chemical shift perturbations in the ubiquitin spectrum, indicating a specific interaction (Figure 1B). Chemical shift mapping on the surface of ubiquitin revealed that the hydrophobic patch surrounding Ile44 is involved in the noncovalent interaction with Ube2S (Figure S2B). Mutation of the isoleucine to alanine (ubiI44A) disrupted the interaction with Ube2S (Figure 1B and Figure 2A). Furthermore, mutating residues in the hydrophobic patch (L8A, I44A, V70A) interfered strongly with the formation of K11-linked ubi2 by Ube2S (Figure 1C). Ube2S extends K11-linked chains on APC/C substrates after initiation by Ube2C (Williamson et al., 2009; Wu et al., 2010). To test whether mutations in ubiquitin interfere with chain elongation Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 771
by Ube2S and APC/C, we bypassed the need for Ube2C by generating a fusion between ubiquitin and the APC/C substrate cyclin A (Ub-L-cycA). Ube2S and APC/C rapidly elongated ubiquitin chains on Ub-L-cycA, which did not require Ube2C and was not inhibited by an excess of inactive Ube2CC114S (Figure 1D). Mutation of the hydrophobic patch of ubiquitin interfered strongly with this activity of Ube2S (Figure 1E; Figure S1D), without affecting charging by E1 (Figure S1F). The same mutations in ubiquitin blocked ubiquitination of APC/C substrates in an assay containing both Ube2C and Ube2S (Figure S1E). Thus, a noncovalent interaction with ubiquitin is required for the ability of Ube2S to assemble K11-linked ubiquitin chains. The Hydrophobic Patch Is Required on the Donor Ubiquitin To determine whether Ube2S interacts with donor or acceptor ubiquitin, we made a ubiquitin mutant lacking its two C-terminal Gly residues, ubiDGG. ubiDGG is not activated by E1 and can only act as acceptor. Ube2S produced dimers between ubiDGG and ubiquitin (ubiDGG-ubi; Figure 1F), and mutation of Lys11 on the acceptor ubiDGG, but not the donor ubiquitin, blocked this reaction (Figures 1F and 1G). The mutation of Leu8, Ile44, or Val70 on the acceptor ubiDGG had no effect on the formation of ubiDGG-ubi dimers (Figure 1F; Figure S1G). Instead, when the hydrophobic patch was mutated on the donor ubiquitin, dimer formation was prevented (Figure 1G). We conclude that Ube2S recognizes the donor ubiquitin. Several aspects of our analysis indicate that this is the only thermodynamically stable interaction between ubiquitin and Ube2S in solution. The NMR-derived binding isotherms are well described by a single-site binding model (Figure S2D); significant chemical shift perturbations map to one contiguous binding region (Figure S2B); and disruption of the donor interface did not result in the population of an alternate binding site (Figure 1B). Variation of the experimental conditions, such as ionic strength and pH, also did not provide evidence for a second binding site (data not shown). Thus, Ube2S forms a noncovalent interface with the hydrophobic patch of the donor ubiquitin, which is required for its activity to promote the formation of K11-linked ubiquitin chains. The Donor Ubiquitin Interacts with Helix aB of Ube2S We identified the donor-binding site on Ube2S by titrating 15Nenriched UBCUbe2S with ubiquitin and measuring 1H-15N HSQC spectra. Significant ubiquitin-induced chemical shift perturbations mapped to a surface region around the C-terminal part of helix aB of Ube2S (Figure 2A; Figure S2B). The same region was found to interact with covalently bound donor ubiquitin (Figure S2A). Mutation of two Ube2S residues in this region (C118A, I121A) impaired ubiquitin binding (Figure 2A; Figure S2C) without affecting the structural integrity of Ube2S (data not shown). The dissociation constant (KD) for this interaction (1.11 ± 0.08 mM or 1.7 ± 0.08 mM for ubiquitin binding to UBCUbe2S or Ube2S; Figure S2D) was comparable to the estimated concentration of donor ubiquitin linked to the Ube2S active site (3 mM; Petroski and Deshaies, 2005). Our results, therefore, suggest that covalently linked donor ubiquitin occupies the noncovalent binding site on Ube2S around helix aB. 772 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
Based on our previous results, we expected the donor interface of Ube2S to be required for activity. Indeed, Ube2SC118A and Ube2SI121A were strongly impaired in ubi2 formation (Figure 2B); K11-linked chain assembly on Ub-L-cycA with APC/C (Figure 2C); or modification of cyclin A in an APC/C assay containing Ube2C and Ube2S (Figure 2D). Disrupting this Ube2S surface did not affect charging by E1 (Figure S2E) or binding to the APC/C (Figure S2F). To test for the importance of donor binding at physiological ubiquitin levels, we generated HeLa cell lines that stably express Ube2S or Ube2SI121A. Endogenous Ube2S was specifically depleted from cells by siRNAs against the 30 untranslated region (UTR) of the Ube2S mRNA, and formation of K11-linked chains was monitored upon exit from mitosis by a K11-linkage-specific antibody. As expected, long K11-linked chains were absent from cells lacking Ube2S, which was rescued by expression of siRNAresistant wild-type (WT)-Ube2S (Figure 2E). By contrast, the donor-binding-deficient Ube2SI121A failed to promote K11linked chain formation. Thus, recognition of the donor ubiquitin by Ube2S is essential for K11-linked ubiquitin chain formation in vitro and in vivo. NMR-Based Docking of the Donor Ubiquitin on Ube2S The binding site for the donor ubiquitin on Ube2S, as defined by chemical shift mapping, makes it plausible that this interaction occurs in cis, i.e., involves the same E2 that the donor is covalently attached to. To test this idea, we determined whether a catalytically inactive Ube2S mutant with an intact donorbinding site (Ube2SC95A) could complement the loss-of-function phenotype of a Ube2S mutant with a defective noncovalent interface (Ube2SC118A, Ube2SI121A). As this was not the case (Figure 2F), the observed noncovalent interaction most likely occurs in cis. To obtain a structural model of the interaction between UBCUbe2S and donor ubiquitin, we used the docking program HADDOCK (de Vries et al., 2007). The NMR chemical shift data were used to specify residues at the interface, and we defined only one explicit distant restraint that required the C-terminal carbon atom of Gly76 of ubiquitin to be close to the Sg atom of Cys95 in Ube2S. HADDOCK produced an ensemble of 200 structures after automated refinement, which were clustered using a backbone root-mean-square deviation (rmsd) cut-off of 7.5 A˚. The resulting three clusters contain 71%, 25.5%, and 2.5% of all docked models, respectively (Table S1). As shown later, structures in cluster 1, but not those of clusters 2 and 3, could be validated by biochemical data. As a representative structure of the Ube2S-donor ubiquitin complex, we selected a model from cluster 1 that among the top 3 according to HADDOCK scoring had the largest buried surface area, the most negative interaction energy, and the smallest number of distant restraint violations (Figure 3A; Table S1). A similar model with a low backbone rmsd of 1.1 A˚ was obtained by a different docking program, Cluspro (Comeau et al., 2007), without restraints (Figure S3A). Our model resembles the structure of the charged E2 Ubc9, when bound to its E3 (Reverter and Lima, 2005), and an NMR-based, docked model of the E2 Ubc1 and ubiquitin (Hamilton et al., 2001).
N
C
αC
αD
αD C
E51
C Cys95
donor ubiquitin
αCat
C-terminal tail
E142 Y141
L138 L137
αC Q49
P125 I121 P123 H122 N124 E126 C118
αB
S127 C95 V98 N97
αCat
R101 K100 D102
αB
D
E
K48 A46 G47
Ube2S mutant: Ube2S APC/C
Ube2S mutant:
ubi~ubi
H68 R42 K6 R72 V70 L69
Ubn-L-cycA
ubi
L71 L8 G76 G75 L73
Ube2S Coomassie
R74
Ub2-L-cycA Ub-L-cycA
C-terminal tail
αA
autoradiography
αA
UBCUbe2S
WT E5 1 Y1 K 4 R1 1A 0 D 1 1A 0 S1 2 A 27 A
B
WT E51K Y141A WT R101A D102A S127A
A
I
N
αC
0 2 0 2 0 2 0 2
E51
siUBE2S 0 2 0 2 time (h)
Ube2S APC/C
αB
WT-ubiΔGG
ubi~ubi ubi
K6
3.1Å
K11-linked chains
Ubn-L-cycA
WT WT WT E51K E51K
ubiquitin mutant:
HeLa+ HeLa+ HeLa+ Ube2SE51K Ube2SD102A Ube2SS127A
Ube2S mutant:
WT K6E K6E WT
ubiquitin mutant:
H INPUT WT K6 E K4 8 T6 E 6 H6 E 8 L7 A 1A R7 2A
G
WT K6E K48E T66E H68A L71A R72A
F
Ube2S Coomassie
Ube2S Ub2-L-cycA Ub-L-cycA autoradiography
/
β-actin Western
Ube2S Silver
Figure 3. Structural Model of the Ube2S-Donor Ubiquitin Complex (A) NMR-based HADDOCK model of the UBCUbe2S-donor ubiquitin complex (cluster 1, no. 3; see Table S1). The C-terminal tail of ubiquitin (cyan) was allowed full flexibility during docking. (B and C) Surface representation of the binding interface on UBCUbe2S (B) and donor ubiquitin (C). Residues that make intermolecular contacts within a radius of 4 A˚ are shown in pink. (D) Ube2S residues at the donor-binding interface are required for formation of ubi2 dimers (ubiubi), as monitored by Coomassie staining. (E) Donor-binding-deficient Ube2S mutants do not promote chain elongation on Ub-L-cycA with APC/C, as analyzed by autoradiography. (F) Ubiquitin residues at the Ube2S interface are required for linkage formation, as seen by Coomassie staining. (G) Ubiquitin residues at the Ube2S interface are required for chain elongation on Ub-L-cycA with APC/C, as analyzed by autoradiography. (H) Donor binding is required for Ube2S activity in cells. HeLa cell lines expressing donor-binding-deficient Ube2S (E51K; D102A; S127A) were depleted of endogenous Ube2S, synchronized in prometaphase (t = 0) or allowed to exit mitosis (t = 2 hr), and tested for K11-linked ubiquitin chains by aK11-western. (I) Charge-swap analysis of the ionic interaction between Lys6 of donor ubiquitin and Glu51 on Ube2S. ubiDGG was mixed with ubiquitin or ubiK6E in the presence of Ube2S or Ube2SE51K. Reactions were monitored by Silver staining. See also Figure S3 and Table S1.
Within our model, the donor ubiquitin docks onto a hydrophobic area on Ube2S, comprising the C-terminal half of helix aB, the C-terminal part of helix aC, and the N-terminal part of helix aD (Figures 3A and 3B). The corresponding interaction surface on donor ubiquitin contains the hydrophobic patch (Figure 3C), which is extended to form a contact area that buries a total of 830 A˚2 on ubiquitin. The model includes ionic contacts between Lys6, Arg42, and Lys48 of ubiquitin, and Glu51, Glu126, and Glu142 of Ube2S (Figure S3B). An ionic contact between Arg74 of ubiquitin and Asp102 of Ube2S serves as a linchpin to guide the C terminus of ubiquitin toward the active site of Ube2S (Figure S3C). The ubiquitin tail is also anchored by hydrogen bonds between the peptide backbone and Ube2S residues close to the active site. The distance between the Sg atom of Cys95 of Ube2S and the C-terminal carbon atom of ubiquitin, 3.9 A˚, is too long for a covalent bond, but small adjustments around the active site of Ube2S could readily close this gap.
Validation of the Ube2S-Donor Ubiquitin Model Based on the selected model for this complex, we designed additional mutations to test the structural details of the predicted interface. We found that altering residues at the binding interface (Ube2S: E51K, R101A, D102A, S127A, Y141A; ubiquitin: K6E, K48E, T66E, H68A, L71A, R72A) interfered with Ube2S activity in vitro (Figures 3D–3G) and, as seen for ubiK6E, disrupted the Ube2S-donor ubiquitin interaction (Figure 2A). Residues that do not make intermolecular contacts (Ube2S: D29, G30, L114, E142; ubiquitin: A46) were not required for activity (data not shown). Introducing mutations into ubiDGG showed that most ubiquitin residues were required in the donor but not the acceptor (Figures S3D and S3E). The role of Lys6 in the acceptor ubiquitin is described below. With the exception of ubiR72A, no Ube2S or ubiquitin mutant was impaired in charging by E1 (Figures S3F and S3G). To confirm this analysis in vivo, we generated cell lines that express Ube2S mutants with Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 773
Ube2S
B
A
1A
E5 1K
I12
ub i-I4 4A C1 18 A
WT
K11R
0
ubiquitin or Ube2S mutant:
20
40
ΔGG
60
Ube2S-I121A
Ube2S K11R I44A
80
100
120
0
20
40
ΔGG
60
80
K11R
100
ΔGG
120 0 30 60 90 120 time (min)
6.25 12.5 25
ubi~ubi
50
ubi
100
longer exposure
150
Ube2S
Coomassie C Coo oom omassie ma m mas a assie e
[ubiΔGG] (μM)
ubi-K11R WT-ubi
200
Coomassie
250
D ubiquitin mutant:
0 5 10 20 30 5 45 5 45 5 45 time (min)
Ubpoly-L-cycA
- UBA + UBA
WT-ubiquitin
WT
-ub
i
C
ub i-I4 4A ub i-V 70 A ub i-I4 4A /V7 0A
Silver
+UBA
1
Ub2-L-cycA Ub-L-cycA
3
4
5 6 7 8 9 11 13 - UBA + UBA
ubi-V70A
Uboligo-L-cycA
2
+UBA
autoradiography
1
2 3 4 number of ubiquitin molecules in chain
Figure 4. Noncovalent Donor Binding Increases the Processivity of Ube2S (A) Donor binding is not required for the K11-linkage specificity of Ube2S. Ube2S or donor-binding mutants were incubated with ubiquitin or ubiK11R, or as indicated with ubiI44A and ubiI44A/K11R. Reactions were incubated longer and at higher ubiquitin concentrations to observe formation of ubi2 and analyzed by Coomassie staining. (B) Donor binding promotes catalysis at low acceptor concentrations. Dimer formation between increasing levels of ubiDGG and ubiK11R or ubiK11R/I44A, respectively, by Ube2S was monitored by Silver staining. (C) Time-course analysis of chain elongation on Ub-L-cycA by APC/CCdh1 and Ube2S in the presence of ubiquitin mutants, as analyzed by autoradiography. (D) Donor binding is required for processive chain formation by Ube2S. Chain elongation on Ub-L-cycA by APC/C and Ube2S was monitored in the presence of the UBA domains of Rad23A. Reactions were performed with ubiquitin or ubiV70A and analyzed by autoradiography (top) and line scanning (bottom).
defective donor-binding interfaces (E51K; D102A; S127A). Importantly, all of these failed to promote the formation of K11linked ubiquitin chains in HeLa cells that lacked endogenous Ube2S (Figure 3H). To further test our model, we used charge-swap analysis to analyze the role of the predicted ion pair between Glu51 of Ube2S and Lys6 of ubiquitin. While the K6E mutation in donor ubiquitin interfered with formation of ubiDGG-ubi dimers, this was rescued by a complementary mutation in Ube2S, Ube2SE51K (Figure 3I). Ube2SE51K did not establish ubi2 formation for other ubiquitin mutants, such as ubiI44A, attesting to the specificity of this rescue (Figure S3H). Together, the mutational studies, charge-swap analysis, and in vivo experiments validate the 774 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
selected NMR-based model for the Ube2S-donor ubiquitin interaction and show its importance for chain formation by this E2. Noncovalent Donor Ubiquitin Binding Is Required for Processive Chain Formation We next determined the role of donor binding for catalysis by Ube2S. It was unlikely that recognizing the donor ubiquitin was important for specificity, and indeed, any ubi2 formed in the presence of ubiI44A was lost upon mutation of K11 (Figure 4A). The same was observed when Ube2S mutants with a defective donor-binding interface (Ube2SI121A; Ube2SC118A; Ube2SE51K) were analyzed for ubi2 formation (Figure 4A). Thus, donor binding does not determine the K11 specificity of Ube2S.
Alternatively, the noncovalent interaction between the donor and Ube2S might prevent a flexible donor molecule from interfering with acceptor recognition. If this is the case, higher concentrations of the acceptor ubiDGG should rescue the defect in ubi2 formation when the Ube2S-donor ubiquitin interface is disturbed. Consistent with this hypothesis, high levels of ubiDGG allowed linkage formation with ubiI44A or Ube2SI121A (Figure 4B). The acceptor concentration required under these conditions was above the endogenous ubiquitin levels in HeLa cells (90 mM) (Ryu et al., 2006), consistent with the lack of Ube2SI121A activity in vivo. These findings suggest that noncovalent binding of the donor ubiquitin facilitates acceptor recognition by Ube2S. Based on these observations, we expected that donor binding would increase the processivity of chain formation by Ube2S. Indeed, a time-resolved analysis of chain elongation on UbL-cycA suggested that Ube2S assembles chains with high processivity (Figure 4C), whereas chain formation occurred in a step-like, distributive fashion if donor-binding was impaired (Figure 4C). To directly measure the processivity of chain elongation, we supplied the reactions with a UBA domain. As previously described (Rape et al., 2006), the UBA domain captures any substrate dissociating from the APC/C, thereby preventing it from rebinding the E3 and revealing the number of ubiquitin molecules transferred in a single substrate-binding event. Ube2S could transfer up to 13 ubiquitin molecules to Ub-LcycA per binding event (Figure 4D, top), whereas less than four molecules of the hydrophobic patch mutant ubiV70A were transferred (Figure 4D, bottom). As this UBA domain only recognizes K11-chains with at least 5 ubiquitin moieties (data not shown), the number of ubiV70A molecules transferred in a single binding event is likely even smaller. Thus, the noncovalent interaction between Ube2S and donor ubiquitin increases the processivity of chain formation, at least in part by facilitating acceptor ubiquitin recognition. Noncovalent Donor Ubiquitin Binding Is a Feature of Chain Elongation in Other E2s To test whether other E2s bind the donor ubiquitin noncovalently, we turned to Ube2R1 and Ube2G2, which extend K48linked chains (Li et al., 2009; Pierce et al., 2009). Disruption of the hydrophobic patch on ubiquitin strongly impaired the formation of K48 linkages by these E2s (Figure S4A), while having no effect on their charging by E1 (Figure S4B). Analogous to our results for Ube2S, we found that the activity of Ube2R1 and Ube2G2 was dependent on recognition of the hydrophobic patch on the donor but not the acceptor ubiquitin (Figure 5A). As revealed by ubiI44A/K48R- and ubiV70A/K48R-double mutants, noncovalent donor binding did not determine linkage specificity (Figure 5B) but was required for catalysis at low substrate concentrations (Figure 5C). To test whether a common E2 surface recognizes the donor ubiquitin, we studied Ube2R1 mutations of sites that are structurally homologous to the Ube2S-donor interface (Figure 5D). These mutations (Ube2R1T122E; Ube2R1L125A; Ube2R1I128E) strongly inhibited the formation of K48 linkages (Figure 5D; Figure S4C), without affecting charging by E1 (Figure S4D). The same mutations also impaired the SCF- and Ube2R1-dependent formation of ubiquitin chains on IkBa (Figure 5E). Thus, similar
surfaces on the conserved E2 fold and ubiquitin are used for chain elongation by E2s of different linkage specificity. The tethering of the donor ubiquitin by an E2, therefore, provides a conserved mechanism to facilitate acceptor recognition. Ube2S Recognizes the TEK-Box in Acceptor Ubiquitin We next identified the binding site for the acceptor ubiquitin on Ube2S. Consistent with previous analyses (Petroski and Deshaies, 2005; Rodrigo-Brenni et al., 2010), acceptor binding was too transient to be detected by NMR, independently of whether the donor ubiquitin had been linked to the Ube2S active site or not (data not shown). We therefore used HADDOCK to dock a second ubiquitin molecule, the acceptor, onto the validated Ube2S-donor ubiquitin complex. The only restraint used for this docking defined the Nz-atom of the acceptor Lys11 to be close to the Sg atom of Cys95 at the Ube2S active site. HADDOCK generated two clusters of models, which were similar in terms of energy and buried surface area (Figure S5; Table S2). Intriguingly, models in cluster 1 orient the TEK-box of ubiquitin toward the active site of Ube2S. The TEK-box is a surface region of ubiquitin that was previously identified to mediate the preference of Ube2C/UbcH10 for assembling K11-linked chains (Jin et al., 2008). We found that mutation of the TEK-box strongly interfered with the ability of Ube2S to synthesize K11-linked ubi2 (Figure 6A), to elongate chains on Ub-L-cycA (Figure 6B), and to modify cyclin A in a full APC/C assay (Figure S6A). Residues outside of the TEK-box (Thr9, Glu16, Lys33) were not required for activity (data not shown). Introducing mutations into ubiDGG revealed that the TEK-box was essential on the acceptor (Figure 6C) but, with exception of Lys6, not the donor ubiquitin (Figure 6D). The TEK-box was dispensable for Ube2S charging by E1 (Figure S6B) or binding of donor ubiquitin to Ube2S (Figure S6C). Thus, the TEK-box of the acceptor ubiquitin is required for K11-linkage formation by Ube2S. On the basis of these results, we performed another docking run that defined four TEK-box residues to be at the interface (Table S2). All HADDOCK solutions grouped into a single cluster reproducing the binding topology of cluster 1 of the previous run (Figure 6E). With backbone rmsd values of 1 A˚, the best models of this ensemble were remarkably similar, and we focused on the top-scoring model (cluster 1, no. 1; Table S2, bottom). In this model of the ternary complex, the acceptor ubiquitin interacts mostly with Ube2S (Figure 6E), with a total buried surface area between the acceptor ubiquitin and the Ube2Sdonor complex of 980 A˚2 (Figures 6F and 6G). The acceptor ubiquitin uses its b strand region to bind a surface of Ube2S comprising the active site helix aCat, the loop region between helices aB and aC, and helix aC (Figure 6F). Although the interaction leads to the burial of few hydrophobic residues, these are not closely packed, and the interface is predominantly electrostatic. In particular, a network of ionic contacts involving Lys6 and Lys63 of ubiquitin, Glu131 and Glu139 of Ube2S, and a series of hydrogen bonds including Glu64 of ubiquitin and Arg135 of Ube2S are key features of the interface (Figure S6D). The electrostatic Ube2S-acceptor interface is consistent with the low affinity between these molecules (Sheinerman and Honig, 2002). Mutating residues in the predicted binding site on Ube2S (N97A, E131K, R135E), but not outside of this interface (K76, Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 775
WT L8A L8A/K48R V70A V70A/K48R
B
C ubiquitin mutant:
WT WT WT I44A/V70A
WT I44A WT I44A WT I44A WT
A
ubi~ubi
ubiquitin
ubi-I44A/V70A
0 5 10 20 30 40 50 60 80 0 5 10 20 30 40 50 60 80 time (min)
25
Ube2R1
0 0 60 60 60 60 60 60 time (min)
50
Coomassie
150
WT L8A L8A/K48R V70A V70A/K48R
Ube2S /
Ube2R1 /
250
ubiquitin mutant:
[ubiquitin] (μM)
ubi
Coomassie
Ube2G2 ubi~ubi
Ube2G2 E 20min INPUT
D
Cys95
I12
8E
A 25 L1
22 T1
αB
WT
Ile121 Cys118
60min ubi-I44A/V70A Ube2R1-I128E ubiquitin Ube2R1
Ube2S
E
Ube2R1
30min INPUT
ubi Coomassie
INPUT
/ Silver
IκBαUBI
0 30 60 0 30 60 0 30 60 0 30 60 time (min)
ubi~ubi
Ile128 Leu125
Ube2R1 mutant:
Cys93
ubi Ube2R1
Thr122
autoradiography
Coomassie
IκBαPO4 IκBα
Figure 5. Noncovalent Donor Binding Is Utilized by E2s Independently of Linkage Specificity (A) Ube2R1 and Ube2G2 require the hydrophobic patch in the donor but not acceptor ubiquitin for K48-linkage formation. Ube2S, Ube2R1, or Ube2G2 and its E3 gp78 were incubated with ubiDGG or ubiDGG/I44A/V70A (purple) and ubiquitin or ubiI44A (blue). Reactions were analyzed by Silver staining. (B) The hydrophobic patch of donor ubiquitin is not required for K48 specificity of Ube2R1 or Ube2G2. ubi2 formation by Ube2R1 or Ube2G2/gp78 with ubiquitin mutants was analyzed by Coomassie staining. (C) Donor binding is required for rapid catalysis by Ube2R1. Time courses of ubi2 formation by Ube2R1 in the presence of increasing concentrations of ubiquitin or ubiI44A were analyzed by Coomassie staining. (D) A similar surface as the donor-binding interface of Ube2S (yellow) is required in Ube2R1 (green; PDB ID: 2OB4). Ube2R1 mutants were analyzed for K48specific ubi2 formation by Coomassie staining. (E) Donor binding is required for processive chain formation by SCFbTrCP and Ube2R1. Phosphorylated IkBa was incubated with SCFbTrCP, Ube2R1 or Ube2R1I128E, and ubiquitin or ubiI44A/V70A and analyzed by autoradiography. See also Figure S4.
N91, E93, K100, E126, E132), impaired production of ubi2 (Figure 6H; data not shown), chain elongation on Ub-L-cycA (Figure 6I), and substrate modification in a full APC/C assay (Figure S6E). As seen in cells expressing Ube2SE131K or Ube2SR135E, these mutations also inhibited formation of K11-linked chains in vivo (Figure 6J). Asn97, Glu131, and Arg135 were not required for Ube2S charging by E1 (Figure S6F) or Ube2S binding to the APC/C (Figure S2F). We next probed the predicted ionic contacts between Ube2S and the acceptor ubiquitin by charge-swap analyses. Our model found the acceptor Lys6 to face Glu131 of Ube2S (Figure 7A). Consistent with this, the loss of ubiDGG-ubi formation caused 776 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
by a K6E mutation in the acceptor ubiDGG could be rescued by Ube2SE131K or Ube2SE131A (Figure 7A; Figure S7A), but not by other Ube2S mutants in this interface (N97A, R135E) (Figure S7B). Thus, the acceptor Lys6 is recognized by Glu131 of Ube2S, whereas the donor Lys6 contacts Glu51 of Ube2S (Figure 3); indeed, Ube2SE51K/E131K completely rescued ubi2 formation by ubiK6E (Figure 7B; Figure S7C). Our model also showed Arg135 of Ube2S in proximity of Glu64 of ubiquitin. Accordingly, the diminished activity of Ube2SR135E was significantly rescued by ubiE64K (Figure S7D), but not other ubiquitin mutants in the proximity of this surface (Figure S7E). Finally, less ubi2 was formed in the presence of ubiK63E, which could be rescued by
WT WT K6A K11A T12A T14E E34A K6E
K63
ubi
G10
Coomassie
αA
αC
UBCUbe2S
N
N
αD
Ub2-L-cycA Ub-L-cycA
Gly76
donor ubiquitin
Lys11
HeLa+ HeLa+ Ube2SE131K Ube2SR135E 0 2 0 2
siUBE2S 0 2 0 2 time (h)
Cys95
αB
acceptor ubiquitin
αA Ube2S
UBC
K11-linked chains
Ubn-L-cycA
C
Ub2-L-cycA Ub-L-cycA
/
Ube2S
Ube2S mutant:
Ube2S APC/C
WT K6A K6E K11A T12A T14E E34A
WT
J INPUT WT N9 7 E1 A 3 R 1 1K 35 E
I C
autoradiography
C
ubi Ube2S
αB
E
Ubn-L-cycA
Ube2S mutant:
ubi~ubi
C-terminal tail αCat
C-terminal tail
ubiquitin mutant:
αD
G76 C95 N97
K11
Coomassie
INPUT WT K6 A K1 1 T1 A 2A T1 4 E3 E 4A
F4 T14 T12
Ube2S
Ube2S APC/C
E139 L138R135 E131 P125 E126 A128 S127 L129
Q2
K6
/
Coomassie
αC
E64
T66
Ube2S
B
H
donor ubiquitin
WT WT WT WT WT WT WT
WT K6A K6E K11A T12A T14E E34A
ubi~ubi
G
F
D ubiquitin mutant:
INPUT WT N97A E131K R135E
A
N
Ube2S β-actin
autoradiography Western
Silver
Figure 6. Acceptor Ubiquitin Recognition by the Ube2S-Donor Ubiquitin Complex (A) The TEK-box in ubiquitin is required for K11-linkage formation by Ube2S. Ube2S was incubated with ubiquitin mutants, and ubi2 formation (ubiubi) was monitored by Coomassie staining. (B) TEK-box mutants in ubiquitin inhibit chain formation on Ub-L-cycA by Ube2S and APC/C, as analyzed by autoradiography. (C) The TEK-box is required on acceptor ubiquitin. ubiDGG mutants (purple) were incubated with ubiquitin (blue) and Ube2S and analyzed by Silver staining. (D) The TEK-box is not required in donor ubiquitin. TEK-box mutants of ubiquitin were mixed with WT-ubiDGG and Ube2S, and reactions were analyzed by Coomassie staining. (E) HADDOCK-based model of the ternary complex between the UBCUbe2S (yellow), donor ubiquitin (blue), and acceptor ubiquitin (pink; cluster 1, no. 1; see Table S2, bottom). (F) Surface representation of the Ube2S-binding interface on acceptor ubiquitin. Contact residues within a radius of 4 A˚ are shown in pink. (G) Surface representation of the acceptor-binding interface on the Ube2S-donor complex. (H) Acceptor binding is required for Ube2S activity. Ube2S mutants were incubated with ubiquitin and analyzed by Coomassie staining. (I) Ube2S residues at the acceptor interface are required for chain elongation by APC/C. The modification of Ub-L-cycA by APC/C and Ube2S mutants was analyzed by autoradiography. (J) Ube2S residues at the acceptor-binding interface are required in vivo. HeLa cell lines expressing Ube2SE131K and Ube2SR135E were treated with siRNAs to deplete endogenous Ube2S, arrested in prometaphase (t = 0 hr) or allowed to exit mitosis (t = 2 hr), and tested for K11-linked chains by aK11-western. See also Figure S5, Figure S6, and Table S2.
mutation of the opposing Glu139 of Ube2S (Figure S7F). The mutational and charge-swap analyses provide strong support for our model of acceptor recognition by Ube2S. Linkage Specificity Is Determined by SubstrateAssisted Catalysis The low affinity of Ube2S for acceptor ubiquitin raised the question of how this E2 achieves the stringent selection of K11 over other linkages. In one scenario, recognition of other Lys residues would be even less favored, with differences in binding energies accounting for the K11 specificity of Ube2S. To address this issue, we carried out docking calculations that placed each of the other Lys residues of ubiquitin in proximity to the active site of charged Ube2S. Among the clusters returned by HADDOCK, several had buried surface areas and
energy characteristics comparable to our K11-centered model (Figure S8; Table S3). This suggests that other Lys residues can be exposed to the active site of Ube2S, yet these binding events do not result in linkage formation. Thus, selective acceptor binding is not sufficient to explain the K11 specificity of Ube2S. Alternatively, the composition of the active site of Ube2S might force the reaction toward K11 linkages. Our models of the ternary complex found that the active site of Ube2S was similar to the E2 Ubc9 (Reverter and Lima, 2005; Figure 7C). For Ubc9, several residues in addition to the active site cysteine were attributed roles in catalysis: Tyr87 and Asn85 of Ubc9 contribute to pKa suppression of the substrate lysine through desolvation. Further, Asn85 serves to stabilize the oxyanion intermediate during ubiquitin transfer, and Tyr87 provides a hydrophobic Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 777
E Ubc9
Ube2S αC
K11
E34
G76
E131K WT WT WT E131K E131K
C93 Ube2S mutant:
Y87
Ube2S Silver
WT WT K6E K6E
Ub2-L-cycA Ub-L-cycA
Ub2-L-cycA Ub-L-cycA
HeLa
autoradiography
ubi~ubi
ubi
ubi
Ube2S
Ube2S
autoradiography
WT (pH7.5) KM: 1166 μM vmax: 4.9 pmol/min kcat: 0.0016 s-1
I 3
Coomassie
HeLa+ HeLa+ Ube2S Ube2SL129E
0 2 0 2 0 2 0 2 0 2 0 2
WT K6E K6E WT
Ubn-L-cycA
ubiquitin mutant:
ubi~ubi
Coomassie
F
E34 E34Q
Ube2S mutant:
vubi2 (pmol/min)
/
G E34 E34A E34K E34D
WT N87A L129E L129A S127D
WT K6E K6E WT
D
Ube2SUbe2S E51K/E131K mutant:
Ubn-L-cycA
N87
WT
B
Ube2S APC/C
N85
C95
ubiquitin mutant:
K524
L129
K6
H
Ube2S APC/C
D127
2.6Å
Ube2S mutant:
INPUT WT E3 4 E3 A 4 E3 K 4D WT E3 4Q
C E131
WT N8 7 L1 A 2 L1 9E 2 S1 9 A 27 D
A
siUBE2S time (h)
2.5 2
E34Q (pH 7.5) KM: 1111 μM vmax: 1.13 pmol/min
1.5 1
kcat: 0.0003 s-1
0.5 0 0
[ubiquitin] (μM)
K11-linked chains
/
Ube2S
250 500 750 1000 1250
E34Q (pH 8) KM: 1592 μM vmax: 3.59 pmol/min kcat: 0.0012 s-1
J
Silver
pH 9
pH 7.5
Ube2S
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 time (min)
β-actin
WT E34Q
Western
T14E I44A/V70A Coomassie
Figure 7. Substrate-Assisted Catalysis Contributes to the K11-Linkage Specificity of Ube2S (A) Charge-swap analysis of the ionic contact between acceptor Lys6 and Glu131 of Ube2S. Ube2S or Ube2SE131K was mixed with ubiDGG or ubiDGG/K6E and reactions were analyzed by Silver staining. (B) Ube2SE51K/E131K rescues mutation of Lys6 in both acceptor and donor ubiquitin. Lys6 was mutated in acceptor ubiDGG (purple) or donor ubiquitin (blue), and ubiDGG-ubi formation by Ube2S or Ube2SE51K/E131K was analyzed by Silver staining. (C) Ube2S (left) and Ubc9 (PDB ID: 2GRN; right) show similar active site constellations. The highest scoring Ube2S model of the HADDOCK run in the absence of ambiguous restraints is shown (Table S2, top; cluster 1, no 1). (D) Candidate active-site residues are required for the activity of Ube2S to catalyze ubi2 formation (ubiubi), as analyzed by Coomassie staining. (E) Active-site residues in Ube2S are required for chain elongation by APC/C. Ub-L-cycA was incubated with APC/CCdh1 and Ube2S mutants and analyzed by autoradiography. (F) Leu129 is required for Ube2S activity in vivo. HeLa cell lines expressing Ube2S or Ube2SL129 were tested for formation of K11-linked chains after endogenous Ube2S was depleted by siRNAs. K11-chain formation in cells arrested in prometaphase or exiting mitosis was monitored by aK11-western. (G) Glu34 of acceptor ubiquitin is required for K11-linkage formation. Ubiquitin mutants were incubated with Ube2S and analyzed by Coomassie staining. (H) Glu34 of acceptor ubiquitin is required for chain elongation by APC/C and Ube2S. The modification of Ub-L-cycA by APC/C, Ube2S, and ubiquitin mutants was analyzed by autoradiography. (I) ubiE34Q displays catalytic, but not binding defects. The rates of ubi2 formation at different concentrations of ubiquitin and ubiE34Q at the indicated pH were determined from two or three independent time courses. Apparent kinetic constants were obtained by fitting the rate constants to a Michaelis-Menten equation. (J) Rescue of ubiE34Q, but not other TEK-box or hydrophobic patch mutants, by increasing the reaction pH. Ubiquitin or indicated mutants were incubated with Ube2S at pH 7.5 (left) or pH 9 (right) and analyzed by Coomassie staining. See also Figure S7, Figure S8, and Table S3.
platform to position the attacking lysine side chain (Yunus and Lima, 2006). Whereas Asn85 of Ubc9 is conserved in Ube2S (Asn87) (Figure 7C), the attacking lysine is likely positioned by Leu129 of Ube2S (Figure 7C). Asn87 and Leu129 are essential for Ube2S activity in vitro (Figures 7D and 7E) and, as seen 778 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
with HeLa cell lines expressing Ube2SL129E, in vivo (Figure 7F). The mutation of Asn87 or Leu129 did not impede charging of Ube2S by E1 (Figure S7G). In addition to Asn85 and Tyr87, Asp127 of Ubc9 was assigned a catalytic role in reducing the pKa of the substrate lysine (Yunus
and Lima, 2006). In Ube2S, this residue is replaced by serine (Ser127), which our models place into a position to interact with the donor ubiquitin rather than activate an acceptor lysine (Figure 3). Strikingly, instead of an E2 residue, our models show an amino acid of ubiquitin, Glu34, to be in an appropriate position to orient the acceptor Lys11 and to promote its desolvation (Figure 7C). The mutation of Glu34 inhibited K11-linkage formation (Figures 7G and 7H), and kinetic analyses found this to be due to a strong reduction in the apparent catalytic rate constant (kcat) but not the Michaelis-Menten constant (KM) (Figure 7I). Thus, a residue in the substrate, Glu34 of ubiquitin, plays an important role in catalysis by Ube2S. Due to its position in the ternary complex, Glu34 is expected to orient Lys11, but not other Lys residues, and to suppress its pKa. The E34Q mutant of ubiquitin might maintain the position of Lys11 but fail to promote its deprotonation. If this assumption is correct, the inability of ubiE34Q to produce ubi2 may be rescued by increasing the pH of the reaction, which facilitates lysine deprotonation. Indeed, Ube2S efficiently linked ubiE34Q molecules at higher pH (Figure 7J), which was due to a change in the apparent kcat, but not KM (Figure 7I). By contrast, an acceptor TEK-box mutant defective in Ube2S binding (ubiT14E) and a donor ubiquitin mutant (ubiI44A/V70A) were inactive at pH 9 (Figure 7J). These findings support the notion that Glu34 of ubiquitin participates in catalysis by suppressing the pKa of the acceptor Lys11. As the catalytic role of Glu34 could be bypassed by increasing the pH, the same treatment might reduce the specificity of Ube2S. Consistent with this hypothesis, Ube2S modified Lys residues in a peptide derived from its C-terminal tail much more efficiently at pH 9 than at pH 7.5 (Figure S7H). Moreover, at pH 9, but not at pH 7.5, Lys63 and Lys48 of ubiquitin could act as acceptor for Ube2S (Figure S7I), although the bulk of linkage formation still occurred through K11 (Figure S7J). These findings further suggest that Ube2S requires a residue in ubiquitin, Glu34, for specific formation of K11 linkages. We conclude that Ube2S promotes linkage-specific ubiquitin chain formation by substrate-assisted catalysis. Conclusions K11- and K48-linked ubiquitin chains are often assembled by single-subunit E2s that cooperate with RING-E3s (Ye and Rape, 2009). Most of these E2s are specific and processive, but how these properties are achieved in the absence of cofactors was poorly understood. Here, we addressed this question by dissecting the mechanism of ubiquitin chain assembly by the K11-specific E2 Ube2S. We found that Ube2S requires a noncovalent interaction with the donor ubiquitin for chain formation. Although the affinity of Ube2S for the donor ubiquitin is weak, this interaction occurs in addition to the covalent thioester bond at the E2 active site. It tethers the donor ubiquitin to the E2, thereby restricting its flexibility and facilitating acceptor recognition. It also places the C terminus of the donor ubiquitin in an optimal position for nucleophilic attack by the acceptor lysine. The Ube2S-donor ubiquitin complex binds the acceptor ubiquitin very transiently through primarily electrostatic interactions. Ube2S recognizes the TEK-box on acceptor ubiquitin, a motif
previously identified as being required for formation of K11 linkages by Ube2C (Jin et al., 2008). As seen in crystal structures of K11-linked ubiquitin dimers (Matsumoto et al., 2010; Bremm et al., 2010), all TEK-box residues in the distal ubiquitin are fully accessible for recognition by Ube2S. The low affinity of Ube2S for the acceptor is in agreement with observations for other E2s (Petroski and Deshaies, 2005; Rodrigo-Brenni et al., 2010) and likely protects cells from spurious chain elongation in the absence of E3s. However, together with our comparative docking analysis, the transient nature of acceptor binding suggests that selective acceptor recognition does not explain the linkage specificity of Ube2S. Indeed, our model of the ternary complex between Ube2S, donor, and acceptor ubiquitin revealed that Ube2S lacks a residue required for suppressing the pKa of the substrate lysine. This function is instead provided by Glu34 of ubiquitin, which is in direct proximity to Lys11. Mutation of Glu34 had strong effects on the apparent kcat, but not the KM, of linkage formation by Ube2S, supporting a role in catalysis. Other Lys residues of ubiquitin do not have a suitably positioned acidic residue when docked into the active site of Ube2S or display features incompatible with catalysis (Figure S8). The same likely applies to Lys residues of APC/C substrates, which may explain why Ube2S is unable to promote chain initiation (Garnett et al., 2009; Williamson et al., 2009). Thus, our findings suggest that formation of a competent catalytic center requires residues of Ube2S and ubiquitin, which only occurs when K11 of the acceptor is exposed to the active site of Ube2S. We conclude that linkage-specific chain assembly by Ube2S occurs through substrate-assisted catalysis. Do other E2 enzymes use similar mechanisms for ubiquitin transfer? We found that noncovalent donor binding is a property shared by E2s with different linkage specificity. Ube2R1 and Ube2G2 also tether the donor ubiquitin for efficient catalysis, but not for K48 specificity, and Ube2R1 uses a similar surface on its UBC domain as Ube2S uses for donor recognition. In addition, the HECT-E3 Nedd4L binds E2-linked donor ubiquitin, a feature required for rapid ubiquitin transfer to the catalytic cysteine of the E3 (Kamadurai et al., 2009); the E3 RanBP2 binds SUMO to restrict its conformational freedom (Reverter and Lima, 2005); and in some cases, a ubiquitin-binding domain can promote E2-dependent ubiquitination reactions (Hoeller et al., 2007). We, therefore, propose that noncovalent donor binding is a general property of ubiquitination enzymes to increase the processivity of substrate modification. Other E2s may also use substrate-assisted catalysis for chain assembly. Mms2-Ubc13 positions Glu64 of ubiquitin close to the E2 active site, and mutation of this residue resulted in a decrease of K63-linkage formation (Eddins et al., 2006). Thus, although acceptor binding to Mms2 helps to orient Lys63 toward the active site of Ubc13, a catalytic ubiquitin residue might increase the specificity of chain formation. Moreover, mutation of a Tyr residue in ubiquitin reduced the catalytic rate of K48-linkage formation by yeast Ubc1 (Rodrigo-Brenni et al., 2010). Together, these findings allow us to propose that several E2 enzymes achieve linkage-specific ubiquitin chain formation through a mechanism of substrate-assisted catalysis. Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 779
EXPERIMENTAL PROCEDURES A detailed methods description can be found in the Extended Experimental Procedures. Reagents Table S4 shows a complete list of all constructs. Protein Purification Most proteins were purified from BL21/DE3 (RIL) cells. E1 was purified from Sf9 cells. For uniform isotopic enrichment, Ube2S and ubiquitin were expressed in M9-medium using 15N-enriched (NH4)2SO4 and/or 13C-enriched glucose. To generate ester-linked complex, 75 mM 15N-enriched UBCUbe2S/C95S and 230 mM unlabeled ubiquitin were incubated at 37 C for 3 hr. The diluted reaction was subjected to two rounds of anion-exchange chromatography. Formation of ubi2 60 mM ubiquitin and/or ubiDGG and 5 mM E2s were incubated with E1 and energy mix at 30 C for 1 hr and analyzed by Coomassie or Silver staining. In assays comparing activity at different pH, Tris/HCl was replaced with 50 mM Bis-tris propane, pH 7.5, 8, or 9. Ube2S Kinetics Assays Time courses of ubi2 formation were performed with different concentrations of WT- or E34Q-ubiquitin. Levels of ubi2 were quantified by Quantity One and compared to a known amount of Ube2S on each gel. Initial velocity rates and kinetic constants were calculated with GraphPad Prism and MichaelisMenten equations. APC/C Ubiquitination Assays 35 S substrates were synthesized by IVT/T. Ub-L-cycA was synthesized in the presence of 175 mM ubiK29R to inhibit the UFD-pathway, which is active in reticulocyte lysate. To purify 35S-Ub-L-cycA, HisCdk2 was bound to NiNTA. IVT/T was added to beads for 3 hr at 4 C. Beads were eluted with imidazole, and Ub-L-cycA/Cdk2-complexes were concentrated with 30 MWCO Microcon filters. APC/C was purified from G1-HeLa extracts and used for ubiquitination as described (Rape et al., 2006). Analysis of Ube2S Activity In Vivo HeLa cells were transfected in 6-well plates with 4 mg Ube2S vectors and Lipofectamine 2000. Twenty-four hours later, 10% of transfected cells were expanded to 10 cm dishes and selected with hygromycin B. Individual hygromycin-resistant colonies were picked with cloning discs and tested for Ube2S expression by western. Cells expressing Ube2S or mutants were transfected with 100 nM siRNA with Oligofectamine and synchronized in prometaphase by thymidine/nocodazole. Samples were taken at 0 hr and 2 hr post-release and processed for aK11-western. Computational Docking Donor docking was performed with HADDOCK 2.1, using crystal structures of UBCUbe2S (Protein Data Bank (PDB) ID: 1ZDN, chain A) and ubiquitin (PDB ID: 1UBQ). Active residues were based on chemical shift perturbation data and solvent accessibility. For thioester linkage between donor and Ube2S, we applied an unambiguous intermolecular distance restraint between C95 of Ube2S and G76 of ubiquitin. Residues 70–76 of ubiquitin were defined fully flexible. To generate a model of the ternary complex, we docked a second ubiquitin onto the selected E2-donor complex (cluster 1, no. 3; Table S1). We initially applied a single unambiguous restraint between K11 of the acceptor and C95 of Ube2S (Table S2, top), followed by a refined run with ambiguous restraints based on functional data (Table S2). Additional donor-docking experiments used ClusPro 2.0 and default parameters. NMR Data were recorded at 25 C on Bruker DRX spectrometers (500, 600, 800, and 900 MHz) and processed with NMRPipe. Backbone chemical shift assign-
780 Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc.
ments for UBCUbe2S and ubiquitin were obtained by standard triple resonance experiments. Titration experiments were performed by mixing stock solutions containing 240 mM 15N-enriched UBCUbe2S or 200 mM 15N-enriched ubiquitin and no or an 6-fold molar excess of unlabeled partner. Phase-sensitive gradient-enhanced 1H-15N HSQC spectra were recorded. To compare chemical shift perturbations, a weighted combined chemical shift difference Dd(1H15N) was calculated. 1 H-15N HSQC experiments of ester-linked complex between 15N-enriched UBCUbe2S/C95S and ubiquitin were recorded with 22 mM complex. ACCESSION NUMBERS The 1H, 15N, and 13C backbone chemical shift assignments for Ube2S (1–156) and ubiquitin have been submitted to the Biological Magnetic Resonance Bank (BMRB), http://www.bmrb.wisc.edu, with accession numbers 17437 and 17439, respectively. SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, eight figures, and four tables and can be found with this article online at doi:10.1016/ j.cell.2011.01.035. ACKNOWLEDGMENTS We thank H.J. Meyer for Ub-L-cycA; A. Williamson for golden extracts; members of the Rape and Kuriyan labs, J. Winger, S. Kassube, and J. Kirsch for discussions; J. Schaletzky for reading the manuscript and suggestions; J. Pelton for help with NMR experiments; D. King and T. Iavarone for mass spectrometry; and the staff at beamline 12.3.1 at LBNL for technical support. NMR instrumentation and operation were supported by NIH-GM 68933, NIH GM68933, NSF BBS 0119304 and NIH RR15756, and NSF BBS 8720134. S.L. is a fellow of The Leukemia & Lymphoma Society. M.R. is a Pew fellow and is supported by NIH GM83064 and an NIH New Innovator Award. Received: July 8, 2010 Revised: November 24, 2010 Accepted: January 31, 2011 Published: March 3, 2011 REFERENCES Bremm, A., Freund, S.M., and Komander, D. (2010). Lys11-linked ubiquitin chains adopt compact conformations and are preferentially hydrolyzed by the deubiquitinase Cezanne. Nat. Struct. Mol. Biol. 17, 939–947. Comeau, S.R., Kozakov, D., Brenke, R., Shen, Y., Beglov, D., and Vajda, S. (2007). ClusPro: Performance in CAPRI rounds 6-11 and the new server. Proteins 69, 781–785. Deshaies, R.J., and Joazeiro, C.A. (2009). RING-domain E3 ubiquitin ligases. Annu. Rev. Biochem. 78, 399–434. de Vries, S.J., van Dijk, A.D., Krzeminski, M., van Dijk, M., Thureau, A., Hsu, V., Wassenaar, T., and Bonvin, A.M. (2007). HADDOCK versus HADDOCK: new features and performance of HADDOCK2.0 on the CAPRI targets. Proteins 69, 726–733. Dikic, I., Wakatsuki, S., and Walters, K.J. (2009). Ubiquitin-binding domains from structures to functions. Nat. Rev. Mol. Cell Biol. 10, 659–671. Eddins, M.J., Carlile, C.M., Gomez, K.M., Pickart, C.M., and Wolberger, C. (2006). Mms2-Ubc13 covalently bound to ubiquitin reveals the structural basis of linkage-specific polyubiquitin chain formation. Nat. Struct. Mol. Biol. 13, 915–920. Garnett, M.J., Mansfeld, J., Godwin, C., Matsusaka, T., Wu, J., Russell, P., Pines, J., and Venkitaraman, A.R. (2009). UBE2S elongates ubiquitin chains on APC/C substrates to promote mitotic exit. Nat. Cell Biol. 11, 1363–1369. Hamilton, K.S., Ellison, M.J., Barber, K.R., Williams, R.S., McKenna, S., Ptak, C., Glover, M., and Shaw, G.S. (2001). Structure of a conjugating
enzyme-ubiquitin thiolester intermediate reveals a novel role for the ubiquitin tail. Structure 9, 897–904. Hoeller, D., Hecker, C.M., Wagner, S., Rogov, V., Do¨tsch, V., and Dikic, I. (2007). E3-independent monoubiquitination of ubiquitin-binding proteins. Mol. Cell 26, 891–898. Jin, L., Williamson, A., Banerjee, S., Phillip, I., and Rape, M. (2008). Mechanism of ubiquitin chain formation by the human Anaphase-Promoting Complex. Cell 133, 653–665. Jung, C.R., Hwang, K.S., Yoo, J., Cho, W.K., Kim, J.M., Kim, W.H., and Im, D.S. (2006). E2-EPF UCP targets pVHL for degradation and associates with tumor growth and metastasis. Nat. Med. 12, 809–816. Kamadurai, H.B., Souphron, J., Scott, D.C., Duda, D.M., Miller, D.J., Stringer, D., Piper, R.C., and Schulman, B.A. (2009). Insights into ubiquitin transfer cascades from a structure of a UbcH5B approximately ubiquitin-HECT (NEDD4L) complex. Mol. Cell 36, 1095–1102. Li, W., Tu, D., Li, L., Wollert, T., Ghirlando, R., Brunger, A.T., and Ye, Y. (2009). Mechanistic insights into active site-associated polyubiquitination by the ubiquitin-conjugating enzyme Ube2g2. Proc. Natl. Acad. Sci. USA 106, 3722–3727. Matsumoto, M.L., Wickliffe, K.E., Dong, K.C., Yu, C., Bosanac, I., Bustos, D., Phu, L., Kirkpatrick, D.S., Hymowitz, S.G., Rape, M., et al. (2010). K11-linked polyubiquitination in cell cycle control revealed by a K11 linkage-specific antibody. Mol. Cell 39, 477–484. Petroski, M.D., and Deshaies, R.J. (2005). Mechanism of lysine 48-linked ubiquitin-chain synthesis by the cullin-RING ubiquitin-ligase complex SCF-Cdc34. Cell 123, 1107–1120. Pierce, N.W., Kleiger, G., Shan, S.O., and Deshaies, R.J. (2009). Detection of sequential polyubiquitylation on a millisecond timescale. Nature 462, 615–619. Rape, M., Reddy, S.K., and Kirschner, M.W. (2006). The processivity of multiubiquitination by the APC determines the order of substrate degradation. Cell 124, 89–103.
Ryu, K.Y., Baker, R.T., and Kopito, R.R. (2006). Ubiquitin-specific protease 2 as a tool for quantification of total ubiquitin levels in biological specimens. Anal. Biochem. 353, 153–155. Schulman, B.A., and Harper, J.W. (2009). Ubiquitin-like protein activation by E1 enzymes: the apex for downstream signalling pathways. Nat. Rev. Mol. Cell Biol. 10, 319–331. Sheinerman, F.B., and Honig, B. (2002). On the role of electrostatic interactions in the design of protein-protein interfaces. J. Mol. Biol. 318, 161–177. Song, L., and Rape, M. (2010). Regulated degradation of spindle assembly factors by the anaphase-promoting complex. Mol. Cell 38, 369–382. VanDemark, A.P., Hofmann, R.M., Tsui, C., Pickart, C.M., and Wolberger, C. (2001). Molecular insights into polyubiquitin chain assembly: crystal structure of the Mms2/Ubc13 heterodimer. Cell 105, 711–720. Wagner, K.W., Sapinoso, L.M., El-Rifai, W., Frierson, H.F., Butz, N., Mestan, J., Hofmann, F., Deveraux, Q.L., and Hampton, G.M. (2004). Overexpression, genomic amplification and therapeutic potential of inhibiting the UbcH10 ubiquitin conjugase in human carcinomas of diverse anatomic origin. Oncogene 3, 6621–6629. Williamson, A., Wickliffe, K.E., Mellone, B.G., Song, L., Karpen, G.H., and Rape, M. (2009). Identification of a physiological E2 module for the human anaphase-promoting complex. Proc. Natl. Acad. Sci. USA 106, 18213–18218. Wu, T., Merbl, Y., Huo, Y., Gallop, J.L., Tzur, A., and Kirschner, M.W. (2010). UBE2S drives elongation of K11-linked ubiquitin chains by the anaphasepromoting complex. Proc. Natl. Acad. Sci. USA 107, 1355–1360. Xu, P., Duong, D.M., Seyfried, N.T., Cheng, D., Xie, Y., Robert, J., Rush, J., Hochstrasser, M., Finley, D., and Peng, J. (2009). Quantitative proteomics reveals the function of unconventional ubiquitin chains in proteasomal degradation. Cell 137, 133–145.
Reverter, D., and Lima, C.D. (2005). Insights into E3 ligase activity revealed by a SUMO-RanGAP1-Ubc9-Nup358 complex. Nature 435, 687–692.
Ye, Y., and Rape, M. (2009). Building ubiquitin chains: E2 enzymes at work. Nat. Rev. Mol. Cell Biol. 10, 755–764.
Rodrigo-Brenni, M.C., Foster, S.A., and Morgan, D.O. (2010). Catalysis of lysine 48-specific ubiquitin chain assembly by residues in E2 and ubiquitin. Mol. Cell 39, 548–559.
Yunus, A.A., and Lima, C.D. (2006). Lysine activation and functional analysis of E2-mediated conjugation in the SUMO pathway. Nat. Struct. Mol. Biol. 13, 491–499.
Cell 144, 769–781, March 4, 2011 ª2011 Elsevier Inc. 781
Yap1 Acts Downstream of a-Catenin to Control Epidermal Proliferation Karin Schlegelmilch,1,2,3,4,8 Morvarid Mohseni,1,2,3,8 Oktay Kirak,5 Jan Pruszak,3,5 J. Renato Rodriguez,5 Dawang Zhou,6 Bridget T. Kreger,7 Valera Vasioukhin,7 Joseph Avruch,6 Thijn R. Brummelkamp,5 and Fernando D. Camargo1,2,3,* 1Stem
Cell Program, Children’s Hospital, Boston, MA 02115, USA of Stem Cell and Regenerative Biology, Harvard University, Cambridge, MA 02138, USA 3Harvard Stem Cell Institute, Cambridge, MA 02138, USA 4Institute for Chemistry/Biochemistry, FU Berlin, Berlin 14195, Germany 5Whitehead Institute for Biomedical Research, Cambridge, MA 02142, USA 6Department of Molecular and Cellular Biology, Massachusetts General Hospital, Boston, MA 02114, USA 7Fred Hutchinson Cancer Research Center, Seattle, WA 18109, USA 8These authors contributed equally to this work *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.031 2Department
SUMMARY
During development and regeneration, proliferation of tissue-specific stem cells is tightly controlled to produce organs of a predetermined size. The molecular determinants of this process remain poorly understood. Here, we investigate the function of Yap1, the transcriptional effector of the Hippo signaling pathway, in skin biology. Using gain- and loss-of-function studies, we show that Yap1 is a critical modulator of epidermal stem cell proliferation and tissue expansion. Yap1 mediates this effect through interaction with TEAD transcription factors. Additionally, our studies reveal that a-catenin, a molecule previously implicated in tumor suppression and cell density sensing in the skin, is an upstream negative regulator of Yap1. a-catenin controls Yap1 activity and phosphorylation by modulating its interaction with 14-3-3 and the PP2A phosphatase. Together, these data identify Yap1 as a determinant of the proliferative capacity of epidermal stem cells and as an important effector of a ‘‘crowd control’’ molecular circuitry in mammalian skin. INTRODUCTION During mammalian development, proliferation and cell death of tissue-specific progenitor and stem cells (SCs) need to be tightly monitored and controlled to produce organs of a predetermined size. Experimental manipulation of SC numbers or activity during embryogenesis can thus have striking effects on the final size of certain organs (Depaepe et al., 2005; Kim et al., 2005; Stanger et al., 2007). Such exquisite regulatory mechanisms also orchestrate homeostasis of adult tissues and regenerative processes whereby the final shape and size of organs can be restored after cellular loss. Though a number of signaling molecules have been implicated in controlling SC proliferation, little is known about 782 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
endogenous mechanisms that ‘‘sense’’ or provide information about organ size. It is also unclear how these mechanisms relay needs for tissue growth to its resident SCs and how these cues are translated into SC proliferation or apoptosis. The Hippo signaling pathway was initially discovered in Drosophila as a potent mechanism that restricts tissue size by limiting cell proliferation and promoting apoptosis. At the core of this pathway is a kinase cascade composed of four tumor suppressors, including Hippo (Hpo) and Warts (Wts) and its regulatory proteins Salvador and Mats (Harvey et al., 2003; Lai et al., 2005; Pantalacci et al., 2003; Wu et al., 2003; Xu et al., 1995). The Hpo-Sav complex phosphorylates and activates the Wts-Mats complex, which in turn phosphorylates and inactivates the transcriptional coactivator Yorkie (Yki) (Huang et al., 2005; Oh and Irvine, 2008). Yki phosphorylation prevents its nuclear translocation (Dong et al., 2007; Oh and Irvine, 2008; Zhao et al., 2007), wherein it acts as a coactivator for the TEAD/TEF family transcription factor Scalloped (Sd) (Wu et al., 2008; Zhang et al., 2008). This core set of Hippo signaling components is highly conserved in mammals, and orthologs of Hpo (Mst1/2), Sav (WW45), Wts (Lats1/2), and Yki (Yap1) exhibit similar biochemical properties in cultured cells (Dong et al., 2007; Oka et al., 2008; Zhao et al., 2007). The importance of Hippo signaling in mammalian organ size control has been studied extensively in the liver, where loss of Mst1/2 or Sav or overexpression of Yap1 lead to hepatomegaly (Camargo et al., 2007; Dong et al., 2007; Lu et al., 2010; Song et al., 2010; Zhou et al., 2009). However, to what extent these predicted epistatic and functional relationships are conserved and active in other tissues remains uncertain (Zhou et al., 2009). Compared with the core kinase cascade regulating Yki phosphorylation, components acting upstream of this complex are less well defined. Earlier work in Drosophila has implicated the apical membrane-associated FERM-domain proteins Merlin (Mer), Expanded (Ex), and Kibra as pathway components upstream of Hpo (Baumgartner et al., 2010; Genevet et al., 2010; Hamaratoglu et al., 2006). Recent studies further implicated the apical transmembrane protein Crumbs (Crb) (Robinson et al., 2010) and the atypical cadherin Fat (Ft) (Hariharan,
2006) as modulators of the fly Hippo pathway. However, with the exception of the mammalian Merlin ortholog NF2 (Zhang et al., 2010), the significance and relevance of these proteins in regulating the activity of Yap1 in vivo in mammals are largely unclear. The identification of physiological upstream regulators of mammalian Hippo signaling could provide important insights into the mechanisms sensing and controlling organ size. The mammalian epidermis is a rapidly regenerating epithelial tissue whose maintenance depends on the self-renewing ability of epidermal SCs residing in the basal layer. After division, shortlived progenitor cells leave the basal layer and move through the suprabasal layers to the tissue surface as they terminally differentiate (Fuchs, 2007). Epidermal growth must be carefully balanced: inadequate proliferation results in thinning of the skin and loss of protection, and excessive growth is characteristic of hyperproliferative disorders. During development, the epidermis must be able to sense needs for expansion and replace basal vacancies, and during regeneration, cells must migrate and proliferate but also sense when to stop after wound closure. It has been postulated that adherens junctions (AJs), cadherin and catenin-based cell adhesion structures, might have a function in sensing epidermal cell density and restricting basal cell proliferation (Lien et al., 2006b). For instance, loss of some, but not all, AJ components, such as a-catenin or p120catenin, triggers severe epidermal hyperproliferation and tumors in transgenic mice (Kobielak and Fuchs, 2006; Perez-Moreno et al., 2006; Vasioukhin et al., 2001). The mechanisms acting downstream of these complexes are not fully understood. Intriguingly, it has been recently found that NF2 can modulate epidermal development via an association with AJs, specifically a-catenin, and the Par3 polarity complex (Gladden et al., 2010). These findings raise the possibility that signaling cues that regulate cell density and cell polarity could converge on the regulation of downstream Hippo signaling components. Here, we utilize gain- and loss-of-function models in mice to show that Yap1 is an essential regulator of epidermal maintenance and SC proliferative capacity. In addition, we identify a-catenin, an AJ component and a known tumor suppressor in the skin, as a crucial upstream negative regulator of Yap1 in epidermal cells in vitro and in vivo. Our work identifies a direct link between a signaling component of cell density-dependent AJs and a transcriptional regulator and provides a paradigm for the regulation of tissue expansion in response to extracellular cues. RESULTS Activation of Yap1 Can Expand the Epidermal SC Compartment In newborn mouse epidermis, Yap1 has a dynamic subcellular localization pattern in the basal compartment where it is present in the nucleus in some cells and in the cytoplasm/membrane of others. Expression of Yap1 is reduced and in a more diffuse nonnuclear pattern in the suprabasal-differentiated cell layers (Figure 1A). This pattern is also observed in cultured human keratinocytes in which high cellular densities lead to Yap1 relocalization out of the nucleus (Figure 1B). Thus, regulation of Yap1 subcellular localization might be important for the switch
between proliferative and terminally differentiating compartments. To test this, we utilized transgenic mice carrying a doxycycline (Dox)-inducible allele of a mutated version of Yap1 (S127A). This mutation results in the enhanced nuclear localization of Yap1 (Camargo et al., 2007; Dong et al., 2007; Zhao et al., 2007). To restrict Yap1 activation to the epidermis, we developed a binary genetic strategy based on the Cre-mediated activation of a reverse tetracycline-dependent transactivator selectively in the progeny of K14-expressing progenitors of stratified epithelia (Belteki et al., 2005; Dassule et al., 2000) (Figure S1A available online). Mice carrying these transgenes will be referred to as Tg hereafter. Eight days after Dox administration, adult Tg mice developed thickening and wrinkling of the skin. Histological analyses demonstrated that, contrary to the typical one-/twocell layer common to adult mouse epidermis, Tg skin developed into a multilayered epithelium (Figure 1C). Staining for the proliferation marker Ki-67 revealed a substantial increase in the number of proliferating basal cells and an extension of the proliferative domain into the suprabasal layers, normally quiescent in control skin (Figure 1D). Tg epidermis contained an abundance of cells expressing progenitor markers such as K5 and K14, whereas expression of the differentiation markers K10 and loricrin was mostly abrogated (Figure 1E and Figure S1B). Additionally, the number of cells expressing p63, a SC marker and essential regulator of stemness in stratified epithelia (Senoo et al., 2007), was amplified in Tg skin (Figure 1F). Tg skin failed to express hair follicle SC markers such as K15, or Sox9 (data not shown). These results suggested that activation of Yap1 induced expansion of an undifferentiated stem/progenitor cell population in the interfollicular epidermis. We then compared the clonogenic capacity of primary keratinocytes isolated 7 days after Dox administration. Skin of Tg mice had a >5-fold increase in the number of colony-forming cells compared to controls after Yap1 activation in vivo (Figure 1E). Moreover, serial plating assays, typically used to distinguish self-renewing SCs versus shorter-lived progenitors, revealed that Yap1 activation enhanced the SC proliferative capacity by 10-fold in secondary and tertiary platings (Figure 1G and Figure S1C). Taken together, these experiments demonstrate that activation of Yap1 promotes the proliferation of epidermal stem and progenitor cells and that its regulation is critical in maintaining normal skin homeostasis. Activation of Yap1 Can Lead to Squamous Cell Carcinoma-like Tumors Tg mice became ill 8 days after Dox induction, likely because of similar dysplastic changes in mucosal tissues such as the tongue (Figure S1D). In order to assess the long-term effects of Yap1 activation, newborn Tg and control skin was grafted onto Nude-mouse recipients followed by Dox administration. Epidermal thickening, hyperkeratosis, and stunted hair growth were evident in Tg-grafted mice shortly after grafting. Beginning at 20 days, Tg grafts developed large tumor-like masses that subsequently ulcerated (Figure 2A). Histological analyses revealed hyperplasia and multiple epidermal invaginations into the dermis as early as 9 days after grafting (Figure 2B). These lesions were epidermal in nature, as judged by staining with a pan-cytokeratin antibody and by their expression of keratin 6, Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 783
A
YAP
C
Ctr
Tg
density
low
B
high
YAP
F
Ctr
Ki67/β-int4
Tg
Tg
200 150
E
Ctrl * Tg
G
Tg
Ctr
100 50 0
Tg
p63/K5
Ctr
P2
4 # colonies / 10 cells
Ctr
# colonies / 10 5 cells
D
50 40
Ctrl Tg
**
30 20 10 0
Figure 1. Activation of Yap1 Leads to Epidermal Stem Cell Expansion (A) Yap1 expression (brown) indicated by immunohistochemical staining on E18.5 murine wild-type epidermis. (B) Immunofluorescence staining for Yap1 shows a dynamic localization pattern in human keratinocytes that is cell density dependent. (C) Hematoxylin and eosin (H&E) staining shows an abnormally thick epidermis and hyperkeratinization in adult Tg mouse skin after 8 days of doxycycline (Dox) induction. (D–E) Immunofluorescence analysis on frozen sections of adult epidermis shows an expansion of the proliferative Ki-67-positive compartment as well as the stem cell compartment (p63-positive, K5-positive) in Tg skin after 8 days of dox induction. Dashed line marks epidermal-dermal junction. (F) Rhodamine B staining of primary mouse keratinocytes, isolated from 8 day Dox-treated Tg mice and cultured for 9 days on feeders, shows a significant expansion of the epidermal stem cell compartment demonstrated by a higher colony-forming efficiency in Tg skin. (G) Rhodamine B staining and colony counts show a significant increase in the self-renewal capacity of colony-forming progenitors after Dox administration measured by serial replating assays (shown is passage 2, P2). Data are presented as mean ±. *p < 0.05; **p value < 0.01. Scale bars, 20 mm. See also Figures S1A–S1C.
which is expressed anomalously in hyperproliferative states (Figure 2C). Donor origin of the tumors was confirmed by their expression of the Yap1 transgene (Figure 2D and Figure S1E). Overall, the morphological perturbations resembled those of squamous cell carcinoma (SCC) in situ in human (Figure 2B). By 42 days, full dermal invasion was evident, and this was associated with the rupture of the basement membrane (Figure 2E). At this stage, tumors resembled human sarcomatoid SCC, characterized by the presence of invasive atypical proliferating spindle cells intermingled with keratinized centers (Figure 2B and Figures S1F and S1G). As predicted from their morphology, invading cells stained mostly negative for keratin markers but expressed the mesenchymal marker vimentin, indicating a potential epithelial-mesenchymal transition, a common occurrence in advanced human SCC (Figure 2E). Tumors were also assessed for their renewal by performing secondary transfers by subcutaneous injection into Nude mice. All transplantations gave rise to large masses of similar morphology as their primary counter784 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
parts (Figure S1H). Our results here demonstrate that proper regulation of Yap1 activity is essential for tumor suppression in the epidermis. Loss of Yap1 Results in Failure of Skin Expansion To test the role of endogenous Yap1 in epidermal biology, we generated a conditional floxed (fl) allele of Yap1 by gene targeting (Figure S2A). Deletion of loxP-flanked exons 1 and 2 of Yap1 after breeding to K14-Cre mice resulted in epidermis lacking Yap1 protein (Figures S2B–S2D). K14-Cre Yap1(fl/fl) mice will be referred to as cKO mice hereafter. cKO mice were either aborted or died shortly after birth. Gross examination of mutant embryos at E18.5 revealed thinner and fragile skin and absence of epidermal tissue covering the distal part of the limbs (Figure 3A). Assessment of skin permeability using the toluidine blue dye revealed a complete loss of epidermal barrier function in the regions described and around the mouth and nose (Figure 3B). Histological analyses of the proximal limb skin of cKO
A
Ctr
Tg 14d
Tg 42d
B
Ctr
Tg 9d
Tg 42d
D
Tg 9d
Yap
Tg 42d
C
CK
K6
Tg 9d
E
Vim
β 4-int
Tg 42d
Figure 2. Activation of Yap1 Leads to Tumor Formation (A) Gross morphology of Ctr and Tg grafts 14 and 42 days after beginning of Dox treatment. Note hyperkeratosis and stunted hair growth in Tg 14d graft and ulceration of skin and subcutaneous tumor mass in Tg 42d graft. (B) H&E staining of Ctr and Tg grafts 9 and 42 days after Dox induction show an invasion of underlying dermis by invasive keratinocytes derived from the Tg graft. Arrow points out to neighboring normal Nude mouse epidermis. White dashed line indicates border of carcinoma to normal Nude dermis. (C) Expression of cytokeratin (CK) and keratin 6 (K6) in Tg grafts treated 9 days with Dox as shown by immunohistochemistry. (D) Immunohistochemistry for Yap1 on Tg grafts with 9 and 42 days Dox treatment reveals epithelial origin of the tumor. Arrow points at Nude mouse epidermis. (E) Immunohistochemistry for vimentin (Vim) and integrin-b4 (b4-int) on Tg grafts with 42 days of Dox induction. Scale bars, 100 mm. See also Figure S1D–S1H.
mice revealed a thinner epidermis, reduced stratum corneum, and disorganized epidermal architecture (Figure 3C). Notably, the basal layer was hypoplastic, and basal cells had lost their columnar organization and adopted a flat morphology, typical of suprabasal differentiated cells (Figure 3C). These phenotypes were also observed in the back skin of cKO mice but were less severe. At E18.5, mutant cells expressing K10, an early differentiation marker, were nearly juxtaposed to the basement membrane in contrast to control skin where the basal layer was more pronounced (Figure 3D). Staining of cKO skin with progenitor markers demonstrated their abnormal morphology and reduction in numbers (Figure 3E). Staining with the proliferation marker phosphohistone H3 (pH3) revealed a >3.5-fold decrease in the number of proliferating basal cells in proximal limb skin of cKO mice (Figure 3F and Figure S2E). No discernible difference in apoptotic cells was detected, as judged by staining with active caspase 3 (data not shown). To directly address whether Yap1 deletion had an effect on the proliferative potential of epidermal SCs, we performed clonogenic assays. These experiments demonstrated a >50-fold decrease in the number of colony-
forming cells and a decrease in colony size in cKO epidermis (Figure 3G and Figure S2F). This reduction in proliferative potential is likely the cause of the absence of skin in distal limb areas, where due to growth demands during development, the proliferation rate at E18.5 is more than 3-fold that of back skin (Figure S2E). These results indicate that Yap1 is required for the maintenance of epidermal proliferative potential during development and, together with our gain-of-function experiments, suggest that levels of active Yap1 are a critical and physiological determinant of epidermal SC proliferative capacity. TEAD Interaction Is Critical for the Role of Yap1 in the Epidermis We next explored whether the ‘‘canonical’’ Hippo pathway components played a role in regulating Yap1 in the epidermis. In mammals, the transcription factors TEAD1–4 have been shown to interact with Yap1 and to mediate some of its proliferative effects in vitro (Vassilev et al., 2001; Zhao et al., 2008). To test whether TEAD factors were also involved in the regulation of epidermal biology, we generated a reporter construct carrying multimerized copies of the consensus TEAD DNA binding Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 785
D
Ctr
cKO
Ctr
B Ctr
C Ctr
cKO
cKO E Ctr
cKO
Lor/K5
cKO
Ctr
4/pH3
# of pH3+ cells / field
K10/ 4 F
cKO
10 8
Ctr cKO
G Ctr
80
Ctrl cKO
60
6 4
cKO
# colonies / 104 cells
A
40
***
20
2 0
0
**
Figure 3. Yap1 Is Required for the Maintenance of the Proliferative Capacity of Epidermal Stem Cells (A) Gross morphology of representative control (Ctr) and cKO E18.5 embryos. Note absence of skin in distal limbs, eyes, and ears and overall thinner skin. (B) Toluidine blue skin barrier assay in E18.5 mice reveals absence of epidermal barrier in limbs, ears, nose, and mouth. (C) H&E staining shows a decrease in epidermal thickness in cKO skin compared to the Ctr (indicated by bracket). Note the flat morphology of basal cells and disorganized epidermal architecture. Dashed line represents epidermal-dermal junction. (D and E) Immunofluorescence on frozen E18.5 epidermis reveals signs of basal cell depletion in cKO skin. Note the reduction of keratin 10 (K10)-negative, keratin 5 (K5)-positive progenitor/stem cells juxtaposed to the basement membrane (b4). Note the background staining for K5 in stratum corneum of cKO. (F) Marked reduction in number of proliferative basal cells in cKO limb skin as measured with an antibody against phosphohistone H3 (pH3). Results represent measurements of at least five different fields in three different animals. (G) Reduction in the number of colony-forming progenitors in E18.5 skin. Rhodamine B stain and colony counts were performed at day 8 after plating and are representative of two independent experiments. Data are presented as mean ± standard deviation (error bars). **p < 0.01; ***p < 0.005. Scale bars, 20 mm. See also Figure S2A–S2F.
sequence (TBS) to monitor their endogenous activity (Figure 4A). TBS reporter expression in a HaCaT keratinocyte cell line was highly responsive to Yap1, as its overexpression augmented reporter signal and its knockdown (KD) reduced reporter activity (Figure 4A). Proliferating cells in the periphery of colonies expressed the highest levels of the reporter, and reporter signal decreased when these cells were cultured at high densities, thus mimicking the observed pattern of Yap1 activity (Figure 4B). Given the functional overlap among the four mammalian TEAD proteins, we decided to test their role in epidermal biology in vivo by genetically disrupting their interaction with Yap1. Through homologous recombination in ES cells, we mutated the Serine at position 79 of Yap1 to Alanine (S79A) (Figures S3A and S3B). The ortholog serine in human Yap1 (S94) is necessary to 786 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
mediate a physical interaction with the TEAD proteins, but not with other DNA-binding cofactors (Zhao et al., 2008). The predicted consequences of this mutation were tested in Yap1fl/S79A mouse embryonic fibroblasts after administration of Cre, and we found that, whereas the overall levels of Yap1 protein were unchanged, its ability to interact with TEAD factors was mostly ablated (Figure 4C). We next generated K14-Cre Yap1fl/S79A (S79A) mice, in which only the Yap1 mutant protein would be expressed in the epidermis of mice starting at E14.5. These mice died at birth and fully phenocopied cKO mice (Figure 4D). S79A animals lacked epidermis in the distal part of the limbs, eyes, and ears. Mutant limb skin was thinner, with a reduced number of proliferating basal cells lying flat above the basal membrane (Figure 4D and Figure S3C). Clonogenic
B
A minP
A 79
Ctr Yapkd
5.0 ***
MFI
2.0 1.0 0.0 1.0 2.0 3.0
80
High Low
60 40
+/S
A
79
fl/S
A
79
+/S
TEAD1
Yap β-tub
20 **
0
2
0 10
3
10
4
10
mCherry
D
fl/S
+ Cre A 79
TEAD1
1349 6033
WCL
100
3.0
% of Cells
Reporter Activity
Basal S127A 4.0
- Cre
mCherry
IP:Yap
10x TBS
C
mCherry
5
10
E
Ctr
S79A
Ctr
S79A
Ctr
S79A
Figure 4. Yap1 Function in the Skin Is Mediated through TEAD Factors (A) A HaCaT keratinocyte cell line carrying multimerized TEAD-binding sequences (TBS) upstream of a minimal promoter (minP) and an mCherry reporter is responsive to expression of Yap1-S127A, whereas RNAi KD of Yap1 downregulates reporter levels. Cells that have been transfected with a scrambled control-RNAi display reporter activity at basal levels. (B) TBS reporter dynamics mimic Yap1 activity/localization. (Top) TBS-mCherry shows a dynamic fluorescence pattern with keratinocytes at the edge of a growing colony expressing higher mCherry levels. (Bottom) Reporter signal decreases significantly at high cellular densities. (C) Physical interaction between Yap1 and TEAD1 is virtually ablated in Adenovirus-Cre (+Cre)-treated Yapfl/S79A embryonic fibroblasts. Immunoprecipitation (IP) was performed with an anti-Yap1 antibody followed by immunoblotting for TEAD1. Loading control b-tubulin (b-tub). (D) Gross morphology of control (Ctr) and K14-Cre Yapfl/S79A E18.5 embryos. Note the absence of skin in distal limbs and thinner skin with flat basal cells. (E) Reduction in the number of colony-forming progenitors and proliferative potential in the skin of E18.5 K14-Cre Yapfl/S79A mice, as demonstrated by colony assays stained with Rhodamine B. Data are presented as mean ± standard deviation (error bars). **p < 0.01; ***p < 0.005. Scale bars, 20 mm. See also Figures S3A–S3C.
assays demonstrated a depletion of SC activity similar to cKO mice (Figure 4E). These results provide genetic evidence that TEADs are likely to be the major transcriptional partners of Yap1 in epidermal proliferation. a-Catenin Interacts with Yap1 and 14-3-3 The mammalian orthologs of the Drosophila hippo kinase, Mst1 and Mst2, function upstream of Yap1 to mediate its phosphorylation and inactivation (Zhao et al., 2007). Deletion of both Mst1 and Mst2 in the liver results in Yap1 activation and hepatomegaly, suggesting that these kinases can act as bona fide
negative regulators in vivo (Song et al., 2010; Zhou et al., 2009). We assessed whether deletion of Mst1/Mst2 in the skin would lead to similar phenotypes as Yap1 activation. Surprisingly, the epidermis of mice carrying a skin-specific deletion of Mst1/Mst2 displayed no abnormalities and no evidence of hyperplasia in mice up to 5 months of age (Figures S4A and S4B). This was accompanied by no changes in Yap1 S127 phosphorylation in Mst1/Mst2 null keratinocytes (Figure S4C). Additionally, KD of Mst1/Mst2 in the HaCaT-TBS reporter line produced no change in reporter activity or Yap1 nuclear localization (Figures S4D and S4E). Similarly, knockdowns of Lats1 and Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 787
Post-IP
D
WW1
WW2
ΔC
WQDP
WLDP
FLTWL
AQDA
ALDA
-----
r
S127
Ct
r Ct
Ya p1
pu t In
Ya p1
IP
A IB: α-cat
A
GST-αCatenin FLAG-Yap
B
S127A FLAG-Yap
+ +
+ + GST−α−catenin IP: GST IB: Flag
+
WCL
IP: GST IB: FLAG
IB: GST IB: Flag
IP: FLAG IB: GST
ap -Y
AG
WCL
IB: FLAG
E
IB: GST
C
Low Density
GST-Yap His-αCatenin MBP-14-3-3 λPPase
+ − − −
+ − − +
+ + − −
+ + − +
-Δ AG
FL
FL
+ + + −
+ + + +
1
C
1/2
W
7A
W
W
G-
A FL
-W
G LA
12
-S
G LA
F
F
G + − + −
His-αCatenin GST-Yap MBP-14-3-3 λPPase IP: His IB: FLAG
PD: GST IB: His
Yap/αcat
2
W
W
G-
A FL
+ + + +
+ + + −
Input CTR
Input Ctr
IB: GST
IB: His IB: MBP
Yap/αcat
F
GST-Yap His-αCatenin MBP-14-3-3 λPPase PD: His IB: GST
Input Ctr
IB: His
Yap αcat
IB: MBP
IB: GST
IB: GST IB: His
− − − −
+ − − −
+ + − −
+ + − +
+ + + −
+ + + +
H
FLAG-Yap FLAG-YapS127A GST-αCatenin His-14-3-3 His-14-3-3K49E
+ − − − −
+ − + − −
+ − + + −
+ − + − +
− + + + −
IP: GST IB: FLAG
Input CTR
Low Density + Ca +2
IB: MBP
IB: GST IB: Flag IB: His
Overlay
Figure 5. a-Catenin Interacts with Yap1 and 14-3-3 (A) Confluent HaCaT cell lysates (input) were used for an IP with an antibody against Yap1 or Pu.1 (Ctrl) and IP and post-IP material blotted with an antibody for a-catenin. (B) 293T cells were cotransfected with the indicated plasmids followed by IP with either GST or FLAG-beads. Co-IPs demonstrate an association between full-length Yap1 and a-catenin and a largely diminished association between a-catenin and the YapS127A mutant. IB, immunoblotting; WCL, whole-cell lysate. (C) Calcium-induced differentiation in primary human keratinocytes leads to translocation of Yap1 into membrane regions that colocalize with a-catenin. (D) Characterization of Yap1 domains that are important for the interaction with a-catenin. (E and F) In vitro pull-down (PD) of recombinant Yap1 and a-catenin in the presence or absence of 14-3-3. (G) In vitro pull-down of recombinant a-catenin and 14-3-3 in the presence of phosphorylated recombinant Yap1. (H) Immunoprecipitation of Yap1 and a-catenin demonstrates that the biochemical interaction is dependent on 14-3-3. Scale bars depicted are 10 mm. See also Figures S4A–S4E, Figures S5A and S5B, and Table S1.
Lats2, the orthologs of Drosophila warts, did not lead to increased reporter activity or enhanced Yap1 nuclear localization (Figures S4D and S4E). These results suggested that Yap1 might be regulated by alternative mechanisms in keratinocytes that are independent of the predicted canonical Hippo pathway kinases. In order to identify potential alternative upstream regulators of Yap1 activity, we performed coimmunoprecipitation (co-IP) experiments followed by mass spectrometry (MS) analysis in high-density HaCaT cultures. Unexpectedly, this screen 788 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
revealed that the most common interaction partner of Yap1 in keratinocytes corresponded to a-catenin, an important tumor suppressor in stratified epithelia (Table S1). We next validated our mass spectrometry data by characterizing the interaction of Yap1 with a-catenin. Endogenous co-IP experiments in confluent HaCaT cells and keratinocytes directly isolated from skin, and pull-downs following transfection of tagged forms of Yap1 and a-catenin in 293T cells further demonstrate the biochemical interaction (Figures 5A and 5B and
Figure S5A). Confocal microscopy studies in primary human keratinocytes corroborate these findings and reveal that Yap1 can begin to colocalize with a-catenin shortly after induction of differentiation following calcium treatment (Figure 5C). With more time in calcium or at high cell densities, nuclear Yap1 is depleted, with some of it colocalizing with a-catenin at the membrane and some accumulating in a cytoplasmic pool (Figure S5B). Phosphorylation of Yap1 at Ser127 creates a 14-3-3 binding site that has been shown to be crucial for the regulation of its activity (Zhao et al., 2007). Remarkably, disruption of 14-3-3mediated regulation by mutation of Yap1 Ser127 to alanine nearly abolished the interaction between Yap1 and a-catenin (Figures 5B and 5D). The participation of 14-3-3 in Yap1 binding to a-catenin is further evidenced by the presence of a large number of 14-3-3 peptides found in our Yap1 mass spectrometry pull-downs (Table S1). We also find that the WW domains of Yap1, previously shown to be important for negative regulation by inhibitory kinases (Oka et al., 2008), are also required for its efficient interaction with a-catenin (Figure 5D). We next examined whether the interaction between a-catenin and Yap1 was direct using in vitro binding assays. Bacterially purified a-catenin and recombinant Yap1 purified from Sf9 cells were coincubated and subjected to pull-down assays. Our results reveal a lack of direct interaction between these two proteins (Figures 5E and 5F). Given that phosphorylation at Yap1 S127 and subsequent 14-3-3 binding are critical for the Yap1 and a-catenin association in cells, we tested whether 14-3-3 could allow for the formation of a tripartite complex in vitro. Indeed, the addition of bacterially purified 14-3-3 resulted in the association of Yap1 and a-catenin (Figures 5E and 5F). As predicted, this association was dependent on Yap1 phosphorylation, as pretreatment of phosphorylated recombinant Yap1 with lambda phosphatase resulted in loss of complex formation (Figures 5E and 5F). Our results also provide evidence for a direct physical interaction between a-catenin and 14-3-3, which can be enhanced by the presence of phoshoYap1 (Figure 5G). Additionally, expression of a dominant-negative 14-3-3 mutant (K49E) in mammalian cells leads to a reduction of Yap1 and a-catenin binding (Figure 5H). Thus, these data uncover a-catenin as a binding partner for Yap1 and demonstrate that 14-3-3 is critical for this biochemical interaction. a-Catenin Regulates Yap1 Localization and Activity To test whether AJs are functionally important for the nonnuclear localization and inactivation of Yap1 at high cellular densities, keratinocyte cultures were treated with the calcium chelator EGTA. After only 1 min of EGTA treatment, the subcellular localization of Yap1 turned predominantly nuclear (Figure 6A and Figure S5C). The rapid kinetics of these observations indicates that AJs play a functional role in inactivating Yap1 and that the same pool of Yap1 can rapidly relocalize to the nucleus. To more specifically dissect the role of individual AJ proteins in regulating Yap1 activity, we depleted their expression in HaCaT and 293T cell lines carrying the TBS reporter. Remarkably, a robust activation of the reporter was observed after a-catenin KD, similar to levels obtained after depletion of the known upstream regulator NF2. A weaker but significant response was obtained after
depletion of p120-catenin, and no reporter activation was observed after KD of b-catenin, E-cadherin, P-cadherin, or both E- and P-cadherins (Figure 6B and Figures S5D and S5E). a-catenin depletion activated the TBS reporter in several other cell lines tested, indicating a conserved phenomenon across epithelial cell types (Figure S5F). Silencing of a-catenin in confluent keratinocytes resulted in Yap1 relocalization to the nucleus and the formation of a Yap1/TEAD nuclear complex (Figure 6C-D). a-catenin depletion was also accompanied by a reduction in the inhibitory phosphorylation at Yap1-S127 and minimal changes in the activation status of the Mst and Lats kinases (Figure 6E), suggesting that the regulatory effects of a-catenin might be independent of these kinases. To extend these observations and to test whether a-catenin could be an important negative regulator of Yap1 in vivo, we examined Yap1 localization in a-catenin mutant tissues. We first analyzed Yap1 localization in epidermis of E18.5 mice with a K14-Cre driven deletion of a-catenin. Skin of control mice displayed the typical heterogeneous subcellular localization pattern in which only a few basal cells display predominant nuclear Yap1 protein (Figure 6F). This pattern was drastically changed in a-catenin mutant epidermis where basal cells were, in their great majority, positive for nuclear Yap1 and showed substantially enhanced nuclear staining (Figure 6F and Figure S5G). Additionally, in some regions, we observed nuclear Yap1 staining extending into the suprabasal cell compartment. Nuclear Yap1 was never seen suprabasally in control tissue (Figure 6F). We next evaluated the localization of Yap1 in a panel of 19 human skin SCCs that expressed varying levels of a-catenin (Figure S5H). As shown in Figure S5H, no or low expression of a-catenin clearly correlated with enhanced nuclear Yap1 localization in these tumors. Having established that a-catenin is crucial for the inactivation of Yap1 in keratinocytes in vitro and in vivo, we tested whether ectopic expression of a-catenin could result in the recruitment of Yap1 to AJs. Remarkably, enforced expression of a-catenin in low-density cultures led to the relocalization of Yap1 to membrane regions containing AJ components (Figure 6G and Figure S6A). To test whether a different AJ protein could mediate this recruitment of Yap1 to the membrane, we performed similar experiments with E-cadherin. Cells that overexpressed E-cadherin, however, did not show localization of Yap1 to membrane regions (Figure S6B). These observations support the idea that a-catenin plays a direct role in the recruitment of Yap1 to AJs. This result, together with the siRNA depletion data (Figure 6B), speaks to the unique ability of a-catenin, and not E/P-cadherin or b-catenin, to regulate Yap1 localization and activity. These findings also argue that Yap1 hyperactivation is not due simply to loss of AJ-mediated cellular adhesion. Genetic ablation of a-catenin in murine epidermis leads to keratinocyte hyperproliferation and squamous cell carcinoma (Kobielak and Fuchs, 2006; Vasioukhin et al., 2001), observations that are highly reminiscent of the Yap1 gain-of-function phenotypes described here. Altogether, our data suggest that the growth-suppressive effects of a-catenin in the epidermis might be mediated through inactivation of Yap1. We tested this prediction experimentally and found that reduction of Yap1 levels in a a-catenin KD HaCaT cell line rescued its Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 789
Figure 6. a-Catenin Regulates Yap1 Localization and Activity (A) Disruption of AJs in high-confluent keratinocytes with EGTA results in Yap1 nuclear localization after 1 min of treatment. (B) Knockdown (KD) of AJ proteins in either HaCaT or 293T cells carrying the TBS-reporter. KD of NF2 is shown as a positive control. (C and D) KD of a-cat in confluent keratinocytes leads to Yap1 nuclear localization, and interaction with its nuclear partner Tead1 (D). (E) siRNA-mediated depletion of a-cat in high-density HaCaT cultures leads to loss of Yap phosphorylation at serine 127 (pYap) and unchanged levels of activated Mst1/2 (pMst) or Lats1/2 (pLats). (F) Immunohistochemistry for Yap1 in Ctrl and K14-Cre conditional a-catenin mutant (cKO) E18.5 epidermis. Note the enhanced nuclear staining in basal and suprabasal cells of cKO mice. (G) Immunofluorescence detection of Yap1 (purple) localization in low-density keratinocytes ectopically expressing GST-a-cat (green). Keratinocytes overexpressing a-cat show Yap1 localization to sites of cell-cell contact, whereas untransfected cells show nuclear staining and an absence of Yap1 staining at the cell membrane. (H) Stable knockdown of a-catenin (a-cat-KD) promotes hyperproliferation in HaCaT cells. However, doxycycline (dox)-induced KD of Yap1 in a-cat-KD cells slows down the rate of cell proliferation to control (Ctr) levels. (I) Ectopic expression of a-cat in a HaCaT cells suppresses cell proliferation (a-catOE). This growth inhibition is rescued by the expression of a Dox-inducible Yap1S127A mutant (Yap1S127A+Dox). Data are mean ± standard deviation (error bars). ***p < 0.001. Scale bars, 20 mm (A and C), 10 mm (F), and 5 mm (G). See also Figures S5C–S5H and S6A–S6E.
hyperproliferative phenotype to the level of control keratinocytes (Figure 6H and Figures S6C–S6E). Conversely, ectopic expression of a-catenin suppressed keratinocyte proliferation (Figure 6I), but expression of Yap1-S127A was able to rescue these inhibitory effects (Figure 6I). a-Catenin Binding Prevents Yap1 Dephosphorylation and Activation Knockdown of a-catenin results in decreased Yap1 S127 phosphorylation, which is likely the cause of increased Yap1 nuclear localization and activity. To begin understanding the potential mechanisms linking a-catenin levels and Yap1 phosphorylation, we performed a Yap1 co-IP followed by MS in a-catenin KD cells. These experiments revealed a selective interaction of 790 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
Yap1 with the catalytic subunit of the protein phosphatase 2A (PP2Ac) in cells with reduced a-catenin levels (data not shown). Co-IP and immunoblotting experiments confirmed this observation (Figure 7A). These results suggested that loss of a-catenin could result in a more efficient association of Yap1 and PP2Ac, which could potentially dephosphorylate Yap1-S127, triggering its activation. PP2A is a highly promiscuous enzyme that catalyzes the dephosphorylation of a wide range of substrates. Purified PP2Ac, in fact, is able to efficiently dephosphorylate Yap1 S127 in vitro in an okadaic acid-dependent manner (Figure 7B). Importantly, reduction of PP2Ac levels in a-catenin KD cells suppresses the increase in TBS reporter activity and S127A phosphorylation observed after depletion of a-catenin (Figure 7C and Figure S7). Thus, PP2Ac acts downstream or in
E
8 7
Fold Activation
r Sc
C
IP: Yap1
aC at -k d Sc r
Input
aC at -k d
A
IB: PP2A IB: 14-3-3
***
5 4 3
**
2
**
1
IB: Yap
0
Ct r NF 2 αC PP at 2 PP Ac + 2A α- c Ca t
+ − + −
+ + + +
+ + + −
PP2A
Yap1
P
+ − + +
P
Yap1
P
Yap1
P P
Yap1
IB: Yap
IB: MBP
TEAD
Yap1
PP2A
Yap1
14-3-3
IB: Yap
PP2A
Yap1
KINASE ?
IB: pYap
IB: pYap
14-3-3
+ − − −
P
5 0.0 0.25 1.0
+ − − +
14-3-3
Ctr OA
Yap1
GST-Yap His-αCatenin MBP-14-3-3 PP2Ac
α catenin α-c en n
14-3-3
+ PP2Ac
P
D
B
α-catenin 14-3-3
IB: αCatenin
Low Cell Density/ α-catenin-null
High Cell Density
6
TEAD
IB: His
Figure 7. a-Catenin Binding Prevents Yap1 Dephosphorylation and Activation (A) PP2Ac associates with Yap1 in a-catenin-depleted HaCaT cells. The association of Yap1 and 14-3-3 is diminished in the absence of a-catenin. (B) Yap1 is dephosphorylated by PP2Ac in vitro. Different concentrations of purified PP2A were incubated with Sf9-purified recombinant Yap1 for 30 min. (C) siRNA KD of a-catenin increases TBS reporter activation and the dephosphorylation of Yap1. Simultaneous KD of a-catenin and PP2Ac suppresses TBS reporter transcriptional activity and increase in Yap1 phosphorylation. (D) Western blot analysis of phospho-Yap levels following an in vitro dephosphorylation competition assay of Yap1 by PP2Ac in the presence of recombinant 14-3-3 and a-catenin. (E) Schematic model of density-dependent and a-catenin-mediated Yap1 activation. Data are presented as mean ± standard deviation (error bars). **p < 0.01; ***p < 0.001. See also Figure S7.
parallel of a-catenin, and the increased Yap1 activity in a-catenin KD cells is dependent on PP2Ac activity. An antagonistic and competitive interaction has been reported between PP2A and 14-3-3 for different phosphosubstrates (Chiang et al., 2003; Martin et al., 2008). Hence, we tested whether KD of a-catenin would result not only in increased association of Yap1 with PP2Ac, but also in reduced binding to 14-3-3. In effect, Yap1 does not seem to be bound to 14-3-3 in a-catenin KD keratinocytes (Figure 7A). To determine whether a Yap1, 14-3-3, and a-catenin complex could be protective against PP2Ac-mediated dephosphorylation, an in vitro dephosphorylation assay was performed in the presence and absence of these complex-forming proteins. Though the presence of 14-3-3 alone was not able to mitigate phosphatase activity, addition of a-catenin to the complex resulted in diminished dephosphorylation (Figure 7D). One of the many PP2A complexes has been previously implicated in the regulation of Hippo signaling by controlling phosphorylation of the Hpo kinase in Drosophila (Ribeiro et al., 2010). Our experiments in mammalian cells extend the role of PP2Ac as a regulator of mammalian Hippo signaling and suggest that a-catenin might regulate Yap1 activity by modulating its ability to be dephosphorylated at S127. DISCUSSION The present findings provide a mechanistic paradigm for the regulation of SC proliferation and tissue expansion in response
to extracellular cues. In this molecular circuitry, cellular ‘‘neighborhood’’ information is provided by density-dependent cellcell junctions (Figure 7E). A critical and unique component of these junctions, a-catenin, acts as a signaling molecule to translate this context-dependent information into SC proliferation and tissue expansion. a-catenin achieves this through modulating the activity of Yap1, the transcriptional coactivator of the sizerestricting Hippo signaling pathway. Yap1 as a Critical Regulator of Epidermal SCs Relatively little is known about the transcriptional networks controlling epidermal SC maintenance. Arguably, only p63 and Tcf3/Tcf4 have been implicated as essential factors for the long-term proliferative capacity of epidermal SCs in vivo (Nguyen et al., 2009); Senoo et al., 2007). Our work here provides loss- and gain-of-function genetic evidence indicating that Yap1 and the TEAD proteins are critical endogenous regulators of epidermal SC and progenitor proliferation. While this manuscript was under review, Zhang et al. reported that transgenic activation of Yap1 in the skin led to an expansion of basal epidermal progenitors and inhibition of terminal differentiation (Zhang et al., 2011). These authors also find that nuclear Yap1 progressively declines with age and that it correlates with the proliferative potential of epidermal progenitors. Our work corroborates and extends these observations, and together they highlight the emerging role of Yap1 and TEAD factors in skin homeostasis. In the future, it will be important to identify Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 791
the transcriptional targets of this complex that mediate its function. Our results also extend previous observations linking Yap1/ TEAD activity to SC function. In the intestine, Yap1 activation leads to the expansion of an undifferentiated progenitor population (Camargo et al., 2007), whereas its deletion causes no obvious defects during development or homeostasis, though it impairs regeneration (Cai et al., 2010). Loss of Yap1 in the liver leads to defective biliary development but unimpaired regeneration after partial hepatectomy (Zhang et al., 2010; F.D.C., unpublished data). Thus, it seems that the requirement for endogenous Yap1 in tissue-specific SCs varies significantly according to the cellular context. The intrinsic need for Yap1/ TEAD activity in epidermal SCs during development is more akin to the role that this complex plays in human embryonic SCs, in which it is required for their self-renewal and suppression of differentiation (Lian et al., 2010). a-Catenin as a Regulator of Yap1 Activity Our findings provide additional insight into the role of a-catenin in growth and tumor suppression. Previously, a-catenin was considered to be a simple structural linker between the cadherin/catenin complex and the actin cytoskeleton. During the past several years, it has become evident that a-catenin also functions as a signaling molecule in a variety of cellular processes in addition to its role in cellular adhesion (Stepniak et al., 2009). Interestingly, ablation of epidermal E- and P-cadherins leads to AJ disruption, defects in intracellular adhesion, and epidermal integrity but no significant increase in cellular proliferation (Tinkle et al., 2008). These findings, together with our data (Figure 6B and Figure S5D), demonstrate that hyperproliferation and Yap1 activation in the absence of a-catenin are not simply due to loss of cell adhesion and epithelial integrity. Additionally, in human cancers, loss of a-catenin is usually a stronger prognostic factor than loss of Ecadherin (Benjamin and Nelson, 2008). Thus, though it seems that AJ components function coordinately in mediating keratinocyte adhesion, they differ in their ability to influence proliferative responses. Particularly, it was shown that aberrant activation of the Ras-MAPK pathway was partially responsible for the hyperproliferative phenotypes observed in a-catenin mutant skin (Vasioukhin et al., 2001). It is still unclear to what extent Yap1, Ras-MAPK, or other signaling pathways contribute to growth suppression mediated by a-catenin in the skin. Whether Yap1 mediates all or only part of a-catenin tumor-suppressive activity awaits the results of epistatic experiments in vivo. 14-3-3 and Yap1 Phosphorylation in the Skin Our data would suggest that 14-3-3 proteins are also critical growth suppressors in the epidermis. The 14-3-3 proteins are a large group of highly conserved and homologous dimeric proteins. A particular isoform, 14-3-3s, is unique among the family in that it is selectively expressed in stratified epithelia. Interestingly, 14-3-3s is barely detectable in basal cells but is abundantly expressed in the suprabasal layer of human epidermis (Dellambra et al., 2000). Downregulation of 14-3-3s in primary keratinocytes results in the expansion and immortalization of cells expressing p63 (Dellambra et al., 2000; Pellegrini et al., 2001). Additionally, the epidermis of mice carrying a domi792 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
nant-negative mutation in 14-3-3s displays hyperplasia, expansion of K14 and K5-expressing progenitor cells, and enhanced proliferative capacity in vitro (Herron et al., 2005; Li et al., 2005). These phenotypes are reminiscent of Yap1 overexpression and a-catenin deletion in the epidermis and further indicate that 14-3-3s is part of a critical molecular circuitry controlling epidermal progenitor expansion. As predicted from work in Drosophila, Mst1/2 and Lats1/2 are critical for Yap1 S127 phosphorylation in a wide array of cell lines (Dong et al., 2007; Oh and Irvine, 2008; Zhao et al., 2007). Our data here demonstrate that Mst1/2, and potentially Lats1/2, are dispensable for Yap1 phosphorylation and activity in the epidermis. Our results parallel those of Zhao et al. and identify an unexpected diversity in the mammalian kinase components that are necessary to phosphorylate Yap1 in relation to what is predicted from work in Drosophila. For instance, in mouse liver, Mst1/2 signal through an unknown intermediary kinase that is distinct from Lats1/2 to control Yap1 phosphorylation. On the other hand, mouse embryonic fibroblasts do not require Mst1/ 2 but do depend on Lats1/2 to phosphorylate Yap1 (Zhao et al., 2007). We have tested whether pharmacological manipulation of MAP kinase and insulin signaling, pathways that were previously shown to be involved in a-catenin signaling (Vasioukhin et al., 2001), had effects on Yap1 phosphorylation. Inhibition of IGFR, JNK, and Src kinases resulted in no relative change in Yap1-S127 phosphorylation (data not shown). Thus, the identity of the kinases that are responsible for Yap1 inactivation in the epidermis remains unknown. Several examples exist in the literature that demonstrate that competition between 14-3-3 proteins and PP2A complexes for a substrate is critical for regulation (Chiang et al., 2003; Martin et al., 2008). Our data suggest that binding of the Yap1/14-3-3 complex to a-catenin would stabilize this complex and/or inhibit access of PP2Ac (Figure 7). When binding to a-catenin is lost, PP2Ac could then more efficiently compete for binding to S127 in Yap1 and could dephosphorylate it, allowing Yap1 to translocate into the nucleus, bind to TEADs, and activate a proproliferative gene expression program (Figure 7). Recently, a-catenin has been found to interact with NF2, another upstream component of the Hippo pathway (Gladden et al., 2010). Interestingly, it was reported that NF2 could link a-catenin to Par3, thus providing a bridge between AJs and polarity complexes. Together with the recent identification of the apical polarity protein Crumbs as a functional regulator of the Hippo pathway (Chen et al., 2010; Grzeschik et al., 2010; Ling et al., 2010; Robinson et al., 2010; Varelas et al., 2010), these findings raise the intriguing possibility that a signaling complex of a-catenin, NF2, and other cell polarity proteins that are associated with the cell membrane could make up an important signaling hub that may allow cells to sense cues such as density and positioning. These complexes then could relay those signals to transcriptional effectors such as Yap1 (see below). The LIM protein Ajuba might provide an additional avenue for crosstalk between AJs and Hippo signaling. Ajuba can interact with a-catenin and has been recently shown to inhibit Yap1 phosphorylation via interaction with Lats (Das Thakur et al., 2010). Further studies are needed to determine whether and to what extent a-catenin, NF2, and polarity complexes in epidermal cells control proliferation through Yap1.
Crowd Control and the Hippo Pathway An important longstanding question in the Hippo signaling field is the identity of extracellular signals that modulate the pathway’s activity. It has been speculated that this pathway could be involved in interpreting morphogen gradients to regulate tissue size, and therefore the existence of a secreted or membrane ligand has been postulated (Harvey and Tapon, 2007). Our data here suggest that, in the epidermis, this regulation is, at least partly, mediated by a-catenin and AJs. Based on the massive overgrowth phenotypes obtained by deletion of a-catenin in the skin and the developing brain, Vasioukhin et al. postulated that AJs could act as molecular biosensors of cell density and positioning (Lien et al., 2006a, 2006b). Our data here support and extend this notion and indicate that Yap1, through its interaction with a-catenin, is a critical mediator of this ‘‘crowd control’’ molecular circuitry in the epidermis (Figure 7). In this model, increased cellular density or differentiation (which is sensed by an increased number of AJs), higher calcium concentration (Menon and Elias, 1991), and selective expression of 14-3-3 in the suprabasal layers limit SC expansion by inactivating Yap1. Low basal cell density, as in a growing embryo or after wounding, would translate into nuclear Yap1 and proliferation. When this molecular network is defective, e.g., by deletion of a-catenin, inactivation of 14-3-3, or activation of Yap1, hyperproliferation and tumors can arise. Further elucidation of the transcriptional program controlled by Yap1 and identification of modulators of its interaction with a-catenin, including the kinase that mediates Yap1 inactivation in keratinocytes, will likely provide insights into the mechanisms that underlie SC maintenance, tissue expansion, and tumorigenesis. EXPERIMENTAL PROCEDURES Generation of Gene-Targeted Mice For generation of cKO mice, a targeting strategy was designed to allow Cremediated deletion of exon 1 and 2 of Yap1. Progeny carrying the targeted Yap1 allele were bred to K14-Cre mice. To generate the mutated Yap1S79A allele, a point mutation was introduced by site-directed PCR mutagenesis in the mYap1-targeting construct substituting a serine at amino acid position 79 (TCC) for an alanine (GCG). Mice were generated in the same way as described for the Yap1fl/fl mice. Mice carrying the Yap1S79A allele were bred to K14-Cre Yap1fl/+ mice to generate K14 Cre Yap1S79A/fl mice. Cell Isolation, In Vitro Culture, and Clonogenic Assays Primary keratinocytes isolated from the epidermis of E18.5 control and transgenic mice were cultured on mitomycin-C-treated 3T3 feeder cells in FAD medium. For clonogenic assays, colonies of primary mouse keratinocytes were fixed and stained with 1% rhodamine B (Sigma) 9 days after initial plating. Skin Engraftments and Barrier Function Assay Full thickness skins of E18.5 K14-Cre Yap1fl/fl and littermate Ctr mice were removed from torsos and spread on a sterile petri dish and briefly stored at 4 C. During this time, each skin graft recipient site was prepared by removal of a patch of full thickness skin on the back of an anesthetized Foxn1nu/ Foxn1nu mouse. For skin barrier assays, newborn mice were rinsed in PBS and successively dehydrated in methanol for 1 min each. Mice were then rehydrated with PBS and stained in 0.1% toluidine blue in PBS, destained with PBS for 20 min, and then photographed. Mass Spectrometry and Proteomic Analysis Approximately 4 mg of precleared high-density HaCaT cells were used for Yap1 immunoprecipitation. Immunoprecipitates were loaded onto
SDS-PAGE and stained with colloidal Coomassie. Mass spectrometry analysis was performed at the proteomics facility at Children’s Hospital Boston. Dephosphorylation Assay Catalytically active recombinant PP2A (PP2Ac) was purchased from Millipore. Recombinant proteins were incubated in the presence of varying concentrations of PP2Ac. For PP2Ac and 14-3-3 competition assays, Yap1, a-catenin, and 14-3-3 were preincubated in 50 mM Tris (pH 7.4), 150 mM NaCl, 5 mM MgCl2, and 0.02% Triton X-100 for 1 hr prior to PP2Ac addition. SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, seven figures, and one table and can be found with this article online at doi:10.1016/j. cell.2011.02.031. ACKNOWLEDGMENTS We thank K. Shrestha and S. Zaghlul for their technical assistance; L. Zon and A. von Giese and members of the Camargo lab for critical discussion; and members of the Gregory laboratory for advice on mass spectrometry experiments. This work was supported by grants from the Stand Up to Cancer-AACR initiative, NIH grant R01 CA131426, and funds from the Whitehead Institute Fellows Program. F.D.C. is a V Foundation Scholar and Pew Scholar in the Biomedical Sciences. Received: August 12, 2010 Revised: January 7, 2011 Accepted: February 14, 2011 Published: March 3, 2011 REFERENCES Baumgartner, R., Poernbacher, I., Buser, N., Hafen, E., and Stocker, H. (2010). The WW domain protein Kibra acts upstream of Hippo in Drosophila. Dev. Cell 18, 309–316. Belteki, G., Haigh, J., Kabacs, N., Haigh, K., Sison, K., Costantini, F., Whitsett, J., Quaggin, S.E., and Nagy, A. (2005). Conditional and inducible transgene expression in mice through the combinatorial use of Cre-mediated recombination and tetracycline induction. Nucleic Acids Res. 33, e51. Benjamin, J.M., and Nelson, W.J. (2008). Bench to bedside and back again: molecular mechanisms of alpha-catenin function and roles in tumorigenesis. Semin. Cancer Biol. 18, 53–64. Cai, J., Zhang, N., Zheng, Y., de Wilde, R.F., Maitra, A., and Pan, D. (2010). The Hippo signaling pathway restricts the oncogenic potential of an intestinal regeneration program. Genes Dev. 24, 2383–2388. Camargo, F.D., Gokhale, S., Johnnidis, J.B., Fu, D., Bell, G.W., Jaenisch, R., and Brummelkamp, T.R. (2007). YAP1 increases organ size and expands undifferentiated progenitor cells. Curr. Biol. 17, 2054–2060. Chen, C.L., Gajewski, K.M., Hamaratoglu, F., Bossuyt, W., Sansores-Garcia, L., Tao, C., and Halder, G. (2010). The apical-basal cell polarity determinant Crumbs regulates Hippo signaling in Drosophila. Proc. Natl. Acad. Sci. USA 107, 15810–15815. Chiang, C.W., Kanies, C., Kim, K.W., Fang, W.B., Parkhurst, C., Xie, M., Henry, T., and Yang, E. (2003). Protein phosphatase 2A dephosphorylation of phosphoserine 112 plays the gatekeeper role for BAD-mediated apoptosis. Mol. Cell. Biol. 23, 6350–6362. Das Thakur, M., Feng, Y., Jagannathan, R., Seppa, M.J., Skeath, J.B., and Longmore, G.D. (2010). Ajuba LIM proteins are negative regulators of the Hippo signaling pathway. Curr. Biol. 20, 657–662. Dassule, H.R., Lewis, P., Bei, M., Maas, R., and McMahon, A.P. (2000). Sonic hedgehog regulates growth and morphogenesis of the tooth. Development 127, 4775–4785. Dellambra, E., Golisano, O., Bondanza, S., Siviero, E., Lacal, P., Molinari, M., D’Atri, S., and De Luca, M. (2000). Downregulation of 14-3-3sigma prevents
Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 793
clonal evolution and leads to immortalization of primary human keratinocytes. J. Cell Biol. 149, 1117–1130.
suppressor that regulates Hippo signaling by binding to Expanded. Proc. Natl. Acad. Sci. USA 107, 10532–10537.
Depaepe, V., Suarez-Gonzalez, N., Dufour, A., Passante, L., Gorski, J.A., Jones, K.R., Ledent, C., and Vanderhaeghen, P. (2005). Ephrin signalling controls brain size by regulating apoptosis of neural progenitors. Nature 435, 1244–1250.
Lu, L., Li, Y., Kim, S.M., Bossuyt, W., Liu, P., Qiu, Q., Wang, Y., Halder, G., Finegold, M.J., Lee, J.S., and Johnson, R.L. (2010). Hippo signaling is a potent in vivo growth and tumor suppressor pathway in the mammalian liver. Proc. Natl. Acad. Sci. USA 107, 1437–1442.
Dong, J., Feldmann, G., Huang, J., Wu, S., Zhang, N., Comerford, S.A., Gayyed, M.F., Anders, R.A., Maitra, A., and Pan, D. (2007). Elucidation of a universal size-control mechanism in Drosophila and mammals. Cell 130, 1120–1133.
Martin, M., Potente, M., Janssens, V., Vertommen, D., Twizere, J.C., Rider, M.H., Goris, J., Dimmeler, S., Kettmann, R., and Dequiedt, F. (2008). Protein phosphatase 2A controls the activity of histone deacetylase 7 during T cell apoptosis and angiogenesis. Proc. Natl. Acad. Sci. USA 105, 4727–4732.
Fuchs, E. (2007). Scratching the surface of skin development. Nature 445, 834–842.
Menon, G.K., and Elias, P.M. (1991). Ultrastructural localization of calcium in psoriatic and normal human epidermis. Arch. Dermatol. 127, 57–63.
Genevet, A., Wehr, M.C., Brain, R., Thompson, B.J., and Tapon, N. (2010). Kibra is a regulator of the Salvador/Warts/Hippo signaling network. Dev. Cell 18, 300–308.
Nguyen, H., Merrill, B.J., Polak, L., Nikolova, M., Rendl, M., Shaver, T.M., Pasolli, H.A., and Fuchs, E. (2009). Tcf3 and Tcf4 are essential for long-term homeostasis of skin epithelia. Nat. Genet. 41, 1068–1075.
Gladden, A.B., Hebert, A.M., Schneeberger, E.E., and McClatchey, A.I. (2010). The NF2 tumor suppressor, Merlin, regulates epidermal development through the establishment of a junctional polarity complex. Dev. Cell 19, 727–739.
Oh, H., and Irvine, K.D. (2008). In vivo regulation of Yorkie phosphorylation and localization. Development 135, 1081–1088.
Grzeschik, N.A., Parsons, L.M., Allott, M.L., Harvey, K.F., and Richardson, H.E. (2010). Lgl, aPKC, and Crumbs regulate the Salvador/Warts/Hippo pathway through two distinct mechanisms. Curr. Biol. 20, 573–581. Hamaratoglu, F., Willecke, M., Kango-Singh, M., Nolo, R., Hyun, E., Tao, C., Jafar-Nejad, H., and Halder, G. (2006). The tumour-suppressor genes NF2/ Merlin and Expanded act through Hippo signalling to regulate cell proliferation and apoptosis. Nat. Cell Biol. 8, 27–36. Hariharan, I.K. (2006). Growth regulation: a beginning for the hippo pathway. Curr. Biol. 16, R1037–R1039. Harvey, K., and Tapon, N. (2007). The Salvador-Warts-Hippo pathway - an emerging tumour-suppressor network. Nat. Rev. Cancer 7, 182–191. Harvey, K.F., Pfleger, C.M., and Hariharan, I.K. (2003). The Drosophila Mst ortholog, hippo, restricts growth and cell proliferation and promotes apoptosis. Cell 114, 457–467. Herron, B.J., Liddell, R.A., Parker, A., Grant, S., Kinne, J., Fisher, J.K., and Siracusa, L.D. (2005). A mutation in stratifin is responsible for the repeated epilation (Er) phenotype in mice. Nat. Genet. 37, 1210–1212.
Oka, T., Mazack, V., and Sudol, M. (2008). Mst2 and Lats kinases regulate apoptotic function of Yes kinase-associated protein (YAP). J. Biol. Chem. 283, 27534–27546. Pantalacci, S., Tapon, N., and Le´opold, P. (2003). The Salvador partner Hippo promotes apoptosis and cell-cycle exit in Drosophila. Nat. Cell Biol. 5, 921–927. Pellegrini, G., Dellambra, E., Golisano, O., Martinelli, E., Fantozzi, I., Bondanza, S., Ponzin, D., McKeon, F., and De Luca, M. (2001). p63 identifies keratinocyte stem cells. Proc. Natl. Acad. Sci. USA 98, 3156–3161. Perez-Moreno, M., Davis, M.A., Wong, E., Pasolli, H.A., Reynolds, A.B., and Fuchs, E. (2006). p120-catenin mediates inflammatory responses in the skin. Cell 124, 631–644. Ribeiro, P.S., Josue´, F., Wepf, A., Wehr, M.C., Rinner, O., Kelly, G., Tapon, N., and Gstaiger, M. (2010). Combined functional genomic and proteomic approaches identify a PP2A complex as a negative regulator of Hippo signaling. Mol. Cell 39, 521–534. Robinson, B.S., Huang, J., Hong, Y., and Moberg, K.H. (2010). Crumbs regulates Salvador/Warts/Hippo signaling in Drosophila via the FERM-domain protein Expanded. Curr. Biol. 20, 582–590.
Huang, J., Wu, S., Barrera, J., Matthews, K., and Pan, D. (2005). The Hippo signaling pathway coordinately regulates cell proliferation and apoptosis by inactivating Yorkie, the Drosophila Homolog of YAP. Cell 122, 421–434.
Senoo, M., Pinto, F., Crum, C.P., and McKeon, F. (2007). p63 Is essential for the proliferative potential of stem cells in stratified epithelia. Cell 129, 523–536.
Kim, K.A., Kakitani, M., Zhao, J., Oshima, T., Tang, T., Binnerts, M., Liu, Y., Boyle, B., Park, E., Emtage, P., et al. (2005). Mitogenic influence of human R-spondin1 on the intestinal epithelium. Science 309, 1256–1259.
Song, H., Mak, K.K., Topol, L., Yun, K., Hu, J., Garrett, L., Chen, Y., Park, O., Chang, J., Simpson, R.M., et al. (2010). Mammalian Mst1 and Mst2 kinases play essential roles in organ size control and tumor suppression. Proc. Natl. Acad. Sci. USA 107, 1431–1436.
Kobielak, A., and Fuchs, E. (2006). Links between alpha-catenin, NF-kappaB, and squamous cell carcinoma in skin. Proc. Natl. Acad. Sci. USA 103, 2322–2327. Lai, Z.C., Wei, X., Shimizu, T., Ramos, E., Rohrbaugh, M., Nikolaidis, N., Ho, L.L., and Li, Y. (2005). Control of cell proliferation and apoptosis by mob as tumor suppressor, mats. Cell 120, 675–685. Li, Q., Lu, Q., Estepa, G., and Verma, I.M. (2005). Identification of 14-3-3sigma mutation causing cutaneous abnormality in repeated-epilation mutant mouse. Proc. Natl. Acad. Sci. USA 102, 15977–15982. Lian, I., Kim, J., Okazawa, H., Zhao, J., Zhao, B., Yu, J., Chinnaiyan, A., Israel, M.A., Goldstein, L.S., Abujarour, R., et al. (2010). The role of YAP transcription coactivator in regulating stem cell self-renewal and differentiation. Genes Dev. 24, 1106–1118.
Stanger, B.Z., Tanaka, A.J., and Melton, D.A. (2007). Organ size is limited by the number of embryonic progenitor cells in the pancreas but not the liver. Nature 445, 886–891. Stepniak, E., Radice, G.L., and Vasioukhin, V. (2009). Adhesive and signaling functions of cadherins and catenins in vertebrate development. Cold Spring Harb. Perspect. Biol. 1, a002949. Tinkle, C.L., Pasolli, H.A., Stokes, N., and Fuchs, E. (2008). New insights into cadherin function in epidermal sheet formation and maintenance of tissue integrity. Proc. Natl. Acad. Sci. USA 105, 15405–15410. Varelas, X., Samavarchi-Tehrani, P., Narimatsu, M., Weiss, A., Cockburn, K., Larsen, B.G., Rossant, J., and Wrana, J.L. (2010). The Crumbs complex couples cell density sensing to Hippo-dependent control of the TGF-b-SMAD pathway. Dev. Cell 19, 831–844.
Lien, W.H., Klezovitch, O., Fernandez, T.E., Delrow, J., and Vasioukhin, V. (2006a). alphaE-catenin controls cerebral cortical size by regulating the hedgehog signaling pathway. Science 311, 1609–1612.
Vasioukhin, V., Bauer, C., Degenstein, L., Wise, B., and Fuchs, E. (2001). Hyperproliferation and defects in epithelial polarity upon conditional ablation of alpha-catenin in skin. Cell 104, 605–617.
Lien, W.H., Klezovitch, O., and Vasioukhin, V. (2006b). Cadherin-catenin proteins in vertebrate development. Curr. Opin. Cell Biol. 18, 499–506.
Vassilev, A., Kaneko, K.J., Shu, H., Zhao, Y., and DePamphilis, M.L. (2001). TEAD/TEF transcription factors utilize the activation domain of YAP65, a Src/Yes-associated protein localized in the cytoplasm. Genes Dev. 15, 1229–1241.
Ling, C., Zheng, Y., Yin, F., Yu, J., Huang, J., Hong, Y., Wu, S., and Pan, D. (2010). The apical transmembrane protein Crumbs functions as a tumor
794 Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc.
Wu, S., Huang, J., Dong, J., and Pan, D. (2003). hippo encodes a Ste-20 family protein kinase that restricts cell proliferation and promotes apoptosis in conjunction with salvador and warts. Cell 114, 445–456. Wu, S., Liu, Y., Zheng, Y., Dong, J., and Pan, D. (2008). The TEAD/TEF family protein Scalloped mediates transcriptional output of the Hippo growth-regulatory pathway. Dev. Cell 14, 388–398. Xu, T., Wang, W., Zhang, S., Stewart, R.A., and Yu, W. (1995). Identifying tumor suppressors in genetic mosaics: the Drosophila lats gene encodes a putative protein kinase. Development 121, 1053–1063. Zhang, H., Pasolli, H.A., and Fuchs, E. (2011). Yes-associated protein (YAP) transcriptional coactivator functions in balancing growth and differentiation in skin. Proc. Natl. Acad. Sci. USA 108, 2270–2275. Zhang, L., Ren, F., Zhang, Q., Chen, Y., Wang, B., and Jiang, J. (2008). The TEAD/TEF family of transcription factor Scalloped mediates Hippo signaling in organ size control. Dev. Cell 14, 377–387.
Zhang, N., Bai, H., David, K.K., Dong, J., Zheng, Y., Cai, J., Giovannini, M., Liu, P., Anders, R.A., and Pan, D. (2010). The Merlin/NF2 tumor suppressor functions through the YAP oncoprotein to regulate tissue homeostasis in mammals. Dev. Cell 19, 27–38. Zhao, B., Wei, X., Li, W., Udan, R.S., Yang, Q., Kim, J., Xie, J., Ikenoue, T., Yu, J., Li, L., et al. (2007). Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control. Genes Dev. 21, 2747–2761. Zhao, B., Ye, X., Yu, J., Li, L., Li, W., Li, S., Yu, J., Lin, J.D., Wang, C.Y., Chinnaiyan, A.M., et al. (2008). TEAD mediates YAP-dependent gene induction and growth control. Genes Dev. 22, 1962–1971. Zhou, D., Conrad, C., Xia, F., Park, J.S., Payer, B., Yin, Y., Lauwers, G.Y., Thasler, W., Lee, J.T., Avruch, J., and Bardeesy, N. (2009). Mst1 and Mst2 maintain hepatocyte quiescence and suppress hepatocellular carcinoma development through inactivation of the Yap1 oncogene. Cancer Cell 16, 425–438.
Cell 144, 782–795, March 4, 2011 ª2011 Elsevier Inc. 795
Endocrine Regulation of Male Fertility by the Skeleton Franck Oury,1,6 Grzegorz Sumara,1,6 Olga Sumara,1 Mathieu Ferron,1 Haixin Chang,3 Charles E. Smith,4 Louis Hermo,4 Susan Suarez,3 Bryan L. Roth,5 Patricia Ducy,2 and Gerard Karsenty1,* 1Department
of Genetics and Development of Pathology and Cell Biology Columbia University, New York, NY 10032, USA 3Department of Biomedical Sciences, Cornell University, T5-006 Veterinary Research Tower, Ithaca, NY 14853, USA 4Department of Anatomy and Cell Biology, McGill University, 3640 University Street, Montreal, Quebec H3A 2B2, Canada 5Department of Pharmacology, Medicinal Chemistry and Natural Products and National Institute of Mental Health Psychoactive Drug Screening Program, Schools of Medicine and Pharmacy, University of North Carolina, Chapel Hill, NC 27599, USA 6These authors contributed equally to this work *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.004 2Department
SUMMARY
Interactions between bone and the reproductive system have until now been thought to be limited to the regulation of bone remodeling by the gonads. We now show that, in males, bone acts as a regulator of fertility. Using coculture assays, we demonstrate that osteoblasts are able to induce testosterone production by the testes, though they fail to influence estrogen production by the ovaries. Analyses of cellspecific loss- and gain-of-function models reveal that the osteoblast-derived hormone osteocalcin performs this endocrine function. By binding to a G protein-coupled receptor expressed in the Leydig cells of the testes, osteocalcin regulates in a CREBdependent manner the expression of enzymes that is required for testosterone synthesis, promoting germ cell survival. This study expands the physiological repertoire of osteocalcin and provides the first evidence that the skeleton is an endocrine regulator of reproduction. INTRODUCTION Bone is a dynamic tissue undergoing modeling during childhood and remodeling throughout adulthood (Harada and Rodan, 2003; Rodan and Martin, 2000). These two processes, referred to hereafter as bone (re)modeling, are characterized by the succession of resorption of mineralized bone by osteoclasts and de novo formation by osteoblasts. Bone (re) modeling is regulated locally by cytokines produced by bone cells and systemically by hormones and neuropeptides (Harada and Rodan, 2003; Karsenty et al., 2009). One of the most powerful hormonal regulations of bone (re)modeling is exerted by the sex steroid hormones that are necessary to maintain bone integrity (Khosla et al., 2001; Nakamura et al., 2007; Riggs 796 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
et al., 1998). The biological importance of this regulation is best exemplified by the fact that gonadal failure triggers bone loss and causes osteoporosis in postmenopausal women (Manolagas et al., 2002; Rodan and Martin, 2000). To date, the study of the interplay between gonads and bone has focused on the mechanism whereby sex steroid hormones affect bone mass accrual (Manolagas et al., 2002; Nakamura et al., 2007). Based on physiological and clinical observations, we hypothesized 10 years ago that bone mass, energy metabolism, and reproduction might be coordinately regulated (Ducy et al., 2000). Testing this hypothesis revealed that bone is an endocrine organ favoring whole-body glucose homeostasis and energy expenditure. These functions of bone are mediated by an osteoblast-specific secreted molecule, osteocalcin, that, when uncarboxylated, acts as a hormone favoring b cell proliferation, insulin secretion, and sensitivity and energy expenditure (Lee et al., 2007). A second gene expressed in osteoblasts, Esp, inhibits endocrine functions of osteocalcin by favoring, through an indirect mechanism, its carboxylation (Ferron et al., 2010; Fulzele et al., 2010). Despite these findings, basic facts about how osteocalcin performs its endocrine function are unknown. Most importantly, the receptor for osteocalcin remains to be determined, as do the signaling pathways triggered by this hormone in target cells. We now show that osteocalcin, in addition to its endocrine role as a regulator of energy homeostasis, favors male fertility. It does so by promoting synthesis by Leydig cells of testosterone, a steroid hormone that is required for many aspects of testicular function (Sinha Hikim and Swerdloff, 1999; Walker, 2009) and has no effect on female fertility. Furthermore, we identify a bona fide receptor for osteocalcin that is expressed and transduces its signal in Leydig cells. Using this tool, we identify genes whose expression is regulated by osteocalcin in these cells and that account for its regulation of male fertility. Our findings expand the biological importance of osteocalcin, begin to unravel its molecular mode of action, and provide the first evidence that the skeleton is an endocrine regulator of fertility.
RESULTS Osteoblasts Enhance Testosterone Production by Leydig Cells In an effort to determine whether osteoblasts or any other cells of mesenchymal origin may regulate the functions of gonads, we asked whether supernatants of mesenchymal cell cultures affect hormone production by testes and/or ovaries (Figure 1A). In a first set of exploratory experiments, we observed that, of all those tested, the supernatant of osteoblast cultures increased testosterone secretion by testes explants to the largest extent (>3-fold) but did not affect estradiol and progesterone secretion by testes or ovaries (Figures 1B–1G). In subsequent experiments, testes explants were replaced by defined cell populations. Because testosterone is produced by Leydig cells, we asked whether osteoblast-derived molecule(s) act(s) directly on Leydig cells by culturing mouse primary Leydig cells in the presence or absence of supernatants of osteoblast cultures or of other mesenchymal cell type cultures. In the conditions of this assay, supernatants of osteoblast cultures were also the only ones that were able to significantly increase testosterone production by Leydig cells (>4-fold) (Figure 1H). These experiments indicate that osteoblasts are the cells of mesenchymal origin that affect testosterone biosynthesis to the largest extent and that they do so through secreted molecule(s) acting on Leydig cells of the testis. This endocrine function of osteoblasts was restricted to androgen production. Osteocalcin Favors Male Fertility by Enhancing Testosterone Production Osteocalcin is a major osteoblast-derived hormone. We had previously noticed that Osteocalcin-deficient male mice (Ocn/) breed poorly (P.D. and G.K., unpublished data), so we tested whether it could be a, or the, osteoblast-derived hormone enhancing testosterone secretion by Leydig cells. Several lines of evidence indicated that that is the case. First, supernatants of wild-type (WT), but not of Ocn/, osteoblast cultures increased testosterone production by Leydig cells (Figure 2A). Second, treating Leydig cells with an increasing amount of uncarboxylated osteocalcin, the active form of the hormone, resulted in a dose-dependent increase in testosterone secretion, although at high concentration, the stimulatory effect of osteocalcin weakened (Figure 2B). Third, injection of osteocalcin in WT mice increased circulating levels of testosterone (Figure 2C). Fourth, we analyzed loss- (Ocn/ mice) and gain-of-function (Esp/ mice) mouse models for osteocalcin (Lee et al., 2007). When Ocn/ males were crossed with WT female mice, the size of the litters was nearly 2-fold smaller than when WT males were crossed with WT females (Figure 2D). Conversely, the number of pups per litter was increased when Esp/ males were bred with WT female mice, although this increase did not reach statistical significance (Figure 2D). The frequency of litters for a period of 8 weeks was also decreased in the case of the loss-of-function model and increased in the gain-of-function model (Figure 2E). Testes size and weight were significantly decreased in Ocn/ and increased in Esp/ mice at 3 months of age. In some of the latter mutant mice, this was caused, in part, by fluid accumulation (Figures 2F and 2G). The weights of
epididymides and seminal vesicles and sperm count were also significantly decreased in Ocn/ and increased in Esp/ mice (Figures 2H–2J). These abnormalities worsened over time (Figures 2G and 2J). Motility of sperm from both WT and Ocn/ males was assessed by videomicroscopy immediately after dissemination from the caudal epididymis or after 2 hr of incubation under conditions known to prepare sperm for fertilization (Sua´rez and Osman, 1987). In both cases, the percentage of motile sperm did not differ between Ocn/ and WT mice (Figure S1A available online). Likewise, the percentage of abnormally shaped or dead sperm was similar in WT and Ocn/ mice (Figures S1B and S1C). Consistent with the fact that osteocalcin favors testosterone synthesis in Leydig cells ex vivo, circulating levels of testosterone were markedly decreased in Ocn/ and increased in Esp/ mice at all time points tested (Figure 2K). Accordingly, circulating levels of luteinizing hormone (LH), a pituitary-derived hormone favoring testosterone synthesis, was increased 2.5-fold in Ocn/ mice (Figure 2L). Taken together, these cell biological and genetic experiments identify osteocalcin as a secreted molecule favoring male fertility by increasing testosterone production by Leydig cells. Circulating progesterone levels were similar in Ocn/ and WT mice, and although circulating levels of estradiol were higher in Ocn/ than in WT mice, they remained within the normal range (from 9.3 to 28.9 ng/ml for nonbreeder mice and from 14.4 to 71.1 ng/ml for breeder mice) (Figure 2K). Estradiol levels were not affected in Esp/ mice. As predicted by the coculture assays, female fertility, ovary weight, morphology of the uterus, follicles number, and circulating levels of sex steroid hormones were normal in Ocn/ female mice (Figures S1D–S1L). Osteocalcin Regulates Male Fertility as an Osteoblast-Derived Hormone To verify that osteocalcin regulates male fertility as an osteoblast-secreted hormone and not as a testis-secreted factor, we performed gene expression and cell-specific gene deletion experiments. When assessing Osteocalcin expression by quantitative PCR (qPCR), we observed that it was more than 750-fold higher in bone than in gonads; accordingly, we failed to detect Osteocalcin transcript or protein in testes by in situ hybridization or western blot analyses (Figures 3A–3C). To be able to trace Osteocalcin-expressing cells in vivo, we knocked the mCherry fluorescent reporter gene into the Ocn locus (Ocn-mCherry mice) (Figures S2A and S2B). Though we observed the expected strong signal in osteoblasts, there was no detectable mCherry fluorescence in testes (Figure 3D). Thus, in multiple assays, we failed to detect Osteocalcin expression in testes. Next, we generated cell-specific loss- and gain-of-function models of osteocalcin by crossing mice harboring floxed alleles of Ocn (Figures S2C and S2D) or Esp with either the a1(I) Collagen-Cre transgenic mice or the Cyp17-iCre transgenic mice to delete genes in osteoblasts or in Leydig cells only, respectively (Bridges et al., 2008; Dacquin et al., 2002). Testes size and weight, epididymides and seminal vesicles weights, sperm count, and circulating testosterone levels were all Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 797
A
Measure sex steroid hormones after 1hour
Collect supernatants after 24 hours
Testis explants or Ovary explants or Primary Leydig cells
Primary: . Osteoblasts . Myoblasts . Fibroblasts . Adipocytes
C
0
40.0 30.0 20.0 10.0 0
tes ocy
um
Ad ip
Ovary explants 15.0
10.0
5.0
s
tes ocy
ast
um edi Con tro lm
Con t
rol m
edi
um
0
Ost eob las ts My obl ast s Ad ipo cyt es
0
edi
Progesterone (ng/ml)
Estradiol (pg/ml)
0.05
Con tro lm
s Fib rob las ts My obl ast s
um
las t
Ad ipo cyt es
G
Ovary explants
Con tro lm edi um Ost eob las ts My obl ast s Ad ipo cyt es
Testosterone (ng/ml)
eob
Con tro lm
F 50.0
0.10
2.0
Ad ip
Ovary explants
0.15
4.0
0
edi
ast s My obl ast s Ad ipo cyt es
Fib rob l
Con tro lm edi um Ost eob las ts
0
E
2.0
6.0
My obl
1.0
4.0
7.0
las ts Fib rob las ts My obl ast s
2.0
6.0
sts
3.0
8.0
Testis explants 8.0
Ost eob la
5.0 4.0
Progesterone (ng/ml)
10.0
eob
*
D
Testis explants
Ost
Testis explants
Estradiol (pg/ml)
Testosterone (ng/ml)
6.0
Ost
B
Leydig cells 1.50
*
1.25 1.00 0.75 0.50 0.25
es
s
ocy t Ad ip
s ast
My obl ast
s ast
Fib rob l
Con tro lm
Ost eob l
um
0
edi
Testosterone (ng/ml/103cells)
H
Figure 1. Osteoblasts Enhance Testosterone Biosynthesis by Leydig Cells (A) Schematic representation of the cell-based assay used to determine the role of various mesenchymal cells in sex steroid hormone production. Various primary mesenchymal cells from mice were cultured in Leydig cell medium, and supernatants were collected after 24 hr. Then, testis or ovary explants or primary Leydig cells were cultured for 1 hr with these supernatants, and radioimmunoassays (RIAs) were performed to measure levels of testosterone, estradiol, or progesterone. (B–D) Testis explants cultured in the presence of supernatants of different mesenchymal cell cultures: RIA measurement of (B) testosterone, (C) estradiol, and (D) progesterone levels.
798 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
2.5
1h
10
g
ng
g
3n
1n
0n
g
ng
edi um Ost eob las ts -Os teo bla sts
Testis Weight (g)
* 0.10
**
*
*
0.05
3 Months
6 Weeks
K
WT (n>10) Ocn-/- (n>10) WT (n>8) Esp-/- (n>8)
WT
Sperm Count (107)
**
Ocn-/-
(Non-Breeders)
*
(Non-Breeders)
4.0
**
**
WT (Breeders)
Esp-/-
2.0
(Breeders)
(n=9)
WT
(n=6)
(Non-Breeders)
0
Esp-/6 Weeks
3 Months
6 Months
(Non-Breeders)
(n=5)
40
*
30 20 10 0
2.70 + - 0.53* 28.09 +- 2.89* 2.07 + - 0.60
12.28 + - 1.59
28.90 + - 3.27
0.49 + - 0.12* 61.60 +- 6.57*
(n=7)
(n=10)
*
60 50
Estradiol Progesterone Testosterone (ρg/mL) (ng/mL) (ng/mL) 13.60 + - 0.72 12.24 +- 0.82 13.14 +- 1.27
(n=10)
Ocn-/-
1.60 1.20 0.80
*
0.40
I
WT (n=10) Ocn-/- (n=10) WT (n=6) Esp-/- (n=6) 500
* 400
*
300 200 100 0
6 Months
(n=9)
(Breeders)
WT
*
6.0
*
(n=10)
(Breeders)
8.0
2.00
0
WT (n=12) Ocn-/- (n=5) WT (n=6) Esp-/- (n=5)
* *
0
J
4.0
0
H
Esp-/-
*
2.0
WT (n>10) Ocn-/- (n>10) WT (n>8) Esp-/- (n>8) 0.15
Ocn-/-
6.0
0
G
Litter Frequency
*
0
Testis Size (3 months) WT
5.0
Seminal Vesicle Weight (mg)
0.25
*
Litter Size
*
*
Ocn -/
Con tro lM
F
0.50
8.0
7.5
WT (n=8) Ocn-/- (n=8) WT (n=8) Esp-/- (n=8)
2.40
10.0
13.34 + - 1.12
21.24 + - 2.27
8.78 + - 1.25
23.50 + - 1.77 * 18.54 +- 1.81
10.32 + - 0.70
2.03 + - 1.35
16.02 + - 1.64
13.46 + - 1.19 * 19.80 +- 1.82
L
WT (n=8) Ocn-/- (n=8) 12.0
Circulating LH level (ng/ml)
0
0.75
*
10.0
E
WT (n=8) Ocn-/- (n=8) WT (n=8) Esp-/- (n=8)
8h
0.25
**
Epididymis Weight (mg)
0.50
*** 1.00
4h
0.75
D 12.5
Testosterone (ng/ml)
* 1.00
C
Leydig cells
1.25
0.3
Testosterone (ng/ml)
B
Leydig cells
1.25
Testosterone (ng/ml)
A
* 10.0 8.0 6.0 4.0 2.0 0
Figure 2. Osteocalcin Favors Male Fertility by Increasing Testosterone Production by Leydig Cells (A) Testosterone production by primary Leydig cells cultured in the presence of supernatants of wild-type (WT) or Ocn/ osteoblast cultures. (B) Testosterone production by primary Leydig cells following stimulation with increasing doses of osteocalcin. (C) Circulating testosterone levels in WT mice 1 hr, 4 hr, and 8 hr after vehicle or osteocalcin injection (3 ng/g of body weight). (D and E) Comparison between the average litter size (D) and frequency (E) generated by WT, Ocn/, or Esp/ male littermate mice crossed with WT females (breeding was tested from 8 to 16 weeks of age). (F–J) Testis size (F), testis weight (G), epididymis weight (H), seminal vesicle weights (I), and sperm counts (J) in Ocn/ and Esp/ compared to WT nonbreeder littermate mice. (K) Circulating sex steroid levels in Ocn/ and Esp/ compared to WT littermate mice. The analyses were performed on breeder and nonbreeder mice. (L) Circulating LH levels in Ocn/ compared to WT nonbreeder littermate mice. Error bars represent SEM. Student’s t test; *p < 0.05; **p < 0.001. See also Figure S1.
reduced in 12-week-old Ocnosb/ mice. None of these parameters were affected in mice lacking Osteocalcin in Leydig cells only (Figures 3E–3I). There was a tight correlation between circulating levels of osteocalcin and testosterone in Ocnosb/ mice
(Figure 3J). Conversely, Esposb/ mice displayed testicular abnormalities identical to those of Esp/ mice. Inactivation of Esp in Sertoli cells, where this gene is expressed (Dacquin et al., 2004; Jamin et al., 2003), had no detectable deleterious
(E–G) Ovary explants cultured in the presence of supernatants of different mesenchymal cells cultures: RIA measurement of (E) testosterone, (F) estradiol, and (G) progesterone levels. (H) Testosterone production by primary Leydig cells cultured in the presence of supernatants of different mesenchymal cell cultures. Error bars represent SEM. Student’s t test; *p < 0.05.
Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 799
B
C Femur Calvaria 15
Osteocalcin 10
500 400
Osteocalcin mCherry β-Actin
50 37
0
* 0.10
0.05
Epididymis Weight (mg)
Sperm Count (107)
4.0
**
3.0 2.0 1.0
40
WT (n=9) Esposb-/- (n=6) WT (n=8) EspSertoli-/- (n=6)
0.45 0.15
0.40 0.35 R2 = 0.65 0.30 0.25
30 20 10
10.0
*
0.10
0.05
8.0
0
25
*
6.0 4.0 2.0
0
0.20
WT (n=9) Esposb-/- (n=6) WT (n=8) EspSertoli-/- (n=6)
L
Sperm Count (107)
K
0.50
*
0
0
0
Testis Weight (g)
Testis Weight (g)
50
5.0
H
WT (n=22) Ocnosb-/- (n=12)
0
I
WT (n=22) Ocnosb-/- (n=12)
WT (n=7) OcnLeydig-/- (n=7)
WT (n=7) OcnLeydig-/- (n=7)
6.0
0.15
Testosterone (ng/ml)
WT (n=22) Ocnosb-/- (n=12)
WT (n=7) OcnLeydig-/- (n=7)
WT (n=7) OcnLeydig-/- (n=7)
J
G
WT (n=6) OcnLeydig-/- (n=7)
400 150 300
Testosterone (%)
WT (n=22) Ocnosb-/- (n=12)
*
200 100 0
M 500 400 300 200 100 0
125 100 75 50
*
25 0
WT (n=9) Esposb-/- (n=6) WT (n=8) EspSertoli-/- (n=6)
N 250
*
WT (n=6) Esposb-/- (n=5) WT (n=7) EspSertoli-/- (n=7)
* Testosterone (%)
F
Seminal Vesicle Weight (mg)
WT (n=22) Ocnosb-/- (n=12)
mCherry
Osteocalcin
Seminal Vesicle Weight (mg)
300 200 100
E
Testis
Bone Bone
765
Ocn-mCherry +/-
D
Testis
Testis
1000 900 800 700 600
Bo ne Tes tis Ov ary
Osteocalcin Expression in WT (/ActinX103)
A
200 150 100 50 0
50 75 100 125 150 Osteocalcin (ng/ml)
Figure 3. Osteocalcin Promotes Male Fertility through Its Expression in Osteoblasts (A) qPCR analysis of Osteocalcin expression in bone, testes, and ovaries of 3-month-old nonbreeder WT mice. (B) Western blot analysis of osteocalcin in femur, calvaria, and testis. (C) In situ hybridization analysis of Osteocalcin expression in bone and testis of 3-month-old WT mice. Arrows show the Osteocalcin expression in bone. (D) Analysis of mCherry fluorescent protein in bone and testis of Osteocalcin-mCherry knockin mice. Arrows show the presence of mCherry fluorescent protein reflecting Osteocalcin. (E–H) Fertility in mice lacking Ocn specifically in osteoblasts (Ocnosb/) or Leydig cells (OcnLeydig/) compared to WT nonbreeder littermates: (E) testes weight, (F) sperm count, (G) epididymis, and (H) seminal vesicle weights. (I) Ratio of circulating testosterone levels measured in WT and Ocnosb/ or in WT and OcnLeydig/ nonbreeder littermate mice. (J) Linear regression representation of circulating testosterone levels versus circulating osteocalcin levels in Ocnosb/ (n = 11) nonbreeder mice. Each dot represents one Ocnosb/ mouse. In WT littermate mice, the levels of osteocalcin varied from 106 to 177 ng/ml (on average, 133 ng/ml). For Ocnosb/, the average osteocalcin level was 68.4 ng/ml. (K–M) Fertility in mice lacking Esp specifically in osteoblasts (Esposb/) or Leydig cells (EspLeydig/) compared to WT nonbreeder littermates: (K) testis weight, (L) sperm count, and (M) seminal vesicle weight. (N) Ratio of circulating testosterone levels measured in WT and Esposb/ or in WT and EspLeydig/ nonbreeder littermate mice. Error bars represent SEM. Student’s t test; *p < 0.05; **p < 0.001. See also Figure S2.
consequence on testis biology, demonstrating that it is through its expression in osteoblasts, not in Sertoli cells, that Esp regulates male fertility (Figures 3K–3N). These experiments therefore indicate that it is only through its expression in osteoblasts that osteocalcin promotes male fertility. Cellular and Molecular Bases of Osteocalcin Regulation of Male Fertility To address this aspect of osteocalcin biology, we first studied the morphology of Leydig cells by immunostaining of 3-b-hydroxyste800 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
roid dehydrogenase (3b-HSD). The number of Leydig cells was not significantly affected by the absence of osteocalcin or Esp, nor was the expression of genes affecting cell proliferation (Figure 4A and data not shown). However, Leydig cells appeared hypotrophic in Ocn/ testes, as determined by the significant decrease of the ratio between the Leydig cells and interstitial areas observed in Ocn/ compared to WT testes (Figures 4B and 4C). Conversely, this ratio was increased in Esp/ testes (Figures 4B and 4C). When germ cells were analyzed through a stereological approach, we observed that the number of spermatocytes and
300 250 200 150 100 50
WT
* 50 40
*
30
Esp-/-
Leydig Cells Area vs Interstitial Area (%)
20 10 0
0
D
E
Apoptotic Nuclei per Tubule
Number of cells (x105) per testis
WT (n=5)
Ocn-/- (n=5)
Sertoli cells
Spermato -gonia
Spermato -cytes
Round spermatids
3.69 + - 0.27
7.98 + - 1.07
31.1 + - 0.68
44.8 + - 2.21
3.36 + - 0.21
7.39 + - 0.79
+ 24.3 + - 1.72** 30.3 - 1.50**
F
WT (n=10) Ocn-/- (n=10) WT (n=4) Esp -/- (n=4) 1.0
*
0.8 0.6 0.4
*
0.2
Control (n=3) osteocalcin stimulation (n=3)
**
Expression Relative to Control (%)
Leydig Cells Number per Leydig Cells Area (nb/mm2)
60
350
C
WT (n=10) Ocn-/- (n=10) WT (n=5) Esp-/- (n=5)
WT
B
WT (n=10) Ocn-/- (n=10) WT (n=5) Esp-/- (n=5)
Ocn-/-
A
0
350
**
300
**
200 150 100 50 0 StAR
H 150
100
* 50
*
* *
0 StAR
Cyp17 Cyp11a 3β-HSD Cyp19 HSD-17
300
WT (n=9) Esp-/- (n=7)
250 200
Expression Relative to WT (%)
WT (n=8) Ocn-/- (n=5)
I Expression Relative to WT (%)
Expression Relative to WT (%)
G
* *
*
*
150 100 50 0 StAR
**
250
300
Cyp17 Cyp11a 3β−HSD Cyp19 HSD-17
WT (n=12) Ocn-/- (n=10) Esp-/- (n=8) Vehicle (n=5) Osteocalcin (n=5)
J
Ocn-/-
WT
Ocn-/-
*
250
*
200
β-Actin
150 100
WT
Cleaved Caspase 3
tACE
*
50
β-Actin
0
Cyp17 Cyp11a 3β−HSD Cyp19 HSD-17
Grth
Figure 4. Cellular and Molecular Events Triggered by Osteocalcin in Leydig Cells (A–C) Histological analyses of Leydig cells in Ocn/ and Esp/ nonbreeder mice. (A) Absolute number of Leydig cells per testis was quantified by the number of 3b-HSD-positive cells. (B) Ratio between Leydig cells (immunopositive for 3b-HSD) versus testis interstitial areas in WT, Ocn/, and Esp/ nonbreeder mice. (C) 3b-HSD immunohistochemistry staining of WT, Ocn/, and Esp/ testes. (D) Quantification of the different testicular cell types in WT and Ocn/ nonbreeder mice. (E) Germ cell apoptosis analysis by TUNEL assay in WT, Ocn/, and Esp/ nonbreeder testes. (F–H) qPCR analysis of the expression of steroidogenic acute regulatory protein (StAR), cholesterol side-chain cleavage enzyme (Cyp11a), cytochrome P-450 17 a (Cyp17), 3-b-hydroxysteroid dehydrogenase (3b-HSD), aromatase enzyme (Cyp19), and 17-b-hydroxysteroid dehydrogenase (HSD-17) in primary Leydig cells treated with 3 ng/ml of osteocalcin (F), in Ocn/ compared to WT nonbreeder littermate testes (G), and in Esp/ compared to WT nonbreeder littermate testes (H). (I) qPCR analysis of Grth/Ddx25 expression in WT, Ocn/, Esp/, and WT nonbreeder mice treated with vehicle or osteocalcin (3 ng/g of body weight). (J) Western blot analysis of cleaved caspase 3 and tACE in WT and Ocn/ nonbreeder testes. Error bars represent SEM. Student’s t test; *p < 0.05; **p < 0.001. See also Figure S3.
round spermatids was significantly decreased in Ocn/ mice (Figure 4D). Consistently, the size of the epithelium in testis tubules was also significantly decreased in these mutant mice, a feature suggesting that osteocalcin regulates, presumably through its effect on testosterone biosynthesis, germ cell number (Figure S3B). Because testosterone inhibits germ cell apoptosis (Brinkworth et al., 1995; Henrikse´n et al., 1995; Sinha Hikim and Swerdloff, 1999), we performed TUNEL assays. Those
showed a 2-fold increase in germ cell apoptosis in Ocn/ compared to WT mice and a 2-fold decrease in Esp/ testes (Figure 4F). Importantly, there was an increase in the number of apoptotic cells in stages VI–VIII of the spermatogenic cycle, the very stages of spermatogenesis in which testosterone affects germ cell apoptosis more efficiently (Sharpe et al., 1992) (Figure S3C). There were, on the other hand, no abnormalities of germ cell proliferation in either Ocn/ or Esp/ mice Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 801
(Figure S3A). Taken together, these data further indicate that the decrease in fertility demonstrated by Ocn/ male mice is caused by a decrease in testosterone levels. To determine how osteocalcin favors testosterone synthesis by Leydig cells, we tested whether it affects expression of the enzymes that are necessary for testosterone biosynthesis such as StAR, Cyp11a, Cyp17, and 3b-HSD. In cell culture, uncarboxylated osteocalcin increased expression of these genes in Leydig cells (Figure 4F). Accordingly, their expression was significantly decreased in Ocn/ and increased in Esp/ testes (Figures 4G and 4H), whereas it was unaffected in Ocn/ and Esp/ ovaries or adrenal glands (Figures S3D–S3G). Of note, there was no change in expression of Cyp19, the gene encoding the testosterone aromatase, or of HSD-17 in Ocn/ and Esp/ testes (Figures 4F–4H). To further support the notion that osteocalcin influences germ cell apoptosis through testosterone, we examined expression of Gonadotropin Regulated Testicular Helicase (Grth) because this gene has emerged as an essential regulator of spermatogenesis that inhibits germ cell apoptosis and whose expression in germ cells and Leydig cells is regulated by testosterone (Dufau and Tsai-Morris, 2007; Tsai-Morris et al., 2007, 2010). Grth expression was decreased in Ocn/ and increased in Esp/ testes (Figure 4I). GRTH inhibits activation of caspase 3, a determinant of apoptosis (Sheng et al., 2006), and favors expression of tACE, a protein favoring germ cell maturation. Consistent with these notions, western blot analyses showed an increase of cleaved caspase 3 accumulation and a decrease of tACE in Ocn/ testes (Figure 4J). Gprc6a, a G Protein-Coupled Receptor, Transduces Osteocalcin Signal in Leydig Cells To begin to elucidate the molecular mode of action of osteocalcin, we next searched for a receptor expressed in Leydig cells that could transduce its signal. To that end, we used a twostep experimental strategy, taking advantage of the fact that osteocalcin regulates fertility only in male mice. First, we defined the signal transduction pathway used by osteocalcin (these experiments were performed in TM3 cells, and not primary Leydig cells, in order to obtain a sufficient amount of extract for analysis). With this aim, we treated the cells with uncarboxylated osteocalcin and assayed for tyrosine phosphorylation, ERK activation, intracellular calcium accumulation, and cAMP production, using in each case an appropriate positive control. Osteocalcin consistently induced cAMP production in Leydig cells to a level comparable to that induced by human chorionic gonadotropin, the positive control, but did not induce tyrosine phosphorylation, ERK activation, or intracellular calcium accumulation in these cells (Figures 5A–5D). At higher concentrations, osteocalcin stimulation of cAMP production weakened. These data implied that the osteocalcin receptor may be a G protein-coupled receptor (GPCR). Hence, in the second step of this experimental strategy, we took advantage of the dichotomy of function of osteocalcin between males and females and asked how many orphan GPCRs were expressed in testes at a level at least 5-fold higher than in ovary. Twenty-two out of 103 orphan GPCRs tested were predominantly expressed in testes; of these 22, four were enriched in Leydig cells (Figures 5E and 5F). Among 802 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
them, Gprc6a caught our attention because its deletion in all cells results in a metabolic and fertility phenotype similar to that of Ocn/ mice (Pi et al., 2008). Immunofluorescence experiments verified that Gprc6a is expressed in Leydig cells in testes, but not in follicular cells of the ovary. Importantly, the same is true in human gonads (Figures 5G and 5H and Figure S4B). Postnatally, Gprc6a expression peaked within the first week of life, when circulating testosterone levels are elevated. Gprc6a expression then decreased before increasing again at 6 weeks of age, when circulating levels of testosterone also rebound (Figure 5I). We also performed binding assays on mouse testes using biotinylated osteocalcin as a ligand. In the conditions of this assay, osteocalcin bound to Leydig cells and the specificity of this binding was confirmed by several criteria. First, there was no signal when using avidin-biotin alone; second, there was no signal either in other cellular compartments of the testicular tubules; third, we could not detect any binding when using Gprc6a-deficient testes; fourth, osteocalcin binding could be competed away by an excess (100-fold) of unlabeled osteocalcin, but not by the same excess of hCG or of other molecules proposed as ligands of Gprc6a (Wellendorph et al., 2005) (Figure 5J). These data identify Gprc6a as an osteocalcin receptor in Leydig cells. To define Gprc6a function in Leydig cells in vivo, we generated Gprc6aLeydig/ mice. Prior to analyzing these Gprc6aLeydig/ mice, we verified that we had deleted Gprc6a, albeit partially (75% of deletion), in Leydig cells, but not in other organs (Figures S5C and S5D). In Gprc6aLeydig/ male mice, testes size and weight, epididymis and seminal vesicle weights, sperm counts and circulating testosterone levels, and Leydig cell area were all reduced, as was the expression of Grth and the three genes controlling testosterone biosynthesis that are regulated by osteocalcin (Figures 6A–6I and Figure S5E). Accordingly, the number of apoptotic germ cells was increased compared to WT testes (Figure 6J). To establish genetically that Gprc6a is a signaling receptor for osteocalcin in Leydig cells, we analyzed compound mutant mice lacking one allele of Ocn and one allele of Gprc6a in Leydig cells only (Ocn+/;Gprc6aLeydig+/ mice). Whether we looked at testes, epididymis and seminal vesicle weights, or sperm count, Ocn+/;Gprc6aLeydig+/ mice had a phenotype identical to that observed in Gprc6aLeydig/ and Ocnosb/ mice (Figures 6A–6I). CREB Is a Transcriptional Effector of Osteocalcin Signaling in Leydig Cells Observations that osteocalcin treatment of Leydig cells increased cAMP production and that the osteocalcin receptor is a GPCR suggested that CREB could mediate osteocalcin functions in these cells. In support of this hypothesis, osteocalcin treatment of Leydig cells favors CREB phosphorylation (Figure 7A). To investigate this further, we generated mice that lack CREB expression specifically in Leydig cells (CrebLeydig/ mice). Twelve-week-old CrebLeydig/ male mice displayed a reduction in testis size and weight, in epididymides and seminal vesicles weights, in sperm count, and in circulating testosterone levels similar to those seen in Ocn/ and Gprc6aLeydig/ mice (Figures 7B–7G and Figure S6). CrebLeydig/ mice also
lin Insu
B
C
250
* * * * **
150
75
** *
5
pERK
+Osteocalcin
30
15
5
60
15
* *
+Vehicle 5
50 ERK
50
30
60
FBS
44
30
15
+Osteocalcin 60
5
15
30
4.0
60 FBS
44 42
Ionophore 10% FBS Control 3ng/ml Osteocalcin 6ng/ml Osteocalcin 30ng/mlOsteocalcin 600ng/ml Osteocalcin 1μg/ml Osteocalcin
3.0
2.0
1.0
β-actin
0 37
6.0
Testis enriched orphan GPCR (n=22)
**
4.0 2.0
Leydig cells enriched orphan GPCR (n=4)
H
hC G
ng 10
1ng
Mouse Anti-Gprc6a
3n g
0n g
0
Human
I
10.0 8.0 6.0 4.0 2.0 0
WT
Gprc6a Expression in Testis (Relative to 4 Week-old)
Anti-IgG
70 Testis Ovary 100 80 60 40 20 0
J
Anti-Gprc6a 10.0
Anti-IgG
60
50
Gprc6a Relative Expression in human (%)
8.0
40
Gp r13 9
Orphans GPCR (n=103)
r11 2
cAMP Production (pmol/107cells)
**
**
30
Testis Leydig cells
Gp r45
12.0 10.0
20
G
Gp rc6 a
E
Gpcrs Expression in Leydig Cells Relative to Testis (Fold)
D
10
F
Gp
pTyrosine
100
+Vehicle
Fluorescence change (fold over baseline)
FBS
10 n g
3 ng
1 ng
Vehic le
A
Gprc6a-/-
WT
**
8.0
6.0 *
b-Osteocalcin
Control
*
WT
4.0
b-Osteocalcin
WT
WT
2.0
1W eek 4W eek s 6W eek s 12 We eks
0 b-Osteocalcin +hCG
b-Osteocalcin +Lysine
b-Osteocalcin +Osteocalcin
Figure 5. G Protein-Coupled Receptor Gprc6a Is a Receptor for Osteocalcin (A) Anti-phosphotyrosine antibody western blot analysis of TM3 Leydig cells treated with increasing concentrations of osteocalcin, or 10% FBS or insulin as positive controls, for 1 min (top). Proteins phosphorylated on tyrosine residues appear in positive controls (asterisks), but not in osteocalcin-treated cells. Equal loading was assessed using an anti-actin antibody (bottom). (B) Western blot analysis of TM3 Leydig cells showing the absence of ERK1/2 phosphorylation upon stimulation with vehicle or osteocalcin. (C) Calcium fluxes in primary Leydig cells upon stimulation with increasing doses of osteocalcin. 10% FBS and ionophore (A23187) were used as positive controls. (D) cAMP production upon osteocalcin stimulation is increased in TM3 Leydig cells. (E) Schematic representation of the results obtained by the differential expression search for osteocalcin receptors. Among the 103 orphan GPCRs expressed in testis and ovary, 22 were predominantly expressed in testis, and only four were enriched in primary Leydig cells compared to the expression in whole testis. (F) Relative expression of Gprc6a, Gpr45, Gpr112, and Gpr139 in Leydig cells compared to whole testis. (G) qPCR analysis of Gprc6a expression in human testis and ovary. (H) Immunofluorescence analysis of Gprc6a expression in mice and human testis coronal sections. Anti-IgG was used as negative control. (I) qPCR analysis of Gprc6a expression in 1-, 4-, 6-, and 12-week-old WT testes. (J) Cross-sections of testes from WT and Gprc6a-deficient mice stained with biotinylated osteocalcin (b-osteocalcin). (Upper-left) WT testis stained with avidinbiotin complex only. (Upper-middle) WT testis stained with 10 nM of b-osteocalcin. (Upper-right) Testis from Gprc6a-deficient mice stained with 10 nM of b-osteocalcin. (Lower-left) WT testis stained with 10 nM of b-osteocalcin in the presence of 1000 nM hCG. (Lower-middle) WT testis stained with 10 nM of b-osteocalcin in the presence of 1000 nM lysine. (Lower-right) WT testis stained with 10 nM of b-osteocalcin in the presence of 1000 nM of unlabeled osteocalcin. Error bars represent SEM. Student’s t test; *p < 0.05, **p < 0.001. See also Figure S4.
demonstrated a strong decrease in the expression of Grth and of the four genes involved in testosterone biosynthesis whose expression is regulated by osteocalcin (Figures 7H and 7I). In agreement with these data, CREB could bind to the promoter
regions of Cyp11a, 3b-HSD, and StAR (Zhang et al., 2005) (Figure 7J). To establish whether CREB acts downstream of Gprc6a in Leydig cells to regulate male fertility, we generated compound heterozygous mice lacking one copy of Creb and one copy of Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 803
WT (n=39)
C
Ocn+/-; Gprc6aLeydig-/- Gprc6a Leydig+/-
F
Ocn+/-; Gprc6aLeydig+/- (n=6)
Ocn+/-; Gprc6aLeydig+/- (n=6)
Ocn+/-; Gprc6aLeydig+/- (n=6)
Ocn+/- (n=21) Gprc6aLeydig+/- (n=16)
Ocn+/- (n=21) Gprc6aLeydig+/- (n=16)
** *
0.10
0.05
Sperm Count (107)
6.0 5.0 4.0 3.0 2.0 1.0 0
G
WT (n=33)
** *
5 0
H
Gprc6aLeydig-/- (n=8)
WT (n=7)
WT (n=7)
Ocn+/-; Gprc6aLeydig+/- (n=6)
Ocn+/-; Gprc6aLeydig+/- (n=3)
Gprc6aLeydig-/- (n=7)
Gprc6aLeydig-/- (n=7)
Ocn+/- (n=21) Gprc6aLeydig+/- (n=16)
Ocn+/- (n=12) Gprc6aLeydig+/- (n=8)
200
* *
150 100 50
I
100 75
*
50 25 0
120
45 40 35 30 25 20 15 10 5 0
Testosterone (%)
300 250
125
Leydig Cells Are per Interstitial Area (%)
350
0
*
100 80 60 40 20 0
J
WT (n=16)
Expression Relative to WT (%)
** **
45 40 35 30 25 20 15 10
Gprc6aLeydig-/- (n=8)
Grth Expression Relative to WT (%)
Seminal Vesicle Weight (mg)
WT (n=39)
Gprc6aLeydig-/- (n=8)
Ocn+/- (n=21) Gprc6aLeydig+/- (n=16)
0
E
WT (n=39)
Gprc6aLeydig-/- (n=8)
0.15
Testis Weight (g)
WT
D
WT (n=39)
Gprc6aLeydig-/- (n=8)
Epididymis Weight (mg)
B
A
Gprc6aLeydig-/- (n=3)
WT (n=7)
Ocn+/-; Gprc6aLeydig+/- (n=3)
Gprc6aLeydig-/- (n=7)
Ocn+/- (n=12) Gprc6aLeydig+/- (n=8)
1.0
100 80
*
*
*
*
60 40 20
Apoptotic Germ Cells per Tubule
120
0.8
*
0.6 0.4 0.2 0
0 StAR
Cyp11a
3β-HSD
Figure 6. Specific Deletion of Gprc6a in Leydig Cells Decreases Male Fertility
(A–E) Fertility in mice lacking Gprc6a in Leydig cells only (Gprc6aLeydig/) or lacking one allele of Ocn or one allele of Gprc6a in Leydig cells only (Ocn+/ or Gprc6aLeydig+/) or in compound heterozygous mice (Ocn+/; Gprc6aLeydig+/) compared to control littermates. (A) Testis size, (B) testis weight, (C) sperm count, and (D and E) epididymis and seminal vesicle weights. (F) qPCR analysis of Grth expression in mice of indicated genotypes. (G) Ratio between Leydig cells (stained by immunohistochemistry of 3b-HSD) versus testis interstitial areas. (H) Ratio of circulating testosterone levels measured in WT and Gprc6aLeydig/ mice. (I) qPCR analysis of StAR, Cyp11a, and 3b-HSD in Gprc6aLeydig/ and Ocn+/; Gprc6aLeydig+/ compared to WT littermate testes. (J) Germ cell apoptosis analysis by TUNEL assay. All of the analyses were performed in nonbreeder mice. Error bars represent SEM. Student’s t test; *p < 0.05; **p < 0.001. See also Figure S5.
804 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
Gprc6a in Leydig cells. The decrease in fertility demonstrated by CrebLeydig+/;Gprc6aLeydig+/ male mice was similar to that observed in CrebLeydig/ or Gprc6aLeydig/ male mice (Figures 7B–7F and 7H) and was not observed in single-heterozygous mutant mice (data not shown). These results identify CREB as a transcriptional mediator of osteocalcin regulation of testosterone biosynthesis in Leydig cells. DISCUSSION This study reveals that bone is a positive regulator of male fertility. This interaction is mediated through the osteoblastderived hormone osteocalcin, which binds to a specific receptor that is present on Leydig cells of the testes and favors testosterone biosynthesis (Figure 7K). Our results, along with those previously published, support the hypothesis that regulation of bone remodeling, energy metabolism, and reproduction are linked (Ducy et al., 2000; Lee et al., 2007). They also demonstrate that bone is a more important regulator of whole-organism physiology than was anticipated. In the last 10 years, the view of bone as a mere assembly of inert calcified tubes has evolved because two independent lines of investigation painted a much more dynamic picture of this tissue. First, the dialog between bone physiology and energy metabolism and the paramount influence of the brain in the control of bone mass accrual became apparent (Karsenty, 2006). Second, it has been realized that bone is an endocrine organ regulating at least two functions, phosphate metabolism and energy homeostasis, through two distinct osteoblastderived hormones, FGF23 and osteocalcin (Fukumoto and Yamashita, 2007; Lee et al., 2007). These two endocrine functions of bone begged the following question: are these the only functions affected by bone in its endocrine capacity? The well-known regulation of bone remodeling by gonads provides an ideal setting to address this question and raises the possibility that bone, in its endocrine capacity, could influence through a feedback mechanism reproductive functions in either gender. We show here that osteoblasts, the bone-forming cells, favor through osteocalcin fertility in male, but not female, mice. Our observation that Ocn/ mice have low circulating testosterone levels despite an increase in circulating LH levels may have several explanations. For instance, it could be that deletion of osteocalcin causes, for unknown reasons, a loss of negative feedback. Alternatively, it may also suggest that LH cannot favor testosterone production in the absence of osteocalcin. Further experiments will be required to address this point. All of our experiments conducted ex vivo and in vivo in lossand gain-of-function models indicate that the main mechanism whereby osteocalcin favors male fertility is by increasing testosterone synthesis in Leydig cells. Testosterone, in turn, supports maturation and prevents apoptosis of germ cells (Henrikse´n et al., 1995; Sinha Hikim and Swerdloff, 1999; Walker, 2009). Because there is no expression of Gprc6a in Sertoli cells or germ cells, it is likely that osteocalcin regulates male fertility by binding to its receptor on Leydig cells. This was confirmed by the analysis of the Leydig cell-specific deletion of Gprc6a and Creb. Testosterone affects germ cell survival by involving other cell types such as Sertoli cells because germ cells do not
express androgen receptor (Bremner et al., 1994; De Gendt et al., 2004; Wang et al., 2003). Unexpectedly, Ocn/ male mice have a higher level of circulating estrogen than WT littermates, even though osteocalcin does not promote estradiol synthesis. The most likely explanation for this mild increase in circulating estradiol levels in the Ocn/ mice is that the increase in the number of adipocytes caused by Osteocalcin inactivation may result in an increase in the aromatization of testosterone into estrogen in fat (Nelson and Bulun, 2001; Simpson et al., 2000; Simpson, 2003). Importantly, this was also observed in mice lacking Gprc6a (Pi et al., 2008). The existence of such a profound influence exerted by a hormone other than LH on sex steroid hormone synthesis raises the question of whether it is a male-specific phenomenon. To date, we have no evidence that, at least in the mouse, the skeleton favors estrogen production in females. This, however, does not rule out the possibility that other peripheral organs may secrete hormones favoring estrogen synthesis. The growing number of functions identified for osteocalcin makes the identification of its receptor all the more important. An unbiased approach based on the ability of osteocalcin to increase cAMP production in Leydig cells and on its dichotomy of function between male and female gonads led to the identification of Gprc6a as the osteocalcin receptor. This orphan receptor, which belongs to the C family of GPCRs (Wellendorph and Bra¨uner-Osborne, 2004), has been proposed to be a receptor for amino acids, for calcium in the presence of osteocalcin as a cofactor, and for androgen (Pi et al., 2005, 2008, 2010). Yet, the possibility that Gprc6a could be a specific receptor for osteocalcin has never been tested through biochemical or genetic means. Here, we provide biochemical and genetic evidence establishing its identity as a specific receptor for osteocalcin in Leydig cells, though we failed to detect an interaction between amino acids and Gprc6a. That Gprc6a is expressed in human and mice Leydig cells in the testes, but not in follicular cells of ovaries, provides a molecular basis for the fact that osteocalcin affects male fertility only. The expression of this receptor is extremely dynamic and peaks at adulthood when testosterone biosynthesis is at its maximum (Feldman et al., 2002; Gray et al., 1991; Quigley, 2002). This observation suggests that the regulation of osteocalcin functions occurs, at least in part, by regulating the expression of its receptor. The identification of an osteocalcin receptor opens the door to a thorough molecular dissection of the mode of action of osteocalcin in various cell types where it is expressed. This may eventually lead to the identification of additional functions for osteocalcin. An obvious question raised by this study is whether the skeleton also regulates male fertility in humans. In the absence of inactivating mutations in the Osteocalcin or Gprc6a genes in humans and of studies correlating circulating levels of uncarboxylated osteocalcin and fertility in the aging male population, this question can only be addressed through indirect means for now. Although we cannot rule out the possibility that the endocrinology of reproduction may be different in this particular aspect between rodents and humans, indirect evidence suggests that the function of osteocalcin described here may be conserved in humans. First, and most importantly, this work Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 805
A
0
2
5
15
30
37 0
2
5
15
30
60
CREB 43 β-Actin
0.15
50
Testis Weight (g)
β-Actin
CrebLeydig+/-; Gprc6aLeydig+/-
CrebLeydig-/-
WT
pCREB 43
WT (n=10) CrebLeydig-/- (n=6) CrebLeydig+/-; Gprc6aLeydig+/- (n=5)
C
B
60
50
* *
0.10
0.05
37 0
* * 30 20 10 0
0 WT (n=10) CrebLeydig-/- (n=6) CrebLeydig+/-; Gprc6aLeydig+/- (n=5) 125 100 75
*
*
50 25
150
*
300
* 200
100 0
100
50
*
0
J
I Expression Relative to WT (%)
Grth Expression Relative to WT (%)
H
WT (n=9) CrebLeydig-/- (n=6)
140
WT (n=10) CrebLeydig-/- (n=6)
120
Cyp11a1 hCre -1364
100
Cyp11a1 hCre -4176
*
80 60
*
3β-HSD hCre -993
* *
3β-HSD hCre-3109
40 StAR hCre-40
20
No Cre
0
0
StAR
Cyp17
Cyp11a
3β-HSD
Bone
Cyp19a
HSD-17
Testis
K Gprc6a
P
P
StAR
CREBCREB Cyp11
Cyp17 3β-HSD
Leydig cell
Basement membrane
Osteocalcin
Apoptosis
Germ Cells
806 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
Spermatogenesis
Testosterone
Input
40
G
IP anti-IgG
2.0
400
WT (n=10) CrebLeydig-/- (n=6) CrebLeydig+/-; Gprc6aLeydig+/- (n=5)
IP anti-CREB
* *
4.0
50
F
Testosterone (%)
Sperm Count (107)
6.0
E
WT (n=10) CrebLeydig-/- (n=6) CrebLeydig+/-; Gprc6aLeydig+/- (n=5)
Seminal Vesicle Weight (mg)
WT (n=10) CrebLeydig-/- (n=6) CrebLeydig+/-; Gprc6aLeydig+/- (n=5)
Epididymis Weight (mg)
D
was, in part, initiated because of a clinical observation made in humans: loss of sex steroid hormones triggers a decrease in bone mass. This led us to test whether feedback regulation of fertility by bone may also occur. Second, Gprc6a is expressed, in human and in mice, in Leydig cells of the testes, but not in ovaries. Hence, the entire signaling cascade from the hormone to the receptor exists in the equivalent organs in humans. Third, the growing number of reports indicating that osteocalcin is a reliable indicator of glucose intolerance, just as it is in mice, strongly suggests that this molecule also acts as a hormone in humans (Kanazawa et al., 2010; Saleem et al., 2010; Yeap et al., 2010). Along these lines, there is no example yet of a molecule being a hormone in the mouse that has abruptly lost this attribute in humans. This is, nevertheless, an aspect of osteocalcin that will need further investigation in the future. EXPERIMENTAL PROCEDURES Mice Generation All experiments were performed on the 129-Sv (Taconic) genetic background. Control littermates were used in all experiments. Mice genotypes were determined by PCR; primer sequences are available upon request. OsteocalcinmCherry knockin, Ocn conditional, and Gprc6a conditional allele generation strategies are described in Figure S2 and Figure S5. Primary Leydig Cells and Testes Explant Preparation Adult mouse Leydig cells were isolated by mechanical dissociation of the testes followed by purification on a 0%–90% Percoll gradient (Hunter et al., 1982; Schumacher et al., 1978). Primary Leydig cells were cultured in Minimal Essential Medium (MEM + GlutaMAX, Invitrogen) supplemented with 1 3 PenStrep, 25 mM HEPES (pH 7.4), and 0.07% BSA at 33 C in 5% CO2. After 3 hr of attaching and starvation, cells were washed once with culture medium and then used for experiments. The testes explant preparation protocol was adapted from (Powlin et al., 1998). Explants were washed three times with PBS and placed in serum-free RPMI medium for 2 hr before being used for experiments. Osteocalcin Stimulation of Leydig Cells or Testes Explants Primary Leydig cells and testes explants were washed three times with PBS and stimulated with different doses of recombinant osteocalcin prepared as previously described (Ferron et al., 2008) or with hCG as a positive control. After 1 hr, an aliquot of medium was collected for measurements of testosterone. Cells were then maintained for 3 additional hr and lysed in 1 ml TRIZOL (Invitrogen) for RNA isolation. Sperm Counts and Hormone Measurements Caudal epididymides were minced in 1 ml PBS, and the number of cells released was counted after 1 hr. The total sperm count was assessed in the final suspension by using a hemocytometer (Dakhova et al., 2009). Circulating levels of testosterone, estradiol (E2), and progesterone were measured by
radioimmunoassay (RIA) from Diagnostic Systems Laboratories (Testosterone RIA DSL-4000, Estradiol RIA DSL-43100, and Progesterone RIA-3900). Histology One testis or ovary from each mouse was randomly selected for molecular analysis, and the other one was used for histology. Specimens were collected, weighed, and fixed in Bouin’s fixative for histological analyses before being dehydrated through graded ethanol, processed for paraffin embedding, and serially sectioned at 5 mm. For histological analysis, testes and ovaries sections were stained with periodic acid-Schiff and counterstained with hematoxylin. TUNEL labeling was performed using the ApopTag Peroxydase In Situ Apoptosis detection kit (Millipore-S7100). Apoptotic indices were determined by counting the total number of TUNEL-positive cells or the number of TUNELpositive germ cells at different stages (Russell et al., 1990). Approximately 500 tubules were counted on at least four cross-sections located at midtestis for each animal. Gene Expression Studies RNA was purified from tissues, primary Leydig cells, or cultured cells using TRIZOL (Invitrogen). RNA isolation, cDNA preparation, and qPCR analysis was carried out following standard protocols. qPCR analyses were performed using specific quantitative PCR primers from SABiosciences (http://www. sabiosciences.com/RT2PCR.php). cAMP Quantification For cAMP measurements, TM3 Leydig cells were plated in 6 cm dishes (107 cells per dish) 1 day before the experiment. Cells were serum starved for 16 hr (in the presence of 0.1% BSA) and then preincubated in the presence of 0.5 mM IBMX for 30 min and stimulated with indicated concentration of osteocalcin also in the presence of 0.5 mM IBMX for 30 min. cAMP concentration was measured with the Parameter cAMP kit (R&D Systems, KGE002). Receptor Binding Assays For binding studies, testes from 8-week-old mice were snap-frozen in liquid nitrogen, and 20 mm thick sections were prepared and desiccated overnight at 4 C under vacuum. On the following day, sections were rehydrated in icecold binding buffer (50 mM TrisHCl [pH 7.4], 10 mM MgCl2, 0.1 mM EDTA, and 0.1% BSA) for 15 min and incubated for 1 hr in the presence of biotinylated osteocalcin. For competition assays, a 100-fold molar excess of unlabeled osteocalcin, glycine, lysine, or hCG was added. After three washes in cold PBS, sections were incubated for 1 hr in the detection system containing 0.1% BSA (ABC Elite, Vector Laboratories), washed again, and incubated with DAB peroxidase substrate kit (Vector Laboratories) according to the manufacturer’s protocol. After a final wash, sections were mounted in waterbased mounting medium. As negative controls, we used sections incubated with the detection system only (ABC Elite and DAB) or Gprc6a/ testis sections (Basura et al., 2008). SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures and six figures and can be found with this article online at doi:10.1016/j.cell. 2011.02.004.
Figure 7. CREB Is a Transcription Factor-Mediating Osteocalcin-Evoked Gene Expression in Leydig Cells (A) Western blot analysis of CREB activation upon stimulation with osteocalcin. (B–F) Fertility in mice lacking Creb in Leydig cells (CrebLeydig/) or of compound heterozygous mice (CrebLeydig+/; Gprc6aLeydig+/) compared to control littermates. (B) Testis size, (C) testis weight, (D) sperm count, and (E and F) epididymis and seminal vesicle weights. (G) Quantification of circulating testosterone levels represented as fold change compared to WT. (H) qPCR analysis of Grth expression in mice of indicated genotypes. (I) qPCR analysis of StAR, Cyp11a, Cyp17, 3b-HSD, Cyp19, and HSD-17 in CrebLeydig/ compared to control littermate testes. (J) Chromatin immunoprecipitation (ChIP) using anti-CREB antibody and unspecific isotype IgG antibody in the TM3 cell line. (K) Model representing current knowledge about the regulation of male fertility by the skeleton. All of the analyses were performed in nonbreeder mice. Error bars represent SEM. Student’s t test; *p < 0.05. See also Figure S6.
Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 807
ACKNOWLEDGMENTS We thank Drs. Behringer and Ko for generously providing the Amh-mis-Cre and Cyp17-iCre mice; Dr. Hampson for generously providing us Gprc6a antibody; Dr. A.F. Parlow for measurements of LH in serum (National Hormone and Peptide Program); Drs. Przedborski and Jackson-Lewis for gracious help with stereological analysis; Drs. Mark, Yadav, Yoshizawa, Allen, and Setola for critical experimental advice; and G. Ren for mouse genotyping. This work was supported by grants from the NIH (G.K.), Fond de la recherche en sante´ du Que´bec (M.F.), and HFSP (G.S. and F.O.). Received: November 18, 2010 Revised: January 12, 2011 Accepted: January 31, 2011 Published online: February 17, 2011
REFERENCES Basura, G.J., Abbas, A.I., O’Donohue, H., Lauder, J.M., Roth, B.L., Walker, P.D., and Manis, P.B. (2008). Ontogeny of serotonin and serotonin2A receptors in rat auditory cortex. Hear. Res. 244, 45–50. Bremner, W.J., Millar, M.R., Sharpe, R.M., and Saunders, P.T. (1994). Immunohistochemical localization of androgen receptors in the rat testis: evidence for stage-dependent expression and regulation by androgens. Endocrinology 135, 1227–1234.
Ferron, M., Wei, J., Yoshizawa, T., Del Fattore, A., DePinho, R.A., Teti, A., Ducy, P., and Karsenty, G. (2010). Insulin signaling in osteoblasts integrates bone remodeling and energy metabolism. Cell 142, 296–308. Fukumoto, S., and Yamashita, T. (2007). FGF23 is a hormone-regulating phosphate metabolism—unique biological characteristics of FGF23. Bone 40, 1190–1195. Fulzele, K., Riddle, R.C., DiGirolamo, D.J., Cao, X., Wan, C., Chen, D., Faugere, M.C., Aja, S., Hussain, M.A., Bru¨ning, J.C., and Clemens, T.L. (2010). Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition. Cell 142, 309–319. Gray, A., Berlin, J.A., McKinlay, J.B., and Longcope, C. (1991). An examination of research design effects on the association of testosterone and male aging: results of a meta-analysis. J. Clin. Epidemiol. 44, 671–684. Harada, S., and Rodan, G.A. (2003). Control of osteoblast function and regulation of bone mass. Nature 423, 349–355. Henrikse´n, K., Hakovirta, H., and Parvinen, M. (1995). Testosterone inhibits and induces apoptosis in rat seminiferous tubules in a stage-specific manner: in situ quantification in squash preparations after administration of ethane dimethane sulfonate. Endocrinology 136, 3285–3291. Hunter, M.G., Magee-Brown, R., Dix, C.J., and Cooke, B.A. (1982). The functional activity of adult mouse Leydig cells in monolayer culture. Effect of lutropin and foetal calf serum. Mol. Cell. Endocrinol. 25, 35–47. Jamin, S.P., Arango, N.A., Mishina, Y., Hanks, M.C., and Behringer, R.R. (2003). Genetic studies of the AMH/MIS signaling pathway for Mu¨llerian duct regression. Mol. Cell. Endocrinol. 211, 15–19.
Bridges, P.J., Koo, Y., Kang, D.W., Hudgins-Spivey, S., Lan, Z.J., Xu, X., DeMayo, F., Cooney, A., and Ko, C. (2008). Generation of Cyp17iCre transgenic mice and their application to conditionally delete estrogen receptor alpha (Esr1) from the ovary and testis. Genesis 46, 499–505.
Kanazawa, I., Yamaguchi, T., Tada, Y., Yamauchi, M., Yano, S., and Sugimoto, T. (2010). Serum osteocalcin level is positively associated with insulin sensitivity and secretion in patients with type 2 diabetes. Bone, in press. Published online December 23, 2010. 10.1016/j.bone.2010.12.020.
Brinkworth, M.H., Weinbauer, G.F., Schlatt, S., and Nieschlag, E. (1995). Identification of male germ cells undergoing apoptosis in adult rats. J. Reprod. Fertil. 105, 25–33.
Karsenty, G. (2006). Convergence between bone and energy homeostases: leptin regulation of bone mass. Cell Metab. 4, 341–348.
Dacquin, R., Starbuck, M., Schinke, T., and Karsenty, G. (2002). Mouse alpha1 (I)-collagen promoter is the best known promoter to drive efficient Cre recombinase expression in osteoblast. Dev. Dyn. 224, 245–251. Dacquin, R., Mee, P.J., Kawaguchi, J., Olmsted-Davis, E.A., Gallagher, J.A., Nichols, J., Lee, K., Karsenty, G., and Smith, A. (2004). Knock-in of nuclear localised beta-galactosidase reveals that the tyrosine phosphatase Ptprv is specifically expressed in cells of the bone collar. Dev. Dyn. 229, 826–834. Dakhova, O., O’Day, D., Kinet, N., Yucer, N., Wiese, M., Shetty, G., and Ducy, P. (2009). Dickkopf-like1 regulates postpubertal spermatocyte apoptosis and testosterone production. Endocrinology 150, 404–412. De Gendt, K., Swinnen, J.V., Saunders, P.T., Schoonjans, L., Dewerchin, M., Devos, A., Tan, K., Atanassova, N., Claessens, F., Le´cureuil, C., et al. (2004). A Sertoli cell-selective knockout of the androgen receptor causes spermatogenic arrest in meiosis. Proc. Natl. Acad. Sci. USA 101, 1327–1332. Ducy, P., Amling, M., Takeda, S., Priemel, M., Schilling, A.F., Beil, F.T., Shen, J., Vinson, C., Rueger, J.M., and Karsenty, G. (2000). Leptin inhibits bone formation through a hypothalamic relay: a central control of bone mass. Cell 100, 197–207. Dufau, M.L., and Tsai-Morris, C.H. (2007). Gonadotropin-regulated testicular helicase (GRTH/DDX25): an essential regulator of spermatogenesis. Trends Endocrinol. Metab. 18, 314–320. Feldman, H.A., Longcope, C., Derby, C.A., Johannes, C.B., Araujo, A.B., Coviello, A.D., Bremner, W.J., and McKinlay, J.B. (2002). Age trends in the level of serum testosterone and other hormones in middle-aged men: longitudinal results from the Massachusetts male aging study. J. Clin. Endocrinol. Metab. 87, 589–598. Ferron, M., Hinoi, E., Karsenty, G., and Ducy, P. (2008). Osteocalcin differentially regulates beta cell and adipocyte gene expression and affects the development of metabolic diseases in wild-type mice. Proc. Natl. Acad. Sci. USA 105, 5266–5270.
808 Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc.
Karsenty, G., Kronenberg, H.M., and Settembre, C. (2009). Genetic control of bone formation. Annu. Rev. Cell Dev. Biol. 25, 629–648. Khosla, S., Melton, L.J., III, Atkinson, E.J., and O’Fallon, W.M. (2001). Relationship of serum sex steroid levels to longitudinal changes in bone density in young versus elderly men. J. Clin. Endocrinol. Metab. 86, 3555–3561. Lee, N.K., Sowa, H., Hinoi, E., Ferron, M., Ahn, J.D., Confavreux, C., Dacquin, R., Mee, P.J., McKee, M.D., Jung, D.Y., et al. (2007). Endocrine regulation of energy metabolism by the skeleton. Cell 130, 456–469. Manolagas, S.C., Kousteni, S., and Jilka, R.L. (2002). Sex steroids and bone. Recent Prog. Horm. Res. 57, 385–409. Nakamura, T., Imai, Y., Matsumoto, T., Sato, S., Takeuchi, K., Igarashi, K., Harada, Y., Azuma, Y., Krust, A., Yamamoto, Y., et al. (2007). Estrogen prevents bone loss via estrogen receptor alpha and induction of Fas ligand in osteoclasts. Cell 130, 811–823. Nelson, L.R., and Bulun, S.E. (2001). Estrogen production and action. J. Am. Acad. Dermatol. 45(3, Suppl), S116–S124. Pi, M., Faber, P., Ekema, G., Jackson, P.D., Ting, A., Wang, N., Fontilla-Poole, M., Mays, R.W., Brunden, K.R., Harrington, J.J., and Quarles, L.D. (2005). Identification of a novel extracellular cation-sensing G-protein-coupled receptor. J. Biol. Chem. 280, 40201–40209. Pi, M., Chen, L., Huang, M.Z., Zhu, W., Ringhofer, B., Luo, J., Christenson, L., Li, B., Zhang, J., Jackson, P.D., et al. (2008). GPRC6A null mice exhibit osteopenia, feminization and metabolic syndrome. PLoS ONE 3, e3858. Pi, M., Parrill, A.L., and Quarles, L.D. (2010). GPRC6A mediates the nongenomic effects of steroids. J Biol. Chem. 285, 39953–39964. Powlin, S.S., Cook, J.C., Novak, S., and O’Connor, J.C. (1998). Ex vivo and in vitro testis and ovary explants: utility for identifying steroid biosynthesis inhibitors and comparison to a Tier I screening battery. Toxicol. Sci. 46, 61–74. Quigley, C.A. (2002). Editorial: The postnatal gonadotropin and sex steroid surge-insights from the androgen insensitivity syndrome. J. Clin. Endocrinol. Metab. 87, 24–28.
Riggs, B.L., O’Fallon, W.M., Muhs, J., O’Connor, M.K., Kumar, R., and Melton, L.J., III. (1998). Long-term effects of calcium supplementation on serum parathyroid hormone level, bone turnover, and bone loss in elderly women. J. Bone Miner. Res. 13, 168–174. Rodan, G.A., and Martin, T.J. (2000). Therapeutic approaches to bone diseases. Science 289, 1508–1514. Russell, L.D., Attlin, R., Sinha Hikim, A., and Clegg, E. (1990). Histological and histopathological evaluation of the testis (Clearwater, FL: Cache River Press). Saleem, U., Mosley, T.H., Jr., and Kullo, I.J. (2010). Serum osteocalcin is associated with measures of insulin resistance, adipokine levels, and the presence of metabolic syndrome. Arterioscler. Thromb. Vasc. Biol. 30, 1474–1478. Schumacher, M., Scha¨fer, G., Holstein, A.F., and Hilz, H. (1978). Rapid isolation of mouse Leydig cells by centrifugation in Percoll density gradients with complete retention of morphological and biochemical integrity. FEBS Lett. 91, 333–338. Sharpe, R.M., Maddocks, S., Millar, M., Kerr, J.B., Saunders, P.T., and McKinnell, C. (1992). Testosterone and spermatogenesis. Identification of stage-specific, androgen-regulated proteins secreted by adult rat seminiferous tubules. J. Androl. 13, 172–184. Sheng, Y., Tsai-Morris, C.H., Gutti, R., Maeda, Y., and Dufau, M.L. (2006). Gonadotropin-regulated testicular RNA helicase (GRTH/Ddx25) is a transport protein involved in gene-specific mRNA export and protein translation during spermatogenesis. J. Biol. Chem. 281, 35048–35056. Simpson, E.R. (2003). Sources of estrogen and their importance. J. Steroid Biochem. Mol. Biol. 86, 225–230. Simpson, E., Rubin, G., Clyne, C., Robertson, K., O’Donnell, L., Jones, M., and Davis, S. (2000). The role of local estrogen biosynthesis in males and females. Trends Endocrinol. Metab. 11, 184–188. Sinha Hikim, A.P., and Swerdloff, R.S. (1999). Hormonal and genetic control of germ cell apoptosis in the testis. Rev. Reprod. 4, 38–47.
Sua´rez, S.S., and Osman, R.A. (1987). Initiation of hyperactivated flagellar bending in mouse sperm within the female reproductive tract. Biol. Reprod. 36, 1191–1198. Tsai-Morris, C.H., Koh, E., Sheng, Y., Maeda, Y., Gutti, R., Namiki, M., and Dufau, M.L. (2007). Polymorphism of the GRTH/DDX25 gene in normal and infertile Japanese men: a missense mutation associated with loss of GRTH phosphorylation. Mol. Hum. Reprod. 13, 887–892. Tsai-Morris, C.H., Sheng, Y., Gutti, R., Li, J., Pickel, J., and Dufau, M.L. (2010). Gonadotropin-regulated testicular RNA helicase (GRTH/DDX25) gene: cellspecific expression and transcriptional regulation by androgen in transgenic mouse testis. J. Cell. Biochem. 109, 1142–1147. Walker, W.H. (2009). Molecular mechanisms of testosterone action in spermatogenesis. Steroids 74, 602–607. Wang, G.M., O’Shaughnessy, P.J., Chubb, C., Robaire, B., and Hardy, M.P. (2003). Effects of insulin-like growth factor I on steroidogenic enzyme expression levels in mouse leydig cells. Endocrinology 144, 5058–5064. Wellendorph, P., and Bra¨uner-Osborne, H. (2004). Molecular cloning, expression, and sequence analysis of GPRC6A, a novel family C G-protein-coupled receptor. Gene 335, 37–46. Wellendorph, P., Hansen, K.B., Balsgaard, A., Greenwood, J.R., Egebjerg, J., and Bra¨uner-Osborne, H. (2005). Deorphanization of GPRC6A: a promiscuous L-alpha-amino acid receptor with preference for basic amino acids. Mol. Pharmacol. 67, 589–597. Yeap, B.B., Chubb, S.A., Flicker, L., McCaul, K.A., Ebeling, P.R., Beilby, J.P., and Norman, P.E. (2010). Reduced serum total osteocalcin is associated with metabolic syndrome in older men via waist circumference, hyperglycemia, and triglyceride levels. Eur. J. Endocrinol. 163, 265–272. Zhang, X., Odom, D.T., Koo, S.H., Conkright, M.D., Canettieri, G., Best, J., Chen, H., Jenner, R., Herbolsheimer, E., Jacobsen, E., et al. (2005). Genome-wide analysis of cAMP-response element binding protein occupancy, phosphorylation, and target gene activation in human tissues. Proc. Natl. Acad. Sci. USA 102, 4459–4464.
Cell 144, 796–809, March 4, 2011 ª2011 Elsevier Inc. 809
Astrocyte-Neuron Lactate Transport Is Required for Long-Term Memory Formation Akinobu Suzuki,1 Sarah A. Stern,1,6 Ozlem Bozdagi,1,2,6 George W. Huntley,1 Ruth H. Walker,3,4 Pierre J. Magistretti,5,* and Cristina M. Alberini1,2,* 1Department
of Neuroscience of Psychiatry Mount Sinai School of Medicine, New York, New York 10029 3Department of Neurology, James J. Peters VAMC, Bronx, NY 10468 4Department of Neurology, Mount Sinai School of Medicine, New York, New York 10029 5Laboratory of Neuroenergetics and Cellular Dynamics, Brain Mind Institute, Ecole Polytechnique Fe ´ de´rale de Lausanne (EPFL) and Center for Psychiatric Neuroscience, University of Lausanne-CHUV, Lausanne, Switzerland 6These authors contributed equally to this work *Correspondence:
[email protected] (C.M.A.), pierre.magistretti@epfl.ch (P.J.M.) DOI 10.1016/j.cell.2011.02.018 2Department
SUMMARY
We report that, in the rat hippocampus, learning leads to a significant increase in extracellular lactate levels that derive from glycogen, an energy reserve selectively localized in astrocytes. Astrocytic glycogen breakdown and lactate release are essential for long-term but not short-term memory formation, and for the maintenance of long-term potentiation (LTP) of synaptic strength elicited in vivo. Disrupting the expression of the astrocytic lactate transporters monocarboxylate transporter 4 (MCT4) or MCT1 causes amnesia, which, like LTP impairment, is rescued by L-lactate but not equicaloric glucose. Disrupting the expression of the neuronal lactate transporter MCT2 also leads to amnesia that is unaffected by either L-lactate or glucose, suggesting that lactate import into neurons is necessary for long-term memory. Glycogenolysis and astrocytic lactate transporters are also critical for the induction of molecular changes required for memory formation, including the induction of phospho-CREB, Arc, and phospho-cofilin. We conclude that astrocyte-neuron lactate transport is required for long-term memory formation. INTRODUCTION Astrocytes have been generally believed to have a mainly supportive function for neurons in the central nervous system (Kettenmann and Ransom, 2005). However, a growing body of evidence suggests that they play many more active roles, including information processing, signal transmission and regulation of neural and synaptic plasticity (Halassa and Haydon, 2010; Henneberger et al., 2010; Perea et al., 2009). Thus, brain functions, including perhaps cognitive ones, may result from 810 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
the concerted action of a neuron–glia network. Little is known about the physiological contribution and relative mechanisms of astrocytes to cognitive functions. Astrocytic adenosine modulates sleep homeostasis and cognitive consequences of sleep loss (Halassa et al., 2009), and some studies have begun investigating the role of astrocytes in learning and memory. Spatial learning and working memory in rats appears to increase the number of astrocytes (Jahanshahi et al., 2008a, 2008b) and bead discrimination memory in a day-old chick requires glycogenolysis (Gibbs et al., 2006), which is known to occur in astrocytes and not in neurons (Magistretti, 2008), suggesting that byproducts of glycogen metabolism are critical for memory formation. One mechanism through which astrocytes and neurons may actively be linked is through metabolic coupling, particularly as a result of neuronal activity (e.g., Magistretti, 2008; Tsacopoulos and Magistretti, 1996). An important source of energy for the brain is glucose entering via transport across the blood brain barrier into both neurons and astrocytes. Astrocytes, but not neurons, store glycogen (Brown et al., 2004; Magistretti, 2008; Vilchez et al., 2007) that can be rapidly converted to pyruvate/ lactate and metabolized in the TCA cycle or used for biosynthesis of glutamate (Hamprecht et al., 2005), as well as also potentially produce glucose (Ghosh et al., 2005). Both glucose and lactate can then be exported to neurons as fuel (Brown et al., 2004; Dringen et al., 1993). Alternatively, astrocytes may utilize glycogen and spare the glucose for neurons’ use (DiNuzzo et al., 2010; Swanson, 1992). Glycogenolysis also results in lactate release from astrocytes (Brown et al., 2004; Dringen et al., 1993), and lactate transport between astrocytes and neurons has been shown from glutamate-stimulated glycolysis (Magistretti et al., 1999; Pellerin and Magistretti, 1994) and/or glycogenolysis stimulated by neuromodulators such as noradrenaline and vasoactive intestinal peptide (VIP) (Magistretti, 2006; Magistretti and Morrison, 1988). The transport of lactate occurs via monocarboxylate transporters (MCTs), proton-linked membrane carriers that transport monocarboxylates including lactate, pyruvate and ketone bodies across the cell membrane. MCT4 is expressed mainly by astrocytes, whereas MCT2 is primarily neuronal and
MCT1 is expressed in astrocytes, endothelial cells of microvessels ependymocytes and oligodendrocytes (Pierre and Pellerin, 2005; Rinholm et al., 2011). The hypothesis of an astrocyteneuron metabolic coupling is attractive because it would underscore the importance of functionally connecting these two cell types during activity subserving brain functions, including perhaps cognitive ones. New learning initiates a cascade of events in the brain that can lead to short and long-term memories. While short-term memories require posttranslational modifications, the consolidation, or stabilization of long-term memories depends upon the activation of a gene cascade and downstream modifications in neurons that store the acquired information (Dudai, 2004; Kandel, 2001; Klann and Sweatt, 2008). Among the best characterized and widely proven gene expression mechanisms known to underlie memory consolidation are the activation of CREB (cAMP response element binding protein)-dependent gene expression (Alberini, 2009; Kandel, 2001) and the translation, at activated synapses, of the immediate early gene Arc (activity-regulated cytoskeletal protein), which is believed to play a key role in actin cytoskeletal dynamics and regulate the membrane expression of AMPA receptors (Bramham et al., 2008). Long-term synaptic plasticity and memory are accompanied by synaptic structural changes, which involve actin polymerization (Chen et al., 2007; Fischer et al., 2004; Mantzura et al., 2009) associated with the phosphorylation of the p21-activated kinase-cofilin cascade. Phosphorylated cofilin is one of the major regulators of F-actin dynamics in spines, promotes cytoskeleton assembly and regulates spine morphology (Bamburg et al., 1999; Chen et al., 2007; Fedulov et al., 2007). Thus, long-term memory formation appears to have high metabolic demands within the underlying active neuronal network. Here, we employed rat inhibitory avoidance (IA) to test the hypothesis that astrocyte-neuron metabolic coupling mechanisms play a critical role in memory formation. RESULTS Astrocytic Glycogen Metabolism Is Required in the Hippocampus for Long-Term Memory Formation To test whether astrocytic glycogenolysis in the hippocampus affects short- and/or long-term memory retention, we bilaterally injected 300 mM (300 pmol) of the inhibitor of glycogen phosphorylation 1,4-dideoxy-1,4-imino-D-arabinitol (DAB) (Walls et al., 2008), either 15 min before or immediately after IA training (Figures 1A–1C). Chicago sky blue injections showed the diffusion, which spread into the dorsal hippocampus approximately 1.7 mm across the injection site (Figure S1). To test for a doseresponse effect, 1000 mM (1000 pmol) of DAB, injected immediately after training, was also investigated (Figure 1C). Student’s t test showed that DAB injected before training did not affect IA acquisition, as the mean latencies to enter the shock compartment during training (acquisition) were similar in all groups (p > 0.05). DAB did not affect short-term memory 1 hr after training (p > 0.05, Student’s t test; Figure 1A). However, DAB injected either before (Figure 1B) or immediately after training (Figure 1C) significantly blocked long-term memory tested at 24 hr (Test 1) in a dose-dependent manner. Retesting
6 days after Test 1 showed that the memory loss persisted (Test 2, p < 0.05, two-way repeated-measure analysis of variance [ANOVA] followed by Newman-Keuls post hoc test; Figure 1B and 1C). The disrupted memory did not reinstate after a foot-shock reminder given one day after Test 2 and tested 24 hr later (Test 3, 15 min before, p < 0.05, Student’s t test; immediately after, p < 0.05, one-way ANOVA followed by Newman-Keuls post hoc test). Retraining of the DAB-injected group, 24 hr after Test 3, resulted in normal memory retention 24 hr later (Test 4), indicating that the hippocampus was functionally intact (Figure 1B and 1C). To begin determining the time course of the DAB effect following training, rats were injected with 300 pmol of DAB 24 hr after training and tested 24 hr (Test 1) and 5 days (Test 2) later (Figure 1D). At both Tests, DAB and vehicle-injected rats had comparable memory retention (p > 0.05, two-way repeated-measure ANOVA), indicating that DAB disrupts longterm memory formation during a limited time window: it has a maximal effect when injected shortly before training but no effect when injected 24 hr after training. Similarly, bilateral hippocampal injections of another glycogen phosphorylase inhibitor, isofagomine (Waagepetersen et al., 2000), 15 min before training, significantly disrupted memory retention 24 hr later compared to vehicle (p < 0.05, Student’s t test; Figure 1E). Thus, astrocytic hippocampal glycogen mobilization is required for long-term, but not short-term, memory formation. Learning-Induced Astrocytic Release of Lactate Is Required for Memory Formation In vitro studies have suggested that glycogenolysis results in lactate release (Brown et al., 2004; Dringen et al., 1993). To test whether glycogenolysis leads to lactate release in vivo after learning, we determined the levels of lactate in the hippocampus of rats before and after training using in vivo microdialysis (Rex et al., 2009). Basal lactate dialysate levels remained constant over time before training (Vehicle, 197.95 ± 55.95 mM; DAB, 216.14 ± 43.62 mM). Training led to a statistically significant increase in lactate dialysate levels compared to baseline concentrations (170%–190%, p < 0.05, two-way repeatedmeasure ANOVA followed by Bonferroni post hoc test; Figure 2A), which was completely abolished by DAB in the perfusion medium, indicating that it resulted from glycogenolysis (p < 0.05, two-way repeated-measure ANOVA followed by Bonferroni post hoc test; Figure 2A). Retention latencies of both DAB- and vehicle-injected groups are shown in Figure S2. We then asked whether the DAB-induced amnesia could be rescued by the administration of exogenous lactate. Fifteen min before training, rats were bilaterally injected into the hippocampus with a combination of DAB or vehicle and either 10 nmol (Figure 2B), 100 nmol L-lactate (Figure 2C) or vehicle. Memory was tested 24 hr (Test 1) and 7 days (Test 2) after training. Acquisition was not affected by treatment (p > 0.05; one-way ANOVA). DAB injection with vehicle disrupted long-term memory retention, confirming our previous results (Figure 1B). Conversely, 100, but not 10 nmol of L-lactate coadministered with DAB significantly rescued the memory loss. Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 811
A
B
Training(0h) Test -15min
Test 1 6d
Acq
Test
500 400 300 200 100 0
Acq
Test 1
300 pmol DAB
Reminder shock
Training(0h) Test 1 24h
Mean Latency (sec.)
500 400 300 200 100 0
C
24h
24h
6d
24h
24h
Test 4
24h
Training(0h) Test 2
Test 1
Test 4 24h
24h 24h
5d
Inject
acq
Test 1
Vehicle
Test 2
300 pmol DAB
Test 3
Test 4
1000 pmol DAB
Mean Latency (sec.)
500 400 300 200 100 0
Test 3
300 pmol DAB
D Retraining
Test 3
Test 2
Test 2
Vehicle
Inject Mean Latency (sec.)
24h
Test 4
Inject
Vehicle
E
24h
Retraining
Test 3
Test 2
-15min 24h
1h
Inject Mean Latency (sec.)
Reminder shock
Training(0h)
500 400 300 200 100 0
Acq
Test 1 Vehicle
Test 2
300 pmol DAB
Training(0h) Test -15min
24h
Mean Latency (sec.)
Inject 500 400 300 200 100 0
Acq
Test
Vehicle
Isofagomine
Figure 1. DAB and Isofagomine Disrupt Long-Term Memory Acquisition (Acq) and retention are expressed as mean latency ± SEM (in seconds, sec). Latency scores, n and detailed statistic are reported in Table S1. See also Figure S1. (A) Hippocampal injections of DAB 15 min before IA training had no effect on short-term memory tested at 1 hr after training (n = 7/group). (B and C) Hippocampal injections of DAB 15 min before ([B] n = 11/group) or immediately after training ([C] n = 7–9/group) disrupted long-term memory at 24 hr (Test 1). The disruption persisted 7 days after training (Test 2), and memory did not recover after a reminder shock (Test 3). DAB-injected rats had normal retention after retraining (Test 4). (D) Hippocampal injections of DAB 24 hr after IA training did not affect long-term memory (n = 8/group). (E) Hippocampal injections of isofagomine 15 min before training disrupted long-term memory (n = 8-9/group). *p < 0.05.
Retesting the rats 7 days after training revealed that the effects of treatments persisted (p < 0.05 for both tests, two-way repeated-measure ANOVA followed by Newman-Keuls post test). To additionally prove that the transport of L-lactate into neurons is critical for the formation of long-term memory, the 812 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
competitive effect of the inactive isomer D-lactate was tested. Bilateral hippocampal injection of 20 nmol of D-lactate, 15 min before IA training, significantly blocked memory retention 24 hr after training compared to vehicle (p < 0.05; Student’s t test; Figure 2D).
B
Training(0h) Test 1
250 -15min
200
100 50 Control
0
Test 2
24h
6d
Inject
150
Mean Latency (sec.)
Concentration (% of baseline)
A
-20 -10
0
DAB
10 20 30 Time (min)
40
50
60
500 400 300 200 100 0
Acq
Test 1
Vehicle + Vehicle
Test 2
DAB + Vehicle
DAB + 10 nmol L-lactate
C
D
Training(0h) Test 1 -15min
Training(0h)
Test 2
24h
Test -15min
6d
Acq
Test 1
Vehicle + Vehicle
Test 2 DAB + Vehicle
Mean Latency (sec.)
Mean Latency (sec.)
500 400 300 200 100 0
24h
Inject
Inject
500 400 300 200 100 0
Acq Vehicle
Test D-lactate
DAB + 100 nmol L-lactate
Figure 2. L-Lactate Rescues the DAB-Induced Memory Impairment Lactate concentration, latency scores, n and detailed statistic are reported in Table S2. (A) Dorsal hippocampal extracellular lactate in freely moving rats infused with either vehicle or DAB. Baseline was collected for 20 min before training (0 min, [) and continued for 50 min. Training resulted in a significant increase in lactate levels compared to baseline (*p < 0.05) that was completely blocked by DAB (# p < 0.05). Data are expressed as % of baseline ± SEM (mean of the first 2 samples set 100%). See also Figure S2. (B–D) Acquisition (Acq) and retention are expressed as mean latency ± SEM (in seconds, sec). (B and C) Hippocampal injection of DAB or vehicle in combination with 10 nmol (B, n = 7/group), 100 nmol L-lactate (C, n = 12/group) or vehicle were performed 15 min before training and memory was tested at 24 hr. 100 nmol but not 10 nmol of L-lactate rescued the memory impairment by DAB (Test 1). The effect persisted at 7 days after training (Test 2). (D) Hippocampal injections of D-lactate 15 min before training disrupted long-term memory (n = 7–8/group). *p < 0.05.
Hence, training results in a rapid glycogenolysis and consequent efflux of lactate in the hippocampus, which is critical for long-term memory formation. Blocking Glycogen Metabolism with DAB Disrupts Maintenance of Hippocampal LTP In Vivo and the Impairment Is Rescued by L-Lactate Long-term potentiation (LTP) is a leading cellular model for memory formation and is induced at CA1 synapses by IA (Whitlock et al., 2006). Accordingly, we next tested the effect of blocking glycogen metabolism on LTP in area CA1 of anesthetized young adult rats. After establishing a baseline of responses at Schaffer collateral-area CA1 (SC-CA1) synapses, LTP of SC-CA1 synapses was induced by a high-frequency tetanic stimulation protocol in the absence or presence of DAB delivered locally to the CA1 neuropil through the recording pipette. In the absence of DAB (the control group), robust LTP was induced (fEPSP slope: 133.85 ± 7.9% of baseline at 120 min posttetanus; Figure 3A, closed circles) as expected (Bozdagi et al., 2007). In contrast, in the presence of DAB,
tetanic stimulation produced a strong initial potentiation, but thereafter, potentiation decayed rapidly to baseline by 75 min in the DAB injected group (Figure 3A, open circles), indicating that DAB blocks the maintenance but not the induction of LTP. To rule out that the effect of DAB on LTP maintenance resulted from non-specific general metabolic demands of the high-frequency tetanic stimulation per se, we tested the effect of DAB on baseline neurotransmission in a group of rats in which LTP was blocked by an i.p. injection of the N-methylD-aspartate receptor (NMDAR) antagonist MK-801 prior to stimulation. Figure 3A shows that when LTP was blocked by MK-801, DAB had no effect on baseline over 120 min following stimulation, supporting the ides that DAB blocked LTP maintenance selectively. We verified that DAB alone, in the absence of tetanic stimulation, had no effects on basal synaptic transmission by examining parameters that reflect normal synaptic function, including the relationship between stimulus strength and the size of the postsynaptic response (input-output relationship; Figure 3B) or paired-pulse facilitation (PPF, Figure 3C). Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 813
B 200
DAB
0.5mV
100
10ms
50 0 -50
DAB
0
50 100 time (min)
150
3
control DAB
2 1 0 0
D
C 2 FP2/FP1
control
150
1.5 1 0.5 0
0.1 0.15 0.2 0.05 Stimulus intensity (mA)
E control DAB after LTP
250 IPI=50 ms
field EPSP (%)
field EPSP (%)
250
200
DAB
150 100 50
control L-lactate+DAB
250 field EPSP (%)
control DAB MK801+DAB
field EPSP slope (mV/ms)
A
200 150 100 50 L-lactate+DAB
control
DAB
0
-50
0 50 time (min)
100
0 -50
0 50 time (min)
100
Figure 3. DAB Impairs In Vivo Hippocampal LTP and the Impairment Is Rescued by L-Lactate (A) Average Field EPSP data recorded for 120 min posttetanus shows that DAB injection (bar) before high frequency stimulation (arrow) blocks LTP (p < 0.05 versus controls at 120 min, n = 4/group). LTP is abolished in animals receiving an intraperitoneal injection of the N-methyl-D-aspartate (NMDA) receptor antagonist MK-801 (3mg/kg) 30 min prior to tetanus (open triangles). Inset: Representative EPSP traces were recorded before and 120 min after (indicated by arrows) LTP induction. Left panel, control; right panel, in the presence of DAB. (B and C) No effect on the relationship between stimulus strength and the size of the postsynaptic response (input-output relationship, B, n = 4/group, p > 0.05) or paired-pulse facilitation (PPF, C, n = 4/group, p > 0.05). (D) DAB injected 30 min after high frequency stimulation did not affect synaptic potentiation (n = 4/group). (E) L-lactate reversed the blocking effect of DAB on LTP (n = 4/group). All data are expressed as mean values ± SEM.
To determine whether DAB has an effect on LTP once it is established, we injected DAB into CA1 30 min following LTP induction. There was no effect of DAB on synaptic potentiation once established, indicating that glycogen phosphorylase function is required during the induction or early phases of LTP (Figure 3D) (average percentage of baseline 60 min after LTP induction: control, 131.39 ± 11.01%, DAB, 139.86 ± 6.8%, p > 0.05, oneway ANOVA). To test if the effect of DAB was due to the astrocytic lactate production, we injected lactate together with DAB prior to LTP induction. Under these conditions, lactate abrogated the blocking effect of DAB on LTP. In the animals injected with lactate and DAB, the increase in fEPSP slope after LTP induction was not significantly different compared to that of control animals (control: 131.39 ± 11.01%, lactate+DAB: 127 ± 13%, of baseline at 60 min posttetanus, p > 0.05, one-way ANOVA; Figure 3E). Lactate Transport through MCT1 and MCT4 Plays an Essential Role in Long-Term Memory Formation We then investigated whether the effect we found with DAB was indeed due to the astrocytic release of lactate. Toward this end, we tested the effects of knocking down the hippocampal expression of the lactate transporters MCT1 and MCT4 using antisense oligodeoxynucleotides (ODNs). In the brain, the MCT1 is expressed mainly in endothelial cells of microvessels, ependymo814 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
cytes astrocytes and oligodendrocytes but not in neurons, whereas the MCT4 is expressed virtually exclusively in astrocytes (Pierre and Pellerin, 2005). Western blot analyses of synaptoneurosomal preparations (Villasana et al., 2006) showed that MCT4 and MCT2 are highly enriched in the synaptic fraction compared to the total lysate; MCT1, although not enriched compared to total lysate, is clearly present (Figure 4A). Hence, all these MCTs may be involved in lactate transport across the astrocytic-synaptic compartments. Rats were bilaterally injected into the hippocampus with MCT1-ODN or a relative scrambled control ODN containing the same base pair sequence but in a randomized order (scrambled-ODN, SC1-ODN) 1 hr before training and were euthanized 12 hr after training. Control rats received the same injections but did not undergo training. An ODN injection into the dorsal hippocampus targets and remains confined to this region (e.g., Taubenfeld et al., 2001a), as also shown by the Chicago blue diffusion shown in Figure S1. Dorsal hippocampal extracts obtained from these animals were analyzed for MCT1 expression levels using quantitative western blot analyses. A one-way ANOVA followed by Newman-Keuls post hoc test showed that the expression of MCT1 was significantly higher in trained compared to untrained rats injected with SC1-ODN (p < 0.05). This induction was completely and selectively blocked by MCT1-ODN (p < 0.05; Figure 4B). In fact, no change in MCT4 and MCT2 concentrations was found
across the same samples (MCT2, p > 0.05; MCT4, p > 0.05; one-way ANOVA followed by Newman-Keuls post hoc test; Figure 4B). Thus, MCT1, but not MCT4 or MCT2 expression is significantly upregulated upon training and MCT1-ODN selectively blocks this enhancement. The efficacy of the MCT1-ODN knock-down was further confirmed with quantitative western blot analyses of hippocampal extracts of trained rats obtained from the area targeted by the injection, that is 1.7 mm around the needle placement. As shown in Figure 4C, MCT1 level was significantly reduced by the MCT1-ODN compared to SC1-ODN (p < 0.05; Student’s t test), whereas no change in MCT2 and MCT4 were found (MCT2, p > 0.05; MCT4, p > 0.05; Student’s t test). The effect of MCT1-ODN-mediated knock down was then assessed on memory retention (Figures 4D and 4E). Hippocampal injection of MCT1-ODN, compared to SC1-ODN, 1 hr before training, had no effect on acquisition (p > 0.05; Student’s t test or one-way ANOVA) or short-term memory tested 1 hr after training (p > 0.05, Student’s t test; Figure 4D). However, it completely disrupted long-term memory 24 hr after training (Test 1). This disruption persisted 7 days after training (Test 2, p < 0.05, two-way repeated-measure ANOVA followed by Newman-Keuls post hoc test; Figure 4E), and memory did not reinstate following a reminder foot shock (Test 3, p < 0.05; Student’s t test). However, after retraining, the MCT1-ODNinjected rats had normal memory retention, indicating that their hippocampi were functioning normally. Thus, blocking the induction of MCT1 evoked by IA training disrupts long-term memory. Given the evidence for a role of lactate in long-term memory formation, we next asked whether exogenous L-lactate could rescue the MCT1-ODN-induced amnesia. Because a memoryenhancing effect has been reported for glucose (McNay and Gold, 2002; Messier, 2004) we also tested the potential rescuing effect of this energy substrate. Rats were injected with MCT1- or SC1-ODN 1 hr before training and again injected, 15 min before training, with either 100 nmol L-lactate, equicaloric 50 nmol glucose or vehicle (Figure 4F). Memory was tested 24 hr (Test 1) and 7 days (Test 2) after training. The acquisition latencies were similar in all groups (p > 0.05; one-way ANOVA). A twoway repeated-measure ANOVA followed by Newman-Keuls post hoc test indicated that MCT1-ODN with vehicle significantly impaired memory at both Test 1 and Test 2 (p > 0.05), thus confirming our previous results shown in Figure 4E. Furthermore, rats injected with SC1-ODN and glucose had a significant increase in memory retention at 24 hr after training compared to SC1-ODN-vehicle-injected rats (p < 0.05). Notably, L-lactate but not glucose significantly rescued memory retention caused by MCT1-ODN (p < 0.05). Because glucose enhanced memory retention, we tested whether higher concentrations of glucose could rescue/enhance the MCT1-induced amnesia. Rats were injected with MCT1ODN or SC1-ODN 1 hr before training and again injected, 15 min before training, with 150 nmol of glucose or vehicle (Figure 4G). A two-way repeated-measure ANOVA followed by Newman-Keuls post hoc test revealed that, as with previous results, MCT1-ODN significant impaired memory at both Test 1 and Test 2 (p < 0.05) and glucose significantly rescued memory
impairment at Test 1 (p < 0.05). However, the effect was transient, as memory was again impaired at Test 2 (p < 0.05). We next determined the effect of knocking down the expression of the astrocytic MCT4 on memory formation (Figure 5). Hippocampal injection of MCT4-ODN 1 hr before training significantly and selectively knocked down MCT4 concentration 12 hr after training, compared to scrambled-ODN (SC4-ODN), as revealed by quantitative western blot analyses of extracts obtained from 1.7 mm around the needle placement (Figure 5A, p < 0.05; Student’s t test). No change in MCT1 and MCT2 expression levels was found across samples (MCT1, p > 0.05; MCT2, p > 0.05; Student’s t test). MCT4-ODN injected 1 hr before training significantly impaired memory retention at both 24 hr (Test 1) and 7 days (Test 2) after training compared to SC4-ODN (p < 0.05, two-way repeated-measure ANOVA followed by Newman-Keuls post hoc test, Figure 5B). Notably, the memory of MCT4-ODN-injected rats was significantly rescued by 100 nmol of L-lactate but not of equicaloric, 50 nmol glucose injected 15 min before training (p < 0.05). There was no effect of treatment on acquisition (p > 0.05; one-way ANOVA). These data suggest that lactate transport through MCT1 and MCT4 is critical for long-term memory formation. Lactate Transport through MCT2 Is Also Required for Long-Term Memory Formation We next tested the effect of antisense-mediated knock-down of the neuronal lactate transporter MCT2 on memory formation (Figure 6). Rats received a bilateral hippocampal injection of MCT2-ODN or scrambled-ODN (SC2-ODN) 1 hr before training and were euthanized 12 hr after training. Quantitative western blots of the hippocampal area targeted by the injection (1.7 mm radius) showed that MCT2 concentrations were significantly reduced by MCT2-ODN compared to SC2-ODN (Figure 6A, p < 0.05; Student’s t test). No change in MCT1 and MCT4 concentrations was found across samples (MCT1, p > 0.05; MCT4, p > 0.05; Student’s t test), indicating that MCT2-ODN selectively downregulates the expression of MCT2. We then examined the effect of knocking down of MCT2 on memory retention. Rats were injected with MCT2-ODN or SC2ODN 1 hr before training. Memory retention was tested 1 or 24 hr after training. Acquisitions were similar in all groups (p > 0.05; Student’s t test). MCT2-ODN did not affect shortterm memory at 1 hr after training (p > 0.05; Student’s t test; Figure 6B). However, it significantly disrupted long-term memory retention at 24 hr after (Test 1) and 7 days after training (Test 2, p < 0.05, two-way repeated-measure ANOVA followed by Newman-Keuls post hoc test; Figure 6C). Memory did not recover following a reminder foot shock (Test 3, p < 0.05; Student’s t test), but the memory of MCT2-ODN-injected rats was normal after retraining (Figure 6C). Hence, blocking the function of MCT2 impairs long-term memory. To determine whether L-lactate and/or glucose rescues the MCT2-ODN-induced memory impairment, MCT2- or SC2-ODN were injected 1 hr before training and 100 nmol of L-lactate, 50 nmol of glucose or vehicle were injected 15 min before training. Memory was tested 1 (Test 1) and 7 days (Test 2) after training. A two-way repeated-measure ANOVA followed by Newman-Keuls post hoc test showed that, confirming our previous Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 815
A
B Euthanize
MCT1
-1h
12h
Inject
MCT2
% of Untrained - SC1-ODN
S
T
MCT1 Training(0h)
140 120 100 80 60 40 20 0
SC1-ODN MCT1-ODN
Untrained Trained
Untrained Trained S M S M
MCT2 % of Untrained - SC1-ODN
MCT1
GFAP
MCT2
PSD95
MCT4 Actin
Actin
140 120 100 80 60 40 20 0
Untrained Trained
S; SC1-ODN M; MCT1-ODN
C
D
Untrained Trained
E
MCT4 Actin
Reminder shock
% of - SC1-ODN
-1h 24h
SC1ODN
MCT1ODN
SC1ODN
MCT1ODN
Test 2
24h
24h
6d
Retraining
Test 3 24h
Test 4 24h
Inject
Inject
MCT4
SC1- MCT1ODN ODN
Test 1
1h
-1h
MCT1 120 100 80 60 40 20 0
Test
Mean Latency (sec.)
MCT2
120 100 80 60 40 20 0
Training(0h)
Mean Latency (sec.)
MCT1
% of - SC1-ODN
MCT2
SC1-ODN MCT1-ODN
% of - SC1-ODN
140 120 100 80 60 40 20 0
Training(0h)
Trained
120 100 80 60 40 20 0
MCT4 % of Untrained - SC1-ODN
MCT4
500 400 300 200 100 0 Acq
Test
SC1-ODN
500 400 300 200 100 0
Acq
Test 1
MCT1-ODN
F
Test 2
SC1-ODN
Test 3
Test 4
MCT1-ODN
G Training(0h)
Training(0h)
-1h
-15min
Inject
Inject
Test 2
24h
Test 1 -1h -15min
6d
500 400 300 200 100 0 Acq
Test 2
24h
6d
Inject Inject
Mean latency (sec.)
Mean latency (sec.)
Test 1
Test 1
Test 2
SC1-ODN + Vehicle
SC1-ODN + L-lactate
SC1-ODN + Glucose
MCT1-ODN + Vehicle
MCT1-ODN + L-lactate
MCT1-ODN + Glucose
500 400 300 200 100 0 Acq
Test 1
SC1-ODN + Vehicle
Test 2 MCT1-ODN + Vehicle
MCT1-ODN + High glucose
Figure 4. Disruption of MCT1 expression impairs long-term memory Mean %, n, latency and detailed statistic are reported in Table S3. (A) MCTs expression performed on total (T) or synaptoneurosomal (S) extracts of dorsal hippocampus. Postsynaptic density-95 (PSD95) is used to show synaptoneurosomal enrichment. Glial fibrillary acidic protein (GFAP) is used as an astrocytic marker. (B) Examples and densitometric quantitative western blot analysis of MCT1, MCT2, and MCT4 carried out in dorsal hippocampal extracts from trained and untrained rats (n = 6/group) injected with MCT1- or SC1-ODN and euthanized 12 hr after training. MCT1, but not MCT2 or MCT4, was significantly induced by training. This induction was selectively blocked by MCT1-ODN injection. Data are expressed as mean percentage ± SEM of the untrained-SC1-ODN (100%) mean values. All MCTs values were normalized to those of actin.
816 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
(p > 0.05; Figure 6D). These results confirm that glucose does not rescue the memory impairment due to astrocytic export of lactate and indicates that lactate import through MCT2 is essential for long-term memory formation.
A
12h
120 100 80 60 40 20 0
120 100 80 60 40 20 0
Trained SC4-ODN MCT4-ODN
MCT4 MCT1 MCT2 Actin
B
SC4- MCT4ODN ODN MCT2
MCT1
SC4- MCT4ODN ODN
% of - SC4-ODN
-1h Inject
MCT4
% of - SC4-ODN
Euthanize
% of - SC4-ODN
Training(0h)
120 100 80 60 40 20 0
SC4- MCT4ODN ODN
Training(0h) Test 1 -1h -15min
Test 2
24h
6d
Mean latency (sec.)
Inject Inject
500 400 300 200 100 0
Acq
Test 1
Test 2
SC4-ODN + Vehicle
MCT4-ODN + Vehicle
MCT4-ODN + L-lactate
MCT4-ODN + Glucose
Figure 5. Disruption of MCT4 Expression Also Impairs Long-Term Memory Mean %, n, latency, and detailed statistic are reported in Table S4. (A) Examples and densitometric analysis of MCT4, MCT1, and MCT2 quantitative western blots of extracts from dorsal hippocampal punches taken around the needle placement from trained rats that received hippocampal MCT4- or SC4-ODN injections and were euthanized 12 hr after training (n = 4/group). Data are expressed as mean percentage ± SEM of trained-SC4ODN (100%) mean values. All MCTs values were normalized to those of actin. (B) Memory acquisition (acq) and retention are expressed as mean latency ± SEM (in seconds, sec). Hippocampal injection of MCT4-ODN disrupted longterm memory and L-lactate but not glucose, rescued it. MCT4- or SC4-ODN were injected 1 hr before training. L-lactate, glucose or vehicle (PBS) were injected 15 min before training. Rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). (n = 10-12/group). *p < 0.05.
data (Figure 6C), MCT2-ODN with vehicle significantly impaired memory retention at both Test 1 and Test 2 (p > 0.05), and both L-lactate and glucose failed to rescue this amnesia
Training-Dependent Induction of Arc, pCREB, and Pcofilin Are Completely Blocked by DAB and Significantly Rescued by L-lactate We tested whether disrupting astrocytic glycogen mobilization influences neuronal hippocampal molecular changes known to be required for memory consolidation in a variety of tasks. Toward this end, we investigated the induction of Arc expression, ser 133 pCREB, which is known to be critical for CREB-dependent gene expression regulation (Alberini, 2009; Mayr and Montminy, 2001) and the phosphorylation of cofilin (pcofilin). The change in expression levels of these markers following training, as well as the effect of hippocampal DAB and L-lactate administrations were determined. Rats were either trained in IA or remained in the homecage (untrained controls). Untrained controls, or rats 15 min before training, received a bilateral injection into the hippocampus of either 300 pmol DAB or vehicle, or DAB and 100 nmol L-lactate. All animals were euthanized either 30 min or 20 hr after training. Thirty minutes after training was used to examine Arc expression, since Arc is known to be rapidly induced (Bramham et al., 2008). Twenty hours after training was used to determine pCREB and pcofilin changes, since they are known to accompany the formation of long-term memory and plasticity for several hours or days (Alberini, 2009; Fedulov et al., 2007). The homecage control rats that received similar treatments were euthanized at paired time points (Figure 7). Quantitative western blot analyses were used to determine Arc, pCREB, CREB, pcofilin, and cofilin concentrations. Quantitative western blot analyses indicated that, as expected, rats that received vehicle injection and underwent IA training (vehicletrained) had a significant increase in hippocampal Arc compared with untrained vehicle-injected rats (p < 0.05, one-way ANOVA followed by Newman-Keuls post hoc test, Figures 7A and 7B). This increase was completely blocked by DAB (p < 0.05). L-lactate significantly rescued this DAB effect compared to vehicle (p < 0.05). Similarly, at 20 hr after training, pCREB and pcofilin were significantly increased in the hippocampus of vehicle-injected, trained rats compared to vehicle-injected untrained (p < 0.05, one-way ANOVA followed by Newman-Keuls post hoc test,
(C) Examples and densitometric analysis of MCT1, MCT2, and MCT4 quantitative western blots of extracts from dorsal hippocampal punches taken around the needle placement from trained rats that received hippocampal MCT1- or SC1-ODN injections and were euthanized 12 hr after training (n = 4/group). Data are expressed as mean percentage ± SEM of trained-SC1-ODN (100%) mean values. All MCTs values were normalized to those of actin. (D–G) Memory acquisition (acq) and retention are expressed as mean latency ± SEM (in seconds, sec). (D) Hippocampal injections of MCT1-ODN 1 hr before training did not affect short-term memory (n = 7/group). (E) Hippocampal injection of MCT1-ODN disrupted long-term memory. MCT1- or SC1-ODN were injected 1 hr before training and rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). The memory disruption persisted at Test 2, and memory did not recover following a reminder shock (Test 3). MCT1-ODN amnesic rats showed normal retention after retraining (Test 4) (n = 10–11/group). (F) L-lactate but not glucose rescued the memory impairment induced by blocking MCT1 expression (n = 7-13/group). MCT1- or SC1-ODN were injected 1 hr before training. L-lactate, glucose or vehicle (PBS) were injected 15 min before training. Rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). (G) High concentration of glucose transiently rescued the MCT1-ODN-induced memory impairment (n = 8-11/group). MCT1- or SC1-ODN were injected 1 hr before training. 150 nmol of glucose or vehicle (PBS) were injected 15 min before training. Rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). *p < 0.05.
Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 817
A
B MCT2
-1h
12h
Inject
120 100 80 60 40 20 0
Trained
-1h
MCT2ODN
% of - SC2-ODN
MCT4 Actin
C
Reminder shock
Training(0h)
Test 1 -1h
24h
120 100 80 60 40 20 0
Test 2 6d
SC2ODN
120 100 80 60 40 20 0
Retraining
Test 3
24h
MCT2ODN
MCT4 % of - SC2-ODN
MCT1
MCT2
24h
24h
SC2ODN
Test 1
Test 2
SC2-ODN
Test 3
MCT2ODN
D
24h
Test 4
MCT2-ODN
Acq
Test MCT2-ODN
Training(0h) Test 1
Mean latency (sec.)
Mean Latency (sec.)
Acq
500 400 300 200 100 0
SC2-ODN
Test 4
Inject
500 400 300 200 100 0
1h
Inject
SC2ODN
SC2-ODN MCT2-ODN
MCT1
Test
Training(0h)
Mean Latency (sec.)
Euthanize
% of - SC2-ODN
Training(0h)
-1h
-15min
Inject
Inject
500 400 300 200 100 0
Acq
Test 2
24h
6d
Test 1
SC2-ODN + Vehicle MCT2-ODN + L-lactate
Test 2 MCT2-ODN + Vehicle MCT2-ODN + Glucose
Figure 6. Disruption of MCT2 Expression Impairs Long-Term Memory Mean %, n, latency, and detailed statistics are reported in Table S5. (A) Examples and densitometric analysis of MCT2, MCT1, and MCT4 quantitative western blots of extracts from dorsal hippocampal punches taken around the needle placement from trained rats that received hippocampal MCT2- or SC2-ODN injections and were euthanized 12 hr after training (n = 4/group). Data are expressed as mean percentage ± SEM of trained-SC2-ODN (100%) mean values. All MCTs values were normalized to those of actin. (B–D) Acquisition (acq) and retention are expressed as mean latency ± SEM (in seconds, sec). (B) Hippocampal injections of MCT2-ODN 1 hr before training did not affect short-term memory tested 1 hr later (n = 8/group). (C) Hippocampal injections of MCT2-ODN disrupted long-term memory. MCT2- or SC2-ODN were injected 1 hr before training. Rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). The memory disruption persisted at Test 2, and memory did not recover following a reminder foot shock (Test 3). The amnesic rats that received the MCT2-ODN showed normal retention after retraining (Test 4) (n = 8/group). (D) Neither L-lactate nor glucose rescued the memory impairment induced by MCT2 disruption (n = 6-8/group). MCT2- or SC2-ODN were injected 1 hr before training. L-lactate, glucose or vehicle (PBS) were injected 15 min before training. Rats were tested 24 hr after training (Test 1) and 6 days later (Test 2). *p < 0.05.
Figures 7A and 7C–7F), whereas the total concentration of CREB and cofilin, respectively, remained unchanged (p > 0.05). Both increases were completely blocked by DAB (p < 0.05). L-lactate injection rescued the effect of DAB and restored a high pCREB and pcofilin expression, which was comparable to that of trained rats that received vehicle injection (p > 0.05). These results were further confirmed by MCT1 knock down treatments. Bilateral hippocampal MCT1-ODNs injection significantly blocked pCREB induction 12 hr after training, compared to SC1-ODN (p < 0.05, one-way ANOVA followed by Newman-Keuls post hoc test, Figure S3). Together, these data indicate that both DAB and downregulation of MCT1 expression block molecular changes induced by learning and required for long-term memory formation as well as relative synaptic structural changes, and that L-lactate is sufficient to fully rescue the effect. 818 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
DISCUSSION Little is known about the contribution of astrocytes to cognitive functions. The results of this study lead to the conclusion that astrocyte-neuron lactate transport plays a critical role in longterm memory formation. The Critical Role of Glycogen Metabolism-Derived Lactate in Long-Term Plasticity and Memory Formation Our results on the DAB-dependent hippocampal lactate release after training, DAB- and isofagomine-induced amnesia rescued by exogenous L-lactate and D-lactate-induced amnesia indicate that learning leads to astrocytic glycogenolysis and lactate release that is critical for long-term memory formation, but dispensable for learning and short-term memory. The DABdependent effects are astrocytic and not neuronal because
A
Untrained
B
Trained
Arc
V/V D/V V/V D/V D/L
Inject
Arc
-15min % of untrained - Vehicle
Actin pCREB CREB pcofilin cofilin Actin
Inject
% of untrained - Vehicle
-15min
200 150 100
Training(0h)
30min
250 200 150 100 50 0 Untrained
D
pCREB
Trained
CREB
Euthanize
Inject -15min
20h % of untrained - Vehicle
C
Euthanize
Training(0h)
Training(0h) 20h
Euthanize
Figure 7. Training-Induced Increase of Arc, pCREB, and Pcofilin Are Completely Blocked by DAB and Significantly Rescued by L-Lactate Mean %, n, and detailed statistic are reported in Table S6. (A–F) Examples (A) and densitometric western blot analysis of Arc (B), pCREB (C), CREB (D), pcofilin (E), and cofilin (F) performed on dorsal hippocampal extracts from trained and untrained rats injected 15 min before training with vehicle, DAB+vehicle or DAB+L-lactate and euthanized 30 min (for Arc) or 20 hr (for all other markers) after training. See also Figure S3. Arc (B), pCREB (C), and pcofilin (E) expression were significantly increased after training. This increase was completely blocked by DAB and rescued by L-lactate. There is no change in expression of CREB (D) and cofilin (F) across samples (n = 4–7/group). Data are expressed as mean percentage ± SEM of untrained, vehicle-injected control (100%) mean values. All proteins values were normalized to those of actin. *p < 0.05.
150 100
% of untrained - Vehicle
% of untrained - Vehicle
aspartate. Glucose also had an effect in rescuing DAB-induced memory impair50 ment, but it was transient and found 0 0 only immediately after the injection but Untrained Trained Untrained Trained not later (Gibbs et al., 2006). The authors concluded that glycogenolysis is imporF E tant to provide neurons with glutamine pcofilin cofilin as the precursor for neuronal glutamate Euthanize Inject Training(0h) Euthanize Inject Training(0h) and GABA and that glycogen is a preferred glutamate precursor during -15min 20h -15min 20h learning in chick (Gibbs et al., 2007). Our results extend these findings to an 150 200 adult rat hippocampal memory and show that the hippocampus is an anatomical 150 100 site where glycogenolysis, around the 100 time of training but not 24 hr later, is 50 required for long-term memory formation. 50 Furthermore, we show (for the first time 0 0 to our knowledge) that learning results Untrained Trained Untrained Trained in glycogenolysis-dependent lactate Vehicle + Vehicle (V/V) DAB + Vehicle (D/V) DAB + L-lactate (D/L) release in the hippocampus critical for memory formation. In line with our data, sensory and cognitive stimuli in humans and animals produce an increase in brain although glycogen can be found in fetal, neonatal, and postnatal lactate (Fray et al., 1996; Urrila et al., 2004). The DAB-induced juvenile neurons up to one month of age (McKenna et al., 2006), it amnesia did not reverse after a foot-shock reminder, suggesting is absent from neurons of the hippocampus and neocortex of that disrupting glycogen metabolism around the time of training adult brain and only found occasionally in large neurons in the persistently impairs memory consolidation or storage. brainstem or in the peripheral nervous system (Magistretti, Our results do not exclude that, in addition to L-lactate, other 2008; Sotelo and Palay, 1968). substrates that crosstalk with or are downstream of lactate An amnestic effect of DAB on memory retention at 30 to 70 min production may rescue or enhance the impaired memory caused after training has been previously reported by Gibbs et al. (2006) by blocking glycogenolysis. For example, glutamine and GABA in a day-old chick bead-discrimination learning paradigm. This may rescue the DAB-induced amnesia, as one of the end-results amnesia was rescued by lactate, glutamine or acetate plus of training-induced lactate production may be indeed the
50
Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 819
increase of glutamine and GABA. Glucose uptake may also result in lactate increase that may contribute to memory consolidation or enhancement. Our data, confirming previous studies (McNay and Gold, 2002; Messier, 2004), show, in fact, that glucose can enhance memory. However, glucose is much less efficient in rescuing the amnesia caused by DAB and its effect is transient, indicating that the end mechanisms of lactate or glucose might be different or at least have different kinetics. Dissecting the effects of lactate downstream to its uptake into neurons shall be the focus of future investigations. We also found similar outcomes when we investigated in vivo hippocampal LTP and conclude that maintenance but not induction of Schaffer collateral-CA1 LTP requires glycogenolysis and is mediated by lactate. Although our results indicate that glycogenolysis is required for LTP and not baseline activity, they do not inform about the nature of the targeted mechanisms and further experiments shell shed light on this issue. These findings add to the evidence provided by a number of other studies that investigated other types of astrocytic mechanisms, including release of D-Serine (Henneberger et al., 2010; Panatier et al., 2006; Zhang et al., 2008), ATP-dependent vesicular release of glutamate (Jourdain et al., 2007) and astroglial gap junctionsmediated delivery of energetic metabolites from blood vessels to distal neurons (Rouach et al., 2008). The Critical Role of Lactate Transport between Astrocytes and Neurons in Memory Formation Disrupting the expression of both the astrocytic and neuronal lactate transporters MCT4 and MCT2, which are highly enriched at synaptic sites, result in amnesia, which is rescued by L-lactate after MCT4 but not MCT2 expression disruption. These results together with the evidence of a necessary role for glycogenolysis-dependent lactate release in the hippocampus strongly support the conclusion that an astrocytic-neuronal lactate transport critically mediates long-term memory formation. Furthermore, learning leads to a significant increase in MCT1 expression, whose disruption completely blocks long-term memory. MCT1 is expressed in endothelial cells of microvessels ependymocytes, oligodendrocytes and astrocytes and, interestingly, was found in the synaptoneurosomal fraction, which consists of resealed presynaptic sac (synaptosome) fused to a larger resealed postsynaptic sac (neurosome) as well as resealed astrocytic processes, as confirmed by the presence of GFAP (Villasana et al., 2006; Williams et al., 2009). GFAP has been generally believed to be absent from the very fine perysynaptic astrocytic compartments; hence, a possible explanation for its detection in synaptoneurosomal fractions is that communications between astrocytes and synapses may not be restricted to the fine astrocytic protrusions but also take place in larger, GFAP-positive processes that are sealed to synaptoneurosomal sacs. In addition, it is possible that MCT1 expressed in the other cells also play a critical role in long-term memory formation. Interestingly, rapid regulation of MCT1 expression has been documented in cell lines incubated with lactate (Hashimoto et al., 2007). Further experiments should shed light on whether the learning-related induction of MCT1 is perisynaptic and how MCT1expressed in different cells and compartments contribute to long-term memory formation. L-lactate, but not 820 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
equicaloric concentrations of glucose, rescues these MCT1- or MCT4-knock-down-dependent amnesia, indicating that glucose and lactate are not functionally interchangeable. However, as discussed above, glucose enhances memory retention in control animals, and higher concentrations of glucose can rescue the MCT1-ODN induced amnesia, but only transiently. Possible explanations for these results are: first, that the glucose-mediated memory enhancement might occur via astrocytic glucose uptake and subsequent glycolytic processing leading to lactate release and that an overload of lactate via high glucose concentrations may overcome the temporallylimited effect of the antisense treatments. Second, glucose entering the neuron directly may produce outcomes overlapping with those resulting from the lactate shuttling (e.g., glutamine and GABA increase). Nevertheless, our data point to the conclusion that glucose partially and less efficiently substitutes for lactate. Thus, we conclude that an astrocyte-neuron lactate transport critically supports neuronal functions, in agreement with what proposed by Pellerin and Magistretti (1994). The astrocyte-neuron lactate transport hypothesis does not contradict the classical belief that glucose is a fundamental metabolic substrate in mature brain (Clarke and Sokoloff, 1994); it suggests that both glucose and lactate have critical roles for adult brain functioning. On one hand, lactate may represent an essential substrate mediating the astrocyte-neuron metabolic coupling during activity-related high-energy demands; on the other, glucose uptake into neurons also occurs as suggested by the expression of glucose transporters of the glut 3 type (Magistretti, 2008). However, recent data by Herrero-Mendez et al. (2009) showed that cultured cortical neurons cannot sustain a high glycolytic rate as glucose is mainly used to maintain their antioxidant status, and studies in the skeletal muscle show that lactate is formed and utilized continuously under fully aerobic conditions and mediates, in addition to energy supply, important functions such as cellular signaling and redox regulation. Thus, in agreement with Brooks (2009), we propose that glycolytic and oxidative pathways should be viewed as linked, as opposed to alternative processes also with the astrocyte-neuron network. Further studies shall shed light on the underlying mechanisms. In agreement with this idea, there is also evidence of compartmentalized energy metabolism between astrocytes and neurons, especially during increased neuronal activity (Kasischke et al., 2004) and of lactate being a crucial aerobic energy substrate that enables neurons to endure activation (Schurr et al., 1999). The Critical Role of Astrocytes in Learning-Induced Molecular and Synaptic Changes We also found that pCREB and Arc expression, as well as pcofilin, which represent essential mechanisms underlying long-term synaptic plasticity and memory formation and their related synaptic structural changes (Alberini, 2009; Bramham et al., 2008) also depend on an intact glycogen metabolism and astrocyte-neuron lactate transport. Both pCREB and Arc are mostly induced in neurons following training or long-term plasticity (Impey et al., 1998; Leutgeb et al., 2005; Taubenfeld et al., 2001b; Vazdarjanova et al., 2006). The complete disruption
of both learning-dependent pCREB and Arc by DAB or MCT1ODN suggests that there is a critical functional link between the astrocytic glycogenolysis and lactate transport and the neuronal gene expression regulation underlying long-term memory formation. The lactate-mediated rescuing indicates that this functional link is mediated by lactate. Hence, it is tempting to speculate that the neurometabolic coupling mediated by lactate supplies the energy needed by activated neurons for their local protein synthesis, degradation and activation of signaling pathways that lead to the regulation of the gene expression cascade that underlies long-term memory. Because lactate has been suggested to play an important role in cellular and organelle redox balance (Brooks, 2009), we may also hypothesize that lactate not only represents an energy source, but also mediates coordinated astrocyte-neuron cell-cell and intracellular responses important for cell signaling (Gordon et al., 2008) and regulation of gene expression. Finally, our data showing learning-induced increased phosphorylation of cofilin is in agreement with and extends previous studies reporting that unsupervised context learning in rats leads to a 30% increase in spine numbers in the hippocampal CA1 region. These spines express pcofilin and have a larger size compared to the average spines of control naive rats (Fedulov et al., 2007). Similarly, pcofilin has also been found upregulated following LTP in spines that have larger postsynaptic densities (PSDs) (Chen et al., 2007). Hence, pcofilin represents a correlative marker of spine morphological changes induced during long-term plasticity and memory. We also showed that the increase in pcofilin depends on an intact glycogen metabolism and lactate transport, strengthening the hypothesis that the astrocyte-neuron lactate transport is critical to enable the synaptic molecular and spine morphological changes underlying long-term memory formation. In conclusion, we propose that astrocyte-neuron lactate transport is essential for long-term synaptic plasticity, long-term memory, and its underlying molecular and synaptic changes. These results may have important implications for memory disorders and pathologies carrying memory or cognitive deficits in general, including neurodegenerative conditions such as Alzheimer’s disease, aging, and dementia. EXPERIMENTAL PROCEDURES Animals All protocols involving the use of animals complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Mt. Sinai School of Medicine Animal Care Committees. See Extended Experimental Procedures for details. Cannulae Implants Bilateral hippocampal surgeries were performed as described by Taubenfeld et al. (2001a). See supplemental experimental procedures for details.
were dissolved in phosphate-buffered saline (PBS [pH 7.4]). Oligodeoxynucleotides were reverse phase cartridge-purified and purchased from Gene Link (Hawthorne, NY). All injections were performed at 1 ml of volume per hippocampal side. See Extended Experimental Procedures for details. Microdialysis Microdialysis was performed as described in Rex et al. (2009), using a flowrate of 3 ml/min and samples were collected every 10 min. Samples were subsequently analyzed using the AbCam Lactate Fluorescence Assay Kit. See Extended Experimental Procedures for details. Synaptoneurosomal Preparation Synaptoneurosomal preparation was carried out as described in Villasana et al. (2006). See Extended Experimental Procedures for details. Western Blot Analysis Primary antibodies [rabbit anti-MCT1, rabbit anti-MCT2, rabbit anti-MCT4, rabbit anti-pCREB and rabbit anti-CREB, rabbit anti-cofilin, mouse antiPSD95 and rabbit anti-GFAP (Millipore, Billerica, MA), rabbit anti-pcofilin (Abcam, Cambridge, MA), and rabbit anti-Arc (Synaptic System, Gottingen, Germany)] were used. Actin (Santa Cruz Biotechnology, Santa Cruz, CA) was used for loading normalization. Western blot were performed as described previously (Taubenfeld et al., 2001b). See Extended Experimental Procedures for details. Electrophysiology Electrophysiological experiments were carried out as previously described (Bozdagi et al., 2007). DAB (300 mM), lactate (100 mM) or lactate+DAB (23 concentration from each) were applied. See Extended Experimental Procedures for details. Statistical Analysis Data were analyzed with one- or two-way repeated-measure ANOVA followed by Newman–Keuls or Bonferroni post hoc test. When two groups were compared, Student’s t test was used. SUPPLEMENTAL INFORMATION Supplemental Information includes Extended Experimental Procedures, three figures, and seven tables and can be found with this article online at doi:10. 1016/j.cell.2011.02.018. ACKNOWLEDGMENTS This work was supported by grants R01-MH065635 and R01-MH074736 to C.M.A., Department of Veterans Affairs Merit Award to R.H.W. and T32MH087004 to S.A.S. We thank Rick Koch, Dhananjay Bambah-Mukku, Dillon Chen, Gabriella Pollonini, and Virginia Gao for technical support; and Reginald Miller and the Center for Comparative Medicine and Surgery facility at Mount Sinai School of Medicine for technical support. P.J.M. is the recipient of the Asterion Chair at Ecole Polytechnique Fe´de´rale de Lausanne (EPFL). Received: March 8, 2010 Revised: September 21, 2010 Accepted: February 8, 2011 Published: March 3, 2011 REFERENCES
Inhibitory Avoidance Inhibitory avoidance (IA) was carried out as described previously (Taubenfeld et al., 2001a). See Extended Experimental Procedures for details.
Alberini, C.M. (2009). Transcription factors in long-term memory and synaptic plasticity. Physiol. Rev. 89, 121–145.
Drug and Oligodeoxynucleotide Injections DAB, sodium L-lactate, sodium D-lactate, D-glucose (Sigma-Aldrich, St. Louis, MO), and isofagomine (Santa Cruz Biotechnology, Santa Cruz, CA)
Bozdagi, O., Nagy, V., Kwei, K.T., and Huntley, G.W. (2007). In vivo roles for matrix metalloproteinase-9 in mature hippocampal synaptic physiology and plasticity. J. Neurophysiol. 98, 334–344.
Bamburg, J.R., McGough, A., and Ono, S. (1999). Putting a new twist on actin: ADF/cofilins modulate actin dynamics. Trends Cell Biol. 9, 364–370.
Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 821
Bramham, C.R., Worley, P.F., Moore, M.J., and Guzowski, J.F. (2008). The immediate early gene arc/arg3.1: regulation, mechanisms, and function. J. Neurosci. 28, 11760–11767.
is controlled by continuous degradation of a key glycolytic enzyme by APC/ C-Cdh1. Nat. Cell Biol. 11, 747–752.
Brooks, G.A. (2009). Cell-cell and intracellular lactate shuttles. J. Physiol. 587, 5591–5600.
Impey, S., Smith, D.M., Obrietan, K., Donahue, R., Wade, C., and Storm, D.R. (1998). Stimulation of cAMP response element (CRE)-mediated transcription during contextual learning. Nat. Neurosci. 1, 595–601.
Brown, A.M., Baltan Tekkok, S., and Ransom, B.R. (2004). Energy transfer from astrocytes to axons: the role of CNS glycogen. Neurochem. Int. 45, 529–536.
Jahanshahi, M., Sadeghi, Y., Hosseini, A., Naghdi, N., and Marjani, A. (2008a). The effect of spatial learning on the number of astrocytes in the CA3 subfield of the rat hippocampus. Singapore Med. J. 49, 388–391.
Chen, L.Y., Rex, C.S., Casale, M.S., Gall, C.M., and Lynch, G. (2007). Changes in synaptic morphology accompany actin signaling during LTP. J. Neurosci. 27, 5363–5372.
Jahanshahi, M., Sadeghi, Y., Hosseini, A., Naghdi, N., and Piriaie, A. (2008b). Working memory learning method and astrocytes number of different subfields of rat’s hippocampus. Am. J. Anim. Vet. Sci. 3, 28–31.
Clarke, D.D., and Sokoloff, L. (1994). Circulating and energy metabolism of the brain. In Basic Neurosciemistry, G.J. Siegle, B.W. Agranoff, R.W. Albers, and P.B. Molinoff, eds. (New York: Raven Press), pp. 645–680.
Jourdain, P., Bergersen, L.H., Bhaukaurally, K., Bezzi, P., Santello, M., Domercq, M., Matute, C., Tonello, F., Gundersen, V., and Volterra, A. (2007). Glutamate exocytosis from astrocytes controls synaptic strength. Nat. Neurosci. 10, 331–339.
DiNuzzo, M., Mangia, S., Maraviglia, B., and Giove, F. (2010). Glycogenolysis in astrocytes supports blood-borne glucose channeling not glycogen-derived lactate shuttling to neurons: evidence from mathematical modeling. J. Cereb. Blood Flow Metab. 30, 1895–1904. Dringen, R., Gebhardt, R., and Hamprecht, B. (1993). Glycogen in astrocytes: possible function as lactate supply for neighboring cells. Brain Res. 623, 208–214. Dudai, Y. (2004). The neurobiology of consolidations, or, how stable is the engram? Annu. Rev. Psychol. 55, 51–86. Fedulov, V., Rex, C.X., Simmons, D.A., Palmer, L., Gall, C.M., and Lynch, G. (2007). Evidence That Long-Term Potentiation Occurs within Individual Hippocampal Synapses during Learning. J. Neurosci. 27, 8031–8039. Fischer, A., Sananbenesi, F., Schrick, C., Spiess, J., and Radulovic, J. (2004). Distinct roles of hippocampal de novo protein synthesis and actin rearrangement in extinction of contextual fear. J. Neurosci. 24, 1962–1966. Fray, A.E., Forsyth, R.J., Boutelle, M.G., and Fillenz, M. (1996). The mechanisms controlling physiologically stimulated changes in rat brain glucose and lactate: a microdialysis study. J. Physiol. 496, 49–57. Ghosh, A., Cheung, Y.Y., Mansfield, B.C., and Chou, J.Y. (2005). Brain contains a functional glucose-6-phosphatase complex capable of endogenous glucose production. J. Biol. Chem. 280, 11114–11119. Gibbs, M.E., Anderson, D.G., and Hertz, L. (2006). Inhibition of glycogenolysis in astrocytes interrupts memory consolidation in young chickens. Glia 54, 214–222.
Kandel, E.R. (2001). The molecular biology of memory storage: a dialogue between genes and synapses. Science 294, 1030–1038. Kasischke, K.A., Vishwasrao, H.D., Fisher, P.J., Zipfel, W.R., and Webb, W.W. (2004). Neural activity triggers neuronal oxidative metabolism followed by astrocytic glycolysis. Science 305, 99–103. Kettenmann, H., and Ransom, B.R. (2005). Neuroglia, Second Edition (New York: Oxford University Press). Klann, E., and Sweatt, J.D. (2008). Altered protein synthesis is a trigger for long-term memory formation. Neurobiol. Learn. Mem. 89, 247–259. Leutgeb, J.K., Frey, J.U., and Behnisch, T. (2005). Single cell analysis of activity-dependent cyclic AMP-responsive element-binding protein phosphorylation during long-lasting long-term potentiation in area CA1 of mature rat hippocampal-organotypic cultures. Neuroscience 131, 601–610. Magistretti, P.J. (2006). Neuron-glia metabolic coupling and plasticity. J. Exp. Biol. 209, 2304–2311. Magistretti, P.J. (2008). Brain energy metabolism. In Fundamental Neuroscience, 3rd, L.R. Squire, D. Berg, F.E. Bloom, S. du Lac, A. Ghosh, and N.C. Spitzer, eds. (San Diego: Academic Press), pp. 271–293. Magistretti, P.J., and Morrison, J.H. (1988). Noradrenaline- and vasoactive intestinal peptide-containing neuronal systems in neocortex: functional convergence with contrasting morphology. Neuroscience 24, 367–378. Magistretti, P.J., Pellerin, L., Rothman, D.L., and Shulman, R.G. (1999). Energy on demand. Science 283, 496–497.
Gibbs, M.E., Lloyd, H.G., Santa, T., and Hertz, L. (2007). Glycogen is a preferred glutamate precursor during learning in 1-day-old chick: biochemical and behavioral evidence. J. Neurosci. Res. 85, 3326–3333.
Mantzura, L., Joelsa, G., and Lamprecht, R. (2009). Actin polymerization in lateral amygdala is essential for fear memory formation. Neurobiol. Learn. Mem. 91, 85–88.
Gordon, G.R., Choi, H.B., Rungta, R.L., Ellis-Davies, G.C., and MacVicar, B.A. (2008). Brain metabolism dictates the polarity of astrocyte control over arterioles. Nature 456, 745–749.
Mayr, B., and Montminy, M. (2001). Transcriptional regulation by the phosphorylation-dependent factor CREB. Nat. Rev. Mol. Cell Biol. 2, 599–609.
Halassa, M.M., Florian, C., Fellin, T., Munoz, J.R., Lee, S.Y., Abel, T., Haydon, P.G., and Frank, M.G. (2009). Astrocytic modulation of sleep homeostasis and cognitive consequences of sleep loss. Neuron 61, 213–219.
McKenna, M.C., Rolf Gruetter, R., Sonnewald, U., Waagepetersen, H.S., and Schousboe, A. (2006). Energy Metabolism of the Brain. In Basic Neurochemistry, G. Siegel, R.W. Albers, S. Brady, and D.L. Price, eds. (London: Elsevier), pp. 531–557.
Halassa, M.M., and Haydon, P.G. (2010). Integrated brain circuits: astrocytic networks modulate neuronal activity and behavior. Annu. Rev. Physiol. 72, 335–355.
McNay, E.C., and Gold, P.E. (2002). Food for thought: fluctuations in brain extracellular glucose provide insight into the mechanisms of memory modulation. Behav. Cogn. Neurosci. Rev. 1, 264–280.
Hamprecht, B., Verleysdonk, S., and Wiesinger, H. (2005). Enzymes of carbohydrate and energy metabolism. In Neuroglia, Second Edition, H. Kettenmann and B.R. Ransom, eds. (New York: Oxford University Press), pp. 202–215.
Messier, C. (2004). Glucose improvement of memory: a review. Eur. J. Pharmacol. 490, 33–57.
Hashimoto, T., Hussien, R., Oommen, S., Gohil, K., and Brooks, G.A. (2007). Lactate sensitive transcription factor network in L6 cells: activation of MCT1 and mitochondrial biogenesis. FASEB J. 21, 2602–2612. Henneberger, C., Papouin, T., Oliet, S.H., and Rusakov, D.A. (2010). Longterm potentiation depends on release of D-serine from astrocytes. Nature 463, 232–236. Herrero-Mendez, A., Almeida, A., Ferna´ndez, E., Maestre, C., Moncada, S., and Bolan˜os, J.P. (2009). The bioenergetic and antioxidant status of neurons
822 Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc.
Panatier, A., Theodosis, D.T., Mothet, J.P., Touquet, B., Pollegioni, L., Poulain, D.A., and Oliet, S.H. (2006). Glia-derived D-serine controls NMDA receptor activity and synaptic memory. Cell 125, 775–784. Pellerin, L., and Magistretti, P.J. (1994). Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc. Natl. Acad. Sci. USA 91, 10625–10629. Perea, G., Navarrete, M., and Araque, A. (2009). Tripartite synapses: astrocytes process and control synaptic information. Trends Neurosci. 32, 421–431.
Pierre, K., and Pellerin, L. (2005). Monocarboxylate transporters in the central nervous system: distribution, regulation and function. J. Neurochem. 94, 1–14. Rex, A., Bert, B., Fink, H., and Voigt, J.P. (2009). Stimulus-dependent changes of extracellular glucose in the rat hippocampus determined by in vivo microdialysis. Physiol. Behav. 98, 467–473. Rinholm, J.E., Hamilton, N.B., Kessarism, N., Richardson, W.D., Bergersen, L.H., and Attwell, D. (2011). Regulation of oligodendrocyte development and myelination by glucose and lactate. J. Neurosci. 31, 538–548.
Vazdarjanova, A., Ramirez-Amaya, V., Insel, N., Plummer, T.K., Rosi, S., Chowdhury, S., Mikhael, D., Worley, P.F., Guzowski, J.F., and Barnes, C.A. (2006). Spatial exploration induces ARC, a plasticity-related immediate-early gene, only in calcium/calmodulin-dependent protein kinase II-positive principal excitatory and inhibitory neurons of the rat forebrain. J. Comp. Neurol. 498, 317–329.
Rouach, N., Koulakoff, A., Abudara, V., Willecke, K., and Giaume, C. (2008). Astroglial metabolic networks sustain hippocampal synaptic transmission. Science 322, 1551–1555.
Vilchez, D., Ros, S., Cifuentes, D., Pujadas, L., Valle`s, J., Garcı´a-Fojeda, B., Criado-Garcı´a, O., Ferna´ndez-Sa´nchez, E., Medran˜o-Ferna´ndez, I., Domı´nguez, J., et al. (2007). Mechanism suppressing glycogen synthesis in neurons and its demise in progressive myoclonus epilepsy. Nat. Neurosci. 10, 1407– 1413.
Schurr, A., Miller, J.J., Payne, R.S., and Rigor, B.M. (1999). An increase in lactate output by brain tissue serves to meet the energy needs of glutamateactivated neurons. J. Neurosci. 19, 34–39.
Villasana, L.E., Klann, E., and Tejada-Simon, M.V. (2006). Rapid isolation of synaptoneurosomes and postsynaptic densities from adult mouse hippocampus. J. Neurosci. Methods 158, 30–36.
Sotelo, C., and Palay, S.L. (1968). The fine structure of the lateral vestibular nucleus in the rat. I. Neurons and neuroglial cells. J. Cell Biol. 36, 151–179.
Waagepetersen, H.S., Westergaard, N., and Schousboe, A. (2000). The effects of isofagomine, a potent glycogen phosphorylase inhibitor, on glycogen metabolism in cultured mouse cortical astrocytes. Neurochem. Int. 36, 435–440.
Swanson, R.A. (1992). Physiologic coupling of glial glycogen metabolism to neuronal activity in brain. Can. J. Physiol. Pharmacol. 70, 138–144. Taubenfeld, S.M., Milekic, M.H., Monti, B., and Alberini, C.M. (2001a). The consolidation of new but not reactivated memory requires hippocampal C/EBPb. Nat. Neurosci. 4, 813–818. Taubenfeld, S.M., Wiig, K.A., Monti, B., Dolan, B., Pollonini, G., and Alberini, C.M. (2001b). Fornix-dependent induction of hippocampal CCAAT enhancer-binding protein b and d co-localizes with phosphorylated cAMP response element-binding protein and accompanies long-term memory consolidation. J. Neurosci. 21, 84–91. Tsacopoulos, M., and Magistretti, P.J. (1996). Metabolic coupling between glia and neurons. J. Neurosci. 16, 877–885. Urrila, A.S., Hakkarainen, A., Heikkinen, S., Vuori, K., Stenberg, D., Ha¨kkinen, A.M., Lundbom, N., and Porkka-Heiskanen, T. (2004). Stimulus-induced brain lactate: effects of aging and prolonged wakefulness. J. Sleep Res. 13, 111–119.
Walls, A.B., Sickmann, H.M., Brown, A., Bouman, S.D., Ransom, B., Schousboe, A., and Waagepetersen, H.S. (2008). Characterization of 1,4-dideoxy-1,4-imino-d-arabinitol (DAB) as an inhibitor of brain glycogen shunt activity. J. Neurochem. 105, 1462–1470. Whitlock, J.R., Heynen, A.J., Shuler, M.G., and Bear, M.F. (2006). Learning induces long-term potentiation in the hippocampus. Science 313, 1093–1097. Williams, C., Mehrian Shai, R., Wu, Y., Hsu, Y.H., Sitzer, T., Spann, B., McCleary, C., Mo, Y., and Miller, C.A. (2009). Transcriptome analysis of synaptoneurosomes identifies neuroplasticity genes overexpressed in incipient Alzheimer’s disease. PLoS ONE 4, e4936. Zhang, Z., Gong, N., Wang, W., Xu, L., and Xu, T.L. (2008). Bell-shaped D-serine actions on hippocampal long-term depression and spatial memory retrieval. Cereb. Cortex 18, 2391–2401.
Cell 144, 810–823, March 4, 2011 ª2011 Elsevier Inc. 823
Correction
Rallying the Exocyst as an Autophagy Scaffold Jean-Claude Farre´ and Suresh Subramani* *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.021
(Cell 144, 172–174; January 21, 2011) In Figure 1 of the above article, the dotted arrow was incorrectly labeled as ‘‘autophagy induction.’’ The label has been corrected to ‘‘autophagy inhibition’’ in Figure 1 reprinted below and in the online version of the article.
824 Cell 144, 824, March 4, 2011 ª2011 Elsevier Inc.
Errata
Functional and Mechanistic Diversity of Distal Transcription Enhancers Michael Bulger* and Mark Groudine *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.026
(Cell 144, 327–339; February 4, 2011) In the above article, the following grant funding should have been included in the Acknowledgments section: M.G. is funded by NIH grants R37 DK44746 and RO1 HL65440. M.B. is funded by NIH grant R01 DK070687.
Role for Dpy-30 in ES Cell-Fate Specification by Regulation of H3K4 Methylation within Bivalent Domains Hao Jiang, Abhijit Shukla, Xiaoling Wang, Wei-yi Chen, Bradley E. Bernstein, and Robert G. Roeder* *Correspondence:
[email protected] DOI 10.1016/j.cell.2011.02.036
(Cell 144, 513–525; February 18, 2011) In the above article, two human Dpy-30 sequences were inadvertently inserted in the rows designated for mouse Dpy-30 sequences in Table S1. The corrected table is now available online.
Cell 144, 825, March 4, 2011 ª2011 Elsevier Inc. 825
Scientific Editor, Cell Cell is seeking a scientist to join its editorial team. The minimum qualification for this position is a PhD in a relevant area of biomedical research, although additional postdoctoral or editorial experience is preferred. This is a superb opportunity for a talented individual to play a critical role in promoting science by helping researchers disseminate their findings to the wider community. As an editor, you would be responsible for assessing submitted research papers and overseeing the refereeing process, and you would commission and edit material for Cell's Leading Edge. You would also travel frequently to scientific conferences to follow developments in research and to establish and maintain close ties with the scientific community. You would have opportunities to pioneer and contribute to new trends in scientific publishing. The key qualities we look for are breadth of scientific interest and the ability to think critically about a wide range of scientific issues. The successful candidate will be highly motivated and creative and able to work independently as well as in a team. This is a full-time, in-house editorial position, based at the Cell Press office in Kendall Square near MIT in Cambridge, Massachusetts. Cell Press offers an attractive salary and benefits package and a stimulating working environment that encourages innovation.
To apply, please visit http://reedelsevier.taleo.net/careersection/51/jobdetail.ftl?lang=en&job=SCI0005W.
Scientific Editor, Molecular Cell Molecular Cell is seeking a full-time scientific editor to join its editorial team. We will consider qualified candidates with scientific expertise in any area that the journal covers. The minimum qualification for this position is a PhD in a relevant area of biomedical research, although additional experience is preferred. This is a superb opportunity for a talented individual to play a critical role in the research community away from the bench. As a scientific editor, you would be responsible for assessing submitted research papers, overseeing the refereeing process, and choosing and commissioning review material. You would also travel frequently to scientific conferences to follow developments in research and establish and maintain close ties with the scientific community. The key qualities we look for are breadth of scientific interest and the ability to think critically about a wide range of scientific issues. The successful candidate will also be highly motivated and creative and able to work independently as well as in a team. This is a full-time in-house editorial position, based at the Cell Press office in Cambridge, Massachusetts. Cell Press offers an attractive salary and benefits package and a stimulating working environment. Applications will be held in the strictest of confidence and will be considered on an ongoing basis until the position is filled.
To apply Please submit a CV and cover letter describing your qualifications, research interests, and reasons for pursuing a career in scientific publishing, as soon as possible, to our online jobs site: http://www.elsevier.com/wps/find/job_search.careers. Click on “search for US jobs” and select “Massachusetts.” Or: http://reedelsevier.taleo.net/careersection/51/jobdetail.ftl?lang=en&job=SCI0005X.
No phone inquiries, please. Cell Press is an equal opportunity/affirmative action employer, M/F/D/V.
Scientific Editor, Cancer Cell Cancer Cell is seeking an additional full-time scientific editor to join its editorial team based in Cambridge, Massachusetts. Cancer Cell publishes studies across broad areas of cancer research with an emphasis on translational research. As a scientific editor, you would be responsible for assessing submitted manuscripts, overseeing the review process, and commissioning and editing review material for the journal. You would travel frequently to scientific conferences and research institutions to follow developments in cancer research and to establish and maintain close ties with the scientific community. This position will also work closely with other aspects of the business, including production, business development, marketing, and commercial sales, and, therefore, provide an excellent entry opportunity to scientific publishing. The minimum qualification for this position is a PhD in a relevant area of cancer research; additional experience at the postdoctoral level is preferred. Previous editorial experience is beneficial but is not required. This is a superb opportunity for a talented individual to play a critical role in the research community away from the bench. The key qualities we look for are breadth of scientific interest and the ability to think critically about a wide range of scientific issues. The successful candidate will also be highly motivated and creative, possess strong communication skills, and be able to work in a team as well as independently. This is a full-time, in-house editorial position, based at the Cell Press office in Cambridge, Massachusetts. Cell Press offers an attractive salary and benefits package and a stimulating working environment. Applications will be held in the strictest of confidence and will be considered on an ongoing basis.
To apply Please submit a cover letter and CV to our online jobs site: http://reedelsevier.taleo.net/careersection/51/jobdetail.ftl?lang=en&job=SCI00065. Please, no phone inquiries. Cell Press is an equal opportunity/affirmative action employer, M/F/D/V.
Editor, Trends in Biotechnology We are seeking to appoint a new Editor for Trends in Biotechnology, to be based in the Cell Press offices in Cambridge, Massachusetts. As Editor of Trends in Biotechnology, you will be responsible for the strategic development and content management of the journal. You will be acquiring and developing the very best editorial content, making use of a network of contacts in academia plus information gathered at international conferences, to ensure that Trends in Biotechnology maintains its market-leading position. This is an exciting and challenging role that provides an opportunity to stay close to the cutting edge of scientific advances while developing a new career away from the bench. You will work in a highly dynamic and collaborative publishing environment that includes 14 Trends titles and 12 Cell Press titles. You will also collaborate with your Cell Press colleagues to maximize quality and efficiency of content commissioning and participate in exciting new non-journal-based initiatives. The minimum qualification is a doctoral degree in a relevant discipline, and postdoctoral training is an advantage. Previous publishing experience is not necessary—we will make sure you get the training and development you need. Good interpersonal skills are essential because the role involves networking in the wider scientific community and collaboration with other parts of the business.
To apply, please visit http://reedelsevier.taleo.net/careersection/51/jobdetail.ftl?lang=en&job=SCI0007Q and submit a CV and cover letter describing your qualifications, research interests, and reasons for pursuing a career in publishing. No phone inquiries, please. Cell Press is an Equal Opportunity Employer. Cell Press offers an attractive salary and benefits package and a stimulating working environment. Applications will be held in the strictest of confidence and considered on an ongoing basis.
Positions Available
Positions Available
60
years of leadership in human genetics research, education and service. 1948–2008 www.ashg.org
.%7
Find Your Ideal Job! så3EARCHåJOBSåBYåKEYWORD å LOCATIONååTYPEå så0OSTåYOURåRESUMEåå ANONYMOUSLYå så#REATEåAå*OBå!LERTåå ANDåLETåYOURåIDEALåå JOBålNDåYOU
careers.cell.com
3TAYå#ONNECTED *OINå!3('å4ODAY "ENElTS¬OF¬-EMBERSHIP så3UBSCRIPTIONåTOå!*(' åNOWåPUBLISHEDåå BYå#ELLå0RESSå så2EDUCEDåREGISTRATIONåFEESåFORåTHEåå !NNUALå-EETINGå så$ISCOUNTEDåSUBSCRIPTIONSåTOå#ELLåå 0RESSåJOURNALS så!CCESSåTOåEDUCATIONALåMATERIALS så)NmUENCEåONåPUBLICåPOLICYåPROCESS
WWWASHGORG
¬,OOK¬!GAIN
$ISCOVER¬-ORE s¬!CCESS¬TO¬THE¬¬#ELL¬0RESS¬PRIMARY¬RESEARCH¬JOURNALS¬AND¬ ¬4RENDS¬REVIEWS¬TITLES ¬ALL¬ON¬THE¬SAME¬PLATFORM s¬)MPROVED ¬MORE¬ROBUST¬ARTICLE¬AND¬AUTHOR¬SEARCH s¬6IDEO ¬ANIMATIONS¬AND¬SOUND¬lLES s¬%ASY¬TO¬NAVIGATE¬HOME¬PAGE ¬ARTICLES¬PAGES¬AND¬ARCHIVE
WWWCELLCOM
Scopus is the largest abstract and citation database of peer-reviewed literature and quality web sources with smart tools to track, analyze and visualize research.
enrich your experience
www.scopus.com
#ALLåFORå0APERS 3UBMITåYOURåMANUSCRIPTåTODAY )MPACT¬ &ACTOR
3OURCEåå4HOMSONå2EUTERSå*OURNALå#ITATIONå2EPORTS
4OPICSåCOVEREDåINCLUDE
å
¬#HANNELS ¬2ECEPTORS ¬AND¬%LECTRICAL¬3IGNALING
¬3PECTROSCOPY ¬)MAGING ¬/THER¬4ECHNIQUES
¬0ROTEINS¬
¬-USCLE¬AND¬#ONTRACTILITY
¬"IOPHYSICAL¬4HEORY¬AND¬-ODELING¬
¬3UPRAMOLECULAR¬!SSEMBLIES
¬#ELL¬"IOPHYSICS
¬"IOENERGETICS
¬0HOTOBIOPHYSICS
¬.UCLEIC¬!CIDS¬
¬-EMBRANES¬
¬%LECTROPHYSIOLOGY
HTTPSUBMITBIOPHYSJORG
SnapShot: Neuromuscular Junction Steven J. Burden Skirball Institute of Biomolecular Medicine, New York University Medical School, New York, NY 10016, USA
Motor axons approach muscles that are prepatterned NEUROMUSCULAR JUNCTION Muscle
Muscle fibers
Bone
Motor axons
Motor axon terminal
Muscle membrane
Motor axons Nucleus AChR transcription
Innervation
Synaptic cleft Mitochondrion
MUSCLE
MOTOR AXON TERMINAL
NMJ Diseases
ACh
CONGENITAL MYASTHENIA
Lrp4
Hypomorphic mutations in these genes impair postsynaptic differentiation. The number of AChRs at synapses is reduced, which compromises the reliability of synaptic transmission and leads to muscle weakness and fatigue.
Agrin MuSK Dok-7 Rapsyn
Agrin
MuSK Antagonizes differentiation Stabilizes differentiation
Mutations in these genes impair synaptic transmission, leading to muscle weakness and fatigue. ChAT encodes the enzyme for synthesizing ACh, and Nav1.4 encodes a voltage-activated sodium channel expressed in skeletal muscle.
AChR AChE ChAT Nav1.4
AChR
MYASTHENIA GRAVIS Auto-antibodies to AChRs
Cause accelerated degradation of AChRs and structural disorganization of the synapse, leading to Myasthenia Gravis
Auto-antibodies to MuSK
Cause structural disorganization of the synapse, leading to Myasthenia Gravis.
P P
P
Rapsyn
Dok-7
Crk/ Crk-L
P P
AChR MuSK Lrp4 AChE Utrophin 826
P
Erm/GABP/ERG
JNK
ETS
Cell 144, March 4, 2011 ©2011 Elsevier Inc.
Rho GTPases Dvl, Pak, Tid1 Abl1/2
DOI 10.1016/j.cell.2011.02.037
Cortactin, D-actinin ACF7, Actin, Myosin
CYTOSOL
See online version for legend and references.
Metabolism & Aging March 27-29, 2011 Cape Cod, Massachusetts, USA
Conference Organizers Prof. David A. Sinclair, Harvard Medical School, Boston, USA Dr. Nir Barzilai, M.D., Albert Einstein College of Medicine, New York, USA Dr. C. Ronald Kahn, Joslin Diabetes Center at Harvard Medical School, Boston, USA Speakers Domenico Accilli, Columbia University, NY, USA Adam Antebi, Max Planck Institute for Biology of Ageing, Germany Dongsheng Cai, Albert Einstein College of Medicine, NY, USA Hassy Cohen, UC Los Angeles, CA, USA Jill Crandall, Albert Einstein College of Medicine, NY, USA Rafael de Cabo, National Institute of Health, MD, USA Andy Dillin, Salk Institute For Biological Studies, CA, USA David J. Glass, Novartis Institutes for BioMedical Research, MA, USA Leonard Guarente, Massachusetts Institute of Technology, MA, USA Pankaj Kapahi, The Buck Institute for Age Research, CA, USA Brian Kennedy, University of Washington, WA, USA James Kirkland, Mayo Clinic, MN, USA Valter D. Longo, UC San Francisco, CA, USA Jim Nelson, UT Health Science Center, TX, USA Eric Ravussin, Pennington Biomedical Research Center, LA, USA Arlan Richardson, UT Health Science Center, TX, USA Randy Strong, UT Health Science Center, TX, USA Marc Tatar, Brown University, RI, USA Heidi Tissenbaum, University of Massachusetts Medical School, MA, USA Eric Verdin, UC San Francisco, CA, USA
The first Cell Symposia meeting of 2011, Metabolism & Aging takes place on March 27 – 29 in the beautiful Cape Cod peninsula at the southern tip of Massachusetts, USA. This meeting aims to bring together scientists with interests in aging and metabolism to further explore how these fields intersect and to identify the most promising future directions. We will hear the latest data from leaders in the field about the key pathways at the level of the cell and the organ, across a range of contexts including model organisms, mammalian systems, and translational studies in primates and humans. Topics will also include how the signalling networks of metabolism and aging connect and communicate and how we can best make use of these connections to improve medicine and society.
Visit the Metabolism & Aging website to: REGISTER
Supporting publications
Submit your poster abstract Study the final programme View our speakers biographies
www.cell-symposia-metabolism-aging.com
Biocharacterized. Dramatic results begin with highly characterized bioactive peptides from Sigma . ®
Sigma Life Science is the leading provider of bioactive peptides for your pathway research. Our highly characterized peptides offer the purity and specificity that your research demands, with all of our specifications and analysis available on the web. View over 250 NEW bioactive peptides at
wherebiobegins.com/newpeptides
biomolecules Sigma and Sigma-Aldrich are registered trademarks of Sigma-Aldrich Biotechnology, L.P. and Sigma-Aldrich Co.
