Methods in Enzymology Vol 164: Ribosomes

C o n t r i b u t o r s to Volume 164 Article numbers are in parentheses followingthe names of contributors. Affiliation...

Author: Jr. Harry F. Noller | Kivie Moldave

16 downloads 915 Views 78MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

C o n t r i b u t o r s to Volume 164 Article numbers are in parentheses followingthe names of contributors. Affiliations listed are eurr~L

STEVENAEo (10), Department of Chemistry,

Yale University, New Haven, Connecticut 06511 ANDREA BARTA(24), lnstitutfftr Biochemie, Universitdt Wien, A-1090 Vienna, Austria ANDREAS BARTETZKO (44), Max-Planck-lnstitut J~r Molekulare Genetik, Abt. Wittman, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany N. V. BELITSINA(43), A. N. Bakh Institute of Biochemistry, Academy of Sciences of the USSR, Moscow, USSR EGBERT BOEr~MA (2), Fritz-Haber-Institut der Max-Planck-Gesellschafl, D-1000 Berlin 33, Federal Republic of Germany ALEXEY A. BOODANOV (29), A. N. Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 119899, USSR ANGELA BORDEN (46), Department of Experimental Biology, Roswell Park Memorial Institute, Buffalo, New York 14263 MILOSLAV BOUBLIK (3), Roche Institute of Molecular Biology, Roche Research Center, Nutley, New Jersey 07110 RICHARD BRIMACOMBE (19), Max-PlanckInstitut fi~r Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany MELISSA A. BUCK (36), NINCDS, National Institutes of Health, Bethesda, Maryland 20892 V. N. BusnuEv (9), Institute of Experimental Cardiology, Cardiology Research Center of the USSR Academy of Medical Sciences, 121552 Moscow, USSR DAVID G. CAMP (26), Department of Biochemistry and Biophysical Sciences, University of Houston, Houston, Texas 77004 PAULINE A. CANN (34), Department of BiD-

logical Chemistry, UCLA School of Medicine, University of California, Los Angeles, Los Angeles, California 90024 MALCOLM S. CAPEL (7, 37), Biology Department, Brookhaven National Laboratory, Upton, New York 11973 Lx-MINa CHANGCHIEN(16, 17), Wadsworth Center for Labs and Research, New York State Department of Health, Empire State Plaza, Albany, New York 12201 C. CmARUTTrNI (20), Institut de Biologie Physico-Chimique, Laboratoire de Chimie Cellulaire, 75005 Paris, France NINA V. CmCHKOVA(29), A. N. Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 119899, USSR JAN CHRISTIANSEN(30, 49), Department of Clinical Chemistry, Bispebjerg Hospital, DK-2400 Copenhagen NF,, Denmark ROBERT CONRAD (16), Department of Biology, Indiana University, Bloomington, Indiana 47405 BARRY S. COOPERMAN(23, 36), Department of Chemistry, University of Pennsylvania, Philadelphia, Pennsylvania 19104 GARY R. CRAVENI (16, 17, 37), Laboratory of Molecular Biology and Department of Genetics, University of Wisconsin, Madison, Wisconsin 53706 ALBERT E. DAHLBERG(47), Section of BiDchemistry, Division of Biology and Medicine, Brown University, Providence, Rhode Island 02912 DIPAK B. DATTA (17, 37), Department of Botany, University of Wisconsin, Madison, Wisconsin 53706 INGRID C. DECKMAN(13), Smith, Kline & 'Deceased.

xiii

xiv

CONTRIBUTORS TO VOLUME 164

French Research Laboratories, Molecular Genetics Research and Development, King of Prussia, Pennsylvania 19406 H.-Y. DENG (11), Clayton Foundation Biochemical Institute, Chemistry Department, University of Texas, Austin, Texas 78712 ROBERT DENMAN (25), Institute for Basic Research in Developmental Disabilities, Staten Island, New York 10314 DAVID E. DRAPER (13, 14), Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218 JAN EGEBJERG (49), Biostructural Chemistry, Kemisk Institut, Aarhus Universitet, 8000 Aarhus C, Denmark MANs EHRENBERG(42), Department of Molecular Biology, University of Uppsala, S- 751 24 Uppsala, Sweden MOHAMED ETTAYEm (46), Department of Experimental Biology, Roswell Park Memorial Institute, Buffalo, New York, 14263 A. EXPERT-BEzANtTON(20), Institut Jacques Monod-C.N.R.S., Laboratoire de Photobiologie Moleculaire, 75251 Paris Cedex 05, France ANTONIO FOCELLA (25), Department of Chemistry, Hoffmann-LaRoche Inc., Nutley, New Jersey 07110 JOACmM FRANK (1), Wadsworth Center for Laboratories and Research, New York State Department of Health, and School of Public Health, State University of New York at Albany, Albany, New York 12201 BETTY FREEBORN (10), Department of Chemistry, Yale University, New Haven, Connecticut 06511 ROVER GARRETT (30, 49), Department of Biostructural Chemistry, Kemisk Institut, Aarhus Universitet, 8000 Aarhus C, Denmark UTE GEIGENMOLLER (45), Max*Planck-Institut far Molekulare Genetik, Abt. Wittman, D-1000 Berlin 33 (Dahlem), Federal Republic of Germany DANIEL T. GEWIRTH (10), Department of Molecular Biophysics and Biochemistry,

Yale University, New Haven, Connecticut 06511 DOHN G. GLITZ (34), Department of Biological Chemistry and Molecular Biology Institute, UCLA School of Medicine, University of California, Los Angeles, Los Angeles, California 90024 LARRY GOLD (27), Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, Colorado 80309 H. U. GORINGER (50), Max-Planck-Institut far Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany A. T. G u D t o v (9), Institute of Protein Research, Academy of Sciences of the USSR, 142292 Pushchino, Moscow Region, USSR JAMES F. HAINFELD (3), Biology Department, Brookhaven National Laboratory, Upton, New York 11973 GEOROE HARAUZ(2), Department of Molecular Biology and Genetics, University of Guelph, Guelph, Ontario, N1G 21 W Canada BOYD HARDESTY(11), Clayton Foundation Biochemical Institute, Chemistry Department, University of Texas, Austin, Texas 78712 DIETER HARTZ (27), Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, Colorado 80309 JOHN E. HEARST(22), Department of Chemistry, University of California, Berkeley, Berkeley, California 94720 HANS A. HEUS (12), Department of Biochemistry, Leiden University, 2333 AL Leiden, The Netherlands WALTER E. HILL (26), Department of Chemistry, University of Montana, Missoula, Montana 59812 PAUL W. HUBER (31), Department of Chemistry, University of Notre Dame, Notre Dame, Indiana 46556 DAVID K. JEMIOLO (47), Biology Depart-


ment, Vassar College, Poughkeepsie, New York 12601 ROZA MARIA KAMP (38), Max-Planck-Institut far Molekulare Genetik, Abt. Wittman, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany JOANNE M. KEAN (14), Division of Biophysics, School of Hygeine and Public Health, Johns Hopkins University, Baltimore, Maryland 21218 ALEXEYM. KOPYLOV(29), A. N. Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 119899, USSR JOACHIM KRIEG (39), Freidrich MiescherInstitut, CH-4002 Basel, Switzerland ERNST KUECHLER (24), lnstitut far Biochemie, Universitdt Wien, A-1090 Vienna, Austria C. G. KURLAND(42), Department of Molecular Biology, University of Uppsala, S-751 24 Uppsala, Sweden APOSTOLOS KYRIATSOULIS(19), I Medizinische Klinik der Johannes, Gutenberg Universitdit, D-6500 Mainz, Federal Republic of Germany LANCE G. LAING(15), Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218 LINDA S. LASATER(34), Department of Biological Chemistry, UCLA School of Medicine, University of California, Los Angeles, California 90024 NEOCLES B. LEONTIS (10), Chemistry Department, Bowling Green State University, Bowling Green, Ohio 43403 ARNOLD LIEBMAN (25), Department of Chemistry, Hoffmann-LaRoche Inc., Nutley, New Jersey07110 ROLF LIETZKE(18), Max-Planck-Institutfar Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany ROLAND LILL (41), Department of Biological Chemistry, School of Medicine, University of California, Los Angeles, Los Angeles, California 90024

XV

SAMUEL E. LIPSON (22), CODON, South

San Francisco, California 94080 DAVID MALAREK (25), Department

of Chemistry, Hoffmann-LaRoche Inc., Nutley, New Jersey 07110 PETER MALY (19), Biochemisches Institut der Universitdt Zftrich, CH-8057 Zarich, Switzerland VALSAN MANDIYAN (3), Roche Institute of Molecular Biology, Roche Research Center, Nutley, New Jersey 07I 10 ALEXANDER S. MANKIN (29), A. N. Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 119899, USSR DAVID S. MCPHEETERS(27), Division of Biology, California Institute of Technology, Pasadena, California 91125 DANESH MOAZED (33), Thimann Laboratories, University of California, Santa Cruz, Santa Cruz, California 95064 PETER B. MOORE (10), Department of Chemistry, Yale University, New Haven, Connecticut 06511 EDWARD A. MORGAN (46), Department of Experimental Biology, Roswell Park Memorial Institute, Buffalo, New York 14263 KNUD H. NIERHAUS (8, 18, 44, 45), Max-

Planck-Institut ff~r Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany CONCEPCION R. NIER~S (37), Laboratory of Molecular Biology, University of Wisconsin, Madison, Wisconsin 53706 KAYOKO NISHI (48), Department of Biological Chemistry, School of Medicine, University of California, Davis, Davis, California 95616 HARRY F. NOLLER (32, 33), Board of Studies in Biology, University of California, Santa Cruz, Santa Cruz, California 95064 PETRA NOWOTNY (8), Max-Planck-Institut )~r Molekulare Genetick, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany VOLKER NOWOTNY (8), Max-Planck-Institut far Molekulare Genetik, Abt. Witt-

xvi


Planck-Institut J~r Molekulare Genetik, mann, D-IO00 Berlin 33 (Dahlem), FedAbt. Wittmann, D-IO00 Berlin 33 (Daheral Republic of Germany lem), Federal Republic of Germany KELVIN NORSE (25), Department of BioJAMES M. ROBERTSON(40), Thimann Labochemistry, Roche Institute of Molecular ratories, University of California, Santa Biology, Roche Research Center, Nutley, Cruz, Santa Cruz, California 95064 New Jersey 07110 JOACHIM SCrINIER (48), Department of BioO. W. ODOM (11), Clayton Foundation Biological Chemistry, School of Medicine, chemical Institute, Chemistry Department, University of California, Davis, Davis, University of Texas, Austin, Texas 78712 California 95616 JAMES OFENOAND(25), Department of BioIVAN N. SHATSKY(5), A. N. Belozersky Labchemistry, Roche Institute of Molecular oratory of Molecular Biology and BioorBiology, Roche Research Center, Nutley, ganic Chemistry, Moscow State UniverNew Jersey 07110 sity, 117234 Moscow, USSR YASUNARI OGIHARA(51), Kihara Institute, CURT D. SIOMUND(46), Department of ExYokohama City University, Yokohama perimental Biology, Roswell Park Memo232, Japan rial Institute, Buffalo, New York 14263 ANDR~E R. OLIVIER (39), Friedrich EVOENY A. SKRIPraN (29), A. N. Belozersky Miescher-Institut, CH-4002 Basel, SwitLaboratory of Molecular Biology and zerland Bioorganic Chemistry, Moscow State UniGARY J. OLSEN (53), Department of Microversity, Moscow 119899, USSR biology, University of Illinois, Urbana, IlliA. S. SPIRIN (28, 43), Institute of Protein nois 61801 Research, Academy of Sciences of the HELEN McKuSFdE OLSON (34), Department USSR, 142292 Pushchino, Moscow Reof Biological Chemistry, UCLA School of gion, USSR Medicine, University of California, Los MICHAEL J. R. STARK (47), Leicester BioAngeles, Los Angeles, California 90024 centre, University of Leicester, Leicester, GERRIT T. OOSTERGETEL(3), Roche InstiEngland tute of Molecular Biology, Roche Research ROLE STEEN (47), Department of Molecular Center, Nutley, New Jersey 07110 Biology, Biomedicum, University of UppHARALD PAULSEN (40), Botanisches Instisala, S-751 24 Uppsala, Sweden tut, Universit~t Mf2nchen, D-8000 Mfmchen 19, Federal Republic of Germany GONTER STEINER (24), Allg. Krankenhaus, II. Med. Universitdtsklinik, A-1090 ANASTASIA PROMBONA (51), Max-PlanckVienna, Austria Institut far Molekulare Genetik, Abt. Wittmann, D-1000 Berlin 33 (Dahlem), Fed- SETH STERN (33), Thimann Laboratories, eral Republic of Germany University of California, Santa Cruz, Santa Cruz, California 95064 MICHAEL RADERMACHER (1), Wadsworth Center for Laboratories and Research, WOLFGANG STIEGE (19), Max-Planck-InstiNew York State Department of Health, Altut f~r Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Fedbany, New York, 12201 eral Republic of Germany V. RAMAKRISHNAN (7), Biology Department, Brookhaven National Laboratory, GEORG STOFFLER(4, 35), InstitutfarMikroUpton, New York 11973 biologie der Medizinische Fakultd~t, Universiti~t Innsbruck, A-6020 Innsbruck, BERNHARD REDL (4), Institut far MikrobioAustria logie der Medizinische Fakult,~t, Universitdt Innsbruck, A-6020 Innsbruck, Austria MARINA STOFFLER-MEILICKE(4, 35), MaxPlanck-Institut ~ r Molekulare Genetik, HANS-JORG RHEINBERGER (45), Max-


Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany ALAP R. SUBRAMANIAN(51), Max-PlanckInstitut fi~r Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany GRACE SUN (10), Department of Chemistry, Yale University, New Haven, Connecticut 06511 WILLIAM E. TAPPRICH (26), Division of Biology and Medicine, Brown University, Providence, Rhode Island 02912 ANCHALEE TASSANAKAJOHN(26), Department of Chemistry, Ramkamhangu University, Bangna, Bangkok 10260, Thailand GEORGE THOMAS (39), Friedrich MiescherInstitut, CH-4002 Basel, Switzerland ROBERT TRAUT (27), Department of Biological Chemistry, School of Medicine, University of California, Davis, Davis, California 95616 MARIN VAN HEEL (2), Fritz-Haber-lnstitut der Max-Planck-Gesellschafi, D-IO00 Berlin 33, Federal Republic of Germany PETER H. VAN KNIPPENBERG (12), Department of Biochemistry, Leiden University, 2333 AL Leiden, The Netherlands BARBARA J. VAN STOLK (32), Thimann Laboratories, University of California, Santa Cruz, Santa Cruz, California 95064 JAILAXMI V. VARTIKAR(13), Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218 VICTOR D. VASILIEV(5), Institute of Protein Research, Academy of Sciences of the USSR, 142292 Pushchino, Moscow Region, USSR ADRIANA VERSCHOOR (1), Wadsworth Center for Laboratories and Research, New York State Department of Health, Albany, New York 12201 BIRTE VESTER (49), Biostructural Chemistry, Kemisk Institut, Aarhus Universitet, 8000 Aarhus C, Denmark HELGA VOSS (8), Max-Planck-lnstitut fi2r Molekulare Genetik, Abt. Wittmann,

xvii

D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany TERENCE WAGENKNECHT (1), Wadsworth Center for Laboratories and Research, New York State Department of Health, Albany, New York 12201 R. WAGNER (50), Institut fflr Physikalische Biologie, Universitlit Dftsseldo~ D-4000 D~sseldorf 1, Federal Republic of Germany JOSEPH S. WALL (3), Biology Department, Brookhaven National Laboratory, Upton, New York 11973 JAN WALLECZEK (4), Max-Planck-lnstitut fiir Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany MICHAEL S. WATERMAN(52), Departments of Mathematics and Molecular Biology, University of Southern California, Los Angeles, California 90089 MARKUS WEDDE (45), Max-Planck-lnstitut J'fir Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany CARL J. WEITZMANN(36), Roche Institute of Molecular Biology, Nutley, New Jersey 07110 SUSANA. WHITE (14), Department of Chemistry, Yale University, New Haven, Connecticut 06511 ERIC WICKSTROM (15), Department of Chemistry, University of South Florida, Tampa, Florida 33620 WOLFGANG WINTERMEYER(40, 41), Instirut fftr Physiologische Chemie, Universitdt Witten/Herdecke, D-5810 Witten, Federal Republic of Germany H. G. WITTMANN(6), Max-Planck-lnstitut fi~r Molekulare Genetik, Abt. Wittmann, D-1000 Berlin (Dahlem), Federal Republic of Germany BRIGITTE WITTMANN-LIEBOLD(38), Max-

Planck-Institut f~r Molekulare Genetik, Abt. Wittmann, D-IO00 Berlin 33 (Dahlem), Federal Republic of Germany PAUL L. WOLLENZIEN(2 l), E. A. Doisy De-

XVlll

CONTRIBUTORS TO VOLUME 1 6 4

partment of Biochemistry, St. Louis University Medical School, St. Louis, Missouri 63104 IRA G. WOOL (31), Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago, Illinois 60637 A. YONATH (6), Max-Planck-Unitfor Structural Molecular Biology, D-2000 Hamburg 52, Federal Republic of Germany, and Wiezmann Institute of Science, Rehovot, Israel

M. M. YusuPov (28), Institute of Protein

Research, Academy of Sciences of the USSR, 142292 Pushchino, Moscow Region, USSR G. Z. YUSUPOVA(TNALINA)(43), Institute of Protein Research, Academy of Sciences of the USSR, 142292 Pushchino, Moscow Region, USSR GLADYS ZENCHOFF (25), Department of Chemistry, Hoffmann-LaRoche Inc., Nutley, New Jersey 07110

Preface The overwhelming structural complexity of ribosomes continues to present a great technical challenge to those who study these interesting ribonucleoprotein particles. Ribosomologists have responded over the years by inventing a wide range of novel biochemical, physicochemical, and genetic approaches, many of which have found widespread application outside the ribosome field. The most recent period of research in this area is no exception as is reflected in the contents of this volume. Among the notable advances that we have seen during this time are a vastly sharpened understanding of the structure of ribosomal RNA and how it may participate in the translation process, major accomplishments in the area of ribosome structure, and the emergence of the ribosome as an evolutionary chronometer. This volume supplements some of the recent ones in this series dealing with various aspects of protein synthesis (XXX, LIX, LX, and 101). Much of the methodology relating to the function of ribosomes in translation, the preparation and characterization of ribosomes and their constituent parts, their interaction with other translational components (protein factors, aminoacyl-tRNAs, mRNAs, nucleotides, etc.), and the properties of cellfree translating systems is to be found in these volumes. Also presented are a number of methods used for the characterization of the changes that occur in ribosomes when they react with other components in the translational system. In this rapidly moving area of research, many procedures are continually being modified and improved, and new methodologies are being developed which allow for more incisive analyses and elaboration of more detailed, reliable information regarding the structure of the particle, its component proteins and ribonucleic acids, and its diverse functional states. This volume includes a variety of methods involving electron microscopy and other biophysical techniques, such as crystallography, neutron scattering, and NMR, procedures for the analysis of protein-RN A or RNARNA interactions by cross-linking, the use of chemical, enzymatic, and immunological probes, as well as functional, kinetic, and genetic approaches for the study of this ribonucleoprotein. These methodologies will contribute to the continued progress toward the elucidation of the structure, function, and regulatory processes that affect this most important complex cellular component, the ribosome. HARRY F. NOLLER, JR. KIVIE MOLDAVE xix

METHODS IN ENZYMOLOGY EDITORS-IN-CHIEF

Sidney P. Colowick and Nathan 0. Kaplan

VOLUMEVIII. Complex Carbohydrates Edited by ELIZABETH F. NEUFELD AND VICTOR GINSBURG VOLUME

IX. CarbohydrateMetabolism

Edited by WILLIS A. WOOD VOLUME

X. Oxidation and Phosphorylation

Edited by RONALD W. ESTABROOK AND MAYNARD E. PULLMAN VOLUME

XI. Enzyme Structure

Edited by C. H. W. HIRS VOLUME

XII. Nucleic Acids (PartsA and B)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE VOLUME

XIII. Citric Acid Cycle

Edited by J. M. LOWENSTEIN VOLUME

XIV. Lipids

Edited by J. M. LQWENSTEIN VOLUME

XV. Steroidsand Terpenoids

Edited by RAYMOND B. CLAYTON VOLUME

XVI. Fast Reactions

Edited by KENNETH KUSTIN VOLUME

XVII. Metabolism of Amino Acids and Amines (PartsA and B)

Edited by HERBERT TABOR AND CELIA WHITE TABOR .. xx111

xxiv

METHODS INENZYMOLOGY

VOLUME XVIII. Vitamins and Coenzymes(PartsA, B, and C) Edited MCDONALD B. MCCORMICKANDLEMUEL D. WRIGHT VOLUME XIX. Proteolytic Enzymes Edited ~~GERTRUDE E. PERLMANNANDLASZLOLORAND VOLUME XX. Nucleic Acids and Protein Synthesis(Part C) Editedby KIVIEMOLDAVEANDLAWRENCEGROSSMAN VOLUME XXI. Nucleic Acids (Part D) Editedby LAWRENCEGROSSMANANDKIVIEMOLDAVE VOLUME XXII. Enzyme Purification and RelatedTechniques Editedby WILLIAM B. JAKOBY VOLUME XXIII. Photosynthesis(Part A) Editedby ANTHONYSANPIETRO VOLUME XXIV. Photosynthesisand Nitrogen Fixation (Part B) Editedby ANTHONY SANPIETRO VOLUME XXV. Enzyme Structure(Part B) EditedbyC. H.W. HIRSANDSERGEN.TIMASHEFF VOLUME XXVI. Enzyme Structure(Part C) EditedbyC. H. W. HIRSANDSERGEN.TIMASHEFF VOLUME XXVII. Enzyme Structure(Part D) Editedby C. H. W. HIRSANDSERGEN.TIMASHEFF VOLUME XXVIII. Complex Carbohydrates(Part B) Editedby VICTORGINSBURG VOLUME XXIX. Nucleic Acids and Protein Synthesis(Part E) Editedby LAWRENCEGROSSMANANDKIVIEMOLDAVE VOLUME XXX. Nucleic Acids and Protein Synthesis(Part F) Editedby KIVIEMOLDAVEANDLAWRENCEGROSSMAN VOLUME XXXI. Biomembranes(Part A) Editedby SIDNEYFLEISCHERANDLESTERPACKER

METHODSINENZYMOLOGY

xxv

VOLUME XxX11. Biomembranes(Part B) Edited ~~SIDNEYFLEISCHERANDLESTERPACKER VOLUME XxX111. Cumulative SubjectIndex Volumes I- XXX Edited ~~MARTHAG.DENNISANDEDWARD A. DENNIS VOLUME XXXIV. Affinity Techniques(Enzyme Purification: Part B) Editedby WILLIAM B. JAKOBY AND MEIRWILCHEK VOLUME XXXV. Lipids (Part B) Editedby JOHN M. LOWENSTEIN VOLUME XXXVI. Hormone Action (Part A: SteroidHormones) Editedby BERT W. O'MALLEYANDJOELG.HARDMAN VOLUME XXXVII. Hormone Action (Part B: PeptideHormones) Editedby BERT W. O'MALLEYANDJOELG.HARDMAN VOLUME XXXVIII. Hormone Action (Part C: Cyclic Nucleotides) Editedby JOELG.HARDMANANDBERT W. O'MALLEY VOLUME XxX1X. Hormone Action (Part D: IsolatedCells, Tissues,and

OrganSystems) Editedby

JOELG.HARDMANANDBERT

W. O'MALLEY

VOLUME XL. Hormone Action (Part E: Nuclear Structureand Function) Editedby BERT W. O'MALLEYANDJOELG.HARDMAN VOLUME XLI. CarbohydrateMetabolism (Part B) Edited by W. A. WOOD VOLUME XLII. CarbohydrateMetabolism (Part C) Edited by W. A. WOOD VOLUME XLIII. Antibiotics Edited by JOHN H. HASH VOLUME XLIV. Immobilized Enzymes Editedby KLAUSMOSBACH VOLUME XLV. Proteolytic Enzymes(Part B) Editedby LASZLO LORAND

xxvi

METHODSINENZYMOLOGY

VOLUME XLVI. Affinity Labeling Edited ~~WILLIAM B. JAKOBY AND MEIR WILCHEK VOLUME XLVII. Enzyme Structure(Part E) EditedbyC. H.W. HIRSANDSERGEN.TIMASHEFF VOLUME XLVIII. Enzyme Structure(Part F) EditedbyC. H. W. HIRSANDSERGEN.TIMASHEFF VOLUME XLIX. Enzyme Structure(Part G) EditedbyC. H.W. HIRSANDSERGEN.TIMASHEFF VOLUME L. Complex Carbohydrates(Part C) Editedby VICTORGINSBURG VOLUME LI. Purine and Pyrimidine NucleotideMetabolism Editedby PATRICIA A. HOFFEEANDMARYELLENJONES VOLUME LII. Biomembranes(Part C: Biological Oxidations) Editedby SIDNEYFLEISCHERANDLESTERPACKER VOLUME LIII. Biomembranes(Part D: Biological Oxidations) Editedby SIDNEYFLEISCHERANDLESTERPACKER VOLUME LIV. Biomembranes(Part E: Biological Oxidations) Editedby SIDNEYFLEISCHERANDLESTERPACKER VOLUME LV. Biomembranes(Part F: Bioenergetics) Editedby SIDNEYFLEISCHERANDLESTERPACKER VOLUME LVI. Biomembranes(Part G: Bioenergetics) Editedby SIDNEYFLEISCHERAND LESTER PACKER VOLUME LVII. Bioluminescenceand Chemiluminescence Editedby MARLENE A. DELUCA VOLUME LVIII. Cell Culture Editedby WILLIAM B. JAKOBYANDIRAPASTAN VOLUME LIX. Nucleic Acids and Protein Synthesis(Part Editedby KIVIEMOLDAVEANDLAWRENCEGROSSMAN

G)

xxvii

METHODSINENZYMOLOGY VOLUME LX. Nucleic Acids and Protein Synthesis(Part H) Editedby KIVIEMOLDAVEANDLAWRENCEGROSSMAN VOLUME 6 1. Enzyme Structure(Part H) EditedbyC. H. W. HIRSANDSERGEN.TIMASHEFF VOLUME 62. Vitamins and Coenzymes(Part D) Editedby DONALD B. MCCORMICKANDLEMUEL

D. WRIGHT

VOLUME 63. Enzyme Kinetics and Mechanism (Part A: Initial Rate and

Inhibitor Methods) DANIEL L. PURICH

Editedby

VOLUME 64. Enzyme Kinetics and Mechanism (Part B: Isotopic Probes

and Complex Enzyme Systems) DANIEL L. PIJRICH

Editedby

VOLUME 65. Nucleic Acids (Part I) Editedby LAWRENCEGROSSMANANDKIVIEMOLDAVE VOLUME 66. Vitamins and &enzymes (Part E) Editedby DONALD B. MCCORMICKANDLEMUEL

D. WRIGHT

VOLUME 67. Vitamins and Coenzymes(Part F) Editedby DONALD B. MCCORMICKANDLEMUEL

D. WRIGHT

VOLUME 68. RecombinantDNA Edited by RAY Wu VOLUME 69. Photosynthesisand Nitrogen Fixation (Part C) Editedby ANTHONYSANPIETRO VOLUME 70. Immunochemical Techniques(Part A) Editedby HELENVANVUNAKISANDJOHN J. LANGONE VOLUME 7 1. Lipids (Part C) Editedby JOHN M. LOWENSTEIN VOLUME 72. Lipids (Part D) Editedby JOHN M. LOWENSTEIN

...

xxvlll

METHODSINENZYMOLOGY

VOLUME 73. Immunochemical Techniques(Part B) Editedby JOHN J. LANGONEANDHELENVANVUNAKIS VOLUME 74. Immunochemical Techniques(Part C) Editedby JOHN J. LANGONEANDHELENVANVUNAKIS

75. Cumulative Subject Index Volumes XxX1, XxX11,

VOLUME

XXXIV-LX Editedby

EDWARDA.DENNISANDMARTHAG.DENNIS

VOLUME 76. Hemoglobins Edited by ERALDO ANTONINI, CHIANCONE

LUIGI ROSSI-BERNARDI, AND EMILIA

VOLUME 77. Detoxication and Drug Metabolism Editedby WILLIAM B. JAKOBY VOLUME 78. Interferons(Part A) Editedby SIDNEYPESTKA VOLUME 79. Interferons(Part B) Editedby SIDNEYPESTKA VOLUME 80. Proteolytic Enzymes(Part C) Editedby LASZLO LORAND VOLUME 81. Biomembranes(Part H: Visual Pigmentsand Purple Mem-

branes,I) Editedby

LESTERPACKER

VOLUME 82. Structural and Contractile Proteins (Part A: Extracellular Matrix) Editedby LEON W. CUNNINGHAMANDDIXIE W. FREDERIK~EN VOLUME 83. Complex Carbohydrates(Part D) Editedby VICTORGINSBURG VOLUME 84. ImmunochemicaI Techniques (Part D: Selected Immunoassays) Editedby JOHN J. LANGONEANDHELENVANVUNAKIS

METHODSINENZYMOLOGY

xxix

VOLUME 85. Structural and Contractile Proteins(Part B: The Contractile

Apparatusand the Cytoskeleton) Editedby DIXIE W. FREDERIKSEN ANDLEON W. CUNNINGHAM VOLUME 86. Prostaglandinsand ArachidonateMetabolites Editedby WILLIAM E.M. LANDS AND WILLIAM L. SMITH VOLUME 87. Enzyme Kinetics and Mechanism (Part C: Intermediates,

Stereochemistry,and Rate Studies) DANIEL L. PURICH

Editedby

VOLUME 88. Biomembranes(Part I: Visual Pigments and Purple Mem-

branes,II) Editedby

LESTERPACKER

VOLUME 89. CarbohydrateMetabolism (Part D) Editedby WILLIS A. WOOD VOLUME 90. CarbohydrateMetabolism (Part E) Editedby WILLIS A. WOOD VOLUME 9 1. Enzyme Structure(Part I) EditedbyC. H.W. HIRS ANDSERGEN.TIMASHEFF VOLUME 92. Immunochemical Techniques(Part E: Monoclonal Antibod-

ies and GeneralImmunoassayMethods) JOHN J. LANGONEANDHELENVANVUNAKIS

Editedby

VOLUME 93. Immunochemical Techniques(Part F: Conventional Anti-

bodies,Fc Receptors,and Cytotoxicity) JOHN J. LANGONEANDHELENVANVUNAKIS

Editedby

VOLUME 94. Polyamines Editedby HERBERTTABORANDCELIAWHITETABOR VOLUME 95. Cumulative SubjectIndex Volumes 6 1- 74,76 - 80 Editedby EDWARD A. DENNISANDMARTHAG.DENNIS VOLUME 96. Biomembranes[Part J: MembraneBiogenesis:Assemblyand

Targeting(GeneralMethods; Eukaryotes)] Editedby

SIDNEYFLEISCHERANDBECCAFLEISCHER

xxx

METHODSINENZYMOLOGY

VOLUME 97. Biomembranes [Part K: Membrane Biogenesis:Assembly

and Targeting(Prokaryotes,Mitochondria, and Chloroplasts)] Editedby


VOLUME 98. Biomembranes(Part L: Membrane Biogenesis:Processing

and Recycling) Editedby

SIDNEYFLEISCHERAND

BECCA FLEISCHER

VOLUME 99. Hormone Action (Part F: Protein Kinases) Editedby JACKIE D. CORBINANDJOELG.HARDMAN VOLUME 100.Recombinant DNA (Part B) Editedby RAY WV, LAWRENCEGROSSMAN,ANDKIVIEMOLDAVE VOLUME 101.RecombinantDNA (Part C) Editedby RAY WV, LAWRENCEGROSSMAN,ANDKIVIEMOLDAVE

102. Hormone Action (Part G: Calmodulin and CalciumBinding Proteins) Editedby ANTHONY R. MEANSANDBERT W. O'MALLEY

VOLUME

VOLUME 103.Hormone Action (Part H: NeuroendocrinePeptides) Editedby P. MICHAELCONN VOLUME 104.Enzyme Purification and RelatedTechniques(Part C) Editedby WILLIAM B. JAKOBY VOLUME 105.OxygenRadicalsin Biological Systems Editedby LESTERPACKER VOLUME 106.PosttranslationalModifications (Part A) Editedby FINN WOLD AND KIVIE MOLDAVE VOLUME 107.Posttranslational Modifications(PartB) Editedby FINN WOLDAND KIVIEMOLDAVE VOLUME 108. Immunochemical Techniques (Part G: Separation and

Characterizationof Lymphoid Cells) Editedby GIOVANNIDISABATO,JOHNJ.LANGONE,AND HELENVANVUNAKIS VOLUME 109.Hormone Action (Part I: PeptideHormones) Editedby LUTZBIRNBAUMERANDBERT W. O'MALLEY

METHODSINENZYMOLOGY

xxxi

VOLUME 110.Steroidsand Isoprenoids(Part A) Edited UPJOHN H. LAWANDHANSC.RILLING VOLUME 111.Steroidsand Isoprenoids(Part B) Edited ~~JOHNH.LAWANDHANSC.RILLING VOLUME 112.Drug and Enzyme Targeting(Part A) Edited ~~KENNETH J. WIDDERANDRALPHGREEN VOLUME 113. Glutamate, Glutamine, Glutathione, and Related Com-

pounds Editedby

ALTON MEISTER

VOLUME 114.Diffraction Methodsfor Biological Macromolecules(Part A) Edited by HAROLD W. WYCKOFF, C. H. W. HIRS, AND SERGE N. TIMASHEFF VOLUME 115.Diffraction Methodsfor BiologicalMacromolecules(Part B) Edited by HAROLD W. WYCKOFF, C. H. W. HIRS, AND SERGE N. TIMASHEFF VOLUME 116.Immunochemical Techniques(Part H: Effectersand Media-

tors of Lymphoid Cell Functions) Editedby GIOVANNI DI SABATO,JOHN J. LANGONE, AND HELEN VAN VUNAKIS VOLUME 117.Enzyme Structure(Part J) EditedbyC. H. W. HIRSANDSERGEN.TIMASHEFF VOLUME 118.Plant Molecular Biology Editedby ARTHUR~EISSBACH AND HERBERT WEISSBACH VOLUME 119.Interferons(Part C) Editedby SIDNEYPESTKA VOLUME 120.Cumulative SubjectIndex Volumes 8 1- 94,96 - 101 VOLUME 121. Immunochemical Techniques(Part I: Hybridoma Technol-

ogy and Monoclonal Antibodies) JOHN J. LANGONEANDHELENVANVUNAKIS

Editedby

VOLUME 122.Vitamins and &enzymes (Part G) Editedby FRANKCHYTILANDDONALD B. MCCORMICK

xxxii

METHODSINENZYMOLOGY

VOLUME 123.Vitamins and &enzymes (Part H) Editedby FRANKCHYTILANDDONALD B. MCCORMICK VOLUME 124.Hormone Action (Part J: NeuroendocrinePeptides) Editedby P. MICHAELCONN VOLUME 125.Biomembranes(Part M: Transport in Bacteria,Mitochon-

dria, and Chloroplasts:GeneralApproachesand Transport Systems) Editedby

SIDNEY FLEISCHER AND BECCAFLEISCHER

VOLUME 126. Biomembranes(Part N: Transport in Bacteria,Mitochon-

dria, and Chloroplasts:Protonmotive Force) Editedby

SIDNEYFLEISCHERAND

BECCAFLEISCHER

VOLUME 127.Biomembranes(Part 0: Protons and Water: Structureand Translocation) Editedby

LESTERPACKER

VOLUME 128. Plasma Lipoproteins (Part A: Preparation,Structure, and Molecular Biology) Editedby JERE P. SEGRESTANDJOHN J. ALBERS VOLUME 129.PlasmaLipoproteins(Part B: Characterization,Cell Biology,

and Metabolism) JOHN J. ALBERSANDJERE P. SEGREST

Editedby

VOLUME 130.Enzyme Structure(Part K) EditedbyC. H. W.HIRSAND SERGEN.TIMASHEFF VOLUME 131.Enzyme Structure(Part L) EditedbyC. H. W. HIRSAND SERGEN.TIMASHEFF VOLUME 132. Immunochemical Techniques (Part J: Phagocytosisand

Cell-MediatedCytotoxicity) Editedby

GIOVANNIDISABATOANDJOHANNESEVERSE

VOLUME 133.Bioluminescenceand Chemiluminescence(Part B) Editedby MARLENEDELUCAANDWILLIAM D. MCELROY VOLUME 134.Structuraland Contractile Proteins(Part C: The Contractile

Apparatusand the Cytoskeleton) RICHARD B. VALLEE

Editedby

METHODSINENZYMOLOGY

... xxxm

VOLUME 135.Immobilized Enzymesand Cells(Part B) Editedby KLAUS MOSBACH VOLUME 136.Immobilized Enzymesand Cells (Part C) Editedby KLAUS MOSBACH VOLUME 137.Immobilized Enzymesand Cells (Part D) Editedby KLAUSMOSBACH VOLUME 138.Complex Carbohydrates(Part E) Editedby VICTORGINSBURG VOLUME 139. Cellular Regulators(Part A: Calcium- and Calmodulin-

Binding Proteins) Editedby

ANTHONY R. MEANSAND P. MICHAELCONN

VOLUME 140.Cumulative SubjectIndex Volumes 102- 119, 121- 134 VOLUME 141.Cellular Regulators(Part B: Calcium and Lipids) Editedby P. MICHAELCONNANDANTHONY R. MEANS VOLUME 142.Metabolism of Aromatic Amino Acids and Amines Editedby SEYMOURKAUFMAN VOLUME 143.Sulfur and Sulfur Amino Acids Editedby WILLIAM B. JAKOBYAND OWENGRIFFITH VOLUME 144. Structural and Contractile Proteins (Part D: Extracellular

Matrix) Editedby

LEON W. CUNNINGHAM

VOLUME 145. Structural and Contractile Proteins (Part E: Extracellular

Matrix) Editedby

LEON W. CUNNINGHAM

VOLUME 146.PeptideGrowth Factors(Part A) Editedby DAVID BARNES AND DAVID A. SIRBASKU VOLUME 147.PeptideGrowth Factors(Part B) Editedby DAVIDBARNESANDDAVID A. SIRBASKU VOLUME 148.Plant Cell Membranes Editedby LESTERPACKERANDROLANDDOUCE

xxxiv

METHODSINENZYMOLOGY

VOLUME 149.Drug and Enzyme Targeting(Part B) Edited ~~RALPHGREENANDKENNETH J. WIDDER VOLUME 150.Immunochemical Techniques(Part K: In Vitro Models of B and T Cell Functions and Lymphoid Cell Receptors) Editedby GIOVANNI DI SABATO VOLUME 151. Molecular Geneticsof Mammalian Cells Editedby MICHAEL M. GOTTESMAN VOLUME 152.Guide to Molecular Cloning Techniques Editedby SHELBY L. BERGERANDALAN R. KIMMEL VOLUME 153.Recombinant DNA (Part D) Editedby RAY Wu ANDLAWRENCEGROSSMAN VOLUME 154.Recombinant DNA (Part E) Editedby RAY WV ANDLAWRENCEGROSSMAN VOLUME 155.RecombinantDNA (Part F) Edited by RAY WV VOLUME 156. Biomembranes(Part P: ATP-Driven Pumps and Related Transport: The Na,K-Pump) Editedby


VOLUME 157. Biomembranes(Part Q: ATP-Driven Pumps and Related

Transport:Calcium, Proton, and PotassiumPumps) Editedby


VOLUME 158.Metalloproteins(Part A) Editedby JAMES F. RIORDANANDBERT

L. VALLEE

VOLUME 159.Initiation and Termination of Cyclic NucleotideAction Editedby JACKIE D. CORBINANDROGER A. JOHNSON VOLUME 160.Biomass(Part A: Celluloseand Hemicellulose) Editedby WILLIS A. WOOD AND SCOTT T. KELLOGG VOLUME 161.Biomass(Part B: Lignin, Pectin,and Chitin) Editedby WILLIS A. WOODANDSCOTT T. KELLOGG

METHODSINENZYMOLOGY

xxxv

VOLUME 162. Immunochemical Techniques (Part L: Chemotaxis and

Inflammation) Editedby

GIOVANNI DI SABATO

VOLUME 163. Immunochemical Techniques (Part M: Chemotaxis and

Inflammation) Editedby

GIOVANNI DI SABATO

VOLUME 164.Ribosomes Editedby HARRY F. NOLLER,JR.ANDKIVIEMOLDAVE VOLUME 165.Microbial Toxins: Tools for Enzymology Editedby SIDNEYHARSHMAN VOLUME 166.Branched-ChainAmino Acids Editedby ROBERTHARRISANDJOHN R. SOKATCH VOLUME 167.Cyanobacteria Editedby LESTERPACKERANDALEXANDERN.GLAZER VOLUME 168. Hormone Action (Part K: NeuroendocrinePeptides)(in

preparation) P. MICHAELCONN

Editedby

VOLUME 169.Platelets:Receptors,Adhesion,Secretion(PartA) (in prepa-

ration) Editedby

JACEKHAWIGER

VOLUME 170.Nucleosomes(in preparation) Editedby PAUL M. WASSARMANANDROGER D. KORNBERG VOLUME 171. Biomembranes(Part R: TransportTheory: Cellsand Model

Membranes)(in preparation) Editedby


VOLUME 172.Biomembranes(Part S: Transport Theory: Membrane Iso-

lation and Characterization)(in preparation) Editedby

SIDNEY FLEISCHER AND BECCAFLEISCHER

[ 1]

IMAGE PROCESSING

OF SINGLE RIBOSOMES IMAGED

BY

EM

3

[ 1] Studying Ribosome Structure by Electron

Microscopy and Computer-Image Processing

By JOACHIM FRANK, MICHAEL RADERMACHER, TERENCE WAGENKNECHT, and ADRIANA VERSCHOOR Introduction T o date, the most detailed knowledge o f ribosomal morphology has been obtained by the combination o f electron microscopy and single-particle averaging ~-3 and three-dimensional (3-D) reconstruction methods. The term single-particle averaging refers to a m e t h o d whereby m a n y (e.g., several hundred) images o f a macromolecular structure appearing in a characteristic orientation are averaged after precise correlation alignment. Thus, this m e t h o d is equivalent to the averaging, by Fourier techniques, o f a micrograph showing a flat, two-dimensional protein crystal. 4-6 The averaging is required because the individual images o f a macromolecular structure obtained in the electron microscope contain a large a m o u n t o f noise, allowing the significant part o f the image to be extracted only from a set o f repeated measurements. Because o f the presence o f systematic variations a m o n g the individual projections o f a structure, e.g., due to variations in orientation or a m o u n t o f staining, a multivariate statistical analysis is frequently necessary before a meaningful average (or several averages) can be formed. 7,8 In the m e t h o d o f single-particle 3-D reconstruction, 9,'° the macromolecule is reconstructed in three dimensions from a large n u m b e r o f projections. In the variant o f this m e t h o d most viable for structure research, these

J. Frank, W. Goldfarb, D. Eisenberg,and T. S. Baker, Ultramicroscopy 3, 283 (1978). 2j. Frank, A. Verschoor, and M. Boublik, Science 214, 1353 (1981). 3j. Frank, A. Verschoor,and T. Wagenknecht, in "New Methodologiesin Studies of Protein Configuration" (T. T. Wu, ed.), p. 36. Van Nostrand-Reinhold, New York, 1985. 4 H. P. Erickson and A. Klu~, Philos. Trans. R. Soc. London, Ser. B 261, 105 (1970). 5p. N. T. Unwin and R. Henderson, J. Mol. Biol. 94, 425 (1975). 6 L. A. Amos, R. Henderson, and P. N. T. Unwin, Prog. Biophys. Mol. Biol. 39, 183 (1982). 7M. van Heel and J. Frank, Ultramicroscopy 6, 187 (1981). s j. Frank and M. van Heel, J. Mol. Biol. 161, 134 (1982). 9 W. Hoppe, J. Gassmann, N. Hunsmann, J. Schramm, and M. Sturm, Hoppe-Seyler's Z. Physiol. Chem. 355, 1483 (1975). lo W. Hoppe and R. Hagerl, in "Computer Processingof Electron MicroscopeImages" (P. W. Hawkes, ed.), p. 127. Spfinger-Vedag Berlin, Federal Republic of Gcrmany, 1980. METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All l~hts of~produclion in any form reaerved.

4

ELECTRON MICROSCOPY

[ 1]

projections come from different particles lying in, or brought into, different orientations providing a wide range of viewing directions. H-m In principle, electron crystallographic techniques6,15are capable of providing structural data on ribosomes. 16,17However, a resolution comparable to that obtained by single-particle reconstruction has not been achieved, mainly because large, well-ordered crystals of ribosomes or ribosomal subunits have been difficult to grow. Electron Microscopy Currently, most electron microscopy studies of ribosomes are done using any one of several modifications of the method of negative staining first developed by Valentine and Green. m Immunoelectron microscopic studies of ribosomes are usually done on specimens that axe sandwiched, along with stain, between two layers of carbon? 9-2~ These so-called double-carbon layer preparations give improved contrast for antibodies bound to ribosomes or ribosomal subunits as compared to the single-carbon layer preparations. Fortunately, these same methods of negative staining can be used to obtain images suitable for the application of single-particle imageprocessing techniques. We have found that, under the proper conditions, the double-carbon layer methods give the most consistent results and provide the most detailed information on the interiors of the particles, whereas the single-layer regions yield better defined outlines of the particles, which are therefore more easily aligned by correlation methods. We prefer the methods described by Tischendorf et al. 2~,22and Boublik et al., ~9 mainly because they usually result in a large proportion of the grid surface being double-layered. II A. Verschoor, J. Frank, M. Radermacher, and T. Wagenknecht, J. Mol. Biol. 178, 677 (1984). 12 M. Radermacher, T. Wagenkneeht, A. Verschoor, and J. Frank, J. Microsc. 141, RPI (1985). ~3M. Radermaeher, T. Wagenkneeht, A. Verschoor, and J. Frank, J. Microsc. 146, 113 (1987); E M B O J. 6, 1107 (1987). t4 G. Harauz and F. P. Ottensmeyer, Ultramicroscopy 12, 309 (1984). t5 R. Henderson and P. N. T. Unwin, Nature (London) 257, 28 (1975). t6 W. Kiihlbrandt and P. N. T. Unwin, J. Mol. Biol. 156, 439 (1982). 17A. Yonath, Trends Biochem. Sci. 9, 227 (1984). ~s R. C. Valentine and M. Green, J. Mol. Biol. 27, 615 (1967). ~9M. Boublik, W. Hellman, and A. K. Kleinschmidt, Cytobiologie 14, 293 (1977). 2o j. A. Lake, J. Mol. Biol. 105, 131 (1976). 2t G. W. Tisehendoff, H. Zeiclahard~ and G. St6ttler, Mol. Gen. Genet. 134, 187 (1974). 22 G. St0mer and M. StOtiler-Meilieke, in "Modern Methods in Protein Chemistry" (H. Tsehesehe, ed.), p. 409. de Gruyter, Berlin, Federal Republic of Germany, 1983.

[ 1]

IMAGE PROCESSING OF SINGLE RIBOSOMES IMAGED BY

EM

5

The conditions required for single-particle averaging and three-dimensional reconstruction are, however, more stringent than those for simple visual analysis of images. Ideally, the particles should be identically stained, they should be embedded in stain (as opposed to just being surrounded by a ring of stain24), and they should not be flattened or collapsed. The double-layer preparations are clearly superior to the single-layer preparations in fulfilling the first of these criteria. However, micrographs obtained with the double-layer technique often contain a very thin layer of stain appearing as a ring of high density surrounding the particles and the particles appear larger than those in the single-layer preparations, implying that some particle flattening may have occurred; we do not analyze micrographs that exhibit these characteristics. We have obtained the highest resolutions with specimens which were prepared by one of the double-carbon layer procedures and in which the stain layer is thick, at least in the immediate vicinity of the particles. In micrographs of this type the stain thins out gradually with distance from the boundaries of the particles such that in areas where the particles are clustered together the stain appears almost uniformly distributed over the grid surface. It should be pointed out that the methodology for obtaining the desired staining characteristics has not been perfected, and often many grids have to be prepared and examined before a suitable set of micrographs is obtained. In order to attain the highest resolution it is desirable to minimize the electron dose applied to the specimen. However, we often observe severe specimen drift (detectable in the optical diffraction patterns of the carbon support films25) when attempting minimal dose microscopy on specimens prepared by the double-carbon layer techniques. It is not unusual to encounter regions of the grid from which virtually all of the micrographs arc unusable because of drift. In order to increase the stability of the specimen so as to minimize drift, we do not use "naked" grids when applying the specimen and carbon support films but, instead, the grids are coated with a thick carbon film densely packed with holes. Only those regions of the micrographs in which the sandwiched specimen is suspended over the holes in the thick carbon film are considered for further analysis. An optimal dose for negatively stained specimens6,26,27 is about 23 Deleted in proof. 24 H. Oettl, R. Hegerl, and W. Hoppe, J. Mol. Biol. 163, 431 (1983). 25 j. Frank, Optik 30, 171 (1969). 26 p. N. T. Unwin, J. Mol. Biol. 87, 657 (1974). 27 T. S. Baker and D. A. Goodenough, J. Cell Biol. 96, 204 (1983).

6

ELECTRON MICROSCOPY

[ 1]

1000 el/nm 2 which, fortunately, can be achieved at medium magnifications (e.g., 40,000-50,000X) without significant loss of signal-to-noise ratio due to electron shot noise. 2s It is feasible to use even lower doses (100 el/nm 2) by recording a pair of micrographs, one at the desired dose and a second at a higher dose and containing sufficient power to determine the translational and rotational alignment parameters for each particle which can then be applied to the low-dose micrograph images. 29 The latter approach might be required, for example, in the analysis of unstained frozen-hydrated ribosomes or subunits. Digitization and Selection of Images The electron micrograph is digitized with a microdensitometer into an array of optical density (OD) values, which are stored on a magnetic tape and read into the computer. 3° In the experimentally important OD range, the OD measured in a given picture clement (pixel) indexed j, p(ri) is proportional to the electron intensity I(rj) at that location. For bright field images of sufficiently thin biological particles obtained in the conventional transmission electron microscope with appropriate defocus, the contrast, i.e., the local variation of the electron intensity A/j ---/j - i relative to the mean intensity i, is a measure of the projected potential distribution. A/j and thus the local OD variation Apj -- pj --/~, is roughly proportional to the total projected mass at location j.31 One of the crucial prerequisites for the image to be interpreted as a projection of the object is that the defocus range be properly selected, such that all spatial frequencies up to the significant resolution (see below) lie essentially in one contrast transfer band. An incorrect choice of focus causes certain spatial frequency bands to be suppressed or transferred with false contrast. 32 Although restoration techniques can be used to correct 2s K. H. Downing and D. A. Grano, Ultramicroscopy 7, 381 (1982). 29 M. Kessel, J. Frank, and W. Goldfarb, J. Supramol. Struct. 14, 405 (1980). 3oj. Frank, in "Advanced Techniques in Electron Microscopy" (J. K. Koehler, ed.), p. 215. Springer-Verlag, Berlin, Federal Republic of Germany, 1973. 31 The transfer of structural information from the object into the image is governed by the phase-contrast transfer theory. The relationship between object projection and image contrast is straightforward only in the resolution range considered here (> 2 nm) and only when a number of simplifying assumptions are made. Among these are that the object is thin, and that its interaction with the electrons can be described as a simple phase shift ("pure phase object"). For details, see, e.g., F. Lenz, in "Electron Microscopy in Material Science" (U. Valdr6, ed.), p. 542. Academic Press, New York, 1971. 32 F. Thon, Z. Naturforsch. 21a, 476 (1966).

[ 1]


EM

7

images having incorrectly placed transfer bands, a3-35 it has proved more practical to select "good" images from a large number of experimental mierographs with varying quality. Other experimental parameters affecting the quality of the image are axial astigmatism and drift. 25 Microgmphs that are to be analyzed are therefore routinely checked in the optical diffractometer for proper defocus and absence of controllable aberrations. 36 In order for details in the range of 2 n m in size--which have been proved to be reproducible in averaged projections of ribosomal subunits prepared by the negative staining technique2--to be represented in the digitized image, the scanning step of the microdensitometer should be equivalent to a value of 0.5 to 0.7 nm. The factor of between 2 and 1.5 × over the 1 n m step theoretically required provides a safety margin that prevents deterioration of image resolution in computational steps requiring interpolation; e.g., image rotation. Although programs for automatic selection of particles from micrograph fields have been d e v e l o p e d , 37-39 interactive cursor-controlled selection by means of a graphics terminal is the most efficient procedure. After this selection, the particle images exist as a set of arrays {pi(rj), j = 1 . . . P} stored in separate files on the disk of the computer, pi(rj) stands for the OD value of the jth pixel in the/th image. The need to scale micrographs arises in many applications, e.g., averaging of projections selected from different micrographs, comparison of different specimens, and 3-D reconstruction from a large number of projections. The scaling procedure is based on the known properties of bright field_images of weak phase objects. 4° Under these conditions, the contrast AIj/I is independent of the mean intensity I. Consequently, when two particles (p~j and P2j, J = 1 . . . P} are selected from different micrographs, comparable density values are obtained by

plj=plj/pl;

p2j = p2j-//~2

(1)

where/~l and/S2 are the mean densities measured by averaging portions of the micrographs that do not contain particles. 33 W. Hoppe, Acta Crystallogr. A 26, 414 (1970). O. Kfibler, M. Hahn, and J. Seredynski, Optik 51, 171 (1978). 35 T. A. Welton, Adv. Electron. Electron Physics 48, 37 (1979). 3~ B. V. Johansen, Princ. Tech. Electron Microsc. 5, 114 (1975). 37 j. Frank and T. Wagcnknecht, Ultramicroscopy 12, 169 (1984). 3s M. van Heel, Ultramicroscopy 7, 331 (1981). 39 D. W. Andrews, A. H. C. Yu, and F. P. Ottensmeyer, Ultramicroscopy 19, 1 (1986). 4o F. Lenz, in "Electron Microscopy in Material Science" (U. Valdr~, ed.), p. 542. Academic Press, New York, 1971.

8

ELECTRON MICROSCOPY

[ 1]

Alignment of Images The objective of the alignment procedure 1,2 is to determine coordinate transformations Ci ----TiRj (Ti translation, Ri rotation matrix) such that N

P

~ [pt(Cir/) - pl,(Crrj)] 2 ~= rain

(2)

i> i' j - - 1

where N is the number of images. This objective is achieved, to a good approximation, through application of an iterative procedure. First, the error sum, or Euclidean distance P

ET, = ~ [pi(C; rj) - pr(rj)] 2

(3)

j--I

is minimized between each image p~ and a suitably chosen "typical" reference image, denoted by p,. Subsequently, E~ = ~ [pi(C] rj) - p° (rj)] 2

(3a)

j-I

where the new reference p ° (rj) -- 1/N Y~_ ~p~(C~ rj) is the average after the first cycle. This procedure is repeated several times until the error sum E~o of Eq. (3a) or a similar measure of discrepancy no longer shows a significant change. Similarly, as for molecular search methods of X-ray crystallography, 4~ the minimization of the error sum for two images p~(r) and p2(r) is achieved by maximizing a function

x(c) = fa pl(Cr)p2(r) dr

(4)

which is obtained by integration over the image domain A. ~ C ) is closely related to the two-dimensional cross-correlation function (CCF), which is used for translation search only (see below). For increased speed, the three-dimensional search (with respect to x, y, and ~b) is split into a two-dimensional search of translation (x,y) yielding T and a one-dimensional search of rotation (angle ~b) yielding R.l'~2 Optimal alignment of an image set may be achieved by combining four basic techniques (Fig. 1). Initially, the time-saving decoupling of rotation and translation searches is accomplished by centration 13 (A) or by the use of a translation-invariant rotation search (B). ~ Subsequently, the transla41 E. Lattman, this series, Vol. 115, p. 55. 42 R. Langer, J. Frank, A. Feltynowski, and W. Hoppe, Ber. Bunsenges. Phys. Chem. 74, 1120 (1970).

[1]


A

EM

9

B

CENTER PARTICLE BY 2-D CROSS-CORRELATION WITH A CENTERED, LOW-PASS FILTERED DISK

SEARCH ORIENTATION BETWEEN ACFs OF PARTICLES

C

D

SEARCH ORIENTATION BETWEEN CENTERED PARTICLES ("DIRECT METHOD")

•

ALIGN BY 2-D CROSS-CORRELATION

FIG. 1. Four components of the single-particle alignment procedure (see text).

tion and rotation parameters are refined by alternation between "direct" rotational alignment (C) 2,43and two-dimensional cross-correlation ( D ) . t'44 Examples for common search strategies are, in operator notation to be read from right to left, • . . [C o D] o [C o D] o [C o D] o [C o A]

and

(5) • . . [D o C] o [D o C] o [D o C] o [D o B]

Starting from the right, each term in brackets stands for one alignment pass that uses the average of the previous pass as reference. As stated above, the initial reference is an original, unaveraged image deemed typical. When several dissimilar types of projections are known to be present, several reference images must be tried for each image, and the selection of the appropriate class is then made on the basis of the highest cross-correlation coefficient (multireference alignment45'4*). Both translational and rotational searches are done by fast Fourier techniques. The values of the translation parameters are determined from 43 M. Steinkilberg and H. Schramm, Hoppe-Seyler's Z. Physiol. Chem. 361, 1363 (1980). 44j. Frank, in "Computer Processing of Electron Microscope Images" (P. W. Hawkes, ed.), p. 187. Springer-Verlag, Berlin, Federal Republic of Germany, 1980. 45 M. van Heel, J.-P. Bretaudiere, and J. Frank, Proc. Int. Congr. Electron Microsc., lOth 1, 563 (1982). 46 M. van Heel, Proc. Fur. Congr. Electron Microsc., 8th 2, 1317 (1984).

10

ELECTRON MICROSCOPY

[ 1]

• (r') = fA pl(r + r')p2(r) dr

(6)

the CCF:

where r' is the argument vector describing the translation between the two images p~ and P2. For the fast computation of the CCF the Fourier theorem is exploited47: O(r') -- ~ - ~ ( 3 ~ (r)]3~*[P2(r)]}

(7)

where 3~ and 3~-~ denote the forward and reverse Fourier transformations, respectively, and the asterisk stands for formation of the complex conjugate. /~(r) and P2(r) are derived from the original images (defined by domain A0 by padding into a field A2 whose size is large enough to avoid circular overlapS:

.~(r) = fp(r)

r ~ Al

r ~ A2

(8)

The value of the constant c is chosen to be equal to the mean of the pixels along the boundary of A~ to minimize the density step. (For bright field images, the mean of all pixels lying within A~ may be used instead with little practical difference.) The vector pointing from the origin of the CCF to its maximum is the vector by which Pu is translated relative to pt. For the computation of a rotational search function R(C~) -- fA W(I r I)f~(C+r)f2(r) dr,

(9)

where C+ is the probing rotation matrix and I41(Irl) a weighting function, the two functionsf~ and f2 to be matched are represented on a polar grid by reinterpolation from the cartesian grid, and the Fourier theorem is used, for increased speed, to compute one-dimensional CCFs along each polar coordinate ring?4 The rotational search function ("rotation function") is then obtained by weighted summation over all one-dimensional CCFs A~')(q~); r = [rl:

R(dp) -- ~, W(r) A(lf(q~)

(10)

¢

(see Crowther,4s where a similar simplification is used). The weighting function is chosen such that ring zones whose contributions are most sensitive to rotation receive highest weights. For algorithm C above (Fig. 1), the functionsft and f2 in Eq. (9) are the images themselves, whereas in +7G. D. Bergland, I E E E Spectrum p. 41 (1969). 48 R. A. Crowther, Int. Sci. Rev. Ser. 13, 10 (1972).

[1]

I M A G E P R O C E S S I N G OF SINGLE RIBOSOMES I M A G E D BY

EM

11

algorithm B the autocorrelation functions (ACF) of the images are used instead, and the search extends over only half the angular range. The rotational search may result in several peaks of comparable size. In particular, algorithm B (Fig. 1) leads to an intrinsic 180-degree ambiguity due to the centrosymmetry of the ACF. Such ambiguities are resolved by computing the CCF for each orientation, and then selecting the rotation angle that produces the maximum CCF peak.49 Averaging After alignment (denoted by the final cumulative coordinate transformations CD, projections relating to the same viewing directions may be averaged: N

p(rj) =

1/N ~

p,(C~'rj),

(11)

i--I

which results in an image p(rj) with greatly reduced noise contributions (by the factor Nm). As the sum of Eq. (11) is formed, the so-called variance map 2 is also computed N

v(rj) ---- E

[Pi(CTrj) -- p(rj )]2

(12)

i--I

This map is very informative as it gives an account of the location of regions where high discrepancies occur among images. This typically happens near the stain boundaries and at sites of the particle where positional changes of flexible components are encountered (Fig. 2). Resolution and Reproducibility of Averaged Projections Data obtained by single-particle averaging techniques lack an in-built resolution measure, in contrast to those obtained from crystals, for which diffraction orders can be counted. For determining the reproducible resolution ("cross-resolution"), two independently derived averages are compared in Fourier space: if F~(k) and F2(k) are the Fourier transforms of such averages, the differential phase residual2

"~I/2

A-~k, Ak)= [~'l[Adp(k)]2[]F'(k)]+[F2(k)[][ [~ak, [IFl(k)l + Ie2(k)l]

(13)

J

49 H. P. Zingsheim, D. C. Neugebauer, F. J. Barrantes, and J. Frank, Proc. Natl. Acad. Sci. U.S.A. "17, 952 (1980).

12

ELECTRON MICROSCOPY

[ 1]

a

b

Fro. 2. Averaging of the crown view of the 50S ribosomal subunit from E. coil [A. Verschoor, J. Frank, and M. Boublik, J. Ultrastruct. Res. 92, 180 (1985)]. (a) Portion of micrograph field from single-layer preparation. Scale bar 50 nm. (b) Examples of selected images (first six frames), and average and variance images (seventh and eighth frames, respectively; N = 60). The average image and variance map are displayed with a limiting resolution of 3 nm. Inspection of the variance map (final frame in b) shows that the singlelayer preparation permits a large amount of variation in the position of the L7/LI2 stalk (arrow) and in the amount of peripheral staining. The tip of the stalk has a low contrast and so low variance despite its large positional variation.

with A~b(k)----phase[Fl(k)]- phase[F2(k)] is computed by summation over rings k + A/c, k = Ikl in the Fourier plane, and plotted as a function of ring radius k. Although the entire curve is needed to accurately describe the reproducibility of the averages, it is convenient to state a tingle figure, k45, for which Ad~(k45, Ak) = 45 ° (Fig. 3). It must be stressed that in electron microscopy of stained specimens, the global equivalent to Eq. (13) obtained by summation from 0 to km~5° is not very informative, since the structure factor F(k) falls off, and the phase difference between two transforms increases rapidly, as a function of k. In addition to the phase residual of Eq. (13), another differential measure of cross-resolution, the Fourier ring correlation (FRC), 5~,52 is fre5o p. N. T. Unwin and A. Klu& J. Mol. Biol. 87, 641 (1974). 51 W. O. Saxton and W. Baumeister, J. Microsc. 127, 127 (1982). ~2M. van Heel, W. Keegstra, W. Schutter, and E. F. J. van Bruggen, in "Structure and Function of Invertebrate Respiratory Proteins" (E. J. Wood, ed.), p. 69. Harwood, 1982.

[ 1]

I M A G E P R O C E S S I N G OF SINGLE RIBOSOMES I M A G E D BY

Phase Residual

EM

13

•

AO

60

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

. . . . . . . . . . . . . . . . . . . . . . . . .

40'

20.

1)4o

'1)2o

A-'

FI~. 3. Resolution determination by phase residual analysis. The particle analyzed is the 50S ribosomal subunit from E. coli prepared by the double-carbon layer method, showing the crown view.t3 (a) Phase residual plot obtained with two independent subset averages (N = 245 each). The 45* limit is reached for a resolution of 2.3 nm (arrow). (b) The two subset averages and the total average (N = 490) limited to the resolution found in (a), displayed as a contoured halftone representation.

quently used: FRC(k, Ak) - -

E F| (k)F2(k)} {[k'ak]

Here the resolution is defined as the spatial frequency for which FRC(k~t, Ak) = [2 Ak/(~k,~t)] 1/2

(14)

(15)

14

ELECTRON MICROSCOPY

[ 1]

Comparisons of k45 and k~t for selected data sets indicate, however, that k~tlt is as a rule larger, and A$(k~it, Ak) is larger than 80 °, suggesting that the FRC criterion may be too liberal for the purpose of resolution estimation.53,54 Statistical Significance of Averages The statistical significance of structural features in an averaged image can be assessed by applying standard statistical tests to the individual image elements) ,55 Assuming Gaussian statistics and that the individual images used to compute the average were correctly aligned, the standard error of the mean for the jth picture element of the average is Em[/~(ri)] = [v(rj)/N] 1/2

( 16)

where v(rj) is the variance map introduced before in Eq. (12). The standard error of the difference between two averaged picture elements is given by Ed [/~1(r~,/~2(rk)] = IV1( r j ) / N 1 + v2(rk)/N2] 1/2

(17)

(N1 and N2 are the numbers of images averaged in each case). The two picture elements may be different elements from a single average (pl ffi/72; j ~ k) or corresponding elements from two different averages (p~ * P2; j - k). The latter case occurs in studies where it is desired, e.g., to map the location of an antibody or other macromolecule (label) bound to a ribosomal subunit by quantitatively comparing averaged images of the labeled and control subunits. Ideally, the only significant differences between corresponding pixels in the two averages will occur at the location of the bound label. As a rule of thumb, we consider a difference between two averaged picture elements to be significant if it exceeds the standard error by a factor of at least three; this factor corresponds to a significance level of 0.2%. It is usually necessary to scale the ODs when comparing averages obtained from different micrographs (see Digitization and Selection of Images). Differences between averages that are strongly dependent upon the scaling parameters should be viewed with suspicion. It should be emphasized that statistically significant differences among averages obtained from different micrographs do not necessarily corre53 M. Radermacher, J. Frank, and C. A. Mannella, Proc. Annu. Meet. Electron Microsc. Soc. Am., 44th, p. 140 (1986). J. Frank, M. Radermacher, T. Wagenknccht, and A. Verschoor, Ann. N.Y. Acad. Sci. 483, 77 (1986). 55 H. P. Zingsheim, F. J. Barrantes, J. Frank, W. H~inicke, and D. C. Neugebauer, Nature (London) 299, 81 (1982).

[1]


EM

15

spond to genuine differences in macromolecular structure. The detected differences could be due, for example, to variation in the mode of staining in the micrographs. When averaged images of modified (e.g., labeled) and control specimens are being compared, ambiguities of this type can be avoided by analyzing micrographs consisting of a mixture of the two specimen types. It is sometimes feasible, using the technique of correspondence analysis (see following section), to distinguish the two populations present in a tingle micrograph. 3 Alternatively, the aligned images of a mixed population can be averaged and a variance map [Eq. (12)] computed; the differences detected by comparing the two homogeneous populations will be visible in the variance map of the mixed population, provided that the optical density differences due to the label are significantly larger than the random deviations in density due to noise among the images. 56 If there is reason to suspect that the individual images used to compute an average were not correctly aligned, then nonparametric statistical tests may be applied to check the validity of features in the average imageY These tests are quite general, making no assumption on the validity of the statistics of the noise.

Multivariate StatisticalAnalysis Images aligned by the methods sketched out above may not represent particleshaving preciselythe same state of preservation or distributionof stain. Flexible components may occur in differcntpositions. "Rocking" movements around thc characteristicorientation may produce variations in the particle'sprojected appearance. For screening the images without reference to a preconceived model of the structure,correspondence analysis,58,59 a branch of multivariatc statisticalanalysis, is used in the way introduced by van Heel and FrankV,8:the basic philosophy of thisapproach isthat aligned images containing P pixclscan be described as vectors in R ~, a P-dimensional vector space, since, by virtue of the alignment, pixcls fallingonto the same grid point refer to the same absolute location in the molecule's coordinate system. Thc end points of thc vectors dcscribing the images form a "cloud." Points lying closely together belong to images that are closely similar. Thus, the structureof the point cloud (i.e.,itsshape and internalstructure) 56T. Wagenknecht and J. Frank, Biochemistry 23, 3383 (1984). 57W. H~inicke,J. Frank, and H. P. Zingsheim, J. Microsc. 133, 223 (1983). 5sL. Lebart, A. Morineau, and K. M. Warwick, "Multivariate DescriptiveStatistical Analysis." Wiley,New York, 1984. ~9j. p. Benzecri, in "Methodologies of Pattern Recognition" (S. Watanabe, ed.), p. 35. Academic Press, New York, 1969.

ELECTRON MICROSCOPY

16

[ 1]

reflects the existence and relative configurations of classes among the images. The multivariate statistical analysis provides a means for identifying, describing, and separating these classes in a quantitative way. In correspondence analysis of aligned particle projections pi(rj), a matrix with the general element f/J=

N

Pi(rfl P

(18)

E E p,(r,)

i--lj--I

is created. Based on a chi-squared distance metric, a covariance matrix with the general term N

Sjj, = E ft'(:J)l/2(fiJ/f]°:J

-

i-I P

1)(f'J')lY2(ftJ'/fi'f'J"-

1)

(19)

N

Y,-= Ef J; j-i

;:.J= F_,Y,J t-i

is formed. The row vectors {f0; J = 1 . . . P) are referred to as image profiles. The chi-squared metric ensures that the multivariate analysis is independent of the scaling of the individual images. Specifically, brightfield images taken of the same object with different exposures differ by a scalar factor only, and are thus represented by the same point. The matrix S of Eq. (19) is analyzed for its eigenvectors, or factors. These factors form a new set of basis vectors (~mj;J = 1 . . . P) in the P-dimensional vector space, which are ranked by the size of their eigenvalues, with the first factor pointing in the direction of largest extension of the data cloud, the second pointing in the direction of largest extension orthogonal to the first, and so forth. Thus these factors may be used to represent the dispersion or shape of the cloud formed by the rescaled images (f0;J - 1 . . . P}, originally in R e, in a space with greatly reduced dimensionality. In this space, each image is described by a rr6~-tupel of coordinates, (x~,; m - 1 . . . m~O. A new, condensed representation (reconstitution) of the images may be obtained in the form of linear combinations of the m , ~ high-ranking factors59: m~

#i(rj) = ~

Iqj.,%.j

(20)

m--O

The vectors {e~j =f.~mj; J---- 1 . . . P} may be thought of as eigenim-

[ 1]


EM

17

ages,~°,6~and the sum in Eq. (20) has a close analogy to a Fourier synthesis from a limited number of orders of reflections. The expression in Eq. (20) can also be formed over any subset of factors, selectively omitting contributions from those factors that express structurally irrelevant features; e.g., variations in the amount and spatial distribution of peripheral staining. 62 To describe the shape of the point cloud formed by the image profile vectors in R e, and to define and isolate homogeneous subsets of the images, one of three methods of interpretation based on the factorial coordinates may be used: (1) inspection of 2-D factor maps, e.g., xi~ versus x,~; (2) identification of dusters and classification; and (3) nonlinear mapping. Which of these methods will be most informative and effective in isolating subsets depends on the type of variations present in the projection set. If strong clustering or a predominant trend occurs, one or two factor maps may be sufficient to identify classes of images (Fig. 4a). If, however, the clusters are separated or stretched in three or more dimensions, then the factor maps may be difficult to interpret. A similar problem arises in the analysis of data exhibiting continuous variationsn especially if linked to a continuous value range of a single parameter (such as tilt angle of a cylindrical particle about the cylinder's axis) n which require determination of the seriation of the data.H The methods dealing with these problems, classification and nonlinear mapping, will be briefly outlined below. Identification of Clusters and Classification Through the alignment of an image series and the subsequent multivariate statistical analysis of the resulting aligned image set, the data are in a form (as points in an up to 8-dimensional vector space, which is the number of factors often used in practice) that allows a variety of clustering procedures to be carried out. The objective is to find clusters, or compact accumulations of points, and to characterize these in terms of their compactness and relative positions. According to their relative locations, the initial clusters can be grouped into classes in a hierarchical manner, and thus the determination of the hierarchy of class associations is required for a full description of the data structure. This second step of cluster analysis is termed classification. Once clusters and classes have been identified, characteristic "average" 6o j._p. Bretaudiere and J. Frank, J. Microsc., 144, 1 (1986). 6~ M. van Heel, in "Pattern Recognition in Practice" (E. S. Gelsema and L. N. Kanal, eds.), Vol. 2. North-Holland, Amsterdam, 1985. 62 j. Frank, A. Verschoor, and M. Boublik, J. Mol. Biol. 161, 107 (1982).

18 a

ELECTRON MICROSCOPY : .......

=,. . . . . . .

. .......................

.( .................

[ I]

-E: ...............................

~ - - - l - - ~ - ~

2 "~ •.1

.: . . . . .

•. \

I ,

,

-e.ele

• I

~ i

G

1 • . . . . . . l ~". . . . . . . . . . . . . . . . . . . . . . . . . . .

,

"~pe ~._

*

i

v

I0 • . "

-0.011

II

i

lira

:

o

0 e1

ll?, %, , " ~"

I1

?:3

M

~'

...ca. : ( " a a " ~ . . . . . . le.~-t) "

./;

. e/.o

! ~ ; : . :

~t I "R"

-0.004

r 9"6

,2

O " ~K .t';~; .........

6 " (~

:

/ ~

: :

2 20

2

~ .................

;J

0o

6F~ 6

7 . '

3 2S

.......

4-~ 6

0.003

a 4 ~ /

:

:

! I ~i~ "

. ~

0.011

: "

L

0.810

F~o. 4. Correspondence analysis and classification of the overlap view range exhibited by the 70S ribosome from E. coliP A total of 204 images were included in this analysis. All results were independently reproduced with a second data set comprising 177 images. (a) Correspondence analysis map of factor 1 versus 2, with assignment of clusters following Diday's multiple partitioning method. Each image is represented by a symbol that stands for the cluster it belongs to. The symbols (1 through 9, A through Z, * beyond Z) follow the ranking of the cluster, cluster 1 being the most populous. The images occur ordered along factor I according to the orientation of the ribosome on the specimen grid.ss For the meaning of the classes l, I], and IH, see (c). The images representative for the 11 clusters with memberships of five and above are shown at the bottom. They were obtained by reconstitution using 8-dimensional cluster center coordinates. ~°(Orientation changes reflected by factor 1 are thought to be produced by rotation of the particle about an axis ronghly vertical in this figure.) (b) Hierarchical ascendant classification dendrngram showing the successive merging

[1]

IMAGE PROCESSING OF SINGLE RIBOSOMES IMAGED BY EM

19

C

III

II

6

4

i

i

?

9

I

I 8

I 10

5

II

1

11

3

2

III

of classes, starting from the 11 most populous clusters, according to the principle of minimum increase of intraclass variance. The distances in v e r t i ~ direction of the dendrogram reflect the distances between aggregated classes. The most sensible cutting level, indicated by a dashed line in (c) generates three classes. These classes are demarcated in (a) by solid lines and are marked by Roman numerals I, II, III. The result indicates that three more or less stable orientations are assumed by the particle in the overlap range of views. The views corresponding to these orientations are typified by the reconstituted class center images I, II, and III shown at the bottom.

20

ELECTRON MICROSCOPY

[1]

images may be computed by using class center coordinates in the reconstitution formula of Eq. (20). As van Heel observes,63 the exact determination of an optimal partitioning of a data set consisting of N points into a given number of compact classes requires a number of trial computations on the order of N!.. Since this number is impractically large for all practical cases, only algorithms for finding local optima have been given in the literature. Of the numerous algorithms proposed, two (discussed in Ref. 58) have been explored and found useful in the analysis of electron microscopic data: hierarchical ascendant classification (HAC) with minimum added variance as merging criterion, 64 and the method of clustering about moving centers combined with multiple cross-partitioning used by Diday. 65 van Heel 45,61,63,66 uses HAC throughout the analysis, starting with the original points as "classes," plus a postprocessing procedure designed to reduce the residual misclassification. Frank et aL 67 use Diday's variant of the method of clustering about moving centers, followed by HAC at the stage where stable clusters have been found. As an example for the application of these techniques, we show the classification67 of the overlap views6s of the 70S ribosome from Escherichia coli (Fig. 4).

Nonlinear M a p p i n g The objective of nonlinear mapping is to display a point distribution that is formed in a three- or higher dimensional space onto a 2-D map such that interpoint distances are optimally preserved. Thus the resulting 2-D map allows the clustering as well as the seriation of the data to be analyzed.69,70 The points in the m~-dimensional space spanned by the factors are first linearly projected onto a 2-D map. Their projected positions are iteratively changed such that the differences between the Euclidean distances in the 2-D map dii, and those in the higher-dimensional space t~ii,are

63 M. van Heel, Ultramicroscopy 13, 165 (1984). S. C. Johnson, Psychometrika 32, 241 (1967). 65 E. Diday, Rev. Fr. Inf. Rech. Oper. 6, 61 (1972). M. van Heel and M. St6fller-Meilicke, EMBO J. 4, 2389 (1985). 67 j. Frank, J.-P. Bretaudiere, J.-M. Carazo, A. Verschoor, and T. Wagenknecht, J. Microsc. 150, 949 (1988). 6a A. Verschoor, J. Frank, T. Wagenknecht, and M. Boublik, J. Mol. Biol. 187, 581 (1986). 69 M. Radermacher and J. Frank, Ultramicroscopy 17, 117 (1985). ToM. Radermacher and J. Frank, Ultramicroscopy 19, 75 (1986).

[ 1]


EM

21

minimized according to an error criterion E: 1 E-

iv

,,

E

E

' 0. 5 The algorithm that determines the locations of the individual proteins from the second moments of the length distribution function is very sensitive not only to the moments themselves, but also to the errors associated with them. For this reason, an error analysis on the second moment is absolutely crucial. The precision error in the second moment that is returned by the indirect Fourier transformation program does not reflect the true error for the following reasons. The analysis depends on the choice of d, the maximum length in the distribution of distances from points in one protein to points in the other. For a given measurement, d is not known a priori. What is done in practice is that the reduced x 2 and the second moment are determined for a range of d values. There is generally a range of d for which x 2 is about the same. In this region, the second moment M e has a range M1 to M2. The overall uncertainty in M Uis given by 5 o~ = tr,~ + (3//2 --M~)2/12 (12) Yet another uncertainty comes from the fact that, due to systematic errors, not all data sets give the same value of reduced x ~. Those data sets that contain such error are less reliable than those that do not, and this uncertainty should be reflected in the error for the second moment. This results in a modification of the error for the second moment as follows: oa = max(l, xa)t72 + (M 2 -- Mr) 2

(13)

The final estimates for the second moment and errors are used in the next step of the analysis, which is to determine actual positions and radii of gyration of the individual proteins. Model Building

Equations (1) and (2) provide the basis for model building based on the second moments, since they relate to second moments directly to the positions and radii of gyration of the proteins. A model can be found by determining the coordinates and radii of gyration that minimize So =

~

[ M o. - ( R 2 + R 2 + (xi -- xi) 2 + (Y~- yj)2 + (z, - zj)2}] (14)

U

This minimization, which involves a nonlinear least squares optimization, is done by the use of Marquardt's algorithm. 2~ One knows a p r i o r i that the 2t p. R. Bevington, "Data Reduction and Error Analysis for the Physical Sciences." McGrawHill, New York, 1969.

[7]

SMALL SUBUNIT QUATERNARY STRUCTURE

129

radii of gyration for the proteins cannot be smaller than those obtained under t h e assumption that the proteins are anhydrous spheres and given their known molecular weights. Accordingly, it is useful to prevent the model from converging on nonphysical values of radii of gyration by constraining the radii of gyration to be greater than the anhydrous sphere values. This is done b y adding penalty terms to the function to be minimized 5,2°,22 so that one now minimizes the function t~

S=S0 + ~ (Ri_goi),

a>0

(15)

where t~ is a positive constant, and R0t is the radius of gyration of the ith protein if it were an anhydrous sphere. We have also included a provisional upper-bound constraint on the radii of gyration equal to three times the lower bound constraint. Provided that one starts with initial guesses for the radii of gyration that are larger than Roi, the algorithm will always yield final estimates that are larger than the lower limit, and less than the upper-bound constraint. In the absence of constraints, nonphysical values for the radii of gyration are often obtained, owing to the fact that the radii of gyration are small compared to the distances between proteins, and their squares are usually of the same magnitude as the errors in the second moments. The errors in the coordinates and radii of gyrations would be obtained in a straightforward way from the minimization routine if no contraints were operative. In practice, about one-half of the radii of gyration are near the limits to which they are constrained. Thus the errors are evaluated by a Monte Carlo procedure, s The model that results from the analysis of the real second moments M o is used to calculate a set of ideal M o values. Then a large number of sets of M o values are generated, assuming that each of the second moments are normally distributed about the ideal with a standard deviation that equals the actual error estimates. Each of these trial data sets is then solved by the model building program, and the values of the parameters obtained for each are used to calculate a distribution mean and standard error. Current Model Figure 5 presents two views of the current model including all proteins except $2 l, contained within contours representing the shape of the 30S subunit, as interpreted by St6tiler and St6ftler-Meihcke.~ About 90 different interprotein distance measurements, made over a span of a decade, are involved in the construction of this map. The average of the uncertainties 22y. Bard, "Nonlinear Parameter Estimation." Academic Press, New York, 1974.

130

OTHER BIOPHYSICAL METHODS

[7]

FI6. 5. Two views of the current model of the structure of the 30S ribosomal subunit. Contours are the subunit shape as interpreted by the Berlin group.' Correlation between neutron map and the shape was obtained by reference to immunoelectron microscopic data and by maximizing the "containment" of ribosomal proteins by the shape contours.

in the coordinates of all proteins is around 10 .At along all three axes. The aggregate error of the map is well within the range of expected errors calculated by Monte Carlo modeling, given the precision of the individual measurements used in constructing the map. Individual proteins are represented as spheres whose volumes are determined by molecular weight, assuming a fixed partial specific volume (anhydrous) of 0.71. The neutron map was aligned with the shape contours by correlating it with the immunoelectron microscopy map.1 Note that all proteins are well confined by the shape contours. None of the calculated positions results in significant collision between different proteins (i.e., we do not map two different proteins into the same position). The calculated aggregate radius of gyration for 30S proteins is 71 A, which agrees exceptionally well with recent experimental measurements. 23 The neutron map is highly consistent with the mapping of surface-exposed antigenic determinants of ribosomal proteins constructed by both the Berlin ~ and UCLA 24 groups. Furthermore, interprotein spatial 23 V. Ramakrishnan, Science 231, 1562 (1986). 24 j. A. Lake, Annu. Rev. Biochem. 54, 507 (1985).

[8]

PREPARATION OF 5 0 S FOR NEUTRON SCATTERING

131

relationships within the map are congruent with the results of bifunctional chemical cross-linking studies, with the notable exception of cross-links involving protein S 13.a Acknowledgments We wish to express our deep appreciation to Drs. P. B. Moore and D. M. Engelman, in whose laboratories this work was carried out. This work was supported by NIH grants AI 09167 to P. B. Moore and A120466 to P. B. Moore and D. M. Engelman, and by the Office of Health and Environmental Research of the United States Department of Energy.

[8] P r e p a r a t i o n a n d A c t i v i t y M e a s u r e m e n t s o f Deuterated 50S Subunits for Neutron-Scattering Analysis

By

PETRA NOWOTNY, VOLKER NOWOTNY, HELGA Moss,

and KNUD H. NIERHAUS Ribosomes are heterogeneous particles with respect to the neutron scattering-length densities of their components. Because of the different proton contents of these components the scattering-length density of ribosomal proteins is less than that of RNA. In order to obtain ribosomal subparticles of like density we homogenized protein and RNA by means of differential deuteration. Both rRNA and protein fractions are prepared from cells grown under different 020 concentrations, so that both fractions adopt a scattering density of near 100% D20. From these deuterated fractions a 50S subunit is reconstituted which is now homogeneous for the neutron beam. If this homogeneous 50S subunit is transferred to a buffer of near 100% D20, it becomes "invisible" to the neutron beam ("glassy ribosome"). This system allows two kinds of measurements to be made using neutron scattering: (1) Integration of one protonated protein permits the measurement of shape parameters of this protein in situ. (2) The integration or binding of two protonated components enables us to measure the distance of mass centers of the respective components. A detailed discussion of the concept of the "glassy ribosome" can be found elsewhere, l V. Nowotny, R. P. May, and K. H. Nierhaus, in "Structure, Function and Genetics of Ribosomes" (B. Hardesty and G. Kramer, eds.). Springer-Verlag, Berlin, Federal Republic of Germany, 1986. METHODS IN ENZYMOLOGY, VOL. 164

Copyright© 1988by AcademicPress,Inc. All n4~htsof reproductionin any formr--,~ervcd.

[8]


131

relationships within the map are congruent with the results of bifunctional chemical cross-linking studies, with the notable exception of cross-links involving protein S 13.a Acknowledgments We wish to express our deep appreciation to Drs. P. B. Moore and D. M. Engelman, in whose laboratories this work was carried out. This work was supported by NIH grants AI 09167 to P. B. Moore and A120466 to P. B. Moore and D. M. Engelman, and by the Office of Health and Environmental Research of the United States Department of Energy.

[8] P r e p a r a t i o n a n d A c t i v i t y M e a s u r e m e n t s o f Deuterated 50S Subunits for Neutron-Scattering Analysis

By

PETRA NOWOTNY, VOLKER NOWOTNY, HELGA Moss,

and KNUD H. NIERHAUS Ribosomes are heterogeneous particles with respect to the neutron scattering-length densities of their components. Because of the different proton contents of these components the scattering-length density of ribosomal proteins is less than that of RNA. In order to obtain ribosomal subparticles of like density we homogenized protein and RNA by means of differential deuteration. Both rRNA and protein fractions are prepared from cells grown under different 020 concentrations, so that both fractions adopt a scattering density of near 100% D20. From these deuterated fractions a 50S subunit is reconstituted which is now homogeneous for the neutron beam. If this homogeneous 50S subunit is transferred to a buffer of near 100% D20, it becomes "invisible" to the neutron beam ("glassy ribosome"). This system allows two kinds of measurements to be made using neutron scattering: (1) Integration of one protonated protein permits the measurement of shape parameters of this protein in situ. (2) The integration or binding of two protonated components enables us to measure the distance of mass centers of the respective components. A detailed discussion of the concept of the "glassy ribosome" can be found elsewhere, l V. Nowotny, R. P. May, and K. H. Nierhaus, in "Structure, Function and Genetics of Ribosomes" (B. Hardesty and G. Kramer, eds.). Springer-Verlag, Berlin, Federal Republic of Germany, 1986. METHODS IN ENZYMOLOGY, VOL. 164

Copyright© 1988by AcademicPress,Inc. All n4~htsof reproductionin any formr--,~ervcd.

132


[8]

The preparation procedures and activity measurements are summarized in the tabulation below, and the methods are described accordingly. References are given for those techniques which have been previously described in detail. Preparation and Activity Measurement of the Ribosomal Components Protonated components Fermentation2 Isolation of ribosomal subunits2 Isolation2 and separation of rRNA (23S + 5S rRNA) Isolation of the total proteins of the 50S subunit (TP50) Isolation of single proteins Deuterated components Fermentation Isolation of ribosomal subunits Isolation and separation of (23S + 5S) rRNA Preparation of TP50 Determination of the concentration of isolated proteins Reconstitution procedure Activity measurements Poly(U)-dependent poly(Phe) synthesis assay Peptidyltransferase assay

Preparation of Protonated Components The fermentation of protonated Escherichia coli cells (D10; RNase I-, Met-, tel A-), the isolation of ribosomal subunits, and the isolation of the total rRNA fraction have been described elsewhere in this series. 2 Separation of 23S and 5S rRNA

Materials 70% (v/v) Phenol: freshly distilled phenol is stored at - 2 0 ° in 10-ml portions; before use 4.3 ml of glass-distilled water is added per portion Bentonite-SF (Serva, Heidelberg, FRG; Cat. No. 14515) 2 K. H. Nierhaus and F. Dohme, this series, Vol. 59, p. 443.

[8]


133

Sodium dodecyl sulfate (SDS; Bio-Rad, Richmond, CA; Cat. No. 161-0301) RNA buffer: 10 m M Tris-HC1, pH 7.6 (4°), 50 m M KCI, 1% (v/v) ethanol Sucrose (BRL, Bethesda, MD; Cat. No. 5503UB) 5% (w/v) sucrose in RNA buffer 20% (w/v) sucrose in RNA buffer Ethanol T~oM4 buffer: 10 m M Tris-HC1, pH 7.5 (4°), 4 m M magnesium acetate Procedure. The (23S + 5S) rRNA is isolated by phenol extraction from 50S subunits containing intact 23S RNA. Sterilized tubes and pipets are used in all steps, and the whole procedure is performed at 4 °. 0.1 volume of 10% (w/v) SDS, 0.05 volume of 2% (w/v) bentonite and 1.2 volume 70% phenol are added to 50S subunits (concentration up to 400 A26ounits/ml). The mixture is shaken vigorously for 8 min and centrifuged for 10 min at 10,000 g. The aqueous phase (upper phase) is mixed with 1 volume of 70% phenol, shaken for 5 min, and centrifuged. The aqueous phase is extracted a third time. One to 2 ml of the RNA-containing aqueous phase is layered onto a sucrose gradient (5-20% sucrose in RNA buffer, SW27 rotor). After centrifugation (17 hr at 20,000 rpm) the gradient is pumped out and the optical density is recorded at 290 rim. 23S and 5S rRNA-containing fractions are pooled and precipitated at - 2 0 * overnight by addition of 2 volumes ethanol. After centrifugation (30 min at 10,000 g) the RNA pellets are resuspended in Tt0M4 buffer at a concentration of about 200 A260 units/ml for the 23S rRNA and about 20 A26o units/ml for the 5S rRNA. The concentrations are measured and the rRNA solutions are stored at - 8 0 ° in small portions. Preparation of the Total Proteins (TP50) from the Large Subunit This method is a modification of that previously described. 3 Materials Magnesium acetate, 1 M Glacial acetic acid Bentonite-SF (Serva, Heidelberg, FRG; Cat. No. 14515) Dowex l-XS, 20-50 mesh Acetone

3 H. Schulze and K. H. Nierhaus, E M B O J . 1, 609 (1982).

134


[8]

Rec4-6U buffer: 20 m M Tris-HC1, pH 7.5 (4°), 4 m M magnesium acetate, 400 m M NH4CI, 2 m M 2-mercaptoethanol, 6 M urea. The buffer is mixed with bentonite-SF (1 g per liter), stirred at 4 ° for 1 hr, and filtered through two layers of Selecta filters (Schleicher & Schuell, Dassel, FRG; 595 1/2, diameter 240 mm), and stored at 4 ° for up to 3 weeks. Alternatively, the 6 M urea solution may be passed through a Dowex ion-exchange column (5 × 30 cm) and then the salts are added Rec4 buffer: same as Rec4-6U but without urea. Therefore the washing procedure with bentonite is omitted

Procedure. Two volumes of glacial acetic acid are added to the 50S suspension (made 0.1 M Mg 2+ by addition of 0.1 volume of magnesium acetate). The mixture is stirred for 45 min at 0 ° and centrifuged at 10,000 g for l0 min at 4 °. Five volumes of acetone are added to the protein supernatant. The precipitated proteins are collected by centrifugation at 10,000 g for 10 min. The pellet is dried under vacuum and dissolved in Rec4-6U buffer to a final concentration of about 200 equivalent units per milliliter (1 equivalent unit of protein is the amount of protein extracted from 1 A26o unit of 50S subunits). The solution is dialyzed overnight against a 200-fold volume Rec4-6U buffer, and four times (each for 45 min) against a 200-fold volume Rec4 buffer. The solution is centrifuged at 6000 g for 5 min, the optical density at 230 nm is determined, and the solution is stored in small portions at - 8 0 °. The following relationships are an approximation: 1 A23o unit of a TP50 solution is equivalent to 220/zg and also to l0 equivalent units (eu) of TPS0.

P r e p a r a t i o n of Single Proteins The proteins of the large ribosomal subunit are separated by conventional chromatographic techniques. 4 We also apply an additional protocol for the protein isolation: the proteins are prefractionated with the LiCl-split technique,4 and we isolate proteins by ion-exchange and reversed-phase high-performance liquid chromatography (HPLC). Various buffer systems are applied. In all cases gradient systems are used. The isolated proteins 4 G. Wystup, H. Teraoka, H. Schulze, H. Hampl, and K. H. Nierhaus, Eur. Z Biochem. 100, 101 (1979).

[8]

PREPARATION OF 50S FOR NEUTRON SCATTERING

135

remain active with respect to their ability to assemble into active 50S subunits~ (total reconstitution). Materials Reversed-phase chromatography Trifluoroacetic acid (TFA; Fluka, Buchs, Switzerland) 2-propanol acetonitrile Buffer A: 0. 1% TFA in water Buffer B: 0.1% TFA in 2-propanol or acetonitrile. The organic solvents used for the HPLC separations are of LiChrosolv grade (Merck, Darmstadt, FRG). Reversed-phase separations are performed on a Nudeosil 300-C4 column (size 250 × 16 mm, pore size 300 A, particle size 7 #m). Guard column: size 30 )< 16 m m (packed with support from Macherey & Nagel, D~iren, FRG). Ion-exchange chromatography Buffer A: 20 m M HEPES, 5 M urea, 2 m M methylamine in water Buffer B: 1 M KC1 added to buffer A. The urea solution is passed through a Dowex ion-exchange column (see above). Ion-exchange chromatography separations were performed on an Ultropac TSK CM-3SW column (size 150 × 21.5mm, particle size 10/zm). Guard column: size 75X 21.5 mm (purchased from LKB, Munich, FRG).

Procedure. Aqueous buffers are filtered through a 0.45/zm type HAWP filter (Millipore, Molsheim, France). The proteins are eluted with a flow rate of 5 - 8 ml/min using gradients4b from 0% B to 100% B. The gradient shape is adapted to the separation problem. The absorption is measured at 230 nm. After HPLC separation, the following steps are performed: The pooled protein fractions obtained with urea-containing solvents are extensively dialyzed against double-distilled water and lyophilized. Protein fractions from the reversed-phase separation are dried directly. The dried proteins are dissolved in Rec4-6U buffer, dialyzed first against 100 volumes of that buffer for 12 hr, and then four times for 45 min against 100 volumes of Rec4 buffer. All dialysis steps are performed at 4 °. Handling of

4~p. Nowotny, H. Eckardt, V. Nowotny, and R. M. Kamp, Chromatographia, in press (1988). 4b R. M. Kamp, A. Bosserhoff, D. Karnp, and B. Wittman-Liebold, J. Chromatogr. 317, 181 (1984).

136


[8]

the purified proteins with respect to purity check and characterization are described in detail in Ref. 4. Preparation of Deuterated Components Fermentation Materials M3 minimal medium: Glucose 40% (w/v) in 84% I)20 or 76% D20, respectively, sterilized at 120* for 20 rain in an autoclave Salt solution5 for a 100-liter fermenter: 150 g NaC1, 200 g (NH4)2SO4, 650 g KH2PO4, 1000 g K2HPO4, pH 7.2, sterilized as above MgC12 solution: 10 g/100 ml 100 liters 84% or 76% D20-water mixture. The exact D20 concentration has to be adjusted with pure I)20 and measurement is controlled with a density measurement. We use a Paar DMA 60 densitometer (Graz, Austria), which allows readings to at least 5 significant digits Procedure. Sterilized 40% glucose, 0.6 ml, in 84% or 76% D20 is added to 50 ml deuterated sterilizeql M3 medium and 0.12 ml 10% (w/v) MgC12 is added. The medium is inoculated with 4 ml solution ofE. coli strain MRE 600 in glycerol and shaken overnight at 37*. This ciJlture is then used to inoculate a deuterated M3 medium of 84% or 76% D20. The culture is added to 450 ml M3 medium with the respective D20 content after addition of 6 ml 40% glucose. The solution is supplemented with 1.2 ml 10% MgC12. The culture is grown for 8 hr and then poured into 100 liters D20 M3 medium with the respective D20 concentration supplemented with 100 ml 10% MgClz solution and 1 liter of 40% glucose. The solution is stirred (1000 ~ m ) and kept at 37 °. The fermenter (Bioengineering, Switzerland) is not aerated to reduce contamination of the I)20 with water vapor from air. The outlet is cooled to reduce a loss of I)20 vapor. Under these conditions the cells grow with generation times of about 1 - 1.5 hr. The fermentation is stopped at an optical density of 1.5 A~5o units per ml (usually reached after 8 - 1 0 hr). The medium is pumped through pipes surrounded by ice-water as a cooling device and the cells are pelleted with a Padberg centrifuge. The medium is collected and recycled. With this procedure 300-400 g (wet weight) cells is obtained with one

P. B. Moore, this series, Vol. 59, p. 639.

[8]


137

fermentation. The cells are collected in a plastic bag, shock frozen, and stored at - 80 °. Recycling of Used D20 Media The media are distilled with a 50-liter Buechi (Buechi, Switzerland) rotavapor and further purified by pumping them through a Millipore Super Q device several times until a resistance of 10,Mf~ is reached. The D20 concentration is reduced after one fermentation by less than 1% D20. This loss is compensated for by the addition of 100% D20 before the next fermentation. Preparation of Deuterated Ribosomal Subunits

Materials Dissociation buffer: 1 0 r a m K2HPO4-KH2PO4, pH 7.5, 1 m M MgCI2, 6 m M 2-mercaptoethanol 40% sucrose (w/v) in dissociation buffer 50% sucrose (w/v) in deionized water Aluminum oxide (Alcoa, Serva, Heidelberg, FRG; Cat. No. 12293) Tl0MloNi0oSH~ buffer: 10 m M Tris-HC1, pH 7.5, 10 m M MgC12, 100 m M NH4C1, 6 m M 2-mercaptoethanol

Procedure. The entire procedure is carded out at 4 °. For the preparation of 70S ribosomes the cells are washed with dissociation buffer (2 ml per gram of cells). The pelleted cells (10 rain at 10,000 g) are mixed with aluminum oxide (Alcoa, 2 g per gram of cells) and ground in a Retsch mill KM I (Retsch, Haar, FRG) for 45 min. One batch (300-400 g) of cells is used. Dissociation buffer is added (1.5 ml per gram of cells). The paste is homogenized for l0 min in the mill. Aluminum oxide is removed by centrifugation (10 min at 10,000 g), and the supernatant is withdrawn. The pelleted aluminum oxide is suspended in dissociation buffer (1.0 ml per gram of cells) and is ground for about 10 min in the Retsch mill until the paste becomes homogeneous. The aluminum oxide is again removed (10 min at 10,000 g). Both supernatants are combined and the cell debris is removed by low-speed centrifugation (60 min at 30,000 g). From the supernatant of this step (S-30) the ribosomes are collected via high-speed centrifugation (14 hr at 27,000 rpm; 45Ti). The ribosomes are suspended in dissociation buffer (0.3 ml per gram of cells). The ribosome solution is divided into portions of about 10,000 A260 units each and stored at - 80°. Usually 200- 300 A26o units of 70S ribosomes are obtained per gram of cells. The subunits are separated with a

138


[8]

zonal run in a B15Ti zonal rotor containing a hyperbolic gradient6 (1600 ml from 6 to 38% sucrose) made from dissociation buffer and 40% sucrose in dissociation buffer. After centrifugation (17 hr at 23,000 rpm) the gradient is pumped out with 50% sucrose in deionized water. Fractions (about 18 ml) containing either small or large subunits are pooled. The subunits are collected via a centrifugation step (22 hr at 38,000 rpm; 45Ti). The pellet of each tube is suspended in 4 ml T~oM~oN10oSH6buffer for the 50S subunit and 2 ml of the same buffer for the 30S subunit by gently shaking in the cold room (4 °). The suspension is cleared by centrifugation (5 min at 10,000 g). Its optical density at 260 nm is measured. The subunit suspension is shock frozen in large portions (10 ml) and stored at - 8 0 °. Isolation of Deuterated (23S + 5S) rRNA The procedure follows the one described above for protonated material. However, special care is taken to obtain maximal yields. Thus, the phenol phase obtained after the first phenol extraction is washed with one volume of TtoM4 buffer which is also used to wash the phenol phases of the second and third phenol extractions. Both aqueous phases obtained after the third phenol extraction (plus washing step) are combined. This modification increases the yield significantly, which reaches 80-90% of the input amount of 50S material. Isolation of Deuterated T P 5 0 Escherichia coli strain KI 2 (D 10) is used for the isolation of protonated TP50. However, we have to use the strain MRE 600 for the isolation of the deuterated TP50, since it grows better in a D20-containing medium. 50S subunits derived from MRE 600 cells which were harvested at 1.5 A65o units per ml contain RNases which destroy the large rRNA during the reconstitution process. Therefore, the RNases have to be removed from the 50S subunits by washing with LiC1 as described below before the total proteins are extracted. Materials 5 M LiTloM~o buffer: 5 M LiC1, 10 m M magnesium acetate, 10 m M Tris-HCl, pH 7.5 (washed with bentonite as described above for Rec4-6U buffer) T~oM~0buffer: 10 m M magnesium acetate, 10 m M Tris-HC1, pH 7.5 6 E. F. Eickenberry, T. A. Biekle, R. R. Traut, and C. A. Price, Eur. J. Biochem. 12, 113 (1970).

[8]


139

Procedure. 50S subunits in TloMtoN~0oSH6 buffer are diluted with TtoMl0 buffer and LiTtoMlo buffer to a final concentration of 100 A26o units/ml and a LiC1 concentration of 0.3 M. This solution is kept at 0 ° for 2 hr. A centrifugation step (7 hr at 40,000 rpm; 45Ti) at 0 ° collects the washed 50S subunits. The supernatant is discarded and the subunits are suspended to a final concentration of about 600 A26o units/ml. A 0.1 volume of 1 M magnesium acetate and 2.2 volumes of glacial acetic acid are added to this suspension. The rest of the procedure follows that described above for protonated material. The yield for the TP50 is (with the approximation 1 A23o unit of TP50 - 10 eu of TP50) in the range of 60-75% of the input amount of 50S subunits.

Determination of the Concentration of an Isolated Protein Two methods are employed: (1) determination of the total nitrogen content in small protein samples, 7 and (2) absorption measurement around 230 nm. 8 Materials

Phenol reagent: 2% phenol, 0.1% sodium nitroprusside in water (this solution can be kept for several weeks at 4 °) Alkaline hypochlorite: 20 m M NaOC1, 2.5 M NaOH (the titer of the hypochlorite solution is controlled iodometrically after preparation; this solution is stable at 4 °) Nitrogen standard: 10 m M (NH4)2SO4 in glass-distilled water is kept in tightly sealed glass bottles KPh buffer: 10 m M potassium phosphate, pH 7.5, 4 m M magnesium acetate, 400 m M KCI BSA standard: l0 mg bovine serum albumin in 100 ml KPh buffer HC104: 72% solution Procedure. Glassware and pipets have to be intensively cleaned. Wash them with hot tap water, immerse them in a solution of hot tap water (about 50 °) with solid K O H (1 M final concentration) for 1 hr, and rewash them with tap water, 10% acetic acid, and finally glass-distilled water at least six times. In order to remove traces of nitrogen-containing components, we extensively dialyze the protein solution against a 100-fold volume of KPh buffer at least four times for 45 min. The samples ( 1 0 - 5 0 p l protein solution with an optical density of about 1 A23o unit/ml) are diluted in 100 pl of KPh buffer, and 25 pl HC104 solution is added. The samples are heated at 215 ° for 1 hr in a thermostat-

L. Jaenicke, Anal. Biochem. 61, 623 0974). B. Ehresmann, P. Imbault, and J. H. Weil, Anal. Biochem. 54, 454 (1973).

140


[8]

ted aluminum block with 4 cm deep holes which hold the Duran test tubes 0 4 × 100 mm). After cooling, 500al glass-distiUed water is added. To start the color reaction 450/~1 phenol reagent is added and then 200250 ]zl NaOCl solution (containing 320 ag NaOC1). After 20 min the absorption is recorded at 614 nm. A buffer sample is treated the same way and used as a control. To reduce error measure at least four different amounts of a given protein solution (for example 10, 25, 40, 55/~l). A nitrogen standard curve is determined from 5, 10, 15, 20, and 25 ]zl of the nitrogen standard solution (equivalent to 140, 280, 420, 560, and 700 ng nitrogen, respectively), and a BSA standard curve plotted from corresponding inputs of the BSA standard solution. The slopes (m) of the regression lines are calculated for the curves: the nitrogen standard curve (m~; A614 reading versus nanograms nitrogen input), the BSA standard c u r v e [ m 2 ; A614reading versus microliters BSA input; the relative nitrogen content of BSA is taken as 15.8% (w/w)], and the sample curve (m3; A614 reading versus microliters sample input), m 2 and m~ are used to calculate the recovery R (as percentage). The measured concentration of BSA is [BSA]= - -

m2 X 100

ml × 15.8 [ng/lzl]

Division by the input concentration [BSA]i multiplied by 100 gives the recovery R in percent. R = ([BSA]=/[BSA]i) × 100 Since m3/m I is the measured nitrogen content of the sample, the protein mass per milliliter (My) is obtained by Mp-

m a x 104 m~ X %N X R [/~g/ml]

The value %N from the ribosomal protein under observation is obtained from the sequence data (see Table I). The Mp values of a sample solution with a concentration of 1 A2aounit per ml (Mp/A230) are given in Table I. The method outlined above yielded absolute values which were compared to a simple absorption measurement. This method s was developed for measuring protein concentrations in protein-RNA mixtures. The protein content is determined by measuring the absorbances at 228.5 and 234.5 nm. The difference between these values (AA =A22a.5-A2u.5) is related to the protein concentration. We have determined that a AA of I is equivalent to 312 ag protein/ml in KPh buffer. The results of both methods are expressed as micrograms per A23o unit in Table I. Although the UV method yields reproducible results, we assume that the nitrogen

[8]

PREPARATION OF 508 FOR NEUTRON SCATTERING

141

determination gives the correct values. Therefore, the latter were used to calculate the molar extinction coefficient at 230 nm (E23o, see Table I). Since we use the term "equivalent unit" (eu) for the input estimates of proteins, we calculate the volume of a ribosomal protein solution which contains 1 eu. One equivalent unit is 36 × Mr × 10-6 gg, assuming that 1 A260 unit of 50S subunits corresponds to 36 pmol. The concentration of the ribosomal protein solution under observation is AA × 312 × 10-3 gg/#l. Therefore, 1 eu is present in 36 × 10-6 Mr 312)< AA X 10-3 X F

36 X 10 - 3 X M r

312XAAXF

K AA/tl

The factor F is the ratio of the results from the nitrogen determination and that of the UV measurement. F and K are presented in Table I for the individual ribosomal proteins. Reconstitution P r o c e d u r e The optimal molar ratio of rRNA to TP50 is determined for each preparation. Ratios of about 1:1.2 with protonated TP50 and ratios of 1:1.2 to 1:1.8 with deuterated TP50 are usually obtained. The slight overstoichiometry of the proteins in the latter case is due to the LiCl-washing procedure described above. Materials Rec4 buffer: 20 m M Tris-HCl, pH 7.5, 4 m M magnesium acetate, 0.2 m M ethylenediaminetetraaceticacid (EDTA), 400 m M NH4C1, 2 m M 2-mercaptoethanol Rec4 adaptation buffer: 110 mMTris-HC1, pH 7.5, 4 mMmagnesium acetate, 0.2 m M EDTA, 4 M NH4C1, 2 m M 2-mercaptoethanol T2oM4o0N4oo buffer: 400 m M magnesium acetate, 400 m M NH4CI, 20 m M Tris-HC1, pH 7.5 Rec20 buffer: same as Rec4, but with 20 m M magnesium acetate 40% sucrose in Rec20 buffer D-Rec20(K) buffer: 2 0 r a m Tris-DCl, pD 7.9, 20 m M MgC12, 400 m M KC1 in 100% D20

Procedure. RNA dissolved in the T~oM4 buffer has to be adapted to the reconstitution milieu with respect to the ion concentrations. For this purpose, 9 parts of RNA solution is mixed with l part of the Rec4-adaptation buffer. RNA solutions of concentrations exceeding 250 A26o units/ml tend to become very viscous under Rec4 ion milieu on ice and are thus stored during the assay preparation period at room temperature. For optimizing

,~

~

~

~

~

~i

--,,6--

~

--

(--i

--

..-;

~

w~

I"-

c~

0

v

~

~v+~

~+~

~ " ~ ~"

~

~

z

v ~'~

Z

oo

~1~

~

~

I" ~

~

~

~

~1~

~o

r-

~-

oo

~,~

',0

/~.

~'-

~'~

I~ -

oc~

(~

o~

t--

r~

~6

oo

~

~

o~

oo

o

Z L~ Z 0

i--

c,

v-;

o6

v

v

v

~.÷I~÷I~÷~÷I~

v

÷I~ v

v

0

A~

v~

144


[8]

the amounts of TP50 a pilot reconstitution is performed. A single assay contains 2.5 A260 units (23S + 5S) RNA and various amounts of proteins (from 2.5 to 4 eu TP50) in 100/d of Rec4 buffer. For neutron experiments 1600A26o units of deuterated (23S + 5S) RNA are reconstituted with proteins in a volume of 60 ml. If one or both of the protonated proteins belong to the group of rRNA binding proteins 9,9a (Ll, L2, L3, L4, (L5), L7/L12, L9, Ll0, Ll l, L14, L15, Ll6, L17, LI8, L20, L22, L23, L24, L25, L28, L29), the following three-step procedure is applied. In the first step, 1600 A26o units of rRNA are mixed with optimal amounts of the rRNA binding protein(s) in a volume of 60 ml minus the volume containing the optimal amounts of deuterated TP50. If one (or both) of the protonated proteins is a nonbinding protein it is added in a 5to 10-fold molar excess over rRNA. The deuterated RNA and protonated protein mixture is incubated for 15 rain at 44 °. For the second step, the optimal amount of deuterated TP50 is added, yielding a total volume of 60 ml. The mixture is incubated for 15 rain at 37 ° which allows for a slow but ordered assembly, ~° and a further incubation follows (44 ° for 30 min), still under Rec4 conditions. For the third step, the Mg 2+ concentration is raised to 20 m M by the addition of 2.54 ml of T2oM~oN4oo buffer and an incubation is performed at 50 ° for 90 min. After the reconstitution process, the sample is layered on a zonal gradient made from 6 to 40% sucrose in Rec20 buffer and centrifuged (B 15Ti rotor, 17 hr at 20,000 rpm). The fractions containing the 50S material are pooled and the 50S subunits collected with a centrifugation step (24 hr at 38,000 rpm; 45Ti rotor). The pelleted material is suspended in about 700/zl D-Rec20(K) buffer. The tubes should be dried with pieces of filter paper to remove any remaining zonal buffer droplets before addition of the deuterated buffer. The samples are dialyzed four times against 20 ml of D-Rec20(K) buffer. The samples are transferred to 3-ml sample containers which are tightly closed to exclude water exchange, and shock-frozen in liquid nitrogen. Prior to the neutron measurements, D-Rec20(K) buffer is added to the samples in order to yield a volume of more than 1120/zl. The samples are centrifuged for 10 min in a table-top centrifuge to remove aggregated material. The supernatant is carefully withdrawn and 1100/zl is transferred into the measuring quartz cell. The remaining material is used for activity 9 R. Roehl and K. H. Nierhaus, Proc. Natl. Acad. Sci. U.S.A. 79, 729 (1982). 9a M. Herold and K. H. Nierhaus, J. BioL Chem. 262, 8826 (1987). io M. Herold, V. Nowotny, E. R. Dabbs, and K. H. Nierhaus, MoL Gen. Genet. 203, 281 (1986).

[8]


145

measurements and the determination of concentration (A26o/ml). The neutron-scattering experiment is described elsewhere.H Activity Measurements The isolated native subunits are tested with the poly(U)-dependent poly(Phe) synthesis system, as are the reconstituted subunits from the pilot reconstitution experiments. The samples for neutron experiments are tested using the peptidyltransferase assay before and after exposure to the neutron beam. This assay is chosen because it is performed at 0 °, thus preventing a possible reactivation of the measured samples during the assay. Poly(U)-Dependent Poly(Phe) Synthesis

Materials Energy mix: 62.5 m M Tris-HC1, pH 7.8, 8.5 m M magnesium acetate, 160 m M NH4C1, 9 m M ATP, 0.3 m M G T P , 30 m M phosphoenolpyruvate, 24 m M 2-mercaptoethanol TloM~oSI-I6buffer: l0 m M Tris-HC1, pH 7.5, l0 m M magnesium acetate, 6 m M 2-mercaptoethanol Pyruvate kinase (Boehringer Mannheim, FRG; Cat. No. 128 163; 1 mg per milliliter of water) Bulk tRNA from E. coli, 20 mg per milliliter Poly(U)-[14C]Phe mix: 500/zM Phe with 100,000 clam [~4C]Phe (500 Ci/mol) and 100 #g poly(U) per 20/A S 150 enzymes in Tl0MtoSI-I6buffer2 30S subunits in T~0MtoNl0oSI-I6 (10 m M Tris-HC1, l0 m M magnesium acetate, 100 m M NH4C1, and 6 m M 2-mercaptoethanol) Glass filters No. 6, diameter 23 m m (Schleicher & Schuell, Dassel, FRG) BSA solution, l g/100 ml water Trichloroacetic acid (TCA), 5% (w/v) Ether-ethanol, 1 : 1 Scintillation cocktail Ready Solv EP (Beckman, Munich, FRG; Cat. No. 158 729)

Procedure. Poly(U) mix sufficient for all 30 assays of one experimental run is freshly prepared: Poly(U)- [14C]Phe mix Energy mix

600 pl 600/d

" K. H. Nierhaus, R. Lietzke, R. P. May, V. Nowotny, H. Schulze, K. Simpson, P. Wurmbach, and H. B. Stuhrmann, Proc. Natl. Acad. Sci. U.S.A. 80, 2889 (1983).

146


S 150 enzymes (optimized for each preparation) plus T~0M~oSH6buffer Pyruvate kinase tRNA 30S subunits (30 A2~ounits) in TloMIoNIooSH6 buffer plus TioMloN10oSI-I~ Poly(U) mix

[8]

950/zl 90 pl 60 pl

100 #1 2400/A

To 80/zl poly(U) mix a 40/zl aliquot of the reconstitution assay solution containing 1 A2~o unit of reconstituted particles is added. The final concentrations are 20 m M Tris-HC1, pH 7.8, 11 m M magnesium acetate, 160 m M NH4C1, 5 m M 2-mercaptoethanol, 83.33/zM phenylalanine, 1.5 m M ATP, 0.05 m M GTP, and 5 m M phosphoenolpyruvate. In addition the 120/zl assay solution contains 100/zg poly(U), 3/zg pyruvate kinase, and 40/tg bulk tRNA from E. coll. After incubation (30 ° for 45 min) 1 drop of BSA solution and 2 ml 5% TCA are added. The samples are heated for 15 min at 90 °C. The polymerized material (longer than Phe4) is collected on glass filters. The filter-adsorbed material is washed twice with 5% TCA and with ether: ethanol (1 : 1). The filters are counted in 4 ml of Readi Solv EP scintillation cocktail. Peptidyltransferase Assay Materials

Ion mix: 1.5 M KC1, 160 m M Tris-HC1, pH 7.8, 40 m M magnesium acetate Ac[3H]Leu-tRNA: bulk tRNA is taken and charged with [3H]Leu. The tRNA mixture is then acetylated 12and stored in small portions (50/zl/tube) with 50,000 cpm//d Ethyl acetate 0.3 M sodium acetate in water saturated with MgSO4, pH 5.5 (at room temperature) Puromycin (Serva, Heidelberg, FRG; Cat. No. 33835) solution: 1 mg/ml in ethanol Ethanol Procedure. For each assay a mix is prepared of 40/tl ion mix, 40/zl H20 with 50,000 cpm Ac[3H]Leu-tRNA (about 1 #1) and 1 A26o unit of 50S particles diluted in 80/zl of Rec20 buffer. The reaction starts with the ~2A. L. Haenni and F. Chapcvill¢, Biochim. Biophys. Acta 114, 135 (1966).

[8]


147

addition of 80/zl puromycin solution. The assay is incubated for 45 min at 0 °. The addition of 150#l of 0.3 M sodium acetate solution stops the reaction. The reaction product is extracted with ethyl acetate. One milliliter ethyl acetate is added to each sample and mixed vigorously (Vortex) for more than 30 sec. After 5 min at 0 ° (phase separation), 700/~l of the upper organic phase is carefully withdrawn, added to 5 ml of a scintillation cocktail, and counted. A sample with native 50S subunits serves as the 100% value; as a blank we take a sample of native 50S where 80 ~l ethanol was added instead of the puromycin solution. Usually, the native 50S subunits convert 40 to 50% of the input radiolabeled material to Ac[aH] Leu-puromycin. Yields and Etiiciencies

Escherichia coli cells are grown in a 100-liter fermenter containing deuterated medium. Each fermentation yields 300 to 400 g cell material. From l g of cells we isolate about 300 A2~o units of crude 70S ribosomes. The separation of the ribosomes is done via zonal centrifugation. Each zonal run is loaded with 10,000 A2~o units of the crude 70S and yields 2,400 A260units of 50S subunits and l, 100,4260 units of 30S. From the 50S subunits the (23S + 5S) rRNA yield is near 80%, whereas only 60 to 70% of the TPS0 can be recovered. The D 2 0 content of the 84% D20 medium is decreased to 83% after fermentation and recycling; it can be raised again to 84% by addition of 6.25 liters of 100% D20 per 100 liters. Likewise, the 76% medium falls to 75% and is raised to 76% again by addition of 4.2 liters of pure D20 per 100 liters. Nucleic acids from the 76% fermentation and proteins from the 84% fermentation batch show a match point at about 90% on the D20 scale. This value allows a contrast variation. 13 The D20 content in the measured sample can easily be varied above and below this 90% value, which would be impossible with buffer solutions, if the particles would match at 100% D20.

For the preparation of one particle, we start with deuterated 1600 `426o units of (23S + 5S) rRNA (67 mg) and the optimal amounts of TP50 (about 2000 eu TPS0; 200 .423o = 44 mg protein). After the separation of unbound material and the collection of reconstituted 50S subunits from the zonal purification run, we find recoveries of 500 to 700 -426o units of 50S subunits (30 to 45 mg), which are diluted to a volume of 1.1 ml and then measured with the camera D11 at the Institut von Laue-Langevin, Grenoble, France. The activities of the measured particles are between 50 and 100% that of native 50S subunits. 13K. Ibel and H. B. Stuhrmann, J. Mol. Biol. 93, 55 (1975).

148


[9]

[9] Nuclear Magnetic Resonance Techniques for Studying Structure and Function of Ribosomes B y V. N. BUSHUEV and A. T. GUDKOV

The ribosome is a multicomponent ribonucleoprotein complex. Its importance in protein biosynthesis and its very complex structure have given rise to numerous studies of ribosomes using different techniques and approaches (see, for example, Refs. 1 and 2). Nonetheless, many details of the structure and function of the ribosome are not well understood and require further investigation. Although high-resolution nuclear magnetic resonance (NMR) spectroscopy finds increasing applications in biology,3,4 until recently it had not been used for studies of the ribosome. The physical principles on which NMR spectroscopy is based are given in numerous monographs and reviews3-6 and are outside the scope of this paper. It should be noted, however, that most progress has been reached in NMR studies of proteins with Mr up to 20,000. Such a situation is primarily due to the complicated interpretation of NMR spectra for large proteins. An increase in the mass (number of amino acid residues) leads to an increase in the number of signals with the result that peaks overlap and dipolar line broadening occurs because of the large correlation time for such proteins. Calculations have shown that, for proteins with Mr 500,000 (if considered to be in a compact spherical form), the linewidth must be more than 100 Hz. 4 Therefore narrow lines in NMR spectra of large proteins can indicate internally mobile protein segments or amino acid residues. The existence of such mobile segments in high-molecular-mass proteins and complexes including ribosomes has been shown. 7- l0 I G. Chambliss et al. (eds.), "Ribosomes: Structure, Function and Genetics." University Park Press, Baltimore,Maryland, 1980. 2This series, Vols. 29, 30, 59, 60, and elsewherein this volume. 3I. D. Campbelland C. M. Dobson,Methods Biochem. Anal. 25, 1 (1979). 40. Jardetzloland G. C. K. Roberts, "NMR in MolecularBiology."AcademicPress, New York, 1981. s K. Wfithrich,"NMR in BiologicalResearch:Peptidesand Protein." North-Holland,Amsterdam, 1976. 6G. Wagner,Q. Rev. Biophys. 16, 1 (1983). J. L. De Wit, M. A. Hemninga,and T. J. Schaafsma,J. Magn. Reson. 31, 97 (1978). s R. Perham, H. Duckworth,R. Jaenicke,and G. C. K. Roberts,Nature (London) 292, 474 (1981). 9T. R. Tritton,FEBS Lett. 120, 141 (1980). ,o V. N. Bushuev,M. L. Metsis, A. D. Morozkin, E. K. Ruuge, N. F. Sepetov, and V. E. Koteliansky,FEBSLett. 189, 276 (1985). METHODS 1N ENZYMOLOGY, VOL. 164

English translation copyright © 1988 by Academic Press, Inc.

[9]

N M R STUDIES OF RIBOSOMES

149

In the case of compact and very tight protein packing in ribosomes, it is impossible to use NMR for studies of such large ribonucleoprotein (RNP) complexes ( - 2.5 X 10*) due to the very broad fines in the NMR spectra. However, in ~H NMR spectra of ribosomes there are narrow lines which are attributed to ribosomal proteins and this permits the use of NMR techniques for the study of some proteins in the ribosome. Materials 70S ribosomes and their subunits were prepared from Escherichia coli MRE 600.~,~2 The L7/L12 proteins were isolated from 50S subunits.t3 50S particles deprived of L7/L12 proteins were obtained by treatment of intact subunits with 50% ethanol. ~4 The complex of L7/LI2 proteins with LI0 protein was isolated from the mixture of individual proteins ~5 by gel filtration using a Sephadex G-100 column. Elongation factor G (EF-G) was prepared from E. coli MRE 600 according to Rohrbach et al. ~6 Uncleavable GTP analogs, [3H] GMPPCP and [3H]GMPPNHP, were from Amersham (England). Buffer. Sodium phosphate, 1 or 2 raM, pH 7.4-7.6 (without correction for isotopic effect) with l0 m M MgC12 and 50 to 175 m M KC1 in heavy w a t e r (2H20).

Sample Preparation The association of ribosomal subunits (percentage of 70S ribosomes) in the samples was checked in an analytical ultracentrifuge UCA-10 (USSR) equipped with absorption optics. Samples where subunit association was less than 80% were not studied. Ribosomes and their subunits were transferred into heavy water either by dialysis or using small (1.5- 2 ml) columns of Sephadex G-25, the latter procedure being timesaving and more economical. The protein complex (LT/L12)4-L10 was transferred into heavy water by dialysis, or using Sephadex G-25 columns, or by dissolving in 2H20 after lyophilization from sodium phosphate buffer with 100-150 m M potassium chloride. In all three cases the NMR spectra were identical. *~ M. I. Lerman, A. S. Spirin, L. P. Gavrilova, and V. F. Golov, J. Mol. Biol. 15, 268 (1966). t2 L. P. Gavrilova, D. A. Ivanov, and A. S. Spirin, J. Mol. Biol. 16, 473 (1966). 13 W. M611er, A. Groene, C. Terhorst, and R. Amons, Eur. £ Biochem. 25, 5 (1972). t4 E. Hamel, H. Koka, and T. Nakamoto, J. Biol. Chem. 247, 805 (1972). 15 A. T. Gudkov, L. G. Tumanova, S. Y. Venyaminov, and N. N. Khechinashvili, FEBSLett. 93, 215 (1978). 16 M. S. Rohrbach, M. E. Demsey, and J. W. Bodley, J. Biol. Chem. 249, 5094 (1974).

150

OTHER BIOPHYSICALMETHODS

[9]

Ribosomal complexes with EF-G were obtained according to InoueYokosawa et al. ~7 and Lin et al. ~s The amount of ribosomal complexes with EF-G and the uncleavable GTP analog was determined by nitrocellulose filter binding techniques~8; the yield was about 50-60%. NMR Measurements ~H NMR spectra were recorded in Bruker WH-360 and Bruker WM-500 spectrometers operating at 360 and 500 MHz, respectively, in the Fourier transform mode, using standard 5-ram ampoules at 22"-27". Chemical shifts were measured in parts per million downfield from the internal reference. 2,2-dimethyl-2-silapentane sodium sulfate. Since the ribosomal ~H NMR spectrum exhibits broad resonance lines (more than 1 kHz), a wide spectral width from 20,000 up to 30,000 Hz was utilized to obtain qualitative spectra and the computer block size was 32K. A 70* flip angle was applied. The cross-saturation method was used to reveal sharp resonances of highly mobile components in the ribosomal spectra. ,9.2o The cross-saturation experiments were performed in the homonuelear gated decoupling mode. A presaturation pulse for 1 sec was applied at a frequency corresponding to the HDO solution signal. A 70* sampling pulse was applied after the saturation pulse. The concentration of 70S ribosomes and their subunits was between 5 and 20 mg/ml. Depending on the concentration, the number of scans was between 2,000 and 60,000. 1H N M R Spectra of Ribosomes and T h e i r Subunits The ~H NMR spectra of E. coli 30S and 50S subunits and 70S ribosomes (Fig. 1A-C) represent a superposition of sharp and broad (wider than 1 kHz) lines, indicating the difference between the mobilities of amino acid side chains of several proteins. Sharp resonance lines are due to the protons of the amino acids belonging to those parts of the polypeptide chain of the protein which exhibit a significantly higher mobility than the motion of the particle as a whole. A full set of spectra for all the individual ribosomal proteins facilitated a preliminary analysis of the ribosomal ~H NMR spectra and this, together with the known data on the structure of ribosomes, t permitted us to single 17 N. Inoue-Yokosawa, C. Ishikawa, and Y. Kaziro, J. Biol. Chem. 249, 4321 (1974). 18 L. Lin and J. W. Bodley, J. Biol. Chem. 251, 1795 (1976). 19K. Akasaka, M. Konrad, and K. Goody, FEBS Lett. 96, 287 (1978). 2o K. Akasaka, J. Magn. Reson. 51, 14 (1983).

[9]


151

/ B

t

12

I0

8

6

4

1

• 2

~

1 0

i

_ l -2

J

~ -4

8 ppm) FIG. 1. 360-MHz IH NMR spectra of ribosomal subparticles at 22 °. (A) 30S subunits, 7 mg/ml, 5700 transients; (B) 50S subunits, 10 mg/ml, 3100 transients; (C) 70S ribosomes, 10 mg/ml, 3000 transients; (D) 50S subunits without LT/L12 proteins, 5.3 mg/ml, 55,000 transients. From Gudkov et al. n

out those proteins for which sharp signals in the ribosomal spectra are most probable. In the case of 30S subunits the sharp lines in their spectra, from a logical standpoint, may be due to S1 protein. The spectra of intact 30S subparticles and those of particles lacking S 1 protein testify in favor of this assumption. 2t Removal of this protein leads to a significant decrease of the intensity of sharp lines in the spectra of 30S subunits, though they do not disappear completely. Consequently, besides S1 protein in the 30S particle there must also be mobile parts in other proteins. The most mobile proteins of the 50S subunit are L7/L12 proteins, 22 four copies of which are present in the particles. 23 Removal of these proteins is readily performed and does not lead to a significant injury of the 50S subparticles. '4 In the spectrum of 50S particles lacking the LT/L12 proteins the intensity of sharp signals is essentially suppressed (cf. Fig. 1B and D) and the ratios between the intensities of different lines are changed. 21 C. A. Cowgill, B. G. Nickols, J. W. Kenny, P. Butler, E. M. Bradbury, and R. R. Traut, J. Biol. Chem. 259, 15257 (1984). 22A. T. Gudkov, G. M. Gongadze, V. N. Bushuev, and M. S. Okon, F E B S Lett. 138, 229 (1982). 23S. J. S. Hardy, Mol. Gen. Genet. 140, 253 (1975).

152


[9]

This is a direct indication that these proteins are the most mobile in the subunit. It should be noted (Fig. 1 A and B) that the sharp signals in the spectra of 50S particles have a higher intensity than those in the spectrum of the small subunit. This is evidence that the 50S particle contains more amino acid residues exhibiting a higher independent mobility than the ribosomal particle as a whole. Comparison of the spectra of 70S ribosomes (Fig. 1C) with the spectra of 30S (Fig. 1A) and 50S (Fig. 1B) subunits reveals that the spectrum of 70S ribosomes virtually coincides with that of 50S particles. This is seen clearly from a comparison of the spectra of 50S particles and of 70S ribosomes (Fig. 2C and D) in which the broad component is suppressed by cross-saturation. Consequently, the principal contribution to the sharp resonance signals in the IH NMR spectrum of 70S ribosomes is from the same proteins as in the spectrum of 50S subunits, i.e., from L7/L12 proteins, while the contribution from S 1 protein of the small subunit is insignificant.

Phe

Phe 54 30

•

J 7.5

x8

I 6.5

I 3

I 2

I 1

J 0

8 (ppm) FIG. 2. 360-MHz ~H NMR spectraat 22 °. (A) L7/LI2 proteins, 3 mg/ml; (B) (L7)4-L10 complex, 2 mg/ml; (C) 50S subunits, I0 mg/ml; (D) 70S ribosomes, I0 mg/ml. Spectra C and D were obtained in a gated decoupling mode; the presaturation pulse was applied for 1 sec at a frequency (f2) corresponding to the tEK) signal. From Gudkov et a[.22

[9]


153

Since the subunits are almost fully associated with each other (from sedimentation data), and signals from the 30S and 50S subunits must be present in the spectrum of 70S ribosomes, it may be asserted that interaction of the ribosomal subunits leads to no significant changes of the most mobile ribosomal protein component. This primarily concerns L7/L12 proteins, while S l protein as a mobile component of the 30S subunit may prove to be less mobile in the 70S ribosomes than in the subunlt. To understand the basis of such a high mobility of L7/LI2 proteins in the ribosomes there must be knowledge of their structural features. It is known that L7/L12 proteins in solution form a stable dimer 13 and that its structure is symmetric.24 Dimerization results from interaction of the Nterminal segment of L7/LI2 proteins,25 while the globular C-terminal moiety in the dimer is free and takes no part in the dimerization. L7/L12 proteins form a complex with L10 protein26 through which they are attached to the ribosome. Spectral analysis of the dimer of L7/LI 2 protein and of the pentameric complex (L7)4-L10 (Fig. 2A and B) reveals that only signals from L7 protein are present in the spectrum of the complex. This may be explained both by the lower molar concentration of L 10 protein in the complex and by broadening of all the signals from L10 protein in the complex. When protein L7/L12 is attached to L10 protein the signals from the amino acid residues in the N-terminal segment undergo the most changes. Signals from the ring protons of residue Phe 3° (7.3 ppm), from the methyl groups of residues Met 14, Me¢ 7, Met 26, and of the N-terminal Ac-Ser~(2 ppm) are broadened to such an extent that they are not observed in the spectrum of the complex. The least changed are signals of residues in the C-terminal globular segment, for example, Phe ~, high-field signals (0.63 and 0.72 ppm), and resonances from the e-CH2 (3 ppm) groups oflysine residues. Of the 13 lysine residues in the protein only two (Lys4 and Lys29)27 are in the N-terminal segment taking part in the dimerization. Thus, in the complex of L10 and L7/L12 the globular part of L7/LI 2 protein is mobile and does not participate in the interaction with L10. We arrive at the same conclusion from a comparison of the spectra of L7/L12 dimers, their complex with Ll0, the spectra of 50S particles and of 70S ribosomes. A comparison of the spectra (Fig. 2B-D) of the pentameric complex (L7)4-L10 with the spectra of 50S subunits and 70S ribosomes points to their extraordinary identity even in detail. Consequently, in the 24 V. N. Busheuv, N. F. Sepetov, and A. T. Gudkov, FEBSLett. 178, 101 (1984). 25 A. T. Gudkov and J. Behlke, Fur. J. Biochem. 90, 309 (1978). 26 A. T. Gudkov, L. G. Tumanova, G. M. Gongadze, and V. N. Bushuev, FEBSLett. 109, 34 (1980). 27 C. Terhorst, W. M611er, R. Laursen, and B. Wittmann-Liebold, FEBS Lett. 28, 325 (1972).

154


[9]

spectra of 50S and 70S particles the C-terminal globular segment of L7 protein also represents a highly mobile domain. Since the spectra of the 50S subparticles and of the 70S ribosomes are almost identical, the conclusion may be drawn that L7/L12 proteins (at least their C-terminal segment) do not participate in association of the ribosome subunits and that in the absence of translation factors and of messenger RNA they exhibit a high mobility within the ribosomes. Effect of Elongation Factor on the Structure of Proteins L7/L12 in the Ribosome Since L7/L 12 proteins have a considerable intramolecular mobility in the ribosome (see above) and affect the function of the elongation factors (for a review, see Ref. 28), it would be interesting to follow the influence of the factors on the properties of L7/L 12 proteins in the ribosome. EF-G is the best suited for this task, since together with the uncleavable GTP analogs it forms quite a stable complex with the ribosome) 7 The spectrum of EF-G in solution (Fig. 3F) has broader and less well resolved lines than the L7/L 12 spectrum in the ribosome (Fig. 3A and Fig. 2C,D). The spectrum of the 50S subunit with added EF-G is an additive sum of their spectra. This is confirmed by the difference spectra (Fig. 3E) obtained by subtraction of the 50S subunit spectrum (Fig. 3A) from the spectrum of the 50S subunit with EF-G (Fig. 3B) which virtually coincides with the EF-G spectrum in solution (Fig. 3F). This result shows that there is no interaction of EF-G with the 50S subunits in the absence of GTP, and this is in line with the known data. There are some changes in the spectrum of 50S subunits as a result of addition of EF-G and GMPPNHP in the range of 0.6- 1.4 ppm. Most of the Val, Leu, lie, and Thr methyl group signals are located in this range. These changes are seen more distinctly in the difference spectrum (Fig. 3D). The positive signals in the spectrum (Fig. 3D) are from the contribution of EF-G. Some of the L7/L12 signals in the spectrum (Fig. 3D) became negative as the intensity of these signals decreased after the interaction of ribosomes with the EF-G and GMPPNHP. Consequently, the interaction of subunits with EF-G leads to immobilization of the proteins L7/L12. The binding of EF-G to the 70S ribosome exerts a much greater effect on the L7/L12 spectrum in situ (Fig. 4). The addition of EF-G to the 70S ribosomes, even without the uncleavable analog of GTP, leads to changes in the narrow signals of L7/L12 analogous to those occurring in the 50S 2sA. Liljas,Prog. Biophys. Mol. Biol. 40, 161 (1982).

[9]


155

f

fi'// i i

[

9

]

I

7

I

I

I

5

L

3

I

\,

i

I

I

0

Ij

-

8 (ppm) FIG. 3. 500-MHz proton NMR spectra at 27 °. (A) 50S subunits, 7.2 mg/ml; (B) the same as A, but with EF-G; (C) the same as A, but with EF-G and GMPPNHP; (D) C minus A difference spectrum; (E) B minus A difference spectrum; (F) spectrum of EF-G. Spectra A - C are normalized to the same ribosomal subparticle concentration as that of the 50S.

particles with EF-G and the GTP analog (Fig. 4D, compare also with Fig. 3D). This indicates that the 70S ribosomes interact weakly with EF-G without GTP, whereas the 50S particles do not interact at all. Such an interaction of the 70S ribosome with EF-G has been shown also by other approaches.~S Addition of EF-G and the GTP analog to 70S ribosomes leads essentiaUy to a decrease in L7/L12 signal intensity in the NMR spectrum (Fig. 4C). This can be seen clearly from the difference spectrum (Fig. 4E) obtained by subtraction of the 70S ribosome spectrum (Fig. 4A) from that of the complex (Fig. 4C). After GMPPNHP addition, virtually all the lines become broader; this is expressed in reversed signals (they become negative) in the Fig. 4E difference spectrum. The positive component in the Fig. 4E spectrum is lower than that in the Fig. 4D spectrum. This can be explained by immobilization of not only L7/L 12 proteins but also of EF-G after formation of the complex between EF-G, ribosomes, and the uncleavable GTP analog.

156


[9]

c

1

9

I

I

7

I

I

I

5

I

3

I

I

I

I

0

I

-I

8 (ppm) FIG. 4. 500-MHz proton NMR spectra. (A) 70S ribosomes, 18 mg/ml; (B) the same as A, but with EF-G; (C) the same as A, but with EF-G and GMPPNHP; (D) B minus A difference spectrum; (E) C minus A difference spectrum. Spectra A - C are normalized to the same ribosomal concentration as that of the 70S.

In the ternary complex of EF-G, ribosomes, and GMPPNHP, the L7/LI2 proteins are the first to change. This is seen from a comparison of the difference spectrum (Fig. 4E) and the spectrum of the L7 dimer (Fig. 2D). In the spectrum (Fig. 4E) there are signals characteristic for L7 protein, namely those from ring protons of Phe 54 (7.2 ppm), signals in the high field (0.63 and 0.72 ppm), and a number of resonances from methyl groups of apolar amino acids (1.4, 1.2, 0.9 ppm) and also the ~-CHz group of lysine (3 ppm). Immobilization of L7/L12 proteins in the ribosome cannot be explained by direct interaction of EF-G with these proteins. In such a case, the effect of EF-G on the mobility of L7/L12 in the 50S subunit as well as in the 70S ribosome would be equal. Neither is there any interaction of EF-G with the pentameric complex of (L7/L 12)4-L 10. It must be considered also that, despite the binding of EF-G in the proximity of the L7/L 12 stalk,z9 it is known that EF-G interacts with the ribosome (though not so well) without L7/L123° and even with 23S RNA. 31 29A. S. Girshovich, T. V. Kurtsldmlia, Y. A, Ovchinnikov, and V. D. Vasiliev, F E B S Lett. 130, 54 (1981). 3oV. E. Koteliansky, S. P. Domogatsky, A. T. Gudkov, and A. S. Spirin, F E B S Left. 73, 6 (1977). 3~A. S. Girshovich, E. S. Bochkareva, and A. T. Gudkov, FEBSLett. 159, 99 (1982).

[9]


157

More probably the interaction of EF-G with L7/LI 2 and, as a result, their immobilization is a consequence of some essential eonformational changes in the ribosomal components after EF-G binding. It should be noted that a very distinct result of interaction was observed for the ternary complex, EF-G-ribosome-uncleavable GTP analog, i.e., this pertains to the functional state of L7/L12 in the ribosome after EF-G binding, but before GTP hydrolysis. There is a possibility of obtaining a complex of the ribosome with EF-G and GDP in the presence of fusidie acid, i.e., to study the complex after GTP hydrolysis. Unfortunately, the numerous resonance signals in the ~H NMR spectrum of fusidic acid prevent a reliable interpretation of the ribosomal ~H NMR spectra in the presence of fusidic acid. At the same time, limited proteolysis of the ribosomal complex with EF-G indicates that proteins L7/L 12 change their conformation in the complex with the uncleavable GTP analog and return to their initial state after GTP hydrolysis in the presence of fusidic acid and EF-G. 32 Since the studied complexes with GTP and EF-G are functional, it can be assumed, from the totality of data, that the ribosome and/or its components in the process of translocation undergo structural changes, at least in the L7/L12 domain, and after translocation and GTP hydrolysis the ribosome returns to the initial state.

Summary The following conclusions can be drawn from the use of NMR techniques for studies of ribosomes: 1. The majority of ribosomal proteins are rigidly fixed within the particles, and the most mobile components in the isolated ribosome are L7/L 12 proteins from the large subunit. 2. Interaction of EF-G with ribosomes results in some changes in ribosomal domains, and, particularly, immobilization of L7/L12 proteins takes place. The changes may pertain to the translocation reaction, since complexes with ribosomes, EF-G, and GTP are functional. The results of these studies using IH NMR show that structural studies with this technique are limited as only a few proteins express their resonances in the 1H NMR spectra (S 1, L7/L12). At the same time such studies are not exhaustive, since only the simplest samples were studied (ribosomes, the ribosomal complex with EF-G). Complexes with other ligands (tRNA, EF-Tu) have not yet been studied. It is also possible to enhance the resolution of ~H NMR techniques with the help of deuterated factors, 32 A. T. Gudkov and G. M. Gongadze, F E B S Left. 176, 32 (1984).

158


[ 10]

ribosomes, and proteins, and to adapt the use of NMR to other nuclei (e.g., the use of fluorinated labels or incorporation of fluoroamino acids into the proteins). Many other approaches using NMR in biology have still to be explored. Therefore it is hoped that the use of NMR techniques will prove to be very useful in studies of the different functional steps of protein biosynthesis.

[10] P r e p a r a t i o n o f 5 S R N A - R e l a t e d M a t e r i a l s f o r Nuclear Magnetic Resonance and Crystallography Studies B y PETER B. MOORE, STEVEN ABO, BETTY FREEBORN, DANIEL T. GEWIRTH, NEOCLES B. LEONTIS, a n d GRACE SUN

For the past several years, we have been investigating the structure of 5S RNA and its protein complexes using NMR and crystallography. Both of these techniques are notorious for the rate at which they consume materials. A typical sample of 5S RNA for IH NMR, for example, contains 20 mg of RNA, and a serious investigation may require dozens of samples. The fact that NMR is nondestructive, and that samples can be reused is helpful, but the need for large amounts of material remains. In this context, bacterial strains capable of overproducing the materials of interest are more than a mere convenience. They make experiments possible which differ qualitatively from those an investigator with finite resources would seriously contemplate if dependent on wild-type strains only. Below are described the techniques we use for purifying and manipulating 5S RNA and its various nucleolytic fragments, taking advantage of the availability of 5S RNA-overproducing strains. (Those interested in preparing 5S RNA from wild-type strains may consult Refs. 1 and 2, which have appeared earlier in this series, or a recent publication from this laboratory. 3) All the preparations described yield NMR-scale quantities of end product.

i R. Monier and J. Feunteun, this series, Vol. 20, p. 494. 2 R. A. Zimmermann, this series, Vol. 59, p. 551. 3 M. J. Kime and P. B. Moore, Nucleic Acids Res. 10, 4973 (1982).

METHODSIN ENZYMOLOGY,VOL. 164

Copyrisht© 1988byAcademicPress,Inc. Allfightsofreproductionin any formreserved.

158


[ 10]

ribosomes, and proteins, and to adapt the use of NMR to other nuclei (e.g., the use of fluorinated labels or incorporation of fluoroamino acids into the proteins). Many other approaches using NMR in biology have still to be explored. Therefore it is hoped that the use of NMR techniques will prove to be very useful in studies of the different functional steps of protein biosynthesis.

[10] P r e p a r a t i o n o f 5 S R N A - R e l a t e d M a t e r i a l s f o r Nuclear Magnetic Resonance and Crystallography Studies B y PETER B. MOORE, STEVEN ABO, BETTY FREEBORN, DANIEL T. GEWIRTH, NEOCLES B. LEONTIS, a n d GRACE SUN

For the past several years, we have been investigating the structure of 5S RNA and its protein complexes using NMR and crystallography. Both of these techniques are notorious for the rate at which they consume materials. A typical sample of 5S RNA for IH NMR, for example, contains 20 mg of RNA, and a serious investigation may require dozens of samples. The fact that NMR is nondestructive, and that samples can be reused is helpful, but the need for large amounts of material remains. In this context, bacterial strains capable of overproducing the materials of interest are more than a mere convenience. They make experiments possible which differ qualitatively from those an investigator with finite resources would seriously contemplate if dependent on wild-type strains only. Below are described the techniques we use for purifying and manipulating 5S RNA and its various nucleolytic fragments, taking advantage of the availability of 5S RNA-overproducing strains. (Those interested in preparing 5S RNA from wild-type strains may consult Refs. 1 and 2, which have appeared earlier in this series, or a recent publication from this laboratory. 3) All the preparations described yield NMR-scale quantities of end product.

i R. Monier and J. Feunteun, this series, Vol. 20, p. 494. 2 R. A. Zimmermann, this series, Vol. 59, p. 551. 3 M. J. Kime and P. B. Moore, Nucleic Acids Res. 10, 4973 (1982).

METHODSIN ENZYMOLOGY,VOL. 164

Copyrisht© 1988byAcademicPress,Inc. Allfightsofreproductionin any formreserved.

[ 10]

5S MATERIALS FOR N M R AND CRYSTALLOGRAPHY

159

Growth of H B 101/pKK5-1 Most of the 5S RNA we use is the product of the overproducing plasmid pKK5-1 which carries the rrnB 5S cistron. 4 For routine production of 5S RNA it is grown in Escherichia coli HB101 using a supplemented L broth as the medium.

Medium (concentration/liter) Bactotryptone Yeast extract NaCI Glucose 1 N NaOH (to set medium to pH 7.4) Adenosine Uridine 0.2% amipicillin

10 g 5g 10 g 1g 1 ml 0.2 g 0.2 g 10 ml

Ampicillin hydrolyzes if autoclaved. Stock solutions are made up, sterilized by filtration, and can be stored for up to 1 week at 4*. AmpiciUin should be added to media only after they have cooled following autoclaving, and media containing ampicillin should not be stored for more than a day or two prior to use. The adenosine and uridine required are dissolved in a modest amount of water, sterilized by filtration, and, like the ampicillin, added to the medium after it has cooled. The medium includes ampicillin to maintain the plasmid in the host, the plasmid being a cartier of a gene for ampicillin resistance, a gene the host lacks. The supplementation with adenosine and uridine ensures that the ability of the cells to synthesize RNA during overproduction is not limited by the supply of precursors. Stocks of HB 101/pKK5-1 are maintained at - 80 ° as glycerol freezes. It is best to start cultures from single colonies picked from L broth-ampicillin plates inoculated with aliquots of the frozen stocks. Cells are grown at 37 ° until the optical density of the culture reaches 1.5 at 550 nm (about 10 9 cells/ml). At that point 50 mg/liter of chloramphenicol is added to stimulate overproduction. We find that 4 hr of"growth" following the addition of chloramphenicol gives the best yield of 5S RNA. These procedures work satisfactorily for production runs in a 2001 fermentor. tSN Labeling of 5S R N A The information extractable from the physical study of macromolecules can be increased substantially if the molecules of interest can be 4j. Brosius, Gene27, 161 (1984).

160

[ 10]


labeled with stable isotopes such as 2H, ~3C,or ~SN. Since substitution levels often must exceed 50% at the locations of interest to be useful, and since heavily enriched precursors are costly, the expense of preparing labeled macromolecules is a major barrier to their use. The benefits of using overproducing strains in this context are obvious. A 10-fold overproducer makes a $500 experiment out of one which would otherwise cost $5000 in isotopes. We have done a number of experiments using 5S RNA uniformly labeled with ~SN. Cells are grown on a minimal medium where the sole source of nitrogen is a cheap, 'SN-containing compound. The one we use is 99% 15N-enriched NH4CI (MSD Isotopes, St. Louis, MO), which currently costs $80/g. The RNA derived from these cells is, of course, 99% 'SN labeled at all positions. HB101, the "normal" host for pKK5-1, does not grow on minimal media. In order to take advantage of pKK5-1 for 15N labeling, it must be transduced into a host which is prototrophic. The one we use is E. coli NG135 which carries the markers strA, gal-$165, and recA 56. 5 The recA trait is important since it reduces the probability that recombination will take place between the plasmid and the chromosome of the host. pKK5-1 DNA was purified from HBI01/pKK5-1 and transduced into NGI35 using standard methods? Transductants were selected on ampicillin-containing plates. The only difficulty in carrying out the transfer of pKK5-1 from HB 101 to NG135 arises from the fact that HB101-produced DNA is restricted by NG135. In order to get a reasonable number of transductants, about 100 times the amount of plasmid DNA called for in standard protocols is required. We have found that glycerol freezes of NG135/pKK5-1 made from cultures grown on minimal medium are not viable while those made with cells grown on broth are. The first step in growing this strain, therefore, is to plate the glycerol freeze material on minimal plates containing ampicillin. A single colony is then used to start the overnight culture.

lSN Medium (concentration~liter) KH2PO4 Na2HPO4 15NH4C1 NaC1 20% glucose

3g 6g 0.3 g 0.5 g 20 ml

5 N. D. F. Grindley and C. M. Joyce, Proc. Natl. acad. Sci. U.S.A. 77, 7176 (1980). 6 T. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloninig A Laboratory Manual." Cold Spring Harbor Lab., Cold Spring Harbor, New York, 1982.

[ 10]

5 S MATERIALS FOR N M R AND CRYSTALLOGRAPHY

1 M MgSO4" 7H20 0.2% ampicillin 0.01 M CaCI2

161

1.0 ml 10 ml 10 ml

The glucose, MgSO4, and CaCI2 solutions must be autoclaved separately. Ampicillin solution is added to the medium after cooling, as usual. The pH of the medium should be about 7.5. The overnight culture is grown in the medium described above except that 1 g/liter of 14NH4C1 is used instead of ~SNH4C1. The day the full-scale culture is to be grown, the cells in the overnight culture are spun down and resuspended in the isotope-containing medium. The level of NH4C1 in this medium is just adequate to support the growth and overproduction required. Cells are allowed to grow at 37 ° until the optical density at 550 nm is 1.5. Chloramphenicol (50 mg/ liter) is then added to induce plasmid replication and overproduction. Two hours of incubation after chloramphenicol addition seems optimal in this case. We find that NG135 must be grown on a shaker. NG135 cells appear to be killed when they are.as vigorously aerated and agitated as they would be in a fermentor. We grow 10 to 20 liters at a time. Purification of 5S R N A The preparation of 5S RNA from chloramphenicol-treated pKK5-1containing cells is straightforward. The cells are broken open; any of the standard methods will do. The ribosomes and cell debris are removed by centrifugation, and the RNA in the supernatant is isolated by phenol extraction. The 5S component of the resulting RNA mixture is purified by chromatography. The protocol outlined below is the one we use for largescale preparations.

Cell Rupture Solutions 10× A: 1 M NH4C1, 0.1 M magnesium acetate, 5 × 10-3 M ethylenediaminetetraacetic acid (EDTA), 3 m M 2-mercaptoethanol, 0.2 M Tris-HC1, pH 7.5 A: the same as above diluted 10-fold with H20 , the pH reset to 7.4, and the mercaptoethanol concentration maintained at 3 m M A convenient tool for breaking bacteria is a device called a "Bead Beater" (Biospec Products, P.O. Box 722, Bartlesville, OK 74005). This instrument is a small, cheap, anaerobic ball mill which uses glass beads as the working abrasive. One hundred and fifty grams of frozen cell paste is thawed in 67 ml of

162


[ 10]

10X A, 200 mg of lysozyme is added, and the mixture allowed to incubate at 4". The reaction is stopped after 30 rain by the addition of 2 ml of 10% sodium deoxycholate, and a few crystals (about 1 mg) of DNase are added, enough to reduce the viscosity to manageable proportions once the cells have ruptured. A single loading consists of the slurry plus 85 ml of 0.1to 0.11-mm diameter glass beads topped off with additional buffer if necessary to fill the chamber. (The chamber must be full to prevent protein denaturation due to foaming.) The jacket surrounding the chamber is filled with ice and five 30-see cycles of beating carded out. The cell suspension is rinsed out of the chamber using 60 ml of A and centrifuged at slow speed (8,000 rpm for 25 rain, at 4*) to remove the beads and cell debris. The pellets are resuspended in A, subjected to a second round of disruption in the Bead Beater, and centrifuged as before. The supernatants from the two cycles are pooled, and the ribosomes spun out in an ultracentrifuge. Two hours of centrifugation in a Ti60 rotor at 60,000 rpm (4°) suffices. The supernatant contains the products of overproduction. The ribosomal pellet is usually discarded because the amount of 5S RNA it represents is too small to be worth recovering. The glass beads can be reused. After extensive washing with water to remove cell debris, the beads are soaked in 1 N HC1 overnight, and then rinsed with water until the pH returns to neutrality. The beads are then put through an overnight soak in 1 N NaOH followed by a water rinse until the pH again returns to neutrality. They are stored after drying in an oven.

Phenol Extraction Solutions Phenol: "Liquified phenol," the standard commercial product, is redistilled to remove the water it contains, and a variety of colored oxidation products. We keep the fraction which distills between 178" and 182". (Boiling chips are essential to prevent bumping.) The redistilled material is a solid at room temperature, and can be stored in glass bottles at - 2 0 " . It is melted when needed in boiling water, or in a water bath. It must be equilibrated with the buffer to be used in the extraction procedure. A separatory funnel is useful for this purpose. Once equilibrated, and saturated with buffer it remains a liquid at room temperature. Equilibrated phenol is prepared fresh every time it is required. SSC-EDTA: 0.15 MNaC1, 15 m M s o d i u m citrate, 15 mMEDTA, pH 7.0. SDS: 10% (w/v) sodium dodecyl sulfate (Pierce) in SSC-EDTA.

[ 10]


163

The RNA contained in the centrifugal superuatant is isolated by phenol extraction. Any of the standard protocols for phenol extraction will work. We do it in the following manner. To the postribosomal supernatant is added 0.1 volumes of SDS solution, and a volume of phenol equal to the volume of the supernatant. The mixture is shaken vigorously for 1 min. We often do this in a l-liter glass bottle with a frosted glass stopper. The suspension is then centrifuged at a few thousand rpm for a minute or two to break the phenol-water emulsion. The upper, aqueous layer is removed, and saved in another bottle on ice. An equal volume of SSC-EDTA is added to the phenol layer, and the mixture shaken, centrifuged, and the water phase saved as before. The two water phases are pooled, and phenolextracted twice more. The nucleic acid in the resulting protein-free solution is recovered by adding 2 volumes of 95% ethanol. The ethanolic solution is allowed to stand for about 1 hr at - 2 0 ° to encourage complete precipitation of RNA. The precipitate is collected by centrifugation at 5000 g for 15 min at 4 °. The RNA pellet is taken up in S-200 buffer (see below), and can be stored indefinitely in that buffer at - 2 0 °. Chromatography Solutions

S-200 buffer: 0.15 M NaCI, 0.1 M sodium acetate, 1% (v/v) methanol, with the pH set to 5.0 using acetic acid The fractionation of crude, low-molecular-weight RNA is conveniently done by chromatography on columns packed with Sephacryl S-200 (Pharmacia). 7 A 5 × 100 cm column is poured and equilibrated with S-200 buffer. A sample in 20 ml or less is applied to the column, and the chromatography carried out at room temperature. We collect 100 16-ml fractions. The RNA elution profile can be monitored by reading the optical density of the undiluted fractions at 300 or 305 nm. A typical example is shown in Fig. 1. When running at rates of around 150 ml/hr, these columns will fractionate up to 10,000 A2~o,m of crude RNA without loss of performance. The resolving power of these S-200 columns degrades after a period of use. The degradation is particularly rapid if the samples being run contain the slightest amount of poorly solubilized material. Columns can be regenerated by washing the Sephacryl according to the manufacturer's instructions. The column is unpacked and the Sephacryl suspended in 0.2 N NaOH at room temperature. The column is then repacked in alkali, and a 7T. H. Kao and D. M. Crothers,Proc. Nate. Acad. Sci. U.S.A. 77, 3360 (1980).

164

[ 10]


25

c

o

tO CM v

20

>,-

---- 15 Z

p0

5 / x . t / PO

,

I ~ " T ,~ I 40 60 FRACTION NUMBER

, 80

FIG. 1. Chromatography of low-molecular-weight RNAs from HBI01/pKK5-1 on Sephacryl S-200. Several thousand A2~o,m of low-molecular-weight RNA were applied to a 5 × 100 cm column of Sephacryl S-200 and eluted with S-200 buffer (see text). The RNA was derived from HB 10 I/pKK5-1 cells following overproduction as described elsewhere. Sixteenmilliliter fractions were collected. The peak at fraction 60 is 5S RNA. The peak at fraction 70 is tRNA.

bed volume or two of S-200 buffer run through it. The column can be used as soon as the pH returns to 5.0. The yield of 5S RNA at the end of the sequence of steps just outlined can be as high as 7.5 nag (150 A26om) per gram wet weight of cells, which is about 20 times the total amount of 5S RNA present in the ribosomes derived from same amount of starting material. (The yield from cells grown on minimal medium is substantially less, about 2.5 mg/liter of starting medium.) The column product is recovered by ethanol precipitation, and redissolved in S-200 buffer. 5S RNA is stored in this buffer at - 2 0 °. It is exclusively in the A form.

Preparation of Fragment 1

Buffers IX: 0.1 MKC1, 5 m M MgC12, 50 mMtris, 90 m M b o r i c acid, pH 7.8 Fragment buffer: 0.1 M NaCI, 3 m M MgCI2, 0.1 g/liter NAN3, 10 m M cacodylic acid, pH 6.0

[ 10]

5S MATERIALS FOR N M R ANI) CRYSTALLOGRAPHY

165

Limited digestion of 5S RNA with RNase A leads to the production of a fragment of the molecule consisting of bases (1- 11, 69-120) of the parent sequence,s The molecule is called fragment 1, and has a structure similar to that of the same sequences in the parent molecule? The protocol for making this material is as follows. 5S RNA is precipitated with ethanol, and brought up in 1X at a concentration of 20 A26om / m l (nominally 1 mg/ml). The solution is left on ice until the temperature is below 4 °. RNase A is then added at levels between 20 and 0.5 #g/ml. The mixture is allowed to incubate for 45 rain on ice. Digestion is terminated by making the solution 0.1% in SDS, and immediately extracting it with phenol. Three extractions are done to ensure complete removal of the enzyme. The main cleavage points in the 5S sequence under the conditions just described are after C-I 1, and C-68. What is controlled by variation of the RNase concentration is the degree to which the fragment produced is also cleaved at the 87- 89 loop. The lower the enzyme concentration, the larger the fraction of molecules in the preparation unclcaved in that loop. The "penalty" for reduced cleavage in the loop is the appearance of molecules in the population whose cleavage point is to the 5' side of C-68. One chooses the RNase level according to the needs of the experiment for which the preparation is being made. Fragment 1 is purified from other components in the digestion mixture by chromatography on Sephadex G-75 in fragment buffer (Fig. 2). The temperature of the column is critical for this purification. If it is too low, the resolution of the column will be poor. If it is too high the strands of the product will begin to dissociate. A reasonable compromise is 37 °. A 2.5 × 100cm (jacketed) column will suffice to purify the products of a 2000 A26onmdigestion. Fractions (5.5 ml) are collected. The buffers used in these columns must be degassed, and then maintained at or above the column temperature to prevent the destruction of the column bed due tO the formation of gas bubbles. The yield of fragment 1 is typically 250-300 A26om per 1000 A26onmof5S RNA digested. Preparation of Fragment 2

Buffer Fragment 2 buffer: 0.1 M KC1, 10 m M magnesium acetate, 50 m M Tris-HC1, 90 m M boric acid, pH 7.8 8 s S. Douthwaite, R. A. Garrett, R. Wagner, and J. Feunteun, Nucleic Acids Res. 6, 2453 (1979). 9 M. J. Kime and P. B. Moore, FEBSLett. 153, 199 (1983).

166

[ 1 O]


Z5 otO 20 od >o~ Z t.tJ ~3

j

b-13.

°

--//

t

20

~"~

I

~

]

t

40 60 FRACTION NUMBER

I

80

FIG. 2. Chromatography of a fragment 1 digest on SephadexG-75. One thousand A26o.. of intact 5S was digested at a weight ratio of RNA to R_Nas¢A of 100:1 under the conditions described in the text for the production of fragment 1. The ethanol-precipitable products of this digestion were chromatographed on a 2.5 × 100 cm column of Sephadcx G-75 at 37" and 5.5 ml fractions were collected. The elution profile is shown. The peak at fraction 45 is

fragment 1. By careful adjustment of the conditions of RNase A digestion a second fragment of 5S RNA can be isolated in reasonable yield. It comprises bases 15- 36 and bases 4 4 - 6 5 of the parent sequence. We call this oligonucleotide fragment 2. It includes most of helix II and all of helix III of native 5S RNA. 10 5S RNA is precipitated with ethanol and brought up in Fragment 2 buffer at a concentration of 20 A ~ o , J m l . RNase A is added at a concentration of 1.1 gg/ml, and the mixture incubated at 0* for 45 rain. Digestion is then terminated by phenol extraction, and the product purified by Sephadex chromatography exactly as described for fragment 1. A typical elution profile is shown in Fig. 3. Both fragment 1 and fragment 2 are produced in such a digest. The fragment 1 component is only slightly nicked in the 87,88,89 loop. Partial Denaturation and Reconstitution of F r a g m e n t 1

Buffers EDTA buffer: 0.1 M NaCI, 2 m M EDTA, 10 m M 2-~r-morpholinoethanesulfonic acid (MES), pH 6 ~oN. B. Leontisand P. B. Moore,Biochemistry 25, 5736 (1986).

[ 10]


167

1

15 E ¢::: O cO ) - 10 I-Z LI.I C~

2

_J I-O_ 0

,,I

20

i l l

60 40 FRACTION NUMBER

I

80

Flo. 3. Chromatographyof a fragment2 digest on Sephadex(3-75. One thousand A2~o of intact 5S RNA was digestedat a weight ratio of RNA to RNase A of 1000:1 under the ionic conditions described for the preparation of fragment 2. The elution profile of the ethanol-precipitableproducts of this digestionare shown. The conditionsfor chromatography were those used for the purificationof fragment 1 (see text). The first (unnumbered)peak is a partial digestionproduct we have not yet characterized.The peak marked "1" is fragment 1. The peak marked "2" is fragment2 material. Reconstitution buffer: 0.1 M KC1, 5 m M MgClx, 10 mMN-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid (HEPES), pH 7 Fragment 1 preparations contain four covalent species we call "strands." Strand I is bases 69-120. Strand II is bases 88 (or 89)-120. Strand III is bases 6 9 - 8 7 , and strand IV is bases 1-11. Fragment 1 preparations are mixtures of the complexes of strands I with IV, and of strands II, III, and IV. When fragment 1 is heated, especially in the absence of Mg 2+, strand III tends to dissociate from the (II,III,IV) complex. This fact can be taken advantage of in a number of ways. It can be used as a method of separating the (I,IV) form of the complex from the (II,III,IV) version. It can also be exploited as a means of producing partially labeled fragment by taking advantage of the fact that (II,IV) readily recombines with III to reconstitute fragment 1. Fragment 1 is precipitated with ethanol and taken up in EDTA buffer. It is then applied to a Sephadex G-100 column equilibrated with the same buffer. Chromatography is carried out at 300-35 ° using warm, degassed EDTA buffer as the eluant. Figure 4 shows a typical elution profile. The performance of this column is crucially dependent on the monovalent

168


[ 1 0]

I+IX

0.5

"tI, l v I 0.4 E

c

O O ro

c3 o

0.3

0.2

tin)

O.1

~--4/

t

50 70 90 FRACTION NUMBER

Fro. 4. Chromatography of fragment 1 on Sephadex G-100 in EDTA. A preparation of fragment 1 obtained by digesting 5S RNA with RNase A at a weight ratio of 800:1 was applied to a 2.5 3< 100 cm column of Sephadex G-100 in 0.1 MNaCl, 2 m M EDTA, 10 m M MES, pH 6, at 35 °. Then 5-ml fractions were collected and read. The peak at fraction 62 is fragment uncleared at the 87,88,89 loop, i.e., the complex of strands I and IV. The peak at fraction 72 consists of molecules containing bases 1- 11, 89-120, strands IV and II, and the third peak, the one at fraction 90, is strand III, bases 69-87.

cation used in the buffer. Na + works; K + does not. The products are recovered by ethanol precipitation as usual. Samples of the (II,IV) complex and strand III are dialyzed (separately) into reconstitution buffer in the cold. They are then mixed to produce a solution which is 12.00D2e e m / m l in II,IV and 8.00D2~o nm/ml in strand III. The mixture is heated at 60 ° for 10 min, and then allowed to cool to room temperature for 45 min. The product is recovered by ethanol precipitation and purified by chromotography on Sephadex G-75 following the protocol prescribed for the initial purification of fragment 1. The material regenerated in this way has normal gel electrophoretic mobility, binds L25 normally, and gives a standard proton NMR spectrum, n 11 D. T. Gewirth, S. R. Abo, N. B. Leontis, and P. B. Moore,

Biochemistry26, 5213 (1987).

[10]

5S MATERIALS FOR N M R

A N D CRYSTALLOGRAPHY

I

I

I

[

L 40

I

I 60

169

I

I

0.4

m

0.3E

c 0 0

0.2-

0.1-

\ I

0

L 20

I

80

FRACTION NUMBER F]o. 5. Chromatography of fragment 1 on Sephadex G-75 in urea. A preparation of fragment 1 like the one used in the experiment depicted in Fig. 4 was dissolved in dissociation buffer, and chromatographed on a 2.5 × 100 em column of Sephadex G-75 at 50*. Then 5.5-ml fractions were collected and read. The oligonucleotide components of the fragment 1 preparation elute in order of molecular weight: strand I (bases 69-120), strand II (bases 89-120), strand III (bases 69-87), and strand IV (bases 1- 11). Reproduced with permission from M. J. Kime, D. T. Gewirth, and P. B. Moore [Biochemistry23, 3559 (1984). Copyright 1984 American Chemical Society.]

Total Dissociation of Fragment 1 Strands

Buffer Dissociation buffer: 8 M urea, 50 m M NaCI, 1 m M E D T A , 10 m M cacodylate, p H 6.0 F r a g m e n t 1 dissociates entirely in urea buffers at elevated temperatures.t2 F r a g m e n t 1 is precipitated with ethanol a n d taken up in dissociation buffer. T h e solution is e h r o m a t o g r a p h e d on Sephadex G-75 or (3-100 in the same buffer at 50*. A typical elution profile is shown in Fig. 5. T h e strands are recovered by ethanol precipitation. 12 M. J. Kime, D. T. Gewirth, a n d P. B. Moore,

Biochemistry23,

3559 (1984).

170


[ 10]

Analytical Gels

Solutions Acrylamide-Bis: 20% (w/v) acrylamide, 1% (w/v) bisacrylamide 10× TBE: 25 m M EDTA, 500 m M Tris-HC1, 500 mMboric acid, pH 8.3 Persulfate: 10% (w/v) ammonium persulfate 10X 5S: 50 m M magnesium acetate, 1 M KC1, 500 m M Tris-HC1, 900 m M boric acid, pH 7.6 The standard method for examining the products of the manipulations described above is acrylamide gel electrophoresis. We run two kinds of gels for routine use, one which permits 5S and its protein complexes to run intact, and a second which is urea containing, and displays the oligonucleotides a sample contains. The native gel mixture contains 50 ml of acrylamide-Bis, 10 ml of 10X 5S, 80 #1 of tetramethylethylenediamine (TEMED), 0.5 ml of persulfate, and enough water to bring the total volume to 100 ml. We usually pour slab gels which are 170 × 170 X 1.5 ram. RNA, 0.1 to 0.3 OI:)26o m, is loaded per lane. The running buffer is 10× diluted 10-fold. Gels are run at 3 V/cm for 16 hr at room temperature using bromphenol blue as the marker dye. The current which runs through these gels is high. In order to keep the pH constant it is necessary to recirculate the running buffer. The composition of the denaturing gel mix is 60 ml of acrylamide-Bis, 10 ml of 10× TBE, 48 g urea, 80 #l of TEMED, and 0.5 ml of persulfate. (The volume of this mixture is about 100 ml.) The loading of these gels is the same as that of the native gels. We usually run them at 15 V/cm at room temperature using 10× TBE diluted l:10 as the running buffer. Bromphenol blue and xylene cyanole are good marker dyes. The electrophoresis is complete in 3 to 4 hr under these conditions. The results of the electrophoresis can be visualized in a number of ways. A gel can be quickly examined by placing it on a TLC plate containing fluorescent dye and illuminating the gel with a hand-held UV lamp which emits at 254 rim. The RNA-containing bands appear as shadows against a bright background. (Wear glasses to protect your eyes from UV light while inspecting your gel!) Gels can be stained for RNA using methylene blue. The gel is soaked in 10% acetic acid for l0 or 15 min, and then transferred into 0.4% methylene blue, 0.2 M acetate, pH 4.7. An hour of staining with agitation is sufficient. The gel is then destained in tap water by diffusion. Usually it can be read fairly well after a few hours of destaining with several changes of water. The stain solution can be reused many times. The sensitivity of methylene blue staining is quite high, higher than

[ 10]

5S MATERIALSFOR NMR AND CRYSTALLOGRAPHY

171

the shadowingmethod; 0.5/tg of RNA in a 1-cm wide band is easily detected. Methylene blue does not stain protein. A dye which will stain protein, but not RNA is Coomassie blue. We use a 0.025% solution of Coomassie blue in a 50: 50 mixture of 7% acetic acid and ethanol. The staining takes several hours, and destaining is by diffusion into ethanol-acetic acid solution. One microgram of protein is readily visualized. Ifa protein which binds to 5S RNA or one of its fragments is added to a lane with RNA in it, the electrophoretic mobility of the RNA on native gels is altered, a fact which provides a simple assay for this interaction. The presence of both protein and RNA in the complex is readily proven using the RNA and protein staining methods described here. 5S-Binding Ribosomal Proteins Fortunately, our work has consumed ribosomal proteins at a rate of less than 50 mg/species/year. The reason this fact is fortunate is that as far as we are aware there are no strains which overproduce any ribosomal protein to an appreciable degree. The techniques for preparing ribosomal proteins are in rapid flux today due to the introduction of high-performance liquid chromatography (HPLC) methods into the field. Below is described a more traditional ion-exchange approach to purifying the 5S-binding proteins from E. coli. It is slower than HPLC methods, if all that is needed is 10 mg of material. On the other hand, unlike HPLC, it is easy to scale up; 50 mg preparations of single proteins are readily accomplished. The three 5S binding proteins of E. coli, LS, L18, and L25, can be purified either from 50S subunit protein or from the protein of 70S particles. The ion-exchange procedures we use for purifying ribosomal proteins are detailed in Vol. 59 of this series, t3 and will be briefly summarized here. Protein is extracted from subunits or whole ribosomes by the LiC1-urea method, and the protein dialyzed exhaustively against 6 M urea, 30 m M methylamine, 6 m M 2-mercaptoethanol, adjusted to pH 5.6 with acetic acid ("Start 5.6"). In a recent run, the protein from 550,000 O D ~ o m of 70S ribosomes (about 12 g of protein) was loaded onto a 1-liter column of carboxymethyl-cellulose (CMC, Whatman CM-52) and eluted with a linear, 0.02 to 0.15 M NaC1 gradient whose total volume was 22 liters. By 0.15 M NaC1, all the 5S proteins are offthe column. Figure 6 shows the elution profile which results when a CMC column with lower loading (5 nag protein/ml of column volume) is run with a gradient of the same slope as described above extending past 0.25 M NaCI. The solid trace shows the elution of 70S protein (OD2ao nm), and the broken ,3 p. B. Moore,this series,Vol. 59, p. 639.

172

OTHER

--

BIOPHYSICAL

METHODS

[ 1 O]

00~0m

....... 3H-CP M

L1

'1

,.

L3

iil

°

o

L19 L22

L?4

•

~

|"

: ~

:,i

0

5o

)0o

,~o

2oo 250 30o FRACTION NUMBER

: . .

,,

55o

, ..........

4oo

45o

FIG. 6. The elution profiles of 70S protein and 50S protein from CMC compared. Nonradioactive 70S protein was mixed with a small amount of ~H-labeled 50S protein, and the mixture resolved by chromatography on CIVICat pH 5.6 in 6 M urea, as described in the text. The NaC! concentration was increased finearly during the elution, and ran from 0 to 0.25 M across the span represented in the figure. 70S protein was detected by its absorption at 230 nm. 50S protein was detected by its radioactivity. The identities of the 50S components in a number of peaks are indicated. 5S-binding proteins arc designated by numbers enclosed in a box.

line trace 50S protein (all counts). The identities of the 50S proteins contributing to a number of the peaks is given. L5 elutes free of all other proteins, as does L25. It is our practice to purify both fractions further by chromatography on Sephadex G-100 in Start 5.6. Following this step, both are sufficiently pure to use for physical studies. L 18 elutes in a mixture which includes three other 50S proteins and $4. $4 is enough larger than LI 8 and the other 50S protons in this fraction that Sephadcx G-100 chromatography will remove it cleanly from the mixture. The fractionation of the four 50S proteins is more tedious. The mixture is dialyzed into 6 M urea, 20 m M phosphoric acid-methylaminc, pH 7.0, 6 m M 2-mercaptoethanol ("Start 7"). It is then run on CMC in the same buffer system. The loading used is the same as that for the initial column. If the first step requires a l-liter column, the pH 7 step does as well. The elution is carded out using a linear gradient of NaC1 running from 0.075 to 0.15 M with about the same slope as the initial pH 5.6 column. Three peaks will be detected at 230 nm. The order of elution is L13 first, followed by L22, and then L18, and L19 as a single peak. The L 1 8 - L I 9 mixture can be resolved on phosphocellulose, but we prefer to do it using reversed-phase HPLC. We use a SynChropak RP-P column (Cl8). The solvent system is 0.1% trifluoroacetic acid in water: 0.1% triflu-

[ 10]


17 3

oroacetic acid in acetonitrile. A full description of the application of HPLC methods to the purification of ribosomal proteins may be found elsewhere in this volume, t4 In order to make use of these proteins, they must first be renatured. L25 is dialyzed into 0.1 MKC1, 10 m M Tris-HC1, pH 7.5, at 4 °. Indeed, L25 recovers its native conformation upon dialysis into almost any physiological buffer, t5 (L25 denatures below pH 6.0.) The other two proteins are more difficult to handle. First they are dialyzed against 6 M guanidine, 6 m M 2-mercaptoethanol, 20 m M cacodylate, pH 7.0, and then dialyzed into 1 M KCI, 6 m M 2-mercaptoethanol, 10 m M cacodylate, pH 7.0. What appears to be important is that the proteins be kept at high ionic strength while the denaturant is removed. Following the second dialysis, the proteins can be transferred into any buffer one wishes by further dialysis. Renatured L5 is quite unstable. It cannot be frozen in physiological buffers unless it is made 50% in glycerol first. Renatured L 18 is more forgiving than renatured L5; it can be frozen and thawed in the absence of glycerol. Concentrated solutions of L 18, however, like similar solutions of L25, tend to form precipitates of (presumably) denatured protein at a slow rate upon incubation at room temperature. Using biuret as a means for estimating protein concentrations, we find that a 1 mg/ml solution of L18 has an optical density of 0.42 at 276 nm, and that a similar solution of L25 has an optical density of 0.38 at the same wavelength. Both these values are close to those one would estimate for these proteins on the basis of their amino acid compositions, taking standard values for the extinction coefficients of the aromatic amino acids, and using their known molecular weights. (We have not measured an extinction coefficient for L5, but calculation based on amino acid composition suggests that a 0.1% solution of L5 should have an extinction coefficient of about 0.7 at 280 nm.) The extinction coefficient of 5S RNA at 260 nm is 8.474 × 105 M - t cm-I. t6 For the smaller fragments of 5S RNA, we estimate concentrations form extinction at 260 nm using the extinction coefficient of the intact molecule multiplied by the ratio of the molecular weights of intact 5S RNA and the fragment in question. Preparation of Samples for N M R

and Crystallization

N M R samples for proton work are typically0.2 to 0.5 ml in volume, and the molecule of interestmust be present at a concentration around 14B. S. Cooperman, C. J. Weitzmann, and M. A. Buck,this volume [36]. 15M. J. Kime, R. G. Ratcliffe,P. B. Moore,and R. J. P. Williams, Eur. J. Biochem. 116, 269 (1981). 16R. Osterberg,B. Sjoberg,and R. A. Garrett, Eur. J. Biochem. 116, 481 (1976).

174


[ 11 ]

1 mM. Since neither 5S RNA nor its proteins are particularly robust, we transfer them into the buffer in which we wish to study them by dialysis at a stage where the concentration is of the order of a tenth of that desired for the final sample. Concentration is achieved by ultrafdtration. A disposable, centrifugal ultrafdtration device has become available in the past few years which is well suited for this purpose. The device is called the Centricon and is manufactured by Amicon. It permits one to concentrate 3 ml down to less than 100 gl, if necessary, with acceptably low losses. Both fragment 1 and its complex with L25 crystallize readily under a wide range of conditions. 17 Samples suitable for crystallization are much less concentrated than those used in NMR. A typical preparation for crystallization has an RNA concentration of 100 OD260, J m l . The ionic conditions are varied to meet the particular purposes of the experimenter, but typically would be 0.1 M KCI, 4 m M magnesium acetate, 5 m M TrisHCI, pH 7.4. Again dialysis is the preferred way of establishing the ionic conditions in a sample. Acknowledgments This work has been supported by grants from the National Institutes of Health to P.B.M. (AI-09167, GM-22778, and GM-32206).

17 S. S. Abdel-Meguid, P. B. Moore, and T. A. Steitz, J. Mol. Biol. 171,207 (1983).

[ 11 ] F l u o r e s c e n c e L a b e l i n g a n d I s o l a t i o n o f L a b e l e d RNA and Ribosomal Proteins B y O . W . O D O M , H . - Y . D E N G , a n d BOYD H A R D E S T Y

The development of highly sensitive fluorimeters to measure steadystate intensity and lifetime of fluorescence coupled with computerized data analysis has vastly enhanced the utility of fluorescence techniques for investigation of systems of biological origin. Estimation of the distance between an energy donor fluorophore and an acceptor by nonradiative energy transfer has proven to be particularly useful. This technique has been used extensively to analyze the structure and function of ribosomes during the reaction steps of protein synthesis. The size of ribosomes, about 220 A in diameter, is well suited to the distances that can be measured by nonradiative energy transfer, about 20 to 80 A with commonly used METHODS IN ENZYMOLOGY,VOL. 164

Copyright© 1988by AcademicPress, Inc. Allfightsofreproductionin any formreserved.

174


[ 11 ]

1 mM. Since neither 5S RNA nor its proteins are particularly robust, we transfer them into the buffer in which we wish to study them by dialysis at a stage where the concentration is of the order of a tenth of that desired for the final sample. Concentration is achieved by ultrafdtration. A disposable, centrifugal ultrafdtration device has become available in the past few years which is well suited for this purpose. The device is called the Centricon and is manufactured by Amicon. It permits one to concentrate 3 ml down to less than 100 gl, if necessary, with acceptably low losses. Both fragment 1 and its complex with L25 crystallize readily under a wide range of conditions. 17 Samples suitable for crystallization are much less concentrated than those used in NMR. A typical preparation for crystallization has an RNA concentration of 100 OD260, J m l . The ionic conditions are varied to meet the particular purposes of the experimenter, but typically would be 0.1 M KCI, 4 m M magnesium acetate, 5 m M TrisHCI, pH 7.4. Again dialysis is the preferred way of establishing the ionic conditions in a sample. Acknowledgments This work has been supported by grants from the National Institutes of Health to P.B.M. (AI-09167, GM-22778, and GM-32206).

17 S. S. Abdel-Meguid, P. B. Moore, and T. A. Steitz, J. Mol. Biol. 171,207 (1983).

[ 11 ] F l u o r e s c e n c e L a b e l i n g a n d I s o l a t i o n o f L a b e l e d RNA and Ribosomal Proteins B y O . W . O D O M , H . - Y . D E N G , a n d BOYD H A R D E S T Y

The development of highly sensitive fluorimeters to measure steadystate intensity and lifetime of fluorescence coupled with computerized data analysis has vastly enhanced the utility of fluorescence techniques for investigation of systems of biological origin. Estimation of the distance between an energy donor fluorophore and an acceptor by nonradiative energy transfer has proven to be particularly useful. This technique has been used extensively to analyze the structure and function of ribosomes during the reaction steps of protein synthesis. The size of ribosomes, about 220 A in diameter, is well suited to the distances that can be measured by nonradiative energy transfer, about 20 to 80 A with commonly used METHODS IN ENZYMOLOGY,VOL. 164

Copyright© 1988by AcademicPress, Inc. Allfightsofreproductionin any formreserved.

[11]

SEPARATION OF RIBOSOMAL PROTEINS AND t R N A s

175

probes. Individual ribosomal proteins or rRNAs can be labeled, then reconstituted into active ribosomes, and probes can be covalently linked to tRNA at a number of specific sites. Several excellent articles 1-a describe in detail the instrumentation, techniques, and data analysis used for distance measurement by nonradiative energy transfer. These aspects will not be treated here. Rather, the focus will be directed toward what in most experimental situations is the most troublesome factor that may compromise the reliability of the measurements. This is incomplete pairing between donor and acceptor probes and spurious positioning of the probes. These conditions may be generated by incomplete or nonspecific labeling, respectively. In many experimental situations, incomplete pairing can be dealt with by determining energy transfer from fluorescence lifetime data. Under favorable circumstances, a two-exponential function can be fit to a fluorescence decay curve so that fluorescence lifetime of the quenched and unquenched species can be evaluated directly, and we have used this approach? Also, increased fluorescence from the energy acceptor can be used with the decrease in donor fluorescence intensity in some circumstances3 However, the most direct and generally the most satisfactory approach is to avoid the problem by using completely, and specifically, labeled components. Techniques for obtaining such ribosomal proteins and tRNAs are described below. Labeling P r o c e d u r e s Precautions must be taken throughout the labeling and isolation procedure to use reagents, solvents, and glassware that are as free as possible of contaminating fluorescent material and degradative enzymes. In this laboratory, distilled water is passed through a deionizing column and then a column of activated charcoal before it is redistilled in glass. Where possible, plastic containers are avoided for solution storage. Glassware washed with detergent is rinsed with hot ethanol before it is used. Eppendorf tips and tubes are never reused. Plastic gloves are worn throughout the labeling procedure to avoid both fluorescent and enzymatic contamination. When working with low concentrations oftRNA or 5S RNA and always with 16S RNA or 23S RNA, glassware is baked at 150 ° for 24 hr to reduce nuclease contamination. 1j. R. Lakowicz, "Principles of Fluorescence Spectroscopy." Plenum, New York, 1983. 2 C. R, Cantor and T. Tao, Proced. Nucleic Acid Res. 2, 31 (1971). 3 R. H. Fairclough and C. R. Cantor, this series, Vol. 48, p. 347. 4 D. Robbins, O. W. Odom, Jr., J. Lynch, G. Kramer, B. Hardesty, R. Liou, and J. Ofengand, Biochemistry 20, 5301 (1981). 5 B. Epe, K. G. Steinhliuser, and P. Woolley, Proc. Natl. Acad. Sci. U.S.A. 80, 2579 (1983).

176

[ 11 ]

OTHER BIOPHYSICALMETHODS TABLE I E. coli Ribosomal Proteins with Cysteine Residues* Large subunit

Small subunit

number

Cysteine position

Total amino acids

2 5 6 10 II b 14 17 27 b 28 b 31

5,187 86 124 70 38 21, 84 100 52 4 16,18,37,40

272 178 176 165 141 123 127 84 77 62

L

S

Cysteine

number

position

Total amino acids

Ib 2 4 8 11 12 13b 14 17b 18 21 b

292,349 86 31 126 69,120 26, 33, 52, 103 84 63 58,63 10 22

557 240 203 129 128 123 117 98 83 74 70

° Work of B. Wittmann-Liebold and co-workers, summarized in B. Wittmann-Liebold, Adv. Protein Chem. 36, 56 (1984). b Mutant E. coli strains have been isolated by Dabbs~4O with ribosomes lacking these proteins or

carrying a defective protein that can easily be removed from the ribosome, as described further in the text.

Labeling of Proteins Reaction of cysteine thiol with a maleimide derivative of a fluorescent probe provides a specific reaction for labeling a ribosomal protein. At pH 7.0 maleimides show a high degree of selectivity for sulfhydryl groups of proteins. 6 Alkyl halides such as iodoacetamide derivatives also can be used, but these provide lower sulfhydryl group specificity, n A number of Escherichia coli ribosomal proteins have one or a small number of cysteine residues as indicated in Table I and thus provide a basis for probe attachment at a specific site. 821, 7 L11, 8 81, 9 L27, and L28 have been successfully labeled in this laboratory. A typical labeling procedure involves incubation of 1 mg of the homogeneous ribosomal protein in 200/tl of 7 M guanidine-HC1, 10 m M N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES)/KOH, pH 7.0, with 1 m M of the fluorogenic maleimide 6G. E. Means and R. E. Feeney, "Chemical Modification of Proteins," pp. 105-118. Holden-Day, San Francisco, California, 1964. O. W. Odom, E. R. Dabbs, C. Dionne, M. Mfiller, and B. Hardesty, Eur. J. Biochem. 142, 261 (1984). s H.-Y. Deng, O. W. Odom, and B. Hardesty, Fur. J. Biochem. 156, 497 (1986). 90. W. Odom, H.-Y. Denf, A. R. Subramanian, and B. Hardesty, Arch. Biochem. Biophys. 230, 178 (1984).

[ 1 1]

SEPARATION OF RIBOSOMAL PROTEINS AND tRNAs

177

derivative at 37 ° for 30 min. Excess labeling reagent is inactivated by adding glutathione to a final concentration of 10 raM, then the sample is passed over a Sephadex G-25 column equilibrated with 7 M urea, 20 m M HEPES/KOH, pH 7.5. Guanidine-HCl and inactivated excess labeling reagent are separated from the labeled protein. Finally, the protein is dialyzed against 30 m M Tris-HCl, pH 7.4, 500 m M KC1, 20 m M magnesium acetate, 1 m M dithiothreitol (30S proteins) or 20 m M Tris-HC1, pH 7.4, 400 m M NH4C1, 4 m M magnesium acetate, 4 m M 2-mercaptoethanol (50S proteins), and then frozen in small aliquots and stored at - 80 ° until used. If the labeled protein is to be purified by high-performance liquid chromatography (HPLC) it is dialyzed against 0.1% trifluoroacetic acid (TFA) in water. For determining stoichiometry of labeling, the protein is digested with proteinase K (50#g/ml in 10 m M Tris-HC1, pH 7.4, 100 m M NH4CI, 10 m M magnesium acetate, 5 m M 2-mercaptoethanol for 30 min at 37 °) and the absorbance due to the fluorophore is compared with that of a standard solution of the cysteine derivative of the probe, to obtain probe concentration. 7 Protein concentration is usually determined by the Amido Black method, ~° using bovine serum albumin as a standard.

Labeling of RNAs Ribosomal RNAs and tRNAs can be labeled at their Y-terminal ends by oxidation of the ribose with periodate, followed by reaction of the resulting dialdehyde with hydrazides or thiosemicarbazides, one molecule of which appears to add to both aldehydes to form a morpholine structure) ~Oxidation of 5S ribosomal RNA or tRNA with periodate is accomplished by incubation at a concentration of 50 A2~/ml or less with 0.09 M sodium periodate in 0.1 M sodium acetate, pH 5.0, for 90 rain at room temperature in the dark. t2 At the end of the incubation KC1 is added to a final concentration of 0.2 M, then the reaction solution is allowed to stand on ice for 10 min before the potassium periodate precipitate is removed by centrifugation for 5 min at 10,000 g. The supernatant is passed over a Sephadex G-25 column equilibrated with 0.1 M sodium acetate, pH 5.0, to remove remaining periodate. 16S or 23S ribosomal RNA is oxidized under somewhat milder conditions by incubating with 0.09 M sodium periodate in 0.1 M sodium phosphate, pH 7.0, for 2 hr at 0 ° in the dark at concentrations of no more than 200 or 400 A260 units/ml, respectively) 2 KC1 is then added and the precipitate of potassium periodate is removed as 10W. Schaffner and C. Weissmann, Anal. Biochem. 56, 502 (1973). 1~F. Hansske, M. Sprinzl, and F. Cramer, Bioorg. Chem. 3, 367 (1974). t20. W. Odom, D. J. Robbins, J. Lynch, D. Dottavio-Martin, G. Kramer, and B. Hardesty, Biochemistry 19, 5947 (1980).

178


[ 1 1]

described for 5S RNA and tRNA, followed by passage over Sephadex G-25 equilibrated with 0.05 M sodium phosphate, pH 7.0, to remove remaining periodate. Labeling of the oxidized RNAs with fluorescein 5'-thiosemicarbazide (FTS) as an example is as follows, n Oxidized 5S RNA or tRNA in 0.1 M sodium acetate, pH 5.0, is incubated with 2 m M FTS in the dark for 2 hr at room temperature, followed by three extractions with equal volumes of 70% phenol to remove most of the unreacted labeling reagent. Then the sample is made 0.1 M in KC1 and precipitated with two volumes of 95% ethanol at - 2 0 °. The ethanol precipitation is repeated twice, after which the sample is dissolved in 10 m M Tris-HCl, pH 7.4, 4 m M magnesium acetate, and frozen at - 8 0 °. Oxidized 16S RNA, in 0.05 M sodium phosphate, pH 7.0, is incubated for 1 hr in the dark at room temperature with 2 m M FTS. Oxidized 23S RNA, in 0.05 M sodium phosphate, pH 7.0, is incubated in the dark for 2 hr at 0 ° with 2 m M F r S . The labeled 16S and 23S RNA samples are then treated as described for labeled 5S RNA and tRNA except that, just prior to ethanol precipitation, two volumes of 0.1 M KC1 is added. This dilution is necessary to prevent precipitation of sodium phosphate by the ethanol. After ethanol precipitation, 23S RNA is dissolved in l0 m M Tris-HC1, pH 7.4, 4 m M magnesium acetate, and 16S RNA in 30 m M Tris-HC1, pH 7.4; both are stored at - 80 °. In addition to the 3' end, tRNA can be specifically labeled at various other sites by reaction with, or removal of, modified bases that have unusual chemical reactivity. Procedures have been developed for replacing dihydrouracils of various tRNAs ta or the wybutine base adjacent to the anticodon of yeast t R N A ) TM for reacting the amino group of the X base in the extra loop of some E. coli tRNAs with isothiocyanates ~ and for reacting thiouridine and, in some cases, pseudouridine with fluorescent alkyl bromides ~6a7 or iodoacetamide derivatives. ~s The procedure for replacing dihydrouracil depends on the labilization of this base by NaBH4 reduction to ureidopropanol, ~9which is hydrolyzed from the tRNA under acidic conditions, as is the wybutine base. 13 The resulting aldehydic C-1 atom of the ribose can then react with amine or hydrazine derivatives. Yeast and E. coli tRNA ~e, which can be used with 13W. Wintermeyer, H.-G. Schleich, and H. G. Zachau, this series, Vol. 59, p. 110. 14O. W. Odom, B. B. Craig, and B. A. Hardesty,Biopolymers 17, 2909 (1978). 15j. A. Plumbridge, H. G. Btiumert, M. Ehrenberg, and R. Rigler, Nucleic Acids Res. 8, 827 (1980). 16C. H. Yang and D. $611,J. Biochem. (Tokyo) 73, 1243(1973). 17C. H. Yang and D. $611,Biochemistry 13, 3615 (1974). 18A. E. Johnson, H. J. Adkins, E. A. Matthews, and C. R. Cantor, J. Mol. Biol. 156, 113 (1982). 19p. Cerutti and N. Miller, J. Mol. Biol. 26, 55 (1967).

[ 11 ]


179

the convenient artificial messenger, polyuridylic acid, both contain two dihydrouracils, at positions 16 and 172o and 16 and 20, 2t respectively. The procedure used in this laboratorY for replacement of dihydrouracil is a modification of the published procedure. 13 tRNA (10-20 Az6o/ml) is reduced with 10mg/ml NaBH4 in 0 . 2 M Bicine [N,N-bis(2hydroxyethyl)glycine], pH 9.0, for 45 min at ambient temperature. During the reduction the pH is maintained at 9.0 by addition of small amounts of acetic acid. At the end of the incubation, excess NaBH4 is destroyed by adding acetic acid to a final pH of 5.0, followed by ethanol precipitation of the reduced tRNA. The labeling of reduced tRNA Pbcwith 7-diethylaminocoumarin-3-carbohydrazide (DCCH), as an example, is as follows. Deacylated tRNA Phc or AcPhe-tRNA in 0.1 M sodium acetate, pH 4.3, is incubated for 2 hr at 37 ° with 2 m M DCCH, added from a 50 m M stock solution in dimethylformamide. At the end of the incubation period, the solution is made 0.1 M in Tris-HCl, the pH is adjusted to 6.0-6.5 with KOH, and the sample is extracted twice with equal volumes of 70% phenol to remove excess DCCH, followed by three ethanol precipitations to remove residual phenol. The labeling with DCCH is usually incomplete, and unlabeled, singly labeled, and a small amount of the doubly labeled tRNA species are separated by reversed-phase chromatography, as described in the next section. Separation and Purification P r o c e d u r e s HPLC was carried out with a Beckman system which included a 421 controller unit for generating elution gradients and a 165 variable wavelength detector with which absorption was monitored at two wavelengths simultaneously. The reversed-phase column was #Bondapak C~s (3.9 m m × 30 cm) from Waters Associates, Milford, MA. R i b o s o m a l Proteins

The labeling of sulfhydryl groups of ribosomal proteins is frequently incomplete and it is desirable to remove unlabeled protein. Of the methods used in this laboratorY, the best involves HPLC in 0.1% trifluoroacetic acid (TFA) over a/tBondapak C18 reversed-phase column. This procedure also has been used to purify unlabeled ribosomal proteins) 2a3 The column is 20u. L. RajBhandaryand S. H. Chang, J. Biol. Chem. 243, 598 (1968). 21B. G. Bah'elland F. Sanger,FEBSLett. 3, 275 (1969). 22A. R. Kerlavage,C. J. Weitzmann,T. Hasan, and B. S. Cooperman,J. Chromatogr. 266, 225 (1983). ~3R. M. Kamp, A. Bosserhoff,D. Kamp, and B. Wittmann-Liebold,J. Chromatogr. 317, 181 (1984).

180

[ 11]

OTHER BIOPHYSICAL METHODS I

0.3

...............

._

1~02

,

~ . °.......

, "'"'"'""+"

u;ul ~ - z 0.1

t

""" ""

.°,*beled

1

O0

..... '......... "*........ '-430

I

2o

1'+

labeled

A

131111 {lg ¢~" O0 (Zl I111

0 o

8

16 24 ELUTION VOLUME, ml

32

40

T'

0

FIG. I. Separation o f E . coli ribosomal protein $21 labeled with fluorescein 5'-maleimide from unlabeled protein. Unlabeled $21 (25/Jg) was m i x e d with labeled S21 (25/Jg) a n d applied in 0.1% a q u e o u s T F A to the C~s c o l u m n described in the text. T h e c o l u m n h a d been equilibrated with the s a m e solution. Elution was with a nonlinear gradient o f aeetonitrile in a n a q u e o u s solution containing 0.1% TFA, as shown. Flow rate was 1 m l / m i n .

equilibrated with 0.1% TFA in water and the labeled protein is dialyzed against this same solution prior to loading. Elution is with a gradient of acetonitrile in an aqueous solution containing 0.1% TFA. Most ribosomal proteins elute between 25 and 50% acetonitrile.2z Labeled derivatives typically elute somewhat later than the unlabeled forms. Separation of unlabeled $21 from $21 labeled with fluorescein 5'-maleimide is shown in Fig. 1. The unlabeled species is eluted at about 25.5%, and the labeled species at about 29.5%, acetonitrile. Up to 0.5 mg of protein has been chromatographed under these conditions with little or no loss in resolution. The fractions from HPLC containing labeled protein are evaporated to dryness (Savant Speedvac concentrator), redissolved in 20 m M Tris-HC1, pH 7.5, 7 M urea, and finally dialyzed against 30 m M Tris-HC1, pH 7.4, 500 m M KC1, 20 m M magnesium acetate, 1 m M dithiothreitol (30S proteins) or 20 m M Tris-HC1, pH 7.4, 400 m M NH4CI, 4 m M magnesium acetate, 4 m M 2-mercaptoethanol (50S proteins). Ribosomal RNA

Labeling ratios approaching unity are usually obtained for 16S and 23S RNA, lessening the need for purification. However, 5S RNA often gives incomplete labeling. The procedure described below for tRNA probably would be applicable to 5S RNA but has not yet been tested with this material.

[ 11 ]


181

tRNA

To our knowledge, the use of Cls reversed-phase HPLC for separation of tRNAs has not been described previously, although it has been used for separating oligodeoxynucleotides,u It appears to represent a significant improvement over several other methods. Chromatography on benzoylated DEAE-cellulose (BD-cellulose) 25,26and RPC-5 chromatography13,~8-27 are frequently used. Although both of these methods involve hydrophobic interactions and the latter is referred to as reversed-phase chromatography, a large portion of the energy involved in binding of tRNA to such columns is due to ionic interactions of the negatively charged phosphate groups of tRNA with the positively charged nitrogen atoms of the BD-cellulose and RPC-5 material. The following procedure appears to separate tRNA species solely on the basis of hydrophobic interactions, presumably involving interaction with the bases of the RNA. Purines have more hydrophobic character than pyrimidines.28 Hartwick et aL 2s developed a procedure for separating ribonucleoside monophosphates on a Ci8 column, the order of elution being CMP, UMP, GMP, and AMP. This report prompted the investigation of the interaction of tRNA with C~8. A /zBondapak C~s column is equilibrated with 20 mM Tris-acetic acid (pH 5- 7 depending on the application), 10 mM magnesium acetate, and 0.4 M NaC1. tRNA samples are applied in the same solution. The column is developed with a methanol gradient between 0 and 60% in a solution containing the salt at the same concentrations. Figure 2 shows the elution profile of unlabeled deacylated yeast tRNA z*e and its acetylated aminoacylated form. Deacylated tRNA l~e elutes from the column at about 22% methanol and Ac[14C]Phe-tRNA at about 28% methanol. Both are eluted as relatively sharp peaks which are well separated. No radioactivity is detected in the first peak. Also shown in Fig. 2 by arrows 1 and 2 are the approximate positions at which E. coli deacylated tRNA ~ and AcPhe-tRNA elute. Unacetylated Phe-tRNA also can be separated from deacylated tRNA. It is eluted at a slightly lower methanol concentration than AcPhe-tRNA at approximately the position indicated by arrow 3. Although BD-cellulose has been used to separate E. coli Phe-tRNA from deacylated tRNAme,29 24 H.-J. Fritz, R. Belagaje, E. L. Brown, R. H. Fritz, R. A. Jones, R. G. Lees, and H. G. Khorana, Biochemistry 17, 1257 (1978). 25 I. C. Gillam, S. Milward, D. Blew, M. Von Tigerstrom, E. Wimmer, and G. M. Tener, Biochemistry 6, 3043 (1967). 26 W. Wintermeyer and H. G. Zaehau, FEBS Left. 18, 214 (1971). 27 R. L. Pearson, J. F. Weiss, and A. D. Kelmers, Biochim. Biophys. Acta 228, 770 (1971). 28 R. A. Hartwiek, S. P. Assenza, and P. R. Brown, J. Chromatogr. 186, 647 (1979). 29 K. L. Roy, A. Bloom, and D. $611, Proced. Nucleic Acid Res. 2, 524 (1971).

182

[ 11]

OTHER BIOPHYSICALMETHODS I

I

I

I

I

I

I

I

'

48,

ti

A

I

I

. . . . . . . . ..~ 4 4

0.8

I

2

i

t

3 ~

/:

0.6

24 d

12

o.,o 0

0

......J: . J ij 8

....... 16

24

32 40 ELUTION VOLUME.ml

48

56

64

FIG. 2. Separation of deacylatedyeast tRNA~ from Ac[14C]Phe-tRNA.Forty A2~ounits of Ac[14C]Phe-tRNA (specific activity 100 mCi/mmol) in 20 mM Tris-acetic acid, pH 5.0, 0.4 MNaCI, 10 mM magnesiumacetate were appliedto the C,8 column equilibratedwith the same solution. Elution was with a gradient of a solution containing 20 mM Tris-acetic acid, pH 5.0, 0.4 M NaCI, 10 mM magnesium acetate, 60% methanol. The gradient shown is expressed as final percentage of methanol. Absorbance of 290 nm was monitored continuously. Fractions of 4 ml were collectedand radioactivitywas determinedwith 10-#1aliquots by scintillationcounting in diphenyloxazole-toluenecounting fluid containing 5% Bio-Solve. Arrows 1 and 2 indicate the approximateelution positions of E. coil deacylatedtRNA~ and AcPhe-tRNA, respectively.Arrow 3 indicates the approximation elution position of yeast Phe-tRNA. separation by the procedure described here provides considerably cleaner resolution of the two species. Shown in Fig. 3 is the separation of deacylated DCCH-labeled tRNA ~ from DCCH-labeled AcPhe-tRNA. The tRNA was labeled with D C C H by the procedure indicated above, then chromatographed on the Cts column, followed by aminoacylation, acetylation, and finally rechromatography on the C~8 column. A small a m o u n t of unlabeled deacylated tRNA and unlabeled AcPhe-tRNA present in the preparation are eluted at about 22 and 28% methanol, respectively. The DCCH-labeled deacylated tRNA is eluted with 34% methanol whereas the DCCH-labeled AcPhe-tRNA is eluted with 37% methanol. BD-cellulose, by contrast, gave no resolution of this pair. The large peak of UV-absorbing material at the beginning of the gradient is ATP, which was coprecipitated with the tRNA by ethanol following the aminoacylation reaction. Figure 4A shows the separation of yeast tRNA Ph* species that are generated during the labeling procedure in which the wybutine base is replaced with proflavine. Yeast t R N A w was labeled with proflavine at

0

[11]

SEPARATION OF RIBOSOMAL PROTEINS AND t R N A s 0.36

0.32

;

i

i

,

!

•

I

'

I

'

I

I

183

I

!

48

-

40 0.28

-

.....................

...-"''"

~

"" ....

:-

32

A

0

0.20 o

24

)

t~w 0.16

>-

of,.) zz

I>

IZlm

(z: ~ 2

(2_

0

,6 ~

0,12

OO

I-

o,)¢n ~m <
500: 1) and a 2-hr reaction period to yield a high percentage of a complete cleavage product which is still active in binding. Activity is lost with the extension of reaction time beyond this point.

[ 17] I s o l a t i o n o f K i n e t i c I n t e r m e d i a t e s in in Vitro Assembly of the Escherichia coli Ribosome Using Cibacron Blue F3GA By DIPAK B. DATTA, LI-MING CHANGCHIEN, and GARY R. CRAVEN In vitro self-assembly of a functional subcellular structure from its dissociated components has been commonly used to elucidate the probable pathway by which the structure assembles in vivo. For a complex structure such as the Escherichia coli ribosome which contains 3 species of RNA and more than 50 different proteins, the study of such an assembly mechanism has been especially intriguing. To elucidate the details of the assembly process several different approaches have been taken. For example, single components have been omitted from the assembly mixture~; assembly with chemically modified proteins 2 or ribosomal RNAs 3 (rRNAs) has been attempted; accessibility of proteins in partially assembled particles to chemical modifying agents has been studied4; and thermodynamic intermediates in the assembly process have been isolatedJ ,5 In all these cases, the assembly intermediates are thermodynamically stable particles which do not provide any clear picture of the kinetic order of addition of the proteins to the growing particle. If we are to gain further insight into the mechanism of ribosome assembly, we must develop techniques which will permit the isolation of many more intermediate particles in the kinetic pathway of assembly. We have therefore sought chemical reagents that might block the addition of proteins to rRNA during in vitro ribosome assembly. We reasoned that, as

M. Nomura and W. A. Held, in "Ribosomes"(M. Nomura, A. Tissieres,and P. Lengyel, eds.), p. 193.Cold Spring Harbor Lab., Cold Spring Harbor, New York, 1974. 2L. Daya-Grojean,J. Reinbolt, O. Pongs, and R. A. Crarr~, FEBSLett. 44, 253 (1974). 3p. L. Schendeland G. R. Craven, NucleicAcids Res. 3, 3001 (1976). 4L.-M. Changchien and G. R. Craven, J. Mol. Biol. 113, 103 (1977). 5K. H. Nierhaus, in "Ribosomes:Structure, Function and Genetics" (G. Chambliss, G. R. Craven, J. Davies, K. Davis, L. Kahan, and M. Nomura, cds.), p. 267. University Park Press, Baltimore, Maryland, 1980. METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All fights ofteproducdon in any form reserved.

270

PROTEIN- RNA INTERACTIONS

[ 17]

fragments. Application of this cleavage technique to protein $4 requires the use of a high ratio of CNBr : Met (> 500: 1) and a 2-hr reaction period to yield a high percentage of a complete cleavage product which is still active in binding. Activity is lost with the extension of reaction time beyond this point.

[ 17] I s o l a t i o n o f K i n e t i c I n t e r m e d i a t e s in in Vitro Assembly of the Escherichia coli Ribosome Using Cibacron Blue F3GA By DIPAK B. DATTA, LI-MING CHANGCHIEN, and GARY R. CRAVEN In vitro self-assembly of a functional subcellular structure from its dissociated components has been commonly used to elucidate the probable pathway by which the structure assembles in vivo. For a complex structure such as the Escherichia coli ribosome which contains 3 species of RNA and more than 50 different proteins, the study of such an assembly mechanism has been especially intriguing. To elucidate the details of the assembly process several different approaches have been taken. For example, single components have been omitted from the assembly mixture~; assembly with chemically modified proteins 2 or ribosomal RNAs 3 (rRNAs) has been attempted; accessibility of proteins in partially assembled particles to chemical modifying agents has been studied4; and thermodynamic intermediates in the assembly process have been isolatedJ ,5 In all these cases, the assembly intermediates are thermodynamically stable particles which do not provide any clear picture of the kinetic order of addition of the proteins to the growing particle. If we are to gain further insight into the mechanism of ribosome assembly, we must develop techniques which will permit the isolation of many more intermediate particles in the kinetic pathway of assembly. We have therefore sought chemical reagents that might block the addition of proteins to rRNA during in vitro ribosome assembly. We reasoned that, as

M. Nomura and W. A. Held, in "Ribosomes"(M. Nomura, A. Tissieres,and P. Lengyel, eds.), p. 193.Cold Spring Harbor Lab., Cold Spring Harbor, New York, 1974. 2L. Daya-Grojean,J. Reinbolt, O. Pongs, and R. A. Crarr~, FEBSLett. 44, 253 (1974). 3p. L. Schendeland G. R. Craven, NucleicAcids Res. 3, 3001 (1976). 4L.-M. Changchien and G. R. Craven, J. Mol. Biol. 113, 103 (1977). 5K. H. Nierhaus, in "Ribosomes:Structure, Function and Genetics" (G. Chambliss, G. R. Craven, J. Davies, K. Davis, L. Kahan, and M. Nomura, cds.), p. 267. University Park Press, Baltimore, Maryland, 1980. METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All fights ofteproducdon in any form reserved.

[17]

TEMPORAL SEQUENCE OF RIBOSOMAL ASSEMBLY

271

S03H

~ L\

03H

'[

'1 0

.I HN

SO3H

N N

( CI

FIo. 1. A planar diagram of the molecular structure of Cibacron Blue F3GA. From left to right there are four parts in the molecule: the sulfonated anthraquinone, the sulfonated benzene, the triazine ring containing the active chlorine, and the terminal sulfonated benzene.

about half of the ribosomal proteins can directly bind the appropriate rRNA, there might be a set of rules followed by all or most of them for associating with the developing protein-RNA complex during assembly. We therefore looked for a generalized inhibitor of protein-RNA interactions. This search led us to investigate the triazine dye Cibacron Blue F3GA as a potential candidate for such an inhibitor. Cibacron Blue F3GA is a complex sulfonated aromatic dye (see Fig. l) that has been found to bind to a large number of nucleotide-binding enzymes.6 This observation has led to the development of the method of dye-ligand affinity chromatography for the purification of enzymes of the kinase, oxidoreductase, synthetase, and transferase classes. These enzymes will bind to a column of agarose or Sepharose that has been chemically bonded with Cibacron Blue; most of them can be eluted with a very low concentration of their respective nucleotide cofactors or with salts at high ionic strength. There is substantial evidence that the dye specifically binds to a domain on the protein called the "dinucleotide fold. ''7 The high flexibility of the structure of the dye molecule probably helps it mimic the structures of different nucleotides and reinforces its binding to the dinucleotide fold of the proteins. Using a column of Cibacron Blue-bound agarose, we have shown that all E. coli ribosomal proteins, when free in solution, bind to the column even in the presence of the high salt concentrations in the reconstitution buffers used in these experiments. We have also shown that when the proteins are part of ribosomes, they do not bind to the column under 6 p. Dean and D. Watson, J. Chromatogr. 165, 301 (1979). E. Stellwagon, R. Cass, S. Thompson, and M. Woody, Nature (London) 257, 716 (1975).

272

PROTEIN-RNA INTERACTIONS

[ 17]

comparable conditions--and that ribosomes are mostly excluded from such a column) In light of these observations, we correctly reasoned that the free dye might inhibit in vitro ribosome assembly if it were prebound to the proteins. Using the following methodology, we discovered some kinetic intermediates in the 30S and 50S ribosome assembly. Preparation of Materials Solutions RB-30 (reconstitution buffer for 30S ribosomal subunits): 30 m M Tris-HCl (pH7.6), 0.33M KC1, 20raM magnesium acetate, 0.5 m M dithiothreitol (DTT) RB-30K: 30 m M Tris-HCl (pH 7.6), 0.5 M KC1, 20 m M magnesium acetate, 0.5 m M DTT RB-30N: 30 m M Tris-HC1 (pH 7.6), 20 m M magnesium acetate, 0.5 m M DTT RB-50 (reconstitution buffer for 50S ribosomal subunits): 20 m M Tris-HC1 (pH 7.4), 0.4 M NH4C1, 4 m M magnesium acetate, 0.5 m M DTT TM: 10 m M Tris-HC1 (pH 7.5), 10 m M magnesium acetate TM-4:l0 m M Tris-HC1 (pH 7.5), 4 m M magnesium acetate Cibacron Blue Stock Solution. A 13.33 m M solution of Cibacron Blue F3GA (Pierce & Co.) is incubated with a 10-fold molar excess of lysine in 20 m M Tris-HC1 (pH 7.5) at 44 ° for l hr for covalent modification of the dye at the site of its C1- component. Unmodified dye may covalently react with lysine groups at its binding site in the proteins. Alternatively Cibacron Blue-C1 can be hydrolyzed to Cibacron Blue-OH by NaOH treatment at 60 ° as described by Moe and Piszkiewicz. 9 Free Cibacron Blue (-C1 or - O H ) , although soluble in water, is not dialyzable in purely aqueous solutions. To facilitate future studies on dye-protein-binding properties, we tested solutions of various organic solvents in water for their effects on dialyzability of the dye. A solution of 250/0 (v/v) dioxane in water was found to give a reasonably good rate of dialysis of the dye. Therefore the aqueous solution of the dye to be used as a stock solution is made 25% in dioxane. The stock solution contains 10 m M Cibacron Blue, 250/0(v/v)dioxane, 20 m M Tris-HC1 (pH 7.5)(and 100 m M lysine if applicable). The final concentration of dioxane in the a D. Datta, L.-M. Changchien, and G. R. Craven, Nucleic Acids Res. 14, 4095 (1986). 9 j. G. Moe and D. Hszkiewicz, Biochemistry 18, 2810 (1979).

[17]


273

reaction mixtures has no effect on ribosome rcconstitution? ° Freshly prepared dye solution is always used. Ribosomal Subunits. The 70S ribosomes and their subunits are prepared following the procedures of Craven and Gupta, ~ and stored in TM at --70 ° . 30S Proteins. Total protein (TP-30) is extracted from the 30S ribosomes with acetic acid, ~2 then dialyzed sequentially against solutions of 30 m M Tris (unneutralized), 3 M urea, 0.5 m M DTT and 20 m M TrisHC1 (pH 8), 6 M urea, 0.5 m M DTT. It is stored in the latter solution at - 7 0 °" 50S Proteins. Total protein (TP-50) is extracted from the 50S ribosomes with acetic acid as in the TP-30 preparation; the rest of the procedure of TP-50 preparation follows the method ofNowotny et al.13 Proteins are stored in RB-50 at - 7 0 ° in aliquots. Ribosomal RNAs. 16S RNA is prepared by phenol extraction of 30S ribosomes. ~4 For preparation of intact 23S RNA, it is necessary to deproteinize 70S ribosomes with phenol and to isolate the 23S RNA by two successive preparative bentonite-treated sucrose gradient centrifugations. Pure 23S RNA from the appropriate peak fractions of the second centrifugation is precipitated by ethanol and dissolved in TM-4. 5S RNA is prepared by the method of Erdmann and Doberer? 5 All RNA solutions are stored in TM-4 at - 7 0 °. Procedures

Method of Reconstitution Prior to reconstitution experiments, the TP-30 solution is dialyzed against RB-30K buffer. The high-salt concentration of this buffer prevents precipitation of proteins when urea is removed. The dialyzed TP-30 is then diluted with RB-30N buffer to bring the KC1 concentration down to the normal level of 0.33 M. The RNAs and proteins are taken separately in 1.5 ml each of the appropriate reconstitution buffer. The RNA solution is heated to 42 ° for 10 rain and cooled to 0 ° before use. For the 30S reconstitution system, 80D26o units of 16S RNA are taken against 2.40D23o units ~0M. A. Cantrell, Ph.D. thesis. University of Wisconsin, Madison, 1977. i1 G. R. Craven and V. Gupta, Proc. Natl. Acad. Sci. U.S.A. 63, 1329 (1970). ~2S. Y. S. Hardy, C. G. Kurland, P. Voynow, and G. Mora, Biochemistry 8, 2897 (1969). ~3V. Nowotny, H.-J. Rheinberger, K. Nierhans, B. Tesche, and R. Amils, Nucleic Acids Res. 8, 989 (1980). 14 p. Traub, S. Mizushima, C. V. Lowry, and M. Nomura, this series, Vol. 20, p. 391. 15 V. A. Erdmann and H. G. Doberer, Mol. Gen. Genet. 114, 89 (1971).

274


[ 17]

of TP-30. For the 50S reconstitution system, 8 OD26o units of a 1 : 1 mixture of 23S and 5S RNAs are taken against 2 OI)23o units of TP-50. The RNA: protein ratio in both systems is 1 : 2. To slow down the rate of self-assembly all experiments are performed at 0 °. The mixtures are incubated for 1 hr and then cleared by centrifugation at 10,000 g for 20 min in a Beckman JA-20 rotor. The RNA-protein complexes are pelleted by ultracentrifugation through a 15-ml cushion of 12% sucrose in appropriate reconstitution buffer at 100,000 g for 17 hr in a Beckman Type 50.2 Ti rotor.

Electrophoretic Analysis RNA-protein pellets are taken up in 100/zl of a solution of 10 m M Tris-HC1 (pH 7.5), 1 m M etl~ylenediaminetetraacetic acid, 8 M Urea, 0.5 m M DTT, treated with RNase, and analyzed for proteins by polyacrylamide gel electrophoresis (PAGE). One-dimensional PAGE is performed following the method of Voynow and Kurland.~6 Two-dimensional PAGE is performed as previously reported ~7 except that a small-sized gel (0.2 i.d. × 12.7 cm) is used in the first dimension and the thickness of the second-dimensional gel is 0.15 cm. Is

Inhibitory Molar Ratio of Dye: Ribosomal Proteins A series of experiments was done with increasing ratios of dye: proteins. Cibacron Blue was added to a solution of proteins in reconstitution buffer in appropriate aliquots, immediately followed by addition of the appropriate RNAs. Reconstitution was allowed to proceed for 1 hr. Unassembled protein:dye complexes were found to tend to aggregate; the low-speed centrifugation (described above) was essential to remove these. Any remaining such complexes, especially at high concentrations of Cibacron Blue, formed a blue ring around the RNA-protein pellet following ultracentrifugation. This ring was carefully removed with a swab stick. The RNA-protein pellets were analyzed for proteins by one-dimensional PAGE. Some typical results are shown in Fig. 2. It clearly shows that at the moderate dye:protein molar ratios of 40:1 to 100:1, all proteins are prevented from incorporation into a ribonucleoprotein complex. The corresponding dye:RNA ratios are 80:1 to 200:1. The proteins with the highest affinity for their respective RNAs compared to their affinity for the dye are $8, L1, and L24. 16p. Voynow and C. G. Kurland, Biochemistry 10, 517 (1971). 17L.-M. Changchien and G. R. Craven, J. Mol. Biol. 125, 43 (1978). ~8A. Lin, E. Collatz, and I. G. Wool, Mol. Gen. Genet. 144, 1 (1976).

[17]

275


SS, S9, SI2, SI4,: SI6,

sfe~

TP-50

0

II

21

52

57

42

47

55

105

MOLES DYE / MOLES TP-50 PROTEIN FIG. 2. One-dimensional PAGE analysis of 30S reconstitution intermediates formed in the presence of Cibacron Blue. T P - 30 proteins were treated with the dye at the dye: protein

molar ratio shown, immediately followed by the addition of 16S RNA, all at 0". The RNA-protein complexeswere isolated and analyzed as describedin the text. From Datta et al.s

Temporal Sequence of Ribosome Assembly From the above experiments, the minimum dye:protein molar ratio required to inhibit most proteins from assembly into ribosomes is found to be 50: 1. Subsequently the dye is used at this ratio to inhibit assembly at desired time points. A series of reconstitution mixtures is made and the dye added to the mixtures at various times after R N A - p r o t e i n mixing. The R N A - p r o t e i n complexes are isolated as described above. (The R N A protein pellets have a faint blue coloration especially at longer times of incubation.) Typical two-dimensional analyses for the 50S reconstitution system are shown in Fig. 3. The gels clearly show a time-dependent appearance of proteins in the intermediate particles. From a large number of such analyses, a kinetic sequence of ribosome assembly can be constructed. Our construction of such assembly sequences for E. coli 30S and 50S ribosome reconstitutions is shown in Fig. 4.

Possible Protein-Stripping Effect on Preformed Particles A possible pitfall in the use of Cibacron Blue as the inhibitor of ribosome assembly is that at relatively high concentrations it may start strip-

276

[ 17]


8M Urea, pH 4.5

)

SDS pH7.1

FIo. 3. Two-dimensional PAGE analysis of 50S reconstitution intermediates isolated following addition of Cibacron Blue at different times during assembly. For the 0 time experiment, the dye was added to the proteins immediately prior to the addition of the RNA. A dye: protein molar ratio of 50 : l was used. From Datta et al.S

[17]

TEMPORAL SEQUENCE OF RIBOSOMALASSEMBLY 30S

ASSEMBLY,(7'

[ S8 $4 S5 S6 SIS ]

50S ASSEMBLY,O* STAGE I

0

0 0 "10

0 "0 0

t x 10 7

1 xlO 9

C

0

b--

8

I1

5 O0

I 100q

I 1500

I 2 000

Tungsten wire temperature (K) Fro. 4. Dependenceof the degree of labeling of ribosomal protein in ribosomes on tungsten wire temperature.(0) Total radioactivityin the sample;(O) radioactivityincorporated into ribosomalproteins. the ribosomes in sucrose gradients (Fig. 5). It is found that ribosomes labeled at a wire temperature of 1000 K retain their original integrity: the radioactivity profile corresponds to the UV adsorption profile in the ribosome region (Fig. 5a) or to that of the ribosomal subunits (Fig. 5b); only the exchangeable tritium, unattached to the proteins, is present in the upper part of the sucrose gradient. Ribosomes labeled at 2000 K aggregate, probably as a consequence of destruction, and radioactivity is observed at the bottom of the tube (Fig. 5b). Identification of Proteins Exposed on the Ribosome Surface

Solutions and Buffers 5% acetic acid Buffer A: 10 m M MgC12, 10 m M NH4C1, 0.1 m M Na2EDTA, 1 m M DTT, 20 m M Tris-HC1, pH 7.4 Buffer B: buffer A without MgC12 containing 1 m M Na2EDTA

Tritium Bombardment of Total Ribosomal Proteins The fluorogram of two-dimensional electrophoresis of the total ribosomal protein labeled under denaturing conditions in 5% acetic acid is

[28]

Az6o 1.0

0.5

435

HOT TRITIUM BOMBARDMENT OF RIBOSOMES

a

cpm

b

_L 5

10

15

C 2x 10 5

1~I0 5

2O

5

~0

Fraction

15

20

5

10

15

20

number

Fie. 5. Sedimentation analysis of ribosomes labeled with hot tritium atoms. Sucrose gradient 4-20%, SW-41 rotor, centrifuged for 12 hr at 20,000 rpm at 4 ° for (b) and (c); 14 hr at 16,000 at 4 ° for (a). Ribosomes labeled at a tungsten wire temperature of 1000 K: (a) sucrose gradient in buffer A (10 m M MgClz), (b) sucrose gradient in buffer B (1 m M MgC12). Ribosomes labeled at a tungsten wire temperature of 2000 K: (c) sucrose gradient in buffer B (l raM"MgCl2). (0) 260 nm absorbance, (O) total radioactivity, (A) radioactivity incorporated into ribosomal proteins.

shown in Fig. 6a. It can be seen that all of the ribosomal proteins are labeled. The majority of proteins are labeled roughly proportionally to their molecular masses. However, there are several exceptions, probably explained either by lower amounts of these proteins on the slab or by their incomplete denaturation. There are no spots on the fluorogram corresponding to degradation products of labeled ribosomal proteins. All of the radioactive spots strictly correspond in mobility to the spots of proteins on the stained slab (cf. Fig. 6a with Fig. 3).

b

!

b FIG. 6. Two-dimensional ele~:trophoresis of ribosomal proteins. Fluorograms of [3H]TP70: (a) labeled in isolated state in 5% acetic acid (150 gg of protein, 3.6 X 106 dpm); (b) labeled in unfolded ribosomal subunits (150 gg of protein, 5 X l06 dpm); (c) labeled in 70S ribosomes (150gg of protein, 0.5 X 106 dpm). (a, b) Reproduced from Ref. 6.

436

CHEMICAL AND ENZYMATIC PROBING METHODS

[9-8]

Tritium Bombardment of Unfolded Ribosomal Subunits Unfolded ribosomal subunits are prepared from nondissociated ribosomes in buffer B. Sedimentation analysis has shown that the particles prepared in this way sediment with sedimentation coefficients of 17S and 25S, indicating a partial unfolding of the ribosomal subunits. ~3,~4All of the proteins are found to be labeled in this sample of unfolded ribosomal subunits (Fig. 6b). However, the distribution of the label between the ribosomal proteins in this case is not proportional to their molecular masses. Differential labeling of proteins in unfolded ribosomal subunits seems to be the result of their different degrees of exposure, which may be the result of protein-protein and RNA-protein interactions within the nucleoprotein.

Tritium Bombardment of Nondissociated Ribosomes The fluorograms of two-dimensional electrophoresis of ribosomal proteins labeled by tritium bombardment of 70S ribosomes are shown in Fig. 6c and 8a. It can be seen that a significant number of the proteins in this case are labeled to a still lesser degree or virtually unlabeled. The decrease in accessibility of ribosomal proteins to tritium atoms in compact 70S ribosomes, in comparison with that of unfolded ribosomal subunits, is likely the result of an increase of protein-protein and RNA-protein interactions and of shielding by folding of the particle. The more exposed proteins in the 70S ribosomes are S1, $4, $7, $9 and/or S11, S12 and/or L20, S13, S18, $20, $21, LI, L5, L6, L7/L12, L9, L10, L11, L16, L17, L24, L26, and L27. The degree of labeling for proteins $7, S13, $9 and/or S11, $20 (L26), $21, L1, L2, L5, L6, L7/L12, L9, L17, L25, and L27 varies somewhat in different ribosome samples (cf. Fig. 6c and Fig. 8a). From the results presented it can be concluded that the most exposed proteins on the surface of the 70S ribosomes comprise about half of the 30S subunit proteins and about one-third of the 50S subunit proteins. The remaining ribosomal proteins are more buried in the ribosome structure. D e p e n d e n c e of Protein Exposure on the Structural State of

Ribosomes

Buffers A: 100 m M NH4C1, 0.1 m M Na2EDTA, 1 m M DTT, 20 m M TrisHC1, pH 7.4, where the concentration of MgCI2 varies from 0 to 10mM ~3L. P. Gavrilova, D. A. Ivanov, and A. S. Spirin, J. Mol. Biol. 16, 473 (1966). ~4R. F. Gesteland, J. Mol. Biol. 18, 356 (1966).

[28]


437

B: 20 m M MgC12, 100 m M KC1, 1 m M DTT, 20 m M Tris-HC1, pH 7.4

Tritium Bombardment of Ribosomal Particles in Buffers with Different Concentrations of Magnesium Ions The dependence of ribosome dissociation on magnesium ion concentration in buffer A as measured by light scattering is shown in Fig. 7. The state of ribosomes in the sample was controlled also by sedimentation in an analytical ultracentrifuge. It has been shown that in the range of magnesium ion concentrations from 10 to 4 mM in the buffer, the samples contain only 70S ribosomes. At concentrations of magnesium ions of 1 mM and below, the samples consist only of ribosomal 30S and 50S subunits. Decreasing the magnesium ion concentration in the buffer to 0.3 m M leads to a decrease of the sedimentation coefficient of 30S subunits to 20S, while that of the 50S subunits does not change. Comparison of surface protein labeling in nondissociated ribosomes (in buffer A with I0 mM MgClz) and in a mixture of ribosomal subunits (in buffer A with 1 m M MgC12) is the main purpose of these experiments. Analysis of fluorograms of several independent experiments has shown that dissociation of ribosomes into subunits is accompanied neither by exposure of additional ribosomal proteins to hot tritium bombardment, nor by a reproducible increase in labeling of the accessible proteins (Fig. 8a,b). It can be concluded that, upon dissociation of 70S ribosomes into subunits, no additional ribosomal proteins become exposed and, hence, there are no proteins on the contacting surfaces of the ribosomal subunits.

o~ c

=

1.5

o~-

1.0

~

0.5

_.J

I

1

I

I

3

I

I

I

5

I

7

I

I

9

Mg 2., mM FIG. 7. D e p e n d e n c e o f ribosome dissociation o n M g 2+ concentration. Measured by light scattering at 400 n m . R i b o s o m e concentration 1 m g / m l . Reproduced from Ref. 6.

438


[28]

Fxo. 8. Two-dimensional clcctrophorcsis of ribosomal proteins. Fluorograms of [3H]TP70 (150/Lg of protein,0.5 to 1.5 × I06 dpm) labeledin ribosomes in bufferscontaining (a) 10 m M MgCI~, (b) I m M MgCI2, (c) 0.5 m M MgCl2, (d) 0.3 m M MgCl 2. (a, b) Reproduced from Ref. 6.

It should be noted, however, that in some samples of ribosomes dissociated into subunits an additional radioactive spot Y is observed (see Fig. 8b,c,d) which has not been identified as a ribosomal protein; this unknown protein may be a novel component located between the ribosomal subunits. Further decrease of magnesium ion concentration to 0.5 m M leads to exposure of proteins $3, $5, $7, and S16 on the surface of the 30S subunits (Fig. 8c). This effect seems to be the consequence of shght changes of subunit structure (though the sedimentation coefficient of the particles remains the same). Decrease of magnesium ion concentration to 0.3 mM results in additional exposure of ribosomal proteins $3, $4, $5, $7, $9 and/or S 11, S 14, and S 18 (Fig. 8d); in these conditions the sedimentation coefficient of 30S subunits decreases to 20S. It is likely that in the latter case the increased exposure of the proteins is caused by partialunfolding of the 30S subunits.

[28]


439

q

SO

o

FIG. 9. Two-dimensionalelectrophoresisof ribosomal proteins. Fluorograms:(a) [3H] TP30, (b) [;H]TPS0, (c) [~H]TP70labeled in the isolated 30S subunits, 50S subunits, and their 70S couples,respectively.

Tritium Bombardment of Isolated Ribosomal Subunits and Their Couples Analysis of fluorograms of two-dimensional electrophoresis of ribosomal proteins labeled in isolated 30S and 50S ribosomal subunits in buffer B shows that the most exposed proteins are S1, $3, $4, $5, $6, $7, $9 and/or SI 1, S13, S14, S18, $20, $21, L1, L5, L7/L12, L9, L10, L11, LI6, L24, L25, and L26 (Fig. 9a,b). The considerable difference in the exposure of several proteins in nonisolated (Fig. 8a) and isolated (Fig. 9a,b) subunits can be explained by some alteration ofsubunit structure during isolation of subunits by zonal centrifugation under conditions of high ionic strength (0.5 M NH4C1) with low Mg2+. 9 A reverse shielding of proteins $3, $5, $7, S14, and S18 occurs upon interaction of isolated 30S and 50S subunits to form 70S ribosomes in buffer B (Fig. 9c). These same proteins become exposed in nonisolated 30S subunits in response to a decrease of magnesium ion concentration in buffer A (see the previous section). Thus, it seems likely that protein shielding upon association of isolated subunits is directly coupled with restoration of the altered 30S subunit structure.

440


[29] Surface

Topography

of Ribosomal

[29]

RNA

B y ALEXEY A. BOGDANOV, NINA V. CHICHKOVA, ALEXEY M. KOPYLOV, ALEXANDER S. MANKIN, and EVGENY A. SKRIPKIN

Ribosomal ribonucleic acids (rRNA) form the very compact central core o f ribosomal subunits, whereas both ribosomal proteins and rRNA segments have been found on their surface (see Ref. 1 for a review). Since there is scarcely any doubt now that rRNA is directly involved in ribosome functioning, investigation o f these exposed r R N A regions is o f great interest. Hence they are likely to be involved in the interactions o f ribosome functional centers with tRNA, m R N A , and i n i f i a t o n factors, as well as in subunit association. In this chapter, we will concentrate on the methods which have been developed in our laboratory for identification o f R N A regions located on the surface o f r R N A both in the isolated state and in ribosomal subunits. I. Site-Specific H y d r o l y s i s o f r R N A - O l i g o d e o x y r i b o n u c l e o t i d e C o m p l e x e s with R N a s e H A. Principle

The idea o f site-specific cleavage o f R N A with RNase H was first formulated by Smirnov. 2 The corresponding experimental procedure for MS2 phage R N A was developed in our laboratory in 1978 as the joint work o f three groups. 3,4 Donis-Keller has described a similar approach for fragmentation o f 5.8S rRNA. 5 Subsequently we have shown that this m e t h o d is very useful for s t r u c t u r e - f u n c t i o n studies o f ribosomes. 6-8 A. A. Bogdanov, A. M. Kopylov, a n d I. N. Shatsky, in "Subcellular Biochemistry" (D.

Roodyn, ed.), Vol. 7, p. 81. Plenum, New York, 1980. 2 V. D. Smimov, personal communication (1976). 3 V. G. Metelev, O. B. Stepanova, N. V. Chiehkova, N. P. Rodionova, V. D. Smirnov, S. L. Bogdanova, N. F. Sergeeva,K. I. Ratmanova, A. A. Bogdanov,Z. A. Shabarova, and J. G. Atabekov, Biol. Nauki (Moscow) 8, 27 (1978). "O. B. Stepanova, V. G. Metelev, N. V. Chiehkova, V. D. Smimov, N. P. Rodionova, J. G. Atabekov, A. A. Bogdanov,and Z. A. Shabarova, FEBS Lett. 103, 197 (1979). 5H. Donis-Keller,Nucleic Acids Res. 7, 179 (1979). 6 A. S. Mankin, E. A. Skripkin, N. V. Chiehkova, A. M. Kopylov, and A. A. Bogdanov, FEBS Lett. 131, 253 (1981). 7 E. A. Skripkin, V. K. Kagramanova, N. V. Chiehkova, A. M. Kopylov, and A. A. Bogdanov, Biokhimiya (Moscow)46, 2250 (198I). 8 G. Z. Gaida, A. Y. Spunde, E. A. Skripkin, V. K. Kagramanova, V. P. Veiko, N. V. Chiehkova, and A. A. Bogdanov,Bioorg. Khim., 8, 1952 (1982). METHODS IN ENZYMOLOGY, VOL. 164

English translation copyright © 1988 by Academic Press, Inc.

[29]

SURFACE TOPOGRAPHY OF rRNA

441

The approach is to select oligodeoxyribonucleotides 6 - 1 0 nucleotides in length that are complementary to given RNA regions, form the corresponding heteroduplexes, and treat them with RNase H, which is the enzyme which specifically cleaves the RNA chain in R N A - D N A hybrids. In many cases, estimation of the chain length of fragments formed by polyacrylamide gel electrophoresis under denaturating conditions is enough to prove that the RNA chain is hydrolyzed at the selected site. For precise localization of cleavage sites, however, direct sequencing of terminal segments of RNA fragments must be performed. It is important that RNase H cleaves R N A - D N A hybrids, leaving Y-terminal hydroxyls and 5'-terminal phosphates which can be labeled in the usual way.

B. Ribonuclease H from Escherichia coli Escherichia coli RNase H is a basic protein with a molecular weight of 17,559. 9

1. Assay. [14C]poly(rA)" poly(dT) in which the specific radioactivity of [~4C]poly(rA) is 105 cpm per microgram is used as a substrate. The standard assay mixture (200 gl) for measuring RNase H activity contains 40 m M Tris-HC1 (pH 7.5), 4 m M MgC12, 1 m M dithiothreitol, 30/tg of nuclease-free BSA per milliliter, 5% glycerol, 20 #g of substrate per milliliter, and aliquots (1-25/tl) of RNase H. After incubation for 20 min at 37 ° the mixture is cooled in ice. The samples (5-10 gl) are mixed with 20 gl of cold BSA solution (4 mg/ml) and 100 gl of 7% trichloroacetic acid (TCA). The precipitates are removed by centrifugation and the radioactivity of the supernatants is measured. One unit of RNase H activity is the amount of enzyme producing 1 nmol of acid-soluble nucleotides in 20 min at 37 ° under these assay conditions. 2. Enzyme Preparation. E. coli RNase H is now available from a number of commercial sources. An E. coli strain which carries cloned RNase H gene and overproduces this enzyme has also been d e s c r i b e d . 9 However, in this study, we have used the enzyme isolated from E. coli MRE 600 according to the method of Darlix. ~° In contrast to the original procedure, in our hands, the major part of the RNase H activity is retained on the DEAE column and the enzyme is eluted from DE52 (Whatman) with a linear gradient of NaCI (0-0.4 M ) in a buffer containing 20 m M Tris-HC1 (pH 7.9), 0.1 m M dithiothreitol, 0.1 m M ethylenediaminetetraacetic acid (EDTA), 10% glycerol. RNase H is stored at - 2 0 * in the same buffer containing 0.1 M NaC1, 0.05M KCI and mixed with an equal vol9 S. Kanaya and R. J. Crouch, J. Biol. Chem. 258, 1276 (1983). 1oj. L. Darlix, Eur. J. Biochem. 51, 369 (1975).

442


[29]

ume of glycerol. The enzyme in this storage buffer is stable for more than a year. Usually, the specific activity of the enzyme is about 2500 units/ml. The enzyme is sufficiently free of DNase and other RNase activities. The RNase H activity falls 5-fold when the NaCI concentration is increased from 0.05 to 0.5 M. Mg 2+ ions are necessary for activity; maxim u m activity is obtained in the presence of 10 m M MgC12; SH reagent must also be present. H

C. Site-Specific Cleavage of rRNA 1. Oligodeoxyribonucleotides. Most of the oligodeoxyribonucleotides used in this study are synthesized by a solid-phase phosphotriester procedure ~2 and carefully purified by ion-exchange and Cls reversed-phase HPLC. They are sequenced according to the method of Maxam and Gilbert. ~3 2. Ribosomal Subunits and rRNA. 70S ribosomes, ribosomal subunits, and rRNA are isolated from E. coli MRE 600 according to conventional techniques. ~4 In addition to SDS-phenol extraction, the 16S rRNA is isolated by treatment of 30S subunits with acetic acid and urea in the presence of high Mg 2+ concentrations. ~5In all cases the rRNA solutions are treated with bentonite. 3. Heteroduplex Formation and RNase H Treatment. The following conditions give a good yield of RNA fragments for most oligodeoxyribonucleotides complementary to accessible rRNA regions. 6/tl H buffer (10 m M Tris-HC1, pH 7.6, 10 m M MgC12, 0.2 MKCI, 0.1 m M dithiothreitol) 2/zl rRNA (2- 5 mg/ml in H buffer) 1 pl oligodeoxyribonucleotide ( 1 - 2 A26o/ml in buffer H) 1 pl RNase H ( 2 - 3 units) Incubation for 1 hr at 4 °. In the case of ribosomal subunits, the RNase H and oligonucleotide concentrations should be increased at least 10-fold. The temperature of the reaction could be raised up to 20* to facilitate the hydrolysis. To stop the reaction, 10/zl of phenol saturated with buffer H is added. t l V. G. Metelev, O. B. Stepanova, N. V. Chichkova, V. D. Smirnov, N. P. Rodionova, I. M. Berzin, N. V. Jansone, E. J. Gren, A. A. Bogdanov, Z. A. Shabarova, and J. G. Atabekov, Mol. Biol. (Engl. Transl.) 14, 200 (1980). 12 A. Rosenthal, D. Ccch, V. P. Veiko, T. S. Orezkaja, E. A. Kuprijanova, and Z. A. Shaharova, Tetrahedron Lett. 24, 1691 (1983). 13 A. M. Maxam and W. Gilbert, this series, Vol. 65, p. 499. 14 R. A. Zimmerman, this series, Vol. 59, p. 551. 15 H.-K. Hochkeppel, E. Spicer, and G. R. Craven, J. Mol. Biol. 101, 155 (1976).

[29]

SURFACE TOPOGRAPHYOF rRNA

443

After phenol extraction, RNA fragments are precipitated with 96% ethanol at - 7 0 °, centrifuged, washed with 80% ethanol, and dissolved in 1/A of water. In some particular cases, due to the lower ability of a given RNA region to form oligodeoxyribonucleotide-RNA hybrids, the optimal conditions for RNase H hydrolysis can be selected by changing the RNA/oligonucleotide ratio, temperature, and reaction time. Alterations in reaction conditions (decreasing oligonucleotide content and increasing temperature, for example are also important if one wishes to get a single break in the RNA chain. This can help to avoid RNA cleavages at sites with a partial oligonucleotide- RNA complementarity. 4. Separation and Isolation of rRNA Fragments. Electrophoresis in a denaturating polyacrylamide gel is used for characterization of products of the rRNA cleavage reaction. One microliter of an rRNA fragment solution in water is mixed with 5/ll of loading buffer (10 m M Tris-borate, pH 8.3, 7 M urea, 0.2 m M EDTA, 0.03% bromphenol blue, 0.03% xylene cyanole). In the case of large fragments, samples are loaded onto 4% polyacrylamide- 7 M urea slab gels and electrophoresed as described by Maxam and Gilbert. 13 For separation of smaller (usually 32P-labeled) fragments, 14- 20% polyacrylamide- 7 M urea gels are used) 6 rRNA fragments have been localized on gels by ultraviolet shadowing or autoradiography and then eluted with 0.3 ml of 0.5 M ammonium acetate, 0.1% SDS, 0.1 m M EDTA, and 50/tl phenol by a crush and soak procedure. 13 For isolation of large amounts of rRNA fragments, ultracentrifugation in appropriate sucrose gradients is used. 5. Identification of Cleavage Sites Procedure A: Sites located near rRNA termini. If the selected regions of rRNA are in close proximity to either 5' or 3' ends one can expect the appearance of rather short oligonucleotides as a result of RNase H treatment. In this case it is reasonable to perform preliminary labeling of the corresponding terminal group of the RNA molecule. However, in the case of ribosomal subunits it is more convenient to label the 3' ends of the released oligonucleotides with [ 5 ' - 3 2 p ] p C p and T4 RNA ligase. The following procedure gives an example of localization of RNase cleavage sites near the 5' end ofE. coli 16S RNA both in the isolated state and in 30S subunits, s The octamer dCAAACTCT complementary to the segment 8 - 15 of 16S RNA is used. 16S RNA ( 15/lg in 15/~1 of 30 m M Tris-HCl, pH 8.0) is incubated with 0.03/~g of alkaline phosphatase (Sigma) for 20 min at 37 o and then, after ,6 D. A. Peattie and W. Gilbert, Proc. Natl. Acad. Sci. U.S.A. 77, 4679 (1980).

444

CHEMICAL A N D ENZYMATIC PROBING METHODS

TABLE I SITE-SPECIFIC CLEAVAGE OF E. coil 16S RNA

Sample

16SRNAC

Complementary complexes and cleavage sites°

5, lO l i pA~OAAOAOU~UOAU

Activity of reconstituted 30S subunits (%)b

N~

TCTCAAAC 5,

5, 30S subunits

l! ° l

pAAAUUGAAGAGUUUGAU

ND

TCTCAAAC 5,

16SRNA°

5' 3~ l C~C~OAOA~A~A

6O

A G T T T C C T T 5, 520 16S RNAc

7 AGCAGCCmGCGGUAAUAC

ND

UCGG C| GCCATTAs, 5' 770

h q

16S RNAc

780 v

AGCGU~AGCAAACA (rA)CCCCTCGTT 5, 5' 770

30S subunits

40 50e

780

AGCGU~AGC~CAf

ND

(rA)CCCCTCGTTT 5,

,6SRNAC

5' l !o5o CO~CA~ACA~O

60

TCTGTCC 5,

16S RNAg

GAACUCAAAGGAGAC AGTTTCCTT 5,

h

[29]

[29]

SURFACE TOPOGRAPHY OF r R N A

445

addition o f E D T A up to 6 raM, heated for 3 rain at 90 °. R N A is phosphorylated at the 5' end with [7-32p]ATP ( > 1000 C i / m m o l , A m e r s h a n ~ a n d T4 p o l y n u d e o t i d e kinase.~7 The radiolabeled R N A is deproteinized twice, precipitated with ethanol, dissolved in buffer H and treated with the oct a m e r a n d R N a s e H as described above (see Section I,C,3). After 1 hr incubation at 20 °, an equal v o l u m e o f formamide, containing 0.05% xylene cyanole a n d b r o m p h e n o l blue is added, and the samples are heated for 1 m i n at 90 ° and layered o n t o a 20% polyacrylamide, 7 M urea, 50 m M Tris-borate ( p H 8.3), l r a M E D T A gel, 1 m m thick. After electrophoresis (40 V / c m ) and radioautography, 16S r R N A fragments are cut out and eluted f r o m the gel. ~3 R N A sequences o f 5'-end-labeled fragments are determined according to the procedure o f Donis-Keller et al. ~7 Nine m i c r o g r a m s o f 30S subunits in 5/11 o f buffer H is preheated for 20 m i n at 37 °, then m i x e d with 2 / d o f a d C A A A C T C T aqueous solution (27-fold m o l a r excess) and 5/11 R N a s e H (12 units), and incubated for 14 hr at 4 °. The reaction mixture is treated with phenol and precipitated with ethanol. 3 ' - T e r m i n i o f 16S R N A fragments are labeled with [5'-32p]pCp a n d T4 R N A ligase ~s and the R N A fragments are subjected to

m7H. Donis-Keller, A. Maxam, and W. Gilbert, Nucleic,4cidsRes. 4, 2527 (1977). is T. E. England, A. G. Bruce, and O. C. Uhlenbeck, this series, Vol. 65, p. 65. a The arrows show the cleavage sites. The forked arrows show the cleavage sites which were determined with an accuracy of-+ 1 nucleotide residue. b The activity of 30S subunits in polyphenylalanine synthesis. 30S subunits were reconstituted from ] 6S RNA with single breaks (see text for details). The activity is expressed as a percentage of that of 30S subunits reconstituted from intact 16S RNA. c 16S RNA isolated by the acetic acid-urea procedure. a ND, not determined. e 30S subunits were reconstituted from cleaved 16S RNA preheated for 3 rain at 50 ° (i.e., after separation of the two halves of 16S RNA). /The exact position of a cleavagesite was not determined. s 16S RNA isolated by the phenol-SDS procedure. h Particles with sedimentation parameters characteristic of 30S subunits cannot be reconstituted from 16S RNA isolated by this procedure.

446


[29]

electrophoresis in a 20% polyacrylamide gel as above. The nucleotide sequences of the RNA fragments are determined according to the method of Peattie.19 The results are shown in Table I. Procedure B: Sites which are 50-300 nucleotides away from rRNA termini. In this case, fragments with free Y-terminal hydroxyls and of the appropriate size for direct RNA sequencing are formed. In the following example the nonanucleotide dTTCCTTTGA, which is partially complementary to several regions in E. coli 16S RNA, is used. We have found that RNase H treatment of 16S RNA isolated by the acetic acid-urea procedure in the presence of this oligodeoxyribonucleotide produces two RNA fragments of about 300 and 1150 nucleofides long, respectively (Fig. l). 16S RNA, 140 #g, extracted from 30S subunits with acetic acid-urea 15 is treated under standard conditions (see Section I,C,3) with RNase H (70 units) in the presence of 2.5 #g of dTTCCTTTGA. After electrophoresis in an 8% polyacrylamide-7 M urea gel, the polynucleotide (about 300 nucleotides long) is isolated from the corresponding band and 32p-labeled with [ 5 ' - 3 2 p ] p C p and T4 RNA ligase. TM The 32p-labeled RNA fragments formed during the ligation reaction are purified by 14% polyacrylamide7 M urea gel electrophoresis and sequenced according to the method of Peattie. ~9 The results (Table I) demonstrate that RNase H cleaved the single phosphodiester bond between nucleotides 301 and 302. Procedure C: Sites located in the middle of rRNA chains. In this case, fragments with nonphosphorylated 3' ends are too large to be sequenced directly, and it is convenient to use a combination of two oligodeoxyribonucleotides, the cleavage site of one of which has been identified independently. Fragments 50-250 nucleotides long can be obtained and their sequences can be determined in the usual way. For example, to identify the cleavage site produced in E. coli 16S RNA with RNase H in the presence of the oligonucleotide dTTTGCTCCCCrA (see Table I) complementary to the region 772-782, the 16S RNA is complexed with this undecanucleotide and the dodecanucleotide dATTACCGCGC~rU complementary to the region 523-524. 2o The complex is treated with RNase H under standard conditions and a fragment about 250 nucleotides long is isolated, labeled with [5'-32p]pCp and T4 RNA ligase, and sequenced according to the method of Peattie. 19 Procedure D: Fragmentation of rRNA in the presence of oligonucleotides with random nucleotide sequences. This approach provides informa19D. A. Peattie, Proc. Natl. Acad. Sci. U.S.A. 76, 1760 (1979). A. S. Mankin, unpubfished results.

[29]

1


2

3

4

5

G

l

447

8

9

FIG. 1. RNase H cleavage ofE. coli 16S RNA. Reaction conditions arc described in the text. Samples were subjc~'tedto ele~trophoresis in 4% polyacrylamide-7 M urea gels. Lane 1, 16S RNA treated with RNas¢ H in the absence of oligonucleotides (control). Lanes 2-4, 8, and 9 show 16S RNA treated with RNase H in the presence of oligonucleotide: lane 2, dCCTGTCT; 3, dTTCCTTTGA; 4, dTTTGCTCCCCA; 8, dCGTCAATTCATTT(complementary to the 16S region (913-925); 9, dATTACCGCGC,CrU. Lane 5 shows RNA size markers (obtained by treatment of 30S subunits with cobra venom endonuclease as described by S. K. Vasilenko, P. Carbon, J.-P. Ebel, and C. E. Ehresmann [J. Mol. Biol. 152, 699 (1981)]. Lane 6, 30S subunits treated with RNase H in the presence of d'I'TTG~CCCA. Lane 7, 30S subunits treated with RNase H in the absence of oligonuclcotides. (See also Table I for cleavage sites.)

448


[29]

tion about the distribution of the most exposed regions of an RNA molecule. After RNase H cleavage, rRNA fragments are labeled, separated, and sequenced as described above. We have studied the fragmentation of E. coli 16S RNA in the presence of a hexadeoxyribonucleotide mixture with the following nucleotide composition: A, 36%; G, 14.9%; C, 27.4%; T, 21.7%. 21 16S RNA is cleaved under standard conditions, the fragments are separated on a 14% polyacrylamide-7M urea gel, labeled with [5'-32P]pCp and T4 RNA ligase, and sequenced according to the Peattie method) 9 Thirty-two fragments have been localized in the 16S RNA polynucleotide chain.21

D. Reconstitution, Purification, and Biological Activity of 30S Subunits Containing Fragmented 16S rRNA 16S rRNA containing a unique nick (see Table I) is reprecipitated with 3 volumes of ethanol from TM buffer (30 m M Tris-HCl, pH 7.6, 20 m M MgC12, 6 m M 2-mercaptoethanol) at least 5 times to eliminate traces of phenol and is resuspended finally in TM buffer to a concentration of 30 A2t~/ml. Before reconstitution, the RNA solution is heated at 37* for 10 min and then cooled to 4 °. Total protein from 30S subunits is dialyzed for 6 hr against 3 changes (100 volumes) of 0.5 TMK buffer (TM buffer containing 0.5 M KC1). For reconstitution, 2 volumes of dialyzed proteins in 0.5 TMK buffer is slowly added to 1 volume of 16S RNA solution (30 A2~/ml) in TM buffer. The final concentration of KC1 is 0.33 M (0.33 TMK) and the concentration of 16S RNA is 10 A26o/ml. The 16S RNA to individual protein ratio is 1 : 1.5. The reconstitution mixture is incubated at 40* for 45 min, then cooled on ice and the reconstituted 30S subunits are precipitated with 0.65 volume of ethanol. After centrifugation, the pellet is dried and resuspended in 0.33 TMK buffer. The solution of 30S subunits is incubated at 40 ° for 20 min, cleared by centrifugation (3 min, 16,000 rpm), and centrifuged through a sucrose gradient (10-30%) in 0.33 TMK buffer at 31,000 rpm for 13 hr in a SW41 rotor (Beckman) or 50,000 rpm for 3.5 hr in a SW 50.1 rotor (Beckman). The gradient is fractionated and the fractions corresponding to 30S subunits are collected and precipitated with 0.65 volume of ethanol. The sedimentation coefficient of reconstituted subunits is determined by analytical centrifugation in a Beckman E ultracentrifuge (AnIT rotor). Poly(U)-directed [3H]polyphenylalanine synthesis is used to assay reconstituted 30S subunit activity. Each assay mixture of 50 #1 contains a 21 G. Z. Gaida, M. I. Kolosov, N. V. Chichkova, and A. A. Bogdanov, Bioorg. Khim. 9, 678 (1983).

[29]

SURFACETOPOGRAPHYOF rRNA

449

solution of 20 mMTris-HC1, pH 7.3, 10 mMMgC12, 0.1 MNH4C1, 1 m M dithiothreitol, 0.1 m M EDTA; 0.03 A260of 30S subunits; 0.06 A26oof 50S subunits; 50/lg [3H]Phe-tRNA with specific activity 9.7 Ci/mmol; 6 #g of protein factors (EF-Tu, EF-G, aminoacyl-tRNA synthetases); 0.4 m M GTP (Fluka); 6/~g of poly(U) (Calbiochem); 2 #l of phosphoenolpyruvate (50 m M solution, Boehringer); 1/tl of phosphoenolpyruvate kinase (10 mg/ml, Boehringer). The mixture is incubated for 5-40 min at 37 °. Aliquots (10/~l) are mixed with l0 ml 5% TCA and 50/A BSA (2 mg/ml), heated for 20 min at 90 °, and filtered through glass fiber filters (Whatman GFF). After washing and drying, the filters are placed in 5 ml of scintillation fluid and counted. E. Comments

The selectivity of fragmentation of RNA with RNase H is comparable with that of DNA hydrolysis with restriction enzymes. This approach works, however, only if the selected RNA region is in single-stranded (or weak double-stranded) conformation and exposed enough to bind with a complementary oligonucleotide. In addition the oligodeoxyribonucleotide-RNA hybrid must be accessible to the enzyme. These limitations turn out to be an advantage for topographical studies of the rRNA surface: since the substrates and effectors of protein biosynthesis reactions are also large in size, the method allows one to identify selectively RNA regions potentially involved in ribosome functioning. We have found that oligonucleotide binding followed by RNase H hydrolysis is quite sensitive to conformational changes of rRNA (compare binding dTTCCTTTGA to two 16S RNA samples isolated by different methods, Table I). Selective formation of hidden breaks in the rRNA chain can be useful for chemical modification of these sites with haptens and their subsequent localization by immune electron microscopy.22 It has to be noted that the method described here is the only approach for direct localization of oligodeoxyribonucleotide binding sites on RNA molecules. II. Determination of Apparent Association Constants for Binding of Oligonucleotides to rRNA by Nonequilibrium Gel Filtration Quantitative estimation of oligonucleotide binding to complementary sites in rRNA may provide valuable information on rRNA conformational changes resulting from interaction with ribosomal proteins, i.e., during reconstitution of ribosomal subunits, as well as on alterations in accessibility of certain regions of rRNA at different stages of ribosome functioning. 22 I. N. Shatsky and V. D. Vasiliev, this volume [5].

450


[29]

A. Principle of the Method and Calculation of Binding Parameters Conventional methods are not applicable to quantitative analysis of binding of oligonueleotides 7 - I 0 nucleotides long to rRNA: we have found that equilibrium cannot be attained after a reasonable time of dialysis through standard membranes. Moreover, equilibrium gel filtration requires a large amount of oligonueleotide. Among the non-equilibrium methods we have tested, such as sucrose gradient ultracentrifugation, filtration through nitrocellulose membranes, and gel filtration, the last approach turned out to be the most convenient. Its key point is the proper selection of filtration conditions (column size, elution rate, and column size to volume ratio) for the separation of the complex and unbound oligonueleotide. Separation should be good enough to permit measurement of the amount of the complex, and at the same time allow the complex to be in contact with the free oligonucleotide as long as possible. Under these conditions, the following method of calculation of the apparent association constant for oligonucleotide binding to rRNA (K') and the number of apparent binding sites (n) can be used. 23 The expression for the binding of oligonucleotide (L) to RNA (R) with n binding sites is R + nL~-RL.

(1)

The dissociation constant (K') will be K " = [R][L]. [RE.]

(2)

Because of possible complex dissociation during gel filtration the following relation is valid: RL. = t¢ RL.* where RL* is the measured amount of the complex in a gel filtration experiment and K is a nonequilibrium eoetiicient. The initial amount of RNA is Ro = R + RL. Thus, from F-xl. (2) it follows that Ro/RL* = ~cK'(1/[L].) + lc

(3)

Plotting Ro/RL* versus 1/[L]., it is possible to determine the tc as value of the ordinate after the extrapolation of 1/[L]. to zero (see Fig. 3). 23 E. A. Skripkin, A. M. Kopylov, A. A. Bogdanov, S. V. Vinogradov, and Y. A. Berlin, Mol. Biol. Rep. 5, 221 0979).

[29]

SURFACE TOPOGRAPHYOF rRNA

451

In logarithmic form Eq. (3) is log[(Ro/RL*) -- l] = --n log[L] + log K" By definition log K " = - n log K ' where K ' is the apparent association constant, log[(Ro/RL*) - 1] -- - n log[L] - n log K '

(4)

Linearization of Eq. (4) gives a straight line (see Fig. 4b). n is the tangent of the slope angle, and the intersection with the abscissa corresponds to log K'. B. An Example: A Comparison of the Accessibility of the 5'-Terminal Region of l 6 S rRNA in Free 16S rRNA, 30S Subunits, and 70S Ribosomes The synthetic oligodeoxyribonucleotide dCAAACTCT which is complementary to the 5'-terminal region 8 - 15 o f E . coli 16S rRNA has been used in this study. It has been shown in separate experiments with RNase H that this oligonucleotide binds to the targeted sequence of 16S rRNA (see Section I,C,6 in this chapter). The oligonucleotide is 32p-labeled at its 5'-end in the usual manner, Is to give a specific radioactivity in the range 5 - 150 X 106 cpm/A26 o. The complementary binding assay is done in a volume of 55/ll. Free 16S rRNA is incubated with the oligonucleotide in 10 m M Tris-HC1 (pH 7.5), 1 MNaC1, and l m M EDTA for 3 min at 60 °, for 10 min at 20 °, and for 3 0 - 4 0 min at 5 °, successively. Gel filtration is then performed on a Sephadex G-50 column (0.27 X 6.9 cm) at 5 ° with an elution rate of 3 ml/hr. For subunits and ribosomes the binding buffer is l0 m.M Tris-HC1 (pH 7.5), 10 m M MgC12, and 100 m M KC1. Two series of binding experiments are done: either incubation of subunits or ribosomes with the oligomer at 37 o for 20 min, then at 5 ° for 40 rain (preheating procedure) or incubation at 5 ° for 40 min. In all cases, ribosomal particles are preactivated at 37 ° for 20 min for subunits, and at 20* for 30 rain for ribosomes in the binding buffer. Gel filtration is performed under the same conditions as for 16S rRNA. The 32p radioactivity of the samples is determined by measuring Cerenkov radiation. The following values are used to calculate the solution molarity: A26o corresponds to 75 pmol of 16S rRNA, to 67 pmol of 30S subunit, to 39 pmol of 50S subunit, to 25 pmol of 70S ribosome, and to 12.5 nmol of octanucleotide.

452


,6

[29]

;

% x

B

a:

4

i

5

10

15

FRACTION NO. FIG. 2. Gel filtration of the complex of dCAAACTCT with 30S subunits. One hundred and sixty picomoles of30S subunits and 65 pmol of octanucleotide were incubated at 37 ° for 20 min and then at 5 ° for 40 min. Gel filtration was done at 50. The yield of the complex was 0.87 pmol. The dotted line represents the result of substraction of the curves corresponding to gel filtration of the octanucleotide with and without 30S subunits.

¢,0 0

-8 E Q,,

8

"0 e--

" 0

< I~ o E

e~

4

2 t

I

I

I 0.1

L

J

I

I

[.OJ

I

OCTAMER TO 30S RATIO IN MIXTURE F]G. 3. Binding curves for the dCAAACTCT-30S subunit interaction at 5 °. (1) 160 pmol subunits, preheating of the mixture at 37 °, and (2) 117 pmol subunits, without preheating. The molar ratio was varied by changing the oligonuclcotide concentration.

[29]


453

10

"7o

t// "f/

/t /'-°,'f i o

t

,4

8

42

I~ I

s.o -1=

t

l

s.5

-2

rT]2~t° (M)

I

6.o -tg[L~]

I

I

6.5

I

7.o

FIG. 4. Binding curve lineadzation for 30S subunits under preheating: (a) According to Eq. (3), assuming either one (1) or two (2) binding sites, r ~ - 110 was determined from extrapolation. (b) According to Eq. (4); the number of apparent binding sites (n == 1.6) and the average apparent association constant (K = 1.3 X 106 M -~) were calculated.

Figures 2 - 4 show the calculations for the preheating procedure for oligonucleotide binding, and the results of binding at 5 o are summarized in Table II. The results show equal accessibility of the 5'-terminal region of 16S rRNA alone, of 30S subunits, and of ribosomes. Separate binding experiments confirmed the specificity of binding: 50S subunits bind the oligonucleotide at least l0 times weaker. The number of apparent binding sites for 16S rRNA in all cases is more than one; this means that other binding sites, TABLE II OCTANUCLEOTIDE BINDING PARAMETERS

Sample 16S rRNA Binding at 5 ° 30S subunit 70S ribosome

Average apparent association constant K ' × 10-~ ( M - t ) a

Number of apparent binding sites n a

2.0

1.6

1.8 1.4

1.6 1.5

a Linearization of Eq. (4) was done by the least squares method; correlation coefficients exceed 0.92.

454


[29]

beside the 5'-terminal region, may exist with partial complementarity. This binding will increase the number n of the apparent binding sites and simultaneously reduce the average apparent association constant K'. C. Comments The method described here, as in the case of all nonequilibrium methods, provides for semiquantitative estimation of parameters of oligonucleotide binding to rRNA. For this reason, comparison of the accessibility of rRNA regions in different states should be done under precisely controlled and equivalent conditions (temperature, column size, oligonucleotide to rRNA ratio, oligonucleotide specific radioactivity, etc.). III. Specific Cleavage of rRNA at 7-Methylguanine Residue The rRNA regions with methylated guanine residues (mTG) occur at very specific locations on the surface of ribosomal subunits. One can assume that they are important for the biological activity of ribosomes. Here we describe the procedure for specific fragmentation of rRNA at mTG. A. Principle The chemistry of the cleavage of tRNA at 7-methylguanine was developed by Wintermeyer and Zachau. 24 It was based on reducing the 7-methylguanine residue with sodium borohydride and subsequent splitting of the polynucleotide chain at the reduced residue with aniline. The same reaction is used as a component of the guanine-specific cleavage reaction in the protocol for chemical sequencing of RNA. 19 However, it was found that this reaction critically depends on the concentration of the 7-methylguanine in the reaction mixture as well as the molar ratio of 7-methylguanine to the other nucleotides at the reduction stage. 25 This finding has explained why large RNA molecules, such as 16S rRNA, could not be cleaved at 7-methylguanine by the original procedure of Wintermeyer and Zachau. To overcome this obstacle, we have modified this procedure by the addition of exogenous 7-methylguanine in the form of carder RNA heavily methylated with dimethyl sulfate.26

24 W. Wintermeyer and H. G. Zachau, FEBS Left. 58, 306 (1975). 25 V. S. Zueva and A. S. Mankin, Biokhimiya (Moscow) 49, 160 (1984). 26 V. S. Zueva, A. S. Mankin, A. A. Bogdanov, and L. A. Baratova, Eur. J. Biochem. 146, 679 (1985).

[29]

SURFACE TOPOGRAPHY OF rRNA

455

B. Procedure 1. Preparing methylated RNA Carrier (Methyl-RNA). One milligram of RNA carrier is dissolved in 300/tl of 50 m M sodium cacodylate, pH 5.5, containing 1 m M EDTA, and 5 gl dimethyl sulfate is added. The mixture is incubated for 5 min at 90 °, and then quickly chilled on ice. Seventy-five microliters of cold 1.0 M Tris-acetate, pH 5.5, containing 1.5 M sodium acetate and 1.0 M 2-mercaptoethanol is added and RNA is precipitated by the addition of 900 gl of ethanol. The RNA pellet is reprecipitated from 0.3 M sodium acetate, pH 5.5, washed with ethanol, dried, and dissolved in water to a final concentration of 10 mg/ml. The methyl-RNA solution is distributed in 20-#1 aliquots and stored at - 2 0 °. 2. Cleavage of E. coli 16S rRNA at 7-Methylguanine. Forty micrograms of 16S rRNA is dissolved in 10 #1 of 0.5 M Tris-HC1, pH 8.0, and 2 gl of the methyl-RNA solution (20 #g) is added and vigorously mixed. Ten microliters of freshly prepared 1 M sodium borohydride solution is added and the reaction mixture incubated for 5 min at 20 ° in darkness. The reaction is stopped with 200/tl of 0.3 M sodium acetate, pH 5.5, and RNA is precipitated with 600 #1 of 96% ethanol. The ethanol-washed pellet is dried and dissolved in 80/A of water. Twenty microliters of aniline/acetic acid mixture (1:3, v/v) is added and incubated for 2 hr at 20 ° in darkness. The addition of the aniline/acetic acid mixture to the RNA solution sometimes leads to formation of a suspension. This does not influence the results of the procedure. RNA is precipitated by addition of 200 gl of 0.3 M sodium acetate, pH 5.5, and 900 gl of ethanol; the RNA pellet is washed with 1 ml of ethanol, dried, and dissolved in 90% formamide containing electrophoresis dyes. 27 The sample is denatured by heating for 1 min at 90 °, chilled on ice, and loaded onto a standard polyacrylamide gel (see Section I,C). Comments The protocol for analytical cleavage of 16S rRNA at m7G can easily be converted to large-scale use by proportionally increasing the amounts of all the reagents, keeping the concentrations unchanged. In the course of RNA cleavage, the methyl-RNA is degraded to rather short fragments which do not interfere with the 16S rRNA fragments either in the context of analytical electrophoresis, or preparative sucrose gradient fractionation. 27 The "postaniline" pellet of the cleaved RNA is not readily dissolved in aqueous solutions. The solubility of the RNA pellet can be increased by ethanol reprecipitation from 90% formamide or 7 M urea solutions.

456


[30]

Acknowledgments The authors are indebted to Drs. N. Sergeeva,V. Veiko, E. Volkov,S. Vinogradov, and V. Zaritova for synthesisof the oligodeoxyribonucleotidesused in this work.

[30] Enzymatic and Chemical Probing of Ribosomal R N A - Protein Interactions By JAN CHRISTIANSEN and ROGER GARRETT

About one third of the eubacterial ribosomal proteins have strong binding sites on the rRNAs and these primary binding proteins appear to have a special role in assembling, organizing, stabilizing, and/or reversibly altering the structure of the RNAs so as to generate a functionally active ribosome. Although the binding sites of many of these proteins have been mapped on the rRNAs, our understanding of their interactions is still rudimentary, and few principles have emerged regarding common structural motifs and/or sequence elements. However, double helices exhibiting irregularities, such as unpaired nucleotides or purine-purine juxtapositions, appear to be a common feature of most binding sites.l The protein-binding sites were first characterized partially by digesting complexes formed with in vivo 32p-labeled RNA with dbonucleases. More recently, various enzymatic and chemical probes, in combination with rapid gel electrophoretic analyses, have been employed to examine these complexes at nucleotide level. The following sections provide a guide to the various problems and pitfalls involved in such experiments and, also, to the type of structural information that can be obtained from them. Detailed experimental procedures for the isolation of ribosomal RNA, proteins, and complexes were described by Zimmermann in Vol. 59 of this series, 2 together with standard methods for testing the specificity, or "nativity," of a given complex. Isolation of Complexes Most studies, to date, have been performed on reconstituted complexes of ribosomal components of Escherichia coll. Table I lists the primary i R. A. Garrett, B. Vester, H. Leffers,P. M. Szrensen, J. Kjems,S. O. Olesen,A. Christensen, J. Christiansen, and S. Douthwait¢, in "Gene Expression" (B. Clark and H. U. Petcl~n, eds.), Alfred Benzon Symla. 19, p. 331. Munksgaard, Col~nhagen, Denmark, 1984. 2R. A. Zimmermann, this series, Vol. 59 [44]. METHODS IN ENZYMOLOGY, VOL. 164

Copyright© 1988by AcademicPress,Inc. All rightsof reproduction in any form reserved.

456


[30]

Acknowledgments The authors are indebted to Drs. N. Sergeeva,V. Veiko, E. Volkov,S. Vinogradov, and V. Zaritova for synthesisof the oligodeoxyribonucleotidesused in this work.

[30] Enzymatic and Chemical Probing of Ribosomal R N A - Protein Interactions By JAN CHRISTIANSEN and ROGER GARRETT

About one third of the eubacterial ribosomal proteins have strong binding sites on the rRNAs and these primary binding proteins appear to have a special role in assembling, organizing, stabilizing, and/or reversibly altering the structure of the RNAs so as to generate a functionally active ribosome. Although the binding sites of many of these proteins have been mapped on the rRNAs, our understanding of their interactions is still rudimentary, and few principles have emerged regarding common structural motifs and/or sequence elements. However, double helices exhibiting irregularities, such as unpaired nucleotides or purine-purine juxtapositions, appear to be a common feature of most binding sites.l The protein-binding sites were first characterized partially by digesting complexes formed with in vivo 32p-labeled RNA with dbonucleases. More recently, various enzymatic and chemical probes, in combination with rapid gel electrophoretic analyses, have been employed to examine these complexes at nucleotide level. The following sections provide a guide to the various problems and pitfalls involved in such experiments and, also, to the type of structural information that can be obtained from them. Detailed experimental procedures for the isolation of ribosomal RNA, proteins, and complexes were described by Zimmermann in Vol. 59 of this series, 2 together with standard methods for testing the specificity, or "nativity," of a given complex. Isolation of Complexes Most studies, to date, have been performed on reconstituted complexes of ribosomal components of Escherichia coll. Table I lists the primary i R. A. Garrett, B. Vester, H. Leffers,P. M. Szrensen, J. Kjems,S. O. Olesen,A. Christensen, J. Christiansen, and S. Douthwait¢, in "Gene Expression" (B. Clark and H. U. Petcl~n, eds.), Alfred Benzon Symla. 19, p. 331. Munksgaard, Col~nhagen, Denmark, 1984. 2R. A. Zimmermann, this series, Vol. 59 [44]. METHODS IN ENZYMOLOGY, VOL. 164

Copyright© 1988by AcademicPress,Inc. All rightsof reproduction in any form reserved.

[30]

PROBING OF r R N A - P R O T E I N

INTERACTIONS

457

TABLE I RNA FRAGMENTSPROTECTEDAGAINSTRIBONUCLEASEDIGESTIONBY RIBOSOMAL PROTEINSa

Protein 16S RNA ES4 ES8 ES15 ES20 23S RNA EL1 EL10 (L12)4 EL 11 EL23 EL24 5S RNA EL25

Ribonuclease

Approximate size

Ref.

A/T t A A Tl

470 35 65 280

3- 5 3, 4 3, 4 3, 4

Tt Tl Tl Tt TI

100 100 60 160 480

6, 7 8 9 6, 10 11, 12

A

40

13

a Those protein- RNA complexes from E. coli that yield ribonuclease-resistant fragments are included and literature references that include experimental conditions are listed. The approximate sizes of the RNA fragments are given (number of nucleotides). With the possible exception of the LI 1 site, they all exhibit internal cuts and yield multiple subfragments on denaturing gels. Primary binding proteins which have so far failed to protect RNA sites are $7, L2, L3, and L4. l The very basic protein L20 has yielded a protected fragment, 6 but since it tends to bind unspecifically to RNA (like LI5) further work is required to establish its significance, t Protein L18 protects almost all of the 5S RNA against RNAse A and Tt digestion. 13

binding proteins that yield protected fragment complexes together with the ribonuclease used. 3- ~3An appropriate literature reference giving the experimental conditions is also presented for each complex. Those proteinRNA complexes which have so far failed to yield ribonuclease-resistant a E. Ungewickell, R. Garrett, C. Ehresmann, P. Stiegler, and P. Fellner, Eur. J. Biochem. 51, 165 (1975). 4 R. A. Zimmermann, G. A. Mackie, A. Muto, R. A. Garrett, E. Ungewickell, C. Ehresmann, P. Stiegler, J. P. Ebel, and P. Fellner, Nucleic Acids Res. 2, 279 (1975). 5 R. A. Garrett, E. Ungewickell, V. Newberry, J. Hunter, and R. Wagner, Cell BioL Int. Rep. 1,487 (1977). 6 C. Branlant, A. Krol, J. Sriwidada, J. P. Ebel, P. Sloof, and R. A. Garrett, F E B S Lett. 52, 195 (1975). 7 p. Sloof, R. Garrett, A. Krol, and C. Branlant, Eur. J. Biochem. 70, 447 (1976). s A. Beauclerk, E. Cundliffe, and J. Dijk, £ Biol. Chem. 259, 6559 (1984). 9 F. J. Schmidt, J. Thompson, K. Lee, J. Dijk, and E. Cundliffe, J. Biol. Chem. 256, 12301 (1981). to B. Vester and R. A. Garrett, J. Mol. Biol. 179, 431 (1984).

458


[30]

complexes are indicated in the footnote to Table I; protein L18 protects most of the 5S RNA. ~3 The protein-RNA fragment complexes are purified under nonequilibrium conditions on sucrose density gradients, by gel filtration, or on polyacrylamide gels. This requires that the dissociation rate constant is below 0.03 hr -~ (hi2 -- 23 hr) to avoid appreciable loss of complex. However, the less stable complexes formed between 23S RNA and proteins L11 and Ll0 (L12)4 have been isolated on nitrocellulose filters. 8'9 A step for renaturing the RNA, before adding the protein, will generally improve the yield of complex. This is particularly important for complexes where the protein recognizes a large and compact RNA structure as occurs, for example, for proteins $4 and L24 that bind within domain I of 16S and 23S RNA, respectively. A recommended general procedure is to heat the RNA component at 50 o for l0 min followed by slow cooling to room temperature, and then add the protein and heat at 33 ° for 40 min. ~4 The importance of monovalent and divalent ions, as well as pH and temperature, has been investigated in detail for $4 and $8-16S RNA, ~5,~6 L24-23S RNA, ~6 and L18 and L25-5S RNA. 17 Each of these complexes requires different optimal binding conditions but all will form in the ribosomal reconstitution buffer containing 30 m M Tris-HC1, 20 m M MgCl 2, 300 m M KC1, 6 m M 2-mercaptoethanol at pH 7.8 and 37 ° . A few complexes containing 5S RNA have been isolated directly from the ribosome or cell with no dissociation step. These include ribosomal protein complexes from Bacillus stearothermophilus, ~a E. coli, ~9 yeast,2° rat, 2~ and the transcription factor TF IIIA-5S RNA complex from Xenopus laevis. 22 However, some of these preparative procedures involve denaturing steps, including EDTA treatment, that are known to cause unspecific complex formation. 23,24 II C. Branlant, J. Sriwidada,A. Krol, and J. P. Ebel, Eur. J. Biochem. 74, 155 (1977). t2 p. Sloof, J. Hunter, R. A. Garrett, and C. Branlant, Nucleic Acids Res. 5, 5303 (1978). ~3S. Douthwaite, R. A. Garrett, R. Wagner, and J. Feunteun, Nucleic Acids Res. 6, 2453 (1979). ~4E. Ungewickell, R. A. Garrett, and M. Le Bret, FEBSLett. 84, 37 (1977). ~5C. Schulte and R. A. GarteR, Mol. Gen. Genet. 119, 345 (1972). ~6C. Schulte, C. A. Morrison, and R. A. Garrett, Biochemistry 13, 1032 (1974). ~7p. Spierer and R. A. Zimmermann, Biochemistry 17, 2474 0978). ~s j. Home and V. A. Erdmann, Mol. Gen. Genet. 119, 337 (1972). ~9U. Chen-Schmeisser and R. A. Garrett, FEBS Lett. 74, 287 (1977). 2o R. N. Nazar, M. Yaguehi, G. E. WiUick, C. F. Rollin, and C. Roy, Eur. J. Biochem. 102, 573 (1979). 2t j. Behlke, H. Weltle, I. Wendel, and H. Bielka, Acta Biol. Meal. Germ. 39, 33 (1980). 22 B. Picard and M. Wegnez, Proc. Natl. Acad. Sci. U.S.A. 76, 241 (1979). 23 U. Chen-Sehmeisser and R. Garrett, Eur. J. Biochem. 69, 401 (1976). 24 I. Newton, J. Rinke, and R. Bdmacombe, FEBSLett. 51, 215 (1975).

[30]

PROBING OF r R N A - PROTEIN INTERACTIONS

459

Analysis of Binding Sites The usual strategy for characterizing a protein site on an RNA molecule is to analyze the reactivity of both free and complexed RNA, treated under the same conditions with various enzymes and chemicals, on denaturing polyacrylamide gels. The reactive nucleotides are deduced by aligning the sample bands with those of coelectrophoresed sequencing tracks. The classical approach is to end label the RNA and monitor strand scissions, 25 but use of reverse transcription to detect reactive sites is undoubtedly superior for examining large RNA fragments and molecules.26,27

In Vivo Labeling Method Most of the protein binding sites on the large RNAs were characterized using in vivo a2p-labeled RNA and the standard Sanger RNA sequencing methods. Relevant references to these methods are included in Table I. This approach now has the disadvantages that it is both laborious and requires large amounts of radioactive isotope. Moreover, often the amount of radioactivity in the purified protein-RNA fragment complex, isolated in a given experiment, was too low for complete analysis. It has an advantage, however, over the more recently developed methods when examining large RNA sites that have incurred multiple internal cuts during preparation, because it eventually yields a complete analysis of the RNA moiety. With the end-labeling methods described below, the molar yields of the subfragments are ditticult to estimate and there is always the possibility that one or more subfragments will exhibit heterogeneous ends or not label.

End Labeling Methods These methods are best suited to naturally occurring RNAs such as 5S RNA or 5.8S RNA that have homogeneous ends. For example, 5S RNA complexes with EL18 and EL25 have been characterized in detail by this approach. 2s-3° However, the techniques have been applied successfully to the analysis of protein-RNA fragment complexes isolated from the large RNAs, in particular the L23 site on 23S RNA. 1° 25 D. A. Peattie and W. Gilbert, Proc. Natl. Acad. Sci. U.S.A. 77, 4679 (1980). 26 D. C. Youvan and J. E. Hearst, Proc. Natl. Acad. Sci. U.S.A. 76, 3751 (1979). 27 A. Barta, G. Steiner, J. Brosius, H. F. Noller, and E. Kuechler, Proc. Natl. Acad. Sci. U.S.A. 81, 3607 (1984). 2s D. A. Peattie, S. Douthwaite, R. A. Garrett, and H. F. Noller, Proc. Natl. Acad. Sci. U.S.A. 78, 7331 (1981). 29 S. Douthwaite, A. Christensen, and R. A. Garrett, Biochemistry 21, 2313 (1982). 3oj. Christiansen and R. A. Garrett, in "Structure, Function and Genetics of Ribosomes" (B. Hardesty and G. Kramer, eds.), p. 397. Springer-Verlag, New York, 1986.

460


32p 5'

A ~

B ~

A

B

[30]

Gel I ~ result

I'-I

FIG. 1. A primary cut (A) is distinguished from a secondary cut (B) by the presence of the former in autoradiograms of both 5'- and Y-end-labeled RNA; the latter is always absent from one of the autoradiograms.

Samples to be probed with ribonucleases should be end labeled before digestion whereas samples to be chemically modified can be labeled before or after modification. The advantage of prelabeling is that it is easier to monitor purification of the complex, or RNA, after modification, whereas the advantage of postlabeling is that the sample may be stored for long periods without deterioration. 3'-End labeling of RNA molecules can be performed directly with [a2p]pCp and RNA ligase, a~ while efficient 5'-end labeling with [7-32p]ATP and polynucleotide kinase requires a prior alkaline phosphatase treatment. However, RNA fragments obtained by partial digestion with RNase T~ or A can be 5'-end labeled directly, while the 3'-end labeling requires a prior phosphatase treatment. Phosphatasetreated RNA should always be purified on a denaturing polyacrylamide gel to prevent labeling at nicked positions. Most laboratories use only one type of end labeling but it should be emphasized that the parallel use of 3'- and 5'-end labeling is a powerful way of distinguishing between primary and secondary effects in enzymatic probing experiments.32,33 This principle is illustrated diagrammatically in Fig. 1 and an example is depicted in Fig. 2 for yeast 5S RNA treated with RNase T 2. Moreover, it is unnecessary to purchase two kinds of radioactive substrate, since it is straightforward to produce [32p]pCp from [7-32p]ATP and Cp; after heat-inactivating the polynucleotide kinase at 65 ° for 10 min and adjusting the buffer. This mixture can be employed for 3'-end labeling of RNA with no further purification step. It is recommended to select for a conformationally homogeneous complex29 and RNA 32 on a nondenaturing gel immediately after probing. The advantages of this are twofold. First, complexes that have altered conformationally as a result of secondary cutting effects, particularly with singlestrand specific ribonucleases, can be eliminated from further analysis. Second, RNA that has dissociated during treatment with, for example, 3t A. G. Bruce and O. G. Uhlenbeck, Nucleic Acids Res. 5, 3665 (1978). 32 S. Douthwaite and R. A. Garrett, Biochemistry 20, 7301 (1981). 33 R. A. Garrett and S. O. Olesen, Biochemistry 21, 4823 (1982).

[30]

PROBING OF r R N A - P R O T E I N INTERACTIONS

51 A

461

3I B C

D E F

T

W

B

C E

G

-'-50

-.-G37 .--40

-G41

----G37 ---30 --,-G52 ~25

....

--20 FIG. 2. Five micrograms of yeast 5S RNA was 5'- or 3'-end labeled, renatured, and digested with RNase T 2. The autoradiogram demonstrates clearly the occurrence of secondary cuts in the regions 22-24 and 32-36 on the 5'-end-labeled sample and at positions 42-44 on the 3'-end-labeled sample. Digestions were performed in 10/tl of TMK buffer for 20 rain at 0* with (A) no ribonuclease, (B) 0.005 unit, (C) 0.01 unit, (D) 0.05 unit, (E) 0.1 unit, and (F) 0.5 unit. The intact RNA band was purified by polyacrylamide gel electrophoresis prior to running the sequencing gel. Ribonuclease T 1 (T) and water hydrolysis (W) tracks are shown for the 5'-end-labeled sample, and a guanosine ((3) chemical sequencing track is given for the 3'-end-labeled sample.

diethyl p y r o c a r b o n a t e (see below) c a n be r e m o v e d . H o w e v e r , r u n n i n g a n o n d e n a t u r i n g gel can also lead to a loss o f i n f o r m a t i o n if a c o n f o r m a tional change, a n d a c o n s e q u e n t alteration in electrophoretic mobility, results f r o m p r i m a r y cutting o r modification effects. C h e m i c a l l y modified c o m p l e x e s or free R N A c a n be subjected to an

462


[30]

additional selection for full-length molecules, on denaturing gels, prior to analysis on sequencing gels. This step is particularly important for the reverse transcriptase method because the primer preferentially anneals to the smaller degradation products. Exchange Reaction

If reconstitution of a specific complex is difficult or end labeling of a native complex is unsatisfactory, it is worth trying to exchange end labeled RNA with the unlabeled RNA in the complex. This approach has been applied successfully to protein-5S RNA complexes from yeast~ and other eukaryotes in 25 m M ethylenediaminetetraacetic acid (EDTA) at pH 7.0. Although the labeled complex is selected, the "nativity" of the resulting interaction is uncertain since multi-contact proteins may attach to a few high-affinity, and Mg2+-independent, sites on 5S RNA while weaker Mg2+dependent interactions are irreversibly lost. EDTA is known to produce unspecific interactions in ribosomal protein-RNA interactions. 23,24 A safer approach, if the protein remains soluble, is to raise the KC1 concentration and subsequently reduce it. This has recently been done successfully, in the presence of a nonionic detergent, for the Xenopus TF IIIA- 5S RNA complex; the specificity of this complex was confirmed by ribonuclease probing studies.35 Modification - Selection

The experiment can be applied to both reconstituted complexes and to those produced by exchange and it is the only appropriate option if the complex dissociates during chemical treatment. The RNA component is modified, complexed with protein, and then the distribution of modified nucleotides between free and complexed RNA is analyzed. 28 If RNA molecules exhibiting a particular modified nucleotide are excluded from the complex, then it is inferred that this nucleotide is involved, directly or indirectly, in binding. When renatured RNA is modified the structural interpretation is fairly straightforward, but difficulties arise when denatured RNA is modified and then renatured, since many modifications will be selected against simply because they inhibit renaturation of the RNA structure. This problem is particularly acute for the diethyl pyrocarbonate reaction that is normally used to probe for accessible N-7 atoms on adenosines; first, because reaction at the N-7 position destroys the imidazole ring 34 R. N. Nazar and A. G. Wildeman, Nucleic Acids Res. 11, 3155 (1984). 35 j. Christiansen, R. S. Brown, B. S. Sproat, and R. A. Garrett, EMBOJ., 6, 453 (1987).

[30]


INTERACTIONS

463

and, second, because the reagent can also modify the exocyclic amino group at position 6. So far, the method has been applied only to 5S RNA 2s where the main experimental problem has been to cleanly separate the complex containing modified RNA from the unbound RNA. This problem arises because RNA that has been partially modified may bind less strongly to the protein and will tend to dissociate during electrophoresis. However, by coelectrophoresing unmodified complex and RNA as markers, the bands can be excised fairly accurately. Examination of the RNA material migrating between the modified complex and unbound RNA can also, of course, yield additional useful information on the binding site.

Hybridization Method This procedure involves hybridizing a restriction fragment of DNA to an RNA region that is known to constitute a protein binding site after it has been modified in the free and complexed state. Nonhybridized RNA is then removed by ribonuclease treatment and the hybridized RNA is subsequently isolated, end labeled, and analyzed as described above. This method, which is described in detail in this volume by Van Stolk and Noller [32], has so far only been applied to the large rRNAs. 36,37 The method is laborious and sometimes it is difficult to end label the hybrids. However, it has a potential advantage over the above-mentioned methods in that a whole protein binding region can be isolated from a large RNA with no internal cuts. Moreover, there are fewer problems with control bands than occur with the reverse transcriptase method described below.

Reverse Transcriptase Method This approach exploits the property of the reverse transcriptase to pause, or terminate, at a modified base or a break in the ribose-phosphate backbone. 26,27,38 For smaller RNAs (< 300 nucleotides) the reverse transcriptase method provides an excellent alternative to those described above. Moveover, the additional inclusion of an aniline/acid-catalyzed strand scission provides the reverse transcriptase procedure with the ability to monitor N-7 modifications in addition to modifications of the pyrimidine nucleus and ribonuclease-generated cuts. 36 B. J. Van Stolk and H. F. Noller, J. Mol. Biol. 180, 151 (1984). 37 A. Andersen, N. Larsen, H. Leffers, J. Kjems, and R. A. Garrett, in "Structure and Dynamics of RNA" (C. Hilbers and P. van Knippenberg, eds.), p. 221. Plenum, New York, 1986. 38 D. Moazed, S. Stem, and H. F. Noller, J. Mol. Biol. 187, 399 (1986).

464


[30]

The method has several advantages over the end-labeling procedures. First,it is unnecessary to label the complex. Second, it is appropriate for probing internal regions of large R N A s (see chapter by Noller et al. [33]) since the end-labeling techniques combined with high-resolutiongels have a practical limit of only about 300 nuclcotidcs from the termini.39 The major disadvantages of the procedure arc that strong termination may occur at naturally occurring modifications within the R N A and that it is necessary to synthesize primers or prepare restrictionfragments so the R N A under study can be covered; however, the combination of t~-[35S]dNTPs and wedge-shaped gels allows the analysis of up to 500 nuclcotidcs per loading.4° W h e n the reverse transcriptaseprocedure is used for small R N A molecules such as 5S R N A , it is preferable to 5'-end label the primer with [~,-a2p]ATP rather than to use ix-labeleddNTPs. Loss of information at the 3'-end of the R N A , where the primer anneals, is unavoidable. Data obtained from the reverse transcriptascprocedure and end-labeling methods should be identical. However, occasionally the former approach revealed double bands particularlyat Kethoxal modification sites and cobra venom (CV) RNase cutting sites;the lattermay be duc to the presence of a 5'-phosphate at the cutting site. R ib o n u cleas e and Chemical P r o b e s

The quality of the selectedprobes is very important. Commercial preparations of ribonuclcases often exhibit protcase activity,so itis essentialto establishthe integrityof the protein component in a complex afterribonuclcase treatment. Highly purified chemical reagents should also be used, invariably. It is also important to check the effectof the reagents on the stabilityof the complex since some of them willmodify the protein component. Thus, kethoxal modifies the guanidino group of argininc, dicthyl pyrocarbonate attacks the imidazolc side chain of histidinc,and dimcthyl sulfate may attack cystcincs;both of the latterreactions have been shown to dissociatecomplexes. A few examples from each group of probes arc listed in Table II together with their spccificiticsand conditions for use. Regardless of the type of probe employed, itis important to use conditions given in Table II that produce single hits per molecule in order to minimize secondary effects.The problem of secondary effectsis greater for ribonuclcascs, and 39 R. A. Crarrett, A. Christensen, and S. Douthwaite, J. Mol. Biol. 179, 689 (1984). 4o j. Egebjerg, H. Leffers, A. Christensen, H. D. Andersen, and R. A. Garrett, J. Mol. Biol. 196, 125 (1987).

[30]

PROBING OF rRNA-PROTEIN INTERACTIONS

465

TABLE II EXPERIMENTALCONDITIONSFOR ENZYMATICAND CHEMICALPROBESa

Probe Ribonucleases T~ T2 A CV Chemicals DMS DEPC Kethoxal CMCT

Amount

Buffer

Volume (/tl)

Time (rain)

Gp~ Ap~, > Np~ Yp~ N~p in helix

0.005 U 0.02 U 0.01 U 0.1 U

TMK TMK TMK TMK

20 20 20 20

30 30 30 30

G(N-7) > A(N-I) > C(N-3) A(N-7) G(N- I,N-2) U(N-3) > G(N-2)

0.3 gl 5 gl I gmol 2.5/zmol

HMK HMK HMK BMK

50 50 50 50

10-30 240 180 150

Specificity

a In all experiments 5 #g of free or complexed RNA is treated at 0". TMK is the ribosomal reconstitution buffer without 2-mercaptoethanol (see text), and HMK and BMK are corresponding buffers where Tris has been replaced by 70 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES) (pH 7.6) and 50 mM potassium borate (pH 8.0), respectively. CV, cobra venom; DMS, dimethyl sulfate; DEPC, diethyl pyrocarbonate; CMCT, 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluene sulfonate.

especially single-strand-specific ones, since they effect strand scission as the primary event. A good rule of thumb is to obtain gel analysis tracks where more than 50% of the RNA molecules is unaffected. However, for certain reactions this may be insufficient. For example, the N-3 on cytidine reacts weakly with dimethyl sulfate so, although gel analyses (after hydrazine treatment and strand scission) may indicate less than one hit per molecule, this could be erroneous due to the concurrent methylation of N-7 on guanosines and N-1 on adenosines. Generally, reactions are performed at 0 ° where the protein-RNA complexes are most stable. However, some chemical reactions are too slow at 0 ° and require a higher temperature; a rough guide is to halve the reaction time for each temperature increase of 10 °. Dimethyl sulfate and diethyl pyrocarbonate modifications are generally performed in 200 #1 cacodylate buffer.25 HEPES is suggested as an alternative buffer (Table II) because it has a higher pK (7.5) and is insufficiently nucleophilic to interfere with the dimethyl sulfate reaction. Although HEPES has a greater buffer capacity than cacodylate at neutral pH, it is nevertheless prudent to check the pH after adding dimethyl sulfate, since the latter invariably contains some sulfuric acid. The reaction volumes given in Table II have also been reduced to facilitate preparation of sampies for nondenaturing polyacrylamide gels.

466


[30]

When dimethyl sulfate is used to probe N-7 positions on guanosines, the modified imidazole ring is reduced with sodium borohydride prior to the aniline/acid-catalyzed strand scission. The reduction reaction is critically dependent on the concentration of N-7 methylated guanosines and their ratio to nonmethylated nucleotides, so the inclusion of a heavily methylated cartier RNA will ensure extensive reduction.4l Additional Probes

The above-mentioned ribonucleases and chemicals are those we have employed extensively in studies of various RNAs and their complexes. However, owing to the high affinity of RNase A for pyrimidine-adenosine linkages, this enzyme tends also to cut in double-helical regions. More useful alternatives, with greater single-strand specificity, may be the cyfidine-specific ribonuclease CL3 from chicken liver and the pyrimidine-specific Bacillus cereus enzyme (R. Nazar, personal communication). Singlestrand-specific enzymes such as RNase S~ and the adenosine-specific RNase U2 are less useful for studying complexes owing to their low pH optima. The purine-specific ot-sarcin enzyme cuts both helices and loops and is potentially useful as an universal probe with a minimum of "blind spots" (see Huber and Wool [3 1]). Its major drawback, however, is its inhibition by Mg2+ which is required for the stability of many complexes. Ethyl nitrosourea ethylates accessible phosphates, and is useful for localizing phosphates involved in higher-order interactions, and for investigating protein-rRNA complexes.42 The advantage of this reagent is its unique specificity, but its low reactivity is a disadvantage (an order of magnitude slower than methylation by dimethyl sulfate) which can lead to RNA degradation problems.-43 Recently, a highly water-soluble psoralen derivative, 8-[(3-(4-methyl-1piperazinyl)propyl)oxy] psoralen, has been employed as a mono-addition reagent."-4~ Psoralen mono-addition is a two-step process: intercalation followed by near UV-induced cyclobutane formation. The derivative 4mV. S. Zueva, A. S. Mankin, A. A. Bogdanov, and L. A. Baratova, Eur. J. Biochem. 146, 679 (1985). 42 j. McDougall and R. N. Nazar, J. Biol. Chem. 258, 5256-5259 (1983). 43 V. V. Vlassov, R. Giege, and J. P. Ebel, FEBSLett. 120, 12 (1980). 44 j. B. Hansen, P. Bjerring, O. Buchardt, P. Ebbesen, A. Kanstup, G. Karup, P. Nielsen, B. Norden, and B. Ygge, J. Med. Chem. 28, 1001 (1985). 4s H. Leffers, J. Egebjerg, A. Andersen, T. Christensen, and R. A. Garrett, J. Mol. Biol., submitted for publication, 4~j. Christensen, Nucleic Acids Res., submitted for publication.

[30]


INTERACTIONS

467

reacts strongly with uridines at or near internal loops but not with regular double helices. The advantage of this probe is that it recognizes flexible RNA regions which can undergo the necessary axial rise per residue; such features often constitute protein-binding sites. 35,4°,4s,~ Interpretation of Polyacrylamide Gel Analyses Establishing which nucleotides have reacted in free RNA and complexes is usually straightforward if sequencing tracks are coelectrophoresed with the samples. In the 3'-end labeling procedure, denatured RNA can be treated with enzymes or chemicals to yield sequencing tracks; with the exception of RNase St and the CV RNase-generated fragments, ribonuclease cuts align with enzyme tracks, while chemically induced sequence tracks align with those produced by chemicals and by RNase S~ and CV RNase. In the reverse transcriptase procedure, sequencing tracks obtained by using dideoxynucleoside triphosphates47 are displaced by one nucleotide relative to the modified positions. It is difficult to quantify digestion and modification data on autoradiograms accurately, and, generally, a semiquantitative + / - system is used for band intensities. A picture of a protein-binding site emerges when the "fragmentation" patterns of the free and complexed RNA are compared for the various probes. Usually, multiple protection effects occur with occasional enhancements. Interpretation of the protection of a ribonuclease cut is less straightforward than for chemical modification owing to the sheer bulk of the ribonucleases. However, free RNA is more flexible conformationally than complexed RNA and some protection effects will result from proteininduced tightening of conformation; enhanced reactivities may reflect an increase in local conformational homogeneity. Most probes exhibit single-strand specificity and the data inevitably have a bias toward loop regions. It is important, therefore, not to dismiss double-helical regions that are "blind spots." Dimethyl sulfate reacts weakly with double helices in the major groove at the N-7 position while the mechanism of recognition of double helices by the CV RNase remains uncertain. The CV RNase employed in our studies was isolated by Vassilenko and Babkina~ and appears to attack double helices primarily while the commercially available RNase Vt from cobra venom may have an additional strong specificity for helical single strands. 49,5° 47 F. Sanger, S. Nieldcn, and A. R. Coulson, Proc. NatL Acad. Sci. U.S.A. 74, 5463 (1977). 4s S. Vassilenko and V. Babkina, Biokhimiya (Moscow) 341, 705 (1965). 49 R. E. Lockard and A. Kumar, Nucleic Acids Res. 9, 5125 (1981). 50 H. B. Lowman and D. E. Draper, J. Biol. Chem. 261, 5396 (1986).

468

C H E M I C A L A N D E N Z Y M A T I C PROBING M E T H O D S

[31]

Concluding R e m a r k s Many protein-binding sites on ribosomal RNA have now been mapped by enzymatic and chemical methods. These approaches have suggested possible protein attachment sites, some of which have been investigated in greater detail by site-directed mutagenesis. 5~,s2 Moreover, previously unidentified sites, such as those of E. coli proteins L2 and L3, have recently been characterized by synthesizing RNA domains using the T7 polymerase system, selectively binding the protein from total ribosomal proteins, and then probing the binding site. 45 What is lacking now is more information on the conserved sequences and structure of the ribosomal proteins. Acknowledgments We thank all current and previous members of this laboratory for their experimental contributions, and SolveigKjaer for her help with the manuscript. 51j. Christiansen, S. Douthwaite, A. Christensen, and R. A. Garrett, EMBO J. 4, 1019 (1985). 52R. J. Gregoryand R. A. Zimmerrnann, NucleicAcids Res. 14, 5761 (1986).

[31] Use of ct-Sarcin to Analyze Ribosomal R N A - Protein Interactions By PAUL W. HUBER and IRA G. WOOL

Galas and Schmitz devised a relatively rapid and precise method to identify a protein-binding domain on a molecule of DNA that they named footprinting. 1 The procedure requires and exploits the placement of a radioactive label at the end of a segment of DNA, the digestion of this DNA to generate a nested set of oligonucleotide fragments (from the free DNA and from the protein-nucleic acid complex), and the analysis of the digestion products alongside sequencing lanes on a polyacrylamide gel. The hydrolysis of the nucleic acid must be carded out in specific circumstances. All, or nearly all, positions must be equally susceptible to the action of the nuclease, and the digestion must be limited, on the average, to no more than one cut per molecule (i.e., "one-hit-kinetics"). This, of course, requires that not all of the molecules be hydrolyzed. If these two conditions are met, a broad distribution of fragment sizes is generated by single cuts of the nucleic acid. Comparison of the digestion ladders of protein-bound nucleic acid and protein-free nucleic acid reveals the site of I D. J. Galas and A. Schmitz, Nucleic Acids Res. 5, 3157 (1978). METHODS IN ENZYMOLOGY, VOL. 164

Copyfisht© 1988by Academic Press,Inc. All fightsof reproduction in any form reserved.

468

C H E M I C A L A N D E N Z Y M A T I C PROBING M E T H O D S

[31]

Concluding R e m a r k s Many protein-binding sites on ribosomal RNA have now been mapped by enzymatic and chemical methods. These approaches have suggested possible protein attachment sites, some of which have been investigated in greater detail by site-directed mutagenesis. 5~,s2 Moreover, previously unidentified sites, such as those of E. coli proteins L2 and L3, have recently been characterized by synthesizing RNA domains using the T7 polymerase system, selectively binding the protein from total ribosomal proteins, and then probing the binding site. 45 What is lacking now is more information on the conserved sequences and structure of the ribosomal proteins. Acknowledgments We thank all current and previous members of this laboratory for their experimental contributions, and SolveigKjaer for her help with the manuscript. 51j. Christiansen, S. Douthwaite, A. Christensen, and R. A. Garrett, EMBO J. 4, 1019 (1985). 52R. J. Gregoryand R. A. Zimmerrnann, NucleicAcids Res. 14, 5761 (1986).

[31] Use of ct-Sarcin to Analyze Ribosomal R N A - Protein Interactions By PAUL W. HUBER and IRA G. WOOL

Galas and Schmitz devised a relatively rapid and precise method to identify a protein-binding domain on a molecule of DNA that they named footprinting. 1 The procedure requires and exploits the placement of a radioactive label at the end of a segment of DNA, the digestion of this DNA to generate a nested set of oligonucleotide fragments (from the free DNA and from the protein-nucleic acid complex), and the analysis of the digestion products alongside sequencing lanes on a polyacrylamide gel. The hydrolysis of the nucleic acid must be carded out in specific circumstances. All, or nearly all, positions must be equally susceptible to the action of the nuclease, and the digestion must be limited, on the average, to no more than one cut per molecule (i.e., "one-hit-kinetics"). This, of course, requires that not all of the molecules be hydrolyzed. If these two conditions are met, a broad distribution of fragment sizes is generated by single cuts of the nucleic acid. Comparison of the digestion ladders of protein-bound nucleic acid and protein-free nucleic acid reveals the site of I D. J. Galas and A. Schmitz, Nucleic Acids Res. 5, 3157 (1978). METHODS IN ENZYMOLOGY, VOL. 164

Copyfisht© 1988by Academic Press,Inc. All fightsof reproduction in any form reserved.

[31]

USE OF O/-SARCINTO FOOTPRINT rRNP COMPLEXES

469

association. This method depends on the physical protection of the DNA by a protein. The successful adaptation of this technique to ribonucleoprotein (RNP) complexes has been thwarted by the unavailability of ribonucleases that can generate an extensive population of oligonucleotide fragments of different lengths. Ribonucleic acids have a great deal of secondary and tertiary structure and most ribonucleases are structure specific, cleaving only single-stranded or, in a small number of cases, only double-stranded RNA. In nondenaturing conditions hydrolysis by a particular ribonuclease is confined to a relatively few sensitive positions in the structure. Thus, neither condition for proper hydrolysis is met: positions are not equally hydrolyzed and secondary cuts occur. t~-Sarcin is a cytotoxic ribonuclease that blocks protein synthesis by inactivating ribosomes. 2-4 The cause of this inactivation is a single cleavage at a position within a highly conserved sequence of the large (23-28S) rRNA. 5 However, a-sarcin can hydrolyze on the 3' side of purines in both single- and double-stranded regions of naked RNA. 6 Moreover, this nuclease is fully active at neutral pH and has no unusual cofactor requirements. These properties suggested to us that c~-sarcin might be a valuable reagent for footprinting protein-RNA complexes. We established the suitability of a-sarcin for this purpose by determining the association site for the three ribosomal proteins (L5, Ll 8, and L25) that bind to Escherichia coli 5S rRNA (Fig. 1).7 Our definition of the L18 and L25 binding sites is in good agreement with several earlier studies; moreover, we succeeded in determining the L5 attachment site which had not been identified before. We have gone on to locate the regions of association of Xenopus transcription factor IIIAs and of rat ribosomal protein L59 on their cognate 5S rRNAs, thus providing an extensive analysis of protein- 5S rRNA interactions. Materials a-Sarcin is produced by a mold, Aspergillusgiganteus MDH 18894; the protein can be purified in high yield directly from the extracellular culture 2 C. Fernandez-Puentes and D. Vazquez, F E B S Lett. 78, 143 (I 977). 3 F. P. Conde, C. Fernandez-Puentes, M. T. V. Montero, and D. Vazquez, FEMSMicrobiol. Lett. 4, 349 (1978). 4 A. N. Hobden and E. Cundliffe, Biochem. J. 170, 57 (1978). s y. Endo and I. G. Wool, J. Biol. Chem. 257, 9054 (1982). 6 y. Endo, P. W. Huber, and I. G. Wool, J. Biol. Chem. 258, 2662 (1983). 7 p. W. Huber and I. G. Wool, Proc. Natl. Acad. Sci. U.S.A. 81, 322 (1984). s p. W. Huber and I. G. Wool, Proc. Natl. Acad. Sci. U.S.A. 83, 1593 (1986). 9 p. W. Huber and I. G. Wool, J. Biol. Chem. 261, 3002 (1986).

I 2

3

4

5

6

78

I

2

3

4

5

6

7

8

Fxo. 1. Protection of5S rRNA from digestion with c~-sarcinby ribosomal protein LIB.7 E. ¢oli ribosomal protein L18 (10 pM) was incubated with renatured radioactive 5S rRNA

(0.4 #M) in buffer (50 mM Tris-HC1, pH 7.6, 285 mM KCI, 6 mM MgCl2) for 30-45 min at 33 °. The mixture was diluted (to 50 mM Tris-HC1, pH 7.6, 95 mM KC1, 2 mM MgC12)and digested with a-sarcin for 15 min at 30 °. Lanes: 1, alkaline hydrolyzate of 5S rRNA; 2, T~ ribonuclease digest of 5S rRNA; 3, 8 #M~-sarcin digest of a LI8-5S rRNA complex; 4, 8 #Ma-sarcin digest of 5S rRNA; 5, 4 p.Ma-sarcin digest o f a L18-5S rRNA complex; 6, 4 #Ma-sarcin digest of5S rRNA; 7, 0.8/~M¢-sarcin digest of5S rRNA in the absence of KCl and MgCl2; 8, 5S rRNA that was not treated with a-sarcin. The 5S rRNA was labeled at the 5' end and the digests were analyzed by electrophoresis on 10% polyacrylamide gels. Brackets enclose regions of 5S rRNA protected by ribosomal protein Ll 8.

[31 ]

USE OF OL-SARCINTO FOOTPRINT rRNP COMPLEXES

471

filtrate.'° Fermentation conditions for optimal production ofa-sarcin have been described.l' The lyophilized product is stable, and solutions of small amounts of the protein dissolved in water at a concentration of 10 mg/ml retain full activity for several months if kept at 5 °. In agreement with Olson and Goerned ° (but in contrast to Schindler and Davies '2) we find that slow freezing at - 2 0 ° causes inactivation of the nuclease; therefore, solutions of ot-sarcin should not be frozen. Methods The concentration of a-sarcin that will generate a broad distribution of oligonucleotides of different lengths in the conditions of a particular experiment must be determined empirically. The range of concentrations of the toxin that are effective is usually quite narrow, spanning only a two- to threefold difference. This determination is complicated by the fact that a-sarcin is inhibited by the very cations that are frequently necessary to stabilize protein-nucleic acid complexes3 ,7 The enzyme is inhibited by monovalent cation in excess of 100 raM, and by divalent cation in excess of 2 mM. This inhibition can be overcome to some extent by increasing the concentration ofa-sarcin. For example, the digestion ofE. coli 5S rRNA in 50 m M Tris-HC1 alone required only 0.8/tM ot-sarcin, whereas in 50 m M Tris-HCl, pH 7.6, 95 m M KC1, 2 m M MgC12, 4 to 8 pal//ct-sarcin was needed (Fig. 1, compare lanes 4, 6, 7). The higher concentrations of a-sarcin do not alter the specificity of the nuclease. It is imperative that most of the nucleic acid be associated with the specific binding protein in order to minimize background interference. We have satisfied this requirement in two ways. In the first, the RNA was carefully renatured and the RNP particles were reconstituted in the presence of concentrations of cations optimal for binding and with a 6- to 20-fold molar excess of protein over RNA to ensure efficient complex formation, i.e., the absence of free RNA. The samples were then diluted to decrease the cation concentrations to levels just sufficient to maintain the stability of the complexes, but low enough to allow efficient hydrolysis by a-sarcin. The second procedure was used where reconstitution was not required. The stable RNP complexes were separated from free nucleic acid by preparative electrophoresis through cylindrical, nondenaturing polyacrylamide gels. s,9 Elution was into a small volume ( > P h e ~ is reasonably well fulfilled.

T h e Split-Factor Titration Here the proofreading of correct and incorrect ternary complexes is investigated separately. When Jw is equal to zero, Eq. (31) reduces to a relation for correct substrate only

f~-- TuokD/J-~

(33)

With no correct Phe flow in the system (J~- = 0), Eq. (31) simplifies to a relation solely for incorrect proofreading fw = (1 -

Jw/J+w)TUokD/J~,

(34)

Experimental conditions corresponding to Eq. (33) are easy to accomplish. To obtain reliable estimates Offw is more demanding. One reason is that in order to get hot TCA-precipitable chains, more than five incorrect aminoacids must be present in the nascent chain. For substrates with high values off,, it might take a very long time to obtain oligopeptide chains that are long enough to precipitate. One way to solve this problem is to use "primer" chains of the correct amino acid. The ribosomes are started with nonlabeled chains of Phe which are long enough so that precipitability is already achieved for the first missense amino acid. If there is a limited amount of correct amino acid when the incubation starts, this will rapidly be consumed in a transient poly(Phe) synthesis. After this Phe burst, only missense amino acids are left andfw can subsequently be determined from the slope of a plot of Jw versus the input concentration (Tuo) of EF-Tu. The split-factor titration is technically simpler than the Phe-Leu plot described above.

630

RIBOSOME FUNCTION AND KINETICS

[42]

The requirements for long messengers are not as critical here. There are, however, objections that can be raised against the split-factor method (see Ehrenberg et al.2°). First, the proofreading parameter for incorrect substrate (Jew)is studied in a missense context where the neighboring peptidyl-tRNA in the P-site is mismatched to its UUU-codon. The correct proofreading parameter (f~) is, in contrast, measured in the context of a properly matched peptidyl-tRNAph~in the P-site. In the Phe-Leu plot, both correct (fe) and incorrect (f~) proofreading parameters are jointly measured in a predominantly sense context, with a correctly matched peptidyltRNA l'h~in the P-site. Here the correct Phe flow always dominates over the flow of incorrect amino acid into the polypeptide. 2°,2+ Second, in split-factor titrations there is a covariation of ternary complex and T u - G D P . Since T u - G D P can also stimulate peptide bond formation, the proofreading parameters (f~ or f,,) may be underestimated. 31 In the Phe-Leu plot the amount of T u - G D P is approximately constant throughout the whole Phe-synthetase titration. Any contribution to the missense incorporation stimulated by T u - G D P is therefore also constant and can be subtracted as background. More detailed investigations of the Tu-GDP-stimulated flow into polypeptide31 indicate that this backflow is relatively small. Careful comparisons between F factor estimates obtained from PheLeu plots and split-factor titrations reveal no difference between the two methods for the tRNA~" and tRNA+L~ isoacceptors. 2° Conditions of Split-Factor Assay The 70S mixture (A) is prepared as in the Phe-Leu plot with one exception. The concentration of EF-Tu is not kept constant but is varied. The modifications in the factor mix (B) in relation to the Phe-Leu plot are as follows. The ratio between the number of phenylalanines in B and the number of active ribosomes in A is adjusted to roughly 15. For 30 pmol of active ribosomes in A, 450 pmol of phenylalanine is typically used in B. About 80% of these phenylalanines appear finally as poly(Phe) and this corresponds to an average primer length of 12 phenylalanines. It is important that the primer length is well above 5 arninoacids per ribosome, since this is the limit of hot TCA precipitability. It is also important that the primers are not too long in relation to the number of triplets in poly(U). Too-long primers lead to a reduction of the number of ribosomes which can elongate incorrect amino acids after the initial Phe burst.

31A.-M. Rojas, manuscript in preparation.

[43]

TEMPLATE-FREE SYNTHESIS OF POLYPEPTIDES

631

The concentration of Phe-synthetase must be kept sufficiently high to allow rapid primer formation. As in the Phe-Leu plot described above, the concentration of tRNA 1~ must be well above the total input concentration of ribosomes. In the absence of EF-Ts the amount of EF-Tu is varied in the range of 40-300 pmol when the incubation time is 15 min and the total Phe primer synthesis 400 amino acids. In the absence of EF-Ts 50 pmol of EF-Tu can give a maximum of about 500 pmol ofPhe peptide bonds in 15 minutes [50 (pmol × 0.011 (sec-l) X 900 (sec)]. At the lowest EF-Tu concentration, therefore, the amount of EF-Tu is just enough for the primer synthesis to reach completion. If 14C-Phe is used for the primers and the incorrect amino acid is 3H-labeled, a proper correction for the influence of the primer synthesis on the incorrect amino acid incorporation can be made from the known amount of incorporated correct amino acid. 2° The range of EF-Tu used plus EF-Ts is generally smaller than the range in the presence of EF-Ts. This makes the concentration range of ternary complex more similar in the two cases.

[43] R i b o s o m a l S y n t h e s i s o f P o l y p e p t i d e s f r o m Aminoacyl-tRNA without Polynucleotide Template

By A. S. SPIRIN, N. V. BELITSINA,and G. Z. YUSUPOVA(TNALINA) It has been demonstrated that the ribosomes ofEscherichia coli can use certain aminoacyl-tRNAs for polypeptide synthesis without a polynucleotide template.l-3 The best substrate for the ribosomal template-free peptide synthesis was lysyl-tRNA. I-4 It has been shown that it is the structure of the tRNA and not the nature of the amino acid residue that determines the ability of the aminoacyl-tRNAL~ to participate in peptide elongation on ribosomes.5 Other aminoacyl-tRNAs that were studied as substrates for the ribosomal peptide synthesis in the absence of the template include seryltRNA, threonyl-tRNA, and aspartyl-tRNA.2,3 Template-free synthesis of N. V. Belitsina, G. Z. Tnalina, a n d A . S. Spirin, FEBSLett. 131, 289 (1981). 2 G. Z. Tnalina, N. V, Bclitsina, and A. S. Spifin, Dokl. Akad. Nauk S.S.S.R. 266, 741 (1982). 3 N. V. Belitsina, G. Z. Tnalina, and A. S. Spirin, BioSystems 15, 233 (1982). 4 G. Z. Yusupova (Tnalina), Y. L. Remme, N. V. Belitsina, and A. S. Spirin, Dokl. Akad. Nauk S.S.S.R. 286, 725 (1986). s G. Z. Yusupova (Tnalina), N. V. Belitsina, and A. S. Spirin, FEBSLett. 206, 142 (1986). METHODS IN ENZYMOL(~Y, VOL. 164

English translationcopyright © 1988 by Academic Press, Inc.

[43]


631

The concentration of Phe-synthetase must be kept sufficiently high to allow rapid primer formation. As in the Phe-Leu plot described above, the concentration of tRNA 1~ must be well above the total input concentration of ribosomes. In the absence of EF-Ts the amount of EF-Tu is varied in the range of 40-300 pmol when the incubation time is 15 min and the total Phe primer synthesis 400 amino acids. In the absence of EF-Ts 50 pmol of EF-Tu can give a maximum of about 500 pmol ofPhe peptide bonds in 15 minutes [50 (pmol × 0.011 (sec-l) X 900 (sec)]. At the lowest EF-Tu concentration, therefore, the amount of EF-Tu is just enough for the primer synthesis to reach completion. If 14C-Phe is used for the primers and the incorrect amino acid is 3H-labeled, a proper correction for the influence of the primer synthesis on the incorrect amino acid incorporation can be made from the known amount of incorporated correct amino acid. 2° The range of EF-Tu used plus EF-Ts is generally smaller than the range in the presence of EF-Ts. This makes the concentration range of ternary complex more similar in the two cases.

[43] R i b o s o m a l S y n t h e s i s o f P o l y p e p t i d e s f r o m Aminoacyl-tRNA without Polynucleotide Template

By A. S. SPIRIN, N. V. BELITSINA,and G. Z. YUSUPOVA(TNALINA) It has been demonstrated that the ribosomes ofEscherichia coli can use certain aminoacyl-tRNAs for polypeptide synthesis without a polynucleotide template.l-3 The best substrate for the ribosomal template-free peptide synthesis was lysyl-tRNA. I-4 It has been shown that it is the structure of the tRNA and not the nature of the amino acid residue that determines the ability of the aminoacyl-tRNAL~ to participate in peptide elongation on ribosomes.5 Other aminoacyl-tRNAs that were studied as substrates for the ribosomal peptide synthesis in the absence of the template include seryltRNA, threonyl-tRNA, and aspartyl-tRNA.2,3 Template-free synthesis of N. V. Belitsina, G. Z. Tnalina, a n d A . S. Spirin, FEBSLett. 131, 289 (1981). 2 G. Z. Tnalina, N. V, Bclitsina, and A. S. Spifin, Dokl. Akad. Nauk S.S.S.R. 266, 741 (1982). 3 N. V. Belitsina, G. Z. Tnalina, and A. S. Spirin, BioSystems 15, 233 (1982). 4 G. Z. Yusupova (Tnalina), Y. L. Remme, N. V. Belitsina, and A. S. Spirin, Dokl. Akad. Nauk S.S.S.R. 286, 725 (1986). s G. Z. Yusupova (Tnalina), N. V. Belitsina, and A. S. Spirin, FEBSLett. 206, 142 (1986). METHODS IN ENZYMOL(~Y, VOL. 164

English translationcopyright © 1988 by Academic Press, Inc.

632


[43]

polypeptides was strongly dependent on the two elongation factors (EF-Tu and EF-G) and GTP) -3 Materials Escherichia coli MRE 600 ribosomes are washed four times with 1 M purified ribosomes are stored in a buffer containing 20 mM, Tris-HCl, 100 mM NH4C1, l0 mM MgCI2, 0.1 mM ethylenediaminetetraaceticacid (EDTA), and 10% glycerol, pH 7.6 (at 37 °) at --70 °. The purified elongation factors, EF-Tu and EF-G, are also isolated from E. coli MRE 600. 8,9 Preparations of total E. coli tRNA (Boehringer-Mannheim), aminoacylated enzymatically with one of ~4C-labeledamino acids, t° are stored at 4* in the lyophilized state. Specific activities of the t4C-labeled amino acids and the ~4C-labeledaminoacyl-tRNAs used are given in Table I. Individual [t4C]lysyl-tRNAL~ is prepared from the total E. coli tRNA acylated with [t4C]lysine by the procedure of affinity chromatography on immobilized EF-Tu of Thermus thermophilus HB8 H (EF-Tu we used was a gift from Dr. M. Garber, Institute of Protein Research, Pushchino); specific activity of [14C]lysyl-tRNAL~ is 1000- 1100 pmol [~4C]lysine/A26o unit. Individual [3H]phenylalanyl-tRNATM is prepared by misacylation of the tRNA TM from E. coli with [3H]phenylalanine (Amersham, 50 Ci/ mmol) using phenylalanyl-tRNA synthetase (ligase) from yeast t2 (this enzyme was a gift from Dr. P. Remy, Institute of Molecular and Cellular Biology, Strasbourg, France); specific activity of [3H]phenylalanyl-tRNAL~ is 750 pmol [3H]phenylalanine/A26o unit. The samples of purified individual aminoacyl-tRNAs are stored in l0 mM NaCH3COO buffer, pH 4.5, at -70*. NH4C1. 6'7 The

Reagents

CM-cellulose (Whatman) Poly(U) (Calbiochem) Poly(A) (Calbiochem)

S. Pestka, J. Biol. Chem. 243, 2810 (1968). 7 R. W. Erbe, M. M. Nau, and P. Leder, J. Mol. Biol. 39, 441 (1969). 8 K. Arai, M. Kawakita, and Y. Kaziro, J. Biol. Chem. 247, 7029 (1972). 9 y. Kaziro, N. Ynoue-Yokosawa, and M. Kawakita, J. Biochem. (Tokyo) 72, 853 (1972). 10L. P. Gavrilova and V. V. Smolyaninov, Mol. Biol. (USSR) 5, 883 (1971). t~ W. Fischer, K.-H. Derwenskus, and M. Sprinzl, Eur. J. Biochem. 125, 143 (1982). t2 j. Wagner and M. Sprinzl, Eur. J. Biochem. 108, 213 (1980).

[43]


633

TABLE 1 SPECIFIC ACTIVITIESOF 14C=LABELEDAMINO ACIDS AND t4C-LABELED AMINOACYL-tRNAs

Amino acid Lysine Arginin~ Glutamic acid Aspartic acida Glutamine Asparaglne Threonine Serine Glycine Methionine Phenylalanine Proline Isoleucine Leucine Valine Alanine

Specific activity of 14C-labeled amino acid (Ci/mol)

[~'C] Aminoacyl-tRNA content per 1 mg of total tRNA (pmol)

336 210 285 140 48 151 228 170 l 18 285 486 280 354 342 280 171

1481 2597 1027 1769 1349 1295 1237 1037 1494 2056 954 813 1943 2056 1737 1276

a "C-Labeled amino acids were from the Institute for Research, Production and Uses of Radioisotopes, Prague, Czechoslovakia; other m~-Iabeled amino acids were from Amersham, England.

Poly(C) (Department for Production of Biologically Active Substances, Novosibirsk, USSR) GTP, Nasalt (Fluka) Tetracycline, chloramphenicol (Department for Production of Biologically Active Substances, Novosibirsk, USSR) Phosphoenolpyruvate (Fluka) Phosphoenolpyruvate kinase (Boehringer-Mannheim) Guanyl-5'-ylmethylene diphosphonate (synthesized by Dr. K. K. Zikherman, Institute of Protein Research, Pushchino) Fusidic acid (Sigma) Phenylboric acid (synthesized by Dr. K. K. Zikherman, Institute of Protein Research, Pushchino) Glass filters GF/F (Whatman) Nitrocellulose filters Synpor No. 6 (Chemapol)

634


[43]

POPOP (Serva) PPO (Serva) Bovine serum albumin (Reanal) Triton X- 100 (Serva) Buffers

Standard buffer for cell-free elongation systems: 10 m M or 12 m M MgCI2, 100 m M NH4CI, 0.1 m M EDTA, 1 m M dithiothreitol, 20 m M Tris-HC1, pH 7.6 (at 37*). In studies of the dependence of polypeptide synthesis on Mg2+ concentrations, MgC12 in the buffer varies from 5 to 20 mM.

Solutions 3 m M pyridine-Acetate buffer, pH 5.2 1 M pyridine-acetate buffer, pH 5.2 Phenol Toluene 96% ethanol Trichloroacetic acid (TCA), 30% and 5% 5% TCA with 0.25% Na2WO4, pH 2.0 NaOH, 1 M and 0.2 M CH3COOH, 1 M Triton X- 100 (Serva) Template-Free Ribosomal Synthesis of Polylysine from Lysyl-tRNA

Use of Total tRNA Aminoacylated with Lysine The mixture containing 20 pmol of ribosomes, 100/~g of total tRNA aminoacylated with [14C]lysine (148 pmol of [14C]lysyl-tRNA), 170 pmol of EF-Tu, 37 pmol of EF-G, 8 - 16 nmol of GTP, 100 nmol of phosphoenolpyruvate, and 1/tg of phosphoenolpyruvate kinase in 50/tl of the standard buffer is incubated at 37 °. The reaction is stopped by adding 2 ml of 5% trichloroacetic acid with 0.25% Na2WO41~; 100 #g of bovine serum albumin is added as a carrier and the suspension is hydrolyzed at 90* for 20 min. The hot acid-insoluble precipitates are collected on GF/F glass filters and their radioaetivities are measured in the standard toluenePPO-POPOP mixture using the Beckman LS-100 or LS-9800 seintillaiton 13 R. S. Gardner, A. J. Wahba, C. Basilio, R. S. Miller, P. Lengyel, and J. F. Speyer, Proc. Natl. Acad, Sci. U.S.A. 48, 2087 (1962).

[43]

TEMPLATE-FREE SYNTHESIS OF POLYPEPTIDES I

635

!

COMPLETE E t~

/

N

E O

_J

Z/

'-5' ,1" 0

-EF-G 10

20

Time, min

FIG. 1. Kinetics of [14C]lysineincorporation into TCA-Na2WO4-insoluble product in the template-free ribosomal system. (0) Complete: +EF-Tu, +EF-G, -I-GTP; ~1) +EF-Tu, -EF-G, +GTP. Incubation at 37 °, 10 mM MgCl2.

spectrometer. Under these conditions dilysines are not precipitated and short oligolysines are not precipitated quantitatively.~4 About 3 - 4 pmol of [~4C]lysine is incorporated into the TCA-Na2WO4-insoluble product after l 0 - 2 0 min of incubation under the conditions described. The kinetic curve of poly[~4C]lysine synthesis is represented in Fig. l as an example. 'Fable II shows that the template-free system under investigation is completely dependent on ribosomes. Two elongation factors, EF-Tu and EF-G, are strictly required for the template-free polylysine synthesis, suggesting the participation of enzymatic binding of lysyl-tRNA and EF-Gpromoted translocation of oligolysyl-tRNA in the process, just as in the normal translation system. The ribosomal template-free system of polypeptide synthesis depends on temperature and stops in the cold. It is noteworthy that this system is much more dependent on GTP regeneration than the usual template-dependent translation system: exclusion of phosphoenolpyruvate from the system blocks polylysine synthesis. This could be an indication that EF-Tu-promoted aminoacyl-tRNA binding to ribosomes without template polynucleotide is less efficient, and more molecules of GTP are expended per one molecule of the aminoacyl-tRNA bound, as compared with template-dependent binding (the :same situation 14M. A. Smith and M. A. Stahmann, Biochem. Biophys. Res. Commun. 13, 251 (1963).

636

[43]

RIBOSOME FUNCTION AND KINETICS TABLE II POLY[14C]LYSINE SYNTHESIS IN THE TEMPLATE-FREE RIBOSOMAL SYSTEMa

No. 1

2 3

a

System of polypff2]lysine synthesis

[14C]Lysine, polymerized (pmol)

Complete, 37 ° -EF-G -EF-Tu -Ribosomes Complete, 37 ° -Phosphoenolpyruvate Complete, 37 ° Complete, 4 °

2.8 0.5 0.2 0.05 3.9 0.45 4.2 0.3

Incubation for 20 rain, 10 m M MgCl2.

is seemingly observed when ribosomes carrying polynueleotide templates bind noncognate aminoacyl-tRNAs 15). Table III shows that polylysine synthesis in the template-free system is very strongly inhibited by all of the same agents which specifically inhibit the natural process of translation: tetracycline, chloramphenicol, phenylTABLE III EFFECT OF SOME INHIBITOR$ ON TEMPLATE-FREE RIBOSOMAL SYNTHESIS OF POLy[laC]LYSlNE

Inhibitor Tetracycline (TC)

Chloramphenicol (CM) Phenylboric acid (PBA) Fusidic acid (FA) Guanyl-5'-ylmethylene diphosphonate (GMPPCP) Poly(U) Poly(C)

Concentration of inhibitor (raM)

Inhibition of polyp4C]lysine synthesis (%)

0.03 0.16 0.25 0.01 0.02 15.0 80.0 0.4 0.4

50.0 83.7 87.5 50.0 86.2 50.0 91.0 95.0 96.8

4.0~ 4.0~

87.2 65.6

a Values given as/zg per 50 #1. ~5D. G. Kakhniashvili, S. K. Smallov, and L. P. Cravrilova, F E B S L e t t . 193, 103 (1986).

[43]


637

boric acid, fusidic acid, and guanyl-5'-ylmethylene diphosphonate. Moreover, it is seen that the template-free system is much more sensitive to some antibiotics, for example, tetracycline, than the template-directed system) ,3 Polylysine synthesis is significantly inhibited by noncognate template polynucleotides, such as poly(U) and poly(C).t The Mg2+-dependence of template-free polylysine synthesis was determined in comparison with poly(A)-directed synthesis by variations of the Mg2+ concentrations in the standard buffer. The results are presented in Fig. 2. The Mg2+ optimum of template-free polylysine synthesis is lower than that of poly(A)-directed synthesis. High Mg2+ concentrations strongly inhibit template-independent synthesis; the latter is completely blocked in the region of 15- 18 mM Mg2+ where the poly(A)-directed synthesis is optimal. It is noteworthy that the curve of the Mg2+ dependence ofpolylysine synthesis in the poly(A)-directed system has a shoulder in the region of low (about 10 mM) Mg2÷ concentration; this can be explained by the concomitant template-free lysine incorporation in the ribosome-synthesized product. Sedimentation analysis of the template-free incubation mixture in the sucrose gradient shows that the ribosomal zone contains newly synthesized

+ Poly(A) -5 E 0,.

N

E 0 Q.

y(A)

._1

(

I

0

I

I

I0

20

[M gZ+], mM FIG. 2. Dependence of poly[t4C]lysine synthesis on Mg2+ concentration. (A) Poly(A)-directed system [+20 gg poly(A)]: (0) poly(A)-independent (template-free) system. Incubation at 37 ° for 30 min.

638

[43]


0.15 70S

-6 E 0.10

N

_E

E

O

,-~---, . . . . . i ~

I

u :I

i o O,y .

_.a

J 12

......

C

•

2

I 8

L

4

~

8

!

moLor rotio Ac ItS'C] Phe-tRNA: ?0S

FIo. 2. Saturation of tight couples under P-site (A) and A-site conditions (B); (0) Binding curves. The inserts show the relative site location of AcPhe-tRNA as measured by the puromycin reaction (30 rain at 0°). (m) P-site location; (A) A-site location. Aliquots contained 13.8 pmol 70S.

A 1.0 .o.--. o

~:. 0.5

215

50

8

z B '~ 1.0

,=o

n _ _

0.5 I

50 time [ h ]

100

FIo. 3. Quantitative puromycin reaction of A-site bound (A) and P-site bound (B) AcPhe-tRNA (ratio of added AcPhe-tRNA to 70S = 8 and 3, respectively). (O) binding of AcPhe,tRNA; (A) puromycin reaction. After the binding reaction bad been performed, one half of the sample was used for the puromycin reaction (incubation time at 0 ° as indicated), the other half was kept at 0 ° and the binding controlled at the indicated times. Aliquots contained 6.15 pmol 70S. From Geigenmfiller et aL'3

PREPARATION OF ACTIVE E. coli RIBOSOMES

[45] A

669

A

1.0 "0 tO

Z r~

.,,..

0.5

l-

t%

< I

B >

I

1 2 m t i o A c [ l ~ C ] P h e - t R N A :70S

I

3

l.O

t-

< 7

' 05 o IE ~J

_.--.O

I

2

4

8 14 ratio f[l~ C ] M e t - t R N A : 7 0 5

2O

FIG. 4. (A) AcPhe-tRNA binding to poly(U)-programmed 70S ribosomes. Aliquots conrained 8.9 pmol 70S ribosomes. (O) Ac[=4C]Phc-tRNA (1400 pmol per A260unit) pufitied by BD-cellulos¢ chromatography as described in this paper. (O) Ac[=4C]Phe-tRNA (1750 pmol per A2~ounit) purified according to Odom et a l l 4 on an HPLC column. (B) Initiation-factor dependent binding of i~14C]Met-tRNAfM~t to 70S ribosomes in the presence of MS2 RNA. One aliquot (100/~1) contained 12pmol 70S, 144pmol MS2 RNA, 12 to 240pmol f[~4C]Met-tRNAfM= (150cpm/pmol), and 5/zl of a crude initiation-factor preparation (142 A2ao/ml) where indicated. The ionic conditions were 12 mM Mg2+, 50 mM NH + and 30 mM K+. The MS2-directed translation system has been described in Funatsu et a l ) 5 f[~4C]Met-tRNAfM= has been prepared according to Kahn et al.~6 The preparation of initiation factors was according to Noll et al.? 7 except that the final precipitation step was omitted. Initiation factors were stored in a buffer containing 20 mM HEPES.KOH, pH 7.5, 1 mM EDTA, 0.5 mM DTE, 10 mM NI-I4CI and 10% glycerol. (O) binding in the absence of initiation factors, and (e) binding in the presence of initiation factors.

670


[45]

location (percentage of P-site and A-site bound material) equivalent to a quantitative reaction is obtained also with a short incubation time, if the precautions listed above are observed.

Results of the Activity Test Site-Specific Saturation. Figure 2 shows site-specific saturation of 70S tight couples with Ac[14C]Phe-tRNA. Under P-site conditions (Fig. 2A), exactly one Ac[~4C]Phe-tRNA molecule per 70S can be bound, whereas under A-site binding conditions (P site blocked with deacylated tRNA, Fig. 2B) the saturation level is somewhat lower (0.75 for the concentration range tested). The site specificity of the bound Ac[~4C]Phe-tRNA is better than 80% over the whole concentration range (see the puromycin reaction depicted in the inserts of Figs. 2A and B). Figure 3 demonstrates that all material bound under P-site conditions can react with puromycin, provided the incubation time is long enough (Fig. 3A), whereas A-site bound material remains essentially puromycin-insensitive during the same time of incubation at 0 ° (Fig. 3B). For comparison, AcPhe-tRNA was purified on a HPLC column 14 (instead of by BD-column chromatography as described in this paper) yielding AcPhe-tRNA essentially free of contaminating deacylated tRNA (1750 pmol/A26o versus 1400 pmol/A260 of the BD-purified AcPhe-tRNA, see Fig. 4A). The same saturation level of about one AcPhe-tRNA per poly(U)-programmed 70S ribosome is obtained. This finding indicates that the content of contaminating deacylated tRNA (at least up to a contamination of 20%) does not affect the saturation level of AcPhe-tRNA in accordance with the exclusion principle for AcPhe-tRNA binding. ~3 The initiation-factor dependent binding of fMet-tRNA to 70S ribosomes in the presence of the natural mRNA MS2 also levels off at 0.9 fMet-tRNA per 70S ribosome (Fig. 4B). 15-17 The binding results clearly demonstrate that 90- 100% of the ribosomes prepared observing the precautions described here participate in tRNA binding to the P site.

~40. W. Odom, H. Y. Deng, and B. Hardesty, this volume, [11]. ~5G. Funatsu, K. H. Nierhaus, and B. Wittmann-Liebold, J. Mol. Biol. 64, 201 (1972). ~6D. Kahn, M. Fromant, G. Fayat, P. Dessen, and S. Blanquet, Eur. J. Biochem. 105, 489 (1980). ~7M. Noll, B. Hapke, M. H. Schreier, and H. Noll, J. Mol. Biol. 75, 281 (1973).

[46]

ANTIBIOTIC RESISTANCE MUTATIONS

673

[46] A n t i b i o t i c R e s i s t a n c e M u t a t i o n s in R i b o s o m a l R N A G e n e s o f E s c h e r i c h i a coli B y CURT

D. SIGMUND, MOHAMED ETTAYEBI, ANGELA BORDEN, and EDWARD A. MORGAN

Dominant or codominant selectable mutations in rRNA genes can identify regions of rRNA involved in specific ribosome functions, provide useful genetic tools for the study of rRNA structure and regulation, and help determine how important antibiotics work. However, selectable mutations in rRNA genes of Escherichia coli had eluded identification until recently because mutations in only one of the seven E. coli rRNA operons provide at best very weak phenotypes. I Our experience with antibiotic resistance mutations in rRNA genes suggests that strong, stable phenotypes are obtained only when approximately 50% of the rRNA in ribosomes is of the mutant type. This requirement can be fulfilled when the mutations are isolated in an rRNA operon on a multicopy plasmid.l-6 In this chapter we outline our strategies for selecting mutants. We describe procedures to map the mutations to small regions of rRNA operons, facilitating identification of the mutations by DNA sequencing. We also describe a primer extension method using rRNA templates and DNA oligonucleotide primers that permits convenient and accurate quantitation of the percent mutant rRNA in an rRNA preparation. This primer extension method is useful for determining the structural and functional capabilities of ribosomes synthesized from a single mutant rRNA operon (e.g., by analysis of total cellular ribosomes or ribosomes on polysomes) and provides a convenient way to determine if newly isolated mutations have base changes at the same position as previously isolated mutations. We also describe how a similar primer extension method can be used to accurately determine the copy number of mutant rRNA genes (e.g., plasmid copy number), permitting reliable quantitative analysis of rRNA gene expression. 1C. D. Sigmund, M. Ettayebi, S. M. Prasad, B. M. Flatow, and E. A. Morgan, in "Sequence Specificity in Transcription and Translation" (R. Calender and L. Gold, eds.), p. 409. Liss, New York, 1985. 2 C. D. Sigmund and E. A. Morgan, Proc. Natl. Acad. Sci. U.S.A. 79, 5602 (1982). 3 L. G. Mark, C. D. Sigmund, and E. A. Morgan, J. Bacteriol. 155, 989 (1983). 4 C. D. Sigmund, M. Ettayebi, and E. A. Morgan, Nucleic Acids Res. 12, 4653 (1984). 5 M. Ettaybei, S. M. Prasad, and E. A. Morgan, J. Bacteriol. 162, 551 (1985). 6 S. Douthwaite, J. B. Prince, and H. F. Noller, Proc. Natl. Acad. Sci. U.S.A. 82, 8330 (1985). METHODS IN ENZYMOLOGY, V O L 164

Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form ret~rved.

674

GENETICS

[46]

Materials Media

For selection of antibiotic resistance mutations we have arbitrarily used LB medium, which consists of 10 g tryptone, 5 g yeast extract, and 5 g NaC1 per liter. For LB agar, 15 g agar is added per liter. Media differences can result in substantial differences in antibiotic sensitivity. Selection of other types of mutants may of course require different media. Colicin E l

Colicin E 1 is a bactericidal protein made by E. coli strains containing the CoE 1 plasmid. Colicin E 1 can be used for selection of plasmids conmining an rRNA operon in a CoE 1 vector because Cole 1 vectors confer immunity to colicin E 1. Colicin E 1 is slightly more difficult to obtain and use than conventional antibiotics, but its use soon becomes routine and problem-free. Colicin E 1 is produced in our laboratory by mitomyein C induction of W3110(ColE1), followed by a simple partial purification scheme involving salt extraction of a cell pellet and ammonium sulfate precipitation. 7 The ammonium sulfate pellet is resuspended in 0.1 M potassium phosphate, pH 7.4, 50% glycerol, and stored over chloroform at - 2 0 °. Typically, 20 ml of colicin E 1 is prepared from 24 liters of culture. Colicin E 1 can also be obtained from Sigma Chemical Company. The amount of coliein E 1 to use is determined by putting a measured drop of colicin E 1 on the agar surface of a petri dish and rapidly mixing in 0.1 ml of an overnight culture of one of the bacterial strains used in the experiment, then spreading the mixture over the surface of the dish. When the appropriate amount of colicin El is used, less than 100 colonies are obtained, and these test as colicin E1 resistant when cross-streaked with colicin E1 and the phage BF23. A resistant mutant lacks the receptor shared by colicin E1 and BF23, a sensitive cell is killed by both agents, and an immune cell (which produces an immunity protein from a CoE1 plasmid) is immune to eolicin E1 but sensitive to BF23. Typically, 75 #1 of colicin E 1 prepared in our laboratory as described above is used per petri dish. Large excesses of colicin E l should not be used, as immunity protein can be overwhelmed by excess colicin E 1. Plasmids

All plasmids used in our procedures are described in Fig. 1. We have successfully isolated mutations in rRNA genes in vivo using the plasmid 7 S. A. Schwartz and D, R. Helinski, J. Biol. Chem. 20, 6318 (1971).

[46]


675

I lkbl

Pvu I

Ps! I

C

pBR322

Hint lr

t I

Hind= ~ S i n a i / " I "Who I CIo I

pACYC~77

Eco RI Pvu Sca I ~ / ~ r~

~

polylinker

~V~ ~ (O~ i" pUCe) EGO RI

Nru Z_ _ _ ~ ' - - - H m d Wl ~ / I I "Born HI Xma If SolI Sph I

pACYCt84

pME 2t

FIG. 1. Plasmid structures. In these drawings, shaded regions represent rRNA operons, white regions indicate flanking E. coli DNA, and black regions indicate vector sequences. pLC7-21 contains rrnH cloned in a ColE1 vector, pRR-1 contains rrnC cloned in a portion of pBR322. The rrnC operon of pRR-1 was derived from pLC22-36, pLC7-21 and pLC22-36 have been previously described [M. E. Kenerley, E. A. Morgan, L. Post, L. Lindahl, and M. Nomura, J. Bacteriol. 132, 931 (1977)]. Small internal segments of the 16S and 23S rRNA genes of pRR- l have been replaced by gene segments of rrnH that include the spectinomycin and erythromycin resistance mutations present in pSPC-I and pERY-1. 4 pACYC177 and pACYC 184 have been previously described [A. C. Y. Chang and S. N. Cohen. J. Bacteriol. 134, 1141 ( 1978)]. pME21 is an approximately 1500-bp HaeIII origin-containing fragment of pRK-6 [D. M. Stalker, R. Kolter, and D. R. Helinski, J. Mol. Biol. 161, 33 (1982)] ligated to a 1345-bp RsaI fragment that contains the alpha-complementing lacZ fragment and polylinker of pUC8 [C. Yanisch-Perron, J. Vieira, and J. Messing, Gene 33, 103 (1985)]. A HincII cartridge containing a chloramphenicol resistance gene [T. J. Close and R. L. Rodriquez, Gene 20, 305 (1982)] was then ligated into a PvuII site within a region of lacZ not required for alpha complementation, completing the construction of pME21.

p L C 7 - 2 1 , w h i c h c o n t a i n s a c o m p l e t e r r n H operon. 2 This p l a s m i d can be mobilized b y the E . coli F fertility factor, a p r o p e r t y essential to the success o f o u r procedures. A n o t h e r favorable p r o p e r t y o f p L C 7 - 2 1 is its nearly ideal c o p y n u m b e r (nearly as great as a n y available p l a s m i d c o n t a i n i n g a n rrn o p e r o n , b u t n o t so high as to reduce cell g r o w t h to a n extent where faster growing derivatives c o n t a i n i n g deletion plasmids rapidly a c c u m u late). U n f a v o r a b l e aspects o f the p l a s m i d are its s o m e w h a t large size a n d the fact that it can o n l y be selected using colicin E 1. H o w e v e r , the use o f

676

GENETICS

[46]

colicin E 1 does have the significant advantage of not directly affecting the ribosome. All ribosome-active antibiotics we have tested seem to act synergistically with other ribosome-active antibiotics, making the design of experiments difficult. Of the antibiotic resistance genes commonly present in plasmids, only ampicillin resistance seems to be a suitable alternative to colicin E1 immunity.

Bacterial Strains Donor bacterial strains should not have chromosomal mutations affecting the ribosome (because these mutations might interfere with phenotypic expression of the desired rRNA mutation), be F + so they can mobilize the plasmid with rRNA genes, be RecA ÷ so that they can be mutagenized, and not have lysogenic prophages lacked by the recipient strain. We have used EM2(F, pLC7-21). EM2 is ilv-1, his-29, pro-2, tsx, trpA-9605, trpR, ara. Recipient strains must be F - and RecA+ so they can be efficient recipients in the mating, and must contain a nonribosomal mutation with a low forward mutation rate to allow selection against the donor strain after mating events. We have used W3110 Nal r, a spontaneous mutant of the prototrophic strain W3110 that is resistant to 100/tg/ml of nalidixic acid. Nalidixic acid-resistant derivatives can be obtained from most strains by plating the cells from 10 ml of overnight culture onto LB agar containing 100/~g/ml of nalidixic acid. Several precautions are advised due to the instability of large recombinant plasmids containing rRNA operons. Culture stocks should be stored at - 70 ° after addition of glycerol to 20%. To prevent the accumulation of deletion and mutant plasmids, experiments should always begin from single-colony isolates obtained by streaking out scrapings from frozen stocks.

M u t a n t Isolation and Characterization We describe below a procedure for the isolation of erythromycin resistance mutations in rRNA genes on pLC7-21. Careful modification of the procedure allows other types of mutations to be isolated. Our procedure has many details that have explicit purposes and are essential to the success of the procedure. These details can best be understood if it is recognized that the procedure must: (1) enable the isolation of mutants that are usually infrequent, (2) enable the isolation of mutants with weak phenotypes, (3) permit phenotypic expression of mutations before strong selective pressure is applied (phenotypic expression requires several cell generations because of the nature of the mutations), and (4) cause the enrichment

[46]


677

of mutations on a plasmid relative to mutations on the chromosome (a feature that is desirable due to the fact that unwanted chromosomal mutations can be much more frequent than plasmid rRNA gene mutations).

Antibiotic Sensitivity Levels The use of antibiotic concentrations near the threshold sensitivity level can assist in the enrichment for mutations and allow recovery of mutants with weak phenotypes. Therefore, it is necessary to carefully test the antibiotic sensitivity of both donor [EM2(F, pLC7-21)] and recipient (W3110 Nal r) strains by plating 0.1 ml of overnight cultures onto different, closely spaced, concentrations of antibiotic. Three concentrations of antibiotic are chosen for subsequent use. The "low" concentration is a concentration that is only partially inhibitory to cell growth, and is the lowest concentration that permits mutant bacteria to form detectable papillae visible on the lawn of partially inhibited cells. For erythromycin, the low concentration is 100/zg/ml, and 1,000 to 10,000 papillae are observed per petri dish. Most of these papillae are the result of cell envelope mutants resistant to this low concentration of erythromycin. An "intermediate" concentration of antibiotic is also chosen, and is a concentration that allows detectable growth of the background lawn of wild-type bacteria, but inhibits lawn formation sufficiently to enable mutant bacteria to form prominent, easily recognized colonies. For erythromycin, this concentration is 200 gg/ml and there are 100 to 1,000 prominent colonies on the weak lawn. A "high" concentration of antibiotic is also chosen, and is a concentration just sufficient to completely prevent growth of the background lawn. For erythromycin, this concentration is 300 gg/ml, and from 0 to 10 mutant colonies appear per petri dish, most or all of which have mutations affecting ribosomal components.

Mutagenesis Portions (0.2 ml) of overnight cultures of EM2(F, pLC7-21) are plated on petri dishes containing colicin E 1 and low and intermediate concentrations of erythromycin. Only the low and intermediate concentrations are used at this stage because the high concentration does not allow sufficient growth for phenotypic expression of new mutations. Immediately after spreading bacteria on the plate, a l-cm diameter filter paper disk saturated with mutagen is placed in the center of the dish. We have used 100/d of ethylmethane sulfonate per dish. Because individual mutagens cause mutations at preferred sites, it may be advantageous in some instances to use other mutagens. Optimal amounts of mutagen should cause a 1 to 2 cm zone of killing around the filter paper disk. We mutagenize on a petri dish

678

GENETICS

[46]

only because this method has been convenient and effective. Mutagenesis of suspensions of cells just prior to plating on selective agent is probably at least as effective when selecting mutants resistant to bacteristatic antibiotics. Mutagenesis in the presence of the "low" or "intermediate" concentrations of a bacteristatic antibiotic or mutagenesis immediately followed by exposure to "low" or "intermediate" concentrations of antibiotic does not completely prevent cell growth after mutagenesis and therefore does not prevent phenotypic expression of mutations. However, in the presence of "low" or "intermediate" concentrations of antibiotic, mutants grow faster. These procedures therefore enable mutagenesis and enrichment for mutations in a single step. However, when selecting mutants using nutritional selections or bactericidal antibiotics, or when seeking mutants that can only be identified by screening procedures, partially inhibitory concentrations of selective agent cannot be used because they do not exist. In these cases, mutagenesis should b~ followed by five or more cell doublings to allow phenotypic expression of mutations before the addition of selective agents.

Mating The mutagenized cells are then washed off the petri dishes and resuspended at an OD55o of 0.2 in LB medium containing the same concentration of erythromycin used in the mutagenesis plate. After growth overnight in this medium to further enrich for erythromycin resistance mutants, the cells are washed with LB medium several times by centrifugation to remove all traces of antibiotics that might otherwise interfere with the mating. The cells are then resuspended in LB medium to an ODss0 of 0.05, grown to an OD550 of 0.2, and 5 ml of this culture is mixed with 5 ml of a growing broth culture of W3110 NaF (also at an OD550 of 0.2) in a 500-ml Erlenmeyer flask. To permit efficient mating, the resulting thin culture layer is incubated overnight at 37 ° without shaking. The resulting transconjugants must then be allowed to phenotypically express mutations present on newly acquired plasmids. Therefore, the two mating mixtures derived from mutagenesis on low and intermediate erythromycin concentrations are each diluted l : 20 in LB containing 20 #g/ml of nalidixic acid and colicin E l, followed by growth overnight at 37 °. For reasons outlined in the above section on mutagenesis, the agent selective for rRNA mutations should be omitted from the culture if the selective agent is not truly bacteristatic. The amount of colicin E1 per 50 ml of liquid culture should be the amount previously determined to be appropriate to use on a single petri dish. Cells from each of the two resulting cultures are then plated on LB agar containing 20 #g/ml nalidixic acid, colicin E 1, and the low, inter-

[46]


679

mediate, and high concentrations of erythromycin. Controls consisting of unmutagenized overnight broth cultures of W3110 Nal r (pLC7-21) and EM2(F, pLC7-21) are also plated on identical media. In the case of erythromycin resistance, the success of the mutagenesis, enrichment, and mating procedures is obvious, as the cultures obtained after mating yield 10 to 100 times more colonies on the intermediate and high concentrations of erythromycin than either control culture. However, a clear enrichment for plasmid mutants at this stage has not been evident in two other cases where mutations causing resistance to antibiotics were nevertheless identified by the following screening procedures.

Colony Screening Numerous individual mutant colonies growing on all three concentrations of erythromycin are picked and restreaked on LB agar containing nalidixic acid, colicin E 1, and the respective concentration of erythromycin. All size classes of colonies should be picked, as the phenotype of rRNA gene mutations cannot be predicted. After restreaking, 1-ml overnight LB cultures are made from single-colony isolates and are tested by a series of simple streak and cross-streak tests. Transconjugants that pass these tests should test as prototrophic, colicin El immune, and BF23 sensitive, and should grow better than W3110 Nal r (pLC7-21) on at least one erythromycin concentration (when tested on the low, intermediate, and high concentrations of erythromycin). Mutants that pass these tests are candidates for having mutant rRNA genes in pLC7-21, and are then directly tested for erythromycin resistance mutations in pLC7-21 as described below.

Plasmid Screening Fresh l-ml LB broth cultures are grown starting with the l-ml LB cultures described above, which have been saved at room temperature for this purpose. Of each culture, 0.5 ml is used to prepare plasmids by a NaOH minilysate procedure. 8 Care is taken to carefully wash the DNA pellet obtained by the first ethanol precipitation step of the minilysate procedure, but all subsequent steps in this DNA preparation procedure can be omitted. The ethanol precipitate is then dried in vacuo, resuspended in 10/zl of water, and used to transform W3110 Nal', selecting colicin immunity on LB agar containing colicin E1 and nalidixic acid. The resulting colonies are then restreaked on the same media. Single colonies are then used to start 1-ml overnight broth cultures. The broth cultures and control cultures of W3110 Nal r (pLC7-21) are tested by streak tests for colicin 8 H. C. Birnboim and J. Doily, Nucleic Acids Res. 7, 1513 (1979).

680

GENETICS

[46]

immunity, BF23 sensitivity, and for erythromycin resistance on low, medium, and high erythromycin concentrations. Once it is confirmed by this method that a mutant plasmid has been isolated, the remaining portions of the 1-ml cultures used for these tests are used to inoculate 10-ml cultures, from which frozen glycerol stocks are prepared. The frozen stocks are the source of inoculums for all subsequent cultures. The antibiotic sensitivity thresholds of mutants should be carefully tested using the antibiotic used in the selection procedure (erythromycin in this example). In addition, any structurally or functionally similar antibiotics and several dissimilar ribosome-active antibiotics should be tested. It is not unusual to isolate mutations that confer very strong resistance to antibiotics structurally dissimilar to the antibiotic used in the selection procedures.4,5 The knowledge resulting from these sensitivity tests assists subsequent genetic manipulation of the mutations, can often clearly distinguish between mutants with different base changes, and can provide a wealth of useful information on the ways that antibiotics interact with ribosomes and affect ribosome function. 4-6 The key features of our selection procedures are the efficient enrichment of plasmid mutations and proper provisions for phenotypic lag. However, certain mutations are not easily obtained by our procedures because the background of unwanted chromosomal mutants is very high and the available selective regimens do not allow efficient enrichment for plasmid mutants. For example, efficient enrichment is prevented by procedural modifications necessary when selecting mutants resistant to bactericidal antibiotics, when using nutritional selections, when seeking mutations that can be detected by a screening procedure but cannot be selected, or when cells with mutant rRNA operons grow much slower under selective pressure than cells with chromosomal mutations. In the absence of favorable enrichment procedures, larger culture sizes and extensive mutant screening procedures may be required to isolate plasmid mutations. In the absence of enrichment steps, we estimate that upward of 109 cells must in some cases be plated on selective media to obtain rRNA gene mutations in plasmids and that unwanted chromosomal mutants can be over 1000 times more frequent than plasmid rRNA gene mutants. Mapping and Sequencing Mutations The identification of point mutations by sequencing the DNA from large regions of rRNA operons is unnecessarily time-consuming and prone to error, whereas sequencing only regions where the mutation is guessed to be located can lead to unreliable conclusions or failure to identify all sequence alterations required for the phenotFpe. It is therefore advantageous to map mutations to small regions prior to sequencing DNA.

[46]

ANTIBIOTICRESISTANCEMUTATIONS

681

Fragment Exchange The most straightforward method used for mapping mutations to smaller regions is the careful reciprocal exchange of restriction nuclease fragments between pLC7-21 and its mutant derivative. The resolution and difficulty of mapping by this method depends on the distribution of restriction nuclease recognition sites. The most significant problem with this method results from the fact that the starting plasmids and the final plasmids resulting from fragment switching have structures that are indistinguishable when analyzed by restriction nucleases. Because of the nature of recombinant DNA methods, the source of all fragments in the final constructions will therefore be ambiguous unless the switching of fragments proceeds through clonally purified intermediate plasmids (deletions or subclones) that are structurally different from the plasmids obtained in each immediately preceding and subsequent step. Because of the need to construct intermediate plasmids, this mapping method is not as fast as marker rescue mapping methods. However, fragment switching methods are the only mapping method that can be used if the mutation does not confer a strong selectable phenotype.

Marker Rescue If a mutation confers a strong selectable phenotype, mapping can proceed using marker rescue methods. In these methods, restriction nuclease fragments of mutant pLC7-21 derivatives are cloned into a plasmid (pACYC 177, pACYC184, or pME21) that can stably replicate in cells also containing pLC7-21. The cloned restriction nuclease fragments must not express a complete mutant rRNA species. If the cloned restriction fragment carries the mutation, the mutation usually can be transferred to pLC7-21 by recombination and cause a selectable phenotype several generations after the recombination event. The presence of a mutation on a cloned restriction fragment is therefore indicated when 0.1 ml of an overnight culture of cells containing both pLC7-21 and the pACYC177, pACYC 184, or pME21 plasmid with the cloned fragment results in many more colonies on LB agar plus erythromycin than does 0.1 ml of a culture of cells containing pLC7-21 and a pACYC177, pACYC184, or pME21 plasmid without an inserted fragment.

Shotgun Marker Rescue As an alternative to the one-by-one cloning of individual restriction fragments and testing of each by marker rescue methods, restriction nuclease fragments generated by restriction enzymes that recognize four or six bases can be shotgun cloned into either pACYC177, pACYC184, or

682

~EN~rICS

[46]

pME21, the resulting plasmids transformed into W3110(pLC7-21), and transformants containing both types of plasmid selected on agar. To facilitate the subsequent identification of cloned fragments, fragments generated by restriction enzymes are usually cloned into sites on the vector that allow precise excision of the cloned fragment. Subsequently, colonies on a petri dish containing 100 to 5,000 colonies resulting from a shotgun cloning experiment are replica plated using velvet pads to a petri dish containing LB agar plus intermediate and high concentrations of erythromycin. On LB agar plus erythromycin, only those colonies in which frequent recombinational transfer of the erythromycin resistance mutation to pLC7-2 l has occurred will grow. After replica plating, irregular patches therefore result from replica-plating colonies containing many erythromycin-resistant recombinants. Therefore, patches growing on petri dishes containing erythromycin usually contain a mutation-bearing fragment cloned in pACYC177, pACYC184, or pME21. Pure cultures are then obtained by restreaking from patches of cells growing on petri dishes containing erythromycin, and are used to prepare plasmid DNA by the minilysate procedure. To obtain cells containing only one plasmid, the DNA is used to transform W3110 Nal r (or JM83 in the case of pME21), selecting only the antibiotic resistance markers of the pACYC177, pACYC184, or pME21 vectors. Transformants are then purified by restreaking and confirmed to lack the colicin immunity conferred by pLC7-21 and to have a gene on the vector inactivated by an insertion. Plasmid DNA is then prepared from cultures of these transformants using the minilysate procedure. To be sure that the purified plasmid DNA contains the erythromycin resistance mutation on the cloned fragment, the pure plasmid is transformed into W3110 Nal r (pLCT-2 l) and analyzed by the marker rescue method appropriate to single cloned restriction fragments, as described above. The pure DNA is also analyzed by restriction nuclease digestion. The size of the cloned DNA fragment and its internal restriction enzyme recognition sites, together with sequence information on rRNA genes,9 allow identification of the cloned fragment. The cloned DNA fragment can then serve as a convenient source of DNA for DNA sequencing.

Sequencing the Mutations Overlapping restriction nuclease fragments containing the mutation can be identified by the mapping methods described above. The region of overlap maps the mutation to a region that can be less than 50 base pairs in length.4-6 Both strands of the entire region containing a mutation, as well 9 j. Brosius, T. J. Dull, D. D. Sleeter, and H. F. Noller, J. Mol. Biol. 148, 107 (1981).

[46]


683

as the corresponding region of the wild-type rRNA operon, should then be sequenced. Other Uses of rRNA Gene Mutations The presently available spectinomycin and erythromycin resistance mutations in 16S and 23S rRNA genes4 can be employed in the analysis of mutations at other positions. For example, the resistance phenotypes are obviously very useful for determining if mutations at other positions in rRNA genes affect ribosome function. These antibiotic resistance mutations may also allow the isolation of informative second site revertants of mutations in rRNA genes that affect ribosome function. In addition, primer extension methods designed around the available antibiotic resistance mutations (see below) can be used to help determine if mutations elsewhere in rRNA operons affect rRNA synthesis, affect the ability of rRNA to participate in ribosome assembly, or affect the functional abilities of properly assembled ribosomes. To facilitate the use of selectable rRNA mutations in other studies, we have constructed pRR-1 (Fig. 1), which contains an rrnC operon with a spectinomycin resistance mutation in the 16S rRNA gene and an erythromycin resistance mutation in the 23S rRNA gene. The spectinomycin resistance mutation is a C to U change at position 1192 in 16S rRNA, whereas the erythromycin resistance mutation is an A to U change at position 2058 in 23S rRNA. 4 Several restriction nuclease fragments of pRR-1, with end points generated by restriction enzymes that cleave pRR-1 only once, have been cloned into M13 phages. Together, these cloned fragments encompass the entire rrnC operon and facilitate oligonucleotide-directed mutagenesis of this plasmid. Primer Extension for Quantitative Measurement We have developed a primer extension method capable of measuring the fraction of 16S or 23S rRNA in crude RNA preparations that is synthesized from rRNA genes with the spectinomycin or erythromycin resistance mutations. The essence of this method is that, with the proper design of synthetic DNA primers and the use of proper reaction conditions, primer extension stops efficiently at one position when the primer is annealed to wild-type rRNA, and stops efficiently at a different position when antibiotic resistant mutant rRNA is the template. As a result, primer extension products of two lengths are generated in proportion to the relative abundance of the two types of templates. Using this primer extension method, single-lane electrophoresis can be used to determine if base

684

GENETICS

[46]

changes anywhere in an rRNA operon affect the abundance of mutant rRNA in total cellular RNA, total ribosomes, or subfractions of ribosomes that have been fractionated on the basis of functional activities (for example, subfractions of ribosomes capable of association with mRNA can be obtained from isolated polysomes, and subfractions of ribosomes capable of peptidyl transfer can be obtained from ribosomes released from polysomes by puromycin). The spectinomycin and erythromycin resistance mutations can therefore serve as "marker" base changes to assess the structural and functional consequences of mutations at known or unknown locations elsewhere in the rRNA operon. These antibiotic resistance mutations provide ideal marker base changes for this purpose because they do not prevent any essential ribosome function (the mutant rRNA permits cell growth in the presence of antibiotics) and because these mutations probably do not cause mutant rRNA to be discriminated against during ribosome assembly (see below). The strongly terminated primer extension method also allows rapid determination of whether newly isolated mutants are identical to previously isolated mutants, as the primer extension products of two mutants will be identical only if they have the same base change.

Primer Extension Design To illustrate the primer extension method, we have used as template RNA total cellular RNA prepared from EM2(pRR-1) growing exponentiaUy in LB broth containing 10/tg/ml of tetracyline. The RNA was extracted from the cells by hot phenol: SDS: EDTA extractiont° followed by two conventional phenol extractions and high-salt precipitation. ~i The design of primers and the resulting primer extension products are shown in Fig. 2. Typical primer extension products are shown in Fig. 3. Using this procedure, termination caused by addition of dideoxyadenosine triphosphate (ddATP) indicates that erythromycin-resistant 23S rRNA is 69% of the total 23S rRNA. Termination caused by ddGTP reveals that spectinomycin-resistant rRNA is 71% of the total 16S rRNA. In good agreement, termination caused by ddATP indicates spectinomycin-resistant rRNA is 74% of the total 16S rRNA. The consistency of the results obtained by this primer extension method indicates that the method is reliable and that there is probably little or no preferential degradation of rRNA due to these antibiotic resistance mutations. ~o R. J. Siehnel and E. A. Morgan, J. Bacteriol. 163, 476 (1985). ~t D. J. Lane, B. Pace, G. J. Olsen, D. A. Stahl, M. L. Sogin, and N. R. Pace, Proc. Natl. Acad. Sci. U.S.A. 82, 6955 0985).

[46]

ANTIBIOTIC

RESISTANCE

MUTATIONS

685

ANALYSIS OF E r r r MUTANT

't

U (in Ery r Mutant) PRIMER 23S rRNA

3 ' -AC'r'I~TGATATCG.~ff~-5 ' 5'-GCAGUGUACCCGCGGCAAOACGGAAAGACCCCOUGAACCUUUACDAUAGCUUGACACUGA-3'

PRIMER EXTENSION PRODUCTS

WILD TYPE S r y r MUTANT

3'-ddATGGGCGCCGTTCTGCCTTTCTGGGGCACTTGGAAATGATATCGAACTG-5t 7 + ddATP, dCTP 3f-ddATTCTGGGGCACTTGGAAATCATATCCAACTG-5 ' ~ dTTP, dGTP

ANALYSIS OF Spcr MUTANT U (in Spe r Mutant) PRIMER I6S rRNA

3' T-AGTTCAGTACTACCC~GAATGC-5' 5'-CCAGUGAUAAACUGGACGAAGGUGGGGAUGACGUCAAGUCAUCAUGGCCCUUACGACCAC-3'

PRIMER

WILD TYPE Spc r MUTANT WILD TYPE Spe r M U T A N T

EXTENSION PRODUCTS

3 ' -ddACTCCAGTTCAGTAGTACCC~CAATCC-5 ' "7 + ddATP, dCTP 3 ' -ddACACTTCAGTAGTACCGGGAATCC-5' ] dTTP, dGTP 3 ' -ddGCAGTTCAGTAGTACCGGGAATGC-5' 7 + ddGTP, dCTT 3'-ddGACCTCCTTCCTCCCCTACTACAGTTCACTACTACCGCCAATCC-5' ] dTTP, dATP

FIG. 2. The sequence of primers for strongly ternfinated primer extension and the regions of 16S and 2 3 S r R N A they hybridize to are shown. When the indicated dideoxynucleotides and deoxynueleotides are included in primer extension reactions, the primer extension products shown are observed when analyzed as described in Fig. 3.

Primer Extension Method

Our primer extension methods are similar, but not identical, to those of Pace and co-workers H Besides obvious strategic differences unique to our strongly terminated primer extension reactions, important modifications include the use of an end-labeled primer to facilitate labeling and quantitation of short primer extension products, differences in other reactants required by this alteration in labeling method, and a higher incubation temperature to reduce the synthesis of artifactual premature termination products. We have used gel-purified primers 22 nucleotides in length (Fig, 2) that were 5'-end-labeled to maximum specific activity using T4 polynucleotide kinase.t2 For sequencing ladders or strongly terminated primer extension ,2 T. Maniatis, E. F. Fritseh,and J. Sambrook,"MolecularCloning:A LaboratoryManual." Cold

Spring Harbor L a b . C o l d Spring Harbor, New York, 1982.

686

GENETICS

[46]

t23ACTG

AC GT CA

Wild Type rRNA - Templote

CA

U

G

W W

G G

W

t ~II1~

g •

Templote

G C C G T

T

Q G

| gw

C

T C

U

Ery r rRNA m

T

•

C

u

T+A

U¸

T

•

T

•

c •

T U

G

•

G

e

G G C

Primer

[46]


687

reactions electrophoresed without adjacent sequencing ladders we use 1.5 × 104 dpm (about 15 ng) of primer per annealing. However, we used only 7% as much primer for the strongly terminated extension products electrophoresed next to the RNA sequencing ladders in Fig. 3 to cause the band intensity of the strongly terminated primer extension products to approximate the band intensity of sequencing ladders. In each annealing reaction, 9 ag of RNA is annealed to primer in 7.5 al reactions containing 100 mM KC1, 50 mM Tris-Cl, pH 8.5. Annealing is accomplished by placing a 0.5-ml microfuge tube containing the annealing mixture in a 5-ml plastic tube full of water preheated to 90 °. After 1 min at 90 o, the entire 5-ml tube and its contents are allowed to slowly air cool at room temperature for 1 hr. Primer extension reactions yielding sequence ladders are obtained using a strategy similar to that used for dideoxynucleotide DNA sequencing. Primer extension is carried out in 5-al reactions containing 1 al of annealing mixture and 4 al of a solution that brings the final concentration of reagents to 25 mM Tris-C1, pH 8.5, 25 mM KC1, 5 mM dithiothreitol, 5 mM MgC12, 100 pM each of all four deoxynucleoside triphosphates, and one didoxynucleoside triphosphate per primer extension reaction, ddATP is used at a final concentration of 10 aM, ddCTP at 25 aM, ddGTP at 12.5 aM, and ddTTP at 20 aM. Each reaction also contains 1 unit of avian myeloblastosis virus (AMV) reverse transcriptase added prior to the addi-

FIo. 3. Autoradiograms of gel-separated primer extension products obtained using rRNA templates prepared from EM2(pRR- 1) and the primer hybridizing near the Eryr mutation in 23S rRNA (see Fig. 2) are shown. The four lanes on the right are conventional primer extension sequencing lanes obtained using this primer-template hybrid. Lane 1 is the primer subjected to ddA-terminated, strongly terminated primer extension conditions in the absence of template RNA. Lane 2 is the same amount of primer as in lane 1, subjected to ddA-terminated, strongly terminated primer extension after annealing to EM2(pRR-1) RNA. Lane 3 has no sample. Two primer extension products are apparent in lane 2, resulting from priming on the two types of templates. Note that the sequencing lanes also detect the base heterogeneity cause by the Eryr mutation, but that sequencing lanes are less suitable for detection and quantitation of mutant rRNA due to spurious termination at certain positions and large sequence-dependent variations in termination efficiency. Sequence-dependent variations in termination efficiency result from the effects of local sequence on the competition between deoxy and dideoxynucleotides for incorporation into DNA. Competition is prevented in strongly terminated primer extension reactions because a didcoxynudeotide is added in the absence of the corresponding deoxynucleotide. When RNA is prepared from strains containing pSS-1, which is a variant of pRR-I that does not possess the Eryr mutation, or from cells containing variants of pRR-1 that do not synthesize rRNA because they lack the rrnC promoter, ddA-terminated, strongly terminated reactions yield only a single primer extension product migrating at the wild-type position, and the sequence ladders read only T at the position of the Eryr mutation.

688

GENETICS

[46]

tion of nucleotides but subsequent to all other reagents. For our strongly terminated primer extension reactions, identical conditions are used except that one deoxynucleotide is completely omitted and 100 # M of the corresponding dideoxynucleotide is added. The reaction mixtures are then incubated for 5 rain at room temperature followed by 30 min at 42 °. Reactions yielding sequencing ladders are then incubated for 15 additional min at 42 ° after addition of 1/zl of 10 m M Tris-C1, pH 8.5, containing 1 m M o f a l l four deoxynucleotide triphosphates and 1 unit of AMV reverse transcriptase. Strongly terminated reactions are incubated for 15 additional min at 42 ° after addition of I/A of 10 m M Tris-C1, pH 8.5, containing 1 m M of three deoxynucleotide triphosphates, 1 m M of a single dideoxynueleotide triphosphate, and 1 unit of AMV reverse transcriptase. After all incubation is completed, 6/tl of 86% deionized formamide, 10 m M EDTA, 0.08% xylene cyanol, 0.08% bromphenol blue is added, the reactions heated for 2 min at 90 °, and a portion of each sample electrophoresed on standard 20% polyacrylamide-urea DNA sequencing gels until the xylene cyanol nears the bottom of the gel. The 22 nucleotide primers migrate 95% as fast as the xylene cyanol dye.

Mutant rRNA Gene Number The primer extension method described above allows a point mutation to be used to accurately determine the percentage of mutant rRNA in the cell. If a similar method is applied to total single-stranded rDNA isolated from a cell population, the percent of mutant rDNA can be measured. The seven nonmutant rRNA genes present in the bacterial chromosome thereby provide an internal copy number reference by which the number of mutant copies can be judged. Thus, if a plasmid contains an rRNA gene with a point mutation, the plasmid copy number can be determined from the relative number of mutant and nonmutant rRNA genes. Measurements of mutant rRNA gene number, coupled with measurements of mutant rRNA, are all that is necessary for most investigations of the effects of cis-acting mutations on rRNA gene regulation or rRNA stability. To measure the copy number of mutant rRNA genes, it is first necessary to prepare total cellular DNA. Generally, we split a single culture and simultaneously prepare DNA and RNA for primer extension analysis, as it is not safe to assume that culture-to-culture variation in plasmid copy number is insignificant. Cells from 50 ml of culture are collected by centrifugation, resuspended in l ml of 10% sucrose, 50 raM Tris-C1, pH 8.0. Then, 0.2 ml of 100 mg/ml lysozyme freshly dissolved in 0.25 M Tris-C1, pH 8.0 is added, and the cell suspension incubated on ice for 10 min. Then, 0.6 ml of 5 M NaCI is added and 0.4 ml of 10% sodium dodeeyl

[46]


689

sulfate quickly and completely mixed in. The resulting mixture is incubated overnight on ice. Then, without removal of cell debris by centrifugation, 0.08 ml of 10 mg/ml ethidium bromide and 1 g of CsCl are added per milliliter oflysate, and the density is adjusted to 1.55 g/ml (refractive index

BstXI + $ma I t

2

3

Ss t X

Ora I

t23

~ 2 3

E

S E

$

FIG. 4. Ethidium bromide-stained agarose gels of total cellular DNA digested with the indicated restriction nucleases. The lanes marked 1 contain 1 #g of pRR-1 DNA, lanes marked 2 contain 10 gg of DNA from the strain EM22 without a plasmid, and lanes marked 3 contain 10 gg of DNA from the strain EM22(pRR-I). S indicates the restriction fragments of pRR-I which contain the Spc" mutation at position 1192 of the 16S rRNA gene, and E indicates the fragments which contain the erythromycin resistance mutation at position 2058 in the 23S rRNA gene. All of the indicated fragments of pRR-I, with the exception of the large fragment produced by Sinai, comigrate with chromosomal rRNA gene fragments. Thus, the mutant rRNA gene copy number is proportional to the percentage of mutant rRNA genes in the comigrating restriction nuclease fragments.

690

GENETICS

[46]

of 1.3860) by the addition of solid CsC1 or water. The resulting viscous mixture is then centrifuged at 45,000 rpm at 15 ° for 60 hr in either a swinging bucket or fixed angle rotor. The DNA is visualized by the use of a long-wave ultraviolet light and is removed through the side of the centrifuge tube using a single syringe. Care is taken to remove all plasmid and chromosomal DNA without losing any DNA. Generally, we remove the entire contents of the tube with the exception of the pellet and pellicle. The ethidium bromide is then removed from the DNA by repeated extraction with water-saturated N-butanol and the CsC1 subsequently removed by dialysis against l0 m M Tris-C1, pH 7.4, 1 m M EDTA. To analyze the copy number of rRNA genes containing the Spc r (position 1192 in 16S rRNA) or the Ery r (position 2058 in 23S rRNA) mutations, the DNA is digested with Sinai, DraI, or BstXI and SstII. The restriction nuclease digests should contain 20 gg/ml of RNase A to prevent contaminating RNA from obscuring restriction nuclease fragments after gel electrophoresis. As a result of these digestions, restriction fragments are produced which contain entirely rDNA sequences from within 16S or 23S rRNA genes. These fragments are therefore uniform in size for all mutant and nonmutant rRNA genes. Ten to 50 #g of digested DNA is then electrophoresed on a 1.2% agarose gel. As can be seen in Fig. 4, restriction fragments from rRNA genes are visible above the background distribution of restriction fragments, even when the cell contains only the seven chromosomal rRNA operons. To determine the percentage of rRNA operons which are mutant, it is then only necessary to purify the restriction fragments which span the region containing the mutations, denature them, and perform primer extension analysis on the single-stranded DNA using the methods described above for rRNA templates. The fragments needed for determining the copy number of rRNA genes containing the Spc ~ and Ery" mutations are identified in Fig. 4. We denature the DNA by heating the annealing mixture to 100 ° immediately prior to annealing. It is likely that any of the commonly used methods for extracting restriction nuclease fragments from agarose gels will work. We have found that the desired fragments can be obtained in good yield and purity by electroeluting the DNA from excised gel pieces using an ISCO model 1750 sample concentrator and the methods recommended by the manufacturer.

[47]

MAXlCELL ANALYSISOF PLASMID-CODEDrRNA

[47] Analysis of Plasmid-Coded Ribosomal Maxicell Techniques

691

RNA

B y DAVID K . JEMIOLO, ROLF STEEN, MICHAEL J. R. STARK,

and

ALBERT E. DAHLBERG

Introduction It is now apparent from numerous studies that the rRNAs have important functional roles in the process of protein synthesis. Studies of mutations in cloned rRNA operons of E. coli have begun to define specific regions involved in particular steps of translation. These mutants have been produced by a variety of methods including deletion, 1 bisulfite, 2 ethylmethane sulfonate, 3 and synthetic oligonucleotide directed mutagenesis. 4 The initial mutants were made in plasmid pKK3535, which contains the rrnB operon cloned into plasmid pBR322. 5 A number of different plasmids are now available, 3,6 including two plasmids which contain inducible promoters PL (pNO2680) 7 and T7 (pAR3056) s that permit the isolation and characterization of lethal mutants. While rRNA mutations in the plasmids often result in altered phenotypes (growth rate or colony morphology), it is important to study the gene products directly, namely the rRNAs. To do this, plasmid gene transcripts have to be distinguished from chromosomal gene transcripts. In this chapter we describe a maxicell technique to accomplish specific labeling of rRNA transcribed from plasmid-borne cloned genes. Also, we discuss an alternate method for specific labeling of plasmid-coded rRNA transcripts in vivo that is based on the use of T7 RNA polymerase and a plasmid-borne T7 late promoter.

R. Gourse, M. Stark, and A. Dahlberg £ Mol. Biol. 159, 397 (1982). 2 C. Zwieb and A. Dahlberg, Nucleic Acids Res. 12, 436 (1984). 3 L. Mark, C. Sigmund, and E. Morgan, £ Bacteriol. 155, 989 (1983). 4 H. Goeringer, R. Wagner, W. Jacob, A. Dahlberg, and C. Zwieb, Nucleic Acids Res. 12, 6935 (1984). 5 j. Brosius, T. Dull, D. Sleeter, and H. Noller, J. Mol. Biol. 148, 107 (1981). 6 R. Steen, D. Jemiolo, R. Skinner, J. Dunn, and A. Dahlberg, Prog. Nucleic Acid Res. MoL BioL 33, 1 (1986). 7 R. Gourse, Y. Takebe, R. Sharrock, and M. Nomura, Proc. NatL Acad. Sci. U.S.A. 82, 1069 (1985). s R. Steen, A. Dahlberg, B. Lade, F. Studier, and J. Dunn, EMBO J. 5, 1099 (1986). METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form reserved

692

GENETICS

[47]

Maxicells

The original maxicell technique developed by Sancar and co-workers9 was designed to specifically radiolabel plasmid-coded proteins. Transcription from the host chromosome was abolished by destroying the host DNA with ultraviolet (UV) irradiation. In order to specifically label rRNA instead of protein we found it necessary to modify this procedure. 1° Several factors made this necessary. There are seven rRNA operons distributed throughout the chromosome in E. coli. These operons are transcribed as primary transcripts containing single copies of the 16S, 23S, and 5S rRNAs and one or more tRNAs. The primary transcript is processed into mature products in a series of steps involving specific ribonucleases, ribosomal proteins, and various RNA-modifying enzymes. These proteins must be present for processing to occur and thus levels of ultraviolet irradiation sufficient to abolish production of these proteins cannot be used. On the other hand, sufficient irradiation is required to stop transcription of host chromosomal rRNA operons. By varying the level of irradiation we have found conditions that block host chromosomal transcription of rRNA operons while allowing plasmid-borne genes to be transcribed. Under these conditions, cells continue to produce proteins and to process the primary transcripts into mature rRNA.

Repair Mechanisms The principle products of ultraviolet irradiation (220 to 300 nm) of DNA are pyrimidine dimers. Escherichia coli has three ways of repairing these lesions. In the process known as photoreactivation the enzyme DNA photolyase, the product of the phr gene, binds to the pyrimidine dimer and catalyses a light-dependent (300 to 500 nm) reversal. The second repair mechanism is based on excision of a strand of DNA containing the dimer by the UVRABC nuclease, the product of the uvrA, uvrB, and uvrC genes. Finally, there is recombinational repair involving genes whose products are known to play roles in homologous recombination such as recA, recB, recC, recF, recJ, recN, and ruv. In order to use ultraviolet irradiation to achieve differential expression of genes, the repair mechanisms discussed above must be inoperative. Because DNA photolyase is light dependent, repair of pyrimidine dimers by photoreactivation can be avoided by keeping cells in the dark. (For cells with mutations in phr, normal lighting can be used.) To avoid excision repair, cells harboring mutations in the genes coding for the UVRABC 9 A. Sancar, A. Hack, and W. Rupp, J. Bacteriol. 137, 692 (1979). l0 M. Stark, R. Goursc, and A. Dahlberg, J. Mol. Biol. 159, 417 (1982).

[47]

MAXICELL ANALYSIS OF PLASMID-CODED

rRNA

693

nuclease can be used. The uvrA and uvrB genes code for two of the proteins that make up the UVRABC nuclease and expression from these genes is controlled by the SOS regulatory system. In this system, a protein repressor, the product of the lex gene, blocks expression of uvrA and uvrB. The LexA protein is in turn controlled by the RecA protein. SOS-inducing treatments such as ultraviolet irradiation result in activation of the RecA protein, a protease, that cleaves and inactivates the LecA repressor protein. Thus excision repair is activated by the ReeA protein and cells that are recA- cannot induce production of the UVRABC nuclease. It is known that the LexA protein also regulates several of the genes involved in recombinational repair. Thus recA- cells cannot repair damaged DNA by recombination. Cell Strain Requirements (Table I)

The minimum requirement for maxicells is a strain that is recA- such as HBI01. H This strain is phr + and must be kept in the dark after irradiation. Cells that are phr'-, such as CSR6039 can be used in normal lighting. CSR603 is recA- and uvrA- and is very sensitive to ultraviolet radiation, requiring only low levels of irradiation to induce chromosomal degradation. However, this cell strain has some undesirable characteristics. First, it is hsdR + and will therefore restrict plasmid DNA from cells that are hsdMsuch as HB 101. ~ (In order to avoid restriction damage, plasmids should be passed through a cell strain that is hsdR- and hsdM + such as MC1061? 2 Alternatively one may use UNC1085, a strain derived by Sancar from CSR603 but deficient in host restriction.) Second, CSR603 cells form colonies that are loose and tend to spread on plates, thus readily coalescing if plated as a dense culture. Third, it is difficult to isolate large amounts of plasmid DNA from CSR603 by standard minilysis techniques. Cell Growth

Cells harboring plasmids are plated from - 2 0 ° glycerol stocks onto LB (Luria broth) ~3 agar with 200 gg/ml ampicillin and incubated at 37 ° overnight. The plates should produce well-separated colonies of uniform size. Irregular-sized colonies should be removed. In addition to being antibiotic-resistant, cells harboring a plasmid with the rRNA operon grow slower than cells with a plasmid lacking the rRNA operon. Mutations in the Bolivarand K. Baekman, this series, Vol. 68, p. 245. ~2M. Casadaban and S. Cohen, J. Mol. Biol. 138, 179 (1980). ~3j. Miller,/n "Experiments in MolecularGenetics," p. 431, Cold Spring Harbor Lab., Cold Spring Harbor, New York, 1972. H F.

694

GENETICS

[47]

TABLE I CELL STRAINREQUIREMENTS E. coli strains HBI01: recA- and F-, pro-, leu-, thi-, lacY-, str', hsdM-, hsdR-, endol- ara-14, galK2, xyl-5, mtl-1, sup44° MCI061: araD139, A (ara leu) 7697, A lacX74, galU-, hsdR-, hsdM +, strAb CSR603: recA1, uvrA6, phr-1, thr-1, leuB6, proA2, argE3, thi-1, ara-14, lacY1, gal-K2, xyl-5, mtl-1, rpsL31, tsx-33, ,~-, supE44, nalA98, F -c BL21(DE3): hsdM-, hsdR-, r/fl, 2 lysogend Plasmids pBR322 is a multicopy ColE1 plasmid,e tet~, amp" pKK3535 is a derivative of pBR322. It carries the rrnB operon cloned into the tetracycline gene at a unique BamHI restriction endonuclease recognition site, amp"/ pNO2680 is a derivative of pKK3535 constructed by Gourse and co-workers.s It has the 2 promoter PL in place of the ribosomal promoters PI and P2. amp" pEJM007 is a low-copy vector derived from pDPT487 and NRI. ~t It carries the rrnB operon of E. coli. cam; str'/spc r pCI857 is a multicopy plasmid derived from miniplasmid PISA/This plasmid is compatible with ColE1 plasmids. It codes for the temperature-sensitive 2 repressor protein CI857. kam" pAR3056 is a high-copy number plasmid (pBR322 derivative) carrying the rrnB operon fused to a T7 late promoter, amp"k aF. Bolivar and K. Backman, this seres, Vol. 68, p. 245. bM. Casadaban and S. Cohen, J. Mol. Biol. 138, 179 (1980) cA. Sancar, A. Hack, and W. Rupp, J. Bacteriol. 137, 692 (1979). aF. Studier and B. Moffatt, J. Mol. Biol. 189, 113 0986). ej. Suttcliife, Cold Spring Harbor Symp. Quant. Biol, 43, 77 (1978). fJ. Brosius, T. Dull, D. Sleeter, and H. Noller, J. Mol. Biol. 148, 107 0981). gR. Gourse, Y. Takebe, R. Sharrock, and M. Nomura, Proc. Natl. Acad. Sci. U.S.A. 82, 1069 (1985). hR. Steen, D. Jemiolo, R. Skinner, J. Dunn, and A. Dahlberg, Prog. Nucleic Acid Res. Mol. Biol. 33, 1 0986). i D. Taylor and S. Cohen, J. Bacteriol. 137, 92 (1979). J E. Remaut, H. Tsao, and W. Hers, Gene 22, 103 (1983). kR. Steen, A. Dahlberg, B. Lade, F. Studier, and J. Dunn, EMBO J. 5, 1099 (1986).

cloned rRNA gene often cause further reduction in cell growth rate. Thus fast-growing clones may be the result of spontaneous alterations in the plasmid. We avoid propagating cells in liquid culture because fast-growing cells are not as easily detected in liquid as on plates. When colonies are about 1 to 2 mm in diameter, the cells are scraped from the plates in 1.5 ml LB and used to inoculate 25 ml of LB with 200 #g/ml ampicillin at a final A~0o of between 0.18 and 0.25. The cells are then grown at 37 ° with good aeration to a final A600 of 0.5. Because the cell density will attenuate ultraviolet radiation, it is important to irradiate cells

[47]

MAXICELL ANALYSIS OF PLASMID-CODED r R N A

695

at a fixed cell density. In the event that cells have overgrown slightly, they can be diluted with broth to the appropriate density. Cells can also be grown in defined medium such as M9 with casamino acids and glucose. ~3

Irradiation of Cultures Transfer 12.5 ml of the culture to a 100 m m × 15 m m sterile, plastic petri dish, add a 1.25 in. stir bar, cover the dish with a cardboard disk to shield the cells, and place it on a magnetic stirrer positioned 25 cm beneath the ultraviolet light source (G15T8 germicidal lamp, Sylvania). Start the culture stirring gently. The stir rate should be sufficient to keep the cells from collecting at the perimeter of the plate yet gently enough to avoid frothing. With the culture stirring, expose the cells directly to the ultraviolet irradiation for the appropriate length of time by removing both the cardboard cover and the top lid of the petri dish. After irradiation, 10 ml of cells is transferred to an 125-ml Erlenmeyer flask and allowed to recover for 2 hr at 37 ° with shaking. For cell strains that are wild type at the phr locus (e.g., HB 101), cells must be kept in the dark after irradiation. This can be accomplished by covering the flask with aluminum foil.

Recovery After 2 hr, D-cycloserine is added (final concentration of 200/~g/ml, freshly prepared in water), and the cells are allowed to recover for an additional 3 hr. D-Cycloserine is a cell wall inhibitor and any cells that have escaped irradiation damage will lyse as they continue to grow. Without cylcoserine, these cells would produce a high background of radiolabeled rRNA from the host chromosome. In cultures that have not been irradiated sufficiently, the ceils will lyse several hours after addition of cycloserine. This fact can be used to determine an approximate exposure time. Cells that have been irradiated for sufficient lengths of time will not be lysed when treated with cycloserine. The exposure times are strain specific and also depend on the media since certain media components absorb ultraviolet radiation. As examples, HB101 in M9 with casamino acids requires approximately 73 see of exposure and CSR603 in LB medium requires about a 10-sec exposure. Throughout the recovery period, background transcription from chromosomal operons decreases. While a 5-hr recovery is sufficient for most purposes, longer times may be required. To monitor background, cells with plasmids lacking a rRNA operon are used. As an example, we use cells with pBR322 because it is the parent plasmid of pKK3535 (pKK3535 contains the rrnB operon of E. coli cloned in a unique BamHI restriction

696

GENETICS

[47]

E Ut~ ..go 0..

0

t-

1

~

O.

2

"~

O.

3

0 IZ

tr) v v

Q..

2

-90.0 p23Sm23S-

-67.0 "ql

-43.0

x 0

-30.0

17S16S-

I

- 20.0 - 14.4 a

b

FI~. 1. (a) Plasmid-dependent rRNA production in maxicells. E. coli cells (strain CSR603) containing either no plasmid (lane 1), pKK3535 (lane 2), or pBR322 (lane 3) were irradiated for 10 sec and labeled with ortho[32P]phosphate for 10 hr as described in the text. rRNA was extracted from lysates and separated by electrophoresis on 3.0%/0.5%acrylamide/ agarose composite gels. Labeled rRNA was located by autoradiography. The labeled bands of 23S and 16S rRNA are present only in cells containing plasmid-borne rRNA genes (pKK3535, lane 2) and not present in cells with either no plasmid (lane 1) or with pBR322 (lane 3). Thus the rRNA transcripts in lane 2 are from plasmid-borne genes. In addition the results show processing of the primary transcripts into mature 16S and 23S rRNAs. (b) Protein synthesis in modified maxiceUs. Aliquots of the irradiated cultures of CSR603, described in (a), containing either no plasmid (lane 1) or pKK3535 0ane 2) were labeled with [3SS]methionine.3sS-Labeledproteins were separated by el~rophoresis on 12.5% acrylamide gels [U. Laemmli, Nature (London) 227, 680 (1970)] and visualiTed by fluorography. The results show that modified maxicells continue to produce proteins coded by host genes.l°

[47]


697

endonuclease recognition site located in the tetracycline gene in pBR322) 5 (see Fig. I a). Alternatively, background transcription can be detected using the mutants STU 1 or SMA 1- 6 with deletions of 371 and 770 bases, respectively, within the structural gene for 16S rRNA in pKK3535. These deletion mutants are transcribed and processed into RNAs that are smaller than 16S rRNA and migrate faster than 16S rRNA in agarose/acrylamide composite gels.~° While the occurrence of these mutant products indicates plasmid gene transcription, the absence of labeled RNA in the region of 16S rRNA indicated lack of chromosomal rDNA transcription. If there is

12345 23S-

A371 FIG. 2. Analysis of ribosomal RNA genes cloned on a low-copy-number vector. The rrnB operon was cloned into a low-copy-number plasmid (pDPT487) to produce pEJM007. Deletion mutations within the structural gene for 16S rRNA of eitber 371 bases (STUI-I) or 53 bases (SALI-72) were then constructed. E. coli strain HB101 was transformed with these plasmids and analyzed. In lane 1, "cells harboring the 371-base deletion, STUI-1, were pulse labeled with ortho[a2p]phosphate for 30 rain and then analyzed for RNA. In the remaining lanes, maxicells were prepared from cells with either STUI-I (lane 2), pEMJ007 (lane 3), pDPT487, parent of pEJM007 without the rmB operon (lane 4), or SALI-72 (lane 5) and labeled for 10 hr as described in the text. Pulse labeling (lane 1), yields bands at 23S rRNA (coded by both host and plasmid genes), 17S and 16S rRNA (coded by host genes), and A371 (coded by plasmid genes). Specific labeling of plasmid borne genes in maxicells is shown in lane 2 where the plasmid-derived A371 transcript is observed in the absence of 17S and 16S rRNA labeling. The presence of labeled rRNA in cells with a cloned rrnB (lane 3) and the absence of labeled rRNA in cells harboring only the plasmid vector (lane 4) again show that only plasmid-borne genes are labeled in maxiceUs.Maxicell analysis of the 53-base deletion in SALI-72 (lane 5) shows that the deletion effects processing of precursor small subunit RNA.6

698

GENETICS

[47]

evidence o f host r D N A transcription then the U V exposure time is increased slightly until this disappears. (See Fig. 2; here a low-copy-number plasmid, pEJM007, is used in place o f p K K 3 5 3 5 . )

Labeling of rRNA After the recovery period, cells are collected by gentle centrifugation at 5000 rpm for 1 rain in the Sorvall SS34 rotor at 3 °. The cells are washed twice with 5 ml o f zero phosphate m e d i u m (ZPM) containing 0.2% phosphate-free casamino acids (see Appendix), 0.2% glucose, and 1.0/zg/ml thiamine, resuspended in 2.0 ml o f the same solution, and transferred to a 30-ml Corex tube. The cell suspension is adjusted to 200/tg/ml D-cycloserine and 50 # M KH2PO 4. Radiolabel is added (50/tCi o f carrier-free ortho[32p]phosphate) and the cells are incubated at 37* with shaking (see Table II). The length o f time o f labeling can be from 20 min to 16 hr. We have observed that processing o f a primary transcript in maxicells is slow. ~° For short label times, less than 1 or 2 hr, most o f the label is found in precursors to 16S r R N A and 23S rRNA. For this reason, we routinely label for 10 to 16 hr.

Lysate Preparation The radiolabeled cells are pelleted and washed at 4 ° with 1.0 ml o f a solution containing 25 m M Tris, p H 7.6, 60 m M KC1, 10 m M MgCI2, 20% (w/v) RNase-free sucrose, and 150/zg/ml lysozyme. The wash steps are carried out in a 1.5-ml E p p e n d o r f centrifuge tube. The cells are resuspended in 25/zl o f this solution and lysed by three cycles o f freezing and thawing. (Samples should be kept at a low temperature during the thaw, i.e., 4°.) Add 175/zl o f a solution containing 25 m M Tris, p H 8.0, 30 m M TABLE II MAXICELL PROTOCOL

1. Dilute cells into 25 ml of media with antibiotic to an At00of 0.2 and grow to final A~00of 0.5 2. Transfer 12.5 ml of the culture to a petri dish, irradiate while stirring 3. Place 10 ml of the irradiated culture into a 25-ml flask and incubate at 37° for 2 hr 4. Add cydoserine to a final concentration of 200/~g/ml and incubate at 37° for 3 hr 5. In a Corex tube collect cells by centrifugation, rinse twice with 5.0 ml of ZPM with 0.2% phosphate-freecasamino acids, 0.2%glucose, and 1.0/zg/mlthiamine and resuspend in 2.0 ml of this solution 6. Add cycloserine(to 200/zg/ml), KI-I2PO4(to 50/zM), and add 50 #Ci of ortho[32p]phosphate. Label at 37° with shaking 7. Collect cells by centrifugation, rinse, and lyse as described in Table IV

[47]


699

KC1, 10 m M MgCI2, 0.2% (w/v) sodium deoxycholate, 0.6% (w/v) Brij 58, and 60 gg/ml DNase I. (The sodium deoxycholate should be from a fresh 1% stock solution.) The samples are kept on ice for 10 rain and then centrifuged in an Eppendorf centrifuge at full speed for 10 rain at 4 ° to pellet the cellular debris. The supernatant is carefully transferred to fresh tubes. Samples are usually divided into 25-/tl aliquots and stored at - 80 o.

Extracting RNA A portion of the sample is diluted 1:4 with a solution containing 25 m M Tris, pH 8.0, 30 m M KC1, and 10 m M MgC12 and adjusted to 0.2% in sodium dodecyl sulfate. An equal volume of a 1:1 mixture of phenol and chloroform is added and the solution is vortexed briefly. The samples are centrifuged in an Eppendorf microcentrifuge for 10 min at room temperature. (Use tubes with screw-top caps and O tings to avoid radioactive spills.) The aqueous layer is carefully removed and analyzed by electrophoresis after being mixed with an equal volume of 40% sucrose with either xylene cyanole or bromphenol blue. Alternatively, the samples can be concentrated by precipitation with ethanol, resuspended in a solution containing 10 m M Tris, pH 8.0, and 1.0 m M EDTA, mixed with sucrose and dye, and electrophoresed.

Electrophoresis Ribosomal RNAs are analyzed by electrophoresis in agarose/acrylamide composite gels using a Tris/borate/EDTA running buffer (see Figs. 1 and 2). To study 70S ribosomes and subunits a Tris/potassium/magnesium chloride-containing buffer is used. 14 In the latter case, maxicell lysates are electrophoresed directly into the gel. In the presence of 10 m M magnesium ion, normal subunits associate to form 70S particles. Some rRNA mutations, however, prevent subunit association) ° Radiolabel in 70S ribosomes may be due to the presence of either labeled 16S rRNA or 23 rRNA or both. This can be determined by a two-dimensional gel method where the first dimension separates ribosomes from subunits and the second dimension separates rRNA, after first deproteinizing the samples within the gel by soaking in SDS (see Fig. 3). Samples can be compared by loading either equal volumes of lysates or equal numbers of counts. Normalization by equal volumes of lysates assumes that the efficiency of lysis of the samples is constant. The normalization by counts assumes that equal counts con'e-

14A. Dahlberg, in "Gel Electrophoresis of Nucleic Acids--A Practical Approach" (D. Richwood and B. Haines, eds.), p. 213. IRL Press, Oxford, England, 1982.

700

GENETICS

[47]

000

J[~k

vv

,

B

23S"

1 7 S ,16S

"-

v ~'

..........~

FIG. 3. Two-dimensionalanalysisof ribosomal particles in maxicell lysates. Cells carrying wild-type plasmid pKK3535 or an uncharacterized mutant derivative defective in processing 17S to 16S rRNA were grown and labeled with 32p as described in the text. After labelinf, aliquots of lysates were electrophoresed in the first-dimension gel of 3%/0.5% acylamide/ agarose, 25 h i m Tris, pH 8.0, 10 m M MgC12, 30 m M KCI (from left to right) to separate 70S

ribosomes and subunits.~° To analyze the RNA content of the ribosomalparticles, the gel was sliced into strips and incubated at room temperature for 1 hr in Tris/borate/EDTA buffer, and then for an additional hour in the same buffer containing 0.2% SDS. The gel strip was then polymerized into a second gel of 3%/0.5% acrylamide/agarose, in Tris/borate/EDTA and electrophoresed (from top to bottom) for 10 hr at 100 V and 3°. Ribosomal particles (70S, 50S and 30S) are identified by the presence of mature 23S and 16S rRNA. (A) pKK3535, (B) mutant plasmid. Note the reduced level of radiolabel present in mutant rRNA (unstable?). Label is present in the 17S form (in the second-dimension gel) and in particles migrating faster than 30S subunits (in the first-dimension gel). spond to an equal n u m b e r o f cells. Only a small fraction o f the total counts are actually incorporated into r R N A . C h e m i c a l l y I n d u c e d Maxicells Plasmid-encoded r R N A transcripts can be specifically labeled by a second m e t h o d that is based on the use o f a T7 late p r o m o t e r and a cloned T7 R N A polymerase gene under control o f the lac UV5 p r o m o t e r ? In the plasmid pAR3056, promoters P1 and P2 have been replaced by a T7 late promoter. This p r o m o t e r is not recognized by the E. coli R N A polymerase but rather by the T7 R N A polymerase, and, in cells lacking T7 R N A polymerase, transcription from the T7 p r o m o t e r does not occur.

[47]


701

An E. coli strain BL21 (DE3) is available that contains one copy of the T7 RNA polymerase gene integrated into the chromosome under control of a lac UV5 promoter. ~5Thus production of T7 RNA polymerase can be chemically induced by addition of IPTG (Isopropyl fl-o-thiogalactopyranoside). Transformation of BL2 I(DE3) with pAR3056 produces a system capable of conditional expression of a cloned rRNA gene. Treating cells with IPTG causes T7 RNA polymerase to be produced which, in turn, transcribes rRNA from the T7 late promoter. The characteristic that makes this system particularly useful for radiolabeling plasmid-coded rRNA is the fact that T7 RNA polymerase is resistant to rifampicin, an inhibitor of E. coli RNA polymerase. In the presence of rifampicin, the only genes that are transcribed are those controlled by a T7 late promoter. The T7 promoter/RNA polymerase system has been successfully employed to label rRNA transcripts. This is accomplished by first inducing the production of RNA polymerase using IPTG. Cells are then treated with rifampicin to block transcription of host genes. Upon addition of ortho[a2p]phosphate, rRNA transcribed from plasmid-borne genes is specifically labeled. To our surprise, processing of the transcripts to mature 16S and 23S rRNA continues even in the presence of rifampicin for up to 35 min. We suggest that, under the condition of overproduction of plasmid-coded rRNA, the mRNAs coding for ribosomal proteins and ribosomal processing enzymes are stable and continue to be translated) Cell Growth

To achieve effective processing of plasmid-coded rRNA the cells must be in late log phase of growth at the time of induction by IPTG (Fig. 4). A culture ofBD2 I(DE3) harboring plasmids derived from pAR3056 is grown overnight to stationary phase in 10 ml of ZPM containing 10% LB ~a and 200 ~tg/ml ampicillin at 37 °. This culture is diluted 1:3 with a medium containing ZPM, 10% LB, 0.2% casamino acids, 0.2% glucose, and 200 #g/ml ampicillin and grown for 2.5 hr at 37 ° with good aeration. Growth is monitored carefully. Under these conditions, cells begin to grow without a lag phase and by 2.5 hr the culture should be in late log phase growth (Aroo=1.2, Klett = 200). It is important to keep the level of glucose low (0.2%) or, alternatively, use glycerol or succinate (0.5%) since excess glucose will inhibit induction by IPTG. ~6

~5F. Studier and B. Moffatt, J. Mol. Biol. 189, i 13 (1986). ~6W. Gilbert and B. Mfiller-Hill, Proc. Natl. Acad. Sci. U.S.A. 56, 1891 (1966).

702

GENETICS

[47]

b

400--

a

O R I --,,- ~

200--

b

--

KLETT

p 2 3 S..~ :!~ W m2 3 S ' "

tO0_ 8O

6040-

1 7 S-,,. Q 1 6S "~

I 2O

I

I 60

I

I 100

I

I 140

I

I 180

I

I 220

I

MINUTES

FIG. 4. Synthesisof rRNA by T7 RNA polymerasein vivo;effectof growthconditions on processing of rRNA. BL21(DE3)/pAR3056 cells were induced with IPTG for 45 rain at mid and late log phases of growth (at points a and b of the growthcurve, respectively).Rifampicin (150/tg/ml) was added, and after 5 rain the cells were radiolabeled for 20 rain with 32p. The RNA was extracted and separated by eleetrophoresis as in Fig. 1. Lanes a and b in the autoradiogramat the fight showrRNAs of cellsinduced at points a and b in the growthcurve.

Induction of T7 RNA Polymerase T7 RNA polymerase production is induced by the addition of IPTG (0.5 m M final concentration) (see Table III). The cells are allowed to continue for 45 min at 37". After this time, rifampicin is added (150/lg/ml final concentration) to inhibit transcription by E. coil R N A polymerase. (Rifampicin is made up as a 10 mg/ml stock in DMSO and stored at -20*.)

Labeling of rt~VA Five minutes after addition of rifampidn, ortho[32p]phosphate is added to a final level of 2 #Ci/ml and the culture is incubated at 37" for 20 rain. The culture is then chilled rapidly to just above the freezing point by transferring the sample to a centrifuge tube at - 70". Cells are harvested by centrifugation at 5,000 rpm for 5 min at 3", washed in ice-cold Z P M (1 ml ZPM per 10 ml original culture), respun, resuspcnded in 25 ~1 Tris/potasslum/magnesium buffer, and lyscd (Table IV). Ribosomes and RNA are analyzed by gel dectrophoresis as described above for maxiceUs.

[47]


703

TABLE III CHEMICALLYINDUCEDMAXICELL

1. Grow cells overnightin ZPM containing 1096LB and 200/zg/ml ampieillin 2. Dilute culture with 3 volumes of ZPM containing 10% LB, 0.2% casamino adds, 0.2% glucose, and 200/zg/ml ampicillin 3. Shake at 37" for 2.5 hr (to late log phase) 4. Add IPTG to 0.5 mM (to induce production of T7 RNA polymemse) and continue to shake at 37* for 45 rain 5. Add rifampicin to 150/zg/mland, after 5 min, add ortho[32p]phosphateto a concentration of 2/zCi/ml and shake for 20 rain at 37* 6. Rapidly cool to just above the freezingpoint by transferring the culture to a centrifugetube at --70" 7. Collect cells by centrifugation, rinse, and lyse as described in Table IV

Discussion The maxicell m e t h o d introduced by Sancar and co-workers9 is a technique designed to specifically label proteins in vivo which are coded by plasmid-borne genes. It represents an alternative to the so-called minicell technique in which labeling occurs in portions o f a cell lacking chromosomal D N A , formed by uneven cell division. The basis o f the maxicell technique is the use o f U V irradiation to damage c h r o m o s o m a l D N A in cells that are unable to repair the damage. While plasmid D N A largely escapes damage due to its small target size, c h r o m o s o m a l D N A is damaged and is eventually degraded. We have modified the maxicell technique to label r R N A coded by plasmid-borne genes. Modifications were necessary for several reasons. Ribosomal R N A is coded by operons that contain a single copy o f each o f the r R N A s (16S, 23S, and 5S) and one or more tRNAs. The opcrons are

TABLE IV CELLLYSlS 1. Wash cells at 4 ° with 1.0 ml of a solution containing 25 mMTris, pH 7.6, 60 mMKCI, 10 mM MgCl2,20% (w/v) RNase-freesucrose, and 150/zg/mllysozymeand resuspend cells in 25/Jl of this solution 2. Lysecells by three cyclesof freezing/thawing 3. Add 175/11 of a solution containing 25 mM Tris, pH 8.0, 30 mM KCI, Ill mM MgCI2, 0.2% (w/v) sodium deoxycholate, 0.6% (w/v) Brij 58, and 60/~g/ml DNase I. (The sodium deoxycholate should be from a fresh 1% stock solution.) Incubate at 4 ° for 10 min 4. Pellet the cellular debris in an Eppendorf centrifuge at full speed for 10 rain. Freeze aliquots at -70* 5. For RNA, extract lysate with 0.1% SDS and phenol/chloroform

704

6F.NETICS

[47]

transcribed as a single primary transcript that must undergo extensive processing to become the mature products. Processing requires all of the ribosomal proteins and various ribonucleases and RNA modification enzymes. The high level of UV irradiation required for analysis of translation products of plasmid-borne genes in maxicells would prevent rRNA processing since the required proteins would not be synthesized. Fortunately, low UV fluency (e.g., 10 sec for CSR603) is sufficient to prevent host rRNA transcription but not production of host mRNAs coding for proteins required for rRNA processing. The reason for this is that the host chromosome is probably not extensively degraded following the relatively moderate UV treatment. However, transcription of rDNA in vitro and in vivo is particularly sensitive to inhibition of DNA gyrase, ~7,~8and superhelicity of rDNA is probably required for it to be expressed efficiently. Therefore it is possible that treatment of plasmid-containing strains with low UV fluency causes preferential relaxation (rather than degradation) of the host chromosome, thereby accounting for the specificity of rRNA transcription of the plasmid-coded rDNA while maintaining the ability to synthesize host-coded proteins. The maxicell system described here has been used successfully to specifically label plasmid-coded rRNA and study the effects of mutations on processing of rRNA and assembly into subunits. While much of our research has been on plasmid systems derived from the multicopy plasmid pBR322, the maxicell technique is also applicable to lower-copy-number vector systems. For example, we are using a low-copy-number plasmid, pEJM007, derived from pDPT487 and NRI, 6,s9 and have no difficulty in specifically labeling the rRNA using the maxicell technique as described here (see Fig. 2). The vectors pKK3535 and pEJM007 have rrnB operons with the wildtype promoters P l and P2. Conditional expression systems are also available in which these promoters are replaced by the 2 promoter, PL,7 or the T7 promoters as described above. Using the PL promoter, expression can be controlled by the temperature-sensitive 2 repressor protein, CI857. At 30 °, the permissive temperature for repression, transcription of the gene is blocked by the repressor while at elevated temperatures (e.g., 42 °) thermal denaturation of the repressor allows transcription to occur. We have used the maxicell technique to label transcripts whose production is controlled b y PL. One may use an E. coli strain (e.g., K5637, M5219, or N4830) in which the gene for CI857 is integrated into the host chromosome and the J7 H. Yang, K. Heller, M. Gellert, and G. Zubay, Proc. Natl. Acad. Sci. U.S.A. 76, 3304 (1979). ~8B. Oostra, A. Van Vliet, G. Ab, and M. Gruber, J. Bacteriol. 148, 782 (1981). ~9D. Taylor and S. Cohen, J. Bacteriol. 137, 92 (1979).

[47]


705

cells produce repressor protein at the level of a single-gene dose. Alternatively, the repressor gene may be introduced into cells such as HB101 on the multicopy plasmid pCI857. 2° This plasmid is compatible with the rrnB-containing plasmids described above and may be maintained stably in the same cell. In this construction the plasmid-coded repressor protein is produced in larger quantities and repression is very effective. The maxicell procedure is as described above. The cells are propagated, irradiated, and allowed to recover at 30 °, the permissive temperature for repression. After the recovery period, however, the cells are shifted to 42 ° and radioisotope is added. At 42 ° the repressor is no longer active, thus allowing transcription to occur from the PL promoter. The T7 maxicell system described here has several advantages over the UV maxicell system. It requires only the addition of chemicals (IPTG and rifampicin) rather than UV irradiation. It is much faster: 1.5 rather than 15 hr. It gives a greater yield of plasmid-coded rRNA; liters ofT7 maxicells may be used in contrast to UV irradiation of 10 ml. Also significant is the high level of transcription by T7 RNA polymerase. The one negative aspect of the T7 system is the reliance on stable mRNA to provide ribosomal proteins for processing and assembly of the rRNA. However, this appears to be quite adequate for the first 30 to 40 min after addition of rifampicin,s While we point out differences between the two maxicell systems with conditional expression plasmids (T7 and PL), they are both equally valuable techniques for providing important information about the processing and assembly of rRNA mutants which, in nonrepressed plasmid systems, would be lethal to the cell. In summary, we have described a modified maxicell technique which accomplishes specific labeling of transcripts of rDNA cloned on plasmid vectors. The transcripts are processed normally and incorporated into ribosomal subunits. The modified maxicell technique may also be used in a vector-host system that allows controlled expression using the PL promoter and a temperature-sensitive repressor protein. A second technique, the chemically induced maxicell system, is also described which uses a T7 late promoter and a cloned T7 RNA polymerase gene controlled by IPTG. The advantages of each system have been discussed. Appendix

Culture Media I. ZPM: 6.06g Tris base, 0.5g sodium citrate, 0.15g KC1, 0.1 g MgSO4" 7H20 per liter and 2/gM FeC13, pH 7.4, with HC1. 20 E. Remaut, H. Tsao, and W. Fiefs, Gene 22, 103 (1983).

706

GENETICS

[48]

II. Phosphate-free casamino acids, 10% stock solution: Add 100 g casamino acids to 500 ml H20, and add MgCI2 to 50 mM. Adjust the pH to 8.4 with ammonium hydroxide and place the solution on ice for 1 to 2 hr. The resulting precipitate is filtered and the pH adjusted to 7.2 with HC1. The solution is adjusted to 1 liter and sterilized by autoclaving. Solutions of 10% casamino acids may have phosphate concentrations around 20 m M before treatment and 20/zM after treatment.

[48] T e m p e r a t u r e - S e n s i t i v e M u t a n t s w i t h A l t e r a t i o n s in Ribosomal Protein L24 and Isolation of Intra- and Extragenic Suppressor Mutants By JOACHIM SCHNIER and KAYOgO NISHI Ribosomal protein L24 has been shown to be one of the first proteins involved in the assembly of 50S subunits in vitro. ~ It binds to the 5'-end of 23S rRNA. 2 However, not very much is known about important domains of the protein nor its precise interaction with ribosomal RNA or other proteins in vivo. Here we describe an approach for the analysis of mutants in the gene rplX for ribosomal protein L24 in vivo. It should be also applicable to other ribosomal proteins. A Nonsense Mutant in rp/X Two different mutants in the gene rplX for ribosomal protein L24 were characterized and used for further mutant isolations. They had been isolated as spontaneous independent mutants from antibiotic-dependent strains. 3,4 Mutant TAI09 has a nonsense codon which alters a codon AAA for lysine at position 20 to the stop codon TAA, resulting in the complete lack of a functional protein. 5 This mutant grows very slowly and gives rise to temperature-sensitive growth. The other mutant, KNS 19, shows an alteration of a ~ codon for glycine to a GAC codon for aspartic acid at position 84 in the protein. This mutant is temperature sensitive in growth.6 1 V. Nowotny and K. Nierhaus, Proc. Natl. Acad. Sci. U.S.A. 79, 7238 (1982). 2 p. Sloof, J. B. Hunter, R. A. Garrett, and C. Branlant, Nucleic Acids Res. 5, 3503 (1978). 3 E. R. Dabbs, J. Bacteriol. 140, 734 (1979). 4 E. R. Dabbs, Mol. Gen. Genet. 187, 453 (1982). s K. Nishi, E. R. Dabbs, and J. Schnler, J. Bacteriol. 163, 890 (1985). 6 K. Nishi and J. Schnier, E M B O J., in press (1986). METHODS IN ENZYMOIA3GY, VOL. 164

Copyright © 1988 by Academic Pr¢~ Inc. All rights of reproduction in any form reserved.

706

GENETICS

[48]

II. Phosphate-free casamino acids, 10% stock solution: Add 100 g casamino acids to 500 ml H20, and add MgCI2 to 50 mM. Adjust the pH to 8.4 with ammonium hydroxide and place the solution on ice for 1 to 2 hr. The resulting precipitate is filtered and the pH adjusted to 7.2 with HC1. The solution is adjusted to 1 liter and sterilized by autoclaving. Solutions of 10% casamino acids may have phosphate concentrations around 20 m M before treatment and 20/zM after treatment.

[48] T e m p e r a t u r e - S e n s i t i v e M u t a n t s w i t h A l t e r a t i o n s in Ribosomal Protein L24 and Isolation of Intra- and Extragenic Suppressor Mutants By JOACHIM SCHNIER and KAYOgO NISHI Ribosomal protein L24 has been shown to be one of the first proteins involved in the assembly of 50S subunits in vitro. ~ It binds to the 5'-end of 23S rRNA. 2 However, not very much is known about important domains of the protein nor its precise interaction with ribosomal RNA or other proteins in vivo. Here we describe an approach for the analysis of mutants in the gene rplX for ribosomal protein L24 in vivo. It should be also applicable to other ribosomal proteins. A Nonsense Mutant in rp/X Two different mutants in the gene rplX for ribosomal protein L24 were characterized and used for further mutant isolations. They had been isolated as spontaneous independent mutants from antibiotic-dependent strains. 3,4 Mutant TAI09 has a nonsense codon which alters a codon AAA for lysine at position 20 to the stop codon TAA, resulting in the complete lack of a functional protein. 5 This mutant grows very slowly and gives rise to temperature-sensitive growth. The other mutant, KNS 19, shows an alteration of a ~ codon for glycine to a GAC codon for aspartic acid at position 84 in the protein. This mutant is temperature sensitive in growth.6 1 V. Nowotny and K. Nierhaus, Proc. Natl. Acad. Sci. U.S.A. 79, 7238 (1982). 2 p. Sloof, J. B. Hunter, R. A. Garrett, and C. Branlant, Nucleic Acids Res. 5, 3503 (1978). 3 E. R. Dabbs, J. Bacteriol. 140, 734 (1979). 4 E. R. Dabbs, Mol. Gen. Genet. 187, 453 (1982). s K. Nishi, E. R. Dabbs, and J. Schnler, J. Bacteriol. 163, 890 (1985). 6 K. Nishi and J. Schnier, E M B O J., in press (1986). METHODS IN ENZYMOIA3GY, VOL. 164

Copyright © 1988 by Academic Pr¢~ Inc. All rights of reproduction in any form reserved.

[48]

TEMPERATURE-SENSITIVE MUTANTS 20 AAA (K)

ATGI

56 GGC (G)

aAC (D) GAA TCA TTA (E)

(S)

(L)

84 ,,GGC (G)

707

ITAA

la-XCI (D) GAG (E)

FIG. 1. Mutations in the gene rplX for ribosomal protein L24.

A mutant with a similar alteration as KNS 19 at position 56 did not show any apparent phenotype and was not used further. All mutations are summarized in Fig. 1. Various temperature-resistant mutants were isolated from the nonsense mutant TA 109 and analyzed. All of them had the common feature that an L24 protein was again detectable in two-dimensional gels. In most of the cases, the migration of these L24 proteins was different from wild type, indicating a variety of pseudorevertants. Revertants which grew slowly at 42 ° had acquired a mutation which was mapped at a different locus from the gene rplX for protein L24. 4 It was suggested that these may be located in genes suppressing nonsense codons, since the original mutation was still maintained. Other fast-growing revertants were mapped within the gene rplX. In order to see the alteration, it was necessary to clone the mutant genes in a simple and fast way. Cloning Strategy for rp/X Missense M u t a n t Genes As mentioned above, the nonsense mutant, TA109, grew very slowly (240 min for the mutant versus 40 min for wild type). Complementation analysis had revealed that faster-growing transformants could be obtained by plasmids containing the wild-type gene for protein L24 that is located on a 1.9-kb DNA fragment. As a vector we used either the mini-F-plasmid pRE4327 with a low copy number or plasmid pACYC184. 8 The doubling time of TA 109 decreased to 90 min using plasmids containing the wildtype gene for protein L24. We, therefore, tried to clone several missense mutants in rp/X by positive selection using faster growth as a phenotype. In fact, when chromosomal DNA fragments from several rplX mutants were cloned into one of the above-mentioned plasmids and TA109 was transformed, we observed some fast-growing colonies. All of these contained 7 R. Eichenlaub, in "Recombinant DNA Research and Virus" (Y. Becker, ed.), p. 39. Martinus Nijhoff, Boston, 1985. s A. C. Y. Chang and S. N. Cohen, J. Bacteriol. 134, 1141 (1978).

708

GENETICS

[48]

plasmids with the corresponding rplX missense mutant fragments. Sequence analysis showed mutations as they are summarized in Fig. 1. This method was also successfully applied to isolate the L24 gene of the missense mutant KNS 19 which showed a temperature-sensitive phenotype. In turn, this mutant was used for cloning various other missense mutant genes for L24 by selecting temperature-resistant growth. Since the L24 gene is located within a number of other ribosomal protein genes, 9 the mutants TA109 and KNS 19 could also be used to isolate mutant genes for other ribosomal proteins by a similar approach. Isolation of Spontaneous r R N A M u t a n t s In contrast to the nonsense mutant, the ts- missense mutant KNS 19 (C,C_~ to GAC at position 84) gave rise to not only intragenic revertants but also a number of extragenic suppressor mutants whose isolation will be described now. A similar missense mutant with a ts- phenotype was described l° which showed a defect in the 50S subunit assembly at 42 °. At the same time, the altered protein L24 showed a decreased affinity to 23S rRNA at all temperatures. We, therefore, thought that it might be possible to isolate temperature-resistant suppressor mutants in an rRNA gene. We made use of the fact that a plasmid, pKK3535, H containing the rrnB operon represses the synthesis of rRNA coded by chromosomal genes, 12 so that most of the rRNA in newly synthesized ribosomes is derived from the plasmid. We transformed mutant KNS 19 with plasmid pKK3535 and selected clones which grew at the nonpermissive temperature 42 °. It turned out to be important to select directly on agar plates containing ampicillin and at the same time at 42 °. By this way, chromosomal mutants which usually occurred at a high frequency could be largely excluded, since only clones containing plasmids could grow on ampicillin plates. After incubation overnight, we obtained two clones from which we isolated plasmid DNA. We retransformed mutant KNS 19 with both plasmids and examined the number of transformants at 30 ° and at 42 °. Both plasmids gave rise to a similar number of transformants at both temperatures. We concluded that the mutation(s) had occurred in the plasmids. The mutations were then localized in the Y-end of the 23S rRNA by DNA sequence and complementation analysis. From the location of the mutations a model was 9 S. R. Jaskunas, A. M. Fallon, and M. Nomura, J. Biol. Chem. 252, 7323 (1977). 1oT. Cabezon, A. Herzog, J. Petre, M. Yaguchi, and A. Bollen, J. Mol. Biol. 116, 361 (1977). 1~j. Brosius, T. J. Dull, D. D. Sleetgr, and H. F. Noller, J. Mol. Biol. 148,107 (1981). t2 S. Jinks-Robertson, R. L. Gourd, and M. Nomura, Cell 33, 865 (1983).

[48]

TEMPERATURE-SENSITIVE MUTANTS

709

deduced. 6 For the first time, a protein-rRNA interaction could be shown genetically. Isolation of Other Extragenic Suppressor M u t a n t s We further isolated and localized extragenic chromosomal suppressor mutants from mutant KNS 19 which showed temperature-resistant growth. For isolation we used two different media. One type of agar plates was rich medium containing L Broth (per liter: l0 g tryptone, 5 g yeast extract, 5 g NaC1) and the other type was minimal medium containing AN salts without NaCl) 3 In this way, we obtained three different extragenic suppressor mutants which were distinguished by their location on the Escherichia coli chromosome. The two mutants selected on the rich medium grew on both media and the one mutant selected on minimal medium grew only on the selection medium. The latter mutant gave rise to mucoid growth and showed UV sensitivity. The mutation was located in or close to the Ion gene coding for protease LA. It could be shown that this mutant is unable to induce the lon gene product at high temperature. We could further confirm that another lon mutation which was introduced into the mutant showed suppressor activity only on plates with minimal medium (K. Nishi and J. Schnier, unpublished observations). We have not yet characterized the other suppressor mutants, so their identity, apart from the map position, is still unknown. In summary, we have shown three ways of isolating mutant genes in ribosomal proteins, rRNA genes, and genes coding for proteins which are functionally correlated with ribosomal proteins. (1) A simple cloning system for cloning ribosomal protein genes was developed. This system can be used for the fast isolation of a number of mutant genes for ribosomal proteins in the neighborhood of the rplX gene for protein L24, like the S 10 or spc operon. (2) A new approach to isolate specific mutations in the ribosomal RNA was introduced. This system is based on a positive selection and could also be used in combination with mutagens. (3) We were able to obtain a number of suppressor mutants of a temperature-sensitive ribosomal mutant that map in genes other than ribosomal protein genes. This method is based on a simple selection using different media. Acknowledgments We thank Dr. H. G. Wittmann for support and thank Dr. E. R. Dabbs for some of the mutants as well as stimulating discussions. ~3B. D. Davis and E. S. Mingioli, J. Bacteriol. 60, 17 (1950).

7 10

GENETICS

[49] P r i m e r - D i r e c t e d

By BIRTE

[49]

Deletions in 5S Ribosomal RNA

VESTER, J A N EGEBJERG, R O G E R G A R R E T T , J A N CHRISTIANSEN

and

Structural and functional studies of 5S RNA, and of its protein complexes, are both important and topical, because of a wide general interest both in mechanisms of protein-RNA recognition and in RNA structurefunction relationships. 5S RNA and its protein complexes are essential for ribosomal function. Parts of the RNA molecule are very accessible on the ribosomal surface and the molecule contains universally conserved nucleotides that may have critical structural and/or functional importance. However, the precise role of the RNA remains to be clarified.1 At present, we have good minimal secondary structural models for eubactefial, archaebacterial, and eukaryotic 5S RNAs based on both phylogenetic and experimental evidence.2 All probably exhibit five main double-helical segments but their tertiary structure remains undetermined. The RNAs show some kingdom-specific features in their secondary structures2 and, also, in their patterns of conserved nucleotides that may be of functional importance. Only the protein-binding sites on the eubactedal 5S RNAs have been mapped in detail1,3; those of the eukaryotic ribosomal proteins and also of the transcription factor TF IIIA have been mapped partially.4,5 In summary, we are left with resolving the RNA tertiary structure, the chemical basis of protein-RNA interactions, and the details of structure-function relationships at a nucleotide level. Site-directed mutagenesis offers an important approach to such problems. For example, the effects of mutations on the conformation of the RNA may yield important insight into the higher-order structure of the RNA, and nucleotides thought to be involved in protein recognition can be altered, or deleted, and their effects on protein binding investigated. Moreover, conserved nucleotides that may be important functionally can be changed or eliminated and their effect on function examined. It is important though to choose your deletion carefully since the amount of work involved in producing and characterizing the mutant RNA is considerable. i R. A. Garrett, S. Douthwaite, and H. F. NoUer, Trends Biochem. Sci. 6, 137 (1981). 2 G. E. Fox, in "The Bacteria" (C. Woese and R. Wolfe, eds.), Vol. 8, p. 257. Academic Press, Orlando, Florida, 1985. 3 j. Christiansen and R. A. Garrett, in "Structure, Function and Genetics of Ribosomes" (B. Hardesty, ed.), p. 253. Springer-Verlag, New York, 1986. 4 p. W. Huber and I. G. Wool, J. Biol. Chem. 261, 3002 (1986). s j. Christiansen, R. S. Brown, B. S. Sproat and R. A. Garrett, E M B O J . 6, 453 (1987). METHODSIN ENZY'MOLOGY,VOL. 164

C o p ~ t © 1988by AcademicPr¢~, Inc. Allright5ofreproductionin any formreserved.

~ I

Sou3A 5SRNA Sou 3A

ISOLATE SAU3A FRAGMENT I

5s ~ ~

DIGEST, BAMHI LIGATE, TRANSFECT SCREEN FOR ORIENTATION

Isolate ssM13 DNA

I ANNEAL PRIMER

EXTEND WITH KLENOWFRAGMENT LIGATE 12-15 °, 16 H

0I

IGEST HINDIII / EC_O_ORI ISOLATE FRAGMENT

5S ~ T2

~ LIGATE

(( pKK 3535 "i

P,

,>

'P2~-'.---.'J'

H i.._.n.nd TIT

0:

S RNA

TRANSFORM SCREEN BY COLONY HYBRIDIZATION SEGREGATE MUTANT PLASMID

Sequence (DNA or RNA) FIG. l. General strategy for effecting primer-directed deletions in 5S RNA. pKK3535 contains the rrnB operon of E. coil as constructed by Brosius et al. 2°

712

GENETICS

[49]

Below we present a detailed procedure (illustrated in Fig. 1) for generating and characterizing deletion mutants in Escherichia coli 5S RNA. There follows a more detailed appraisal of the 5S RNA expression system in plasmids, deletion repair mechanisms in the cell, the possible occurrence of DNA rearrangements, and methods for determining structural and functional effects of the mutations. We conclude with a section on special problems associated with producing mutations in the large rRNAs of

E. coli. Experimental Procedure

Phosphorylation of the Oligonucleotide Hybridization Probe 1. Dry down 20 #1 [?-32p]ATP (3000 Ci/mmol). 2. Dissolve in 10/zl buffer A [50 mMTris-HCl, pH 9.5, 10 mMMgCI2, 5 m M dithiothreitol (DTT), 0.4% glycerol] containing 7 pmol oligonucleotide and 1 unit T4 polynucleotide kinase (P-L Biochemicals). 3. Incubate at 37 ° for 35 min.

Mutagenic Primer 1. Dissolve 200 pmol oligonueleotide in 20/tl buffer A containing 0.1 m M ATP and 3.5 units T4 polynucleotide kinase. 2. Incubate at 37 ° for 35 min.

Isolation of Phosphorylated Oligonucleotide The mutagenic primer is phosphorylated prior to ligation. This also has the advantage that the difference in mobility between the phosphorylated and nonphosphorylated form in polyacrylamide gels can be exploited during purification of the former. An aliquot of labeled probe is added, prior to electrophoresis, to facilitate detection and the remainder can be run in the same gel to remove [~,-32p]ATPand thus the background in the screening procedure: 1. Add an aliquot (0.5 pl) of the labeled hybridization probe to the solution of phosphorylated mutagenic primer. 2. Electrophorese the samples (mutagenic primer and labeled probe) on a 20% polyacrylamide-7 M urea gel (50 m M Tris-borate, pH 8.3, 1 m M EDTA). 3. Autoradiograph for 2 - 5 rain. 4. Excise bands and soak each in 200/tl H20 overnight to extract the oligonucleotide. Care should be taken not to excise the nonphosphorylated oligonucleotide.

[49]

PRIMER-DIRECTED DELETIONS IN 5S RIBOSOMAL R N A

713

Purification of the M13 Template 1. Innoculate 1.5 ml LB medium (10 g tryptone, 5 g yeast extract, 5 g NaC1 per liter) in a 10-ml tube with 15/tl fresh overnight culture of JM 101 cells. Incubate for 30 min. 2. Toothpick a single plaque into the 1.5 ml of innoculated medium and incubate by shaking vigorously at 37 ° for 5 hr. 3. Spin for 5 min at 8000 rpm and room temperature and transfer 1.2 ml of the supernatant to an Eppendorf tube. 4. Add 150/zl 25% PEG-6000, 3 M NaC1, and mix by inverting the tube several times. 5. Leave for 10 min at room temperature. Spin for 5 min in a microfuge and discard the supernatant. 6. Dissolve the pellet in 100/~1 T2oE~ buffer (20 m M Tris-HCl, pH 7.5, 1 m M EDTA). 7. Extract twice with phenol (saturated in TE) and twice with chloroform. 8. Add sodium acetate to 0.25 M, precipitate with 2.5 volumes ethanol, wash, and dry. 9. Dissolve the pellet in 25/tl T~0Eo.~buffer, pH 7.5, quick-freeze, and store at - 80 °.

Primer Anneafing 1. Mix 3/A M13 template (-0.4 pmol), 20#1 phosphorylated mutagenic primer (-20 pmol), and 2/tl 3 M sodium acetate followed by 68/zl ethanol. (This step also removes urea from the primer.) 2. Precipitate, wash, and dry. 3. Dissolve the pellet in 20/~1 annealing buffer (10 m M Tris-HCl, pH 8.0, 10 m M MgCI2, 50 mMNaC1). 4. Place the Eppendorf tube in a small beaker of hot water (70°), and cool slowly to room temperature ( - 3 0 - 4 5 min).

Extension and Ligation To the annealing solution, add 1. 7.5/11 4X extension/ligation buffer (13 m M Tris-HC1, pH 8.0, 13 m M MgCI2, 20 m M DTT, 0.7 m M dNTPs, 4 m M ATP). 2. 5 Weiss units DNA ligase (2/A 2.5 U//d, Amersham). 3. One unit Klenow fragment (0.3/tl 3 U/ld, Amersham). Then mix and incubate at 12- 15 ° for 16 hr.

714

GENETICS

[49]

Isolation of Heteroduplex Fragment 1. Add 70/~1 T2oEl0 buffer, pH 8.0. 2. Extract twice with phenol, once with chloroform, and precipitate with ethanol. 3. Redissolve the dried pellet in 20/d of the appropriate restriction enzyme buffer. 4. Add 1-5 units of restriction enzymes (e.g., HindlII and Eco RI; see Fig. 1). Digest for 1- 2 hr. 5. Add 2/tl 20% Ficoll loading buffer and electrophorese in 0.8-1.0% agarose gel containing 0.5 #g/ml ethidium bromide. 6. Excise fragment, crush gel piece, and freeze at -70* for 5 min, thaw at 37°; add one-third volume phenol and shake for 2.5 hr. 7. Centrifuge and extract the aqueous phase twice with chloroform to remove ethidium bromide. 8. Ethanol precipitate and redissolve in 20/zl TloEo.l buffer, pH 7.5.

Ligation into pKK 3535-Derived Expression System The 5S RNA expression system is considered in detail in a separate section below. 1. Mix 4/A of fragment and 50 ng vector in a total volume of 10/tl ligase buffer (50 m M Tris-HC1, pH 7.4, 10 m M MgC12, 10 m M DTT, 1 m M spermidine, 1 m M ATP). 2. Add 0.2 Weiss unit DNA ligase. 3. Ligate at 8 - 10* for 5 - 12 hr.

Transformation Any transformation procedure is probably adequate, but owing to the ease of using frozen competent cells we recommend the method described by Hanahan. 6 The transformed cells are plated on agar containing 50/~g/ ml ampicillin.

Colony Immobilization on a Filter 1. When colonies are 0.5-1 m m in diameter, place a filter (Gene Screen) on the surface of the agar plate and avoid trapping bubbles. 2. Transfer the filter with colony-side up to an agar plate containing chloramphenicol (150/tg/ml) and incubate at 37 ° overnight. 3. Incubate the master plate at 37* until colonies reappear. 6 D. Hanahan, in "DNA Cloning" (D. M. Glover, ed.), Vol. 1, p. 109. IRL Press, Oxford, England, 1985.

[49]

PRIMER-DIRECTEDDELETIONS IN 5S RIBOSOMAL R N A

715

4. Place the replica filter with colony-side up on Whatman 3 MM paper saturated with 0.5 M NaOH, 1.5 M NaCI, and leave for 15 min at room temperature. 5. Transfer filter to Whatman 3 M M paper saturated with 1 M TrisHC1, pH 7.0, 1.5 MNaC1 for 2 - 3 min. 6. Immerse filter in 2 × SSC buffer, pH 7.2 (20 × SSC: 3 M NaC1, 0.3 M sodium citrate) for 15 sec. 7. Air-dry and bake filter at 80 ° for 2 hr. 8. To remove bacterial debris, wash the filter in 3 × SSC containing 0.1% sodium dodecyl sulfate (SDS) using 50 ml/filter at 65* for 1 hr. 9. Change solution twice and wash at 65 ° for 0.75 hr both times (eventually rub filters with fingers, wear gloves). 10. Rinse the filter with 200 ml 2 X SSC. The filter can be stored dry at 4 ° .

Screening by Hybridization 1. Use 2 ml of hybridization solution per filter. The solution can be made up from 0.2 ml 50 × Denhardt's solution (1% Ficoll, 1% polyvinylpyrrolidone, 1% bovine serum albumin), 0.4 ml 30× TSE buffer (0.45 M Tris-HC1, pH 7.5, 4.5 M NaCI, 0.3 M EDTA), 1 ml 20% dextran sulfate, 0.2 ml 5% SDS, and 0.2 ml hybridization probe. 2. Hybridize overnight ( 14-18 hr) at an appropriate temperature. Hybridization temperature (T) = (number of G-C bp) X 4 ° + (number of A-T bp) X 2 ° - 4 °. If the calculated temperature is higher than 55 °, then hybridize at 50 ° . 3. Wash the filter at room temperature with 150 ml of 6 × SSC three times for a total of 20 min. 4. Air-dry and mark the filter. 5. Autoradiograph the filter for 2 - 6 hr using an intensifying screen. 6. The positive colonies should appear as dark spots on the film. If there are no clear differences in the spot intensities, wash the filter with 6× SSC at Td = (numbers of G-C bp) X 4 * + (numbers of A-T bp) × 2 °. (The dissociation temperature is not, as stated in the "Wallace Rule," a simple function of base content, but also a function of sequence.) If there are still no differences then wash at increasing temperatures. 7. Pick out positive colonies and isolate plasmid DNA.

Segregation of Mutant Plasmid The segregation step is important because ~50% of the plasmids are wild-type.

716

GENETICS

[49]

1. Transform competent cells under dilute DNA conditions to avoid more than one plasmid entering the same cell. 7 2. Either screen the colonies as described above or pick 10-20 colonies onto a master plate and use this for screening. Isolation o f 5 S R N A

1. Grow to a cell density of A650 -- 0.5 in LB medium. 2. Add adenosine + uridine (300 #g/ml) and thiamin (6/zg/ml), chloramphenicol (100 #g/ml). 3. Amplify while vigorously shaking for 8 hr. 4. Harvest cells, wash in 100 m M sodium acetate, pH 5.0, 10 m M magnesium acetate, and resuspend in washing buffer (one-fortieth of a volume of growth medium). 5. Add one volume phenol saturated in 0.3 M sodium acetate, pH 6.0, vortex for 1 min, and spin for 5 min at 5000 rpm. 6. Reextract the aqueous layer three times with phenol and once with chloroform and precipitate with ethanol. 7. Purify the 5S RNA on a 10% polyacrylamide-7 M u r e a gel containing 50 m M Tris-borate, pH 8.3, 1 m M E D T A . Sequencing

The mutation can be verified either by sequencing the DNA s,9 or the RNA.l°,ll Expression Plasmid for 5S R N A Figure 2 depicts the coding region and flanking sequences of the 5S RNA expression plasmid that employs the promoters and terminators of the rrnB operon of E. coli. This construction yields about 2 mg of mutant 5S RNA per liter of culture, in the presence of chloramphenicol, after extraction and polyacrylamide gel purification. Excision of the 16S RNA and most of the 23S RNA genes eliminates regulation by ribosome feedback, ~2 thus allowing expression to occur according to the gene dosage. A similar expression plasmid was developed by Brosius and Noller (unpub7D. Hanahan,J. Mol. Biol. 166, 557 (1983). s F. Sanger,A. R. Coulson,B. G. Barrell,A. J. H. Smith, and B. A. Roe, .1. Mol. Biol. 143, 161 (1980). 9A. M. Maxamand W. Gilbert,this series, Vol. 65 [57]. loD. A. Peattieand W. Gilbert, Proc. NatL Acad. Sci. U.S.A, 77, 4679 (1980). HH. Donis-Keller,A. M. Maxam,and W. Gilbert,Nucleic Acids Res. 4, 2527 (1977). 12S. Jinks-Robertson,R. L. Gourse, and M. Nomura,Ce1133, 865 (1983).

[49]


100bp

HindlTl B o m H I

Sau3A

V/A V / / / / / / / / / / / / / / / / / / / , ' ~ P1

P2

16S

t

M13 t i n k e r

23S

717

EcoRI

[////,4 SS

T1

T2

t

pBR 322

M13 l i n k e r

FIG. 2. Map of the coding region and flanking sequences in the 5S RNA expression plasmid. Ps and Ts indicate the promoters and terminators, respectively, that occur in the rrnB operon ofE. coli.

lished results). The distance between promoters and the 5S RNA gene is not critical as long as the primary transcript is able to adopt a structure recognized by RNase E. ~3 Thus, the 80 bp in front of the 5S RNA gene should be intact, otherwise 5'-end processing may prove dit~cult. 5S RNA expressed on this plasmid exhibits a dinucleotide extension at the Y-end and a heterogeneous 5'-end. This incomplete processing results from chloramphenicol inhibition of protein synthesis. The terminal extensions are absent if 5S RNA is isolated from nonamplified cells, but then the proportion of wild-type RNA will increase. The amount of chromosomecoded wild-type RNA in the mutant RNA, isolated from amplified cells, is about 10%. Primary transcripts are processed by RNase E in the absence of protein synthesis. Therefore, a particular mutant RNA will be produced even if it has lost its ribosomal protein-binding ability. Thus a low yield of mutant RNA probably reflects a decreased stability in vivo. Some desired mutations may seriously impede ribosomal function and, therefore, cell growth. Then it will be necessary to back-clone the heteroduplex that is constructed in vitro, as nonexpressed DNA and, subsequently, to express the mutant DNA after a second cloning step and establish whether, indeed, the mutation has adverse effects in vivo. If a mutation is deleterious in vivo or, alternatively, if it is very important to obtain pure and mature 5S RNA, it may be an advantage to employ an inducible promoter based on the T7 RNA polymerase system ~4 or the temperature-inducible PL system from phage 2.15 Thus even unstable 5S RNA deletion mutants, which cannot be purified in the expression system based on ribosomal promotors, can be transcribed in the in vivo T7 polymerase system. They can be purified in 5-10-fold higher amounts than stable 5S RNA from the above mentioned system. The purity of the mutant 5S RNA from the T7 promotor system is also higher (> 95-98%). 13 M. K. Roy, B. Singh, B. K. Ray, and D. Apirion, Eur. J. Biochem. 131, 119 (1983). ~4F. W. Studier and B. A. Moffatt, J. Mol. Biol. 189, 113 (1986). t5 R. L. Goursc, Y. Takebe, R. A. Sharrock, and M. Nomura, Proc. Natl. Acad. Sci. U.S.A. 82, 1069 (1985).

718

GENETICS

[49]

The ends of the mutated 5S RNAs are extended as occurs for chloroamphenicol-amplified cells (unpublished observations). Efficiency of Mutagenesis and Mismatch Repair Although several procedures have been described that yield high mutation frequencies for nucleotide substitutions, we have always observed very low frequencies when they are applied to nucleotide deletions. To obviate this problem we employed the back-cloning procedure (Fig. l) that reduces the methylation bias caused by dam mcthylation of the template strand.t6 In addition, it both prevents incorrect primer annealing to the M 13 vector and enables nonlethal mutations to be introduced directly into the expression system. Nevertheless, there seem to be differences between the repair of the template and of the in vitro synthesized strand, probably because GATC sites occur within the back-cloned fragment. After back-cloning into an M 13 vector we obtained substitution frequencies (using a 17-mer) in mutL strains of about 50%, while deletion frequencies (also using a 17-mer) were less than 1%. The substitution frequency in mutL + strains was 3 - 5%. Therefore, the mutL product must play a role in the repair of substitutions but not deletions. Back-cloning into a pKK3535 derivative and transforming into HB101 (recA-) cells gave a higher deletion frequency of about 3%; this suggests that there may be an advantage in back-cloning into plasmids rather than M 13 vectors.17 A general back-cloning procedure has been described that employs double primersJ s In our hands, this yields substitution frequencies of less than 1% for 5S RNA, which probably reflects poor annealing of the primer to the rDNA part of the M 13 ssDNA. This in turn is probably caused by the formation of very stable secondary structure in the single-stranded rDNA. DNA R e a r r a n g e m e n t Artifacts After constructing a desired mutant and expressing the mutant RNA, the altered DNA or RNA sequence-- preferably both - - should be verified. Reliance on the initial hybridization screening, especially for deletions, is inadequate because such mutational events can lead to gross DNA rearrangements. For example, an attempt to remove nucleotide G69 resulted ~6B. Kramer, W. Kramer, and H. J. Fritz, Cell 38, 879 (1984). ~7There are now commercially available mu~genesis kits which circumvent the mismatchrepair systems. ~s K. Norris, F. Norris, L. Christiansen, and N. Fiil, Nucleic Acids Res. 1 I, 5103 (I 983).

[49]


719

in a large deletion of both the 3'-end of the 5S RNA gene and the terminators. Cells expressing this plamid grew very slowly, probably due to runaway transcription that interfered with plasmid replication. Another mutagenesis experiment, designed to delete A66, yielded a 50-bp insertion. A partially duplicated 5S RNA was produced that was stable in vivo; its sequence and putative secondary structure are depicted in Fig. 3. Structure- Function Effects of Mutations Two general approaches can be invoked to investigate changes in structure or function caused by a particular mutation. The in vitro approach includes gel electrophoresis to examine conformational changes that will

~ so c c CAC c

C C

C C A /,o UG C

u G,A,u

5'

3,u PUcc

t

~

/ A "GP'--'C, U G 120 I"1 A (y

~c

Gu~

r,G

U

,

~ ~°

c

CAAG

A

, ~C ',..C~/A.

'~ C " I ° C A G U A c G ' ~ ' ."o G , GG ~ GUc, cG~ 11o" A ./G..~ C A- -~.yyG c .~o

AC

G ~o '~AG u

G - C ~as

G- C AG- C loo

U- A

5o I

A G AG'c-3o G CU

UA c C "

90

C

uCAGA

A

G~

, U - uG" .Gcc c.,",,•

U G' G eo ~" A u'G,C,c~ u G U'A~ AGc GC CG 70

FIG. 3. Sequence and putative secondary structure of a partially duplicated 5S RNA containing one A66 deletion (A) and one nondelction (~). Numbers refer to the 5S RNA sequence and nucleotides 2 5 - 7 2 (except A66) are duplicated. Two noncoded uridines that occur between the two repeated sequences are boxed. Helices I, II, and III of the general 5S RNA structure can form. The secondary structure is supported by probing data with ribonu-

cleases.

720

GENETICS

[49]

be reflected in mobility shifts, followed by probing of the RNA with chemicals and/or ribonucleases. Mutant RNA conformers can also be examined by spectroscopic methods such as nuclear magnetic resonance, infrared spectroscopy, and circular dichroism. A simple development of such studies would be to test the mutant's ability to interact with ribosomal proteins LS, L18, and L25 and to draw inferences concerning the modes of protein-RNA interaction. A straightforward in vivo approach is to examine the effect of the mutation on growth rate, preferably in both rich and minimal media. This could then lead on to elucidating any functional effects of the mutant RNA after assembly into polysomes. Mutations in Large rRNAs We have also made mutations in the large ribosomal RNAs of E. coli, in particular nucleotide substitutions, within the central part of domain V of 23S RNA that is associated with peptidyl transfer. Apart from the general problems of site-directed mutagenesis these studies also revealed an additional problem, specific to the large RNA genes, that relates to the back-cloning of a mutagenized DNA fragment. Owing to the lack of appropriate unique restriction sites it is difficult to back-clone the fragment into its exact position in the rRNA operon. This problem can be overcome by making the mutations directly in a gapped-duplex form of the plasmid. ~9 pKK35352° is digested with a restriction enzyme (PvulI) to yield two unequal fragments. The smaller one is religated to form a small plasmid containing the origin of replication, the Amp gene, and the last 2 kb of the rrnB operon. A 773-bp fragment containing the site to be mutagenized is then removed from the smaller plasmid. The remainder of the plasmid is heated and renatured with the same plasmid that has been linearized at a unique PstI site within the Amp gene. After establishing that the gapped-duplex is formed, the mutation is induced in the singlestranded area using the procedure described above for 5S RNA. After screening for mutants the plasmid is linearized and then cloned back into the rRNA operon using either a constitutive, or an inducible, expression system. Substitution frequencies within the 23S RNA gene of the rrnB operon are about 10%. If there are suitable restriction enzyme sites to generate the singlestranded region in pKK3535, and if the mutations are not lethal, then this ~9S. Inouyc and M. Inouye, in "Synthesis and Applications of DNA and RNA" (S. Narang, ed.). Academic Press, San Diego, 1987. 20j. Brosius, A. Ulldch, M. A. Raker, A. Gray, T. J. Dull, R. R. Gutell, and H. F. Noller, Plasmid6, 112 (1981).

[50]

5S RNA STRUCTUREAND FUNCTION

721

plasmid method has the considerable advantage that no cloning step is necessary. The method involving construction of a small plasmid involves some cloning but still avoids problems associated with back-cloning from the M 13 vector. Acknowledgments We thank SolveigKjaer and Arne Lindahl for their help with the manuscript.

[50] 5S RNA Structure

and Function

By H. U. GORIN6ER and R. WA6~ER Introduction 5S RNA from Escherichia coli has been the target of intensive research for more than a decade. Consequently, a number of reviews have appeared in the past summarizing the early findings on the structure and function of this molecule? -4 Due to the development of new techniques, more information about this interesting RNA species has been accumulated and some of our views on the structure and especially on the function have to be revised, although we are still far from a complete understanding of the molecule. The studies performed so far have been concentrated on two aspects, structural and functional. The structural predictions, starting from the comparison of evolutionary conserved elements, 5 which were verified and extended by biochemical data, have merged into a reasonable secondary structure, widely accepted today. In contrast, attempts to explore the functional participation of 5S RNA in the process of translation have been a matter of much more controversy. According to an early hypothesis, 6 the highly conserved 5S RNA sequence CGAAC was considered to interact i V. A. Erdmann, Prog. Nucleic Acid Res. Mol. Biol. 18, 45 (1976). 2 R. A. Garrett, S. Douthwaite, and H. F. Noller, Trends Biochem. Sci. 6, 137 (1981). 3 R. Monier, in "Ribosomes" (M. Nomura, A. Tissieres, and P. Lengyel, eds.), pp. 141 - 168. Cold Spring Harbor Lab. Cold Spring Harbor, New York, 1974. 4 R. A. Zimmermann, in "Ribosomes: Their Structure, Function and Genetics" (G. Chambliss, G. R. Craven, J. Davies, K. Davis, L. Kahan, and M. Nomura, eds.), pp. 135-170. University Park Press, Baltimore, Maryland, 1979. 5 G. E. Fox and C. R. Woese, Nature (London) 256, 505 (1975). 6 B. G. Forget and S. N. Weissman, Science 158, 1695 (1967). METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form r-~erved.

[50]


721

plasmid method has the considerable advantage that no cloning step is necessary. The method involving construction of a small plasmid involves some cloning but still avoids problems associated with back-cloning from the M 13 vector. Acknowledgments We thank SolveigKjaer and Arne Lindahl for their help with the manuscript.

[50] 5S RNA Structure

and Function

By H. U. GORIN6ER and R. WA6~ER Introduction 5S RNA from Escherichia coli has been the target of intensive research for more than a decade. Consequently, a number of reviews have appeared in the past summarizing the early findings on the structure and function of this molecule? -4 Due to the development of new techniques, more information about this interesting RNA species has been accumulated and some of our views on the structure and especially on the function have to be revised, although we are still far from a complete understanding of the molecule. The studies performed so far have been concentrated on two aspects, structural and functional. The structural predictions, starting from the comparison of evolutionary conserved elements, 5 which were verified and extended by biochemical data, have merged into a reasonable secondary structure, widely accepted today. In contrast, attempts to explore the functional participation of 5S RNA in the process of translation have been a matter of much more controversy. According to an early hypothesis, 6 the highly conserved 5S RNA sequence CGAAC was considered to interact i V. A. Erdmann, Prog. Nucleic Acid Res. Mol. Biol. 18, 45 (1976). 2 R. A. Garrett, S. Douthwaite, and H. F. Noller, Trends Biochem. Sci. 6, 137 (1981). 3 R. Monier, in "Ribosomes" (M. Nomura, A. Tissieres, and P. Lengyel, eds.), pp. 141 - 168. Cold Spring Harbor Lab. Cold Spring Harbor, New York, 1974. 4 R. A. Zimmermann, in "Ribosomes: Their Structure, Function and Genetics" (G. Chambliss, G. R. Craven, J. Davies, K. Davis, L. Kahan, and M. Nomura, eds.), pp. 135-170. University Park Press, Baltimore, Maryland, 1979. 5 G. E. Fox and C. R. Woese, Nature (London) 256, 505 (1975). 6 B. G. Forget and S. N. Weissman, Science 158, 1695 (1967). METHODS IN ENZYMOLOGY, VOL. 164

Copyright © 1988 by Academic Press, Inc. All rights of reproduction in any form r-~erved.

722

GENETICS

[50]

directly with the constant sequence GT~K~G found in tRNA. 7,8 Experiments where the corresponding 5S RNA sequence has been deleted clearly disprove such an interaction as an obligate step in translation. 9,~° A direct involvement of the 5S RNA in ribosomal function might nevertheless, be anticipated due to the following notions: (1) The presence of 5S RNA and the direct binding proteins is a prerequisite for in vitro reconstruction of functional ribosomes. H (2) An exposed nucleotide sequence of 5S RNA is strongly shielded from chemical modification when ribosomes contain bound tRNA, demonstrating either a close spatial neighborhood to the tRNA-binding site or a direct participation in the binding process. (3) 5S RNA undergoes a structural alteration when tRNA is bound to the ribosome. ~2 The structural and functional studies overlap at this point. The notion that 5S RNA exists in a number of different conformers which are interconvertible, and most probably related to different functional states,~3-~5 demonstrates the intrinsic relation of the structure and function of this molecule, two aspects which cannot be regarded separately. Here we describe the use of some earlier and some more recently developed approaches which we have employed for the investigation of the 5S RNA molecule. The studies have helped to add new and important facts regarding structure and also to understand more about the functional implications of 5S RNA for the ribosome and'the mechanism of translation. The first part of this chapter is divided into biochemical approaches, including limited enzymatic digestion and chemical modification studies as well as cross-linking. The second part describes genetic approaches, where we report on the construction, expression, and structural as well as functional characterization of 5S RNA mutants. We would like to draw the attention of the reader to a similar approach described in this volume [49].

M. Sprinzl, T. Wagner, S. Lorenz, and V. A. Erdmann, Biochemistry 15, 3031 (1976). 8 V. A. Erdmann, M. Sprinzl, and O. Pongs, Biochem. Biophys. Res. Commun. 54, 942 (1973). 9 B. Pace, E. A. Matthews, K. D. Johnson, C. R. Cantor, and N. R. Pace, Proc. Natl. Acad. Sci. U.S.A. 79, 36 (1982). ~oL. Zagorska, J. Van Duin, H. F. Noller, B. Pace, K. D. Johnson, and N. R. Pace, J. Biol. Chem. 259, 2798 (1984). H F. Dohme and K. H. Nierhaus, Proc. Natl. Acad. Sci. U.S.A. 73, 2221 (1976). ~2H. U. GOringer, S. Bertram, and R. Wagner, J. Biol. Chem. 259, 491 (1984). t3 T. Kao and D. M. Crothers, Proc. Natl. Acad. Sci. U.S.A. 77, 3360 (1980). t4 M. J. Kime and P. B. Moore, Nucleic Acids Res. 10, 4973 (1982). t5 D. Rabin, T. Kao, and D. M. Crothers, J. Biol. Chem. 258, 10813 (1983).

[50]


723

Biochemical Approaches

Preparation of 5S RNA Both large- and small-scale preparations of 5S RNA from E. coli are begun either from alumina-ground cells or isolated 70S ribosomes. The samples are deproteinized by three consecutive phenol extractions using phenol saturated with 100 m M ammonium acetate, pH 5.5, 20 m M sodium borate, 10 m M ethylenediaminetetraacetic acid (EDTA), and 0.1% sodium dodecyl sulfate (SDS). The optical density of the solution to be extracted should not exceed 100 A~0 units per milliliter. Residual phenol is extracted by two ether extractions. The RNA is concentrated by precipitation with 2.5 volumes ethanol and separated on 5 to 30% sucrose gradients (SW 27 rotor, 25,000 rpm, 16 hr, 15"). After gradient separation, the 5S RNA is further purified from tRNA and breakdown products by preparative gel electrophoresis on 12% polyacrylamide gels in the presence of 8 M urea (40 V/cm; xylene cyanole marker dye: 10 cm). The 5S RNA is visualized by UV shadowing and extracted from the gel with 250 m M ammonium acetate, 20 m M sodium borate, 1 m M EDTA in the presence of an equal volume of phenol, equilibrated with the same buffer. Preparations exceeding 20 A26ounits of 5S RNA are alternatively separated on Sephadex G-75 columns using 150 m M sodium chloride, 15 m M sodium citrate, pH 7.5, and 5 m M E D T A as equilibration and elution buffer. The 5S RNA is concentrated by ethanol precipitation and dissolved at concentrations of 0.1 mg/ml in either 10 m M Tris-HC1, pH 6.9, 10 m M magnesium chloride, 20 m M sodium borate, or 250 m M Tris-HC1, pH 7.8, 7 M urea, 20 m M sodium borate, depending on whether the A- or B-conformer is to be investigated.

Isolation of 5S RNA A- and B-Conformers For structural investigations it is important that the material in question is homogeneous and not a mixture of different conformers. Although this is an important prerequisite, it can not normally be achieved completely. In case of the 5S RNA from E. coli, however, two stable conformers can be separated and investigated independently,~6-1s although they are probably not completely homogeneous. Both conformers can be obtained in high yields by renaturation and denaturation procedures. They ,6 M. Aubert, G. Bellemare, and R. Monier, Biochimie 55, 135 (1973). 17 H. F. Noller and R. A. Garrett, J. Mol. Biol. 132, 621 (1979). msH. U. G6ringer, C. Szymkowiak, and R. Wagner, Eur. J. Biochem. 144, 25 (1984).

AT~ L

KA .

~,

BT1 .

KB.-~.~-~.

L

G41 _jG54 ~-G56 ~-G61 ~- G64 --G69 ~-G72 ---G75 ~-G76

XC--~

i

I

".~--XC

I

w

u

I

U89--

w4w

Wm qD It ill

--C93

--G96 --G98

lIB

+-- I l l - - G l o o

EB

~

i--Glo

2

lib

qJ UIO3 - --GIo 5

lip

I l l - - Glos 6

- - G~o7

Ii FIG. 1. Limited RNase Tt and nuclease St digestion of 5S RNA A- and B-forms. RNA fragments were separated on 12% acrylamide, 8 M urea gels. For the St experiment, results obtained with 3'- and 5'-labeled RNA are shown and the corresponding patterns are labeled a and b, respectively. A and B denote 5S RNA A- or B-forms. The T, concentrations for the experiments shown were from left to right: 0.32, 0.64, and 1.3 X 10-3 units; the corresponding St concentrations in (a) are 0.4 and 0.9 units and in (b) 0.2 units. K stands for unhydro-

KAAS1

K BS,

b

KA

--XC

XC ..~

--

U95

U 103----

C 110-BPB'-~

-BPB

lyzed control samples. L shows a random alkali ladder. Nucleotide numbers are indicated at the margin. XC and BPB denote the marker dyes xylene cyanole and bromphenol blue. Bands labeled with an asterisk are either caused by secondary cuts or show cuts in the control. Published in similar form in Ref. 18.

726

GENETICS

[50]

can be separated electrophoretically and are relatively stable at room temperature. Pure 5S RNA A-conformer is obtained by incubating the 5S RNA samples in l0 m M Tris-HC1, pH 6.9, 10 m M magnesium chloride, 20 m M sodium borate at 57 ° for 5 rain followed by slow cooling to room temperature (1 ° per minute). To obtain pure B-form, the 5S RNA is incubated in 250 m M Tris-HC1, pH 7.8, 7 M urea, 20 m M sodium borate for 45 rain at 23 ° followed by quick cooling to - 15 °. The efficiency of the interconversion of the two conformers can be analyzed by nondenaturating gel electrophoresis on 8% polyacrylamide gels in 40 m M Tris-acetate buffer, pH 8.3.12

Radioactive End Labeling of 5S RNA 3'-End Labeling. 5S RNA is labeled at the Y-end with [5'-32p]pCp and polynucleotide ligase according to Bruce and Uhlenbeck? 9 The reaction is performed in the presence of 20 m M sodium borate and an ATP concentration adjusted to 50/zM. Y-End Labeling. The 5'-termini of 5S RNA are labeled with [7-32p]ATP and phage T4 polynucleotide kinase after removing the 5'phosphate with alkaline phosphatase. 2° Limited Enzymatic Digestion For structural studies employing limited enzymatic digestion between 1 and 2 × 105 cpm of 3'- or 5'-end-labeled 5S RNA (specific activity 400 TBq/mol) together with 3/tg carrier tRNA are incubated with the different enzymes (see Fig. 1) under the following conditions.

RNase 7"1Digestion Enzyme range: 0.3- 1.3 × 10-3 units Incubation buffer: 100 m M ammonium acetate, pH 5.5, 20 m M sodium borate Incubation time: 10 rain Incubation temperature: 0 ° Reaction volume: 10/tl

Nuclease St Digestion Enzyme range: 0.2-2 units Incubation buffer: 30 m M sodium acetate, pH 4.6, 50 mM sodium chloride, 1 m M zinc chloride, 50/0(v/v) glycerol m9A. G. Bruce and O. G. Uhlenbeck, Nucleic Acids Res. 5, 3665 (1978). 2o S. Douthwaite and R. A. Garrett, Biochemistry 20, 7301 (1981).

[50]

5S RNA STRUCTURE AND FUNCTION

727

Incubation time: 15 min Incubation temperature: 37 ° Reaction volume: 23 #1

Nuclease V~RNase Digestion Enzyme range: 0.1 - 1 unit 2~ Incubation buffer: 50 m M Tris-HC1, pH 6.9, 5 m M magnesium chloride, 50 m M sodium chloride Incubation time: 30 min Incubation temperature: 25 ° Reaction volume: 35 #l

Chemical Modification The same amounts of radioactive 5S RNA and tRNA carrier are employed for the chemical modifications as for the limited enzymatic digestion studies (see Fig. 2).

Diethyl PyrocarbonateReaction Reagent concentration: 5/tl DEP Reaction buffer: I00 m M ammonium acetate, pH 5.5, 20 m M sodium borate Reaction times: 10, 30, and 60 min Reaction temperature: 37 ° Reaction volume: 50/tl The reaction is stopped by the addition of 20/tl of 1.5 M sodium acetate and precipitation of the RNA with 2.5 volumes ethanol. The modification is followed by an aniline-catalyzed chain scission reaction according to Peattie. 22

Dimethyl Sulfate Reaction Reagent concentrations: 0.1, 0.2, and 1/tl DMS Reaction buffer: 100 m M ammonium acetate, pH 5.5, 20 m M sodium borate Reaction times: 3, 5, and 10 min Reaction temperature: 37 ° Reaction volume: 50/~1 The reaction is stopped by adding 30/~1 1 M Tris-acetate, pH 7.5, 2 M 2-mercaptoethanol, 1.5 M sodium acetate and precipitation with 2.5 vol21 M. Digweed, T. Pieler, D. Kluwe, L. Schuster, R. Walker, and V. A. Erdmann, Eur. J. Biochem. 154, 31 (1986). 2: D. A. Peattie, Biochemistry 76, 1760 (1979).

728

GENETICS

KA A.CVE

L

[50]

KB BCVE

C27 C37 G54

tGse ~C 70 C71 U77 -4-, XC

U87

- C92 " C93 - A94

- G98

- UlO 3

-.,li

BPB

FIG. 2. Limited V~ nuclease digestion and DEP and DMS modifications of 5S RNA Aand B-forms. (a) Modification of adenosines with DEP; (b) DMS modifications under condi-

[50]

5S R N A STRUCTURE AND FUNCTION

a

KA ~

A-DEP

KS.

729

B~,DEP .

A39 - -

A66--

A73 .

A78--

.

.

.

• .....

-,*.-- XC

--A94

A99-AlO1- -

A~04~

AIO 9 - -

~"

BPIB

IPB

tions specific for cytidines. DEP reaction times were from left to right: 10, 30, and 60 rain. The C modification reaction was performed for 3 min. The V~ nuclease hydrolysis shown in the right panel was performed with 0.4 and 1 unit of enzyme, respectively. The RNA for all experiments shown was labeled at the 3'-end. Separation was the same as in Fig. 1. Published in similar form in Ref. 18.

730

GENETICS

[50]

umes ethanol. The RNA pellet is washed with 100 #1 0.3 M sodium acetate and precipitated again. The modification is followed by a hydrazine reaction and an aniline-catalyzed chain cleavage as described by Peattie 22 and Douthwaite and Garrett. 2° Separation of Reaction Products. The RNA fragments from the limited enzymatic hydrolysis and chemical modification reactions are separated on 12% polyacrylamide, 8 M urea gels as described ~s and autoradiographed (see Table I).

Inter- and Intramolecular Cross-Linking Topographical information can be obtained by intramolecular crosslinking of 5S RNA or intermolecular cross-linking of 5S RNA and neighboring proteins, using for instance the bifunctionai reagent phenyldiglyoxal (PDG). Reactions can be performed with isolated 5S RNA, 5S RNA-protein complexes, or intact ribosomes. Reaction conditions have been established where the structure of the reacted components are not noticeably affected, and the different conformers are not interconverted.

TABLE I REACTIONS SPECIFIC FOR A- OR ]3- FORM

Reactivity Position

A-form

C12

A15

(+)

GI6 C19 G20 U22 C27 A34 C36 C37 A39 U40 G41 C42 C43 A46 A50 A52 A53

(+)

B-form (+)

SI

(+)

s,

(+) ++ ++ ++ "4-/ ++ +++ J J | / ~ ', J

(+) ++/++ -H'/++ +-1-/-I-4++ /"~"

+++ +++ +++

+/

Cut

+ (+) ++/(+)

V~ V~ VI S~ V,/DMS V1 V~/DMS V I/DMS DEP St TI/S I

DMS DMS St DEP DEP $1/DEP

Remarks

nd/

nd/

[50]

5S R N A STRUCTURE AND FUNCTION TABLE I

731

(Continued)

REACTIONS SPECIFIC FOR A- OR ]3- FORM

Reactivity Position G54 G56 A57 A58 A59 G61 C63 G64 U65 A66 G69 C70 C71 G72 U74 G75 G76 U77 U82 U87 C88 U89 C92 C93 A94 U95 G96 G98 A99 G 100 A101 GI02 U 103 AI04 GI05 G106 G107

A-form

+/ + ++

+

B-form +/+ + + ++/+ ++ ++ ++ + 4"+ ++ +/+ + + (+) ++ + + +

4-+ + -I+/+ + (+) /+ (+)

+ (+)

+ +/4-++ +/+ +/+ +/+ +/++ + +/+ + ++ ++ ++ + + +

Cut T l/V 1 Tl DEP S~/DEP Sl Tl V t/DMS T, Sl $1 V,/T l V, V1 Tl Sl TI T1 Vt Vi Vl S~/DMS * V~ */V~ V~/DEP Si/Vt TI/Vi TI/VI Sl T t/Vi Sl Tl Vl SI Tl Ti Tl

Remarks /nd

A:V l ; B:DMS

nd nd nd

Control Control

nd

a The relative reactivities are given as (+), very weak; +, weak; ++, medium; and + + + , strong. They are based on the band intensities as revealed by densitometer scanning. If more than one intensity symbol is given it refers to different enzymes which are given in the same order in the next line. nd, not double stranded. An asterisk indicates that the cut is found reproducibly in the control RNA showing a strong A or B form specificity.

732

GENETICS

[50]

The cross-linking results are therefore not falsified by the presence of several conformers.23-25 Cross-Linking Reaction. Isolated 5S RNA, 5S RNA-protein complexes, or intact ribosomes can be used for the cross-linking reaction where the RNA is either unlabeled, in vivo 32p-labeled, or end-labeled employing polynucleotide kinase or ligase. Reagent concentration: 2.4 m M PDG, freshly dissolved in 70 m M sodium cacodylate, pH 7.2, 20 m M magnesium chloride, 20 m M sodium borate, 0.3 M KC1 Reagent buffer: 70 m M sodium cacodylate, pH 7.2, 20 m M magnesium chloride, 20 m M sodium borate, 0.3 M KCI Reaction time: 2 to 4 hr Reaction temperature: 37 ° Reaction volume: 25/tl per A260unit of RNA or ribosomes RNA sequence analysis of the cross-linked components is performed according to Wagner and Garrett, 23 Hancock and Wagner, 24 or Szymkowiak and Wagner, 25 where detailed descriptions are given. Results and Discussion

5S RNA Secondary Structure Examples showing the results of the limited enzymatic digestion studies and the chemical modification of the 5S RNA A- and B-conformers are presented in Figs. 1 and 2. A summary of the data obtained from a series of such experiments is given in Table I. To avoid artifacts due to secondary cutting and structural alterations as a consequence of early modification or hydrolysis events, two precautions are followed: (1) All limited enzymatic hydrolysis experiments are performed with 3'- and 5'-end-labeled 5S RNA samples and only those results obtained in both types of experiments are considered as valid. (2) Enzyme ranges or modification kinetics are chosen keeping the number of hits per 5S RNA molecule close to one. The structural conclusions drawn from these results have led to detailed secondary structural models of the 5S RNA A- and B-forms (see Fig. 3).

5S RNA Tertiary Structure In addition to secondary structural information a number of nucleotides can be inferred from the data presented in Table I which are probably 23 R. Wagner and R. A. Garret't, Nucleic Acids Res. 5, 4065 (1978). 24j. Hancock and R. Wagner, Nucleic Acids Res. 10, 1257 (1982). 2s C. Szymkowiak and R. Wagner, Nucleic Acids Res. 13, 3953 (1985).

[50]

5S RNA STRUCTURE AND FUNCTION

733

involved in tertiary interactions. Evidence for such a notion is obtained if one takes into account the different specifieities of the structural probes. In particular, the reaction of the V~ enzyme in combination with the action of single-strand-specific probes point to the fact that the corresponding nucleotides are involved in tertiary interactions V~ recognizes base-paired stem regions and certain tertiary base-base hydrogen-bonded regions. 26 Examples are C27 (DMS reactive and V~ cut in the B-form) or G54, G69, and G98 (T~ cuts and Vi cuts in the B-form). The results from intramolecular cross-linking studies are especially helpful for the construction of any tertiary structural model. Isolated 5S RNA in the A-form results specifically in the high-yield cross-link between bases G41 and G72. 24 The same cross-link is not found in the B-form of the molecule. The spatial neighborhood of these two nucleotides should be accounted for in any 5S RNA tertiary structural model. The combination of the structural data presented in Table I, the crosslinking results, and the application of a recently discovered folding principle for RNA has led to the proposal of pseudoknotted tertiary structural models for the 5S RNA A- and B-forms (see Fig. 4). The arguments in favor of such structures are outlined in detail by Grringer and Wagner. 27 5 S R N A - Protein Interactions

Most of our knowledge of the RNA-protein contact sites between 5S RNA and the direct binding proteins L5, L18, and L25 is based on limited enzymatic digestion and chemical modification studies of isolated 5S RNA-protein complexes. These data have been described and reviewed extensively in the past. 28-3° The methods of investigation outlined here are well suited for such studies because under the limited enzymatic digestion, chemical modification, as well as the cross-linking conditions, no noticeable structural alterations can be observed for the components to be investigated. As an example, we would like to mention the results of an in situ cross-linking experiment where a contact region of the ribosomal protein L25 is identified within the 5S RNA structure. A fragment comprising the sequence region U103 to Ul20 was covalently linked to the protein, demonstrating a close spatial neighborhood between the corresponding nucleotides and the protein L25 within the ribosome. Two other fragments (U89 to G106 and A34 to G51) showed marked alterations in the chemical modification pattern but no evidence for a direct protein contact was 26 R. E. Lockard and A. Kumar, Nucleic Acids Res. 9, 5125 (1981). 27 H. U. Grringer and R. Wagner, Nucleic Acids Res. 14, 7473 (1986). 2s S. Douthwaite, R. A. Garrett, R. Wagner, and J. Feunteun, Nucleic Acids Res. 7, 2453 (1979). 29 R. A. Garrett and H. F. Noller, J. Mol. Biol. 132, 637 (1979). 3o S. R. Douthwaite, A. Christensen, and R. A. Garrett, Biochemistry 21, 2313 (1982).

A-U-G ~--

A

C

C

/

40 C RECA

LEXA AMFC LPP HIS] (s. t.

PORI-r £ORI- [ SPOT 42 RN HI RNA ALAS TRY3

GL~S TUFB

TYRT

LEUL t R N A SUPB-E RRNABPI RRNGPI RRNDPI RRNEPI RRNXP1 RRNABP2 RRNGP2 RRNDE~P2

STR SR3 Si~ RF~)A RPLJ RPOB

TTTCTACAAAAC~tt gata TGTGCAGTTTATGgt cca4 TGCTATCCTGAC~ttgt c4 C T G A ~ a CCATCAAAAAAATAttct c a l ~ I CAAGGTAGAATGCTtt c ~

GAC ~ c A T ~ ATI"ACAAAAA~ ATGCT(~ZAA~t g a ~ AAOGCATACGG-T ' At t1 ~

ta CA CAGAACA1 t act~ACAGCATAACIGTAI

~atmAACGCATCGCCAATC {~G'ICGCATOG'GAIL-TI~tao~c{{GTAACGC~fACAT(I(~'

~ A T ~ ct~CXXZAGCrFfATACGGI T&~AOAAAAAAG~{ta a g ~ A G ' D 3 T A C ~ GI]GCGCAAAC~a act~GC ' IEZ~GAAGL-qI~AOC CCAGTCAAGAAAA~t t ctt~FFfO]CA~CAG-q

AGCCAGCCFCGA~ IAG~fl]~'_ _ATAAA~IIA~ TAAAAAACTAAC~tt gtca (L~G'TOCCGCTfAtaa a t c a t A ~ A T A C G T I ATGCAATFITITAGttgcat~ CTCGCATGTC~ta aat~ACTTGATGCC C'fAOIGCXIA~

TCTCAACGTAAC/~tt t acai GGCGCGTCATTT~ta ~ a ~ ( ~ ]

TCGATAATTAACTAt tg a c ~ A/~AAAACCACt ag I t U m ~ A G I ] A ~ ~AAAAAGAGGt t g a c ~ CAAGI~ATAq~ca a a ~ ] O ] C ~ A k ~ f TTITAAATI'IIIL'TC t t gtc~ CGGAATAACTCCCtat BtmECCACCACTGACACG( T'I~ATA'I"I-II-I(]C~t t g t c a1 C6~AATAACTCCCtataat~CCACCACTGACACI% GA~AAAAAAATAC~gt gcaa AAA~A11]C~ t at Bat~TTGAGACG~ CTGCAAI-I-I-I-IL%At t gcg ~ GCGGAGAACTCCCtataat~CCTCCATCGACACGC ATG~AI il-l-it.XI~t tgtctn TGAGCCGACTCCCtat aat~CCTCCATCGACACL-'C ' G-G ' GAAGG~tat t a ~ A C A ~ ( ] C G C GCAAAAATAAAII~t tgact TGTAGC AAGCAAAGAAAII~ttgact TGTAG(]GGGAAC~t attat ~ A C A C C ~ C ~ CL'I~AAA2"I'CA~t gact~ &~AAGA~A A A ~ a tatIGCCA _C~C~]G~I~GACAG TCGq'I~TATATI~ttgac~ Ul-iq-it.~GCATC(Imtaaaat~ T C L ~ C A T A I CCGTTTAITTTII~taccca ATCCTTGAAGO~tat aat~ T C G A T ~ TACTAGCAAT/~tt~c~ ~ A A G ~ t ataatGL-GCGI]~CTTGTCGI GT'IIZIGGTTGAG~tagat t AGO]AGCCAATCIT '~ TI~Z!C~]ATA~ {IG{]GGq~;A~ acaat~TTA~ACGTAT~ ~-TAAAC'TAA~ CGACTTAATATA(~ A O I I ~ A ~ A A A T G G T I ' I ~ A ~

FIO. 3. The sequences and patterns creating the two major peaks in Fig. 2. T h e f i g h t - h a n d p a t t e r n has score 2 1 . 8 3 and is the famous - 10 c o n s e n s u s , T A T A A T . The leit-hand pattern h a s s c o r e 1 7 . 1 7 and is the - 3 5 consensus p a t t e r n .

It is convenient to understand these methods with an example. A subset of 30 E. coli promoter sequences 24 is aligned on known m R N A start sites. The sequences are approximately 60 bases long. They are analyzed with W = 9, k = 6, and maximum mismatches m m = 2. The resulting plot appears in Fig. 2. The sequences, their descriptions, and the patterns causing the highest peak appear in Fig. 3. The rightmost pattern is TATAAT (the famous - l 0 pattern) which has 8 exact occurrences, 11 with l mm, and 7 with 2 mm. The leftmost pattern is the - 3 5 pattern, which has its strongest representation in these data as TTGCCA; it has 7 occurrences with l m m and 17 with 2 mm. It is possible to analyze the sequences with the sequences written in the RY, or any other alphabet. There are no patterns found in the RY alphabet which are not found in the usual four-letter alphabet. Finally, similar ideas can be used to find consensus palindromes in a set of sequences. Since protein-binding sites sometimes are palindromes, this is a useful tool. 24 D. K. Hawley and W. R. McClure, Nucleic Acids Res. 11, 2237 (1983).

[52]

COMPUTER ANALYSIS OF NUCLEIC ACID SEQUENCES

779

Secondary Structure The firstt R N A sequence by Holley et al.2~ in 1965 was published with a secondary structure inferred from the sequence. Since that time many other secondary structureshave been presented and the problem of how to estimate structure by experimental as well as theoretical techniques has frequently been addressed. The current structures of 16S and 23S R N A were found by a combination of both approaches.9,1°Here of course I restrictdiscussion of computer methods for prediction of secondary structure. Two major approaches have emerged for the computer analysis. The first employs dynamic programming, in an algorithm closely related to the sequence comparison algorithms presented above. Tinoco et al.26 presented a base pair matrix [with (i,j) entry = l if base i will pair base j] which led them to consideration of minimum energy structures. Dynamic programming algorithms came later and are the current method of choice for a single sequence. The second class of methods employs consensus ideas, tracing back to early deduction of a common tRNA structure. 27 Woese, Noller, and colleagues have advanced and refined these methods, and I discuss a mathematical version of this consensus analysis below. If the data consist of a single sequence then the dynamic programming approach is recommended. On the other hand, if the data are a set of sequences suspected to have common structural elements, then the consensus method can succeed in cases when dynamic programming cannot.

Folding by Dynamic Programming The application of dynamic programming to secondary structure prediction was begun by two groups. The approach of one group had the advantage of incorporating general loop, bulge, and base pairing free energy functions; the disadvantage was the building up of complex structures from simpler ones. 2s,29 The other group did optimization in one pass but only found structures with maximum base pairing. 3° Since the presentation of the first algorithms, Zuker has become the 25 R. W. Holley, J. Apgar, G. A. Everett, J. T. Madison, M. Marquisee, S. H. Merrill, J. R. Penswick, and A. Zamir, Science 147, 1462 (1965). 26 I. Tinoco, O. C. Uhlenbeck, and M. D. Levine, Nature (London) 230, 362 (1971). 27 M. Levitt, Nature (London) 224, 759 (1969). 28M. S. Waterman, Adv. Math. Suppl. Studies 1, 167 (1978). 29 M. S. Waterman and T. F. Smith, Math. Biosci. 42, 257 (1978). 3°R. Nussinov, G. Pieczenik, J. R. Griggs, and D. J. Kleitman, SIAMJ. Appl. Math. 35, 68 (1978).

780

MATHEMATICAL AND COMPUTER ANALYSIS

[52]

leading figure with his useful dynamic programming codes (see Zuker and Sankoff for a review), al He has combined realistic energy functions into a single pass algorithm that is quite efficient. His program runs in time proportional to n 3, where n = sequence length, and it requires storage n 2. Fully rigorous prediction takes exponential time (which is unacceptable) and n2 storage. Recently it was shown 32 that, by increasing storage to n 3, the exponential time can be reduced to n 4. None of this should deeply concern someone with a sequence to fold. Zuker's efficient and useful code is recommended. To understand why complicated programs are needed to study RNA folding, I briefly consider the number of candidate structures. 33 If F(n) is the number of secondary structures for a sequence of length n, it is required that F(0) = F(1) = F(2) = F(3) = 1; that is, there must be at least 2 bases in an end loop. Since secondary structures do not include knotted structures, a recursion is obtained:

F(n + l ) = F(n) +

~

F ( j --1)F(n - j),

n>3

1 ~j'~n--2

This formula counts all possible structures, forcing pairs between base j and base n + 1. The recursion can be shown 33 for large n to behave like F(n)-[(1 + 42)/rOt/2n-3/2(1 +

42P

For n = 150, F(n) --- 1.2 X 1054. Even with all allowances for the overcount as compared to real sequences, these large numbers show that an algorithm is needed. Surprisingly, the dynamic programming algorithms are based on logic similar to that for the counting. The maximum number of base pairs algorithm goes like this: let M~j = maximum number of base pairs in the sequence segment from base i to base j. Take Mi,j known for 0 - i,j < n. Then add base n + 1. If, in the optimal structure base n + 1 is unpaired, M~.n+ ~ = Mr,,. Otherwise, base n + 1 is paired to j , where 1 -< j - n - 2. Then Ml.n+l = M i j - i + 1 + Mj+l, n Here the 1 counts the new base pair between j and n + 1, while the other two terms are optimal for the other two pieces of sequence. To collect this into a recursion, let ~ij = 1 if bases i and j can pair and 0 otherwise. Then Mij = max{Mi,n; maxl.~j~n_2{Mld_l + 1 + gj+l,n}(~i,j} 3t M. Zuker and D. Sankoff, Bull. Math. Biol. 46, 591 (1984). 32 M. S. Waterman and T. F. Smith, Adv. Appl. Math., 7, 455 (1986). 33 p. R. Stein and M. S. Waterman, Discrete Math. 26, 261 (1978).

[59-]

COMPUTER ANALYSIS OF NUCLEIC ACID SEQUENCES

781

5' U

u

/G C/

C G~

C

G/

u

A/

I G

C/

G C/

C

G" G A

C

c

GU A

G

C

II

vC

U~ G C G G G AAC j G C G G UG

CA

G

III

C

CU

mJllll

A C~

G U

C% G G C% '~G U G C U

U

C G

A U

o

G A C U A A A G

C C

A

A

G

C

U G

G G

C

c c c

IIII

G A G A G C~ U G C C G C G% G A A U~, U A G A IV

C G A

C U

C AG

Fro. 4. A secondary structure for E. coli 5S RNA.

All the complication of algorithms, coding, and running times comes in converting this simple, elegant idea to handle the various free energy functions associated with base pairs, bulges, interior looos, and multibranch loops. This is a difficult task!

Folding by Consensus The consensus methods of folding are sometimes referred to as comparative methods. Levitt 27 in 1969 gave an analysis of the known tRNA sequences by this approach. In contrast with his impressive results, the dynamic programming codes currently fold about 50% of tRNAs into a cloverleaf. More recently, comparative methods were used by Woese, Noller, and colleagues 9,m to solve 16S and 23S structures. I now describe programs and mathematics to fold a set of RNAs. The ideas are based on the insights of Woese and Noller but differ from their methods by being able to perform systematic, complete, and explicitly defined searches. The algorithms will be illustrated by 34 5S sequences from E. coli and related sequences that were obtained from a collection of Olsen and Pace and which can be found in GenBank. A 5S model for E. coli is taken from the literature ~ and is shown in Fig. 4 for reference. 34 B. Lewin, "Genes," 3rd Ed. Wiley, New York, 1987.

782

MATHEMATICAL AND COMPUTER ANALYSIS

[52]

This analysis is very different from dynamic programming: here it is desired to find many "common" helices of a certain size and quality. No minimum energy calculations are made. The base pairs A * U, G * C, and G * U are allowed. These helices are allowed to shift in location with reference to the sequences in some fixed alignment. Two windows are placed on the sequences and it is these windows which determine the shifting. For example with WINDOW 1 • **AUGG******

WINDOW 2 *********CCGU

the 4-base pair helix UoccAUCis -~ formed and the patterns could appear anywhere in their windows. Window positions determine approximate helix position while window width determines the amount of shifting allowed. It is not required that the helices in the various sequences be composed of the same base pairs. These features will be illustrated with the 5S sequence set. The sequences must be aligned initially. Obvious features to align on are the right and left ends of the sequences. This can be done in three ways: (1) align on left ends, (2) align on right ends, and (3) align on both ends, leaving gaps in the center of the shorter sequences. Other features to align on include known biological features or highly significant long patterns common to all the sequences. In our sequences such a pattern (cgaac) will prove useful. These various alignments are explored for common patterns of folding. In Fig. 5, the longest common pattern is seen to be cga, shown in lowercase letters in the figure. The pattern cgaac is in 32 of 34 sequence while ccgaac is in 31 of 34. Notice the small amount of shifting to achieve the alignment. The expected length E common to 32 of 34 random sequences 15 is given by E = log

-~

* 12032 + log

+ 0.577 log(e) - ~ = 2.77

and tr = 0.29 with all logs to the base 4. a~ Therefore cgaac occurs almost 8 as above expected. The first analysis of secondary structure now takes place. There are many ways to place two windows on the sequences. To organize the analysis, first fix the separation of the windows. First with no (0) separation, i.e., adjacent windows, move the two windows across the sequence set. Then increase the separation, moving across the set at each fixed separation until the windows are at the maximum separation. In each position the number of helices found is plotted. Mispairs are allowed. Thus each separation produces a graph. All these graphs are superimposed so that interesting peaks, representing a larger number of helices, can be

II I1"¸1

.+ i

i i i

0

~

g

Methods in Enzymology Vol 164: Ribosomes

Methods in Enzymology Vol 307: Confocal Microscopy

Methods in Enzymology Vol 426: Integrins

METHODS IN ENZYMOLOGY - GLYCOBIOLOGY (Vol. 415)

Methods in Enzymology Vol 416: Glycomics

Methods in Enzymology Vol 417: Functional Glycomics

Methods in Enzymology Vol 43: Antibiotics

Methods in Enzymology Vol 450: Fluorescence Spectroscopy

Methods in Enzymology Vol 416: Glycomics

Methods in Enzymology Vol 290: Molecular Chaparones

Methods in Enzymology Vol 19: Proteolytic Enzymes

Methods in Enzymology Vol 424: RNA Editing

Methods in Enzymology Vol 392: RNA Interference

Methods in Enzymology Vol 425: RNA Modification

Methods in Enzymology. Liposomes

Methods in Enzymology. Liposomes

Numerical Computer Methods, Part E (Methods in Enzymology Vol 384)

Methods in Enzymology (Liposomes )