Deutsche Forschungsgemeinschaft Microbial Fundamentals of Biotechnology
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright © 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
Deutsche Forschungsgemeinschaft
Microbial Fundamentals of Biotechnology Final report of the collaborative research centre 323, „Mikrobielle Grundlagen der Biotechnologie: Struktur, Biosynthese und Wirkung mikrobieller Stoffe“, 1986 – 1999 Edited by Volkmar Braun and Friedrich Götz Collaborative Research Centres
Deutsche Forschungsgemeinschaft Kennedyallee 40, D-53175 Bonn, Federal Republic of Germany Postal address: D-53175 Bonn Phone: ++49/228/885-1 Telefax: ++49/228/885-2777 E-Mail: (X.400): S = postmaster, P = dfg, A = d400, C = de E-Mail (Internet RFC 822):
[email protected] Internet: http://www.dfg.de
This book was carefully produced. Nevertheless, editor, authors and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Cover: Crystal structure of E.coli FhuA with bound rifamycin CGP 4832 as determined by Ferguson et al. Structure 9:707-16 (2001)
Library of Congress Card No.: applied for
A catalogue record for this book is available from the British Library.
Die Deutsche Bibliothek ± CIP Cataloguing-in-Publication Data A catalogue record for this publication is available from Die Deutsche Bibliothek ISBN 3-527-30615-3
WILEY-VCH Verlag GmbH, Weinheim (Federal Republic of Germany), 2001 Printed on acid-free and chlorine-free paper. All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form ± by photoprinting, microfilm, or any other means ± nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition: ProSatz Unger, Weinheim Printing: betz-druck gmbh, Darmstadt Bookbindung: J. Schåffer GmbH & Co. KG, Grçnstadt Printed in the Federal Republic of Germany
Contents
Antibiotics and Other Biologically Active Microbial Metabolites 1
1.1 1.2 1.3 1.4 1.5 1.6 2 2.1 2.2 2.3
3 3.1 3.2 3.3 3.4 3.5 3.6 3.7
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Volkmar Braun and Friedrich Götz Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antibiotic research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The unique features of microbial iron transport . . . . . . . . . . . . . . . . Transport of bacterial proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane components and membrane polarization . . . . . . . . . . . . Chemistry of microbial peptides and proteins . . . . . . . . . . . . . . . . . Summary of short-term projects of the collaborative research centre Screening for New Secondary Metabolites from Microorganisms Hans-Peter Fiedler and Hans Zähner Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Screening methods and novel compounds . . . . . . . . . . . . . . . . . . . . Increasing structural diversity by directed fermentations . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biosynthesis of the Lantibiotics Epidermin and Gallidermin . . . . . Friedrich Götz and Günther Jung History of lantibiotics and lantibiotic research in Tübingen . . . . . . . Primary structure and proposed maturation of epidermin in staphylococci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genetic organization and regulation of the epidermin genes . . . . . Isolation and characterization of genetically engineered gallidermin and epidermin analogues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Function of the epidermin immunity genes epiFEG . . . . . . . . . . . . . Inactivation and characterization of the epidermin leader peptidase EpiP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The flavoenzyme EpiD and formation of peptidyl-aminoenethiolates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 3 3 6 9 10 12 14 16 16 18 41 45 47 52 52 55 56 59 66 72 74 V
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
Contents 3.8
3.9
4 4.1 4.2 4.3 4.4 4.5
5 5.1 5.2 5.3 5.4 5.5 5.6
6
6.1 6.2 6.3 6.4
7
7.1 7.2 7.3
VI
Incorporation of d-alanine into S. aureus teichoic acids confers resistance to defensins, protegrins, and other antimicrobial peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fermentation of Lantibiotics Epidermin and Gallidermin . Uwe Theobald Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Strains for gallidermin/epidermin production . . . . . . . . . . . . Disadvantages during gallidermin process development . . . Gallidermin – a lantibiotic and its way towards industrial production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83 86 88
......
93
...... ...... ......
93 94 94
...... ...... ......
95 99 100
Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 Christiane Bormann Introduction: nikkomycins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Isolation of nikkomycin biosynthetic genes . . . . . . . . . . . . . . . . . . . Isolation of the nikkomycin gene cluster and expression in Streptomyces lividans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organization of the nikkomycin gene cluster . . . . . . . . . . . . . . . . . . Roles of the nik genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transcriptional organization and regulation of the nik cluster . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
102
Glycosylated Antibiotics: Studies on Genes Involved in Deoxysugar Formation, Modification and Attachment, and their Use in Combinatorial Biosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andreas Bechthold Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cloning of the avilamycin, landomycin, urdamycin, and granaticin biosynthetic gene clusters . . . . . . . . . . . . . . . . . . . . . Organization of avilamycin, landomycin, urdamycin, and granaticin biosynthetic genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . New genetically engineered natural compounds . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of the Biosynthesis of Glycopeptide Antibiotics: Basis for Creating New Structures by Combinatorial Biosynthesis . . . . . . . Stefan Pelzer and Wolfgang Wohlleben Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
102 104 108 109 110 120 121
124 124 127 128 132 136
139 139 141 149 149
Contents 8
8.1 8.2 8.3
Homologous Recombination and the Induction of the SOS-Response in Antibiotic Producing Streptomycetes Günther Muth and Wolfgang Wohlleben Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mutational analysis of the S. lividans recA gene . . . . . . Regulation of RecA activity . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.........
151
. . . .
151 152 156 160
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Membrane Processes 9
9.1 9.2
9.3
9.4 9.5 9.6
10 10.1 10.2 10.3 10.4
11
11.1 11.2 11.3 11.4
Regulated Transport and Signal Transfer Channels involved in Bacterial Iron Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Volkmar Braun and Helmut Killmann Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Fhu proteins catalyze active transport of ferrichrome and the antibiotic albomycin across the outer membrane and the cytoplasmic membrane of E. coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transduction of energy from the cytoplasmic membrane into the outer membrane for the activation of FhuA as a transporter and phage receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport of ferrichrome across the cytoplasmic membrane . . . . . . . Ferric-carboxylate transport system of Morganella morganii Volkmar Braun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport of ferric iron ions by the Sfu system of Serratia marcescens Volkmar Braun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iron Transport in Gram-negative and Gram-positive Bacteria . Klaus Hantke Ferric iron transport in bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . Ferrous-iron transport systems (Feo) of E. coli . . . . . . . . . . . . . . Regulation of iron transport and metabolism . . . . . . . . . . . . . . . An [2Fe-2S] protein is involved in ferrioxamine B utilization . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
163 163
165
175 178 181
182 183 183
...
188
... ... ... ... ...
188 194 195 198 201
Regulation of the Ferric-Citrate Transport System by a Novel Transmembrane Transcription Control . . . . . . . . . . . . . . . . . . . . . . Volkmar Braun and Sabine Enz Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transport of Fe3+ is mediated by citrate . . . . . . . . . . . . . . . . . . . . . . Transcription initiation by a signaling cascade from the cell surface into the cytoplasm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Iron regulation of fecIR and fecABCDE transcription . . . . . . . . . . . .
205 205 205 207 209 VII
Contents
12 12.1 12.2 12.3 12.4 12.5 12.6
13
13.1 13.2 13.3
14
14.1 14.2 14.3 14.4 14.5 14.6
15
15.1 15.2 15.3 15.4
VIII
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
209 211
Structure, Function, Import, and Immunity of Colicins . . . . . . . . . Volkmar Braun and Helmut Pilsl Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Colicin M inhibits murein biosynthesis and thus displays a unique activity among the colicins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Colicins 5 and 10 are taken up by a novel mechanism . . . . . . . . . . . Colicins evolved by the exchange of DNA fragments which precisely defined functional domains . . . . . . . . . . . . . . . . . . . . . . . . Pore-forming colicins are inactivated by the cognate immunity proteins shortly before the formation of the transmembrane pores . Pesticin is a muramidase which is inactivated by the immunity protein in the periplasm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
212
Structure, Activity, Activation, and Secretion of the Serratia marcescens Hemolysin/Cytolysin . . . . . . . . . . . . . . . . . . . . . . . Volkmar Braun and Ralf Hertle Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characterization of the S. marcescens hemolysin (ShlA) . . . . . . Pathogenicity of S. marcescens hemolysin/cytolysin . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
213 214 215 217 218 220 220
...
222
... ... ... ...
222 224 231 235
Staphylococcal Lipases: Molecular Characterization and Use as an Expression and Secretion System . . . . . . . . . . . . . . . . . . . . . . . . Friedrich Götz and Ralf Rosenstein Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular organization of staphylococcal lipases . . . . . . . . . . . . . . . Biochemical characterization of staphylococcal lipases . . . . . . . . . . Role of the pro-peptide region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The use of ShyL as expression and secretion system . . . . . . . . . . . . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Moritz von Rechenberg, Waldemar Vollmer, and Joachim-Volker Höltje The murein sacculus, a “growing” molecule . . . . . . . . . . . . Murein growth is accompanied by massive turnover . . . . . . Enlargement and division of a stress bearing structure . . . . Interaction of murein hydrolases and synthases as indicated by affinity chromatography . . . . . . . . . . . . . . . .
212
238 238 239 241 244 244 246 246
......
249
...... ...... ......
249 252 253
......
254
Contents 15.5 15.6 15.7 15.8
Dimerization of the bifunctional transpeptidase/transglycosylase PBP1B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reconstitution of the core particle of a murein synthesizing machinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proposed structure of a hypothetical holoenzyme of murein synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recent insights in the mechanism of growth of the murein sacculus reveal novel targets for antibiotics . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The Changing Path of Hopanoid Research: From Condensing Lipids to New Membrane Enzymes . . . . . . . . . . . . . . . . . . . . . . Karl Poralla 16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.2 The cyclization reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.3 Purification of squalene cyclases . . . . . . . . . . . . . . . . . . . . . . . . . 16.4 Properties of purified cyclases . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.5 Cloning of squalene-hopene cyclases . . . . . . . . . . . . . . . . . . . . . 16.6 Properties of SHC sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.7 The structure of squalene-hopene cyclase . . . . . . . . . . . . . . . . . 16.8 Site directed mutagenesis of squalene-hopene cyclase . . . . . . . 16.9 Hopanoid biosynthesis gene clusters . . . . . . . . . . . . . . . . . . . . . 16.10 Miscellaneous results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.11 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
255 256 256 259 260
16
...
263
... ... ... ... ... ... ... ... ... ... ... ... ...
263 265 266 267 270 271 272 274 278 279 280 280 281
17
Genetic and Biochemical Analysis of the Biosynthesis of the Orange Carotenoid Staphyloxanthin of Staphylococcus aureus . . Friedrich Götz 17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2 Cloning of the carotenoid biosynthetic genes from S. aureus Newman in S. carnosus and E. coli . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3 Function of CrtM and CrtN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.4 Identification of carotenoids in S. carnosus (pOC21), E. coli (pUG1), and S. aureus Newman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.5 Identification of dehydrosqualene in E. coli (pUG1) and E. coli (UG9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.6 Squalene is very likely no substrate for CrtN, the proposed dehydrosqualene desaturase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.7 The crt operon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.8 Homology of CrtO, CrtP, and CrtQ . . . . . . . . . . . . . . . . . . . . . . . . . . 17.9 Construction of crtM mutants of S. aureus strain Newman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.10 sB-regulated promoter of the crt operon from S. aureus strain Newman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
284 284 285 286 287 287 288 288 289 289 290 IX
Contents 17.11 The carotenoid biosynthesis genes . . . . . . . . . . . . . . . . . . . . . . . . . . 17.12 Function of the pigments in S. aureus strain Newman . . . . . . . . . . . 17.13 Distribution of pigment biosynthesis genes among staphylococcal species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
290 292
Second Messenger Systems in Paramecium . . . . . . . . . . . . . . . . . . Joachim E. Schultz and Jürgen Linder Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Identification and characterization of cGMP and cAMP second messenger signaling systems in Paramecium . . . . . . . . . . . . . . . . . . Biochemical properties of an adenylyl cyclase . . . . . . . . . . . . . . . . . A guanylyl cyclase disguised as an adenylyl cyclase . . . . . . . . . . . . On the way to an adenylyl cyclase with an intrinsic ion conductance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Downstream of second messengers . . . . . . . . . . . . . . . . . . . . . . . . . In vivo screening of bacterial secondary metabolites . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
295
18 18.1 18.2 18.3 18.4 18.5 18.6 18.7
292 293
295 296 300 302 306 308 311 312 312
Chemical Synthesis and Structure Elucidation 19
19.1 19.2 19.3 19.4
Structure Elucidation and Chemical Synthesis of Microbial Metabolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roderich D. Süßmuth, Jörg Metzger, and Günther Jung Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Development of methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structure elucidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.....
319
. . . . .
. . . . .
319 320 325 337 339
... ... ... ... ... ... ... ... ... ... ...
345 345 345 348 349 349 355 364 364 365 365
. . . . .
. . . . .
. . . . .
Documentation 20 20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9 20.10 X
Documentation of the Collaborative Research Centre 323 List of institutes involved . . . . . . . . . . . . . . . . . . . . . . . . . . . List of supported project areas . . . . . . . . . . . . . . . . . . . . . . Promotion of members of the collaborative research centre Recruitment of new project leaders . . . . . . . . . . . . . . . . . . . Alphabetical list of members and participants . . . . . . . . . . Support of young scientists . . . . . . . . . . . . . . . . . . . . . . . . . Alphabetical list of guests . . . . . . . . . . . . . . . . . . . . . . . . . . International cooperation . . . . . . . . . . . . . . . . . . . . . . . . . . . International conferences . . . . . . . . . . . . . . . . . . . . . . . . . . Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . .
... ... ... ... ... ... ... ... ... ... ...
Preface
This book summarizes the scientific contributions of members of the collaborative research centre 323. Tübingen is one of the very few academic places worldwide where microbes have been systematically screened for biologically active metabolites. This research started in 1964 when the first chair of microbiology was created at the University of Tübingen and was led by Hans Zähner. Research on antibiotics in Tübingen continued in the 1980s, when most pharmaceutical companies had abandoned the development of new antibiotics as more than 100 antibiotics were already available to treat seemingly all relevant microbial infections. We now know that this decision was premature. Bacterial antibiotic resistance was already emerging and continued to progress at an increasing pace, resulting in multi-resistant pathogens which now can only be controlled by newly developed antibiotics. The collaborative research centre 323 was an ideal instrument for bringing together scientists of different disciplines and defining common interests. It consisted of chairs and groups of Microbiology/Biotechnology, Microbiology/Membrane Physiology, Microbial Genetics, Pharmaceutical Chemistry, Pharmaceutical Biology, and Organic Chemistry of the University and various groups of the Max-Planck Institutes of Developmental Biology (Biochemistry Department), and Infection Biology. The members of the collaborative research centre met regularly in seminars, which led to very successful co-operations in nearly all scientific projects. Striking examples include the lantibiotic research, the identification of new siderophores in pathogenic microorganisms, the screening/isolation/and subsequent structure elucidation of new antibiotics, and the synthesis of defined substrates for iron transport analysis or hemolysin function. The extremely fruitful cooperation between microbiologists and chemists is documented by co-authorship in numerous publications and is specified in more detail in the “Research Projects”. What is the secret of success of a collaborative research centre? There are several reasons why the scientific outcome of a collaborative research centre is usually more than the sum of the individual projects. First of all, the duration of a collaborative research centre is long-term. Our collaborative research centre 323 was designed to run for 15 years, which allowed the realization of long-term XI Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
Preface future-oriented projects. Research at universities is normally hampered by short-term grants and fellowships. The collaborative research centre enabled researchers to delve into scientific topics without the persistent fear of premature termination of the project, which would render the field open for competitors to “harvest the fruit”. Secondly, successful scientific cooperation is a matter of trust, competence, and bilateral benefit. At a single university or research institution, it is not easy to find a configuration that meets these prerequisites. A foundation of trust takes time to develop and is largely dependent on individuals. The environment of our collaborative research centre 323 facilitated and stimulated scientific cooperation. The importance of the collaborative research centre for our place at the frontier of science may be best illustrated by the recruitment of the scientific staff to meet the requirements of the collaborative research centre 323. It is no exaggeration to say that through the efforts of the members of the collaborative research centre 323, fundamental and important results were achieved in a number of areas, and many colleagues worldwide were influenced and stimulated by the achievements of the collaborative research centre. We are deeply indebted to the Deutsche Forschungsgemeinschaft for continuous support and to its reviewers for dedicated evaluation of the general concept of the collaborative research centre and the individual research projects. We would also like to thank the University of Tübingen and the Ministry for Science of Baden-Württemberg for their understanding and financial support. Volkmar Braun
XII
Antibiotics and Other Biologically Active Microbial Metabolites
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
1 Introduction Volkmar Braun* and Friedrich Götz**
Abstract
A summary of the major scientific achievements in the antibiotic, the membrane-traffic, and chemistry projects of the collaborative research centre 323 from 1986 to 1999 will be presented in the Introduction, which is followed by a more detailed description of the research projects of 1995 to 1999.
1.1 Antibiotic research
Because of the threatening spread of multi-resistant pathogenic microorganisms, the search for and development of new antibiotics (anti-infective drugs) has become more compelling than ever before. Vancomycin-resistant Staphylococcus aureus strains have already been isolated in various countries, and one can foresee that we are facing an increased morbidity and mortality due to treatment failure. Therefore, the search for new anti-infectives and lead compounds and the development of new strategies will always be important. Antibiotics are considered secondary metabolites, which, at least under laboratory conditions, do not participate in the primary metabolism essential for microbial growth. Their role in the natural environment has always been an issue in the collaborative research centre. Antibiotics inhibit microorganisms com-
* Mikrobiologie/Membranphysiologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen ** Mikrobielle Genetik, Universität Tübingen, Waldhäuser Str. 70/8, D-72076 Tübingen
3 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
1 Introduction peting for the same ecological niche. It was proposed that secondary metabolism also forms the basis for the evolution of new metabolic pathways. Through elucidation of the mode of action of various antibiotics, a large number of distinct metabolic activities of both prokaryotic and eukaryotic organisms were identified. Because of the beneficial activities of compounds related to antibiotics, such as siderophores, which deliver iron to microorganisms, antibiotics can no longer be regarded as secondary metabolites. Antibiotics with structures similar to those of iron complexes were found to be actively transported into target cells. This finding led to the concept of increasing the efficacy of antibiotics by linking them to actively transported compounds.
1.1.1 Screening and fermentation Antibiotic research in Tübingen involved the screening of many newly isolated microorganisms for novel antibiotics and production of the antibiotics in amounts sufficient for structure elucidation and utility testing. In the course of the collaborative research centre, a state-of-the-art fermentation technology was maintained, and new assay systems, and analytical and synthetic tools were developed. Metabolite production was increased by optimizing the fermentation conditions, aided by a sophisticated on-line analysis of the products released into the culture media. An extensive antibiotic database was created to differentiate new compounds from known compounds at an early stage of the investigation. The spectrum of products usually formed by the microorganisms was modified by directed fermentation and more recently by genetic means, such as mutation and combinatorial pathway recombination. The concept was not a large-scale screening in the sense that thousands of strains and compounds were tested per day. Rather, a selective screening was employed using special growth conditions, novel test systems, and various strains, some of which synthesized a number of different secondary metabolites. The screening was also not target-oriented, and an unbiased selection procedure was used. Some test systems were established within the collaborative research centre, and others were used in cooperation with research groups outside the collaborative research centre. Potential applications were focused not only on antibacterial compounds, antifungal products and therapeutic compounds were also considered. Apart from the biotechnological purpose of searching for new microbial metabolites, the wealth of compounds found demonstrated the extremely high metabolic potential of microorganisms. For example, in the original nikkomycinproducing Streptomyces strain, 14 nikkomycin variants were determined and mutagenesis resulted in 24 additional derivatives. The metabolic diversity of Streptomycetes along with their capability of differentiation is reflected by the size of their genome, which is at the upper limit of bacterial genomes. 4
1.1 Antibiotic research
1.1.2 Marketed compounds developed under the collaborative research centre Among the four commercially most interesting compounds developed under the collaborative research centre, only gallidermin and avilamycin are antibacterial compounds; the latter is employed as a food additive. Desferrioxamine B (trade name Desferral) is used to treat iron-overload diseases, phosphinothricin (trade name BASTA) is a herbicide, and the nikkomycins are potent inhibitors of chitin synthases and thus kill fungi, insects, and acarids without toxic side effects on mammals. Nikkomycin Z is currently in the second stage of clinical trial.
1.1.3 Lantibiotics The lantibiotic era in Tübingen began in 1985 with the determination of the structure of epidermin, a peptide antibiotic isolated from Staphylococcus epidermidis. Four research teams in the collaborative research centre worked to reveal this new class of natural products with a novel biosynthesis pathway. Tübingen, where the name “lantibiotic” was coined, has been a stronghold in this research area over the years. The functions of most of the proteins and enzymes involved in biosynthesis were elucidated, although the mechanism of the key reaction, lanthionine formation, is still unknown. Since chemical synthesis of large amounts of epidermin and gallidermin is currently impossible, great efforts were spent to improve the fermentation process, the scale up, and the downstream processing.
1.1.4 Glycosyl antibiotics and regulation of biosynthesis New derivatives of the glycosylated antibiotics avilamycin, landomycin, urdamycin, and granaticin were generated. The gene clusters involved in the biosynthesis of each of the antibiotics were analyzed, and the functions of most of genes were identified. Especially those genes involved in deoxy-sugar formation, modification, and attachment were used to create novel natural products. A series of new genetically engineered natural compounds were created by inactivation or overexpression of certain genes. Balhimycin is a glycopeptide with properties similar to those of vancomycin, which is often used to combat bacterial infections when no other antibiotic is effective. Balhimycin has the same heptapeptide core as vancomycin and dif5
1 Introduction fers in the glycosylation pattern. A cloning system for the producing strain was developed, and a number of genes were identified by using DNA probes based on consensus sequences of the typical biosynthetic enzymes, such as those encoding peptide synthetases and glycosyltransferases. Most of the balhimycin biosynthesis gene cluster was sequenced, and the function of a number of genes analyzed. The door is now open for the generation of hybrid antibiotics using the combinatorial biosynthesis strategy. Antibiotic production of mycelia-forming Streptomycetes is controlled by a complex regulatory network that allows the cells to sense different growth conditions and to react to these changes by producing antibiotics. Antibiotic production of Streptomyces coelicolor was shown to be affected by cell density, nutritional limitations, nutritional shiftdown, imbalance in metabolism, and different kinds of stress. The key player of the SOS response in Streptomyces lividans, RecA, was investigated.
1.1.5 Antibiotics and transport Membrane studies were important for gaining knowledge on the entry of antibiotics into cells, resistance via permeability barriers, and active export systems, and for identifying novel targets and altering the antibiotics to exert their detrimental function. One antibiotic studied is albomycin, which is actively taken up by cells through an iron siderophore (ferrichrome) transport system. Inside the target cell, the antibiotically active portion is cleaved off the carrier, which is released from the cells. Active transport reduces the minimal inhibitory concentration to the lowest value known for an antibiotic that kills Escherichia coli. Another example is phosphinothricyl-alanyl-alanine, which is transported into cells by an oligopeptide system; the antibiotic is released from the peptide carrier by intracellular proteases to generate phosphinothricin, which then inhibits glutamine synthetase. These findings prompted research by the collaborative research centre and by pharmaceutical companies aimed at developing antimicrobial compounds that are taken up by transporters.
1.2 The unique features of microbial iron transport
It was clear from the beginning that iron transport systems must have special features not shared by transport systems of any other nutrient. Under oxic conditions, iron occurs as the completely insoluble Fe3+. Since iron is an important 6
1.2 The unique features of microbial iron transport cofactor of redox enzymes, iron shortage must be overcome by microorganisms; they handle this by synthesizing iron-complexing compounds of low molecular weight, designated siderophores (originally called sideramines, siderochromes). The antibiotic-screening group of the collaborative research centre also had a long-standing interest in microbial iron complexes because sideromycins, potent antibiotics, belong to this class of compounds. Studies of the uptake of sideromycins, the intracellular metabolism, and their mode of action were essential for understanding antibiotic activity. In addition, it was clear that the iron supply must be carefully balanced since iron overload is toxic due to iron-catalyzed radical formation, which results in the destruction of DNA, proteins, and membrane lipids. Therefore, transport of iron-loaded siderophores and sideromycins and regulation of siderophore synthesis and transport were the focus of the iron projects.
1.2.1 Iron transport through the outer membrane of E. coli and other pathogenic bacteria Novel iron transport and regulatory mechanisms were expected and also found. Transport of Fe3+-siderophores was unique in several respects. Transport across the outer membrane of Gram-negative bacteria consumes energy, which is provided by the proton-motive force of the cytoplasmic membrane. Energy transfer from the cytoplasmic membrane to the outer membrane became an important research topic. A major breakthrough was the identification of the proteins involved in energy transfer: TonB, ExbB, and ExbD (Ton system). The Ton system was extensively characterized at molecular, biochemical, and structural levels. Seven E. coli K-12 Fe3+-siderophore transport systems were identified, and those of ferrichrome and ferric citrate were studied in detail. The receptors undergo conformational changes upon substrate binding and through interaction with the energized Ton system, as supported by the analysis of the crystal structure of the FhuA transporter. Upon binding of ferrichrome to FhuA close to the cell surface, a strong structural transition occurs. The long-range conformational change takes place across most of the molecule and the width of the outer membrane. The link to antibiotics is provided by FhuA, which serves as the active transporter of two antibiotics: albomycin is structurally related to ferrichrome; rifamycin CGP 4832 is structurally unrelated to ferrichrome. Surprisingly, both occupy the same position as ferrichrome on FhuA. Iron siderophores transporters homologous to the E. coli K-12 outer membrane transporters were identified in Yersinia enterocolitica and Morganella morganii. The ferrioxamine B transport system of the highly pathogenic Y. enterocolitica O8 strain explains the occurrence of yersiniosis upon treatment of patients suffering from iron overload with Deferral (mesylate salt of desferri-ferrioxamine B). In addition, a siderophore, designated yersiniabactin, was detected in the culture supernatants of highly pathogenic strains. Yersiniabactin 7
1 Introduction was isolated in amounts sufficient for determination of its novel structure. Furthermore, the genes of the entire heme transport system of Y. enterocolitica were cloned and sequenced, and functions were assigned to the encoded proteins. This was the first characterized heme transport system. Characterization of the iron transport systems of Serratia marcescens was initiated by the finding that transcription of the hemolysin genes is iron-regulated. These studies revealed a plethora of Fe3+-siderophore transport systems, one of which transports Fe3+ across the cytoplasmic membrane without involvement of a siderophore. Other research groups later related this system to the uptake of iron delivered by human transferrin to a variety of human pathogenic bacteria.
1.2.2 Iron transport through the cytoplasmic membrane Transport of Fe3+-siderophores, heme, and Fe3+ across the cytoplasmic membrane is catalyzed by ABC transporters, which consist of a periplasmic binding protein, one or two integral membrane proteins, and a cytoplasmic ATPase. ABC transporters represent the most frequently occurring transport systems in bacteria. Regions of interaction between the periplasmic binding protein and the cytoplasmic membrane transporter were shown for the first time with FhuB/D. The same type of ferrichrome transport system was shown to occur in Bacillus subtilis, a Gram-positive bacterium, which lacks a periplasm. Here, a protein similar to the periplasmic binding protein of Gram-negative bacteria is linked by a lipid anchor of the murein lipoprotein type to the cytoplasmic membrane.
1.2.3 Iron transport regulation Fur was the first iron regulatory gene to be mapped, cloned, and sequenced. Fur functions as an oligomer and binds when loaded with Fe2+ to iron-regulated promoters and inhibits transcription. An assay for the identification of Fur-regulated promoters was developed. The ferric citrate transport system displays the particular property that it is not only repressed by iron, but is induced by ferric citrate. Ferric citrate binds to the outer membrane FecA transport protein; this binding initiates a signal that is transmitted by the FecR protein across the cytoplasmic membrane. In the cytoplasm, FecI is converted to an active sigma factor, which in turn transcribes the fecABCDE transport genes. In this dual stepwise control, first iron limitation is recognized and subsequently the transport system is synthesized only when the cognate substrate is in the culture medium. 8
1.3 Transport of bacterial proteins Under anoxic conditions, bacteria may acquire Fe2+, which is much more soluble than Fe3+ and does not require chelating agents. Feo of E. coli, the only Fe2+ transport system characterized at the molecular level, is encoded by three genes; one gene, feoB, is very likely involved in energizing the transport by nucleotide triphosphate hydrolysis. Mutants in feo were shown to be attenuated in the mouse gut.
1.2.4 Intracellular iron metabolism Very little is known about the intracellular iron metabolism in bacteria. The atypical [2Fe-2S] protein FhuF was characterized; the cysteine residues that bind the iron-sulfur center were identified by amino acid replacement studies. FhuF mutants no longer utilize ferrioxamine B as an iron source, which suggests that FhuF may be involved in iron mobilization from ferrioxamine B. The two ironregulated genes sufS and sufD play a role in utilization of ferrioxamine B as an iron source and possibly in intracellular iron metabolism. Sequence similarities of SufS to NifS suggest that SufS is involved in the formation of the iron-sulfur center of FhuF.
1.3 Transport of bacterial proteins
1.3.1 Transport of colicins and toxins The activities, import, immunity, and evolution of bacterial protein toxins were studied. The genes of eight colicins and pesticin were cloned and sequenced. Cells that synthesize the toxins are protected by immunity proteins with a high specificity for the cognate colicin. Most of the colicins are released from cells by lysis proteins that are encoded downstream of the activity and immunity genes. Colicins can be subdivided into the N-terminal translocation region, the central receptor recognition region, and the C-terminal activity and immunity regions. Comparison of amino acid sequences clearly demonstrated evolution of the pore-forming colicins by exchange of DNA fragments that encode functional domains. How and when the immunity proteins in the cytoplasmic membrane inactivate the pore-forming colicins was a major question. The transmembrane topology of the immunity proteins and the regions of interaction with the colicins in9
1 Introduction dicate that the colicins are inactivated shortly before the pores are opened. In contrast, colicin M, which inhibits murein and O-antigen biosynthesis by interfering with C55-lipid carrier regeneration, and pesticin, which degrades murein by a mechanism similar to that of lysozyme, are inactivated by their immunity proteins in the periplasm before they reach their targets. The hemolysin/cytolysin (ShlA) of Serratia marcescens is activated by a single protein (ShlB) in the outer membrane through a novel mechanism during secretion. The hemolysin is a large protein that remains in a non-hemolytic form in the periplasm of cells that synthesize no ShlB protein. The N-terminal portion of ShlA is important for activation and secretion, the central portion for binding to erythrocytes, and the C-terminus for the formation of small pores in the membrane of erythrocytes, leukocytes, and epithelial cells. ShlA represents one of the very few cases where a major phospholipid of a biomembrane also serves as a cofactor for activity. ShlB has the potential to form pores through which ShlA might be exported.
1.3.2 Transport of staphylococcal (phospho)lipases Five different staphylococcal lipase genes of S. aureus, S. epidermidis, and Staphylococcus hyicus were cloned and sequenced. All corresponding proteins are organized as pre-pro-enzymes in which the pro-region comprises between 207 and 267 amino acids. The pro-region acts as an intramolecular chaperone that facilitates translocation of the native lipase; the pro-peptide can also translocate a number of completely unrelated proteins fused to it. The pro-region protects the proteins from proteolytic degradation. The lipase pro-peptide-based expression and secretion system is used by an increasing number of groups for production of human proteins and peptides in Staphylococcus carnosus, a food-grade microorganism for which a cloning system was developed.
1.4 Membrane components and membrane polarization
1.4.1 Biosynthesis of triterpenes in bacteria There is a tremendous variety of triterpenes in the plant kingdom; a single higher plant always contains several types of triterpenes. The triterpenoic hopanoids found in a large number of Gram-positive and Gram-negative bacteria 10
1.4 Membrane components and membrane polarization show less structural variability. In some bacteria, hopanoid biosynthesis genes are present, but are not expressed under laboratory conditions; therefore, an even wider range of bacteria may synthesize hopanoids. The study of triterpene biosynthesis and the structural variation of triterpenes in nature was approached by investigating the membrane-bound squalene-hopene cyclase of Alicyclobacillus acidocaldarius, which proved to be easier to work with than that of plants. The encoding gene was cloned, sequenced, and expressed, and the gene product was purified and characterized. Comparisons of the amino acid sequence with those of other triterpene cyclases revealed a conserved 16amino acid repeat. Interestingly, the highly purified A. acidocaldarius squalenehopene cyclase forms minor products of mostly tetracyclic structure; this finding was important for a better understanding of the cyclase reaction mechanism. The studies formed the basis for the determination of the crystal structure, which, together with the crystal structures of two sesquiterpene cyclases, are the first three-dimensional structures of terpene cyclases. The squalene-hopene cyclase is a monotopic membrane-bound enzyme. Knowledge of the structure allowed site-directed mutagenesis of specific residues in the catalytic cavity. Some mutant squalene-hopene cyclases significantly increased the synthesis of tetracyclic and bicyclic byproducts; a “new” cyclase in which a leucine in the central cavity is replaced by lysine produced a bicyclic compound. Additional genes involved in hopanoid biosynthesis were detected upstream of the squalene-hopene cyclase genes of Bradyrhizobium japonicum, Zymomonas mobilis, and Methylococcus capsulatus; the first bacterial gene found to encode a squalene synthase is among them. In the aerial mycelium of Streptomyces coelicolor, a differentiation-dependent formation of hopanoids was found; hopanoids are not formed in substrate mycelium and when cultures are grown in liquid.
1.4.2 Biosynthesis of staphyloxanthin The yellow to orange colony color of S. aureus is one of the classical species criteria. The main pigment is staphyloxanthin, a C30-carotenoid that is integrated into the cytoplasmic membrane. The genes involved in the biosynthesis of staphyloxanthin were identified and analyzed. Through the creation of deletion mutants and the analysis of the intermediary compounds formed, a biosynthetic pathway was postulated. The function of staphyloxanthin is still unclear; however, the expression of its gene responds to the stress sigma factor, SigB, which suggests that staphyloxanthin is necessary for survival under certain conditions.
11
1 Introduction
1.4.3 Signal transduction by cAMP and cGMP The initial steps in signal transduction in Paramecium involving the second messengers cAMP and cGMP were characterized. Unlike in metazoans, where hormones as first messengers elicit intracellular second messenger formation, in the ciliate Paramecium abrupt changes in the cell’s membrane potential activated second messenger biosynthesis. Characterized behavioral mutants of the ciliate with defined defects in electrogenesis showed that cAMP generation depends on a K+-outward current, whereas cGMP formation is enhanced by a depolarizing Ca2+-inward current. Analysis of clones carrying genes of the respective protozoan nucleotide triphosphate cyclases demonstrated the presence of an adenylyl cyclase embedded in a protein background that strongly resembles a potassium ion channel. Most surprisingly, the guanylyl cyclase is disguised in a membrane topology identical to that of canonical mammalian adenylyl cyclases and, in addition, carries an extended N-terminus that closely resembles a P-type ATPase unit with a total of ten transmembrane-spanning helices. These findings obtained with the ciliate Paramecium open new vistas on the structural and functional evolution of nucleotide triphosphate cyclases and provide Rosetta Stone sequences to decipher novel binding/regulating partnerships.
1.5 Chemistry of microbial peptides and proteins
During the course of the collaborative research centre, innovative analytical and synthetic methods were introduced. These methods allowed high-level biochemical investigations to solve microbiological research problems in interdisciplinary co-operations. Very often the analytical and synthetic work was even decisive for the success of projects.
1.5.1 Structure determinations The major contributions of the chemistry group were the structure determinations of a large number of antibiotics, and the synthesis of precursors, which allowed the elucidation of the activities of enzymes involved in biosynthesis. The 3-D structures of gallidermin and actagardine are the basis for the model of their mode of action, which very recently became of increased interest due to the inhibitory activity on peptidoglycan biosynthesis. One of the most unusual and in12
1.5 Chemistry of microbial peptides and proteins teresting peptide structures ever found is the 43-peptide antibiotic microcin B17, elucidated by multidimensional NMR of the 13C-, 15N-labeled polypeptide containing eight oxazole and thiazole rings in its backbone. This gyrase (topoisomerase II) inhibitor is ribosomally synthesized as a precursor peptide which is post-translationally modified. To elucidate such structures at that time, innovative instrumental methods, such as greatly improved NMR methods, HPLC-ESI-MS, and the Edman sequencer coupled to an ESI-mass spectrometer, and novel chemical transformations had to be introduced. Unusual peptide structures can now be sequenced using very small amounts of samples. The number and complexity of the elucidated natural products increased considerably, e. g. lipoglycopeptides and other unusual peptides, and new nonpeptidic metabolites, such as siderophores, macrolides, polyols, lactam antibiotics, and steroidal antibiotics. Recently, the complex structure of CDA (calciumdependent peptide antibiotic), a peptide pheromone carrying a thiolactone ring, of intermediates in nikkomycin biosynthesis, and of the first linear glycopeptide precursors in the biosynthesis of the antibiotic balhymicin were elucidated.
1.5.2 Peptide chemistry, peptide libraries, and mass spectrometric analysis The continuous improvements in parallel automated synthesis of peptides, peptide mimetics, and peptide libraries contributed extraordinarily to structure activity studies in various groups of the collaborative research centre. The outstanding synthetic capabilities combined with novel achievements in library analytics by ESI-MS, HPLC-ESI-MS, ICR-MS, and Edman pool sequencing stimulated collaborations in microbiology and immunology, in and outside of Tübingen. The binding regions of the gating loop of FhuA were identified using synthetic peptides. A number of enzyme activities (e. g. oxidative decarboxylase in the epidermin biosynthesis) and binding domains of proteins to the cell wall (e. g. autolysin) could only be studied with the aid of synthetic peptides and peptide libraries. Furthermore, many new proteins were characterized by LCMS and Edman sequencing, circular dichroism, peptide mapping, and antipeptide antibodies, such as the novel antifungal protein from Streptomyces. The recent introduction of a high-resolution, Fourier-transform, ion cyclotron resonance mass spectrometer will provide powerful technologies for further fruitful research between microbiologists and organic chemists.
13
1 Introduction
1.6 Summary of short-term projects of the collaborative research centre
The collaborative research centre was designed to run for 15 years. During this period, the details of the scientific program changed; however, the basic concept was maintained. Detailed descriptions of the results are contained in the research reports of the collaborative research centre from 1986–87, 1988–1990, 1990–1993, and 1993–1995. This book contains the detailed reports of 1995–1999. The following paragraphs describe the contributions made in short-term projects by the groups of the listed project leaders. Karl-Dieter Entian. Major contributions to the cloning and sequencing of the epidermin and the pep5 lantibiotic biosynthesis gene clusters were made. Heterologous proteins and peptides in yeast were synthesized. Bernd Hamprecht. Intercellular communication in the human nerve system was studied. A method was developed and successfully applied to the quantitative determination of adenosine binding to neural adenosine receptors, which paved the way for the isolation of adenosine receptors. The activities of glycogen phosphorylase, creatine kinase, and sorbitol dehydrogenase were determined in an attempt to analyze metabolic processes regulated by cyclic AMP in astroglia-enriched primary cultures. The group was further involved in the study of the role carnosine plays in the brain, taurine transport, and the mode of action of bradykinin. The neuronal cell cultures were used to analyze the activities of products isolated in the microbial screening programs. Thomas F. Meyer. Secretion of the IgA protease by Neisseria gonorrhoeae was investigated, and a novel mechanism was discovered. The mode of action of the translocating b-domain in the outer membrane was studied, and the domain was fused with heterologous proteins that became exposed at the cell surface. The OmpT protease was identified as the enzyme that degrades the fused proteins, which results in decreased yields. The formation of disulfide bonds by oxidation in the periplasm was shown to prevent secretion by locking fusion proteins in a secretion-incompetent conformation. The b-domain is therefore suitable for exposing antigens at the cell surface with the aim to produce antibodies and to stimulate the human immune system. Johannes Pohlner. The a-domain of the IgA protease of N. gonorrhoeae was studied since it contains a sequence of basic amino acids found in proteins that enter the nucleus of eukaryotic cells. The group also studied the post-translational processing of the IgA protease polypeptide to form the protease proper, which is released into the culture medium along with the a-domain and the bdomain that reside in the outer membrane. Rainer Haas. The VacA cytotoxin of Helicobacter pylori was characterized. The growth conditions for the production of the toxin were established, the vacA structural gene was cloned and sequenced, the occurrence of vacA in var14
1.6 Summary of short-term projects of the collaborative research centre ious Helicobacter strains was determined, vacA mutants were isolated and characterized, and the secretion mechanism of VacA was studied. In addition, tools for the genetic analysis of Helicobacter were developed. Susanne Klumpp. The protein phosphatases type 1, 2A, and 2C of Paramecium were studied. The genes of the phosphatases 1 and 2C were cloned and sequenced. The three phosphatases were purified to electrophoretic homogeneity, and functions were ascribed to protein domains. As part of a collaboration, the biochemistry of sensory transduction in Paramecium was studied.
15
2 Screening for New Secondary Metabolites from Microorganisms Hans-Peter Fiedler* and Hans Zähner
2.1 Introduction
Originally screening for secondary metabolites was focused on antibacterial compounds. Later on the screening was extended to antifungal, antiviral and antitumor activity and today it has expanded to human medicine, animal health, and plant protection. The initial idea, using only natural products produced by microorganisms has been replaced by the search for novel lead structures, accompanied by the development of novel targets in all application fields. Still, the most prominent source for novel leads is found in nature and especially in the secondary metabolism of microorganisms [1]. The new lead compounds can be used for derivatisation programs or as platform for chemical synthesis. Therefore, the screening for novel secondary metabolites received an increased interest in the last 15 years. More than hundred new test systems are described till today for applications in pharmaceutical and agricultural fields [2]. These in vitro-assays are mostly based on key enzymes or receptors and differ from classical antibiotic assays by the following aspects: . Proteases in microbial samples or extracts lead to false positive results by degradation of assay enzymes or protein receptors. . Numerous assays are sensitive to metabolites from the intermediary metabolism, which are found in variable concentrations in all microbial cultures. . The assays are sensitive to infections, osmotic conditions, and changes in the pH value. . The assays are selective for a distinct mode of action within a cascade and do not take account to the whole cascade.
* Mikrobiologisches Institut, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
16 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
2.1 Introduction . Not all new developed assays are suitable for automated robot screening. . Quite a number of assays are sensitive to already known compounds. . None of the target directed assays is suitable to detect chemical diversity in microbial cultures or extracts. Within the collaborative research centre 323 we developed screening strategies which aimed not only on classical antibacterial activity, but also on antifungal activity, on activities which are involved in differentiation processes and on detection of novel siderophores. However, our main screening strategy is based on physico-chemical methods to detect a maximal number of novel secondary metabolites in freshly isolated Actinomycete strains. The so-called “chemical screening” which is based on thin-layer chromatography and staining reagents was first introduced by Hamao Umezawa [3] and continued a few years later by Satoshi Omura, who detected staurosporin by this assay which his most prominent compound found by chemical screening [4]. Hans Zähner modified this method with respect to staining reagents, sample preparation, and variation of the cultivation conditions of the isolated strains [5, 6]. The detection of a biological activity comes only second order. A new dimension of insights into the chemical diversity of produced secondary metabolites was the coupling of high-performance liquid chromatography with computer-assisted diode array detection (HPLC-DAD) and the construction of a database of antibiotics and other natural products based on HPLC and UV-visible absorbance spectral libraries (HPLC-UV-Vis Database) by HansPeter Fiedler [7]. This efficient method allowed the identification of known compounds in raw extracts at a very early stage of investigations or permitted a classification of the compound by comparing UV-visible spectra data. The efficiency of HPLC-DAD screening technique was extended by HPLC-ESI-MS analysis in co-operation with the members of the collaborative research centre Prof. Günther Jung and Prof. Jörg Metzger. The additional information of the molecular mass permitted a more accurate search in commercially available chemical databases. The goal of our strategy was to detect a novel compound which then was isolated and broadly tested for its biological activity, considering that each secondary metabolite will have an activity. A further advantage of chemical screening was achieved by testing pure compounds in the assays. Quite a number of test systems are not compatible with microbial cultures or crude extracts and need the application of pure compounds. Nevertheless, we have not sufficient assay systems available and for many new compounds we have so far not found any biological activity despite intensive co-operation with various pharmaceutical and agrobiological companies. We expect that the “Naturstoffpool” (sponsored since 1996 by BMBF and German pharmaceutical companies) comprising a collection of compounds will be tested by the end of 1999 by a large variety of molecular robot assays. For the analysis of the structure-activity relationship of secondary metabolites we increased the number of original compounds by “directed fermentations” and feeding, by modification of precursors during the production phase, or, by “mutasynthesis” using blocked mutants. 17
2 Screening for New Secondary Metabolites from Microorganisms In the following subchapters all new secondary metabolites are described which were detected in various screening programs during 1986 and 1999 in the groups of Prof. Hans Zähner and Prof. Hans-Peter Fiedler within the collaborative research centre 323.
2.2 Screening methods and novel compounds 2.2.1 Classical screening for antimicrobial activity The classical agar plate diffusion assay for the detection of antimicrobial agents produced by Gram-positive and Gram-negative bacteria, yeasts or filamentous fungi was applied only until 1988. By this assay system we found pyridazomycin [8] and chlorotetain [9]. Both antibiotics show a selective antifungal activity. Pyridazomycin is distinguished by the unusual pyridazine ring that was not described before in microbial secondary metabolites, chlorotetain is a dipeptide containing an unusual chlorinated amino acid. A screening for growth inhibitors against Bacillus subtilis revealed a novel peptide antibiotic named aborycin [10]. The structures of the novel antifungal antibiotics are shown in Fig. 2.1.
Cl+ N H2N
N
O COO+ H3N
Cl
H
H2C
CH3
O +
Pyridazomycin
H3N
Streptomyces violaceoniger sp. griseofuscus Tü 2557
CH
CO
NH
CH
H -
COO
Chlorotetain Bacillus subtilis ATCC 6633
HO
Trp
Phe
Val
Val
Ala
S S Cys
Aborycin Streptomyces griseoflavus Tü 4072
Ala
Tyr
Cys
Asn
Gly
P he
Cys
Asp
S
Ser
Cys
Gly
Leu Ile
Gly
S
Gly
Figure 2.1: Microbial secondary metabolites detected by classical screening for antimicrobial activity.
18
2.2 Screening methods and novel compounds
2.2.2 Screening for antibiotics causing morphological changes of hyphae of Botrytis cinerea This assay is based on both, growth inhibition of Botrytis cinerea and morphological changes of the hyphae, a so-called “bulging effect“. By this assay substances with antifungal action in the presence of polyene antibiotics were found. Nikkomycin Z and X [11, review] were the most prominent antibiotics analysed in the group of Prof. Hans Zähner during the collaborative research centre 76. Nikkomycin Z is a potent inhibitor of chitin synthase and is non-toxic for humans. It was several years under intensive investigation as acaricide for agricultural use at BAYER AG but cancelled in 1984 because of too high costs and its too narrow application in plant protection. From 1994 till 1998 nikkomycin Z was developed as an antimycotic agent for therapy of histoplasmosis, blastomycosis and coccidoidomycosis in human medicine by Shaman Pharmaceuticals in the USA and passed to the second clinical trial. The Botrytis-assay was continued during the beginning of the collaborative research centre 323 and resulted in the detection of galbonolides [12, 13], four new members of antifungal macrolide antibiotics. Their structures are shown in Fig. 2.2. HO
OH
O CH3
R HO
H3C
O
CH2
A B
CH3
O
CH3
: R = OCH3 : R = CH3
OH
H3C H3C
O
CH2
O
CH3
OH CH3
O
CH3
OH
H3CO O
H3C
CH2
C
O
CH3
CH3
O
CH3
D
Galbonolides A-D Streptomyces galbus Tü 2253
Figure 2.2:
Microbial secondary metabolites detected in the Botrytis assay.
2.2.3 Screening for novel siderophores With the exception of lactobacilli all other microorganisms are dependent on the uptake of external iron ions to supply their iron-containing enzymes. Because of the extreme water insolubility of Fe3+-ions all microorganisms have developed very efficient iron-chelating compounds and specific iron-uptake systems. Iron-chelating compounds are for example trihydroxamates, catecholes, tricarboxylates, and other compounds which are able to chelate iron. In the course of the collaborative research centre 323 the following novel chelating metabolites were isolated: 19
2 Screening for New Secondary Metabolites from Microorganisms Maduraferrin was isolated from strain Actinomadura madurae DSM 43067 after detection by an HPLC assay [14]. The complexing centres are a salicylamide moiety, a hydroxamic acid group and an acid hydrazide group. The highly hydrophilic carboxylate-type siderophores staphyloferrins A and B were isolated from Staphylococcus hyicus DSM 20459 grown under strong iron-restricted conditions [15–18]. Both compounds are strictly iron-regulated. Staphyloferrin A consists of two molecules citric acid, each linked to D-ornithine by an amino bond, whereas staphyloferrin B consists of 2,3-diaminopropionic acid, citrate, ethylenediamine and 2-ketoglutaric acid. From Bacillus sp. strain DSM 6940 was isolated besides schizokinen the new dihydroxamate siderophore schizokinen B, in which citrate is replaced by aconitate [19]. Rhizoferrin is a novel carboxylate-type siderophore which was isolated in collaboration with the groups of Prof. Winkelmann and Prof. Jung from Rhizopus microsporus and other fungi of the Mucorales [20, 21]. Rhizoferrin is similar in structure to staphyloferrin A. In case of rhizoferrin, D-ornithine is replaced by putrescin as bridge. From a highly virulent Yersinia enterocolitica strain H1852 a siderophore named yersinibactin was isolated [22]. The novel compound contains a benzene and a thiazolidine ring, as well as two thiazoline rings. It forms stable complexes with trivalent cations such as iron and gallium. While we investigated the fermentation of S,S-ethylenediamine disuccinic acid (EDDS) with Amycolatopsis orientalis strain we found out that EDDS is not an iron but a zinc chelator which opens new biotechnological applications. A novel, not ferrioxamine-type siderophore named amycolachrome was isolated that is similar in structure to the fungal ferrichrome-hexapeptides [23]. The structures of the isolated new siderophores are shown in Fig. 2.3.
2.2.4 Screening for secondary metabolites involved in differentiation processes Actinomycetes are characterised by complex differentiation processes. The search for metabolites which influence these processes is of general importance because they give insight in the tricky sequence and regulation of this dramatic event in the live cycle of these organisms. Prof. Heinz Wolf in the group of Prof. Zähner developed a screening system that allows detection of compounds that stimulate formation of aerial mycelium in Streptomycetes and he isolated hormaomycin from Streptomyces griseoflavus W384, a novel peptide-lactone antibiotic [24, 25]. Hormaomycin induces not only aerial mycelium formation but also antibiotic production in these organisms. Germicidins A and B were isolated from Streptomyces viridochromogenes NRRL B-1551 [26]. Germicidin A is the first known autoregulative inhibitor of spore germination in the genus Streptomyces. Another novel peptide, streptofactin, was isolated from the nikkomycin producer Streptomyces tendae Tü 901/8 c in the group of Prof. Fiedler. This bio20
2.2 Screening methods and novel compounds HO O
H
O
HN
NH N
O
O
OH N
N
H N
O
NH
O
COOH
H CH3
H N H
OH
HOOC
COOH
Maduraferrin
HOOC
Rhizoferrin
Actinomadura madurae DSM 43067
Cunninghamella elegans, Rhizopus microsporus
COOH
H
H N
O
HO
OH
O
O
H N
OH O
N H COOH
H N
O
NH
COOH O
HN
O
HO
OH
HOOC
COOH
H2N
O
COOH
HOOC
COOH
B
A Staphyloferrins
Staphylococcus hyicus DSM 20459
O
O N H
N
CH3
CH3
OH
OH
OH
COOH
OH
H N
N
N
N
CH3
CH3
S
O
Schizokinen B
Yersiniabactin
Bacillus sp. DSM 6940
Yersinia enterocolitica
OH
HO
N
CH2
O C
CH2
CH2
C
O CH
H2C
Figure 2.3:
CH
C
O
NH
O
CH2 CH CH2
NH
C
CH
NH
C
CH3
O
OH
CH2 H2C
N O
O
CH
CH2
C
O
C
OH
CH2
NH
C NH
H2C H3C
CH
NH
COOH S
H3C
S
O
N
OH
CH2 HO
Amycolachrome Amycolatopsis orientalis
N H
Novel iron-chelating compounds.
21
2 Screening for New Secondary Metabolites from Microorganisms surfactant plays a structural role in aerial mycelium development of Streptomycetes and supports the erection of aerial hyphae by lowering the surface tension of water films enclosing the colonies. Mass spectrometry results and amino acid analysis revealed the peptide sequence H2N-Leu-Leu-Ala-Val-Ala-Leu-Lys-Thr and a molecular mass of 1021 Daltons, including a further valine. The missing small part with 94 Daltons of the molecule is bound to the N-terminal end of the peptide [27]. Streptofactin is the first peptide described having structurally and autoregulatory functions. The structures of the isolated secondary metabolites involved in differentiation processes are summarised in Fig. 2.4.
CH3 H3C
CH2
HC
H C
O
C
H N
O HC C
CH3
CH
NH O C HC
C O
C
O
C
CH
C O CH CH
C
NH
O
H3C
C CH2
O2N
NO2
NH
NH CH3
CH2
N H
CH3
O
CH NH C
HO
O
N
Hormaomycin Cl
Streptomyces griseoflavus W-384
OH
OH CH3
H3C
O
O
H3C
CH3 H3C
O
O
H3C
Germicidin A
Germicidin B
Streptomyces viridochromogenes NRRL B-1551
Figure 2.4:
22
New secondary metabolites involved in differentiation processes.
2.2 Screening methods and novel compounds
2.2.5 Chemical screening by TLC, monitoring coloured secondary metabolites Concentrated extracts of culture filtrates and mycelia of Streptomyces strains were separated on silica gel TLC. Such strains were investigated whose extracts showed coloured spots on TLC. The assay lead to the detection of various novel anthraquinone, phenazine and polyene antibiotics. Urdamycins A–F were the most prominent secondary metabolites detected by this method [28–32]. These novel angucycline antibiotics, produced by Streptomyces fradiae Tü 2717, are biologically active against Gram-positive bacteria and show a strong cytotoxic activity against stem cells of murine L1210 leukaemia. For the para-quinone metabolites cinnaquinone and di-cinnaquinone, which were isolated from Streptomyces griseoflavus ssp. thermodiastaticus Tü 2484, no biological activities have been detected so far [33, 34]. Two dark green substances, the esmeraldines A and B, were isolated from Streptomyces antibioticus 2706 [35]. They formally derive by condensation of two phenazine residues of the saphenic acid family. They don’t have any antibacterial activity, but esmeraldine B shows a cytotoxic activity against various tumor cell lines. From Streptomyces violaceus Tü 3556 the new naphthoquinone complex naphthgeranines A–D was isolated [36], from which naphthgeranines A and B show a weak antibacterial and antifungal activity, whereas A, B and C have a moderate cytocidal activity against various tumor cell lines. In addition, strain Tü 3556 produced the new naphthoquinone compounds naphtherythrins D–F. The bright-yellow polyene carboxylic acid serpentene which shows an antibacterial activity against Bacillus subtilis was isolated from Streptomyces sp. Tü 3851 [37]. Remarkable regarding the structure is the benzene ring nearly in the middle of the molecule. The structures of the isolated secondary metabolites screened by this approach are summarised in Fig. 2.5.
2.2.6 Chemical screening by TLC, monitoring fluorescent secondary metabolites Two novel metabolites were detected regarding their blue fluorescence on TLC plates by irradiation with UV light. Pyridindolol glycosides were isolated from Streptomyces parvulus Tü 2480 [38]. No biological activity could be observed of all three compounds. Depsichlorins, isolated from Streptomyces antibioticus ssp. griseorubinosus Tü 1661, represent a group of new cyclopeptide antibiotics which show biological activity against Gram-positive and Gram-negative bacteria [39, 40]. The structures of pyridindolol glycosides and depsichlorins are summarised in Fig. 2.6. 23
2 Screening for New Secondary Metabolites from Microorganisms
O
OH
CH3
Urdamycins Streptomyces fradiae Tü 2717 CH3
O
O
O
OH
CH3 O
HO
OH
O H3C
CH3 O
HO
O
A : R = H
O
HO
R OH
O
E : R = SCH3 O
O
CH3 OH
CH3 O
HO O H3C
CH3 HO
B
O
HO
O
OH
O
O
O
CH3
OH
O R
O
CH3
O O
OH
CH3 O
HO
OH
O H3C
CH3
O
O
HO
O
HO
OH
HO
O
C: R = H N
D: R =
O
O
CH3
CH3
O O
OH
CH3 O
HO
OH
OH
O H3C
CH3 HO HO
O O
OH
O
O
OH
F
Figure 2.5: New secondary metabolites by chemical screening using TLC and monitoring colored compounds.
24
2.2 Screening methods and novel compounds O
O COOH
HO
O H2N
NH2 O
COOH
HO
NH2 OH
HOOC
di-Cinnaquinone
O
Cinnaquinone
O
Streptomyces griseoflavus ssp. thermodiastaticus Tü 2486
H3C CH3 H3C
HN N
HOOC
O
N
R O
CH3 H3C
HN N
HOOC
O
N
OH
O
N
N
COOH
A
COOH
B Esmeraldins
Streptomyces antibioticus Tü 2706
a
R = C13H27
b
R =
(CH2)10
CH CH2 CH3
c
R =
(CH2)12
CH CH3
d
R = C15H31
CH3 CH3 (CH2)13
e
R =
f
R = C16H33
g
R = (CH2)14
CH2R1
CH CH3
O
CH3
HO
R4
O
h R = C17H33 i
H
CH CH3 CH3
OH
R1
R2
R3
R4
A
H
H
H
H
B
OH
H
H
H
C
OH
OH
H
H
D
OH
OH
OH
H
R2
R3 H CH3 CH3
O
R = C17H31
CH2OH OH
O
E
HO
O OH
CH3 R2
C
CH3
Naphthgeranins Streptomyces violaceus Tü 3556
CH3 OH
O R1
O
NH CH3 HOOC
O
CH3
O
R1 D E F
R2
CHO H H H H OH
COOH
CH3
Naphtherythrins D-F Streptomyces violaceus Tü 3556
Serpentene Streptomyces sp. Tü 3851
Fig. 2.5 continued
25
2 Screening for New Secondary Metabolites from Microorganisms
OR1 N H H
N
Pyridindolol glucosides
OR2
Streptomyces parvulus Tü 2480
Tü 2480 F2
H
Tü 2480 F3
H
HO
CH
O
O
C
C
HO
H
OH OH
O
H
OH OH OH
O
H
O CH
H3C
CH
CH
O
C
O
C
N
OH
CH
CH
O HO
CH2
Y
C
O
X
C O H3C
OAc
NH
Y
CH NH
C O
N
C
CH3 CH2 CH3 CH3 CH3 O
O
Leu N
CH3
Depsichlorins Streptomyces antibioticus ssp. griseorubinosus Tü 1661
O OAc
B
H
NH
Cl
A
OH OH
O
CH3
Cl
X
OH
HO
OH
HO
Tü 2480 F4
R3
R2
R1
OR3
Homo-Ile N
CH3
O OAc
C
CH3
Leu N
CH3
O OAc
D
CH3
Homo-Ile N
CH3 O
Figure 2.6: New secondary metabolites by chemical screening using TLC and monitoring fluorescent compounds.
26
2.2 Screening methods and novel compounds
2.2.7 Chemical screening by TLC and Ehrlich reagent Ehrlich reagent reacts mainly with primary amines and the products appear as red-violet zones within few seconds on the TLC. This reagent was successfully applied for detection of pyrrol-3-yl-2-propenoic acid and pyrrol-3-yl-2-propenamide, two further non-active secondary metabolites isolated from Streptomyces parvulus Tü 2480, the producer of pyridindolol glucosides [41]. The group of pyrrolams are four biosynthetically new pyrrolozidinones produced by Streptomyces olivaceus Tü 3082 [42]. They show no antibacterial and antifungal activities, but a weak herbicidal activity against wheat and rice seedlings. Pyrrolam influences the embryonic development of the fish Brachydanio rerio. Obsurolides A2 and A3 produced by Streptomyces viridochromogenes Tü 2580 represent a novel class of phosphodiesterase inhibitors [43]; they have no growth inhibiting potency against bacteria, yeasts and filamentous fungi. Two new phenylpentadienamides were detected in Streptomyces sp. Tü 3946 by orange spots on the TLC stained with Ehrlich reagent, 5-(4-aminophenyl)penta-2,4-dienamide and N2-[5-(4-aminophenyl)penta-2,4-dienoyl]-L-glutamine [44]. Both secondary metabolites show no antibacterial and antifungal activities. The structures of the novel secondary metabolites are summarised in Fig. 2.7.
2.2.8 Chemical screening by TLC and blue tetrazolium staining reagent Blue tetrazolium is a relatively specific derivatisation reagent for steroids and reducing compounds. Blue or violet coloured zones are formed on a light background on the TLC sheet. From Streptomyces aurantiogriseus Tü 3149 a compound was isolated which revealed a yellow-orange colour by staining with blue tetrazolium. Because of its stimulation of aerial mycelium and spore formation of Streptomyces glaucescens, the compound was named differolid [45]. No growth inhibiting activity against bacteria, yeasts and filamentous fungi was observed. A further blue tetrazolium positive compound, (2S,3R,4R,6R)-2,3,4-trihydroxy-6-methylcyclohexanone, was isolated from Streptomyces phaeochromogenes ssp. venezuelae Tü 3154 and Streptomyces albus Tü 3226 [46]. The compound shows no biological activity to bacteria and fungi. A new member of natural compounds having a thiotetronic acid structure was isolated from Streptomyces olivaceus Tü 3010 [47]. The secondary metabolite (2S)-4-ethyl-2,5-dihydro-3-hydroxy-2-[(1E)-2-methyl-1,3-butadienyl]-5-oxo2-thienylacetamide shows antibacterial activity especially against Streptomyces strains. From Streptomyces griseoflavus Tü 2880 the bright yellow colabomycins A–C were isolated which represent new members of the manumycin group [48, 49]. They react with blue tetrazolium as brown spots, with vanillin-sulphuric 27
2 Screening for New Secondary Metabolites from Microorganisms H
COR
R
N
N H
O
Pyrrol-3-yl-2-propenoic acid: R = OH Pyrrol-3-yl-2-propenamide: R = NH2
N O
Pyrrolam
A : R = OH B : R = OCH 3
Streptomyces parvulus Tü 2480
C : R = O
CH O CH2 CH3 CH3
Streptomyces olivaceus Tü 3082 O O
5-(4-Aminophenyl)penta-2,4-dienamide
N H
R
CH3
H2N O
HO
Obscurolides A2:: R = CHO A3:: R = CH2OH Streptomyces viridochromogenes Tü 2580
NH2
N2-(5-(4-aminophenyl)penta-2,4-dienoyl)-L-glutamine H2N O COOH NH
Streptomyces sp. Tü 3946
Figure 2.7: agent.
CONH2
New secondary metabolites by chemical screening using TLC and Ehrlich re-
acid as dark violet and with molybdatophosphoric acid as black spots, indicating their reducing character. The main compound, colabomycin A, is active against Gram-positive bacteria and shows a cytotoxic activity against stem cells of murine L1210 leukaemia. In collaboration with Hoechst AG and Prof. Fiedler, seven musacin compounds were detected in extracts of Streptomyces griseoviridis FH-S 1832 on TLC plates. The compounds were detected by blue tetrazolium chloride (showing a blue-violet colour), by anisaldehyde, orcinol, and Ehrlich’s reagent, respectively. The determination of their structure revealed that six of the seven compounds were new [50]; musacin C shows an anthelmintic activity against Caenorhabditis elegans and Trichostrongylus colubriformis. The structures of the isolated secondary metabolites are summarised in Fig. 2.8.
28
2.2 Screening methods and novel compounds O
H
O
H3C
H
OH
O O
O
OH OH
Differolid
(2S,3R,4R,6R)-2,3,4-trihydroxy6-methylcyclohexanone
Streptomyces aurantiogriseus Tü 3149
Streptomyces phaechromogenes ssp. venezuelae Tü 3154
O
H N
H O
O
H
CH2
OH
CH3 CH3 OH
O C
CH2
H2N
S
CH2 CH3 O
Colabomycin A Thiotetronic acid Tü 3010
Streptomyces griseoflavus Tü 2880
O
Streptomyces olivaceus Tü 3010
NH HO
O OH
A
OH
HO
O
CH3 O
OH
OH
O
OH
B
O
O
CH3 O
OH
OH
OH O
HO
CH3
C O
D
OCH3
H3C
O
OH
O
O
O
O
OH HO
F
H3C OH
Musacins Streptomyces griseoviridis FH-S 1832
Figure 2.8: New secondary metabolites by chemical screening using TLC and blue tetrazolium staining reagent.
29
2 Screening for New Secondary Metabolites from Microorganisms
2.2.9 Chemical screening by TLC and anisaldehyde and orcinol reagent With anisaldehyde-sulphuric acid reagent sugars, steroids, and terpenes can be detected. After heating the stained TLC sheets, a great variety of coloured spots from violet, blue, grey to green were formed on a weakly ochre coloured background. The same strain, Streptomyces griseoviridis FH-S 1832, that showed blue violet musacin spots on the TLC plate when sprayed with blue tetrazolium chloride, showed another pattern of spots with altered Rf values and colour, when the plate was sprayed with anisaldehyde and orcinol reagent, respectively. Besides cineromycin B, three new members of the cineromycin group of macrolide antibiotics were isolated [50]. The cineromycins showed weak activity against Gram-positive bacteria; no further biological activities have yet been observed. The structures of the new cineromycins are summarised in Fig. 2.9. O HOH2C
H3C
CH3
H3C
O
HO
H
H3C
CH3
CH3
OH
OH
CH3
H3C
O
Dehydrocineromycin B
O
H
CH3 OH
CH3
H3C
O
Oxycineromycin B
HO
O
O
2,3-Dihydrocineromycin B
Streptomyces griseoviridis FH-S 1832
Figure 2.9: New secondary metabolites by chemical screening using TLC and anisaldehyde or orcinol staining reagent.
2.2.10 Chemical screening by TLC and vanillin-sulphuric acid staining reagent Vanillin-sulphuric acid reacts relatively specific with higher alcohols, phenols and steroids. Coloured zones are produced on a pale background on the TLC sheet. In the mycelium of the colabomycin producing strain Streptomyces griseoflavus Tü 2880, a further compound, called 2880-II, was detected by vanillin-sulphuric acid staining reagent, resulting in a dark brown spot on the TLC sheet. The compound is related to ferulic acid and shows no antibacterial and antifungal activity [51]. (3S,5R,6E,8E)-Deca-6,8-diene-1,3,5-triol and (3S,6E,8E)-1,3-dihydroxydeca6,8-diene-5-one were isolated from the (3S,8E)-1,3-dihydroxydec-8-en-5-one 30
2.2 Screening methods and novel compounds OH
OH
OCH3
OH OH
H3C
(3S,5R,5E,8E)-Deca-6,8-diene-1,3,5-triol
O
O
NH
HO
O
OH OH
H3C
(3S,6E,8E)-1,3-Dihydroxydeca-6,8-diene-5-one 2880-II Streptomyces griseoflavus Tü 2880
Streptomyces fimbriatus Tü 2335
Figure 2.10: New secondary metabolites by chemical screening using TLC and vanillinsulphuric acid staining reagent.
producer Streptomyces fimbriatus Tü 2335. All compounds are inactive against bacteria and fungi [52]. The structures of the isolated secondary metabolites are summarised in Fig. 2.10.
2.2.11 Screening for new secondary metabolites by polystyrene resin fermentation Addition of the non-polar polystyrene resins Amberlite XAD-16 or XAD-1180 to growing cultures of microorganisms, preferably at the end of the growth phase, enables the absorption of unstable intermediate products or stimulates the producing organism to an altered metabolite pattern. The naphthgeranine and naphtherythrine producer Streptomyces violaceus Tü 3556 (see Section 2.2.5) synthesised the novel series of naphtherythrins A–C, when Amberlite XAD-1180 was added to growing cultures after 36 hours of incubation [53]. The main compounds, naphtherythrins A and B show a biological activity against Gram-positive bacteria and a weak activity against fungi. Under the same conditions Streptomyces exfoliates Tü 1424 produced a group of three new naphthoquinone antibiotics named exfoliamycins [54, 55]. They inhibit growth of Gram-positive bacteria, whereas Gram-negative bacteria and fungi are not sensitive against these antibiotics. The structures of secondary metabolites produced during polystyrene resin fermentations are summarised in Fig. 2.11.
31
2 Screening for New Secondary Metabolites from Microorganisms CH3
H3C
CH3
H3C
OH
O
O
OH
O
O OH
O
O
O
HOOC
O
N H O
HOOC
O
A : R = CH3 B: R=H
O
O
N R
O
CH2OH
O
C Naphtherythrins A-C Streptomyces violaceus Tü 3556
Exfoliamycines Streptomyces exfoliatus Tü 1424 CH3
CH3 HO OH
OR
HO
O
O H3C
OH OH O
CH2OH
OH
O
O
O
CH2OH
H3C O
Exfoliamycin R=H 3-O-Methylexfoliamycin R = CH3
O
Anhydroexfoliamycin
Figure 2.11: New secondary metabolites by screening using polystyrene resin fermentation.
2.2.12 Screening for new secondary metabolites by HPLC and photoconductivity detection Photoconductivity detection has a complete different detection window than UV-Vis spectroscopy and offers the detection of new secondary metabolites in culture filtrates and extracts of microorganisms by HPLC analysis. Streptomyces antibioticus Tü 99, who is known as a producer of chlorothricin, juglomycins A and B, ketomycin, nikkomycins Z and J, as well as nocardamine, was reinvestigated using a HPLC photoconductivity screening system. With this method we could detect four new butenolides [56]. The compounds, which are summarised in Fig. 2.12, show a weak antibiotic activity against Pseudomonas aeruginosa and also a weak inhibition of the chitinase from Serratia marcescens.
32
2.2 Screening methods and novel compounds OH
OH
O
H3C
O
H3C O
H3C H3C
HO
H
Tü 99-1 OH
Tü 99-2 OH
O
H3C
O
H3C H3C H
O
H3C H3C
HO
O
H3C H3C H3C
Tü 99-3
O
Tü 99-4 Tü 99 Butenolides
Streptomyces antibioticus Tü 99
Figure 2.12:
New secondary metabolites by HPLC-photoconductivity screening.
2.2.13 Screening for new secondary metabolites by HPLC and diode-array detection This method represents a modification of the classical chemical screening procedure. TLC and staining reagents were replaced by the more efficient reversedphase HPLC technique coupled with computerised diode-array detection (HPLC-DAD). Commercially available antibiotics and secondary metabolites from our institute pool were analysed with identical standardised HPLC conditions as culture filtrates and raw extracts from freshly isolated Actinomycete strains. Retention times and UV-visible absorbance spectra of references and biological samples were stored in libraries of a HPLC-UV-Vis-Database [7]. Till today more than 600 secondary metabolites, mostly antibiotics, are stored in the database. The technique was first used for detection of minor congeners and for characterisation of blocked mutants and intermediate products of biosynthetic pathways, and since 1990 as screening method for identification of new secondary metabolites in freshly isolated strains. All new secondary metabolites resulting from this screening strategy are summarised in Fig. 2.13. In raw extracts from the elloramycin producer Streptomyces olivaceus Tü 2353 five minor congeners, elloramycins B–F, were detected by HPLC-DAD and determined in structure [57, 58]. All elloramycins are strongly active against Streptomyces strains. As expected, the less methylated elloramycin B shows the best activity against Gram-positive bacteria. The new tetracenomycins B3 and D3 were detected in a blocked mutant of the elloramycin producer, Streptomyces olivaceus Tü 2353-R [59]. The main 33
2 Screening for New Secondary Metabolites from Microorganisms R2 O
O
O
OH
CH3
O OCH3
H3CO OH
OH
O O
R1
CH3
O
OCH3 OR3 O
CH3 O
O
OH
CH3
OR4
R1
R2
R3
R4
B
H
H
H
CH3
C
H
CH3
CH3
H
D F
H
CH3
H
CH3
OH
CH3
CH3
CH3
O OCH3
H3CO O
OH
Elloramycins B-F
O O
Streptomyces olivaceus Tü 2353 CH3
O
E
OCH3 OCH3
OCH3
O RO
OH
COOH OH
O
OH
CH3
B3 R = CH3 D3 R = H Tetracenomycins B3 and D3 H3C
Streptomyces olivaceus Tü 2353-R
O
H3C
N
N
N
N
COOH
COOH
6-Acetylphenazine-1carboxylic acid
H3C N
Saphenic acid methyl ether
O
R O
N
A
Saphenyl fatty acid esters
R = 12-methyltridecanoic acid
B
R = tetradecanoic acid
D
R = 12-methyltetradecanoic acid
E
R = 14-methylpentadecanoic acid
G
R = hexadecanoic acid
I
R = 14-methylhexadecanoic acid
J
R = unsaturated C18 acid
K
R = 16- methylheptadecanoic acid
COOH
Streptomyces antibioticus Tü 2706
Figure 2.13:
34
OCH3
New secondary metabolites by HPLC-diode-array screening.
2.2 Screening methods and novel compounds
A =
HN
B = O
O
O
CHO HN N
C =
O
HN
O
N
R2
R6 R3 R4
C
O
O R1
CH CH CH C NH CH N
O
NH2
OH
R5 OH OH
O
CHO HN
A =
HN
B = O
N
O
N
R2 CH3
O
C
O
CH CH CH C NH CH N
R1 O
NH2
OH
OH OH
O
CHO HN
A =
HN
B = O
N
O
N
OH O HO
C
O
CH2 CH C NH CH
R1 O
NH2
Nikkomycins Streptomyces tendae Tü 901
NH
OH OH
Fig. 2.13 continued
35
2 Screening for New Secondary Metabolites from Microorganisms
R1
R2
O HN
Sz
H
O
N
CHO
HN
Sx O
R2 C
HN O
O
O
OH
OH
OH N
CHO
HN
Sox
O
N O
Soz
R1
OH
H
O
OH
Nikkomycins
N
Streptomyces tendae Tü 901
O
O
O
O
R
R
N H
N H
CH3
A1 : R = COOH A4 : R = CH2OCH3
HO
CH3
HO
B2 : R = CHO B3 : R = CH2OH B4 : R = CH2OCH3 O
COOR2 O
OH R1
N H
N H
OHC CH3 HO
C : R1 = CHO; R2 = H C2 methyl ester: R1 = CHO; R2 = CH3
D2
Obscurolides Streptomyces viridochromogenes Tü 2580
OH
O CH3
Juglomycin Z
O HO
Streptomyces tendae Tü 901
Fig. 2.13 continued
36
H
COOH
CH3 O
2.2 Screening methods and novel compounds CH2OH O
3
O
HO
OH
CH2OH
CH2OH
5
O
CH3
1-(3-Indolyl)-2,3-dihydroxypropan-1-one
Naphthgeranine F Streptomyces violaceus Tü 3556
Streptomyces violaceus Tü 3556
N
H3C C
N
H H N C
O
CH3 O CH3
O C
N C H H
HOCH2
HO O
C
OH
N H
O
H C
C N CH
S
O H C C
C
CH CH3 O H3C CH3
C H
N
N
C H
O
C
OH
SCH3
CH2 N
CH3
CH3 CH
C
CH3 O
H C
H N C
CH3
CH2OH C H
N H
O C
N
O N
Echinoserin Streptomyces tendae Tü 4031
O
O
NH2
OH
O
NH2
O
O
O
O
O H
O O
COOH
H3C
OR
Dioxolid A R = H Dioxolid B R = COCH3
NH2
Dioxolid D
H
OR
OH
Dehydrodioxolid A R = H Dehydrodioxolid B R = COCH3
para-Hydroxybenzamide Streptomyces tendae Tü 4042
Streptomyces tendae Tü 4042
OH
CH3
CH3
CH3
COOH
O O
OCH3
OH
O
1-Hydroxy-4-methoxy-2-naphthoic acid Streptosporangium cinnabarinum ATCC 31213
OH
O
Spirofungin OH
CH3
CH3
Streptomyces violaceusniger Tü 4113
Fig. 2.13 continued
37
2 Screening for New Secondary Metabolites from Microorganisms OH
OH
OH
H3C
OH
10 5
HO
CH3
HO 20
15
O OR2
CH3
O
1
O H3C
O 1'
R1
O 2'
3'
OH
NH
A
35
OH
NH
OH NH
D
NH2
45
CH3
O
R1
40
H NH2+
CH3
CH3
NH2
45
C
30
H
O
OH
O
NH2
45
25
R2
Kanchanamycins Streptomyces olivaceus Tü 4018
O O
H3C OH
HO
(E)-4-oxonon-2-enoic acid
HO O
Streptomyces olivaceus Tü 4018
O
2E,4Z,7Z-decatrienoic acid
2E,4Z-decadienoic acid
Streptomyces viridochromogenes Tü 6105
CH3
O OH O H3C
CH3
O HO H3C
OH O
O
O
OH O
CH3
O
OH O
CH3
OH O
O
CH3
O O O O
H3C
O HO
H3C
O
Tigloside Amycolatopsis sp.
Fig. 2.13 continued
38
CH3
O CH3
OH O
CH3
2.2 Screening methods and novel compounds Simocyclinones
O
Streptomyces antibioticus TuÈ 6040
CH3
O HO
Simocyclinones A
O
A1: R = H A2: R = OH OH
OH
O
Simocyclinones B B1: R1 = H B3: R1 = OH B4: R1 = OH
R2 = H R2 = H R2 = COCH3
R2
H3C
R
CH3
O HO
O
O
OH
O
OH
HO OH
R1
OH
Simocyclinones C C2: R1 = OH C3: R1 = H C4: R1 = OH
R2 = H R2 = COCH3 R2 = COCH3
O R2
OH
H3C
CH3
O HO
O
O
O
OH
O
O
OH
O
R1
OH
R3 O
O
OH
O R2
HN OH
H3C
O
CH3
HO
O
O
O
OH
O
O
OH
O
R1
OH
Simocyclinones D D2: D3: D4: D6: D7: D8:
R1 = OH R1 = H R1 = OH R1 = OH R1 = H R1 = OH
R2 = H R2 = COCH3 R2 = COCH3 R2 = H R2 = COCH3 R2 = COCH3
R3 = H R3 = H R3 = H R3 = Cl R3 = Cl R3 = Cl
O
OH
O
O
CH3 CH3
OH
O
OH
Kyanomycin
HN
O
O C16-18H33-37
O
O OH
OH
O O
Fig. 2.13 continued
P
Nonomuria sp. NN22303
C13-16H27-33 O
39
2 Screening for New Secondary Metabolites from Microorganisms compound B3 is antibiotically inactive against Gram-positive and Gram-negative bacteria, but D3 shows a moderate activity against Bacillus subtilis and Arthrobacter aurescens. The importance of the new compounds is based in their role as key intermediates and in the elucidation of the biosynthetic pathway of elloramycins and tetracenomycins. Seven phenazine compounds were isolated from Streptomyces antibioticus Tü 2706. Besides saphenamycin, saphenic acid and tubermycin B, three new phenazines were detected by HPLC-DAD, 6-acetylphenazine-1-carboxylic acid, saphenic acid methyl ether and a group of eight saphenyl fatty acid esters [60]. A great success of HPLC-DAD screening was the identification of new nikkomycin compounds in mutants of Streptomyces tendae Tü 901. Twenty new nikkomycins were detected by this method [61–67] allowing intensive studies on structure activity relationships [68] and getting new insights in the biosynthetic pathway of nikkomycins. A further new compound from the juglomycin family, juglomycin Z, was detected in the nikkomycin producing strain Streptomyces tendae Tü 901, when the organism was grown under modified nutrition conditions [69]. The naphthoquinone antibiotic shows biological activity against Gram-positive and Gram-negative bacteria and against yeasts. A series of eight new obscurolides was detected besides the main compounds A2 and A3 in Streptomyces viridochromogenes Tü 2580 [43, 70]. B4 is the most active obscurolide in the phosphodiesterase assay. All obscurolides revealed no growth inhibiting potency against bacteria, yeasts and filamentous fungi. In the naphthgeranine producing strain Streptomyces violaceus Tü 3556 besides tubermycin B and 1-phenazinecarboxylate, a new minor congener naphthgeranine F, as well as 1-(3-indolyl)-2,3-dihydroxypropan-1-one, which was not previously described as a natural product, were detected [71]. Naphthgeranine F showed a similar antibacterial activity against Gram-positive bacteria as naphthgeranine C, the main compound in fermentations of strain Tü 3556, whereas 1-(3-indolyl)-2,3-dihydroxypropan-1-one shows no biological activity. A new member of the quinoxaline group antibiotics, echinoserine, was detected in strain Streptomyces tendae Tü 4031 [72]. The new compound is a noncyclic form of echinomycin, but is not a biosynthetic precursor. Echinoserine is less antibiotically active than echinomycin. Dioxolides, a novel class of secondary metabolites, were detected in the culture filtrate of Streptomyces tendae Tü 4042 [73]. Besides dioxolides, which consist of an unusual substituted dioxolane ring, para-hydroxybenzamide was detected, which was not yet described as a natural product. All compounds show no biological activity against Gram-positive and Gram-negative bacteria, yeasts and filamentous fungi. The kanchanamycins, a group of novel 36-membered polyol macrolide antibiotics were detected in Streptomyces olivaceus Tü 4018 [74, 75]. The compounds show antibacterial and antifungal activities, and are especially effective against Pseudomonas. In the same strain the fatty acid (E)-4-oxonon-2-enoic acid was detected, isolated and determined in structure [76]. This new second40
2.3 Increasing structural diversity by directed fermentations ary metabolite shows an antibacterial activity against various Gram-positive and Gram-negative bacteria, especially against Staphylococcus aureus, but not against yeasts and other fungi with the exception of Paecilomyces variotii. Streptosporangium cinnabarinum ATCC 31213 was characterised as a producer of the known antibiotics 43,334 and 43,596, but also as a producer of the new naphthalene compound 1-hydroxy-4-methoxy-2-naphthoic acid, which shows herbicidal activity in the Lemna minor assay [77]. In the culture filtrates and extracts of Streptomyces violaceusniger Tü 4113 the new secondary metabolite spirofungin was detected, a compound having a polyketide-spiroketal structure that shows various antifungal activities, particularly against yeasts [78]. Tigloside, a new tigloylated tetrasaccharide was detected in Amycolatopsis sp. NN0 21702 [79]. This secondary metabolite shows unusual structural elements, which have never before been isolated from Actinomycete strains. Until now, no biological activity could be observed. Two fatty acids, (2E,4Z)-decadienoic acid and (2E,4Z,7Z)-decatrienoic acid, the latter one being described for the first time as a natural product, were detected in the culture filtrate of Streptomyces viridochromogenes Tü 6105 [80]. Both metabolites show strong herbicidal activities against Lemna minor and Lepidium sativum. Simocyclinones, a novel group of angucyclinones, were detected during the fermentation of Streptomyces antibioticus Tü 6040 [81–84]. They are novel natural “hybrid” polyketide antibiotics and consist of an unusual angucyclinone ring with a tetraene side chain and a coumarin ring. Simocyclinones can be subdivided in A-, B-, C- and D-series regarding structures and UV-visible spectra. Compounds of the D-series are distinguished by an activity against Gram-positive bacteria and against several tumor cell lines. Kyanomycin, a blue-coloured secondary metabolite, was detected in the mycelium extract of Nonomuria sp. by HPLC-DAD and HPLC-ESI-MS screening and determined as an unusual anthracycline-phosphatidylethanolamine hybrid that shows weak antibacterial activity [85].
2.3 Increasing structural diversity by directed fermentations 2.3.1 Biomodification of saphenamycin and esmeraldine Streptomyces antibioticus Tü 2706 produces the antibacterial and antitumor active saphenamycin, a phenazine compound, and the cytotoxic active esmeraldine B (see Fig. 2.5). Esmeraldine is the condensation product of one molecule saphenic acid and one molecule saphenamycin. As saphenamycin is a methylsalicylic acid ester of saphenic acid, esmeraldine B shows the same methylsalicyclic acid side-chain than saphenamycin. It was of interest to investigate whether strain Tü 2706 modifies this side chain when derivatives of methylsalicylic acid 41
2 Screening for New Secondary Metabolites from Microorganisms were fed during directed fermentations, and if modified saphenamycins and esmeraldins show an altered antibacterial and antitumor spectrum. Biomodification was achieved by feeding acetylsalicylic acid, 3-methylsalicylic acid, 4-methylsalicylic acid, 5-fluorosalicylic acid, 5-chlorosalicylic acid and 5-bromosalicylic acid, resulting in six new saphenamycin and six new esmeraldine compounds. No incorporation into the molecules was achieved by feeding 5-iodosalicylic acid, 5-bromo-4-hydroxysalicylic acid, 3-hydroxysalicylic acid, 3-methoxysalicylic acid, 3,5-dinitrosalicyclic acid, 4-aminosalicylic acid, 5sulfosalicylic acid, and thiosalicylic acid [86]. The structural modifications are summarised in Fig. 2.14. All saphenamycin derivatives show an antibacterial activity, however, only 4-methylsaphenamycin is as active as saphenamycin. The cytotoxic activity towards an urinary bladder carcinoma cell line was lower in case of the derivatives than with saphenamycin. In comparison to esmeraldine B, 4-methyl- and 5-fluoro-esmeraldine B show an altered antibacterial spectrum and an increased antibacterial activity. 3-Methylesmeraldine B shows a higher cytotoxic activity against the tested tumor cell line than the original esmeraldine B [86].
2.3.2 Biomodification of ferrioxamines Streptomyces olivaceus Tü 2718 is naturally overproducing the iron-chelating compound desferrioxamine E, also known as the antibiotically active nocardamine. Optimisation of the fermentation conditions yielded in amounts of more than 10 g per litre desferrioxamine E. The siderophore consists of three mole succinic acid and three mole L-lysine, forming a trihydroxamate ring. Desferrioxamine E shows the strongest iron(III)-chelating complexing constant related to desferrioxamines A–I. The specificity for the incorporation of diamines containing two to six carbon atoms was investigated by feeding the following diamines: 1,2-diamine ethane, 1,3-diamine propane, 1,4-diamine butane (putrescine), 1,5-diamine pentane (cadaverine), 1,6-diamine hexane, 1,7-diamine heptane, 1,8-diamine octane, 1,5-diamine ethylether, S-2-aminoethyl cysteine, and N-glycin-1,2-ethylenediamine. The incorporation was monitored by HPLC-DAD that allowed the detection of all modified ferrioxamines [87]. Diamines with a space of more than four carbon atoms between the amino groups were not incorporated into the molecule, such as diamines with more than six carbon atoms. All other diamines were incorporated and led to the isolation and identification of 13 new ferrioxamines besides ferrioxamine D2 and E [88], and were determined in their differences of iron-complexation [89]. The new desferrioxamine structures are shown in Fig. 2.15.
42
2.3 Increasing structural diversity by directed fermentations R2 R3
R1 H3C
O
N
R4 O
OR5
N COOH
R2 R3
R1 CH3 H3C
HN HOOC
N
N
O
R4 O
OR5
N COOH
Figure 2.14: Biomodifications of saphenamycin and esmeraldine B by directed fermentations with Streptomyces antibioticus Tü 2706.
43
2 Screening for New Secondary Metabolites from Microorganisms O NH
C
R3
(CH2)2 N
C OH
O
C (CH2)2 C
O
O
HO R4 N
R1 NH
N
R2
NH
O C (CH2)2 C O
Figure 2.15: Desferrioxamine structures produced by directed fermentations with strain Streptomyces olivaceus Tü 2717.
44
Acknowledgments
2.3.3 Biomodification of rhizoferrins Although rhizoferrin represents a fairly simple molecule that consists of two molecules of citric acid linked to 1,4-diaminobutane through two amide bonds (see Fig. 2.3), it may have potential application in biotechnology due to its appreciable metal-binding properties and the ability to be easily degraded by various microorganisms. The specificity of the biosynthetic enzymes of the rhizoferrin producing fungus Cunninghamella elegans was investigated by modifying both, the chain length of the diamine and the citric acid part of the molecule [21]. Variations in chain length of the diamine backbone were very well tolerated. Branching of functionalization in b-position to the amino group of the diamine compounds was also accepted. However, it was not possible to introduce aamino acids. Therefore, the biosynthesis of rhizoferrin must be different to the biosynthesis of staphyloferrin A, produced by staphylococci. Neither by feeding of D-ornithine nor by application of inhibitors of ornithine decarboxylase, with and without simultaneous addition of D-ornithine, it was possible to detect even trace amounts of staphyloferrin A in Cunninghamella elegans. A higher degree of enzyme specificity was involved in the formation of the activated citryl species since analogues of citric acid are more difficult to be incorporated into rhizoferrin analogues than diamines. The structures are summarised in Fig. 2.16 (series A are structures modified in the diamine part, series B are structures modified in the citric acid part). All derivatives obtained by directed fermentations showed similar iron-chelating properties.
Acknowledgments Following scientists from the “Mikrobiologie/Antibiotika” group contributed to the success in search for novel secondary metabolites: Post-docs: J. Bielecki, C. Bormann, Z. Chen, H. Decker, H. Drautz, U. Fauth, M. Harder, T. Hörner, J. Müller, A. Plaga, O. Potterat, T. Schüz, and U. Theobald. Doctoral students: M. Alverado-Kirigin, N. Andres, S. Blum, M. Brandl, D. Braun, K. Burkhardt, I. Cebulla, H.-J. Cullmann, H. Haag, U. Hartjen, H. Hoff, W. Huhn, C. Isselhorst-Scharr, O. Jung,W. Katzer,W. Kuhn, J. Meiwes, F. Petersen, C. Pfefferle, U. Pfefferle, P. Reuschenbach, M. Richter, J. Schimana, P. Schneider, U. Schneider, A. Seiffert, J. Stümpfel, M. Tschierske, B. Wahl, and F. Walz. The excellent collaboration with the groups of Walter Keller-Schierlein (ETH Zürich), Axel Zeeck and Jürgen Rohr (Universität Göttingen), Günther Jung and Jörg W. Metzger (Universität Tübingen), Wilfried König (Universität Hamburg), Urs Séquin (Universität Basel), and Gerhard Bringmann (Universität Würzburg), which were involved in structure elucidation of isolated secondary metabolites, is gratefully acknowledged. 45
2 Screening for New Secondary Metabolites from Microorganisms A-Series
B-Series
H N
O
H N
COO-
C
HO -
OOC
OH
O
O
C
C
-
COO-
H N
H N
COO-
C
HO
OH
COO-
C
-
H N
H N
O
O
HO OOC
C
-
COO-
H N
H N
COOH
H -
OOC
O
C
-
OOC
CH3
O
HO
C
-
OOC
OH
-
COO-
H N
COOCH3
HO -
OOC
O
C
-
OOC
Monomethylmonodesoxyrhizoferrin
H N
COO-
OOC
COO-
OOC
Oxahomorhizoferrin
C
-
H N
O
C
-
OH
O
C
Didesoxyrhizoferrin
COO-
C
OOC
COO-
OOC
Norrhizoferrin
O
HO -
H N
O
C
OOC
H
O
Monodesoxyrhizoferrin O
-
COO-
COO-
OOC
Homorhizoferrin
O
H N
H N
OOC
2-Methylhomorhizoferrin
H N
H N
O
C
COOCH3
COO-
H3C -
OOC
O
C
-
OOC
Dimethyldidesoxyrhizoferrin
O H N
O
C
H N
COOOH
COO-
HO -
OOC
O
C
-
OOC
2-Oxorhizoferrin
Figure 2.16: Rhizoferrin structures produced by directed fermentations with Cunninghamella elegans.
46
References
References
1. Bérdy, J. (1995) Are actinomycetes exhausted as a source of secondary metabolites? Biotekhnologiya 8, 13–34. 2. Hill, D. C., Wrigley, S. K., and Nisbet, L. J. (1998) Novel screen methodologies for identification of new microbial metabolites with pharmaceutical activity. Adv. Biochem. Engin. Biotechnol. 59, 73–121. 3. Umezawa, S., Tsuchiya, T., Tatsuta, K., Horiuchi, Y., Usui, T., Umezawa, H., Hamada, M., and Yagi, A. (1970) A new antibiotic, dienomycin. I. Screening method, isolation and chemical studies. J. Antibiot. 23, 20–27. 4. Omura, S., Iwai, Y., Hirano, A., Nakagawa, A., Awaya, J., Tsuchiya, H., Takahashi, Y., and Masuma, R. (1977) A new alkaloid AM-2282 of Streptomyces origin, taxonomy, fermentation, isolation and preliminary characterization. J. Antibiot. 30, 275–282. 5. Drautz, H. and Zähner, H. (1981) A new method of screening for secondary metabolites in Streptomycetes. In: Actinomycetes, Zbl. Bakt. Suppl., Vol. 11 (Schaal, K. P. and Pulverer, G., eds), Gustav Fischer Verlag, Stuttgart, pp. 515–519. 6. Zähner, H. and Drautz, H. (1985) Methods in the search for new secondary metabolites from microorganisms – a comparison. In: VIII International Symposium on Medicinal Chemistry, Proceedings Vol 2 (Dahlbohm, R. and Nilsson, J. L. G., eds.), Swedish Pharmaceutical Press, Stockholm, pp. 49–68. 7. Fiedler, H.-P. (1993) Screening for secondary metabolites by HPLC and UV-visible absorbance spectral libraries. Nat. Prod. Lett. 2, 119–128. 8. Grote, R., Chen, Y., Zeeck, A., Chen, Z., and Zähner, H. (1988) Pyridazomycin, a new antifungal antibiotic produced by Streptomyces violaceoniger. J. Antibiot. 41, 595–601. 9. Rapp, C., Jung, G., Katzer, W., and Loeffler, W. (1988) Chlorotetain aus Bacillus subtilis, ein antifungisches Dipeptid mit einer ungewöhnlichen chlorierten Aminosäure. Angew. Chem. 100, 1801–1802. 10. Potterat, O., Stephan, H., Metzger, J. W., Gnau, V., Zähner, H., and Jung, G. (1994) Aborycin – a tricyclic 21-peptide antibiotic isolated from Streptomyces griseoflavus. Liebigs Ann. Chem. 1994, 741–743. 11. Fiedler, H.-P., Schüz, T., and Decker, H. (1993) An overview of nikkomycins: history, biochemistry, and applications. In: Cutaneous Antifungal Agents: Compounds in Clinical Practice and Development (Rippon, J. W. and Fromtling, R. A., eds), Marcel Dekker, New York, pp. 325–352. 12. Fauth, U., Zähner, H., Mühlenfeld, A., and Achenbach, H. (1986) Galbonolides A and B – two non-glycosidic antifungal macrolides. J. Antibiot. 39, 1760–1764. 13. Achenbach, H., Mühlenfeld, A., Fauth, U., and Zähner H. (1988) The galbonolids. Novel, powerful antifungal macrolides from Streptomyces galbus ssp. eurothermus. Ann. New York Acad. Sci. 544, 128–140. 14. Keller-Schierlein, W., Hagman, L., Zähner, H., and Huhn, W. (1988) Maduraferrin, ein neuartiger Siderophor aus Actinomadura madurae. Helv. Chim. Acta 71, 1528–1540. 15. Meiwes, J., Fiedler, H.-P., Haag, H., Zähner, H., Konetschny-Rapp, S., and Jung, G. (1990) Isolation and characterization of staphyloferrin A, a compound with siderophore activity from Staphylococcus hyicus DSM 20459. FEMS Microbiol. Lett. 67, 201–206. 16. Konetschny-Rapp, S., Jung, G., Meiwes, J., and Zähner, H. (1990) Staphyloferrin A: a structurally new siderophore from staphylococci. Eur. J. Biochem. 191, 65–74. 17. Drechsel, H., Freund, S., Nicholson, G., Haag, H., Jung, O., Zähner, H., and Jung, G. (1993) Purification and chemical characterization of staphyloferrin B, a hydrophilic siderophore from staphylococci. BioMetals 6, 185–191.
47
2 Screening for New Secondary Metabolites from Microorganisms 18. Haag, H., Fiedler, H.-P., Meiwes, J., Drechsel, H., Jung, G., and Zähner, H. (1994) Isolation and biological characterization of staphyloferrin B, a compound with siderophore activity from staphylococci. FEMS Microbiol. Lett. 115, 125–130. 19. Seiffert, A. and Zähner, H. (1992) Siderophore production in alkalophilic bacilli. In: Dechema Biotechnology Conferences, Vol. 5 (Kreysa, G. and Driesel, A. J., eds.), VCH, Weinheim, pp. 725–728. 20. Drechsel, H., Metzger, J., Freund, S., Jung, G., Boelart, J. R., and Winkelmann, G. (1991) Rhizoferrin – a novel siderophore from the fungus Rhizopus mircrosporus var. rhizopodiformis. BioMetals 4, 238–243. 21. Drechsel, H., Tschierske, M., Thieken, A., Jung, G., Zähner, H., and Winkelmann, G. (1995) The carboxylate type siderophore rhizoferrin and its analogs produced by directed fermentation. J. Industr. Microbiol. 14, 105–112. 22. Drechsel, H., Stephan, H., Lotz, R., Haag, H., Zähner, H., Hantke, K., and Jung, G. (1995) Structure elucidation of yersiniabactin, a siderophore from highly virulent Yersinia strains. Liebigs Ann. 1995, 1727–1733. 23. Cebulla. I. (1995) Gewinnung komplexbildender Substanzen mittels Amycolatopsis orientalis. Doctoral Thesis, Universität Tübingen. 24. Andres, N., Wolf, H., Zähner, H., Rössner, E., and Zeeck, A. (1989) Hormaomycin, ein neues Peptid-lacton mit morphogener Aktivität auf Streptomyceten. Helv. Chim. Acta 72, 426–437. 25. Andres, N., Wolf, H., and Zähner, H. (1990) Hormaomycin, a new peptide lactone antibiotic effective in inducing cytodifferentiation and antibiotic biosynthesis in some Streptomyces species. Z. Naturforsch. 45 c, 851–855. 26. Petersen, F., Zähner, H., Metzger, J. W., Freund, S., and Hummel, R.-P. (1993) Germicidin, an autoregulative germination inhibitor of Streptomyces viridochromogenes NRRL B-1551. J. Antibiot. 46, 1126–1138. 27. Richter, M., Willey, J. M., Süssmuth, R., Jung, G., and Fiedler, H.-P. (1998) Streptofactin, a novel biosurfactant with aerial mycelium inducing acitivity from Streptomyces tendae Tü 901/8 c. FEMS Microbiol. Lett. 163, 165–171. 28. Drautz, H., Zähner, H., Rohr, J., and Zeeck, A. (1986) Urdamycins, new angucycline antibiotics from Streptomyces fradiae. I. Isolation, characterization and biological properties. J. Antibiot. 39, 1657–1669. 29. Zeeck, A., Rohr, J., Sheldrick, G. M., Jones, P. G., and Paulus, E. F. (1986). Structure of the new antibiotic and cytotoxic indicator substance, urdamycin A. J. Chem. Res. (Synopsis) 1986, 104–105. 30. Rohr, J. and Zeeck, A. (1987) Urdamycins, new angucycline antibiotics from Streptomyces fradiae. II. Structural studies of urdamycins B to F. J. Antibiot. 40, 459–467. 31. Rohr, J., Zeeck, A., and Floss, H. G. (1988) Urdamycins, new angucycline antibiotics from Streptomyces fradiae. II. The structures of urdamycins C and D. J. Antibiot. 41, 126–129. 32. Rohr, J., Beale, J. M., and Floss, H. (1989) Urdamycins, new angucycline antibiotics from Streptomyces fradiae. IV. Biosynthetic studies of urdamycins A–D. J. Antibiot. 42, 1151–1157. 33. Korff, U. (1988) Strukturaufklärung, Synthese und Derivatisierung eines Pigments aus dem Stamm Tü 2484. Doctoral Thesis, Universität Tübingen. 34. Alvarado-Kirigin, M. (1990) Monomeres und dimeres Cinnachinon aus Streptomyces griseoflavus (Tü 2484). Doctoral Thesis, Universität Tübingen. 35. Keller-Schierlein, W., Geiger, A., Zähner, H., and Brandl, M. (1988) Die Esmeraldine A und B, tief grüne Farbstoffe aus Streptomyces antibioticus, Stamm TÜ 2706. Helv. Chim. Acta 71, 2058–2070. 36. Wessels, P., Göhrt, A., Zeeck, A., Drautz, H., and Zähner, H. (1991) Naphthgeranines, new naphthoquinone antibiotics from Streptomyces sp. J. Antibiot. 44, 1013–1018. 37. Ritzau, M., Drautz, H., Zähner, H., and Zeeck, A. (1993) Serpentene, a novel polyene carboxylic acid from Streptomyces. Liebigs Ann. Chem. 1993, 433–435.
48
References 38. Hagmann, L., Keller-Schierlein, W., Wahl, B., and Zähner, H. (1988) Pyridindolol glucosides from Streptomyces parvulus. J. Antibiot. 41, 289–295. 39. Gröne, C. (1991) Strukturaufklärung der Depsichlorine: Neue Depsipeptid-Antibiotika aus Streptomyces Tü 1661. Doctoral Thesis, Universität Göttingen. 40. Cullmann, H. J. (1994) Depsichlorine und andere Metabolite aus Streptomyces antibioticus Tü 1661. Doctoral Thesis, Universität Tübingen. 41. Keller-Schierlein, W., Müller, A., Hagmann, L., Schneider, U., and Zähner, H. (1985) (E)-3-(1H-Pyrrol-3-yl)-2-propensäure und (E)-3-(1H-Pyrrol-3-yl)-2-propensäureamid aus Streptomyces parvulus. Helv. Chim. Acta 68, 559–562. 42. Grote, R., Zeeck, A., Stümpfel, J., and Zähner, H. (1990) Pyrrolams, new pyrrolizidinones produced by Streptomyces olivaceus. Liebigs Ann. Chem. 1990, 525–530. 43. Hoff, H., Drautz, H., Fiedler, H.-P., Zähner, H., Schultz, J. E., Keller-Schierlein, W., Philipps, S., Ritzau, M., and Zeeck, A. (1992) Obscurolides, a novel class of phosphodiesterase inhibitors from Streptomyces. I. Production, isolation, structural elucidation and biological activity of obscurolides A1 to A4. J. Antibiot. 45, 1096–1107. 44. Potterat, O., Zähner, H., Metzger, J. W., and Freund, S. (1994) 5-Phenylpentadienoicacid derivatives from Streptomyces sp. Helv. Chim. Acta 77, 569–574. 45. Keller-Schierlein,W., Bahnmüller, U., Dobler, M., Bielecki, J., Stümpfel, J., and Zähner, H. (1986) Isolierung und Strukturaufklärung von Differolid. Helv. Chim. Acta 69, 1833–1836. 46. Müller, A., Keller-Schierlein, W., Bielecki, J., Rak, G., Stümpfel, J., and Zähner, H. (1988) (2S,3R,4R,6R)-2,3,4-Trihydroxy-6-methylcyclohexanon aus zwei ActinomycetenStämmen. Helv. Chim. Acta 69, 1829–1832. 47. Rapp, C., Jung, G., Isselhorst-Scharr, C., and Zähner, H. (1988) A new member of the class of antibiotics with thiotetronic acid structure isolated from Streptomyces olivaceus TÜ 3010. Liebigs Ann. Chem. 1988, 1043–1047. 48. Grote, R., Zeeck, A., Drautz, H., and Zähner, H. (1988) Colabomycins, new antibiotics of the manumycin group from Streptomyces griseoflavus. I. Isolation, characterization and biological properties. J. Antibiot. 41, 1178–1185. 49. Grote, R., Zeeck, A., and Beale Jr., J. M. (1988) Colabomycins, new antibiotics of the manumycin group from Streptomyces griseoflavus. II. Structure of colabomycin A. J. Antibiot. 41, 1186–1195. 50. Burkhardt, K., Fiedler, H.-P., Grabley, S., Thiericke, R., and Zeeck, A. (1996) New cineromycins and musacins obtained by metabolite pattern analysis of Streptomyces griseoviridis (FH-S 1832). I. Taxonomy, fermentation, isolation and biological activity. J. Antibiot. 49, 432–437. 51. Grote, R., Zeeck, A., Drautz, H., and Zähner, H. (1988) 2880-II, a metabolite related to ferulic acid from Streptomyces griseoflavus. J. Antibiot. 41, 1275–1276. 52. Niederer, D., Séquin, U., Drautz, H., and Zähner, H. (1990) (3S,5R,6E,8E)-Deca-6,8dien-1,3,5-triol, ein neues Stoffwechselprodukt von Streptomyces fimbriatus (Millard et Burr 1926). Helv. Chim. Acta 73, 2129–2135. 53. Wessels, P. (1990) Strukturaufklärung neuer Naphthochinon-Antibiotika aus Streptomyces Tü 3556 und Beiträge zum chemischen Screening mit extrem thermophilen Archaebakterien. Doctoral Thesis, Universität Göttingen. 54. Potterat, O., Zähner, H., Volkmann, C., and Zeeck, A. (1993) Exfoliamycin and related metabolites, new naphthoquinone antibiotics from Streptomyces exfoliatus. J. Antibiot. 46, 346–349. 55. Volkmann, C., Zeeck, A., Potterat, O., Zähner, H., Bohnen, F.-M., and Herbst-Irmer, R. (1995) The structures of the exfoliamycins. J. Antibiot. 48, 431–432. 56. Braun, D., Pauli, N., Séquin, U., and Zähner, H. (1995) New butenolides from the photoconductivity screening of Streptomyces antibioticus (Waksman and Woodruff). Waksman and Henrici 1948. FEMS Microbiol. Lett. 126, 37–42. 57. Fiedler, H.-P., Rohr, A., and Zeeck, A. (1986) Elloramycins B, C, D, E and F: minor congeners of the elloramycin producer Streptomyces olivaceus. J. Antibiot. 39, 856–859.
49
2 Screening for New Secondary Metabolites from Microorganisms 58. Fiedler, H.-P. (1986) Identification of new elloramycins, anthracycline-like antibiotics, in biological cultures by high-performance liquid chromatography and diode array detection. J. Chromatogr. 361, 432–436. 59. Rohr, J., Eick, S., Zeeck, A., Reuschenbach, P., Zähner, H., and Fiedler, H.-P. (1988) Tetracenomycins B3 and D3, key intermediates of the elloramycin and tetracenomycin C biosynthesis. J. Antibiot. 41, 1066–1073. 60. Geiger, A., Keller-Schierlein, W., Brandl, M., and Zähner, H. (1988) Phenazines from Streptomyces antibioticus strain TÜ 2706. J. Antibiot. 41, 1542–1551. 61. König, W. A., Hahn, H., Rathmann, R., Hass, W., Keckeisen, A., Hagenmaier, H., Bormann, C., Dehler, W., Kurth, R., and Zähner, H. (1996) Drei neue Aminosäuren aus dem Nikkomycin-Komplex – Strukturaufklärung und Synthese. Liebigs Ann. Chem. 1986, 407–421. 62. Decker, H., Bormann, C., Fiedler, H.-P, Zähner, H., Heitsch, H., and König, W. A. (1989) Isolation of new nikkomycins from Streptomyces tendae. J. Antibiot. 42, 230– 235. 63. Heitsch, H., König, W. A., Decker, H., Bormann, C., Fiedler, H.-P., and Zähner, H. (1989) Structure of the new nikkomycins pseudo-Z and pseudo-J. J. Antibiot. 42, 711–717. 64. Bormann, C., Mattern, S., Schrempf, H., Fiedler, H.-P., and Zähner, H. (1989) Isolation of Streptomyces tendae mutants with an altered nikkomycin spectrum. J. Antibiot. 42, 913–918. 65. Decker, H., Walz, F., Bormann, C., Zähner, H., and Fiedler, H.-P. (1990) Nikkomycins Wz and Wx, new chitin synthetase inhibitors from Streptomyces tendae. J. Antibiot. 43, 43–48. 66. Decker, H., Pfefferle, U., Bormann, C., Zähner, H., Fiedler, H.-P., van Pée, K.-H., Rieck, M., and König, W. A. (1991) Enzymatic bromination of nikkomycin Z. J. Antibiot. 44, 626–634. 67. Schüz, T. C., Fiedler, H.-P., Zähner, H., Rieck, M., and König, W A. (1992) Nikkomycins Sz, Sx, Soz and Sox, new intermediates associated to the nikkomycin biosynthesis. J. Antibiot. 45, 199–206. 68. Decker, H., Zähner, H., Heitsch, H., König, W. A., and Fiedler, H.-P. (1991) Structureactivity relationships of the nikkomycins. J. Gen. Microbiol. 137, 1805–1813. 69. Fiedler, H.-P., Kulik, A., Schüz, T. C., Volkmann, C., and Zeeck, A. (1994) Juglomycin Z, a new naphthoquinone antibiotic from Streptomyces tendae. J. Antibiot. 47, 1166– 1122. 70. Ritzau, M., Philipps, S., Zeeck, A., Hoff, H., and Zähner, H. (1993) Obscurolides, a novel class of phosphodiesterase inhibitors from Streptomyces. II. Minor components belonging to the obscurolide B to D series. J. Antibiot. 46, 1625–1628. 71. Volkmann, C., Hartjen, U., Zeeck, A., and Fiedler, H.-P. (1995) Naphthgeranine F, a minor congener of the naphthgeranine group produced by Streptomyces violaceus. J. Antibiot. 48, 522–524. 72. Blum, S., Fiedler, H.-P., Groth, I., Kempter, C., Stephan, H., Nicholson, G., Metzger, J. W., and Jung, G. (1995) Echinoserine, a new member of the quinoxaline group, produced by Streptomyces tendae. J. Antibiot. 48, 619–625. 73. Blum, S., Groth, I., Rohr, J., and Fiedler, H.-P. (1996) Dioxolides, novel secondary metabolites from Streptomyces tendae. J. Basic Microbiol. 36, 19–25. 74. Fiedler, H.-P., Nega, M., Pfefferle, C., Groth, I., Kempter, C., Stephan, H., and Metzger, J. W. (1996) Kanchanamycins, new polyol macrolide antibiotics produced by Streptomyces olivaceus Tü 4018. I. Taxonomy, fermentation, isolation and biological activities. J. Antibiot. 49, 101–107. 75. Stephan, H., Kempter, C., Metzger, J. W., Jung, G., Potterat, O., Pfefferle, C., and Fiedler, H.-P. (1996) Kanchanamycins, new polyol macrolide antibiotics produced by Streptomyces olivaceus Tü 4018. II. Structure elucidation. J. Antibiot. 49, 109–113.
50
References 76. Pfefferle, C., Kempter, C., Metzger, J. W., and Fiedler, H.-P. (1996) (E)-4-oxonon-2enoic acid, an antibiotically active fatty acid produced by Streptomyces olivaceus Tü 4018. J. Antibiot. 49, 135–137. 77. Pfefferle, C., Breinholt, J., Gürtler, H., and Fiedler, H.-P. (1997) 1-Hydroxy-4-methoxy2-naphthoic acid, a herbicidal compound produced by Streptosporangium cinnabarinum ATCC 31213. J. Antibiot. 50, 1067–1068. 78. Höltzel, A., Kempter, C., Metzger, J. W., Jung, G., Groth, I., Fritz, T., and Fiedler, H.P. (1998) Spirofungin, a new antifungal antibiotic from Streptomyces violaceusniger Tü 4113. J. Antibiot. 51, 487–495. 79. Breinholt, J., Kulik, A., Gürtler, H., and Fiedler, H.-P. (1998) Tigloside: a new tigloylated tetrasaccharide from Amycolatopsis sp. Acta Chem. Scand. 52, 1239–1242. 80. Maier, A., Müller, J., Schneider, P., Fiedler, H.-P., Groth, I., Tayman, F. S. K., Teltschik, F., Günther, C., and Bringmann, G. (1999) (2E,4Z)-Decadienoic acid and (2E,4Z,7Z)decatrienoic acid, two herbicidal metabolites from Streptomyces viridochromogenes Tü 6105. Pesticide Sci. 55, 733–739. 81. Schimana, J., Fiedler, H.-P., Groth, I., Süßmuth, R., Beil, W., Walker, M., and Zeeck, A. (2000) Simocyclinones, novel cytostatic angucyclinone antibiotics produced by Streptomyces antibioticus Tü 6040. I. Taxonomy, fermentation, isolation and biological activities. J. Antibiot. 53, 779–787 82. Walker, M., Zeeck, A., Schimana, J., and Fiedler, H.-P. (2000) Simocyclinones, novel cytostatic angucyclinone antibiotics produced by Streptomyces antibioticus Tü 6040. II. Structure determination and biosynthesis. J. Antibiot., in press. 83. Walker, M., Schimana, J., Süßmuth, R., Beil, W., Fiedler, H.-P., and Zeeck, A. (2000) New simocyclinones of the A-, B-, C- and D-series, novel angucyclinone antibiotics from Streptomyces antibioticus Tü 6040. J. Antibiot., in press. 84. Schimana, J., Walker, M., Zeeck, A., and Fiedler, H.-P. (2000) Simocyclinones: diversity of metabolites is dependent on fermentation conditions. J. Ind. Microbiol. Biotechnol., in press. 85. Pfefferle, C., Breinholt, J., Olsen, C. E., Kroppenstedt, R. M., Gürtler, H., and Fiedler, H.-P. (2000) Kyanomycin, a complex of unusual anthracycline-phospholipid hybrid from Nonomuria species. J. Nat. Prod. 63, 295–298. 86. Kuhn, W. (1993) Untersuchungen zur Produktion der Esmeraldine und der Physiologie von Streptomyces antibioticus. Doctoral Thesis, Universität Tübingen. 87. Fiedler, H.-P., Meiwes, J., Werner, I., Konetschny-Rapp, S., and Jung, G. (1990) Identification of new ferrioxamines by HPLC and diode array detection. J. Chromatogr. 513, 255–262. 88. Meiwes, J., Fiedler, H.-P., Zähner, H., Konetschny-Rapp, S., and Jung, G. (1990) Production of desferrioxamine E and new analogues by directed fermentation and feeding fermentation. Appl. Microbiol. Biotechnol. 32, 505–510. 89. Konetschny-Rapp, S., Jung, G., Raymond, K. N., Meiwes, J., and Zähner, H. (1992) Solution thermodynamics of the ferric complexes of new desferrioxamine siderophores obtained by directed fermentations. J. Am. Chem. Soc. 114, 2224–2230.
51
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin Friedrich Götz* and Günther Jung**
3.1 History of lantibiotics and lantibiotic research in Tübingen
The first lantibiotic structure elucidated was that of nisin [23], and nisin was probably the first lantibiotic identified [65]. Nisin is produced by the cheese starter culture organism Lactococcus lactis subsp. lactis and has an antimicrobial effect on a broad variety of Gram-positive bacteria [23]. The pioneering research of E. Gross and coworkers demonstrated that the peptide antibiotics nisin and subtilin, the latter produced by Bacillus subtilis, actually contain lanthionine (Lan) and 3-methyllanthionine (MeLan) as well as (Dha) and (Dhb), confirming previous hypotheses [6, 21, 22]. The lantibiotic era in Tübingen began in 1985 with the publication of the peptide sequence of epidermin isolated from Staphylococcus epidermidis. With the elucidation of the structure of epidermin, it became clear that it is a heterodet tetracyclic 22-amino acid (2164 Da), amide peptide [1] that contains one residue each of Dhb and MeLan and two residues of Lan. The fourth cyclic structure results from the novel C-terminal mono-carboxy, di-amino acid, AviCys. It was not clear at that time how these amino acid structures arose and how the complex structures of nisin and epidermin are synthesized. Are they non-ribosomally synthesized, as are gramidicin and valinomycin, which are typically synthesized by large multi-enzyme complexes in the cell and for which no structural genes exists [27, 33], or are they synthesized ribosomally as a precursor peptide, which is subsequently posttranslationally modified? In 1987, when Friedrich Götz took over the chair in Mikrobielle Genetik in Tübingen, he and his coworkers started to unravel the biosynthetic principles of
* Mikrobielle Genetik, Universität Tübingen, Waldhäuser Str. 70/8, D-72076 Tübingen ** Institut für Organische Chemie, Universität Tübingen, Auf der Morgenstelle 18, D-72076 Tübingen
52 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
3.1 History of lantibiotics and lantibiotic research in Tübingen epidermin. In co-operation with Karl-Dieter Entian’s group, a gene corresponding to the epidermin amino acid sequence was identified on a 54-kb plasmid of the epidermin-producing strain [72]. These pioneering results brought lantibiotic research a big step forward. For the first time it was shown that a lantibiotic is ribosomally synthesized, and clues on the organization of epidermin as a precursor protein and the processing and post-translational modification steps were found. In the following years, the principles of the genetic organization and biosynthesis first found with epidermin were subsequently also found with other lantibiotics – all other lantibiotics studied later were also shown to be ribosomally synthesized. To date, antibiotics have been found in many Gram-positive genera, such as Bacillus, Lactococcus, Pediococcus, Staphylococcus, Streptococcus, and Streptomyces. Lantibiotics have not been found in a Gram-negative species. Subsequent studies of lantibiotics have revealed that they are post-translationally modified at specific positions to give rise to the large number of modified amino acids found in these peptides. In addition, the peptides are produced with a leader peptide, which is removed during maturation, and are transported by specific transport-related proteins out of the cell. Still other lantibiotic-specific proteins are involved in the genetic regulation of biosynthesis and generation of the specific producer-cell self-protection mechanism(s) frequently observed. The name “lantibiotics” was coined in Tübingen and refers to the rapidly expanding group of antibiotic-like peptides that contain the non-protein amino acids lanthionine and 3-methyllanthionine [72]. The discussion whether epidermin and nisin should be regarded as antibiotics or as bacteriocins was intense. The small size and compact structure of the compounds and the wide range of antibacterial activity against most Gram-positive and some Gram-negative bacteria favors consideration as antibiotics – bacteriocins are usually larger proteins with a narrow range of activity, such as colicins. On the other hand, typical amino-acid-derived antibiotics, such as penicillin and gramicidin, are non-ribosomally synthesized. With the knowledge available today, it is clear that lantibiotics fulfill criteria of both, antibiotics and bacteriocins and cannot be placed in one or the other category. The name lantibiotic belies the full extent and complexity of this class of bacterial peptides. A strong and very fruitful collaboration was developed with Hans-Georg Sahl (Bonn) and his coworkers. Since 1983, they had been studying the mode of action of Pep5 and nisin, and later also of gallidermin and other lantibiotics. Earlier studies suggested that nisin could interfere with the biosynthesis of the bacterial cell wall [46, 64]. This appears to be indeed true, as recent studies show [13]. However, the binding to cell wall precursors does not explain the observed bactericidal effect of lantibiotics. Type A lantibiotics form non-specific transmembrane pores in an energy-dependent fashion and allow efflux of preaccumulated intracellular components [66, 69, 73]. In 1984, Friedrich Götz, then at the Technical University in München, isolated an antibiotic compound from a Staphylococcus gallinarum strain, which belongs to a species previously described by Devriese et al. [17]. Preliminary 53
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin studies showed that this antibiotic compound exerts a rather broad activity against Gram-positive bacteria. However, it was not until he moved to Tübingen that the compound was isolated in the group of Hans Zähner and the structure was elucidated in the group of Günther Jung [28]. The compound was named gallidermin after the species name of the producing staphylococcus. Gallidermin is a structural analogue of epidermin, but is slightly more active than epidermin. Today, after long and laborious work to optimize the fermentation and thus increase gallidermin production approximately hundred-fold [29], S. gallinarum Tü 3928 is used for the biotechnological production of gallidermin by the groups of Peter Fiedler (fermentation) and Rolf Werner (downstream processing; Boehringer Ingelheim, Biberach). To say that in the years following 1986, Tübingen became a stronghold in lantibiotic research, by setting the pace of the research, inspiring many other groups, and initiating quite a number of national and international co-operations, is no exaggeration. Three international workshops on lantibiotics have already taken place. Günther Jung (Tübingen) initiated together with HansGeorg Sahl (Bonn) the first lantibiotic workshop, held in April 1991 in Bad Honnef. The contributions were published under the title “Nisin and novel lantibiotics” with Jung and Sahl as editors (Leiden: Escom). The second workshop on “Lantibiotics: a unique group of antibiotic peptides” was organized by Ruud Konigs and Cees Hilbers (Nijmegen, The Netherlands) and held in Arnhem in November 1994. The contributions were published in Antonie van Leeuwenhoek, vol. 69 in 1996. The third workshop on “Lantibiotics and related modified antibiotic peptides” was organized by Friedrich Götz, R. Jack, Günther Jung, and Hans-Georg Sahl and was held in Blaubeuren in April 1998. The ever present international interest in the lantibiotic subject is best documented by the increasing number of applicants and participants from workshop to workshop. A number of potential applications have been found for the mature lantibiotics, including the use as an anti-infective in medical and veterinary areas, in food, beverage and cosmetic preservation, and as regulators of both, human immune function and blood pressure [26, 50, 62, 67]. Lantibiotics represent new lead structures, but their chemical synthesis is too complicated and costly; therefore, the only way to produce large enough quantities for marketing is through biotechnology. In the following sections, more detailed information on lantibiotic research is presented. Since this is a final report on all achievements obtained in the framework of the collaborative research centre 323, we will focus on the results obtained by the groups in Tübingen. However, the picture would be incomplete if major achievements of other groups were not included. The topics on novel structures, mechanism(s) of biosynthesis, genetic organization and regulation, biological activities and mode of actions as well as potential applications for this fascinating, novel class of bacterial-derived, biologically active peptides will be addressed.
54
3.2 Primary structure and proposed maturation of epidermin in staphylococci
3.2 Primary structure and proposed maturation of epidermin in staphylococci
The nucleotide sequence of the epidermin structural gene, epiA, revealed that epidermin is part of a pre-peptide that consists of a 30-amino-acid N-terminal leader peptide and a 22-amino-acid C-terminal propeptide [72]. A comparison of the chemical structure of epidermin [2] with the peptide sequence deduced from epiA indicates that this propeptide undergoes several modifications (Fig. 3.1) before it is transported out of the cell. At each pair of positions where the mature epidermin contains a lanthionine bridge, the precursor peptide contains one serine or threonine and one cysteine. Lanthionine is proposed to be formed in a two-step process involving dehydration of serine and threonine and the subsequent addition of a cysteine thiol group [10, 24].
Figure 3.1: Posttranslational modifications of the epidermin pre-peptide. (A) Model of lanthionine synthesis: dehydration of serine (a) is followed by a nucleophilic addition reaction of a thiol group, thereby forming a thioether bridge (b); the formation of meso-lanthionine from threonine is analogous. (B) Biosynthesis of epidermin: dehydration (a), nucleophilic addition (b), oxidative decarboxylation (c), and removal of the leader peptide (d). Abu, aminobutyric acid; Dhb, dehydrobutyrine.
55
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.3 Genetic organization and regulation of the epidermin genes
3.3.1 Organization and function of epidermin genes The gene cluster for biosynthesis of the lantibiotic epidermin is composed of eleven genes in five transcription units. The genes are involved in 1) synthesis of the epidermin precursor peptide, 2) modification of the precursor peptide, 3) secretion of epidermin, 4) cleavage of the leader peptide, 5) immunity of the producer to epidermin, and 6) regulation of the gene cluster. Four of the transcription units are activated by the DNA-binding regulator EpiQ. The factors or conditions leading to activation of the EpiQ protein or its expression remain unclear. The gene cluster for epidermin biosynthesis, located on plasmid pTü32 of S. epidermidis Tü 3298 [72], contains genes for all activities proposed to be involved in the biosynthetic pathway (Fig. 3.2). Like all lantibiotics, epidermin is ribosomally synthesized and the gene cluster accordingly contains the structural gene epiA, which encodes the precursor peptide [70, 72]. In the same operon, the genes for modification reactions leading to the formation of thioether bridges (epiB and epiC) [4, 41, 58] and to oxidative decarboxylation of the C-terminus (epiD) [42, 43] are encoded. Between epiA and epiB, a weak terminatorlike structure reduces the expression of the subsequent genes [56, 70] whose products are required only in catalytic amounts. In a second transcriptional unit, the extracellular leader peptidase EpiP [20] and the regulator protein EpiQ [56] are encoded. Upstream of epiA, the individually expressed genes epiT and epiH are located [60]. The deduced product of epiT shares sequence similarity with an ABC-type transporter that is involved in secretion of certain peptides and proteins [75]. The coding sequence
Figure 3.2: Organization of the epidermin gene cluster encoded on plasmid pTü32. Epidermin genes are shown as white arrows; flanking genes with putative functions in plasmid replication (rep) and maintenance (resolvase) are indicated in gray. Gene functions: epiA, structural gene; epiB and epiC, dehydration and lanthionine formation; epiD, flavoprotein (oxidative decarboxylation); epiQ, activator; epiP, pro-epidermin processing protease; epiH and epiT, translocation (export) of pro-epidermin; epiFEG, epidermin immunity. Arrows indicate promoters that are activated by EpiQ. The epiT gene is incomplete due to a deletion and frame-shift mutation, but its function can be complemented by the intact gallidermin gene, gdmT.
56
3.3 Genetic organization and regulation of the epidermin genes is disrupted by two frame-shift deletions, and it is very questionable whether epiT has a function in epidermin biosynthesis. Transcriptional reporter gene fusions have demonstrated, however, that epiT is expressed (see below). The homologous gene gdmT from the gene cluster of the closely related lantibiotic gallidermin, has an intact sequence; it mediates an increase of epidermin production in the heterologous host Staphylococcus carnosus [60] and thereby substantiates the proposed capacity of GdmT to secrete epidermin or gallidermin. The gene adjacent to gdmT – gdmH – is also necessary for increased epidermin production; GdmH may be an accessory factor for secretion. EpiH and GdmH are hydrophobic proteins without conspicuous similarities to other proteins; they furthermore exert a limited level of immunity to epidermin [57]. Upstream of epiH, three cotranscribed genes, epiF, epiE, and epiG, are encoded. They mediate resistance (immunity) to epidermin, and their products very likely constitute the subunits of an ABC transporter that expels the harmful epidermin molecules from the cytoplasmic membrane [54, 57]. The epidermin genes are flanked by genes which apparently encode factors for replication and stability of plasmid pTü32 (Fig. 3.2). The res gene product has a high level of identity to plasmid resolvases, whose function is the resolution of cointegrates [58]. The open reading frame upstream of epiP shares significant similarity with plasmid replication genes from various Gram-positive bacteria (M. Hille and A. Peschel, unpublished). The epidermin gene cluster thus seems to be complete; however, several chromosomally encoded genes are also very likely involved in epidermin biosynthesis.
3.3.2 EpiQ is an unusual transcriptional regulator EpiQ shares similarities with DNA-binding proteins of the response regulator family [70] and binds to the epiABCD promoter region to activate expression [56]. Unlike conventional response regulators, EpiQ does not contain the conserved aspartic acid residue that is phosphorylated by a corresponding sensor kinase to activate the regulator [76]. Since a corresponding kinase gene is absent from the epidermin gene cluster, it is very questionable whether EpiQ is phosphorylated at all. Several other lantibiotic gene clusters encode a conventional kinase and regulator pair [75]; therefore, the regulator gene epiQ is unusual. The twocomponent regulatory systems of nisin and subtilin activate the biosynthetic genes in response to the extracellular concentration of the respective lantibiotic [36]. Nisin and subtilin thus act as bacteriocin-like antimicrobial peptides and as quorum-sensing peptide pheromones that control their own expression. Synthesis of epidermin seems to be regulated differently. Regulating agents or conditions have not yet been identified, but EpiQ seems to be active under all laboratory conditions tested so far (A. Peschel, unpublished).
57
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.3.3 Regulation of epidermin genes
Lipase act ivit y ( U/ g cell dry wt .)
Trans-activation experiments with transcriptional reporter gene fusions have demonstrated that EpiQ controls most of the transcriptional units in the epidermin gene cluster; the promoters of the epiFEG, epiH, epiT, and epiABCD gene clusters are activated [56, 58, 60] (Fig. 3.3). All these promoter regions have a palindromic sequence motif (agAaAATTAC – 6 bp – GTAATTtTct) located immediately upstream of the only weakly conserved –35 regions. The motif upstream of the epiABCD promoter is the EpiQ binding site [56]. The EpiQ operator is also present in the gallidermin gene clusters at the same positions, and the promoters are also sensitive to EpiQ. - E p iQ
1000
+ E pi Q 100
10
1
Pep i FE G
Pep i H
Pep i T
Pep i AB C D
Pep i PQ
Figure 3.3: Promoter activities of various epidermin genes. Extracellular lipase activities of S. carnosus strains carrying a promoter test plasmid containing the lipase gene as a reporter under the control of the indicated epidermin gene promoters (Pepi). Lipase expression was determined in the presence and absence of the activator gene, epiQ, encoded on a second plasmid.
The only transcriptional unit not controlled by EpiQ is the epiPQ operon (Fig. 3.2); EpiQ therefore exerts no autoregulation. The epidermin gene cluster contains additional promoters within the operon structures that are not controlled by EpiQ, e. g. in front of epiC, epiD, and epiQ. The epiD promoter has a particularly high activity (Peschel et al., unpublished). The regulation of the epidermin genes remains elusive since no activating agents have been found. Recent literature illustrates the importance of global regulatory relays in Gram-positive bacteria [31]. Quorum-sensing systems control the production of antimicrobial substances in B. subtilis, and it is conceivable that a similar, yet unknown system controls the epidermin genes. A further possibility is the involvement of special sigma factors that couple epidermin production to certain growth phases or environmental conditions. In this respect, it is interesting to note that epidermin production occurs mainly during the exponential growth phase and is obviously switched off in the stationary phase [54]. Since EpiQ con58
3.4 Isolation and characterization of genetically engineered gallidermin trols most of the epidermin genes, the promoters controlling its own expression are likely targets for a chromosomally encoded regulatory system. Chromosomally encoded proteins may also be involved in secretion of epidermin, thereby substituting for the defective epiT, and may be involved in the modification reactions.
3.4 Isolation and characterization of genetically engineered gallidermin and epidermin analogues
Gallidermin (Gdm) and epidermin (Epi) are highly similar. To study the substrate specificity of the modifying enzymes and to find variants of the lantibiotics with altered or new biological activities, we exchanged certain amino acids of the gallidermin and epidermin structural genes, gdmA and epiA, respectively, by sitespecific mutagenesis [53]. No epidermin/gallidermin analogues are found in the supernatant when 1) the hydroxyamino acids involved in thioether amino acid formation are substituted by non-hydroxyamino acids (S3N and S19A); 2) cysteine residues involved in thioether bridging are deleted (C21, C22 and C22); or 3) a ring amino acid is substituted by an amino acid with a completely different character (G10E and Y20G). Production is greatly decreased when serine residues involved in thioether amino acid formation are exchanged by threonine residues (S16T, S19T). A number of conservative exchanges at positions 6, 12, or 14 of the gallidermin backbone are tolerated and lead to analogues with altered biological properties, such as an enhanced antimicrobial activity (L6V) or a remarkable resistance to proteolytic degradation (A12L and Dhb14P). The T14S substitution leads to the simultaneous production of two gallidermin species formed by an incomplete posttranslational modification (dehydration) of the S14 residue. The fully modified Dhb14Dha analogue has antimicrobial activity similar to gallidermin, whereas the Dhb14S analogue is less active. Both peptides are more sensitive to tryptic cleavage than gallidermin. The construction and characterization of the various analogues are described in more detail in the following sections.
3.4.1 Characterization of two Epi – mutants and development of a host-vector system for expression of wild type and mutated gdmA and epiA genes We isolated a series of Epi – mutants of S. epidermidis Tü 3298 by ethyl-methanesulfonate(EMS) mutagenesis; two mutants (EMS5 and 6) carry mutations in the epiA region [4]. Both mutants can be complemented to an Epi+ phenotype by 59
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin transformation with a plasmid carrying epiA, which indicates that other epi biosynthetic genes are functional. Single point mutations in epiA result in an S3N substitution (ring A) in EMS5 and a G10E substitution (ring B) in EMS6 (Fig. 3.1). Analytical HPLC analyses indicated that no epidermin analogue or precursor peptide is present in the culture supernatants of EMS5 and EMS6 [53]. We chose the mutant S. epidermidis EMS6 as an expression host for the synthesis of epidermin and gallidermin analogues. To test this mutant for its ability to synthesize the heterologous gallidermin, a 1.6-kb SalI/EcoRI subfragment (encoding gdmA and a part of gdmB and of gdmT) of the 4.8-kb EcoRI chromosomal fragment of S. gallinarum Tü 3928 [71] was cloned into the polylinker region of the staphylococcal plasmid pT181mcs. S. carnosus TM300 was transformed with the resultant plasmid, pTgdmA [53]. After isolation, the plasmid was transferred into S. epidermidis EMS6 by electroporation. This indirect transformation method is necessary because S. epidermidis Tü 3298 and derivative strains are poorly transformed with ligation products; only ccc plasmids are electroporated efficiently. EMS6 transformants carrying plasmid pTgdmA formed halos on plates containing Micrococcus luteus, and analyses revealed that only fully modified gallidermin is produced.
3.4.2 Characterization of gallidermin and epidermin analogues generated by site-directed mutagenesis The gallidermin and epidermin analogues were isolated from the culture supernatant of various S. epidermidis EMS6 clones, purified, and analyzed by analytical HPLC, electrospray mass spectrometry (ES-MS), and continuous Edman sequencing. MIC values and tryptic sensitivity were determined for all gallidermin analogues, except Gdm L6V/S16T and epidermin S19T because these peptides were only produced in trace amounts. The results are summarized in Table 3.1 [53].
3.4.2.1 Mutations in ring A Since Gdm (L6) is more active than Epi (I6) against various Gram-positive bacteria (23), we generated two more analogues of gallidermin with mutations at position 6: L6V and L6G. The production and modification of these two analogues are not impaired. The L6G analogue is less active than gallidermin, regardless of the indicator strain employed. The L6V analogue is twice as active as gallidermin against Micrococcus luteus and Corynebacterium glutamicum, and just as active against Arthrobacter cristallopoietes and B. subtilis. Antimicrobial activity against S. aureus is reduced fivefold (Table 3.1). The L6G analogue is almost as sensitive to trypsin as gallidermin, whereas the L6V analogue is more resistant. 60
3.4 Isolation and characterization of genetically engineered gallidermin Table 3.1: Antimicrobial activities of gallidermin and its analogues against various Grampositive indicator strains [53]. Gdm + derivatives
Gdm Gdm Dhb14Dha Gdm Dhb14S Gdm Dhb14A Gdm Dhb14P Gdm A12L Gdm L6G Gdm L6V a
Minimal inhibitory concentration (lg/ml)a Micrococcus Arthrobacter luteus cristallopoietes
Corynebacterium glutamicum
Bacillus subtilis
Staphyloc. aureus Cowan I
0.004 0.004 0.008 0.004 0.12 0.15 0.008 0.002
0.065 0.065 0.13 0.065 0.52 0.13 0.13 0.032
3.0 3.0 15.0 >20.0 >20.0 10.0 8.0 3.0
4.0 5.0 25.0 30.0 >30.0 12.0 30.0 20.0
0.005 0.005 0.01 0.02 0.08 0.02 0.04 0.005
MIC values obtained for M. luteus, A. cristallopoietes, and C. glutamicum are located within the nanomolar range (0.002–0.52 mg/ml; *1–240 nM), whereas those observed for B. subtilis and S. aureus are located within the micromolar range (3.0 to 630 mg/ml; 1.4 to 614 mM).
3.4.2.2 Mutations in ring C and D To determine the importance of the C-terminal bicyclic structure (intertwined rings C and D) for epidermin/gallidermin biosynthesis, we exchanged various amino acids in this region. Using the gdmA-L6V gene, we created the double mutant Gdm L6V/S16T. The production of this analogue is only 8% of that observed for Gdm L6V. Molecular mass determinations of Gdm L6V/S16T revealed that the T16 residue is dehydrated. However, it was not possible to verify the formation of the 3-methyllanthionine in ring C by ES-MS. Gdm L6V/S16T has antimicrobial activity; however, the MIC values were not determined because of the very low production of this analogue. The S19T codon of EpiA was replaced in order to create a posttranslationally formed S-(2-aminovinyl)-2-methyl-d-cysteine residue as found in mersacidin. This analogue has antimicrobial activity and is also produced in very low amounts. The production is only 0.4% of that observed for the gallidermin-producing EMS6 (pTgdmA) clone, which produces approximately 12 mg Gdm/l. ES-MS analysis revealed an average molecular mass of 2177 Da, indicating that T19 is dehydrated and that the C-terminus is decarboxylated. It was also not possible to verify the formation of the S-(2-aminovinyl)-2-methyl-d-cysteine residue by ES-MS; however, thioether bridge formation is very likely because of the observed chemical instability of the enethiol structure [42, 43]. To prevent formation of the S-(2-aminovinyl)-d-cysteine residue in ring D, we generated an S19A (pTepiA-S19A) substitution and a C22 deletion (pTepiADC22). These mutations could possibly lead to the formation of alternative thioether bridges. However, the corresponding EMS6 clones form no halos on 61
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin plates of M. luteus, and no epidermin analogue or precursor form was detected by analytical HPLC. In addition no antimicrobial activity and no extracellular product were detected when the last two C residues (pTepiA-DC21, C22) were deleted. An amino acid substitution not directly involved in thioether bridge formation of ring D was created by replacing the bulky Y20 residue with a G (pCUgdmA-Y20G). The corresponding EMS6 clone has no antimicrobial activity, and no gallidermin analogue or precursor peptide was detected by analytical HPLC.
3.4.2.3 Mutations in the flexible middle region (A12 to G15) With the A12L substitution, a more bulky residue was introduced in the flexible middle region. The production of this analogue is comparable to that of gallidermin; however, antimicrobial activity against most of the test strains is reduced threefold (Table 3.1). A striking feature of this analogue is its high resistance to tryptic cleavage. Further amino acid alterations mainly focused on the 2,3-unsaturated Dhb14 residue to investigate whether the reactive C=C double bond is important for biological activity, as suggested for the Dha5 residue of subtilin and nisin [38, 48], or whether this residue plays a role in stabilizing a distinct structure required for pore-forming activity, as reported for the P residue in the channelforming peptides alamethicin and melittin [78]. With the T14S substitution, two analogues were produced in nearly equimolar ratios, as judged from the respective peak areas. By ES-MS and continuous Edman degradation, the two peptides were identified as gallidermin analogues possessing either a dehydrated (Dha14) or an unmodified S14 residue (Fig. 3.1A, B). Antimicrobial activity of Gdm Dhb14Dha is similar to that of gallidermin, whereas the Dhb14S analogue is less active, especially against B. subtilis and S. aureus (Table 3.1). Both analogues were more sensitive to tryptic cleavage than gallidermin; the Dhb14S analogue was one of the most sensitive analogues, being completely cleaved after 30 min. To remove the reactive C=C double bond, a Dhb14A analogue was created. Slightly more of this analogue is produced as compared to gallidermin. Antimicrobial activity of Gdm Dhb14A is similar to that of gallidermin against M. luteus and C. glutamicum and approximately seven-fold less against B. subtilis and S. aureus (Table 3.1). Like Dhb14S, the Dhb14A analogue is more sensitive to tryptic cleavage and is also fully degraded within 30 min. The Dhb14P substitution was expected to cause the strongest conformational change of the middle region. Production of this analogue is comparable to that of gallidermin. The antimicrobial activity is generally reduced 8- to 30-fold (Table 3.1). This analogue has a pronounced resistance to tryptic cleavage.
62
3.4 Isolation and characterization of genetically engineered gallidermin
3.4.3 Overview of the characteristics of gallidermin and epidermin analogues We demonstrated that although epidermin and gallidermin are produced by two different staphylococcal species, gallidermin can be synthesized successfully using the heterologous S. epidermidis EMS6 as expression host and the cloned gdmA gene. The gdmA promoter is controlled by the transcriptional activator EpiQ with an efficiency similar to the epiA promoter [60], and the enzymes involved in epidermin biosynthesis (including maturation and secretion) function efficiently with the gdmA gene product. Mutagenesis of amino acid positions that are directly involved in thioether amino acid formation result in the loss of or a large decrease in production. Loss of production (i. e. no HPLC-detectable extracellular product) is observed with the analogues containing the S19A substitution, the deletion of the C22 residue (ring D, Fig. 3.1), and the S3N substitution (ring A). Thus, it can be hypothesized that biosynthesis and/or secretion of epidermin and gallidermin are severely impaired when formation of only one of the four thioether bridges is prevented. There is also no production observed with the analogues containing the G10E (ring B) and the Y20G (ring D) substitutions, which suggests that these mutations interfere with thioether bridge formation even though the substituted residues are not directly involved. The replacement of the G10 residue for a bulky and negatively charged E residue may impose steric hindrance on thioether bridge formation at ring B. 2D-NMR studies of gallidermin [18] revealed that ring B, which is identical to ring B of nisin and subtilin, adopts a typical b-turn type II conformation. This specific conformation would be disturbed by a G10E exchange and, as a consequence, thioether bridge formation might be affected, regardless whether this reaction is enzyme-catalyzed or occurs spontaneously in a Michael-additionlike reaction [77]. Recent investigations of the substrate specificity of EpiD [43] indicate that a precursor molecule possessing a G residue instead of the Y20 residue is not a substrate of EpiD. This result suggests that at least oxidative decarboxylation of the last C residue of the mutant Y20G precursor peptide is prevented. For all mutations that lead to the loss of production, further investigations will determine whether only secretion or whether also other stages of biosynthesis are blocked. It also cannot entirely be ruled out that partially modified precursor peptides are rapidly degraded by the EMS6 host. The only thioether amino acid mutations that result in the production of antimicrobially active substances are the S19T (ring D) and the S16T (ring C) substitutions. The extremely low production of Epi S19T and Gdm L6V/S16T could be explained by assuming that dehydration of T16 and T19 is inefficient and that only fully modified (dehydrated and subsequently thioether-bridged) precursor peptides are efficiently secreted. The simultaneous production of Gdm Dhb14S and Gdm Dhb14Dha is an indication of different efficiencies of Sand T-dehydration. For the applied use of peptide antibiotics, an advantageous trait would be resistance to proteolytic degradation. Epidermin and gallidermin already natu63
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin rally have a rather high resistance. Among various peptidases tested (pronase, trypsin, pepsin, thermolysin, collagenase, and carboxypeptidases A and B), only pronase and trypsin cleave epidermin [2]. Proteolytic degradation by the other peptidases is sterically hindered by the thioether bridges. In one study, purified gallidermin and analogues were tested for their sensitivity to tryptic cleavage in order to screen for analogues with an increased resistance to proteolytic degradation and, as a consequence, a prolonged period of action. All gallidermin analogues are cleaved by trypsin only in the central part of the molecule between K13 and the residue at position 14, which is in agreement with earlier reports on gallidermin [28] and epidermin [2]. Tryptic cleavage at the second putative tryptic cleavage site, the K4-F5 bond (ring A), is sterically hindered, at least under the conditions employed. 2D-NMR studies of the gallidermin molecule [18, 19] revealed that the rigid rings A/B and C/D are connected by a flexible middle region (A12 to G15). The conformation (e. g. flexibility of the peptide backbone) of this region appears to be important for antimicrobial activity and resistance to tryptic cleavage. Gallidermin analogues with a supposed decreased flexibility of the middle region, such as Gdm Dhb14P and Gdm A12L, are highly resistant to tryptic cleavage, whereas their antimicrobial activity against various Gram-positive indicator bacteria is greatly reduced (Table 3.1). According to modeling studies of the gallidermin/trypsin interaction [26], the gallidermin molecule must adopt a bent structure in order to fit into the catalytic cleft of the trypsin molecule. This bent structure is supported by the flexibility of the peptide backbone within the middle region. Thus, the high resistance to tryptic cleavage observed for Gdm Dhb14P and Gdm A12L may reflect a decreased flexibility of the middle region, impeding the formation of the bent structure. A P residue following a K residue usually has a strong negative influence on trypsin action. This also may explain the high resistance to tryptic cleavage observed for Gdm Dhb14P (tryptic cleavage site: K13–P14). However, a negative influence of a preceding L residue is not known [28], which suggests that the large increase in resistance to tryptic cleavage observed for Gdm A12L (tryptic cleavage site identical to that of gallidermin) is mainly caused by a restricted flexibility of the middle region. The K13–Dhb14 bond of gallidermin is cleaved by trypsin with a strikingly lower efficiency than bonds of K to normal protein amino acids [28]. Thus, an increased substrate affinity of the trypsin molecule and an increased flexibility of the peptide backbone within the middle region, which may lead to a greater accessibility of the tryptic cleavage site, may be responsible for the greatly reduced resistance to tryptic cleavage observed for Gdm Dhb14A and Gdm Dhb14S (tryptic cleavage sites: K13–A14 and K13–S14, respectively). This effect was strikingly less enhanced for Gdm Dhb14Dha (tryptic cleavage site: K13– Dha14). Interestingly, the tryptic sensitivity of Gdm L6V and Gdm L6G differ, although these gallidermin analogues are not altered in the vicinity of the tryptic cleavage site; this observation possibly indicates a slight overall change in conformation of the L6V and the L6G analogue. 64
3.4 Isolation and characterization of genetically engineered gallidermin Of all the gallidermin derivatives altered within the middle region, only the Dhb14Dha analogue has antimicrobial activity similar to that of gallidermin (Table 3.1). This may indicate the importance of the 2,3-didehydroamino acids for maintaining a structure required for efficient pore formation. A critical role of the polypeptide backbone flexibility for biological activity has also been suggested for the channel-forming peptides alamethicin and melittin [78]. With Gdm Dhb14A, a decrease in antimicrobial activity is only observed for three of five indicator strains (Table 3.1). Thus, one can speculate that gallidermin can exert its antimicrobial activity by different mechanisms: a Dhb/Dha-dependent and a Dhb/Dha-independent mechanism. Subtilin exerts its antimicrobial activity against Bacillus by two different mechanisms. The inhibitory effect on Bacillus spore germination is dependent on the Dha5 residue and is most likely caused by a Michael-type reaction of the didehydro residue with nucleophilic membrane sulfhydryl groups [48]. The ability to lyse vegetative Bacillus cells, however, is independent of the Dha5 residue [47]. Thus, our results [53] and the results obtained previously by mutagenesis of other type A lantibiotics strongly indicate that the molecular mechanisms by which these peptides exert their antimicrobial activity may differ and that a general function of the 2,3-didehydroamino acids for biological activity is not yet known. The various test strains varied greatly in their susceptibility to gallidermin and its analogues (Table 3.1), possibly because of differences in the membrane phospholipid composition [25, 35] or different membrane potentials [34, 68]. Both factors are crucial for the formation of voltage-dependent transmembrane pores. Furthermore, the composition of the bacterial cell wall might influence the kinetics of pore formation. Cationic lantibiotics interact with polyanions (e. g. teichoic acids) of the cell wall [11] and with membrane-bound cell wall precursor molecules [64]. Binding of lantibiotics to these components may influence the kinetics of pore formation, and may also be responsible for secondary killing mechanisms observed for Pep5 and nisin, such as activation of autolytic enzymes and inhibition of murein synthesis [64]. Recently it was indeed demonstrated that an increased negatively charged cell wall leads to an increased sensitivity to gallidermin and other positively charged antimicrobial peptides [59]. Since B. subtilis and S. aureus produce a variety of exoproteases, the strikingly lower susceptibility of these test strains (Table 3.1) might be partially due to proteolytic degradation of gallidermin and its analogues. In summary, we successfully engineered gallidermin and epidermin analogues with increased antimicrobial activities and/or resistance to proteolytic degradation. Furthermore, we obtained valuable information about the structural elements required for proper biosynthesis of epidermin and gallidermin. NMR analyses of the generated analogues and black lipid membrane studies may provide further information on the structure/activity relationship of these lantibiotics and may pave the way for target-oriented peptide engineering.
65
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.5 Function of the epidermin immunity genes epiFEG
Epidermin, gallidermin, Pep5, subtilin, and nisin are the best-studied members of type A group of lantibiotics, which are all synthesized by Gram-positive bacteria. Type A lantibiotics act by forming membrane-potential-dependent pores in the cytoplasmic membrane of bacteria [5, 68, 69]. Because of this activity, self-protection against the lantibiotic is of vital importance to the producing organism. Several genes responsible for producer self-protection have been found in the gene clusters of the respective systems: 1) nisI and spaI, which code for lipoproteins [32, 37]; 2) epiF, epiE and epiG [57] which are homologous to nisF, nisE, and nisG of the nisin system [74] and to spaF and spaG of the subtilin system [32], all of which code for ABC transporters; and 3) pepI, which does not share any sequence similarity to any known genes [63]. Analysis of the amino acid sequences suggests that the EpiFEG-type transporters and the lipoproteins NisI and SpaI are membrane-located; this has been experimentally shown for PepI [63]. The mechanism of action has not been elucidated for any of the gene products. The sequence similarity of the EpiFEG proteins to that of ABC transporters led to the proposal that the EpiFEG-type proteins act by transporting the lantibiotic out of the cytoplasmic membrane, thus keeping its concentration below a critical level and preventing pore formation. Two conceivable directions of transport have been discussed: import into the cell for inactivation by proteolytic cleavage and export into the surrounding medium. As shown by heterologous expression experiments in S. carnosus, the selfprotection (immunity) of the epidermin-producing strain S. epidermidis Tü 3298 against the pore-forming lantibiotic epidermin is mediated by an ABC transporter composed of the EpiF, EpiE, and EpiG proteins. We developed a sensitive assay based on HPLC analysis of the substrate gallidermin in cell supernatants to investigate the mechanism of the EpiFEG transporter. Our results indicate that the EpiFEG transporter works by expelling the lantibiotic from the cytoplasmic membrane into the surrounding medium with a high substrate specificity. Thus, the EpiFEG transporter functions according to the “hydrophobic vacuum cleaner” mechanism [12]. Furthermore, we showed that the gallidermin derivative L6G has an EpiE-dependent enhanced activity. These results will be covered in more detail in the following sections. To our knowledge, the EpiFEG transporter is the first of its kind to be investigated that has a pore-forming peptide as substrate. Future research on the EpiFEG transporter will focus on the question whether the EpiFEG transporter can substitute in S. epidermidis Tü 3298 for the non-functional EpiT transporter, whose task is to export completely modified pre-epidermin (with the leader peptide still present) out of the producer cell.
66
3.5 Function of the epidermin immunity genes epiFEG
3.5.1 The epiFEG genes confer immunity to the epidermin producer We characterized a DNA region located upstream of the structural gene epiA that mediates immunity and increased epidermin production. The sequence of a 2.6-kb DNA fragment revealed three ORFs (Fig. 3.2), epiF, E, and G, which form an operon [57]. In the cloning host S. carnosus, the three genes mediate an increased tolerance to epidermin, and the highest level of immunity (sevenfold) is achieved with S. carnosus carrying epiFEG and epiQ (Fig. 3.4). The promoter of the first gene epiF responds to the activator protein EpiQ and contains a palindromic sequence similar to the EpiQ-binding site of the epiA promoter, which is also activated by EpiQ. Inactivation of epiF, E, or G results in the complete loss of the immunity phenotype. An epidermin-sensitive S. epidermidis Tü 3298 mutant is complemented by a DNA fragment containing all three genes. When the epiFEG genes are cloned together with plasmid pTepi14, which contains the biosynthetic genes epiABCDQP, the production of epidermin is approximately fivefold higher. The deduced amino acid sequences of EpiF, E, and G are similar in sequence and proposed structure to the components of various ABC transporter systems. EpiF is a hydrophilic protein with conserved ATP-binding sites, while EpiE and G have six alternating hydrophobic regions and very likely constitute the integral membrane domains. When EpiF is overproduced in S. carnosus, it is at least partially associated with the cytoplasmic membrane.
Figure 3.4: The epiFEG genes confer immunity against gallidermin in the heterologous host S. carnosus. MIC values were determined with S. carnosus TM300 harboring all or only two of the epiFEG genes (in all possible combinations) on plasmid pRB473. Black bars represent MIC values of strains harboring in addition the positive regulator of the epidermin biosynthetic system, epiQ, on plasmid pTepiQ10.
67
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.5.2 Transport mechanism In most cases, it has been assumed that ABC transporters reported to be involved in resistance to non-pore-forming antibiotics have to export the antibiotic from the cytoplasm to the surrounding medium. For an antibiotic whose target is located in the cytoplasm, this mechanism seems self-evident, whereas for a pore former whose target is the cytoplasmic membrane, import into the cell as well as export to the surrounding medium are both conceivable mechanisms for removing the antibiotic from its target (Fig. 3.5).
Figure 3.5: Two alternative models for the mechanism of the EpiFEG transporter. The model shows the EpiE and EpiG proteins situated within the cytoplasmic membrane and an EpiF dimer bound to the complex at the cytoplasmic side of the membrane. EpiF harbors the ATPase binding site, as predicted by hydropathy plots and sequence alignments (Peschel and Götz 1996). Black arrows illustrate the mechanism model 1: the export of gallidermin into the surrounding medium. Every exported molecule in model 1 is likely to be able to reintegrate into the membrane, thereby causing an incessant expulsion process. Gray arrows illustrate model 2: the import of gallidermin into the cytoplasm, where one would expect proteolytic degradation of the lantibiotic.
To investigate the direction of transport mediated by the EpiFEG transporter, it was not possible to use membrane vesicle systems because the cytoplasmic orientation of the ATPase site in the EpiF part of the transporter constitutes the only means to discriminate between right-side-out and inside-out vesicles and because the substrate gallidermin is known to cause efflux of small molecules such as ATP out of the cell. This would, unfortunately, lead to an equal 68
3.5 Function of the epidermin immunity genes epiFEG distribution of ATP in the internal and external fluid. We therefore developed an assay in which the amount of the substrate gallidermin remaining in the supernatant of whole cells incubated with gallidermin was determined. In this transporter assay, a concentration of gallidermin in the supernatant of cells expressing the epiFEG genes higher than that of a control strain would suggest that the EpiFEG transporter works by export; a lower concentration in the supernatant would suggest that the transporter works by import. All experiments were performed with S. carnosus TM300 as a host for heterologous expression of all or some of the epiFEG genes. Gallidermin was quantified by HPLC detection. Our results indicate that model 1 shown in Fig. 3.5 represents the mechanism used by the EpiFEG transporter. It is not known whether the transporter recognizes the substrate in its monomeric or oligomeric form; both possibilities are illustrated in Fig. 3.5. Optimal results were observed with rather low gallidermin concentrations, and HPLC quantification was optimized and resulted in a detection limit of about 20 ng at the very specific wavelength of 266 nm (absorption of C-terminal aminovinylcysteine in gallidermin and all derivatives). At 2 mg gallidermin/ml, the extracellular gallidermin content is about fourfold higher for the epiFEG-expressing strain than for the control strain. At higher and lower concentrations of applied gallidermin, the difference is less pronounced. To exclude the possibility that the observed difference in extracellular gallidermin concentration is caused by non-specific interaction of gallidermin with one of the EpiFEG proteins and not by the functional transporter, the epiFEGexpressing strain was compared with strains harboring the genes encoding only two of the three transporter components (epiF, epiE and epiG) in all three possible combinations. The gallidermin concentration in the assay supernatant of each of the three strains showed a similar basal level, whereas in the epiFEGexpressing strain, this concentration was about fourfold higher, demonstrating that the observed effect is not caused by non-specific interaction with one of the protein components of the EpiFEG transporter (Fig. 3.6).
3.5.2.1 Energy dependence of transport Energy dependence of the EpiFEG transporter could not be directly studied using ATPase inhibitors because destruction of the membrane potential would also result in inhibition of gallidermin activity. Instead, dependence of the EpiFEG-mediated transport on the glucose concentration was determined. Glucose was the only energy source in the incubation buffer. In the absence of glucose, transport is strongly reduced. Increases in the concentration of glucose up to 1% result in increases in the transport efficiency; at concentrations of glucose higher than 1%, transport efficiency does not increase further. These results are in accordance with the expected energy dependence of the ATP-consuming EpiFEG transporter.
69
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
Figure 3.6: Comparison of the effect of the entire EpiFEG transporter with that of the transporter protein components. The transporter assay was carried out with S. carnosus strains harboring plasmid pTepiQ10, which contains the positive regulator of the epidermin system, and two (“FE”, “EG”, “FG”), none of (“473”), or all of (“FEG”) the epiF, epiE and epiG genes on plasmid pRB473. Gallidermin was used as substrate at a concentration of 2 mg/ml. The values of the determined extracellular gallidermin content from five samples for each strain were averaged.
3.5.2.2 Specificity of the transporter Eight gallidermin derivatives with single amino acid substitutions were used to determine the specificity of the EpiFEG transporter in the transporter assay. In addition to epidermin (gallidermin L6I), seven other derivatives originating from site-specific mutagenesis of the gallidermin structural gene gdmA [53] were investigated. Transporter efficacy was defined as the value obtained with the epiFEG-expressing strain divided by the value obtained with the control strain. This factor is in the range of 3 to 5 for each substance, with the exception of gallidermin Dhb14P and gallidermin L6G, where there is almost no detectable effect. No differences among the test strains were detected with nisin, which suggests that nisin is not a substrate of the EpiFEG transporter. The most prominent feature of gallidermin Dhb14P is that the flexibility of the so-called hinge region is strongly decreased; gallidermin L6G was the most hydrophilic derivative investigated (as concluded from reversed-phase HPLC retention). The three-step model for the mode of action of nisin and type A lantibiotics in general includes adhesion to the outer surface of the cell membrane mediated by electrostatic interaction between the cationic peptide and anionic charges on the surface of the cytoplasmic membrane, integration into the membrane with a DC or DpH present where hydrophobicity and flexibility are considered to be important, and formation of oligomeric pores [51, 52]. Thus, gallidermin Dhb14P and gallidermin L6G appear to be impaired in the integration into the mem70
3.5 Function of the epidermin immunity genes epiFEG brane, and this is presumably also the cause for their low bactericidal activity. Adhesion to the cytoplasmic membrane is not likely to be affected in any of the gallidermin derivatives tested because the charge of the molecule is not changed. Since specifically gallidermin Dhb14P and gallidermin L6G are by far the poorest substrates among all gallidermin derivatives investigated, we assume that integration of the substrate into the membrane is important for EpiFEG activity. This suggests that the substrate binding site of the EpiFEG transporter is within the membrane-spanning part of the protein complex, an assumption that is further supported by the interaction of gallidermin L6G with the internal membrane protein EpiE (see below). MIC values of gallidermin, epidermin, gallidermin derivatives, and nisin of S. carnosus (pRBepiFEG/pTepiQ10) and S. carnosus (pRB473/pTepiQ10) were also determined. No difference in the MIC values of nisin among the test strains was detected. Gallidermin, epidermin, and the gallidermin derivatives Dhb14S, Dhb14Dha, and A12L are more active against the control strain, whereas the activity of gallidermin Dhb14P is only slightly influenced by the presence of the epiFEG genes in the test strain. In contrast, the MICs of gallidermin derivatives L6V and Dhb14A are not influenced by the presence of the epiFEG genes in the strain; however, they are good substrates of the EpiFEG transporter, as shown in the transporter assay. While conditions in the transporter assay were selected to optimize EpiFEG transporter efficacy, this was not the case in the complex medium used for MIC determinations. Conditions in the complex medium seemed to suppress the interaction of EpiFEG with some derivatives.
3.5.2.3 Interaction of EpiE with gallidermin L6G Paradoxically, the activity of gallidermin L6G is clearly higher in the presence of the epiFEG genes in the test strain. Further experiments were designed to investigate this phenomenon. The MIC of gallidermin L6G of S. carnosus (pRBepiFE/pTepiQ10), S. carnosus (pRBepiEG/pTepiQ10), S. carnosus (pRBepiFG/ pTepiQ10), and the strains used before was determined (Table 3.2). The activity of gallidermin L6G is higher whenever the epiE gene is expressed in the strain; the expression of the other two genes (epiF and epiG) is not necessary for this effect. The activity of gallidermin L6G is also higher against S. carnosus (pTXepiE), in which the epiE gene is under the regulation of a xylose-inducible promoter, than against the control strain S. carnosus (pTX16). The integration of gallidermin L6G into the cytoplasmic membrane is usually impaired because of its comparatively low hydrophobicity, yet the activity of gallidermin L6G against strains that expressed the epiE gene was always relatively high (Table 3.2). As an explanation for this phenomenon, we propose a direct protein-protein interaction of gallidermin L6G with EpiE; this interaction would also suggest a participation of EpiE in substrate binding in general. Gallidermin L6G probably binds to the substrate binding site on EpiE and remains bound there, forming a nucleus to which other gallidermin L6G molecules adhere to form a pore. Without this interaction, gallidermin L6G would be too hydrophilic to form a pore. 71
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin Table 3.2: Activity of gallidermin L6G [53]. Strain
MIC a (mg/ml)
S. carnosus (pRB473/pTepiQ10) S. carnosus (pRBepiFEG/pTepiQ10) S. carnosus (pRBepiFE/pTepiQ10) S. carnosus (pRBepiEG/pTepiQ10) S. carnosus (pRBepiFG/pTepiQ10) S. carnosus pTX16 S. carnosus pTXepiE
1.0 0.1 0.1 0.1 1.0 1.0 0.2
a
MIC values of gallidermin L6G (Ottenwälder et al. 1995) of various strains were determined as described in the experimental procedures to investigate the gallidermin L6GEpiE interaction. Strains S. carnosus pTXepiE, in which epiE is under the control of a xylose-inducible promoter, and S. carnosus pTX16 as a control were grown in basic medium without glucose and with 0.5% xylose.
3.6 Inactivation and characterization of the epidermin leader peptidase EpiP
The sequences of lantibiotic leader peptides differ from Sec-dependent protein export signal sequences. Type A lantibiotics have a negatively charged, amphiphilic a-helix that is removed at the characteristic processing-site P–2–Q/R–1 ;X+1 (Fig. 3.7). In nearly all lantibiotic gene clusters, a serine protease is encoded, which is proposed to mediate precursor processing. These proteases are usually located in the cytoplasm. In contrast, NisP, the nisin leader peptidase, and EpiP contain a signal sequence, indicating that they act extracellularly. Therefore, the location of the processing of the various lantibiotics may differ [75]. The serine protease EpiP from S. epidermidis Tü 3298 catalyzes the extracellular processing of the epidermin precursor peptide, as shown in experiments where epiP in the xylose-regulated expression vector pCX15 in S. carnosus is
Figure 3.7: The processing site of modified pre-epidermin. The arrow indicates the processing site of the secreted EpiP protease.
72
3.6 Inactivation and characterization of the epidermin leader peptidase EpiP overexpressed. The cleavage of the unmodified EpiA precursor peptide to leader peptide and pro-epidermin by EpiP-containing culture filtrate of S. carnosus (pCX15epiP) was followed by reversed-phase chromatography and subsequent electrospray mass spectrometry [20]. The epidermin leader peptidase gene epiP was inactivated so that the final intermediates of epidermin biosynthesis could be isolated to characterize the mechanism of precursor processing. The isolated precursor peptides may also serve as natural substrates for EpiP in future experiments.
3.6.1 epiP gene replacement in S. epidermidis Tü 3298 To construct a plasmid for the replacement of epiP, 1 kb upstream and 1 kb downstream of the gene were amplified by PCR and sequenced. The two amplified fragments were ligated into pBT2, a temperature-sensitive shuttle vector [14], flanking an erythromycin resistance cassette. The resulting plasmid was introduced into S. epidermidis Tü 3298 by electroporation. A homologous recombination event deleted the entire epiP gene, leaving the epiP promoter intact to ensure expression of the regulator epiQ [56] downstream of epiP (Fig. 3.2); there is no terminator structure in the erythromycin resistance cassette that could inhibit transcription of epiQ. The antimicrobial activity of S. epidermidis Tü 3298DepiP mutant strains on agar plates containing the epidermin-sensitive M. luteus was strongly decreased. By transforming the mutant with an epiP-expressing plasmid, epidermin production could be reconstituted (Kies and Götz, unpublished).
3.6.2 Detection and isolation of epidermin precursor peptides Epidermin and epidermin precursor peptides were purified by reversed-phase chromatography of the supernatant from S. epidermidis Tü 3298DepiP grown in synthetic medium. A silver-stained nitrocellulose blot of an SDS-polyacrylamide gel with fractions from the reversed-phase chromatography showed peptides with an apparent molecular weight similar to that of mature epidermin as well as peptides with an apparent molecular weight similar to that of the N-terminally cleaved precursor peptides isolated from the S. epidermidis Tü 3298 wild type strain grown in defined medium. These preliminary results indicate that in S. epidermidis Tü 3298DepiP, the same or at least a similar processing at the Nterminus occurs as in the S. epidermidis Tü 3298 wild type strain. To test this hypothesis, the precursor peptides isolated from S. epidermidis Tü 3298DepiP were analyzed by mass spectrometry and Edman degradation. 73
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin From S. carnosus (pTepiABCDQ) grown in synthetic medium, completely modified epidermin precursors could be isolated that were N-terminally processed at different positions within the leader peptide, i. e. before amino acids 17, 19, or 20 (Fig. 3.7). Mature epidermin was also produced. The peptides were characterized by Edman degradation and mass spectrometry. The same peptide mixture was detected in S. epidermidis TÜ 3298 (wild type strain) when grown in synthetic medium, where EpiP shows a significant reduction in activity because of the low buffer capacity of this medium, which becomes rather acidic. The precursor peptides isolated are not antimicrobially active against the epidermin-sensitive M. luteus. Treatment with endoprotease ArgC, a protease that simulates the EpiP reaction in vitro, leads to antimicrobially active, mature epidermin, as shown by mass spectrometry and in the M. luteus bioassay. In conclusion, the serine protease EpiP catalyzes the last modification step in epidermin biosynthesis – the processing of the epidermin precursor peptide. Due to the inactivation of epiP in S. epidermidis Tü 3298DepiP, it was possible to detect the final intermediate of epidermin biosynthesis, i. e. the completely modified precursor peptide. The precursor had undergone proteolytic cleavage within the leader peptide. Similar results have been observed for the S. epidermidis Tü 3298 wild type strain in synthetic medium, where EpiP shows decreased activity, and for S. carnosus (pTepiABCDQ), which indicate a two-step processing of the epidermin precursor peptide.
3.7 The flavoenzyme EpiD and formation of peptidylaminoenethiolates
One of the goals of lantibiotic research is the analysis of the enzymatic mechanisms involved in the modification process from pro-epidermin to pre-epidermin. The key function is certainly associated with the EpiB and EpiC enzymes, which are apparently involved in dehydration of serine and threonine residues and in the thioether bridge formation. Homologous protein sequences that occur in all other lantibiotic genes are referred to as LanB and LanC. Several groups are studying the enzymatic mechanism of these two enzymes; however, to date, no clean in vitro reaction has been achieved. It was speculated that the enzymatic reaction is strictly oxygen sensitive and/or that unusual co-factors are involved. Epidermin and gallidermin modification involves a third enzyme, EpiD, whose enzymatic function was extensively studied by Thomas Kupke (see below). An epiD homologue has so far only been found in the mersacidin biosynthesis gene cluster [9], but not the nisin, pep5, or subtilisin gene clusters.
74
3.7 The flavoenzyme EpiD and formation of peptidyl-aminoenethiolates
3.7.1 Preparation of precursor peptide EpiA To investigate the enzymes involved in epidermin biosynthesis, it is necessary to produce sufficient amounts of pre-epidermin (EpiA) as a substrate and to design methods to detect EpiA. EpiA was therefore expressed in Escherichia coli using a malE-epiA fusion. The malE gene encodes the maltose-binding protein MBP. The fusion protein MBP-EpiA was expressed from pIH902-epiA and purified in one step by amylose affinity chromatography. EpiA was cleaved from MBP-EpiA by factor Xa, and the identity of purified EpiA was confirmed by ES-MS and amino acid sequencing. Upon prolonged incubation, factor Xa not only cleaves EpiA from the fusion protein, but also less efficiently cleaves EpiA internally between R–1 and I+1. The internal factor Xa cleavage site of EpiA was masked by altering the sequence -A–4-E-P-R–1- to -A–4-E-P-Q–1- by site-directed mutagenesis. Since the mutant peptide EpiA R–1Q) has properties different than EpiA (the molecular mass is 28 Da and the isoelectric point is approximately 1.5 pH units lower), it is used as a control peptide in incubation experiments. Anti-EpiA antisera were raised to detect EpiA [44].
3.7.2 Construction and purification of MBP-EpiD fusion proteins According to the nucleotide sequence, epiD encodes a 181-amino acid protein. In order to obtain more information on the function and biochemical properties of EpiD, epiD was expressed in E. coli using the maltose-binding protein (MBP) fusion system. The soluble 67-kDa MBP-EpiD fusion protein was purified in one step by amylose affinity chromatography and was used to raise polyclonal antibodies directed against EpiD. The MBP-EpiD fusion protein was yellow and identified as a flavoprotein by its absorption spectrum with maxima at 277, 378, and 449 nm (compare Fig. 3.8). The coenzyme is very tightly, but not covalently attached. It could only be removed by TCA extraction of the fusion protein and was identified by thin-layer chromatography (TLC) as FMN. S. epidermidis Tü 3298/EMS11 [4] carries a mutation within epiD, designated epiD*, which was PCR-amplified using purified pTü32 from the mutant as a template. This DNA fragment was inserted in the StuI site of the pIH902 polylinker. Plasmids were isolated from four E. coli TB1 (pIH902-epiD*) clones, and the entire epiD* region was sequenced. The epiD* start codon immediately follows the factor Xa cleavage sequence and all four isolated plasmids have a point mutation in codon 93, substituting the wild type GGT (Gly) with GAT (Asp) in epiD*. This G–A transition concurs with the use of EMS as the mutagenizing agent. MBP-EpiD* is not yellow and does not exhibit the characteristic absorption maxima of MBP-EpiD when similar concentrations of fusion proteins are used; it was necessary to dilute MBP-EpiD because of the low amount of soluble MBP75
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
Figure 3.8:
Absorption spectrum of purified EpiD.
EpiD*. MBP-EpiD* does not bind FMN. Since MBP-EpiD* differs from MBPEpiD only in the substitution of Gly93 by Asp, Gly93 may be important for FMN binding [45].
3.7.3 Characterization of purified EpiD EpiD was expressed using the T7 RNA polymerase promoter system in E. coli and was purified to homogeneity in three column chromatography steps: DEAE-Sepharose Fast Flow, Mono-Q, and Phenylsuperose chromatography. The first 18 N-terminal amino acids of purified EpiD were determined by Edman degradation. The amino acid sequence correlates exactly with the amino acids (Met-Tyr-Gly-Lys-…) deduced from the epiD sequence [70]. Purified EpiD is yellow. The absorption spectrum exhibits maxima at 274, 382 and 453 nm, which are characteristic for flavoproteins in the oxidized state (Fig. 3.8). The liberated flavin component was analyzed by TLC and ES-MS. The mass of the flavin coenzyme was determined to be 455.5 Da (457 Da in a second experiment), which closely agrees with the theoretical value of 456.3 Da and confirms that the flavin component is FMN and not FAD (theoretical mass 785.6 Da). The average molecular mass of EpiD was determined to be 20,827 +/– 5 Da, which is in close agreement with the theoretical value of 20,825 Da calculated for the amino acid sequence derived from the nucleotide sequence, indicating that EpiD is not covalently modified.
76
3.7 The flavoenzyme EpiD and formation of peptidyl-aminoenethiolates
3.7.4 Model of oxidative decarboxylation of peptides Since flavin coenzymes are normally involved in oxidation-reduction reactions, EpiD belongs to the class of oxidoreductases. There is only one obvious reaction involved in epidermin biosynthesis that requires an oxidoreductase activity: The synthesis of S-((Z)-2-aminovinyl)-D-cysteine [1] includes the removal of two reducing equivalents from a -C–C- group to form a -C=C- group. Regarding the proposed order of the modification reactions [72], it was assumed that the four lanthionine rings are formed first, followed by the oxidation of the C-terminal lanthionine by EpiD and the subsequent decarboxylation reaction. In the course of the oxidation, EpiD-FMN becomes reduced. According to this first published hypothesis, EpiD catalyzes the final step in the modification of pre-epidermin, and its reaction product is then processed by the leader peptidase, forming mature epidermin [45]. Thioether formation is only possible if dehydrated serine and threonine residues are present, but in principle, oxidative decarboxylation could be independent of the other modification reactions. Therefore, the following alternative model of the C-terminal oxidative decarboxylation of peptides has been suggested (Fig. 3.9): in the first reaction, EpiD catalyzes the removal of two reducing equivalents from the C-terminal cysteine residue of unmodified EpiA. A double bond is formed, and FMN is reduced to FMNH2. The C-terminal carboxyl group is then removed, catalyzed either by EpiD or occurring spontaneously [42]. The occurrence of a reaction between EpiD and the unmodified precursor peptide EpiA was investigated.
O
EpiD-FMN
HN CH C OH
CO2
EpiD-FMNH2
O HN C C OH
HN C
CH2
CH
CH
SH
SH
SH
H
Figure 3.9: Model of the C-terminal oxidative decarboxylation of unmodified precursor peptide EpiA and the role of flavoprotein EpiD. The C-terminal cysteine residue of the precursor peptide EpiA is shown. After oxidation and decarboxylation, the peptidyl-aminoenethiols are formed.
3.7.5 Interaction between EpiD and EpiA EpiA or EpiAR–1Q was coupled to an NHS-activated HiTrap column in order to purify the enzymes involved in epidermin biosynthesis by affinity chromatography. Binding studies were carried out with purified EpiD or extracts of induced E. coli K38 (pGP1–2, pT7–5epiD) cells. In both cases, the migration of EpiD is re77
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin tarded only under reducing conditions. The applied EpiD focuses as a sharp yellow band on the EpiA-HiTrap column. Other proteins of the cell extract are not retarded, which indicates the specificity of the interaction. Even under reducing conditions, the interaction is weak, and EpiD already elutes with the loading buffer containing no salt. In a control experiment, a column without any coupled peptide was used, and as expected, the migration of EpiD was not retarded [40, 43].
3.7.6 Reaction of EpiD with EpiA and pro-epidermin Under reducing conditions, purified EpiD reacts with unmodified precursor peptide EpiA. Several products were identified after separation of the reaction mixture by reversed-phase chromatography and ES-MS analysis. The product formed depends on the incubation time. The first product formed, product 3 (the products were designated according to increasing hydrophobicity), is more hydrophobic than the unmodified peptide EpiA and has a molecular mass of 5,579–5,585 Da (values obtained from several experiments), i. e. 45–48 Da less than EpiA. In order to determine the mass difference between EpiA and product 3 with greater accuracy, the peptides were measured together, giving a mass difference of 46.5 Da. Furthermore, product 3 has an absorbance at 260 and 280 nm in 0.1% TFA/H2O/acetonitrile, higher than that of EpiA [42]. This enzyme reaction was also detected using crude cell extracts of S. carnosus (pTepi14) and EpiA [40]. This clone has been used for heterologous expression of epidermin [3, 70]. With longer incubation times, two additional, less hydrophobic peptides with molecular masses of 5,524–5,528 Da (product 1; 102–105 Da less than EpiA) and 5,579–5,585 Da (product 2; 45–48 Da less than EpiA) were identified. Product 3 (oxidatively decarboxylated peptide) is unstable and is non-enzymatically converted to products 1 and 2. Product 2 has the same mass as product 3, but its absorbance at 260 and 280 nm is comparable to that of unmodified peptide EpiA. Based on its molecular mass, product 1 is either the peptide EpiA(M1-C51) (lacking the last cysteine residue; 103.2 Da less) or the peptide EpiA(M1-C51)-NH2 (104.1 Da less) [42]. All known lantibiotics are synthesized as pre-peptides with an N-terminal leader peptide. It has been proposed that all the processing signals are in the leader region of the pre-peptides [15]. Thus, the pre-peptides would be recognized as substrates by binding of the leader peptides to the enzymes involved in posttranslational modifications. To test this hypothesis, unmodified pro-epidermin (obtained by factor Xa cleavage of EpiA) was used as a substrate for the flavoprotein EpiD. Even with this substrate, the reaction occurs. Hence, a leader peptide is not required for substrate recognition by EpiD. The primary reaction product was analyzed by Edman degradation to detect unmodified amino acid residues. A pattern of pmol amounts of the PTH-amino acids Ile+1 to Tyr+20 similar to that of unmodified pro-epidermin was obtained, which indicates that amino acids 1 to 78
3.7 The flavoenzyme EpiD and formation of peptidyl-aminoenethiolates 20 are unmodified and that one of the last two cysteine residues (Cys+21, Cys+22) is modified [42]. Recently, a more detailed analysis of the reaction product has become possible using tandem mass spectrometry and NMR methods.
3.7.7 In vivo reaction of affinity-tag-labeled epidermin precursor peptide with the flavoenzyme EpiD The genes encoding the His-tag-labeled epidermin precursor peptide EpiA and the flavoenzyme EpiD or the mutant protein EpiD-G93D, which lacks the coenzyme, were co-expressed and the proteins were synthesized in vivo in E. coli. Only in the presence of EpiD was the precursor peptide converted to a reaction product with a decrease in mass of 44–46 Da. This result confirms the in vitro experiments carried out with purified EpiA and purified EpiD from S. epidermidis. In the presence of EpiD, the amount of purified (modified) peptide EpiA was several-fold higher than in the presence of EpiD-G93D, indicating that the stabilization of EpiA against proteolysis is due to an interaction with EpiD or to the presence of the C-terminal modification [39].
3.7.8 Determination of the substrate specificity of EpiD using mutant precursor peptides and chemically synthesized peptides The precursor peptide EpiAC52S, which is altered by gene mutation and contains a C-terminal serine, was used to test the production of enols by reaction with EpiD [31]. No reaction occurred, which indicates the necessity for a Cterminal cysteine residue. However, not all peptides with a C-terminal cysteine residue, e. g. the peptide EpiA M1-C51, were a substrate of EpiD. EpiAS49A was a substrate for oxidative decarboxylation, providing the first hint that EpiD has no absolute substrate specificity. The leader peptide of the precursor peptide EpiA has no significant influence on the reaction with EpiD. Synthetic peptides with increasingly larger deletions of the amino-terminus were used to determine the minimal size of the substrate. A weak reaction was even observed for the tetrapeptide SYCC. The heptapeptide SFNSYCC was used to study the substrate specificity of EpiD further (Table 3.3). The serine and cysteine residues of this heptapeptide form the Cterminal bicyclic structure of mature epidermin, and it could not be excluded that amino acid exchanges in this peptide have an influence on the reaction with EpiD [31]. Interestingly, the penultimate cysteine residue can be exchanged with at least a serine or threonine residue. Modification or exchange of the C-terminal cysteine residue results in loss of the reaction with EpiD. The peptides SFNSYCS, SFNSYCM, and SFNSYC were no substrates of EpiD, which confirms the results 79
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin Table 3.3: Determination of the substrate specificity of the flavoprotein EpiD [43]. Peptide
Reaction with EpiD
EpiA, EpiAR-1Q, K-EpiA EpiAS+19A EpiAC+22S EpiAdesC+22
+ + – –
proepidermin synthetic proepidermin AKTGSFNSYCC SFNSYCC FNSYCC NSYCC SYCC YCC
+ + + + + + + –(?)
AFNSYCC SANSYCC SFASYCC SFNAYCC SFNSACC SFNSYGC
+ + + + – –
SFNSYCS SFNSYC SFNSYCC-NH2 SFNSYCC(Et) SFNSYCHCy SFNSYCM
– – – – – –
SFNSYSC SFNSYTC
+ +
SFNSFCC SFNSWCC
+ +
SFNYYCC SYNSYCC SWNSYCC
+ + +
TLTSECIC
–
obtained with the mutant precursor peptides. The analogue with a C-terminal ethyl-thioether structure [SFNSYCC(Et)], the amide SFNSYCC-NH2, and the peptide SFNSYCHcy (C-terminal homocysteine residue) were not substrates of EpiD. Mersacidin is another lantibiotic containing a C-terminal -NH–CH=CH–Sgroup [61], probably formed by modification of a C-terminal cysteine residue. The C-terminal peptide TLTSECIC of the mersacidin precursor peptide is not a substrate of EpiD [43]. 80
3.7 The flavoenzyme EpiD and formation of peptidyl-aminoenethiolates
3.7.9 Tandem mass spectra of modified SFNSYTC and SFNSYSC The sequence and structure of peptides can be analyzed with tandem mass spectrometry (collision-induced dissociation). In collision-induced dissociation experiments, peptides are preferentially cleaved at the peptide bonds between NH and CO, resulting in amino-terminal Bn fragments and C-terminal Yn fragments [7, 8]. To verify that the C-terminal cysteine residue of the reaction products is modified, product ion scans of SFNSYTC and SFNSYSC and the corresponding reaction products were recorded. These peptides were used to exclude intramolecular disulfide bridge formation in the peptide SFNSYCC. The B1–B6 fragments of the peptide and its reaction product are identical, proving that the C-terminal amino acid residue is modified. SFNSYTC and SFNSYSC and their reaction products differ from each other by 14 mass units in their B6 fragments, showing the S/T exchange. The mass difference between the modified peptide and its B6 fragment is 75 mass units; the mass difference between the unmodified peptide and its B6 fragment is 121 mass units. It was, therefore, possible to identify the reaction products by neutral loss mass spectrometry [43].
3.7.10 The application of neutral loss mass spectrometry to determine the substrate specificity of EpiD Tandem mass spectrometry methods have already been used to determine the composition of synthetic multicomponent peptide mixtures. For example, O-tertbutylated by-products of peptide libraries were identified by neutral loss scans [49]. In the constant neutral loss scan, the first and second analyzer of the mass spectrometer are scanned together such that there is a constant mass difference between the ions transmitted by the two analyzers. Under these conditions, only ions that lose a neutral fragment with a mass corresponding to the chosen mass difference will be detected. Heptapeptide sublibraries with one variable amino acid residue at positions 1 to 7 of the peptide substrate S1FNSYCC7 were synthesized and incubated with EpiD. Peptides were identified by their masses using product ion scans and neutral loss scans, and by comparison of the mass spectra obtained after various incubation times. The heptapeptides with a single amino acid substitution at positions 1 to 4 are substrates of EpiD. Not all amino acid residue substitutions at positions 5 to 7 of the peptide led to active substrates; therefore, the last three amino acids determine the substrate specificity for EpiD. For the sublibraries SFNSX5CC and SFNSYX6C, the reaction products were identified by neutral loss mass spectrometry of the peptide mixtures. The tyrosine residue at position 5 can be replaced by the hydrophobic amino acid residues V, I/L, (M), F, and W; it is not possible to differentiate between Ile and Leu by ES-MS. The cysteine residue at position 6 can be replaced by A, S, V, (I/L), and T. Pep81
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin tides containing amino acid residues with acidic or basic side chains at position 5 or 6 are not substrates of EpiD. In position 7 of the peptide substrates, only cysteine is accepted [43].
3.7.11 Structure analysis of the reaction products by two-dimensional NMR methods Due to the complexity of the NMR spectra, the introduction of a 13C isotope in the C-terminal cysteine residues was necessary to obtain the interesting signals using heteronuclear correlation experiments. Hence, the peptide KKSFNSYTC was synthesized by solid-phase synthesis using cysteine labeled with 13C at the b-carbon atom. The two lysine residues at the N-terminus were introduced to increase the solubility of the substrate in water. Two crosspeaks occur in the HSQC spectrum of the educt KKSFNSYTC. The first signal is due to the 13C-labeled b-carbon, while the second is due to the methyl groups in Tris[tris(hydroxymethyl)aminomethane]. The NMR data, the UV-VIS spectra, the MS-MS experiments, the observed reaction product with Ellman’s reagent, and the mass difference between educt and product of 46 Da indicate that the product contains a C-terminal thioenol [30].
3.7.12 The pKa value of the enethiol side chain of the reaction products is lower than that of the thiol side chain of peptides The UV-VIS spectra of the reaction products of EpiD are pH-dependent, indicating that the enethiol side chain is converted to an enethiolate anion. The pKa value of the enethiol group was determined to be 6.0 and is substantially lower than the pKa value of the thiol side chain of cysteine residues. This increased acid strength of the enethiol side chain compared to that of the thiol group is attributed to the resonance stabilization of the negative charge of the anion [39].
3.7.13 Overview of the function of EpiD In conclusion, the formation of the lantibiotic epidermin from the precursor peptide EpiA includes the oxidative decarboxylation of the C-terminal cysteine residue of EpiA to a (Z)-enethiol catalyzed by the FMN-containing enzyme EpiD. This oxidative decarboxylation reaction was the first analyzed posttranslational 82
3.8 Incorporation of d-alanine into S. aureus teichoic acids confers resistance modification reaction involved in lantibiotic biosynthesis. Two reducing equivalents from the C-terminal cysteine residue of EpiA are removed, a double bond is formed, and the coenzyme FMN is reduced to FMNH2. The decarboxylation occurs spontaneously or is catalyzed by EpiD. The in vivo conversion of (His)6tag-labeled precursor peptide EpiA by EpiD was demonstrated. The (Z)-enethiol derivative is the intermediate in the formation of the C-terminal S-((Z)-2-aminovinyl)-D-cysteine residue of epidermin. The enethiol structure has been confirmed by UV-VIS spectroscopy, mass spectrometry, tandem mass spectrometry, two-dimensional NMR spectroscopy, and conversion of the enethiol to a mixed disulfide with 5,5'-dithiobis(2-nitrobenzoic acid) (Ellman’s reagent). At physiological pH, the dominant form of the reaction product is the enethiolate anion (“peptidyl-aminoenethiolates”), and the pKa of the enethiol group is 6.0. EpiD has a low substrate specificity, and most of the peptides with the sequence (V/I/ L/(M)/F/Y/W)-(A/S/V/T/C/(I/L))-C at the carboxy terminus are substrates of EpiD, as elucidated by analysis of the reaction of EpiD with single peptides and peptide libraries. Amino acid residues of EpiD involved in FMN binding or substrate binding, or important for the catalytic action are currently identified by analyzing analogues generated by site-directed mutagenesis. In addition, the structure of EpiD has been determined by X-ray crystallography (Kupke et al. unpublished). The overall aim is the detailed molecular analysis of the flavoenzyme EpiD and related enzymes, and a more detailed investigation of the role of EpiD in epidermin biosynthesis. It is still not known how reduced EpiD is re-oxidized in vivo and whether EpiD forms a complex with the enzymes EpiB and EpiC. The chemistry of the peptidyl-aminoenethiolate reaction products has to be studied in order to develop these peptides as enzyme inhibitors.
3.8 Incorporation of d-alanine into S. aureus teichoic acids confers resistance to defensins, protegrins, and other antimicrobial peptides
Why some staphylococcal strains are more tolerant to gallidermin than other strains has always been of interest. In order to answer this question, Andreas Peschel [59] isolated S. aureus and S. xylosus transposon-insertion mutants that were hypersensitive to gallidermin, and found that genes hit by the transposon are involved in teichoic acid modification.
83
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.8.1 Identification and sequence analysis of the S. aureus and S. xylosus dlt operon S. aureus Sa113 and the coagulase-negative S. xylosus C2 a have high innate tolerances towards several antimicrobial peptides. To investigate the resistance mechanism, S. xylosus C2 a was mutagenized with Tn917, and the resulting transposon-insertion mutants were analyzed for reduced growth on agar plates containing the antimicrobial peptide gallidermin. The nucleotide sequence upstream and downstream of the transposon from seven clones whose growth is specifically reduced in the presence of gallidermin was determined. In all seven mutants, the transposon had integrated into the same determinant of 4.6 kb, which encodes four open reading frames arranged in an operon-like structure, followed by a typical terminator (Fig. 3.10). The open reading frames showed sequence similarity to the Lactobacillus casei and B. subtilis dltABCD operons, which are responsible for esterification of teichoic acids with d-alanine [16, 55]. The growth rates in the absence of gallidermin and the microscopic appearance of the mutants were indistinguishable from those of the wild type strain.
Figure 3.10: cocci.
Genetic organization and proposed function of the dlt operon in staphylo-
3.8.2 Disruption of the S. aureus dlt operon and analysis of the d-alanine content in lipoteichoic acids (LTA) and wall teichoic acids (WTA) The dltA gene of S. aureus Sa113 was replaced by a spectinomycin resistance gene (spc) by homologous recombination, thereby producing the galliderminsensitive strain AG1. WTA and LTA of S. aureus and S. xylosus wild type strains and mutants were isolated, and the molar ratios of d-alanine to phosphorus were determined (Table 3.1). In the wild type strains, 75% (S. aureus) and 95% (S. xylosus) of the alditol phosphate residues in LTA are esterified with d-ala84
3.8 Incorporation of d-alanine into S. aureus teichoic acids confers resistance nine, while only 51% (S. aureus) and 15% (S. xylosus) are esterified in WTA. In the dlt mutants, no d-alanine is detected in LTA or WTA, indicating that the pathway for d-alanine incorporation is inactivated by the spectinomycin resistance gene and transposon insertions. When the mutant strains are complemented with plasmid pRBdlt1, which carries the dlt operon, normal or slightly increased amounts of d-alanine are found in LTA and WTA. Transformation of the wild type strains with pRBdlt1 results in an increase of d-alanine in LTA and WTA by 5–18% [59].
3.8.3 Sensitivity towards antimicrobial peptides The minimal inhibitory concentrations of gallidermin and of several other membrane-damaging antimicrobial peptides were determined for the S. aureus and S. xylosus wild type and mutant strains. The mutants are sensitive to a variety of antimicrobial peptides that bear a positive net charge [59]. The sensitivity of the S. aureus mutant to defensin from human neutrophils and to protegrins 3 and 5 from porcine leukocytes is at least 10- to 23-fold higher. Factors of 7–12 were determined with tachyplesin 1 and 3 from hemocytes of the horseshoe crab and to a variant of magainin II from the skin of the clawed frog. The tolerance towards the lanthionine-containing bacterial peptides gallidermin from S. gallinarum and nisin from L. lactis is 8–50-fold lower; very similar results were obtained with the S. xylosus strains. The increased sensitivity of dlt mutants seems to be restricted to cationic peptides since no considerable differences are observed in the inhibitory concentrations of the neutral peptide gramicidin D from Bacillus brevis. Furthermore, the mutants are not sensitive to cationic polylysine, indicating that cationic properties are not sufficient for activity of a peptide against staphylococci lacking d-alanine esters in their teichoic acids [59]. The results indicate that the general surface charge of staphylococci can be severely influenced by the degree of esterification with d-alanine. Loss of dalanine esterification is not deleterious to the bacterium, but has other pleiotropic effects. One such effect is that mutants become hypersensitive to a large variety of cationic peptides. Other effects are currently being studied. One can speculate whether the D-alanination of teichoic acids is a means whereby S. aureus overcomes the human defensin attack to which they are confronted during the first defense regimen.
85
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
3.9 Conclusions
The collaborative research centre 323 allowed the lantibiotic research to boom, particularly in Tübingen. The fermentation group (H. Zähner and P. Fiedler), the organic chemistry group (G. Jung) and microbial genetics group (F. Götz) complemented each other ideally. The fermentation group optimized the fermentation process of the producing strains and provided sufficient amounts of the compounds to the chemists for structure determination and mass analysis, and the molecular biologists studied the biosynthetic genes and the functions of the corresponding enzymes together with the organic chemists. This concerted action was the secret of success. Quite a few of the Tübingen achievements can be regarded as milestones in lantibiotic research, such as: . Determination of the chemical structure of epidermin (Allgaier et al. 1986) . First finding of an antibiotic substance (later named gallidermin) in Staphylococcus gallinarum (F. Götz 1984) . First identification of the structural gene (epiA) encoding the lantibiotic precursor for epidermin; and the first proof that epidermin, and as we now know, all lantibiotics, are ribosomally synthesized (Schnell et al. 1989) . Determination of the NMR structure of gallidermin (Freund et al. 1991) . First cloning and nucleotide sequence of lantibiotic (epidermin) biosynthesis genes: epiB, C, D, P, and Q (Schnell et al. 1992, Augustin et al. 1992) . First characterization of a novel flavin-mononucleotide-dependent decarboxylase encoded by the epiD gene (Kupke et al. 1992) . Regulation of epidermin biosynthesis by the activator EpiQ (Peschel et al. 1993) . Isolation and characterization of genetically engineered gallidermin and epidermin analogues (Ottenwälder et al. 1995) . Further optimization of gallidermin production leading to a productivity of 300 mg/liter culture supernatant (Kempf et al. 1997) . Secretion mechanism of the lantibiotics epidermin and gallidermin (Peschel et al. 1997) . Unique producer self-protection against the lantibiotic epidermin by the ABC transporter EpiFEG (Peschel et al. 1996; Otto et al. 1998) . Nearly complete X-ray structure of the lantibiotic biosynthesis enzyme, the flavoprotein EpiD (Kupke et al., in progress) In Fig. 3.11 an overview of the biosynthetic pathway of epidermin (gallidermin) is presented. There are still many open questions that deserve attention. For example, the underlying enzyme mechanisms for the key reaction, the lanthionine formation, is still unsolved and it is not known which cofactors are involved. The environmental conditions under which the activator EpiQ is functional are unknown; there are no indications for a two-component system like that of nisin or subtilisin. Finally, the question remains why Gram-positive bac86
3.9 Conclusions
Figure 3.11: Pathway of epidermin biosynthesis.
teria produce such lantibiotics – what is their real function? It is unlikely that the antibiotic activity is the only function since it only becomes obvious when larger amounts of the lantibiotic are produced. Many isolates produce only minute amounts, which might, however, be sufficient for communication with other bacteria. Most of the lantibiotics interact with the cytoplasmic membrane; some, such as mersacidin and gallidermin, bind to the cell wall precursor lipid I and, as recently shown, interact with the teichoic acids to an extent that depends on the degree of negative charges. Perhaps they represent a measure for the fine tuning of cell wall structures. Since genes nearly identical to the epidermin biosynthesis genes are also present in S. aureus, one can ask whether they have an influence on infection and survival in the host. Where the lantibiotic biosynthesis genes originate is not known. The nucleotide sequences of most lantibiotic genes suggest that the genes are either plasmid encoded (e. g. epidermin) or part of a transposon (e. g. nisin). They might be relicts of bacteriophages that used lantibiotics for a better penetration of their genome into the target cells; in this respect, a pore-forming activity could be helpful. Whether any of the lantibiotics were once used as antibiotics or not, the study of these unique peptides sheds light on various areas of microbiology.
87
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin
References
1. Allgaier, H., Jung, G., Werner, R. G., Schneider, U., and Zähner, H. (1985) Elucidation of the structure of epidermin, a ribosomally synthesized, tetracyclic heterodetic polypeptide antibiotic. Angew. Chem. Int. Ed. Engl. 24, 1051–1053. 2. Allgaier, H., Jung, G., Werner, R. G., Schneider, U., and Zähner, H. (1986) Epidermin: sequencing of a heterodet tetracyclic 21-peptide amide antibiotic. Eur. J. Biochem. 160, 9–22. 3. Augustin, J. (1991) Dissertation. Eberhard-Karls-Universität Tübingen. 4. Augustin, J., Rosenstein, R.,Wieland, B., Schneider, U., Schnell, N., Engelke, G., Entian, K. D., and Götz, F. (1992) Genetic analysis of epidermin biosynthetic genes and epidermin-negative mutants of Staphylococcus epidermidis. Eur. J. Biochem. 204, 1149–1154. 5. Benz, R., Jung, G., and Sahl, H.-G. (1991) Mechanism of channel formation by lantibiotics in black lipid membranes. In: Jung, G. and Sahl, H. G. (eds.), Nisin and novel lantibiotics. Leiden: Escom, pp. 359–372. 6. Berridge, N., Newton, N. J., and Abraham, E. P. (1952) Purification and nature of the antibiotic nisin. Biochem. J. 52, 529–535. 7. Biemann, H. P. and Erikson, R. L. (1990) Abnormal protein kinase C down regulation and reduced substrate levels in non-phorbol ester-responsive 3T3-TNR9 cells. Mol. Cell Biol. 10, 2122–2132. 8. Biemann, K. (1992) Mass spectrometry of peptides and proteins. Annu. Rev. Biochem. 61, 977–1010. 9. Bierbaum, G., Brötz, H., Koller, K. P., and Sahl, H. G. (1995) Cloning, sequencing and production of the lantibiotic mersacidin. FEMS Microbiol. Lett. 127, 121–126. 10. Bierbaum, G., Götz, F., Peschel, A., Kupke, T., van de Kamp, M., and Sahl, H.-G. (1996) The biosynthesis of the lantibiotics epidermin, gallidermin, pep5 and epilancin K7. Antonie Van Leeuwenhoek. 69, 119–127. 11. Bierbaum, G. and Sahl, H. G. (1987) Autolytic system of Staphylococcus simulans 22: influence of cationic peptides on activity of N-acetylmuramoyl-L-alanine amidase. J. Bacteriol. 169, 5452–5458. 12. Bolhuis, H., van Veen, H. W., Poolman, B., Driessen, A. J., and Konings, W. N. (1997) Mechanisms of multidrug transporters. FEMS Microbiol. Rev. 21, 55–84. 13. Brötz, H., Josten, M., Wiedemann, I., Schneider, U., Götz, F., Bierbaum, G., and Sahl, H. G. (1998) Role of lipid-bound peptidoglycan precursors in the formation of pores by nisin, epidermin and other lantibiotics. Mol. Microbiol. 30, 317–327. 14. Brückner, R. (1997) Gene replacement in Staphylococcus carnosus and Staphylococcus xylosus. FEMS Microbiol. Lett. 151, 1–8. 15. Buchman, G. W., Banerjee, S., and Hansen, J. N. (1988) Structure, expression and evolution of a gene encoding the precursor of nisin, a small protein antibiotic. J. Biol. Chem. 263, 16260–16266. 16. Debabov, D. V., Heaton, M. P., Zhang, Q., Stewart, K. D., Lambalot, R. H., and Neuhaus, F. C. (1996) The d-alanyl carrier protein in Lactobacillus casei: cloning, sequencing and expression of dltC. J. Bacteriol. 178, 2869–3876. 17. Devriese, L. A., Poutrel, B., Kilpper-Bälz, R., and Schleifer, K. H. (1983) Staphylococcus gallinarum and Staphylococcus caprea, two new species from animals. Int. J. Syst. Bact. 33, 480–486. 18. Freund, S., Jung, G., Gutbrod, O., Folkers, G., and Gibbons, W. A. (1991) The threedimensional solution structure of gallidermin determined by NMR-based molecular graphics. In: Jung, G. and Sahl, H.-G. (eds.), Nisin and novel lantibiotics. Leiden: Escom, pp. 91–102.
88
References 19. Freund, S., Jung, G., Gutbrod, O., Folkers, G., Gibbons, W. A., Allgaier, H., and Werner, R. (1991) The solution structure of the lantibiotic gallidermin. Biopolymers 31, 803–811. 20. Geißler, S., Götz, F., and Kupke, T. (1996) Serine protease EpiP from Staphylococcus epidermidis catalyzes the processing of the epidermin precursor peptide. J. Bacteriol. 178, 284–288. 21. Gross, E., Kiltz, H., and Nebelin, E. (1973) Subtilin. VI. Die Struktur des Subtilins. Hoppe-Seyler’s Z. Physiol. Chem. 354, 810–812. 22. Gross, E. and Morell. J. L. (1971) The structure of nisin. J. Am. Chem. Soc. 93, 4634– 4635. 23. Hurst, A. (1966) Biosynthesis of the antibiotic nisin by Streptococcus lactis organisms. J. Gen. Microbiol. 44, 209–220. 24. Ingram, L. C. (1969) Synthesis of the antibiotic nisin: formation of lanthionine and bmethyl-lanthionine. Biochim. Biophys. Acta. 184, 216–219. 25. Jack, R., Benz, R., Tagg, J., and Sahl, H. G. (1994) The mode of action of SA-FF22, a lantibiotic isolated from Streptococcus pyogenes strain FF22. Eur. J. Biochem. 219, 699–705. 26. Jung, G. (1991) Lantibiotics – ribosomally synthesized biologically active polypeptides containing sulfide bridges and a,b-didehydroamino acids. Angew. Chem. Int. Ed. Engl. 30, 1051–1068. 27. Katz, E. and Demain, A. L. (1977) The peptide antibiotics of Bacillus: chemistry, biogenesis, and possible functions. Bacteriol. Rev. 41, 449–474. 28. Kellner, R., Jung, G., Hörner, T., Zähner, H., Schnell, N., Entian, K. D., and Götz, F. (1988) Gallidermin: a new lanthionine-containing polypeptide antibiotic. Eur. J. Biochem. 177, 53–59. 29. Kempf, M., Theobald, U., and Fiedler, H.-P. (1999) Economic improvement of the fermentative production of gallidermin by Staphylococcus gallinarum. Biotechnol. Lett. 21, 663–667. 30. Kempter, C., Kupke, T., Kaiser, D., Metzger, J. W., and Jung, G. (1996) Thioenols from peptidyl-cysteines: oxidative decarboxylation of a 13C labeled substrate. Angew. Chem. Int. Ed. Engl. 35, 2104–2107. 31. Kleerebezem, M., Quadri, L. E. N., Kuipers, O. P., and de Vos, W. M. (1997) Quorum sensing by peptide pheromones and two-component signal-transducing systems in Gram-positive bacteria. Mol. Microbiol. 24, 895–904. 32. Klein, C. and Entian, K. D. (1994) Genes involved in self-protection against the lantibiotic subtilin produced by Bacillus subtilis ATCC 6633. Appl. Environ. Microbiol. 60, 2793–2801. 33. Kleinkauf, H. and von Döhren, H. (1990) Nonribosomal biosynthesis of peptide antibiotics. Eur. J. Biochem. 192, 1–15. 34. Kordel, M. and Sahl, H. G. (1986) Susceptibility of bacterial, eukaryotic and artificial membranes to the disruptive action of the cationic peptide Pep5 and nisin. FEMS Microbiol. Lett. 34, 139–144. 35. Kordel, M., Schuller, F., and Sahl, H. G. (1989) Interaction of the pore forming-peptide antibiotics Pep 5, nisin and subtilin with non-energized liposomes. FEBS Lett. 244, 99–102. 36. Kuipers, O. P., Beerthuyzen, M. M., de Ruyter, P. G. G. A., Luesink, E. J., and de Vos, W. M. (1995) Autoregulation of nisin biosynthesis in Lactococcus lactis by signal transduction. J. Biol. Chem. 270, 27299–27304. 37. Kuipers, O. P., Beerthuyzen, M. M., Siezen, R. J., and de Vos, W. M. (1993) Characterization of the nisin gene cluster nisABTCIPR of Lactococcus lactis. Eur. J. Biochem. 216, 281–291. 38. Kuipers, O. P., Rollema, H. S., Yap, W. M., Boot, H. J., Siezen, R. J., and de Vos, W. M. (1992) Engineering dehydrated amino acid residues in the antimicrobial peptide nisin. J. Biol. Chem. 267, 24340–2436.
89
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin 39. Kupke, T. and Götz, F. (1997) The enethiolate anion reaction products of EpiD: pKa value of the enethiol side chain is lower than that of the thiol side chain of peptides. J. Biol. Chem. 272, 4759–4762. 40. Kupke, T. and Götz, F. (1996) Expression, purification, and characterization of EpiC, an enzyme involved in the biosynthesis of the lantibiotic epidermin, and sequence analysis of Staphylococcus epidermidis epiC mutants. J. Bacteriol. 178, 1335–1340. 41. Kupke, T. and Götz, F. (1996) Post-translational modifications of lantibiotics. Antonie van Leeuwenhoek. 69, 139–150. 42. Kupke, T., Kempter, C., Gnau, V., Jung, G., and Götz, F. (1994) Mass spectroscopic analysis of a novel enzymatic reaction. J. Biol. Chem. 269, 5653–5659. 43. Kupke, T., Kempter, C., Jung, G., and Götz, F. (1995) Oxidative decarboxylation of peptides catalyzed by flavoprotein EpiD: determination of substrate specificity using peptide libraries and neutral loss mass spectrometry. J. Biol. Chem. 270, 11282– 11289. 44. Kupke, T., Stevanovic, S., Ottenwälder, B., Metzger, J. W., Jung, G., and Götz, F. (1993) Purification and characterization of EpiA, the peptide substrate for posttranslational modifications involved in epidermin biosynthesis. FEMS Microbiol. Lett. 112, 43–48. 45. Kupke, T., Stevanovic, S., Sahl, H. G., and Götz, F. (1992) Purification and characterization of EpiD, a flavoprotein involved in the biosynthesis of the lantibiotic epidermin. J. Bacteriol. 174, 5354–5361. 46. Linnett, P. E. and Strominger, J. L. (1973) Additional antibiotic inhibitors of peptidoglycan synthesis. Antimicrob. Agents Chemother. 4, 231–236. 47. Liu, W. and Hansen, J. N. (1993) The antimicrobial effect of a structural variant of subtilin against outgrowing Bacillus cereus T spores and vegetative cells occurs by different mechanism. Appl. Environ. Microbiol. 59, 648–651. 48. Liu, W. and Hansen, J. N. (1992) Enhancement of the chemical and antimicrobial properties of subtilin by site-directed mutagenesis. J. Biol. Chem. 267, 25078–25085. 49. Metzger, J. W., Kempter, C., Wiesmuller, K. H., and Jung, G. (1994) Electrospray mass spectrometry and tandem mass spectrometry of synthetic multicomponent peptide mixtures: determination of composition and purity. Anal. Biochem. 219, 261–277. 50. Molitor, E. and Sahl, H.-G. (1991) Applications of nisin: a literature survey. In: Jung, G. and Sahl, H.-G. (eds.), Nisin and novel lantibiotics. Leiden: Escom, pp. 434–439. 51. Moll, G. N., Konings, W. N., and Driessen, A. J. (1998) The lantibiotic nisin induces transmembrane movement of a fluorescent phospholipid. J. Bacteriol. 180, 6565–6570. 52. Moll, G. N., Roberts, G. C., Konings, W. N., and Driessen, A. J. (1996) Mechanism of lantibiotic-induced pore-formation. Antonie Van Leeuwenhoek. 69, 185–191. 53. Ottenwälder, B., Kupke, T., Brecht, S., Gnau, V., Metzger, J., Jung, G., and Götz, F. (1995) Isolation and characterization of genetically engineered gallidermin and epidermin analogs. Appl. Environ. Microbiol. 61, 3894–3903. 54. Otto, M., Peschel, A., and Götz, F. (1998) Producer self-protection against the lantibiotic epidermin by the ABC transporter EpiFEG of Staphylococcus epidermidis Tü3298. FEMS Microbiol. Lett. 166, 203–211. 55. Perego, M., Glaser, P., Minutello, A., Strauch, M. A., Leopold, K., and Fischer, W. (1995) Incorporation of d-alanine into lipoteichoic acid and wall teichoic acid in Bacillus subtilis. J. Biol. Chem. 270, 15598–15606. 56. Peschel, A., Augustin, J., Kupke, T., Stevanovic, S., and Götz, F. (1993) Regulation of epidermin biosynthetic genes by EpiQ. Mol. Microbiol. 9, 31–39. 57. Peschel, A. and Götz, F. (1996) Analysis of the Staphylococcus epidermidis genes epiF, E, and G involved in epidermin immunity. J. Bacteriol. 178, 531–536. 58. Peschel, A., Ottenwälder, B., and Götz, F. (1996) Inducible production and cellular location of the epidermin biosynthetic enzyme EpiB using an improved staphylococcal expression system. FEMS Microbiol. Lett. 137, 279–284.
90
References 59. Peschel, A., Otto, M., Jack, R. W., Kalbacher, H., Jung, G., and Götz, F. (1999) Inactivation of the dlt operon in Staphylococcus aureus confers sensitivity to defensins, protegrins, and other antimicrobial peptides. J. Biol. Chem. 274, 8405–8410. 60. Peschel, A., Schnell, N., Hille, M., Entian, K.-D., and Götz, F. (1997) Secretion of the lantibiotics epidermin and gallidermin: sequence analysis of the genes gdmT and gdmH, their influence on epidermin production and their regulation by EpiQ. Mol. Gen. Genet. 254, 312–318. 61. Prasch, T., Naumann, T., Markert, R. L., Sattler, M., Schubert, W., Schaal, S., Bauch, M., Kogler, H., and Griesinger, C. (1997) Constitution and solution conformation of the antibiotic mersacidin determined by NMR and molecular dynamics. Eur. J. Biochem. 244, 501–512. 62. Rayman, M. K., Aris, B., and Hurst, A. (1981) Nisin: a possible alternative or adjunct to nitrite in the preservation of meats. Appl. Environ. Microbiol. 41, 375–380. 63. Reis, M., Eschbach-Bludau, M., Iglesias-Wind, M. I., Kupke, T., and Sahl, H. G. (1994) Producer immunity towards the lantibiotic Pep5: identification of the immunity gene pepI and localization and functional analysis of its gene product. Appl. Environ. Microbiol. 60, 2876–2883. 64. Reisinger, P., Seidel, H., Tschesche, H., and Hammes, W. P. (1980) The effect of nisin on murein synthesis. Arch. Microbiol. 127, 187–193. 65. Rogers, L. A. and Whittier, E. O. (1928) The inhibitory effect of Streptococcus lactis on Lactobacillus bulgarius. J. Bacteriol. 16, 321–325. 66. Ruhr, E. and Sahl, H. G. (1985) Mode of action of the peptide antibiotic nisin and influence on the membrane potential of whole cells and on cytoplasmic and artificial membrane vesicles. Antimicrob. Agents Chemother. 27, 841–845. 67. Sahl, H.-G., Jack, R. W., and Bierbaum, G. (1995) Biosynthesis and biological activities of lantibiotics with unique post-translational modifications. Eur. J. Biochem. 230, 827–853. 68. Sahl, H. G. (1991) Pore formation in bacterial membranes by cationic lantibiotics. In: Jung, G. and Sahl, H. G. (eds.), Nisin and novel lantibiotics. Leiden: Escom, pp. 347– 358. 69. Sahl, H. G. and Brandis, H. (1983) Efflux of low Mr substances from the cytoplasm of sensitive cells caused by the staphylococcin-like agent pep5. FEMS Microbiol. Lett. 16, 75–79. 70. Schnell, N., Engelke, G., Augustin, J., Rosenstein, R., Ungermann, V., Götz, F., and Entian, K. D. (1992) Analysis of genes involved in the biosynthesis of lantibiotic epidermin. Eur. J. Biochem. 204, 57–68. 71. Schnell, N., Entian, K. D., Götz, F., Horner, T., Kellner, R., and Jung, G. (1989) Structural gene isolation and prepeptide sequence of gallidermin, a new lanthionine containing antibiotic. FEMS Microbiol. Lett. 49, 263–267. 72. Schnell, N., Entian, K. D., Schneider, U., Götz, F., Zähner, H., Kellner, R., and Jung, G. (1988) Prepeptide sequence of epidermin, a ribosomally synthesized antibiotic with four sulphide-rings. Nature 333, 276–278. 73. Schüller, F., Benz, R., and Sahl, H. G. (1989) The peptide antibiotic subtilin acts by formation of voltage-dependent multi-state pores in bacterial and artificial membranes. Eur. J. Biochem. 182, 181–186. 74. Siegers, K. and Entian, K.-D. (1995) Genes involved in immunity to the lantibiotic nisin produced by Lactococcus lactis 6F3. Appl. Environ. Microbiol. 61, 1082–1089. 75. Siezen, R. J., Kuipers, O. P., and de Vos, W. M. (1996) Comparison of lantibiotic gene clusters and encoded proteins. Antonie Van Leeuwenhoek. 69, 171–184. 76. Stock, J. B., Ninfa, A. J., and Stock, A. M. (1989) Protein phosphorylation and regulation of adaptive response in bacteria. Microbiol. Rev. 53, 450–490. 77. Vogel, H., Nilsson, L., Rigler, R., Meder, S., Boheim, G., Beck, W., Kurth, H. H., and Jung, G. (1993) Structural fluctuations between two conformational states of a trans-
91
3 Biosynthesis of the Lantibiotics Epidermin and Gallidermin membrane helical peptide are related to its channel-forming properties in planar lipid membranes. Eur. J. Biochem. 212, 305–313. 78. Weil, H. P., Beck-Sickinger, A. G., Metzger, J., Stevanovic, S., Jung, G., Josten, M., and Sahl, H. G. (1990) Biosynthesis of the lantibiotic Pep5: isolation and characterization of a prepeptide containing dehydroamino acids. Eur. J. Biochem. 194, 217–223.
92
4 Fermentation of Lantibiotics Epidermin and Gallidermin Uwe Theobald*
4.1 Introduction
In recent years the research of the lantibiotics epidermin and gallidermin focused on two topics, mainly the more genetically based work on the biosynthesis of those antimicrobial peptides [1–3] and their production in bioreactors, respectively. Since antibiotics serve a vital role in modern medicine, fermentation technology represents a commercially significant part [4, 5]. Additionally, the need for an industrial production of new bioactive compounds become more and more important due to the increasing problem of multi-resistence in bacteria [6]. Both lantibiotics, gallidermin produced by Staphylococcus gallinarum Tü 3928 [7] and epidermin produced by Staphylococcus epidermidis Tü 3298 [8] exhibit strong activity against Gram-positive bacteria, particularly against propioni bacteria which are involved in the acne disease. Other favourable properties for a large-scale production are the comparable biological activity to renowned antibiotics in current clinical practice like erythromycin or fusidin because of the bactericidal impressiveness of both compounds against actively growing and non-dividing bacteria, the advantage for treatment of endocarditis, abscesses or skin infections. Finally, gallidermin or epidermin are eligible alternatives to vancomycin, so that these compounds gained pharmaceutical interest. The remarkable pharmacological properties led to extensive investigations during the last decade to optimise the production process of these drugs for industrial purpose [9–12]. However, each biotechnological production process is an unique operation. So, every step of this process needs optimisation for a promising conversion into an industrial standard. Major crucial points and optimisation steps known in general are the variation of media components or the design of special media compositions [13], the development of a suitable and reproducible process [14], the scale-up problems including oxygen transfer and mixing in the
* LSMW GmbH, Roßbachstraße 38, D-70499 Stuttgart
93 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
4 Fermentation of Lantibiotics Epidermin and Gallidermin bioreactor [15], and the implementation of down stream processing steps to the process [16]. Hence, a complete summary of data concerning the fermentative production of gallidermin or epidermin has to reflect all optimisation steps of a fermentation – from strain cultivation to large scale production. Therefore, this chapter will focus on several different topics of an entire antibiotic production process that enabled efficient metabolite production.
4.2 Strains for gallidermin/epidermin production
The fermentative production of epidermin or gallidermin was possible with different microbial systems. Besides the mentioned wild type strains epidermin could be produced by a recombinant strain of Staphylococcus carnosus. In contrast to those wild type strains, Staphylococcus carnosus is a “food-grade” strain which is applied in starter cultures. Beyond that the strain Staphylococcus carnosus TM 300 was successfully used for heterologous expression of epidermin in a two-plasmid system. With the recombinant strain Staphylococcus carnosus TM 300 (pTepiMA+, pRBgepSI), containing the biosynthesis genes for epidermin production and immunity against its own antibiotic, nearly 70 % of the wild type production yield were observed. A problem of the two-plasmid system was the insufficient plasmid stability, so that most work on process development and optimisation was carried out with the wild type strains S. gallinarum Tü 3928 and S. epidermidis Tü 3298. According to the somewhat higher activity against living and non-dividing Gram-positive bacteria, this review will focus on the antibiotic gallidermin solely, although a lot of successful work was done for epidermin, too [17–21].
4.3 Disadvantages during gallidermin process development
Regarding the development of a gallidermin production process over a couple of years a few problems were generally noticed. First of all, a exceedingly strain instability concerning product excretion was observed, though biomass formation remained still stable. This fluctuating lantibiotic production yield was contradictory to a reproducible and standardised process. A second crucial point 94
4.4 Gallidermin – a lantibiotic and its way towards industrial production was the composition of the production medium. Several investigations dealt with the search for a defined and synthetic medium for staphylococci, but the biomass and product concentrations gained with the developed medium [22] could not compete with any of the tested complex media. Best production yield could be observed with a medium based upon meat extract. Meat extract was supposed to be the only suitable complex medium component that led to appropriate product yields [9] but is first of high costs for production or downstream processing and second carried a possible risk of prion infections such as bovine spongiform encephalopathy or Creutzfeld-Jakob-disease [23–25].
4.4 Gallidermin – a lantibiotic and its way towards industrial production
The basis for a microbial process is the stock culture for storage of the talented strain. This first step of the production could be identified as prime cause for the problem of the mentioned instability in product formation during fermentation [26]. A new mathematical parameter (hs-value) based upon the commonly analysed parameters biomass and product concentration was introduced to facilitate the optimisation of the media used for this cultivation of stock cultures on agar slants. The hs-value (high and stable product concentration) reduced the amount of data generated in optimisation experiments to one single value for each medium composition and allowed an assessment of any medium formulation with regard to reproducibility and product formation [26]. Figure 4.1 illustrates the time course of the product concentration over several passages found in the corresponding liquid cultures after 24 hours of incubation (A) and the corresponding hs-values (B) obtained exemplary from four different stock culture media (out of 53 different formulations tested) during several passages. Another indispensable optimisation step was the design of a new medium composition. An analysis of different complex compounds [13] suggested the possibility for a successful use of yeast extract (without risk of prion infection). In contrast to earlier results [9, 27] the investigations revealed that a special yeast extract could efficiently replace meat extract in the production medium of S. gallinarum Tü 3928. In addition to that the special yeast extract (Ohly KAT) applied in a three-fold lower concentration led to approximately 20 % higher product yields compared to the old medium composition [28]. Figure 4.2 shows a comparison of the gallidermin concentrations obtained with these media in batch fermentations. Since the Staphylococcus species are very halotolerant in general, the effect of salt (cation) is very important for growth and production [9–11, 27]. To optimise this parameter, different cations in various concentrations were 95
4 Fermentation of Lantibiotics Epidermin and Gallidermin
Figure 4.1: Gallidermin concentrations (A) and corresponding hs-values (B) obtained from a suitability test shown for four different stock culture media during several passages.
tested [12, 28–30]. Best results were gained with CaCl2 in very high concentrations of 30 to 40 g/l whereas the other cations like NaCl showed only slight influence. Figure 4.3 indicated, that there is an optimum in CaCl2 concentration whereas the NaCl concentration is only of slight influence upon product formation. However, Peschel and co-workers demonstrated that cationic teichonic acids at the staphylococcal cell surface are involved in gallidermin resistance [31]. This may be due to an ionic interaction between the teichonic acids and the gallidermin molecule that also has a positive net charge (3-fold). Although the mechanism of resistance is not yet clear in detail [1] it is possible that Ca2+ 96
4.4 Gallidermin – a lantibiotic and its way towards industrial production
Figure 4.2: Comparison of gallidermin formation by Staphylococcus gallinarum Tü 3928 grown in the previously used medium (open symbols) and the new developed production medium (full symbols). Both cultures were inoculated from the same agar plate.
Figure 4.3: Maximal gallidermin concentrations in the culture broth of Staphylococcus gallinarum Tü 3928 in dependency on different NaCl/CaCl2 proportions in the production medium.
97
4 Fermentation of Lantibiotics Epidermin and Gallidermin probably may be involved in the immunity as well. Hence, the components of the new medium were roughly (qualitatively) adjusted. Final optimisation were carried out with a computer-aided procedure. Genetic algorithms are able to optimise simultaneously a multi-parameter system like a medium composition (quantitatively optimisation) [32]. Several fermentation modes (batch, fed-batch, continuous) were tested. Preliminary experiments in continuous cultivations resulted in lower gallidermin yields compared to those gained in batch or fed-batch cultures. Best results were determined in batch processes or fed-batch processes with pulses of several amino acids or maltose during the antibiotic production phase [33]. A comparison of different types of bioreactor (dialysis fermenter, airlift reactor, stirred tank and loop reactor) showed no significant differences in growth and production. The effect of dissolved oxygen tension seemed to be more critical [11] than the type of reactor. The dialysis reactor consisted of two chambers which are separated by a dialysis membrane with an exclusion limit of 10 000 Dalton [10]. This reactor has the advantage of on-line product separation during the production phase but has the crucial drawback that no scale-up procedure is possible. Hence, a transfer from laboratory scale to industrial scale seemed to be out of question at that time. Scale-up investigations were focused on the batch and fed-batch processes in stirred tank reactors, which enabled a reliable scale-up procedure from 0.2 litre to 200 litre scale [34]. Best results in large scale fermentations were achieved by addition of maltose during the late production phase as shown in Fig. 4.4.
Figure 4.4: Concentrations of gallidermin (full symbols) and cell dry weight (open symbols) observed during a scale-up from 20 to 200 litres bioreactor. The arrow marks the transition from 20 to 200 litres and the time when maltose is pulsed into the 200 litres reactor; the grey area shows the range of the maximal gallidermin concentration (including errors) expected in a 200 litres bioreactor without maltose addition.
98
4.5 Conclusion The downstream processing (product recovery from the fermentation broth) was successfully optimised by product adsorption onto resin (Amberlite XAD-1180) [35]. An integration of the product separation step was performed by direct addition of resin into the bioreactor during fermentation or addition of an cross-flow filtration step between bioreactor and adsorber column. This resulted in a cell retention by the cross-flow module during the production phase. Product containing filtrate was harvested afterwards onto an adsorber column with XAD resin. So, the overall production time could be prolonged.
4.5 Conclusion
Extensive investigations led to a new method for strain storage and cultivation, the design of a new production medium, process development and the ability for scale-up into technical scale. An integration of a filtration step during production permitted an on-line product harvesting. Finally, it should be emphasised that higher production yields were obtained (more than 300 mg/l) and costs for medium components were drastically reduced more than 15-fold from 72 Euro per gram gallidermin to approximately 4.6 Euro per gram. Figure 4.5 summarised the reduction of medium costs and the increase in gallidermin concentration during the three optimisation steps (replacement of ingredients, rough adjustment of salt concentrations and final optimisation using a computer program).
Figure 4.5: Evolution of medium costs (columns) and product yields during three optimisation steps (see text).
99
4 Fermentation of Lantibiotics Epidermin and Gallidermin
References
1. Jack, R. W., Bierbaum, G., and Sahl, H.-G. (1998) Lantibiotics and related peptides. Springer-Verlag, Berlin. 2. Bierbaum, G., Götz, F., Peschel, A., Kupke, T., van de Kamp, M., and Sahl, H.-G. (1996) The biosynthesis of the lantibiotics epidermin, gallidermin, pep5 and epilancin K7. Antonie van Leeuwenhoek. 69, 119–127. 3. Kupke, T. and Götz, F. (1996) Post-translational modifications of lantibiotics. Antonie van Leeuwenhoek. 69, 139–150. 4. Strohl, W. R. (1997) Industrial antibiotics: Today and the future. In: Biotechnology of Antibiotics (ed. W. R. Strohl), Marcel Dekker, New York, pp 1–48. 5. Neijssel, O. M., Teixeira de Mattos, M. J., and Tempest, D. W. (1993) Overproduction of metabolites. In: Biotechnology (eds. H.-J. Rehm and G. Reed), Vol 1 Biological fundamentals (ed. H. Sahm), VCH, Weinheim, pp 163–188. 6. Dennensen, P. J. W., Bonten, M. J. M. and Weinstein, R. A. (1998) Multiresistant bacteria as a hospital epidemic problem. Annals of Medicine 30, 176–185. 7. Kellner, R., Jung, G., Hörner, T., Zähner, H., Schnell, N., Entian, K.-D., and Götz, F. (1988) Gallidermin: a new lanthionine-containing polypeptide antibiotic. Eur. J. Biochem. 177, 53–59. 8. Allgaier, H., Jung, G., Werner, R. G., Schneider, U., and Zähner, H. (1986) Epidermin: a sequencing of a heterodet tetracyclic 21-peptide amide antibiotic. Eur. J. Biochem. 160, 9–22. 9. Hörner, T., Ungermann, V., Zähner, H., Fiedler, H.-P., Utz, R., Kellner, R., and Jung, G. (1990) Comparative studies on the fermentative production of lantibiotics by staphylococci. Appl. Microb. Biotechnol. 32, 511–517. 10. Ungermann, V., Goeke, K., Fiedler, H.-P., and Zähner, H. (1991) Optimisation of fermentation and purification of gallidermin and epidermin. In: Nisin and novel lantibiotics (eds. G. Jung and H.-G. Sahl), Escom, Leiden, pp 410–421. 11. Kempf, M., Theobald, U., and Fiedler, H.-P. (1997) Influence of dissolved O2 on the fermentative production of gallidermin by Staphylococcus gallinarum. Biotech. Lett. 19, 1063–1065. 12. Breckel, A., Harder, M., Fiedler, H.-P., and Zähner, H. (1995) Production of gallidermin by Staphylococcus gallinarum Tü 3928. In: Biochemical Engineering 3 (ed. R. D. Schmid), Kurz & Co, Stuttgart, pp 62–66. 13. Greasham, R. L. (1993) Media for microbial fermentations. In: Biotechnology (eds. H.J. Rehm and G. Reed), Vol 3 Bioprocessing (ed. G. Stephanopoulos), VCH, Weinheim, pp 127–140. 14. Fiechter, A. (1986) Bioprocess development. In: Overproduction of microbial metabolism (eds. Z. Vanek and Z. Hostalek), Butterworths, Boston, pp 231–259. 15. Reuss, M. (1993) Oxygen transfer and mixing: scale-up implications. In: Biotechnology (eds. H.-J. Rehm and G. Reed), Vol 3 Bioprocessing (ed. G. Stephanopoulos), VCH, Weinheim, pp 185–217. 16. Spears, R. (1993) Overview of downstream processing. In: Biotechnology (eds. H.-J. Rehm and G. Reed), Vol 3 Bioprocessing (ed. G. Stephanopoulos), VCH, Weinheim, pp 39–55. 17. Jung, G., Allgaier, H., Kellner, R., Schneider, U., Hörner, T., Zähner, H., and Werner R. G. (1987) Isolation, purification and structure elucidation of epidermin, a ribosomally synthesized polypeptide antibiotic. In: Biochemical Engineering 1 (eds. H. Chmiel, W. P. Hammes and J. E. Bailey), Gustav Fischer, Stuttgart, pp 494–497. 18. Werner, R. G., Zähner, H., Jung, G., Allgaier, H., and Schneider, U. (1985) Antibio-
100
References
19.
20. 21.
22.
23. 24. 25. 26.
27. 28.
29.
30. 31.
32. 33.
34.
35.
tisches Polypeptid, Verfahren zu seiner Herstellung und seine Verwendung. European patent No. 0 181 578 B1. Werner, R. G., Zähner, H., Jung, G., Hörner, T., Kellner, R., and Fiedler, H.-P. (1989) Verfahren zur Gewinnung, Isolierung und Reinigung von Epidermin. European patent No. 0 350 810 B1. Allgaier, H., Hentschel, N., Walter, J. and Werner R. G. (1992) Isolation and purification of lantibiotics. European patent No. 0 508 371 A1. Ungermann, V., Hörner, T., Utz, R., Fiedler, H.-P., and Zähner, H. (1991) Comparative studies on the fermentation of lantibiotics, produced by Staphylococci. In: Biochemical Engineering 2 (eds. M. Reuss, H. Chmiel, E.-D. Gilles, and H.-J. Knackmuss), Gustav Fischer, Stuttgart, pp 301–305. Mollenkopf, F. (1998) Untersuchungen zur Epiderminbiosynthese bei Staphylococcus epidermidis Tü 3298 und zum Wachstum von Staphylokokken. Dissertation Universität Tübingen. Collee, J. G. and Bradley, R. (1997) BSE: A decade on – part 1. Lancet 349, 636–641. Collee, J. G. and Bradley, R. (1997) BSE: A decade on – part 2. Lancet 349, 715–721. Ironside, J. W. (1996) Creutzfeld-Jakob disease. Brain Path 6, 379–388. Theobald, U. and Kempf, M. (1998) A novel tool for medium optimisation and characterization in the early stages of a metabolite production process. Biotechnol. Techniques 12, 893–897. Jung, G., Kellner, R., Zähner, H., Götz, F., Hörner, T., Werner, R. G., and Allgaier, H. (1989) Antibiotic. European patent No. 0 342 486 B1. Kempf, M., Theobald, U., and Fiedler, H.-P. (1999) Economic improvement of the fermentative production of gallidermin by Staphylococcus gallinarum. Biotech. Lett. 21, 663–667. Ungermann, V. (1992) Untersuchungen zur Produktbildung und Scale-up der Produktion von Gallidermin einem Lanthioninhaltigen Peptidantibiotikum aus Staphylococcus gallinarum Tü 3928. Dissertation Universität Tübingen. Hörner, T. (1989) Fermentation und Isolierung Lanthioninhaltiger Polypeptidantibiotika aus Staphylococcen. Dissertation Universität Tübingen. Peschel, A., Otto, M., and Götz, F. (1998) Incorporation of D-alanine into staphylococcal teichonic acids confers resistance to antimicrobial peptides from bacteria, animals and humans. Lecture (No. KA021) held at the VAAM Jahrestagung, (22.–25. 3. 1998), Frankfurt. Holland, J.H. (1975) Adaption in natural and artificial systems. The University of Michigan Press Ann Arbor, Michigan. Kempf, M., Theobald, U., and Fiedler, H.-P. (1999) Correlation between the consumption of amino acids and the production of the antibiotic gallidermin by Staphylococcus gallinarum. Biotech. Lett. (accepted for publication). Kempf, M., Theobald, U., and Fiedler, H.-P. (2000) Production of the antibiotic gallidermin by Staphylococcus gallinarum – Development of a scale-up procedure. Biotech. Lett. (submitted for publication). Allgaier, H., Walter, J., Schlüter, M., and Werner, R. G. (1991) Strategy for purification of lantibiotics. In: Nisin and novel lantibiotics (eds. G. Jung and H.-G. Sahl), Escom, Leiden, pp 422–433.
101
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 Christiane Bormann*
5.1 Introduction: nikkomycins
Nikkomycins are peptidyl nucleoside antibiotics that are potent and specific inhibitors of chitin synthetases. They are structurally similar to UDP-N-acetylglucosamine, the natural substrate of these enzymes, and inhibition occurs in a competitive manner [1, 2]. Structures of nikkomycins are shown in Fig. 5.1. Nikkomycins X and Z, the main compounds produced by Streptomyces tendae Tü 901, exhibit high antifungal, insecticidal, and acaricidal activity [3]. Since their toxicity towards mammals and bees is very low or not detectable, and since nikkomycins are easily degraded in nature, they are potentially useful in agriculture or as therapeutic antifungal agents in humans. Nikkomycin Z shows significant activity towards the highly chitinous, pathogenic, dimorphic fungi Coccidioides immitis and Blastomyces dermatitidis [4]. Polyoxins, antibiotics related to nikkomycins, are commercially produced by Streptomyces cacaoi and applied as agricultural fungicides in Japan (for a review, see [5]). Nikkomycins X and Z are composed of the unusual amino acid hydroxypyridylhomothreonine (HPHT; nikkomycin D) and a peptidically linked nucleoside moiety. The nucleoside moiety comprises an aminohexuronic acid N-glycosidically linked to 4-formyl-4-imidazolin-2-one, forming nikkomycin Cx , or to uracil, forming nikkomycin Cz. Minor components of the culture filtrate of S. tendae Tü 901 are nikkomycins I and J, which are structures analogous to nikkomycins X and Z that contain glutamic acid peptidically bound to the 6'-carboxyl group of the aminohexuronic acid. Nikkomycins Cx , Cz, and D are also detected in culture filtrates and arise by hydrolytic cleavage of nikkomycins X and Z. These three compounds neither act as chitin synthetase inhibitors nor display biological activity. Nikkomycins Kx, Kz, Ox, and Oz (Fig. 5.1), which will be dis-
* Mikrobiologie/Biotechnologie, Universität Tübingen, Auf der Morgenstelle 15, D-72076 Tübingen
102 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
5.1 Introduction: nikkomycins
R3 C O R2 NH CH
O R1
OH OH A
O
HN
N
N
O
HO
OH
NH2
OH
R2
CH CH CH C N
C O
CHO
HN
R1
B
N
O
CH3
NH2
CH CH2 CH C
H
O
D
OH
R2
NH2
CH CH2 CH C
HO N
O
nikkomycin
R1
R2
R3
X
A
A
OH
Z
B
A
OH
I
A
A
Glu
J
B
A
Glu
Cx Cz
A B
B B
OH OH
Kx
A
C
OH
Kz
B
C
OH
Ox
A
D
OH
Oz
B
D
OH
D
HO
NH2
OH
CH CH CH C OH N
CH3
O
Figure 5.1: Structures of nikkomycins.
103
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 cussed in a later section, are synthesized by mutants derived from S. tendae Tü 901/8 c by UV treatment and chemical mutagenesis [6]. Biosynthesis of nikkomycins can be divided into two parts: the nucleoside and peptidyl moieties are synthesized in separate pathways and then are linked by peptide bonds [6]. Figure 5.2 summarizes the steps of the nikkomycin biosynthetic pathway based on isotopic labeling and chemical characterization of pathway intermediates and shunt products. It also includes data on the biosynthesis of the polyoxin nucleoside, which contains an aminohexuronic acid moiety identical to that of nikkomycin nucleosides. Isono and coworkers [7] have shown that the polyoxin nucleoside arises from uridine and carbon-3 of phosphoenolpyruvate, and they proposed the reaction mechanism as a condensation of uridine and phosphoenolpyruvate to octofuranuloseuronic acid as the intermediate, followed by oxidative elimination of carbon-7' and carbon-8' and introduction of an amino group on carbon-5'. Isolation of octosyl acids, shunt metabolites derived from the postulated intermediate, supports this hypothesis. Analogues of octosyl acids, nikkomycins Sx and Sz, have been isolated from nikkomycin-producing S. tendae, and therefore the same biosynthetic pathway has been suggested for the nikkomycin nucleosides nikkomycins Cx and Cz [8]. Histidine is the precursor of the imidazolone base [9]; l-lysine is incorporated via picolinic acid into the pyridyl moiety and the attached carbon atom of nikkomycin D [10, 11]. Pyridylhomothreonine (PHT, nikkomycin E) and 4-pyridyl-2-oxo4-hydroxy-isovalerate (POHIV), which have been isolated from S. tendae culture filtrate [12, 13], are potential biosynthetic precursors of HPHT.
5.2 Isolation of nikkomycin biosynthetic genes
A number of approaches have been successfully used to isolate genes of antibiotic biosynthetic pathways from several Streptomyces strains. The strategy of cloning antibiotic-resistance genes, which are usually clustered with antibiotic biosynthetic genes, is not applicable to nikkomycin genes because the producing strain lacks the nikkomycin target site. Complementation of S. tendae mutants blocked in nikkomycin synthesis has led to the isolation of a 9.4-kb fragment that complements a non-producing mutant to nikkomycin Cx, Cz, and Kx synthesis [14]. However, structural genes of the nikkomycin pathway have not been identified. Therefore, the nikkomycin genes were cloned by identifying gene products involved in nikkomycin synthesis, microsequencing the N-termini to design oligonucleotide probes, and cloning the corresponding genes using these probes [15]. Nikkomycin biosynthetic enzymes were identified in gene expression studies using two-dimensional gel electrophoresis to separate cellular proteins. Initially, gene expression was analyzed in S. tendae wild type and mutant strains 104
Figure 5.2: Nikkomycin biosynthetic pathway based on the incorporation of labeled precursors and chemical characterization of pathway intermediates and shunt products; included are data on the biosynthesis of the polyoxin nucleoside (for references, see text). POHIV, 4-pyridyl-2-oxo-4-hydroxyisovalerate; PHT, pyridylhomothreonine.
105
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901
Figure 5.3: Two dimensional gels of silver-stained cell extracts prepared from wild type Tü 901/8 c (A) and nikkomycin non-producing mutant Tü 901/NP13 (B). Mycelia were harvested in the stationary phase after 27.5 h of incubation. Proteins P1–P10 observed in the Tü 901/8c extract (A) and their corresponding positions in the Tü 901/NP13 extract (B) are indicated by arrows. These proteins were identified on the basis of three experiments performed with each producing and non-producing S. tendae strain. Positions of protein molecular mass standards are shown on the left. The acidic end of the gel is on the left, and the basic on the right.
blocked in nikkomycin biosynthesis [6]. Mutants which were able to synthesize nikkomycins but which differed from the wild type in the spectrum of nikkomycin structures formed had a protein pattern identical to that of the wild type. In contrast, protein profiles of mutants which did not synthesize any known nikkomycin structure differed from the wild type pattern as well as from each other. Selection of protein spots that were present in the producing strains and absent in each non-producing mutant led to the identification of ten gene products (P1–P10) (Fig. 5.3). As nikkomycin biosynthesis is regulated by the growth phase, the kinetics of the appearance of proteins P1–P10 was investigated during growth of S. tendae Tü 901/8 c in production medium (Fig. 5.4). Growth of S. tendae is characterized by a biphasic growth curve with two phases of rapid growth separated by a 1.5 to 2 h period in which growth slows down. Nikkomycin production begins in the transition to the stationary phase, after approximately 27 h after inoculation. Proteins P1–P6 and P10 were detected in extracts of mycelia harvested in the second exponential growth phase after 22.5 h of inoculation, and the amount of all of proteins P1–P10 increased to maximum levels at the early stationary phase and remained constant throughout the stationary phase. Synthesis of proteins P1–P10 preceded nikkomycin production, as would be expected for nikkomycin biosynthetic enzymes. N-terminal amino acid sequences were obtained for six of the ten identified proteins. Oligonucleotide probes designed from the N-terminal sequences of proteins P4, P5, and P8 gave positive signals of similar intensities with more 106
5.2 Isolation of nikkomycin biosynthetic genes
Figure 5.4: Time course of nikkomycin production (A) and gene expression (B) in S. tendae Tü 901/8 c cultivated in production medium. Nikkomycin Z plus X (g) in the culture filtrate compared to growth as determined by dry weight (&). Arrows indicate time points at which mycelia were harvested to prepare protein extracts for 2-D gel electrophoresis. (B) Enlarged sections of silver-stained 2-D gels of extracts of S. tendae Tü 901/8c harvested in the exponential phase (17.5 h), the late exponential phase (22.5 h), and the stationary phase (27.5 h). The position of proteins P1, P2, P3, P4, P7, P8, and P9 are marked by arrows.
than ten bands of digested genomic DNA and therefore were unsuitable for isolating the corresponding genes. A mixture of oligonucleotide probes designed from proteins P1 and P2, which have identical N-terminal sequences except that protein P1 has three additional amino acid residues at its N-terminus, and probes designed from protein P6 led to the isolation of genomic DNA fragments containing the binding sites for the probes used. The isolated fragments, an 8-kb BamHI fragment and a 6.5-kb PvuII fragment, appeared to overlap by 1.5-kb, and the former contained the binding sites for both oligonucleotide probes used. DNA sequence analysis revealed that the 8-kb BamHI fragment contains two open reading frames (ORFs) whose deduced N-terminal amino acid sequences are identical to that of proteins P1/P2 and that of protein P6. As gene-disruption mutants in the ORF encoding proteins P1/P2 (designated nikJ) and encoding protein P6 (designated nikI) failed to synthesize nikkomycins, it was evident that nikkomycin biosynthesis genes had been isolated. 107
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901
5.3 Isolation of the nikkomycin gene cluster and expression in Streptomyces lividans
In bacteria, antibiotic biosynthetic genes are usually clustered. In order to isolate the entire nikkomycin gene cluster, a genomic library of S. tendae Tü 901/8 c was constructed in the Escherichia coli/Streptomyces shuttle cosmid pKC505 [16] and screened with the 8-kb BamHI fragment containing the nikI and nikJ genes. Hybridizing cosmids containing inserts of approximately 30 kb were mapped with restriction enzymes. Cosmids p24/32 and p9/43 carried the recognition sites for oligonucleotide probes designed from the N-terminal amino acid sequences of proteins P1/P2, P4, P5, P6, and P8, and of proteins P1/P2, P5, P6, and P8, respectively. S. lividans TK23, which does not produce nikkomycins, was used as a host to express the cloned nikkomycin biosynthetic genes. The nik cluster was not isolated on a single plasmid, as shown by the lack of synthesis of nikkomycins I, J, X, and Z by S. lividans TK23 transformants containing one of the recombinant plasmids. The nik cluster was cloned on two plasmids, p24/32 and p9/43, which carry inserts of about 31 and 27 kb, respectively, 15 kb of which overlap (Fig. 5.5). To facilitate selection for both plasmids, the apramycin resistance gene of p9/43 was removed by restriction with XhoI, which did not cut within the insert, and replaced with the aphII gene from Tn5. One of the apramycin- and neomycin-resistant S. lividans TK23 transformants synthesized relatively large amounts of nikkomycins I, J, X, and Z, approximately 50 % of that synthesized by S. tendae Tü 901/8 c. This transformant contained a large (>100-kb) plasmid that had formed by recombination between homologous regions of the two plasmids, either the vector pKC505 regions or the homologous regions of the inserts [17]. 27 kb
p9/43 31 kb
p24/32
8 kb
probe
B
S
S B
S
BS
B
S
BB
B
B BS
S
B
S
genome VM5 VM8
VM4
orfR
nikV U T
S R Q P2
P1
A B C
VM6 VM1/2
D
E
F G
I
J
K
LM N
O
Figure 5.5: Organization of the nik gene cluster and restriction map of the cloned chromosomal region containing the nik cluster. The structural genes (nik) and the regulatory gene (orfR) of the nik cluster are indicated by arrows. Wavy arrows indicate transcripts. The boxes below the genome map indicate the recognition sites for oligonucleotide probes VM1/2, VM4, VM5, VM6, and VM8, designed from proteins P1/2, P4, P5, P6, and P8. The hybridization probe used to screen the S. tendae gene library and the inserts of cosmid p24/32 and p9/43 are also indicated. B, BamHI; S, SacI.
108
5.4 Organization of the nikkomycin gene cluster
5.4 Organization of the nikkomycin gene cluster
The nucleotide sequence of an approximately 36-kb genomic region containing the nik cluster has been determined, and a series of 23 ORFs comprising a region of 29 kb identified. The translational start points of the ORFs have been tentatively located using the following criteria: (a) the G+C content in the third position of codons, (b) the location of a potential ribosome binding site at a suitable distance from the putative translation start [18], and (c) the observed similarities of the deduced amino acid sequence with proteins in databases. Each of the deduced proteins of the nikJ, nikS, nikA, nikI, and nikC genes had an N-terminal amino acid sequence identical to that determined for proteins P1/P2, P4, P5, P6, and P8, respectively. The most relevant features of the nik cluster and the encoded proteins deduced from the nucleotide sequence are summarized in Table 5.1, and the organization of the nik cluster is shown in Fig. 5.5. The ORFs of the nik cluster are arranged in three sets of adjacent genes with an intergenic spacing of <92 bp, and the stop codon of several ORFs overlaps with Table 5.1: Relevant features of the nik cluster and the encoded proteins deduced from the DNA sequence. ORF
G+C [mol%]
Residues/MW
Predicted/known function
nikA nikB nikC nikD nikE nikF nikG nikI nikJ nikK nikL nikM nikN
69.3 72.2 72.3 71.0 74.6 71.5 74.5 68.4 70.1 73.9 74.7 70.4 75.8
296/31,429 357/37,146 426/47,072 389/43,055 561/60080 410/45,908 64/6,565 218/24,489 453/50,422 374/40,164 240/25,895 213/24,086 550/57,238
Aldehyde dehydrogenase Synthetase Aminotransferase Oxidase Ligase Cytochrome P450 Ferredoxin
nikO nikP1 nikP2 nikQ nikR nikS nikT
73.4 73.2 75.5 73.1 76.7 70.8 73.6
471/50,345 677/73,061 272/29,514 396/44,479 226/24,991 424/46,873 440/46,227
nikU nikV orfR
63.8 69.2 74.0
155/16,549 423/46,077 1062/114,701
Aminotransferase
Membrane transport protein Enoylpyruvate transferase Peptide synthetase Thioesterase Cytochrome P450 Phosphoribosyltransferase Ligase Acyl carrier protein and aminotransferase Mutase, small subunit Mutase, large subunit Regulatory protein
109
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 the start codon of the subsequent gene, implicating translational coupling. The nikA-nikG gene cluster and the nikI-nikO gene clusters are separated by 944 bp and are oriented in the same direction, while the nikP1-nikV gene cluster is divergently transcribed; the intergenic region between nikP1 and nikA is 273 bp. The orfR gene, which encodes a pathway-specific activator protein for the nik biosynthesis genes is located at the left end of the nik cluster and is divergently oriented relative to the nikP1-nikV genes, with a 447-bp intergenic spacing.
5.5 Roles of the nik genes
The likely roles of the nik genes in nikkomycin biosynthesis as discussed below have been established from: (a) the nikkomycins structures synthesized by strains with inactivated nik genes, (b) sequence comparisons, and (c) enzymatic activities of nik gene products. Except for the nikG and nikU genes, whose roles seemed to be obvious, each gene of the nik cluster was inactivated by inserting a kanamycin resistance cassette via double cross-over homologous recombination. Novel nikkomycin structures accumulated by nik::aphII mutants were isolated from culture broth and their chemical structure (Fig. 5.6) was elucidated. O
R: NH 2
OH
COOH
CH CH CH C NH N
CH3
R
CHO
HN O
N
HN O
N
CH O
(2)
(1)
O
OH OH CHO
HN CHO
HN O
N H
O
HOCH2
N
O
(3) OH OH
(4)
Figure 5.6: Novel nikkomycin structures and biosynthetic intermediates synthesized by nik gene insertion mutants. Nikkomycins Lx (1) and Lz (2) are synthesized by nikF::aphII; 4-formyl-4-imidazolin-2-one (3) by nikK::aphII and fluorouracil-resistant nikR::aphII; and ribofuranosyl-4-formyl-4-imidazolin-2-one (4) by nikO::aphII.
110
5.5 Roles of the nik genes Table 5.2: Nikkomycin structures synthesized by mutants containing an inactivated nik gene. Mutant
Nikkomycin structures
Mutant
Nikkomycin structures
nikA::aphII nikB::aphII nikC::aphII nikD::aphII nikE::aphII nikF::aphII nikI::aphII nikJ::aphII nikK::aphII nikL::aphII nikM::aphII nikN::aphII nikO::aphII
Cx, Cz Cx, Cz Cx, Cz Cx, Cz Cx, Cz Lx , Lz RT 2.7 RT 2.7 Sx, Sz, (3) RT 0.75, Sx, Sz Sx , Sz Sx , Sz (4), RT 2.7
nikP1::aphII nikP2::aphII nikQ::aphII nikR::aphII nikS::aphII nikT::aphII nikV::aphII
Cx*, Cz I*, J, X*, Z RT 2.7, J, Z I*, J, X*, Z Cx, Cz Cx, Cz Cx, Cz, Kx, Kz, Ox, Oz
orfR::aphII
none
Structures of nikkomycins, intermediates, and shunt products are shown in Figs. 5.1, 5.2, and 5.6. Numbers in parentheses refer to structures shown in Fig. 5.6. RT 0.75 and RT 2.7 indicate unknown compounds that might comprise nikkomycin structures; RT refers to retention time in HPLC analyses. * reduced amount; approximately 60 % of wild type nikkomycin X production.
Table 5.2 summarizes the nikkomycin structures synthesized by gene-insertion mutants. Except for the nikJ::aphII and nikP1::aphII mutants, all mutants were able to synthesize the wild type nikkomycin spectrum by genetic complementation or by feeding suitable biosynthetic intermediates, which excluded the possibility that the mutant phenotype was caused by a polar effect of the integrated kanamycin cassette on downstream genes. Table 5.1 summarizes the predicted function of the deduced Nik proteins based on results of database searches. The nikC, nikT, and nikR genes have been expressed in E. coli, and the gene products catalyze the reactions shown in Fig. 5.7.
5.5.1 Biosynthetic pathway of the peptidyl moiety Figure 5.8 shows the proposed biosynthetic pathway for the peptidyl moiety HPHT [19]. The genes needed for its synthesis are all the genes of the nikAnikG operon and the clustered nikT-nikV genes of the nikP1-nikV operon (Fig. 5.5). Except for the nikF::aphII mutant, all mutants inactivated in one of these genes fail to synthesize biologically active nikkomycins, but accumulate the nucleoside structures nikkomycins Cx and Cz. The initial reaction in the HPHT biosynthetic pathway is catalyzed by NikC, which has been characterized as l-lysine-2-aminotransferase [20]. NikC 111
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 OH
OH
R
R O
O
NH2
O OH
(1)
O
NH2
O
OH
(2)
CH CH C C N
CH3
C
O
NikC
N H2
H 2O
OH
PLP
NH 2
O
phenylalanine PLP O
OH
N
O OH
phenylpyruvate
OH
NH2
CH CH CH C
NikT
N
CH3
O OH
O HN
P
(3)
O CH2
O O O
OH OH
P
P
+
O
PPi
HN O
N H
NikR
P
O CH2
N O
OH OH
Figure 5.7: Reactions catalyzed in vitro by recombinant nik gene products. Reactions (1), (2), and (3) are catalyzed by NikC, NikT, and NikR, respectively. PLP, pyridoxalphosphate.
removes the a-amino group from lysine to create piperideine-2-carboxylic acid, the cyclic and dehydrated form of 2-oxo-6-aminohexanoic acid (Fig. 5.7, reaction (1)). NikC belongs to a family of pyridoxamine/pyridoxal-phosphate-dependent enzymes, such as dehydrases involved in the formation of dideoxyhexoses of the Gram-negative cell wall and aminotransferases involved in deoxyamino sugar biosynthesis in secondary metabolism [21, 22]. NikC is a secondary metabolite enzyme that is specific for nikkomycin biosynthesis and is not required for l-lysine catabolism in S. tendae. The next step, the oxidation of piperideine-2-carboxylic acid to picolinic acid, is predicted to be catalyzed by the nikD-encoded protein, as feeding of picolinic acid to nikD mutants restores the ability to synthesize nikkomycins I, J, X, and Z. The NikD protein belongs to the family of monomeric sarcosine oxidases that contain FAD as the coenzyme. Members of this family metabolize sarcosine, l-pipecolic acid, and l-proline [23]. These latter two substrates are oxidized to D-piperideine-6-carboxylic acid and D-pyrrolidine-5-carboxylic acid, respectively. Piperideine-2-carboxylic acid is structurally similar to these compounds. A reaction mechanism similar to that of sarcosine oxidases is proposed 112
NH 2
Figure 5.8:
HO
L-lysine
NH 2
O
N
O
CH
OH
NikC
O
NH2
N
HPHT
CH3
COOH N picolinic acid
COOH
O2+2H+
N
CH
OH
2-oxoglutaric acid 2-oxoglutamic acid
NH2
R
PHT
O
S CoA
O
R
NH2
O
OH
N
CO2
O H
O
NikB
POHIV
CH 3
CH C H C COOH
OH
2-oxo-3-methylsuccinic acid
CH3
N
C
pyridine-2carbaldehyde
CoA-SH NAD+
NikA
+
HC C COOH
HOOC
NikT
O
OH
NikU, NikV
C H CH COOH
CH 3
N
C
NADH + H
picolinic acid-CoA
H 2O AMP + PPi
NikE
CoA-SH ATP
HOOC C H2 CH2 C COOH
NikD
4 [H]
NikF,NikG
H 2O
piperideine-2carboxylic acid
H 2O
C H CH COOH
NH2
OH
Proposed biosynthetic pathway for the peptidyl moiety HPHT.
O
OH
R
O
OH
O
R
5.5 Roles of the nik genes
113
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 for NikD, except that oxidation of piperideine-2-carboxylic acid to picolinic acid requires the removal of two electron pairs. The NikD protein contains an ADPbinding, bab-fold fingerprint sequence at its N-terminus, with an aspartate at position 1, which has been found in several FAD-containing enzymes and seems to be important for the interaction with FAD [24]. This finding implies that NikD might use FAD as a coenzyme. The presumed role of the nikE gene product is the activation of picolinic acid to form picolinic acid-coenzyme A, the predicted substrate for NikA. This proposed reaction rests on the similarity of NikE to adenylate-forming enzymes. NikE is most similar to enzymes that adenylate aromatic acids, such as SnbA, which activates 3-hydroxy-picolinic acid in the pristinamycin biosynthetic pathway [25], and 2,3-dihydroxybenzoate-AMP ligases of enterobactin biosynthesis [26] and other EntE homologs. NikE contains the conserved core sequences needed for ATP-binding and hydrolysis and for adenylate formation [27]. As feeding of HPHT to nikD mutants restores the ability to synthesize nikkomycins I, J, X, and Z, it is clear that NikE functions in the nikkomycin biosynthetic pathway and is not responsible for the formation of peptide bonds. The role of the nikA- and nikB-encoded proteins is deduced from their significant sequence similarity to aldehyde dehydrogenases (acylating) (ADAs) and 4-hydroxy-2-oxovalerate aldolases (HOAs), respectively, of the meta-cleavage pathway for the metabolism of aromatic hydrocarbons. The aldolase releases pyruvate and acetaldehyde, and the acylating dehydrogenase converts acetaldehyde to acetyl-CoA using NAD+ and CoA as cofactors. The mechanistically reverse reactions are postulated for NikA and NikB to form the carbon-carbon bond, thereby yielding the entire skeleton of HPHT. The proposed role of NikA is to catalyze the reduction of picolinic acid-CoA to pyridine 2-carbaldehyde with the release of coenzyme A. The NikA protein has an ADP-binding, bab-fold fingerprint sequence at its N-terminus [28], indicating that NAD(P)H might function as a hydrogen donor. NikB is proposed to catalyze a Claisen condensation of pyridine 2-carbaldehyde and a 2-oxo acid compound that will be discussed below. Two regions of NikB also exhibit similarity with regions of NifV homocitrate synthetases, which catalyze condensation of acetyl-CoA with 2-oxoglutarate to form homocitrate. The reaction mechanism of carbon-carbon bond formation by NifV homologs, also suggested for NikB, is proposed to be similar to that of carbon-carbon-cleavage by HOAs [29]. Interestingly, the order of the nikA and nikB genes in the nikA-nikG operon is the same as that of genes encoding ADA and HOA in the meta-pathway gene clusters. The order of these genes is strictly conserved; the ADA-encoding gene is located directly upstream of the HOA-encoding gene. The two proteins are suggested to form a pre-adapted catabolic module, which is thought to be assembled with other modules of the meta-cleavage operons in Pseudomonas [30]. NikA and NikB represent the first proteins shown to have significant sequence similarities along their entire length with ADAs and HOAs, respectively, that are not involved in meta degradation of aromatic hydrocarbons. These features indicate a common evolutionary origin. The presumed role of nikU and nikV is to catalyze the isomerization of 2oxoglutaric acid to yield 2-oxo-3-methylsuccinic acid, which is the predicted 2114
5.5 Roles of the nik genes oxo acid substrate in the NikB-mediated condensation reaction that yields the skeleton for HPHT with the release of carbon dioxide [31]. This hypothesis is based on the significant sequence similarity of NikU and NikV to methylaspartate mutase S and E chains, respectively, from Corynebacterium. The coenzyme B12-dependent mutase from Corynebacterium is composed of two components (S and E) and converts L-glutamate to threo-b-methyl-L-aspartate [32]. The analogous reaction is proposed for the putative mutase composed of NikV and NikU. The phenotype of nikV mutants is compatible with the proposed role for NikV, as the nikV mutants accumulate nikkomycins Kx , Kz, Ox , and Oz (Fig. 5.1) in addition to the nucleoside moieties nikkomycins Cx and Cz. Compared to HPHT, the peptidyl moiety of these nikkomycins lacks the methyl group at carbon-3. The phenotype of the nikV mutants can be rationalized by assuming that the lack of the nikV gene leads to an increased intracellular concentration of 2oxalacetate or pyruvate, which might be substrates for NikB, yielding the peptidyl moiety of nikkomycins Kx, Kz , Ox, and Oz . The nikT gene encodes a protein which consists of two domains [31]. The N-terminal region (approximately 70 amino acid residues) exhibits 60 % sequence similarity to acyl carrier proteins and contains a putative 4'-phosphopantetheine binding sequence. This domain of NikT is predicted to be involved in peptide bond formation (Section 5.5.3). The other part of NikT (amino acid 113 to 440) reveals significant sequence similarity to aminotransferase class-II and contains a putative pyridoxal-phosphate attachment site. As recombinant NikT transaminates POHIV in vitro to yield PHT (Fig. 5.7, reaction (2); shown in collaboration with K. Liebetrau, University of Münster), the role proposed for NikT is to catalyze this transamination reaction. The final reaction in the HPHT biosynthetic pathway is presumed to be catalyzed by the nikF-encoded cytochrome P450. The NikF protein is responsible for introducing the hydroxy group into the pyridyl ring since disruption of the nikF gene abolishes nikkomycin I, J, X, and Z synthesis, but leads to the formation of nikkomycin Lx and Lz, homologs of nikkomycins X and Z that contain an unhydroxylated pyridyl residue ((1), (2) in Fig. 5.6). PHT is assumed to be the substrate for NikF since PHT has been detected in culture filtrates of S. tendae [12]. NikF is a member of the Class I cytochrome P450 monooxygenases, which are reduced via an iron-sulfur protein by a flavin-containing reductase that accepts electrons from NAD(P)H. The ferredoxin-encoding gene nikG is located immediately downstream of the nikF gene, implying that NikG might be the in vivo electron transport protein for NikF.
5.5.2 Biosynthetic pathway for the nucleoside moiety Figure 5.9 summarizes the information on the nucleoside biosynthetic pathway. The genes nikQ and nikR, arranged in this order within the nikP1-nikV operon, are predicted to be responsible for the formation of 5'-phosphoribofuranosyl-4115
Figure 5.9: and Cz.
Proposed biosynthetic pathway for the nucleoside moieties nikkomycins Cx
5.5 Roles of the nik genes formyl-4-imidazolin-2-one, the precursor for nikkomycin Cx [33]. The genes nikO, nikI, nikJ, nikL, nikM, and nikK, which are located in a single operon, are proposed to be responsible for the formation of the aminohexuronic acid portion of the nucleoside skeletons originating from 5'-phosphoribofuranosyl-4-formyl4-imidazolin-2-one, uridine-monophosphate, and phosphoenolpyruvate [34]. The presumed role of the nikQ-encoded cytochrome P450 monooxygenase, the conversion of L-histidine to 4-formyl-4-imidazolin-2-one, has been deduced from the following findings. The nikQ mutants fail to produce nikkomycins X and I, which contain the imidazolone base, but synthesize nikkomycins Z and J, which contain uracil as the base. Feeding of 4-formyl-4-imidazolin-2-one ((3) in Fig. 5.6) isolated from culture filtrate of a nikK mutant restores the ability of nikQ mutants to synthesize nikkomycins X and J. The nikR gene product is thought to catalyze the transfer of 4-formyl-4imidazolin-2-one to 5-phosphoribosyl-1-pyrophosphate with the release of pyrophosphate, yielding 5'-phospho-ribofuranosyl-4-formyl-4-imidazolin-2-one. The NikR protein has significant sequence similarity to uracil-phosphoribosyl transferases (UPRTases), which catalyze the conversion of uracil and 5-phosphoribosyl-1-pyrophosphate to uridine-monophosphate in the uracil salvage pathway. This reaction is also catalyzed in vitro by recombinant NikR (Fig. 5.7, reaction (3)). The nikR mutants still synthesize nikkomycins containing the imidazolone base (X, I), but at levels of approximately 60 % of wild type levels. Fluorouracilresistant nikR mutants that have no detectable UPRTase activity fail to produce nikkomycins X and I, but synthesize wild type levels of the uracil-containing nikkomycins Z and J. In addition, 4-formyl-4-imidazolin-2-one (Fig. 5.6, structure (3)) is accumulated in culture filtrate. This indicates that the primary metabolism enzyme UPRTase can partly substitute for NikR in nikkomycin biosynthesis in nikR mutants. The proposed role of the nikO gene product is the transfer of phosphoenolpyruvate to the oxygen atom of carbon-5' of the riboside structures, yielding in further steps octofuranuloseuronic acids. NikO has significant sequence similarity to UDP-N-acetylglucosamine enolpyruvyl transferases, which catalyze the first committed step in murein biosynthesis for the bacterial cell wall, and to a minor extent to 5-enol-pyruvylshikimate-3-phosphate synthases of the shikimate pathway. These enzymes catalyze the transfer of phosphoenolpyruvate to a substrate. The nikO mutants accumulate ribofuranosyl-4-formyl-4-imidazolin2-one ((4) in Fig. 5.6), compatible with the role proposed for NikO [35]. The gene products of nikL, nikM, and nikK are involved in the nucleoside biosynthetic pathway after the nikO gene product, as all mutants inactivated in one of these genes accumulate nikkomycins Sx and Sz, which derive from the predicted octofuranuloseuronic acids (Fig. 5.2). While the amino acid sequence of NikL and NikM gave no clue to their role, NikK has significant sequence similarity to histidinol-phosphate aminotransferases. Based on this finding, NikK is proposed to introduce the amino group to 5'-oxohexuronic acid compounds, which are assumed to be the direct precursors for nikkomycins Cx and Cz. Less information is available on the role and order of the nikI and nikJ gene products in the nucleoside biosynthetic pathway. The nikI and nikJ mu117
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901 tants do not accumulate nikkomycins Sx and Sz , but accumulate the unknown compound RT 2.7, which might be a biosynthetic intermediate. Since the nikJ mutants could not be complemented genetically, a polar effect of the inserted kanamycin cassette on the downstream located genes cannot be excluded. However, this seems to be unlikely since the nikJ phenotype differs from that of the nikK and nikL mutants. The nikJ gene product displays limited sequence similarity with methyltransferases involved in the biosynthesis of bialaphos and fortimycin, and with magnesium-protoporphyrin monomethylester oxidative cyclases [36]. The former enzymes use methylcobalamin as the methyl donor. Methylcobalamin and protoporphyrin are structurally similar, and the conserved amino acids might participate in binding or recognizing these cofactors. This implies that NikJ might contain cobalamin as a cofactor and might be responsible for the removal of carbon-7' and/or carbon-8' from the predicted octofuranuloseuronic acid intermediates.
5.5.3 Peptide bond formation The genes nikS, nikT, nikP1, and nikP2 are proposed to catalyze the peptide bond formation between the peptidyl moiety HPHT and the nucleoside structures nikkomycins Cx and Cz, yielding nikkomycin X and Z, respectively (Fig. 5.10) [31]. The NikS protein belongs to a superfamily of enzymes that is characterized by a specific ATP-binding structure. Members of this superfamily reveal strong structural similarity, but limited sequence similarity and catalyze carboxylate-amine/thiol ligation reactions. NikS is proposed to transfer POHIV through
Figure 5.10: Predicted reactions for peptide bond formation. The nikT and nikP1 gene products are represented as boxes, and their domains are indicated by bars. Substrates are attached to the 4'-phosphopantetheine prosthetic group by a thioester bond. R, 4-formyl-4-imidazolin-2-on, uracil, nikX, Z, nikkomycins X, Z.
118
5.5 Roles of the nik genes the phosphate to the 4'-phosphopantetheinyl group attached to the acyl carrier domain of NikT. Aminotransferase reaction and hydroxylation reaction (Fig. 5.8) could involve substrates bound to NikT. nikT and nikS mutants accumulate nikkomycins Cx and Cz and were not complemented to nikkomycin X/Z formation by feeding HPHT to the production medium. In addition, the finding that HPHT has never been found to be accumulated as a biosynthetic intermediate in any of the S. tendae Tü 901 mutants, is compatible with the roles proposed for NikS and NikT. The nikP1 gene encodes a peptide synthetase containing all the core sequences necessary for ATP-binding and adenylate-formation, and also displays the conserved binding site for the 4'-phosphopantetheinyl prosthetic group [27]. The nikP1 mutants accumulate nikkomycins Cx and Cz ; HPHT is not detected in culture filtrates. NikP1 is predicted to activate the nucleoside moieties, nikkomycin Cx and Cz , and catalyze peptide bond formation between them and HPHT attached to the acyl carrier domain of NikT. The nikP2 gene product resembles thioesterases required for the termination step in non-ribosomal peptide synthesis, the cyclization reaction in the biosynthesis of cyclic peptides, or the removal of the synthesized peptide from the peptide synthetase. The nikP2 mutants do not have a noteworthy phenotype; they synthesize the entire wild type nikkomycin spectrum. This finding indicates that another cytoplasmic thioesterase might substitute for NikP2 in nikkomycin synthesis. The mechanism of the formation of the second peptide bond between carbon-6' and glutamic acid in the tripeptidyl nikkomycins J and I is still a mystery since there are no more genes available within the nik cluster to encode for a peptide synthetase.
5.5.4 Export The protein deduced from the nikN gene displays limited sequence similarity to some putative membrane transport proteins. The finding that NikN is predicted to have six transmembrane helices and a binding site for ATP/GTP suggest that the nikN gene might be involved in the export of di- and tripeptidyl nikkomycins.
119
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901
5.6 Transcriptional organization and regulation of the nik cluster
The 29-kb nik cluster contains three polycistronic operons and the monocistronically transcribed regulatory gene orfR (Fig. 5.5) [19, 31, 34, 37]. Transcriptional analyses have revealed that transcripts of the nikA-nikG genes, the nikI-nikO genes, and the nikP2-nikV genes first become evident at the beginning of the second rapid growth phase (Fig. 5.4), about five hours before nikkomycins become detectable in the culture broth, while transcription of the orfR gene is initiated approximately two hours earlier. Levels of all transcripts increase markedly during the transition to the stationary phase, and high levels are maintained during the stationary phase. Transcription of the nikA-nikG, the nikInikO, and the nikP1-nikV operons is activated by the orfR-encoded protein, as insertion of the kanamycin cassette in the 5'-half of the gene abolishes transcription of the three operons in these mutants. Heptameric repeat sequences detected in the promoter regions of these operons provide potential binding sites for OrfR. OrfR has a molecular mass of 114,701 Da (1062 amino acids), deduced from the nucleotide sequence and determined by SDS-PAGE of recombinant protein expressed in E. coli [37]. OrfR has a potential ATP/GTP-binding site at position 344 to 351, and its N-terminal half (amino acid residues 1 to 275) shows a high level of sequence similarity to members of the family of Streptomyces antibiotic regulatory proteins (SARPs), which are characterized by an OmpRlike DNA-binding fold at their N-terminus [38]. This family includes ActII-Orf4 (225 amino acids) from S. coelicolor A3(2) and DnrI (272 amino acids) from S. peucetius, pathway-specific regulatory proteins that activate transcription of actinorhodin and daunorubicin biosynthesis genes, respectively. The AfsR protein (993 amino acids), a pleiotropic regulatory protein from S. coelicolor A3(2) that effects several biosynthetic pathways, is the only member of the SARPs that has a large size similar to that of OrfR. In addition, AfsR contains a P-loop and has been shown to be phosphorylated by the afsK gene product, a serine/threonine protein kinase. The open question is which signal triggers expression of the orfR gene. In S. tendae Tü 901/8 c, the switch of metabolism to nikkomycin biosynthesis appears to take place during the short phase of reduced growth between the two exponential growth phases.
120
References
References
1. U. Dähn, H. Hagenmaier, H. Höhne, W. A. König, and G. Wolf (1976) Stoffwechselprodukte von Mikroorganismen. 154. Mitteilung. Nikkomycin, ein neuer Hemmstoff der Chitinsynthese bei Pilzen. Arch. Microbiol. 107, 143–160. 2. G. U. Brillinger (1979) Metabolic products of microorganisms. 181. Chitin synthetase from fungi, a test model for substances with insecticidal properties. Arch. Microbiol. 121, 71–74. 3. G. Zoebelein and U. Kniehase (1985) Labor-, Gewächshaus- und Freilanduntersuchungen von Nikkomycinen gegen Insekten und Milben. Pflanzenschutz-Nachrichten Bayer 38, 203–304. 4. R. F. Hector, B. L. Zimmer, and D. Pappagianis (1990) Positive interaction of nikkomycins and azoles against Candida albicans in vitro and in vivo. Antimicrob. Agents Chemother. 34, 587–593. 5. K. Isono and S. Suzuki (1979) The polyoxins, pyrimidine nucleoside peptide antibiotics inhibiting fungal cell wall biosynthesis. Heterocycles 13, 333–351. 6. C. Bormann, S. Mattern, H. Schrempf, H.-P. Fiedler, and H. Zähner (1989) Isolation of Streptomyces tendae mutants with an altered nikkomycin spectrum. J. Antibiot. 42, 913–918. 7. K. Isono, T. Sato, K. Hirasawa, S. Funayama, and S. Suzuki (1978) Biosynthesis of the nucleoside skeleton of polyoxins. Am. Chem. Soc. 100, 3937–3939. 8. T. Schüz, H.-P. Fiedler, H. Zähner, H. Rieck, and W. A. König (1992) Metabolic products of microorganisms. 263. Nikkomycins Sz, Sx, Soz and Sox, new intermediates associated to the nikkomycin biosynthesis of Streptomyces tendae. J. Antibiot. 45, 199–206. 9. R.-M. Schmidt, H. Pape, and M. Junack (1986) Biosynthesis of 4-formyl-4-imidazoline-2-on, the heterocyclic base of nikkomycin X. Naturforsch. 41 c, 135–140. 10. U. Sommer (1983) Untersuchungen zur Biosynthese von Nikkomycin D. PhD thesis, Universität Münster. 11. O. Kellner (1987) Biosynthese von Hydroxypyridylhomothreonin – Aminosäure der Nikkomycine. PhD thesis, Universität Münster. 12. W. A. König, H. Hahn, R. Rathmann, W. Hass, A. Keckeisen, H. Hagenmaier, C. Bormann, W. Dehler, R. Kurth, and H. Zähner (1986) Drei neue Aminosäuren aus dem Nikkomycin-Komplex – Strukturaufklärung und Synthese. Liebigs Ann. Chem. 1986, 407–421. 13. I. Hinterlang (1992) Isolierung, Strukturaufklärung, und Stoffwechsel eines Picolinsäuremetaboliten aus Streptomyces tendae. PhD thesis, Universität Münster. 14. C. Bormann, K. Aberle, H.-P. Fiedler, and H. Schrempf (1990) Genetic complementation of Streptomyces tendae deficient in nikkomycin production. Appl. Microbiol. Biotechnol. 32, 424–430. 15. V. Möhrle, U. Roos, and C. Bormann (1995) Identification of cellular proteins involved in nikkomycin production in Streptomyces tendae Tü901. Mol. Microbiol. 15, 561–571. 16. M. A. Richardson, S. Kuhstoss, P. Solenberg, N. A. Schaus, and R. N. Rao (1987) A new shuttle cosmid vector, pKC505, for streptomycetes: its use in the cloning of three different spiramycin-resistance genes from a Streptomyces ambofaciens library. Gene 61, 231–241. 17. C. Bormann, V. Möhrle, and C. Bruntner (1996) Cloning and heterologous expression of the entire set of structural genes for nikkomycin synthesis from Streptomyces tendae Tü 901 in Streptomyces lividans. J. Bacteriol. 178, 1216–1218. 18. M. J. Bibb (1986) Gene expression in Streptomyces: Nucleotide sequences involved in initiation of transcription and translation. In: G. Szabó, S. Biró, and M. Goodfellow
121
5 Genetics of Nikkomycin Production in Streptomyces tendae Tü 901
19.
20.
21.
22. 23.
24.
25.
26.
27.
28.
29. 30.
31.
32.
33.
34.
(ed.): Biological, biochemical and biomedical aspects of Actinomycetes. Part A, 25–34. Akademiai Kiadó, Budapest, Hungary. C. Bruntner, B. Lauer, W. Schwarz, V. Möhrle, and C. Bormann (1999) Molecular characterization of co-transcribed genes from Streptomyces tendae Tü901 involved in the biosynthesis of the peptidyl moiety of the peptidyl nucleoside antibiotic nikkomycin. Mol. Gen. Genet. 262, 90–101. C. Bruntner and C. Bormann (1998) The Streptomyces tendae Tü901 l-lysine 2-aminotransferase catalyzes the initial reaction in nikkomycin D biosynthesis. Eur. J. Biochem. 254, 347–355. J. S. Thorson, F. L. Stanley, H. Liu, C. R. Hutchinson (1993) Biosynthesis of 3,6-dideoxyhexoses: new mechanistic reflections upon 2,6-dideoxy, 4,6-dideoxy, and amino sugar construction. J. Am. Chem. Soc. 115, 6993–6994. W. Piepersberg (1994) Pathway engineering in secondary metabolite-producing actinomycetes. Crit. Rev. Biotechnol. 14, 251–285. B. E. Reuber, C. Karl, S. A. Reimann, and S. J. Mihalik (1997) Cloning and functional expression of a mammalian gene for a peroxisomal sarcosine oxidase. J. Biol. Chem. 272, 6766–6776. Y. Nishiya and T. Imanaka (1996) Analysis of interaction between the Arthrobacter sarcosine oxidase and the coenzmye flavin adenine dinucleotide by site-directed mutagenesis. Appl. Environ. Microbiol. 62, 2405–2410. V. De Crécy-Lagard, V. Blanc, P. Gil, L. Naudin, S. Lorenzon, A. Famechon, N. Bamas-Jacques, J. Crouzet, and D. Thibaut (1997) Pristinamycin I biosynthesis in Streptomyces pristinaespiralis: molecular characterization of the first two structural peptide synthetase genes. J. Bacteriol. 179, 705–713. F. W. Rusnak, S. Faraci, and C. T. Walsh (1989) Subcloning, expression and purification of the enterobactin biosynthetic enzymes 2,3-dihydroxybenzoate-AMP ligase: demonstration of enzymes-bound (2,3-dihydroxybenzoyl)adenylate product. Biochemistry 28, 6827–6835. T. Stachelhaus and M. A. Marahiel (1995) Modular structure of genes encoding multifunctional peptide synthetases required for non-ribosomal peptide synthesis. FEMS Microbiol. Lett. 125, 3–14. R. K. Wierenga, P. Terpstra, and W. G. J. Hol (1986) Prediction of the occurrence of the ADP-binding bab-fold in proteins using amino acid sequence fingerprint. J. Mol. Biol. 187, 101–107. J. Powlowski and V. Shingler (1994) Genetics and biochemistry of phenol degradation by Pseudomonas sp. CF600. Biodegradation 5, 219–236. S. Aemprapa and P. A. Williams (1998) Implications of the xylQ gene of the TOL plasmid pWW102 for the evolution of aromatic catabolic pathways. Microbiology 144, 1387–1396. B. Lauer, R. Russwurm, W. Schwarz, A. Kálmanánczhelyi, C. Bruntner, A. Rosemeier, and C. Bormann (2001) Molecular characterization of co-transcribed genes from Streptomyces tendae Tü901 involved in the biosynthesis of the peptidyl moiety and assembly of the peptidyl nucleoside antibiotic nikkomycin. Mol. Gen. Genet. 264, 662–673. D. E. Holloway and E. N. G. Marsh (1993) Cloning and sequencing of glutamate mutase component E from Clostridium tetanomorphum. Organization of the mut genes. FEBS Letts. 317, 44–48. B. Lauer, R. Russwurm, and C. Bormann (2000) Molecular characterization of two genes from Streptomyces tendae Tü901 required for the formation of the 4-formyl-4imidazolin-2-one-containing nucleoside moiety of the peptidyl nucleoside antibiotic nikkomycin. Eur. J. Biochem. 267, 1698–1706. C. Bormann, R. Russwurm, W. Schwarz, R. Bräuning, B. Lauer, V. Möhrle, and U. Roos (2001) Molecular characterization of co-transcribed genes from Streptomyces tendae
122
References
35. 36.
37. 38.
Tü901 required for the formation of the nucleoside moiety of the peptidyl nucleoside antibiotic nikkomycin. In preparation. B. Lauer, R. Süßmuth, D. Kaiser, G. Jung, and C. Bormann (2000) A putative enolpyruvyltransferase gene involved in nikkomycin biosynthesis. J. Antibiot. 53, 385–392. T. Kuzuyama, T. Seki, T. Dairi, T. Hidaka, and H. Seto (1995) Nucleotide sequence of fortimicin KL1 methyltransferase gene isolated from Micromonospora olivasterospora, and comparison of its deduced amino acid sequence with those of methyltransferases involved in the biosynthesis of bialaphos and fosfomycin. J. Antibiot. 48, 1191–1193. W. Schwarz and C. Bormann (2001) Transcriptional regulation of the nikkomycin biosynthesis gene cluster by a pathway-specific regulatory gene. In preparation. A. Wietzorrek and M. Bibb (1997) A novel family of proteins that regulates antibiotic production in streptomycetes appears to contain an OmpR-like DNA-binding fold. Mol. Microbiol. 25, 1177–1184.
123
6 Glycosylated Antibiotics: Studies on Genes Involved in Deoxysugar Formation, Modification and Attachment, and their Use in Combinatorial Biosynthesis Andreas Bechthold*
6.1 Introduction
6.1.1 Combinatorial biosynthesis Advances in recombinant DNA technology have enabled the cloning and expression of many biosynthetic gene clusters from different biological sources. Gene clusters encoding proteins catalyzing the biosynthesis of different natural bioactive products have been isolated and characterized from plants, fungi and bacteria. From the first isolation of antibiotic biosynthetic genes from a Streptomycete [1–3] and the cloning and expression of entire biosynthetic pathways of antibiotic producing Streptomycetes [4, 5], a lot of progress has been achieved. Functions have been assigned to many gene products in the biosynthesis of different bioactive compounds. More recently, researchers in this area including molecular biologists and bioorganic chemists have started to use these different genes to alter the structure of natural compounds by genetic engineering or to combine genes from different biosynthetic pathways. This new technology named combinatorial biosynthesis [6, 7] resulted in the formation of novel natural products.
6.1.2 Glycosylated antibiotics Deoxysugars are an important class of carbohydrates occurring in plants, fungi and microorganisms. In microorganisms deoxygenated sugars can be found as elements of lipopolysaccharides (LPS), extracellular polysaccharides (EPS) and antibiotics.
* Pharmazeutische Biologie, Universität Freiburg, Stefan-Meier-Str. 9, D-79104 Freiburg
124 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
6.1 Introduction Traditionally, the sugar moieties of antibiotics have been viewed as molecular elements that control the pharmacokinetics of a drug. This view is beginning to change. Recent results clearly show that deoxygenated sugars actively contribute as recognition elements to the mechanism of action of the respective drug [8, 9]. In 1994 the author of this manuscript initiated investigations on the genetic analysis of avilamycin, landomycin and urdamycin biosynthesis and continued a cooperation on granaticin biosynthesis. Our research was motivated by the believe that genes, especially glycosyltransferase genes, are very important tools for combinatorial biosynthesis.
6.1.2.1 Avilamycin The avilamycins which are produced by Streptomyces viridochromogenes Tü 57 (S. viridochromogenes Tü 57) are oligosaccharide antibiotics and belong to the orthosomycin group of antibiotics [10, 11]. Avilamycins as well as other important members of the orthosomycins contain a dichloroisoeverninic acid moiety, as well as one or more orthoester linkages which are associated with carbohydrate residues. Components of the saccharide side chain are D-olivose, 2-deoxyD-evalose, 4-O-methyl-D-fucose, 2,6-Di-O-methyl-D-mannose, L-lyxose and eurekanic acid. Main product of S. viridochromogenes Tü 57 is avilamycin A (Fig. 6.1). The compound SCH27899, another orthosomycin, shows excellent activity against Gram-positive bacteria and is presently being tested for its possible use against human infectious diseases [12, 13].
6.1.2.2 Landomycin S. cyanogenus S136 is the producer of landomycins, consisting of a benz[a]anthraquinone-type aglycone moiety and a varying phenol-glycosidically linked oligosaccharide chain [14]. Landomycin A, the principal metabolite of S. cyanogenus, is the largest angucycline so far described, and the most active landomycin [15]. Its deoxysugar moiety is an unusual hexasaccharide consisting of four D-olivose and two L-rhodinose residues (Fig. 6.1). Besides landomycin A S. cyanogenus S136 produces the related congeners landomycin B and landomycin D. They differ from landomycin A in the length of the sugar side chain. In various tests landomycin A showed interesting antitumor activities, in particular against prostate cancer cell lines. This antitumor activity was discussed to result from interactions with DNA, for which the hexasaccharide chain plays an important role.
125
Figure 6.1:
Structure of avilamycin A, landomycin A, urdamycin A and granaticin.
6 Glycosylated Antibiotics: Studies on Genes
126
6.2 Cloning of the avilamycin, landomycin, urdamycin 6.1.2.3 Urdamycin Urdamycins are produced by S. fradiae. Like landomycins they are polyketides belonging to the angucyline type of antibiotics [16]. In difference to landomycins urdamycins mostly differ in the structure of the polyketide moiety. Main product of S. fradiae is urdamycin A (Fig. 6.1), minor compounds are urdamycin B, C, D, E, F, G and H. Urdamycin A consists of aquayamycin which contains a C-glycosidically linked D-olivose and three additional O-glycosidically linked deoxysugars: two L-rhodinoses, and another D-olivose. Urdamycin A shows only some anticancer and antimicrobial activity.
6.1.2.4 Granaticin Granaticin made by S. violaceoruber Tü 22 is a member of a class of polyketide antibiotics known as benzoisochromanequinones with activity against Gram-positive bacteria and P-388 mouse leukemia [17]. Granaticin contains a D-olivose moiety attached to the polyketide moiety via two carbon-carbon bonds at C-9 and C-10 (Fig. 6.1). A second sugar, L-rhodinose, is attached to D-olivose by a conventional glycosidic bond.
6.2 Cloning of the avilamycin, landomycin, urdamycin, and granaticin biosynthetic gene clusters
6.2.1 Development of a PCR method to clone dNDP-hexose 4,6-dehydratase genes In the past the cloning of biosynthetic gene clusters started with the shotgun cloning of random fragments of DNA from a Streptomyces strain, producing a given antibiotic. Mutants of the strain blocked at a step in the biosynthesis of the antibiotic were transformed with these fragments. Fragments restoring antibiotic production were used for further investigations. In another approach resistance genes of antibiotic producers were cloned in a heterologous host. This led to the identification of linked biosynthetic genes. More recently genes with known functions were used as probes in Southern hybridization to identify homologous genes in other organisms. The most popular gene probes were actI and actIII from S. coelicolor and strD from S. griseus [18, 19]. In our projects a different approach to clone the biosynthetic gene clusters of avilamycin, landomycin, urdamycin and granaticin was performed. The com127
6 Glycosylated Antibiotics: Studies on Genes parison of the sequence of known dTDP-glucose 4,6-dehydratases revealed several regions of high similarity. Based on these sequences two oligonucleotide primers were synthesized taking into account the codon usage of Actinomycetes. These primers were used to amplify DNA fragments from our Actinomycetes. PCR fragments obtained were subcloned and sequenced. The deduced amino acid sequences of the isolated fragments revealed similarity to each other and to the dNDP-glucose dehydratase (StrE) isolated from S. griseus N-2-3-11 [20].
6.2.2 Preparation and screening of cosmid libraries For the construction of gene libraries from our producer strains, DNA fragments of 25–40 kb were ligated into pOJ446. DNA was packed into phages by using the Gigapack Packaging Extract GoldTM from Stratagene (Heidelberg, Germany). The phages were used to transduce E. coli XL1-Blue-MRF '. For screening of the cosmid libraries, DNA fragments obtained by PCR amplification were used as probes. Several cosmid clones hybridizing to the probes were isolated in each case [21].
6.2.3 Identification of the biosynthetic gene clusters by gene inactivation and expression experiments Gene inactivation has been used to show that we in fact had cloned the urdamycin- and avilamycin biosynthetic gene cluster [22, 23], and expression of genes in heterologous strains was used to show that cosmid clones isolated from S. violaceoruber Tü 22 and S. cyanogenus S136 contained the entire granaticinand landomycin biosynthetic gene cluster, respectively [24, 25].
6.3 Organization of avilamycin, landomycin, urdamycin, and granaticin biosynthetic genes
6.3.1 Genes involved in avilamycin biosynthesis More than 42 kb of the avilamycin biosynthetic gene cluster have been sequenced (Fig. 6.2, Table 6.1 a) [26]. So far eight deoxysugar biosynthetic genes (aviQ1 aviQ2, aviQ3, aviD, aviE1, aviE2, aviE3, and aviJ), four glycosyltransfer128
Figure 6.2: Organization of antibiotic biosynthetic gene clusters. Genes are indicated as arrows orientated in the direction of transcription. Genes involved in deoxysugar biosynthesis, modification or in glycosyltransfer are shown as black arrows. A detailed description of the function of these genes is given in the text.
6 Glycosylated Antibiotics: Studies on Genes Table 6.1 a:
Putative functions of genes of the avilamycin biosynthetic gene cluster.
Gene
Deduced function
aviM aviH aviO1, aviO2 aviG1, aviG2, aviG3, aviG4, aviG5 aviB1, aviB2 aviD aviE1, aviE2, aviE3 aviQ1, aviQ2, aviQ3 aviJ1 aviGT1, aviGT2, aviGT3, aviGT4 aviRb, aviRa aviGTP, aviReg1, aviReg2 aviT, aviN, aviABC1+aviABC2 ?
orsellinic acid synthase (PKS) halogenase (tailoring) oxygenase (tailoring) methyltransferase (tailoring) acetoin-dehydrogenase (tailoring) dNDP-hexose synthase (sugar) dNDP-hexose-4,6-dehydratase (sugar) dNDP-hexose-4-epimerase (sugar) oxidoreductase (sugar) glycosyltransferase resistance protein regulatory protein transporter unknown function
ase genes (aviGT1, aviGT2, aviGT3, and aviGT4) and four putative sugar methyltransferase genes (aviG1, aviG2, aviG3, and aviG5) were detected. AviD and AviE1 are probably involved in the biosynthesis of the important key intermediate NDP-4-keto-6-deoxy-D-glucose, which can be converted to the 2,6-dideoxysugar components of avilamycin A, namely ring B (D-olivose), C (D-olivose) and D (D-evalose). AviQ1, AviQ2 and AviQ3 are similar to epimerases (AviQ1, after expression in E. coli, showed UDP-glucose 4-epimerase activity), AviE2 and AviE3 to NDP-D-glucose 4,6-dehydratases and AviJ to oxidoreductases. AviM is an orsellinic acid synthase and AviH seems to be involved in halogenation. AviG4 might code for a methyltransferase involved in the biosynthesis of the dichloroisoeverninic acid moiety. AviO1 and AviO2 resemble oxygenases and might catalyze the formation of the orthoesters. Further genes, aviB1, aviB2, aviRa, aviRb, aviGTP1, aviReg1, aviReg2, aviT, aviN, aviABC1, and aviABC2 are either involved in resistance, transport or regulation. The function of many other genes remains unknown [26, 27].
6.3.2 Genes involved in landomycin A biosynthesis The landomycin gene cluster from S. cyanogenus S136 contained several genes involved in deoxysugar- and polyketide biosynthesis (Fig. 6.2, Table 6.1 b). The formation of dTDP-D-glucose from glucose-1-phosphate is catalyzed by LanG which was shown after expression of lanG in E. coli. dTDP-D-glucose might then be converted to TDP-4-keto-6-deoxy-D-glucose by LanH. Deoxygenation at C-2 is probably catalyzed by LanS leading to a hypothetical intermediate which can be reduced by LanT. TDP-4-keto-2,6-dideoxy-D-glucose might be a 130
6.3 Organization of avilamycin, landomycin, urdamycin Table 6.1 b: Putative functions of genes of the landomycin, urdamycin and granaticin biosynthetic gene cluster. Gene
Deduced function
lanA, urdA, graORF-1 lanB, urdB, graORF-2 lanC, urdC, graORF-3 lan F, lanL, urdF, urdL, graORF-18, graORF-4, graORF-33 lanD, lanN, lanO, lanV, lanZ4, urdD, urdN, urdO, graORF-5, graORF-6, graORF-34 lanE, lanM, lanZ5, urdE, urdM, graORF-21, graORF-29 lanP lanG, lanZ2, urdG, graORF-16 lanH, urdH, graORF-17 lanQ, urdQ, graORF-23
b-ketoacyl-ACP synthase (PKS) chain length factor (PKS) acyl carrier protein (PKS) cyclase (PKS)
lanR, lanZ3, urdR, graORF-22 lanS, urdS, graORF-27 lanT, urdT, graORF-26
reductase (PKS)
oxygenase (tailoring) decarboxylase (PKS) dNDP-hexose-synthetase (sugar) dNDP-hexose-4,6-dehydratase (sugar) dNDP-4-keto-6-deoxy-hexose-3-dehydratase (sugar) dNDP-hexose-4-ketoreductase (sugar) dNDP-4-keto-6-deoxy-hexose-2,3-dehydratase (sugar) oxidoreductase (involved in C-2 deoxygenation) (sugar) dNDP-hexose-3,5-epimerase (sugar) glycosyltransferase
lanZ1, urdZ1, graORF-25 lanGT1, lanGT2, lanGT3, lanGT4, urdGT1 a, urdGT1 b, urdGT1 c, urdGT2, graORF-14 lanK, urdK, graORF-7, graORF-9, regulator graORF-10, graORF-11, graORF-19, graORF-20, graORF-37 lanJ, urdJ, urdJ2, graORF-15 transporter lanU, lanZ6, lanX, graORF-8, graORF-12, ? (unknown) graORF-13, graORF-24, graORF-28, graORF-30, graORF-31, graORF-32, graORF35, graORF-36
central intermediate in the biosynthesis of both TDP-L-rhodinose and TDP-Dolivose. A reduction step at C-4 (LanR) is required to complete the biosynthesis of TDP-D-olivose. 3,5-Epimerization (LanZ1), dehydroxylation at C-3 (LanQ) and reduction at C-4 (LanZ3) would lead to the formation of TDP-L-rhodinose. So far four genes (lanGT1–lanGT4) coding for putative glycosyltransferases have been detected in the biosynthetic gene cluster. Genes involved in the biosynthesis of the polyketide moiety are lanA, lanB, lanC, lanD, lanF, lanL and lanP. Further genes are either involved in tailoring reactions (lanE, lanM, lanZ5, lanN, lanO, lanV, lanZ4) or can be viewed as regulatory (lanK) or transporter (lanJ) genes (Fig. 6.2) [25, 28].
131
6 Glycosylated Antibiotics: Studies on Genes
6.3.3 Genes involved in urdamycin A biosynthesis So far eight genes (urdZ1, urdG, urdH, urdZ3, urdQ, urdR, urdS, urdT) probably involved in deoxysugar formation have been identified in the urdamycin biosynthetic gene cluster (Fig. 6.2, Table 6.1 b). The deduced amino acid sequences of these genes are very similar to the deduced amino acid sequences of genes from the landomycin cluster indicating similar functions. Four putative glycosyltransferase genes (urdGT1 a, urdGT1 b, urdGT1 c, and urdGT2), which have been detected in the cluster, are involved in the formation and attachment of the trisaccharide chain as well as the attachment of L-rhodinose at C12 b. Genes involved in the biosynthesis of the polyketide moiety or in tailoring reactions have also been detected in the urdamycin cluster. Products of these genes (urdA, urdB, urdC, urdD, urdF, urdL, urdE, urdM, urdO) are again very similar to the deduced amino acid sequences of genes from the landomycin cluster, and UrdJ and UrdJ2 resemble antibiotic transporters [26, 27, 29–31].
6.3.4 Genes involved in granaticin biosynthesis As in the landomycin and urdamycin cluster several genes involved in the formation of the polyketide and sugar moieties have been sequenced (Fig. 6.2, Table 6.1 b). The deduced amino acid sequences are very similar to proteins from the landomycin and urdamycin cluster [17, 24].
6.4 New genetically engineered natural compounds
Gene inactivation experiments by introducing in-frame deletions or frame-shift mutations into the genes and combinatorial biosynthetic approaches were carried out. New molecules were created and functions of genes could be determined.
132
6.4 New genetically engineered natural compounds
6.4.1 Avilamycin 6.4.1.1 New avilamycin derivatives through gene inactivation experiments To improve the water solubility of avilamycin A we started a project to inactivate methyltransferase genes of the avilamycin cluster. So far aviG1 and aviG4 have been deleted. Extracts of mutants were analyzed by HPLC-UV and by HPLCMS and were also analyzed for antibiotic activity (bioassay). Experimentally obtained data indicate that aviG1 is encoding a C-methyltransferase and that inactivation of aviG1 is leading to a less antibiotically active derivative. Inactivation of aviG4, which is probably involved in methylation of ring A of avilamycin A was leading to the formation of a more hydrophilic compound which showed high antimicrobial activity.
6.4.1.2 Production of orsellinic acid by expression of aviM The expression of one gene of the avilamycin biosynthetic gene cluster (aviM) in S. lividans was leading to the production of orsellinic acid (Fig. 6.3), a structural component of avilamycin A [22].
6.4.2 Landomycin and urdamycin 6.4.2.1 Accumulation of intermediates of the urdamycin biosynthesis after expression of a cosmid containing genes of the landomycin cluster Cosmid H2-26, containing genes of the landomycin cluster was transformed into the urdamycin producer S. fradiae. Transformants containing the cosmid were accumulating aquayamycin and compound 100-2. The structures of both compounds indicated that both compounds were not produced through combinatorial biosynthesis. Both compounds are likely to be intermediates of urdamycin A biosynthesis and a competition between glycosyltransferases of the urdamycinand landomycin cluster might be responsible for the accumulation of both compounds (Fig. 6.3) [25].
6.4.2.2 New compounds by inactivation of biosynthetic genes Several genes of the urdamycin cluster have been deleted by gene inactivation experiments. Mutants were accumulating new compounds instead of urdamycin A. (Table 6.2, Fig. 6.3) [29–31]. From these experiments the function to several genes can be assigned. 133
6 Glycosylated Antibiotics: Studies on Genes Table 6.2: Production of new antibiotics by gene inactivation- and gene expression experiments. Deleted gene
Gene expressed in the corresponding mutant
Products
urdGT2 urdGT2 urdGT1 a
– urdGT2 –
urdGT1 a urdGT1 b, urdInt and urdGT1 c urdGT1 b, urdInt and urdGT1 c urdGT1 a, urdGT1 b, urdInt and urdGT1 c urdGT1 a, urdGT1 b, urdInt and urdGT1 c urdGT1 a, urdGT1 b, urdInt and urdGT1 c
urdGT1 a
urdamycin I, urdamycin J, urdamycin K urdamycin A 12 b-desrhodinosyl-urdamycin C, 12 b-desrhodinosyl-urdamycin D, 12 b-desrhodinosyl-urdamycin F, urdamycinon D, urdamycin B urdamycin A
–
100-2
urdGT1 c
urdamycin G, urdamycin A
–
aquayamycin
urdGT1 c
12 b-desrhodinosyl-urdamycin G
urdGT1 b and urdGT1 c
12 b-desrhodinosyl-urdamycin G, 12 b-desrhodinosyl-urdamycin A
lanGT3
12 b-desrhodinosyl-urdamycin G
– – – – – – –
urdamycin I, urdamycin J rabelomycin aquayamycin urdamycin I, urdamycin J (no product) aquayamycin aquayamycin
urdGT1 a, urdGT1 b, urdInt and urdGT1 c urdGT1 a, urdGT1 b, urdInt, urdGT1 c and urdGT2 urdM urdR urdZ3 urdS urdZ1 urdQ
UrdGT2 must catalyze the earliest glycosyltransfer step in the urdamycin biosynthetic pathway which is the C-glycosyltransfer of an activated D-olivose. UrdGT1 a is involved in the glycosylation of aquayamycin during urdamycin A biosynthesis and UrdM is an oxygenase involved in oxygenation at position 12 b of urdamycin A.
134
135
Figure 6.3:
Structure of compounds generated during this study.
6 Glycosylated Antibiotics: Studies on Genes 6.4.2.3 New compounds by expression of single genes All glycosyltransferase genes of the urdamycin and landomycin biosynthetic gene cluster have been expressed in glycosyltransferase mutants of S. fradiae. The expression of urdGT1 c and lanGT4 (independent experiments) in a mutant, in which urdGT1 a, urdGT1 b, and urdGT1 c had been deleted and which was accumulating compound 100-2, resulted in the production of 12 b-desrhodinosyl-urdamycin G. After coexpression of urdGT1 c and urdGT1 b the accumulation of 12 b-desrhodinosyl-urdamycin A was detected (Table 6.2, Fig. 6.3). Thus UrdGT1 c is catalyzing the transfer of an activated L-rhodinose to 100–2 during urdamycin biosynthesis and UrdGT1 b is attaching the final sugar [31]. LanGT4 seems to catalyze the attachment of L-rhodinose and might act twice during landomycin biosynthesis.
References
1. Feitelson, J. S. and Hopwood, D. A. (1983) Cloning of a streptomyces gene for an Omethyltransferase involved in antibiotic biosynthesis. Mol. Gen. Genet. 190, 394–398. 2. Gil, J. A. and Hopwood, D. A. (1983) Cloning and expression of a p-aminobenzoic acid synthetase gene of the candicidin-producing Streptomyces griseus. Gene 25, 119–132. 3. Jones, G. and Hopwood, D. A. (1984) Molecular cloning and expression of the phenoxazinone synthase gene from Streptomyces antibioticus. J. Biol. Chem. 259, 2703–2714. 4. Malpartida, F. and Hopwood, D. A. (1984) Molecular cloning of the whole biosynthetic pathway of a Streptomyces antibiotic and its expression in a heterologous host. Nature 309, 462–464. 5. Motamedi, H. and Hutchinson, C. R. (1987) Cloning and heterologous expression of a gene cluster for the biosynthesis of tetracenomycin C, the anthracycline antitumor antibiotic of Streptomyces glaucescens. Proc. Natl. Acad. Sci. USA 84, 4445–4449. 6. Rohr, J. (1995) Kombinatorische Biosynthese – eine Zukunftsstrategie. Angew. Chem. 107, 963–966. 7. Khosla, C. and Zawada, R. J. X. (1996) Generation of polyketide libraries via combinatorial biosynthesis. Trends in Biotechnology 14, 335–341. 8. Kirschning, A., Bechthold, A., and Rohr, J. (1997) Chemical and biochemical aspects of deoxysugars and deoxysugar oligosaccharides. Topics Current Chemistry 188, 1–84. 9. Trefzer, A., Salas, J., and Bechthold, A. (1999) Genes and enzymes involved in deoxysugar biosynthesis. Nat. Prod. Rep. 16, 283–299. 10. Buzzetti, F., Eisenberg, F., Grant, H. N., Keller-Schierlein, W., Voser, W., and Zähner H. (1968) Avilamycin. Experientia 24(4), 320–323. 11. Mertz, J. L., Peloso, J. S., Barker, B. J., Babbitt, G. E., Occolowitz, J. L., Simson, V. L., and Kline, R. M. (1986) Isolation and structural identification of nine avilamycins. J. Antibiot. 39(7), 877–887. 12. Hoffmeister, D. and Bechthold, A. (1999) Halten wir Schritt mit den Mikroben? – Die Suche nach neuen Antibiotika. PZ Prisma 3, 185–194.
136
References 13. Urban, C., Mariano, N., Mosinka-Snipas, K., Wadee, C., Chahrour, T., and Rahal, J. J. (1996) Comparative in-vitro activity of SCH 27899, a novel everninomycin, and vancomycin. J. Antimicrob. Chemother. 37(2), 361–364. 14. Henkel, T., Rohr, J., Beale, J. M., and Schwenen, L. (1990) Landomycins. New angucycline antibiotics from Streptomyces sp. I. Structural studies on landomycins A–D. J. Antibiot. 43, 492–503. 15. Depenbrock, H., Bornschlegl, S., Peter, R., Rohr, J., Schmid, P., Schweighart, P., Block, T., Rastetter, J., and Hanauske A. R. (1996) Assessment of antitumor activity of landomycin A (NSC 6399187-A). Ann. Hematol. 73 (Suppl. II), A80, 316. 16. Drautz, H., Zähner, H., Rohr, J., and Zeeck, A. (1986) Metabolic products of microorganisms. 234. Urdamycins, new angucycline antibiotics from Streptomyces fradiae. 1. Isolation, characterization and biological properties. J. Antibiot. 39, 1657–1669. 17. Bechthold, A., Sohng, J. K., Smith, T. M., and Floss, H. G. (1995) Identification of Streptomyces violaceoruber Tü22 genes involved in the biosynthesis of granaticin. Mol. Gen. Genet. 248(5), 610–620. 18. Bechthold, A. and Salas, J. A. (1999) Combinatorial biosynthesis of microbial metabolites. In: Combinatorial organic chemistry (Jung, G. ed.), Wiley-VCH Weinheim, 381– 407. 19. Hopwood, D. A. (1997) Genetic contribution to understanding polyketide synthases. Chem. Rev. 97, 2465–2497. 20. Decker, H., Gaisser, S., Pelzer, S., Schneider, P., Westrich, L., Wohlleben, W., and Bechthold, A. (1996) A general approach for cloning and characterization of dNDPglucose dehydratase genes from actinomycetes. FEMS Microbiol. Lett. 141, 195–201. 21. Westrich, L., Gaisser, S., Reichenstein, B., and Bechthold, A. (1997) Preparation and screening of cosmid libraries from Streptomyces strains using the rapid DNA ligation kit, digoxigenin labeling and detection kit from Boehringer Mannheim. Biochemica 101, 30–32. 22. Gaisser, S., Trefzer, A., Stockert, S., Kirschning, A., and Bechthold, A. (1997) Cloning of an avilamycin biosynthetic gene cluster from Streptomyces viridochromogenes Tü57. J. Bacteriol. 179, 6271–6278. 23. Decker, H. and Haag, S. (1995) Cloning and characterization of a polyketide synthase gene from Streptomyces fradiae Tü2717, which carries the genes for biosynthesis of the angucycline antibiotic urdamycin A and a gene probably involved in its oxygenation. J. Bacteriol. 177, 6126–6136. 24. Ichinose, K., Bedford, D. J., Tornus, D., Bechthold, A., Bibb, M. J., Revill, W. P., Floss, H. G., and Hopwood, D. A. (1998) The granaticin biosynthetic gene cluster of Streptomyces violaceoruber Tü22: sequence analysis and expression in a heterologous host. Chemistry & Biology 5 (11), 647–659. 25. Westrich, L., Domann, S., Faust, B., Bedford, D., Hopwood, D. A., and Bechthold, A. (1999) Cloning and characterization of the landomycin biosynthetic gene cluster of Streptomyces cyanogenus S136. FEMS Microbiol. Lett. 170, 381–387. 26. Bechthold, A., Domann, S., Faust, B., Hoffmeister, D., Stockert, S., Trefzer, A., Weitnauer, G., and Westrich, L. (1999) Glycosidierte Naturstoffe, Perspektiven für die Kombinatorische Biosynthese. Chemotherapie Journal 4, 130–135. 27. Bechthold, A. and Rohr, J. (1999) Oligosaccharide antibiotics, perspective for combinatorial biosynthesis. In: New aspects in Bioorganic Chemistry (Diederichsen, U., Lindhorst, Th. K., Wessjohann, L., Westermann, B., ed.), Wiley-VCH Weinheim, pp 313– 321. 28. Wohlert, S., Bechthold, A., Beninga, C., Henkel, T., Holzenkämpfer, M., Kirschning, A., Oelkers, C., Weber, S., Weißbach, U., Westrich, L., and Rohr, J. (1999) Investigation on the biosynthesis of landomycin A. In: New aspects in Bioorganic Chemistry (Diederichsen, U., Lindhorst, Th.K., Wessjohann, L. and Westermann, B., ed), Wiley-VCH Weinheim, pp. 305–312.
137
6 Glycosylated Antibiotics: Studies on Genes 29. Künzel, E., Faust, B., Oelkers, C., Bearden, D., Westrich, L., Bechthold, A., and Rohr, J. (1999) The urdGt2 gene encodes a glycosyltransferase responsible for the Cglycosyltransfer of activated D-olivose, the earliest glycosyl transfer step in the urdamycin biosynthesis. J. Am. Chem. Soc., accepted for publication. 30. Faust, B., Hoffmeister, D., Weitnauer, G., Westrich, L., Haag, S., Schneider, P., Decker, H., Künzel, E., Rohr, J., and Bechthold, A. (1999) Two new tailoring enzymes, a glycosyltransferase and an oxygenase, involved in biosynthesis of the angucycline antibiotic urdamycin A in Streptomyces fradiae. Microbiology, accepted for publication. 31. Trefzer, A., Hoffmeister, D., Westrich, L., Weitnauer, G., Stockert, S., Künzel, E., Rohr, J., Fuchser, J., Bindseil, K., and Bechthold, A.(2000) Function of glycosyltransferase genes involved in the biosynthesis of urdamycin A. Chemistry & Biology 7, 13–142.
138
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics: Basis for Creating New Structures by Combinatorial Biosynthesis Stefan Pelzer and Wolfgang Wohlleben*
7.1 Introduction
Together with gentamicin the glycopeptide antibiotic vancomycin (Fig. 7.1) and the structurally related lipoglycopeptide antibiotic teicoplanin are considered to be the last lines of defence against a variety of serious infections caused by Gram-positive organisms, such as enterococci, methicillin-resistant Staphylococcus aureus (MRSA), and Clostridium difficile [1]. Their annual turnover amounts to about one billion US$ [2], which demonstrates their high social and economic value. Glycopeptide antibiotics inhibit the synthesis of the cell wall of Gram-positive organisms. Their primary molecular target is the D-alanyl-D-alanine (D-AlaD-Ala) terminus of growing peptidoglycan. Dimers of vancomycin bind specifically to D-Ala-D-Ala inhibiting simultaneously the two key-steps of the peptidoglycan biosynthesis, the transglycosylase and the transpeptidase reaction [2]. In the last ten years numerous vancomycin resistant enterococci (VRE) emerged in clinical isolates [3] and spread rapidly. These organisms avoid their cell deaths by modifying the drug’s target, specifically by modifying it to the depsipeptide D-alanyl-D-lactat (D-Ala-D-Lac). The affinity of vancomycin to DAla-D-Lac is reduced significantly (factor 1000). The fear that this high-level resistance may eventually spread to methicillin-resistant Staphylococcus aureus (MRSA) has prompted the search for new or modified glycopeptide antibiotics. One way to generate new glycopeptide antibiotics is the manipulation of biosynthetic genes by combinatorial biosynthesis. In order to employ this approach successfully, corresponding gene clusters have to be identified and analysed. Until now only putative genes of the vancomycin producer A. orientalis
* Mikrobiologie/Biotechnologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
139 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics
Figure 7.1: Chemical structures of the glycopeptide antibiotics balhimycin, vancomycin and chloroeremomycin [17].
140
7.2 Results C329.4 [4] and the chloroeremomycin producer A. orientalis A82846 were described [5]. However, the role of the sequenced genes was not elucidated. A glycopeptide with similar properties as vancomycin is balhimycin, which is produced by Amycolatopsis mediterranei DSM5908 [6]. Balhimycin shares the heptapeptide core of vancomycin and differs only in the glycosylation pattern (Fig. 7.1). In the following we present the establishment of a transformation protocol and the construction of vectors for manipulation of the balhimycin producer, the isolation of the balhimycin gene cluster, and the functional analyses of different biosynthetic genes.
7.2 Results
7.2.1 Establishment of a gene cloning system for the balhimycin producer Amycolatopsis mediterranei DSM5908 Although transformation techniques were described for the vancomycin producer A. orientalis ATCC 19795 [7] and the producer of the teicoplanin-like antibiotic A47934, S. toyocaensis [8], the application for generating glycopeptide antibiotic mutants was never reported. For the establishment of a gene cloning system for the balhimycin producer A. mediterranei DSM5908 a transformation method was developed first. In order to monitor successful transformation events, DNA of the Amycolatopsis phage W2 [9], which was able to infect the balhimycin producer was employed. Using W2-DNA in transfection experiments, the three different methods described for the transformation of Amycolatopsis strains were tested for their applicability in A. mediterranei DSM5908. Neither various modifications of the PEG induced protoplast transformation method, which was reported as an efficient method for A. orientalis [8], nor electroporation methods described for transformation of A. mediterranei and A. orientalis [10] enabled introduction of W2-phage DNA into A. mediterranei DSM5908. Only after modifying the direct transformation method [11], which was previously described for the rifamycin producer A. mediterranei [12] and for A. methanolica [13], transformants of the balhimycin producer were obtained. Using the optimised protocol it was not possible to establish any of the described Amycolatopsis vectors or broad host range Streptomyces vectors [11]. Therefore, novel non-replicative vectors for integration mutagenesis were developed. Since the balhimycin producer showed a naturally high resistance against the customary actinomycetes resistance markers thiostrepton and kanamycin, the erythromycin-resistance gene ermE [14] and the chloramphenicol-resistance gene cat [15] were used to construct the vectors pSP1 and pSP2 [11]. 141
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics The application of the vectors in gene disruption, gene replacement and in-frame deletion experiments revealed, that the integration frequencies were influenced by several parameters [11]: Highest integration frequencies were observed when the DNA was isolated from the dam/dcm E. coli strain JM110 and PEG 1000 from NBS Biologicals was employed. The efficiency of integration depended directly on the size of the cloned insert. Plasmids with fragments smaller than 1 kb were difficult to integrate. In gene replacement experiments a high double crossover rate (31%) was demonstrated.
7.2.2 Identification of biosynthetic genes by reverse genetic approaches The application of standard strategies for isolation of antibiotic biosynthetic genes like cloning of resistance genes or complementation of non-producing mutants was not possible. Therefore, different reverse genetic approaches were chosen considering characteristic features of glycopeptide antibiotics like the heptapeptide backbone or the sugar residues.
7.2.2.1 Isolation of different peptide synthetase genes using conserved oligonucleotide probes Since the central core heptapeptide of balhimycin includes five non-proteinogenic amino acids, a non-ribosomal synthesis by peptide synthetases is most likely. Peptide synthetases of bacterial and fungal origins are highly conserved and show a modular organisation. Each module activates the cognate amino acid and incorporates it in the peptide product by a stepwise elongation [16]. Comparison of the amino acid sequences of peptide synthetase modules from various microorganisms showed five strongly conserved regions [11]. These conserved motives were used to design oligonucleotide probes, which were adapted to the Streptomyces codon usage. Using these gene probes in plaque hybridisation experiments we identified two different clusters of peptide synthetase genes in a l library of A. mediterranei. Gene disruption experiments revealed, that both different peptide synthetase gene clusters are not involved in balhimycin production [8]. Today we know that genes encoding peptide synthetases are common in Actinomycetes and that the balhimycin producer carries at least four different peptide synthetase gene clusters, two of them with unknown function.
142
7.2 Results 7.2.2.2 Identification of a glycosyltransferase gene fragment involved in balhimycin biosynthesis Another reverse genetic approach considered, that balhimycin includes two sugar groups, a glucose and the rare aminosugar dehydrovancosamine (Fig. 7.1). Such sugars are normally attached to the aglycon by the action of glycosyltransferases. To design PCR primers for the amplification of balhimycin-specific glycosyltransferase gene fragments, sequences of glycosyltransferases probably involved in the biosynthesis of vancomycin and chloroeremomycin [4] were compared and two conserved regions were chosen. Using conserved oligonucleotides we were able to amplify a 900 bp DNA fragment (“bgtfB”) from genomic DNA of the balhimycin producer in an optimised PCR-protocol [17]. DNA sequencing and comparison of the deduced gene product showed significant similarities to various glycosyltransferases. To prove that the glycosyltransferase gene is involved in balhimycin biosynthesis, a gene disruption experiment was performed. An internal fragment of the bgtfB gene was cloned into the gene disruption vector pSP1. This vector was used to inactivate the chromosomal bgtfB gene by homologous recombination (Fig. 7.2). In bioassays, the mutant (HD1) was able to synthesise a substance which was still antibiotically active. HPLC-MS and MS-MS analyses showed that HD1 produced at least four different compounds, which were not glycosylated anymore [17, 18]. Thus the first glycopeptide biosynthetic mutant was created by a targeted approach and the role of bgtfB in balhimycin biosynthesis was proved.
7.2.3 Analysis of the balhimycin biosynthetic gene cluster Since antibiotic biosynthetic genes including resistance and regulatory genes are normally physically linked in the genome of Streptomycetes [19] isolation of the whole balhimycin gene cluster was intended by using a gene fragment of the glycosyltransferase bgtfB as probe.
7.2.3.1 Isolation and sequencing of the major part of the balhimycin biosynthetic gene cluster Hybridisation experiments with the glycosyltransferase PCR fragment (“bgtfB”) led to the identification of twelve cosmids in an A. mediterranei cosmid library. By DNA sequence, one cosmid (16.1) was characterised in detail (manuscript in preparation). An ORF analysis revealed, that a region of 46 kb carries 26 complete and 1 incomplete putative ORFs (Fig. 7.3). Computer-assisted homology searches with the sequences of the deduced gene products showed significant 143
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics
Figure 7.2: Inactivation of the glycosyltransferase gene bgtfB by gene disruption resulting in mutant strain HD1. HPLC and MS-MS analyses showed that HD1 produced four different balhimycin derivatives which are all not glycosylated [17]. The chemical structure of one derivative (HD-1142) is shown in detail.
similarities to many gene products, which may participate in balhimycin biosynthesis like peptide synthetases, glycosyltransferases, cytochrome P450 mono-oxygenases, sugar biosynthetic enzymes, and an enzyme similar to chalcone/stilbene synthases of plants. The organisation of the major part of the balhimycin gene cluster is very similar to the organisation described for the putative chloroeremomycin gene cluster [5]. Differences between the two clusters are on one hand the presence 144
7.2 Results
Figure 7.3: Organisation of the major part of the balhimycin biosynthetic gene cluster and the proposed function of the encoding ORFs.
of an additional ORF (ORF 7, Fig. 7.3) in the balhimycin cluster showing similarities to Na/H antiporter. On the other hand, ORF 9 in the balhimycin cluster (Fig. 7.3), showing significant similarities to NDP-hexose-4-ketoreductases carries an in-frame deletion of 226 amino acids in comparison with the corresponding intact ORF of the chloroeremomycin cluster (manuscript in preparation). This corresponds to the occurrence of the keto-sugar residue dehydrovancosamine in balhimycin and the reduced 4-epi-vancosamine in chloroeremomycin, respectively (Fig. 7.1).
145
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics 7.2.3.2 Characterisation of the function of different balhimycin biosynthetic genes by gene inactivation experiments Cytochrome P450 mono-oxygenases are involved in the coupling of the aromatic side chains of the balhimycin heptapeptide In the major part of the balhimycin cluster four ORFs (OxyA–D, Fig. 7.3) with remarkable similarities to cytochrome P450 mono-oxygenase were identified. Integration of a plasmid between oxyA and oxyB resulted in the balhimycin mutant SP1–1, which was not able to synthesise an antibiotically active compound [17]. Structural analyses by HPLC-MS, fragmentation studies, and amino acid analyses demonstrated, that SP1–1 produced two balhimycin precursors (SP969 and SP1134; Fig. 7.4), which were linear (unbridged) and not glycosylated [18, 20]. These results showed clearly that oxygenases are involved in the coupling of the aromatic side chains, which are essential to hold the peptide backbone in a rigid, cup-shaped form. This structure is important for the antibiotic activity, because only this form interacts with the D-Ala-D-Ala terminus of the cell wall precursor [2]. The halogenase BhaA is responsible for chlorination of balhimycin The product of the bhaA gene (Fig. 7.3) showed significant similarities (42%) to PrnC, a new type of non-heme halogenase of Pseudomonas fluorescens which is involved in chlorination of pyrrolnitrin [21]. Since balhimycin carries two chlorine atoms at the aromatic side chains of amino acid two and six, the participation of BhaA in chlorination of balhimycin is plausible. To prove this, a bhaA in-frame deletion-mutant (PH4) was constructed using a two-step strategy [22]. HPLC-MS analyses and the analysis of the isotopic pattern clearly showed, that the mutant PH4 was not able to synthesise chlorinated balhimycin-derivatives (Fig. 7.5). Thus bhaA is responsible for the incorporation of both chlorine atoms into the glycopeptide antibiotic. Bioassays revealed that dechlorobalhimycin is still antibiotically active (manuscript in preparation). The non-ribosomal biosynthesis of the balhimycin backbone: Eight peptide synthetase modules are involved in the biosynthesis of a heptapeptide For the non-ribosomal biosynthesis of a heptapeptide seven modules were expected. Until now, three peptide synthetase genes (bpsB–D) encoding five peptide synthetase modules were characterised in the balhimycin biosynthetic gene cluster (manuscript in preparation, Fig. 7.3). From comparison with the putative chloroeremomycin biosynthetic gene cluster [5], we conclude that a fourth peptide synthetase gene (“bpsA”) encoding three further modules is located in front of the peptide synthetase bpsB. Surprisingly, altogether eight peptide synthetase modules are expected to be located within the balhimycin biosynthetic gene cluster. 146
7.2 Results
Figure 7.4: Construction of mutant strain SP1–1 by integration mutagenesis between the cytochrome P450 mono-oxygenase genes oxyA and oxyB. HPLC-MS and MS-MS revealed that the mutant produced a linear balhimycin precursor (SP-1134) which is antibiotically inactive showing that oxygenases are involved in coupling of the aromatic rings [17, 18].
The function of three peptide synthetase genes (bpsB–D) was investigated by gene inactivation experiments and by analysing the enzymatic activities of overexpressed peptide synthetase modules (manuscript in preparation). All mutants were unable to synthesise an antibiotically active compound. The successful complementation of the bpsD replacement-mutant, encoding the eighth module, with a copy of the native bpsD gene showed that polar effects of the gene replacement could be excluded. 147
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics
Figure 7.5: Construction of the mutant PH4 which carries an in-frame deletion in the halogenase gene bhaA. HPLC-MS analysis and determination of the isotopic pattern revealed that the mutant produces the antibiotically active compound dechloro-balhimycin indicating that bhaA is responsible for the chlorinating steps [21].
The results of the mutational analyses were confirmed by biochemical experiments showing that the modules 4, 5 and 7 activated the expected amino acid (manuscript in preparation). Moreover, these experiments demonstrated clearly that also the eighth module is involved in the biosynthesis of the heptapeptide core. 148
References
7.3 Outlook
The balhimycin biosynthesis includes further interesting biosynthetic steps. First experiments showed that the css gene, encoding a gene product similar to plant chalcone/stilbene synthases, is presumably involved in the biosynthesis of the seventh amino acid 3,5-dihydroxyphenyl-glycine. Furthermore, the balhimycin producer and the identified genes are suitable tools for the generation of hybrid antibiotics by rational combinatorial biosynthesis.
References
1. Cunha, B. A. (1995) Vancomycin. Med. Clin. N. Am. 79, 817–831. 2. Williams, D. H. and Bardsley, B. (1999) Die Vancomycin-Antibiotika und der Kampf gegen resistente Bakterien. Angew. Chem. 111, 1264–1286. 3. Arthur, M., Depardieu, F., Reynolds, P., and Courvalin, P. (1996) Quantitative analysis of the metabolism of soluble cytoplasmic peptidoglycan precursors of glycopeptideresistant enterococci. Mol. Microbiol. 21, 33–44. 4. Solenberg, P. J., Matsushima, P., Stack, D. R., Wilkie, S. C., Thompson, R. C., and Baltz, R. H. (1997) Production of hybrid glycopeptide antibiotics in vitro and in Streptomyces toyocaensis. Chem. Biol. 4, 195–202. 5. Van Wageningen, A. M. A., Kirkpatrick, P. N., Williams, D. H., Harris, B. R., Kershaw, J. K., Lennard, N. J., Jones, M., Jones, S. J. M., and Solenberg, P. J. (1998) Sequencing and analysis of genes involved in the biosynthesis of a vancomycin group antibiotic. Chem. Biol. 5, 155–162. 6. Nadkarni, S. R., Patel, M. V., Chaterjee, S., Vijayakumar, E. K. S., Desikan, K. R., Blumbach, J., and Ganguli, B. N. (1994) Balhimycin, a new glycopeptide antibiotic produced by Amycolatopsis sp. Y-86,21022. J. Antibiot. 47, 334–341. 7. Matsushima, P., McHenney, M., and Baltz, R. H. (1987) Efficient transformation of Amycolatopsis orientalis (Nocardia orientalis) protoplasts by Streptomyces plasmids. J. Bacteriol. 169, 2298–2300. 8. Matsushima, P. and Baltz, R. H. (1996) A gene cloning system for Streptomyces toyocaensis. Microbiology 142, 261–267. 9. Lechevalier, M. P., Prauser, H., Labeda, D. P., and Ruan, J.-S. (1986) Two new genera of nocardioform actinomycetes: Amycolata gen. nov. and Amycolatopsis gen. nov. Int. J. Syst. Bacteriol. 36, 29–37. 10. Lal, R., Lal, S., Grund, E., and Eichenlaub, R. (1991) Construction of a hybrid plasmid capable of replication in Amycolatopsis mediterranei. Appl. Environ. Microbiol. 57, 665–671. 11. Pelzer, S., Reichert, W., Huppert, M., Heckmann, D., and Wohlleben, W. (1997) Cloning and analysis of a peptide synthetase gene of the balhimycin producer Amycolatopsis mediterranei DSM5908 and development of a gene disruption/replacement system. J. Biotechnol. 56, 115–128.
149
7 Analysis of the Biosynthesis of Glycopeptide Antibiotics 12. Madon, J. and Hütter, R. (1991) Transformation system for Amycolatopsis (Nocardia) mediterranei: Direct transformation of mycelium with plasmid DNA. J. Bacteriol. 173, 6325–6331. 13. Vrijbloed, J. W., Madon, J., and Dijkhuizen, L. (1995) Transformation of the methylotrophic actinomycete Amycolatopsis methanolica with plasmid DNA: Stimulatory effect of a pMEA300-encoded gene. Plasmid 34, 96–104. 14. Uchiyama, H. and Weisblum, B. (1985) N-methyl transferase of Streptomyces erythraeus that confers resistance to the macrolide-lincosamide-streptogramin B antibiotics: Amino acid sequence and its homology to cognate R-factor enzymes from pathogenic bacilli and cocci. Gene 38, 103–110. 15. Gil, J. A., Kieser, H. M., and Hopwood, D. A. (1985) Cloning of a chloramphenicol acetyltransferase gene of Streptomyces acrimycini and its expression in Streptomyces and Escherichia coli. Gene 38, 1–8. 16. Konz, D. and Marahiel, M. A. (1999) How do peptide synthetases generate structural diversity? Chem. Biol. 6, R39–R48. 17. Pelzer, S., Süßmuth, R., Heckmann, D., Recktenwald, J., Huber, P., Jung, G., and Wohlleben, W. (1999) Identification and analysis of the balhimycin biosynthetic gene cluster and its use for manipulating glycopeptide biosynthesis in Amycolatopsis mediterranei DSM5908. Antimicrob. Agents Chemother. 43, 1565–1573. 18. Süßmuth, R., Metzger, J., and Jung, G., Chapter 19 this edition. 19. Chater, K. F. (1990) The improving prospects for yield increase by genetic engineering in antibiotic-producing streptomycetes. Biotechnology 8, 115–121. 20. Süßmuth, R., Pelzer, S., Nicholson, G., Walk, T., Wohlleben, W., and Jung, G. (1999). New advances in the biosynthesis of glycopeptide antibiotics of the vancomycin type from Amycolatopsis mediterranei. Angew. Chem. Int. Ed. Engl. 38, 1976–1979. 21. Kirner, S., Hammer, P. E., Hill, D. S., Altmann, A., Fischer, I., Weislo, L. J., Lanahan, M., van Pée, K.-H., and Ligon, J. M. (1998) Functions encoded by pyrrolnitrin biosynthetic genes from Pseudomonas fluorescens. J. Bacteriol. 180, 1939–1943. 22. Pelzer, S., Huber, P., Süßmuth, R., Recktenwald, J., Heckmann, D., and Wohlleben, W. (1999) Nucleic acid fragment and vector comprising a halogenase, and a process for halogenating chemical compounds. US patent application 19926770.7
150
8 Homologous Recombination and the Induction of the SOS-Response in Antibiotic Producing Streptomycetes Günther Muth and Wolfgang Wohlleben*
8.1 Introduction
Antibiotic production of mycel forming streptomycetes is controlled by a complex regulatory network allowing Streptomyces to sense different growth conditions and to react to changes in the environment by the production of antibiotics [2]. Antibiotic production of S. coelicolor was shown to be affected by cell density, nutritional limitations, nutritional shiftdown, imbalance in metabolism and by different kinds of stress [2]. To sense and to respond to these environmental conditions Streptomyces evolved a set of pleiotropic regulatory factors such as catabolite repression, sigma factor heterogeneity, two component systems, and pleiotropic regulatory proteins which bind small diffusable butyrolacton autoregulators. These regulatory systems were shown to crosstalk to each other, and they might be also linked to the basic cellular stress response mechanisms as stringent response (ppGpp level), N-limitation (NTR) or SOS-response to trigger antibiotic production. On the other site antibiotic production is also affected by the genetic instability of Streptomyces. Genome rearrangements and deletion of chromosomal ends were shown to result in the deletion of antibiotic pathways [8]. Recently it was demonstrated that the amplification process [14] and the deletion of chromosomal ends [4] depended on the RecA protein. RecA is the key enzyme in homologous recombination and in the induction of the SOS-response by supporting autocleavage of the LexA repressor (reviewed in [15], [7]). The recA genes of a variety of bacteria have been characterized. An alignment of the Streptomyces RecA proteins to RecA sequences from other bacteria revealed a high conservation [6]. Only the C-terminus of the Streptomyces RecA protein was about 20 amino acids longer and showed
* Mikrobiologie/Biotechnologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
151 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
8 Homologous Recombination and the Induction of the SOS-Response no similarity to the C-termini of other RecA proteins [8]. RecA seems to play a particular role in Streptomyces. Overexpression of recA is detrimental to Streptomyces. It was only possible to reintroduce the recA gene into S. lividans on a single-copy plasmid, but not when located on a multi-copy plasmid. If recA was placed under control of an inducible promoter, induction resulted in growth impairment (unpublished data). In contrast to other bacteria, where recA could be inactivated without major problems, the recA gene of Streptomyces seemed to be essential for viability. Extensive attempts to generate recA mutants either by UV-mutagenesis [5] or by gene disruption experiments [9, 1] only allowed the isolation of mutants with residual activity. The gene disruption mutant FrecD3 which lacked the very last 86 amino acids was UV-sensitive, its recombination activity was drastically reduced and genetic instability was enhanced [9]. The recA mutant FRECD3 lost the chromosomal end containing the chloramphenicol (CM) resistance gene 70 times more frequently than the S. lividans wild type [14]. Thus it was concluded that the recombination activity of RecA might be required for the repair of single-stranded breaks occurring during the replication of the Streptomyces chromosome. Due to the linearity of the Streptomyces chromosome a single-stranded break would cause the collapse of the replication fork and result in the deletion of the respective chromosomal end. To analyze the role of RecA in Streptomyces we characterized the S. lividans recA gene by site directed mutagenesis and studied the modulation of RecA activity by the LexA repressor and by the RecX protein.
8.2 Mutational analysis of the S. lividans recA gene
8.2.1 Site directed mutagenesis of recA to discriminate the different biochemical activities of RecA A great variety of different E. coli recA mutants have been characterized in great detail. Certain amino acid exchanges were shown to differentially affect the distinct biochemical activities of the RecA protein (reviewed in [7]). Some of these amino acid residues were found to be highly conserved in other RecA proteins [3]. To identify which activity of RecA, recombination or coprotease activity was responsible for the toxic effects of recA overexpression/deficiency, several amino acid residues of the S. lividans RecA protein were exchanged (Pro67 ?Arg, Gly157 ?Asp, Arg169 ?His, Arg243 ?His) by site directed mutagenesis in analogy to well characterized E. coli recA mutants (manuscript in preparation). 152
8.2 Mutational analysis of the S. lividans recA gene A
B
Figure 8.1: 3D-model of the S. lividans RecA protein. The structure was modeled in analogy to the crystal structure of the E. coli RecA protein [11] using the Swissprod DB viewer and insight II (MSI). (A) Binding of a RecA hexamer to the DNA. (B) Location of the amino acid residues that have been mutated by site directed mutagenesis within the DNA binding domain of a RecA hexamer.
153
8 Homologous Recombination and the Induction of the SOS-Response Since the S. lividans recA gene was fully able to complement an E. coli recA deletion mutant, the effect of the different amino acid exchanges was investigated in E. coli and subsequently in the S. lividans recA mutant SV64 (see below). The mutated recA genes were placed under control of the lacZ promoter and introduced into the E. coli recA deletion strain DK1. UV-resistance was analyzed to indicate proficiency for induction of the SOS-response and recombinational repair. The ability to perform homologous recombination was studied in HFR mating assaying the mobilization of a chromosomal tetracycline resistance marker. The mutated S. lividans RecA proteins conferred only in part the expected phenotype. Whereas the E. coli strain containing Pro67/Arg was UV-sensitive, the respective S. lividans strain was UV-resistant as the wild type. Obviously the mutant protein was defective in supporting autocleavage of the E. coli LexA repressor, but proficient to interact with the S. lividans LexA homologue. Exchange of the Gly157/Asp and Arg243/His in the S. lividans RecA protein showed a very similar phenotype as reported for the respective E. coli mutants (constitutive coprotease, reduced recombination activity). Whereas the Arg169/ His substitution in the E. coli RecA protein still allowed homologous recombination at a reduced ratio, the Streptomyces RecA-Arg169/His protein was completely defective in homologous recombination. Thus, the amino acid exchanges in the S. lividans RecA protein had only in part the effects of the respective aa substitutions of the E. coli RecA. This shows that despite the high evolutionary conservation of RecA proteins the enzymatic activity is differentially affected by analogous amino acid substitutions.
8.2.2 3D-modeling of the mutated RecA proteins The X-ray structure of the E. coli RecA protein was elucidated [10]. Due to the high similarity of the S. lividans RecA protein to the E. coli RecA protein it was possible to model the 3D-structure of the S. lividans protein and to simulate the effect of the above described amino acid exchanges on the 3D-structure of the RecA filaments. Computer modeling was done using insight II (MSI) and the SWISSPROD DB viewer3.5b3. Surprisingly we found that all the mutated residues located at the inner site of RecA filaments (manuscript in preparation), probably interacting with DNA (Fig. 8.1, s. p. 153). This suggests, that small differences in the DNA-binding activity of the RecA filaments are responsible for the selective inactivation of the different biochemical functions, as LexA cleavage, recombinational repair and homologous recombination. This also explains why analogous amino acid exchanges in the S. lividans RecA protein conferred only in part the same phenotype as observed for the respective E. coli recA mutants.
154
8.2 Mutational analysis of the S. lividans recA gene
8.2.3 Construction of a recA replacement mutant Since extensive attempts to inactivate recA by gene disruption or gene replacement failed [9], we intended to replace the chromosomal recA gene while providing a second copy of recA (Fig. 8.2). The plasmid-encoded copy of recA was placed under control of the tipA promoter to allow inducible recA expression. Following the replacement of the chromosomal recA gene by the aphII casette, the recA expression vector (pEXrecA) was eliminated by plasmid incompatibil-
Figure 8.2: Construction of the S. lividans recA gene replacement mutant SV64. After elimination of the temperature-sensitive replacement plasmid and selection for the replacement of the chromosomal recA gene by the aphII cassette, the expression plasmid pEXrecA was removed using the incompatible plasmid pIJ920.
155
8 Homologous Recombination and the Induction of the SOS-Response ity: After transformation of the recA mutant with plasmid pIJ920, which carried the same replicon as pEXrecA, several colonies could be selected that had lost pEXrecA. By PCR analysis and Southern hybridization the lack of recA sequences was proved. Thus we were able to delete the chromosomal copy of recA, previously thought to be indispensable [13]. There are two possible explanations why we were able to generate a completely defective recA mutant by this procedure whereas it was not possible by the classical protocol: 1) Due to the lack of chi sequences or other unknown reasons the recA containing DNA-fragment is only poorly recombinogenic. Overexpression of the RecA protein, however, could result in an enhanced recombination activity that allows the recombination of poor substrates. 2) The recA mutant had acquired an additional, suppressing mutation which overcomes the RecA deficiency. Since the plasmid-encoded copy of recA which is under control of the tipA promoter might be just sufficient to override the toxic effect of recA deficiency, but might not be able to complement recA with wild type efficiency, a selection pressure could exist to select for such suppressing mutations. The recA mutant SV64 represents a so-called whi mutant, that is able to erect aerial mycelial, but is deficient in sporulation. Even after complementation with the wild type recA this sporulation defect is still present, whereas UV-sensitivity and recombination activity are fully complemented. This might be a clear indication that SV64 had acquired an additional mutation that could suppress the toxic effects of recA deficiency.
8.3 Regulation of RecA activity 8.3.1 Transcriptional regulation of recA by the LexA repressor In all bacteria recA is regulated by the SOS-repressor LexA. LexA binds to so called SOS Boxes (Cheo Box) in the promoter region of the genes of the SOS-response. In the putative promoter region of recA there is a sequence GAACATCCATTC which resembles the B. subtilis SOS Box GAACNNNNGTTC/T. To analyze transcriptional regulation of recA by LexA, we expressed the S. coelicolor lexA gene in E. coli as a His6-fusionprotein and purified LexA by Ni2+-NTA chromatography. A 109 bp fragment containing the putative SOS Box of the S. lividans recA gene was amplified by PCR and 3'-labeled with digoxigenin. After incubation with LexA the reaction mixture was separated on a 4% Tris-Borat-PAGE and blotted onto a nylon membrane and visualized with Anti-Dig-antibody-conjugate. The retardation of the SOS box containing fragment demonstrated that the Streptomyces LexA is able to bind the proposed SOS box [13]. 156
8.3 Regulation of RecA activity
8.3.2 Interaction of RecA with RecX 8.3.2.1 Construction of a recX mutant Downstream of the S. lividans recA gene lies an orf with significant similarity to bacterial recX genes. The function of RecX which is in many bacteria translationally coupled to recA is unknown. Since in P. aeruginosa overexpression of recA was lethal if recX was not coexpressed a regulatory role on RecA activity was suggested. To analyze the function of RecX, a recX mutant of S. lividans was constructed. A cloned recX gene was disrupted by the insertion of a tsr cassette into the single BclI site. Subsequently, the chromosomal copy of recX was replaced by the inactivated recX copy. The resulting mutant (SVX1) represents the first bacterial recX mutant that is well defined and had no additional defects [12]. SVX1 displayed the following phenotype: Homologous recombination was not affected. UV-resistance and genetic instability (determined by the percentage of CM sensitive segregants) were identical to that of S. lividans WT. However, the recX mutant showed a clearly reduced colony size (Fig. 8.3). An even more drastic phenotype was observed after overexpression of recA. In a S. lividans TK64 culture carrying the recA expression plasmid pEXrecA the colony titer under inducing conditions was about 60 % of that of uninduced cultures, indicating the toxic effect of recA overexpression. In the recX mutant however, no single colony could grow on thiostrepton-containing medium [12]. This showed that induction of the tipA promoter resulting in the overexpression of recA was lethal in the absence of RecX in S. lividans.
Figure 8.3: Morphology of the recX mutant SVX1. The recX gene of S. lividans was disrupted by a thiostrepton resistance cassette. The colony size of the resulting mutant SVX1 was approximately 30 % (area) of that of the wild type strain.
157
8 Homologous Recombination and the Induction of the SOS-Response To confirm that the overexpression of recA is responsible for the toxic effect in the SVX1 mutant, we expressed an inactive RecA protein (Arg169-His) in the mutant. The mutated recA gene was not able to complement an E. coli recA deficient mutant with regard to recombination efficiency and UV-resistance, whereas the wild type RecA protein could. The inactive recA gene was inserted into the expression plasmid under control of the tipA promoter and transferred into the mutant SVX1. After induction with thiostrepton 95% growth in comparison to non-induced conditions was observed. Obviously, overexpression of an inactive RecA protein was tolerated in the recX mutant [12]. This demonstrated that the toxic effect of recA expression is caused by the enzymatic activity of the RecA protein and not by unspecific effects of protein overexpression.
8.3.2.2 Transcriptional analysis of the recX gene In contrast to the situation in other bacteria, where recX overlaps or lies immediately downstream of recA, the recX gene of S. lividans is separated from recA by a fragment of 220 bp. The distance and a potential termination structure (DE –26.2 kJ) 65 bp downstream of recA suggested that these two genes were not cotranscribed. To assess whether both genes were expressed from a single transcript a reverse transcription polymerase chain reaction (RT-PCR) analysis was performed. Since recA is regulated by the SOS-repressor LexA, RNA was isolated at different intervals after induction with the DNA damaging methyl methanesulfonate (MMS, 25 µg ml–1). The RT reaction was performed using random nonamer oligonucleotides. Primers for subsequent PCR were designed for the specific amplification of recA, recX or recA-recX transcripts. The suitability of the primers chosen for RT-PCR was demonstrated by PCR on genomic DNA as template (Fig. 8.4, lanes “DNA”). To confirm the absence of contaminating DNA, a control PCR was performed using RNA as template. From uninduced cultures no recX transcript and only a weak band indicating basal expression of the recA gene were detected (Fig. 8.4, panel A and C, lane 0'). 20 minutes after induction the intensity of the recA-specific band increased, demonstrating induction of the recA gene during the SOS-response. Transcription of the recX gene, however, was not detectable even 20 minutes after induction. After 40 minutes and 60 minutes, expression of recA reached its maximum. Concomitantly a fragment corresponding to the recA-recX cotranscript was amplified. The recA-recX cotranscript, however, was less abundant (~10 %) than the recA transcript alone [12]. The terminator downstream of recA is probably responsible in ensuring the different transcript levels within the recAX operon.
8.3.2.3 Putative role of the RecX protein in modulating RecA activity The phenotype of the recX mutant SVX1 indicated, that RecX is required to overcome the toxic effects of recA overexpression. To analyze whether RecX 158
8.3 Regulation of RecA activity
Figure 8.4: RT-PCR analysis of the recAX operon. Following induction with MMS, RNA was isolated from S. lividans TK64 (panel A, B, C) and SVX1 (panel D) at 0', 20', 40' and 60' minutes and converted to DNA using AMV reverse transcriptase. Subsequently, PCR was performed using different primer combinations to detect the specific transcripts. Panel A and D, internal recA primers; panel B, recA-recX primers; panel C, internal recX primers; lane “DNA”: control PCR using genomic DNA as template.
functions by regulating down recA transcription, RT-PCR analysis was performed with RNA isolated from SVX1. However, recA transcription was not significantly affected in the recX mutant [12], and an induction pattern very similar to that of the wild type was observed following DNA damage (Fig. 8.4 panel D). Because RecX did not seem to influence transcription of recA we propose an interaction of RecX with the RecA protein. Since all of the biochemical functions of RecA are directly affected by the DNA binding an alteration of the binding characteristics of RecA filaments by the interaction with RecX might efficiently interfere with the specific activity of RecA.
159
8 Homologous Recombination and the Induction of the SOS-Response
References
1. Aigle, B., A. C. Holl, J. F. Angulo, P. Leblond, and B. Decaris (1997) Characterization of two Streptomyces ambofaciens recA mutants: Identification of the RecA protein by immunoblotting. FEMS Microbiol. Lett. 149, 181–187. 2. Bibb, M. (1996) The regulation of antibiotic production in Streptomyces coelicolor A3(2). Microbiology 142, 1335–1344. 3. Brendel, V., L. Brocchieri, S. J. Sandler, A. J. Clark, and S. Karlin (1997) Evolutionary comparisons of RecA-like proteins across all major kingdoms of living organisms. J. Mol. Evol. 44, 528–541. 4. Fischer, G., T. Wenner, B. Decaris, and P. Leblond (1998) Chromosomal arm replacement generates a high level of intraspecific polymorphism in the terminal inverted repeats of the linear chromosomal DNA of Streptomyces ambofaciens. Proc. Natl. Acad. Sci. U.S.A. 95, 14296–14301. 5. Harold, R. J. and D. A. Hopwood (1970) Ultraviolet-sensitive mutants of Streptomyces coelicolor: I. Phenotypic characterization. Mutat. Res. 10, 427–438. 6. Karlin, S. and L. Brocchieri (1996) Evolutionary conservation of recA genes in relation to protein structure and function. J. Bacteriol. 178, 1881–1894. 7. Kowalczykowski, S. C., D. A. Dixon, A. K. Eggleston, S. D. Lauder, and W. M. Rehrauer (1994) Biochemistry of homologous recombination in Escherichia coli. Microbiol. Rev. 58, 401–465. 8. Möhrle, V., U. Roos, and C. Bormann (1995) Identification of cellular proteins involved in nikkomycin production in Streptomyces tendae Tü901. Mol. Microbiol. 15, 561– 571. 9. Muth, G., D. Frese, A. Kleber, and W. Wohlleben (1997) Mutational analysis of the Streptomyces lividans recA gene suggests that only mutants with residual activity remain viable. Mol. Gen. Genet. 255, 420–428. 10. Story, R. M. and T. A. Steitz (1992) Structure of the recA protein-ADP complex. Nature 355, 374–376. 11. Story, R. M., I. T. Weber, and T. A. Steitz (1992) The structure of the E. coli recA protein monomer and polymer. Nature 355, 318–325. 12. Vierling, S., Weber, T., Wohlleben, W., and Muth, G. (2000) Transcriptional and mutational analysis of the S.lividans recX gene and its interference with RecA activity. J. Bacteriol. 182, 4005–4011. 13. Vierling, S., Weber, T., Wohlleben, W., and Muth, G. (2001) Evidence that an additional mutation is required to tolerate insertional inactivation of the Streptomyces lividans recA gene. J. Bacteriol. 183, 4374–4381. 14. Volff, J. N. and J. Altenbuchner (1997) Influence of disruption of the recA gene on genetic instability and genome rearrangement in Streptomyces lividans. J. Bacteriol. 179, 2440–2445. 15. West, S. (1992) Enzymes and molecular mechanism of genetic recombination. Annu. Rev. Biochem. 61,603–640.
160
Membrane Processes
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
9 Regulated Transport and Signal Transfer Channels involved in Bacterial Iron Supply Volkmar Braun* and Helmut Killmann
9.1 Introduction
During studies on the structure and function of the phage receptor TonA (for phage T1, T one), we discovered in 1975 that the receptor participates in the transport of the iron complex ferrichrome [1]; TonA was renamed FhuA for ferric hydroxamate (ferrichrome) uptake. At that time it was known that the product of an additional gene, termed tonB, is required for the FhuA-dependent infection of Escherichia coli by certain phages, and we were soon able to show that the electrochemical potential of the cytoplasmic membrane is somehow transmitted by TonB to the outer membrane protein FhuA and serves to activate FhuA [2]. The study of energy-consuming active import not only of iron complexes, but also of bacterial protein toxins (colicins), to gain insight into the function of protein transporters and the transfer of energy from one membrane to an adjacent membrane, became a major topic of our laboratory. The Fe2+/Fe3+ pair displays a wide range of redox potentials from –300 to +700 mV, depending on the iron ligands and the protein environment. All organisms, with the exception of certain lactobacilli, take advantage of this unusually wide range of electron transport capacity. However, despite its high abundance in nature, iron is difficult to acquire by most organisms. Under oxic conditions and at the physiological pH of 7.0, the concentration of free ferric ions in equilibrium with the ferric hydroxide polymer is in the order of 10–12 µM. The growth-promoting concentration of iron for microbes is approximately 10–1 µM. Bacteria contain about 105 iron ions per cell. Since bacteria usually reach densities of 109 cells per ml, they require 1014 iron ions per generation, which sharply contrasts the available 103 free ions. Bacteria have solved the iron supply problem by synthesizing iron-chelating compounds, called siderophores, which
* Mikrobiologie/Biotechnologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
163 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
9 Regulated Transport and Signal Transfer Channels they secrete and take up again after loading with Fe3+. The bacteria employ highly specific and efficient transport systems for ferric siderophore uptake. In addition to the siderophores they synthesize themselves, bacteria can use siderophores synthesized and secreted by other bacteria and fungi or iron sources of their hosts (Fig. 9.1). In the human body, iron is bound to proteins: transferrin in the serum, lactoferrin in secretory fluids and in granules of polymorphonuclear leukocytes, and intracellular ferritin. Bacteria take up the free iron in equilibrium with transferrin and lactoferrin via siderophores, or they take up free heme, heme from hemoglobin and hemopexin, and iron from transferrin and lactoferrin without the involvement of siderophores. Heme and iron is released from the host proteins at the bacterial cell surface.
Figure 9.1: Iron transport systems of Gram-negative bacteria. Transport proteins located in the outer membrane (OM), periplasm (PP), and in the cytoplasmic membrane (CC) are symbolized by filled barrels. Siderophores are produced by some species and released as indicated by the dashed arrow. Acquisition of ferric iron from lactoferrin, transferrin, heme, hemoglobin, hemopexin, and ferric siderophores, or of ferrous iron is indicated by solid arrows. The Ton complex composed of TonB, ExbB, ExbD is indicated by empty barrels. Fe2+ is transported by the Feo system consisting of a large and two small transport proteins.
164
9.2 The Fhu proteins catalyze active transport of ferrichrome E. coli K-12, the most widely used laboratory strain, synthesizes six distinct siderophore-mediated Fe3+ transport systems, each of which recognizes a single type of ferric siderophore, and one Fe2+ transport system. However, E. coli K-12 synthesizes only one siderophore, enterobactin (cyclic trimer of 2,3-dihydroxybenzoylserine, also designated enterochelin). Two of the transport systems take up the biosynthetic precursors of enterobactin, 2,3-dihydroxybenzoic acid and 2,3-dihydroxybenzoylserine [3]. One transport system is specific for ferric citrate, although the concentration of citrate released from E. coli K-12 into the growth medium under laboratory conditions (not more than 15 µM) is not sufficient to induce the system (threshold concentration: 100 µM citrate). Citrate is an exogenous siderophore, as are ferrichrome and coprogen of fungal origin, all of which are also used as siderophores by E. coli. Pathogenic E. coli strains frequently synthesize aerobactin and a ferric-aerobactin-specific transport protein, whose genes are encoded on plasmids. The many iron transport systems in E. coli and in other bacteria probably reflect the iron sources to which the bacteria are exposed in their various environments to which they have to be able to adapt. In the following, the systems of iron transport via ferrichrome in E. coli and Bacillus subtilis, citrate in E. coli, and heme in Yersinia enterocolitica will be described as they represent the first systems studied and those studied in most detail and are typical for Fe3+-siderophore transport systems in these organisms and in other bacteria. The reader is referred to recent references for more comprehensive reviews [4–14].
9.2 The Fhu proteins catalyze active transport of ferrichrome and the antibiotic albomycin across the outer membrane and the cytoplasmic membrane of E. coli
9.2.1 Transport across the outer membrane – the need for receptor proteins Gram-negative bacteria such as E. coli are surrounded by an outer membrane, the inner cytoplasmic membrane, and the periplasm between the two membranes (Fig. 9.1). Most substrates diffuse across the outer membrane through the water-filled channels of the porins, which in E. coli have an exclusion limit of approximately 600 Da [15]. In contrast, the Fe3+-siderophores, heme, and vitamin B12 are actively transported across the outer membrane by proteins (Fig. 9.1). The Fe3+-siderophores with molecular masses of 700–1000 Da are too large to diffuse through the porins. The concentration of Fe3+-siderophores is 165
9 Regulated Transport and Signal Transfer Channels also too low for a sufficient iron supply if one considers the strong dilution of the exported siderophores in the environment of the cells. Competition of the siderophores for iron with other strong iron-chelating compounds, such as siderophores produced by bacteria and fungi that cannot be used by a particular strain, or proteins produced by the infected hosts, e. g. transferrin, lactoferrin, and ferritins, further lowers the Fe3+-siderophore concentration. Fe3+-siderophores that make contact with bacterial cells are captured by their cognate receptor proteins. By this means, the Fe3+-siderophores are extracted from the medium and concentrated at the bacterial cell surface. The low KD of the receptors (below 0.1 µM) favors binding of the Fe3+-siderophores. A function similar to that of the receptor proteins at the cell surface of Gram-negative bacteria is displayed by the binding proteins, which are anchored by a lipid moiety of the murein-lipoprotein type to the outer face of the cytoplasmic membrane of Gram-positive bacteria, as demonstrated for the ferrichrome transport in B. subtilis [16]. In Gram-positive bacteria, the Fe3+-siderophores diffuse through the multi-layered murein and teichoic acids and gain direct access to the Fe3+-siderophore transport systems in the cytoplasmic membrane.
9.2.2 The FhuA transport protein forms a regulated channel in the outer membrane of E. coli FhuA was originally chosen and continuously studied since it is a multifunctional protein, and the mechanisms of energy-dependent phage infection, active transport of ferric siderophores and the antibiotics albomycin and rifamycin CGP 4832, and the import of the protein toxin colicin M and of the peptide toxin microcin J25 by FhuA can be studied. At the beginning of the studies, it was known that a function encoded by the tonA gene, later designated as the fhuA gene [17], is required for infection by the phages T1, T5, and f80 and for the sensitivity of cells to colicin M. In 1975, we and the Neilands group independently showed that FhuA is involved in ferrichrome transport [1, 18] and that mutants resistant to albomycin are mostly impaired in albomycin and ferrichrome transport. This provided a convenient way to characterize the transport across the outer membrane and the cytoplasmic membrane. Transport genes and their products were identified, and the transport proteins were localized. Evidence for a channel in the FhuA outer membrane protein involved in ferrichrome transport has been obtained by excision of DNA fragments from the fhuA structural gene. Deletion of residues 322–355 (FhuAD322–355) results in a stable protein that is integrated into the outer membrane and contains an open channel through which ferrichrome enters cells at a rate proportional to the external ferrichrome concentration, without showing saturation at higher ferrichrome concentrations. Transport across the cytoplasmic membrane is not rate limiting. At a low ferrichrome concentration (1 µM), FhuAD322–355 supports iron uptake much less than the FhuA wild type. At higher ferrichrome concen166
9.2 The Fhu proteins catalyze active transport of ferrichrome trations (above 7 µM), iron uptake through FhuAD322–355 exceeds uptake via the FhuA wild type, as one would expect for a diffusion-controlled process. FhuAD322–355 virtually does not bind ferrichrome. FhuAD322–355 incorporated into artificial lipid bilayers (black lipid membranes) forms open channels with a single channel conductance more than three times as high as that of E. coli porins [19]. Cells that synthesize FhuAD322–355 can grow on maltotetraose and maltopentaose in the absence of the LamB protein through which maltodextrins diffuse across the outer membrane, and they are sensitive to SDS and bacitracin [19] because the maltodextrins and noxious agents diffuse through the outer membranes via the open FhuA channels. Segment 322–355 of FhuA is exposed to the cell surface, as revealed by proteolysis of FhuA by added proteases after insertion of a tetrapeptide or a hexadecapeptide after residue 321, which results in ferrichrome-transport-active FhuA derivatives [20]. FhuA contains four cysteine residues that form disulfide bridges [21]. After cleavage of the disulfide bridge with mercaptoethanol between the two cysteines at positions 318 and 329, FhuA in living cells reacts weakly with biotinmaleimide; however, a newly introduced cysteine at position 336 is highly reactive, which demonstrates its exposure near the cell surface [21]. Deletion of a single amino acid, Asp 348, inactivates ferrichrome transport activity of FhuA and therefore is in a region important for FhuA activity [22].
9.2.3 FhuA as phage receptor FhuA not only catalyzes transport of ferrichrome and of the structurally related antibiotic albomycin, but also serves as a receptor for the infection of a number of phages and for killing by toxic proteins (Fig. 9.2). The large variety of ligands that bind to the multifunctional FhuA protein makes FhuA a particularly attractive receptor for studying structure-function relationships. Cells that synthesize FhuAD322–355 and FhuAD335–355 are resistant to the phages T1, T5, and f80 (UC-1 was not tested) and to colicin M (microcin 25 was not tested). This shows that this segment is involved in binding of the phages and in binding and uptake of colicin M. The binding sites of the phages in the region 316–355 were demonstrated by competitive peptide mapping [23]. This method avoids long-range conformational changes, caused by mutations in the gating loop, at regions outside the gating loop that may be involved in phage binding. Synthetic hexapeptides that span the entire region were employed, and those identical to sequences of three segments inhibited infection by phages T1, T5, and f80. Similar results were obtained with pentapeptides comprising residues 316–320, 332–337, 348–351, and with selected tetrapeptides (Killmann, H., unpublished results). The hexapeptides not only interfere with phage binding, but inactivate the phages by triggering DNA release from the phage head [23]. The peptides mimic binding of phage T5 to the entire isolated FhuA protein, which causes DNA release from 167
9 Regulated Transport and Signal Transfer Channels
168
9.2 The Fhu proteins catalyze active transport of ferrichrome the phage as shown by electron microscopy. The peptides also trigger release of DNA from phage f80, in contrast to the entire FhuA molecule, which does not cause DNA release from f80 [23, 24] or f80 inactivation [23]. DNA release caused by the peptides requires input of thermal energy since, under the conditions used, all phages are inactivated at 37 8C and half of them remain intact at 20 8C. After the release of DNA caused by FhuA, phage T5 largely maintains its shape. In contrast, the peptides cause a collapse of the phage head and release of the phage tail. The stronger effect of the peptides compared to that of FhuA can be explained by their flexible conformation, which enables them to adapt to the surface of the phage tail region with which the phages bind to FhuA. One of the peptide conformations is the one that the FhuA segment assumes when energized by the proton-motive force across the cytoplasmic membrane via the Ton complex (see Section 9.3). This energized conformation is required for binding or uptake of all FhuA ligands except phage T5. In contrast, the entire FhuA protein assumes a defined conformation with less flexibility.
9.2.4 Sequence comparison of the FhuA proteins of various Enterobacteriaceae reveal conserved sites important for FhuA activity Comparison of the ferrichrome transporters of E. coli K-12 [25, 26], Salmonella paratyphi, Salmonella typhimurium, and Pantoea agglomerans (formerly Erwinia herbicola) [27] support the division of the E. coli FhuA segment 322–355 into two regions. Mutant FhuAD322–336 still transports ferrichrome, but is resistant to phage T1, and FhuAD335–355 does not transport ferrichrome and is resistant to the phages T1, T5 and f80. FhuA of S. typhimurium lacks residues 321–337 (E. coli FhuA numbering) and P. agglomerans lacks residues 318–331 and 338– 341 [27]. Both strains are resistant to the E. coli phages, but transport ferrichrome. S. paratyphi lacks none of the amino acids, transports ferrichrome, and is sensitive to the E. coli phages. The entire segment 322–355 is therefore required for infection by the phages, but residues 322–336 are dispensable for ferrichrome transport.
3 Figure 9.2: Upper panel (A): Crystal structure of the FhuA protein in the unloaded and albomycin-loaded form [31]. Note the large structural change that occurs at the periplasmic side of FhuA when loaded with albomycin. Lower panel (B): Ligands of the multifunctional FhuA protein in the outer membrane (OM) and transmembrane topology of the TonB, ExbB, and ExbD proteins anchored to the cytoplasmic membrane (CM) and extending into the periplasm (PP). A predicted gating loop that controls the permeability of FhuA is indicated in the cork domain. Interaction of TonB with FhuA is proposed to occur via a region in which residue 160 resides. N designates the N-terminus, and C designates the C-terminus of the proteins. T5, T1, f80, and UC-1 designate phages that use FhuA as a receptor to infect E. coli cells.
169
9 Regulated Transport and Signal Transfer Channels Deletion of fragment 236–248 of the E. coli FhuA [27], which is identical in all four species, abolishes ferrichrome binding and transport. As will be shown below, this region contains a ferrichrome binding site.
9.2.5 Crystal structure of the FhuA transporter in free form and loaded with ferrichrome Initial attempts of the laboratory of Prof. W. Welte of the University of Konstanz to crystallize our cloned FhuA protein yielded only crystals with a diffraction of 8 Å. Later, Andrew Ferguson of the same laboratory obtained crystals with a diffraction of 2.6 Å of a FhuA derivative with a His-tag at residue 405 [28, 29]. Locher et al. [30] succeeded in the determination of the crystal structure of our cloned FhuA protein purified without a His-tag [30]. According to the results of these two groups, FhuA assumes a novel structure (Fig. 9.2, left panel) in which residues 160–714 form a bbarrel composed of 22 antiparallel b-strands. The b-barrel is closed by residues 19– 159, which form a globular structure and enter the b-barrel from the periplasmic side. Since this portion entirely closes the b-barrel, it has been designated the cork or plug. Residues 1–18 are not seen and presumably are flexible. The crystal structure reveals that the above-described segment 322–336 indeed forms a loop (L4), which is the most prominent loop at the cell surface, consistent with the finding that L4 serves as a binding site for the phages and colicin M. In the crystal structure, ferrichrome is bound at a site located outside the outer membrane bilayer (Fig. 9.2, right panel shown with albomycin instead of ferrichrome). Four amino acid residues of the cork and six residues of the b-barrel come close enough to ferrichrome to form hydrogen bonds and van der Waals contacts (Table 9.1). The aromatic residues lining the inner walls of the external pocket of FhuA probably extract ferrichrome from the external medium. Upon binding of ferrichrome, a portion of the cork moves by 1.7 Å towards ferrichrome, accompanied by a large structural transition at the periplasmic side where a short a-helix of the cork is unwound, and Glu19 moves 17.3 Å away from the position it occupies in ferrichrome-unloaded FhuA. Despite this large structural change across the entire FhuA molecule and the thickness of the outer membrane, the b-barrel is not converted to an open channel.
9.2.6 FhuA-albomycin is the first example of an antibiotic-proteintransporter crystal structure Most antibiotics diffuse into bacteria. Their efficiency, as measured by the minimal inhibitory concentration (MIC), is determined by the diffusion rate and the activity at the target sites. Gram-negative bacteria are usually less sensitive to 170
9.2 The Fhu proteins catalyze active transport of ferrichrome Table 9.1:
Interactions of FhuA with its ligands a. Ligand
Residue
Ferricrocin
Albomycin b
Albomycin c
Arg 81
Arg 81
Arg 81
Gly 99 Gln 100
Gly 99 Gln 100
Gly 99 Gln 100
Tyr 116 Tyr 244 Trp 246 Tyr 313 Tyr 315
Tyr 116 Tyr 244 Trp 246 Tyr 313 Tyr 315 Lys 344 Phe 391
Phe 391
Tyr 393
Phe 693 a b c
Phe 557 Phe 558 Phe 693
Phe 115 Tyr 116 Tyr 244 Trp 246 Tyr 313
Phe 391 Tyr 393 Tyr 423 Gln 505
Phe 693
CGP 4832
Glu 98 Gly 99 Gln 100 Ser 101 Phe 115 Tyr 116 Tyr 244 Trp 246 Tyr 313 Tyr 315 Lys 344 Phe 391 Gly 392 Tyr 423 Gln 505
Phe 693 Tyr 696
Distance within 4 Å forming hydrogen bonds and van der Waals contacts; Extended conformation; Compact conformation.
antibiotics than Gram-positive bacteria because they contain an outer membrane that functions as a permeability barrier. However, if antibiotics are actively transported across the outer membrane, their MIC may be lower in Gramnegative bacteria than in Gram-positive bacteria because the antibiotic is accumulated in the periplasm and forms a steep concentration gradient across the cytoplasmic membrane into the cytoplasm, which enhances the diffusion rate, or it may even be actively transported across the cytoplasmic membrane. Both uptake routes have been investigated. The crystal structure of FhuA loaded with albomycin (Fig. 9.2, right panel) reveals that the Fe3+-hydroxamate portion of albomycin occupies the same site on FhuA and is bound by the same amino acid chains as ferrichrome [31] (Table 9.1). In Table 9.1, ferricrocin is shown instead of ferrichrome; ferricrocin has the same structure as ferrichrome except that one of the glycine residues is replaced by a serine residue. Ferricrocin is produced by Aspergillae, functions as an Fe3+-ligand, and is transported as well as ferrichrome. The predominant binding sites of albomycin on FhuA are by far aromatic residues (69%). Binding of the thioribosyl pyrimidine moiety occurs in the external pocket and involves residues Phe 115, Lys 344, Tyr 393, Tyr 423, and Gln 505. These additional bind171
9 Regulated Transport and Signal Transfer Channels ing sites of albomycin as compared to those of ferricrocin do not prevent release of albomycin from FhuA and transport through FhuA. The albomycin transport rate is half the transport rate of ferrichrome, but it is not clear whether the lower rate is caused by transport across the outer membrane or across the cytoplasmic membrane. The structure of the FhuA-albomycin co-crystal also reveals the hitherto unknown conformation of albomycin and the conformation in the transport-competent form. The most unexpected result is the existence of two albomycin conformations in the crystal – an extended and a compact conformation. Both conformations fit into the external cavity of FhuA and occupy different amino acid ligands (Table 9.1). The solvent-exposed external cavity of FhuA is sufficiently large to accommodate the voluminous side chain bound to the Fe3+-hydroxamate moiety of albomycin. With the modular composition of albomycin, in which the iron carrier is linked by a peptide linker to the antibiotically active thioribosyl pyrimidine, nature provides a clue of how to design highly efficient antibiotics that are actively transported into bacteria. Such antibiotics could be synthetically assembled from Fe3+-hydroxamates, which fit into the active center of the transporters, and from an antibiotic that diffuses too slowly into cells to be useful itself as a drug. The FhuA-albomycin structure demonstrates that the water-filled cavities in transporters can tolerate rather large antibiotics that are structurally unrelated to the carrier. The tolerance to the antibiotic structure is not confined to FhuA; albomycin is also transported very well across the cytoplasmic membrane and during this process is recognized by the FhuD and the FhuB proteins.
9.2.7 Crystal structure of FhuA with bound rifamycin CGP 4832 In 1987, a group at Ciba-Geigy reported on a semisynthetic rifamycin derivative, CGP 4832, with an activity against many Gram-negative bacteria 200-fold higher than unmodified rifamycin [32]. It was then shown with our mutants [33] that CGP 4832 is transported by FhuA across the outer membrane of E. coli and that TonB activity is required [34]. Mutants in the fhuBCD genes, which encode the proteins required for active transport of ferrichrome across the cytoplasmic membrane, display unaltered CGP 4832 sensitivity. Our attempts to find additional transport mutants only revealed mutations in fhuA and tonB exbB exbD, which suggests that CGP 4832 crosses the cytoplasmic membrane by diffusion rather than by transport [35]. The use of FhuA as transporter for CGP 4832 is surprising since CGP 4832 does not contain iron and has no structural resemblance to ferrichrome. Therefore, it was particularly of interest to determine the crystal structure of FhuA loaded with CGP 4832 [35]. Analysis of the X-ray diffraction data reveals that CGP 4832 largely occupies the site in FhuA that is also used by ferrichrome (Table 9.1). Interestingly, the amino acid residues Gly 99, Tyr 116, and Tyr 244, which also bind ferri172
9.2 The Fhu proteins catalyze active transport of ferrichrome chrome, recognize those side chains in which CGP 4832 differs from unmodified rifamycin. In contrast to ferrichrome and albomycin, CGP 4832 in the crystal does not cause the large structural change in the periplasmically oriented pocket of FhuA. This does not seem to be caused by restriction of FhuA movements in the crystal since binding of CGP 4832 to FhuA in solution does not result in intrinsic FhuA tryptophan fluorescence quenching [35], as is observed when FhuA binds ferrichrome [35, 36]. This finding has an impact on the concept of how FhuA interacts with TonB since, as discussed above, the large structural transition is thought to facilitate interaction of FhuA with TonB. Since transport of CGP 4932 depends on TonB, interaction of TonB with FhuA may occur in the absence of the structural change. Analysis of a FhuA deletion derivative described below supports this conclusion.
9.2.8 The b-barrel domain of FhuAD5–160 is sufficient for TonB-dependent FhuA activities The FhuA crystal structure reveals that residues 160 to 714 of the mature protein form a b-barrel that is closed from the periplasmic side by the globular N-proximal fragment, residues 1 to 159 designate the cork. We deleted the cork with the idea that the resulting FhuAD5–160 protein might form an open channel [37]. To be active, the remaining b-barrel domain has to be stable and exported across the cytoplasmic membrane into the outer membrane. If the b-barrel lacking the cork does not collapse, one could expect that it would form a permanently open channel larger than the channels of the previously described FhuAD322–355 and FhuAD335–355 deletion derivatives [19, 38]. Such a channel would allow diffusion, but no active transport of ferrichrome, and would confer sensitivity to SDS and antibiotics that are too large to diffuse through the porin channels. One could anticipate that FhuAD5–160 might still function as a receptor for the TonB-independent infection by phage T5, provided loop 4 retains its conformation. In fact, deletion of the cork results in a stable protein, FhuAD5– 160, that is incorporated into the outer membrane. Cells that synthesize FhuAD5–160 display a higher sensitivity to large antibiotics such as erythromycin, rifamycin, bacitracin, and vancomycin, and grow on maltotetraose and maltopentaose in the absence of LamB. High concentrations of ferrichrome support growth of a tonB mutant that synthesizes FhuAD5–160. These results demonstrate non-specific diffusion of compounds across the outer membrane of cells that synthesize FhuAD5–160 [37]. However, growth of a FhuAD5–160 tonB + strain occurs at low ferrichrome concentrations, and ferrichrome is transported at about 45% of the FhuA wild type rate, despite the lack of ferrichrome binding sites provided by the cork. FhuAD5–160 confers sensitivity to phages T1 and f80 and the uptake of colicin M at levels as high or nearly as high through wild type FhuA; FhuAD5–160 also 173
9 Regulated Transport and Signal Transfer Channels confers sensitivity to albomycin and rifamycin CGP 4832. These data confirm the high activity of FhuAD5–160 and its strong dependence on TonB [37], despite the lack of the TonB box (residues 7 to 11) previously implicated in the interaction of FhuA with TonB. FhuAD5–160 still functions as a specific transporter and as a phage receptor, and sites in addition to the TonB box are involved in the TonBmediated response of FhuA to the proton gradient of the cytoplasmic membrane. It is proposed that TonB interacts with the TonB box of FhuA and with the b-barrel to release ferrichrome from the FhuA binding sites and to open the channel in FhuA. For transport of ferrichrome through the open channel of FhuAD5–160, interaction of TonB with the b-barrel is sufficient to release ferrichrome from the residual binding sites at the b-barrel and to induce the active conformation of the L4 loop at the cell surface for infection by the TonB-dependent phages T1 and f80. For the infection by phages T1 and f80, induction of the infection-competent conformation of loop 4 by TonB cannot be mediated by the cork domain, but must be transmitted by the b-barrel to the ferrichrome binding pocket close to the cell surface. This implies that the interaction of FhuA with TonB is propagated from the periplasm through the b-barrel up to loop 4 [37]. The activity of FhuAD5–160 sheds light on the mode of action of wild type FhuA. There is more than a single interaction site between FhuA and TonB. Interaction of TonB with the TonB box of FhuA might impose the structural change in the cork that is required to open a channel; without this interaction, the cork tightly closes the channel lumen of FhuA. This structural change may contribute to the release of ferrichrome from its FhuA binding site. However, the structural change in the cork might only alter the orientation of the cork amino acids that contribute to ferrichrome binding and not affect the amino acids at the b-barrel that bind ferrichrome. For the complete release of ferrichrome, the b-barrel also changes its structure in response to TonB. As our results demonstrate, TonB can activate FhuAD5–160 exclusively through the b-barrel, and we propose that this occurs by a conformational transition that changes the orientation of the aromatic residues such that they no longer bind ferrichrome. These findings imply that the b-barrel does not form such a rigid structure that it cannot change the conformation to alter its activity. Export of FhuAD5–160 across the cytoplasmic membrane and insertion into the outer membrane may also shed light on the assembly of wild type FhuA. Obviously, the FhuA barrel can be assembled in the absence of the cork; this suggests that the barrel is constructed first and then the cork is introduced into the pre-formed barrel. This conclusion is supported by the stability of the FhuAD5–160 b-barrel to intracellular and added proteases as high as the stability of wild type FhuA. FhuA, like many other outer membrane proteins, contains at its C-terminus a phenylalanine; in the outer membrane protein PhoE, this amino acid is important for the formation of an assembly-competent folded monomer. In FhuA, the C-terminal Phe residue is part of the barrel, which may fold in the periplasm and insert into the outer membrane prior to the insertion of the cork.
174
9.3 Transduction of energy from the cytoplasmic membrane
9.3 Transduction of energy from the cytoplasmic membrane into the outer membrane for the activation of FhuA as a transporter and phage receptor
Resistance of cells that synthesize wild type FhuA to SDS and large antibiotics and the failure of the cells to grow on maltodextrins when LamB is lacking indicate that FhuA does not form a permanently open channel with an inner diameter larger than that of the porins. Cells that synthesize the FhuAD5–160, FhuAD322–355, or FhuAD335–355 deletion protein have an increased antibiotic sensitivity, allow diffusion of ferrichrome, and grow on maltodextrins in the absence of LamB, which indicate that the FhuA channels are open. The question arises how the FhuA wild type channel is opened physiologically. The first indication how this may happen and still the best evidence obtained comes from studies on the infection of phage T1 [2]. Phage-resistant cells were mutated in two genes, designated tonA and tonB (ton from T one). tonA was later renamed fhuA [17] to indicate its physiological activity in ferric hydroxamate (ferrichrome) uptake. Early studies at the time when phage genetics opened a field now known as molecular biology revealed that phage T1 adsorbs to tonB mutants reversibly and that cellular energy is required for irreversible adsorption accompanied by infection. Phage T1 does not bind to tonA (fhuA) mutants, suggesting that FhuA determines the primary binding site and TonB defines a later step in infection (cited in [2]). The nature of the energy required for irreversible adsorption of phage T1 (and f80) was then shown to be the electrochemical potential of the cytoplasmic membrane [2]. Correlation of the energy and TonB requirement suggests that TonB is somehow involved in coupling the energy conserved in the transmembrane potential of the cytoplasmic membrane to irreversible adsorption of phage T1 to the outer membrane. Later, the fhuA gene product was shown to be a protein in the outer membrane [39] and that phage T5, which does not require energy and TonB for productive adsorption (infection), is inactivated by the isolated protein, in contrast to phages T1 and f80, which do require TonB in energized cells [40]. Phage T5 inactivation is prevented by colicin M, which suggests a common binding site for phage T5 and colicin M [1]. This has been proven with point mutations and deletions in the region 316–355 and by competitive peptide mapping [19, 22, 23, 38]. The relationship between the adsorption step and the requirements for energy and the TonB activity was shown by isolating phage T1 host mutants that infected tonB mutants. Infection by these mutants demonstrates that transfer of DNA across the outer membrane does not require TonB activity. The energized cytoplasmic membrane induces an infection-competent conformation in FhuA through the activity of TonB. This conformation is recognized by phages T1 and f80, while phage T5 recognizes the unenergized conformation. Evidence for two FhuA conformations has also been obtained by competition studies between phage T5 and ferrichrome at FhuA. Ferrichrome inhibits phage T5 adsorption to FhuA of unenergized cells 175
9 Regulated Transport and Signal Transfer Channels (tonB mutants or energy-deprived tonB+ cells) more strongly than adsorption to FhuA of energized cells [41]. It is likely that all Ton-dependent receptors undergo similar conformational changes upon energization, which results in translocation of the ferric siderophores across the outer membrane. Transport of ferrichrome across the outer membrane into the periplasm of E. coli was determined with a fhuB mutant devoid of ferrichrome transport across the cytoplasmic membrane. The periplasmic FhuD binding protein had to be overproduced to measure accumulation of radioactive [55Fe3+]ferrichrome in the periplasm [42]. About 8000 ferrichrome molecules per cell bind to FhuD and can be chased with non-radioactive ferrichrome. Under the same conditions, 100 000 ferrichrome molecules are taken up into the cytoplasm of a fhuB+ strain and can no longer be chased because iron is released from ferrichrome and incorporated into heme and non-heme iron proteins and into the undefined iron pool of the cell. For energization of outer membrane transport, the electrochemical potential (proton-motive force) of the cytoplasmic membrane is required. Three proteins are known to be involved in the transduction of energy from the cytoplasmic membrane into the outer membrane, where FhuA is activated to transport ferrichrome, albomycin, rifamycin CGP 4832, colicin M, and microcin J25, and to serve in the energized state as receptor for the infection by phages T1 and f80 (Fig. 9.2). The proteins are TonB, ExbB, and ExbD. tonB mutants are devoid of all FhuA activities, except infection by phage T5, while exbB and exbD mutants display residual activities. For this reason, an accessory role was ascribed to the ExbB and ExbD proteins until it was shown that two other proteins, TolQ and TolR, can partially substitute for ExbB and ExbD, respectively [43–45]. exbB tolQ and exbD tolR double mutants are both completely inactive, as are tonB deletion mutants. tonB mutants do not accumulate ferrichrome in the periplasm of a fhuB mutant [42], and uptake in cells that synthesize FhuAD322–355 does not require TonB, which suggests that TonB is required only for active transport across the outer membrane. The transmembrane topology of TonB (Fig. 9.2) further supports its role as an energy transducer between the cytoplasmic and the outer membranes. TonB is anchored by the N-terminal end to the cytoplasmic membrane and extends into the periplasm [46]. Interaction of TonB with outer membrane receptors has been demonstrated by mutations close to the N-terminal end of receptors, in the so-called TonB box, which are suppressed by mutations at residue 160 of TonB [47–49]. Furthermore, overproduced FhuA prevents proteolytic degradation of overproduced TonB in cells only when both proteins functionally interact with each other [50]. Evidence for a structural complex between TonB, ExbB, and ExbD is based on the inhibition by ExbB of TonB and ExbD degradation by cellular proteases [46, 51]. A fragment of TonB, consisting of only 44 N-terminal residues, is still stabilized by ExbB [52]. Since this fragment consists mainly of the portion of TonB that spans the cytoplasmic membrane, it is likely that interaction between the two proteins occurs within or close to the cytoplasmic membrane. Weak suppression of a TonB derivative lacking Val-17 by an Ala-39-to-Glu mutation in the first transmembrane region of ExbB further indicates functional in176
9.3 Transduction of energy from the cytoplasmic membrane teraction of the two proteins in the cytoplasmic membrane [53]. Furthermore, TonB and ExbB can be chemically cross-linked to each other in cells [53]. In vitro evidence for interaction of the three proteins has been obtained by binding isolated ExbB which carried a His-tag at the C-terminal end to a nickel-nitrilotriacetate agarose column. His-tag ExbB specifically retains ExbD and TonB on the column [54]. Although the N-terminal transmembrane segment of TonB and the region around residue 160 are important functional regions of TonB, they are certainly not the only regions that determine TonB activity. Mutations in the C-terminal region beyond residue 160 (E. coli TonB consists of 239 residues) strongly affect uptake of ferrichrome, sensitivity to certain colicins and phages [52, 55], and the activity of certain mutated FhuA proteins with phage T5 [56]. ExbD is, like TonB, anchored by the N-terminal region in the cytoplasmic membrane (Fig. 9.2) and extends into the periplasm [57]. Replacement of Asp 25, the only charged amino acid in the transmembrane region, by Asn results in an inactive ExbD [54]. It is conceivable that Asp 25 interacts with His 20 of TonB, and when replaced by Asn, largely inactivates TonB [52]. Substitution of Leu 132 by Asn also completely inactivates ExbD [54]. Like TonB, ExbD contains important functional sites in the cytoplasmic membrane and in the periplasm. In contrast to TonB and ExbD, ExbB spans the cytoplasmic membrane three times (Fig. 9.2) and most of the protein is located in the cytoplasm [58]. The ExbB protein may be the key of the device that senses the electrochemical potential of the cytoplasmic membrane by the TonB-ExbB-ExbD protein complex. Although infection by phage T5 does not require TonB and phage T5 multiplies in tonB deletion mutants as well as in tonB + cells, certain fhuA point mutants display a strongly altered sensitivity to phage T5 when combined with certain tonB point mutants. For example, the sensitivity of cells that synthesize FhuA(L106P, DD348) to phage T5 is 100-fold less than the sensitivity of cells that synthesize wild type FhuA. Sensitivity is fully restored by replacing wild type TonB by TonB(R204H). By contrast, FhuA(L106P) confers full phage T5 sensitivity when combined with wild type TonB and is reduced three orders of magnitude when combined with TonB(G174R,V178I). These data are consistent with the proposal that various TonB derivatives impose distinct conformations on the FhuA variants that are differently suitable for phage T5 infection [56]. TonB mutant proteins and TonB proteins of different strains confer various sensitivities to TonB-dependent ligands. For example, TonB of Serratia marcescens [59] confers sensitivity of E. coli to colicins B and M, but not to colicin Ia [52, 55]. Transfer of energy from the cytoplasmic membrane into the outer membrane via the TonB-ExbB-ExbD protein complex can be envisioned to occur by an allosteric mechanism. TonB assumes an energized conformation through the action of ExbB and ExbD induced by the proton-motive force of the cytoplasmic membrane. Interaction of energized TonB with outer membrane receptor induces a conformational change that opens the receptor channel and lowers the affinity of the receptor for ferrichrome, albomycin, CGP 4832, colicin M, and mi177
9 Regulated Transport and Signal Transfer Channels crocin J25. The FhuA ligand is released from the receptor and diffuses through the channel into the periplasm, where ferrichrome and albomycin bind to the FhuD binding protein. It is not known whether diffusion through the open FhuA channel takes place vectorially, or whether the FhuA ligand can also be released into the culture medium, from where it has to adsorb again to FhuA to be transported across the outer membrane. At low concentrations, the rate of energy-dependent ferrichrome transport is much higher than the diffusion rate through the permanently open channels of the FhuA deletion derivatives. This could mean that the channel opens only towards the periplasm and prevents escape of the ferric siderophore into the culture medium. The higher rate could also be caused by binding of the ferric siderophore to wild type FhuA, which does not occur with the deletion derivatives. Binding of the ferrichrome to FhuD imposes a concentration gradient from the outside to the inside that facilitates diffusion from the cell surface into the periplasm. Binding of ferrichrome induces a conformational change in FhuA, as shown in vitro by the crystal structures of FhuA in the ferrichrome-loaded and unloaded form, in vivo by a change in the pattern of proteolytic FhuA degradation products [60], by prevention of trypsin cleavage at Lys-67 of isolated FhuA, and by inhibition of binding of certain monoclonal antibodies to FhuA [61]. It is conceivable that ferrichrome binding triggers energization via the Ton system. However, it is questionable whether this is a requirement for FhuA energization because all the other Ton-dependent FhuA ligands would also have to induce the same or a very similar FhuA conformation capable of interacting with the Ton system. The concept of how FhuA might be activated by the proton-motive force through the action of the TonB-ExbB-ExbD protein complex probably applies to all Ton-dependent active transport processes across the outer membrane depicted in Fig. 9.1.
9.4 Transport of ferrichrome across the cytoplasmic membrane
Three types of systems for transport of iron across the cytoplasmic membrane are found in bacteria. They mediate the import of ferrous iron, of ferric iron in the ionic form, and of ferric iron coupled to siderophores or heme. The latter two systems are members of the periplasmic-binding-protein-dependent transport (ABC transporter or traffic ATPase) composed of a periplasmic binding protein, one or two different integral membrane proteins, and one or two different ATPases that face the cytoplasm and supply the systems with energy. The almost identical design suggests a common origin of all ABC transporters. 178
9.4 Transport of ferrichrome across the cytoplasmic membrane The FhuD binding protein of E. coli accepts a number of structurally different siderophores of the hydroxamate type (e. g. ferrichrome, coprogen, aerobactin, ferrioxamine B, shizokinen, rhodotorulic acid) and the antibiotic albomycin [62, 63]. FhuD is synthesized as a precursor with a typical signal sequence and then is processed and exported into the periplasmic space [64–67]. Binding of iron(III) hydroxamates to the mature FhuD protein has been shown by three types of experiments. First, accumulation of [55Fe3+]-ferrichrome in the periplasm of intact cells was shown in an FhuD-overproducing strain, which, due to a mutation in the integral membrane protein FhuB, is unable to translocate the substrate into the cytoplasm. In a second assay, radiolabeled FhuD from the periplasm is protected against proteolytic degradation by proteinase K and trypsin. This protection is exclusively observed in the presence of those ferric hydroxamates that support growth of the bacterial cells under iron-limiting conditions [42]. In a third approach, FhuD was isolated and purified as a “His-tag“-labeled derivative on a Ni-chelate resin. The dissociation constants of ferric hydroxamates were estimated from the concentration-dependent decrease in the intrinsic fluorescence intensity of “His-tag”-FhuD: 0.4 µM for ferric aerobactin, 1.0 µM for ferrichrome, 0.3 µM for ferric coprogen, and 5.4 µM for the antibiotic albomycin. Ferrichrome A, ferrioxamine B, and ferrioxamine E, which are taken up poorly via the Fhu system, display dissociation constants of 79, 36 and 42 µM, respectively [63]. FhuD delivers the ferric hydroxamates to the FhuB transport protein in the cytoplasmic membrane. Three experimental approaches have indicated a physical interaction of FhuD with FhuB: 1) in spheroplasts, FhuD protects radioactively labeled overproduced FhuB from being degraded by trypsin and proteinase K [62], 2) “His-tag”-FhuD added to spheroplasts is chemically cross-linked to overproduced radiolabeled FhuB [63], and 3) peptides of 10 and 20 amino acid residues identical in sequence to two transmembrane regions, two periplasmic loops and two cytoplasmic loops (Fig. 9.3) bind specifically to FhuD and inhibit ferrichrome transport [68]. The competitive peptide binding experiments represent a novel approach in this field and were only possible after construction of FhuAD322–355 [19], which rendered the outer membrane permeable to the added peptides. These data also imply a model of the FhuB structure with an open cylinder at the periplasmic side and a closed configuration at the cytoplasmic side. FhuD inserts into the cylinder and delivers its substrates to FhuB. FhuD comes very close to the proposed binding side of FhuC (Fig. 9.3) and may trigger in the substrate-loaded form ATP hydrolysis by FhuC. Induction of ATP hydrolysis by substrate-loaded binding protein has been demonstrated with reconstituted maltose and histidine transport systems [69, 70]. However, the model derived from these studies proposes a transmembrane-signaling across the entire thickness of the cytoplasmic membrane to accommodate the activation of the cytoplasmic ATPase by the periplasmic binding protein. In our model, FhuD approaches FhuC in the FhuB transmembrane channel so closely that the proteins can interact directly or through a short FhuB peptide region. The transmembrane topology model of FhuB as shown in Fig. 9.3 is derived from genetically constructed FhuB-b-lactamase hybrid proteins in which 179
9 Regulated Transport and Signal Transfer Channels
Figure 9.3: Proposed transmembrane topology of the FhuB Fe3+-hydroxamate transporter in the cytoplasmic membrane [71]. The sites of interaction with the FhuD protein, as revealed by binding of synthetic FhuB peptides to isolated FhuD protein and competition of ferrichrome transport by the peptides [68], are indicated by black bars in the cytoplasmic membrane (CM), and by asterisks in periplasmic (PP) turns and in cytoplasmic (CP) turns.
increasing N-proximal segments of FhuB were fused to the BlaM b-lactamase devoid of its signal sequence. Cells become resistant to ampicillin when the fusion site is located in a periplasmic site of FhuB. In addition, a multiple sequence alignment of 33 transmembrane ABC transporter proteins was used to construct the model [71]. FhuB is unique among the integral membrane proteins in that it is about twice the size (70 kDa) of integral membrane proteins of other ABC transporters. It consists of two domains that display significant sequence similarity to each other. The FhuB halves are connected by a linker region of a few amino acid residues. The linker can be cleaved and the two halves assemble to form an active transporter [72]. Both halves, FhuB(N) and FhuB(C), are essential for transport; no activity is observed with one individually expressed domain. Several areas of striking sequence similarity are found in the primary structures of hydrophobic components of ABC transporters. The so-called “conserved region” (CR), located in the last third of the polypeptide chains, includes an invariant glycine residue at a distance of about 100 amino acids from the Cterminus [73, 74]. The conserved Gly is identical with (or corresponds to) a conserved Gly, contained in the “E A A - - - G - - - - - - - - - I - L P” motif defined by Dassa and Hofnung [75]. Point mutations at two corresponding glycine residues located within FhuB[N] at position 226 and FhuB[C] at position 595 decrease transport activity. The FhuC amino acid sequence contains two highly conserved sequences typical for ATPases, the so-called “Walker A” and “Walker B” consensus motifs. 180
9.5 Ferric-carboxylate transport system of Morganella morganii A total loss of function in all these FhuC derivatives suggests that FhuC indeed acts as an ATP-hydrolase, thereby energizing the transport process, most likely by inducing conformational changes in the components of the permease complex that might open a channel [76]. Interaction of FhuC with FhuB has been demonstrated by dominant negative effects on the transport of FhuC derivatives with single amino acid replacements in the putative ATP-binding domains [77].
9.5 Ferric-carboxylate transport system of Morganella morganii Volkmar Braun
M. morganii cells are able to take up the fungal siderophore rhizoferrin. Two genes essential for the utilization of this polyhydroxycarboxylate siderophore have been identified. They encode an outer membrane protein and a periplasmic protein named RumA and RumB, respectively (rhizoferrin uptake into Morganella). RumA displays striking sequence similarities to FecA from E. coli and to other TonB-dependent receptors; RumB shows similarity to binding proteins of the siderophore family. rumA and rumB have the same transcription polarity and are probably cotranscribed from an iron-regulated promoter upstream of rumA. A predicted Fur regulatory sequence upstream of rumA has been confirmed by using the Fur titration assay. Analysis of a 10-kb sequence flanking rumA and rumB upstream and downstream has revealed seven additional open reading frames for which no role in ferric rhizoferrin uptake can be discerned. Thus, rumA and rumB form an isolated operon, and additional genes required for the uptake of ferric rhizoferrin across the cytoplasmic membrane must map at chromosomal sites distinct from rumA and rumB [78].
181
9 Regulated Transport and Signal Transfer Channels
9.6 Transport of ferric iron ions by the Sfu system of Serratia marcescens Volkmar Braun
The first ferric iron transport system was characterized in Serratia marcescens (Sfu) [79]. In iron-depleted medium, the plasmid-encoded sfuABC genes confer growth to an E. coli K-12 strain that does not synthesize enterobactin, the only siderophore formed by wild type E. coli K-12, and to E. coli mutants unable to transport ferric enterobactin [80]. The Sfu system also stimulates growth of tonB and exbB mutants, and no genes encoding outer membrane proteins could be identified in the S. marcescens genome, thereby excluding an active ferric siderophore transport system across the outer membrane. Ferric iron solubilized in phosphate buffer or oxalate buffer is transported equally well, which indicates that there is no requirement for a specific ligand. At very low iron concentrations, ferric citrate stimulates iron uptake via the Sfu system in E. coli, provided that the FecA outer membrane protein and the Ton system are active. Under these conditions, ferric citrate is actively transported across the outer membrane by the ferric citrate transport system, and iron is released from citrate and is further transported by the Sfu system. Sequence analysis of a 4.8-kb fragment of the S. marcescens chromosome revealed three genes, sfuA, sfuB, and sfuC, which are arranged in an operon [79]. The sfuC start codon overlaps the sfuB stop codon, indicating a regulation of protein synthesis by translational coupling. Upstream of the sfu operon, a putative Fur box overlaps a putative promoter region. The Sfu proteins constitute a typical ABC transporter of which the SfuA protein is localized in the periplasm and the very hydrophobic SfuB protein is located in the cytoplasmic membrane; the SfuC protein is degraded too fast to be located, but contains the two Walker nucleotide binding motifs of bacterial traffic ATPases. Through sequence analysis, homologous genes organized in operons have been found in Neisseria gonorrhoeae (fbp) [81], Haemophilus influenzae (hit) [82], Yersinia enterocolitica (yfu) [83], and Actinobacillus pleuropneumoniae (afu) [84]. These systems transport ferric iron. In Neisseria meningitidis, DNase I footprinting experiments suggest the binding of two Fur repressor dimers upstream of fbpA at two putative Fur consensus sequences that overlap the potential –35 region [85].
182
References
Acknowledgments
The authors thank the many coworkers listed in the reference list who made important contributions to the results described in these chapters. In particular we acknowledge the original work of Wolfgang Köster on transport across the cytoplasmic membrane. The financial support of the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie made this work possible.
References
1. Hantke, K. and Braun, V. (1975) Membrane receptor dependent iron transport in Escherichia coli. FEBS Microbiol. Lett. 49, 301–305. 2. Hancock, R. E. and Braun, V. (1976) Nature of the energy requirement for the irreversible adsorption of bacteriophages T1 and f80 to Escherichia coli. J. Bacteriol. 125, 409–415. 3. Hantke, K. (1990) Dihydroxybenzoylserine – a siderophore for E. coli. FEMS Microbiol. Lett. 67, 5–8. 4. Braun, V. and Hantke, K. (1991) Genetics of bacterial iron transport. In: Handbook of microbial iron chelates (Winkelmann, G., ed.) pp 107–138. 5. Crosa, J. H. (1984) The relationship of plasmid-mediated iron transport and bacterial virulence. Annu. Rev. Microbiol. 38, 69–89. 6. Expert, D., Enard, C., and Masclaux, C. (1996) The role of iron plant host-pathogen interactions. Trends Microbiol. 4, 232–236. 7. Genco, C. A. and Desai, P. J. (1996) Iron acquisition in the pathogenic Neisseria. Trends Microbiol. 4, 179–184. 8. Cornelissen, C. N. and Sparling, F. P. (1994) Iron piracy: acquisition of transferrinbound iron by bacterial pathogens. Mol. Microbiol. 14, 843–850. 9. Kingsley, R., Rabsch, W., Stephens, P., Roberts, M., Reissbrodt, R., and Williams, P. H. (1995) Iron supplying systems of Salmonella in diagnostics, epidemiology and infection. FEMS Immunol. Med. Microbiol. 11, 257–264. 10. Gray-Owen, S. D. and Schryvers, A. B. (1996) Bacterial transferrin and lactoferrin receptors. Trends Microbiol. 4, 185–191. 11. Tao, X., Schiering, N., Zeng, H.-Y., Ringe, D., and Murphy, J. R. (1994) Iron, DtxR, and the regulation of diphtheria toxin expression. Mol. Microbiol. 14, 191–197. 12. Payne, S. M. (1993) Iron acquisition in microbial pathogenesis. Trends Microbiol. 1, 66–69. 13. Wooldrigde, K. G. and Williams, P. H. (1993) Iron uptake mechanisms of pathogenic bacteria. FEMS Microbiol. Rev. 12, 325–348. 14. Perry, R. D. (1993) Acquisition and storage of inorganic iron and hemin by the yersiniae. Trends Microbiol. 1, 142–147. 15. Nikaido, H. (1996) Outer membrane. In: Escherichia coli and Salmonella. (Neidhardt, F. C., ed.), ASM Press, Washington, D.C., pp 29–47.
183
9 Regulated Transport and Signal Transfer Channels 16. Schneider, R. and Hantke, K. (1993) Iron-hydroxamate uptake systems in Bacillus subtilis: identification of a lipoprotein as a part of a binding protein dependent transport system. Mol. Microbiol. 8, 111–121. 17. Kadner, R. J., Heller, K., Coulton, J. W., and Braun, V. (1980) Genetic control of hydroxamate-mediated iron uptake in Escherichia coli. J. Bacteriol. 143, 256–264. 18. Luckey, M., Wayne, R., and Neilands, J. B. (1975) In vitro competition between ferrichrome and phage for the outer membrane T5 receptor complex of Escherichia coli. Biochem. Biophys. Res. Commun. 64, 687–693. 19. Killmann, H., Benz, R., and Braun, V. (1993) Conversion of the FhuA transport protein into a diffusion channel through the outer membrane of Escherichia coli. EMBO J. 12, 3007–3016. 20. Koebnik, R. and Braun, V. (1993) Insertion derivatives containing segments of up to 16 amino acids identify surface- and periplasm-exposed regions of the FhuA outer membrane receptor of Escherichia coli K-12. J. Bacteriol. 175, 826–839. 21. Bös, C., Lorenzen, D., and Braun, V. (1997) Specific in vivo labeling of cell-surface-exposed protein loops: reactive cysteines in the predicted gating loop indicate ferrichrome binding to the FhuA protein of Escherichia coli K-12. J. Bacteriol. 180, 605– 613. 22. Killmann, H. and Braun, V. (1992) An aspartate deletion mutation defines a binding site of the multifunctional FhuA outer membrane receptor of Escherichia coli K-12. J. Bacteriol. 174, 3479–3486. 23. Killmann, H., Videnov, G., Jung, G., Schwarz, H., and Braun, V. (1995) Identification of receptor binding sites by competitive peptide mapping: phages T1, T5, and f80 and colicin M bind to the gating loop of FhuA. J. Bacteriol. 177, 694–698. 24. Boulanger, P., le Maire, M., Bonhivers, M., Dubois, S., Desmadril, M., and Letellier, L. (1996) Purification and structural and functional characterization of FhuA, a transporter of the Escherichia coli outer membrane. Biochemistry 35, 14216–14224. 25. Burkhardt, R. (1988) Molekulare Charakterisierung der Gene fhuA, fhuC und fhuD des Ferri-Hydroxamat-Transportsystems bei E. coli. Thesis work, University of Tübingen. 26. Coulton, J. W., Mason, P., Cameron, D. R., Carmel, G., Jean, R., and Rode, H. N. (1986) Protein fusions of b-galactosidase to the ferrichrome-iron receptor of Escherichia coli K-12. J. Bacteriol. 165, 181–192. 27. Killmann, H., Herrmann, C., Wolff, H., and Braun, V. (1998) Identification of a new site for ferrichrome transport by comparison of the FhuA proteins of Escherichia coli, Salmonella paratyphi B, Salmonella typhimurium, and Pantoea agglomerans. J. Bacteriol. 180, 3845–3852. 28. Ferguson, A. D., Hofmann, E., Coulton, J. W., Diederichs, K., and Welte, W. (1998) Siderophore-mediated iron transport: crystal structure of FhuA with bound lipopolysaccharide. Science 282, 2215–2220. 29. Braun, V. (1998) Pumping iron through cell membranes. Science 282, 2202–2203. 30. Locher, K. P., Rees, B., Koebnik, R., Mitschler, A., Moulinier, L., Rosenbusch, J. P., and Moras, D. (1998) Transmembrane signaling across the ligand-gated FhuA receptor: crystal structures of free and ferrichrome-bound states reveal allosteric changes. Cell 95, 771–778. 31. Ferguson, A. D., Braun, V., Fiedler, H.-P., Coulton, J. W., Diederichs, K., and Welte, W. (2000) Crystal structure of the antibiotic albomycin in complex with the outer membrane transporter FhuA. Submitted. 32. Wehrli, W., Zimmermann, W., Klump, W., Tosch, W., Fischer, W, and Zak, O. (1987) CGP 4832, a semisynthetic rifamycin derivative highly active against some Gram-negative bacteria. J. Antibiotics 40, 11733–1739. 33. Kadner, R. J., Heller, K., Coulton, J. W., and Braun, V. (1980) Genetic control of hydroxamate-mediated iron uptake in Escherichia coli. J. Bacteriol. 143, 256–264.
184
References 34. Pugsley, P. A., Zimmermann, W., and Wehrli, W. (1987) Highly efficient uptake of a rifamycin derivative via FhuA-TonB-dependent uptake route in Escherichia coli. J. Gen. Microbiol. 133, 3505–3511. 35. Ferguson, A. D., Ködding, J., Bös, C., Diederichs, K., Walker, G., Coulton, J. W., Braun, V., and Welte, W. (2000) Crystal structure of a semi-synthetic rifamycin derivative in complex with the active outer membrane transporter FhuA from Escherichia coli K-12. Submitted. 36. Locher, K. P. and Rosenbusch, J. P. (1997) Oligomeric states and siderophore binding of the ligand-gated FhuA protein that forms channels across the Escherichia coli outer membranes. Eur. J. Biochem. 247, 770–775. 37. Braun, M., Killmann, H., and Braun, V. (1999) The b-barrel domain of FhuAD5–160 is sufficient for all TonB-dependent FhuA activities of Escherichia coli. Mol. Microbiol. 33, 1037–1049. 38. Killmann, H., Benz, R., and Braun, V. (1996) Properties of the FhuA channel in the Escherichia coli outer membrane after deletion of FhuA portions within and outside the predicted gating loop. J. Bacteriol. 178, 6913–6920. 39. Braun, V. and Wolff, H. (1973) Characterization of the receptor protein for phage T5 and colicin M in the outer membrane of E. coli B. FEBS Lett. 34, 77–80. 40. Braun, V., Schaller, K., and Wolff, H. (1973) A common receptor protein for phage T5 and colicin M in the outer membrane of Escherichia coli B. Biochem. Biophys. Acta 323, 87–97. 41. Hantke, K. and Braun, V. (1978) Functional interaction of the tonA/tonB receptor system in Escherichia coli. J. Bacteriol. 135, 190–197. 42. Köster, W. and Braun, V. (1990) Iron(III)hydroxamate transport into Escherichia coli. Substrate binding to the periplasmic FhuD protein. J. Biol. Chem. 265, 21407–21410. 43. Braun, V. (1989) The structurally related exbB and tolQ genes are interchangeable in conferring tonB-dependent colicin, bacteriophage, and albomycin sensitivity. J. Bacteriol. 171, 6387–6390. 44. Braun, V. and Herrmann, C. (1993) Evolutionary relationship of uptake systems for biopolymers in Escherichia coli: cross-complementation between the TonB-ExbBExbD and the TolA-TolQ-TolR proteins. Mol. Microbiol. 8, 261–268. 45. Bradbeer, C. (1993) The proton motive force drives the outer membrane transport of cobalamin in Escherichia coli. J. Bacteriol. 175, 3146–3150. 46. Skare, J. T., Ahmer, B. M. M., Seachord, C. L., Darveau, R. P., and Postle, K. (1993) Energy transduction between membranes. J. Biol. Chem. 268, 16302–16308. 47. Heller, K. J., Kadner, R. J., and Günter, K. (1988) Suppression of the btuB451 mutation by mutations in the tonB gene suggests a direct interaction between TonB and TonB-dependent receptor proteins in the outer membrane of Escherichia coli. Gene 64, 147–153. 48. Schöffler, H. and Braun, V. (1989) Transport across the outer membrane of Escherichia coli K-12 via the FhuA receptor is regulated by the TonB protein of the cytoplasmic membrane. Mol. Gen. Genet. 217, 378–383. 49. Bell, P. E., Nau, C. D., Brown, J. T., Konisky, J., and Kadner, R. J. (1990) Genetic suppression demonstrates interaction of TonB protein with outer membrane transport proteins in Escherichia coli. J. Bacteriol. 172, 3826–3829. 50. Günter, K. and Braun,V. (1990) In vivo evidence for FhuA outer membrane receptor interaction with the TonB inner membrane protein of Escherichia coli. FEBS Lett. 274, 85–88. 51. Fischer, E., Günter, K., and Braun, V. (1989) Involvement of ExbB and TonB in transport across the outer membrane of Escherichia coli: phenotypic complementation of exb mutants by overexpressed tonB and physical stabilization of TonB by ExbB. J. Bacteriol. 171, 5127–5134. 52. Traub, I., Gaisser, S., and Braun, V. (1993) Activity domains of the TonB protein. Mol. Microbiol. 8, 409–423.
185
9 Regulated Transport and Signal Transfer Channels 53. Larsen, R. A., Thomas, M. G., Wood, G. E., and Postle, K. (1994) Partial suppression of an Escherichia coli TonB transmembrane domain mutation (DV17) by a missense mutation in ExbB. Mol. Microbiol. 13, 627–640. 54. Braun, V., Gaisser, S., Herrmann, C., Kampfenkel, K., Killmann, H., and Traub, I. (1996). Energy-coupled transport across the outer membrane of Escherichia coli: ExbB binds ExbD and TonB in vitro, and leucine132 in the periplasmic region and aspartate25 in the transmembrane region are important for ExbD activity. J. Bacteriol. 178, 2836–2845. 55. Traub, I. and Braun, V. (1994) Energy-coupled colicin transport through the outer membrane of Escherichia coli K-12: mutated TonB proteins alter receptor activities and colicin uptake. FEMS Microbiol. Lett. 119, 65–70. 56. Killmann, H. and Braun, V. (1994) Energy-dependent receptor activities of Escherichia coli K-12: mutated TonB proteins alter FhuA receptor activities to phages T5, T1, f80 and to colicin M. FEMS Microbiol. Lett. 119, 71–76. 57. Kampfenkel, K. and Braun, V. (1992) Membrane topology of the Escherichia coli ExbD protein. J. Bacteriol. 174, 5485–5487. 58. Kampfenkel, K. and Braun, V. (1993) Topology of the ExbB protein in the cytoplasmic membrane of Escherichia coli. J. Biol. Chem. 268, 6050–6057. 59. Gaisser, S. and Braun, V. (1991) The tonB gene of Serratia marcescens: sequence, activity and partial complementation of Escherichia coli tonB mutants. Mol. Microbiol. 5, 2777–2787. 60. Hoffmann, H., Fischer, E., Schwarz, H., and Braun, V. (1986) Overproduction of the proFhuA outer membrane receptor protein of Escherichia coli K-12: isolation, properties, and immunocytochemical localization at the inner side of the cytoplasmic membrane. Arch. Microbiol. 145, 334–341. 61. Moeck, G. S., Tawa, P., Xiang, H., Ismail, A. A., Turnbull, J. L., and Coulton, J. W. (1996) Ligand-induced conformational change in the ferrichrome-iron receptor of Escherichia coli K-12. Mol. Microbiol. 22, 459–471. 62. Rohrbach, M. R., Braun, V., and Köster, W. (1995) Ferrichrome transport in Escherichia coli K-12: altered substrate specificity of mutated periplasmic FhuD and interaction of FhuD with the integral membrane protein FhuB. J. Bacteriol. 177, 7186–7193. 63. Rohrbach, M. R., Paul, S., and Köster, W. (1995) In vivo reconstitution of an active siderophore transport system by a binding protein derivative lacking a signal sequence. Mol. Gen. Genet. 248, 33–42. 64. Fecker, L. and Braun, V. (1983) Cloning and expression of the fhu genes involved in iron(III)-hydroxamate uptake by E. coli. J. Bacteriol. 156, 1301–1314. 65. Burkhardt, R. and Braun, V. (1987) Nucleotide sequence of fhuC and fhuD genes involved in iron(III)-hydroxamate transport: domains in FhuC homologous to ATP binding proteins. Mol. Gen. Genet. 209, 49–55. 66. Coulton, J. W., Mason, P., and Allat, D. D. (1987) fhuC and fhuD genes for iron(III)ferrichrome transport into Escherichia coli K-12. J. Bacteriol. 169, 3844–3849. 67. Köster, W. and Braun, V. (1989) Iron-hydroxamate transport into Escherichia coli K12: localization of FhuD in the periplasm and of FhuB in the cytoplasmic membrane. Mol. Gen. Genet. 217, 233–239. 68. Mademidis, A., Killmann, H., Kraas, W., Flechsler, I., Jung, G., and Braun, V. (1997) ATP-dependent transporters: competitive peptide mapping reveals functional binding sites of the periplasmic FhuD protein at a periplasmic and a transmembrane region extending in a cytoplasmic loop of the integral FhuB membrane protein of the ferric hydroxamate transport system in Escherichia coli. Mol. Microbiol. 26, 1109–1123. 69. Bishop, L., Agbayani, Jr., Ambudkar, S. V., Maloney, P. C., and Ames, G. F.-L. (1989) Reconstitution of a bacterial periplasmic permease in proteoliposomes and demonstration of ATP hydrolysis concomitant with transport. Proc. Natl. Acad. Sci. USA 86, 6953–6957.
186
References 70. Davidson, A. L., Shuman, H. A., and Nikaido, H. (1992) Mechanism of maltose transport in Escherichia coli: transmembrane signaling by periplasmic binding proteins. Proc. Natl. Acad. Sci. USA 89, 2360–2364. 71. Groeger, W. and Köster, W. (1998) Transmembrane topology of the two FhuB domains representing the hydrophobic components of bacterial ABC transporters involved in the uptake of siderophores, heme and vitamin B12 . Microbiology 144, 2759–2769. 72. Köster, W. and Braun, V. (1990) Iron(III)hydroxamate transport of Escherichia coli: restoration of iron supply by coexpression of the N- and C-terminal halves of the cytoplasmic membrane protein FhuB cloned on separate plasmids. Mol. Gen. Genet. 223, 379–384. 73. Köster, W. (1991) Iron(III)hydroxamate transport across the cytoplasmic membrane of Escherichia coli. Biol. Metals 4, 23–32. 74. Saurin, W., Köster, W., and Dassa, E. (1994) Bacterial binding protein-dependent permeases: characterization of distinctive signatures for functionally related integral membrane proteins. Mol. Microbiol. 12, 993–1004. 75. Dassa, E. and Hofnung, M. (1985) Sequence of malG gene in E. coli K 12: homologies between integral membrane components from binding protein-dependent transport systems. EMBO J. 4, 2287–2293. 76. Becker, K., Köster, W., and Braun, V. (1990) Iron(III)hydroxamate transport of Escherichia coli K-12: single amino acid replacements at potential ATP-binding sites inactivate the FhuC protein. Mol. Gen. Genet. 223, 159–162. 77. Schultz-Hauser, G., Köster, W., Schwarz, H., and Braun, V. (1992) Iron(III)hydroxamate transport in Escherichia coli K-12. FhuB-mediated membrane association of the FhuC protein and negative complementation of fhuC mutants. J. Bacteriol. 174, 2305–2311. 78. Kühn, S., Braun, V., and Köster, W. (1996) Ferric rhizoferrin uptake into Morganella morganii: characterization of genes involved in the uptake of a polyhydroxycarboxylate siderophore. J. Bacteriol. 178 (2), 496–504. 79. Angerer, A., Gaisser, S., and Braun, V. (1990) Nucleotide sequences of the sfuA, sfuB, and sfuC genes of Serratia marcescens suggest a periplasmic-binding-protein-dependent iron transport mechanism. J. Bacteriol. 172 (2), 572–578. 80. Zimmermann, L., Angerer, A., and Braun, V. (1989) Mechanistically novel iron(III) transport system in Serratia marcescens. J. Bacteriol. 171, 238–243. 81. Adhikari, P., Berish, S. A., Nowalk, A. J., Veraldi, K. L., Morse, S. A., and Mietzner, T. A. (1996) The fbpABC locus of Neisseria gonorrhoeae functions in the periplasm-tocytosol transport of iron. J. Bacteriol. 178, 2145–2149. 82. Adhikari, P., Kirby, S. D., Nowalk, A. J., Veraldi, K. L., Schryvers, A. B., and Mietzner, T. A. (1995) Biochemical characterization of a Haemophilus influenzae periplasmic iron transport operon. J. Biol. Chem. 270 (42), 25142–25149. 83. Saaken, E. M. and Heesemann, J. (1997) unpublished. 84. Chin, N., Frey, J., Chang, C.-F., and Chang, Y.-F. (1996) Identification of a locus involved in the utilization of iron by Actinobacillus pleuropneumoniae. FEMS Microbiol. Lett. 143, 1–6. 85. Desai, P. J., Angerer, A., and Genco, C. A. (1996) Analysis of Fur binding to operator sequences within the Neisseria gonorrhoeae fbpA promoter. J. Bacteriol. 178 (16), 5020–5023.
187
10 Iron Transport in Gram-negative and Gram-positive Bacteria Klaus Hantke*
10.1 Ferric iron transport in bacteria
10.1.1 Siderophore-iron uptake in Escherichia coli In the 1970s it was observed that six proteins in the outer membrane of E. coli K-12 in the range of 74 to 83 kDa were overproduced when the cells were grown under low-iron stress conditions. All these proteins, as we now know, are receptors for siderophores (iron carriers), and some are also receptors for phages and/or colicins: FhuA for ferrichrome, phage T1, and colicin M (see Chapter 9); FepA for enterochelin (enterobactin) and colicin B; FecA for ferric citrate (see Chapter 11); and Cir for colicin I [1]. Later, in a systematic study of iron-regulated genes, FhuE, the receptor protein for the fungal siderophore coprogen, and Fiu, the receptor for catecholates and colicins G and H, were identified [2]. Analysis of the nucleotide sequence of the E. coli genome has revealed a seventh siderophore receptor gene, whose gene product possibly corresponds to a minor protein band observed between FhuE and FhuA [2]. This protein has not been characterized. In a BLAST search, the closest relative of this putative receptor was shown to be FyuA, the receptor for yersiniabactin, a siderophore of Yersinia pestis (see Section 10.1.2). When these studies first began, it was a surprise to find so many different iron-siderophore transport systems in one organism. Since then, a similar multitude of ferric iron transport systems has been found in most other bacteria that live under oxic conditions. More than 110 siderophore type receptors are now found in the NCBI database of non-redundant protein sequences – another indication of the ubiquity of this protein type in Gram-negative bacteria.
* Mikrobiologie/Membranphysiologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen, Germany
188 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
10.1 Ferric iron transport in bacteria 10.1.1.1 FhuE is a receptor for ferri-coprogen The FhuE protein, a receptor for coprogen, has been characterized by sequencing and transport studies [3, 4]. It is similar to FhuA and other hydroxamate and pyochelin siderophore receptors. Comparison of the amino acid sequence of FhuE with that of other receptors led to the discovery of an important binding site, the TonB box. The TonB protein interacts with all siderophore receptors in the outer membrane to transduce energy from the cytoplasmic membrane to the receptor to allow release of the bound siderophore into the periplasm. A site for the TonB interaction, the TonB box, has been identified in the N-terminal end of the receptors (see Chapter 9). A mutation,V8P, in the TonB box of FhuE leads to the inability of the cells to grow on coprogen or ferrioxamine B. Different suppressor mutations were isolated in vivo by selecting for growth on coprogen as sole iron source. Interestingly, the suppressor mutations isolated are found only in FhuE (V8L, V8Q, and V8R), and not in TonB, in contrast to the case with FhuA and BtuB (see Chapter 9).
10.1.1.2 Ferric-catecholate transport systems Most E. coli strains and many Enterobacteriaceae produce the catecholate-type siderophore enterochelin (enterobactin), which is a cyclic trimer composed of 2,3dihydroxy-N-benzoyl-serine. Enterochelin transport has mainly been studied by groups in the USA [5]. We have investigated the functions of the two iron-regulated outer membrane proteins, Cir and Fiu, which were originally defined as colicin receptors of E. coli [6, 2]. Their in vivo function is the uptake of ferric catecholates and ferric dihydroxybenzoyl serine, a degradation product of enterochelin [7]. Both receptors recognize an astonishingly broad spectrum of catecholate derivatives, including catecholate-containing cephalosporin derivatives that are transported into the periplasm, the location of their action. Transport of the catecholate-cephalosporins as opposed to diffusion of the cephalosporins reduces the minimal inhibitory concentration about 100-fold [4]. Receptors with a similar substrate specificity are present in Yersinia (see Section 10.1.2) and many Gram-negative bacteria. In Pseudomonas aeruginosa these receptors make the cephalosporin-resistant strains highly sensitive to catecholate-cephalosporins [8].
10.1.2 Iron uptake in Yersinia enterocolitica Y. enterocolitica is a close, but less virulent relative of Yersinia pestis, the causative agent of bubonic plague. Highly virulent serotypes (O8 and O13) and less virulent serotypes (O3 and O9) of Y. enterocolitica have been distinguished. One reason for these differences in virulence is the ability to produce a siderophore, which we have called yersiniabactin. 189
10 Iron Transport in Gram-negative and Gram-positive Bacteria The existence of this siderophore has been a matter of debate. In 1975, Wake et al. [9] claimed that Y. pestis produces a siderophore related to virulence. However, this was questioned by a study where it was not possible to isolate a siderophore with biological activity. Later it was shown in a plate test that siderophore production of Y. enterocolitica is correlated to virulence [10]. The characterization of the siderophore yersiniabactin specifically produced by virulent Yersinia is described below.
10.1.2.1 Siderophore-mediated iron uptake 10.1.2.1.1 Yersiniabactin The siderophore produced by Y. enterocolitica WA-C O8 is unable to crossfeed siderophore-dependent E. coli K-12 strains. In collaboration with H. Zähner and P. Fiedler (see Chapter 2), H. Haag isolated and purified the siderophore yersiniabactin, which was shown to be relatively labile. We isolated a fur mutant using the manganese enrichment technique [11] described in Section 10.3.1. Fur is the iron(II)-dependent repressor of iron uptake systems and also regulates siderophore biosynthesis. As expected, the fur mutant of Y. enterocolitica constitutively produced yersiniabactin. For wild type strains, it is difficult to maintain low-iron growth conditions for optimal siderophore production without retarding growth. In contrast, the fur mutant allowed siderophore production under ironrich growth conditions, which was helpful for the large-scale productions necessary for the structural elucidation of yersiniabactin. The structure of yersiniabactin (Fig. 10.1) was determined in close collaboration and is described in Chapters 2 and 19. The purified siderophore was used to prove that 1) yersiniabactin is a siderophore of Y. enterocolitica, and 2) the receptor FyuA is a 65-kDa protein [12]. Pesticin is a colicin-like protein, and its production had been used in the early 1950s as a characteristic marker of Y. pestis. It was observed that virulent Yersinia were attenuated when they became pesticin resistant. This is now explained by the genetic instability of the virulence island encoding yersiniabactin synthesis and transport. Rare strains of E. coli are sensitive to pesticin. A tonB mutation renders such a strain insensitive to pesticin, which provided a hint that the receptor was a siderophore receptor. Using pesticin-resistant mutants of Y. enterocolitica, it was possible to identify the yersiniabactin siderophore receptor FyuA as the pesticin receptor [12, 13]. Pesticin, with its mode of action, belongs to the rare bacteriocins with a muramidase activity [14]. In line with this, the pesticin immunity protein, which protects the producing cells from the toxic action of pesticin, is found in the periplasm [15]. The pesticin plasmid is a small, colicin-E1-like plasmid that also encodes the plasminogen activator that may be responsible for the highly invasive fulminate character of Y. pestis infections. One hypothesis is that the bacteriocin stabilizes the presence of the plasmid in the population.
190
10.1 Ferric iron transport in bacteria
Figure 10.1: Structures of the siderophores yersiniabactin from Y. enterocolitica and Y. pestis, anguibactin from V. anguillarum, and the fungal siderophore desferri-ferrichrome.
10.1.2.1.2 Ferrichrome Like many other bacteria, Y. enterocolitica is able to utilize ferrichrome (Fig. 10.1) as an iron source. Using the antibiotic albomycin, a structural analogue of ferrichrome, it was possible to isolate a mutant unable to utilize ferrichrome. The mutation is in the fcuA gene, and the FcuA protein (approximately 70 kDa) is found in the outer membrane. The ferrichrome receptors of the Enterobacteriaceae Salmonella typhimurium, Salmonella paratyphi, Pantoea agglomerans (formerly Enterobacter agglomerans), and E. coli have been compared and are very similar [16]. The sequence analysis of the receptor FcuA of Y. enterocolitica provided surprising results. The closest relative of FcuA in the siderophore receptor protein family was AngR, the receptor for anguibactin in Vibrio anguillarum. Anguibactin is a siderophore with similarities to yersiniabactin (Fig. 10.1), which is not structurally related to ferrichrome [17]. This is an example where a few changes in the receptor may lead to a new receptor specificity, illustrating that it is not always possible to predict the substrate specificity based on sequence similarities. 191
10 Iron Transport in Gram-negative and Gram-positive Bacteria 10.1.2.1.3 Catecholate Y. enterocolitica does not produce enterochelin, the characteristic siderophore found in many enterobacteria. Recent studies [19] have indicated that Y. enterocolitica WA contains only a truncated fep-ent operon, which may allow transport of catecholes but not synthesis of enterochelin. This explains the fact that Y. enterocolitica is able to utilize several catecholes. A catecholate-specific receptor mutant with a mutation in the gene cccA has been isolated using a catecholatecephalosporin. The receptor protein is in the outer membrane and has an apparent molecular mass of 65 kDa [18]. In the Sanger sequencing project of Y. pestis an open reading frame with similarity to Cir, a catecholate receptor of E. coli, is found which may be the cccA gene. 10.1.2.1.4 Ferrioxamine Ferrioxamines are siderophores produced by many Streptomyces strains and also by some Gram-negative species, such as Pseudomonas putida and the Enterobacteriaceae Pantoea agglomerans and Hafnia alvei. Ferrioxamine B enhances the virulence of the Y. enterocolitica serotypes O3 and O9 in a mouse infection model, but has no effect on the pathogenicity of highly virulent strains. This has been interpreted to mean that these highly virulent strains gain enough iron through their yersiniabactin transport system for growth in their host. In humans, Desferal (mesylate salt of desferri-ferrioxamine B) is used therapeutically to chelate iron in iron-overload disease. Such patients are prone to Yersinia infections, which may be the result of two synergistic effects of ferrioxamine B: 1) the stimulation of growth of Y. enterocolitica by increasing the iron supply via ferrioxamine B and 2) the immunosuppressing nature of Desferal, which weakens the host defense [20]. The ferrioxamine receptor gene (foxA) of Y. enterocolitica has been cloned and sequenced, and FoxA has been found to be similar to FhuA, the ferrichrome receptor of E. coli. The sequence similarities to porins and some rules derived from the porin structural studies have been used to propose the first bbarrel model of a siderophore receptor [21], which now has to be revised in the light of the FhuA and FepA crystal structures (see Chapter 9).
10.1.2.2 Heme-iron uptake in Y. enterocolitica In vertebrates, heme is the most abundant iron carrier because of the high hemoglobin content of erythrocytes. Heme is released by tissue damage and desquamation of epithelial cells, which is important for bacteria colonizing the mucosa. Heme can be used as an iron source by several bacteria [22]. The uptake systems of mainly pathogenic bacteria have been studied and have revealed an astonishing multitude of substrates and ways how bacteria extract heme from different sources. An overview is given in [23]. We characterized the first heme uptake system of Y. enterocolitica at the molecular level. This system strongly resembles a typical siderophore uptake 192
10.1 Ferric iron transport in bacteria system [24, 25]. The six genes hemPRSTUV seem to constitute an operon. HemP is a small protein that may have regulatory functions, and HemR is the outer membrane receptor for heme. The sequence of HemR shows similarities to the TonB-dependent siderophore receptors. In addition, it has been shown with a Y. enterocolitica tonB mutant that the heme uptake is TonB dependent [24]. HemS most likely helps to degrade heme and to liberate the iron inside the cell. The three genes hemTUV code for a binding-protein-dependent transport system across the cytoplasmic membrane. HemT is the periplasmic binding protein, HemU is the integral membrane protein, and HemV provides the energy as an ATPase for the transport process [25]. Heme transport systems from E. coli O157:H7 [27], Shigella dysenteriae serotype 1 [26] and Vibrio cholerae [28], which are similar to the heme transport system of Y. enterocolitica have been cloned and sequenced.
10.1.3 Hydroxamate-iron transport in Bacillus subtilis Gram-positive bacteria such as B. subtilis do not contain an outer membrane and therefore also have no periplasm between an outer membrane and a cytoplasmic membrane. The question has arisen whether they contain iron transport systems of the ABC transporter type, which include a periplasmic binding protein in Gram-negative bacteria. The genes of the B. subtilis fhu operon, which encode the ferrichrome transport system, are organized in two divergent transcription units with overlapping promoter regions that are regulated by iron and Fur. fhuD encodes a binding protein with a signal sequence and a cysteine residue at the N-terminus of the mature protein, which are typical for lipoproteins of the murein lipoprotein type. It is therefore assumed that the FhuD protein functions as a binding protein and is anchored to the cell surface by a covalently linked lipid. fhuD is oriented in the direction opposite to that of fhuB, fhuG, and fhuC, which most likely form an operon. Unlike transport via FhuB in E. coli, transport across the cytoplasmic membrane of B. subtilis is mediated by two integral membrane proteins, FhuB and FhuG, each of which is half the size of the E. coli FhuB. fhuC, the last gene in the operon, encodes an ATPase. A second putative siderophore binding protein has been detected by sequence analysis. The high sequence similarity to FhuD suggests a function as a binding protein for an iron(III) hydroxamate, most likely ferrioxamine B, or possibly schizokinen.
193
10 Iron Transport in Gram-negative and Gram-positive Bacteria
10.2 Ferrous-iron transport systems (Feo) of E. coli
Enterobacteriaceae are facultatively anaerobic bacteria. Under anoxic conditions, ferrous iron is stable, and many bacteria reduce ferric iron to ferrous iron under anoxic conditions. Fe2+ is more soluble than Fe3+ and this allows its transport without being complexed by ligands. A ferrous-iron transport system has been found in E. coli, but the encoding genes could not be cloned on a plasmid, most likely due to the high toxicity of Fe2+ under oxic conditions. Three genes, feoABC, have been identified [29]. feoA and feoC encode two small proteins of unknown function, each with a molecular mass <10 kDa. feoB encodes an 84-kDa protein located in the cytoplasmic membrane. Analysis of the FeoB sequence shows a typical nucleotide-binding motif at the N-terminal end of the protein, which led to the assumption that ferrous iron uptake is driven by ATP hydrolysis. More recent comparisons of this domain with proteins in the databases revealed similarities to the Era/Obg/Ras protein family, some members of which have been shown to be GTP-binding proteins. If GTP is recognized by FeoB, it will be interesting to determine whether this domain has an energizing or, more likely, a regulatory function. Mutants in feoA or feoB take up ferrous iron at a greatly reduced rate. In addition, feo mutants are derepressed for many Fur-regulated genes. These results indicate that ferrous iron transport contributes to the iron supply of the cells also under oxic conditions [29]. In a mouse model, it has been shown that feo helps E. coli to colonize the gut [30]. feo genes have also been identified in S. typhimurium [31] and in some other Gram-negative bacteria, and genes similar to feoB are found in Methanococcus janaschii [32] and other anaerobic archaea. Mg2+ is taken up by the constitutively expressed CorA protein, which is located in the cytoplasmic membrane and which has been studied in E. coli and in S. typhimurium [33]. As the name of the gene corA (cobalt resistance) indicates, cobalt can be used for the selection of mutations in this gene. This reflects that also Co2+, Mn2+, and Ni2+ can be taken up by this transport system; this may be toxic for the cells when high concentrations of these metals are in the medium. Ferrous iron is also a substrate for this transporter, which may be the often-mentioned “low affinity” iron uptake system of E. coli [34].
194
10.3 Regulation of iron transport and metabolism
10.3 Regulation of iron transport and metabolism
Empirically, it has long been known that the iron supply of bacteria may have a strong influence on the metabolism of bacterial cells. One famous example is the observation in the 1930s that low iron concentrations stimulate the production of diphtheria toxin by Corynebacterium diphtheriae [35]. This type of regulation was later also observed for many other toxins and virulence factors of pathogenic bacteria [36]. Another early observation was that under iron-limiting conditions, many microorganisms secrete chelating substances that produce red to brown colored iron complexes. When the function of these substances as Fe3+-carriers was recognized, they were termed siderophores [37]. How the cells sense their iron needs became clear with the introduction of molecular biology techniques. The first mutation in the iron regulatory gene fur (ferric iron uptake regulation) was isolated in 1978 in S. typhimurium [38]. Later we isolated a similar mutant in E. coli [39], which allowed the cloning and sequencing of the fur gene [40]. The development of the reporter gene technology by Casadaban [41] was important for the study of iron regulation. The expression of the lac gene is put under the control of the promoter of the gene of interest. The regulation of this reporter gene under different growth conditions, for instance +/– iron, can be seen directly on MacConkey lactose plates or with X-gal as an indicator of b-galactosidase activity. A promoter of an iron uptake gene responds to low-iron growth conditions by greatly increasing expression – the colonies of the clone are red on MacConkey lactose plates due to the increased expression of the lac gene under the control of the promoter; with high-iron conditions, the colonies are white [42]. Mutants that produce red colonies on high-iron MacConkey plates contain mutations either in the regulatory gene fur or in the feo genes, which encode a ferrous-iron transport system.
10.3.1 Regulation of gene transcription by Fe2+-Fur The Fur protein of E. coli is a repressor of 148 amino acid residues [40], 12 of which are histidines. The high content of histidines allows the isolation of the Fur protein by chelate-affinity chromatography [43, 44]. The histidines and carboxylate groups of aspartate or glutamate mainly in the C-terminal domain of Fur probably bind the corepressor Fe2+ [45]. However, manganese resistant mutants in E. coli were often found to be mutated in the fur gene. We assume that Mn2+ at high concentrations binds to the Fur repressor instead of Fe2+. This inhibits iron uptake systems although the cell may be in need for iron [45 a]. Mutants with a defect Fur protein have constitutively derepressed iron uptake systems and may grow better than their parents in the presence of Mn2+. This 195
10 Iron Transport in Gram-negative and Gram-positive Bacteria method has also been used with success in other bacteria to isolate fur mutants. 1 H-NMR studies with the isolated Fur protein have indicated that metal binding depends on both, metal ion concentration and proton concentration [46]. This may be a link to the observed influence of Fur on the acid tolerance response, which has been shown to be defective in Fur mutants of S. typhimurium [47]. The function of the four cysteines in the C-terminal domain of Fur is not clear; however, they seem not to bind Fe2+ directly [48, 49]. Two cysteines (C93 and C95) have been shown to bind zinc [50]. Footprinting experiments using the promoter region of the aerobactin biosynthesis genes have shown that the binding of Fur to DNA is mediated by divalent ions, such as Co2+ and Mn2+ [51]. Because of the instability of Fe2+ under oxic conditions, only a few experiments have been conducted with this natural partner of Fur. The protected region on the DNA is a 19-bp degenerate palindrome. A similar sequence is found in the promoter region of fur, and autoregulation has been demonstrated [52]. Similar sequences are found in the promoter regions of many other iron-regulated genes and are assumed to be binding sites for Fur; this site is termed the Fur box or the iron box. An overview on Furmediated regulation of gene transcription is given in [53]. When high-copynumber plasmids containing cloned Fur binding sites are introduced into the cells, Fur is removed from its chromosomal binding sites, and derepression of a reporter gene under the control of an iron-regulated promoter is observed. We developed this Fur titration assay and used it to clone new iron-regulated genes that have a Fur box in the promoter region. In addition, we observed that the presence of genes coding for iron-binding proteins could lead to derepression of the reporter gene by lowering the regulatory concentration of iron in the cell [53]. Since Fur boxes from different bacteria are similar to the Fur box of E. coli, this system has also been used to clone iron-regulated genes from other bacteria. Regulation by Fur has been studied by isolating mutants with mutations in the fur gene that are able to inhibit repression of iron-regulated genes by the chromosomally encoded Fur protein. Studies with such mutants indicated for the first time that Fur acts as an oligomer. Defective Fur protein can dimerize with wild type Fur protein encoded on the chromosome, which leads to an inactive dimer (negative complementation) unable to bind to the Fur box on the DNA [44]. Further proof for the binding of a Fur oligomer to DNA has been obtained by electron microscopy and atomic force microscopy of Fur-DNA complexes. At higher concentrations, Fur polymerizes to a multimer along the DNA [54], which is in accordance with the results of DNA-footprinting experiments [51]. Many regulatory proteins have a typical helix-turn-helix motif at their DNA binding site. Such a motif for Fur has not been identified using structure prediction programs. Studies with Fur-lambda cI hybrid proteins have shown that the N-terminal domain of Fur contains the DNA binding site [55]. This result agrees with the predicted structural similarities of the N-terminal domain of Fur to the LexA helix-turn-helix DNA-binding motif [56]. Fur-like proteins have been found in many other bacteria. A recent BLAST search has identified about 65 homologs; a Fur-like function has been demon196
10.3 Regulation of iron transport and metabolism strated for about 20 of these proteins. These results show that this principle of iron regulation is widely distributed among bacteria. In the low-GC Gram-positive bacteria, such as in Bacillus subtilis, three Fur-like genes have been identified in the genome sequencing project. In addition, the promoter structure of several iron-related genes has led to the postulation that a Fur-like regulator should also exist in B. subtilis [57]. Helman and coworkers [58] then showed that one such regulator has a Fur-like activity, one is a zinc-dependent repressor called Zur, and one, PerR, regulates the oxidative response. Gram-positive bacteria with a high GC content, such as Corynebacteriae and Streptomyces, seem to be the only exception. In these genera, Fur-like proteins have been identified that seem to regulate only the oxidative-stress response genes and not iron-related genes. In these organisms, DtxR-like proteins, another type of iron-regulatory proteins, fulfill the function of Fur in regulating iron uptake and the biosynthesis of siderophores. Proteins of the Fur family recognize different divalent cations and have fulfilled very different functions during evolution: regulation of iron transport and metabolism, regulation of Zn2+ transport and metabolism, and regulation of the oxidative-stress response in connection with cations such as iron or manganese. These different activities of Fur-like proteins makes it difficult to predict from sequence similarities alone the in vivo activities of the proteins. A sequence comparison of most of the better-known Fur proteins is shown as an unrooted phylogenetic tree in Fig. 10.2. Fur proteins that act as iron-transport regulators in pseudomonads and enteric bacteria form a defined subgroup, while the other Fur-like proteins seem to be very diverse. It is apparent that the oxidative stress responsive proteins FurA from Mycobacterium marinum, FurS from Streptomyces reticulum, and PerR from B. subtilis are not closely related. The observed regulation of heme biosynthesis by Irr in Bradyrhizobium japonicum may also be related to the oxidative-response regulation since catalases and the respiratory chain contain heme. Also the two identified Zur proteins for zinc regulation from B. subtilis and E. coli are not closely related. However, further characterization of these regulons is necessary to determine which metals act as corepressor and which genes are regulated by these proteins. The two cysteines which bind zinc in E. coli Fur are well conserved in all Fur proteins with exception of Fur from some pseudomonads. Further research has to find out if the Fur protein has been a zinc regulator which in evolution got the function of an iron regulator.
197
10 Iron Transport in Gram-negative and Gram-positive Bacteria
Figure 10.2: Unrooted tree of the Fur protein family. Only some of the corresponding proteins have been shown to regulate iron transport. Fur-like proteins with a demonstrated different function are indicated: in some Gram-positive bacteria, Fur-like proteins seem to regulate mainly oxidative response genes2and Zur from E. coli and B. subtilis regulate zinc uptake&.
10.4 An [2Fe-2S] protein is involved in ferrioxamine B utilization
Very little is known about the mobilization of iron from siderophores inside the cell. It is generally assumed that the iron is reduced to ferrous iron and then bound by an unknown component of the cytoplasm. We assume that the FhuF protein is somehow involved in these processes. In E. coli, a fhuF-lacZ operon fusion has been used as a reporter to study iron regulation since this fusion reacts very sensitively to slight changes in the iron concentration of the medium. The only phenotype observed for fhuF mutations is a diminished ability of the cells to use ferrioxamine B as an iron source. This siderophore is a poor iron source for E. coli K-12 [3] because this strain lacks a specific receptor for ferrioxamine B in the outer membrane; this receptor is found in Yersinia and other enterobacteria (see Section 10.1.2.1.4). 198
10.4 An [2Fe-2S] protein is involved in ferrioxamine B utilization In an attempt to understand the function of FhuF, the protein has been purified and characterized [59]. The fhuF gene has been identified as open reading frame f262 b at 99.2 min on the genome sequence map of E. coli K-12. The FhuF protein was labeled with a His-tag and purified. Based on sulfur determinations and Mössbauer and EPR spectroscopy, FhuF has been identified as an [2Fe-2S] protein. The g values (gx = 1.886, gy = 1.961, gz = 1.994) and some of the Mössbauer parameters of FhuF obtained (oxidized protein as isolated: DEQ,4.2K = 0.474 mm s–1 ; Fe3+ (reduced protein): DEQ = 0.978 mm s–1) are not typical for common [2Fe-2S] proteins and indicate that FhuF has unusual structural properties. The amino acid sequence of FhuF does not show any similarities to known [2Fe-2S] proteins. By site-directed mutagenesis, each of the six cysteines of FhuF was replaced by serine. EPR of the six reduced mutant proteins revealed that the terminal cysteine residues 244, 245, 256, and 259 form the [2Fe-2S]Cys4 cluster. Mutants having the Cys-to-Ser replacement at positions 244, 245, 256, or 259 do not complement a fhuF mutant. The motif Cys-CysXaa10-Cys-Xaa2-Cys in FhuF differs considerably from the motif Cys-Xaa2-CysXaa9–15-Cys-Xaa2-Cys found in other [2Fe-2S] proteins. The unusual Cys-Cys terminal group of the cluster may explain the atypical EPR and Mössbauer spectra of the FhuF protein; possibly the tetrahedral symmetry at the ferric ion site is distorted. The phenotype of fhuF mutants and the structural features of the FhuF protein suggest that FhuF is involved in the reduction of ferric iron in cytoplasmic ferrioxamine B. Mutants with an FhuF –-like phenotype were selected with the aim of further defining the function of FhuF in ferrioxamine B uptake. Two mutants, sufS::MudI and sufD::MudI, have the same phenotype as a fhuF mutant, namely the inability to use ferrioxamine B as an iron source in the plate assay. The sufS gene and the sufD gene were shown to be regulated by the iron-dependent Fur repressor. Sequence analysis has revealed that the sufS open reading frame corresponds to orf_f406. The protein SufS belongs to the family of NifS-like proteins, which supply sulfur for [Fe-S] centers. The NifS protein is required for the assembly of the [Fe-S] cluster of the nitrogenase in Azotobacter vinelandii. Three open reading frames coding for proteins with similarities to NifS have been identified in the E. coli K-12 genome (Fig. 10.3). One is the NifS-like protein, now called IscS, which has been purified and described as a pyridoxal-phosphate-dependent cysteine desulfurase by Flint [60] and is encoded at 57 min on the genetic map of E. coli. A second E. coli protein of this family has been characterized as a selenocysteine lyase (called Csd) with cysteine sulfinate desulfinase activity. The corresponding gene has been cloned and mapped at 63 min on the E. coli genetic map. The similarity of NifS-like proteins has been discussed extensively by Mihara et al. [61], who pointed out that Csd and NifS are members of two subfamilies. SufS and Csd show 43% amino acid identity, and SufS and IscS have 22% identity, which reflects that SufS also belongs to the Csd group of the NifS-like proteins. Examination of the genes in the neighborhood of sufS have revealed an operon structure. The product of the first gene sufA is similar to IscA from E. coli and Azotobacter vinelandii; IscA is encoded in the iscSUA gene cluster (Fig. 199
10 Iron Transport in Gram-negative and Gram-positive Bacteria
Figure 10.3: Gene clusters with similarities to the E. coli sufABCDSE genes. Similar genes are indicated with the same shading: sufA is similar to iscA; sufS is similar to csd, iscS, and nifS; sufE is similar to ygdK; and part of nifU is similar to iscU.
10.3). Genes encoding counterparts of IscSUA have been found near nifS in A. vinelandii and also in different bacteria and eukaryotes. The function of IscA and IscU is not known, but it is assumed that the proteins from these operons synthesize iron-sulfur centers of various proteins. An obvious counterpart for IscU or NifU is missing in the SufABCDSE cluster. However, SufE and YgdK from E. coli have 35% amino acid identity. ygdK is an open reading frame directly downstream of csd, which again points to a relationship between Csd and SufS. Since the insertion of MudI into sufS and sufD only had an effect on ferrioxamine B utilization, no other vital functions seem to depend solely on these genes. In this early stage of the investigation, it is not known whether the sufD product is directly involved in ferrioxamine utilization or whether the Mu insertion has a polar effect on the downstream genes, sufSE. However, it is interesting to note that these genes are regulated by Fur and iron, as observed for fhuF. Since the FhuF protein contains an [2Fe-2S] center, we assume that SufS may be responsible for the assembly of this iron-sulfur center. A mutation in the upstream sufD gene (orf_f423) causes the same phenotype. The T7 expression system and a His-tag allow the isolation of the FhuF protein from a wild type strain in good yield. In contrast, overproduction of the protein in a DsufD strain has failed. Radioactive labeling of N-His-FhuF with [35S]methionine has shown that the protein is unstable in the DsufD mutant [62]. The instability of FhuF could have various reasons. FhuF may be part of the Suf complex and is unstable as a single protein. Another interpretation, which we prefer, is that the Suf complex is necessary for the building of the “distorted” [2Fe-2S] center in FhuF and that the apo-FhuF protein is unstable in the cell. However, this instability must be very peculiar since the various FhuF Cys?Ser mutant proteins which did not contain a [2Fe-2S] center have been isolated in relatively large amounts [59].
200
References
References
1. Braun, V. and Hantke, K. (1981) Bacterial cell surface receptors. In: Ghosh, B.K. (ed.), Organization of procaryotic cell membranes. CRC Press: Boca Raton, Florida, pp. 1– 74. 2. Hantke, K. (1983) Identification of an iron uptake system specific for coprogen and rhodotorulic acid in Escherichia coli K-12. Mol. Gen. Genet. 191, 301–306. 3. Sauer, M., Hantke, K., and Braun, V. (1987) Ferric-coprogen receptor FhuE of Escherichia coli: processing and sequence common to all TonB dependent outer membrane receptor proteins. J. Bacteriol. 169, 2044–2049. 4. Curtis, N. A., Eisenstadt, R. L., East, S. J., Cornford, R. J., Walker, L. A., and White, A. J. (1988) Iron-regulated outer membrane proteins of Escherichia coli K-12 and mechanism of action of catechol-substituted cephalosporins. Antimicrob. Agents Chemother. 32, 1879–1886. 5. Earhart, C. F. (1996) Uptake and metabolism of iron and molybdenum. In: Neidhart, F.C. (ed.), Escherichia coli and Salmonella, 2nd edn. ASM Press: Washington, D.C., pp. 1075–1090. 6. Braun, V. and Hantke, K. (1982) Receptor dependent transport systems in Escherichia coli for iron complexes and vitamin B12. In:. Martonosi, A. N (ed.), Membrane and transport. Plenum Press: New York, pp. 107–113. 7. Hantke, K. (1990) Dihydroxybenzoylserine – a siderophore for E. coli. FEMS Microbiol. Lett. 67, 5–8. 8. Maejima, T., Inoue, M., and Mitsuhashi, S. (1991) In vitro antibacterial activity of KP736, a new cephem antibiotic. Antimicrob. Agents Chemother. 35, 104–110. 9. Wake, A., Misawa, M., and Matsui, A. (1975) Siderochrome production by Yersinia pestis and its relation to virulence. Infect. Immun. 12, 1211–1213. 10. Heesemann, J. (1987) Chromosomal encoded siderophores are required for mouse virulence of enteropathogenic Yersinia species. FEMS Microbiol. Lett. 48, 229–233. 11. Hantke, K. (1987) Selection procedure for deregulated iron transport mutants (fur) in Escherichia coli K-12: fur not only affects iron metabolism. Mol. Gen. Genet. 210, 135–139. 12. Haag, H., Hantke, K., Drechsel, H., Stojiljkovic, I., Jung, G., and Zähner, H. (1993) Purification of yersiniabactin: a siderophore and possible virulence factor of Yersinia enterocolitica. J. Gen. Microbiol. 139, 2159–2165. 13. Heesemann, J., Hantke, K., Vocke, T., Saken, E., Rakin, A., Stojiljkovic, I., and Berner, R. (1993) Virulence of Yersinia enterocolitica is closely associated with siderophore production, expression of an iron-repressible outer membrane polypeptide of 65,000 Da and pesticin sensitivity. Mol. Microbiol. 8, 397–408. 14. Vollmer, W., Pilsl, H., Hantke, K., Höltje, J. V., and Braun, V. (1997) Pesticin displays muramidase activity. J. Bacteriol. 179, 1580–1583. 15. Pilsl, H., Killmann, H., Hantke, K., and Braun, V. (1996) Periplasmic location of the pesticin immunity protein suggests inactivation of pesticin in the periplasm. J. Bacteriol. 178, 2431–2435. 16. Killmann, H., Herrmann, C., Wolff, H., and Braun, V. (1998) Identification of a new site for ferrichrome transport by comparison of the FhuA proteins of Escherichia coli, Salmonella paratyphi B, Salmonella typhimurium, and Pantoea agglomerans. J. Bacteriol. 180, 3845–3852. 17. Koebnik, R., Hantke, K., and Braun, V. (1993) The TonB-dependent ferrichrome receptor FcuA of Yersinia enterocolitica: evidence against a strict co-evolution of receptor structure and substrate specificity. Mol. Microbiol. 7, 383–393.
201
10 Iron Transport in Gram-negative and Gram-positive Bacteria 18. Bäumler, A., Koebnik, R., Stojiljkovic, I., Heesemann, J., Braun, V., and Hantke, K. (1993) Survey on newly characterized iron uptake systems of Yersinia enterocolitica. Zbl. Bakt. 278, 416–424. 19. Schubert, S., Fischer, D., and Heesemann, J. (1999) Ferric enterochelin transport in Yersinia enterocolitica: molecular and evolutionary aspects. J. Bacteriol. 181, 6387– 6395. 20. Autenrieth, I., Hantke, K., and Heesemann, J. (1991) Immunosuppression of the host and delivery of iron to the pathogen: a possible dual role of siderophores in the pathogenesis of microbial infections? Med. Microbiol. Immunol. (Berlin) 180, 135–141. 21. Bäumler, A. J. and Hantke, K. (1992) Ferrioxamine uptake in Yersinia enterocolitica: characterization of the receptor protein FoxA. Mol. Microbiol. 6, 1309–1321. 22. Otto, B. R., Verweij van Vught, A. M., and MacLaren, D. M. (1992) Transferrins and heme-compounds as iron sources for pathogenic bacteria. Crit. Rev. Microbiol. 18, 217–233. 23. Braun, V., Hantke, K., and Köster, W. (1998) Bacterial iron transport: mechanisms, genetics, and regulation. Met. Ions. Biol. Syst. 35, 67–145. 24. Stojiljkovic, I. and Hantke, K. (1992) Hemin uptake system of Yersinia enterocolitica: similarities with other TonB-dependent systems in Gram-negative bacteria. EMBO J. 11, 4359–4367. 25. Stojiljkovic, I. and Hantke, K. (1994) Transport of haemin across the cytoplasmic membrane through a haemin-specific periplasmic binding-protein-dependent transport system in Yersinia enterocolitica. Mol. Microbiol. 13, 719–732. 26. Mills, M. and Payne, S.-M. (1995) Genetics and regulation of heme iron transport in Shigella dysenteriae and detection of an analogous system in Escherichia coli O157:H7. J. Bacteriol. 177, 3004–3009. 27. Torres, A. G. and Payne, S. M. (1997) Haem iron transport system in enterohaemorrhagic Escherichia coli O157:H7. Mol. Microbiol. 23, 825–833. 28. Henderson, D. P. and Payne, S. M. (1993) Cloning and characterization of the Vibrio cholerae genes encoding the utilization of iron from haemin and haemoglobin. Mol. Microbiol. 7, 461–469. 29. Kammler, M., Schön, C., and Hantke, K. (1993) Characterization of the ferrous iron uptake system of Escherichia coli. J. Bacteriol. 175, 6212–6219. 30. Stojiljkovic, I., Cobeljic, M., and Hantke, K. (1993) Escherichia coli K-12 ferrous iron uptake mutants are impaired in their ability to colonize the mouse intestine. FEMS Microbiol. Lett. 108, 111–115. 31. Tsolis, R. M., Bäumler, A. J., Heffron, F., and Stojiljkovic, I. (1996) Contribution of TonB and Feo mediated iron uptake to growth of Salmonella typhimurium in the mouse. Infect. Immun. 64, 4549–4556. 32. Bult, C. J., White, O., Olsen, G. J., Zhou, L., Fleischmann, R. D., Sutton, G. G., Blake, J. A., FitzGerald, L. M., Clayton, R. A., Gocayne, J. D., Kerlavage, A. R., Dougherty, B. A., Tomb, J. F., Adams, M. D., Reich, C. I., Overbeek, R., Kirkness, E. F., Weinstock, K. G., Merrick, J. M., Glodek, A., Scott, J. L., Geoghagen, N. S. M., Weidman, J. F., Fuhrmann, J. L., Venter, J. C. et al. (1996) Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science 273, 1058–1073. 33. Gibson, M. M., Bagga, D. A., Miller, C. G., and Maguire, M. E. (1991) Magnesium transport in Salmonella typhimurium: the influence of new mutations conferring Co2+ resistance on the CorA Mg2+ transport system. Mol. Microbiol. 5, 2753–2762. 34. Hantke, K. (1997) Ferrous iron uptake by a magnesium transport system is toxic for Escherichia coli and Salmonella typhimurium. J. Bacteriol. 179, 6201–6204. 35. Locke, A. and Main, E. R. (1931) The relation of copper and iron to the production of toxin and enzyme action. J. Infect. Dis. 48, 419–435.
202
References 36. Litwin, C. M. and Calderwood, S. B. (1993) Role of iron in regulation of virulence genes. Clin. Microbiol. Rev. 6, 137–149. 37. Lankford, C. E. (1973) Bacterial assimilation of iron. Crit. Rev. Microbiol. 2, 273–331. 38. Ernst, J. F., Bennett, R. L., and Rothfield, L. R. (1997) Constitutive expression of the iron-enterochelin and ferrichrome uptake systems in a mutant strain of Salmonella typhimurium. J. Bacteriol. 135, 928–934. 39. Hantke, K. (1981) Regulation of ferric iron transport in Escherichia coli K-12: isolation of a constitutive mutant. Mol. Gen. Genet. 182, 288–292. 40. Schäffer, S., Hantke, K., and Braun, V. (1985) Nucleotide sequence of the iron regulatory gene fur. Mol. Gen. Genet. 200, 110–113. 41. Casadaban, M. J. and Cohen, S. N. (1979) Lactose genes fused to exogenous promoters in one step using a Mulac bacteriophage: in vitro probe for transcriptional control sequences. Proc. Natl. Acad. Sci. USA 76, 4530–4533. 42. Hantke, K. (1983) Identification of an iron uptake system specific for coprogen and rhodotorulic acid in Escherichia coli K-12. Mol. Gen. Genet. 191, 301–306. 43. Wee, S., Neilands, J. B., Bittner, M. L., Hemming, B. C., Haymore, B. L., and Seetharam, R. (1988) Expression, isolation and properties of Fur (ferric uptake regulation) protein of Escherichia coli K 12. Biol. Met. 1, 62–68. 44. Braun, V., Schäffer, S., Hantke, K., and Tröger, W. (1990) Regulation of gene expression by iron. In: The molecular basis of bacterial metabolism. Colloquium Moosbach 41 Springer, Berlin. pp. 164–179. 45. Hamed, M. Y. and Neilands, J. B. (1993) An electron spin resonance study of the Mn(II) and Cu(II) complexes of the Fur repressor protein. Binding of the ferric uptake regulation repressor protein (Fur) to Mn(II), Fe(II), Co(II), and Cu(II) ions as co-repressors: electronic absorption, equilibrium, and 57Fe Mössbauer studies. J. Inorg. Biochem. 50, 193–210. 45 a. Hantke, K. (1987) Selection procedure for deregulated iron transport mutants (fur) in Escherichia coli K-12: Fur not only affects iron metabolism. Mol. Gen. Genet., 210, 135–139 46. Saito, T., Duly, D., and Williams, R. J. (1991) The histidines of the iron-uptake regulation protein, Fur. Eur. J. Biochem. 197, 39–42. 47. Hall, H. K. and Foster, J. W. (1996) The role of fur in the acid tolerance response of Salmonella typhimurium is physiologically and genetically separable from its role in iron acquisition. J. Bacteriol. 178, 5683–5691. 48. Saito, T., Wormald, M. R., and Williams, R. J. (1991) Some structural features of the iron-uptake regulation protein. Eur. J. Biochem. 197, 29–38. 49. Hamed, M. Y. and Neilands, J. B. (1994) An electron spin resonance study of the Mn(II) and Cu(II) complexes of the Fur repressor protein. J. Inorg. Biochem. 53, 235– 248. 50. Jacquamet, L., Aberdam, D., Adrait, A., Hazemann, J. L., Latour, J. M, and MichaudSoret, I. (1998) X-ray absorption spectroscopy of a new zinc site in the fur protein from Escherichia coli. Biochemistry 37, 2564–2571. 51. De Lorenzo, V., Wee, S., Herrero, M., and Neilands, J. B. (1987) Operator sequences of the aerobactin operon of plasmid ColV-K30 binding the ferric uptake regulation (fur) repressor. J. Bacteriol. 169, 2624–2630. 52. De Lorenzo, V., Herrero, M., Giovannini, F., and Neilands, J. B. (1988) Fur (ferric uptake regulation) protein and CAP (catabolite-activator protein) modulate transcription of fur gene in Escherichia coli. Eur. J. Biochem. 173, 537–546. 53. Stojiljkovic, I., Bäumler, A. J., and Hantke, K. (1994) Fur regulon in gram-negative bacteria: identification and characterization of new iron-regulated Escherichia coli genes by a fur titration assay. J. Mol. Biol. 236, 531–545. 54. Le Cam, E., Frechon, D., Barray, M., Fourcade, A., and Delain, E. (1994) Observation of binding and polymerization of Fur repressor onto operator-containing DNA
203
10 Iron Transport in Gram-negative and Gram-positive Bacteria
55. 56.
57.
58.
59.
60.
61.
62.
with electron and atomic force microscopes. Proc. Natl. Acad. Sci. USA 91, 11816– 11820. Stojiljkovic, I. and Hantke, K. (1995) Functional domains of the Escherichia coli ferric uptake regulator protein (Fur). Mol. Gen. Genet. 247, 199–205. Holm, L., Sander, C., Rüterjans, H., Schnarr, M., Fogh, R., Boelens, R., and Kaptein, R. (1994) LexA repressor and iron uptake regulator from Escherichia coli: new members of the CAP-like DNA binding domain superfamily. Protein Eng. 7, 1449–1453. Schneider, R. and Hantke, K. (1993) Iron-hydroxamate uptake systems in Bacillus subtilis: identification of a lipoprotein as a part of a binding protein dependent transport system. Mol. Microbiol. 8, 111–121. Gaballa, A. and Helmann, J. D. (1998) Identification of a zinc-specific metalloregulatory protein, Zur, controlling zinc transport operons in Bacillus subtilis. J. Bacteriol. 180, 5815–5821. Muller, K., Matzanke, B. F., Schunemann, V., Trautwein, A. X., and Hantke, K. (1998) FhuF, an iron-regulated protein of Escherichia coli with a new type of [2Fe-2S] center. Eur. J. Biochem. 258, 1001–1008. Flint, D. H. (1996) Escherichia coli contains a protein that is homologous in function and N-terminal sequence to the protein encoded by the nifS gene of Azotobacter vinelandii and that can participate in the synthesis of the Fe-S cluster of dihydroxy-acid dehydratase. J. Biol. Chem. 271, 16068–16074. Mihara, H., Kurihara, T., Yoshimura, T., Soda, K., and Esaki, N. (1997) Cysteine sulfinate desulfinase, a NIFS-like protein of Escherichia coli with selenocysteine lyase and cysteine desulfurase activities. Gene cloning, purification, and characterization of a novel pyridoxal enzyme. J. Biol. Chem. 272, 22417–22424. Patzer, S. I. and Hantke, K. (1999) SufS is a NifS-like protein, and SufD is necessary for stability of the [2Fe-2S] FhuF protein in Escherichia coli. J. Bacteriol. 181, 3307– 3309.
204
11 Regulation of the Ferric-Citrate Transport System by a Novel Transmembrane Transcription Control Volkmar Braun* and Sabine Enz
11.1 Introduction
Citrate does not serve as a carbon source for Escherichia coli K-12 since it is not taken up by the cells; however, Fe3+ delivered as a citrate complex is actively taken up by E. coli. Our investigation of the transport of radiolabeled [55Fe3+] [14C]citrate has revealed uptake of iron and only residual uptake of citrate, indicating that only iron and not the iron complex enters the cytoplasm. Yet ferric citrate induces transcription of ferric-siderophore transport genes and is the only ferric siderophore known to do so in E. coli. The question arose how an inducer initiates transcription of transport genes in the cytoplasm when it does not enter the cytoplasm.
11.2 Transport of Fe3+ is mediated by citrate
A ferric-citrate transport system has been characterized only from E. coli (Fig. 11.1). Since a 20-fold surplus of citrate over iron is required to obtain a predominantly low molecular weight form of ferric citrate, the transport-active composition of ferric citrate is not certain, but most likely consists of ferric dicitrate [1, 2]. Ferric citrate is transported across the outer membrane via the FecA protein at the expense of the electrochemical potential of the cytoplasmic membrane, mediated by the Ton system. Once in the periplasm, ferric citrate binds
* Mikrobiologie/Membranphysiologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
205 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
11 Regulation of the Ferric-Citrate Transport System
Figure 11.1: (A) Arrangement of the fec operon at 97.3 minutes on the E. coli chromosome. Pfec denotes the promoter induced by ferric citrate, Pfur denotes the promoters repressed by the Fe2+-loaded Fur protein. (B) Location of the signaling and transport proteins and the proposed transport and signal pathways. N, N-terminus of FecA; OM, outer membrane; PP, periplasm; CM, cytoplasmic membrane; CP, cytoplasm; RNAP, RNA polymerase; aa, amino acids.
to the FecB protein. Only iron seems to be transported further across the cytoplasmic membrane since ten times more radioactive iron than radioactive citrate is firmly associated with the cells, and iron transport, but not citrate uptake, is inducible. It is not known where iron dissociates from citrate. This could occur at the FecB protein because FecB binds iron in the absence of citrate. Iron is transported across the cytoplasmic membrane with the help of two integral membrane proteins, FecC and FecD, together with the FecE protein, which is presumably bound to the inside surface of the cytoplasmic membrane. FecE contains two Walker motifs of nucleotide-binding proteins and can be photoaffinity-labeled with [32P]8-azido-ATP [3]. These results indicate that iron delivered as ferric citrate is transported across the cytoplasmic membrane by a mechanism that is typical for ABC transporters. FecB delivers ferric iron to the FecC/FecD 206
11.3 Transcription initiation by a signaling cascade transport proteins in the cytoplasmic membrane, and subsequent translocation of iron across the cytoplasmic membrane is energized by ATP hydrolysis catalyzed by FecE.
11.3 Transcription initiation by a signaling cascade from the cell surface into the cytoplasm
Not intracellular ferric citrate, but extracellular ferric citrate serves as an inducer of the ferric-citrate transport system. fecBCDE mutants impaired in the transport of iron across the cytoplasmic membrane are fully inducible, but fecA, tonB, exbB, or exbD mutants (the latter in combination with tolQ or tolR mutations), which are devoid of ferric-citrate transport across the outer membrane, are not inducible [4, 5]. The obvious conclusion that entry of ferric citrate into the periplasm is required for induction has been ruled out by supplying to a fecA null mutant growth-promoting concentrations of ferric dicitrate (molecular mass 434 Da) that enter the periplasm by diffusion through the porins. No induction of fec transport genes is observed [6], and the transport genes encoding cytoplasmic membrane transport activities have to be constitutively overexpressed from a multicopy plasmid to provide sufficient amounts of transport proteins. A direct involvement of FecA in induction has been shown with fecA missense mutants, which induce fec transcription, but do not transport ferric citrate, and which express fec genes constitutively in the absence of ferric citrate [6]. In contrast to other E. coli ferric siderophore receptors, the FecA protein contains an N-terminal extension. When this extra peptide (residues 47–101; numbering includes the signal peptide) is removed by genetic means, induction is abolished but FecA fully retains transport activity. Overexpressed N-terminal FecA(1–100) and FecA(1–127) fragments inhibit induction, but not transport [5]. This proves that the N-terminus of mature FecA (741 amino acids; signal peptide comprises 33 residues) extending from residue 1 to approximately 120 participates in signal transduction. The N-proximal portion of FecA is localized in the periplasm [5] and most likely interacts with the FecR regulatory protein, of which residues 101–317 are contained in the periplasm, residues 86–100 span the cytoplasmic membrane, and the N-terminal portion is in the cytoplasm [7, 8]. The transmembrane topology of FecR suggests that it transmits the signal, elicited by ferric citrate on FecA, across the cytoplasmic membrane into the cytoplasm. Cytoplasmic N-terminal fragments of FecR, the smallest of which consists of 59 residues, induce fec transport gene transcription constitutively [7], showing that the cytoplasmic segment participates in transcription regulation and the periplasmic region responds to the signal. Signal transduction is inhibited by carbonylcyanide-m-chlorophenylhydrazone (CCCP), which dissipates the elec207
11 Regulation of the Ferric-Citrate Transport System trochemical potential of the cytoplasmic membrane [5]. The energized cytoplasmic membrane is required for signal transduction across the outer membrane, but not for signal transduction across the cytoplasmic membrane, as shown by the lack of effect of CCCP on the transcription of the fec transport genes in the absence of ferric citrate and TonB in the fecA4 mutant which constitutively transcribes the fec transport genes [5]. The TonB-ExbB-ExbD complex transmits the energy from the cytoplasmic membrane to the outer membrane for signal transduction, as for ferric citrate transport across the outer membrane. The molecular mechanisms underlying the two metabolic processes do not have to be the same since the fecA4 mutant signals but does not transport. FecR does not directly act on the promoter upstream of the fecA gene that regulates transcription of fecA and of the fecBCDE genes downstream of fecA (Fig. 11.1). The fecI gene upstream of fecR encodes a sigma factor with an amino acid sequence only slightly similar to that of s70 factors. The sigma factors of FecI type are widespread in different genera and are designated as ECF (extracytoplasmic functions) factors because they all seem to be involved in extracytoplasmic metabolic activities [9]. However, no other ECF regulatory system has been studied to the extent as that of the ferric citrate regulation. Purified FecI mediates specific binding of the RNA polymerase core enzyme to the promoter region upstream of fecA, as revealed by DNA-mobility band-shift experiments, and promotes fecA transcription in vitro [10]. Band shifting of fecA-promoter DNA caused by cell lysates requires synthesis of FecA, FecI, and FecR and growth of cells with ferric citrate [10], which implies that activated FecI remains active during disruption of the cells and the band-shift assay. Interaction between the FecA, FecR, and FecI signaling proteins has been demonstrated by utilizing two methods. In in vitro binding assays, FecA is retained by FecR His-tagged at the N-terminus [(His)10-FecR] and bound to a NiNTA agarose column and is co-eluted with [(His)10-FecR]; FecI is retained by FecR His-tagged at the C-terminus [FecR-(His)6] and is co-eluted with [FecR(His)6] from the column. An N-terminally truncated, induction-negative but transport-active FecA protein does not bind to (His)10-FecR. In the in vivo assay, the FecA-, FecR-, and FecI-interacting domains have been determined using the bacterial two-hybrid Lex-based system. FecA1–79 interacts with FecR101–317, and FecR1–85 interacts with FecI1–173 [11]. Moreover, mutations in the cytoplasmic region of FecR are suppressed by FecI(Ser15?Ala) and FecI(His20?Glu) mutations (A. Stiefel, unpublished results). These data clearly support a model that proposes interaction of the periplasmic N-terminus of FecA with the periplasmic C-terminal portion of FecR, and interaction of the cytoplasmic N-terminus of FecR with the N-terminus of FecI, which results in FecI activation.
208
Acknowledgments
11.4 Iron regulation of fecIR and fecABCDE transcription
fecIR and fecABCDE form separate transcripts [12]. Transcription of fecIR is repressed by iron and the Fur protein, but is uninfluenced by ferric citrate, while fecABCDE transcription is regulated by iron and Fur and by ferric citrate via FecI and FecR [13]. FecI and FecR regulate fec transport genes transcription, but they display no autoregulation [14]. The iron transport genes are regulated by the internal iron concentration and by external ferric citrate. This is a very economical way of using ferric citrate as an iron source. When iron is not needed or when ferric citrate is not present, the transport system can be almost totally shut off by cytoplasmic iron. When the iron concentration in the cytoplasm falls below a certain limit, the fecIR genes are transcribed. To turn on the ferric citrate transport system, the carrier has to be in the culture medium. The FecA and FecR proteins do not only have regulatory functions, they also have vectorial activities in that they transmit information through three cell compartments (Fig. 11.2). Binding of ferric citrate to FecA induces a signal that is transferred from the cell surface into the periplasm and across the cytoplasmic membrane into the cytoplasm. It is likely that the information flux involves coupled conformational changes of FecA and FecR. The conceptual question that remains to be answered is how FecI is activated by FecR. No phosphorylation or proteolytic processing of FecI has been found. Transmembrane transcription control of the E. coli Fec type is not confined to the regulation of ferric-citrate transport. A regulatory system most likely similar to the fec transcription control has been found in Pseudomonas putida. P. putida WCS358 transports ferric pseudobactin BN8. An outer membrane protein, PupB, transports the ferric pseudobactins BN7 and BN8, and synthesis of PupB is induced by BN7 and BN8. It has also been shown that the N-terminus of PupB is involved in induction, presumably via the PupI and PupR regulatory proteins, which are homologous to FecI and FecR [15]. The sequenced genome of Pseudomonas aeruginosa reveals several sets of genes that are homologous to fecI fecR, the heme transport system of Serratia marcescens seems to be regulated by fecI fecR homologs (C. Wandersmann, personal communication), and fecI fecR homologs exist in Bordetella pertussis. Fec regulation may become the paradigm of other similarly regulated gene transcription devices.
Acknowledgments
The important contributions of Annemarie Angerer are gratefully acknowledged. 209
11 Regulation of the Ferric-Citrate Transport System
Figure 11.2: Model of transmembrane signaling for transcriptional control of the fec system. The model indicates activation of FecR by FecA loaded with ferric citrate and activation of FecI by active FecR. The iron-loaded Fur repressor binds to the promoters, termed Pfur, upstream of the fecIR genes and the fecABCDE genes. It is not known whether FecI and Fe2+Fur compete physically at the fecABCDE promoter.
210
References
References
1. Spiro, T. G., Bates, G., and Saltmann, P. (1967) The hydrolytic polymerization of ferric citrate. II. The influence of excess citrate. J. Am. Chem. Soc. 89, 5559–5562. 2. Hussein, S., Hantke, K., and Braun, V. (1981) Citrate dependent iron transport system in Escherichia coli K-12. Eur. J. Biochem. 117, 431–4337. 3. Schultz-Hauser, G., Van Hove, B., and Braun, V. (1992) 8-Azido-ATP labeling of the FecE protein of the Escherichia coli iron citrate transport system. FEMS Microbiol. Lett. 95, 231–234. 4. Zimmermann, L., Hantke, K., and Braun, V. (1984) Exogenous induction of the ferric dicitrate transport system of Escherichia coli K-12. J. Bacteriol. 159, 271–277. 5. Kim, I., Stiefel, A., Plantör, S., Angerer, A., and Braun, V. (1997) Transcription induction of the ferric citrate transport genes via the N-terminus of the FecA outer membrane protein, the Ton system and the electrochemical potential of the cytoplasmic membrane. Mol. Microbiol. 23, 333–344. 6. Härle, C., Kim, I., Angerer, A., and Braun, V. (1995) Signal transfer through three compartments: transcription initiation of the Escherichia coli ferric citrate transport system from the cell surface. EMBO J. 7, 1430–1438. 7. Ochs, M., Veitinger, S., Kim, I., Welz, D., Angerer, A., and Braun, V. (1995) Regulation of citrate-dependent iron transport of Escherichia coli: FecR is required for transcription activation by FecI. Mol. Microbiol. 15, 119–132. 8. Welz, D. and Braun, V. (1998) Ferric citrate transport of Escherichia coli: functional regions of the FecR transmembrane regulatory protein. J. Bacteriol. 180, 2387–2394 9. Lonetto, M. A., Brown, K. L., Rudd, K. E., and Buttner, M. J. (1994) Analysis of the Streptomyces coelicolor sigE gene reveals the existence of a subfamily of eubacterial RNA polymerase s factors involved in the regulation of extracytoplasmic functions. Proc. Natl. Acad. Sci. USA 91, 7573–7577. 10. Angerer, A., Enz, S., Ochs, M., and Braun, V. (1995) Transcriptional regulation of ferric citrate transport in Escherichia coli K-12. FecI belongs to a new subfamily of s70-type factors that respond to extracytoplasmic stimuli. Mol. Microbiol. 18, 163–174. 11. Enz, S., Mahren, S., Stroeher, U. H., and Braun, V. (1999) Surface signaling in ferric citrate transport gene induction: interaction of the FecA, FecR and FecI regulatory proteins. J. Bacteriol. 181, 637–646. 12. Enz, S., Braun, V., and Crosa, J. (1995) Transcription of the region encoding the ferric dicitrate transport system in Escherichia coli: similarity between promoters for fecA and for extracytoplasmic function sigma factors. Gene 163, 13–18. 13. Angerer, A. and Braun, V. (1998) Iron regulates transcription of the Escherichia coli ferric citrate transport genes directly and through the transcription initiation proteins. Arch. Microbiol. 169, 483–490. 14. Ochs, M., Angerer, A., Enz, S., and Braun, V. (1996) Surface signaling in transcriptional regulation of the ferric citrate transport system of Escherichia coli: mutational analysis of the alternative sigma factor FecI supports its essential role in fec transport gene transcription. Mol. Gen. Genet. 250, 455–465. 15. Koster, M., Van Klompenburg, W., Bitter, W., Leong, J., and Weisbeek, P. (1994) Role for the outer membrane ferric-siderophore receptor PubB in signal transduction across the bacterial cell envelope. EMBO J. 13, 2805–2813.
211
12 Structure, Function, Import, and Immunity of Colicins Volkmar Braun* and Helmut Pilsl
12.1 Introduction
Colicins are of particular interest regarding their entry into sensitive cells since due to the impermeability of cell membranes for proteins, colicin import is a highly specific and active cellular process. Our studies were mainly concerned with the import of colicins across the outer membrane, and we studied primarily colicins with target sites within the cytoplasmic membrane into which they apparently integrate spontaneously once they have reached the periplasm. The colicins were used as tools to get insights into the receptor-mediated and energydependent protein import in which either the Ton or the Tol system is involved. Colicins are proteins with molecular weights between 27 000 and 75 000 that are synthesized by approximately 50 % of E. coli strains isolated from natural sources. They kill other E. coli strains by forming pores in the cytoplasmic membrane or by degrading DNA or ribosomal 16S RNA [1]. We have sequenced the activity, immunity and lysis genes of colicins B, D, M, K, S4, U, 5, 10, and of pesticin and determined their mode of action, their import into sensitive cells and the immunity of the producer strains. These results contributed to the subdivision of pore-forming colicins into two groups, colicins A, B, N, and U and colicins E1, Ia, Ib, K, 5, and 10. The immunity proteins of the A-group form four transmembrane segments and the immunity proteins of the E1 group form three transmembrane segments in the cytoplasmic membrane.
* Mikrobiologie/Membranphysiologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
212 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
12.2 Colicin M inhibits murein biosynthesis
12.2 Colicin M inhibits murein biosynthesis and thus displays a unique activity among the colicins
We have studied colicin M because it uses FhuA as receptor and the Ton system for uptake. It also does not belong to the pore-forming colicins and the nuclease colicins. Light microscopy and electron microscopy revealed early cell lysis [2] which is caused by inhibition of murein biosynthesis [3] through inhibition of the regeneration of the bactoprenylphosphate from bactoprenylpyrophosphate [4]. Since the same carrier lipid that transfers the murein precursors across the cytoplasmic membrane also transfers lipopolysaccharide O-antigen precursors across the cytoplasmic membrane, colicin M inhibits also O-antigen biosynthesis [5]. These studies have revealed a unique mode of action of colicin M which differs from the action of any other colicin studied. Cells that synthesize colicin M are not lysed by colicin M from inside or when added from outside the cells. Immunity is conferred by an immunity protein (Cmi) which is contained in the periplasm and anchored by the N-proximal end to the cytoplasmic membrane. The hydrophobic sequence of Cmi from residues 3–23 serves to translocate residues 24–117 of Cmi into the periplasm and anchors Cmi to the cytoplasmic membrane. Replacement of residues 1–23 of Cmi by the hydrophobic amino acid sequence 1–42, that anchors the penicillin binding protein to the cytoplasmic membrane, results in active Cmi that protects cells against colicin M [6]. This result demonstrates that translocation and anchoring of Cmi are not sequence-specific. Substitution of residues 1–23 by the cleavable signal sequence of the BlaM b-lactamase results in an active Cmi which is located in the periplasm. Removal of residues 1–23 without replacement by a hydrophobic amino acid sequence results in an inactive Cmi which is located in the cytoplasm. Apparently, the immunity protein inactivates colicin M in the periplasm before it integrates into the cytoplasmic membrane where it inhibits lipid carrier regeneration. Colicin M is usually encoded together with colicin B on large self-transmissible plasmids [7]. Sequencing of both colicins reveals the gene order cba cbi cma cmi (colicin B activity gene, colicin B immunity gene, colicin M activity gene, colicin M immunity gene). The SOS transcription control region, typical for colicin genes, is located upstream of the cba gene, and the cma gene is transcribed from the same promoter. cba and cma have the same, cbi and cmi have the opposite transcription polarity. The colicin BM operon encodes no lysis gene [8] which is usually found downstream of the colicin immunity genes and which determines a small lysis protein that facilitates release of colicins from producing cells. Accordingly, 90 % of colicins B and M remain cell-associated. Colicin B was purified to electrophoretic homogeneity and represents a single polypeptide chain in contrast to a previously published paper. It belongs to the pore-forming colicins and forms the most stable single channels of all colicins in artificial lipid bilayer membranes [9]. 213
12 Structure, Function, Import, and Immunity of Colicins The uptake of colicins B and M requires the Ton system. Both colicins contain a TonB box sequence close to the N-terminal end [10, 11]. Amino acid replacements in the TonB box sequences inactivate the colicins because they are not taken up [12, 13]. The mutations Q160L and Q160K in TonB partially restore the uptake of the mutated colicins. Suppression of a mutation in one protein by a mutation in another protein is taken as evidence for an interaction of both proteins through the mutated sites. The cellular receptor proteins FhuA and FepA also contain TonB box sequences close to the N-terminus. For FhuA it has been shown that mutations in the TonB box inactivate FhuA with regard to its TonBdependent activities but not with regard to the TonB-independent receptor activity for phage T5 [14]. The same TonB mutations Q160L and Q160K partially restore the TonB-dependent FhuA activities. These results demonstrate a dual participation of the Ton system for colicin M transport which is the first case where this has been shown. Since we could also demonstrate suppression of TonB box mutations in colicin B by TonB mutations it is likely that all colicins taken up by TonB-dependent receptors use the TonB activity in two transport steps.
12.3 Colicins 5 and 10 are taken up by a novel mechanism
In 1992 Bradley and Howard published a paper in which they provided genetic evidence that colicins 5 and 10 use the TonB-independent outer membrane receptor protein Tsx but require TonB for uptake [15]. This finding stood in contrast to what we had found with colicins B and M. Therefore, we sequenced the activity, immunity and lysis genes of colicin 5 [16] and the closely related colicin 10 [17]. In fact, both colicins contain a TonB box close to the N-terminus which indicates a TonB-dependent uptake. Replacement of the N-terminal sequence 1–67 of colicin 10 by the N-terminal sequence 1 to 58 of colicin E1 confers to the hybrid protein a Tol-dependent uptake mechanism typical for colicin E1. Replacement of the N-terminal colicin E1 sequence by the N-terminal colicin 10 sequence conferred to the hybrid protein a Ton-dependent uptake. These experiments demonstrate that the N-terminal regions are solely responsible for the Ton and Tol-dependent uptake routes, respectively, and that receptors need not to function in a TonB-dependent way to take up colicins by the Ton system. For these colicins it is sufficient that only they interact with the TonB protein. In addition, these data support the structure of colicins composed of independent functional domains.
214
12.4 Colicins evolved by the exchange of DNA fragments
12.4 Colicins evolved by the exchange of DNA fragments which precisely defined functional domains
The modular structure of colicins was first deduced from the comparison of the amino acid sequences of colicin D and colicin B [18]. The colicin D activity and immunity genes were sequenced because colicin D shares the outer membrane receptor and the TonB dependence of uptake with colicin B but it does not form pores as colicin B. Its mode of action is still unclear. The N-terminal sequences of both proteins determine TonB-dependent uptake and receptor specificity and are almost identical (96%) but the C-terminal regions that determine the killing activities are completely different. The conclusion that during evolution colicins were assembled from DNA fragments that encode functional regions was strongly supported by comparison of the pore-forming colicins K and S4 with colicins 5 and 10 [19, 20]. The high degree of sequence identity of functionally equivalent regions can be seen in Fig. 12.1A. Analysis of the nucleotide sequence of the E. coli colicin S4 determinant revealed 76% identity to the pore-forming domain of the colicin A protein, 77% to the colicin A immunity protein, and 82% identity to the colicin A lysis protein. The N-terminal region, which is responsible for the Tol-dependent uptake of colicin S4, has 93.8% identity to the N-terminal region of colicin K. By contrast, the predicted receptor binding domain shows no sequence similarities to other colicins (Fig. 12.1B) which reflects its unique binding to OmpW, a minor protein of unknown function in the outer membrane of E. coli. Colicin S4 is another example for the assembly of colicins from DNA fragments that encode functional domains. A novel colicin, designated colicin U, was found in two Shigella boydii strains of serovars 1 and 8 [21]. Colicin U is active against bacterial strains of the genera Escherichia and Shigella. Colicin U displays sequence similarities to various colicins (Fig. 12.2). The N-terminal sequence of 130 amino acids has 54% identity to the N-terminal sequence of bacteriocin 28 b produced by Serratia marcescens. Furthermore, the N-terminal 36 amino acids share striking sequence identity (83%) to colicin A. Although the C-terminal pore-forming sequence of colicin U shows the highest degree of identity (73%) to the poreforming C-terminal sequence of colicin B, the immunity protein, which interacts with the same region, displays a higher degree of sequence similarity to the immunity protein of colicin A (45%) than to the immunity protein of colicin B (30.5%). Immunity specificity is probably conferred by a short sequence that ranges from residue 571 to residue 599 of colicin U; this sequence is not similar to that of colicin B. Uptake of colicin U by sensitive cells is mediated by the OmpA protein, the OmpF porin, and core lipopolysaccharide, and is dependent on the TolA, B, Q, and R proteins [21].
215
12 Structure, Function, Import, and Immunity of Colicins
Figure 12.1: (A) Differences in the nucleotide sequences, indicated by vertical bars and as percent of identity, of the genes encoding the activity (termed a), immunity (termed i) and lysis genes (termed l) of colicins 10 and K, as compared to those of colicin 5 [19]. The functional domains required for TonB- and TolA-dependent uptake of the colicins via the Tsx/TolC and the OmpA/OmpF outer membrane receptor proteins, and the pore-forming domains are indicated. (B) Percent of identity of the nucleotide sequences (within the bars) and the encoded proteins (above the bars) of colicin A as compared to colicin S4 [20]. The bars indicate the length of the genes and the arrows the transcription polarity. Note the identical gene arrangements of the colicin determinants.
216
12.5 Pore-forming colicins are inactivated by the cognate immunity proteins
Figure 12.2: Dendrogram of the sequence alignment of the pore-forming colicin proteins encoded by the activity genes [21]. Note that the sequence similarities are not uniform along the polypeptide chains but vary strongly for different functional domains, as shown in Fig. 12.1.
12.5 Pore-forming colicins are inactivated by the cognate immunity proteins shortly before the formation of the transmembrane pores
The immunity proteins of the pore-forming colicins reside in the cytoplasmic membrane, the target site of the colicins. The question is how they inactivate the colicins. We have studied this question by identifying possible binding sites on the immunity proteins and the colicins, and by subcellular localization of the immunity protein binding sites. Mutational analysis of colicin 5 and exchange of DNA fragments between the most closely related colicins 5 and 10, and between their immunity proteins localize the regions that determine the reaction specificity between colicin 5 and its immunity protein to residues 405 to 428 of colicin 5. This region corresponds to the amphipatic a-helix of the pore-forming colicins E1 and Ia. The specificity-conferring residues 55–58 and 68–75 of the immunity protein are localized in the cytoplasmic loop and the inner leaflet of the cytoplasmic membrane. The localization of the reactive regions of the immunity protein and the colicin at the inner side of the cytoplasmic membrane suggests that the immunity protein inactivates colicin 5 shortly before the lethal colicin pores in the cytoplasmic membrane are opened. This finding is probably representative for all pore-forming colicins that belong to the colicin 5 subgroup. 217
12 Structure, Function, Import, and Immunity of Colicins The site of interaction of colicin U with the immunity protein was localized in the tip of the hydrophobic hairpin of the C-terminus which inserts into the cytoplasmic membrane [22]. Deletion of the tip resulted in a fully active colicin that was no longer recognized by the cognate immunity protein. Replacement of eight residues at the tip of colicin U by the corresponding sequence of colicin B resulted in a colicin hybrid that was inactivated by the colicin U immunity and the colicin B immunity protein to the same degree that corresponded to 10 % of the inactivation of wild type colicin U. Since the hairpin forms a loop between helices 8 and 9 of colicin U and inserts deeply into the cytoplasmic membrane interaction of the colicin with its immunity protein probably occurs after the colicin has inserted in the cytoplasmic membrane. The transmembrane topology of the immunity proteins of colicins U and 5 was determined and represents two types with four and three transmembrane helices, respectively (Fig. 12.3). An astonishing finding was made when the sequence DGTGW, which we have proposed to be involved in Tol-dependent uptake of colicin U, was replaced by the TonB box DTMVV of colicin B. Uptake of the mutated colicin U now depended on the Ton system but still required the Tol system (TolQRA) but no longer TolB. Obviously, the Ton system functionally replaced the TolB activity [23]. This hybrid protein is the first example of a colicin whose uptake depends on the Ton and the Tol system. This finding strengthens our results which demonstrate the evolutionary relationship between the Ton and the Tol system [24–26].
12.6 Pesticin is a muramidase which is inactivated by the immunity protein in the periplasm
The sequence of the pesticin gene reveals a bacterial toxin that has no sequence homology to any of the colicins except the TonB box at the N-terminus which is identical to the TonB box of colicin B [27]. However, the nucleotide sequences which flank the pesticin activity and immunity genes are highly homologous to the sequences flanking the colicin 10 activity, immunity, and lysis genes. This result is another indication for our notion that during evolution complete bacteriocin determinants and DNA fragments encoding functional domains were exchanged between plasmids by recombination. In the case of the pesticin determinant containing only a rudimentary fragment of a lysis gene, recombination may have destroyed the lysis gene. Pesticin is a muramidase with a specificity as the lysozymes cleave the glycan strands of murein between the C1 position of muramic acid and the C4 position of N-acetylglucosamine [28]. The immunity protein is contained in the periplasm and may inactivate pesticin before it reaches its target. 218
Figure 12.3: Predicted transmembrane topology of the colicin 5 immunity protein (left) and the colicin U immunity protein (right). The transmembrane helices (H) are interconnected by loops (L). PP, periplasm; CM, cytoplasmic membrane; CP, cytoplasm. Arrows indicate the locations of fusion sites with the BlaM b-lactamase used to determine the transmembrane arrangement of the immunity proteins [16, 22].
12.6 Pesticin is a muramidase which is inactivated by the immunity protein
219
12 Structure, Function, Import, and Immunity of Colicins
Acknowledgments
The authors acknowledge the important contributions of Robin Harkness and Tobias Ölschläger.
References
1. V. Braun, H. Pilsl, and P. Gross (1994) Colicins: structures, modes of action, transfer through membranes, and evolution. Arch. Microbiol. 161, 199–206. 2. V. Braun, K. Schaller, and M. R. Wabl (1974) Isolation, characterization, and action of colicin M. Antimicrob. Agents Chemother. 5, 520–533. 3. K. Schaller, J. V. Höltje, and V. Braun (1982) Colicin M is an inhibitor of murein biosynthesis. J. Bacteriol. 152, 994–1000. 4. R. E. Harkness and V. Braun (1989) Colicin M inhibits peptidoglycan biosynthesis by interfering with lipid carrier recycling. J. Biol. Chem. 264, 6177–6182. 5. R. E. Harkness and V. Braun (1989) Inhibition of lipopolysaccharide O-antigen synthesis by colicin M. J. Biol. Chem. 264, 14716–14722. 6. P. Gross and V. Braun (1996) Colicin M is inactivated during import by its immunity protein. Mol. Gen. Genet. 251, 388–396. 7. T. Ölschläger, E. Schramm, and V. Braun (1984) Cloning and expression of the activity and immunity genes of colicins B and M on ColBM plasmids. Mol. Gen. Genet. 196, 482–487. 8. G. Thumm, T. Ölschläger, and V. Braun (1988) Plasmid pColBM-Cl139 does not encode a colicin lysis protein but contains sequences highly homologous to the D protein (resolvase) and the oriV region of the miniF plasmid. Plasmid 20, 75–82. 9. U. Pressler, V. Braun, B. Wittmann-Liebold, and R. Benz (1986) Structural and functional properties of colicin B. J. Biol. Chem. 261, 2654–2659. 10. E. Schramm, J. Mende, V. Braun, and R. M. Kamp (1987) Nucleotide sequence of the colicin B activity gene cba: consensus pentapeptide among TonB-dependent colicins and receptors. J. Bacteriol. 169, 3350–3357. 11. J. Köck, T. Ölschläger, R. M. Kamp, and V. Braun (1987) Primary structure of colicin M, an inhibitor of murein biosynthesis. J. Bacteriol. 169, 3358–3361. 12. J. Mende and V. Braun (1990) Import-defective colicin B derivatives mutated in the TonB box. Mol. Microbiol. 4, 1523–1533. 13. H. Pilsl, C. Glaser, P. Gross, H. Killmann, T. Ölschläger, and V. Braun (1993) Domains of colicin M involved in uptake and activity. Mol. Gen. Genet. 240,103–112. 14. H. Schöffler and V. Braun (1989) Transport across the outer membrane of Escherichia coli K12 via the FhuA receptor is regulated by the TonB protein of the cytoplasmic membrane. Mol. Gen. Genet. 217, 378–383. 15. D. E. Bradley and S. P. Howard (1992) A new colicin that adsorbs to outer-membrane protein Tsx but is dependent on the tonB instead of the tolQ membrane transport system. J. Gen. Microbiol. 138, 2721–2724. 16. H. Pilsl and V. Braun (1995) Evidence that the immunity protein inactivates colicin 5
220
References
17.
18. 19. 20.
21. 22.
23. 24.
25.
26.
27.
28.
immediately prior to the formation of the transmembrane channel. J. Bacteriol. 177, 6966–6972. H. Pilsl and V. Braun (1995) Novel colicin 10: assignment of four domains to TonBand TolC-dependent uptake via the Tsx receptor and to pore formation. Mol. Microbiol. 16, 57–67. U. Roos, R. E. Harkness, and V. Braun (1989) Assembly of colicin genes from a few DNA fragments. Nucleotide sequence of colicin D. Mol. Microbiol. 3, 891–902. H. Pilsl and Braun V. (1995) Strong function-related homology between the poreforming colicins K and 5. J. Bacteriol. 177, 6973–6977. H. Pilsl, D. Smajs, and V. Braun (1999) Characterization of colicin S4 and its receptor, OmpW, a minor protein of the Escherichia coli outer membrane. J. Bacteriol. 181, 3578–3581. D. Smajs, H. Pilsl, and V. Braun (1997) Colicin U, a novel colicin produced by Shigella boydii. J. Bacteriol. 179, 4919–4928. H. Pilsl, D. Smajs, and V. Braun (1998) The tip of the hydrophobic hairpin of colicin U is dispensable for colicin U activity but is important for interaction with the immunity protein. J. Bacteriol. 180, 4111–4115. H. Pilsl and V. Braun (1998) The Ton system can functionally replace the TolB protein in the uptake of a mutated colicin U. FEMS Microbiol. Lett. 164, 363–367. K. Eick-Helmerich and V. Braun (1989) Import of biopolymers into Escherichia coli: nucleotide sequences of the exbB and exbD genes are homologous to those of the tolQ and tolR genes, respectively. J. Bacteriol. 171, 5117–5126. V. Braun (1989) The structurally related exbB and tolQ genes are interchangeable in conferring tonB-dependent colicin, bacteriophage, and albomycin sensitivity. J. Bacteriol. 171, 6387–6390. V. Braun and C. Herrmann (1993) Evolutionary relationship of uptake systems for biopolymers in Escherichia coli: cross-complementation between the TonB-ExbB-ExbD and the TolA-TolQ-TolR proteins. Mol. Microbiol. 8, 261–268. H. Pilsl, H. Killmann, K. Hantke, and V. Braun (1996) Periplasmic location of the pesticin immunity protein suggests inactivation of pesticin in the periplasm. J. Bacteriol. 178, 2431–2435. W. Vollmer, H. Pilsl, K. Hantke, J. V. Höltje, and V. Braun (1997) Pesticin displays muramidase activity. J. Bacteriol. 179, 1580–1583.
221
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens Hemolysin/Cytolysin Volkmar Braun* and Ralf Hertle
13.1 Introduction
When the characterization of the cytolysin (hemolysin) of Serratia marcescens started [1], it was known that S. marcescens secretes a number of exoenzymes, proteases, and chitinases, a lipase and a nuclease, in contrast to the Escherichia coli K-12 laboratory strain, which does not secrete any proteins. Both organisms belong to the Enterobacteriaceae, and we have been interested in determining the secretory activities of S. marcescens that are lacking in E. coli. We biochemically characterized an exoprotease of S. marcescens [2], and during these studies, we examined whether the unspecific exoprotease lyses erythrocytes to obtain a convenient and sensitive assay for the protease activity. All S. marcescens strains tested lysed erythrocytes – not by the exoprotease, but by an unknown activity. A hemolytic activity on blood agar plates had occasionally been mentioned in the literature, but no data on a hemolysin were available. In a book on the genus Serratia for clinical microbiologists, infectious disease physicians, and epidemiologists, the terms hemolysin and cytolysin are not mentioned, which shows that hemolysis was not considered as a trait of Serratia [3]. The hemolysin was almost neglected because of the very small lysis zones around colonies on standard blood agar plates, which, as we now know, are caused by 1) a low diffusion due to the high molecular weight of the hemolysin [4], 2) its immediate aggregation upon release from the cells [5], 3) degradation by exoproteases, and 4) synthesis only under iron-limiting growth conditions [6]. On blood agar plates S. marcescens can easily meet its iron requirement by active transport through ferric siderophores, heme [7], and hemoglobin [8], resulting in repression of hemolysin synthesis. Eleven S. marcescens strains and two S. liquefaciens strains tested were found to be hemolytic [9], and a subsequent
* Mikrobiologie/Membranphysiologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
222 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
13.1 Introduction survey of nosocomial S. marcescens strains identified 58 hemolytic strains among 60 strains tested [10]. DNA probes of the S. marcescens hemolysin genes do not hybridize to DNA of E. coli, Salmonella typhimurium, Proteus mirabilis, Proteus vulgaris, Citrobacter freundii, Enterobacter cloacae, Klebsiella aerogenes, Klebsiella pneumoniae, Shigella dysenteriae, Yersinia enterocolitica, Yersinia pseudotuberculosis, Listeria sp., Aeromonas sp., Legionella sp., and Meningococcus sp. [9], and polyclonal antibodies against the S. marcescens hemolysin do not react with the rather similar hemolysin of P. mirabilis. Lack of DNA cross-hybridization and antibody cross-reaction makes the hemolysin a suitable marker for the diagnosis of S. marcescens and S. liquefaciens, which are the most frequent Serratia strains among clinical isolates. The S. marcescens hemolysin (ShlA) represents a new type of hemolysin and has been studied at the molecular level in great detail with regard to structure, activation, and secretion. It has nothing in common with the hemolysins (RTX toxins) of the E. coli type [11, 12]. After we had characterized the S. marcescens hemolysin on the molecular level, hemolysins homologous to the S. marcescens hemolysin were found in P. mirabilis and P. vulgaris [13], Haemophilus ducreyi [14], and Edwardsiella tarda [15]. Furthermore, sequence motifs shown to be important for activity and secretion of the S. marcescens hemolysin [16] are also contained in the filamentous hemagglutinin FHA of Bordetella pertussis [17], the hemopexin HxuA/HxuB of Haemophilus influenzae [18], and the adhesins HMW2A/HMW2B of Haemophilus influenzae [19] (Table 13.1). Like the hemolysins of S. marcescens and the related hemolysins, the latter proteins also require a protein for secretion across the outer membrane; these secretory proteins display sequence similarity to the S. marcescens secretory protein. Thus, the S. marcescens hemolysin forms the prototype of a new class of hemolysins and of a new secretory mechanism. Table 13.1: Common sequence motifs related to functionally essential regions of the Serratia marcescens hemolysin. Serratia marcescens hemolysin ShlA Edwardsiella tarda hemolysin EthA Proteus mirabilis hemolysin HpmA Haemophilus ducreyi hemolysin HhdA Bordetella pertussis filamentous hemagglutinin FHA Serratia marcescens hemolysin ShlA Edwardsiella tarda hemolysin EthA Proteus mirabilis hemolysin HpmA Haemophilus ducreyi hemolysin HhdA Bordetella pertussis filamentous hemagglutinin FHA Haemophilus influenzae Type b heme:hemopexin HxuA Haemophilus influenzae surface protein HMW
68-ANPNL 68-ANPNL 68-ANSNL 66-ANPHL 65-KNPNL 109-NPNGIS 108-NPNGIT 109-NPNGIT 107-NPNGMS 94-NPNGIS 86-NPNGVI 150-NPNGIT
Numbering corresponds to the proteins after cleavage of the signal peptide, which is unusually long for FHA (71 residues) and not known for HMW.
223
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens
13.2 Characterization of the S. marcescens hemolysin (ShlA)
13.2.1 ShlA forms pores in eukaryotic plasma membranes, but not in prokaryotic plasma membranes ShlA causes hemolysis by forming pores in erythrocyte membranes [4]. The onset of hemolysis is progressively retarded and the hemolysis rate diminishes when oligosaccharides of increasing molecular weight are added to the assay at 30 mM concentration, which corresponds to the internal osmotic pressure of erythrocytes. Maltoheptaose (Mr 1,152) is highly protective, and dextran 4 (Mr 4,000) prevents hemolysis. Removal of dextran 4 and the surplus of hemolysin immediately results in hemolysis by ShlA, which had been inserted into the erythrocyte membrane in the presence of dextran 4. Prevention of hemolysis by oligosaccharides demonstrates formation of ShlA pores of a limited size range through which water and smaller oligosaccharides flow into the erythrocytes and cause osmolysis. The larger oligosaccharides prevent osmolysis because they are too large to enter the erythrocytes through the ShlA pores and thus counterbalance the internal osmotic pressure of the erythrocytes. Comparison of the ShlA pores with the structurally defined pores (heptamers) of the Staphylococcus aureus a-toxin has revealed that the ShlA pores are smaller and vary in size and are larger at 28 than at 0 8C [20]. Protease digestion experiments have revealed how ShlA is inserted into the erythrocyte membrane. The hemolysin integrates into the erythrocytes such that it is not cleaved by trypsin added to erythrocytes and sealed right-side-out erythrocyte ghosts [21]. In unsealed ghosts and inside-out vesicles, ShlA is cleaved by trypsin, demonstrating accessibility of ShlA to trypsin from the inside of the erythrocytes. The most sensitive cleavage site results in a 143-kDa and a 19.5-kDa fragment, both of which remain in the erythrocyte membrane. This site is close to the C-terminus of ShlA. Upon longer incubations, trypsin yields fragments of 138, 89, and 58 kDa. Only a few ShlA sites in the C-terminal half of ShlA, all of which are exposed to the inside of the erythrocytes, are amenable by trypsin. A genetically engineered N-terminal 72-kDa ShlA fragment adsorbs to but does not integrate into the erythrocyte membrane, as shown by its complete degradation by trypsin. For insertion of ShlA into erythrocyte membranes, ShlA must be modified by the outer membrane protein ShlB, and adsorption of the 72-kDa ShlA depends on co-synthesis with ShlB (Fig. 13.1). ShlA forms pores in erythrocytes from various animal sources. This raises the question why ShlA does not kill bacterial producer cells. Activation during secretion across the outer membrane by ShlB could be a means to protect the bacterial producer cells. This theory has been tested by transforming stable protoplast-type L-forms of Proteus mirabilis lacking an outer membrane with a 224
13.2 Characterization of the S. marcescens hemolysin (ShlA)
Figure 13.1: Arrangement and transcriptional polarity of the shlA and shlB genes. ShlB inserts into the outer membrane (OM), and activates and secretes ShlA (a), and the Cterminally truncated ShlA-255 (c). In the absence of ShlB non-hemolytic ShlA* remains in the periplasm (b).
plasmid carrying the shlA shlB genes [22]. Inactive ShlA* is secreted by the Lform cells, and ShlB is associated with the cytoplasmic membrane. Addition of hemolytic ShlA to the L-form cells has no effect, which suggests that the prokaryotic cytoplasmic membrane is resistant to ShlA. In artificial lipid bilayers, ShlA forms water-filled channels with an inner diameter of 1–3 nm, depending on how ShlA is prepared [20]. The data suggest that pre-formed ShlA monomers and dimers can insert into erythrocyte membranes and that larger oligomers may form in the erythrocyte membranes at 28 8C. However, oligomerization within the erythrocyte membrane is not required for pore formation, which occurs more rapidly at 0 8C than at 28 8C [21]. At 0 8C the lateral mobility of integral membrane proteins is greatly reduced so that oligomerization cannot contribute much to pore formation.
13.2.2 Two proteins determine the S. marcescens hemolytic activity Bacterial protein toxins are necessarily chimeric proteins – they must be hydrophilic to be soluble when released from the bacteria, but they also have to be hydrophobic to enter into or through the plasma membrane of eukaryotic cells. 225
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens These properties frequently result in aggregation of the toxins in aqueous solution, and the S. marcescens hemolysin is no exception; this makes its biochemical characterization difficult [1]. The breakthrough in the precise description of many bacterial toxins came with the recombinant DNA techniques, which yielded accurate data on the number of proteins involved, their molecular weights, and the transcriptional regulation of protein synthesis, and which yielded sufficient quantities of the proteins for functional studies. However, in the case of the S. marcescens hemolysin, biochemical experiments could only be performed after it was discovered that ShlA can be kept in the soluble form in 6 M urea without loss of activity and that precipitated ShlA aggregates can be solubilized in 6 M urea in an active form. ShlA even withstands irreversible denaturation by 10 % trichloroacetic acid, as shown by the hemolytic activity of TCA-precipitated ShlA solubilized in 6 M urea. In contrast, these treatments denature ShlB; however, ShlB, in contrast to ShlA, can be solubilized in an active form in mild detergents, such as octylglucoside. The hemolysin was characterized in transformants of E. coli K-12 that carried both genes encoding ShlA and ShlB (shlA shlB), shlA or shlB, or mutated shlA and shlB on plasmids. In the cases examined, the properties of the E. coli transformants agreed with the properties of wild type S. marcescens. Genetically, the hemolysin trait of S. marcescens is very stable. We have never observed spontaneous non-hemolytic mutants. A 7.5-kb chromosomal DNA fragment of S. marcescens renders E. coli K-12 transformants hemolytic. The nucleotide sequence of the DNA fragment reveals two open reading frames, named shlA and shlB, which are transcribed from shlB to shlA (Fig. 13.1). shlA encodes the hemolysin, and shlB encodes an outer membrane protein that is required for the secretion of the hemolytic ShlA protein across the outer membrane into the culture medium. Mature ShlA is composed of 1578 amino acids, and mature ShlB contains 539 amino acids. Both proteins are synthesized as precursors that contain signal peptides of 30 and 18 amino acids, respectively, which are cleaved off during export of the polypeptides across the cytoplasmic membrane [23]. In S. marcescens, synthesis of the hemolysin is repressed by iron. In rich media, iron limitation by the iron-chelator 2,2'-dipyridyl (0.3 mM) strongly increases hemolysin synthesis [5, 6] despite the numerous iron supply systems that exist in this organism [7]. In E. coli, transcription of the shlB shlA genes is co-regulated by the Fur protein [24], which, when loaded with Fe2+, functions as a transcriptional repressor of iron-controlled genes. Fur-Fe2+ binds to the Fur box, a DNA consensus sequence composed of 19 nucleotides; a similar sequence is located in the –30 region of the promoter upstream of shlB. A fur deletion mutant of E. coli, transformed with the shlA shlB genes, produces tenfold more hemolysin than a fur wild type strain grown in a medium with sufficient iron [6].
226
13.2 Characterization of the S. marcescens hemolysin (ShlA)
13.2.3 Domains in ShlA responsible for secretion, binding to eukaryotic membranes, and pore formation The long polypeptide of 1608 amino acids encoded by the shlA gene can be divided into four functional regions. The N-terminal signal sequence serves for the Sec-dependent export of the ShlA protein [11]. The functional regions of the mature form have been identified by progressive C-terminal truncation of the ShlA protein through genetic means. Reduction by 25% results in a hemolysin that displays 20 % of the wild type hemolysis rate. Shorter ShlA fragments cause only residual or no hemolysis [23]. A 72-kDa fragment (wild type mature protein: 160,000 kDa) still binds to erythrocytes, but does not lyse them. The fragment can be degraded by trypsin, whereas wild type hemolysin becomes trypsin-resistant upon integration into the erythrocyte membrane [21]. An N-terminal fragment of 238 amino acids (ShlA-238) is the shortest fragment that can still be secreted. ShlA-238 also converts inactive ShlA (ShlA*), isolated from the periplasm of an ShlB-negative strain, to an active hemolysin. This kind of activation, which we term complementation because of its similarity to the a-complementation of b-galactosidase, was discovered when an E. coli shlB shlA' transformant that secretes a 269-residue ShlA fragment and an E. coli shlA transformant were streaked in parallel on a blood agar plate. None of the individual transformants were hemolytic since the ShlA-269 fragment secreted by the ShlB protein is non-hemolytic and ShlA*, which remains in the periplasm in the absence of ShlB, is also non-hemolytic. After incubation for several days, a zone of hemolysis appeared around the ShlA*-producing cells, mainly on the side where the ShlA-269 cells were located. Apparently, ShlA-269 diffused to the ShlA* cells and gained access to ShlA*, a fraction of which was unspecifically released by partial lysis of some of the cells during the extended incubation period. This experiment was then repeated with an ShlA-269-containing spent medium; ShlA* in a crude cell extract was rendered hemolytic. Activation of ShlA* by ShlA-269 is reversible, as shown by the inactivation of ShlA* after chromatographic separation of ShlA* and ShlA-269 [25]. Complementation can also be achieved with the smaller ShlA-238 and even with a trypsin degradation fragment of ShlA-269 that consists of only 149 N-terminal residues of ShlA [14]. These data demonstrate that ShlA-238 contains all the information for secretion and activation of ShlA*. For complementation, secretion of ShlA-238 by ShlB is required; periplasmic ShlA-238 synthesized in the absence of ShlB does not activate ShlA*. Conversion of ShlA* to ShlA by the trypsin fragment also occurs only when the trypsin fragment is isolated from a secreted ShlA polypeptide. Phospholipase A2 inactivates ShlA-255 so that it no longer complements ShlA* to ShlA; this indicates binding of PE to the N-terminus of ShlA, a requirement for the activation by ShlB and for the hemolytic activity [26]. ShlA contains the sequence ANPN twice; this sequence rarely occurs in proteins. One of the first asparagine residues of the two tetrapeptides, N-69 or N-109, was replaced by isoleucine, and N-69 was also replaced by lysine. All three mutant proteins are not secreted and are not hemolytic in whole cells and 227
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens after extraction with 6 M urea. The high specificity of these mutations is demonstrated by a mutant with a substitution of N-111 to isoleucine in which secretion and hemolytic activity is fully retained. shlA mutants with deletions covering N69 (ShlAD68–97) and N-109 (ShlAD99–117) are also not secreted and are nonhemolytic. The ShlA derivatives carrying point mutations and deletions gain activity by in vitro complementation with ShlA-269 [14]. These results clearly indicate that the N-terminal region of ShlA carries the information for secretion and activation by ShlB. Superhemolytic ShlA mutants have been isolated by treatment of plasmidencoded shlA with hydroxylamine [27]. Three mutants with hemolysis rates 7to 20-fold higher than that of wild type ShlA were studied in some detail. Two mutants carry single amino acid replacements, glycine to aspartate at position 326 (G326D) and serine to asparagine at position 386 (S386N), and the third mutant contains two mutations (G326D and N236D). The higher activity of the mutant ShlA proteins is mainly due to a greatly reduced aggregation. The half-life of wild type ShlA activity in the spent medium is 2.5 min and that of the mutant ShlA proteins is 10, 30, and 40 min. The superhemolytic mutants differ most strongly from wild type ShlA by their failure to cause hemolysis at 0 8C when dissolved in 6 M urea or 6 M guanidinium chloride to prevent spontaneous precipitation. At 0 8C, adsorption of the mutant ShlA proteins to erythrocytes is greatly reduced. At 20 8C, the mutant ShlA proteins in 6 M urea or 6 M guanidinium chloride lyse erythrocytes with rates similar to that of wild type ShlA. Since residues G-326 and S-386, and perhaps N-236, contribute to the aggregation of ShlA, they define important sites of ShlA activity. The strong effects displayed by these mutants are surprising if one considers the conservative nature of the amino acid replacements, and in addition the large size of ShlA.
13.2.4 The dual function of ShlB in secretion and phosphatidylethanolamine-dependent activation of ShlA In an experiment to localize the hemolysin formed by an E. coli shlA shlB transformant, 480 hemolytic units were determined in the spent medium, 205 units were associated with the cell surface, and 250 units were within the cells. Using radiolabeled ShlA, 59% of the label was found in the spent medium and 41% was associated with the cells. In the absence of ShlB, inactive ShlA* remains completely within cells. Immunogold electron microscopy with anti-ShlA antibodies identified most of the ShlA* in the periplasm and some in the cytoplasm. The hemolytic activity found in ShlA*-producing cells is 0.1% of the total activity of ShlA-producing cells [5]. These experiments demonstrate that ShlB is required for activation of ShlA* and for secretion of ShlA. If activation of ShlA* is a catalytic reaction and ShlB forms pores for secretion of ShlA, much less ShlB than ShlA should fulfill these two functions. This is clearly not the case because ShlA* cannot be activated by ShlB when the crude 228
13.2 Characterization of the S. marcescens hemolysin (ShlA) cell extracts used contain much more ShlA* than ShlB. In addition, the ShlB clone used synthesizes an N-terminal ShlA fragment that is activated by ShlB, which presumably further decreases the amount of ShlB available to activate ShlA*. To demonstrate in vitro activation of ShlA* by ShlB uncoupled from secretion, the genes of both proteins have to be overexpressed and the proteins have to be synthesized in similar amounts [25]; this reflects the in vivo situation where shlB is transcribed prior to shlA and at least as much ShlB is formed as ShlA. However, when both ShlA* and ShlB are highly purified by ion-exchange column chromatography, no active ShlA is formed. The missing component in the assay is phosphatidylethanolamine (PE), which is the major phospholipid (90 %) of the E. coli outer membrane. Phosphatidylserine, the biosynthetic precursor of PE, is similarly active, but does not occur in the outer membrane; phosphatidylglycerol has 10 % of the PE activity; and phosphatidylcholine, cardiolipin, lyso-PE, phosphatidic acid, lipopolysaccharide, and various detergents have no effect [26]. Approximately four PE molecules bind so tightly to ShlA* without the help of ShlB that they remain bound to ShlA* during SDS-PAGE and are only removed by thin-layer chromatography with organic solvents. Binding of PE to ShlA* does not convert ShlA* to hemolytic ShlA. ShlB has to be added for ShlA*-PE activation. Removal of the fatty acid at the C2 position of PE by phospholipase A2 inactivates ShlA and the resulting lyso-PE dissociates from ShlA. Evidence for the in vivo relevance of these in vitro results comes from experiments with an E. coli mutant devoid of PE due to a mutation in the pss gene, which encodes phosphatidylserine synthase [28]. The hemolytic activity of the pss mutant transformed with an shlA-shlB-encoding plasmid is 9% of the wild type activity and probably arises from the highly elevated phosphatidylglycerol content (46%), which can partially substitute for PE. This inactive ShlA, termed ShlA8 to differentiate it from the periplasmic ShlA*, is contained in the culture medium of the pss mutant, and the activity amounts to 16% (measured by complementation, see below) of the total hemolytic activity of a PEsynthesizing strain [26].
13.2.5 ShlB has the potential to form membrane pores through which ShlA might be secreted To examine the question whether ShlB forms a pore through which ShlA is secreted across the outer membrane, ShlB and deletion derivatives of ShlB were added to artificial lipid bilayer membranes and the change in the membrane conductance was measured. This procedure has been successfully applied to FhuA, an outer membrane receptor for various bacterial viruses, bacterial toxins, and the iron carrier ferrichrome; the conductance was not increased unless a surface-exposed loop of FhuA was deleted, which converted the FhuA closed channel into a permanently open channel [29]. 229
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens Various fragments of ShlB were excised by genetic engineering. The resulting proteins were extracted in octylglucoside from the outer membrane and then highly purified by two ion-exchange column chromatographies to remove completely the porins, which have a high propensity to form channels in lipid bilayer membranes. Only two deletion derivatives could be solubilized and purified. Wild type ShlB causes an increase of the conductance of 1 nS with an irregular frequency and after a few milliseconds, the conductance decreases to the zero level, which indicates that ShlB has the potential to form a channel [30]. ShlBD87–153 and ShlBD65–168 increase the membrane conductance stepwise by 1.2 nS, which demonstrates formation of rather stable single channels. Frequently, two to three channel opening events can be observed, followed by one or two closing steps, and the open periods last longer than the closed periods. The deletion derivatives support the channel-forming properties of ShlB, and we propose that ShlB forms a closed channel which can be opened when ShlA is secreted. A topology model of ShlB predicts 20 transmembrane regions interconnected by loops at the cell surface and short turns in the periplasm [30]. This model is based on the reaction of an antigenic epitope inserted at one of 22 sites along the ShlB polypeptide. Intact cells of 16 epitope mutants react with the monoclonal antibody, which demonstrates accessibility of the epitope at the cell surface. In six mutants, the antibody reacts only with the isolated outer membrane, which indicates a periplasmic or transmembrane location of the epitope. A computer-assisted program [31] for the prediction of membrane-spanning b-strands of outer membrane proteins localizes the six latter epitope sites within the membrane half oriented toward the periplasm. According to this model, the deletions in the two ShlB mutants that form rather stable channels in artificial lipid bilayer membranes comprise portions of the two largest ShlB loops at the cell surface, extending from residues 60 to 97 and 120 to 213, two transmembrane segments, and one periplasmic turn. Secretion-competent but activation-incompetent ShlB mutants have been isolated in which secretion is uncoupled from activation. Two tetrapeptide insertion mutants created with a TAB linker at position 136 or 224 of mature ShlB and a deletion mutant (ShlBD154–252) secrete inactive ShlA. A TAB linker mutant at position 332 has a low secretion activity and contains small amounts of cell-bound active hemolysin. The secreted non-hemolytic ShlA proteins are completely degraded by trypsin, in contrast to hemolytic ShlA, which is cleaved into two fragments of 60 and 100 kDa. Secreted non-hemolytic ShlA is converted in vitro into hemolytic ShlA by isolated wild type ShlB and by complementation with an N-terminal ShlA fragment of 255 residues (ShlA-255). Other mutants secrete different amounts of active hemolysin. The secretion-competent but activation-negative mutants define sites in ShlB which are important for activation. That active hemolysin remains with the cells suggests that the region around residue 332 is involved in secretion [32].
230
13.3 Pathogenicity of S. marcescens hemolysin/cytolysin
13.3 Pathogenicity of S. marcescens hemolysin/cytolysin
The hemolysin of S. marcescens is not only responsible for the pathogenicity of this bacterial strain. S. marcescens is an important opportunistic pathogen that causes respiratory and urinary tract infections, bacteremia, endocarditis, keratitis, arthritis, and meningitis [33, 34]. In Section 13.2, the biochemical characterization of the hemolytic activity was presented and the activity was designated as a hemolysin. However, this hemolysin also is a cytolysin, damages tissues and causes the release of the inflammatory mediators leucotrienes (LTB4 and LTC4) from leukocytes and the release of histamine from rat mast cells [35, 36]. ShlA also contributes to the uropathogenicity of the pathogenic E. coli 536/21 after transformation with the S. marcescens shlA shlB genes [37]. The major factor of this observed pathogenicity is the hemolysin/cytolysin ShlA; however, the pathogenicity of Serratia strains is a multi-factorial process and also includes urease, fimbriae, proteases, lipase, and undefined determinants that facilitate invasion. These factors act in concert, and the resulting effects are adherence, host cell invasion, cytotoxic effects, and final cytolysis. Because the hemolytic activity is mainly cell associated, its effect arises predominantly after adherence of the bacteria to the host cell tissue.
13.3.1 Adherence of S. marcescens Colonization of epithelial tissue requires adherence of bacteria to the target cells. S. marcescens strains produce type 1 fimbriae [38, 39] or US5 pili [40] that are involved in adherence to epithelial cells. Non-fimbriated mutants are markedly decreased in their ability to adhere. Fimbriae of S. marcescens also contribute to superoxide production in neutrophils [35, 36] and phagocytosis [41]. Radiolabeled bacteria adhere to purified granulocytes and are subsequently phagocytized. The superoxide production, determined by chemiluminescence, is dependent on the fimbriae and also on hemolysin production [35]. However, S. marcescens also secretes exoproteases and lipases, which also induce chemiluminescence. The mutant strain W1436 lacks exoprotease and lipase activity, but is still active in chemiluminescence – although less than the wild type strain 5817 – and is more active in causing histamine release [35]. It is not clear whether these activities are only related to hemolysin. With purified ShlA, no oxidative burst can be detected; instead, granulocytes continuously increase chemiluminescence [42]. These data indicate that the oxidative burst detected with viable hemolytic S. marcescens strains is the sum of cytotoxic effects of various products secreted by the bacteria and not only of the cytolysin. The S. marcescens hemolysin contributes to colonization of the urinary tract epithelium in an experimental rat model [37]. An E. coli 536/21 transfor231
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens mant that produces ShlA is five times more efficient in colonization than the ShlA-negative recipient strain, but it is not clear whether ShlA acts as a cellbound adhesin or facilitates the colonization by its cytolytic activity. On epithelial cell cultures (HEp-2, HeLa), ShlA-producing S. marcescens strains W1128 or CDC04:H4, for example, are adherent [43]. A non-adherent strain, E. coli BL21, transformed with the shlAB genes, which lead to the production of 30-fold more hemolysin than the Serratia wild type, is not adherent. Therefore, it is clear that adherence is not primarily mediated by the hemolysin. The adherence of various S. marcescens strains tested in our laboratory is mainly determined by mannose-sensitive type I fimbriae. We were able to distinguish between specific adhesion on epithelial cells and unspecific adherence to polystyrene. LPS rough mutants showed enhanced adherence even in the presence of mannose, indicating that not only type I fimbriae but also additional factors, such as LPS, are involved in adherence.
13.3.2 Invasion of S. marcescens in epithelial cells S. marcescens invades tissue cells [42]. Using the gentamycin protection assay, cells of the intracellular S. marcescens wild type strain W225, and of strains W1126 and CDC04:H4 have been isolated from HEp-2, HeLa, and RT112 cells. Inhibition of eukaryotic protein synthesis by cycloheximide or alteration of the cytoskeleton by cytochalasin D does not reduce the uptake of the bacterial cells. The cytotoxicity of these strains mediated by ShlA reduces the number of viable bacteria protected against gentamycin because the bacteria lyse their host cells and become amenable to gentamycin. Isogenic hemolytic-negative S. marcescens W225 and W1128 strains, constructed by site-directed mutagenesis [43], display a lower invasiveness than the ShlA wild type and are less cytotoxic in the LDH release assay. From the data above, it is concluded that adherence mediated by fimbriae or pili of the bacteria directs the hemolysin to the target membranes and the pore-forming toxin gains access to the membrane, but the hemolysin does not mediate adherence by itself. More likely, pore formation may play a role in invasion, but renders invasion studies difficult due to ShlA cytotoxicity.
13.3.3 Cytotoxic effects and cytolysis by the S. marcescens hemolysin ShlA is inactive on keratinocytes, endothelial cells, and monocytes, all of which are specific target cells for other bacterial toxins, such as the staphylococcal a-toxin [44–46]. ShlA induces ATP depletion and potassium efflux in epithelial cells and in fibroblasts [47]. The depletion of cellular ATP is very strong even 232
13.3 Pathogenicity of S. marcescens hemolysin/cytolysin with sublytic doses of ShlA. Treatment of HEp-2 or HeLa cells with 1 µg ShlA/ ml decreases the initial ATP level to 10 % within 30 min. In parallel, intracellular potassium is released into the supernatant through pores in the plasma membrane formed by ShlA. This is the signal for the Na/K-ATPase in the membrane to transport potassium into the cell, but the leakage through the ShlA pores is greater than the import. In an attempt to restore the intracellular potassium pool, ATP is consumed. This effect on the Hep-2 and HeLa cells, as on erythrocytes, is also observed with inactive ShlA* complemented with ShlA255. Pores formed by ShlA in the eukaryotic cytoplasmic membrane are smaller than the 1 to 3 nm estimated for ShlA pores in artificial black lipid membranes and in erythrocytes. The pores do not support the influx of propidium iodide (Mr 668) and trypan blue, as measured by flow cytometry with cells depleted to 10 % of the initial ATP level. ATP depletion is not affected by osmoprotection with oligosaccharides, which prevent ShlA-mediated cells lysis [47]. The observed ATP depletion caused by pore formation is reversible up to 80 % of the initial ATP level. Depletion of the initial ATP level to less than 20 % decreases the recuperation in HEp-2 cells (40 % restored) because the cells begin to lyse. Restoration of the ATP level presumably occurs by the repair or closure of the ShlA-produced pores in the cytoplasmic membrane. The repair requires protein synthesis, as evidenced by the failure in the presence of cycloheximide [47].
13.3.4 ShlA-mediated vacuolation The osmotic imbalance caused by ShlA, indicated by potassium efflux, ATP depletion, and cell swelling, results in vacuolation and is observed with epithelial cells, but not with fibroblasts [47]. Small vacuoles appear all over the cytoplasm; upon prolonged incubation, the small vacuoles fuse to form large vacuoles with irregular shape that fill the entire cytoplasm (Fig. 13.2). No lysis, as determined by the LDH assay, can be seen during vacuolation. Isolated ShlA and hemolytic S. marcescens strains cause vacuolation, which is thought to be the result of an osmotically driven influx of water [47]. Oligosaccharides with a molecular mass up to 1152 Da (maltoheptaose) strongly reduce vacuolation and cytotoxicity, and oligosaccharides larger than 1400 Da (dextrin 15) completely inhibit vacuolation. The data demonstrate formation of pores by ShlA and a pore size in nucleated cells smaller than that in erythrocytes. Removal of the oligosaccharides results in vacuolation and cytolysis of ShlA-pretreated cells. Unexpectedly, bacteria found in the large vacuoles formed in the late stage of infection are highly mobile. The reason for this is unknown, but the bacteria may undergo a transition to a highly mobile swarming state during the infection process. In contrast to vacuoles formed by the Helicobacter pylori cytotoxin VacA [48] vacuoles induced by ShlA are not acidified, vacuolation is not inhibited by bafilomycin, and vacuolation cannot been reversed [47]. 233
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens
Figure 13.2: Photomicrographs of untreated HEp-2 cells (A) and vacuolation of HEp-2 cells after 10 min (B) and 45 min (C) treatment with ShlA. Cells were cultured in a 24-well plate until they reached subconfluence. Medium was supplemented with 100 µl of a bacterial culture supernatant containing ShlA (30 HU/ml). After incubation at 37 8C under an atmosphere of 5 % CO2, cells were examined by phase-contrast microscopy at 320 x magnification.
234
References
References
1. Braun, V., Günther, H., Neuss, B., and Tautz, C. (1985) Hemolytic activity of Serratia marcescens. Arch. Microbiol. 141, 371–376. 2. Schmitz, G. and Braun, V. (1985) Cell-bound and secreted proteases of Serratia marcescens. J. Bacteriol. 161, 1002–1009. 3. von Graevenitz, A. and Rubin, S. J. (1980) The genus Serratia. CRC Press, Boca Raton, Florida. 4. Braun, V., Neuss, B., Ruan, Y., Schiebel, E., Schöffler, H., and Jander, G. (1987) Identification of the Serratia marcescens hemolysin determinant by cloning into Escherichia coli. J. Bacteriol. 169, 2113–2120. 5. Schiebel, E., Schwarz, H., and Braun, V. (1989) Subcellular location and unique secretion of the hemolysin of Serratia marcescens. J. Biol. Chem. 264, 16311–16320. 6. Poole, K. and Braun, V. (1988) Influence of growth temperature and lipopolysaccharide on hemolytic activity of Serratia marcescens. J. Bacteriol. 170, 5146–5152. 7. Angerer, A., Klupp, B., and Braun, V. (1992) Iron transport systems of Serratia marcescens. J. Bacteriol. 174, 1378–1387. 8. Ghigo, J. M., Letoffe, S., and Wandersman, C. (1997) A new type of hemophore-dependent heme acquisition system of Serratia marcescens reconstituted in Escherichia coli. J. Bacteriol. 179, 3572–3579. 9. Ruan, Y. and Braun, V. (1990) Hemolysin as a marker for Serratia. Arch. Microbiol. 154, 221–225. 10. Carbonell, G. V. and Vidotto, M. C. (1992) Virulence factors in Serratia marcescens: cell-bound hemolysin and aerobactin. J. Med. Biol. Res. 25, 1–8. 11. Braun, V. and Focareta, T. (1991) Pore-forming bacterial protein hemolysins (cytolysins). Crit. Rev. Microbiol. 18, 115–158. 12. Braun, V., Schönherr, R., and Hobbie, S. (1993) Enterobacterial hemolysins: activation, secretion and pore formation. Trends. Microbiol. 1, 211–216. 13. Uphoff, T. S. and Welch, R. A. (1990) Nucleotide sequencing of the Proteus mirabilis calcium-independent hemolysin genes (hpmA and hpmB) reveals sequence similarity with the Serratia marcescens hemolysin genes (shlA and shlB). J. Bacteriol. 172, 1206–1216. 14. Palmer, K. L. and Munson, R. S., Jr. (1995) Cloning and characterization of the genes encoding the haemolysin of Haemophilus ducreyi. Mol. Microbiol. 18, 821–830. 15. Hirono, I., Tange, N., and Aoki, T. (1997) Iron-regulated haemolysin gene from Edwardsiella tarda. Mol. Microbiol. 24, 851–856. 16. Schönherr, R., Tsolis, R., Focareta, T., and Braun, V. (1993) Amino acid replacements in the Serratia marcescens haemolysin ShlA define sites involved in activation and secretion. Mol. Microbiol. 9, 1229–1237. 17. Willems, R. J., Geuijen, C., van der Heide, H. G. J., Renauld, G., Bertin, P., van den Akker, W. M. R., Locht, C., and Mooi, F. R. (1994) Mutational analysis of the Bordetella pertussis fim/fha gene cluster: Identification of a gene with sequence similarities to hemolysin accessory genes involved in export of FHA. Mol. Microbiol. 11, 337–347. 18. Cope, L., Yongev, R., Müller-Eberhard, U., and Hansen, E. C. (1995) A gene cluster involved in the utilization of both, free heme and heme: hemopexin by Haemophilus influenzae type b. J. Bacteriol. 177, 2644–2653. 19. Barenkamp, S. J. and Geme, J. W. S. (1994) Genes encoding high-molecular-weight adhesion proteins of nontypeable Haemophilus influenzae are part of gene clusters. Infect. Immun. 62, 3320–3328. 20. Schönherr, R., Hilger, M., Broer, S., Benz, R., and Braun, V. (1994) Interaction of Serra-
235
13 Structure, Activity, Activation, and Secretion of the Serratia marcescens
21. 22.
23. 24.
25.
26.
27. 28.
29.
30.
31. 32. 33. 34.
35.
36.
37. 38.
39.
tia marcescens hemolysin (ShlA) with artificial and erythrocyte membranes: Demonstration of the formation of aqueous multistate channels. Eur. J. Biochem. 223, 655– 663. Schiebel, E. and Braun, V. (1989) Integration of the Serratia marcescens haemolysin into human erythrocyte membranes. Mol. Microbiol. 3, 445–453. Sieben, S., Hertle, R., Gumpert, J., and Braun, V. (1998) The Serratia marcescens hemolysin is secreted but not activated by stable protoplast-type L-forms of Proteus mirabilis. Arch. Microbiol. 170, 436–442 Poole, K., Schiebel, E., and Braun, V. (1988) Molecular characterization of the hemolysin determinant of Serratia marcescens. J. Bacteriol. 170, 3177–3188. Schäffer J., König W., Braun V., and Goebel W. (1988) Comparison of four hemolysinproducing organisms (Escherichia coli, Serratia marcescens, Aeromonas hydrophila, and Listeria monocytogenes) for release of inflammatory mediators from various cells. J. Clin. Microbiol. 26, 544–551 Ondraczek, R., Hobbie, S., and Braun, V. (1992) In vitro activation of the Serratia marcescens hemolysin through modification and complementation. J. Bacteriol. 174, 5086–5094. Hertle, R., Brutsche, S., Groeger, W., Hobbie, S., Koch, W., Könninger, U., and Braun, V. (1997) Specific phosphatidylethanolamine dependence of Serratia marcescens cytotoxin activity. Mol. Microbiol. 26, 853–865. Hilger, M. and Braun, V. (1995) Superlytic hemolysin mutants of Serratia marcescens. J. Bacteriol. 177, 7202–7209. DeChavigny A., Heacock P. N., and Dowhan W. (1991) Sequence and inactivation of the pss gene of Escherichia coli. Phosphatidylethanolamine may not be essential for cell viability. J. Biol. Chem. 266, 5323–5332. Killmann, H., Benz, R., and Braun, V. (1993) Conversion of the FhuA transport protein into a diffusion channel through the outer membrane of Escherichia coli. EMBO J. 12, 3007–3016. Könninger, U., Hobbie, S., Benz, R., and Braun, V. (1999) The haemolysin-secreting ShlB protein of the outer membrane of Serratia marcescens: determination of surfaceexposed residues and formation of ion-permeable pores by ShlB mutants in artificial lipid bilayer membranes. Mol. Microbiol. 32, 1212–1225. Schirmer, T., and Cowan, S. W.(1993) Prediction of membrane-spanning b-strands and its application to maltophorin. Protein Sci. 2, 1361–1363. Yang, F. L., and Braun, V. (1999) Unpublished results. Lyerly, D. M. and Kreger, A. S. (1983) Importance of Serratia protease in the pathogenesis of experimental Serratia marcescens pneumonia. Infect. Immun. 40, 113–119 Maki, D. G., Hennekens, C. G., Philips, C. W., Shaw, W. V., and Bennet, J. V. (1973) Nosocomial urinary tract infection with Serratia marcescens. J. Infect. Dis. 128, 579– 587. König, W., Faltin, Y., Scheffer J., Schöffler, H., and Braun, V. (1987) Role of cell-bound hemolysin as a pathogenicity factor for Serratia infections. Infect. Immun. 55, 2554– 2561. Scheffer, J., König, W., Braun, V., and Goebel, W. (1988) Comparison of four hemolysin-producing organisms (Escherichia coli, Serratia marcescens, Aeromonas hydrophila, and Listeria monocytogenes) for release of inflammatory mediators from various cells. J. Clin. Microbiol. 26, 544–551. Marre, R., Hacker, J., and Braun, V. (1989) The cell-bound hemolysin of Serratia marcescens contributes to uropathogenicity. Microb. Pathog. 7, 153–156. Yamamoto, T., Ariyoshi, A., and Amako, K. (1985) Fimbriae-mediated adherence of Serratia marcescens strain US5 to human urinary bladder surface. Microbiol. Immunol. 29, 677–681. Leranoz, S., Orus, P., Berlanga, M., Dalet, F., and Vinas, M. (1997) New fimbrial adhe-
236
References
40. 41.
42. 43. 44. 45.
46. 47. 48.
sins of Serratia marcescens isolated from urinary tract infections: description and properties. J. Urol. 157, 694–698. Kono, K., Yamamoto, T., Kuroiwa, A., and Amoko, K.(1984) Purification and characterization of Serratia marcescens US5 pili. Infect. Immun. 46, 295–300. Mizunoe, Y., Matsumoto, T., Haraoka, M., Sakumoto, M., Kubo, S., and Kumazawa, J. (1995) Effect of pili of Serratia marcescens on superoxide production and phagocytosis of human polymorphonuclear leukocytes. J. Urol. 154, 1227–1230. Hertle, R. (1998) Unpublished results. Hertle, R., Weingardt-Kocher, S., and Schwarz, H. Serratia marcescens strain W1128 is invasive in epithelial cells in culture. In preparation. Bhakdi, S. and Martin, E. (1991) Superoxide generation by human neutrophils induced by low doses of Escherichia coli hemolysin. Infect. Immun. 59, 2955–2962. Walev, I., Martin, E., Jonas, D., Mohamadzadeh, M., Müller-Klieser, W., Kunz, L., and Bhakdi, S. (1993) Staphylococcal alpha-toxin kills human keratinocytes by permeabilizing the plasma membrane for monovalent ions. Infect. Immun. 61, 4972–4979. Walev, I., Reske, K., Palmer, M., Valeva, A., and Bhakdi, S. (1995) Potassium-inhibited processing of IL-1b in human monocytes. EMBO J. 14, 1607–1614. Hertle, R., Hilger, M., Weingardt-Kocher, S., and Walev, I. (1999) Cytotoxic action of Serratia marcescens hemolysin on human epithelial cells. Infect. Immun. 67, 817–825. Cover, T. L., Reddy, L. Y., and Blaser, M. J. (1993) Effects of ATPase inhibitors on the response of HeLa cells to Helicobacter pylori vacuolating toxin. Infect. Immun. 61, 1427–1431.
237
14 Staphylococcal Lipases: Molecular Characterization and Use as an Expression and Secretion System Friedrich Götz* and Ralf Rosenstein
14.1 Introduction
The first bacterial lipase gene which has been ever cloned and sequenced was the lipase gene from Staphylococcus hyicus [1]. Our interest in staphylococcal lipases was focused in various directions: we planned to use the staphylococcal lipase as an expression and secretion system for the food grade S. carnosus, to study its molecular and biochemical properties, and to study the role of these enzymes in infection. Among the various exo-enzymatic activities of staphylococci, the acylglycerol-hydrolyzing activity of lipases and of esterases belongs to the most frequently detected activities. Lipases or glycerol ester hydrolases (EC 3.1.1.3) are defined as enzymes that hydrolyze emulsions of lipids with longchain fatty acids and that show interfacial activation, i. e. a sharp increase in activity when the solubility limit of the substrate is reached. An uncertainty exists as to whether the staphylococcal lipolytic enzymes should be classified as lipases or as esterases [2]. Despite the lack of clarity, the designation “lipase” for the staphylococcal lipolytic enzymes is commonly accepted in the literature and will also be used here. The importance of staphylococcal lipases, like other microbial lipases, results from their significance in bacterial lipid metabolism and their involvement in pathogenic processes, and also because they are valuable tools in biotechnology [3]. Their potential as biocatalysts is based on enzymatic features, e. g. regio- and enantio-specificity, a broad substrate specificity, and the ability to catalyze not only the hydrolysis, but also the synthesis of fatty acid compounds. The increasing interest in lipases is reflected by the numerous reviews on this topic published during the past few years; some of the reviews cover a broad spectrum of bacterial lipases (see, for example, [3, 4]). We will focus here on the molecular characterization of staphylococcal lipases and their use as expression and secretion system.
* Mikrobielle Genetik, Universität Tübingen, Waldhäuser Str. 70/8, D-72076 Tübingen
238 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
14.2 Molecular organization of staphylococcal lipases
14.2 Molecular organization of staphylococcal lipases
Up to the present, the nucleotide sequences of nine lipase genes from six different staphylococcal species have been published. Three are derived from S. epidermidis (two from S. epidermidis 9 and one from S. epidermidis RP62A), two from S. aureus (from strains NCTC 8530 and PS54), and one each from S. haemolyticus L62, S. hyicus DSM 20459, S. warneri 863, and S. xylosus DSM 20266 [1, 5–11]. For convenience, the following abbreviations will be used for the lipases from these strains: S. aureus NCTC 8530, SAL-1; S. aureus PS54, SAL-2; S. epidermidis 9 (GehC), SEL-1; S. epidermidis 9 (GehD), SEL-2; S. epidermidis RP62A, SEL-3; S. haemolyticus, SHaL; S. hyicus, SHyL; S. warneri, SWL; and S. xylosus, SXL. Five of the sequences were determined in our laboratory. Since the first description of the primary structure of a bacterial lipase [1], numerous studies on the molecular and biochemical properties of SHyL have been undertaken, making SHyL the best-studied staphylococcal lipase. From the nucleotide sequence a protein of 641 amino acids with a predicted molecular mass of 71,382 Da was deduced. Analysis of the supernatant of the donor strain S. hyicus by SDS-PAGE revealed a lipase with an apparent size of 46 kDa. When the SHyL gene was cloned in S. carnosus TM300, a cloning host with low extracellular proteolytic activity, only an 86-kDa form was detected in the supernatant. The N-terminal amino acid sequences of the 86- and 46-kDa forms were determined and a comparison with the SHyL sequence revealed that both forms were derived from the same precursor protein. The sequence of the 86-kDa form of SHyL secreted by S. carnosus starts with Asn39, immediately after a predicted cleavage site (-Ala36-Glu37-Ala38-) for signal peptidase I. The 46-kDa form produced by S. hyicus starts with Val246 [12]. It soon became clear that SHyL is secreted into the medium as a pro-form (pro-SHyL), which is subsequently processed to the mature lipase (mature SHyL). This processing apparently does not occur in the heterologous host S. carnosus. Based on the N-terminal sequences of the lipase forms found in the supernatants of S. hyicus and S. carnosus, it was evident that SHyL is translated as a 641-amino acid precursor protein with a signal peptide of 38 amino acids and a pro-peptide of 207 amino acids, which is processed to the mature lipase of 396 amino acids (Fig. 14.1). All staphylococcal lipases are translated as a pre-pro-enzyme with a leader peptide of 35 to 38 amino acids, followed by a pro-sequence (207 to 321 amino acids) and the mature form, i. e. the active lipase that normally appears in the supernatant of the producing Staphylococcus strain (383 to 396 amino acids). A multiple alignment of the lipase sequences shows a remarkable sequence conservation in the region covering the signal peptides, with a motif containing the perfectly conserved residues Ser, Ile, Arg, and Lys, designated as the SIRKmotif [1, 13]. It is still unclear, whether the strong conservation of this motif in the signal peptide of the staphylococcal lipases reflects a biological function in the secretion process. 239
14 Staphylococcal Lipases: Molecular Characterization
Figure 14.1: Organization of the S. hyicus lipase, ShyL, as an example for the pre-prostructure of staphylococcal lipases. sp, signal peptide; pp, pro-peptide; M, mature enzyme. The residues forming the catalytic triad (S124, D314, H355) are indicated. The amino acids responsible for the phospholipase (region around E295, K298, and S356) activity are indicated below. Amino acids that are involved in calcium binding (D109/112) are indicated. pro-ShyL is processed by the extracellular protease ShpII between T-V246 ; thus the mature ShyL starts with V1. Near the processing site single restriction sites have been inserted facilitating fusion with heterologous genes.
The various pro-peptides have no striking similarities at the sequence level, but are distinguished by their overall hydrophilic character. This hydrophilicity of the pro-peptide is most probably the reason for the observed differences between the calculated molecular weights of the pro-lipases and the masses estimated from SDS-PAGE, with the latter being significantly higher [1, 5, 13]. The molecular masses of the mature lipases, however, show a much better agreement between theoretical and observed values. All known staphylococcal lipases reveal the highest sequence similarity to each other in their mature parts, with identities ranging from 50 to 81%. The serine, aspartic acid, and histidine residues that are presumably involved in the lipolytic catalytic triad are well conserved in all sequences. The active site of lipases consists of three amino acid residues, Ser-Asp (Glu)-His, which form the catalytic triad. These amino acids always appear in this order in the amino acid sequence of the lipases, but they are distantly located from each other. In the tertiary structure, however, they are arranged near each other and constitute the active site. The candidates for the amino acids involved in lipolysis, as revealed by sequence similarity, are Ser124, Asp314, and His355 of the SHyL sequence. Site-directed mutagenesis of any of these amino acids results either in a drastically reduced lipase activity or in a complete loss of lipolysis. Since secretion or substrate specificity of SHyL are not hampered by the mutations, these amino acids are directly involved in catalysis [14]. This supports the hypothesis that SHyL is a serine hydrolase with the catalytic triad comprised of Ser124-His355-Asp314. The corresponding amino acids are conserved in all known staphylococcal lipases. 240
14.3 Biochemical characterization of staphylococcal lipases In addition, a P-loop consensus sequence, -[AG]-x4-G-K-[ST]-, is found in all staphylococcal lipases except SEL-2 and SHaL. The P-loop motif commonly occurs in ATP- or GTP-binding proteins [15]. It is not known, whether this domain plays a functional role in the lipases.
14.2.1 Processing of the pro-form The proteolytic processing of pro-SHyL has been studied in more detail [16–18]. Since the 86-kDa pro-SHyL purified from S. carnosus is processed by culture supernatants of S. hyicus, it became apparent that the processing of the pro-form by an extracellular protease occurs after secretion. In a co-fermentation of S. hyicus with an S. carnosus recombinant strain that produces pro-SHyL, several intermediate degradation products of 72, 55, and 46 kDa were produced during the processing of the pro-form to the mature lipase [12]. The 86-kDa pro-form and all intermediary products had lipase activity. In the later growth phase, products even smaller than 46 kDa that showed no lipolytic activity were observed, indicating that a degradation beyond Val246 affects the catalytic function. The proform and the 72- and 55-kDa products were predominant in the earlier growth phases, whereas the smaller forms dominated in the stationary phase. This indicates a stepwise processing starting at the N-terminus of pro-SHyL. The proteolytic activities in the supernatant of S. hyicus were analyzed. Two proteases, ShpI and ShpII, were identified; ShpII proved to be involved in pro-SHyL processing [19, 20]. ShpII, a 34-kDa protein with the highest activities at pH 7.4 and 55 8C, is a neutral metalloprotease that is strongly inhibited by chelating compounds. In vitro, ShpII cleaves pro-SHyL between Thr245 and Val246. In accordance with the proposed role and the biochemical properties of this protease, the in vivo processing of the pro-lipase by S. hyicus could be inhibited in the presence of 300 µM EDTA [19]. A similar processing scheme seems to be involved in the maturation of the other staphylococcal lipases since in most cases 40- to 48-kDa lipase forms have been detected in the supernatants of the corresponding donor strains.
14.3 Biochemical characterization of staphylococcal lipases
The biochemical characterization of the staphylococcal lipases will be briefly mentioned. We had a very close cooperation with the late Prof. Dr. Bert Verheij (Rijksuniversiteit te Utrecht Centre for Biomembran and Lipid Enzymology, Niederlande), and this cooperation is still ongoing with his remaining coworkers. 241
14 Staphylococcal Lipases: Molecular Characterization
14.3.1 Ca2+-dependency The activity of most of the staphylococcal lipases increases in the presence of Ca2+, with EDTA correspondingly acting as an inhibitor. SAL-1, SEL-3, SHyL, and SWL require Ca2+ for full enzymatic activity. Correspondingly, chelating compounds, e. g. EDTA or EGTA, act as inhibitors of these lipases. For SAL-1, Ca2+ can be replaced by strontium or barium without loss of activity [21]. It was shown that Ca2+ is most probably necessary for stabilizing the three-dimensional structure of the lipase during catalysis [21] and that two aspartate residues are responsible for calcium binding in SHyL (Fig. 14.1). Site-directed mutagenesis of these residues results in a loss of calcium binding, rendering the corresponding mutant lipases still active at room temperature, but inactive at higher temperatures [22].
14.3.2 pH optimum SAL-1 and SEL-3 are active over a broad pH range, with an optimum around pH 6 [11, 21]. Accordingly, both lipases are stable under acidic conditions, whereas they are inactivated at pH values above 10. This preference of acidic conditions is quite unusual among bacterial lipases, which in most cases exert their highest activities at alkaline pH. For SHyL, a pH optimum of 8.5 has been reported [23].
14.3.3 Substrate preferences The differences in pH dependency of SAL-1, SEL-3, and SHyL are also reflected in the substrate preferences of these lipases. SAL-1 and SEL-3 exhibit a strong preference for glycerides with short-chain fatty acids. Both lipases have a significant bias towards substrate molecules with butyric acid esterified to glycerol, pnitrophenol, or umbelliferone. Corresponding ester compounds with an acyl chain length of one methyl group above or below this size, e. g. triacetylglycerol or tripentanoylglycerol, are poorly hydrolyzed by these enzymes [11, 21, 23]. Until now, SHyL is unique among bacterial lipases in having a very broad substrate spectrum ranging from lipids of various chain lengths to phospholipids and lysophospholipids [11].
242
14.3 Biochemical characterization of staphylococcal lipases
14.3.4 Molecular basis of the phospholipase activity of SHyL Several studies have been undertaken to identify elements in the primary structures of SHyL that are responsible for its exceptional enzymological properties [13, 24–26]. By construction of a hybrid lipase in which the C-terminal 146 amino acids of SHyL are replaced by 145 amino acids from the C-terminus of SAL-1, it was demonstrated that the structural elements providing the phospholipase activity must reside within the exchanged element [13]. Attempts were then made to more narrowly define the regions involved in phospholipase activity and chain-length selectivity of SHyL by van Kampen et al. [24–26]. Various chimeras between SHyL and SAL-1 were generated by in vivo recombination and were tested for activity on phospholipids and p-nitrophenyl esters of different chain lengths. Three elements in the C-terminal region of SHyL necessary for phospholipase activity were identified. Furthermore, a central element of about 70 amino acids was shown to be essential for the chain-length selectivity of this enzyme. In a more recent study, small stretches of amino acids were exchanged between SAL-1 and a synthesized part of SHyL comprising the previously identified elements to localize single amino acid residues involved in phospholipase activity. A serine immediately following the catalytically active histidine had already been shown to be involved in hydrolysis of phospholipids; van Kampen et al. could now identify a stretch of polar amino acids (position 293–300) which, when exchanged for the corresponding, more hydrophobic region of SAL-1, led to a drastic decrease of phospholipase activity. Two essential residues, E295 and K298, were identified within this stretch by introducing point mutations; K298 was shown to be the major determinant for phospholipase activity. Interestingly, SAL-1 was made 23-fold more active towards phospholipids by the introduction of the reverse mutations, thus supporting the determined role for these amino acid residues. The authors concluded from their results that the polar stretch between amino acids 293 and 300 lies within a substrate binding pocket and is involved in the interaction with the polar head group of phospholipids [26].
14.3.5 The catalytic mechanism Several efforts have been made to crystallize staphylococcal lipases, but the resulting crystals have been of poor quality [27]. Experiments with covalent inhibitors that inhibit SHyL only in the presence of micelles and results obtained with the substrate analogue p-nitrophenyl-N-alkylcarbamate further support the hypothesis that SHyL has a lid covering its active site [28, 29].
243
14 Staphylococcal Lipases: Molecular Characterization
14.4 Role of the pro-peptide region
According to the sequence comparisons, all staphylococcal lipases are predicted to be primarily synthesized as pre-pro-lipases. While the function of the leader peptide in secretion is obvious, the role of the pro-peptide remained unclear. One hypothesis proposed a function in masking the enzyme activity until the secretion process is completed in order to protect the producing cell from detrimental effects of the lipase activity. However, the pro-form of SHyL is almost as active as the mature protein and the pro-lipase can be synthesized intracellularly without hampering the vital functions of the producing strain (G. Thumm, personal communication). Another possible function is the involvement in secretion, probably as an intramolecular chaperon. In order to evaluate whether the two functions of the pro-peptide are represented by different sections of the pro-peptide, several derivatives of SHyL with deletions in the pro-peptide region were constructed and tested for lipase production in S. carnosus [18]. The results obtained with these lipase mutants indicate that the SHyL pro-peptide may have two functional domains with each one located in one half of the pro-region. The N-terminal part seems to be important for lipase activity and the C-terminal part for translocation and stability. A stabilizing effect of the SHyL pro-peptide has also been observed in an experiment where OmpA of E. coli is fused to the pre-pro-portion of SHyL; in contrast to the construct without the lipase secretion signals, no proteolytic degradation occurred after secretion by S. carnosus [30]. A number of experiments designed to address the question whether the pro-peptide could act also in trans indicated that the pro-region has to be covalently attached to the mature protein in order to exert its beneficial effects on translocation, stability, and activity (G. Thumm, personal communication).
14.5 The use of ShyL as expression and secretion system
The role of the pro-peptide region of SHyL was more thoroughly investigated. A plasmid carrying the cloned gene encoding SHyL was used for the construction of various secretion vectors [16]. The Escherichia coli b-lactamase gene (bla) lacking its own signal sequence was fused at various sites along the gene regions encoding the pro-peptide and mature forms of SHyL. The amount of b-lactamase secreted into the supernatant of S. carnosus clones harboring the recombinant plasmids was measured; only those constructs having at least 160 aa of the SHyL pre-pro-region fused to b-lactamase secreted an amount of fusion protein com244
14.5 The use of ShyL as expression and secretion system parable to that of native lipase. All hybrid proteins having a smaller portion of the lipase fused to b-lactamase remained in the membrane fraction, which indicates a defect in the translocation, and were prone to extensive proteolytic degradation. These results support a dual role for the SHyL pro-peptide: an involvement in protein translocation and a role in stabilization against proteolytic degradation. Because of these encouraging results, we fused the pre-pro-portion of SHyL to several other heterologous proteins such as pro-insulin [31] and malaria antigen [32]. In all cases, the heterologous proteins were successfully and in good quantities secreted by S. carnosus, thereby showing that the results obtained with E. coli b-lactamase are generally applicable. Based on the SHyL-gene, we also have constructed a series of plasmid vectors for gene expression and cloning in staphylococci. pPS11 is a promoter probe plasmid containing a promoterless SHyL-gene. Insertion of a promoter-bearing DNA fragment at the single BamHI site turns on SHyL-gene expression [33]. The lipase activity can be easily determined to estimate the strength of the inserted promoter. pPS11 served also as a basis for the construction of vectors which allow xylose-inducible gene expression in S. carnosus. In plasmid pCX15, the SHyL-gene is under transcriptional control of the repressor, XylR, with the XylR target sequence the xylA promoter/operator. Again, the single BamHI site in front of the SHyL-gene RBS also makes it possible to put other promoterless genes under transcriptional control of XylR [33]. The expression vector pCX15 was the basis for improvement and for further heterologous protein secretion. In Gram-positive bacteria, a number of surface proteins are covalently anchored to the cell wall by an ubiquitous mechanism involving a specific, C-terminal sorting signal [34]. This reaction is catalyzed by the sortase. Normally, non-enzymatic proteins such as IgG- or fibronectin-binding proteins are linked in this way to the cell wall. We asked the question whether an enzyme, such as ShyL, can be covalently anchored to the cell wall in an active state [35]. To achieve cell wall immobilization ShyL was fused with the C-terminal region of S. aureus fibronectin binding protein B (FnBPB). Indeed, expression of the hybrid protein in S. carnosus resulted in efficient cell wall anchoring of enzymatically active lipase. In early stationary cells, 95% of total lipase activity were covalently anchored to the cell wall, 5 % were found to be set free into the culture supernatant. The cell wall-immobilized lipase (approximately 10 000 molecules per cell) retained more than 80 % of the specific activity compared to the unmodified S. hyicus lipase secreted by S. carnosus cells. After solubilization of the hybrid protein by lysostaphin treatment [36] of the cell wall, its specific activity was indistinguishable from that of the unmodified lipase. These results demonstrated for the first time that it is possible to immobilize soluble enzymes on the cell wall of S. carnosus without radically altering their catalytic activity [35]. Even if the pentaglycine anchor structure of the cell wall was modified anchoring to the cell wall was not affected [37]. In the meantime, several other groups [38] are using the ShyL expression and secretion system successfully for heterologous protein secretion in S. carnosus which has been worked out during the collaborative research centre 323. 245
14 Staphylococcal Lipases: Molecular Characterization
14.6 Concluding remarks
Although our knowledge of staphylococcal lipases has steadily increased during the past years, many aspects of these interesting enzymes remain to be investigated. Very little is known about the regulation of lipase synthesis in staphylococci. There are controversial reports whether lipase is subject of the global regulatory system agr (accessory gene regulator), which is known to regulate the expression of the genes of several exoproteins and cell-wall-associated proteins in staphylococci or of the alternative sigma factor SigB. Staphylococcal lipases and especially the ShyL expression and secretion system have a good potential being applied in biotechnology. However, we have to spend more work on overproducing the enzymes.
References
1. Götz, F., Popp, F., Korn, E., and Schleifer, K. H. (1985) Complete nucleotide sequence of the lipase gene from Staphylococcus hyicus cloned in Staphylococcus carnosus. Nucleic Acids Res. 13, 5895–906. 2. Brune, K. A. and Götz, F. (1992) in: Microbial degradation of natural products, pp. 243–266 (Winkelmann, G., Ed.) VCH, Weinheim. 3. Jaeger, K.-E., Dijkstra, B. W., and Reetz, M. T. (1999) Bacterial Biocatalysts: Molecular Biology, Three-Dimensional Structures, and Biotechnological Applications of Lipases. Annu. Rev. Microbiol. 53, 315–51. 4. Arpigny, J. L. and Jaeger, K. E. (1999) Bacterial lipolytic enzymes: classification and properties [In Process Citation]. Biochem. J. 343, Pt. 1, 177–83. 5. Farrell, A. M., Foster, T. J., and Holland, K. T. (1993) Molecular analysis and expression of the lipase of Staphylococcus epidermidis. J. Gen. Microbiol. 139, 267–77. 6. Farrell, A .M., Longshaw, C. M., and Holland, K. T. (1998) GenBank Entry, Accession No. AF090142. 7. Lee, C. Y. and Iandolo, J. J. (1986) Lysogenic conversion of staphylococcal lipase is caused by insertion of the bacteriophage L54 a genome into the lipase structural gene. J. Bacteriol. 166, 385–91. 8. Oh, B., Kim, H., Lee, J., Kang, S., and Oh, T. (1999) Staphylococcus haemolyticus lipase: biochemical properties, substrate specificity and gene cloning [In Process Citation]. FEMS Microbiol. Lett. 179, 385–92. 9. Rosenstein, R. and Götz, F. (1999) GenBank Entry, Accession No. AF208229. 10. Rosenstein, R. and Götz, F. (1999) GenBank Entry, Accession No. AF208033. 11. Simons, J. W., van Kampen, M. D., Riel, S., Götz, F., Egmond, M. R., and Verheij, H. M. (1998) Cloning, purification and characterisation of the lipase from Staphylococcus epidermidis – comparison of the substrate selectivity with those of other microbial lipases. Eur. J. Biochem. 253, 675–83. 12. Wenzig, E., Lottspeich, F., Verheij, B., De Haas, G. H., and Götz, F. (1990) Extracellu-
246
References
13.
14.
15. 16. 17. 18.
19.
20. 21.
22.
23.
24.
25.
26. 27.
28.
29.
30.
lare processing of the Staphylococcus hyicus lipase. Biochem. (Life Sci. Adv.) 9, 47– 56. Nikoleit, K., Rosenstein, R., Verheij, H. M., and Götz, F. (1995) Comparative biochemical and molecular analysis of the Staphylococcus hyicus, Staphylococcus aureus and a hybrid lipase. Indication for a C-terminal phospholipase domain. Eur. J. Biochem. 228, 732–8. Jäger, S., Demleitner, G., and Götz, F. (1992) Lipase of Staphylococcus hyicus: analysis of the catalytic triad by site-directed mutagenesis. FEMS Microbiol. Lett. 79, 249– 54. Saraste, M., Sibbald, P. R., and Wittinghofer, A. (1990) The P-loop: a common motif in ATP- and GTP-binding proteins. Trends Biochem. Sci. 15, 430–4. Liebl, W. and Götz, F. (1986) Studies on lipase directed export of Escherichia coli beta-lactamase in Staphylococcus carnosus. Mol. Gen. Genet. 204, 166–73. Götz, F., Verheij, H. M., and Rosenstein, R. (1998) Staphylococcal lipases: molecular characterisation, secretion, and processing. Chem. Phys. Lipids 93, 15–25. Demleitner, G. and Götz, F. (1994) Evidence for importance of the Staphylococcus hyicus lipase pro-peptide in lipase secretion, stability and activity. FEMS Microbiol. Lett. 121, 189–97. Ayora, S., Lindgren, P. E., and Götz, F. (1994) Biochemical properties of a novel metalloprotease from Staphylococcus hyicus subsp. hyicus involved in extracellular lipase processing. J. Bacteriol. 176, 3218–23. Ayora, S. and Götz, F. (1994) Genetic and biochemical properties of an extracellular neutral metalloprotease from Staphylococcus hyicus. Mol. Gen. Genet. 242, 421–30. Simons, J. W., Adams, H., Cox, R. C., Dekker, N., Götz, F., Slotboom, A. J., and Verheij, H. M. (1996) The lipase from Staphylococcus aureus. Expression in Escherichia coli, large-scale purification and comparison of substrate specificity to Staphylococcus hyicus lipase. Eur. J. Biochem. 242, 760–9. Simons, J. W., Cox, R. C., Egmond, M. R., and Verheij, H. M. (1999) Rational design of alpha-keto triglyceride analogues as inhibitors for Staphylococcus hyicus lipase. Biochemistry 38, 6346–51. van Oort, M. G., Deveer, A. M., Dijkman, R., Tjeenk, M. L., Verheij, H. M., de Haas, G. H., Wenzig, E., and Götz, F. (1989) Purification and substrate specificity of Staphylococcus hyicus lipase. Biochemistry 28, 9278–85. van Kampen, M. D., Simons, J. W., Dekker, N., Egmond, M. R., and Verheij, H. M. (1998) The phospholipase activity of Staphylococcus hyicus lipase strongly depends on a single Ser to Val mutation. Chem. Phys. Lipids 93, 39–45. van Kampen, M. D., Dekker, N., Egmond, M. R., and Verheij, H. M. (1998) Substrate specificity of Staphylococcus hyicus lipase and Staphylococcus aureus lipase as studied by in vivo chimeragenesis. Biochemistry 37, 3459–66. van Kampen, M. D., Verheij, H. M., and Egmond, M. R. (1999) Modifying the substrate specificity of staphylococcal lipases. Biochemistry 38, 9524–32. Ransac, S., Blaauw, M., Dijkstra, B. W., Slotboom, A. T., Boots, J. W., and Verheij, H. M. (1995) Crystallization and preliminary X-ray analysis of a lipase from Staphylococcus hyicus. J. Struct. Biol. 114, 153–5. Simons, J. W., Boots, J. W., Kats, M. P., Slotboom, A. J., Egmond, M. R., and Verheij, H. M. (1997) Dissecting the catalytic mechanism of staphylococcal lipases using carbamate substrates: chain length selectivity, interfacial activation, and cofactor dependence. Biochemistry 36, 14539–50. Leuveling Tjeenk, M., Bulsink, Y. B., Slotboom, A. J., Verheij, H. M., de Haas, G. H., Demleitner, G., and Götz, F. (1994) Inactivation of Staphylococcus hyicus lipase by hexadecylsulfonyl fluoride: evidence for an active site serine. Protein Eng. 7, 579–83. Meens, J., Herbort, M., Klein, M., and Freudl, R. (1997) Use of the pre-pro part of Staphylococcus hyicus lipase as a carrier for secretion of Escherichia coli outer mem-
247
14 Staphylococcal Lipases: Molecular Characterization
31. 32. 33.
34.
35. 36.
37.
38.
brane protein A (OmpA) prevents proteolytic degradation of OmpA by cell-associated protease(s) in two different gram-positive bacteria. Appl. Environ. Microbiol. 63, 2814–20. Knorr, R. (1992) in: Mikrobielle Genetik, Eberhard-Karls-Universität, Tübingen. Samuelson, P. et al. (1995) Cell surface display of recombinant proteins on Staphylococcus carnosus. J. Bacteriol. 177, 1470–1476. Wieland, K. P., Wieland, B., and Götz, F. (1995) A promoter-screening plasmid and xylose-inducible, glucose-repressible expression vectors for Staphylococcus carnosus. Gene 158, 91–6. Mazmanian, S. K., Liu, G., Ton-That, H., and Schneewind, O. (1999) Staphylococcus aureus sortase, an enzyme that anchors surface proteins to the cell wall. Science 285, 760–3. Strauss, A. and Götz, F. (1996) In vivo immobilization of enzymatically active polypeptides on the cell surface of Staphylococcus carnosus. Mol. Microbiol. 21, 491–500. Thumm, G. and Götz, F. (1997) Studies on prolysostaphin processing and characterization of the lysostaphin immunity factor (Lif) of Staphylococcus simulans biovar staphylolyticus. Mol. Microbiol. 23, 1251–65. Strauss, A., Thumm, G., and Götz, F. (1998) Influence of Lif, the lysostaphin immunity factor, on acceptors of surface proteins and cell wall sorting efficiency in Staphylococcus carnosus. J. Bacteriol. 180, 4960–2. Samuelson, P., Cano, F., Robert, A., and Stahl, S. (1999) Engineering of a Staphylococcus carnosus surface display system by substitution or deletion of a Staphylococcus hyicus lipase propeptide. FEMS Microbiol. Lett. 179, 131–139.
248
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli Moritz von Rechenberg, Waldemar Vollmer, and Joachim-Volker Höltje*
15.1 The murein sacculus, a “growing” molecule
Because of the high intracellular turgor pressure impinging on the cytoplasmic membrane most bacterial cell walls are mechanically reinforced by a unique cross-linked biopolymer called murein (peptidoglycan) [1–3]. Murein consists of poly-(GlcNAc-b-1,4-MurNAc)glycan strands, which are substituted at the lactyl group of the muramic acid by short peptides. The presence of a diamino-amino acid (diaminopimelic acid in the case of E. coli) makes it possible that two peptide moieties can be cross-linked. Cross-linkage is achieved by transpeptidation of the terminal carboxyl group of one peptide side chain to the free e-amino group of the diamino-amino acid present in the neighboring peptide side chain (see Fig. 15.1). Thereby the typical murein netting is formed. Importantly, the murein net is tailored into a bagshaped structure, called sacculus, which completely encloses the cell and thus establishes a kind of exoskeleton. It follows that growth of the bacterial cell depends on the enlargement and division of the murein sacculus [1, 2, 4, 5]. These processes must be considered to be one of the greatest technical challenges for the cell because of two major problems that have to be solved. First, a structure under extreme tension has to be enlarged and divided without allowing the sacculus to burst. This process is further aggravated since the sacculus of E. coli is believed to be extremely thin, in major parts being a monolayer only [6]. Second, the shape of the sacculus that determines the species-specific morphology of the bacterium needs to be maintained and transmitted precisely from generation to generation. Recently, as will be discussed below, experimental evidence has been obtained suggesting that this is accomplished by a multienzyme complex, a murein synthesizing machinery that follows a particular growth strategy [4, 7].
* Max-Planck-Institut für Entwicklungsbiologie, Abt. Biochemie, Spemannstraße 35, D-72076 Tübingen
249 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
250 Ala
Ala
A 2 pm-Glu-Ala
A2
pm
la -A lu -G
lu-Ala
Ala A2 pm-Glu-Ala
Ala
Ala Glu A 2pm
Ala
Ala
Ala Glu A 2 pm
murein triplet
stress-bearing layer
murein turnover (release of docking strand)
NH2
Ala-A 2 pm-Glu-Ala
Ala HN 2-A 2 pm Glu Ala
Ala-Glu-A 2 pm Ala
docking strand
A 2 pm-Glu-Ala
A 2 pm-G
Glu Ala
HN2 - A 2 pm
NH 2
Ala-Glu-A 2 pm-Ala
Ala-Glu-A 2 pm
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli
Ala
Ala
Ala-Glu-A 2 pm Ala A 2 pm-Glu-Ala
Ala
Ala-Glu-A 2 pm Ala
Ala
Ala Glu A 2pm
Glu Ala
HN 2- A 2 pm
Ala Ala-Glu-A 2 pm Ala-A 2 pm-Glu-Ala
Figure 15.1: Schematic representation of the three-for-one growth mechanism of monolayered murein. The rods indicate the glycan strands; A2pm stands for diaminopimelic acid. The dotted arrows point to the peptide bonds that have to be cleaved by endopeptidases to allow for the insertion of the murein triplet into the layer under stress.
Ala-Glu-A 2 pm Ala A 2 pm-Glu-Ala
Ala
Ala A 2 pm-Glu-Ala
Ala
Ala Glu A2 pm
Ala-Glu-A 2 pm Ala A 2 pm-Glu-Ala
Glu Ala
HN 2- A2 pm
Ala
Ala-Glu-A 2 pm-Ala
Ala-A 2 pm-Glu-Ala
15.1 The murein sacculus, a “growing” molecule
251
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli
15.2 Murein growth is accompanied by massive turnover
Growth of the murein sacculus of E. coli has been studied by pulse-chase labeling in a diaminopimelic acid auxotrophic strain using 3H-diaminopimelic acid [8–10]. It became clear that about 40–50 % of the incorporated label is released during one generation. However, only 4–8% of the turnover products are lost to the medium [8, 9], the bulk of the muropeptides accumulating in the periplasmic space is taken up across the cytoplasmic membrane into the cytoplasm and reused for the synthesis of new murein precursor molecules [9, 11, 12]. The recycling of the turnover material is very efficient and quite a number of proteins are involved (Fig. 15.2). Specific uptake systems have been identified for the up-
Figure 15.2:
252
Murein turnover and recycling in E. coli.
15.3 Enlargement and division of a stress bearing structure take of non-degraded muropeptides (AmpG) [13] as well as for free peptides (OppA and MppA) [11, 12]. The released turnover products are degraded by amidases, glucosaminidases and L,D-carboxypeptidases, both in the periplasm (AmiA) [14] and cytoplasm (AmpD, NagZ, LdcA) [15–17]. The free peptides are then hooked by a specific ligase (Mpl) to UDP-MurNAc [18]. Recently, we have shown that it is essential that only tripeptides are added to UDP-MurNAc and no tetrapeptides [17]. In the latter case the resulting UDP-MurNAc-tetrapeptide intermediate cannot be further processed to the pentapeptide precursor since the final step is the addition of a D-alanyl-D-alanine dipeptide to the tripeptide intermediate. Increased incorporation of tetrapeptide precursors into the murein sacculus results in a decrease of the cross-linkage and causes autolysis of the cells. A deletion mutant in the L,D-carboxypeptidase LdcA was found to start lysing when entering the stationary phase of growth. Therefore, the activity of LdcA is essential for growth of E. coli [17].
15.3 Enlargement and division of a stress bearing structure
The fate of the labeled murein before its release was analyzed by completely hydrolyzing the murein with lysozyme which was followed by separation of the degradation products (muropeptides) by reversed phase high-pressure liquid chromatography [3, 10]. It was found that the structure, i. e. the muropeptide composition, of the murein changes dramatically during growth. In particular one structure attracted our interest: cross-linked muropeptides of the type tris(GlcNAc-b-1,4-MurNAc-peptide), called trimers, which represent the connecting piece of three glycan strands. For structural reasons the three glycan strands at this site have to be placed in two levels, that is one strand either above or below of the plane defined by the two other strands (Harald Labischinski, personal communication). Therefore, one wonders what the meaning of such a structure could be in a monolayered murein net. It is a minor component (about 5%), but interestingly it is completely cleaved in one generation; thus it represents an intermediate structure [10]. In order to explain murein turnover, the function of the trimers and to explain how the netting of the monolayered stress bearing murein sacculus might be enlarged by a mechanism that avoids any risk of lysis, we proposed a strategy quite analogous to the established growth mechanism of Gram-positive rods [7]. The basic concept of the strategy has been characterized by the motto “make before break” [19]. Accordingly, new material is firstly hooked underneath the stress-bearing layer(s) before it is inserted as a result of the cleavage of some critical bonds in the stressed murein to which it has been attached. The outcome of such a mechanism is that the murein sacculus grows from the inside to the outside. Whereas complete new layers are added to the stress bearing layers of the 253
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli thick murein sacculus of Gram-positive bacteria, we propose that only small patches of new material are added to the thin, monolayered murein sacculus of Gram-negative bacteria such as Escherichia coli. The newly added patch could be a murein triplet, that is three murein strands cross-linked to one another (Fig. 15.1). The triplet is hooked to the peptidyl moieties belonging to the neighboring murein strands to the right and left of a single strand, called the docking strand, in the growing murein sacculus. Specific hydrolysis of the docking strand results in the triplet being pulled into the stress bearing layer of the sacculus. Hence, three new strands replace one old strand. Such a three-for-one growth mechanism is in accordance with the degree of murein turnover. The model also explains the function of the trimeric muropeptide structure: it represents the attachment site of the incoming new murein triplets. In addition, it becomes clear why the trimers are all cleaved during growth: this is needed to release the docking strand from the murein netting (see Fig. 15.1). In analogy to the growth mechanism of the multi-layered murein of Gram-positive bacteria, the new triplets are added from the site of the cytoplasmic membrane to which the murein polymerases are anchored, whereas the old docking strand is degraded from the opposite side by enzymes bound to the outer membrane. It has recently been detected that most of the major murein hydrolases of E. coli, the membrane-bound lytic transglycosylases MltA and MltB, are lipoproteins residing in the outer membrane [20, 21]. To find out how the action of the murein polymerases and hydrolases is coordinated is of utmost importance for our understanding of the controlled growth of the bacterial cell wall. As discussed below, it seems that both classes of enzymes are combined in a multienzyme complex.
15.4 Interaction of murein hydrolases and synthases as indicated by affinity chromatography
In order to obtain some experimental data whether the murein hydrolyzing enzymes directly interact with the murein polymerases, specific affinity chromatography was employed to detect protein-protein interactions. This was done by using various enzymes expected to be involved in murein growth, as specific ligands for affinity chromatography. The enzymes were coupled to activated Sepharose [22–27]. A summary of the obtained results is given in Table 15.1. It was found that lytic transglycosylases, the soluble Slt70 and the membrane-bound MltA and MltB and probably also D,D-endopeptidases show specific affinity to some murein polymerases. This included penicillin-sensitive enzymes that belong to the family of penicillin-binding proteins (PBPs). Each of Slt70, MltA and MltB coupled to Sepharose specifically retained the bifunctional transpeptidase/ transglycosylases, PBP1B and PBP1C, as well as the monofunctional transpepti254
15.5 Dimerization of the bifunctional transpeptidase/transglycosylase PBP1B Table 15.1:
Summary of the results obtained by affinity chromatography.
Retained protein (specificity a)
PBP1A (TG/TP) PBP1B (TG/TP) PBP1C (TG/TP?) PBP2 (TP) PBP3 (TP) PBP4 (EP) PBP5 (CP) PBP6 (CP) PBP6B (CP) PBP7/8 (EP) Slt70 (LT) MltA (LT) MltB (LT) a b
Slt70
Immobilized proteins MltA MltB PBP1C
PBP7
– + + – + – – – – + / – –
+b + + + + – – – – – – / –
– + + + + + – – – – – + –
– + + – + – – – – – – – /
– + / +b + + – – – – – + –
TG/TP, bifunctional transglycosylase/transpeptidase; TP, transpeptidase; EP, endopeptidase; CP, carboxypeptidase; LT, lytic transglycosylase; in the presence of periplasmic proteins.
dase PBP3. MltA showed additional affinity to the bifunctional PBP1A and the monofunctional transpeptidase PBP2. At least one structural protein seems to interact with this group of enzymes [25, 26]. A protein called MipA (MltA interacting protein) could be detected, that not only specifically interacted with MltA but in addition with the murein polymerase PBP1B. No enzymatic activity could be demonstrated for MipA. The specific protein-protein interactions of murein polymerases and hydrolases found in vitro may indicate that a multienzyme complex is formed in vivo. This possibility was therefore studied in more detail employing additional methods.
15.5 Dimerization of the bifunctional transpeptidase/ transglycosylase PBP1B
A major disadvantage of the affinity chromatography method is that binding constants cannot be determined. Therefore, surface plasmon resonance (SPR) experiments were performed using a BIAcore 2000 equipment. In the case of the affinity chromatography experiments the proteins were coupled to activated Sepharose through primary amino groups present in the proteins. As a result, the immobilized ligands were oriented in a random fashion on the surface of the 255
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli matrix material. By contrast, a more controlled method was established for the BIAcore measurements. We took advantage of the covalent binding of the PBPs to penicillin. Ampicillin that carries a free amino group was first linked to the surface of a biosensor chip by a standard amino coupling method. The immobilized ampicillin could then be used to capture and covalently bind PBPs, even from unfractionated protein preparations [25, 27]. The method not only results in uniformly oriented ligands but in addition stabilizes the protein by the presence of a substrate analogue in the active site of the enzyme protein. An interesting finding was that PBP1B binds to immobilized PBP1B (Fig. 15.3 a), indicating dimerization, as has previously been shown by other methods [28].
15.6 Reconstitution of the core particle of a murein synthesizing machinery
When the interaction of PBP1B with MltA as observed by MltA-Sepharose affinity chromatography [26] was investigated by the SPR method, no specific signal could be obtained. It was concluded that the interaction depended on additional specific factors besides PBP1B and MltA. Indeed, addition of MipA resulted in a signal that indicated the formation of a complex between the murein polymerase PBP1B, MipA, and the murein hydrolase MltA. MipA probably functions as a scaffolding protein in the assembly of the protein complex. The kinetics of the formation of the trimeric complex indicates a binding constant of about 0.850 ± 0.057 µM [25, 27] (Figs. 15.3 b and c). This value is comparable with those for cell adhesion molecules.
15.7 Proposed structure of a hypothetical holoenzyme of murein synthesis
On the basis of the results obtained by affinity chromatography and SPR studies a hypothetical multienzyme complex can be compiled that could enlarge the murein sacculus by the safe three-for-one growth strategy [4, 7, 29]. It is assumed that murein polymerizing and murein hydrolyzing enzymes are combined in a kind of “yin yang complex”. As mentioned above, protein-protein interactions between the bifunctional transpeptidase/transglycosylase PBP1B, the transpeptidases PBP3 and PBP2 as well as the D,D-endopeptidases PBP4 and 256
15.7 Proposed structure of a hypothetical holoenzyme of murein synthesis
Figure 15.3: Kinetic measurements of the binding of (a) PBP1B and (b) of the simultaneous binding of MipA together with MltA to immobilized PBP1B. Panel a: At a flow rate of 10 ml/min 100 ml PBP1B (3 mg/ml) were injected to differently modified flow cells. Curve 1: PBP1B immobilized to ampicillin coated CM5 sensorchips, prepared as described previously [25]; curve 2: PBP1A immobilized to ampicillin coated CM5 sensorchips; curve 3: as a control the ampicillin coated sensorchip was digested with b-lactamase; curve 4: the sensorchip was blocked with ethanolamine. Panel b: Different amounts of an equimolar mixture of MipA and MltA (ranging from 30 nM to 2.46 mM) were injected at a flow rate of 10ml/min to a PBP1B loaded sensor chip (see panel a). Panel c: A Scatchard analysis of the experiment depicted in b is shown.
257
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli PBP7 and the lytic transglycosylases Slt70, MltA, and MltB have been demonstrated partly by affinity chromatography, partly by BIAcore measurements, and to some extent by both methods. In addition, a complex consisting of PBP1B and MltA has been reconstituted on the surface of a biosensor chip in the presence of MipA. Dimerization of PBP1B, PBP3, and PBP4 has been obtained by various methods in different laboratories [28, 30]. Therefore, the following speculative complex is supported by quite a number of experimental results (Fig. 15.4). It is assumed that the synthesizing part of the complex consists of a dimer of a bifunctional transpeptidase/transglycosylase (TP/TG) that synthesizes two murein strands and crosslinks these strands by transpeptidation from two sides with a single strand independently synthesized by a transglycosylase (TG). The product would be three murein strands cross-linked with one another, a murein triplet. Also part of the synthesizing sub-complex would be a dimer of a monofunctional transpeptidase such as PBP3 or PBP2 (TP). The transpeptidases would be responsible for hooking the murein triplet by transpeptidation under-
LT
docking strand
EP Ala-Glu-A 2 pm
Ala-A 2 pm-Glu-Ala NH2
Ala-Glu-A 2 pm-Ala
stress-bearing layer
A 2 pm-Glu-Ala
NH 2
Ala
TP
Ala HN 2- A 2 pm Glu Ala
Ala Ala HN2 - A 2 pm
Glu Ala
TP/TG
Ala Glu A 2pm
murein triplet
Ala Ala A2 pm-Glu-Ala
Ala Glu A 2 pm Ala Ala-Glu-A 2 pm Ala
TG Figure 15.4: Hypothetical multienzyme complex supposed to be involved in growth of the murein sacculus of E. coli. The symbols are the same as in Fig. 15.1. The circles represent the enzyme molecules: LT, lytic transglycosylase; EP, D,D-endopeptidase; TP, D,Dtranspeptidase; TP/TG, D,D-transpeptidase/transglycosylase; TG, transglycosylase. The triangle indicates the presence of structural proteins in the complex, such as the recently identified MipA [25].
258
15.8 Recent insights in the mechanism of growth of the murein sacculus neath a murein strand in the murein sacculus. The murein degrading part of the complex would be responsible for the specific hydrolysis of the docking strand. A most efficient degradation would be accomplished by a concerted action of lytic transglycosylases (LT) and endopeptidases (EP). Since PBP2 is known to be responsible for cell elongation and PBP3 for septum formation, it is assumed that two similar complexes exist that differ by the presence of PBP2 in the cell elongation machine and PBP3 in the cell division machine, respectively [4].
15.8 Recent insights in the mechanism of growth of the murein sacculus reveal novel targets for antibiotics
Being an essential bacterial structure and being unique to prokaryotes, the murein sacculus and murein metabolism in general remain among the most attractive targets for antibiotics. Despite of this advantage a surprisingly low number of new targets have been established in the last years that are involved in murein metabolism. Even more surprising is that no simple method has been set up that would allow a high throughput screen for inhibitors of the transglycosylation reaction. This step is responsible for the polymerization of the precursors to murein strands, a process that is as important for the formation of the murein sacculus as the transpeptidation reaction, which is the well known target of the still most important b-lactam antibiotics. Studying different possibilities to immobilize murein-metabolizing enzymes, we found that PBPs 1A, B, and C bind to moenomycin coupled to affigel [27]. This finding prompted us to establish an assay that is based on a competition of test compounds for the binding of moenomycin to the transglycosylation site of the essential PBPs 1A and 1B [31]. The assay, which can be adapted to high throughput screens, might help to identify novel structures that could be the basis for new murein synthesis inhibitors of similar therapeutic importance as penicillins. It is obvious that the assembly pathway of the multienzyme complex, a multistep process with a number of sites that can be inhibited, could be a target for powerful antibacterial agents. The observed dimerization of PBP1B is likely to be a crucial step in the formation of the multienzyme complex. We therefore considered this reaction to be well suited for a screening of inhibitors by employing the established BIAcore assay [27, 32]. In collaboration with G. Jung and K.-H. Wiesmüller (Institute for Chemistry, University of Tübingen) first experiments were done with chemically synthesized peptides from a library that consisted of random peptides from five up to 15 amino acids. Binding of PBP1B to immobilized PBP1B was followed in the presence of various peptide size classes at a concentration of 400 mg/ml. A significant decrease of the control signal was found with peptides larger than eight amino acids. The nonapeptides 259
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli caused an inhibition by more than 60 %. To get an idea of the specific structure (sequence) that interferes with the dimerization, a library of nonapeptides was tested. It turned out that a specific position of leucine but not lysine or glutamate in the nonapeptide was most effective in interfering with the dimerization of PBP1B. As shown in Table 15.2, the nonapeptide X4LeuX4 was also inhibiting the formation of the trimeric complex consisting of PBP1B, MltA, and MipA. Unfortunately, when tested in ether-permeabilized cells, the peptides did not affect the rate of murein synthesis. This might be explained by the fact that the ethertreated cells do not grow and therefore synthesize murein only with the help of the existing, already assembled complexes. It is possible that the peptides do block the formation of the complex but cannot dissociate preexisting complexes. Additional in vivo experiments are needed to characterize the antibacterial potency of these peptides. The relatively simple BIAcore-based assay system represents a promising approach to screen for inhibitors of the formation of the murein synthesizing multienzyme complex and thus for general inhibitors of murein synthesis and bacterial growth. Table 15.2: Complex formation in the presence of peptides. Peptide (400 mg/ml)
none X4LX4 X4KX4
Formation of protein complex (%) PBP1B/MipA/MltA
MipA/MltA
100 18 108
100 100 95
References
1. Park, J. T. (1996) The murein sacculus. In: Escherichia coli and Salmonella (Neidhardt, F. C., ed.), ASM Press, Washington, pp. 48–57. 2. Weidel, W. and Pelzer, H. (1964) Bagshaped macromolecules – a new outlook on bacterial cell walls. Advan. Enzymol. 26, 193–232. 3. Glauner, B., Höltje, J.-V., and Schwarz, U. (1988) The composition of the murein of Escherichia coli. J. Biol. Chem. 263, 10088–10095. 4. Höltje, J.-V. (1998) Growth of the stress-bearing and shape-maintaining murein sacculus of Escherichia coli. Microbiol. Mol. Biol. Rev. 62, 181–203. 5. Nanninga, N. (1991) Cell division and peptidoglycan assembly in Escherichia coli. Mol. Microbiol. 5, 791–795. 6. Labischinski, H., Goodell, E. W., Goodell, A., and Hochberg, M. L. (1991) Direct proof of a “more-than single-layered” peptidoglycan architecture of Escherichia coli W7: a neutron small-angle scattering study. J. Bacteriol. 173, 751–756.
260
References 7. Höltje, J.-V. (1996) A hypothetical holoenzyme involved in the replication of the murein sacculus of Escherichia coli. Microbiol. 142, 1911–1918. 8. Goodell, E. W. and Schwarz, U. (1985) Release of cell wall peptides into culture medium by exponentially growing Escherichia coli. J. Bacteriol. 162, 391–397. 9. Goodell, E. W. (1985) Recycling of murein by Escherichia coli. J. Bacteriol. 163, 305– 310. 10. Glauner, B. and Höltje, J.-V. (1990) Growth pattern of the murein sacculus of Escherichia coli. J. Biol. Chem. 265, 18988–18996. 11. Goodell, E. W. and Higgins, C. F. (1987) Uptake of cell wall peptides by Salmonella typhimurium and Escherichia coli. J. Bacteriol. 169, 3861–3865. 12. Park, J. T. (1993) Turnover and recycling of the murein sacculus in oligopeptide permease-negative strains of Escherichia coli: indirect evidence for an alternative permease system and for a monolayered sacculus. J. Bacteriol. 175, 7–11. 13. Lindquist, S., Weston-Hafer, K., Schmidt, H., Pul, C., Korfmann, G., Erikson, J. Sanders, C., Martin, H. H., and Normark, S. (1993) AmpG, a signal transducer in chromosomal b-lactamase induction. Mol. Microbiol. 9, 703–715. 14. van Heijenoort, J., Parquet, C, Flouret, B., and van Heijenoort, Y. (1975) Envelopebound N-acetylmuramyl-L-alanine amidase of Escherichia coli K 12. Purification and properties of the enzyme. Eur. J. Biochem. 58, 611–619. 15. Höltje, J.-V., Kopp, U., Ursinus, A., and Wiedemann, B. (1994) The negative regulator of b-lactamase induction AmpD is a N-acetyl-anhydromuramyl-L-amidase. FEMS Microbiol. Lett. 122, 159–164. 16. Park, J. T. (1996) The convergence of murein recycling research with b-lactamase research. Microb. Drug Res. 2, 105–112. 17. Templin, M. F., Ursinus, A., and Höltje, J.-V. (1999) A defect in cell wall recycling triggers autolysis during the stationary growth phase of Escherichia coli. EMBO J. 18, 4108–4117. 18. Mengin-Lecreulx, D., van Heijenoort, J., and Park, J. T. (1996) Identification of the Mpl gene encoding UDP-N-acetylmuramate-L-alanyl-gamma-D-glutamyl-meso-diaminopimelate ligase in Escherichia coli and its role in recycling of cell wall peptidoglycan. J. Bacteriol. 178, 5347–5352. 19. Koch, A. L. and Doyle, R. J. (1985) Inside-to-outside growth and turnover of the wall of gram-positive rods. J. Theor. Biol. 117, 137–157. 20. Ehlert, K., Höltje, J.-V., and Templin, M. F. (1995) Cloning and expression of a murein hydrolase lipoprotein from Escherichia coli. Mol. Microbiol. 16, 761–768. 21. Lommatzsch, J., Templin, M, Kraft, A, Vollmer, W., and Höltje, J.-V. (1997) Outer membrane localization of murein hydrolases: MltA, a third lipoprotein lytic transglycosylase in Escherichia coli. J. Bacteriol. 179, 5465–5470. 22. Romeis, T. and Höltje, J.-V. (1994) Specific interaction of penicillin-binding proteins 3 and 7/8 with the soluble lytic transglycosylase in Escherichia coli. J. Biol. Chem. 269, 21603–21607. 23. von Rechenberg, M., Ursinus, A., and Höltje, J.-V. (1996) Affinity chromatography as a means to study multi-enzyme-complexes involved in murein synthesis. Microb. Drug Resis. 2, 155–157. 24. Schiffer, G. and Höltje, J.-V. (1999) Cloning and characterization of PBP1C, a third member of the multimodular class A penicillin-binding proteins of Escherichia coli. J. Biol. Chem. 274, 32031–32039. 25. Vollmer, W., von Rechenberg, M., and Höltje, J.-V. (1999) Demonstration of molecular interactions between the murein polymerase PBP1B, the lytic transglycosylase MltA, and the scaffolding protein MipA of Escherichia coli. J. Biol. Chem. 274, 6726–6734. 26. Vollmer, W. (1998) Thesis, University of Tübingen. 27. von Rechenberg, M. (1998) Thesis, University of Tübingen. 28. Zijderveld, C. A., Aarsman, M. E., den Blaauwen, T., and Nanninga, N. (1991) Peni-
261
15 A Multienzyme Complex Involved in Murein Synthesis of Escherichia coli
29. 30. 31. 32.
cillin-binding protein 1B of Escherichia coli exists in dimeric forms. J. Bacteriol. 173, 5740–5746. Höltje, J.-V. (1996) Molecular interplay of murein synthases and murein hydrolases in Escherichia coli. Microb. Drug Resis. 2, 99–103. Goffin, C. and Ghuysen, J.-M. (1998) Multimodular penicillin binding proteins: An enigmatic family of orthologs and paralogs. Microbiol. Mol. Biol. Rev. 62, 1079–1093. Vollmer, W. and Höltje, J.-V. (2000) A simple screen for murein transglycosylase inhibitors. Antimicrob. Agents Chemother. 44, 1181–1185. von Rechenberg, M., Wiesmüller, K-H., Jung, G., and Höltje, J.-V., in preparation.
262
16 The Changing Path of Hopanoid Research: From Condensing Lipids to New Membrane Enzymes Karl Poralla*
16.1 Introduction
A group of important, nearly ubiquitous components in eukaryotic membranes are the sterols, e. g. cholesterol 1, ergosterol, stigmasterol, or sitosterol. In addition to their role as precursors for hormones etc., they condense the lipid part of cell membranes to physiologically appropriate values of viscosity. This contrasts with the prokaryotes, where with very few exceptions sterols are absent. In prokaryotes the condensing function in the membrane is very often adopted by hopanoids [31, 32]. Hopanoids 2, 3, 5 show structural similarity to sterols. Both classes of compounds are polycyclic triterpenoids with one or more hydrophilic groups. Their rigid, polycyclic structure and hydrophobicity is a prerequisite for their condensing function. Glycerolipids, in contrast, have mobile acyl chains and they therefore comprise the fluid (less viscous) fraction of the membrane lipids.
* Mikrobiologie/Biotechnologie, Universität Tübingen, Auf der Morgenstelle 28, D-72076 Tübingen
263 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
16 The Changing Path of Hopanoid Research
At the beginning of our work hopanoids had been detected in a few bacterial species. Since then they have been found in about 50 % of the bacteria examined, bringing the total to about 100 bacterial strains and species [31]. Potential hopanoid synthesizing bacteria were brought to light in two genome sequencing projects, for Rhizobium NGR234 [33] and Streptomyces coelicolor [34]. In these bacteria the hopanoids are not synthesized under conventional laboratory conditions (unpublished results). This review of our work starting from about 1990 will focus mainly on the cyclization of squalene to hopene, describing studies from the purification of squalene-hopene cyclase to its X-ray structure [19, 42] and to directed mutagenesis at the active site leading to new enzymes. It will also describe how this work has contributed to the understanding of cyclic triterpene formation, a classical field in natural product chemistry. It will also touch on the question of the biosynthetic precursor of hopene and what is known about the elongation reactions in hopanoid biosynthesis, leading for example to tetrahydroxybacteriohopane 5.
264
16.2 The cyclization reaction
16.2 The cyclization reaction
In biosynthesis of sterols and hopanoids the key reaction is the cyclization of the linear polyene compounds squalene and (3S)-2,3-oxidosqualene to hopene and sterols (Fig. 16.1). An acidic group at the active sites of the corresponding cyclases will initiate the reactions by protonation. A carbocationic cyclization cascade will follow and finally the positive charge will be eliminated by proton abstraction, or hydroxyl ion from water will be added. In the case of sterol cyclization ring B of the protosteryl cation is in the strained boat conformation. After the formation of the protosteryl cation, a rearrangement of hydride and methyl groups takes place and finally a proton is eliminated. From mere chemophysical considerations it is evident that the cyclization has to occur in or at the membrane because the substrates are very hydrophobic and are therefore dissolved in the hydrophobic inner part of the membrane. It is also evident that the “unordered” squalene and oxidosqualene molecules must be very specifically folded (chaperoned) just prior to cyclization. From chemical considerations it is obvious that a carbocationic cyclization cascade has to occur in a shielded milieu where side reactions, especially with water molecules, are excluded. In the enzymatic assay it can easily be observed that the cyclization reaction runs to completion without any addition of an energy-rich compound, because the reaction is highly exergonic. The cyclization reaction ranks among the
Figure 16.1: Reactions catalyzed by squalene-hopene cyclase (SHC) and 3S-(2,3)-oxidosqualene-lanosterol cyclase (OLaC).
265
16 The Changing Path of Hopanoid Research most complex one-step reactions in chemistry and biochemistry. In hopene formation twelve covalent bonds are altered, five cycles formed and nine stereocenters established. Despite the high number of stereocenters the specific formation of one main product out of more than 128 possible stereoisomers is accomplished. The squalene-hopene cyclase (= SHC) is the best studied example of a triterpene cyclase and opens the door to understand the reactions of more than hundred triterpene cyclases which contribute to the high diversity of natural products in higher plants and ferns.
16.3 Purification of squalene cyclases
When we started our biochemical work on hopene cyclization no triterpene cyclase had been successfully purified to homogeneity and no gene of a triterpene cyclase had been cloned. In 1987 we started to purify several SHCs in order to clone and sequence the respective genes and to determine conserved amino acids for site directed mutagenesis. The ultimate aim was to alter product specificity of the SHC and thereby to understand the catalysis of hopane skeleton formation. The Alicyclobacillus acidocaldarius SHC was the first triterpene cyclase for which purification successfully has been achieved [1]. The decision to investigate the SHC from this bacterial species (growth optimum at 60 oC) turned out to be a lucky choice; the purified enzyme was easy to handle and stable at room temperature. Because the triterpene cyclases are membrane bound enzymes a critical step in all purification procedures was the solubilization of the enzyme by various detergents from the membrane fraction and to find the appropriate detergent for each cyclase. Triton-X-100, CHAPS, or octyl-thioglucoside proved to be very efficient detergents [1–3]. Worthwhile to mention is the purification of Rhodopseudomonas palustris SHC on a Blue Sepharose column in the last step [8]; Blue Sepharose is normally used for purification of enzymes with a NAD+/ NADP+ binding site. A summary of the different purification procedures is given in Table 16.1. We observed that each cyclase purification is an individual procedure reflecting the different membrane environments and specific requirements of the enzymes. In a later stage of the project we purified cloned A. acidocaldarius SHC after insertion of a His-tag at the N-terminus by Ni2+-agarose affinity chromatography [16]. Critical for the purification was the development of a rapid enzyme assay procedure. We started with a radio thin-layer chromatography method using tritiated squalene. The reaction products were scraped off the thin-layer plates and the radioactivity was counted [35]. Later, we switched to a thin-layer chromatography scanning procedure which proved to be very inefficient because of 266
16.4 Properties of purified cyclases Table 16.1: Solubilization and applied adsorbents for purification of bacterial SHCs and squalene-tetrahymanol cyclase from Tetrahymena thermophila. A. acidocaldarius
Rh. palustris
Z. mobilis
T. thermophila
Solubilization in: Triton X-100
CHAPS
Triton X-100
Octylthioglucoside
Octyl Sepharose Isoelectric foc. Sephacryl S-300 DEAE Cellulose Preparative PAGE
DEAE Trisacryl Hydroxyapatite Mono Q
Column chromatographic method: DEAE cellulose DEAE Cellulose Phenyl Sepharose Octyl Sepharose Sephacryl S-500 Blue Sepharose
the low radiation energy of the tritium label [6]. Finally, we developed a gaschromatographic assay with unlabelled squalene. This is now the established method for the SHC assay [1]. We purified the SHC to homogeneity from A. acidocaldarius and R. palustris [1, 8]. Numerous attempts to purify Zymomonas mobilis SHC failed [6]. The squalene-tetrahymanol cyclase from the protozoon Tetrahymena thermophila has been efficiently purified in a cooperation project with Guy Ourisson [3]. Tetrahymanol 4 is an isomer of diplopterol 3, and therefore, squalene-tetrahymanol cyclase is a member of the triterpene cyclases.
16.4 Properties of purified cyclases
The molecular weight of all the cyclases purified by our group was about 70 kDa [1, 3, 6, 8]. This was in accordance with the size of the genes. This value contrasted very much to the values for oxidosqualene-sterol cyclases (= OSCs) published by other groups which were in the range of 25 to 50 kDa. The correct value for OSCs is in fact in the range of 86 kDa [36, 37]. SHCs possess a certain degree of unspecificity in that they yield several products in vitro and also in vivo. Besides hopene 2 (= hop-21,29-ene) it produces 5–10 % diploterol 3 (= hopane-21-ol). This percentage is independent of the pH in the assay mixture. R. palustris SHC produces even 50 % of diplopterol and no tetrahymanol 4, although diplopterol and tetrahymanol have been isolated from cells [2], and therefore a production of tetrahymanol in vitro has been expected. The unspecificity of the SHC is supposed to be biologically meaningful, because diplopterol and presumably also tetrahymanol are able to condense membranes similar to elongated hopanoids. In contrast, the hydrocarbon 267
16 The Changing Path of Hopanoid Research hopene (unable to condense membranes) is probably used in biosynthetic reactions leading to the elongated hopanoids. With highly purified A. acidocaldarius SHC preparations containing no lipid contaminations it was possible to observe traces of additional side products [22]. In a GC/MS measurement about twelve minor products were detected. Their relative amounts are mostly 0.5% that of hopene. Since they have molecular masses of 410, they are supposed to be isomers of squalene and hopene. These side products turned out not to be artifacts, resulting from the removal of the SHC from the natural membrane environment or attachment of the His-tag. They were also detected in the hydrocarbon fraction of A. acidocaldarius. The structures of five minor hydrocarbons 7, 8, 9, 10, 11 were elucidated [22]. The 17-isodammaradienes 8 and 10 have yet not been found in nature and the rest are natural products, mainly from ferns. Four of these products are tetracyclic with a five-membered ring D. Therefore, probably during ring D formation, the six-membered ring intermediate competes with a five-membered intermediate (Markovnikov cyclization) and this intermediate leads to dead-end products. Furthermore, compounds 7, 8, and 9 show hydride shifts and eupha-7,24diene 11 shows two additional methyl shifts before termination of the reaction.
268
16.4 Properties of purified cyclases Such hydride and methyl shifts do also occur in the formation of sterols. The double bond introduced by the final deprotonation occurs at different C-atoms. This result points to a certain unspecificity of the deprotonation reaction and adventitious deprotonation sites may be postulated. The unspecificity of the deprotonation is conceivable if one takes into account the high acidity of C-atoms adjacent to the carbocation. The side products demonstrate the potential of SHC to produce other and even new cyclization products. Such catalytic promiscuity may pave the way for the evolution of new enzymes [38]. By intricate mutagenesis it should be possible to shift the SHC catalysis to new main products. Thereby, new enzymes will be obtained. Furthermore, the unspecifity of SHC draws attention to the possibility that the high diversity of triterpenes in plants is not necessarily correlated with a high diversity of cyclases. With the purified A. acidocaldarius SHC we measured a few kinetic properties. All kinetic constants described are apparent because all reactions occur in a micellar system; the enzyme, substrate, and inhibitors do not freely diffuse in a homogeneous phase. The apparent Km for squalene is 9 mM [1]. Slightly higher are the values for the cyclases of T. thermophila and Z. mobilis [3, 6]. The turnover number is in the range of 0.3 sec–1. This is a very low value, but it is conceivable by the X-ray structure of SHC and the complex reaction mechanism [16]. We found that two similar groups of common detergents, namely n-alkyltrimethylammonium halides and n-alkyldimethylamine-N-oxides, e. g. 6, are efficient mechanism-based inhibitors of SHC [1]. By the N-methyl groups and the positive charge these compounds have a distant relationship to the protonated squalene during catalysis. The Ki for the inhibitors with a C12-side chain turned out to be 0.32 and 0.14 mM, respectively. This values are 30 to 60-fold lower than the Km for squalene. The kinetic data show that these are competitive inhibitors. Later, lauryldimethylamine-N-oxide 6 (= LDAO) has been used as a substrate model in the X-ray structure elucidation of SHC [19]. The circular dichroism measurements demonstrated that the A. acidocaldarius SHC is predominantly composed of a-helices and loops and only a very low b-sheet proportion [16]. This measurements were confirmed by the X-ray structure [19]. The circular dichroism measurements further showed that the point mutants (concerning Asp374, Asp376, and Asp377) of SHC retain their secondary structure.
269
16 The Changing Path of Hopanoid Research
16.5 Cloning of squalene-hopene cyclases
We cloned the shc gene from A. acidocaldarius [4], A. acidoterrestris [14], Z. mobilis [12], Bradyrhizobium japonicum [15], R. palustris [14, accession no. EMBL Y09979], and Methylococcus capsulatus [21]. To clone several SHCs was important for identifying conserved amino acid residues and was the fundament for directed mutagenesis. Additional aspects were important for the choice of the organisms: there was a chance to clone in addition to shc genes also genes for squalene-tetrahymanol (in R. palustris) and oxidosqualene-lanosterol (in M. capsulatus) cyclase. Different approaches were chosen for cloning the shc genes. For A. acidocaldarius cyclase the path of reverse genetics had to be taken because no triterpene cyclase had been sequenced at that time [4]. From the purified cyclase, we sequenced the N-terminus and synthesized corresponding DNA-probes. Starting with these probes, transformed E. coli clones were isolated and were shown to form functional SHC. For cloning Z. mobilis shc, fragments of A. acidocaldarius shc were used as heterologous probes in hybridisation experiments [12]. Despite the taxonomical (phylogenetic) distance and different GC percentages of genomic DNA in both organisms, a shc containing clone was identified. Upstream of shc the first hopanoid biosynthesis gene cluster was detected [23]. The same PCR strategy was applied in the case of shc from A. acidoterrestris [14]. In the case of B. japonicum a fragment of shc was produced by PCR and sequenced. In a second step, a cosmid-library was checked by PCR with homologous primers designed according to the above fragment for a shc positive clone [15]. From M. capsulatus producing hopanoids and sterols, a fragment of shc was amplified. This PCR-product was used as a probe to isolate the shc gene [21]. The expressed gene was able to convert squalene to hopene and diplopterol. But no conversion was observed with oxidosqualene as substrate. The search by hybridisation techniques with shc fragments as probes for a sterol cyclase gene was not successful. Therefore, synthesis of sterol in this bacterium remains a mystery. One may speculate that the M. capsulatus SHC will be posttranslationally altered by an unknown mechanism to form lanosterol from oxidosqualene.
270
16.6 Properties of SHC sequences
16.6 Properties of SHC sequences
SHCs are 631–658 amino acids long. The degree of similarity of SHCs roughly parallels the phylogenetic similarity of the corresponding bacteria. Therefore, horizontal gene transfer is a rare event for the genes of hopanoid biosynthesis. SHCs are distantly related by 27% identity and 35% similarity to sterol cyclases, and both form two separate clusters. It will be of great interest to elucidate also the sequences of plant hopanoid cyclases. This will clarify the question whether they are directly derived from bacterial SHCs or from plant OSCs. Lupeol cyclase producing a pentacyclic triterpene originates from cycloartenol cyclases and is not a member of the SHC-cluster [25]. With the help of peptide sequences of rat liver OSC [39], a non-tandem repeat could be identified in SHCs and OSCs at about the same positions in each sequence (Fig. 16.2) [9]. The consensus sequence of the repeat is: R/K A/G X X F/Y/W L X X X Q X X X G X W This motif is called QW-repeat because of the regular occurrence of Q (Gln) and W (Trp) at the C-terminus. This repeat is highly specific for triterpene cyclases and therefore has an indicative value. In a later section it will be demonstrated that the QW-repeat is connected to the stabilization and not to the catalysis of SHC. Shortly after the discovery of the QW-repeat it was very suggestive to relate it to catalysis, since a substrate with a repeat unit (isoprene) would be polycyclized by an enzyme repeat element [7]. A second motif was identified in SHCs and OSCs. Starting from peptidedigest of rat liver OSC which reacted covalently with 29-methylidene-2,3-oxidosqualene 15 it was possible to identify a DCTA-motif in OSC [39] and the similar
Figure 16.2: Schematic representation of motifs in squalene-hopene cyclase (SHC) and oxidosqualene cyclases (OSC). The black boxes represent the QW-motifs. For historical reasons numbering is started from the C-terminus. QW5 b occurs in Gram-positive bacteria only and QW5 c is truncated at the N-terminus. The shaded boxes represent the Asp containing motifs.
271
16 The Changing Path of Hopanoid Research DDTA-motif in SHC. They are located at corresponding sequence positions (Fig. 16.2). Thus we have two criteria at hand to identify a triterpene cyclase, similarity and two common motifs. It is interesting to note that two truncated QW-motifs and the DDTA-motif occur in kaurene synthase A and abietadiene synthase [40, 41]. These diterpene cyclases catalyze a cyclization reaction related to triterpene cyclases. Substrate for these synthases is geranylgeranyl diphosphate. The bicyclization reaction is not started by a dephosphorylation to generate the first carbocation. The generation of the first carbocation occurs in the same manner as in triterpene cyclases. In contrary, mono- and sesquiterpene synthases (cyclases) start the cyclization reaction by cleavage of the diphosphate group for the generation of the initial carbocation. These synthases have no similarity or motifs in common with triterpene cyclases.
16.7 The structure of squalene-hopene cyclase
With the cloned shc from A. acidocaldarius, the mutant Asp376Cys, the purification procedure, and enzyme assay, Ulrich Wendt in the group of Georg Schulz in Freiburg achieved crystallization and X-ray structure elucidation at 2.9 and 2.0 Å resolution [18, 19, 24, 42]. The SHC is a membrane enzyme and had to be co-crystallized with detergent molecules. Therefore, crystallization was not an easy task. The enzyme is built mainly from a-helices, loops and turns; also few short b-sheet regions occur (Fig. 16.3 A). The a-helices form two distinctive domains. As shown in Fig. 16.3 A, domain 1 is formed by six inner and six outer a-helices (a6 – a6 barrel). Domain 2 constitutes a less regular barrel of eleven a-helices. The active site is located in a large central cavity sandwiched by both domains. A channel structure with a constriction at the end originates at a hydrophobic surface area (plateau) and leads to the cavity (Fig. 16.3 B). The functional SHC is a dimer formed by two identical monomers [19, 42]. How does this dimeric enzyme interact with the cytoplasmic membrane? Probably the dimeric SHC dips with its hydrophobic areas (1600 A2 each) around the channel entrance from the cytoplasmic side into the hydrophobic part of the membrane (Fig. 16.3 C). Integral membrane proteins that submerge from one side into the non-polar part of the membrane without protruding through it were defined as “monotopic”. By X-ray structure analysis two further monotopic membrane proteins were identified, namely, two prostaglandin-H2 synthase isoenzymes which have no common sequence features to SHC [43]. The basic amino acids surrounding mainly the non-polar plateau interact with phosphoand sulfolipids of the A. acidocaldarius membrane and thereby reinforce the hydrophobic membrane/cyclase interaction. 272
16.7 The structure of squalene-hopene cyclase
A
B
C Figure 16.3: (A) Structure of the monomer of Alicyclobacillus acidocaldarius squalenehopene cyclase [41]. Yellow, inner helices; red, outer helices; cyan, b-structure; green, QW-motifs; brown, the competitive inhibitor LDAO 6; blue, a group of amino acids at the channel constriction and Asp376 at the “top” of the catalytic cavity near the methyl groups of LDAO. (B) Cross section of the channel and the catalytic cavity of squalenehopene cyclase. LDAO, cyan; acid residues (red), basic residues (blue), neutral residues (grey), and hydrophobic residues (yellow). The white arrow indicates the channel. (C) The homodimeric structure of squalene-hopene cyclase and its hypothetical location in the membrane. The arrows point to the approximate position of the channel entrances; gray bars, the polar parts of the membrane.
273
16 The Changing Path of Hopanoid Research It may be hypothesized that squalene diffuses in a twisted, unordered conformation in the cytoplasmic membrane and is decoiled when it enters into the narrow constriction of the SHC channel. Then it may be threaded into the catalytic cavity where it will be properly folded. The catalytic cavity is lined predominately by eight aromatic amino acids which shape a steric matrix for the folding process and stabilize the intermediate carbocations by their p-electrons or their electron pairs at the OH-groups of Tyr. Very probably, Asp376 is the protonating residue [16, 44]. Asp377 and Asp374 will facilitate this process by stabilization of the first carbocation. The protonation starts the cyclization cascade which finally ends in a deprotonation. Responsible for deprotonation is a cluster of polar amino acids (Gln262:Glu45:Glu93:Arg127) together with some water molecules at the “base” of the catalytic cavity. One water molecule has a catalytic function for proton abstraction and will sometimes also spend its OH – to the last hopenyl cation, thereby forming diplopterol. It is not clear by which way the product finally leaves the cavity. Possibly, the product will leave at the end of the exergonic reaction the cavity by passing the channel constriction, which is surrounded by mobile loops. Details of the reaction mechanism will be clarified when a co-crystal of the SHC with squalene or hopene is available. Since the above mentioned QW-motifs are located at the surface of SHC, they are not directly involved in catalysis. Six of eight QW-motifs occur in domain 1. The C-terminal moiety of the QW-motif is part of a loop connecting an outer helix to the next inner helix. This part is forming a net of hydrogen bonds with the preceding outer helix and the following inner and outer helices. Thereby the outer helices of domain 1 are not only connected via the peptide chain but also via a net of at least seven hydrogen bonds of the QW-motif. Thus the barrel of helices is highly stabilized. Likewise two additional QW-motifs in the domain 2 fulfill the same role. This stabilization seems necessary because the hopene cyclization is highly exergonic, producing about 200 kJ/mol (calculated value; see [19]). This value corresponds to the free energy of about seven phosphate bonds.
16.8 Site directed mutagenesis of squalene-hopene cyclase
On the basis of the elucidated X-ray structure of SHC we mutated amino acids in the catalytic cavity (Fig. 16.4). This cavity can be divided into three subsites according to Wendt and Schulz: a protonating site, a site for folding (chaperoning) of squalene and, simultaneously, for stabilizing intermediate carbocations, and a site for deprotonation [19, 42]. A likely region for protonating is 374DVDDTA379, and Asp376 is the most probable residue for donating the proton because it is in the neighborhood to C3 of the modeled hopene molecule. Furthermore, the carboxyl group of 274
16.8 Site directed mutagenesis of squalene-hopene cyclase
Figure 16.4: Schematic view of the catalytic cavity of squalene-hopene cyclase with a modeled hopene molecule (in dark) [41]. The amino acids line the cavity and some mentioned in the text are shown in gray. The intermediate carbocations in the hopene molecule are shown as black dots. The b-site of hopene is upside down.
Asp376 is pointing from the b-site to C3. Indeed, it was demonstrated that hopene is protonated at the b-H of C3. The experiment was performed in pure D2O and nearly 100 % of the b-H at C3 were deuterated. This experiment proves that the proton from the deprotonation reaction is not used for protonation (W. Eisenreich, A. Bacher, and K. Poralla, unpublished). The process of protonation is supposed to be assisted by the hydrogen-bonded couple Asp376:His451 and by bound water between Asp376 and the OH-group of Tyr495. Furthermore, it can be suggested that the couple Asp374:Asp377 carries a negative charge that stabilizes the proton at Asp376. The results in Table 16.2 Table 16.2: Mutant SHCs and their properties. Protonation site: Asp374Gln Asp376Cys Asp376Glu Asp377Glu His451Ala
slightly reduced activity; wild type product pattern very low activity; wild type product pattern reduced activity; wild type product pattern no activity reduced activity; wild type product pattern
Folding site: Trp169Phe Trp312His Tyr420Ala Phe601Ala Leu607Lys Tyr609Phe Phe365Leu
reduced activity; altered product pattern very low activity; new main product wild type activity; altered product pattern wild type activity; altered product pattern very low activity; bicyclic main product wild type activity; mainly tetracyclic products reduced activity; wild type product pattern
275
16 The Changing Path of Hopanoid Research show that Asp376 and Asp377 play an important functional role. The exchange Asp377Glu abolishes activity and Asp376Glu lowers the activity to 10 %. Asp374 possibly takes part in the reaction, but the carboxy group is not essential since Asp374Asn only slightly reduces the activity (Table 16.2). The importance of His451 has to be doubted because the exchange His451Ala leads to no inactivation. From the effects with the mutations we draw the conclusions that an additive and cooperative effect leads to the protonating activity of Asp376. This explains why in most cases a residual activity remains after mutagenesis of the “partners” of Asp376. Further arguments for the protonating function are its conservation in OSCs and their inactivation results merely from mutation of this specific Asp residue [44]. Effects easier to interpret resulted from several mutations at the folding site of the catalytic cavity. These mutations showed distinct alterations of the product pattern. In detail we will discuss four mutations. 1) Tyr420Ala: This mutant produces besides hopene also bicyclic and tricyclic compounds [29]. a-Polypodatetraene 12 (23% of the hopene peak) and g-polypodatetraene 13 (10 %) are deprotonated at different C-atoms in ring B. Also traces of the tricyclic malabaricatriene 14 were detected. This result showed that the protruding OH-group of Tyr420 is stabilizing the incipient carbocations at C8 and also at C13 (hopene 2 numbering). When this stabilization is abolished a premature termination of the cyclization cascade will frequently occur. 2) Phe601Ala: This mutant produces besides hopene also the isodammarene 8 (40 % of the hopene peak) [28]. The p-electrons of Phe601 are directed towards the C17 carbocation. When they are lacking the cyclization will end to a significant percentage with the production of compound 8. This compound has yet not been found in nature; but it is known as a minor product of SHC [22], and the same stereochemistry at C17 has been observed in the transformation of 29methylidene-2,3-oxidosqualene 15 to 16 [20]. The same compound has also been found with mutation Trp169Phe [46]. The above results show clearly that directed mutagenesis will lead to new products. The approach for the formation of new compounds may be further elaborated by the use of altered substrates. 3) Leu607Lys: This mutation leads to a new, although low enzyme activity. Leu607 is protruding into the middle part of the cavity pointing to C8. In the case of mutation Leu607Lys a premature termination occurs by the introduced basic amino acid. Lys607 will probably accept a proton from C7, thereby forming g-polypodatetraene 13 by a type of Hofmann elimination reaction (S. Schmitz and K. Poralla, unpublished). It is very interesting, that in this mutant a single bicyclic compound is formed, as compared to mutant Tyr420Ala. 4) Tyr609Phe: This mutant produces significantly more tetracyclic compounds 8, 9, 10 than hopene ,and it is therefore justified to designate it as a new enzyme. Presumably, Tyr609 stabilizes the cation at C13 by the single electron pairs at the OH-group together with other aromatic residues. This is a critical point in 276
16.8 Site directed mutagenesis of squalene-hopene cyclase the cyclization cascade because this stabilization is important for the unfavorable anti-Markovnikov cyclization of ring D. Because optimal stabilization is absent in mutant Tyr609Phe a favorable Markovnikov cyclization will take place for ring D and therefore a high percentage of tetracycles with a five-membered ring D will be formed. On the other hand, it may be sterically argued. Phe is slightly smaller as compared to Tyr. Therefore, the folded squalene alters its position relative to Phe601 and Trp169 which now insufficiently stabilize the cation at C13 with consequently premature termination of the cyclization cascade [30]. Other mutants are listed in Table 16.2 for which the side products have not been identified.
277
16 The Changing Path of Hopanoid Research
16.9 Hopanoid biosynthesis gene clusters
In B. japonicum, Z. mobilis, and M. capsulatus to shc additional ORFs corresponding to genes for hopanoid biosynthesis (hpn genes) were sequenced (see Fig. 16.5) [21, 23]. Very similar clusters do also occur in Rhizobium NGR234 and Streptomyces coelicolor [33, 34]. In general, the ORFs overlap with their putative stop and start codons. Putative termination sites were found downstream of shc. The hpnC gene has been characterized as a squalene synthase gene [23]. E. coli transformed with hpnC was able to produce squalene; in E. coli the precursor farnesyl diphosphate is produced but not squalene. The reducing cofactor for this first bacterial squalene-synthase has still to be characterized in a cell-free system. Yet unknown is the function of hpnD which shows similarity to genes for squalene, dehydrosqualene, and phythoene synthases; the latter ones are involved in carotenoid biosynthesis. Neither B. japonicum nor Z. mobilis contain carotenoids. Presumably, hpnD codes for a dehydrosqualene synthase which may be reduced by HpnE (similar to phythoene dehydrogenase) to squalene. These two pathways for squalene formation possibly use two different reducing agents. This is perhaps of value for bacteria which have a high requirement of reducing power for N2-fixation. The upstream adjacent gene to shc in M. capsulatus was also characterized as a squalene synthase. This squalene synthase and the squalene synthase of B. japonicum have a very low identity of 32%. The four synthases of B. japonicum and Z. mobilis belong in the phylogenetic tree to a cluster of prokaryotic and eukaryotic phythoene synthases. We suppose therefore that some bacterial squalene synthases originate from phytoene synthases which have changed their function to squalene synthases.
Figure 16.5: The DNA regions from Zymomonas mobilis and Bradyrhizobium japonicum coding for hopanoid biosynthesis (hpn) genes. Features of the genes: hpnA, putative oxidoreductase of sugars; hpnB, putative glycosyl transferase; hpnC, gene for squalene synthase; hpnD, putative squalene or dehydrosqualene synthase; hpnE, similar to phytoene desaturase; hpnF, squalene-hopene cyclase.
278
16.10 Miscellaneous results The ORFs hpnA and hpnB seem to be connected with the elongation reactions in hopanoid biosynthesis. Elongated hopanoids contain at least one pentose and some additionally also a derivative of glucose. HpnA has similarity to UDP-glucose-4-epimerase and HpnB to glycosyl transferases [23]. The hopanoid biosynthesic gene cluster in Rhizobium NGR234 has the same structure as compared to B. japonicum. Both clusters are differently regulated. Under laboratory conditions the cluster of B. japonicum is expressed [13], whereas the cluster of Rhizobium NGR234 is not expressed (E. Kannenberg, unpublished results).
16.10 Miscellaneous results
Since the detection of hopanoids in Frankia [45] there existed a greater interest for hopanoid formation in N2-fixing bacteria. In cooperation we detected hopanoids in two Beijerinckia species and Azotobacter vinelandii and we also found that free-living B. japonicum and some relatives produce hopanoids [11, 13, 17]. It is tempting to speculate that the condensing function of hopanoids lowers the O2diffusion across the cell membrane and thereby prevents nitrogenase inactivation. Normally, tetrahymanol 4, an isomer of diplopterol 3 is found in lower eukaryotes. To our surprise it was also detected in the bacterium R. palustris [2]. In B. japonicum, besides tetrahymanol, methylated tetrahymanol species do exist (M. Perzl, J.-M. Bravo, E. Kannenberg, M. Rohmer, and K. Poralla, unpublished results). Organic geochemists are highly interested in this result, because now tetrahymanol and gammacerans (possessing no OH-group) in sediments are no longer indicative for the former occurrence of lower eukaryotes. In Staphylococcus aureus a dehydrosqualene synthase gene has been characterized as a member of the C30-carotenoid biosynthesis cluster [10]. Dehydrosqualene is a homologue to phythoene, and therefore the corresponding synthases are similar to squalene synthase which is an essential enzyme in hopanoid biosynthesis. As was mentioned above, wild type and mutant SHCs may be useful for the synthesis of new and unusual cyclic isoprenoids. An illustration for this is given by the transformation of a 29-methylidene-2,3-oxidosqualene 15 by SHC to a tetracyclic isodammarenoid 16 containing an additional cylohexane system [20]. In the genome project for Streptomyces coelicolor A3(2) an isoprenoid biosynthesis cluster of genes was described [34]. In this organism also two ORFs are occurring with significant similarity to phythoene and squalene synthases. The cluster also includes a gene for SHC. In liquid culture S. coelicolor does not produce hopanoids, but hopanoids are present when it sporulates on agar medium. All tested white mutants (B, G, and J) which are able to produce aerial mycelium but no spores synthesize hopanoids. Mutants with the property to pro279
16 The Changing Path of Hopanoid Research duce no aerial mycelium (bald) can be split into two groups. Only bald mutants (A, C, D, G, and H) which are defective in the production of an extracellular signal cascade are also deficient in the production of hopanoids, whereas bald B produces at least minute amounts of hopanoids (K. Poralla, G. Muth, and T. Härtner, unpublished results).
16.11 Outlook
Besides the described major achievements several important questions still remain to be solved. For the solution of these questions bacteria will be helpful for which genetic methods are established, as it is the case with B. japonicum and S. coelicolor. With the symbiotic B. japonicum the following questions should be answered: 1) is hopanoid synthesis essential during growth outside the host plant, and 2) what is the role for hopanoids in symbiosis? With S. coelicolor the question will be answered if hopanoid biosynthesis is essential for sporulation by insertion mutagenesis of the shc gene. Nothing is known about the regulation of the hopanoid biosynthesis cluster in B. japonicum and in Rhizobium NGR234. Also unclear is the observation that some organisms have two genes similar to squalene synthase. In our opinion only a single gene seems sufficient for hopanoid biosynthesis. A new straight forward strategy for mutagenesis is an evolutionary approach. Hopefully, this will lead to the isolation of altered SHCs with new substrate and product specificity. Plasmids may be constructed for Saccharomyces cerevisiae mutants deficient in oxidosqualene-lanosterol cyclase (erg7). By this way mutant SHCs may be isolated forming lanosterol and thus repeating in a short time the evolution from hopene to sterol cyclases. In this type of experiments it will be figured out which alterations are necessary unbiased by mutations not related to catalysis. This goal was one motivation for the ongoing research.
Acknowledgments
Since 1990 the following students forwarded the described project: Peter Gärtner, Dietmar Ochs, Cord Tappe, Gisela Kleemann, Jörg Saar, Corinna Feil, Michael Perzl, Annette Tippelt, Ok-Byung Choi, Angelika Rahn, Thorsten Merkofer, Susanne Schmitz, Christine Füll, and Michael Knigge. Most important for 280
References many tasks was Thomas Härtner. I am thankful to them, who contributed so much. Also my French colleagues helped a lot: Guy Ourisson, Michel Rohmer, and Catherine Pale-Grosdemange. Corinna Kaletta, Glenn D. Prestwich, Ikuro Abe, Hermann Sahm, and Georg Sprenger made essential contributions to the project. I am also thankful to Ulrich Wendt and Georg Schulz who enormously strengthened the project by their X-ray analysis. Günther Jung (collaborative research centre 323) was always very helpful. Without the financial help of the DFG this project would have never been started and continued. I like to remember the comments of the DFG-referees. I wish to thank Elmar L. Kannenberg for his input of ideas and critical reading of the manuscript.
References
1. Ochs, D., Kellner, R., and Poralla, K. (1990) Properties of purified squalene-hopene cyclase from Bacillus acidocaldarius. Europ. J .Biochem. 194, 75–80. 2. Kleemann, G., Poralla, K., Englert, G., Kjosen, H., Liaasen-Jensen, S., Neunlist, S., and Rohmer, M. (1990) Tetrahymanol from the phototrophic bacterium Rhodopseudomonas palustris, first report of a gammacerane triterpenoid from a prokaryote. J. Gen. Microbiol. 136, 2551–2553. 3. Saar, J., Kader, J.-C., Poralla, K., and Ourisson, G. (1991) Purification and some properties of the squalene-tetrahymanol cyclase from Tetrahymena thermophila. Biochim. Biophys. Acta 1075, 93–101. 4. Ochs, D., Kaletta, C., Entian, K.-D., Beck-Sickinger, A., and Poralla, K. (1992) Cloning, expression, and sequencing of squalene-hopene cyclase, a key enzyme in triterpenoid metabolism. J. Bacteriol. 174, 298–302. 5. Michel, G. P. F., Neuß, B., Tappe, C. H., and Baratti, J. (1992) Isolation and characterization of Zymomonas mobilis mutants resistant to octadecyltrimethylammonium chloride, a detergent acting on hopanoid-producing bacteria. Arch. Microbiol. 157, 116– 124. 6. C. Tappe (1993) Squalen-Hopen-Cyclasen. Reinigung, Charakterisierung und Inhibitor-Experimente. Dissertation, University of Tübingen. 7. Poralla, K. (1994) The possible role of a repetitive amino acid motif in evolution of triterpenoid cyclases. Bioorg. Med. Chem. Lett. 4, 285–290. 8. Kleemann, K., Kellner, R., and Poralla, K. (1994) Purification and properties of the squalene-hopene cyclase from Rhodopseudomonas palustris, a purple non-sulfur bacterium producing hopanoids and tetrahymanol. Biochim. Biophys. Acta 1210, 317–320. 9. Poralla, K., Hewelt, A., Prestwich, G.D., Abe, I., Reipen, I., and Sprenger, G. (1994) A specific amino acid motif in squalene and oxidosqualene cyclases. Trends Biochem. Sci. 19, 157–158. 10. Wieland, B., Feil, C., Maercker, E.-G., Thumm, G., Lechner, M., Bravo, J.-M., Poralla, K., and Götz, F. (1994) Genetic analysis of the biosynthesis of the yellow carotenoid 4,4'-diaponeurosporene of Staphylococcus aureus. J. Bacteriol. 176, 7719–7726. 11. Llopiz, C., Vilcheze, P., Neunlist, S., Poralla, K., and Rohmer, M. (1994) Prokaryotic triterpenoids. New hopanoids from the nitrogen fixing bacteria Azotobacter vinelandii, Beijerinckia indica and Beijerinckia mobilis. Microbiology 140, 2749–2753
281
16 The Changing Path of Hopanoid Research 12. Reipen, I. G., Poralla, K., Sahm, H., and Sprenger, G. A. (1995) Zymomonas mobilis squalene-hopene cyclase gene (shc): cloning, DNA sequence analysis, and expression in Escherichia coli. Microbiology 141, 155–161. 13. Kannenberg, E. L., Perzl, M., and Härtner, T. (1995) The occurrence of hopanoid lipids in Bradyrhizobium. FEMS Microbiol. Lett. 127, 255–262. 14. Choi, O.-B. (1995) Experimente zur Klonierung, Sequenzierung und Expression der Squalen-Hopen-Cyclasen aus Rhodopseudomonas palustris und Alicyclobacillus acidoterrestris. Dissertation, University of Tübingen. 15. Perzl, M., Müller, P., Poralla, K., and Kannenberg E. L. (1997) Squalene-hopene cyclase from Bradyrhizobium japonicum cloning, expression, sequence analysis and comparison to other triterpenoid cyclases. Microbiology 143, 1235–1242. 16. Feil, C., Süssmuth, R., Jung, G., and Poralla, K. (1996) Squalene-hopene cyclase. Sitedirected mutagenesis of putative active-site residues. Europ. J. Biochem. 242, 51–55. 17. Kannenberg, E. L., Perzl, M., Müller, P., Härtner, T., and Poralla, K. (1996) Hopanoid lipids in Bradyrhizobium and other plant associated bacteria and cloning of the Bradyrhizobium squalene-hopene cyclase gene. Plant and Soil 186, 107–112. 18. Wendt, K. U., Feil, C., Lenhart, A., Poralla, K., and Schulz, G. E. (1997) Crystallisation and preliminary X-ray crystallographic analysis of squalene-hopene cyclase from Alicyclobacillus acidocaldarius. Protein Sci. 6, 722–724. 19. Wendt, K. U., Poralla, K., and Schulz; G. E. (1997) Structure and function of a squalene cyclase. Science 277, 1811–1815. 20. Abe, I., Dang, T., Zheng, Y. F., Madden, B. A., Feil, C., Poralla, K., and Prestwich, G. D. (1997) Cyclization of (3S)29-methylidene-2,3-oxidosqualene by bacterial squalenehopene cyclase. Irreversible enzyme inactivation and isolation of an unnatural dammarenoid. J. Am. Chem. Soc. 119, 11333–11334. 21. Tippelt, A., Jahnke, L., and Poralla, K. (1998) Squalene-hopene cyclase from Methylococcus capsulatus (Bath). A bacterium producing hopanoids and steroids. Biochim. Biophys. Acta 1391, 223–232. 22. Pale-Grosdemange, C., Feil, C., Rohmer, M., and Poralla, K. (1998) Occurrence of cationic intermediates and deficient control during enzymatic cyclization of squalene into hopanoids. Angew. Chem. (Intern. Ed.) 37, 2237–2240. 23. Perzl, M., Reipen, I. G., Schmitz, S., Poralla, K.,Sahm, H., Sprenger, G. A., and Kannenberg, E. L. (1998) Cloning of conserved genes from Zymomonas mobilis and Bradyrhizobium japonicum that function in the biosynthesis of hopanoid lipids. Biochim. Biophys. Acta 1393, 108–118. 24. Poralla, K.(1998) Die Struktur der Squalen-Hopen-Cyclase. BIOspektrum 4, 32–34. 25. Poralla, K. (1999) Cycloartenol and other triterpene cyclases. In: Comprehensive Natural Products Chemistry. (Barton, D., Nakanishi, K., eds), Elsevier Amsterdam, Vol. 2, 299–319. 26. Kannenberg, E., Härtner, T., Perzl, M., Schmitz, S., and Poralla, K. (1999) Hopanoid lipid content of Bradyrhizobium bacteria is dependent on culture conditions. In: Highlights of Nitrogen Fixation Research (Martínez, Hernández, ed.), Kluwer/Plenum Publishers New York, pp 41–44. 27. Kannenberg, E. and Poralla, K. (1999) Hopanoid biosynthesis and function in bacteria. Naturwissenschaften 86,168–176. 28. Merkofer, T., Pale-Grosdemange, C., Wendt, K. U., Rohmer, M., and Poralla, K.(1999) Altered product pattern of a squalene-hopene cyclase by mutagenesis of active site residues. Tetrahedron Lett. 40, 2121–2124. 29. Pale-Grosdemange, C., Merkofer, T., Rohmer, M., and Poralla, K. (1999) Production of bicyclic and tricyclic triterpenes by mutated squalene-hopene cyclase. Tetrahedron Lett. 40, 6009–6012 30. Füll, C. and Poralla, K. (2000) Conserved Tyr residues determine functions of Alicyclobacillus acicocaldarius squalene-hopene cyclase. FEMS Microbiol. Lett. 183, 225–228.
282
References 31. Ourisson, G., Rohmer, M., and Poralla, K. (1987) Prokaryotic hopanoids and other polyterpenoid sterol surrogates. Annu. Rev. Microbiol. 41, 301–333. 32. Sahm, H., Rohmer, M., Bringer-Meyer, S., Sprenger, G. A., and Welle, R. (1993) Biochemistry and physiology of hopanoids in bacteria. Adv. Microbial Physiol. 35, 247– 273. 33. Freiberg, C., Fallay, R., Bairoch, A., Broughton, W. J., Rosenthal, A., and Perret, X. (1997) Molecular basis of symbiosis between Rhizobium and legumes. Nature 387, 394–401. 34. http://www.sanger.ac.uk/Projects/S_coelicolor/ 35. Seckler, B. and Poralla, K. (1986) Characterization and partial purification of squalene-hopene cyclase from Bacillus acidocaldarius. Biochim. Biophys. Acta 881, 356– 363. 36. Buntel, C. J. and Griffin, J. H. (1992) Nucleotide and deduced amino acid sequence of the oxidosqualene cyclase from Candida albicans. J. Am. Chem. Soc. 114, 9711– 9713. 37. Corey, E. J., Matsuda, S. P. T., and Bartel, B. (1994) Molecular cloning, characterization, and overexpression of ERG7, the Saccharomyces cerevisiae gen encoding lanosterol synthase. Proc. Natl. Acad. Sci. USA 91, 2211–2215. 38. O’Brien, P. J. and Herschlag, D. (1999) Catalytic promiscuity and the evolution of new enzymatic activities. Chemistry & Biology 6, R91–R105. 39. Abe, I. and Prestwich, G. D. (1994) Active site mapping of affinity-labeled rat oxidosqualene cyclase. J. Biol. Chem. 269, 802–804. 40. Sun, T.-P. and Kamiya, Y. (1994) The Arabidopsis GA1 locus encodes the cyclase entkaurene synthase A of gibberellin biosynthesis. Plant Cell 6, 1509–1518. 41. Stofer-Vogel, B., Wildung, M. R., Vogel, G., and Croteau, R. (1996) Abietadiene synthase from grand fir (Abies grandis) – cDNA isolation, characterization, and bacterial expression of a bifunctional diterpene cyclase involved in resin acid biosynthesis. J. Biol. Chem. 271, 23262–23268. 42. Wendt, K. U., Lenhart, A., and Schulz, G. E. (1999) The structure of the membrane protein squalene-hopene cyclase at 2.0 A resolution. J. Mol. Biol. 286, 175–187. 43. Picot, D., Loll, P. J., and Garavito, R. M. (1994) The X-ray crystal structure of the membrane protein prostagladin H2 synthase-1. Nature 376, 243–249. 44. Corey, E. J., Cheng, H., Hunter Baker, C., Matsuda, S. P. T., Li, D., and Song, X. (1997) Methodology for the preparation of pure recombinant S. cerevisiae synthase using a baculovirus expression system. Evidence that oxirane cleavage and A-ring formation are concerted in the biosynthesis of lanosterol 2,3-oxidosqualene. J. Am. Chem. Soc. 119, 1277–1288. 45. Berry, A. M., Moreau, R. A., and Jones, A. D. (1991) Bacteriohopanetetrol abundant lipid in Frankia cells and in nitrogen-fixing tissue. Plant Physiol. 95, 111–115. 46. Sato, T., Abe, and T., Hoshino, T. (1998) On the cyclization mechanism of squalene: a ring expansion process of the five-membered D-ring intermediate. Chem. Commun. 1999, 2617–2618.
283
17 Genetic and Biochemical Analysis of the Biosynthesis of the Orange Carotenoid Staphyloxanthin of Staphylococcus aureus Friedrich Götz*
17.1 Introduction
The yellow to orange colony color of Staphylococcus aureus is one of the classical species criteria. As early as 1882, Ogston connected the yellow-orange appearance of pus with the color of the infecting microorganisms [1]. Later it was shown that the pigment should not be the only basis for classification, since it is not a very stable character. Pigmentation is usually apparent after 18 to 24 h of growth at 37 8C, but is more pronounced when cultures are held at room temperature for further 24 to 48 h. In particular those S. aureus strains isolated from bovine or which are multiply antibiotic resistant, are yellow-pigmented. Nonpigmented (“white”) derivatives of S. aureus are often found in subcultures of stored organisms, and non-pigmented clones can arise in pigmented colonies, giving a sectored appearance [2]. Non-pigmented variants are more susceptible to desiccation and to linolenic acid than the corresponding wild type strains [3]. Treatment with nitrosoguanidine can result in irreversible loss of the capacity to synthesize pigment [4]. Although pigment production is a rather unstable character, it has been ruled out that the respective genes are encoded on typical plasmids [3]. Marshall and Wilmoth [5] isolated the pigments of S. aureus S41 and determined their chemical structure, identifying 17 compounds which are all triterpenoid carotenoids possessing a C30- instead of the C40-carotenoid structure found in most other organisms. The main pigment is staphyloxanthin, an alpha-Dglucopyranosyl-1-O-(4,4'-diaponeurosporene-4-oate)6-O-(12-methyltetradecanoate), in which glucose is esterified with both, triterpenoid carotenoid carboxylic acid and a C15 fatty acid. The postulated biosynthetic pathway of staphyloxanthin starts with a head-head condensation of two molecules of farnesyl diphosphate to form dehy-
* Mikrobielle Genetik, Universität Tübingen, Waldhäuser Str. 70/8, D-72076 Tübingen
284 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
17.2 Cloning of the carotenoid biosynthetic genes from S. aureus Newman drosqualene (= 4,4'-diapophytoene) or squalene. 4,4'-Diaponeurosporene is formed by three or four successive dehydrogenation steps and is the first yellow carotenoid intermediate. The oxidation of the terminal methyl group of 4,4'-diaponeurosporene to form the carboxylic acid 4,4'-diaponeurosporene-4-oic acid proceeds via the aldehyde 4,4'-diaponeurosporene-4-al. The acyl compound staphyloxanthin, as the final orange pigment in S. aureus, is formed via glucosyl-4,4'-diaponeurosporenoate [5]. This pathway has been verified by the analysis of various S. aureus mutants, in which intermediary products either were absent or accumulated [6]. The aim of our studies was to genetically characterize the steps of staphyloxanthin biosynthesis in staphylococci. Having identified the genes we are able to create deletion mutants and to study the function and regulation of staphyloxanthin. We chose S. carnosus TM300 as the cloning host because the colonies of this strain are white and should therefore not produce staphyloxanthin or colored intermediates. Staphyloxanthin-producing clones of S. carnosus harboring pOC1 were easily detected by their orange-colored colonies. In a first paper we described the characterization of two genes, crtM and crtN, involved in the biosynthesis of the yellow carotenoid 4,4'-diaponeurosporene and proposed a pathway for its biosynthesis by analysis of intermediary products in the various clones [7]. In a second paper (submitted) we analyzed the function of three further genes crtO, crtP, and crtQ. All five genes have the same orientation and form an operon (crtOPQMN) with a sB-dependent promoter upstream of crtO and a termination region downstream of crtN. The analysis of the sequences of CrtO, CrtP, and CrtQ and the spectrophotometric analyses of the carotenoids produced by Staphylococcus carnosus clones carrying crtMNQ, or crtMNOQ, or crtMNOPQ suggest that CrtP is responsible for the oxidation of 4,4'-diaponeurosporene to 4,4'-diaponeurosporen-4-oic acid and that the glucosylation of the carbon acid is catalyzed by CrtQ. All five genes including crtO were shown to be required for the formation of staphyloxanthine [8]. We also have investigated the survival of S. aureus strain Newman and its non-pigmented mutant DcrtM after exposure to UV light and to oleic acid to determine the function of the carotenoid. However, so far we have not got any clue as to the function of staphyloxanthin, therefore, these studies will be continued.
17.2 Cloning of the carotenoid biosynthetic genes from S. aureus Newman in S. carnosus and E. coli
S. aureus Newman chromosomal DNA was partially digested with MboI. DNA fragments of 5 to 25 kb were ligated to BamHI-digested pCA44. The ligated DNA was transformed into the non-pigmented S. carnosus TM300; several yel285
17 Genetic and Biochemical Analysis of the Biosynthesis low- and orange-pigmented S. carnosus colonies were detected. Restriction analysis of one isolated plasmid, pOC1, revealed that the orange-pigmented S. carnosus clone carried a 12-kb DNA insert in that plasmid [7]. Subcloning of fragments of the 12-kb insert in E. coli and S. carnosus led to yellow and non-pigmented clones in both organisms. The smallest fragment which caused yellow pigmentation was a 3.5-kb fragment containing the crtM and crtN genes. Expression of these two genes in E. coli (pUG1) or S. carnosus (pOC21) led to the formation of a yellow pigment. Analysis of this pigment revealed that it was the deep-yellow carotenoid 4,4'-diaponeurosporene. This is the major pigment produced by Staphylococcus aureus Newman which is after prolonged cultivation in part converted to the orange end product, staphyloxanthin.
17.3 Function of CrtM and CrtN
CrtM encodes a 254 aa protein of Mr 30,121 which is rather hydrophobic (39% hydrophobic aa). The deduced amino acid sequence of CrtM was compared with other enzymes involved in carotenoid biosynthesis. There are three boxes of pronounced similarities with the phytoene synthase of Synechococcus [9] and the squalene synthases of two Erwinia species, Saccharomyces cerevisia and of human. One domain represents the postulated prenyl (farnesyl) diphosphate binding motif [10]. CrtN encodes a 448 aa protein of Mr 50,853. This protein is rather hydrophobic too (38% hydrophobic aa), however, obvious membrane-spanning domains characteristic of integral membrane proteins are absent. The deduced amino acid sequence was compared with other carotenoid biosynthetic enzymes. CrtN shows a pronounced similarity (30–35%) to the phytoene desaturases (CrtI) of Erwinia herbicola and Rhodobacter capsulatus. The NH2-terminus of CrtN (aa positions 2–29) is distinguished by a conserved amino acid sequence with homology to FAD-, NAD(P)-binding domains of a series of dehydrogenases and oxidases. The phytoene desaturases catalyze the dehydrogenation reactions from phytoene to neurosporene (lycopene).
286
17.5 Identification of dehydrosqualene in E. coli (pUG1) and E. coli (UG9)
17.4 Identification of carotenoids in S. carnosus (pOC21), E. coli (pUG1), and S. aureus Newman
In the yellow-pigmented E. coli (pUG1) 4,4'-diaponeurosporene was identified by its characteristic absorption spectrum as a main carotenoid. 4,4'-diapo7,8,11,12-tetrahydrolycopene and 4,4'-diapolycopene were present in lower amounts. These results agree with the presence of the two genes, crtM and crtN, on plasmid pUG1. In radio-TLC, with farnesyl diphosphate as a substrate and E. coli (pUG1) extracts, dehydrosqualene and 4,4'-diaponeurosporene are produced; indicating that the proposed dehydrosqualene synthase (CrtM) is catalyzing the condensation of two farnesyl diphosphates to dehydrosqualene, and the proposed dehydrosqualene desaturase (CrtN) is catalyzing all oxidation steps between dehydrosqualene and 4,4'-diaponeurosporene. In the yellow-colored S. carnosus (pOC21) we found essentially the same carotenoid spectrum as in E. coli (pUG1), identifying 4,4'-diaponeurosporene as the main product at 22.3 µg/g cell dry weight. In S. aureus Newman extracts we found that the concentration of staphyloxanthin was only 50 % of that of 4,4'-diaponeurosporene. This can be explained by the observation that 4,4'-diaponeurosporene is already formed after 12 h cultivation, while staphyloxanthin is produced later after 24 h incubation. The amount of the main carotenoids was estimated spectrophotometrically using their specific extinction coefficients.
17.5 Identification of dehydrosqualene in E. coli (pUG1) and E. coli (UG9)
In E. coli JM83 extracts no squalene was detectable. On the other hand, we identified squalene in S. carnosus TM300 extracts, which is in agreement with earlier observations [11]. This indicates that a squalene synthase must be chromosomally encoded in S. carnosus. The finding of squalene in S. carnosus extracts raised the question as to the enzymatic activity of CrtM which exhibits similarities with phytoene and squalene synthases. In order to answer the question whether the C30-carotenoid biosynthesis starts with squalene or, in analogy to the C40-carotenoid biosynthesis, with dehydrosqualene (= 4,4'-diapophytoene) we constructed plasmid pUG9 containing only the intact crtM gene. In enzyme assays with cell free extracts of E. coli (pUG9) we never obtained squalene, irrespectively of the presence of NAD(P)H. However, we could identify a compound with a slightly lower Rf -value compared to squalene 287
17 Genetic and Biochemical Analysis of the Biosynthesis which was found in a very hydrophobic fraction of the lipid extraction of E. coli (pUG9). The samples were separated by HPLC and analyzed by diode array spectroscopy. Only one peak occurred, showing an absorption maximum at 287 nm and shoulders at 275 and 297 nm. Based on the characteristic absorption spectrum [6] we identified this compound as dehydrosqualene. No dehydrosqualene was detected in control experiments with E. coli (pUC19). The isolated dehydrosqualene revealed in GLC/MS-analysis two peaks of similar retention times with molecular ions at 408 m/z, corresponding to C30H48. The fragmentation patterns are nearly identical, suggesting the two compounds to be the cis- and trans-isomers of dehydrosqualene. Since there is no squalene found in the E. coli (pUG9) extracts these results show that CrtM represents a dehydrosqualene synthase.
17.6 Squalene is very likely no substrate for CrtN, the proposed dehydrosqualene desaturase
With squalene as a substrate and with cell extracts of either E. coli (pUG1) or S. carnosus (pOC21) we never observed in an in vitro assay the formation of 4,4'diaponeurosporene, irrespectively of the presence of cofactors (such as FAD, NADP, NAD, FMN) or various divalent metal ions. We therefore think that the proposed dehydrosqualene desaturase is specific for dehydrosqualene.
17.7 The crt operon
The caroteinoid biosynthesis genes on plasmid pOC1 were further sequenced, and an operon containing five genes (crtOPQMN) was identified; in addition to the known genes crtM and crtN three further genes upstream of crtM were found. By analysis of deletion mutants it turned out that all five genes are necessary for the formation of the orange staphyloxanthin. The organization of the operon is shown in Fig. 17.1.
Figure 17.1: Organization of the staphyloxanthin biosynthesis genes of Staphylococcus aureus strain Newman.
288
17.9 Construction of crtM mutants of S. aureus strain Newman
17.8 Homology of CrtO, CrtP, and CrtQ
CrtO (Mr 20.3 kDa) shows no similarities to any sequences of other carotenoid biosynthesis proteins known. However, the deduced amino acid sequence of CrtP (Mr 50.8 kDa) shows similarities to sequences of diapophytoene dehydrogenase of Heliobacillus mobilis (identity: 29%; similarity: 93%) [Takaichi 1997 #34], phytoene dehydrogenases from Myxococcus xanthus (identity: 30 %; similarity: 63%) [12], Erwinia herbicola (identity: 28%; similarity: 62%) [13], Phycomyces blakesleeanus (identity: 28%; similarity: 61%) [14], and of CrtN of S. aureus strain Newman [7]. CrtQ (Mr 42.5 kDa) shows similarities to the sequences of galactosyl and glycosyl transferases of Bradyrhizobium japonicum (identity: 30 %; similarity: 62%) [15], and to the processive glycosyl transferases of S. aureus (identity: 23%; similarity: 41%) [16] and S. epidermidis (identity: 22%; similarity: 41%) [17]. Five strictly conserved amino acids are involved in the catalytic function of processive glycosyltransferases. A model for beta-glycosyl transferase shows that the catalytic amino acids are organized in two domains: domain D1 contains one amino acid and domain D2 contains four amino acids [18]. Non-processive glycosyl transferases (transferases that transfer only one monosaccharide) have only the one conserved amino acid in domain D1. The alignment of the CrtQ sequence with that of the processive glycosyltransferase IcaA from S. aureus [16] and the corresponding gene from S. epidermidis [17] shows that CrtQ has only the conserved amino acid in domain D1. No sequence similarities to the strictly conserved amino acids of domain D2 were detected. Therefore, CrtQ appears to be a member of the non-processive glycosyl transferase family.
17.9 Construction of crtM mutants of S. aureus strain Newman
By using plasmid pRB573SXDcrtM, the crtM gene in the chromosome of S. aureus strain Newman was exchanged with the gene encoding chloramphenicol transferase (cat) through a double-crossover event. The resulting mutant, S. aureus strain Newman DcrtM did not produce C30-carotenoids and formed colorless colonies on TSB agar plates. This loss of pigmentation also showed that no alternative pigment biosynthesis pathway exists in this strain. No differences between the growth of the wild type strain and of the DcrtM mutant were observed. The inserted cat gene, which lacks a promoter, was also used as a reporter gene to study the regulation of the pigment biosynthesis genes in S. aureus 289
17 Genetic and Biochemical Analysis of the Biosynthesis strain Newman. In the DcrtM mutant, the cat gene is under the control of the promoter of the crt operon. Therefore, the transcriptional regulation of the crt operon was studied by following Cat activity during the growth of the culture. The Cat activity increased significantly at the beginning and during the stationary phase, which is in agreement with earlier observations of a marked increase in the pigmentation of the wild type strain only after 24 to 36 h of growth [7].
17.10 sB-regulated promoter of the crt operon from S. aureus strain Newman
The loss of the sigB gene and the regulatory genes rsbV and rsbW leads to a loss of pigmentation in S. aureus strain Newman. In addition, the colorless S. aureus strain RN4220 carrying plasmid pIK57 (a pTX15 derivative with an inducible sigB gene [19]) is pigmented after sigB induction. Based on these results, a sB-regulated crt promoter has been hypothesized [19, 20]. Further characterization of the crt operon promoter region indicated that indeed the crt operon promoter is regulated by sB. The transcription start point of the crt operon was identified using primer-extension analysis.
17.11 The carotenoid biosynthesis genes
All of the carotenoid biosynthesis genes were cloned together into vector pTX15 under the control of the xylA promoter, forming plasmid pTXSX. With this system, it was possible to induce a strong pigment formation in S. carnosus. To investigate the function of carotenoid synthesis genes and the metabolic products, subsets of genes were cloned into pTX15 behind the xylA promoter, forming plasmids pTXMNPQ (carrying crtMNPQ), pTXMNQ (carrying crtMNQ), and pTXMN (carrying crtMN). After cultivation of the recombinant strains, the carotenoids were extracted and separated by preparative TLC, and the absorption spectrum of each individual pigment band was recorded. From the data obtained, it was possible to assign the gene products of crtP, crtQ, and crtO to the steps of the hypothetical biosynthesis pathway (Fig. 17.2). As shown previously, the gene products of crtM and crtN are responsible for the formation of the first yellowcolored C30-carotenoid, 4,4'-diaponeurosporene. The pigments produced by 290
17.11 The carotenoid biosynthesis genes
Figure 17.2: Proposed enzymatic pathway of the S. aureus staphyloxanthin biosynthesis: CrtM, dehydrosqualene synthase; CrtN dehydrosqualene desaturase (major intermediary product is 4,4'-diaponeurosporene); CrtP, diapophytoene dehydrogenase; CrtQ, non-processive glycosyl transferase; CrtO, esterification with a fatty acid (last step in staphyloxanthin biosynthesis).
S. carnosus (pTXMNQ) and S. carnosus (pTXMN) were similar, and differences only in the non-pigmented compounds were observed after exposing the developed TLC plates to UV light. This change can be explained by an unspecific glycosylation by CrtQ. 4,4'-Diaponeurosporene was only modified when the crtP gene product was also present in S. carnosus (pTMNPQ). The spectral data indicated that CrtP catalyzes the oxidation of a terminal methyl side group, forming 4,4'-diaponeurosporen-4-oic acid, and the carboxyl group is then glycosylated by CrtQ; the last step in the pathway is the esterification with a fatty acid by CrtO.
291
17 Genetic and Biochemical Analysis of the Biosynthesis
17.12 Function of the pigments in S. aureus strain Newman
The survival of S. aureus strain Newman and its non-pigmented mutant DcrtM after exposure to UV light and to oleic acid was tested to determine the function of the carotenoid. After treatment with UV light, the survival of S. aureus strain Newman DcrtM and DsigB mutants was only slightly lower than that of the wild type strain. This shows that pigmentation does not protect against UV light. The oleic-acid-killing assay was previously used by Chamberlain et al. [21] in their studies of the pigmented S. aureus strain 18Z and an undefined colorless mutant strain 18Z-76. Mutant 18Z-76 has a higher sensitivity to oleic acid than the wild type strain. The authors concluded that there is a relationship between pigment production and the resistance of S. aureus to the cell-damaging influence of oleic acid. In our tests, there was no difference in the sensitivity of S. aureus strain Newman wild type and the DcrtM mutant to treatment with oleic acid. The pigmentation is therefore probably not responsible for the higher resistance of the colorless mutant 18Z-76 to oleic acid. The difference seen was probably due to a mutation in a gene that controls pigment formation such as sigB, rsbU, or rsbW as well as other function(s) that rendered the cells more sensitive to oleic acid.
17.13 Distribution of pigment biosynthesis genes among staphylococcal species
The distribution of the C30-carotenoid biosynthesis genes among staphylococci was examined using the S. aureus strain Newman crtM gene, which encodes the key enzyme for the synthesis of the C30-carotenoid, as a probe in cross-species DNA hybridization experiments. Seven species formed pigmented colonies and gave a positive signal with the crtM probe.
292
References
References
1. Kloos, W., Schleifer, K. H., and Götz, F. (1991) in: The Prokaryotes, pp. 1369–1420 (Balows, A., Trüper, H. G., Dworkin, M., Harder, W., and Schleifer, K.H., Eds.) Springer Verlag, London. 2. Servin-Massieu, M. (1961) Spontaneous appearance of sectored colonies in Staphylococcus aureus. J. Bacteriol. 82, 316–317. 3. Grinsted, J. and Lacey, R. W. (1973) Ecological and genetic implications of pigmentation in Staphylococcus aureus. J. Gen. Microbiol. 75, 259–67. 4. Altenbern, R. A. (1967) Genetic studies of pigmentation of Staphylococcus aureus. Can. J. Microbiol. 13, 389–95. 5. Marshall, J. H. and Wilmoth, G. J. (1981) Pigments of Staphylococcus aureus, a series of triterpenoid carotenoids. J. Bacteriol. 147, 900–913. 6. Marshall, J. H. and Wilmoth, G. J. (1981) Proposed pathway of triterpenoid carotenoid biosynthesis in Staphylococcus aureus: evidence from a study of mutants. J. Bacteriol. 147, 914–919. 7. Wieland, B., Feil, C., Gloria Maercker, E., Thumm, G., Lechner, M., Bravo, J. M., Poralla, K., and Götz, F. (1994) Genetic and biochemical analyses of the biosynthesis of the yellow carotenoid 4,4'-diaponeurosporene of Staphylococcus aureus. J. Bacteriol. 176, 7719–7726. 8. Wieland, K.-P. and Götz, F. (2000) Organisation and function of the staphyloxanthin biosynthesis genes of Staphylococcus aureus. submitted. 9. Chamovitz, D., Misawa, N., Sandmann, G., and Hirschberg, J. (1992) Molecular cloning and expression in Escherichia coli of a cyanobacterial gene coding for phytoene synthase, a carotenoid biosynthesis enzyme. FEBS Lett. 296, 305–10. 10. Jennings, S. M., Tsay, Y. H., Fisch, T. M., and Robinson, G. W. (1991) Molecular cloning and characterization of the yeast gene for squalene synthetase. Proc. Natl. Acad. Sci. USA 88, 6038–42. 11. Suzue, G., Tsukada, K., Nakai, C., and Tanaka, S. (1968) Presence of squalene in Staphylococcus. Arch. Biochem. Biophys. 123, 644. 12. Ramakrishnan, L., Tran, H. T., Federspiel, N. A., and Falkow, S. (1997) A crtB homolog essential for photochromogenicity in Mycobacterium marinum: isolation, characterization, and gene disruption via homologous recombination. J. Bacteriol. 179, 5862–8. 13. Matsumura, H., Takeyama, H., Kusakabe, E., Burgess, J. G., and Matsunaga, T. (1997) Cloning, sequencing and expressing the carotenoid biosynthesis genes, lycopene cyclase and phytoene desaturase, from the aerobic photosynthetic bacterium Erythrobacter longus sp. strain Och101 in Escherichia coli. Gene 189, 169–74. 14. Ruiz-Hidalgo, M. J., Benito, E. P., Sandmann, G., and Eslava, A. P. (1997) The phytoene dehydrogenase gene of Phycomyces: regulation of its expression by blue light and vitamin A. Mol. Gen. Genet. 253, 734–44. 15. Cohen, J. L. and Miller, K. J. (1991) A novel membrane-bound glucosyltransferase from Bradyrhizobium japonicum. J. Bacteriol. 173, 4271–6. 16. Cramton, S. E., Gerke, C., Schnell, N. F., Nichols, W. W., and Götz, F. (1999) The intercellular adhesion (ica) locus is present in Staphylococcus aureus and is required for biofilm formation. Infect. Immun. 67, 5427–5433. 17. Heilmann, C., Schweitzer, O., Gerke, C., Vanittanakom, N., Mack, D., and Götz, F. (1996) Molecular basis of intercellular adhesion in the biofilm-forming Staphylococcus epidermidis. Mol. Microbiol. 20, 1083–1091. 18. Saxena, I. M., Brown, R. M., Jr., Fevre, M., Geremia, R. A., and Henrissat, B. (1995)
293
17 Genetic and Biochemical Analysis of the Biosynthesis Multidomain architecture of beta-glycosyl transferases: implications for mechanism of action. J. Bacteriol. 177, 1419–1424. 19. Kullik, I., Giachino, P., and Fuchs, T. (1998) Deletion of the alternative sigma factor sigmaB in Staphylococcus aureus reveals its function as a global regulator of virulence genes. J. Bacteriol. 180, 4814–4820. 20. Kullik, I. and Giachino, P. (1997) The alternative sigma factor sigmaB in Staphylococcus aureus: regulation of the sigB operon in response to growth phase and heat shock. Arch. Microbiol. 167, 151–159. 21. Chamberlain, N. R., Mehrtens, B. G., Xiong, Z., Kapral, F. A., Boardman, J. L., and Rearick, J. I. (1991) Correlation of carotenoid production, decreased membrane fluidity, and resistance to oleic acid killing in Staphylococcus aureus 18Z. Infect. Immun. 59, 4332–4337.
294
18 Second Messenger Systems in Paramecium Joachim E. Schultz* and Jürgen Linder
18.1 Introduction
The world of microbes has proven to be a large and mainly unopened chest, which contains a rich panoply of metabolites representing chemically highly diverse structures. So far perhaps less than 1% of all bacteria thought to exist has been investigated in this respect. For those engaged in the development of novel lead compounds for therapeutic purposes the bacterial producers of secondary metabolites turned out to be gold mines. Yet, discriminating between real nuggets and fool’s gold has become a more and more difficult task. We are facing the fact that isolation and structural elucidation of novel structures is only one side of the coin, the other being the daunting task to screen for a biological activity, which may be of potential therapeutic value. In a somewhat daring and certainly novel approach we wished to use a well known protozoan unicell, Paramecium, as a eukaryotic model to assay mixtures of and purified compounds of microbial origin for an activity on neurophysiological membrane events. The rationale for this approach was based on several lines of thinking. First, considering the universality of governing principles apparent in modern biology we felt that there exist universal mechanisms of excitation, adaptation and sensory transduction in nature, which can be examined in a lower eukaryote such as Paramecium. Second, Paramecium has been a favorite neurobiological model cell to study behavior because the genetics and the electrophysiology are well understood [1–4]. Like neuronal cells, Paramecium has an excitable membrane and generates action potentials, which are graded to the strength of the stimulus. In addition, the stereotypic swimming behavior of this ciliated creature is a direct correlate of electrophysiological events and can, therefore, be easily observed under a stereomicroscope without obligatory measurements
* Pharmazeutische Biochemie, Universität Tübingen, Auf der Morgenstelle 8, D-72076 Tübingen
295 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
18 Second Messenger Systems in Paramecium of electrogenesis by impaling the cells with microelectrodes [3]. An avoiding response, i. e. backward swimming, is always associated with a depolarizing Ca2+inward current whereas hyperpolarization is correlated with accelerated forward swimming. Third, Paramecium responds to chemical, mechanical and thermal cues, i. e. the machinery for different sensory modalities is present. Finally, Paramecium is a fresh-water protozoan, which displays several typical protozoan features known from parasitic protozoa such as a variable glycoprotein surface coat. It was our hope that data generated with this harmless protozoan model may become applicable to some of the dreaded parasitic protozoans, which are responsible for a high fraction of diseases afflicting human beings. As early as 1980, we established as a first step conditions for an axenic massculture of Paramecium, which allowed the production of a cell mass sufficient for protein biochemistry to be carried out [5, 6]. This allowed the identification and characterization of biochemical targets, which may be applicable for screening purposes. This should permit then to use a triple strategy for testing novel compounds of bacterial origin with respect to unknown biological activities: 1) in vivo studies such as observation of an intrinsic activity to elicit stereotypically characterized changes in swimming patterns, which would indicate an interpretable action on electrogenesis in Paramecium or an activity (or toxicity) to impair behavioral responses to canonical stimuli such as sustained depolarization caused by addition of Ba2+ and high K+ or hyperpolarization caused by dilution of the external K+ or an increase of the external Ca2+ (calcium paradox) [7] concentrations; 2) ex vivo studies by measuring the effect of compounds on the intracellular generation of the second messengers cAMP and cGMP, which are biochemical correlates of electrical events in this ciliate [8–12], and 3) in vitro studies, in which the effect of novel compounds or compound mixtures on individual, isolated or recombinant components of the second messenger signal transduction system, such as adenylyl and guanylyl cyclases, phosphoprotein phosphatases, phosphodiesterases, and protein kinases, is examined [13–15].
18.2 Identification and characterization of cGMP and cAMP second messenger signaling systems in Paramecium
18.2.1 Regulation of cGMP levels Cyclic nucleotides are used universally as second messengers. In metazoan cells regulation of cyclic nucleotide biosynthesis routinely involves a first messenger, that interacts with an extracellularly oriented membrane receptor, which in turn 296
18.2 Identification and characterization of cGMP and cAMP second messenger couples to adenylyl cyclase via heterotrimeric G-proteins or directly activates a guanylyl cyclase catalyst by a receptor intrinsic to the membrane-bound enzyme. Soluble guanylyl cyclases are activated by the intracellular first messenger nitrogen oxide radical [16, 17]. In a free-living freshwater protozoan signal sensing and signal responses can be expected to concern overwhelmingly environmental changes such as in the nutrient situation, the ion composition of the pool or a predator/prey relationship. We wished to determine, which stimuli affect and regulate intracellular cAMP and cGMP formation in the ciliate Paramecium. Basically, we found that sudden changes in the external ion composition are the most reliable effectors of cyclic nucleotide biosynthesis in the ciliate. cGMP levels are regulated by a depolarizing Ca2+-inward current elicited, e. g. by the addition of Ba2+-ions or an increase in the concentration of external K+, i. e. the Ca2+-inward current serves a dual role as a simple charge carrier and, conceptually, as a first messenger, which activates a membrane bound guanylyl cyclase [10]. The Ba2+-activated Ca2+-inward current is responsible for a shortlived 6-fold increase in intracellular cGMP in conjunction with a behavioral avoiding response, i. e. repeated bouts of backward swimming (Fig. 18.1). The importance of this current for cGMP biosynthesis was identified by using Paramecium mutants with identified defects in Ca2+-channel regulation. The pawn mutants are unable to respond to a depolarizing stimulus with backward swimming because the voltage-operated Ca2+-channel cannot be opened and remains shut [1, 18–20]. The biochemical readout of the missing Ca2+-inward current in these mutant cells is a lack of cGMP generation although the guanylyl cyclase enzyme activity in the mutants seems to be perfectly normal (Fig. 18.1) [10]. Paramecium
Figure 18.1: Depolarization as a hormone surrogate stimulates cGMP formation in Paramecium. Time-dependent increase in intracellular cGMP concentrations in Paramecium tetraurelia as stimulated by a BaCl2-elicited depolarization (wild type 51s (*), double mutant pawn A/pawn B (t) and mutant dancer (&)). For stimulation 250 µl 6 mM BaCl2 in equilibration buffer (5 mM KCl, 50 µM CaCl2, 10 mM MOPS, pH 7.2) were added to 250 µl of cells. Note the transient response in wild type cells, the extended response in dancer mutant cells which have a defect in Ca2+-channel closure, and the absence of a response in pawn mutant cells which have a defect in opening the voltage-gated Ca2+-channel (adapted from [10]).
297
18 Second Messenger Systems in Paramecium dancer mutants have a genetic defect opposite to the pawn mutation, i. e. they lack the ability to quickly close the activated Ca2+-channel and consequently show extended bouts of ciliary reversal even upon minor depolarizing stimuli [21, 22]. The biochemical correlate of this genetic defect is an exaggerated and sustained intracellular cGMP formation upon depolarization (Fig. 18.1) [10]. We have characterized intensively the enzymatic properties of a particulate guanylyl cyclase of Paramecium, which is localized to a large extent in the excitable ciliary membrane [23, 24]. The enzyme is stimulated by micromolar concentrations of Ca2+. This effect most likely is mediated by an accessory protein initially assumed to be calmodulin [24]. Several experimental observations put this conclusion into doubt, e. g. µM concentrations of exclusively protozoan calmodulins are required and mammalian calmodulins were inactive despite extensive sequence identities. The protein, which actually mediates the Ca2+-effect on the guanylyl cyclase may be a minor, heat-stable, as yet unidentified protein contaminant, which is present in calmodulin preparations specifically from Paramecium and Tetrahymena. Certainly, further work is required to elucidate the details of the Ca2+-regulation of the protozoan guanylyl cyclases.
18.2.2 Regulation of cAMP levels The regulation of intracellular cAMP formation in Paramecium was even more astounding. Control of cAMP levels in Paramecium is not governed by hormones, but by changes in the concentration of extracellular ions that cause hyperpolarization [9, 11, 25]. As a freshwater ciliate, Paramecium must actively control its K+-resting conductance to maintain the resting membrane potential within narrow limits in shifting environments. Therefore, the extent of a hyperpolarization that is caused by dilution of external cations, is dependent on the composition of the equilibration buffer, which determines the existing resting conductance, mainly a K+-conductance. We investigated whether the K+-resting conductance itself may be coupled to activation of a membrane-bound adenylyl cyclase. When Paramecia are adapted for 4 hours to a solution containing 16 mM external K+, in which the resting conductance is high, an instantaneous dilution of K+ to 2 mM causes a rapid, about four-fold rise of cAMP levels starting within milliseconds [9]. It reaches its maximum within 5 to 10 s and declines to a new steady state level after about 240 s. (Fig. 18.2 A). This biochemical response is accompanied by a fast swimming response, which persists for longer periods of time than the actual cAMP response indicating that both must not be coupled. Interestingly, in cells adapted to 1 mM K+, in which the resting conductance is low, dilution of K+ to one eight of its original concentration does not enhance cAMP formation. Similarly, the behavioral response is almost absent. Apparently the stimulation of cAMP formation is correlated with hyperpolarizing changes of the resting membrane potential and suggests that both are coupled. In fact, hyperpolarization and cAMP production were quantitatively related to 298
18.2 Identification and characterization of cGMP and cAMP second messenger
Figure 18.2: Hyperpolarization as a hormone surrogate stimulates cAMP formation in Paramecium. (A) Time-dependent formation of intracellular cAMP elicited by hyperpolarization. Cells were equilibrated in a buffer containing 16 mM KCl, 0.5 mM CaCl2, 10 mM MOPS, pH 7.2. For stimulation, KCl was diluted to one eighth of its original concentration by transfer of 50 µl of cells into 350 µl of potassium-free buffer. (B) Hyperpolarization-dose response curve for cAMP formation. Cells equilibrated in 16 mM KCl were stimulated by dilution of external KCl proportionally as indicated on the abscissa and cAMP levels were determined 5 s later. (C) Hyperpolarization-stimulated cAMP formation in Paramecium wild type and K+-channel mutant cells (restless). Cells were equilibrated in 4 mM KCl and stimulated by an eight-fold dilution of KCl to 0.5 mM. cAMP was determined after 5 s (open bars: control cells; black bars: dilution-stimulated cells). Addition of the K+-channel blocker tetraethylammonium (8 mM; patterned bar) with the dilution buffer abolished the exaggerated response in restless mutant cells (data are adapted from [11]).
299
18 Second Messenger Systems in Paramecium each other. When graded stimuli were applied to Paramecia by dilution of external K+ to different extents, hyperpolarization response curves for cAMP formation were obtained, which were essentially equivalent to hormone dose response curves observed in mammalian cells and tissue preparations (Fig. 18.2 B). This means that conceptually hyperpolarization of Paramecium is a stimulus, which corresponds to a hormonal adenylyl cyclase stimulation in a metazoan cell. To investigate the directness of coupling between the K+-resting conductance and cAMP generation we used classical K+-channel blockers. Tetraethylammonium, Cs+-ions or quinine, which block K+-channels in many cells including Paramecium, inhibited cAMP production upon dilution of external K+ dosedependently (Fig. 18.2 C). Further, a mutant of Paramecium, restless, is overreactive to low K+. The mutant cannot control its K+-resting conductance and behaves like a K+-electrode [20, 26]. E.g. at 4 mM K+-restless has a membrane resistance of 30 MO compared to 60 MO in wild type cells. When incubated in low K+ medium it will die because of the incessant loss of cellular K+. Consequently, restless adapted to 4 mM K+ responds to an eight-fold dilution of K+ with a large accumulation of cAMP whereas wild type cells do not (Fig. 18.2 C). The shift in the K+-equilibrium potential as brought about by dilution is sufficient to transiently hyperpolarize the cell. In principle, a net ion current is not required. Nevertheless, a sudden hyperpolarization affects those inwardly and outwardly directed ion currents, which are at equilibrium at rest, i. e. are electrically balanced. Application of the Nernst equation indicates that only a K+-efflux is possible that may flow through K+-channels, which carry the major resting conductance, and is afflicted in restless mutant cells. This resting current then appears to serve a non-electrical function in that it is involved in regulating cAMP formation.
18.3 Biochemical properties of an adenylyl cyclase
The above data were of particular interest considering the proposed membrane topology of mammalian adenylyl cyclases. The currently accepted topology comprises two hydrophobic membrane cassettes of six a-helical transmembrane spans, designated M1 and M2, and a catalytic center formed by two conserved cytoplasmic domains of approximately 250 amino acids, C1 a and C2 a, that are downstream of M1 and M2, respectively. C1 a and C2 a precede poorly conserved cytosolic segments of variable length, designated C1 b and C2 b, the latter is missing in some isoforms. The overall topology of mammalian adenylyl cyclases can thus be abbreviated as M1C1 abM2C2 a(b) [17] (see Fig. 18.3 A, right, for a model depiction). Immediately after this membrane topology was recognized it was speculated that the bulky membrane anchors of the adenylyl cyclase may 300
18.3 Biochemical properties of an adenylyl cyclase
Figure 18.3: (A) Putative membrane topology of the cloned guanylyl cyclase from Paramecium (Tetrahymena and Plasmodium). The transmembrane spans are depicted as barrels and the C-terminal cytosolic portions of the guanylyl cyclase are shown doughnut shaped for the C1 a/C2 heterodimer [31]. Like in several mammalian adenylyl cyclase isoforms, a distinct C2 b region is absent in the Paramecium guanylyl cyclase. The shorthand nomenclature used for mammalian adenylyl cyclases has been adopted. M1 and M2 designate the two membrane anchors with six transmembrane spans, the C1 a- and C2positioned domains form the catalytic center. (B) A local alignment of the catalytic C1 a and C2 a-positioned domains of the Paramecium guanylyl cyclase with the corresponding mammalian adenylyl (bovine type VII) and guanylyl cyclase (rat soluble) sequences shows that the C1 a and C2-positioned cytosolic domains of the Paramecium guanylyl cyclase are inverted. The signature motifs GDCY and TYMA are shaded, substrate defining amino acid residues are enlarged. Data are adapted from [48].
301
18 Second Messenger Systems in Paramecium possess an intrinsic transport activity [27]. Similar topologies were known to be present in the multi-drug resistance gene, various membrane transporters such as ABC-transporters and voltage-operated ion channels. Quite obviously, our lines of thinking were influenced by these finding. We hypothesized that the membrane-bound adenylyl cyclase of Paramecium may have retained an ancestral ion conductance. Therefore, we axenically cultivated Paramecia in bioreactors of up to 200 liters [6]. Because the cilia constitute only 1% of cellular protein, yet contain 50 % of adenylyl cyclase activity, they were used as starting material for adenylyl cyclase purification. About 20 ng of purified adenylyl cyclase were obtained from 170 g of Paramecium with a specific activity of 25 µmol/mg/min [11]. Analysis by SDS-polyacrylamide gel electrophoresis showed a single band at 96 kDa, i. e. in a range compatible with an adenylyl cyclase of mammalian membrane topology. Next, we tested the purified enzyme in artificial lipid-bilayer membranes for pore-forming and ion-conductance capacity. When purified enzyme was used for reconstitution, an ion conductance of 320 ± 60 pS was measured [11]. The relative permeability ratios for K+: Cs+: Na+: Ca2+: Li+: Mg2+ were determined to be 1 : 1 : 0.5 : 0.3 : 0.25 : 0.2, respectively. Tetraethylammonium did not pass the pore. We were unable to reliably test whether K+-channel blockers affected the reconstituted ion conductance. The observed conductance was impermeable for anions such as chloride or acetate. The enzymatic activity and pore-forming capacity were strictly interdependent. All purification steps that interfered with adenylyl cyclase activity also impaired pore-forming activity, which co-purified through six distinct separation steps. This data strongly suggested that the protozan adenylyl cyclase is not only regulated by the resting membrane potential, but is the transmembrane ion pore, which is responsible for setting the resting potential of the cell. Thus, it may possess an intrinsic secondary, regulatory function compatible with the metazoan adenylyl cyclase membrane topology. Obviously, cloning of the gene corresponding to the protozoan adenylyl cyclase would clarify the exact relationship of this cyclase to its metazoan congeners and eventually yield clues what a presumably primordial ion conductance of the adenylyl cyclase membrane anchor may look like.
18.4 A guanylyl cyclase disguised as an adenylyl cyclase
Initially, we were determined to purify enough protein to obtain amino acid sequence information, which could be used to develop degenerate primers and clone the gene using PCR and a Paramecium cDNA library as a template. All efforts toward that goal were, however, unsuccessful. This was highly discouraging because despite the universality of cAMP as a second messenger, at least three phylogenetically separate classes of adenylyl cyclases exist (a forth and possibly a fifth class are currently discussed) and notwithstanding the above ar302
18.4 A guanylyl cyclase disguised as an adenylyl cyclase guments there was no certainty that the presumed adenylyl cyclase of Paramecium would belong to the class III isoform family present in all eukaryotes. The three classes of adenylyl cyclases, which are defined on the basis of sequence identities and similarities in their putative catalytic domains, share no sequence similarity and the current view is that they evolved independently [17, 28]. The class I adenylyl cyclases evolved in bacteria, e. g. in E. coli or Yersinia and are represented by cytosolic or membrane-attached forms. Class II isoforms comprise the so-called bacterial toxins from Bordetella pertussis and Bacillus anthracis. The extracellularly secreted enzymes of up to 200 kDa are bifunctional [29, 30]. Notably, these toxins are linked with a faintly hemolysin-like domain, which inserts itself as a membrane pore into the eukaryotic cell membrane and uses this gate to deliver the catalytic domain to the cytosol. The class III cyclases, comprising adenylyl as well as guanylyl cyclases, are present in essentially all eukaryotes looked at so far. However, members of this cyclase family also occur in bacteria such as Brevibacterium, Mycobacterium, Stigmatella and Anabaena [28, 31]. Most recently, it was reported that a soluble adenylyl cyclase related to class III enzymes is present in rat testis in a very large protein background of unknown function [32]. This indicates that class III adenylyl cyclase catalysts may occur in different protein backgrounds connected to additional functional and potentially regulatory or regulated domains. For a homology cloning approach we used a 328 bp genomic DNA sequence from Paramecium, which displayed 29% amino acid identity to the catalytic C2 a region of a rat type III adenylyl cyclase isoform (kindly provided by C. Russel and R. D. Hinrichsen) [33], i. e. we set out to clone an adenylyl cyclase, which would conform to the mammalian membrane topology M1C1 abM2C2 a(b). Using cDNA library screening we obtained a huge 7.2 kb DNA fragment with a 3'-TGA stop, a poly-A tail and a continuous open reading frame extending all the way to the 5'-end. A genomic DNA library was used to obtain the ATG start, which was just 12 bp upstream. The completed clone coded for a protein of 2412 amino acids with a calculated molecular mass of 282.6 kDa (Fig. 18.3 A; GenBank accession number AJ238859). Hydrophobicity analysis indicated five domains with a total of 22 putative transmembrane spans. Similarity searches demonstrated that the protein consisted of two large units: an N-terminal half of 1319 amino acids (155 kDa) with topological and amino acid sequence similarities to P-type ATPases and a C-terminal half of 982 amino acids (115 kDa) with a membrane topology identical to the prototypical mammalian adenylyl cyclase and unequivocal sequence similarity in the catalytic loops (Fig.18.3A). Both units were linked by a polypeptide of about 100 amino acids. To exclude the possibility of an artifact cDNA accidentally joined during library preparation, we verified the genomic structure by cloning a 2.5 kb EcoRI fragment from gDNA, which was intronless and fully covered the intradomain region. Thus we unequivocally established the presence of this gene at the gDNA level. The Paramecium cyclase unit contained hydrophobic M1 and M2 regions with six transmembrane spans each, which were followed by C1 a- and C2 a-positioned domains with sequence similarity to the catalytic loops of metazoan 303
18 Second Messenger Systems in Paramecium adenylyl cyclases (Fig. 18.3 A). The C1 a-positioned domain was followed by a hydrophilic stretch of 221 amino acids reminiscent of a C1 b region, a distinct C2 b region was absent. All mammalian adenylyl cyclases have GDCY as a signature sequence in the C1 a domain. In the protozoan cyclase GDCY is present in the C2 a-positioned domain (2276–2279), i. e. close to the C-terminal. Similarly, the motif TYMA is invariant in all C2 a regions of mammalian adenylyl cyclases whereas in the ciliate cyclase it was located in the C1 a-positioned domain (1636–1639), i. e. toward the 5'-end (Fig. 18.3 B). Further, mammalian adenylyl cyclases contain a VKGKG motif in their C2 a catalytic region. The first lysine presumably binds the g-phosphate of ATP [34, 35] and is essential for catalysis. In the Paramecium gene a similar motif, 1793AKGKG1797, was found in the C1 a-positioned domain. Therefore, we concluded that the C1 a- and C2 a-positioned loops of the Paramecium cyclase were inverted compared to mammalian adenylyl cyclases resulting in a (mammalian nomenclature) M1C2 aC1 bM2C1 a architecture. In 1997/1998, the amino acids in the heterodimeric catalytic pocket of metazoan adenylyl and guanylyl cyclases, which are important for substrate specificity, K, D, and Q in adenylyl cyclases, E, C, and R in guanylyl cyclases, have been determined by X-ray crystallography and site-directed mutagenesis [34– 37]. To our surprise, in the Paramecium cyclase these crucial amino acids are guanylyl cyclase-like with a substitution of C by S1698 (Fig. 18.3B). On the other hand, several amino acids, which form a second, non-catalytic pocket in mammalian adenylyl cyclases, are conserved in the protozoan cyclase, i. e. were like those in adenylyl cyclases. These results imply that the cloned gene represents a novel guanylyl cyclase, which by topology and primary structure is most closely related to metazoan adenylyl cyclases. P-type ion transport ATPases were the only proteins with significant similarity to the N-terminal half of this Paramecium nucleotide triphosphate cyclase. The predicted membrane topology was identical to that of P-type ATPases, i. e. two cassettes of two putative transmembrane spans in the N-terminal region (amino acids 63–110 and 349–407) and one set of six transmembrane spans close to the C-terminus (amino acids 1121–1319; Fig. 18.3A) [38, 39]. In the cytosolic domains of the P-type ATPase family, which has more than 150 members, several sequence blocks are conserved [40]. Four of these align to corresponding cytosolic regions of the Paramecium guanylyl cyclase. DKTGT(L/I)T, which is located within the large cytoplasmic loop, is an invariant signature sequence in all P-type ATPases [41]. For ion transport to occur the aspartate must be phosphorylated [42]. In the Paramecium protein DKTGTLT was conserved at the equivalent position (amino acids 461–467). A conserved GDGXND motif present in the hinge region of P-type ATPases was 1025GDSFSD1030 in the Paramecium gene and the (TSND)GE(SNT) block in the transduction domain was retained with a significant E to N change as 190SGNT193 virtually excluding ATPase function [41, 43]. Finally, the ATP-binding domain in nearly all eukaryotic P-type ATPases involves an indispensable aspartate in a block of moderately conserved amino acids [44]. The corresponding amino acid in the Paramecium P-type ATPase-like domain was Glu848. These decisive deviations of the Paramecium 304
18.4 A guanylyl cyclase disguised as an adenylyl cyclase sequence from the ATPase consensus imply that this guanylyl cyclase domain has no ATPase activity and probably has adopted another function. Indeed, ATPases have been cloned from Paramecium, which conform in all respects to the canonical structure [45, 46]. So far, transfection of Paramecium is not possible. Yet, heterologous expression of Paramecium genes is impossible because it uses the universal stops TAA and TAG for glutamine [47]. We changed all 99 TAA and TAG codons to CAA/CAG by site-directed mutagenesis. The “repaired” cDNA was inserted into the bicistronic expression vector pIRES1neo. However, upon transfection of HEK293 cells we obtained no successful transcription, i. e. no functional mRNA. We assumed that the problem may be caused by the high AT-content of the Paramecium guanylyl cyclase gene (66%). Therefore, we resynthesized the ciliate cDNA coding for M1 and M2 using the standard mammalian codon usage [48]. The partially synthetic gene in pIRESneo was transfected into HEK293 cells and yielded a G418-selectable cell population. Cell homogenates had membrane-bound guanylyl cyclase activities of up to 150 pmol 7 mg–1 7 min–1. The KM for MgGTP was 50 µM. Adenylyl cyclase activity was not observed with MgATP as a substrate, yet was detectable using Mn2+ as a metal cofactor. Forskolin did not enhance enzyme activity. These results provided the final proof that the cyclase domain of the cloned gene constituted a guanylyl cyclase in the disguise of what hitherto had been considered to be a prototypical mammalian adenylyl cyclase topology. Up to now, we were unable to express the full-length Paramecium guanylyl cyclase gene. Accordingly, we are unable to speak to the role of the P-type ATPase domain and a potential functional interplay between both domains. We may get a hint by examining the pore and cation-binding helices TM4, 5, and 6. TM4 and 5 differ substantially from the corresponding helices of ion transporters and the ciliate TM6 lacks amino acids, which supposedly are involved in cationbinding [39]. We believe that the protozoan “pore region” does not bind inorganic ions. Perhaps it associates with organic molecules as does another family of P-type ATPases, the putative aminophospholipid transporters. The protozoan P-type ATPase-like domains would then constitute a receptor-like entity reminiscent of the atrial-natriuretic receptor-guanylyl cyclase couple in mammals. However, presently this must be marked as a speculation. Using PCR we detected several additional genes coding for guanylyl cyclases of the same structure (see below). Is the occurrence of this disguised guanylyl cyclase unique to Paramecium or do similar enzymes exist in other organisms? We were not surprised to be able to clone similar genes from the ciliate Tetrahymena because regulation of second messenger cyclic nucleotides has been shown to be essentially identical to Paramecium [25; GenBank accession #AJ238858]. In addition, we searched in databases for the presence of guanylyl cyclases with this architecture in other organisms. In the GenBank we detected a 2.797 bp fragment designated as an adenylyl cyclase from chromosome 11 of Plasmodium falciparum (accession No. U33118). The open reading frame codes for a cluster of six transmembrane spans (M2) and a C2 a-positioned domain. At nucleotide 1966 a substrate-specifying arginine is encoded, which strongly indicates that this gene, contrary to the given definition, 305
18 Second Messenger Systems in Paramecium codes for a guanylyl cyclase with mammalian adenylyl cyclase topology. This deposition did not specify a C1 a-positioned domain. Further, the data, which have been released, yet not edited, from the Malaria Genome Project, show the presence of a guanylyl cyclase sequence on chromosome 13 of Plasmodium falciparum, which exactly matched the Paramecium guanylyl cyclase topology, i. e. it contained a P-type ATPase-like domain fused with a guanylyl cyclase of mammalian adenylyl cyclase topology with the C1 a and C2 a positions inverted [48]. The alignment of the C1 a-positioned region with the Paramecium and Tetrahymena guanylyl cyclases clearly identified GTP as the substrate. The data allow us to discuss evolutionary steps during the development of nucleotide triphosphate cyclases. The fact that cyclases exist with catalytic C1 a and C2 a domains arranged in both ways, supports the hypothesis that initially membrane-anchored monomers formed a homodimeric adenylyl cyclase. This structure is still in existence in several bacteria and may actually represent a common ancestor of mammalian and ciliate nucleotide triphosphate cyclases [31, 49]. Upon gene duplication separate evolution led to the formation of a heterodimeric adenylyl cyclase in eukaryotes before the monomers were fused in a single peptide chain. Evidently, the order of subsequent monomer linkage occurred either way. Only after this event did the protozoan guanylyl cyclase evolve by a few point mutations in the purine binding pocket from an ancestor adenylyl cyclase module. All guanylyl cyclase isoforms detected in Paramecium, Tetrahymena and Plasmodium likely evolved from a common ancestor. On the other hand, the soluble heterodimeric guanylyl cyclases in mammals must have diverged from an ancestral cyclase homodimer before the bulky membrane anchors were added and before heterodimers evolved because the guanylyl cyclases have retained identical amino acids at positions, which determine substrate specificity and at equivalent positions of a non-catalytic pocket (see Fig. 18.2) [35]. In summary then, the protozoan guanylyl cyclases evolved from a membrane-anchored, heterodimeric adenylyl cyclase with inverted catalytic centers whereas mammalian guanylyl cyclases most likely descended from a primordial cytosolic cyclase homodimer. The acquisition by the protozoan guanylyl cyclases of an ATPase-like domain may have occurred prior to or after the change in substrate specificity [48].
18.5 On the way to an adenylyl cyclase with an intrinsic ion conductance
Irrespective of the novelty of the above findings we realized that we missed the original target of our intended studies, cloning of a ciliate adenylyl cyclase with an ion transport capacity [10, 50]. Therefore, we began a renewed search for the 306
18.5 On the way to an adenylyl cyclase with an intrinsic ion conductance protozan enzyme, which converts ATP to cAMP. In keeping with the initial idea of the existence of a mammalian class III adenylyl cyclase with an ancestral ion conductance we started a homology cloning search to detect a cyclase, in which the decisive, substrate defining amino acids located in the C1 a-positioned domain would define ATP as a substrate. Sequencing about 100 PCR clones yielded only another dozen of guanylyl cyclase isozymes, which had a serine, not an aspartate, as one of the three crucial amino acids that specify GTP as a substrate, i. e. these genes coded for guanylyl cyclases. The extremely conserved regions around the GTP-specifying regions (–20 to +2) were nearly identical in all protozoans variants and highly conserved with respect to the corresponding region in mammalian adenylyl cyclases. The anterior 40 amino acids were diverged (21–93% identity) and had approximately the same degree of diversity as that observed in the nine mammalian adenylyl cyclase isoforms [48]. None of the PCR products identified a potential adenylyl cyclase. Another candidate adenylyl cyclase prototype with a potential ion conductance may be the bacterial class II toxins as exemplified by the secreted enzymes from Bordetella pertussis and Bacillus anthracis [28, 30]. Both enzymes share little sequence identity. As mentioned above these cyclase toxins form a membrane pore in the eukaryotic target cell, which is used to deliver the catalytic domain into the cytosol. We developed degenerate oligonucleotide primer pairs using the Paramecium codon usage, which covered the two most conserved sequence stretches. A Paramecium cDNA library was used as a template. Analysis of about 100 PCR products subcloned into pBluescript did not indicate the presence of a class II adenylyl cyclase in Paramecium. We reasoned that class I adenylyl cyclases, which so far have been exclusively detected as soluble bacterial enzymes, were unlikely candidates for a ciliate isoform with an ion conductance and did not warrant a homology cloning approach. Next we centered our attention on bacterial class III adenylyl cyclases, which share to some extent identities in their catalytic domains with mammalian adenylyl cyclases. Some bacterial isoforms are linked or unlinked homodimers and have different membrane anchors. Using this approach we obtained a cDNA clone from Paramecium, which may constitute an adenylyl cyclase monomer. The cDNA codes for a protein, which is partitioned into several domains. Structural predictions indicate the presence of an ion channel, potentially a K+-conductance, an adenylyl cyclase catalytic domain, and a C-terminal tetratricopeptide segment, which could pin the monomer unit to structural proteins underneath the cell membrane. The skewed codon usage of Paramecium precludes heterologous expression, and we will be faced with the same problems alluded to above to remanufacture this gene for expression in heterologous systems, i. e. in HEK293 and Sf9 insect cells and as a cRNA in Xenopus oocytes. The functional proof of the existence of a Paramecium adenylyl cyclase containing an intrinsic, regulatory ion conductance has, therefore, not yet been established.
307
18 Second Messenger Systems in Paramecium
18.6 Downstream of second messengers
After the generation of cyclic nucleotides the predominant intracellular chain of events involves reversible phosphorylation of specific target proteins by respective cylic AMP- and cGMP-activated protein kinases and dephosphorylation by protein phosphatases. Several investigations have shown that regulation of ion channels can proceed by reversible phosphorylation. Thus, Paramecium with a number of behaviorally, electrophysiologically and genetically characterized ion channels may in fact represent a good model system to study. We and others have detected several protein kinases in Paramecium, among those also cAMP and cGMP-dependent varieties [51–55]. We have then focussed our work on protein phosphatases as these enzymes have received much less attention than protein kinases. A large fraction of cytosolic protein phosphatases in eukaryotes has been categorized in four major groups. This classification, which was originally based on enzymological criteria using specific peptide inhibitors, activators and substrate proteins, defined protein phosphatases type 1, 2 A, 2 B (calcineurin) and 2 C. Molecular cloning revealed that protein phosphatases types 1, 2 A, and 2 B belong into one gene family, whereas the type 2C enzymes are defined as a separate gene family. Initially, our work was spurned by the surprising identification of the dinoflagellate polyketal okadaic acid as a highly potent and specific inhibitor of protein phosphatase type 1 and 2 A [56]. We speculated that phosphorylation/dephosphorylation cycles may be decisive in opening and closing the voltage-operated Ca2+-channel. First, we demonstrated by protein biochemistry the presence of a protein phosphatase type 1 in Paramecium and the specific inhibition of the isolated enzyme by okadaic acid [14, 57]. At 1 µM of the inhibitor the protozoan type protein phosphatase was essentially blocked whereas other protein phosphatases present in Paramecium, such as protein phosphatase 2 A and 2 C were not affected. The failure of the toxin to inhibit protein phosphatase 2 A came as a considerable surprise. Protein phosphatases type 1 and 2 A are members of the same gene family and the mammalian enzymes share about 50 % sequence identity. Whatever the explanation for this may be, it was fortuitous because it meant that okadaic acid is a specific inhibitor of protein phosphatase type 1 in Paramecium. This unique property allowed us to prove as an exemplary case the validity of our initial approach to use Paramecium as an in vivo model for screening of novel compounds. We demonstrated that okadaic acid did not affect the normal forward swimming pattern of Paramecium [14]. An increase in the external K+-concentration from 1 to 20 mM triggered backward swimming as expected from a depolarization-activated Ca2+-influx. In the absence of okadaic acid the backward swimming response was transient and >90 % of stimulated cells had returned to the forward swimming mode within 20 s (Fig. 18.4). In the presence of 34 µM okadaic acid, however, >95 % of cells were still swimming backward 2 min after the increase of K+. This dramatic effect required a 15 min preincubation with okadaic acid and, as in metazoans, higher concentrations of okadaic acid compared to those needed to inhibit protein phosphatase type 1 in 308
18.6 Downstream of second messengers
Figure 18.4: Okadaic acid, a specific protein phosphatase type 1 inhibitor, impairs Ca2+channel closure in Paramecium in vivo. Cells were equilibrated in buffer containing 1 mM KCl, 50 µM CaCl2, 10 mM MOPS, pH 7.2. Okadaic acid (final concentration 34 µM in 1% DMSO) was added 15 min prior to stimulation. Swimming behavior was monitored microscopically. Cells were depolarized by addition of KCl (20 mM). Okadaic acid sustained the Ca2+-influx as measured by the prolonged backward swimming response. Addition of 75 µM EGTA abruptly ended the avoiding reaction due to chelation of external calcium (from [14]).
vitro. This may be due to higher local phosphatase concentrations in vivo compared to those present in in vitro assays, as well as to slow and insufficient diffusion of okadaic acid across the thickly protein-coated surface of Paramecium. The effect of okadaic acid most likely was directed at the voltage-operated Ca2+channel because chelation of external Ca2+ by addition of EGTA 2 min after K+depolarization in the presence of okadaic acid caused an immediate termination of backward swimming and resumption of forward movement, i. e. okadaic acid treatment enhanced the influx of Ca2+ and did not affect the motile apparatus of the cilia itself [14]. The effects of okadaic acid in Paramecium are compatible with a model for regulation of the voltage-operated Ca2+-channel in the ciliary membrane, which accounts for all the behavioral, electrophysiological and biochemical observations. With the membrane potential at rest the Ca2+-channel is in a closed state, which is susceptible to activation by depolarization. Depolarization triggers a conformational change of the channel converting it to an open state within 1–2 ms, which leads to a depolarizing Ca2+-influx and drives the reversal of the ciliary beat, hence backward swimming, by an okadaic acid-insensitive mechanism. We propose that the Ca2+-channel is phosphorylated in the closed groundstate, which is inaccessible to dephosphorylation. By contrast, the open state of the channel is dephosphorylated rapidly converting it to an inactivated, dephosphorylated state, which does not permit further Ca2+-influx. This 309
18 Second Messenger Systems in Paramecium reaction sequence is responsible for termination of the backward swimming response after about 10 s in the absence of okadaic acid. In the presence of okadaic acid protein phosphatase 1 is inhibited, dephosphorylation attenuated and a Ca2+-inward current is sustained, which is sufficient to support ciliary reversal unless Ca2+-influx is prevented by chelation of extracellular Ca2+. The reconversion of the inactivated dephosphorylated state to the closed phosphorylated state, which resensitizes the channel to depolarizing stimuli, is a slow process requiring 5–10 min at ambient temperature. These data, though encouraging, also demonstrate that characterization of a molecular target prior to in vivo screening can be a most useful way to define in vivo reaction cascades and support the tendency to approach drug development by reverse pharmacology. The extent of overall conservation among protein phosphatase 2 C enzymes from different phyla, for which sequence data are available, is limited. The considerable sequence disparities may indicate a functional diversity in different phyla. In Paramecium, the protein phosphatase 2C is membrane-associated, a considerable fraction of the enzyme is localized to the cilia, the highly specialized motile organelle of the ciliate [58]. Microsequencing of six tryptic peptides of the purified enzyme revealed a relationship to other PP2C isoforms. The Paramecium PP2C gene was obtained using degenerate oligonucleotide primers. The gene coded for a 33 kDa-protein with 300 amino acids, i. e. it is one of the smallest PP2C isoforms [58]. A C-terminal truncation by about 80 amino acids is responsible for the small size of the Paramecium PP2C compared to isoforms from other organisms. We defined three core regions of high conservation to be present in all PP2C enzymes. These account for about 25% of the protozoan primary sequence. After mutation of nine ciliate Q codons (TAA) to CAA the Paramecium gene was expressed as an active protein in E. coli. The catalytic core region was defined by N- and C-terminal deletions to represent 284 amino acids, i. e. the ciliate PP2C mainly comprises the catalytic core [59]. The data support the notion that the three-dimensional structure of the Paramecium PP2C is identical to that determined with the recombinant human PP2C despite an overall meager amino acid conservation (31%), which is more or less restricted to the three conserved regions mentioned above. In the X-ray structure an unordered loop region is in front of the b9-strand [60]. We introduced a factor Xa protease cutting site (IEGRA) in this domain to define the catalytically active center by generation of proteolytically truncated versions of the PP2C. Surprisingly, this construct, in which the sequence QLII (212–215) was changed to IEGR, thus generating in conjunction with Ala216 a factor Xa cutting site, was inactive. Individual amino acid exchanges at each of the changed positions showed that exclusively the I214G conversion was responsible for the loss of function (Table 18.1). In an I214A variant 62% of wild type activity was retained, and an I214L product was as active as the wild type enzyme. Strikingly, the adjacent isoleucine 215 was not crucial for enzymatic activity. Neither an I215R nor an I215G conversion abolished phosphatase activity. Similar mutation experiments were then carried out with the bovine type 2Ca protein phosphatase, which contains a hydrophobic valine at an equivalent position. A V215G conversion also inactivated the mammalian enzyme (Table 18.1). In this respect our 310
18.7 In vivo screening of bacterial secondary metabolites Table 18.1: A single amino acid replacement (I214G) in a region without defined secondary structure [60] abolishes enzymatic activity of Paramecium protein phosphatase 2C. A corresponding replacement in the bovine type 2Ca isoform (V215G) also leads to an inactive enzyme (adapted from [59]). Enzyme variant
Specific activity (mU/mg)
PtPP2C wild type PtPP2C L213E PtPP2C I214G PtPP2C I214A PtPP2C I214L PtPP2C I215R
15.7 12.3 0 9.7 19.2 11.2
bov. PP2Ca wild type bov. PP2Ca V215G
2.5 0
data clearly go beyond the picture, which has emerged from the crystal structure of PP2C. We have no clue why a mutation that replaces a hydrophobic amino acid at position 214 (Paramecium PP2C numbering) with glycine results in inactivation. The loss of a hydrophobic side chain may alter the stability or a directed movability of this part of the protein, which has no defined tertiary structure in the crystal, such that the active site centers cannot sufficiently stabilize an enzyme/substrate transition state. In contrast to PP1/PP2A, there are currently no specific inhibitors available for protein phosphatase 2C. So far, we have tested a number of synthetic and biological compounds, yet could not unequivocally identify a specific inhibitor of PP2C, which would aid to pinpoint a specific physiological function for this protein phosphatase like okadaic acid did for PP1. We also examined the subcellular localization of PP2C [59]. The major fraction is cytosolic, partly it is associated with the macronucleus, and a minor part is structurally bound to the cilia where it was associated with microtubulus and dynein, i. e. it was clearly targeted to the ciliary motor, implicating a motor component as one of the PP2C substrates. It seems reasonable to consider that the cytoplasmic and nuclear localization of PP2C may be related to processes involving cellular cargo transport [59].
18.7 In vivo screening of bacterial secondary metabolites
We tested more than 100 bacterial secondary metabolites for their toxicity and registered qualitatively and quantitatively by digitized motion analysis changes in the swimming behavior of Paramecium. In part, the assays were marred by 311
18 Second Messenger Systems in Paramecium the poor solubility of the compounds and the need to use organic solvents, mainly dimethylsulfoxide, for bath application. Although we have identified several compounds, e. g. depsichlorin, Tü 3580 and Tü 3586, to distinctly affect evoked swimming patterns of the ciliate, clear-cut actions were not obtained, which would have permitted to deduce a mechanism of action on a single component of the second messenger signal transduction pathway.
Acknowledgments
Supported by the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie. Sequence data for P. falciparum chromosome 13 was obtained from The Sanger Centre website at http://www.sanger.ac.uk/Projects/ P_falciparum/. Sequencing of P. falciparum chromosome 13 was accomplished as part of the Malaria Genome Project with support by The Wellcome Trust.
References
1. Kung, C., Chang, S.-Y., Satow, Y., van Houten, J., and Hansma, H. (1975) The genetic dissection of behavior in Paramecium. Science 188, 898–904. 2. Hinrichsen, R. D. and Schultz, J. E. (1988) Paramecium: A model system for the study of excitable cells. Trends in Neurosciences 11, 27–32. 3. Machemer, H. (1989) Cellular behavior modulated by ions: electrophysiological implications. J. Protozool. 36, 463–487. 4. Machemer, H. and de Peyer, J. (1977) Swimming sensory cells: electrical membrane parameters, receptor properties and motor control of ciliated protozoa. Verh. Dtsch. Zool. Ges. 1977, 86–110. 5. Thiele, J., Honer-Schmid, O., Wahl, J., Kleefeld, G., and Schultz, J. E. (1980) A new method for axenic mass cultivation of Paramecium tetraurelia. J. Protozool. 27, 118– 121. 6. Schönefeld, U., Alfermann, F. W., and Schultz, J. E (1986) Economic mass cultivation of Paramecium tetraurelia on a 200 liter scale. J. Protozool. 33, 222–225. 7. Machemer-Röhnisch, S. and Machemer, H. (1989) A Ca paradox: electric and behavioral responses of Paramecium following changes in external ion concentration. Eur. J. Protist. 25, 45–59. 8. Schultz, J. E., Boheim, G., Gierlich, D., Hanke, W., von Hirschhausen, R., Kleefeld, G., Klumpp, S., Otto, M. K., and Schönefeld, U. (1984) Cyclic nucleotides and calcium in Paramecium: a neurobiological model organism. Hormones and Cell Regulation 8, 99–112.
312
References 9. Schultz, J. E., Grünemund, R., von Hirschhausen, R., and Schönefeld, U. (1984) Ionic regulation of cyclic AMP levels in Paramecium tetraurelia “in vivo”. FEBS Lett. 167, 113–116. 10. Schultz J. E., Pohl, T., and Klumpp, S. (1986) Voltage-gated calcium entry into Paramecium coupled to intraciliary cyclic GMP formation. Nature 322, 271–273. 11. Schultz, J. E., Klumpp, S., Benz, R., Schürhof-Goeters, W., and Schmid, A. (1992) Regulation of adenylyl cyclase from Paramecium by an intrinsic potassium conductance. Science 255, 600–603. 12. Schultz, J. E. and Schade, U. (1989) Ca2+-channel activation and inactivation in Paramecium biochemically measured by cyclic GMP production. J. Membr. Biol. 109, 259–267. 13. Klumpp, S., Jung, G., and Schultz, J. E. (1984) Activation of the calcium/calmodulindependent guanylate cyclase from Paramecium by polypeptide antibiotics and melittin. Biochim. Biophys. Acta 800, 145–151. 14. Klumpp, S., Cohen, P., and Schultz, J. E. (1990) Okadaic acid, an inhibitor of protein phosphatase 1 in Paramecium causes sustained Ca2+-dependent backward swimming in response to depolarizing stimuli. EMBO J. 9, 685–690. 15. Hoff, H., Drautz, H., Fiedler, H.-P., Zähner, H., Schultz, J. E., Keller-Schierlein, W., Philipps, S., Ritzau, M., and Zeeck, A. (1992) Metabolic products of microorganisms. 261. Obscurolides, a novel class of phosphodiesterase inhibitors from Streptomyces. I. Production, isolation, structural elucidation and biological activity of obscurolides A1 to A4. J. Antibiotics 45, 1096–1107. 16. Schulz, S., Yuen, P. S., and Garbers, D. L. (1991) The expanding guanylyl cyclases. Trends Pharmacol. Sci. 12, 116–120. 17. Sunahara, K. S., Dessauer, C. W., and Gilman, A. G. (1996) Complexity and diversity of mammalian adenylyl cyclases. Annu. Rev. Pharmacol. Toxicol. 36, 461–480. 18. Kung, C. and Saimi, Y. (1982) The physiological basis of taxes in Paramecium. Annu. Rev. Physiol. 44, 519–544. 19. Oertel, D., Schein, S. J., and Kung, C. (1977)Separation of membrane currents using a Paramecium mutant. Nature 268, 120–124. 20. Saimi, Y. and Kung, C. (1987) Behavioral genetics of Paramecium. Annu. Rev. Genet. 21, 47–65. 21. Hinrichsen, R. D. and Saimi, Y. (1984) A mutation that alters properties of the Ca2+channel in Paramecium tetraurelia. J. Physiol. Lond. 271, 397–410. 22. Hinrichsen, R. D., Saimi, Y., and Kung, C. (1984) Mutants with altered Ca2+-channel properties in Paramecium tetraurelia: isolation, characterization and genetic analysis. Genetics 108, 545–558. 23. Klumpp, S. and Schultz, J. E. (1982) Characterization of a Ca2+-dependent guanylate cyclase in the excitable ciliary membrane from Paramecium. Eur. J. Biochem. 124, 317–324. 24. Klumpp, S., Kleefeld, G., and Schultz, J. E. (1983) Calcium/Calmodulin regulated guanylate cyclase of the excitable ciliary membrane from Paramecium: Dissociation of calmodulin by La3+: Calmodulin specificity and properties of the reconstituted guanylate cyclase. J. Biol. Chem. 258, 12455–12459. 25. Schultz, J. E. and Schönborn, C. (1994) Cyclic AMP formation in Tetrahymena controlled by a K+-conductance. FEBS Lett. 356, 322–326. 26. Richard, E. A., Hinrichsen, R. D., and Kung, C. (1985) A single gene mutation that affects a potassium conductance and resting membrane potential in Paramecium. J. Neurogenet. 2, 239–252. 27. Krupinski, J., Coussen, F., Bakalyar, H., Tang, W.-J., Feinstein, P. G., Orth, K., Slaughter, C., Reed, R., and Gilman, A. G. (1989) Adenylyl cyclase amino acid sequence: Possible channel- or transporter-like structure. Science 244, 1558–1564. 28. Barzu, O. and Danchin, A. (1994) Adenylyl cyclases: A heterogenous class of ATP-utilizing enzymes. Progr. Nucl. Acid Res. and Mol. Biol. 49, 241–283.
313
18 Second Messenger Systems in Paramecium 29. Brownlie, R. M., Coote, J. G., Parton, R., Schultz, J. E., Rogel, A., and Hanski, E. (1988) Cloning of the adenylate cyclase genetic determinant of Bordetella pertussis and its expression in Escherichia coli and B. pertussis. Microb. Pathol. 4, 335–344. 30. Rogel, A., Schultz, J. E., Brownlie, R. M., Coote, J. G., Parton, R., and Hanski, E. (1989) Bordetella pertussis adenylate cyclase: Purification and characterization of the toxic form of the enzyme. EMBO J. 8, 2755–2760. 31. Tang, W.-J. and Hurley, J. H. (1998) Catalytic mechanism and regulation of mammalian adenylyl cyclases. Mol. Pharmacol. 54, 231–240. 32. Buck, J., Sinclair, M. L., Schapal, L., Cann, M. J., and Levin, L. R. (1999) Cytosolic adenylyl cyclase defines a unique signaling molecule in mammals. Proc. Natl. Acad. Sci. USA 96, 79–84. 33. Hinrichsen, R. D., Fraga, D., and Russell, C. (1995) The regulation of calcium in Paramecium. Adv. Second Messenger Phosphoprotein Res. 30, 311–338. 34. Zhang, G., Liu, Y., Ruoho, A. E., and Hurley, J. H. (1997) Structure of the adenylyl cyclase catalytic core. Nature 386, 247–254. 35. Tesmer, J. J. G, Sunahara, R. K., Gilman, A. G., and Sprang, S. R. (1997) Crystal structure of the catalytic domains of adenylyl cyclase in a complex with GsaGTPgS. Science 278, 1907–1916. 36. Tucker, C. L., Hurley, J. H., Miller, T. R., and Hurley, J. B. (1998) Two amino acid substitutions convert a guanylyl cyclase, RetGC-1, into an adenylyl cyclase. Proc. Natl. Acad. Sci. USA 95, 5993–5997. 37. Sunahara, R. K., Beuve, A., Tesmer, J. J. G., Sprang, S. R., Garbers, D. L., and Gilman, A. G. (1998) Exchange of substrate and inhibitor specificities between adenylyl and guanylyl cyclases. J. Biol. Chem. 273, 16332–16338. 38. Stokes, D. L., Taylor, W. R., and Green, N. M. (1994) Structure, transmembrane topology and helix packing of P-type ion pumps. FEBS Lett. 346, 32–38. 39. Zhang, P. J., Toyoshima, C., Yonekura, K., Green, N. M., and Stokes, D. L. (1998) Structure of the calcium pump from sarcoplasmic reticulum at 8-Å resolution. Nature 392, 835–839. 40. Henikoff, S. and Henikoff, J. G. (1992) Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. USA 89, 10915–10919. 41. Fagan, M. J. and Saier, M. J. Jr (1994) P-type ATPases of eukaryotes and bacteria: sequence analyses and construction of phylogenetic trees. J. Mol. Evol. 38, 57–99. 42. Allen, G. and Green, N. M. (1976) A 31-residue peptide from the active site of the [Ca2+]-transporting adenosine triphosphatase of rabbit sarcoplasmic reticulum. FEBS Lett. 63, 188–191. 43. Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990) Functional consequences of mutations of conserved amino acids in the beta-strand domain of the Ca2+-ATPase of sarcoplasmic reticulum. J. Biol. Chem. 265, 14088–14092. 44. Clarke, D. M., Loo, T. W., and MacLennan, D. H. (1990) Functional consequences of alterations to amino acids located in the nucleotide binding domain of the Ca2+-ATPase of sarcoplasmic reticulum. J. Biol. Chem. 265, 22223–22227. 45. Elwess, N. L. and van Houten, J. L. (1997) Cloning and molecular analysis of the plasma membrane Ca2+-ATPase gene in Paramecium tetraurelia. J. Eukaryot. Microbiol. 44, 250–257. 46. Hauser, K., Pavlovic, N., Kissmehl, R., and Plattner, H. (1998) Molecular characterization of a sarco(endo)plasmic reticulum Ca2+-ATPase gene from Paramecium tetraurelia and localization of its gene product to sub-plasmalemmal calcium stores. Biochem. J. 334, 31–38. 47. Preer, J. R. Jr., Preer, L. B., Rudman, B. M., and Barnett, A. J. (1991) Deviation from the universal code shown by the gene for surface protein 51A in Paramecium. Nature 314, 188–190. 48. Linder, J. U., Engel, P., Reimer, A., Krüger, T., Plattner, H., Schultz, A., and Schultz, J.
314
References
49.
50.
51.
52.
53. 54.
55.
56. 57.
58.
59.
60.
E. (1999) Guanylyl cyclases with the topology of mammalian adenylyl cyclases and an N-terminal P-type ATPase-like domain in Paramecium, Tetrahymena and Plasmodium. EMBO J. 18, 4222–4232. Coudart-Cavalli, M. P., Sismeiro, O., and Danchin, A. (1997) Bifunctional structure of two adenylyl cyclases from the myxobacterium Stigmatella aurantiaca. Biochimie 79, 757–767. Schultz, J. E., Uhl, D., and Klumpp, S. (1987) Characterization of an ionically regulated adenylate cyclase from the excitable ciliary membrane of Paramecium. Biochem. J. 246, 187–192. Schultz, J. E. and Jantzen, H. M. (1980) Cyclic nucleotide-dependent protein kinases from cilia of Paramecium tetraurelia: Partial purification and characterization. FEBS Lett. 116, 75–78. Eistetter, H., Seckler, B., Bryniok, D., and Schultz, J. E. (1983) An electrophoretic analysis of phosphorylation of endogenous proteins of cilia from Paramecium tetraurelia. Eur. J. Cell Biol. 31, 220–226. Hochstrasser, M. and Nelson, D. L. (1989) Cyclic AMP-dependent protein kinase in Paramecium tetraurelia. J. Biol. Chem. 264, 14510–14518. Bonini, N. M., Evans, T. C., Miglietta, L. A. P., and Nelson, D. L. (1991) The regulation of ciliary motility in Paramecium by Ca2+ and cyclic nucleotides. Adv. Second Messenger and Phosphoprotein Res. 23, 227–272. Son, M., Gundersen, R. E., and Nelson, D. L. (1993) A second member of the novel Ca2+-dependent protein kinase family from Paramecium tetraurelia. J. Biol. Chem. 268, 5940–5948. Bialojan, C. and Takai, A. (1988) Inhibitory effect of a marine-sponge toxin, okadaic acid, on protein phosphatases. Specificity and kinetics. Biochem. J. 256, 283–290. Friderich, G., Klumpp, S., Russell, C. B., Hinrichsen, R. D., Kellner, R., and Schultz, J. E. (1992) Purification, characterization and structure of protein phosphatase 1 from the cilia of Paramecium tetraurelia. Eur. J. Biochem. 209, 43–49. Klumpp, S., Hanke, C., Donella-Deana, A., Beyer, A., Kellner, R., and Schultz, J. E. (1994) A membrane-bound protein phosphatase type 2C from Paramecium: Purification, characterization and cloning. J. Biol. Chem. 269, 32774–32780. Grothe, K., Hanke, C., Momayezi, M., Kissmehl, R., Plattner, H., and Schultz, J. E. (1998) Functional characterization and localization of protein phosphatase 2C from Paramecium. J. Biol. Chem. 273, 19167–19172. Das, A. K., Helps, N. R., Cohen, P. T. W., and Bartford, D. (1996) Crystal structure of the protein serine/threonine phosphatase 2C at 2.0 angstrom resolution. EMBO J. 15, 6798–6809.
315
Chemical Synthesis and Structure Elucidation
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites Roderich D. Süßmuth, Jörg Metzger, and Günther Jung*
19.1 Introduction
The introduction of the collaborative research centre 323 opened a fruitful and interesting cooperation between biologists and chemists, wherefrom both sides have greatly benefited. This successful cooperation profited from the expertise of the co-working chemists in the fields of biochemical analytics and chemical synthesis. From the chemist’s view three achievements extraordinarily contributed to the work of the collaborative research centre 323. The first was the introduction of automated parallel peptide synthesis. The development of robotic systems made the synthesis of hundreds of different peptides within several days feasible and enormously accelerated the production of peptide libraries and peptidomimetics. The second important innovation was the electrospray mass spectrometer (ESI-MS) installed in 1989, and the successive development and performance of so called hyphenated analytical techniques, such as HPLC-ESI-MS. The introduction of the electrospray mass spectrometer at that time completely revolutionized the possibilities of analyzing biological samples and extracts. Finally, the installation of the 600 MHz NMR spectrometer enabled the determination of peptide and small protein 3D-structures. However, the impetus by these developments was given not only to the groups directly cooperating within the collaborative research centre 323, but also to new cooperations in related fields e. g. immunology (collaborative research centre 510) and finally the introduction of combinatorial chemistry in Tübingen (BMBF project 03 D 0037). The resonance towards the introduction of these new techniques reached far beyond the cooperating research groups of
* Institut für Organische Chemie, Universität Tübingen, Auf der Morgenstelle 18, D-72076 Tübingen.
319 Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites microbiology in and outside of Tübingen. Due to the early introduction of these new methods in Tübingen, some key reviews [1–4], scientific contributions and books were published, which had impact on the research of related fields. The following gives an overview only of the most significant experimental scientific contributions within the collaborative research centre 323.
19.2 Development of methods
19.2.1 Peptide synthesis Shortly after the invention of the solid phase synthesis by Nobel laureate R. B. Merrifield in 1963, the Institute of Organic Chemistry introduced this new technique at the University of Tübingen (G. Jung, E. Bayer in 1965). Automated peptide synthesizers were introduced in 1989. As one of the first academic groups in Europe G. Jung’s group used these synthesizers for parallel synthesis of hundreds of peptides in mg-quantities. As a consequence of this development and after the introduction of the library concept, partly or fully randomized peptide libraries were synthesized. From the rapidly established key technologies in Tübingen, cooperating research groups in biochemistry, immunology, and sensorics gained considerable profit from protein and substrate mapping experiments, the finding of sequence motifs of MHC ligands and T cell epitopes, and chemosensor developments. Peptides or peptide libraries synthesized in parallel were extensively used in the collaborative research centre 323, e. g. with the groups of V. Braun [5] and F. Götz [6], but also in cooperation with the immunology group of H.-G. Rammensee and many external research groups within Europe. A further important branch was the expertise on the synthesis of unusual peptides, e. g. alamethicin in 1984 and lipopeptides vaccines in 1986. This continuous knowledge was decisive for the success of several projects, e. g. synthesis of heterocyclic backbone modifying amino acids as a part of the microcin B17 structure [7] and the total synthesis of microcin B17 in 1996 [8].
320
19.2 Development of methods
19.2.2 Characterization of peptide libraries with electrospray mass spectrometry The collaborative research centre 323 obtained the first ESI-MS instrument in Europe. This opened new ways for the mass spectrometric characterization of natural products, especially of peptides (Fig. 19.1) and proteins, but also of oligonucleotides, oligosaccharides, and polymers. In the projects of J. Metzger (project C3) and G. Jung (project C2) several hyphenated techniques, e. g. HPLC-MS (Fig. 19.2) and autosampler-ESI-MS were established. The first mass spectrometric investigation of randomized peptide libraries has been performed [9] and further extensive investigations of these libraries confirmed ESI-MS as a method of choice for synthesis control of mixtures containing few to several thousands of components [10]. Moreover, the first LC-ESI-MS characterization of a 9-mer peptide library was performed, revealing side reactions, incomplete side chain deprotection, etc. resulting from peptide synthesis and cleavage from the resin [10, 11]. It was shown that mass spectrometric tandem experiments such as daughter ion scan, parent ion scan and neutral loss scan are suitable methods for the detection of undesired side reactions and byproducts. In further studies mass spectra of the synthetic peptide libraries were compared with theoretically expected
[M+H] + 959
Rel. Int. (%)
100 75
945
50 25
973
1001 987
931
[M+H+56]+ (= tBu)
*
1015 1030 1044 1058 1019
0
Rel. Int. (%)
100
[M+H+56]+ (= tBu) 1015 1001
75 50 25
987
1030 1043 1058 1075 1089
0
Figure 19.1: Identification of tert-butylated peptides with electrospray-tandem mass spectrometry in the octapeptide library SNYTFX1X2X3(X1 = T,I,E,S)(X2 = N,K,Q)(X3 = L,M,I,V). Top: Q1-spectrum of the mixture displaying the [M+H]+-signals of the theoretically expected 48 peptides and the tert-butylated byproducts. Bottom: Selective detection of the tert-butylated byproducts using neutral loss scan (loss of isobutene –56).
321
Rel. intensity [%]
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
Rel. intensity [%]
100 75 50 25 0 100 75 50 25 0 0.0
total ion chromatogram m/z 800-1200
16.1 16.4
ion chromatogram m/z 1041
5.0
15.2 14.5 10.0
15.0
retention time [min]
Figure 19.2: Total ion chromatogram of an HPLC-MS run containing a 48 peptide mixture. The peptides were separated on a nucleosil C-18-column (2 6 100 mm, 5 µm; linear gradient 5–20 % B in 40 min; A 0,1% TFA aq., B 0.1% TFA in acetonitrile). Top: The total ion chromatogram displays the elution of 48 peptides in the mass range of m/z 800–1200. Bottom: The ion chromatogram of m/z 1041 displays the elution of eight peptides with identical nominal mass (isobaric peptides).
mass distributions, using computer-assisted programs, which have been developed especially for this purpose [12]. With the help of these programs, which implement a basis set of proteinogenic amino acids, protecting groups, as well as user-defined amino acid residues, the purity of a peptide library can be qualitatively estimated. These fundamental investigations on peptide library analysis formed the basis for subsequent contributions of other groups, e. g. on combinatorially synthesized small molecule libraries reported by Rebek et al. [13–15].
19.2.3 Automated sequential Edman degradation combined with ESI-MS detection The major focus of the group of G. Jung and his coworkers has been the structure determination of complex peptides. One of the most advanced instrumental sequencing methods is Edman degradation, which is still of essential importance in determining the primary structure of lantibiotics. However, post-translational modifications, e. g. N-terminal blocking and intra-chain bridging via lanthionines as they occur in lantibiotics made the application of standard protocols impossible. In the case of lanthionines, e. g. nisin, bearing a,b-didehydroamino acids sequence abortion occurs due to desamination, thus resisting routine Edman degradation. As a consequence important information about sequences could not be obtained. A second problem arose for lanthionine amino acids, because for these amino acids no PTH-derivative could be detected. To circumvent 322
19.2 Development of methods both problems, a chemical derivatization procedure prior to Edman sequencing was developed [16], which enabled the complete determination of the amino acid sequences of the lantibiotics gallidermin and Pep5. However, analysis of other peptides with unusual amino acids with standard Edman degradation faced problems since the obtained information was not sufficient for compound characterization because no PTH-standards are available for modified amino acids, and their Edman degradation products are often overseen or misinterpreted. The routine identification by HPLC retention times is only standardized for the 19 essential amino acids. As a consequence, the sequencer-ESI-MS-coupling was developed for more sophisticated sequence analysis [17] (Fig. 19.3). The continuous effluent of the HPLC of the Edman se-
Glu, Gly
100
Rel. Int. (%)
UV
75
Gln Thr
Ala
Tyr
Asn Asp Ser
50
Val Trp, His Met Pro Phe Lys Arg Leu Ile
25 0
Trp, His
100
Rel. Int. (%)
TIC
75
Tyr
Gln, Thr Glu, Gly
50
Asn
25
Asp Ser
0
5,0 238
Edman-sequenator
Ala 10,0 475
Phe Lys Leu
Arg Met Val Pro 15,0 713 Time (min)/Scan
Ile
20,0 950
25,0 1187
ESI - mass spectrometer
Figure 19.3: Sequencer-ESI-MS coupling. UV-absorbance chromatogram and total ion chromatogram (TIC) of an optimized gradient of PTH-standard amino acids (despite Cys).
323
Relative Intensity (%)
6,0
200
209,4
[M+NH 4
250
]+
8,0
350
10,0
10,5
dmtu
m/z
9,2
300
385,2
[2M+H] +
12,0
400
402,0
450
[2M+NH 4]+
PTH-Gly
14,0 t / min
Relative Intensity (%)
200
17,1
dmptu
16,0
0
25
50
75
100
18,0
250 m/z
20,0
20,1
300
21,4
350
349,0
22,0
22,5 22,1 23,0
[M+NH 4]+
400
24,0
450
PTH-DidehydroCht 333,0
331,0
[M+H] +
Figure 19.4: Sequencer-ESI-MS coupling: UV chromatogram of an Edman degradation cycle of 3-chloro-b-hydroxy-tyrosine (Cht) and corresponding mass spectra of identified PTH-derivatives.
0
25
50
75
100
0
25
50
75
100
Rel. Int. (%)
324
192,6
[M+H] +
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
19.3 Structure elucidation quencer was directly introduced into the ESI source. The on-line coupling guarantees a complete overview over all products after the degradation cycle. As an additional and most important information, the total ion chromatogram (TIC) contains the molecular mass data of the PTH-amino acids. With this method, peptides bearing unusual amino acids, e. g. 4-hydroxyphenylglycine (Hpg), 3,5dihydroxy-phenylalanine, ornithine, lanthionine, [2,3]-didehydro-asparagine, etc. have been analyzed. This method turned out to be practicable for a great variety of linear and cyclic peptides, e. g. lantibiotics or even microheterogenous peptides. Furthermore, because of the extremely low amounts of sample required this method is suitable for the analysis of natural metabolic peptides of microbial and mammalian origin. The sequencing method based on mass spectrometric detection was of importance in the structure elucidation of CDA, a calcium dependent antibiotic [18], confirming its primary structure, which was not clearly solved from the sequence information obtained by two-dimensional HMBC- and NOESY-NMR spectra. Moreover, this method is generally applied for the structure elucidation of new peptides bearing post-translational modifications. The application of the sequencer-ESI-MS coupling sometimes leads to surprising results which are, however, clearly explicable through the mass spectrometric data. For example, Fig. 19.4 displays the degradation cycle of 3-chloro-bhydroxy-tyrosine (Cht), an amino acid which was found during the structure elucidation of intermediates of the balhimycin biosynthesis [19]. Almost no PTH-derivative of 3-chloro-b-hydroxy-tyrosine was detected. Instead, as a major product PTH-glycine and minor amounts of the 2,3-didehydro-3-chloro-tyrosine PTH-derivative were unequivocally identified. We could explain these results by assuming a retro-aldol reaction under basic conditions (formation of the PTC-peptide) and a dehydration under acid conditions (formation of the ATZ-amino acid and conversion to the PTH-amino acid) during Edman degradation (Fig. 19.5).
19.3 Structure elucidation
Over the past years several dozen structures of a variety of different metabolites from microbial sources have been determined. They comprise molecules of a wide range of compound classes, e. g. antibiotics, siderophores, and signaling molecules. They were either isolated in the course of a targeted screening by the cooperating groups, or represented biosynthetic intermediates produced by mutants genetically generated with the methods of molecular biology. As tools for structure elucidation a variety of analytical methods have been used: electrospray mass spectrometry (ESI-MS) and respective hyphenation techniques, 2-dimensional NMR spectroscopy, GC- and LC-chromatograpy, chiral GC-MS, amino acid analysis, Edman degradation and CD spectroscopy. 325
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
Figure 19.5: Retro-aldol and dehydration reaction for the Edman degradation of 3-chlorob-hydroxy-tyrosine.
19.3.1 Structure elucidation of metabolites of microbial origin The following chapter comprises a selection of some of the structures belonging to different classes of metabolites. Compounds presented had no greater impact on the main research projects which mainly concerned siderophores and lantibiotics. However, they represent interesting structures and some of these compounds developed into research fields of other groups. The following peptidic compounds have been characterized (Fig. 19.6): the rhizocticins, small phosphono-oligopeptides [20], the fengycins lipopeptide antibiotics [21] and the chlorotetains, which all were isolated from Bacillus subtilis strains [22], the lipoglycopeptidic herbicolins from Erwinia herbicola [23], echinoserin from Streptomyces tendae [24], and aborycin, a tricyclic 21-peptide anti326
19.3 Structure elucidation
Figure 19.6:
Peptidic metabolites isolated from different microorganisms.
biotic from Streptomyces griseoflavus [25]. For the latter antibiotic the complete solution 3D-structure was also determined by NMR spectroscopy. In addition, the structures of non-peptidic compounds (Fig. 19.7), e. g. of polyol macrolide antibiotic kanchanamycin from Streptomyces olivaceus [26], the antifungal lactam antibiotic maltophilin produced by Stenotrophomonas maltophilia [27], spirofungin from Streptomyces violaceusniger [28], and boophiline from Boophilus microplus [29] were determined. The early structure determinations were done with the help of FAB-MS, and 1H- and 13C-NMR spectroscopy. However, with the availability of ESI-MS and straightforward 2-dimensional NMR-techniques, the structure elucidation was facilitated and speeded up significantly.
327
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
Figure 19.7:
328
Non-peptidic metabolites isolated from different microorganisms.
19.3 Structure elucidation 19.3.1.1 Siderophores Rhizoferrin from Rhizopus microsporus [30], ferrirhodin from Botrytis cinerea [31], yersiniabactin from Yersinia enterocolitica [32], and staphyloferrin A from Staphylococcus hyicus [33] were isolated and their structure determined. Staphyloferrin A consists of two citric acid moieties and D-ornithin (Fig. 19.8). From dicitrylputrescin contained in rhizoferrin, a number of analogs has been obtained by directed fermentation [34]. The chirality of these analogs and their iron complexation properties have been characterized spectroscopically. The complexation modes of citryl-containing carboxylate siderophores for a number of different transition metal ions have been studied with rhizoferrin as a model compound [35]. From Ralstonia pickettii DSM 6297, the enantiomer of fungal rhizoferrin has been isolated [36, 37]. This opened up a new field of research on the chirality in citric acid-containing carboxylate siderophores and comparative studies on the biosynthetic pathways in fungal and bacterial siderophore producers. New developments in the structure elucidation, synthesis, and function of peptide siderophores have been summarized in a recent review by H. Drechsel and G. Jung [38].
Figure 19.8:
Isolated and characterized siderophores.
329
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites 19.3.1.2 Lantibiotics The lantibiotics represent a class of antimicrobials extensively studied at the University of Tübingen through the cooperation of the groups of Prof. Jung, Prof. Entian, Prof. Götz, and Prof. Zähner. They are polypeptide antibiotics, produced by Gram-positive bacteria which contain intra-chain thioether bridges formed by lanthionine or methyllanthionine. These covalent thioether links impose conformational constraints and stability against protease degradation. Within the collaborative research centre 323 the lantibiotics Pep5 [39], epidermin [40], actagardine [41], gallidermin [42], and SA-FF22 [43], were elucidated as shown in Fig. 19.9.
S
S
Ala Ser Gly Trp Val Ala Abu Leu Abu Leu Glu Ala Gly Abu Val Ile Ala Ala Ala S
S
Actagardin
S
O
S
S
Ile Ala Ala Lys Phe Ile Ala Abu Pro Gly Ala Ala Lys Dhb Gly Ala Phe Asn Ala Tyr Ala
H N S
Epidermin S
S
S
Ile Ala Ala Lys Phe Leu Ala Abu Pro Gly Ala Ala Lys Dhb Gly Ala Phe Asn Ala Tyr Ala
H N S
Gallidermin S
O
S
Ala Gly Pro Ala Ile Arg Ala Ala Val Lys Gln Ala Gln Lys Dhb Leu Lys Ala Dhb Arg Leu Phe Abu Val Ala Ala Lys Gly Lys Asn Lys Ala Lys
O
S
Pep5 Figure 19.9: Sequences and thioether bridging patterns of lantibiotics elucidated by chemical and enzymatic degradation, mass spectrometry, and 2-dimensional NMR spectroscopy.
The structures of epidermin and Pep5 had to be determined by enzymatic degradation and derivatization to smaller, sequencable subfragments. With 2-dimensional NMR-experiments, using a 600 MHz NMR-spectrometer, in addition to ESI-MS and sequencer-MS, 3-dimensional solution structures of duramycin B, C [44], actagardin [45], and gallidermin [46] were determined by distance geometry and restrained-molecular-mechanics calculations (Fig. 19.10). 330
19.3 Structure elucidation
Figure 19.10:
Solution structures of the lantibiotics gallidermin and actagardine.
Knowledge of the lantibiotics structures prompted microbiologists to investigate the biosynthesis of the lantibiotics which is described in Chapter 3.
19.3.1.3 Microcin B17 The structure determination of the 43-peptide antibiotic microcin B17 (Fig. 19.11) from E. coli revealed the structure of the first known gyrase inhibitor (topoisomerase II inhibitor) of peptidic nature. The structure elucidation turned out to be extremely sophisticated because of eight backbone modifications and a nonaglycine sequence. The post-translational backbone modifications [47, 48] of bicyclic oxazole/thiazole-rings protected the major part of peptide from Edman degradation. Fully 13C,15N-labeled microcin B17 had to be prepared for crucial heteronuclear 2D-NMR experiments on the backbone of the heterocyclic polypeptide. In subsequent work, synthesis of the novel oxazole and thiazole amino acids had to be achieved [7], which ended up in the total synthesis of microcin B17 in 1996 [8]. The total synthesis of microcin B17 was performed on solid support using fragment condensations in order to form the nonaglycine sequence, which was not accessible by stepwise successive standard solution phase chemistry. The purity of the crude 43-peptide was about 50 %. Chemical data and antibiotic activity were identical with those of natural microcin B17 [8]. In addition to the natural gyrase inhibitor some analogs have been synthesized for structure-activity studies. The structure of microcin B17 formed the basis to postulate a biosynthesis mechanism for the post-translational modification leading to the oxazole/thiazole rings [47, 48]. The investigation of the microcin biosynthesis is still pursued [49–51].
N
O VGIGGGGGGGGG
S
GGQGG N
N
S
S
G N
N
S SN N
O
O
GGNG N
O
G N
Figure 19.11: Structure of microcin B17 [48].
331
GSHI
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites 19.3.1.4 CDA (Calcium Dependent Antibiotic) Further interesting secondary metabolites are the peptides of the CDA (Calcium Dependent Antibiotic) group, isolated from Streptomyces coelicolor. These cyclic peptides inhibit the growth of Gram-positive bacteria in the presence of Ca2+-ions. The cyclic 11-mer peptide antibiotic (Fig. 19.12) consists of seven proteinogenic and four non-proteinogenic amino acids: D-tryptophan, 4-D-hydroxyphenylglycine, phospho-asparagine and D-methylglutamic acid [18]. Initial attempts to perform Edman degradation failed due to N-terminal blocking with an epoxyhexanoyl group which was identified in a later period of structure elucidation by NMR. Therefore, the peptide was treated with BNPS-skatol, which cleaves the peptide at the N-terminal bond of tryptophan residues. The peptide fragments obtained were further investigated by Edman degradation using the above mentioned sequencer-ESI-MS coupling. The structure elucidation of CDA required the combination of a sophisticated set of analytical methods (HPLC-MS, MS/MS, 2D-NMR) and chemical modification reactions. At present the biosynthesis of CDA is investigated by the groups of Hopwood and Marahiel [52].
Figure 19.12:
Structure of CDA [18].
19.3.1.5 Peptide pheromones of Staphylococcus epidermidis The agr quorum-sensing system (accessory gene regulator) is responsible for the regulation of several virulence factors in staphylococci. Novick et al. [53, 54] characterized the agr-system of S. aureus strains and in accordance with the genomic sequence data, a released peptide factor was presumed to activate the agr-system. Indeed, small amounts of peptides have been isolated from S. aureus strains to perform Edman degradation and mass spectrometry. A conserved cysteine was presumed to form a thiollactone with the carboxy terminus of the peptide. However, insufficient amounts were obtained for a full structure and function analysis. Since we and F. Götz’s group were also unable to isolate sufficient amounts from S. epidermidis strains, we decided to synthesize the proposed molecule. We showed that the cyclic thiollactone DSVc[CASYF] (Fig. 19.13) indeed activated the agr-system in nanomolar concentrations [55]. Non-cyclic peptides were completely inactive, and shortened or elongated cyclic thiollactones 332
19.3 Structure elucidation
Figure 19.13:
Structure of the quorum sensing signal peptide DSVc[CASYF] [55].
(GDSVc[CASYF] and SVc[CASYF]) showed a far lower activity as DSVc[CASYF]. Furthermore, we showed that the S. epidermidis thiollactone peptide DSVc [CASYF] inhibited the activation of the agr-system of S. aureus strains [55]. The replacement of the bridging cysteine, forming the thiollactone bond, by serine (DSVc[SASYF]) or 1,3-diamino propionic acid (DSVc[DprASYF]) had only little influenced the inhibition or activation of the agr-system. Administration of such thiollactone peptides against S. aureus infected mice reduced the infection [56]. It remains to be shown whether such quorumsensing blockers are attractive lead structures for the development of antibiotic drugs aimed at treating staphylococcal infections.
19.3.2 Investigation of biosyntheses 19.3.2.1 Nikkomycin biosynthesis Nikkomycin biosynthesis has been of major interest, because nikkomycins act as competitive inhibitors of chitin synthetases from fungi and insects. Nikkomycins might be attractive drugs and highly diverse lead structures for the agrochemical and pharmaceutical industries. Earlier works on structure elucidation and synthesis of nikkomycine derivatives have been published by W. A. König and H. Hagenmaier [57, 58] in cooperation with H. Zähner’s group. More recently, in cooperation with C. Bormann, biosynthesis of nikkomycins, produced by Streptomyces tendae Tü 901 was investigated by characterization of novel biosynthetic intermediates formed by mutants inactivated in specific nikkomycin-biosynthesis genes. Additionally, mutasynthetically generated compounds from genetically engineered mutants have been characterized. Within this cooperation the function of the nikF nikkomycin biosynthesis gene encoding a P450 monooxygenase was elucidated [59]. Nikkomycins Lx 333
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
Figure 19.14:
Structure of isolated nikkomycins Lx [59] and Bx [60].
and Lz produced by NikF inactivated mutants are analogous structures to nikkomycins X and Z formed by the parent strain, but lack the hydroxy group at the pyridyl residue (Fig. 19.14). The structure of the biosynthetic intermediate formed by a nikO-inactivated mutant was determined as ribofuranosyl-4-formyl4-imidazolone, which represents a novel nucleoside. This finding indicated that the nikO encoded putative enolpyruvyl transferase catalyzes the initial step in the biosynthesis of the nucleoside moieties of nikkomycin. Structure elucidation of two biosynthetic intermediates isolated from the culture filtrate of mutants blocked in the biosynthesis of the nucleoside moieties is in progress. The structure of a novel intermediate produced by a nikK-mutant that is unable to introduce the amino group to the nucleoside moiety was determined as 4-formyl-4imidazolin-2-one. This is the base which is incorporated to yield nikkomycins containing this base. Feeding benzoic acid to a mutant deficient in the nikC gene which encodes lysine-2-aminotransferase catalyzing the initial step in the biosynthesis of the peptidyl moiety led to production of nikkomycin Bx and Bz [60]. These nikkomycins, which are the most potent compounds among known nikkomycins revealed good activity against Candida albicans.
19.3.2.2 Epidermin biosynthesis The characterization and structure elucidation of lantibiotics had an impact on the microbiologists within and outside the collaborative research centre 323. The gene cluster for the biosynthesis of epidermin, epiA-D, epiP, and epiQ was sequenced. The prepeptide EpiA, consisting of 52 amino acids, is ribosomally 334
19.3 Structure elucidation synthesized and transformed into the biologically active 22mer peptide by several post-translational modifications, including oxidative decarboxylation, dehydration, formation of thioether bridges, and cleavage of the leader peptide. In cooperation with T. Kupke and F. Götz the enzyme EpiD has been isolated and characterized. With EpiA substrates it has been shown that EpiD catalyzes the oxidative decarboxylation (–H2, –CO2) of C-terminal cysteine residues to form S-[(Z)-2-aminovinyl]-D-cysteine [61]. This hitherto unknown enzymatic oxidative decarboxylation reaction of EpiA substrates with EpiD has been detected by electrospray mass spectrometry. Subsequently, the substrate specificity of this reaction has been determined with precursor peptides of mutants and chemically synthesized 7mer peptides and peptide libraries [62]. The substrate specificity of EpiD was clearly shown by neutral loss scan experiments using randomized peptide libraries. This was the first application of neutral loss scan to identify products of an enzymatic reaction. Moreover, kinetic studies of the conversion reaction have been performed with electrospray mass spectrometry. The enzymatic reaction of EpiD with 13C-labeled peptide substrate has been followed by HSQC-NMR spectroscopy [63] (Fig. 19.15).
O H H C N C COOH 13
H C H O H
SH
H
C N C 13
HS
C H
675 min 375 min 195 min 75 min 0 min
δ / ppm
120
100
80
60
40
20
Figure 19.15: Reaction of the 13C-labeled substrate KKSFNSYTC with the enzyme EpiD in projections along the F1 (13C) axis of the HSQC spectra recorded during the course of the reaction. The product signal at d(13C) = 120.5 ppm increases in intensity with time [63].
335
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites 19.3.2.3 Biosynthesis of glycopeptides: balhimycin Paying attention to the dramatically increasing number of antibiotic resistant microorganisms, vancomycin and the related family of glycopeptides constitute antibiotics of the last resort, especially in the case of infections caused by methicillin-resistant staphylococci. The cyclic glycopeptides are also of interest as model receptors [64] for the investigation of ligand-receptor interactions, and they are widely used as chiral selectors in chromatography [65]. Considering the stereochemistry of the glycopeptides, the total synthesis of these molecules is one of the most challenging enterprises for synthetically working chemists [66, 67]. The target molecule of the group of W. Wohlleben and S. Pelzer is balhimycin, a vancomycin-type glycopeptide antibiotic, produced by Amycolatopsis mediterranei, from which the biosynthesis gene cluster was sequenced [68]. After identification of the biosynthesis gene cluster, in 1998 the first gene-dis-
Figure 19.16: Structures of the linear biosynthesis intermediates of balhimycin, a vancomycin type tricyclic glyco-peptide antibiotic [19].
336
19.4 Summary ruption mutants were cloned in the oxygenase genes oxyA and oxyB and in the glycosyltransferase gene bgtfB. In contrast to the glycosyltransferase mutant, the oxygenase mutants showed no antibiotic activity. The culture filtrates from both types of mutants were investigated, and according to the characteristic isotopic pattern of the two-fold chlorination of the detected compounds they were assigned to the expected biosynthesis intermediates. The compounds were isolated and investigated with ESI-MS, Edman degradation, amino acid analysis, and 2-dimensional NMR experiments [19]. The compounds produced by the oxygenase mutants revealed the first known linear peptide intermediates of glycopeptides (Fig. 19.16). They give a first insight into the biosynthesis pathway of the glycopeptide antibiotics [19, 68] (Fig. 19.17). The compounds produced by the bgtfB-mutant showed a complete lack of glycosylation. In-frame mutations of each single gene, of the oxygenases OxyA/B/C and the glycosyltransferases BgtfA/B/C are currently generated. The culture filtrates from the mutant strains are investigated with HPLC-ESI-MS for the presence of the biosynthesis intermediates, and the structures of these metabolites are determined after purification. The goal of the ongoing cooperation is to inactivate the halogenase gene bhaA to unravel the so far unknown function of a haloperoxidase and to demonstrate its substrate specificities.
19.4 Summary
During the past decade of the collaborative research centre 323, efficient analytical and synthetic methods were developed and applied, e. g. ESI-MS and coupling techniques, multiple peptide synthesis, 2-dimensional NMR spectroscopy, and molecular modeling. With these new methods it was possible to perform a high level of biological and biochemical research. Frequently, cooperation with biologically working groups showed that an expertise in both analytical chemistry and synthesis was extremely valuable or even decisive for the success of a project. Either the synthetic and analytical work supported the value and meaning of biological results or paved the way for the biological research.This fruitful constellation, the close cooperation between microbiology and chemistry resulted in numerous journal publications. With the installation of the Fourier-transform-ion-cyclotron resonance mass spectrometer in 1998 it has been once more possible to establish a modern analytical technique at the Institute of Organic Chemistry [69]. The installation of these key technologies is extremely important to keep pace with the rapidly growing requirements of biologists for powerful analytical methods.
337
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
O HO
polyketidsynthase “?” aminotransferase “?”
NH2
Cl
HO
OH
Cl
HO
OH HO
OH
OH O
HO
3,5-dihydroxyphenylglycine
O H N
N H
O H N
N H
O
O
NH2
N H O
H2N O
SP-969
OH
peptidesynthetases “?“ Cl
HO
OH
Cl
HO
OH O
OH O
H N
HO
O H N
N H
N H
O
HO
O H N
O
O
H2N O
OH OH
oxygenases oxyA/B/C
NH2
N H
SP-1134
-6H O HO
R1 = -H / -CH3 R2 = -H / -OH
Cl
OH O
R2
Cl O
O HN HO2C
O
H N
N H
O
NH
N H
O
H2N
R1
O
H N
N H
O
glycosyltransferases bgtfA/B/C
HO
OH
HD-1112/1126 HD-1128/1142
OH
HO HO
HO O
O
CH2OH Cl
O
H2N
O
O
O H3C CH 3 O
O O HN HO2C
OH
Cl O
H N
N H
O
N H
O
H N O
H2N
N H
CH3 NH
O HO
OH
OH
Balhimycin
Figure 19.17: Proposed biosynthesis pathway of glycopeptides with the example of balhimycin.
338
References
References
1. Jung, G. (1991) Lantibiotics, ribosomally synthesized biologically active polypeptides containing sulfide bridges and a,b-didehydroamino acids. Angew. Chem. 103, 1067– 1084; Angew. Chem. Int. Ed. Engl. 30, 1051–1068. 2. Jung, G. and Sahl, H.-G. (1991) (eds.) Nisin und novel lantibiotics, Escom, Leiden. 3. Jung, G. and Beck-Sickinger, A. G. (1992) Multiple peptide synthesis methods and their applications. Angew. Chem. 104, 375–391; Angew. Chem. Int. Ed. Engl. 31, 367– 383. 4. Früchtel, J. and Jung, G. (1996) Organic chemistry on solid supports. Angew. Chem. 108, 19–46; Angew. Chem. Int. Ed. Engl. 35, 17–42. 5. Killmann, H., Videnov, G., Jung, G., Schwarz, H., and Braun, V. (1995) Identification of receptor binding sites by competitive peptide mapping: phages T1, T5, and 0/80 and colicin M bind to the gating loop of FhuA. J. Bacteriol. 177, 694–698. 6. Peschel, A., Otto, M., Jack, R. W., Kalbacher, H., Jung, G., and Götz, F. (1999) Inactivation of the dlt operon in Staphylococcus aureus confers sensitivity to defensins, protegrins, and other antimicrobial peptides. J. Biol. Chem. 274, 8405–8410. 7. Videnov, G., Kaiser, D., Kempter, C., and Jung, G. (1996) Synthesis of naturally occurring, conformationally restricted oxazole and thiazole containing di- and tripeptide mimetics. Angew. Chem. 108, 1604–1607; Angew. Chem. Int. Ed. Engl. 35, 1503– 1506. 8. Videnov, G., Kaiser, D., Kempter, C., Brooks, M., and Jung, G. (1996) Synthesis of the DNA gyrase inhibitor Microcin B17, a 43-peptide antibiotic with eight heteroaromatic rings in the backbone. Angew. Chem. 108, 1607–1609; Angew. Chem. Int. Ed. Engl. 35, 1506–1508. 9. Stevanovic, S., Wiesmüller, K.-H., Metzger, J., Beck-Sickinger, A. G., and Jung, G. (1993) Natural and synthetic peptide pools: characterization by sequencing and electrospray mass spectrometry. Bioorg. Med. Chem. Lett. 3, 431–436. 10. Metzger, J. W., Wiesmüller, K.-H., Gnau, V., Brünjes, J., and Jung, G. (1993) Ionspraymass spectrometry and high performance liquid chromatography-mass spectrometry of synthetic peptide libraries. Angew. Chem. 105, 901–903; Angew. Chem. Int. Ed. Engl. 32, 894–896. 11. Metzger, J. W., Kempter, C., Wiesmüller, K.-H., and Jung, G. (1994) Electrospray mass spectrometry and tandem mass spectrometry of multi-component peptide mixtures: determination of composition and purity. Anal. Biochem. 219, 261–277. 12. Kienle, S., Wiesmüller, K.-H., Brünjes, J., Metzger, J. W., and Jung, G. (1997) MSPep: A computer program for the interpretation of mass spectra of peptide libraries. Fresenius J. Anal. Chem. 359, 10–14. 13. Carell, T., Wintner, E. A., Sutherland, A. J., Rebek Jr., J., Dunayevskiy, Y. M., and Vouros, P. (1995) New promise in combinatorial chemistry: synthesis, characterization, and screening of small-molecule libraries in solution. Chem. Biol. 2, 171–183. 14. Dunayevskiy, Y., Vouros, P., Carell, T., Wintner, E. A., and Rebek Jr., J. (1995) Characterization of the complexity of small-molecule libraries of electrospray ionization mass spectrometry. Anal. Chem. 67, 2906–2915. 15. Dunayevskiy,Y. M.,Vouros, P., Wintner, E. A., Shipps, G. W., Carell, T., and Rebek Jr., J., (1996) Application of capillary electrophoresis-electrospray ionization mass spectometry in the determination of molecular diversity. Proc. Natl. Acad. Sci. USA 93, 6152–6157. 16. Meyer, H. E., Heber, M., Eisermann, B., Korte, H., Metzger, J. W., and Jung, G. (1994) Sequence analysis of lantibiotics: chemical derivatization procedures allow a fast access to complete Edman degradation. Anal. Biochem. 223, 185–190.
339
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites 17. Walk, T. B., Süßmuth, R., Kempter, C., Gnau, V., Jack, R. W., and Jung, G. (1999) Identification of unusual amino acids in peptides using automated sequential Edman degradation by electrospray-ionization mass spectrometry. Biopolymers 49, 329–340. 18. Kempter, C., Kaiser, D., Haag, S., Nicholson, G., Gnau, V., Walk, T., Gierling, K.-H., Decker, H., Zähner, H., Jung, G., and Metzger, J. W. (1997) CDA: Calcium-dependent peptide antibiotics from Streptomyces coelicolor A3(2), containing unusual residues. Angew. Chem. 109, 510–513; Angew. Chem. Int. Ed. Engl. 36, 498–501. 19. Süßmuth, R., Pelzer, S., Nicholson, G., Walk, T., Wohlleben, W., and Jung, G. (1999) New advances in the biosynthesis of glycopeptide antibiotics of the vancomycin type from Amycolatopsis mediterranei. Angew. Chem. 111, 2096–2099; Angew. Chem. Int. Ed. Engl. 38, 1976–1979. 20. Rapp, C., Jung, G., Kugler, M., and Loeffler, W. (1988), Rhizocticins – new phosphono-oligopeptides with antifungal activity. Liebigs Ann. Chem., 655–661. 21. Vanittanakom, N., Loeffler, W., Koch, U., and Jung, G. (1986), Fengycin – a novel antifungal lipopeptide antibiotic produced by Bacillus subtilis F-29–3. J. Antibiot. 39, 888–901. 22. Rapp, C., Jung, G., Katzer, W., and Loeffler, W. (1988) Chlorotetain from Bacillus subtilis, a antifungal dipeptide with an unusual chlorinated amino acid, Angew. Chem. 100, 1801–1802; Angew. Chem. Int. Ed. Engl. 27, 1733–1734. 23. Aydin, M., Lucht, N., König, W. A., Lupp, R., Jung, G., and Winkelmann, G. (1985) Structure elucidation of the peptide antibiotics herbicolin A and B. Liebigs Ann. Chem., 2285–2300. 24. Blum, S., Fiedler, H.–P., Groth, I., Kempter, C., Stephan, H., Nicholson, G., Metzger, J. W., and Jung, G. (1995) Biosynthetic capacities of Actinomycetes. Echinoserin, a new member of the quinoxaline group, produced by Streptomyces tendae. J. Antibiot. 48, 619–625. 25. Potterat, O., Stephan, H., Metzger, J. W., Gnau, V., Zähner, H., and Jung, G. (1994), Aborycin – a tricyclic 21-peptide antibiotic isolated from Streptomyces griseoflaveus. Liebigs Ann. Chem., 741–743. 26. Stephan, H., Kempter, C., Metzger, J. W., Jung, G., Potterat, O., Pfefferle, C., and Fiedler, H.-P. (1996) Kanchanamycins, new polyol macrolide antibiotics produced by Streptomyces olivaceus Tü 4018, II. Structure elucidation. J. Antibiotics 49, 765–769. 27. Jakobi, M., Winkelmann, G., Kaiser, D., Kempter, C., Jung, G., Berg, G., and Bahl, H. (1996) Maltophilin: A new antifungal compound produced by Stenotrophomonas maltophilia R3089. J. Antibiotics 49, 1101–1104. 28. Höltzel, A., Kempter, C., Metzger, J. W., Jung, G., Groth, I., Fritz, T., and Fiedler, H.P. (1998) Spirofungin, a new antigungal antibiotic from Streptomyces violaceusniger Tü 4113. J. Antibiotics 51, 699–707. 29. Potterat, O., Hostettmann, K., Höltzel, A., Jung, G., Dichl, P.-A., and Petrini, O. (1997) Boophiline, an antimicrobial sterol amide from the cattle tick Boophilus microplus. Helv. Chim. Acta 80, 2066–2072. 30. Drechsel, H., Metzger, J., Freund, S., Jung, G., Boelaert, J. R., and Winkelmann, G. (1991). Rhizoferrin – a novel siderophore from the fungus Rhizopus microsporus var. rhizopodiformis. BioMetals 4, 238–243. 31. Konetschny-Rapp, S., Jung, G., Huschka, H.-G., and Winkelmann, G. (1988) Isolation and identification of the principal siderophore of the plant pathogenic fungus Botrytis cinerea. BioMetals 1, 90–98. 32. Drechsel, H., Stephan, H., Lotz, R., Haag, H., Zähner, H., Handtke, K., and Jung, G. (1995). Structure elucidation of yersiniabactin, a siderophore from highly virulent yersinia strains. Liebigs Ann. Chem., 1727–1733. 33. Konetschny-Rapp, S., Jung, G., Meiwes, J., and Zähner, H. (1990) Staphyloferrin A: A structurally new siderophore from Staphylococci. Eur. J. Biochem. 191, 65–74. 34. Tschierske, M., Drechsel, H., Jung, G., and Zähner, H. (1996) Production of rhizofer-
340
References
35.
36. 37.
38. 39.
40.
41.
42.
43.
44. 45.
46.
47.
48.
49.
50. 51.
52.
rin and new analogues obtained by directed fermentation. Appl. Microbiol. Biotechnol. 45, 664–670. Carrano, C. J., Drechsel, H., Kaiser, D., Jung, G., Matzanke, B., Winkelmann, G., Rochelle, N., and Albrecht-Gary, A. M. (1996) Coordination chemistry of the carboxylate type siderophore rhizoferrin: The iron(III) complex and its metal analogs. Inorg. Chem. 35, 6429–6436. Drechsel, H., Jung, G., and Winkelmann, G. (1992) Stereochemical characterization of rhizoferrin and identification of its dehydration products. BioMetals 5, 141–148. Münzinger, M., Taraz, K., Budzikiewicz, H., Drechsel, H., Heymann, P., Winkelmann, G., and Meyer, J.-M. (1999) S,S-Rhizoferrin (enantio-rhizoferrin) – a siderophore of Ralstonia (Pseudomonas) pickettii DSM 6297 – the optical antipode of R,R-rhizoferrin isolated from fungi. BioMetals 12, 189–193. Drechsel, H. and Jung, G. (1998) Peptide siderophores. J. Peptide Science 4, 147–181. R. Kellner, G. Jung, M. Josten, C. Kaletta, K.-D. Entian, and H.-G. Sahl (1989) Pep5: Structure elucidation of a large lantibiotic. Angew. Chem. 101, 618–621; Angew. Chem. Int. Ed. Engl. 28, 616–619. Allgaier, H., Jung, G., Werner, R. G., Schneider, U., and Zähner, H. (1985), Structure elucidation of epidermin, a ribosomally synthetized, heterodet tetracyclic polypeptide antibiotic. Angew. Chem. 97, 1052–1054; Angew. Chem. Int. Ed. Engl. 24, 1051–1053. Zimmermann, N., Metzger, J. W., and Jung, G. (1995), The tetracyclic lantibiotic actagardine, 1H-NMR and 13C-NMR assignments and revised primary structure. Eur. J. Biochem. 228, 786–797. Kellner, R., Jung, G., Hörner, T., Zähner, H., Schnell, N., Entian, K.-D., and Götz, F. (1988) Gallidermin, a new lanthionine-containing polypeptide antibiotic. Eur. J. Biochem. 177, 53–59. Jack, R. W., Carne, A., Metzger, J., Stevanovic, S., Sahl, H.-G., Jung, G., and Tagg, J. (1994), Elucidation of the structure of SA-FF22, a lanthionine-containing antibacterial peptide produced by Streptococcus pyogenes strain FF22. Eur. J. Biochem. 220, 455– 462. Zimmermann, N., Freund, S., Fredenhagen, A., and Jung, G. (1993) Solution structures of the lantibiotics duramycin B and C. Eur. J. Biochem. 216, 419–428. Zimmermann, N. and Jung, G. (1997) The three-dimensional solution structure of the lantibiotic murein-biosynthesis inhibitor actagardine determined by NMR. Eur. J. Biochem. 246, 809–819. Freund, S., Jung, G., Gutbrod, O., Folkers, G., Gibbons, W. A., Allgaier, H., and Werner, R. (1991) The solution structure of the lantibiotic gallidermin. Biopolymers 31, 803–811. Bayer, A., Freund, S., Nicholson, G., and Jung, G. (1993) Posttranslational backbone modification in the ribosomal biosynthesis of the glycine-rich antibiotic microcin B17. Angew. Chem. 105, 1410–1413; Angew. Chem. Int. Ed. Engl. 32, 1336–1339. Bayer, A., Freund, S., and Jung, G. (1995) Post-translational heterocyclic backbone modifications in the 43-peptide antibiotic microcin B17, structure elucidation and NMR study of a 13C, 15N-labelled gyrase inhibitor. Eur. J. Biochem. 234, 414–426. Genilloud, O., Moreno, F., and Kolter, R. (1989) DNA sequence, products, and transcriptional pattern of the genes involved in production of the DNA replication inhibitor microcin B17. J. Bacteriol. 171, 1126–1135. Li, Y.-M., Milne, J. C., Madison, L. L., Kolter, R., and Walsh, C. T. (1996) From peptide precursors to oxazole-thiazole containing peptide antibiotics. Science 274, 1188–1193. Roy, R. S., Allen, O., and Walsh, C. T. (1999) Expressed protein ligation to probe regiospecifity of heterocyclization in the peptide antibiotic microcin B17. Chem. Biol. 6, 789–799. Chong, P. P., Podmore, S. M., Kieser, H. M., Redenbach, M., Turgay, K., Marahiel, M., Hopwood, D. A., and Smith, C. P. (1998) CDA: Physical identification of a chromoso-
341
19 Structure Elucidation and Chemical Synthesis of Microbial Metabolites
53. 54. 55. 56.
57. 58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
mal locus encoding biosynthetic genes for the lipopeptide calcium-dependent antibiotic (CDA) of Streptomyces coelicolor A3(2). Microbiology 144, 193–199. Ji, G., Beavis, R., and Novick, R. P. (1995) Cell density control of staphylococcal virulence mediated by an octapeptide pheromone. Proc. Natl. Acad. Sci. 92, 12055–12059. Ji, G., Beavis, R., and Novick, R. P. (1997) Bacterial interference caused by autoinducing peptide variants. Science 276, 2027–2030. Otto, M., Süßmuth, R., Jung, G., and Götz, F. (1998) Structure of the pheromone peptide of the Staphylococcus epidermidis agr system. FEBS Letters 424, 89–94. Mayville, P., Ji, G., Beavis, R., Yang, H., Goger, M., Novick, R. P., and Muir, T. W. (1999) Structure-activity analysis of synthetic autoinducing thiolactone peptides from Staphylococcus aureus responsible for virulence. Proc. Natl. Acad. Sci. 96, 1218–1223. Hagenmaier, H., Keckeisen, A., Zähner, H., and König, W. A. (1979) Structure elucidation of the nucleoside antibiotic nikkomycin X, Liebigs Ann. Chem. 1494–1502. Zimmermann, G., Hass, W., Faasch, H., Schmalle, H., and König, W. A. (1985) Synthesis of pure stereoisomers of the N-terminal amino acid of Nikkomycin B. Liebigs Ann. Chem. 2165–2177. Bormann, C., Lauer, B., Kalmanczhelyi, A., Süssmuth, R., and Jung, G. (1999) Novel nikkomycins Lx and Lz produced by genetically engineered Streptomyces tendae Tu901. J. Antibiotics 52, 582–585. Bormann, C., Kalmanczhelyi, A., Süssmuth, R., and Jung, G. (1999) Production of nikkomycins Bx and Bz by mutasynthesis with genetically engineered Streptomyces tendae Tu901. J. Antibiotics 52, 102–108. Kupke, T., Kempter, C., Gnau, V., Jung, G., and Götz, F. (1994) Mass spectrometric analysis of a novel enzymatic reaction: Oxidative decarboxylation of the lantibiotic precursor peptide EpiA catalyzed by the flavoprotein EpiD. J. Biol. Chem. 269, 5653– 5659. Kupke, T., Kempter, C., Jung, G., and Götz, F. (1995) Oxidative decarboxylation of peptides catalyzed by flavoprotein Epi D: Determination of substrate specificity using peptide libraries and neutral loss mass spectrometry. J. Biol. Chem. 270, 11282–11289. Kempter, C., Kupke, T., Kaiser, D., Metzger, J. W., and Jung, G. (1996) Thioenols from peptidyl-cysteines: Oxidative decarboxylation of a 13C-labeled substrate. Angew. Chem. 108, 2235–2238; Angew. Chem. Int. Ed. Engl. 35, 2104–2107. Chu, Y.-H., Dunajevskiy, Y.M., Kirby, D. P., Vouros, P., and Karger, B. L. (1996) Affinity capillary electrophoresis-mass spectrometry for screening combinatorial libraries. J. Am. Chem. Soc. 118, 7827–7834. Armstrong, D. W., Rundlett, K. L., and Chen, J. R. (1994) Evaluation of the macrocyclic antibiotic vancomycin as a chiral selector for capillary electrophoresis. Chirality 6, 496–509. Evans, D. A., Wood, M. R., Trotter, B. W., Richardson, T. I., Barrow, J. C., and Katz, J. L. (1998) Total syntheses of vancomycin and eremomycin aglycons. Angew. Chem. Int. Ed. Engl. 37, 2700–2704. Nicolaou, K. C., Boddy, C. N. C., Bräse, S., and Winssinger, N. (1999) Chemistry, biology, and medicine of the glycopeptide antibiotics. Angew. Chem. Int. Ed. Engl. 38, 2097–2152. Pelzer, S., Süßmuth, R., Heckmann, D., Recktenwald, J., Huber, P., Jung, G., and Wohlleben, W. (1999) Identification and analysis of the balhimycin biosynthetic gene cluster and its use for manipulating glycopeptide biosynthesis in Amycolatopsis mediterranei DSM5908. Antimicrob. Agents Chemother. 43, 1565–1573. Walk, T. B., Trautwein, A. W., Richter, H., and Jung, G. (1999), ESI Fourier transform ion cyclotron resonance mass spectrometry (ESI-FT-ICR-MS): A rapid high-resolution analytical method for combinatorial compound libraries. Angew. Chem. 111, 1877– 1880. Angew. Chem. Int. Ed. Engl. 38, 1763–1765.
342
Documentation
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
20 Documentation of the Collaborative Research Centre 323
20.1 List of institutes involved
Lehrstuhl für Mikrobiologie/Biotechnologie, Universität Tübingen Lehrstuhl für Mikrobiologie/Membranphysiologie, Universität Tübingen Lehrstuhl für Mikrobielle Genetik, Universität Tübingen Lehrstuhl für Physiologische Chemie/Biochemie, Universität Tübingen Lehrstuhl für Pharmazeutische Chemie, Universität Tübingen Lehrstuhl für Pharmazeutische Biologie, Universität Tübingen Medizinisch-Naturwissenschaftliches Forschungszentrum, Universität Tübingen Max-Planck-Institut für Entwicklungsbiologie, Abt. Biochemie, Tübingen Max Planck-Institut für Biologie, Abt. Infektionsbiologie, Tübingen Institut für Organische Chemie, Universität Tübingen Lehrstuhl für Hydrochemie und Hydrobiologie, Universität Stuttgart Robert Koch-Institut des Bundesgesundheitsamtes, Wernigerode
20.2 List of supported project areas
Project area A TP A1 TP A2
Mikrobieller Sekundärstoffwechsel Prof. Ing. Hans Zähner, Mikrobiologie/Biotechnologie Fermentation, Aufarbeitung und Analytik niedermolekularer Metabolite Fermentationstechnik, Naturstoffisolierung, Naturstoffanalytik
1986–1996 1986–1988 1989–1990 345
Microbial Fundamentals of Biotechnology. DFG, Deutsche Forschungsgemeinschaft Copyright # 2001 WILEY-VCH Verlag GmbH, Weinheim ISBN: 3-527-30615-3
20 Documentation of the Collaborative Research Centre 323 Grundlagen der Produktionsoptimierung und Analytik mikrobieller Sekundärmetabolite 1991–1993 Fermentation und Produktionsoptimierung mikrobieller Naturstoffe (Mikrobiologie/Biotechnologie) 1994–1999 Prof. Dr. Hans-Peter Fiedler, Mikrobiologie/Biotechnologie TP A3
Biotechnologische Produkte von Peptiden in Hefe PD Dr. Karl-Dieter Entian, Biochemie
TP A4
Genetik und Biochemie der Nikkomycin-Biosynthese in Streptomyces tendae Genetische und biochemische Charakterisierung eines antifungisch wirksamen Proteins aus Streptomyces tendae (Mikrobielle Genetik) Biosynthese des Antibiotikums Nikkomycin und extrazelluläre Proteine bei Streptomyceten Dr. Christiane Bormann, Mikrobiologie/Biotechnologie
TP A5
Hopanoid-Cyclasen Synthese und Funktion von Hopanoiden Prof. Dr. Karl Poralla, Mikrobiologie/Biotechnologie
TP A10 Molekulargenetische und biochemische Analyse der Sekundärmetabolit-Biosynthese in Aktinomyceten Prof. Dr. Wolfgang Wohlleben, Mikrobiologie/ Biotechnologie TP A11 Molekularbiologische Untersuchungen zur Biosynthese des Cytostatikums Landomycin A, gebildet von Streptomyces cyanogenus (DSM 5087) – Untersuchungen zum Aufbau der Zuckerseitenkette – Untersuchungen zur Biosynthese der Desoxyzucker L-Rhodinose und D-Olivose Dr. Andreas Bechthold, Pharmazeutische Biologie TP A12 Kontrolle der autolytischen Enzyme in Escherichia coli Prof. Dr. Joachim-Volker Höltje, MPI für Entwicklungsbiologie, Abt. Biochemie
1986–1988
1991–1993
1997–1999 1991–1996 1997–1999
1997–1999
1997–1999 1997–1999
Project area B TP B1
346
Metabolische Wechselwirkungen zwischen Pround Eukaryonten 1986–1996 Eisen als Umweltsignal Prof. Dr. Volkmar Braun und PD Dr. Klaus Hantke, Mikrobiologie/Membranphysiologie Regulierte Transport- und Signaltransfer-Kanäle in der äußeren Membran Gram-negativer Bakterien 1997–1999 Prof. Dr. Volkmar Braun, Mikrobiologie/Membranphysiologie
20.2 List of supported project areas TP B2
TP B3
TP B4
TP B5
TP B6
Regulation der interzellulären Kommunikation in Zellen des Nervensystems durch eu- und prokaryontische Wirkstoffe Auffindung, Isolierung und Charakterisierung von second messenger-Systeme regulierenden bakteriellen Wirkstoffen Prof. Dr. Bernd Hamprecht, Biochemie Molekulare Mechanismen der Aktivierung von Makrophagen zur Phagosytose und Bakterizidie durch mikrobielle Oberflächenkomponenten und synthetische Analoga Prof. Dr. Wolfgang Bessler, Mikrobiologie/ Membranphysiologie Chemie der Reizverarbeitung in Paramecium Prof. Dr. Joachim Schultz, Pharmazie PD Dr. Susanne Klumpp, Pharmazie Funktion von second messengern in Paramecium Prof. Dr. Joachim Schultz, Pharmazeutische Chemie Molekularbiologische Charakterisierung und Regulation von Exoproteinen und Exopeptiden bei Stapylokokken Genetische und biochemische Charakterisierung von Exopeptiden und Staphylokokken Biosynthese und Regulation von Exoproteinen, Exopeptiden und Membranpigmenten bei Staphylokokken Genetische und biochemische Charakterisierung von Exopeptiden/Exoproteinen und Carotinoiden bei Staphylokokken Prof. Dr. Friedrich Götz, Mikrobielle Genetik Eisentransport bei Gram-positiven und Gramnegativen Bakterien Prof. Dr. Klaus Hantke, Mikrobiologie/ Membranphysiologie
TP B7
Proteinphosphatasen in Paramecium PD Dr. Susanne Klumpp, Pharmazeutische Chemie
TP B8
Analyse der IgA Protease b-Domäne, ein Vehikel für den Proteinexport durch die äußere Membran von Gram-negativen Bakterien Mechanismus und Bedeutung der natürlichen Transformationskompetenz bei Neisserien PD Dr. Thomas Meyer, MPI Biologie
1986–1990
1991–1991
1986–1988
1986–1990
1997–1999
1988–1990 1991–1993
1994–1996
1997–1999
1991–1999
1991–1995
1991–1993 1994–1996
347
20 Documentation of the Collaborative Research Centre 323 TP B9
TP B10
Intrazelluläre Freisetzung und Transport von A-Proteinen pathogener Neisserien in eukaryontischen Zellen PD Dr. Johannes Pohlner, MPI Biologie Genetische und biochemische Charakterisierung des Cytotoxins von Helicobacter pylori Dr. Rainer Haas, MPI Biologie
1991–1993
1994–1996
Project area C – Chemical structure elucidation TP C2
Chemie von Antibiotika und Immunmodulatoren Chemie der Mikroorganismen Peptid- und Proteinchemie der Mikroorganismen Prof. Dr. Günther Jung, Organische Chemie
1986–1990 1991–1996 1997–1999
TP C3
Naturstoffaufklärung Analytik und Strukturaufklärung von Naturstoffen PD Dr. Jörg Metzger, Organische Chemie, später Hydrochemie/Hydrobiologie
1991–1993 1994–1999
TP YE1 Dr. Reissbrodt, RKI Wernigerode
1992–1993
TP YW1 Prof. Braun, Tübingen, Mikrobiologie
1992–1993
20.3 Promotion of members of the collaborative research centre
The following project leaders left for other positions outside the collaborative research centre: Annette Beck-Sickinger, C4, University of Basel. Wolfgang Bessler, C3, Immunology, University of Freiburg. Karl-Dieter Entian, C3, Microbiology, University of Frankfurt. Rainer Haas, C3, Medical Microbiology, University of Munich. Knut Heller, C2, Microbiology, University of Konstanz. Susanne Klumpp, C3, Pharmaceutical Chemistry, University of Marburg. Jörg Metzger, C4, University of Stuttgart. Thomas F. Meyer, C4, Infection Biology, MPI Infection Biology, Berlin. Johannes Pohlner, Evotec BioSystems GmbH, Hamburg. Oliver Potterat, C2, University of Lausanne.
348
20.5 Alphabetical list of members and participants
20.4 Recruitment of new project leaders
The project leaders who left were replaced by the following new project leaders: Andreas Bechthold, Pharmaceutical Biology, University of Tübingen. Christiane Bormann, Microbial Genetics/Microbiology-Biotechnology, University of Tübingen. Klaus Hantke, Microbiology-Membrane Physiology, University of Tübingen. Joachim-Volker Höltje, Biochemistry, MPI of Developmental Biology, Tübingen. Karl Poralla, Microbiology-Biotechnology, University of Tübingen. Wolfgang Wohlleben, Microbiology-Biotechnology, University of Tübingen.
20.5 Alphabetical list of members and participants
Name
Prename
Allgaier*, Hermann Alvarado*, Maria Andres*, Nikolaus Angerer*, Annemarie Anton*, Helga Auge, Ulrike Augustin*, Johannes Bauch, Angela Baumgartner*, Angelika Barten*, Roland Bayer*, Anja Bayer, Ernst Bechthold, Andreas Beck-Sickinger, Annette Beitz*, Eric Bessler, Wolfgang Beyer*, Angelika Bielecki, Jarek Blanck*,Wolfgang Blind*, Brigitte Bös*, Christoph Bormann, Christiane** Bormann, Christiane Braun*, Dieter Braun,Volkmar Breisch*, Monika
Acad. degree
Institute
Project area
Participation
Dipl. Chem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Apothekerin Dipl. Biol. Dipl. Biol. Dipl. Biol. Apothekerin Dipl. Biol. Dipl. Biol. Dr. rer. nat. PD Dr. rer. nat. Dipl. Chem. Apotheker Prof. Dr. Apothekerin Dr. rer. nat. Dipl. Biol. Dipl. Ökotroph. Dipl. Chem. Dr. rer. nat. Dr. rer. nat. Dipl. Biol. Prof. Dr. Dipl. Biol.
Organ. Chemie Biologie Biologie Biologie Pharmazie Biologie Biologie MPI, Infektions. Biologie Biologie Organ. Chemie Phys. Chem. Pharmazeut. Bio. Organ. Chemie Pharmazie Biologie Pharmazie Biologie Biologie Phys. Chem. Biologie Biologie Biologie Biologie Biologie Phys. Chem.
C2 A1 A1 B1 B4 B5 B5 B9 B4 B8 C2 B2 A11 C2 B4 B3 B4 A1/GW A1 B2 B1 A1 A4 A1 B1 B2
86–90 86–89 88–89 88–91 91–95 99 88–90 93 86– 94–96 88–94 86–87 97–99 88 94–95 86–88 89 86 86–87 87–91 96–97 86–90 91–99 90–93 86–99 86
349
20 Documentation of the Collaborative Research Centre 323
Name
Prename
Acad. degree
Institute
Project area
Participation
Britten* Uwe Brooks*, Mark Brückner, Reinhold Bruntner*, Christina Brutsche*, Sandra Burckhard*, Renate Cebulla*, Ingeborg Chehadeh*, Heidi Choi*, Ok-Byung Cullmann*, Hans-Jürgen Decker*, Heinrich Demleitner*, Gaby Deres*, Karl Domann*, Silvie Drautz, Hannelore Drechsel*, Hartmut Dringen*, Ralf Dürr, Hansjörg Eick-Helmrich*, Katrin Engel*, Peter Entian**, Karl-Dieter Enz*, Sabine Fauth*, Ursula Faust*, Bettina Feil*, Corinna Fels*, Johannes Flechsler*, Insa Fleckenstein*, Burkhard Fiedler**, Hans-Peter Fiedler,* Waltraud Fischer, Eckhard Franz, Brigitte Freund*, Stefan Freund*, Wolf-Dietrich Friedl, Anette Freund, Stefan Freund*, Wolf-Dieter Fridrich*, Gerald Frosch, Ingrid Früchtel*, Jörg Fussenegger*, Martin Gaisser*, Sabine Gierlich, Doris Götz, Friedrich Groeger*,Wolfram Groß, Matthias Grothe*, Kirsten Gombert*, Frank Groß*, Patricia Günther*, Karola Guo*,Yinglan
Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dipl. Chem. Dipl. Biochem. Dipl. Chem. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dr. rer. nat. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biochem. Prof. Dr. Dipl. Biol. Dr. rer. nat. Dipl. Chem. Dipl. Chem. Dipl. Biochem. Dipl. Biochem. Dipl. Chem. Dipl. Biochem. Apotheker Dr. rer. nat. Dipl. Chem Dipl. Biol. Dipl. Biol. Apothekerin Prof. Dr. Dipl. Biol. Dipl. Biol. Apothekerin Dipl. Biochem. Dipl. Biol. Dr. rer. nat. Dipl. Chem.
Biologie Organ. Chemie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Organ. Chemie Biologie Biologie Organ. Chemie Phys. Chemie Organ. Chemie Biologie Pharmazie Phys. Chem. Biologie Biologie Pharmazie Biologie Biologie Organ. Chemie Organ. Chemie Biologie Biologie Biologie Organ. Chemie Organ. Chemie Pharmazie Phys. Chem. Organ. Chemie Pharmazie Pharmazie Biologie Organ. Chemie Biologie Biologie Pharmazie Biologie Biologie Biologie Pharmazie Organ. Chemie Biologie Biologie Pharmazie
A1 C2 B5 A4 B1 B1 A1 B1 A5 A1 A1 B5 C2 A11 A1 C2 B2 C2 B1 B4 A3 B1 A1 A11 A5 A1 C2 C2 A2 B1 B1 C2 C2 B4 B2 C2 B4 B4 A1 C2 B9 B6 B4 B5 B1 B5 B4 C2 B6 B1 B4
89–91 94–99 89–99 95–97 97–97 86 92–95 86 92–95 92–94 86–95 88–92 88–92 97–98 86–92 89–99 88–90 Stip. 86–89 95 86–88 95–99 86–90 97–98 93–96 91–93 93–98 93–99 86–99 86–87 86 88 88–93 87–91 86–87 88 86–91 89 86–87 94–97 93–95 91–93 86 86–99 98–98 98–99 95–98 90–91 93–95 87–89 86–99
350
20.5 Alphabetical list of members and participants
Name
Prename
Acad. degree
Institute
Project area
Participation
Haag*, Hubert Haag*, Sabine Habeck*, Martina Häsler*, Peter Hambach, Kristina Hamprecht, Bernd Handschuh, Dieter Hansen*, Martin Hantke**, Klaus Hantke, Klaus Harder, Michael, Harkness, Robin Hartjen*, Uwe Hatzelmann, Armin Heckmann*, Dorothee Heidel, Martina Hertle, Ralf Heuermann*, Dorothee Hilger*, Martina Hille*, Matthias Hobbie*, Silke Höltje, Joachim-Volker, Höltzel*, Alexandra Hörner*, Thomas Hörr*, Ingmar Hoff*, Hubert Hofmann,* Hans-Joachim Hoffmann*, Helmut Hoffmann*, Thomas Holz, Bärbel Huhn*, Wolfgang Hummel*, Rolf-Peter Hwang-Kim*, In-Sook Ihlenfeldt*, Hans-Georg Isselhorst-Scharr*, Caroline Jack, Ralph-Wilson Jung, Günther Jung*, Oliver Kabatek* Ursula Kaiser*, Dietmar Kammler*, Meike Kannenberg, Elmar Kapitza, Susanne Katzer*,Werner Kellner*, Roland Kempf, Markus Kempter*, Christoph Kern, Armin Kies, Stefanie Killmann*, Helmut
Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Apothekerin Prof. Dr. Dr. rer. nat. Apotheker Prof. Dr. Prof. Dr. Dr. rer. nat. Dr. rer. nat. Dipl. Biol. Dipl. Biochem. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Prof. Dr. Dipl. Chem. Dr. rer nat. Dipl. Biol. Dipl. Biol. Apotheker Dipl. Biol. Dipl.Chem. Dipl. Biol. Dipl. Biol. Dipl. Biochem. Dipl. Biol. Dipl. Chem.
Biologie Biologie Biologie Biologie Pharmazie Biochemie Biochemie Pharmazie Biologie Biologie Biologie Biologie Biologie Pharmazie Biologie Pharmazie Biologie Biologie Biologie Biologie Biologie Biochemie Organ. Chemie Biologie Organ. Chemie Biologie Pharmazie Biologie Pharmazie Biologie Biologie Organ. Chemie Biologie Organ. Chemie
A1 A1 B1 A1 B4 B2 B2 B4 B1 B6 A2 B1 A2 B4 A10 B4 B1 B10 B1 B5 B1 A12 C2 A2 C2 A1 B4 B1 B4 A1 A1 C2 B1 C2
89–92 92–95 98–98 86–88 98–98 86–91 86–91 98–99 86–90 91–99 94–95 86–88 89–90 86 95–98 86 96–99 95–96 94–95 96–99 90–93 97–99 95–99 86–90 95–99 88–90 86–89 86 95–97 90–94 86–87 87 91–93 90–95
Dipl. Biol. Dr. rer. nat. Prof. Dr. Dipl. Biol Dipl. Biol. Dipl. Chem. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biol. Dipl. Chem. Dr. rer. nat. Dipl. Biol. Dr. rer. nat.
Biologie Organ. Chemie Biochemiker Biologie Organ. Chemie Organ. Chemie Biologie Biologie Organ. Chemie Biologie Organ. Chemie Biologie Organ. Chemie Organ. Chemie Biologie Biologie
A1 C2/GW C2 A1 C2 C2 B6 A5 C2 A1 C2 A2 C2 C2 B5 B1
88–90 91–99 86–99 91–94 86–87 95–98 92–94 98–99 98–99 87–91 86–89 97 93–96 86 98–99 90–99
351
20 Documentation of the Collaborative Research Centre 323
Name
Prename
Kim*, In-Sook Klauser, Thomas Kleemann*, Gisela Klumpp**, Susanne Knigge, Michael Koebnik*, Ralf Könninger*, Ulrich Köster*,Wolfgang Krämer, Joachim Kremer*, Stephan Krismer*, Bernhard Koch*, Ulrike Köster**,Wolfgang Konetschny-Rapp*, Silvia Koss, Dieter Krüger*, Thomas Kugler*, Martin Kühn*, Sabine Kupke*, Thomas Langenberg, Uwe Langer*, Monika Lauer*, Bettina Lechner*, Max Leipert, Dietmar Linder*, Jürgen Lipps*, Hans-Peter Lohmann, Susanne Maerker, Christian Mahnke*, Marion Marquardt*, Udo Mayer*, Irene Merkofer, Thorsten Mehrkühler, Christian Meiwes*, Johannes Mende*, Jasmin Metzler*, Monika Metzger**, Jörg Metzger, Jörg Meyer, Thomas Möhrle*,Volker Möller, Andreas Müller*, Kerstin Müller, Judith Mutard, Denise Muth, Günther Neubauer, Heike Noda, Shigeru Ochs*, Dietmar Ochs*, Martina Odenbreit, Stefan Ölschläger*, Thobias
352
Acad. degree
Institute
Dipl. Biol. Biologie Dr. rer. nat. MPI-Biologie Dipl. Biol. Biologie Dr. rer. nat. Pharmazie Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biol Biologie Dipl. Biol. Biologie Dipl. Biol. MPI Infekt. Biol. Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Chem. Organ. Chemie Dr. rer. nat. Biologie Dr. rer. nat. Organ Chemie Dipl. Biol. Biologie Dipl. Chem. Pharmazie Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biochem. Biologie Dipl. Biochem. MPI Infekt. Biol. Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biol Biologie Dipl. Chem. Organ. Chemie Dipl. Chem. Pharmazie Apotheker Pharmazie Dr. rer. nat. Phys. Chem. Dipl. Biol. MPI Infekt. Biol. Dipl. Biol. Biologie Dipl. Chem. Organ. Chem. Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biol. Phys. Chemie Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biol. Biologie Dr. rer. nat Organ. Chemie Dr. rer. nat. Organ. Chemie Dr. rer. nat. MPI Infekt. Biol. Dipl. Biol. Biologie Dipl. Chem. Phys. Chem. Dipl. Biol. Biologie Dipl.Lebensm.Ing. MPI Biochem. Apothekerin Pharmazie Dr. rer. nat. Biologie Dipl. Biol. Biologie Dr. med. vet. Pharmazie Dipl. Biol. Biologie Dipl. Biol. Biologie Dipl. Biol. MPI, Biochem. Dipl. Biol. Biologie
Project area
Participation
B1 B8 A5 B4 A5 B1 B1 B1 B8 A1 B5 C2 B1 C2 B6 B4 A1 B1 B5 B9 A1 A4 B5 C2 B4 B4 B2 B9 A1 C2 B1 A5 B2 A1 B1 A1 C2 C3 B8 A4 B2 B6 A12 B4 A10 A4 B4 A5 B1 B10 B1
93 91–93 89–92 91–95 99 90–91 94–98 86–87 91–93 87–90 95–99 86–88 86–90 86–90 94–97 93–97 86–86 91–95 89–95 91 92–96 95–99 89 94–98 98–99 89 86 91 86–87 98–99 89–90 96–98 89–91 87–90 86–87 86–89 86–90 91–99 91 91–95 86–90 94–97 99 91 95–99 96 86 90 91–95 94–95 86–90
20.5 Alphabetical list of members and participants
Name
Prename
Ötzelberger, Karin Ondrazcek, Roland Ottenwälder, Birgit Otto*, Michael Otto*, Susanne Pantel, Iris Patzer*, Silke Peschel, Andreas Perzl*, Michael Petersen*, Frank Pfefferle*, Uwe Pfeiffer, Brigitte Pilsl*, Holger Pfeifer,Volker Plaga*, Armin Plantoer*, Stefan Pohlner**, Johannes Potterat, Oliver Poralla, Karl Preßler*, Uwe Pultar*, Thomas Probst, Katrin Rabenhorst*, Jürgen Rak, Gabriele Rauch*, Beatrix Rapp* Claudius Rechenberg*, Moritz von Reeger, Eva Reissbrodt, Rolf Reuschenbach,*, Peter Reuschenbach*, Petra Reutter*, Felix Rexer, Hans-Ulrich Richter*, Monika Röhl*, Franz Rösch, Hartmut Rohling*, Anette Roos*, Margareta Roos*, Ulrich Rose, Andreas Rosenstein, Ralph Ruan,Yuan Russwurm*, Roland Sauer*, Martin Sauter, Simon Schade, Uwe Schäffer*, Sven Scharr*, Jürgen Schaude*, Renate Schiller*, Max Schiffer*, Guido
Acad. degree
Institute
Project area
Participation
Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dr. rer. nat. Dipl. Biochem. Dr. rer. nat. Dipl. Biol. Dr. rer. nat. Dr. rer. nat. Prof. Dr. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Chem. Dipl. Biochem. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biochem. Dr. rer. nat. Dipl. Biol Dipl. Biol. Dipl. Biol. Dipl. Biol. Apotheker Dipl. Biol. Dipl. Biol. Dipl. Biochem. Dipl. Biol. Dr. rer. nat.
MPI Biochem. Biologie Biologie Biologie Pharmazie Biologie Biologie Biologie Biologie Biologie Biologie Phys. Chemie Biologie Biologie Biologie Biologie MPI Biochem. Biologie Biologie Biologie Biologie Organ. Chemie Biologie Biologie Biologie Organ. Chemie MPI Biochem. Biologie R. Koch Inst. Biologie Biologie Organ. Chemie Biologie Biologie Biologie Pharmazie Biologie Biologie Biologie Phys. Chem. Biologie Biologie Biologie Biologie Biologie Pharmazie Biologie Biologie Organ. Chemie Phys. Chemie Biochemie
B9 B1 B5 B5 B7 A4 B6 B5 A5 A1 A2 B2 B1 A10 A2 B1 B9 A1 A5 B1 A1 C2 A1 A1 A1 C2 A12 B6 YE1 A2 A1 C3 A10 A2 A1 B4 A10 A1 B5 A3 B5 B1 A4 B1 B1 B4 B1 A1 C2 B2 A12
91–93 91 95–96 93–97 90–93 95 94–99 94–95 93–99 88–91 89–90 86–90 94–98 99 86–90 98 91–93 91–94 91–99 86 86–88 99 86–87 87–90 89–91 86–88 97–99 97 91–93 86–88 86 96–99 97–99 95 86–87 90– 97 89–92 91–93 86 86–99 92 97–99 86–87 99 86–90 86–87 87–91 87–89 86–90 97–98
353
20 Documentation of the Collaborative Research Centre 323
Name
Prename
Schmidt, Günther Schmitz, Susanne Schneider*, Ursula Schneider, Richard Schnell, Norbert Schöffler*, Harald Schön*, Claudia Schönborn, Christoph Schönefeld*, Ulrich Schönherr*, Roland Schuerhoff-Göters*, Wilhelm Schüz*, Traugott Schüz*, Traugott Schultz*, Gabriele Schultz, Joachim Schwarz*,Wolfgang Schwartz, Dirk Seiffert*, Andreas Selke*, Dagmar Sommer*, Patricia Sorg, Gerhard Sprengler, Siegried Stahl, Bernd Stanger*, Andrea Staudenmaier*, Horst Steinlen*, Siegfried Stephan*, Holger Stefanovic*, Stefan Stegmann*, Evi Stiefel*, Alfred Stümpfel*, Joachim Süßmuth*, Roderich Surovoy, Andrey Tappe*, Cord-Henning Tejmar-Kolar, Liana Templin, Markus Teufel*, Pia Theobald, Uwe Thumm*, Günter Tippelt*, Anette Traub*, Irene Trefzer, Axel Troeger*,Wilfried Tschen, Shu Yuan Tschierske*, Martin Ungermann*,Volker Völkel*; Helge Voges, Brigitte Voges, Klaus-Peter Vollmer*,Waldemar
354
Acad. degree
Institute
Project area
Participation
Dipl. Biochem. Dipl. Biochem. Apothekerin Dipl. Biol. Dipl. Biochem. Dipl. Biol. Dipl. Biol. Apotheker Dipl. Biochem. Dipl. Biol.
Phys. Chemie Biologie Biologie Biologie Phys. Chemie Biologie Biologie Pharmazie Pharmazie Biologie
A3 A5 A1 B1 A3 B1 B1+B6 B4 B4 B1
89 97–99 86 94–95 89 86–89 87–90 90–91 86–88 92–93
Apotheker Dipl. Biol. Dr. rer. nat. Dipl. Biol. Prof. Dr. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Chem. Dipl. Biochem. Dipl. Biol. Dipl. Biol. Dipl. Biochem. Dipl. Chem. Dipl. Biochem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dr. rer. nat. Dipl. Biochem. Dr. rer. nat. Dr. rer. nat. Dipl. Biol. Dr. Ing. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Chem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biochem. Dr. rer. nat. Dr. rer. nat Dr. rer. nat.
Pharmazie Biologie Biologie Biologie Pharmazie Biologie Biologie Biologie Pharmazie Biologie Organ. Chemie Phys. Chemie Phys. Chemie Biologie Biologie Pharmazie Organ. Chemie Organ. Chemie Biologie Biologie Biologie Organ. Chemie Organ. Chemie Botan. Instit. Biologie Biochemie Biologie Biologie Biologie Biologie Biologie Pharmazie Organ. Chemie Biologie Biologie Biologie Pharmazie Organ. Chemie Organ. Chemie Biochemie
B4 A1 A2 B1 B4 A4 A10 A1 B4 A4 C3 B2 B2 A1 B1 B4 C2 C2 A10 B1 A1 C2 C2/GW A5 A1 A12 B5 A2 B5 A5 B1 A11 C2 A1 A1 A1 B4 C2 C2 A12
86–90 86 87–94 88–92 86–99 97–99 95–99 90–92 91–95 94–95 98–99 86–88 86–87 86 86 89–92 91–95 89–92 98–99 97–99 86–88 95–99 90–97 91–93 86 97–99 89–92 95–99 88–93 96–97 90–92 98–99 89–91 91 92–94 88–90 92–96 86–88 86–88 97–99
20.6 Support of young scientists
Name
Prename
Van hove*, Brundhild von der Mülbe, Florian Veitinger*, Sabine Vierling*, Silke Videnov, Georgi Walker, Georg Walz*, Franz Wang-Tschen*, Shu-Yuan Wasiliu*, Michal Weber, Tillmann Weitnauer, Gabriele Welz*, Dietrich Wieland*, Bernd Wieland*, Karsten-Peter Wiesinger, Heiner Wiesmüller*, Karl-Heinz Witke, Claudia Woelk*, Uwe Wohlleben,Wolfgang Zähner, Hans Ziegelmaier-Kemmling,* Dagmar Zimmermann, Luitgard Zimmermann, Norbert
Acad. degree
Institute
Project area
Participation
Dipl. Biol. Dipl. Biochem. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dipl. Biol. Dr. rer. nat. Dipl. Chemiker Dipl. Biol. Dipl. Biol. Prof. rer. nat. Prof. Dr. Ing.
Biologie Organ. Chemie Biologie Biologie Organ. Chemie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Biologie Phys. Chem. Organ. Chemie Biologie MPI Infektions. Biologie Biologie
B1 C2 B1 A10 C2/GW B1 A2 B1 A1 A10 A11 B1 B5 B5 B2 C2 B5 B8 A10 A1
89 97–99 89–92 97–99 91–95 97–99 86–88 91–92 86–86 99 98–99 92–95 89–93 96–99 86–91 86–92 91–94 95–96 95–99 86–96
Dipl. Biol. Dr. rer. nat. Dipl. Chem.
Biologie Biologie Organ. Chemie
A1 B1 C2
86–90 86 91
* Graduation with PhD during the collaborative research centre 323. ** Habilitation during the collaborative research centre 323.
20.6 Support of young scientists
List of promotions resulting from the collaborative research centre TP A1 Alvarado, Maria (1990): Monomeres und dimeres Cinnachinon aus Streptomyces griseoflavus (Tü 2482). Andres, Nikolaus (1989): Hormaomycin, ein Peptidlacton mit morphogener Wirkung auf Streptomyceten. Blank, Wolfgang (1987): Untersuchungen zur biologischen Modifikation der Bafilomycine. Braun, Dieter (1993): Enzymatische Halogenierung von mikrobiellen Metaboliten und Etablierung eines Photokonduktivitätsscreening. 355
20 Documentation of the Collaborative Research Centre 323 Cebulla, Ingeborg (1995): Gewinnung komplexbildender Substanzen mittels Amycolatopsis orientalis. Cullmann, Hans Jürgen (1994): Depsichlorine und andere Metabolite aus Streptomyces antibioticus Tü 1661. Decker, Heinrich (1989): Untersuchung zur Struktur-Wirkungsbeziehung der Nikkomycine und Isolierung neuer Nikkomycine. Fauth, Ursula (1987): Galbonolid A und B, neue antifungische Makrolid-Antibiotika. Fels, Johannes (1994): Suche nach Chitinase-Inhibitoren und Charakterisierung der Hemmstoffe aus Streptomyces tendae Tü 2774 und Streptomyces sp. Tü 3566. Haag, Hubert (1992): Neue Eisenkomplexbildner aus Staphylokokken und Yersinia enterocolitica. Screening, Fermentation, Isolierung und Charakterisierung. Haag, Sabine (1996): Gentransfer zur Produktion neuer Naturstoffe am Beispiel der Tetracenomycine und Urdamycine und CDA – ein Calcium-abhängiges Antibiotikum aus Streptomyces coelicolor A3(2). Häsler, Peter (1988): Streptonolide. Sekundärmetabolite aus Streptomyces olivaceus Tü 2108. Hoff, Hubert (1990): Obskurolide: neue Butyrolactone aus Streptomyceten. Huhn, Wolfgang (1986): Maduraferrin, ein neues Siderophor aus Actinomadura madurae. Screening, Isolierung und Fermentation. Isselhorst-Scharr, Caroline (1995): Tü 3010/1. Ein neues Thiolactonantibiotikum aus Streptomyces olivaceus ssp. gelaticus Tü 3010. Jung, Oliver (1993): Screening nach neuen Siderophoren und die Charakterisierung eines hochaffinen Eisenaufnahmesystems bei Staphylococcus aureus. Katzer, Werner (1991): Bromo- und Chlorotetain aus Bacillus amyloliquefaciens. Kremer, Stefan (1990): Untersuchungen zur Biosynthese von ungewöhnlichen Macrolidantibiotika am Beispiel der Bafilomycine. Kugler, Martin (1986): Rhizocticin. Ein neues Antibiotikum aus Bacillus subtilis ATCC 6633. Langer, Monika (1996): Biotinantagonisten und Siderophore aus Streptomyces griseoflavus ssp. griseoflavus. Mahnke, Marion (1990): Inhibitoren der Lysin-N6-Hydroxylase aus Escherichia coli MM 128. Screening, System und Charakterisierung. Meiwes, Johannes (1989): Neue Siderophore von Staphylococcen und Streptomyceten. Metzler, Monika (1989): Untersuchungen an sekundären Metaboliten verschiedener Streptomyceten. Petersen, Frank (1991): Germicidin B – ein autoregulatorischer Keimungshemmstoff aus Streptomyces viridochromogenes. Plaga, Armin (1990): Studien zur mikrobiologischen Produktion von Phosphinothricin aus Streptomyces viridochromogenes Tü 494. Pultar, Thomas (1988): Lysolipin X und I. Fermentation, Isolierung, biologische Wirkung und Interaktion mit Mg2+. Rabenhorst, Jürgen (1986): Valclavam – ein antifungisches b-Lactam. 356
20.6 Support of young scientists Rauch, Beatrix (1992): Hormaomycin und andere differenzierungsaktive Substanzen aus Streptomyceten. Reuschenbach, Peter (1986): Studien zur Fermentation der phosphorsäuretriesterhaltigen Lactone aus Tü 1718. Röhl, Franz (1986): Untersuchungen zur Wirkungsweise von Clavam Antibiotica. Roos, Margareta (1994): Untersuchungen zur Differenzierung bei Streptomyces griseus und Streptomyces antibioticus. Scharr, Jürgen (1993): Hohe Zelldichten bei Bacillus thuringiensis var. israelensis. Schüz, Traugott (1990): Pelletbildung bei Streptomyces tendae Tü 901/S2566 und verfahrenstechnische Optimierung der Nikkomycin-Fermentation. Seiffert, Andreas (1992): Untersuchungen zum Eisentransport bei verschiedenen Bakterien. Stümpfel, Joachim (1988): Pyrrolam. Ein gamma-Lactam aus Streptomyces olivaceus Tü 3082. Fermentation, Isolierung und biologische Wirkung. Tschierske, Martin (1994): Studien zur Produktion von Rhizoferrin und analogen Verbindungen mit Cunninghamella elegans. Ziegelmaier-Kemmling, Dagmar (1990): Vergleichende Untersuchungen zur Wirkungsweise von Tetramsäure-Verbindungen. TP A2 Incorporated in TP A1 TP A4 Bruntner, Christina (1997): Molekularbiologische Untersuchungen zur Biosynthese von Nikkomycin D in Streptomyces tendae TÜ 901. Lauer, Bettina (1997): Charakterisierung von Nikkomycin-Biosynthesegenen und genetische Manipulation des Biosyntheseweges in Streptomyces tendae TÜ 901. Möhrle, Volker (1995): Klonierung und Charakterisierung von Nikkomycin-Biosynthesegenen aus Streptomyces tendae TÜ 901. Roos, Ulrich (1993): Histidin-Aminotransferase-Aktivität in Streptomyces tendae: Korrelation zur Nikkomycin-Produktion, Reinigung und Charakterisierung. Schwarz, Wolfgang (2000): Untersuchungen zur transkriptionellen Regulation der Nikkomycin Synthese in Streptomyces tendae Tü 901. Sommer, Patricia (1995): Klonierung, Sequenzierung und Charakterisierung eines Lipasegens aus Streptomyces cinnamomeus TÜ 89. Russwurm, Roland (2000): Genetische Untersuchungen zur Nikkomycin-Biosynthese in Streptomyces tendae TÜ 901. Tesch, Cornelia (1995): Klonierung, Sequenzierung und Charakterisierung eines Esterasegens aus Streptomyces diastatochromogenes TÜ 20. TP A5 Choi, Ok-Byung (1995): Experimente zur Klonierung, Sequenzierung und Expression der Squalen-Hopen-Cyclase aus Rhodopseudomonas palustris und Alicyclobacillus acidoterrestris. 357
20 Documentation of the Collaborative Research Centre 323 Feil, Corinna (1996): Ortsspezifische Mutagenese zur Identifizierung katalytisch aktiver Aminosäuren der Squalen-Hopen-Cyclase von Alicyclobacillus acidocaldarius. Kleemann, Gisela (1992): Hopanoidgehalt und Fettsäuremuster zweier Rhodopseudomonas-Arten und Reinigung der Squalen-Hopen-Cyclase aus Rhodopseudomonas palustris. Perzl, Michael (1996): Biochemische und molekularbiologische Untersuchungen zur Hopanoid-Biosynthese in Bradyrhizobium und zur Tetrahymanol-Biosynthese in dem Ciliaten Tetrahymena. Schmitz, Susanne (2000): Charakterisierung des Hopanoidbiosynthese-Operons aus Bradyrhizobium japonicum und der Squalen-Hopen-Cyclase aus Alicyclobacillus acidocaldarius. Tappe, Cord Henning (1993): Squalen-Hopen-Cyclasen: Reinigung, Charakterisierung und Inhibitor-Experimente. TP A10 Burger, Annette (1997): Isolierung und Charakterisierung einer Genregion mit Zellteilungs- und Differenzierungsgenen aus Streptomyces coelicolor A3(2). Maas, Ruth Maria (1997): Molekulargenetische Analyse des Plasmids pSG5 aus Streptomyces ghanaensis DSM 2932. Schwartz, Dirk (1997): Molekulargenetische Analyse der Phosphinothricin-Tripeptid-Biosynthese in Streptomyces viridochromogenes Tü 494. Stegmann, Efthimia (1999): Molekulargenetische und biochemische Untersuchungen des EDDS-Produzenten Amycolatopsis japonicum MG417-CF17. Vierling, Silke (2000): Molekulargenetische Analyse des recA/recX-Operons in Streptomyces lividans TK 64. TP A11 Gaisser, Sibylle (1998): Molekularbiologische und biochemische Untersuchungen zur Avilamycin-Biosynthese und Resistenz in Streptomyces viridochromogenes Tü57. Doman, Silvie (1998): Untersuchungen zum Wirkmechanismus der Landomycine und zur Biosynthese der Desoxyzucker D-Olivose und L-Rhodinose in den Angucyclin-Antibiotika Landomycin A und Urdamycin A. Faust, Bettina (1998): Untersuchungen zur Biosynthese von Urdamycin A und Herstellung neuer Naturstoffe mittels molekularbiologischer Methoden. TP A12 Schiffer, Guido (1998): Identifizierung und funktionale Charakterisierung des Penicillin-Bindeproteins 1C als Mureinpolymerase in Escherichia coli. Rechenberg, Moritz von (1998): Untersuchungen der Protein-Protein Wechselwirkungen zwischen Murein Hydrolasen und Penicillin-bindenden Proteinen in Escherichi coli. Vollmer, Waldemar (1998): Identifizierung und Charakterisierung eines Strukturproteins in Multienzymkomplexen aus Mureinsynthasen und Mureinhydrolasen in Escherichia coli. 358
20.6 Support of young scientists TP B1 Hoffmann, Helmut (1986): Proform und reifes FhuA Rezeptorprotein von E. coli K-12: Reindarstellung und biologische Eigenschaften. Köster ,Wolfgang (1986): Eisenhydroxamattransport von E. coli: Nukleotidsequenz des fhuB Gens. Identifizierung und Lokalisierung des FhuB Proteins. Ölschläger, Tobias (1986): Genetische Analyse des colM Lokus und Nukleotidsequenz des Colicin M Immunitätsproteins. Burkhardt, Renate (1987): Molekulare Charakterisierung der Gene fhuA, fhuC und fhuD des Ferri-Hydroxamat-Transportsystems bei E. coli. Fiedler, Waltraud (1987): Physiologische Bedeutung periplasmatischer Oligosaccharide (membrane derived oligosaccharides, MDO) bei Escherichia coli K-12: Untersuchungen an Mutanten der MDO Biosynthese. Preßler, Uwe (1987): Genetische Charakterisierung des Bindeprotein-abhängigen Eisen-Dicitrattransports. Eick-Helmerich, Katrin (1989): Untersuchungen zur Struktur und Funktion der Gene exbB und exbD bei Escherichia coli K-12. Sauer, Martin (1989): Eisen(III)-Aufnahme von Escherichia coli K12: Nukleotidsequenz des fhuE Gens und Untersuchungen zur Funktion konservierter Bereiche bei TonB-abhängigen Rezeptoren. Schäffer, Sven (1989): Untersuchung des Fur-Eisenrepressors von Escherichia coli K12. Schöffler, Harald (1989): Transport durch die äußere Membran von Escherichia coli. Staudenmaier, Horst (1989): Exogene Induktion des Eisendicitrat-Transportsystems von Escherichia coli. Günter, Karola (1990): Zur Intermembran-Kopplungsfunktion des TonB Proteins von Escherichia coli. Mende, Jasmin (1990): Colicin B: Untersuchungen zum Exportverhalten Colicin-produzierender Zellen und zur TonB-abhängigen Aufnahme durch die äußere Membran. Schön, Claudia (1990): Untersuchung des Aufnahmesystems für zweiwertiges Eisen bei Escherichia coli. Van hove, Brunhilde (1990): Bindeprotein-abhängiger Transport und Transmembranregulation des Eisen(III)Dicitrat-Transportsystems. Chehade, Heidi (1990): Untersuchungen zur Wechselwirkung von Escherichia coli Wildstämmen und Mutanten in umweltregulierten Genen bei bakteriziden Komponenten von Humanserum. Angerer, Annemarie (1991): Ein neues Eisentransportsystem in Serratia marcescens. Gaisser, Sabine (1992): Das TonB Gen aus Serratia marcescens. Koebnik, Ralf (1992): Ferrichrom-Aufnahme in Bakterien: Membrantopologie des Ferrichromrezeptors von Escherichia coli und Struktur des Ferrichromrezeptors und des TonB Proteins von Yersinia enterocolitica. Schultz-Hauser, Gabriele (1992): FhuC, die konservierte Komponente des Eisen(III) Hydroxamat-Transports. Traub, Irene (1992): Untersuchung funktioneller Domänen des TonB-Proteins von Escherichia coli. 359
20 Documentation of the Collaborative Research Centre 323 Veitinger, Sabine (1992): Das Eisen(III)-Dicitrat-Transportssystem: Kartierung auf dem Chromosom von Escherichia coli K-12 und Untersuchungen zur Transmembran-Regulation. Hobbie, Silke (1993): Untersuchungen zur Struktur und Funktion des ShlB Proteins, der transportierenden und aktivierenden Komponente des Hämolysins von Serratia marcescens. Killmann, Helmut (1993): Eisen(III)Hydroxamat-Transport in Escherichia coli. Funktionsdomänen des Rezeptor-Proteins FhuA und Untersuchungen zur dreidimensionalen Grundstruktur des Proteins in der äußeren Membran. Schönherr, Roland (1993): Aktivierung und Sekretion des Hämolysins von Serratia marcescens. Ochs, Martina (1994): Untersuchungen zur Regulation des Eisen-Dicitrat-Transportsystems von Escherichia coli K-12. Kim, In-Sook (1995): Exogene Signaltransduktion des Eisen(III)-Dicitrat-Transportsystems von Escherichia coli K-12. Kühn, Sabine (1995): Rhizoferrin-Aufnahme in Morganella morganii. Groß, Patricia (1996): Untersuchungen zur Struktur und Funktion des Colicin M-Immunitätsproteins (Cmi). Pilsl, Holger (1996): Domänenstruktur, Evolution und Immunität der Colicindeterminanten 5, 10 und K von Escherichia coli und der Pesticindeterminanten von Yersinia pestis. Bös, Christof (1997): In vivo Charakterisierung des Ferrichrom-Rezeptor-Proteins FhuA aus Escherichia coli mittels thiolspezifischer Reagenzien. Brutsche, Sandra (1997): SigX – ein neuer Sigmafaktor der ECF-s70 Subfamilie aus Bacillus subtilis. Enz, Sabine (1997): Transkriptionsregulation des Eisen(III) Dicitrat-Transportsystems von Escherichia coli K-12. Welz, Dieter (1997): Funktion und Lokalisierung der Induktionsproteine FecI und FecR von Escherichia coli K-12. Groeger, Wolfram (1998): Eisen(III)-Hydroxamat-Transport in Escherichia coli: Topologie des integralen Transportproteins FhuB. Habeck, Martina (1998): Energiekopplung durch TonB im Eisen(III)DicitratTransportsystem von Escherichia coli K-12.
TP B4 Hirschhausen, Heinrich Reginhard von (1986): Die Phosphodiesterasen in Paramecium. Schade, Uwe (1988): Zur physiologischen Rolle von cyclischem 3',5'-Guanosinmono-phosphat in Paramecium tetraurelia. Hofmann, Hans-Joachim (1990): Anreicherung und Charakterisierung einer membranständigen, ciliären Guanylatcyclase aus Paramecium tetraurelia. Schürhoff-Goeters, Wilhelm (1990): Zur Entstehung cyclischen AMPs infolge hyperpolarisierender Pufferveränderungen bei Paramecium tetraurelia. Freund, Wolf-Dietrich (1991): Inositol-Phospholipide und Inositol-Phosphate in Paramecium tetraurelia. 360
20.6 Support of young scientists Steinlen, Siegfried (1992): Untersuchungen zur Regulation der partikulären Guanylatcyclasen der Ciliaten Paramecium und Tetrahymena durch Calmodulin. Friderich, Gerald (1992): Reinigung und Charakterisierung der Proteinphosphatase Typ 1 aus den Cilien von Paramecium tetraurelia. Völkel, Helge (1992): Adenylatcyclasen aus ciliären Geweben und dem retinalen Pigmentepithel. Beyer, Angelika (1993): Proteinphosphatase Typ 2C aus Paramecium tetraurelia: Lokalisation, Isolierung, Teilsequenzierung und Charakterisierung. Schönborn, Christoph (1993): Untersuchungen zur Regulation von cyclischem Adenosin 3',5'-monophosphat bei Paramecium tetraurelia und Ionenkanalmutanten. Otto, Susanne (1994): Reinigung und Charakterisierung einer Adenylatcyclase der Retina. Völkel, Helge (1995): Adenylatcyclasen aus ciliären Geweben und dem retinalen Pigmentepithel. Selke, Dagmar (1995): Proteinphosphatase 2C aus der Rinderretina – Aufreinigung, Charakterisierung, Klonierung und Expression. Hanke, Cordula (1996): Proteinphosphatase Typ 2C aus Paramecium tetraurelia. Guo, Yinglan (1996): Zur Ca2+-abhängigen Regulation der Bildung von cGMP, des Schwimmverhaltens und der Lokalisation von Ca-Kanälen in Paramecium tetraurelia. Anton, Helga (1996): Proteinphosphatase Typ1 aus der Rinderretina – Reinigung, Charakterisierung und Expression. Linder, Jürgen (1997): Klonierung einer Adenylatcyclase aus Paramecium. Beitz, Eric (1997): Zur cAMP-Signaltransduktion des retinalen Pigmentepithels des Rindes und des Innenohrs der Ratte. Krüger, Thomas (1997): Klonierung von Adenylatcyclasen aus dem Ciliaten Tetrahymena pyriformis. Grothe, Kirsten (1998): Mutationsanalyse und Lokalisation der Proteinphosphatase Typ 2C aus Paramecium tetraurelia. Hoffmann, Thomas (1999): Membranständige Guanylatcyclasen aus Paramecium und Tetrahymena: Klonierung und bakterielle Expression der katalytischen Bereiche. Engel, Peter (1999): Klonierung und Expression einer Guanylatcyclase aus Paramecium tetraurelia. TP B5 Demleitner, Gabi (1992): Die Exolipase von Staphylococcus hyicus: Identifizierung der katalytisch aktiven Aminosäuren und Untersuchungen zur Funktion des Propeptids. Krismer, Bernhard (1999): Studium der Funktion der sekretierten Proteine SceA und SceB, Analyse des Galaktoseoperons galRKET und Kontruktion von Sekretions- und Expressionsvektoren in Staphylococcus carnosus. Kupke, Thomas (1992): Posttranslationelle Modifikation bakterieller Peptide – proteinchemische Untersuchungen zur Epiderminbiosynthese. Lechner, Max (1989): Klonierung und Charakterisierung des Gens für die Phosphatidylinositol-spezifische Phospolipase C von Bacillus thuringiensis. 361
20 Documentation of the Collaborative Research Centre 323 Otto, Michael (1997): Epidermin: Biochemische Untersuchungen zur Biosynthese, Regulation und Immunität. Teufel, Pia (1993): Isolierung; Sequenzierung und Charakterisierung einer Metalloprotease aus Staphylococcus epidermidis und Sekretions- und Regulationsstudien bei Staphylococcus carnosus. Thumm, Günther (1996): Molekularbiologische Charakterisierung von Lysostaphin und Lysostaphin-Immunitäts-Faktor (Lif). Wieland, Bernd (1993): Der xylA Promotor aus Staphylococcus xylosus als Grundlage der transkriptionellen Regulation von Genen in Staphylococcus carnosus. Wieland, Karsten-Peter (1999): Organisation und Genexpression der Carotinoid-Biosynthesegene aus Staphylococcus aureus Newman und Untersuchungen zur Funktion von Staphyloxanthin. TP B6 Kammler, Meike (1994): Sequenzierung und Charakterisierung des Aufnahmesystems für zweiwertiges Eisen von Escherichia coli. Müller, Kerstin (1997): FhuF, ein neuartiges, eisenreguliertes Eisen-SchwefelProtein von Escherichia coli. Patzer, Silke (1999): Regulation durch Metallionen in Escherichia coli, insbesondere Identifizierung und Charakterisierung des hochaffinen Zink-Aufnahmesystems ZnuABC und des zinkabhängigen Regulators Zur. TP C2 Bayer, Anja (1993): Produktion, Isolierung und Strukturaufklärung des glycinreichen Polypeptidantibiotikums Microcin B17. Brooks, Marc (1999): Neue DNA-Gyrase und Humane Type II DNA-Topisomerase-Inhibitoren. Deres, Karl (1992): MHC-Klasse-1-restringierte Peptide. Drechsel, Hartmut (1993): Strukturaufklärung neuer Siderophore vom Carboxylat-Typ und Synthese von Siderophor-Antibiotika-Konjugaten. Flechsler, Insa (1998): Transfektion von Nukleinsäuren mit neuen kationischen Lipiden und Vergleich verschiedener Antisense-Strategien. Fleckenstein, Burkhard (1997): Kombinatorisch aufgebaute Peptidkollektionen zum Studium der HLA-Klasse II-Peptid-Interaktion und der Antigenerkennung autoreaktiver, humaner T-Zellen. Höltzel, Alexandra (1999): Structure Elucidation of Secondary Metabolites and of the Loop Sequence 316–333 of the ThuA Receptor. Hörr, Ingmar (1999): RNA-Vakzine zur Induktion von spezifischen cytotoxischen T-Lymphozyten und Antikörpern. Ihlenfeldt, Hans-Georg (1995): Die B-und T-Zell-Epitope des Hepatitis C-Virus. Kabatek, Ursula (1987): Eine neuartige Teichonsäure aus Streptomyces venezuelae. Kaiser, Dietmar (1998): Strukturaufklärung und Konformationsanalyse biologisch aktiver Sekundärmetabolite. Kellner, Roland (1989): Lantibiotika – ribosomal synthetisierte Polypeptidantibiotika mit Sulfidbrücken und Dehydroaminosäuren. 362
20.6 Support of young scientists Kempter, Christoph (1996): Strukturaufklärung mikrobieller Sekundärmetabolite durch Elektrospray-Massenspektrometrie und mehrdimensionale Kernresonanzspektroskopie. Koch, Ulricke (1988): Fengycin, Strukturaufklärung eines mikroheterogenen Lipopeptolidantibiotikums. Konetschny-Rapp, Sylvia (1990): Neue mikrobielle Eisenkomplexbildner, Screening, Isolierung, Strukturaufklärung und komplexchemische Untersuchungen. Metzger, Jörg (1988): Immunstimulierende Lipopeptide als Membrananker für Haptene und biologisch aktive Wirkstoffe. Rapp, Claudius (1988): Neue antifungische Phosphono- und Chloro-Oligopeptide aus Bacillus subtilis und ein neues Thiolactonantibiotika aus Streptomyces olivaceus – Isolierung und Strukturaufklärung. Stephan, Holger (1995): Strukturaufklärung mikrobieller Sekundärstoffwechselprodukte durch mehrdimensionale Kernresonanzspektroskopie. Stevanovic, Stefan (1992): Multiple Sequenzanalyse – Ein neuer Ansatz in der Peptidsequenzierung: Isolierung, Synthese und Strukturaufklärung von mikrobiellen Metaboliten aus Amycolatopsis mediterranei, Staphylococcus epidermidis und Streptomyces lividans. List of habilitations resulting from the collaborative research centre Bormann, Christiane (1997): Biosynthese des Antibiotikums Nikkomycin. Brückner, Reinhold (1997): Katobolitrepression in Staphylokokken. Bechthold, Andreas (1998): Streptomycetengenetik, Grundlage für die Herstellung neuer Antibiotika: Klonierung und Charakterisierung der Biosynthesegencluster von Avilamycin, Landomycin, Urdamycin und Granaticin. Fiedler, Hans-Peter (1988): Isolierung und Analytik niedermolekularer mikrobieller Sekundärmetabolite. Klumpp, Susanne (1989): Charakterisierung von Phosphatasen und Entdeckung eines Inhibitor-Proteins. Kupke, Thomas (1998): Mikrobielle Enzyme – neuartige Reaktionen und Katalysemechanismen. Metzger, Jörg (1994): Naturstoffanalytik und Strukturaufklärung mit modernen Methoden der Massenspektrometrie. Graduiertenkollegs – „Mikrobiologie“ – „Analytische Chemie“ – „Zellbiologie in der Medizin“
363
20 Documentation of the Collaborative Research Centre 323
20.7 Alphabetical list of guests
Barnet, James Prof. Bielecki, Jarek Dr., Univ. Warsaw Focareta, Antonio Dr.,University of Adelaide Howard, Stephen Peter, University of Regina Kálmanánczhelyi, Attila, Bukarest Kim, In-Sook Dr. rer. nat Jack, Ralph-Wilson Dr. rer. nat., Univ. of Otago Milla, Paola, Univ. Turin Noda, Shigeru Dr. rer. nat., Univ. Gießen Potterat, Olivier, Dr. rer. nat. Reinhard, Peter Dr. rer.nat., Univ. Canberra Sarem, Aslani Dr. rer. nat., Univ. Würzburg Shi, Liangru Dr. Smaijs, David Dr., Masaryk University Stojiljkovic, Igor Dr. med., Univ. Zagreb Surovoy, Andrey Dr. med., Shemaykin-Inst. Moscow Tschen, Shu-Yuan Dr. rer. nat., Chengdu Videnov, Georgi Dr. rer. nat., Molecular Biology, Sofia
A1-Zähner A1-Zähner B1-Braun B1-Braun A4-Bormann B1-Braun C2-Jung A5-Poralla B4-Schultz A1-Zähner B2-Hamprecht B2-Hamprecht A1-Zähner B1-Braun B1-Braun C2-Jung A1-Zähner
USA Poland Australia Canada Bulgaria South-Korea N. Zealand Italy (China) Switzerland Australia Egypt P. R. China Czech Rep. Croatia Russia P. R. China
87–88 86 89–90 99 91 91 90–2000 98 86 91–92 86 88 90 97 92–93 90–97 91
C2-Jung
Bulgaria
91–95
20.8 International cooperation
(A5) Prof. M. Rohmer, University of Strasbourg, France Prof. L. Cattel, Institute of Applied Pharmacie, University of Turin, Italy Prof. G. D. Prestwich, Dep. Chemistry, University at Stony Brook, New York (A11) Prof. Dr. J. A. Salas, Oviedo, Spain Prof. Dr. P. Leadlay, Cambridge, UK Prof. Dr. K. Ichinose, Tokyo, Japan Prof. Dr. H. G. Floss, Seattle, USA Eli-Lilly and Company Limited, Hampshire, UK Combinature-Biopharm AG, Berlin, Germany Glaxo Wellcome, Stevenage, UK (A12) Prof. Dr. Miguel A. de Pedro, Laboratory for Cell Envelopes, Centro de Biologia Molecular „Severo Ochoa“, Facultad de Ciencias UAM, Campus de Cantoblanco, 28049 Madrid, Spain 364
20.10 Funding (C2) Nijmegen SON Research Center for Molecular Structure, Design and Synthesis University of Nijmegen, The Netherlands Shemyakin Institute of Bioorganic Chemistry, University of Moscow, Russia Laboratoire Nationale de Santé, Luxembourg Laboratory of Microbial Technology, Agricultural University of Norway, As, Norway Ecole Polytechnique Federale de Lausanne, Department de Chimie, EPFLEcublens, Lausanne, Switzerland Groupe RdMN, Institut Pasteur de Lille, Lille, France German-Israel Foundation for Scientific Research & Development, Weizmann Institute Rehovot, Israel
20.9 International conferences
Sekundärmetabolite aus Mikroorganismen. Tübingen 1989. Microbial Secondary Metabolism. Interlaken 1994. „Vectorial Transport Across Bacterial Membranes“. Tübingen 1995. 1. Import, Export and Sorting 2. Protein and Peptide Channels Biologie der Actinomyceten. Tübingen 1996. Plasmide und Genregulation. Blaubeuren 1997. Tübinger/Göttinger Gespräche zur Chemie von Mikroorganismen. Blaubeuren, each year.
20.10 Funding
The collaborative research centre 323 has been supported by grants of the Deutsche Forschungsgemeinschaft totalling DM 35 678 000 in the period 1986– 1999.
365
