Molecular Biotechnology of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases: Vol 88 (Advances in Biochemical Engineering Biotechnology)

88 Advances in Biochemical Engineering / Biotechnology Series Editor: T. Scheper Editorial Board: W. Babel · I. Endo · ...

Author: Axel A. Brakhage

337 downloads 921 Views 4MB Size Report

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

DOWNLOAD PDF

88 Advances in Biochemical Engineering / Biotechnology Series Editor: T. Scheper

Editorial Board: W. Babel · I. Endo · S.-O. Enfors · A. Fiechter · M. Hoare · W.-S. Hu B. Mattiasson · J. Nielsen · H. Sahm · K. Schügerl · G. Stephanopoulos U. von Stockar · G.T. Tsao · C. Wandrey · J.-J. Zhong

Advances in Biochemical Engineering/Biotechnology Series Editor: T. Scheper Recently Published and Forthcoming Volumes

Recent Progress of Biochemical and Biomedical Engineering in Japan II Volume Editor: Kobayashi, T. Vol. 91, 2004 Recent Progress of Biochemical and Biomedical Engineering in Japan I Volume Editor: Kobayashi, T. Vol. 90, 2004 Physiological Stress Responses in Bioprocesses Volume Editor: Enfors, S.-O. Vol. 89, 2004 Molecular Biotechnology of Fungal b -Lactam Antibiotics and Related Peptide Synthetases Volume Editor: Brakhage, A. Vol. 88, 2004 Biomanufacturing Volume Editor: Zhong, J.-J. Vol. 87, 2004 New Trends and Developments in Biochemical Engineering Vol. 86, 2004 Biotechnology in India II Volume Editors: Ghose, T.K., Ghosh, P. Vol. 85, 2003 Biotechnology in India I Volume Editors: Ghose, T.K., Ghosh, P. Vol. 84, 2003 Proteomics of Microorganisms Volume Editors: Hecker, M., Müllner, S. Vol. 83, 2003 Biomethanation II Volume Editor: Ahring, B.K. Vol. 82, 2003

Biomethanation I Volume Editor: Ahring, B.K. Vol. 81, 2003 Process Integration in Biochemical Engineering Volume Editors: von Stockar, U., van der Wielen, L.A.M. Vol. 80, 2003 Microbial Production of l-Amino Acids Volume Editors: Faurie, R., Thommel J. Vol. 79, 2003 Phytoremediation Volume Editor: Tsao, D.T. Vol. 78, 2003 Chip Technology Volume Editor: Hoheisel, J. Vol. 77, 2002 Modern Advances in Chromatography Volume Editor: Freitag, R. Vol. 76, 2002 History and Trends in Bioprocessing and Biotransformation Vol. 75, 2002 Tools and Applications of Biochemical Engineering Science Volume Editors: Schügerl, K., Zeng, A.-P. Vol. 74, 2002 Metabolic Engineering Volume Editor: Nielsen, J. Vol. 73, 2001

Molecular Biotechnology of Fungal b -Lactam Antibiotics and Related Peptide Synthetases Volume Editor : A. A. Brakhage

With contributions by Q. Al-Abdallah · M.S. Barber · M.A. van den Berg · R.A.L. Bovenberg · A.A. Brakhage · J. Casqueiro · H. von Döhren · A.J.M. Driessen · A. Eliasson · M.E. Evers · A. Gehrke · U. Giesecke · N. Gunnarsson · B. Hoff · U. Kück · J.F. Martín · W. Minas · J. Nielsen · H. Plattner · A. Reichert · E.K. Schmitt · P. Spröte · H. Trip · A. Tüncher · R.V. Ullán

2 3

Advances in Biochemical Engineering/Biotechnology reviews actual trends in modern biotechnology. Its aim is to cover all aspects of this interdisciplinary technology where knowledge, methods and expertise are required for chemistry, biochemistry, micro-biology, genetics, chemical engineering and computer science. Special volumes are dedicated to selected topics which focus on new biotechnological products and new processes for their synthesis and purification. They give the state-of-the-art of a topic in a comprehensive way thus being a valuable source for the next 3–5 years. It also discusses new discoveries and applications. In general, special volumes are edited by well known guest editors. The series editor and publisher will however always be pleased to receive suggestions and supplementary information. Manuscripts are accepted in English. In references Advances in Biochemical Engineering/Biotechnology is abbreviated as Adv Biochem Engin/Biotechnol as a journal. Visit the ABE home page at springeronline.com

Library of Congress Control Card Number 2004109334

ISSN 0724-6145 ISBN 3-540-22032-1 DOI 10.1007/b12867 Springer Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer-Verlag. Violations are liable to prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springeronline.com © Springer-Verlag Berlin Heidelberg 2004 Printed in Germany The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typesetting: Fotosatz-Service Köhler GmbH, Würzburg Cover: KünkelLopka GmbH, Heidelberg; design & production GmbH, Heidelberg Printed on acid-free paper

02/3020mh – 5 4 3 2 1 0

Series Editor Professor Dr. T. Scheper Institute of Technical Chemistry University of Hannover Callinstraße 3 30167 Hannover, Germany [email protected]

Volume Editor Professor Dr. Axel A. Brakhage Institute of Microbiology University of Hannover Schneiderberg 50 30167 Hannover, Germany [email protected]

Editorial Board Prof. Dr. W. Babel

Prof. Dr. I. Endo

Section of Environmental Microbiology Leipzig-Halle GmbH Permoserstraße 15 04318 Leipzig, Germany [email protected]

Faculty of Agriculture Dept. of Bioproductive Science Laboratory of Applied Microbiology Utsunomiya University Mine-cho 350, Utsunomiya-shi Tochigi 321-8505, Japan [email protected]

Prof. Dr. S.-O. Enfors

Prof. Dr. A. Fiechter

Department of Biochemistry and Biotechnology Royal Institute of Technology Teknikringen 34 100 44 Stockholm, Sweden [email protected]

Institute of Biotechnology Eidgenössische Technische Hochschule ETH-Hönggerberg 8093 Zürich, Switzerland [email protected]

Prof. Dr. M. Hoare

Prof. W.-S. Hu

Department of Biochemical Engineering University College London Torrington Place London, WC1E 7JE, UK [email protected]

Chemical Engineering and Materials Science University of Minnesota 421 Washington Avenue SE Minneapolis, MN 55455-0132, USA [email protected]

VI

Editorial Board

Prof. Dr. B. Mattiasson

Prof. J. Nielsen

Department of Biotechnology Chemical Center, Lund University P.O. Box 124, 221 00 Lund, Sweden [email protected]

Center for Process Biotechnology Technical University of Denmark Building 223 2800 Lyngby, Denmark [email protected]

Prof. Dr. H. Sahm

Prof. Dr. K. Schügerl

Institute of Biotechnolgy Forschungszentrum Jülich GmbH 52425 Jülich, Germany [email protected]

Institute of Technical Chemistry University of Hannover, Callinstraße 3 30167 Hannover, Germany [email protected]

Prof. Dr. G. Stephanopoulos

Prof. Dr. U. von Stockar

Department of Chemical Engineering Massachusetts Institute of Technology Cambridge, MA 02139-4307, USA [email protected]

Laboratoire de Génie Chimique et Biologique (LGCB), Départment de Chimie Swiss Federal Institute of Technology Lausanne 1015 Lausanne, Switzerland [email protected]

Prof. Dr. G.T. Tsao

Prof. Dr. C. Wandrey

Director Lab. of Renewable Resources Eng. A.A. Potter Eng. Center Purdue University West Lafayette, IN 47907, USA [email protected]

Institute of Biotechnology Forschungszentrum Jülich GmbH 52425 Jülich, Germany [email protected]

Prof. Dr. J.-J. Zhong State Key Laboratory of Bioreactor Engineering East China University of Science and Technology 130 Meilong Road Shanghai 200237, China [email protected]

Advances in Biochemical Engineering/Biotechnology Also Available Electronically

For all customers who have a standing order to Advances in Biochemical Engineering/Biotechnology, we offer the electronic version via SpringerLink free of charge. Please contact your librarian who can receive a password for free access to the full articles by registering at: springerlink.com If you do not have a subscription, you can still view the tables of contents of the volumes and the abstract of each article by going to the SpringerLink Homepage, clicking on “Browse by Online Libraries”, then “Chemical Sciences”, and finally choose Advances in Biochemical Engineering/Biotechnology. You will find information about the – Editorial Board – Aims and Scope – Instructions for Authors – Sample Contribution at springeronline.com using the search function.

Attention all Users of the “Springer Handbook of Enzymes”

Information on this handbook can be found on the internet at springeronline.com A complete list of all enzyme entries either as an alphabetical Name Index or as the EC-Number Index is available at the above mentioned URL.You can download and print them free of charge. A complete list of all synonyms (more than 25,000 entries) used for the enyzmes is available in print form (ISBN 3-540-41830-X).

Save 15% We recommend a standing order for the series to ensure you automatically receive all volumes and all supplements and save 15% on the list price.

Preface

Preface

The concept of one microorganism killing another was introduced by Pasteur who coined the term antibiosis in 1877, but it was much later that this concept was realised in the form of an actual antibiotic. In 1929, the microbiologist Alexander Fleming published his observation about the inhibition of the growth of Staphylococcus aureus on an agar plate contaminated with Penicillium notatum. Three years later, it was shown that the growth inhibition was due to penicillin. The work was taken up further at Oxford University by pathologist Howard Florey and biochemist Ernst Chain. The first clinical trials with penicillin were undertaken in 1941. During the late 1940s the fungus Cephalosporium acremonium (now renamed Acremonium chrysogenum) was isolated from the sea at Cagliari (Italy) by Guiseppi Brotzu. This fungus was found to produce a b-lactam compound designated cephalosporin. The discovery of antibiotics for clinical use started with a b-lactam compound and is perhaps the most important discovery in the history of therapeutic medicine. The application of antibiotics to the therapy of infectious diseases may conceivably have saved more lives than any other medical therapy. The success of b-lactams in the treatment of infectious disease is due to their high specificity and their low toxicity. Despite a growing number of antibiotics and the incidence of penicillin-resistant isolates, b-lactams are still by far the most frequently used antibiotic. In this volume, it was my aim to get together leading scientists in the area of research on fungal b-lactam antibiotics who cover the most recent developments in all areas of research on this important group of compounds. Both the economic aspects and the industrial production of fungal b-lactam antibiotics are summarised in the chapter by Barber et al. Because fungal b-lactam antibiotics are of great clinical importance, the biochemistry and genetics of their biosyntheses are well elucidated which is summarised in several chapters in this book. From an academic point of view the analysis of the regulation of b-lactam biosynthesis represents the most advanced system for elucidating the regulation of a fungal secondary metabolism gene cluster. Furthermore, with regard to applied aspects, this knowledge is of great value to strategies for increasing production levels by the use of gene regulators. Because penicillin and cephalosporin are produced by different fungi and differ in the later steps of the biosynthesis, the regulation of penicil-

X

Preface

lin biosynthesis is described by Brakhage et al. and that of cephalosporin biosynthesis by Schmitt et al. Data presented in these chapters show that we are far from having a complete picture of the regulation of fungal b-lactam biosyntheses.Within the last few years, several studies have indicated that the fungal b-lactam biosynthesis genes are controlled by a complex regulatory network.A comparison with the regulatory mechanisms (regulatory proteins and DNA elements) involved in the regulation of genes of primary metabolism in lower eukaryotes is thus of great interest. Furthermore, such investigations have contributed to the elucidation of signals leading to the production of b-lactams, their physiological meaning for the producing fungi and can be expected to have a major impact on rational strain improvement programs. Recently, the knowledge of the whole cephalosporin biosynthesis was completed by cloning and characterisation of the missing isopenicillin N epimerase system which is reported by Martin et al. Investigations in recent years have shown that the various steps of b-lactam biosynthesis occur in different compartments.This finding is important not only for academic reasons but also for strategies to produce new compounds by metabolic engineering. Therefore, the current knowledge is reviewed by Evers et al. A major aspect of all biotechnological processes concerns their metabolic basis which, for antibiotics, includes the control of fluxes towards antibiotics and the role of primary metabolism in production of antibiotics. This important area is covered by the chapter by Gunnarsson et al. Among the constant challenges in managing bacterial infections are the outbreak of new infectious diseases and the evolution of known commensal and pathogenic bacteria to problem status by acquisition of new resistant determinants. Therefore, there is increasing pressure to provide more and superior antibiotics. A promising approach to identifying novel compounds is based on combinatorial biosynthesis. Because the biosynthesis of b-lactam antibiotics involves a non-ribosomal peptide synthetase (NRPS) which is rather well characterised, this NRPS represents a good starting point for combinatorial biosynthesis of fungal compounds. In addition to the biochemistry of b-lactam biosynthesis, this aspect is discussed in detail in the chapter by Hans von Döhren. Because this book covers the current knowledge of all main aspects of fungal b-lactam antibiotics, I hope it will be a useful reference source for both applied investigators and basic research scientists. I am deeply indebted to the authors of the chapters in this volume for their intelligent and diligent efforts which made this joint project possible. The care and energy with which they approached this work are gratefully acknowledged. I thank the series editor Thomas Scheper for his enthusiasm for preparing this volume and for his excellent scientific co-operation. The editorial and production staff of Springer Verlag are gratefully acknowledged for the fruitful and professional collaboration. Hannover, August 2004

Axel A. Brakhage

Contents

Regulation of Cephalosporin Biosynthesis E.K. Schmitt · B. Hoff · U. Kück . . . . . . . . . . . . . . . . . . . . . . .

1

Regulation of Penicillin Biosynthesis in Filamentous Fungi A.A. Brakhage · P. Spröte · Q. Al-Abdallah · A. Gehrke · H. Plattner · A. Tüncher . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Novel Genes Involved in Cephalosporin Biosynthesis: The Three-component Isopenicillin N Epimerase System J.F. Martín · R.V. Ullán · J. Casqueiro . . . . . . . . . . . . . . . . . . . . 91 Compartmentalization and Transport in b -Lactam Antibiotics Biosynthesis M.E. Evers · H. Trip · M.A. van den Berg · R.A.L. Bovenberg · A.J.M. Driessen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Control of Fluxes Towards Antibiotics and the Role of Primary Metabolism in Production of Antibiotics N. Gunnarsson · A. Eliasson · J. Nielsen . . . . . . . . . . . . . . . . . . . 137 Industrial Enzymatic Production of Cephalosporin-Based b -Lactams M.S. Barber · U. Giesecke · A. Reichert · W. Minas . . . . . . . . . . . . . 179 Biochemistry and General Genetics of Nonribosomal Peptide Synthetases in Fungi H. von Döhren . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 Author Index Volumes 51–88 . . . . . . . . . . . . . . . . . . . . . . . . 265 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Adv Biochem Engin/Biotechnol (2004) 88: 1– 43 DOI 10.1007/b99256 © Springer-Verlag Berlin Heidelberg 2004

Regulation of Cephalosporin Biosynthesis Esther K. Schmitt 1 · Birgit Hoff 2 · Ulrich Kück 2 (✉) 1 2

Novartis Pharma AG, NPU, 4002 Basel, Switzerland Ruhr-Universität Bochum, Lehrstuhl für Allgemeine und Molekulare Botanik, 44780 Bochum, Germany [email protected]

1

Introduction

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

2

2 2.1 2.2 2.3

Precursors and Competing Pathways . . . . . . . . . . . . . . . . . . . . . Marks a Biosynthesis Branch Point . . . L-Valine as a Metabolic Signal . . . . . . . . . . . . . . . . . . . . . . . . . Non-Conventional Biosynthesis of L-Cysteine . . . . . . . . . . . . . . . . .

3 3 5 5

3 3.1 3.1.1 3.2 3.2.1

Biosynthesis of Cephalosporin . . . . . . . . General b-Lactam Biosynthesis . . . . . . . Cellular Localization and Structure of IPNS . Cephalosporin Specific Biosynthesis . . . . . Final Reaction of Cephalosporin Biosynthesis

. . . . .

7 8 10 11 12

4 4.1 4.2

Structural Genes of Cephalosporin Biosynthesis . . . . . . . . . . . . . . . “Early” Cephalosporin Genes . . . . . . . . . . . . . . . . . . . . . . . . . . “Late” Cephalosporin Genes . . . . . . . . . . . . . . . . . . . . . . . . . .

13 15 17

5 5.1 5.2 5.3

Multiple Layers of Control . . . . . . . . . . . . . . . . . . . . . Transcript Level . . . . . . . . . . . . . . . . . . . . . . . . . . . Enzyme Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . Correlation Between Secondary Metabolism and Morphogenesis

. . . .

. . . .

. . . .

. . . .

. . . .

18 18 20 20

6 6.1 6.2 6.3 6.4

Transcription Factors as Activators and Repressors of Cephalosporin Biosynthesis . . . . . . . . . . . . . . . . . . . . . PACC – pH-Dependent Transcriptional Control . . . . . . . . . . . . CRE1 – A Glucose Repressor Protein . . . . . . . . . . . . . . . . . . CPCR1 – Cephalosporin C Regulator 1 . . . . . . . . . . . . . . . . Comparison of Cephalosporin and Penicillin Biosynthesis Regulation

. . . .

. . . . .

. . . . .

. . . . .

22 22 24 26 30

7

Molecular Differences in Production Strains . . . . . . . . . . . . . . . . .

30

8 8.1 8.2

Examples of Molecular Engineering of A. chrysogenum . . . . . . . . . . . Genetic Tools for Molecular Engineering . . . . . . . . . . . . . . . . . . . Optimization of Cephalosporin C Biosynthesis . . . . . . . . . . . . . . . .

33 33 35

9

Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

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

38

L-a-Aminoadipic Acid (L-a-AAA)

References

. . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . .

. . . . .

. . . . .

. . . . .

. . . . .

2

E. K. Schmitt et al.

Abstract The filamentous fungus Acremonium chrysogenum is the natural producer of the b-lactam antibiotic cephalosporin C and is as such used worldwide in major biotechnical applications. Albeit its profound industrial importance, there is still a limited understanding about the molecular mechanisms regulating cephalosporin biosynthesis in this fungus. This review focuses on various regulatory levels of cephalosporin biosynthesis. In addition to precursor and antibiotic biosynthesis, molecular genetic characteristics of cephalosporin biosynthesis genes and the knowledge of multiple layers of their regulatory expressional control, as well as the function of activators or repressors on cephalosporin biosynthesis are jointly being surveyed. Furthermore, this review summarizes (i) molecular features, which distinguish strains with different production levels and (ii) examples of molecular engineering approaches to A. chrysogenum. Keywords Acremonium chrysogenum · Cephalosporin · Gene regulation · Transcription factors · Genetic engineering

1 Introduction Cephalosporin C and its semisynthetic derivatives are very potent and widely used b-lactam antibiotics of general and applied interest. However, the knowledge of the molecular regulation of b-lactam biosynthesis in the corresponding host is still limited. In the case of cephalosporin biosynthesis, even the total number of involved biosynthesis genes is not known and has yet to be identified. Cephalosporin is exclusively produced by Acremonium chrysogenum (syn. Cephalosporium acremonium), but compared to other filamentous fungi, genetic manipulation of this fungus is rather difficult. Acremonium chrysogenum belongs to the Deuteromycetes, which lack a sexual cycle and are thus not accessible for any conventional genetic analysis. In addition, this fungus produces only very few conidiospores, which in other biotechnically relevant fungi are the preferred cells for DNA-mediated transformations. In 1945, A. chrysogenum was first isolated from Sardinian coastal seawater by Prof. Brotzu. Brotzu was also the first to describe the antibiotic effect of extracts generated from this fungus and, some years later, the structure of the active compound was determined [1]. Cephalosporin C was shown to be active against Gram-positive as well as Gram-negative bacteria. Today, A. chrysogenum is cultured worldwide to yield approximately 2500 tons of cephalosporin derivatives. Semisynthetic derivatives are mainly used as broad-spectrum antibiotics for the treatment of bacterial infections. In biotechnical applications, intensive strain improvement programs resulted in production strains that yield a significantly higher titer of the antibiotic than wild-type strains. Approximately 40 years of mutation and selection cycles separate today’s industrial strains from the genetic potential of the original isolates. For basic as well as for applied research, the comparison of wild-type and production strains is of specific interest when differences of

Regulation of Cephalosporin Biosynthesis

3

cephalosporin biosynthesis regulation are being investigated. A deeper knowledge of regulatory changes that occurred during strain improvement of cephalosporin production strains can be highly valuable for the directed improvement of novel, so far not optimized, fungal antibiotic producers by genetic engineering. Future work will show whether or not further significant improvements of cephalosporin production strains are feasible. One perspective is a combined approach, which uses genetic engineering techniques together with conventional strain improvement procedures. The following sections focus on molecular and genetic mechanisms of cephalosporin biosynthesis that were elucidated in recent years. This review starts with a summary of precursors of cephalosporin biosynthesis and their competing pathways, followed by an overview of the biosynthesis and the structural genes involved in the production of cephalosporin C. Then an outline of regulatory parameters and mechanisms is given, and the transcriptional control of the biosynthesis genes by transcription factors is detailed in section 6. The last two sections deal with the molecular differences that occurred during classical strain improvement of industrial strains and attempts to use a rational approach via molecular engineering.

2 Precursors and Competing Pathways The biosynthesis of all occurring b-lactams is primarily based on the three amino acids L-a-aminoadipic acid (L-a-AAA), L-cysteine and L-valine. These amino acids play also an important role in the regulation of the cephalosporin C biosynthesis. L-Cysteine and L-valine are ubiquitous amino acids, whereas the non-proteinogenic amino acid L-a-AAA is synthesized as an intermediate in the L-lysine biosynthesis pathway. 2.1 a -AAA) Marks a Biosynthesis Branch Point a -Aminoadipic Acid (L-a L-a In fungi, the non-proteinogenic amino acid L-a-AAA is synthesized by a specific aminoadipate pathway, which leads to the formation of lysine, whereas in b-lactam producing bacteria, a specific pathway for the formation of L-a-AAA has been identified (reviewed in [2, 3]). The L-lysine biosynthesis pathway in higher fungi, including A. chrysogenum, starts with the condensation of a-ketoglutarate and acetyl-CoA to form homocitrate, which is then subjected to isomerization, oxidative decarboxylation and amination to yield L-a-AAA. Subsequently, this precursor amino acid is converted into a-AA-d-semialdehyde by the action of the a-aminoadipate reductase (a-AAR) to finally form L-lysine [4–6]. Furthermore, L-a-AAA can also be obtained for b-lactam biosynthesis by reversal of the last steps of the L-ly-

4


sine biosynthesis pathway; however, the influence of this catabolic pathway on cephalosporin production remains to be shown [7]. Since L-a-AAA marks the branch point between cephalosporin and the competing L-lysine biosynthesis pathway, its intracellular availability is an important parameter in the regulation of cephalosporin biosynthesis. Mehta et al. [8] showed that L-lysine concentrations reduce the synthesis of cephalosporin C in A. chrysogenum and that this inhibition is derepressed by L-a-AAA. Furthermore, recent studies demonstrated that L-lysine concentrations inhibit a-aminoadipate reductase (a-AAR) activity but do not repress its synthesis [9]. These results and the fact that L-lysine caused inhibition of the homocitrate synthase in Penicillium chrysogenum indicated that the L-aAAA pool available for b-lactam production is reduced by L-lysine through feedback inhibition or through repression of several L-lysine biosynthesis genes and enzymes [10]. The initiation of the ACV tripeptide formation depends not only on the availability of L-a-AAA but also on the affinity of the two enzymes for this intermediate. The a-aminoadipate reductase (a-AAR) encoded by the lys2 gene of A. chrysogenum acts as a key enzyme in the branched pathway for lysine and cephalosporin C biosynthesis, since it competes with ACVS for their common substrate L-a-AAA. a-AAR catalyzes the activation and reduction of L-a-AAA to its a-AA-d-semialdehyde using NADPH as cofactor [11, 12]. Hijarrubia et al. [9] revealed that a lower a-AAR activity could be detected in high cephalosporin producing strains of A. chrysogenum. It was suggested that this lower activity might lead to channeling of L-a-AAA towards the formation of cephalosporin. These results concur with the increased availability of the precursor amino acid L-a-AAA, suggesting that more L-a-AAA is shifted from the primary metabolism (lysine formation) to a higher cephalosporin yield in production strains [13]. Furthermore, the a-AAR activity peaked during the growth phase preceding the onset of cephalosporin production and then drastically decreased. At the end of the growth phase, a metabolic switch appears to occur that correlates with an increased availability of L-a-AAA for its use as precursor of cephalosporin production. This switch also coincides with the beginning of mycelium fragmentation into arthrospores in A. chrysogenum [9]. Immunoblotting analysis has shown a strong negative effect of nitrate on a-AAR formation. A possible explanation could be the requirement of large amounts of NADPH by the nitrate reductase [14]. Such activity would constitute a competitive inhibitor for the reduction of L-a-AAA to its semialdehyde. The possible reversal of the nitrate effect by lysine addition [9] can be explained by the well-known fact that lysine represses nitrate uptake as well as the metabolic route from nitrate to ammonium [15, 16]. Thus, the L-a-AAA biosynthesis pathway in A. chrysogenum is regulated by several control mechanisms such as the feedback inhibition at the a-AAR or homocitrate synthase. However, there is a decided lack of knowledge concerning the L-lysine pathway and its influence on the cephalosporin C production in A. chrysogenum.


5

2.2 L-Valine as a Metabolic Signal Another crucial factor for the initiation of the ACV tripeptide formation is the availability of the precursor amino acid L-valine. The biosynthesis pathway of this ubiquitous amino acid is closely connected to the biosynthesis of leucine. Valine biosynthesis comprises four enzymatic steps with pyruvate as precursor metabolite. Two moles of pyruvate are converted to the intermediate a-acetolactate, which is then reduced to a, b-dihydroxyisovalerate and ketoisovalerate to finally form L-valine. In A. chrysogenum, high levels of L-valine result in a feedback inhibition of the first reaction step catalyzed by acetohydroxy acid synthase [17]. So far, no further data on the regulation of the L-valine biosynthesis pathway and its competing effect on the cephalosporin C biosynthesis have become available. 2.3 Non-Conventional Biosynthesis of L-Cysteine Another limiting step for cephalosporin C biosynthesis is the availability of the amino acid L-cysteine, which can generally be formed through four different biosynthesis pathways (reviewed in [18–20]). In the direct sulfhydrylation pathway, reduced sulfur is incorporated into the intermediate O-acetyl-L-serine to give L-cysteine, whereas in the transsulfuration pathway, sulfide incorporation is catalyzed by O-acetylhomoserine sulfhydrylase. The third possibility is the reverse transsulfuration in which the sulfur of L-methionine is transferred to L-cysteine via four intermediates [21] (see Fig. 1). The incorporated sulfur is known to be the efficient precursor of the sulfur atom contained in cephalosporin C [22]. In addition, L-cysteine is synthesized by the so-called autotrophic pathway, which leads to the assimilation of inorganic sulfur via serine O-acetyltransferase and O-acetylserine sulfhydrylase [23, 24]. All of these pathways seem to exist in A. chrysogenum [19]. However, results of mutant analysis showed that the fungus prefers to generate L-cysteine for optimal cephalosporin C biosynthesis via the reverse transsulfuration pathway, which has been detailed in Fig. 1 [25], and to a certain extent via the autotrophic pathway [26]. The relative contributions of the two pathways to the cephalosporin C biosynthesis are still to be determined. High levels of methionine, particularly the D-isomer, significantly stimulate the synthesis of b-lactam antibiotics. In methionine-supplemented cultures of A. chrysogenum, a two to threefold increase in cephalosporin C titers was determined [27].Additionally, a transient enlargement of the endogenous pool of methionine has been observed in advance of cephalosporin C formation, and the specific biosynthesis seemed to be proportional to the intracellular D-methionine concentration [28]. The addition of high levels of methionine is necessary to achieve optimum cephalosporin C biosynthesis, possibly due to methionine degradation by the intracellular amino acid oxidases [29, 30].

Fig. 1 Biosynthesis of L-cysteine in A. chrysogenum. ‘Reverse transsulfuration’ is the preferred pathway to generate L-cysteine in A. chrysogenum. Alternatively, sulfate assimilation is used, while ‘transsulfuration’ and ‘direct sulfhydrylation’ seem to exist in A. chrysogenum, but are not used for L-cysteine biosynthesis


7

Early analyses have shown that the enzyme cystathionine-g-lyase, which catalyzes the conversion of cystathionine to L-cysteine in the reverse transsulfuration is crucial for the methionine induced titer-enhancing effect. This reaction was proposed to induce the transfer of L-cysteine from the primary metabolism to the cephalosporin C biosynthesis pathway [31]. In recent studies, the so-called mecB gene encoding cystathionine-g-lyase was cloned from A. chrysogenum. The encoded protein was shown to be functional by complementing the Aspergillus nidulans C47 mutant, which is defective in cystathionine-g-lyase activity. The expression of the mecB gene is not regulated by the addition of DL-methionine [32]. Targeted inactivation of the mecB gene indicated that the supply of L-cysteine through the reverse transsulfuration pathway is required for high-level cephalosporin C production but not for low-level biosynthesis proving that the essential L-cysteine is obtained from both the autotrophic and the reverse transsulfuration pathways [33]. The supply of methionine results in the complete repression of sulfate assimilation [34]. mecB-disruption did not affect the methionine induction of the cephalosporin C biosynthesis genes. Thus, their expression is not mediated by a putative regulatory mechanism exerted by cystathionine-g-lyase, but the induction may be triggered by methionine itself or by a catabolite derived from methionine [33]. Amplification of the mecB gene and the resulting overproduction of the cystathionine-g-lyase in moderate doses lead to an increased cephalosporin C formation, whereas high cystathionine-g-lyase activity is likely to produce high intracellular levels of L-cysteine, which are known to be toxic and inhibit b-lactam synthesizing enzymes [35]. Taken together, methionine presumably has a double effect on cephalosporin C biosynthesis in A. chrysogenum. On the one hand it seems to be the main supplier of L-cysteine via the reverse transsulfuration pathway and on the other hand it has an induction effect on cephalosporin biosynthesis genes (reviewed in [36, 37]).

3 Biosynthesis of Cephalosporin Cephalosporins are members of the large group of b-lactam antibiotics, which inhibit the growth of Gram-negative as well as Gram-positive microorganisms at already low concentrations. b-lactam antibiotics are specified by the typical cephem nucleus and are produced by a wide variety of microorganisms, including the filamentous fungus A. chrysogenum, Gram-positive streptomycetes and a small number of Gram-negative bacteria (reviewed in [38]). All of them produce b-lactams essentially through the same biosynthesis pathway, which is chemically and kinetically well characterized owing to the considerable industrial potential of these antibiotics [39].

8


3.1 General b -Lactam Biosynthesis As shown in Fig. 2, the first step of cephalosporin biosynthesis results in the formation of the tripeptide d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (ACV) from the amino acid precursors and is catalyzed by a single multifunctional enzyme designated d-(L-a-aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS). ACVS are monomers with a molecular mass of about 420 kDa, which function similarly to other peptide synthetases from bacterial or fungal sources. They mediate the non-ribosomal synthesis of peptides via a multiple carrier thiotemplate mechanism [40–42]. ACVS contains three repeated modules with conserved amino acid sequences [43]. Each module consists of functional domains for amino acid recognition, activation and thiolation. During condensation, peptide bond formation occurs from the amino to the carboxy terminus of the peptide. In addition, the last module of the ACVS contains an epimerization module, which is involved in the conversion of the activated intermediates ([41], reviewed [44]). A detailed analysis showed that ACVS catalyzes the activation of the carboxyl group of the first amino acid in the presence of Mg2+ and ATP by the formation of the corresponding aminoacyl adenylate and the release of pyrophosphate [45]. This step is followed by the transfer of the activated carboxyl group to the 4¢-phosphopanthetheine cofactor to generate the thioester bond between the enzyme and the amino acid. This thioesterified amino acid represents the target for nucleophilic attack by the amino group of the second amino acid, resulting in the formation of the first peptide bond between the Laminoadipic acid and L-cysteine. The resulting dipeptidyl intermediate remains bonded to the enzyme. After condensation of the dipeptide with the third

Table 1 Designation of genes, which have been isolated and characterized from Acremonium chrysogenum

Gene abbreviation

Product

pcb AB (syn. acvA) pcbC (syn. ipnA) cefD1 cefD2 cefEF cefG lys2 mecB cpcR1 cre1 pacC

δ-(L-α-aminoadipyl)-L-cysteinyl-D-valine synthetase isopenicillin N synthase acyl-CoA-synthetase acyl-CoA-racemase deacetoxycephalosporin C/deacetylcephalosporin C synthetase acetyl-CoA: deacetylcephalosporin C acetyltransferase α-aminoadipate reductase cystathionine-γ-lyase cephalosporin C regulator 1 carbon catabolite repressor CRE1 pH-dependent transcription factor PACC


9

Fig. 2 Cephalosporin C biosynthesis, which exclusively occurs in A. chrysogenum. Biosynthesis genes as well as their products were framed.With the exception of the predicted gene encoding a thioesterase, all others have been cloned. For details see main text

10


amino acid, L-valine is epimerized at the tripeptide stage to its D-enantiomer and is followed by the formation of the final ACV tripeptide. The selective release of the tripeptide with the correct LLD configuration is arranged via the integrated thioesterase domain in the C-terminal region of ACVS [41, 46, 47]. The second reaction is a key step in the cephalosporin biosynthesis pathway, which implies the cyclization of the linear ACV tripeptide to form the first bioactive intermediate isopenicillin N (IPN) (see also Fig. 2). This reaction is mediated by the isopenicillin N synthase (IPNS), a nonheme monoferrous-dependent oxidase of a molecular mass of about 38 kDa, which binds ferrous iron, uses dioxygen as co-substrate and ascorbate as electron donor to form the bicyclic nucleus [48, 49]. In a unique enzymatic reaction, IPNS catalyzes the transfer of four hydrogen atoms from the precursor ACV tripeptide to dioxygen associated with the desaturative ring closure and the formation of two water molecules [38, 49, 50]. X-ray crystallography determined that the four-membered b-lactam ring system is primarily formed in conjunction with a highly oxidized iron-oxo (ferryl) group, which then mediates the closure of the corresponding thiazolidine ring [51, 52]. 3.1.1 Cellular Localization and Structure of IPNS The IPNS enzyme is localized in the cytoplasm as a soluble protein [53]. It exists in two interconvertible forms, one is an oxidized state forming a disulfide linkage and the other exists in a reduced state [54]. IPNS consists of a catalytic center containing a highly conserved H-Xaa-D-(53–57 residues)-Xaa-H motif for iron coordination and of a specific substrate-binding pocket with a common R-X-S motif crucial for its catalytic activity [49, 55–57].A third amino acid residue tyrosine (189–191) is also involved in binding of the valine carboxylate moiety of the ACV tripeptide, but it is not as crucial as the R-X-S motif [58]. Analysis of the crystal structure has shown that the active site is unusually buried within an eight-stranded “jelly-roll” motif and lined by hydrophobic residues [49]. This structural characteristic of the IPNS proteins and many other keto-acid-dependent oxygenases is probably necessary for the isolation of the reactive complex and of subsequent intermediates from the external environment. Combined application of Mössbauer electron paramagnetic resonance as well as nuclear magnetic resonance spectroscopy, has determined a mechanism for isopenicillin N formation. This involves direct ligation of ACV to the active iron site of the IPNS via the corresponding cysteinyl thiol, or more precisely, via the sulfur atom of the ACV [59, 60] and the creation of a vacant iron coordination site into which dioxygen may bind. The binding of ACV leads to the initiation of the reaction and the replacement of the amino acid residue Q 330 side chain, which coordinates the metal in absence of a substrate. Subsequently, iron-dioxygen and iron-oxo species remove the essential hydrogens from ACV [49, 55, 61]. Thus, in the generated Fe2+:ACV:IPNS complex, three of


11

the five coordination sites are occupied with protein ligands. The remaining two sites are filled by a water molecule and the ACV thiolate resulting in a penta-coordinated iron active site (reviewed in [62]). In this reaction step, only the thiol form of the ACV tripeptide serves as a substrate, the spontaneously formed bis-disulfide state shows no binding activity [63]. Due to the broad substrate specificity of IPNS in particular with alterations in the L-AAA as well as with the valine residue of ACV tripeptide, it is possible to generate new penicillins/cephalosporins in vivo or to generally improve the enzyme activity [64]. 3.2 Cephalosporin Specific Biosynthesis The formation of isopenicillin N is the branch point of penicillin and cephalosporin biosynthesis. The reaction step to follow is shown in Fig. 2 and leads to the formation of penicillin N. This step establishes the pathway that is specific for the synthesis of cephalosporins. An epimerization reaction is involved that converts isopenicillin N to penicillin N, which, despite its industrial relevance, had remained uncharacterized for a long time. Konomi et al. [65] has first shown epimerase activity in cell-free extracts of the cephalosporin C producing fungus A. chrysogenum, but it was suggested that the epimerizing enzyme was extremely instable preventing purification of the protein [66–68]. Until recently, no further data on the fungal enzyme have been obtained. Since all known cephalosporin biosynthesis genes of A. chrysogenum are clustered in two separate loci, Ullán et al. [69] suggested that the gene encoding the enzyme involved in the conversion of isopenicillin N into penicillin N might be located in one of the cephalosporin gene clusters. A transcriptional analysis of a 9 kb region located downstream of the pcbC gene revealed the presence of two open reading frames that were cloned and sequenced on both strands. ORF1 corresponds to the gene designated cefD1 and encodes a protein with a molecular mass of about 71 kDa, which shows a high degree of similarity to long chain acyl-CoA synthetases, particularly to those from Homo sapiens (26.3% identity), Rattus norvecigus or Mus musculus (25.5% identity). The encoded protein contains all characteristic motifs of the acyl-CoA ligases involved in the activation of the carboxyl moiety of fatty acids or amino acids [70]. The second identified gene designated cefD2 encodes a protein with a deduced molecular mass of 41.4 kDa, which is similar to a-methyl-acyl-CoA racemases from H. sapiens (42.1% identity) or M. musculus (39.4% identity) [71–73]. Based on the identified homology of the CEFD1 and CEFD2 proteins with known eukaryotic enzymes, it seems feasible to establish a mechanism for the A. chrysogenum two-component epimerization system which is different from epimerizations found in prokaryotes. Such systems have been reported to be involved for example in the inversion of 2-arylpropionic acids (e.g. ibuprofen), which is an important group of non-steroidal anti-inflammatory drugs in hu-

12


mans [74–76]. Therefore, it was suggested that the epimerization reaction in the cephalosporin biosynthesis pathway begins with the activation of the substrate isopenicillin N to its CoA-thioester by the acyl-CoA-synthetase. The product of the cefD2 gene, the a-methylacyl-CoA racemase, catalyzes the epimerization of isopenicillinyl-CoA to D-enantiomer penicillinyl-CoA. Finally, the required hydrolysis of the CoA-thioesters seems to occur in a nonstereoselective manner by different thioesterases [75]. The resulting product, penicillin N, is the direct precursor of all cephalosporins and cephamycins and, thus, available as a substrate for further reactions in the biosynthesis pathway. The next committed step of the cephalosporin pathway leads to the conversion of penicillin N to deacetoxycephalosporin C (DAOC) by expanding the five-membered thiazolidine ring to the six-membered dihydrothiazine ring characteristic for the class of cephalosporins. This reaction is catalyzed by DAOC synthetase, which ensures the required expandase function [77]. In the following reaction of the biosynthesis pathway, DAOC hydroxylase, also designated deacetylcephalosporin C synthetase, catalyzes the incorporation of an oxygen atom from O2 into the exocyclic methyl moiety at the C-3 atom of DAOC thus forming deacetylcephalosporin C (DAC) (reviewed in [78–81]). The enzymatic expansion of the five-membered thiazolidine ring was first observed in cell-free extracts of the cephalosporin C producer A. chrysogenum [77, 82]. The enzyme involved is responsible for the two-step reaction in A. chrysogenum, which leads to the conversion of penicillin N to deacetylcephalosporin C, while in streptomycetes like Streptomyces clavuligerus [83, 84], the two enzymatic activities could be distinctly separated by anion-exchange chromatography [85]. Analysis of the amino acid sequence of the DAOC/DAC synthetase of A. chrysogenum revealed a ten amino acid region containing a cysteine residue at position 100, which is 50% identical to the corresponding region containing the cysteine residue at position 106 of isopenicillin N synthetase. This region is of special interest because the cysteine residue of the IPNS is important for substrate binding and specific activity [86]. Thus, it seems to be possible that the corresponding residue C-100 of the DAOC expandase/hydroxylase may either be directly or indirectly involved in substrate binding [87]. The existent sulfhydryl groups in the enzyme were apparently essential for both ring expansion and hydroxylation [80]. In addition to penicillin N, the DAOC/DAC synthetase exhibits a diverse substrate specificity, which differs in the efficiency of ring expansion [88]. 3.2.1 Final Reaction of Cephalosporin Biosynthesis In the last reaction of the cephalosporin biosynthesis pathway, the transfer of an acetyl moiety from the acetyl coenzyme A to the hydroxyl group on the sulfur-containing ring of deacetylcephalosporin C leads to the formation of the


13

final product cephalosporin C, which possesses high antibiotic activity [89–92]. This acetylation reaction is catalyzed by the acetyl-coenzyme A (CoA):DAC acetyltransferase, which behaves like a soluble cytosolic enzyme without any known targeting signals or other indications for compartmentalization [93]. The pure enzyme shows a molecular mass of about 50 kDa and the Nterminal end possesses the sequence M-P-S-A-Q-V-A-R-L, which perfectly matches the deduced amino acid sequence starting at the first ATG codon. This enzyme seems to be a monomer, which shows no dissociation into subunits. The amino acid sequence of the A. chrysogenum acetyl-CoA:DAC acetyltransferase reveals significant similarity with sequences of several O-acetyltransferases, especially with homoserine-O-acetyltransferases of the fungi Saccharomyces cerevisiae and Ascobolus immersus (55.8% and 48.5% identity) [94, 95]. This is probably due to the structural similarity of the exocyclic CH2OH moiety in DAC and the homoserine molecules [96]. Even though this similarity results from mutant analysis, it suggests that there are two independent Oacetyltransferases for DAC and homoserine in A. chrysogenum [19]. The acetylation reaction of DAC to cephalosporin C seems to be very inefficient in most strains of A. chrysogenum. The cefG gene is expressed very poorly when compared with other genes of the pathway [96, 97] and it is well known that high levels of DAC accumulated in many cephalosporin C producing strains [90]. Consequently, the conversion of DAC to cephalosporin C seems to be the limiting step in the pathway.

4 Structural Genes of Cephalosporin Biosynthesis As shown in Fig. 3, the genes involved in cephalosporin biosynthesis are organized in at least two clusters in A. chrysogenum. The pcbAB and pcbC genes as well as the newly discovered cefD1 and cefD2 genes are linked in the socalled “early” cephalosporin cluster. The “late” cluster contains the cefEF and cefG genes, which are involved in the last two steps of the biosynthesis pathway (reviewed in [69, 98]). In most strains, the “early” cluster could be mapped to chromosome VI and the “late” cluster to chromosome II [99, 100]. Analysis of high cephalosporin producing strains, such as C10, has shown a different localization of the biosynthesis gene cluster on chromosomes I and VII [101], which indicates that significant chromosome rearrangements have occurred during strain improvement (reviewed in [98]). Both cephalosporin clusters are available as a single copy in the genome of all analyzed A. chrysogenum strains. The cluster formation of biosynthesis genes in many antibiotic producing organisms gives rise to the hypothesis that linkage has occurred during evolution conferring an ecological selective advantage [102]. Furthermore, the or-

Fig. 3 Gene organization of the ‘early’ and ‘late’ cephalosporin biosynthesis genes in A. chrysogenum. Intronic sequences are kept in white and the arrows indicate the direction of transcription

14 E. K. Schmitt et al.


15

ganization of the biosynthesis genes into large operons controlled by a single promoter could allow a coordinated regulation of the biosynthesis genes [103]. However, in eukaryotic fungi like A. chrysogenum, b-lactam biosynthesis genes are transcribed separately and are expressed through different promoters (reviewed in [50]). In this light, gene expression does not seem to be coordinated as a result of genomic linkage and, hence, it seems more likely that cluster formation reflects only a common ancestral origin. 4.1 “Early” Cephalosporin Genes ACVS, the first acting enzyme of the cephalosporin C pathway in A. chrysogenum, is encoded by a single structural gene designated pcbAB (syn. acvA) with a size of about 11 kb. The pcbAB gene was first identified and cloned in P. chrysogenum by complementation of mutants blocked in penicillin biosynthesis and by transcriptional mapping of the genome [104]. For the cephalosporin C producer A. chrysogenum, a 32 kb DNA fragment was identified in several phages using the pcbAB and pcbC genes of P. chrysogenum as heterologous probes [43]. Complementation studies using the npe5 strain from P. chrysogenum carrying a defective pcbAB gene confirmed in vivo that the functional pcbAB gene is located on a 15.6 kb fragment. Northern analysis of total RNA using probes internal to the pcbAB gene identified a transcript of 11.4 kb. It has been shown that the pcbAB open reading frame of 11136 bp, which matched the 11.4 kb transcript initiation and termination regions, was located upstream of the pcbC gene. The ORF does not contain any intron sequences and the translational start codon of the gene is not yet clearly defined, because attempts to obtain the Nterminal amino acid sequence have been unsuccessful so far [105, 106]. In A. chrysogenum, the pcbAB gene is separated by an intergenic region of 1233 bp from the pcbC gene and is divergently orientated with respect to pcbC [43]. In industrial strains of A. chrysogenum, the expression of the pcbAB gene was much weaker compared to that of pcbC [107]. Furthermore, differences may even exist in the temporal expression among genes of the same cluster. In A. chrysogenum, the pcbAB gene seems to be coordinately regulated with the pcbC gene, whereas the later genes of the pathway appear to be sequentially induced [108–110]. Disruption of the pcbAB gene in A. chrysogenum resulted in loss of ACVS activity without affecting the other cephalosporin biosynthesis genes [110]. The corresponding pcbAB genes were cloned and sequenced from different prokaryotic and eukaryotic microbial b-lactam producers [2, 104, 111–114]. A comparison of the nucleotide sequences encoding the three repeated domains [115] demonstrated a high similarity among the fungal and bacterial pcbAB genes. The fungal domains showed on average 71% nucleotide sequence identity to each other, whereas fungal and bacterial domains revealed about 48% identity. Only little similarity was found between domain-

16


separating regions. These results imply a close relationship between all pcbAB genes [116, 117]. The second structural gene of the “early” cephalosporin gene cluster in A. chrysogenum is the pcbC (syn. ipnA) gene, which encodes the IPNS enzyme. The pcbC gene is divergently linked by an intergenic promoter region to the pcbAB gene. The pcbC gene of A. chrysogenum was the first b-lactam biosynthesis gene to be cloned and sequenced [118]. This was achieved by purification of the IPNS enzyme, determination of the N-terminal amino acid sequence and the design of two pools of oligonucleotides, which contain all possible sequences encoded by two short peptides of the IPNS amino terminus. Using this DNA as probe for screening a cosmid library, it was possible to identify a clone possessing an ORF with a size of 1014 bp, which encodes a protein of 338 amino acids. Expression of this ORF in E. coli resulted in IPNS activity of the corresponding cell extracts [118]. The pcbC gene does not contain any introns and the corresponding transcript size was determined to lie between 1.15 and 1.5 kb. While primer extension established two major [–56 and –77] and at least two minor transcription start sites [–58 and –78], the corresponding values obtained from S1 endonuclease mapping were –51/–73 and –54/–80/–97, respectively [118, 119]. These transcription initiation sites appeared as major and minor pairs on either side of one of the pyrimidine-rich blocks, which characterize the promoter sequence. After identification of the pcbC gene in A. chrysogenum, the corresponding structural genes have been cloned and sequenced from several different fungi and bacteria, such as P. chrysogenum, A. nidulans or S. clavuligerus (reviewed in [3, 38]).Alignments using pcbC sequences from different prokaryotic and eukaryotic organisms revealed a degree of identity greater than 60%. Most of the sequence identity is scattered throughout the protein, which makes it difficult to identify functionally important domains [57, 120, 121]. The “early” cephalosporin gene cluster was completed with the newly discovered genes cefD1 and cefD2, which encode two enzymes that act in a twoprotein system for formation of the cephalosporin C intermediate penicillin N. Identification of the two genes was obtained by transcriptional studies of a 9 kb region located downstream of the pcbC gene. Analysis was performed using RNA extracted from mycelia of A. chrysogenum strain C10 grown for 48 h, since at this time the expression of the other cephalosporin C biosynthesis genes is known to be high. A 5.8 kb subfragment containing two open reading frames was cloned and sequenced on both strands. ORF1 corresponded to the gene cefD1 with a size of 2193 nucleotides and was interrupted by the presence of five introns with sizes varying between 28 and 150 bp. The corresponding transcript revealed a size of 2 kb. This gene was also cloned from a previously constructed cDNA library [122] and the sequence confirmed the presence of five introns. The cefD2 gene consists of 1146 nucleotides and is interrupted by the presence of a single intron with a size of 92 bp. RT-PCR and sequence analysis have shown that the intron had been removed at the splicing sites corresponding to nucleotides 64–157 relative to the


17

ATG translation initiation codon. The corresponding transcript could be detected in the 48 h cultures with a size of 1.2 kb. The cefD1 and cefD2 genes, which are located closely downstream the pcbC gene are expressed in opposite orientation from a bi-directional promoter region with a size of 1515 bp, which is characteristic for cephalosporin C biosynthesis genes. Functional analyses of the cefD1 and cefD2 genes have been performed by targeted inactivation of both genes using DNA-mediated transformation and resulting in strains lacking iso-penicillin epimerase activity [69]. 4.2 “Late” Cephalosporin Genes The cefEF gene of A. chrysogenum, which encodes DAOC/DAC synthetase is one of the two genes organized in the “late”cephalosporin cluster. Isolation and characterization of the gene was achieved by the design of oligonucleotide probes based on the amino acid sequence of the purified DAOC/DAC synthetase and a subsequent screen of a cosmid genomic library of A. chrysogenum. One ORF with a size of 996 bp coding for a protein of 332 amino acids matched the sequence predicted for the peptide fragments. After expression of this ORF in E. coli, cell extracts harbored both expandase and hydroxylase activities [87]. The cefEF gene does not possess any intron sequences suggesting a prokaryotic origin. In antibiotic producing streptomycetes, a clearly different system was detected. There are two different genes designated cefE and cefF encoding two different enzymes namely DAOC expandase and hydroxylase. Both genes are linked together with the cefD gene in a single cluster (reviewed in [38]). In A. chrysogenum, the cefEF gene is closely linked to the cefG gene, but it is transcribed in the opposite direction. The intergenic region with a size of 938 bp contains the promoters for both genes [96, 123]. The cefG gene of A. chrysogenum encoding the last enzyme of the cephalosporin C pathway, namely the acetyl-CoA:DAC acetyltransferase, was cloned and sequenced independently by three research groups [96, 97, 124, 125]. Mathison and co-workers [97] achieved the cefG gene isolation by sequencing the ambient region of the cefEF gene and identification of an open reading frame. An alternative for gene cloning was the screening of an A. chrysogenum lambda phage library with a probe specific for the cefEF gene. Northern blotting and DNA sequence analysis revealed the existence of the cefG gene close to the cefEF gene [96]. In both cases, the identity of the cefG gene was demonstrated by complementation of A. chrysogenum mutants, which are deficient in acetyl-CoA:DAC acetyltransferase activity. In addition, overexpression of the gene in Aspergillus niger, which lacks these genes, demonstrated such an activity. Matsuda et al. [124] used another strategy by screening a cDNA library with oligonucleotides based on the N-terminal sequence of the corresponding acetyl-CoA:DAC acetyltransferase enzyme. In this case, the identity of cefG has been proven by gene disruption experiments resulting in strains that failed to produce cephalosporin C but accumulated its precursor DAC.

18


Based on the identification of three different ATG translation-initiation codons in the cefG gene, different sizes for the open reading frame and the resulting proteins have been proposed [96, 97, 124]. This was elucidated by Velasco and co-workers [126], who synthesized all estimated proteins in E. coli and purified the native A. chrysogenum acetyl-CoA:DAC acetyltransferase using immunoaffinity chromatography. The cefG gene has a size of about 1.3 kb and contains two intronic sequences as demonstrated by sequencing of its cDNA [97, 124, 125].

5 Multiple Layers of Control The complexity of the biosynthesis of cephalosporin C and its precursors implicates different layers of regulation. In fact, evidence is available for regulatory mechanisms that act on the transcript level of biosynthesis genes and on the activity of the enzymes involved. Furthermore, cellular investigations suggest a correlation between cephalosporin C biosynthesis and mycelial morphology and differentiation. In addition, the uptake of precursors, the compartmentalization of biosynthesis and the export of cephalosporin are regulatory processes, which influence the overall production of cephalosporin C (see separate sections for details). 5.1 Transcript Level Until the early 1990s, almost all of our knowledge of the regulation of cephalosporin biosynthesis was derived from measurements of the product and intermediates. With cloning and sequencing of the cephalosporin biosynthesis genes, a new era started in as much as detailed analysis of the biosynthesis by monitoring transcript levels was facilitated. Smith et al. [119] mapped the transcription start points of the pcbC gene and analyzed the transcript level of pcbC during a seven day fermentation. The transcript level was not constitutive and found to be highest between day two and day four. Their results suggested a transcriptional regulation of the pcbC gene. This assumption was later confirmed and more refined analyses revealed several external parameters that act at the transcriptional level. Velasco et al. [27] used wild-type and two industrial strains from A. chrysogenum to establish the influence of methionine on the transcript levels of the four biosynthesis genes pcbAB, pcbC, cefEF and cefG. Previous reports indicated that methionine has a stimulatory effect on the production of cephalosporin C, and a higher enzyme activity in these cultures has been described before [127, 128]. The inhibition of de novo protein synthesis by addition of cycloheximide prevented increased antibiotic production that usually occurs in cultures supplemented with methionine only. It was concluded that methionine does not


19

simply stimulate already present enzymatic activity, but rather acts on another, earlier level. The assumed transcriptional regulation could be confirmed in the wild-type strain for the pcbAB and pcbC genes, which exhibited increased transcript levels in the presence of methionine.Velasco et al. [27] also reported that in the two industrial strains C10 and CW19, the transcript level of the cefEF gene was increased by methionine supplementation. However, for all strains the highest induction rate could be detected for the pcbAB transcript. In this context, Velasco et al. [27] detected several consensus CANNTG sequences in the intergenic region of the A. chrysogenum biosynthesis genes pcbAB and pcbC, which are recognized by members of the basic region-helixloop-helix (bHLH) protein family. Some of these known transcription factors are involved in the transcriptional control of the sulfur network in S. cerevisiae [129–132]. Thus, the authors suggested that a member of the bHLH proteins might mediate these methionine-inducing effects. Besides methionine, the influence of other factors such as carbon source and ambient pH on the transcript levels of the biosynthesis genes was investigated. Jekosch and Kück [133] were able to show that the pcbC transcript in a wildtype strain of A. chrysogenum was completely repressed in the presence of 6.3% glucose. The amount of transcript correlated well with the amount of isopenicillin N synthase in a Western blot analysis indicating the importance of transcriptional regulation in biosynthesis. However, in the semi-producer strain A3/2, the transcription of the pcbC gene was not repressed by glucose. The higher transcript levels of all biosynthesis genes in the improved strain allowed the analysis of the cefEF gene in addition to the pcbC gene. A clear reduction of the cefEF transcript and protein level in the presence of 6.3% glucose was observed in strain A3/2. A pH-dependent transcription of the pcbC gene was reported for both industrial and wild-type strains. In the wild-type A. chrysogenum strain, highest transcript levels could be detected under neutral and mild alkaline conditions at pH 7 and 8, whereas in the semi-producer A3/2, the pH optimum was at pH 6 [123]. So far, no detailed analysis has yet been published, which might illustrate transcriptional regulation with respect to all parameters, also including different nitrogen sources, known to influence cephalosporin C production [134]. The expression of the cefG gene is limiting for cephalosporin C production in all studied strains of A. chrysogenum [135]. Only a very weak transcript of about 1.4 kb, which corresponds to the cefG gene, could be detected in A. chrysogenum cells grown in a defined production medium for 48 h and 96 h [96, 97]. The fact that the acetyl-CoA:DAC acetyltransferase showed high protein levels in cultures at 72 and 96 h, which decreased dramatically thereafter, corresponds with the late conversion of DAC to cephalosporin C during the fermentation process [126, 136]. Thus, cefG seems to be expressed at a later stage of fermentation and at a lower transcriptional level than the cefEF gene suggesting a different control mechanism for these genes, although they are expressed divergently from the same promoter region [96]. Furthermore, transcriptional analysis revealed that the cefG gene appears not to be the target of

20


glucose-dependent regulation [137], and its expression is not significantly stimulated by the addition of methionine unlike that of other cephalosporin C biosynthesis genes [27]. All these examples indicate that the transcriptional regulation of the biosynthesis genes is an important aspect of the regulation of cephalosporin C production. Undoubtedly, major regulatory effects in the biosynthesis of cephalosporin C result from transcriptional changes. One striking observation is that all industrial producer strains of A. chrysogenum have increased transcript levels from all biosynthesis genes. Nevertheless, in addition to transcriptional control, other regulatory levels exist in the biosynthesis of cephalosporin C and the modification of enzyme activity will be described in the following section. 5.2 Enzyme Activity As described above, glucose has a significant impact on transcription of the pcbC and the cefEF gene. There is no negative effect of glucose on the activity of purified isopenicillin N synthase or resting cells (e.g. [138]). In contrast,ACV synthetase activity is directly inhibited by glucose in vitro [139]. Actually, the inhibitory effect on ACV synthetase, the first enzyme of the biosynthesis, results from glyceraldehyde 3-phosphate and not from glucose itself [140]. Another enzyme with relevance for cephalosporin biosynthesis that shows inhibited enzyme activity under certain circumstances is a-aminoadipate reductase encoded by the lys2 gene in A. chrysogenum. This enzyme is involved in lysine biosynthesis and competes with ACV synthetase for the precursor aaminoadipic acid. a-Aminoadipate reductase activity was quantified in the presence of 0 to 10 mmol/L lysine. A wild-type strain from A. chrysogenum showed 80% inhibition of a-aminoadipate reductase activity in extracts of 1 mmol/L lysine [9]. The inhibition of this enzyme of the lysine pathway is of relevance for cephalosporin C biosynthesis, because a high intracellular level of a-aminoadipic acid is required for efficient antibiotic production. It has been described that in vitro, the energy requirement for tripeptide formation through ACV synthetase is rather high under unfavorable conditions, which could be caused, e.g. through limited amino acid concentrations. At saturated conditions, the consumption amounts to 3 ATPs per ACV tripeptide, whereas under unfavorable conditions it can be more than 20 ATPs [141]. 5.3 Correlation Between Secondary Metabolism and Morphogenesis The biosynthesis of secondary metabolites in filamentous fungi is often associated with cell differentiation and development. In Aspergillus nidulans, there is a link between biosynthesis of secondary metabolites and asexual sporulation. In recent years, the involvement of a common G-protein-mediated growth


21

pathway has been demonstrated (reviewed in [142]). The a-subunit FadA of the trimeric G-protein binds GTP in its active form and then favors vegetative growth by inhibiting conidiogenesis. Upon inactivation of the G-protein through intrinsic GTPase activity, the inhibitory effect on signaling cascades to sporulation and toxin biosynthesis is released. Interestingly, a dominant activating fadA allele stimulates the expression of the pcbC gene and penicillin biosynthesis in A. nidulans [143]. It cannot be excluded that an FadA homologue also influences b-lactam biosynthesis in A. chrysogenum, as such a G-protein-mediated regulation of secondary metabolite production has already been described for different fungi other than A. nidulans. Examples are cyclopiazonic acid and aflatoxin biosynthesis in Aspergillus flavus, trichothecene production in Fusarium sporotrichioides and pigment synthesis in Cryphonectria parasitica. The influence of an active G-protein can be both positive and negative depending on the respective biosynthesis (overview see [142]). In A. chrysogenum, sporulation is very weak and little is known about a possible coupling to cephalosporin biosynthesis. Bartoshevich et al. [144] describe three differentiation types for A. chrysogenum and their correlation with cephalosporin production. Type 1 is the transition from the vegetative stage into a reproductive one with the formation of conidia. In this reproductive stage, cephalosporin production is lowered. It should be noted that conidia are usually not formed by high titer strains of A. chrysogenum. Type 2 is described for the late stages of development and characterized by the formation of arthrospores with thick cell walls and probably retarded metabolism. These arthrospores may be considered as simplified reproductive spores serving the survival of the organism under stress conditions and are accompanied by a lowered production of cephalosporin C. Type 3 differentiation is a multi-stage transformation of the mycelial organization into swollen fragments or yeastlike cells, which are capable of periodical polycyclic development. This alternating mycelial and yeast-like organization is most pronounced under conditions of high cephalosporin production [144]. It has been known for a long time that the phase of hyphal differentiation coincides with the maximum rate of cephalosporin synthesis and that methionine enhances fragmentation and antibiotic production [22, 145, 146]. The stimulatory effect of methionine on the transcription of biosynthesis genes was mentioned earlier, and the pleiotropic action of methionine in A. chrysogenum has been reviewed by Martín and Demain [37]. Norleucine, a non-sulfur analogue of methionine, also stimulated cephalosporin C production and mimics the methionine’s effect on mycelial morphology [147]. The capability of yeast-like cells to produce high amounts of cephalosporin C might rely on the alternative respiration pathway in A. chrysogenum. This cytochrome-independent and cyanide-insensitive respiration seems to be an obligate feature of yeast-like, but not filamentous cells and is important when cytochrome-dependent respiration cannot completely regenerate the reduced coenzyme [148]. It was also reported that the alternative respiration exhibits a

22


more than two-fold increase when A. chrysogenum was grown on sugars or soybean oil. The addition of soybean oil even doubled the specific production of cephalosporin C [149]. However, no strict correlation and interdependency exists between mycelial fragmentation and cephalosporin production rate.Allosamidin, a potent chitinases inhibitor, retarded the fragmentation of hyphae but did not affect cephalosporin C production [150]. An analysis of carbon source, growth rate and antibiotic synthesis revealed that the fragmentation has a causal relationship with growth rate. Low growth rates may weaken the hyphae and agitation could possibly cause breakages [151].

6 Transcription Factors as Activators and Repressors of Cephalosporin Biosynthesis The isolation and analysis of transcription factors from filamentous fungi began about 15 years ago. First examples came from two model organisms Aspergillus nidulans and Neurospora crassa. Transcription factors like CREA from A. nidulans, which acts as a major glucose repressor [152] and NIT2 from N. crassa that regulates structural genes of nitrogen metabolisms are involved in primary metabolism. A. nidulans is not only a model organism but also a penicillin producer and many investigations on the regulation of b-lactam biosynthesis were performed in this fungus over the last several years. Due to low titer, penicillin production in A. nidulans is not of any biotechnical interest and, therefore, other fungi such as Penicillium chrysogenum are used for penicillin production. The first transcription factor, which was isolated from this fungus is the PACC protein [153]. The pacC gene was cloned using the heterologous sequence from A. nidulans as hybridization probe. Transcription factors from A. chrysogenum have only recently been isolated and can be divided into two groups: the zinc finger proteins PACC and CRE1 already known from other fungal b-lactam producers, and the RFX transcription factor CPCR1 initially discovered in Acremonium chrysogenum. As shown in Fig. 4, a detailed sequence analysis revealed that all promoter sequences of cephalosporin biosynthesis genes contain potential DNA-binding sites for all of these transcription factors. 6.1 PACC – pH-Dependent Transcriptional Control Many filamentous fungi are capable of surviving and growing in a broad range of ambient pH, which might be as acidic as 2 or as alkaline as 10. Apart from their ability of homeostasis, they adapt the secretion of enzymes and secondary metabolites in response to the respective pH environment. Penicillins and


23

Fig. 4 Transcription factor binding sites in the divergently orientated promoter sequences from the pcbAB-pcbC, cefD1-cefD2, and cefEF-cefG cephalosporin biosynthesis genes. Bars indicate recognition sites for the transcription factor CPCR1 (grey), CRE1 (black), and PACC (white). Those binding sites, which have been shown to exist experimentally, are marked by asterisks. The CPCR1 consensus binding site is highly complex. Therefore, it is not certain that all predicted sites show in vitro the expected binding activity

cephalosporins are produced in elevated amounts under alkaline ambient pH. The intensive study of mutants with defects in pH regulation led to the isolation of the pacC gene from A. nidulans by complementation of a mutant strain [154]. PACC is a zinc finger transcription factor of the C2H2-type with three zinc fingers. The success in A. nidulans was followed by the isolation of pacC genes from P. chrysogenum, A. niger and A. chrysogenum [123, 153, 155]. The pacC gene from A. chrysogenum (pacCAc) was isolated from a lambda genomic library using the P. chrysogenum pacC gene (pacCPc) as a heterologous DNA probe encompassing the highly conserved zinc finger region. Sequencing of a 3-kb fragment allowed the identification of a DNA fragment encoding an ORF of 621 amino acids. The ORF is interrupted by three introns of which two are located in the zinc finger region. PacCAc is 20 and 57 amino acids shorter than the corresponding genes from the penicillin producers A. nidulans and P. chrysogenum, respectively. PACCAc shows approximately 35% sequence identity to other PACC proteins, which are much more alike with about 60% identity. The observed differences are consistent with the taxonomic classification of the three fungi: Acremonium belongs to the Pyrenomycetes, whereas Aspergillus and Penicillium are Plectomycetes. Southern analysis revealed that pacC is a single copy gene and is located on identical restriction fragments in wild-type and semi-industrial strains of A. chrysogenum [123].

24


PACC proteins are zinc finger transcription factors with highly conserved DNA recognition positions in the first two of the three zinc fingers [156]. They bind to a consensus binding site 5¢GCCAAG3¢ with high affinity in vitro. In A. chrysogenum the pcbAB-pcbC promoter region contains two binding sites for PACC (see Fig. 4). Both are recognized efficiently in vitro by an E. coli synthesized protein fragment of PACCAc encompassing the DNA-binding zinc finger region [123]. The promoter region between the cefEF and cefG genes, which is specific to cephalosporin C biosynthesis and A. chrysogenum also contains two PACC binding sites that are recognized in vitro (see Fig. 4). Experimental results from A. chrysogenum suggest that the PACCAc protein functions in the regulation of cephalosporin gene expression in a pH dependent manner. In wildtype strains of A. chrysogenum, the expression of cephalosporin biosynthesis genes is stronger under alkaline conditions, which probably results from an activated PACC protein [123]. The PACC protein is activated through proteolytic processing, which results in the mature, shorter polypeptide. In Aspergillus, a detailed study of the regulation of gene expression in a pH dependent manner resulted in the identification of the Pal signal transduction pathway (e.g. [157]). Each of the six pal gene products is required for the proteolytic processing of the PACC transcription factor in Aspergillus. Only the processed and shorter form of the protein can act as a transcriptional activator [158, 159]. The activated PACC protein is involved in the pH dependent regulation of many genes and probably also of the pacC gene itself. The existence of five putative PACC binding sites in the promoter of the pacC gene from A. chrysogenum suggests a strong autoregulation of the gene. A similar conclusion has been drawn for the pacC genes from P. chrysogenum and A. niger [153, 155]. As already stated, the regulation of gene expression in a pH-dependent manner is not restricted to b-lactam biosynthesis only. Many different genes and pathways in fungi and yeasts are regulated in response to ambient pH, and PACC is a key player in all of these processes. Although PACC is not a specific regulator of b-lactam antibiotics or cephalosporin biosynthesis, it has a general effect on gene expression of biosynthesis genes. 6.2 CRE1 – A Glucose Repressor Protein Another important parameter for fungal growth and the induction of secondary metabolism is the available carbon source. It has been described for all b-lactam producing fungi that growth is often promoted by glucose, but that higher concentrations of glucose have a negative effect on antibiotic production (reviewed in [160]). This negative effect is suspected to stem from transcriptional and post-transcriptional mechanisms. There are also a number of differences between the three b-lactam producers P. chrysogenum, A. nidulans and A. chrysogenum regarding the extent of glucose repression for certain biosynthesis genes and enzyme activities.


25

In A. chrysogenum, it was reported that the enzyme activity of the gene products from pcbAB, pcbC and cefEF decreased in the presence of 6.3% glucose and cephalosporin production was reduced [138, 139, 161]. Jekosch and Kück [133] showed that in the wild-type strain, both the pcbC and the cefEF gene are transcriptional repressed in the presence of glucose. In A. nidulans and Trichoderma reesei, repression of gene transcription by glucose is regulated by the carbon catabolite repressors CREA (syn. CRE1) [152, 162].A PCR-based approach using degenerative primers derived from amino acid sequences of published CRE proteins led to the amplification of a partial cre1 gene from A. chrysogenum. The gene fragment was used to screen a lambda genomic library, and a 2.9 kb fragment carrying the complete cre1 gene could be isolated.An intronless ORF of 1218 bp codes for 406 amino acids. The deduced CRE1 protein sequence showed an overall similarity of 69% to the T. reesei CRE1 and 56% to the A. nidulans CREA [137]. CRE proteins contain two C2H2-type zinc fingers and recognize a consensus binding motif 5¢-SYGGRG-3¢ in a context-dependent manner. In addition to the zinc fingers, A. chrysogenum CRE1 carries all conserved domains, which were previously described for the T. reesei CRE1 [162]. These include in particular the acidic and regulatory regions, which have been shown to be involved in the regulation of the DNA-binding ability by phosphorylation [163]. The cre1 gene is a single copy gene in wild-type and producer strains of A. chrysogenum and comparison of both strains showed no chromosomal rearrangement within the cre1 gene region [137]. The promoters of the pcbC and the cefEF gene contain several putative binding sites for CRE1 through which the transcription factor might repress these genes. This idea is supported by results obtained with A. chrysogenum transformants that contain ectopically integrated multiple copies of the cre1 gene [133]. In the wild-type strain, the pcbC and the cefEF gene are repressed by glucose. This repression does not significantly differ in transformants carrying multiple copies of the cre1 gene. However, a similar approach using the semiproducer strain A3/2 yielded in changed transcript levels. In strain A3/2, the pcbC gene is no longer subject to glucose repression. However, in transformants with multiple cre1 gene copies, transcript levels of pcbC are lower in the presence of glucose indicating a restored glucose repression mechanism due to insertion of multiple copies of the cre1 gene. The cefEF gene is glucose-repressed in strain A3/2 and in the corresponding cre1-transformants with the repression being more pronounced in the transformants described above [133]. These experiments indicate that in A. chrysogenum, the CRE1 transcription factor acts as a carbon catabolite repressor on the biosynthesis genes of cephalosporin. Like PACC, CRE1 is also regulating its own gene expression via binding sites in the promoter region of cre1. This autoregulation is of special interest in A. chrysogenum since in contrast to A. nidulans and T. reesei, transcript levels of the cre1 gene increased in the wild-type strain in the presence of glucose. In the wild-type strain ATCC14553, the cre1 transcript level was increased about sixfold after two days of cultivation in the presence of glucose. In contrast, changes in the transcript levels in the semi-producer strain A3/2 could not be observed,

26


thus indicating the absence of a glucose regulation of cre1 transcriptional expression [137]. 6.3 CPCR1 – Cephalosporin C Regulator 1 PACC and CRE1 are known to be involved in the transcriptional regulation of b-lactam biosynthesis in various filamentous fungi. These could also be isolated from and investigated in A. chrysogenum. A different approach led to the identification of the CPCR1 (Cephalosporin C Regulator 1) transcription factor in A. chrysogenum. A 24-bp sequence from the pcbAB-pcbC promoter was used in the yeast one-hybrid-system to isolate cDNAs from A. chrysogenum encoding DNA-binding proteins that interact with the promoter sequence. The sequence is located –441 to –418 relative to the translational start of the pcbC gene and contains a CCAAT-box and an imperfect palindrome. A cDNA was identified in the one-hybrid-screen that encodes a polypeptide of 788 amino acids. Analysis of the genomic DNA extended the ORF of the cpcR1 gene to 830 amino acids and revealed the position of two short introns. Southern analysis of genomic DNA from wild-type and semi-producer strains with a cpcR1 gene probe detected only a single hybridizing band of identical size. This indicates that cpcR1 is a single copy gene in A. chrysogenum [164]. CPCR1 is the first member of the RFX-family of transcription factors in filamentous fungi. RFX proteins form a subfamily of the winged-helix proteins that are characterized by a DNA-binding domain of the helix-turn-helix type [165, 166]. Another novel RFX gene was isolated from P. chrysogenum through sequence homology with cpcR1 [164]. PcRFX1 and CPCR1 share about 29% amino acid identity (Fig. 5).A data library search of completely sequenced fungal genomes showed that cpcR1 homologues are present in Neurospora crassa, Magnaporthe grisea, as well as in Aspergillus nidulans. From the amino acid sequence comparison in Fig. 5 can be concluded that all predicted polypeptides show highest homology with regard to the DNA-binding and dimerization domain. A similar degree of identity exists also to the yeast RFX proteins SAK1 from Schizosaccharomyces pombe and CRT1 from Saccharomyces cerevisiae [167, 168]. The yeast proteins are involved in DNA repair and meiotic divisions, whereas the human members of the RFX family function in a tissue and lineage specific manner. RFX5 is an interesting example as it is part of a protein complex that regulates the immune response. Therefore, mutation in RFX5 can lead to severe defects in the immune system (e.g. [169, 170]). CPCR1 is a typical member of the RFX/winged-helix family. Besides the RFXtype DNA-binding domain at amino acid positions 224 to 298, it contains a characteristic C-terminal dimerization domain. The C-terminus is necessary for homodimerization of CPCR1 (Fig. 5). Truncation of the dimerization domain results not only in the inability of CPCR1 to form homodimers but also in a loss of its DNA-binding activity. Thus, CPCR1 only binds DNA in a dimeric state [164]. Interestingly, cpcR1 homologues have been found in DNA sequence data


27

Fig. 5 Alignment of primary amino acid sequences of predicted CPCR1 polypeptides from different fungal species. Using available sequences from data libraries, cpcR1 homologous genes have been identified in five different fungi. The DNA-binding domain and the dimerization domain are underlined with black or grey bars, respectively. Identical residues in all (black) or in four (grey) sequences are shaded.Abbreviations: Ac, Acremonium chrysogenum; An, Aspergillus nidulans; Mg, Magnaporthe grisea; Nc, Neurospora crassa; Pc, Penicillium chrysogenum. For Fig. 5 see also following pages

28

Fig. 5 (continued)



29

Fig. 5 (continued)

from different fungal sources (Fig. 5) indicating that RFX proteins fulfill a different regulatory function which is not restricted to b-lactam biosynthesis. For the functional elucidation of CPCR1, several fungal transformants with varying cpcR1 gene copy numbers were generated [171]. A knockout strain showed changed pcbC transcript levels indicating a direct involvement of CPCR1 in transcriptional regulation of this cephalosporin biosynthesis gene. Reporter gene analysis of pcbC promoter derivatives with deleted CPCR1 binding sites revealed a clear dependence of reporter gene activity on functional binding sites. The deletion of two CPCR1 binding sites resulted in the total loss of reporter gene activity after cultivation of seven days. Interestingly, the levels for cephalosporin C were not significantly altered in the cpcR1 knockout strain. However, the amount of the biosynthesis intermediate penicillin N was drastically reduced. This underlines the assumption that CPCR1 is involved in the regulation of the early biosynthesis genes such as pcbC. The observed reduction of penicillin N was completely reverted in a strain with a complemented cpcR1 gene [171]. To summarize, CPCR1 is a transcription factor, which is involved in the regulation of cephalosporin C biosynthesis. The deletion of the cpcR1 gene does not prevent the production of cephalosporin C, but has significant effects on the level of antibiotic biosynthesis. CPCR1 probably binds to the promoter of the pcbC gene as a dimer with a molecular mass of nearly 200 kDa and it is most likely that one or several other regulatory proteins interact with CPCR1. Thus, the full function of CPCR1 and putative additional transcription factors or mediator proteins has first to be identified before the complete scenario of pcbC gene regulation can be pictured. So far, it has not been investigated whether PACC or CRE1 are protein interaction partners of CPCR1 under defined physiological conditions.

30


6.4 Comparison of Cephalosporin and Penicillin Biosynthesis Regulation Penicillin and cephalosporin biosynthesis share the pcbAB and the pcbC genes encoding enzymes for the first catalytic steps resulting in the formation of isopenicillin N. Although differences between the b-lactam producers exist in promoter length and structure, and in the transcriptional response to environmental parameters, it is not unlikely that a basic set of similar transcription factors is involved in the regulation of b-lactam biosynthesis in other fungi as well. Even if a transcription factor and its target gene are conserved between two fungi, it still remains to be clarified if all regulatory details are identical. One good example is the PACC transcription factor. Differences have been observed between P. chrysogenum and A. nidulans with regard to the transcriptional response of b-lactam biosynthesis genes to the ambient pH [153]. Nevertheless, in A. chrysogenum are some additional transcription factor candidates with a role of regulating cephalosporin biosynthesis genes, which have not been described so far. Both candidates are not specific for penicillin biosynthesis but function in a more general way. General transcription factors can act on a broad range of promoters of target genes that are involved in unrelated pathways. One example is the AREA family of fungal transcription factors including NIT2 and NRE that regulate nitrogen control in A. nidulans, N. crassa and P. chrysogenum. It was shown that the NRE transcription factor binds to promoter sequences from penicillin biosynthesis genes [172]. The second example of a general transcription factor is the HAP-complex from A. nidulans. This multi-protein-complex, which has been designated PENR1 [173] is involved in the regulation of many genes, but also binds to the pcbC promoter of A. nidulans. Finally, there is the question of how much could be learnt from noncephalosporin producing fungi. Recently, we have discovered sequences with similarity to b-lactam biosynthesis genes in the genomic sequence of the human pathogen Aspergillus fumigatus [174]. This finding invites to speculate whether or not cephalosporin biosynthesis genes could be residual in genomes of anamorphs or teleomorphs of A. chrysogenum.Additionally, it remains to be determined whether or not these species have retained regulatory systems for b-lactam biosynthesis genes, similar to those of A. chrysogenum.

7 Molecular Differences in Production Strains All A. chrysogenum strains that are currently used for the production of cephalosporin C have been derived from the Brotzu isolate found in 1945. Repeating cycles of mutation and selection methods resulted in strains that produce under ideal fermentation conditions more than 20 g/L [39]. Selection was aimed at achieving highest possible production rates under conditions suitable


31

for fermentation in a cost efficient way. Bearing this in mind, most changes inevitably occurred at the molecular level that have accumulated in these strains. In general, transcript levels of biosynthesis genes are significantly higher in producer strains, an observation, which has also been made for optimized P. chrysogenum strains. However, there are striking differences between penicillin and cephalosporin producer strains. For A. chrysogenum producer strains, no report is available stating that the gene copy number of biosynthesis genes has increased as it is the case for the penicillin biosynthesis cluster. This has been amplified by a factor of between 5 to greater than 10 in most production strains of P. chrysogenum (e.g. [175]). One explanation for this finding is that cephalosporin biosynthesis genes are not located in a single cluster, but rather are distributed over two different chromosomes. However, several investigations have shown that the transcript level of biosynthesis genes is increased and, thus, changes must have occurred at the level of transcriptional regulation of structural genes. There are two main explanations for these findings: mutations in the promoter region may be responsible for example for a different recruitment of transcriptional activator proteins and/or molecular changes occurred in the regulatory genes and, consequently, in the corresponding proteins. To date, no changes in the copy number of the identified regulatory genes of cephalosporin C biosynthesis have been observed. All transcription factor genes of A. chrysogenum described so far, namely cpcR1, cre1 and pacC, seem to be single copy genes in wild-type and producer strains. From hybridization analysis with restricted genomic DNA can be concluded that no major intrachromosomal DNA rearrangements have taken place when producer strains were generated from wild-type strains [123, 137, 164]. However, it is not known whether mutations occurred in the transcription factors, which might for example increase the transcriptional activation capacity or their interaction with other regulatory factors. Other relatively constant parameters are promoter sequences. It was described that no significant changes could be detected when the pcbC promoter DNA sequences of A. chrysogenum strains with different antibiotic production rates were compared [176]. A similar result was obtained for the promoter regions of penicillin biosynthesis genes from wild-type and producer strains [175]. Nevertheless, a number of changes have already been described. Relatively soon after the discovery of the biosynthesis genes, the chromosomal localization of the gene clusters was investigated using pulsed-field gel electrophoresis. This technique allows the separation of intact chromosomes with a size of up to 10 Mb. Walz and Kück [177] have reported that different A. chrysogenum strains were indistinguishable with respect to restriction fragment patterns, but six out of eight chromosomes differed in size. In addition, the rDNA gene cluster seemed to be relocated from its original site at chromosome II in wild-type strains to chromosome VII in an improved production strain. Another investigation revealed chromosome changes only in the minority of strains from a lineage with improved cephalosporin C production. In one strain, the size of the chromosome was altered on which the pcbC gene is

32


located [101]. Until now, it could not be clarified whether these chromosomal rearrangements have occurred accidentally during industrial strain improvement programs and whether they are at least partially responsible for titer improvement through increased gene expression rates of the translocated biosynthesis genes. The increasing knowledge of transcription of biosynthesis genes and the regulatory proteins involved in this process has led to the discovery of some interesting differences in production strains.As already mentioned, the transcript levels in production strains are generally increased. Besides different transcript levels, improved strains often show an altered regulation in response to certain parameters. One example is provided by Velasco et al. [27]. They report that in production strain C10, the supplementation of methionine correlates with an increased transcription of the pcbAB, pcbC and cefEF genes. This is in contrast to the Brotzu strain, where only pcbAB and pcbC transcript levels are higher in the presence of methionine. This suggests that changes might have occurred in the regulation of the cefEF gene during strain improvement. Another example is related to the transcription factor PACC and the pH-dependent transcriptional control. In the semi-producer strain A3/2 of A. chrysogenum, transcript analysis revealed a pH optimum for pcbC and cefEF transcript levels at pH 6 [123]. This is in contrast to the general observation that higher yields of b-lactam antibiotics are obtained at pH 7 to 8. Indeed, in a nonoptimized strain, the highest level of the pcbC transcript could be detected at pH 7 to 8. The change of the optimum pH to pH 6 is justified for industrial fermentation of A. chrysogenum, as this process is usually run at a slightly acid pH to increase the stability of the product in the fermentation broth. Thus, classical strain improvement resulted in an adaptation of the transcriptional regulation towards an optimized antibiotic production under the employed fermentation conditions. Another interesting observation was that not only the transcript level was increased at pH 6, but that the transcription rate showed a real optimum curve and was lower at an alkaline pH. If this altered pH-dependent transcription still depends on PACC and the related PAL signal cascade, the pH sensing for the induction of the cascade must have changed in the production strain. The last example is derived from investigations of carbon source regulation and the transcription factor CRE1, a glucose repressor protein of the zinc finger type. One difference between the wild-type strain and strains with enhanced antibiotic production is the regulation of the cre1 gene itself. The transcript level of the gene encoding the repressor protein CRE1 is increased sixfold in the presence of glucose in a wild-type strain [137]. Interestingly, this glucosedependent transcriptional upregulation of cre1 does not take place in strain A3/2 with improved production of cephalosporin C. This deregulation of the glucose repressor gene might be related to the increased antibiotic production rate. cre1 promoter sequences of the two strains were determined and found to be completely identical, indicating that the deregulation of the cre1 gene does not result from mutations in its own promoter region. The idea of an altered


33

glucose regulation in improved strains is supported by the finding that in the production strain, the pcbC gene is no longer repressed by glucose as it is the case in the wild-type strain [133]. Northern analysis revealed a complete reduction of the pcbC transcript level in the wild-type strain in the presence of 6.3% glucose and in the producer strain a reduction to less than 50% for the cefEF transcript.As already mentioned, the level of the pcbC transcript was not reduced in the producer strain. The involvement of CRE1 in the glucose effect was shown by the transfer of several copies of the repressor gene cre1 into the producer strain. The resulting transformants have a glucose-dependent regulation of the pcbC and the cefEF gene. This suggests that transcription factors are important targets, which have been directly or indirectly subjected to alterations during strain improvement.

8 Examples of Molecular Engineering of A. chrysogenum The directed manipulation of genetic material of an organism with the aim to change its biosynthesis capabilities can be regarded as an alternative to classical strain improvement to complement current strain breeding strategies. Compared to classic approaches of titer improvement, molecular engineering of biosynthesis genes requires much more knowledge of the relevant molecular details. 8.1 Genetic Tools for Molecular Engineering One prerequisite for molecular engineering is an established set of genetic tools. For A. chrysogenum, the available tools are still very limited. There are few examples for strong or inducible promoters for the expression of homologous and heterologous genes, selection markers for transformation, plasmids and methods for efficient homologous integration of genes and gene disruptions. When the codon usage of A. chrysogenum was investigated in 1999, the DNA sequences of only 19 nuclear genes were accessible in public databases [178]. Today, the number of known genes from A. chrysogenum is approximately 30 of which 6 are cephalosporin C biosynthesis genes and 3 encode transcription factors that are involved in the regulation of the biosynthesis genes. In the 1980s, an efficient integrative transformation system was described for A. chrysogenum [179]. Transformation usually results in the ectopic integration of plasmid DNA at one or several genomic loci due to illegitimate recombination. So far, only three different selection markers have been used repeatedly: the bacterial genes for hygromycin B and phleomycin resistance (e.g. [171, 180]), and a mutated version of the b-tubulin gene from A. chrysogenum providing a homologous transformation system [181]. The substitution of phenylalanine by tyrosine at codon 167 results in a mutated b-tubulin gene, which conveys benomyl resistance as a dominant selection marker. The mu-

34


tated b-tubulin gene can be expressed using its own promoter and flanking regions. Transformation of a gel-purified DNA fragment encompassing the btubulin selection marker, but no bacterial DNA sequences from cloning vectors, is feasible and results in A. chrysogenum transformants without integrated heterologous DNA [182]. This aspect of the homologous transformation system is important with respect to governmental restrictions concerning the use of recombinant strains in biotechnical production processes. It is also possible to integrate a DNA fragment or a plasmid without a suitable selection marker in the genomic DNA of A. chrysogenum when a co-transformation experiment is conducted together with a vector harboring a dominant selection marker [107, 182]. Further molecular tools that are important for studying the regulation of cephalosporin C biosynthesis genes include reporter genes. Menne et al. [107] fused the intergenic region between the pcbAB and the pcbC gene with the two reporter genes lacZ and gusA and compared the specific enzyme activity of A. chrysogenum transformants harboring the four different gene fusions. They could show that the specific activity of the b-galactosidase encoded by the lacZ gene is higher than the enzyme activity obtained with the gusA gene, making the lacZ gene more suitable for the analysis of weak promoters in A. chrysogenum. The ability to disrupt a gene is of high importance for molecular engineering. The disrupted gene can be for example a structural gene of the biosynthesis resulting in truncated biosynthesis, or a regulatory gene whose product is involved in the regulation of cephalosporin C biosynthesis. In the latter case, gene disruption can alter the transcription of one or several biosynthesis genes. So far, only a few examples of gene disruption in A. chrysogenum are available, because the required homologous recombination is a rare event in this filamentous fungus. The first investigation used the disruption of the pcbC gene to determine that 3 kb are the required length of homologous DNA sequences at both sides of the resistance cassette to yield knock-out transformants [183]. Velasco et al. [184] also used several kb of homologous DNA flanking the resistance cassette for their disruption of the cefEF gene in A. chrysogenum. The disruption of a gene encoding a biosynthesis enzyme results in the accumulation of pathway intermediates, which can often be detected using bioassays, allowing a fast identification of the desired gene disruption transformant [183, 184]. The disruption of a regulatory gene often lacks a phenotype, which in principle could be used to identify the desired knockout transformants. Therefore, a PCR strategy is applied in order to detect the knockout strain. This approach was followed for the transcription factor gene cpcR1 [171]. For the disruption of the mecB gene encoding cystathione-g-lyase, which is involved in cysteine synthesis, a doublemarker technique was used. Transformants with a correct double-crossover were hygromycin-resistant and phleomycin-sensitive, whereas ectopic integration led to transformants with resistance against both antibiotics [33]. The tools available for the investigation and alteration of cephalosporin C biosynthesis can also be utilized to establish the synthesis of other products in


35

A. chrysogenum. So far, a few heterologous proteins were synthesized in different strains of A. chrysogenum, but no secondary metabolites besides cephalosporin derivatives. Examples are alkaline proteases from Fusarium sp., human lysozyme and recombinant hirudin [185–187]. 8.2 Optimization of Cephalosporin C Biosynthesis The determination of rate-limiting steps in the biosynthesis is the first goal of a rational approach to improve cephalosporin titer. This can be performed by looking for intermediates that accumulate in fermentation broth, measurements of specific enzyme activity and the comparison of wild-type and hightiter production strains [188]. Several biosynthesis genes have been transformed into A. chrysogenum to yield strains with a higher copy number of these genes. The amplification of the pcbC gene did not result in significantly increased cephalosporin C production indicating that the cyclase activity is not rate-limiting in b-lactam biosynthesis [100]. In contrast to the results with the pcbC gene, an increase in the copy number of the cefG gene had a positive effect on cephalosporin C titer. Mathison and co-workers [97] cloned the cefG gene encoding the acetyl transferase, which catalyzes the last step in the biosynthesis. By transforming the cefG gene in the A. chrysogenum cefG mutant M40, they restored the synthesis of cephalosporin C and observed a correlation between cefG copy number and cephalosporin C titer. The transformation of a wild-type strain with up to five additional copies of the cefG gene increased the cephalosporin C titer from 0.625 mg/mL to 1.9 mg/mL. Mathison et al. [97] concluded that at least in the wild-type strain, the acetyl transferase activity is a rate-limiting step. In another investigation, the cefG gene was expressed from the homologous promoter and from four heterologous promoters. This study used for example promoters from the gpd gene of A. nidulans as constitutive promoters and the pcbC promoter from P. chrysogenum as the non-constitutive one [135]. In general, a higher steady-state transcript level of the cefG gene was observed in all transformants. Transformants of the producer strain C10 showed a doubled acetyl transferase activity when the cefG gene was fused to the pcbC promoter of P. chrysogenum, and a better conversion of deacetylcephalosporin C to cephalosporin C.Again, it was concluded that the expression of the cefG gene is limiting for cephalosporin C biosynthesis. A similar amplification of gene copy number with the goal to increase the corresponding enzymatic activity was tried for a gene, which is involved in precursor synthesis. L-cysteine is a precursor of the ACV tripeptide in cephalosporin C biosynthesis and can be supplied through an autotrophic pathway and through the reverse transsulfuration pathway. In the latter case, it is produced from methionine via cystathionine, which is split into cysteine and a-ketobutyrate enzymatically by cystathionine-g-lyase. This enzyme is en-

36


coded by the mecB gene and is required for high-level cephalosporin production [33]. Some transformants with multiple copies of the mecB gene showed higher cystathionine-g-lyase activity and one transformant produced higher amounts of cephalosporins [35]. It was concluded that moderately increased levels of cystathionine-g-lyase stimulate cephalosporin production, but very high levels are deleterious for growth and production. Titer improvement was also reached in transformants of A. chrysogenum with a bacterial hemoglobin gene. The oxygen-binding heme protein from the bacterium Vitreoscilla has been synthesized in fungal transformants to improve the oxygen supply during fermentation [189]. It is not known, whether oxygen supply directly affects the three oxidation reactions in cephalosporin biosynthesis or indirectly benefits the production by a more efficient overall metabolism. Several transformants expressed the heme gene under the control of the strong constitutive TR1 promoter from Trichoderma reesei and produced a higher cephalosporin titer than control strains in batch culture experiments. Ten out of 17 transformants with the Vitreoscilla heme gene produced 7–64% higher cephalosporin C levels than the non-transformed control strain C10 [189]. In principle, transcription factors are promising candidates for use in molecular engineering. Overexpression of a transcriptional activator gene or the disruption of a repressor gene are relatively simple scenarios. More sophisticated approaches could be in vitro optimized transcription factors, e.g. with stronger transactivation capacities or different DNA-binding specificities. Due to the limited knowledge of transcription factors from A. chrysogenum, the future will show whether this kind of experimentation can considerably contribute to cephalosporin titer improvement. First results obtained from A. chrysogenum transformants with altered transcription factor gene copies are listed above. Besides titer improvement, an objective of gene amplification in genetically engineered strains can be the reduction of an intermediate, which may be an undesirable by-product. The cefEF gene encodes a bifunctional expandase/hydroxylase that converts penicillin N to deacetoxycephalosporin C and then to deacetylcephalosporin. Deacetoxycephalosporin C accumulates in the fermentation broth to a concentration of 1–2% of the final cephalosporin C yield and is an undesired contaminant in the extraction process [190]. Genetically engineered strains with increased copy number of the cefEF gene resulted in the reduction of desacetoxycephalosporin C to 50% or less of the control. The cefEF gene was expressed from its own promoter and Southern analysis indicated the integration of only a single additional gene copy. The reduction of desacetoxycephalosporin content relative to the cephalosporin C production was verified in upscale fermentation of 30,000 L, but the total production of cephalosporin C was not increased significantly [190]. Another reason for the construction of genetically engineered strains is the attempt to produce cephalosporin derivatives that are more suitable for chemical modifications than cephalosporin C itself. Semisynthetic cephalosporins are made from 7-aminodeacetoxycephalosporanic acid (7-ADCA) or 7-


37

aminocephalosporanic acid (7-ACA), which can be derived enzymatically or chemically from cephalosporin C or penicillin G. The direct conversion of cephalosporin C into 7-ACA in A. chrysogenum was performed by transforming two heterologous genes into one recipient strain. The genes encoding Damino acid oxidase from the fungus Fusarium solani and glutaryl acylase from the bacterium Pseudomonas diminuta expanded the biosynthesis potential in the engineered strain. However, the amounts of 7-ACA were detectable but not commercially significant [191]. Recently, a different approach was followed where desacetoxycephalosporin (DAOC) from the fermentation broth of A. chrysogenum was used as starting material for the production of 7-ADCA. In order to accumulate DAOC in A. chrysogenum, a two step approach was used. In the first step, the cefEF gene encoding the bifunctional expandase/hydroxylase was disrupted and subsequent transformants of A. chrysogenum accumulated penicillin N. In the following step, a gene fusion was integrated into the DcefEF strain, which consisted of the pcbC promoter from Penicillium chrysogenum and the cefE gene from S. clavuligerus [184]. The cefE gene from the bacterial cephem-producer encodes the expandase enzyme that catalyzes the ring-expansion step in cephalosporin biosynthesis. Resulting recombinant strains were tested for production of DAOC in bioassays analyzing penicillinase-resistant inhibition of E. coli growth. HPLC analysis of the most promising transformant revealed a DAOC production of 75–80% of the total b-lactams produced by the parental production strain that was used for genetic engineering. It is worth mentioning that about 20% of the b-lactams produced by engineered strains accumulate in the fermentation broth as penicillin N indicating that the heterologous expandase activity is a rate limiting step. The accumulation of penicillin N might be reduced by a higher expression rate of the cefE gene or a classical strain improvement program starting with the engineered strain [184]. The purified DAOC from the fermentation broth of A. chrysogenum was bioconverted into 7-ADCA in two enzymatic steps using D-amino acid oxidase from the basidiomyceteous fungus Rhodotorula gracilis and the bacterial glutaryl acylase originating from Acinetobacter spec.

9 Outlook After the development of different molecular tools for Acremonium chrysogenum, one of the major interests in this field was to elucidate molecular changes that have occurred in strains during production improvement programs. The simple idea in the beginning that gene copy number, DNA rearrangements or promoter sequence mutations are mainly responsible for different gene expression levels has not yet been confirmed. Instead, one of the more interesting lessons learned from extensive molecular investigations in recent years has been the realization that multiple layers of control exist in

38


cephalosporin biosynthesis. Therefore, future efforts will focus on deciphering regulatory networks that control cephalosporin biosynthesis. Thus, the development of genomic tools including genome-wide location and expression analysis can be foreseen to allow the simultaneous interrogation of the expression of thousands of genes in a high-throughput fashion. Microarray analysis responds to physiological or genetic changes and will provide indispensable information that ultimately may lead to improved strain development programs. Knowledge of this kind seems to be the necessary prerequisite together with conventional procedures for the efficient generation of novel strains, which are constantly required in competitive production processes. Acknowledgements We thank E. Jung for the preparation of the manuscript, E. Szczypka for the artwork and D. Janus and J. Dreyer for help in preparing some of the figures. The authors’ work is supported by Sandoz GmbH, Kundl, Austria.

References 1. Newton GGF, Abraham EP (1955) Nature 175:548 2. Brakhage AA (1998) Microbiol Mol Biol Rev 62:547 3. Brakhage AA, Caruso ML (2004) In: Kück U (ed) The Mycota II, 2nd edn. Springer, Berlin Heidelberg New York, pp 317–353 4. Bhattacharjee JK (1985) Crit Rev Microbiol 12:131 5. Demain AL (1983) In: Solomon NA (ed) Antibiotics containing the β-lactam structure. Springer, Berlin Heidelberg New York, p 189 6. Nishida H, Nishiyama M (2000) J Mol Evol 51:299 7. Esmahan C,Alvarez E, Montenegro E, Martín JF (1994) Appl Environ Microbiol 60:1705 8. Mehta RJ, Speth JL, Nash CH (1979) Eur J Appl Microbiol Biotechnol 8:177 9. Hijarrubia MJ, Aparicio JF, Martín JF (2002) Appl Microbiol Biotechnol 59:270 10. Luengo JM, Revilla G, López-Nieto MJ, Villanueva JR, Martín JF (1980) J Bacteriol 144:869 11. Hijarrubia MJ, Aparicio JF, Casqueiro J, Martín JF (2001) Mol Gen Genet 264:755 12. Sinha AK, Bhattacharjee JK (1971) Biochem J 125:743 13. Elander RP (1983) In: Demain AL, Solomon NA (eds) Antibiotics containing the b-lactam structure. Springer, Berlin Heidelberg New York, p 97 14. Haas H, Marx F, Graessle S, Stoffler G (1996) Biochim Biophys Acta 1309:81 15. Crawford NM, Arst HN (1993) Annu Rev Genet 27:115 16. Marzluf GA (1997) Microbiol Mol Biol Rev 61:17 17. Matsumura S, Suzuki M (1986) Agric Biol Chem 50:505 18. Martín JF, Aharonowitz Y (1983) In: Demain AL, Solomon NA (eds) Antibiotics containing the β-lactam structure, vol 1. Springer, Berlin Heidelberg New York, p 229 19. Treichler HJ, Liersch M, Nüesch J, Döbeli H (1979) In: Sebek OK, Laskin AI (eds) Genetics of industrial microorganisms. American Society for Microbiology, Washington DC, p 97 20. Umbarger HE (1978) Annu Rev Biochem 47:532 21. Cherest H, Surdin-Kerjan Y (1992) Genetics 130:51 22. Caltrider NP, Niss HF (1966) Appl Microbiol 14:746 23. Döbeli H, Nüesch J (1980) Antimicrob Agents Chemother 18:111 24. Paszewski A, Brzywczy J, Natorff R (1994) Prog Ind Microbiol 29:299


39

25. Drew SW, Demain AL (1975) Antimicrob Agents Chemother 8:5 26. Suzuki M, Fujisawa Y, Uchida M (1980) Agric Biol Chem 44:19957 27. Velasco J, Gutiérrez S, Fernandez FJ, Marcos AT, Arenos C, Martín JF (1994) J Bacteriol 176:985 28. Matsumura M, Imanaka T, Yoshida T, Taguchi H (1978) J Ferment Technol 56:345 29. Benz F, Liersch M, Nuesch J, Treichler HJ (1971) Eur J Biochem 20:81 30. Nüesch J, Treichler HJ, Liersch M (1973) In: Vanek ZHZ, Cudlin J (eds) Genetics of industrial microoorganisms, vol 2. Academia, Prague, p 309 31. Gygax D, Döbeli N, Nüesch J (1980) Experientia 36:487 32. Marcos AT, Kosalková K, Cardoza RE, Fierro F, Gutiérrez S, Martín JF (2001) Mol Gen Genet 264:746 33. Liu G, Casqueiro J, Bañuelos O, Cardoza RE, Gutiérrez S, Martín JF (2001) J Bacteriol 183:1765 34. Lewandowska M, Paszewski A (1988) Acta Microbiol Pol 37:17 35. Kosalková K, Marcos AT, Martín JF (2001) J Ind Microbiol Biotechnol 27:252 36. Demain AL, Zhang J (1998) Crit Rev Biotechnol 18:283 37. Martín JF, Demain AL (2002) Trends Biotech 20:502 38. Aharonowitz Y, Cohen G (1992) Annu Rev Microbiol 46:461 39. Elander RP (2003) Appl Microbiol Biotechnol 61:385 40. Aharonowitz Y, Bergmeyer J, Cantoral JM, Cohen G, Demain A, Fink U, Kinghorn J, Kleinkauf H, MacCabe A, Palissa H, Pfeifer E, Schwecke T, van Liempt H, van Döhren H, Wolfe S, Zhang J (1993) Biotechnology (NY) 11:807 41. von Döhren H, Keller U, Vater J, Zocher R (1997) Chem Rev 97:2675 42. Stein T, Vater J, Kruft V, Otto A, Wittmann-Liebold B, Franke P, Panico M, McDowell R, Morris HR (1996) J Biol Chem 271:15428 43. Gutiérrez S, Diéz B, Montenegro E, Martín JF (1991) J Bacteriol 173:2354 44. Martín JF (2000) J Antibiot 53:1008 45. van Liempt H, von Döhren H, Kleinkauf H (1989) J Biol Chem 264:3680 46. Schwecke T,Aharonowitz Y, Palissa H, von Döhren H, Kleinkauf H, van Liempt H (1992) Eur J Biochem 205:687 47. Kallow W, Neuhof T, Arezi B, Jungblut P, von Döhren H (1997) FEBS Lett 414:74 48. Nüesch J, Heim J, Treichler HJ (1987) Ann Rev Microbiol 41:51 49. Roach PL, Clifton IJ, Fulop V, Harlos K, Barton GJ, Hajdu J, Andersson I, Schofield CJ, Baldwin JE (1995) Nature 375:700 50. Brakhage AA, Turner G (1995) In: Kück U (ed) The Mycota II. Genetics and biotechnology. Springer, Berlin Heidelberg New York, p 263 51. Baldwin JE,Adlington RM, Morony SE, Field LD, Ting HH (1984) Chem Soc Chem Commun 1984:984 52. Burzlaff NI, Rutledge PJ, Clifton IJ, Hensgens CM, Pickford M,Adlington RM, Roach PL, Baldwin JE (1999) Nature 401:721 53. Müller WH, van der Krift TP, Krouwer AJJ,Wösten HAB, van der Voort LHM, Smaal EB, Verkleij AJ (1991) EMBO J 10:489 54. Baldwin JE, Gagnon J, Ting HH (1985) FEBS Lett 188:253 55. Roach PL, Clifton IJ, Hensgens CM, Shibata N, Schofield CJ, Hajdu J, Baldwin JE (1997) Nature 387:827 56. Borovok I, Landmann O, Kreisberg-Zakarin R, Aharonowitz Y, Cohen G (1996) Biochemistry 35:1981 57. Sim TS, Loke P (2000) Appl Microbiol Biotechnol 54:1 58. Loke P, Sim TS (2002) J Biochem 127:585 59. Orville AM, Chen VJ, Kriauciunas A, Harpel MR, Fox BG, Munck E, Lipscomb JD (1992) Biochemistry 31:4602

40


60. Randall CR, Zang Y, True AE, Que L, Charnok JM, Garner CD, Fujishima Y, Schofield CJ, Baldwin JE (1993) Biochemistry 32:6664 61. Chen VJ, Orville AM, Harpel MR, Frolik CA, Surerus KK, Münck E, Lipscomb JD (1989) J Biol Chem 264:21677 62. Kreisberg-Zakarin R, Borovok I,Yanko M,Aharonowitz Y, Cohen G (1999) Antonie van Leeuwenhoek 75:33 63. Perry D, Abraham EP, Baldwin JE (1988) Biochem J 255:345 64. Baldwin JE, Abraham EP (1988) Nat Prod Rep 5:129 65. Konomi T, Herchen S, Baldwin JE, Yoshida M, Hunt NA, Demain AL (1979) Biochem J 184:427 66. Sawada Y, Baldwin JE, Singh PD, Solomon NA, Demain AL (1980) Antimicrob Agents Chemother 18:465 67. Baldwin JE, Keeping JW, Singh PD, Vallejo CA (1981) Biochem J 194:649 68. Jayatilake GS, Huddleston JA, Abraham EP (1981) Biochem J 194:645 69. Ullán RV, Casqueiro J, Bañuelos O, Fernández FJ, Gutiérrez S, Martín JF (2002) J Biol Chem 277:46216 70. Turgay K, Krause M, Marahiel MA (1992) Mol Microbiol 6:529 71. Reichel C, Brugger R, Bang H, Geisslinger G, Brune K (1997) Mol Pharmacol 51:576 72. Schmitz W, Albers C, Fingerhut R, Conzelmann E (1995) Eur J Biochem 231:815 73. Schmitz W, Helander HM, Hiltunen JK, Conzelmann E (1997) Biochem J 326:883 74. Caldwell J, Hutt AJ, Fournel-Gigleux S (1988) Biochem Pharmacol 37:105 75. Knihinicki RD, Day RO, Williams KM (1991) Biochem Pharmacol 42:1905 76. Shieh WR, Chen CS (1993) J Biol Chem 268:3487 77. Kohsaka M, Demain AL (1976) Biochem Biophys Res Commun 70:465 78. Schofield CJ, Baldwin JE, Byford MF, Clifton I, Hajdu J, Hensgens C, Roach PL (1997) Curr Opin Struct Biol 7:857 79. Baldwin JE, Adlington RM, Coates JB, Crabbe MJ, Crouch NP, Keeping JW, Knight GC, Schofield CJ, Ting HH,Vallejo CA, Thorniley M,Abraham EP (1987) Biochem J 245:831 80. Dotzlaf JE, Yeh WK (1987) J Bacteriol 169:1611 81. Scheidegger A, Kuenzi MT, Nüesch J (1984) J Antibiot (Tokyo) 37:522 82. Yoshida M, Konomi T, Kohsaka M, Baldwin JE, Herchen S, Singh P, Hunt NA, Demain AL (1978) Proc Natl Acad Sci USA 75:6253 83. Rollins MJ, Westlake DW, Wolfe S, Jensen SE (1988) Can J Microbiol 34:1196 84. Cortes J, Martín JF, Castro JM, Laiz L, Liras P (1987) J Gen Microbiol 133:3165 85. Jensen SE, Westlake DW, Wolfe S (1985) J Antibiot (Tokyo) 38:263 86. Samson SM, Chapman JL, Belagaje R, Queener SW, Ingolia TD (1987) Proc Natl Acad Sci USA 84:5705 87. Samson SM, Dotzlaf JE, Slisz ML, Becker GW, van Frank RM, Veal LE, Yeh WK, Miller JR, Queener SW, Ingolia TD (1987) Biotechnol 5:1207 88. Chin HS, Sim J, Seah KI, Sim TS (2003) FEMS Microbiol Lett 218:251 89. Felix H, Nüesch J, Wehrli W (1980) Anal Biochem 103:81 90. Fujisawa Y, Shirafuji H, Kida M, Nara K, Yoneda M, Kanzaki T (1973) Nature New Biol 246:154 91. Fujisawa Y, Kanzaki T (1975) Agric Biol Chem 39:2043 92. Liersch M, Nüesch J, Treichler J (1976) In: MacDonald KD (ed) Second International Symposium on the Genetics of Industrial Microorganisms. Academic Press, London, p 179 93. van de Kamp M, Driessen AJ, Konings WN (1999) Antonie Van Leeuwenhoek 75:41 94. Goyon C, Faugeron G, Rossignol JL (1988) Genetics 63:297 95. Langin T, Faugeron G, Goyon C, Nicolas A, Rossignol JL (1986) Genetics 49:283 96. Gutiérrez S, Velasco J, Fernández FJ, Martín JF (1992) J Bacteriol 174:3056


41

97. Mathison L, Soliday C, Stepan T, Aldrich T, Rambosek J (1993) Curr Genet 23:33 98. Gutiérrez S, Fierro F, Casqueiro J, Martín JF (1999) Antonie van Leeuwenhoek 75:81 99. Skatrud PL (1991) In: Bennett JW, Lasure LL (eds) More gene manipulation in fungi. Academic Press, New York, NY, p 363 100. Skatrud PL, Queener SW (1989) Gene 79:331 101. Smith AW, Collins K, Ramsden M, Fox HM, Peberdy JF (1991) Curr Genet 19:235 102. Martín JF, Liras P (1989) Annu Rev Microbiol 43:173 103. Seno ET, Baltz RH (1989) In: Shapiro S (ed) Regulation of secondary metabolism in Actinomycetes. CRC Press, Boca Raton, Fla, p 1 104. Diéz B, Gutiérrez S, Barredo JL, van Solingen P, van der Voort LH, Martín JF (1990) J Biol Chem 265:16358 105. Baldwin JE, Bird JW, Field RA, O’Callaghan NM, Schofield CJ,Willis AC (1991) J Antibiot 44:241 106. MacCabe AP, van Liempt H, Palissa H, Unkles SE, Riach MB, Pfeifer E, von Döhren H, Kinghorn JR (1991) J Biol Chem 266:12646 107. Menne S, Walz M, Kück U (1994) Appl Microbiol Biotechnol 42:57 108. Ramos FR, López-Nieto MJ, Martín JF (1986) FEMS Microbiol Lett 35:123 109. Ramsden M, McQuade BA, Saunders K, Turner MK, Harford S (1989) Gene 85:267 110. Hoskins JA, O’Callaghan N, Queener SW, Cantwell CA, Wood JS, Chen VJ, Skatrud PL (1990) Curr Genet 18:523 111. Jensen SE, Demain AL (1995) In: Vining LC, Stuttard C (eds) Genetics and biochemistry of antibiotic production. Butterworth-Heinemann, Newton, Mass, p 239 112. Kimura H, Miyashita H, Sumino Y (1996) Appl Microbiol Biotechnol 45:490 113. Martín JF, Gutiérrez S, Demain AL (1997) In: Anke T (ed) Fungal biotechnology. Antibiotics. Chapman and Hall, Weinheim, p 91 114. Smith DJ, Earl AJ, Turner G (1990) EMBO J 9:2743 115. Kleinkauf H, von Döhren H (1996) Eur J Biochem 236:335 116. Martín JF (1991) Ann N Y Acad Sci 646:193 117. Martín JF, Gutiérrez S, Montenegro E, Coque JJR, Fernández FJ, Velasco J, Gil S, Fierro F, Calzada JG, Cardoza RE, Liras P (1992) In: Ladish MR, Bose A (eds) Harnessing biotechnology for the 21st century. Proceedings of the ninth International Biotechnology Symposium. American Chemical Society, Washington, DC, p 131 118. Samson SM, Blankenship DT, Chapman JL, Perry D, Skatrud PL, van Frank RM, Abraham EP, Baldwin JE, Queener SW, Ingolia TD (1985) Nature 318:191 119. Smith AW, Ramsden M, Dobson MJ, Harford S, Peberdy JF (1990) BioTechnol 8:237 120. Martín JF, Gutiérrez S (1995) Antonie van Leeuwenhoek 67:181 121. Peñalva MA, Moya A, Dopazo J, Ramón D (1990) Proc R Soc Lond B Biol Sci 241:164 122. Velasco J, Gutiérrez S, Casqueiro J, Fierro F, Campoy S, Martín JF (2001) Appl Microbiol Biotechnol 57:300 123. Schmitt EK, Kempken R, Kück U (2001) Mol Genet Genomics 265:508 124. Matsuda A, Sugiura H, Matsuyama K, Matsumoto H, Ichikawa S, Komatsu K (1992) Biochem Biophys Res Commun 186:40 125. Matsuda A, Sugiura H, Matsuyama K, Matsumoto H, Ichikawa S, Komatsu K (1992) Biochem Biophys Res Commun 182:995 126. Velasco J, Gutiérrez S, Campoy S, Martín JF (1999) Biochem J 337:379 127. Sawada Y, Konomi T, Solomon NA, Demain AL (1980) FEMS Microbiol Lett 9:281 128. Zhang JY, Banko G, Wolfe S, Demain AL (1987) J Ind Microbiol Biotechnol 2:251 129. Baker RE, Masison DC (1990) Mol Cell Biol 10:2458 130. Cai M, Davis RW (1990) Cell 61:437 131. Mellor J, Jiang W, Funk M, Rathjen J, Barnes CA, Hinz T, Hegemann JH, Philippsen P (1990) EMBO J 9:4017

42 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169.

E. K. Schmitt et al. Thomas D, Jacquemin I, Surdin-Kerjan Y (1992) Mol Cell Biol 12:1719 Jekosch K, Kück U (2000) Appl Microbiol Biotechnol 54:556 Shen YQ, Heim J, Solomon NA, Wolfe S, Demain AL (1984) J Antibiot (Tokyo) 37:503 Gutiérrez S, Velasco J, Marcos AT, Fernández FJ, Fierro F, Barredo JL, Diéz B, Martín JF (1997) Appl Microbiol Biotechnol 48:606 Zanca DM, Martín JF (1983) J Antibiot 36:700 Jekosch K, Kück U (2000) Curr Genet 37:388 Heim J, Shen YQ, Wolfe S, Demain AL (1984) Appl Microbiol Biotechnol 19:232 Zhang J, Wolfe S, Demain AL (1989) Curr Microbiol 18:361 Zhang J, Demain AL (1992) Arch Microbiol 158:364 Kallow W, von Döhren H, Kleinkauf H (1998) Biochemistry 37:5947 Calvo AM, Wilson RA, Bok JW, Keller NP (2002) Microbiol Mol Biol Rev 66:447 Tag A, Hicks J, Garifullina G, Ake C, Phillips TD, Beremand M, Keller N (2000) Mol Microbiol 38:658 Bartoshevich YE, Zaslavskaya PL, Novak MJ,Yudina OD (1990) J Basic Microbiol 30:313 Nash CH, Huber FM (1971) Appl Microbiol 22:6 Karaffa L, Sándor E, Kozma J, Szentirmai A (1997) Process Biochem 32:495 Drew SW, Winstanley DJ, Demain AL (1976) Appl Environ Microbiol 31:143 Karaffa L, Sándor E, Kozma J, Szentirmai A (1996) Biotech Letters 18:701 Karaffa L, Sándor E, Kozma J, Kubicek CP, Szentirmai A (1999) Appl Microbiol Biotechnol 51:633 Sándor E, Pusztahelyi T, Karaffa L, Karanyi Z, Pócsi I, Biro S, Szentirmai A, Pócsi I (1998) FEMS Microbiol Lett 164:231 Sándor E, Szentirmai A, Paul GC, Colin RT, Pócsi I, Karaffa L (2001) Can J Microbiol 47:801 Dowzer CEA, Kelly JM (1989) Curr Genet 15:457 Suárez T, Peñalva MA (1996) Mol Microbiol 20:529 Tilburn J, Sarkar S,Widdick DA, Espeso EA, Orejas M, Mungroo J, Peñalva MA,Arst HN (1995) EMBO J 14:779 MacCabe AP, van den Hombergh JPTW, Tilburn J, Arst HN, Visser J (1996) Mol Gen Genet 250:367 Espeso EA, Tilburn J, Sánchez-Pulido L, Brown CV, Valencia A, Arst HN, Peñalva MA (1997) J Mol Biol 274:466 Negrete-Urtasun S, Reiter W, Diéz E, Denison SH, Tilburn J, Espeso EA, Peñalva MA, Arst HN (1999) Mol Microbiol 33:994 Caddick MX, Brownlee AG, Arst HN (1986) Mol Gen Genet 203:346 Orejas M, Espeso EA, Tilburn J, Sarkar S, Arst HN, Peñalva MA (1995) Genes Dev 9: 1622 Martín JF, Casqueiro J, Kosalková K, Marcos AT, Gutiérrez S (1999) Antonie Van Leeuwenhoek 75:21 Behmer CJ, Demain AL (1983) Curr Microbiol 8:107 Strauss J, Mach RL, Zeilinger S, Hartler G, Stoffler G, Wolschek M, Kubicek CP (1995) FEBS Lett 376:103 Cziferszky A, Mach RL, Kubicek CP (2002) J Biol Chem 277:14688 Schmitt EK, Kück U (2000) J Biol Chem 275:9348 Emery P, Durand B, Mach B, Reith W (1996) Nucl Acids Res 24:803 Gajiwala KS, Chen H, Cornille F, Roques BP, Reith W, Mach B, Burley SK (2000) Nature 403:916 Wu SY, McLeod M (1995) Mol Cell Biol 15:1479 Huang M, Zhou Z, Elledge SJ (1998) Cell 94:565 Iwama A, Pan J, Zhang P, Reith W, Mach B, Tenen DG, Sun Z (1999) Mol Cell Biol 19:3940


43

170. Waldburger JM, Masternak K, Muhlethaler-Mottet A,Villard J, Peretti W, Landmann S, Reith W (2000) Immunol Rev 178:148 171. Schmitt EK, Bunse A, Janus D, Hoff B, Friedlin E, Kürnsteiner H, Kück U (2004) Eukaryotic Cell 3:121 172. Haas H, Marzluf GA (1995) Curr Genet 28:177 173. Litzka O, Papagiannopolous P, Davis MA, Hynes MJ, Brakhage AA (1998) Eur J Biochem 251:758 174. Kück U, Pöggeler S (unpublished) 175. Newbert RW, Barton B, Greaves P, Harper J, Turner G (1997) J Ind Microbiol Biotechnol 19:18 176. Jekosch K, Kück U (unpublished) 177. Walz M, Kück U (1991) Curr Genet 19:73 178. Jekosch K, Kück U (1999) Fungal Genet Newsl 46:11 179. Skatrud PL, Queener SW, Carr LG, Fischer DL (1987) Curr Genet 12:337 180. Kück U, Walz M, Mohr G, Mracek M (1989) Appl Microbiol Biotechnol 31:358 181. Nowak C, Kück U (1994) Curr Genet 25:34 182. Nowak C, Radzio R, Kück U (1995) Appl Microbiol Biotechnol 43:1077 183. Walz M, Kück U (1993) Curr Genet 24:421 184. Velasco J, Adrio JL, Moreno MA, Díez B, Soler G, Barredo JL (2000) Nature Biotech 18:857 185. Morita S, Kuriyama M, Nakatsu M, Kitano K (1994) Biosci Biotechnol Biochem 58:627 186. Morita S, Kuriyama M, Nakatsu M, Suzuki M, Kitano K (1995) J Biotechnol 42:1 187. Radzio R, Kück U (1997) Appl Microbiol Biotechnol 48:58 188. Usher JJ, Hughes DW, Lewis MA, Chiang SD (1992) J Ind Microbiol Biotechnol 10:157 189. DeModena AL, Gutíerrez S,Velasco J, Fernández FJ, Fachini RA, Galazzo JL, Hughes DE, Martín JF (1993) Bio Technol 11:926 190. Basch J, Chiang S-JD (1998) J Ind Microbiol Biotechnol 20:344 191. Isogai T, Fukagawa M,Armori I, Iwami M, Kojo H, Ono T, Ueda Y, Kohasaka M, Imanaka H (1991) BioTechnol 9:188

Received: February 2004


Regulation of Penicillin Biosynthesis in Filamentous Fungi Axel A. Brakhage 1, 2 (✉) · Petra Spröte 1 · Qusai Al-Abdallah 1 · Alexander Gehrke 1 · Hans Plattner 1 · André Tüncher 1 1

2

University of Hannover, Institute of Microbiology, Schneiderberg 50, 30167 Hannover, Germany [email protected]

1 1.1 1.2 1.3

Introduction . . . . . . . . . . . . . . . . . . . . . Fungi as Producers of b-Lactam Antibiotics . . . . . Antibiotics as Secondary Metabolites . . . . . . . . General Aspects Concerning b-Lactam Biosyntheses

. . . .

47 47 47 48

2

Biosynthesis of Penicillins and Cephalosporins: An Outline . . . . . . . . .

49

3 3.1 3.2 3.3 3.3.1 3.3.2

Molecular Genetics of Penicillin and Cephalosporin Biosynthesis in Fungi Genetic Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clustering of Biosynthesis Genes . . . . . . . . . . . . . . . . . . . . . . . Structural Genes and Proteins . . . . . . . . . . . . . . . . . . . . . . . . Genes Common to Penicillin and Cephalosporin-Producing Fungi . . . . Gene Specific for Penicillin Biosynthesis: aatA (penDE) Encoding Acyl Coenzyme A:Isopenicillin N Acyltransferase . . . . . . . . . . . . . . . . Compartmentation of Gene Products and Transport of Penicillins . . . . Molecular Regulation of b-Lactam Biosynthesis Genes . . . . . . . . . . . General Aspects of the Elucidation of the Regulation of Secondary Metabolism Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Promoter Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carbon Source Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . pH Regulation Mediated by the Transcriptional Activator PACC . . . . . . Nitrogen Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amino Acids as Mediators of Regulation . . . . . . . . . . . . . . . . . . Influence of Oxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The CCAAT-Box Binding Protein Complex AnCF . . . . . . . . . . . . . . The A. nidulans bHLH Protein AnBH1 . . . . . . . . . . . . . . . . . . . . Velvet A (veA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Cephalosporin C Regulator CPCR1 Identified in A. chrysogenum Is very Likely also Present in Both A. nidulans and P. chrysogenum . . . . Recessive Trans-Acting Mutations Affecting the Expression of Penicillin Biosynthesis Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G-Protein-Mediated Signal Transduction . . . . . . . . . . . . . . . . . . Post-Transcriptional Regulation . . . . . . . . . . . . . . . . . . . . . . . Regulation of Penicillin Biosynthesis in Fungal Production Strains . . . . Evolution of b-Lactam Biosynthesis Genes in Fungi . . . . . . . . . . . .

. . . .

52 52 52 53 53

. . .

58 60 60

. . . . . . . . . .

60 61 62 63 67 68 69 69 71 72

.

73

. . . . .

73 74 74 75 76

3.4 3.5 3.5.1 3.5.2 3.5.3 3.5.4 3.5.5 3.5.6 3.5.7 3.5.8 3.5.9 3.5.10 3.5.11 3.5.12 3.5.13 3.5.14 3.6 3.7

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

46

A. A. Brakhage et al.

4 4.1 4.2

Applied Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Increase of Expression of Penicillin Biosynthesis Genes . . . . . . . . . . . Genetic Engineering of b-Lactam Biosynthesis Pathways . . . . . . . . . .

77 78 79

5

Future Prospects

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

79

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

79

References

Abstract The b-lactam antibiotic penicillin is one of the mainly used antibiotics for the therapy of infectious diseases. It is produced as end product by some filamentous fungi only, most notably by Aspergillus (Emericella) nidulans and Penicillium chrysogenum. The penicillin biosynthesis is catalysed by three enzymes which are encoded by the following three genes: acvA (pcbAB), ipnA (pcbC) and aatA (penDE). The genes are organised into a gene cluster. Although the production of secondary metabolites as penicillin is not essential for the direct survival of the producing organisms, several studies indicated that the penicillin biosynthesis genes are controlled by a complex regulatory network, e.g. by the ambient pH, carbon source, amino acids, nitrogen etc. A comparison with the regulatory mechanisms (regulatory proteins and DNA elements) involved in the regulation of genes of primary metabolism in lower eukaryotes is thus of great interest. This has already led to the elucidation of new regulatory mechanisms. Positively acting regulators have been identified such as the pH dependent transcriptional regulator PACC, the CCAAT-binding complex AnCF and seem also to be represented by recessive trans-acting mutations of A. nidulans (prgA1, prgB1, npeE1) and P. chrysogenum (carried by mutants Npe2 and Npe3). In addition, repressors like AnBH1 and VeA are involved in the regulation. Furthermore, such investigations have contributed to the elucidation of signals leading to the production of penicillin and can be expected to have a major impact on rational strain improvement programs. Keywords Penicillin biosynthesis · Regulation of penicillin biosynthesis · Aspergillus nidulans · Penicillium chrysogenum List of Abbreviations 6-APA 6-Aminopenicillanic acid A Adenine Å Ångstrom AA Amino acids AAA Aminoadipic acid ACV d-(L-a-Aminoadipyl)-L-cysteine-D-valine AF Aflatoxin AMP Adenosine monophosphate Arg Arginine Asp Aspartic acid ATP Adenosine triphosphate b-GAL b-Galactosidase b-GLU b-Glucuronidase bp Base-pairs bHLH Basic-region helix-loop-helix C Carbon C Cytosine

Regulation of Penicillin Biosynthesis in Filamentous Fungi Cys D DAC DAOC DNA EMSA G Gln Gly h His IPN kbp kDa mg mL mmol mRNA nt ORF RT-PCR SDS-PAGE ST T Thr Val

47

Cysteine Deletion Deacetylcephalosporin C Deacetoxycephalosporin C Deoxyribose nucleic acid Electrophoretic mobility shift assay Guanine Glutamine Glycine Hours Histidine Isopenicillin N Kilo base-pairs Kilo Dalton Milligram Millilitre Millimolar Messenger ribonucleic acid Nucleotide Open reading frame Reverse transcription polymerase chain reaction Sodium dodecylsulfate polyacrylamide gel electrophoresis Sterigmatocystin Thymine Threonine Valine

1 Introduction 1.1 Fungi as Producers of b -Lactam Antibiotics A literature survey covering more than 23,000 microbial products possessing some biological activity, i.e. antifungal, antibacterial, antiviral, cytotoxic and immunosuppressive, shows that the producing strains are mainly from the fungal kingdom (ca. 42%), followed by strains belonging to the genus Streptomyces (32.1%) [1]. Hence, fungi are one of the most important sources of bioactive compounds. 1.2 Antibiotics as Secondary Metabolites The metabolism of fungi can be divided into two parts, the primary metabolism which provides the cells with energy and chemical precursors which are essential for growth and reproduction of the organisms, and the secondary metabolism which seems to possess no obvious function in cell growth [2]. Com-

48


pounds with antibiotic activity mainly belong to the group of secondary metabolites. Fungi produce numerous secondary metabolites which show antibiotic activity against various microorganisms, antiviral or antitumour and/or fungicidal activity. Some of the secondary metabolites, however, are too toxic for therapeutic applications and are therefore classified as mycotoxins some of which show mutagenic or even carcinogenic potential [3]. 1.3 General Aspects Concerning b -Lactam Biosyntheses The discovery of antibiotics is perhaps the most important discovery in the history of therapeutic medicine. It may conceivably have saved more lives than any other medical therapy [4]. The modern antibiotic therapy started with the discovery of a b-lactam antibiotic in 1929, when Alexander Fleming published his observation about the inhibition of growth of Staphylococcus aureus on an agar plate contaminated with Penicillium notatum [5]. This discovery led to the development of the b-lactam penicillin, the first clinically used antibiotic. During the late 1940s the fungus Cephalosporium acremonium (renamed to Acremonium chrysogenum) was isolated from the sea at Cagliari (Italy) by Guiseppi Brotzu [6]. The discovery of cephalosporin C generated a whole new group of clinically significant b-lactams. The success of b-lactams in the treatment of infectious disease is due to their high specificity and their low toxicity. Despite a growing number of antibiotics and the incidence of penicillin-resistant isolates, b-lactams are still by far the most frequently used antibiotic [7, 8]. It is only in the past 20 years that the biosynthesis pathways leading to penicillins and cephalosporins have been elucidated. This is in part due to the fact that industrial production of penicillin and cephalosporin was achieved with Penicillium chrysogenum and Acremonium chrysogenum (syn. Cephalosporium acremonium), respectively. These fungi, however, belong to the deuteromycetes which are in general difficult to analyse genetically. Currently, the greatest progress in elucidation of the molecular regulation of biosyntheses of b-lactams in fungi has been made in the penicillin-producer Aspergillus (Emericella) nidulans, since this fungus is an ascomycete with a sexual cycle. Hence, classical genetic techniques can be applied to A. nidulans [9] and as the result, a detailed genetic map is available [10]. The genome sequence of A. nidulans is publicly available (www.broad.mit.edu/annotation/fungi/aspergillus/index.html). Together with molecular techniques, this facilitated a thorough analysis of the genetic regulation of metabolic pathways, including that of penicillin biosynthesis [11–14]. According to their chemical structures b-lactams can be classified into five groups (Fig. 1). All of these compounds have in common the four-membered b-lactam ring. Apart from the monolactams, which have a single ring only, blactams consist of a bicyclic ring system. The ability to synthesise b-lactams is wide-spread in nature. It was found in some fungi, but also in some Gram-pos-

Regulation of Penicillin Biosynthesis in Filamentous Fungi

49

itive and Gram-negative bacteria (Fig. 1). However, whereas for the production of the hydrophilic cephalosporins organisms belonging to all three groups were described, the hydrophobic penicillins are only produced as end-product by filamentous fungi (Fig. 1). For the remaining groups of b-lactams listed in Fig. 1, so far only bacterial producers have been reported. The number of prokaryotic and eukaryotic microorganisms able to synthesize b-lactam antibiotics is continuously increasing [13, 15]. The biosynthesis of b-lactam compounds and their molecular genetics was subject to several recent reviews [13–19]. In particular, the molecular biology of b-lactam biosynthesis in fungi has seen a tremendous increase in knowledge within the last few years. The regulation of the penicillin biosynthesis will be considered mainly in the remainder of this chapter. For regulatory aspects concerning the cephalosporin biosynthesis, the article of Schmitt et al. in this volume is recommended.

2 Biosynthesis of Penicillins and Cephalosporins: An Outline Penicillins and cephalosporins belong chemically to the group of b-lactam antibiotics. The biosynthesis of both penicillins and cephalosporins have the first two steps in common [13] (Fig. 2). All naturally occurring penicillins and cephalosporins produced by eukaryotic or prokaryotic microorganisms are synthesised from the same three amino acid, L-a-aminoadipic acid (L-a-AAA), L-cysteine and L-valine (Fig. 2). In fungi, the non-proteinogenic amino acid La-AAA is derived from the fungus specific aminoadipate pathway which leads to formation of lysine. It can also be provided by catabolic degradation of lysine although the contribution of this pathway to penicillin biosynthesis has not been clarified yet. In bacteria, a specific pathway for formation of L-a-AAA for b-lactam biosynthesis has been found [13]. In the first reaction of the cephalosporin and penicillin biosynthesis pathway, the amino acid precursors are condensed to the tripeptide d-(L-aaminoadipyl)-L-cysteine-D-valine (ACV). This reaction is catalysed by a single enzyme, d-(L-a-aminoadipyl)-L-cysteine-D-valine synthetase (ACVS) (see below). ACVS is encoded by a single structural gene designated acvA (pcbAB) (Fig. 2). In the second step, oxidative ring closure of the linear tripeptide leads to formation of a bicyclic ring, i.e., the four-membered b-lactam ring fused to the five-membered thiazolidine ring which is characteristic of all penicillins. The resulting compound isopenicillin N (IPN) possesses weak antibiotic activity and is thus the first bioactive intermediate of both penicillin and cephalosporin pathways. This reaction is catalysed by isopenicillin N synthase (IPNS) encoded by the ipnA (pcbC) gene (see below). IPN is the branch point of penicillin and cephalosporin biosyntheses (Fig. 2). In the third and final step of penicillin biosynthesis, the hydrophilic L-aAAA side chain of IPN is exchanged for a hydrophobic acyl group catalysed by

50


Fig. 1 Naturally occurring classes of b-lactam antibiotics essentially according to O’Sullivan and Sykes [180] and as shown in Aharonowitz et al. [51] and Brakhage [13]

Fig. 2 Biosyntheses of penicillin, cephalosporin C and cephamycin C. Gene and organism names are printed in italics, names of enzymes in capital letters. L-a-AAA is an intermediate of the L-lysine biosynthetic pathway but can also be provided by catabolic degradation of L-lysine. The penicillin biosynthesis occurs in fungi only, whereas cephalosporins are synthesised in both fungi, e.g. cephalosporin C by A. chrysogenum, and bacteria, e.g. cephamycin C by S. clavuligerus. Abbreviations: ACV, d-(L-a-aminoadipyl)-L-cysteine-D-valine; DAC, deacetylcephalosporin C; DAOC, deacetoxycephalosporin C; IPN, isopenicillin N; L-a-AAA, L-a-aminoadipic acid

52


acyl coenzyme A:isopenicillin N acyltransferase (IAT). The corresponding gene was designated aatA (penDE). In natural habitats penicillins such as penicillin F and K, which contain D3-hexenoic acid and octenoic acid as side chains, respectively, are synthesised. By supplying the cultivation medium with phenylacetic or phenoxyacetic acid, the synthesis can be directed mainly towards penicillin G and V, respectively [17] (Fig. 2). The side chain precursors have to be activated before they become substrates for the IAT. It is generally believed that the activated forms of the side chains consist of their CoA-thioesters, but the mechanism behind this activation is still not fully elucidated [20] (see below). The formation of hydrophobic penicillins has been reported in fungi only, notably P. chrysogenum and A. nidulans, whereas the hydrophilic cephalosporins are produced by both fungi and bacteria, e.g. A. chrysogenum and Streptomyces clavuligerus, respectively (Fig. 2).

3 Molecular Genetics of Penicillin and Cephalosporin Biosynthesis in Fungi 3.1 Genetic Nomenclature Before their identification, the putative genes encoding ACVS were designated pcbA (penicillin cephalosporin biosynthesis) and pcbB, because it was believed that two enzymes were involved in the formation of an AC dipeptide and the final ACV tripeptide, respectively (Fig. 2) [21, 22]. Cloning and sequencing of the corresponding gene revealed, however, that a single polypeptide encoded by a single gene is responsible for the formation of the ACV tripeptide. Publications reporting the DNA sequence of the P. chrysogenum, A. nidulans and A. chrysogenum genes named the gene acvA, which reflected the involvement of one genetic locus in the synthesis of ACVS [23–26] or, pcbAB derived from the combination of pcbA and pcbB [27, 28]. The gene encoding IAT was named penDE or aat. In this review the gene is designated aatA, reflecting both the correct genetic nomenclature and that one genetic locus encodes the enzyme. The IPNS gene was named ipnA [14]. The alternative names are shown in parentheses at the beginning of the relevant sections. 3.2 Clustering of Biosynthesis Genes So far as we know, in bacteria and fungi all structural genes of b-lactam biosyntheses are clustered (Fig. 3). The penicillin biosynthesis genes form a singe cluster, whereas in A. chrysogenum, two clusters containing the cephalosporin biosynthesis genes were identified. By contrast in cephamycin C producing bacteria the cephamycin biosynthesis genes are organised into a single cluster (Fig. 3) [23, 25, 28–31]. The linkage of antibiotic-biosynthesis genes is a well-


53

known phenomenon in many antibiotic-producing organisms. It has been speculated that linkage has occurred during evolution owing to an ecological selective advantage [32]. Seno and Baltz [33] have suggested that coordinated regulation of antibiotic-biosynthesis genes could be achieved by organising the genes into large operons controlled by a single promoter. For example, genes of the actinorhodin biosynthesis pathway in Streptomyces coelicolor are clustered and expressed in several polycistronic messages [34]. In eukaryotic fungi, however, b-lactam biosynthesis genes are transcribed separately, and are expressed from different promoters [14]. Hence, in fungi, there is no obvious need for clustering and it thus seems more likely that linkage reflects a common ancestral origin (see below). However, there is no evidence that the aatA gene has a close relative in modern prokaryotes, even though it is part of the cluster. This fact supports the hypothesis that linkage might also confer an ecological advantage to the eukaryotic fungi in their natural habitat, although the reason for this is not yet understood. 3.3 Structural Genes and Proteins 3.3.1 Genes Common to Penicillin and Cephalosporin-Producing Fungi 3.3.1.1 a -Aminoadipyl)-L-Cysteinyl-D-Valine Synthetase (ACVS) acvA (pcbAB) Encoding d (L-a The first reaction which has been shown for the biosyntheses of penicillin and cephalosporin/cephamycin is the formation of the d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (ACV) tripeptide. All of the reactions required for synthesis of the tripeptide are catalysed by a single enzyme, d-(L-a-aminoadipyl)-L-cysteine-D-valine synthetase (ACVS) which is encoded by the acvA (pcbAB) gene (Fig. 2). Thus, the ACV tripeptide is formed via a non-ribosomal enzyme thiotemplate mechanism from its amino acid precursors. This is similar in many aspects to the synthesis of other microbial peptides [35–37] (see chapter von Döhren et al.). The first isolation of an ACVS protein was achieved by van Liempt et al. [38] who partially purified ACVS of A. nidulans 118-fold. Since then,ACVS enzymes have been purified from different organisms, including P. chrysogenum, S. clavuligerus, A. chrysogenum and N. lactamdurans [35, 36, 39].Although not entirely clarified, it is believed that ACVS multienzymes are monomers. They exhibit different catalytic activities such as the specific recognition of the three amino acid precursors and their activation, peptide bond formation, isomerisation of the L-valine moiety to the D-form etc.As in ribosomal peptide biosynthesis, the carboxyl function of the amino acid is activated by the formation of a mixed anhydride with the a-phosphate of ATP, resulting in the release of pyrophosphate [38]. After activation of an amino acid, the formed aminoacyl

54


adenylate is cleaved by the action of an enzyme thiol, resulting in formation of a thioester bond between the enzyme (at an appropriate location on the enzyme) and the amino acid, and in the release of AMP. These thioesterified amino acids are high-energy intermediates which are the targets for a nucleophilic attack by the amino group of a second amino acid, resulting in the formation of a peptide bond. As in the ribosome, the nascent peptide grows from the amino-terminus to the carboxy-terminus and the intermediate peptides remain bound (as thioesters) to the enzyme. Substrate specificity is less strict than in protein synthesis, since a variety of tripeptide analogs are known [13, 35]. L-Valine is apparently epimerized to the D-form at the tripeptide stage since no D-valine intermediate has been detected (Fig. 2) [36, 40]. Each ACVS is encoded by a single structural gene (designated acvA or pcbAB) with a size of more than 11 kb (Table 1). The translational start codons of the acvA genes of all fungi are putative because attempts to obtain the N-terminal amino acid sequence proved to be unsuccessful [24, 41]. The genes were cloned and sequenced from P. chrysogenum, A. nidulans, A. chrysogenum and bacterial cephamycin producers such as N. lactamdurans, S. clavuligerus and Lysobacter lactamgenus [13, 15, 19, 26, 27, 42]. Even in fungi, the ORF is not interrupted by introns. Fungal acvA genes are divergently oriented to the ipnA genes (Fig. 3). The genes are separated by about 1 kbp. Sizes of the intergenic regions between both genes vary slightly among the different fungi (Table 1). The order of the biosynthesis of the AAA-Cys-Val tripeptide is believed to reflect the linear organisation of the ACVS in AAA-, Cys- and Val-activating domains [36]. Sequencing of the ACVS structural genes (Table 1) revealed that in the three repeated regions of about 600 amino acids of each ACVS some similarity to 4¢-phosphopantetheine attachment sites described for polyketide synthases (i.e. DSL) is evident [24]. This seems to reflect the attachment of multiple cofactors to ACVS. Because a single phosphopantetheine arm is sufficient for activity of fatty acid synthases, the finding of several phosphopantetheine attachment sites suggest a modified mechanism for the thiotemplate pathway to polypeptides (multiple cofactor model) [24, 43–45]. Although the relevance of all three pantetheine attachment sites of ACVSs has not been proved experimentally yet, it is currently believed that peptide assembly is accomplished by transfer of acyl intermediates between adjacent cofactors [44, 45]. Recently, a putative 4¢-phosphopantetheinyl transferase which is essential for penicillin biosynthesis in A. nidulans was characterised. It is encoded by the npgA/cfwA gene. Mutations in this gene led to defects in growth and pigmentation. Furthermore, the mutant did not produce penicillin [46, 47]. In the carboxyl-terminal region of ACVS enzymes, sequence similarities to the thioesterase active site region, GXSXG, have been found which would be required to release the generated tripeptide from the enzyme [24]. ACV synthetases are of special interest since they represent a route for peptide bond formation independent of the ribosome and allow the incorporation


55

Fig. 3 b-Lactam biosynthesis gene cluster in fungi and bacteria. The A. chrysogenum genes cefD1 and cefD2 are located next to the pcbAB and pcbC gene [181]. Bacterial genes with fungal homologs are boxed. The transcriptional orientation and the transcript units (Bacteria), as far as it has been determined, are indicated by arrows below the boxes. Arrows between boxes (Bacteria) and arrows with broken lines below boxes mark the orientation of genes. ORF specifies an open reading frame whose function is unknown. Abbreviations not mentioned in the text: cmcT, transmembrane protein; pbp, penicillin-binding protein; bla, b-lactamase; blp, showing similarity to the extracellular b-lactamase inhibitory protein BLIP; ORF, open reading frame [14]

56


Table 1 acvA (pcbAB). d(L-a-aminoadipyl]-L-cysteinyl-D-valine synthetase

A. nidulans

DNA (bp)

No Mr of aa

11,310

3770

Transcript- Transcript- InDoRef. trons mains size (kb) starta

422,486 > 9.5

P. chrysogenum OLI13 11,328a 3776a 423,996a AS-P-78 11,376 3792 425,971 11.5

Major: –230 Minor: –317 –195 –188

–

3

[23, 24]

n.d. n.d.

– –

3 33

[25, 26] [27]

a Correction of

published sequence. Translation start moved upstream by 90 bps (Brakhage and Turner, 1995).

of many non-proteinogenic amino acids [48]. Furthermore, since different parts of peptide synthetases are specific for certain amino acids, this can be used to engineer genetically new peptide synthetases producing new compounds, possibly with new pharmacological activities [49] (see chapter von Döhren et al.). 3.3.1.2 ipnA (pcbC) Encoding Isopenicillin N Synthase (IPNS) The second step of the penicillin/cephalosporin biosynthesis, i.e., the cyclisation of the linear ACV tripeptide to the bicyclic isopenicillin N (IPN), is catalysed by isopenicillin N synthase (IPNS), a nonheme Fe(II)-dependent oxidase (Fig. 2) (Table 2). The enzyme formally catalyses the removal of four hydrogen equivalents of the ACV tripeptide in a desaturative ring closure with concomitant reduction of dioxygen to water [15, 50, 51]. The IPNS reaction requires ferrous iron, molecular oxygen as cosubstrate and ascorbate as electron donor to form the b-lactam and thiazolidine ring of IPN [22]. IPNS was purified to homogeneity from A. chrysogenum [52–54] and has subsequently been obtained from P. chrysogenum, A. nidulans, several actinomycetes such as S. clavuligerus, S. lipmanii, N. lactamdurans, and the Gramnegative bacterium Flavobacterium sp. [15]. Only the free thiol form of ACV serves as a substrate, the bis-disulfide dimer, which is spontaneously formed, being inactive [55]. In P. chrysogenum, a broad-range disulfide reductase belonging to the thioredoxin family of oxidoreductases was found which efficiently reduced bis-ACV to the thiol monomer.When coupled to IPNS in vitro, it converted bis-ACV to IPN and was therefore suggested to play a role in penicillin biosynthesis [56]. The crystal structure of the A. nidulans IPNS was solved at a resolution of 2.5 Å and 1.3 Å complexed with manganese [50], and with


57

Fe2+ and substrate [57], respectively. The active-structure shows the manganese ion attached to four protein ligands (His 214, Asp 216, His 270, Gln 330) and bears two water molecules occupying coordination sites directed into a hydrophobic cavity within the protein [50]. The Fe(II):ACV:IPNS structure has one protein molecule with ferrous ion and ACV bound at the active site. The side chain of Gln 330, which coordinates the metal in the absence of substrate, is replaced by the ACV thiolate [57]. In the substrate complex, three of the five coordination sites are filled with protein ligands: His214, His270 and Asp216 [58]. The remaining two sites are occupied by a water molecule (at position 398) and the ACV thiolate. Such a structural characteristic (an iron-binding site within an unreactive hydrophobic substrate binding cavity) is probably a requirement for this class of enzyme, as it results in the isolation of the reactive complex and subsequent intermediates from the external environment. Thus, the reaction can be channelled along a single path, avoiding the many side reactions potentially open to the highly reactive species resulting from the reduction of dioxygen at the metal [50]. Data on the mechanism of the IPNS reaction suggests that initial formation of the b-lactam ring is followed by closure of the thiazolidine ring [59]. The current model of the catalytic mechanism can be found in Roach et al. [50, 57]. IPNS shows broad substrate specificity in particular with alterations in the L-a-AAA moiety and the valine residue of ACV. This finding has an ingenious use in creating novel penicillins from ACV analogs although cyclisation of unnatural tripeptides occurs at lower efficiency [60, 61]. The genes encoding IPNS enzymes are designated ipnA (pcbC) (Table 2). ipnA (pcbC) genes have been isolated from different fungi and bacteria such as A. chrysogenum, A. nidulans, P. chrysogenum, S. clavuligerus, S. griseus, S. lipmanii, Flavobacterium sp., N. lactamdurans etc. [15, 62, 63]. The properties of the fungal genes and their corresponding deduced amino acid sequences are summarized in Table 2. In contrast to bacteria, in fungi ipnA and acvA are bidirectionally oriented (Fig. 3). Fungal IPNS genes identified until now do not possess introns (Table 2).

Table 2 ipnA (pcbC). Isopenicillin N synthase

DNA (bp)

No of aa

Mr

Transcript- Transcript size (kb) starta

A. nidulans

993

331

37,480

~1.7

Major: –106 –

P. chrysogenum

993

331

38,012

1.1

Major: –11

a

Introns Ref.

–

[88, 170, 182, 183] [86, 161, 162]

Values for transcript starts were determined by primer extension or by S1 mapping.

58


3.3.2 Gene Specific for Penicillin Biosynthesis: aatA (penDE) Encoding Acyl Coenzyme A:Isopenicillin N Acyltransferase The third and final reaction of penicillin biosynthesis, which does not occur in cephalosporin biosynthesis and has been found in fungi only, is catalysed by acyl coenzyme A:isopenicillin N acyltransferase (IAT). The hydrophilic L-a-AAA side chain is exchanged for a hydrophobic acyl group, e.g. phenylacetyl in penicillin G (Fig. 2). IAT shows a broad substrate specificity [19, 64, 65]. By addition of appropriate precursor molecules, the fermentation can be directed towards a specific penicillin, e.g. for production of penicillin G, phenylacetic acid is added, for production of penicillin V, phenoxy acetic acid (Fig. 2). Once the precursor has been taken-up, it must be activated to its CoA thioester. A possible candidate to carry out this reaction is the acetyl-CoA synthetase (ACS) which was purified from P. chrysogenum and its structural gene acuA (=facA of A. nidulans) cloned. It was shown that the ACS enzymes of both P. chrysogenum and A. nidulans have the capability to catalyse in vitro the activation (to their CoA thioesters) of some of the side chain precursors required for the production of several penicillins by these fungi [65]. Putatively different ACS-like enzymes have been described as well [66, 67]. In addition, a specific phenylacetic acid-activating ACS-like enzyme designated phenylacetic acid-CoA ligase was isolated and its encoding gene cloned [20, International patent WO97/02349]. Phenylacetic acid activation by ACS appears to be poor [68], and disruption of the acuA gene does not affect penicillin biosynthesis [69, International patent WO92/07079]. Furthermore, overproduction of phenylacetic acid-CoA ligase, however, does not seem to result in a higher penicillin production, either. It is interesting to note, that overproduction of the pcl gene from Pseudomonas putida U encoding a phenylacetic acid-CoA ligase which most probably resides in the fungal cytosol, increased penicillin production two-fold indicating a possible role for a cytosolic enzyme [20, 70]. An efficient penicillin production is apparently hampered by the degradation of the precursors such as phenylacetic acid. A gene of A. nidulans was cloned designated phacA which encodes a cytochrome P450 monooxygenase. The enzyme catalyses the 2-hydroxylation of phenylacetate. It is involved in the degradation of phenylacetate to fumarate and acetoacetate. phacA disruption increased penicillin production three- to fivefold, indicating that catabolism competes with antibiotic biosynthesis for phenylacetate [71]. The corresponding gene of P. chrysogenum (pahA) was also cloned. In contrast to A. nidulans, P. chrysogenum is unable to use phenylacetic acid as sole carbon source. This block in phenylacetic acid catabolism could be originated by inactivation or strong reduction of PAHA activity. However, interestingly, PAHA activity displayed an inverse correlation with the penicillin productivity of the P. chrysogenum strains studied. Comparison of pahA genes of several strains revealed that an L181F mutation was responsible for the reduced function of PAHA in


59

present industrial strains compared with the wild-type NRRL1951. The mutation was tracked down to strain Wisconsin 49–133 [72]. A two step enzymatic process for conversion of IPN to penicillin G by IAT has been proposed [73]. In the first step, IPN is deacylated to 6-aminopenicillanic acid (6-APA), which in the second step is acylated to penicillin G through addition of a phenylacetyl group from its CoA derivative (Fig. 2). Thus, two enzymatic functions are required, an isopenicillin-N amidohydrolase and acyl-CoA:6-aminopenicillanic acid acyltransferase activity. The cloning and sequencing of the aatA (penDE) gene encoding IAT revealed that the P. chrysogenum enzyme has the required activities [64]. The properties of the P. chrysogenum and A. nidulans aatA (penDE) genes are summarised in Table 3. In contrast to the other penicillin biosynthesis genes (acvA and ipnA) the aatA genes contain three introns in both organisms at similar positions [74–76]. No DNA sequence homologous to the aatA gene of P. chrysogenum was found in the genome of three different strains of A. chrysogenum and actinomycetes [19]. This finding is consistent with the notion that 6-aminopenicillanic acid: acyltransferase activity which is also carried out by IAT is lacking in A. chrysogenum and other cephalosporin producers [77]. Therefore, these organisms do not produce penicillin G or any other penicillins with a hydrophobic side chain. The active form of the IAT enzyme results from processing of the 40 kDa monomeric precursor to a heterodimer containing subunits of 11 and 29 kDa [19, 65, 78]. Both subunits are required for activity [79]. In P. chrysogenum, it was shown that the processing event that generated the two subunits from the 40 kDa precursor polypeptide occurred between Gly102/Cys103 [80]. Additional investigations suggest that the formation of recombinant IAT involves cooperative folding events between the subunits and IAT hydrolysis is an autocatalytic event [79]. Site-directed mutagenesis of the aatA gene and production of the mutant enzyme in E. coli revealed that Cys103 is required for IAT proenzyme cleavage. Whether this requirement reflects a direct participation of Cys103 in cleavage or as part of a cleavage recognition site has not been clarified yet. However, it cannot be entirely excluded yet that Cys103 is involved in IAT enzyme activity because all of these experiments were based on the deTable 3 aatA (penDE). Acyl coenzyme A: isopenicillin N acyltransferase

DNA (bp)

No of aa

Mr

Transcript- Transcriptsize (kb) start

Introns Ref.

A. nidulans

1237

357

39,240

1.2

–61 (–60) –52, –82

3

[75, 76, 85, 184, 185]

P. chrysogenum

1274

357

39,943

1.15

n.d.

3

[74, 76, 186]

60


tection of enzyme specific activity [81]. The encoded amino acid sequence in the cleavage site is identical in P. chrysogenum and A. nidulans (Arg-AspGly...Cys-Thr-Thr) [80–82]. 3.4 Compartmentation of Gene Products and Transport of Penicillins The penicillin biosynthesis pathway occurs in different compartments of the cell, as reviewed by Driessen and colleagues (see chapter Evers et al.). 3.5 Molecular Regulation of b -Lactam Biosynthesis Genes 3.5.1 General Aspects of the Elucidation of the Regulation of Secondary Metabolism Genes One of the problems in elucidating the transcriptional control of b-lactam biosynthesis genes is that the physiological meaning of the production of blactams for the producing fungi is not entirely understood. It is generally accepted that penicillin and cephalosporin act as antibiotics in the soil against competing bacteria but an experimental proof of this assumption is difficult to obtain. Hence, as long as the physiological meaning of b-lactams for the producing fungi is not fully understood it is not possible to predict the regulatory circuits involved in the regulation of b-lactam biosyntheses. Therefore, an alternative strategy is based on the identification of regulatory proteins and to elucidate to which regulatory circuits these proteins belong. By this means, it should be feasible to unravel the physiology behind the production of b-lactams in fungi. As for most genes, transcriptional control is a major determinant of the appearance of their products. In case of the b-lactam biosynthesis genes in fungi, there are some studies directly measuring steady state levels of mRNA and, in addition, studies using reporter gene fusions. In the latter case, the promoter regions of the b-lactam biosynthesis genes including codons encoding some of their N-terminal amino acids were fused in frame with the Escherichia coli reporter genes lacZ or uidA encoding b-galactosidase (b-GAL) and b-glucuronidase (b-GLU), respectively. The most sophisticated system offers A. nidulans. This fungus has no significant endogenous b-GLU activity and, in addition, there are mutants available which have no endogenous b-GAL activity. Furthermore, this fungus has, in contrast to the deuteromycetes P. chrysogenum and A. chrysogenum which are employed for industrial production of penicillin and cephalosporin, respectively, a well defined sexual cycle facilitating genetic analyses [14]. Although by using gene fusions it cannot be entirely excluded that also posttranscriptional regulation contributes to the overall measurement of the re-


61

porter enzymes, in this review the data is formally used as indicative for transcriptional regulation. Although penicillin is a secondary metabolite, in A. nidulans and P. chrysogenum with lactose as the carbon source its production occurs right from the beginning of the fermentation run.A strict separation of trophophase (growth phase) and idiophase (metabolite production phase), which has been observed in many antibiotic-producing bacterial cultures, was absent [83–85]. This is consistent with the notion that in fermentation medium, A. nidulans acvA and ipnA gene fusions were expressed for up to 68 h and 46 h, respectively [83]. In contrast, aatA expression was only detected for about 24 h [85]. However, in P. chrysogenum ipnA steady-state mRNA levels increased with the age of the culture, indicating preferential transcription of the gene at late growth times [86]. This is consistent with the observation that expression of an ipnA-uidA gene fusion was only detectable after 24 h in fermentation medium with lactose as the carbon source. In contrast, an acvA-uidA gene fusion seemed to be expressed from the beginning of a fermentation run [87]. Hence, there might even be differences in the temporal expression among genes of the same cluster as well as among genes of clusters in different fungi. 3.5.2 Promoter Structures Studies to analyse promoters of the penicillin biosynthesis genes have been reported of both fungi, i.e., A. nidulans and P. chrysogenum. Deletion analyses revealed that the intergenic region between acvA and ipnA of A. nidulans contains several regions containing cis-acting DNA elements. Furthermore, the promoters of both genes are, at least in part, physically overlapping and share common cis-acting elements [88, 89]. In P. chrysogenum, a deletion analysis of the acvA (pcbAB) promoter region showed that at least three regions are important for regulation under the conditions tested. Together with biochemical assays, such as EMSAs and uracil interference assays a TTAGTAA motif was identified. Point mutations and deletions of the entire TTAGTAA sequence which is a target site, e.g. for BAS2 (PHO2) in Saccharomyces cerevisiae, supported the involvement of this sequence in the binding of a transcriptional activator whose biochemical nature is unknown yet. Furthermore, it was shown that this sequence is required for high level expression of the acvA gene [90] (see below). The promoter strengths of b-lactam biosynthesis genes are rather different. On the basis of reporter gene fusions, it became evident that in all three fungi, i.e. P. chrysogenum, A. nidulans and A. chrysogenum expression of acvA was much weaker compared to that of ipnA [83, 85, 87, 91]. The low expression of acvA is, at least in wild-type strains of A. nidulans, rate-limiting for penicillin production because overexpression of acvA led to drastically increased production of penicillin [92], while similar overexpression of ipnA and aatA did not [75]. In A. nidulans, it was also shown that aatA had lower expression than ipnA [83, 85].

62


3.5.3 Carbon Source Regulation Industrial production of penicillin was usually carried out by using lactose as the carbon source (C-source), which gave the highest penicillin titre. The use of excess glucose leads to a drastic reduction of the penicillin titre [83, 93, 94]. This problem is partially overcome by feeding subrepressing doses of glucose and by the use of lactose as C-source [95]. Since in general the fungus grows better with glucose than with lactose [83], the production of penicillin appears to be favoured by sub-optimal growth conditions. C-source regulation seems to act at several points of the penicillin biosynthesis: (i) in P. chrysogenum flux of L-a-AAA to d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (ACV) [96]; (ii) both, transcriptional and post-transcriptional regulation of penicillin biosynthesis genes [13]. The expression of both the acvA (pcbAB) and ipnA (pcbC) gene of P. chrysogenum strain Q176, both measured by the use of the uidA reporter gene, was repressed by glucose [87]. In P. chrysogenum, a deletion analysis of the acvA gene promoter together with EMSAs using protein crude extract led to the suggestion of a putative region which could be responsible for glucose repression [90]. Renno et al. [84] claimed, however, that steady state mRNA levels of all three P. chrysogenum penicillin biosynthesis genes were highest during rapid growth when considerable levels of glucose were present. This shows that measurement of carbon regulation depends, at least in part, on the experimental approach used. In A. nidulans, results obtained with reporter gene fusions showed that the expression of the ipnA gene was repressed when glucose or sucrose was used instead of lactose as the C-source during fermentation [83, 93]. This was further supported by the finding that the IPNS specific activity was drastically reduced in glucose-grown mycelia [83]. The repression of ipnA expression by repressing C-sources occurs, at least in part, at the transcriptional level because the steady state level of ipnA mRNA decreased when mycelia were cultivated with repressing C-sources, such as sucrose [93]. Unexpectedly, in A. nidulans the expression of both acvA and aatA reporter gene fusions was only slightly, if at all, repressed by glucose in fermentation medium [83, 85]. However, the specific activity of the aatA gene product, IAT, was reduced in mycelia grown with glucose instead of lactose [83, 85]. This suggests that the glucose regulation of IAT takes place, at least in part, post-transcriptionally (see below). In contrast to the penicillin production strain AS-P-78 of P. chrysogenum investigated by Revilla et al. [97], the IAT specific activity of both A. nidulans and the P. chrysogenum wild-type strain NRRL1951 was clearly reduced in glucosegrown cultures [83]. To study the molecular basis of C-source regulation, several mutants of A. nidulans carrying previously characterized loci affecting glucose repression of several genes of the primary metabolism (creAd-1, creB304, creC302) [98, 99] were analysed. In these mutants, penicillin production was still reduced by glucose [83, 100]. However, in extreme loss-of-function mutations in creA slightly


63

derepressed ipnA steady state transcript levels were observed [93]. This was consistent with a deletion analysis of the ipnA promoter, demonstrating that a cis-acting DNA region crucial to sucrose repression maps between –1334 and –966 relative to the transcriptional start site of the gene [88]. A single CREA binding site was detected in this region, which was protected in DNase I footprint analysis using a GST::CREA protein which contained amino acids 35–240 of CREA [101]. However, the analysis of the expression of an ipnA-lacZ gene fusion revealed that the identified putative CREA binding site is not functional in vivo [101], making it unlikely that CREA plays a role in C-source repression of penicillin biosynthesis. This also agrees with the finding that acetate and glycerol, which are repressing and derepressing C-sources, respectively, of some primary metabolism genes in the creA-mediated circuit of carbon catabolite repression, behaved opposite to what would be expected from creA control. The use of acetate led to increased steady state levels of ipnA transcript and penicillin titres and glycerol led to the opposite effects, i.e. decreased ipnA transcript levels and penicillin titres [101]. Additional experiments further excluded the possibility of a direct involvement of creB and creC mutations on C-source repression of ipnA transcription [83, 100, 102]. Thus, in A. nidulans the mechanism(s) of regulation of penicillin biosynthesis by repressing C-sources remains to be elucidated. (iii) activation of side chain precursors; glucose was also found to cause inactivation of P. chrysogenum acetyl-CoA synthetase which has the capability to catalyse the activation (to their CoA thioesters) of some of the side chain precursors required for the production of several penicillins in vitro [68] (see above). Previously, it was reported that the uptake of side chain precursors of phenylacetic acid was regulated by glucose [103]. Recently, it was shown, however, that phenylacetic acid passes the plasma membrane via passive diffusion of the protonated species [104], thus excluding that the uptake could be regulated by the available C-source. 3.5.4 pH Regulation Mediated by the Transcriptional Activator PACC Penicillin production is subject to regulation by ambient pH [101, 105]. Wildtype strains of A. nidulans can grow in media over the pH range of 2.5–10.5 [106]. There was markedly more penicillin in the culture broth when the pH value of the medium was kept constant at 8.1 than at 6.5 or 5.1 [105]. The analysis of the molecular basis of this phenomenon showed that the transcriptional regulator PACC is the key player of the pH regulation (Fig. 4). The DNA sequence of the pacC gene consists of 2172 bp interrupted by 2 introns of 85 and 53 bp, respectively. The 678 residue-derived protein (Mr 72,939) revealed that PACC contains three putative Cys2His2 zinc fingers. At alkaline ambient pH, PACC activates transcription of alkaline-expressed genes, e.g., of the alkaline phosphatase and protease genes palD and prtA, respectively, and also of the penicillin biosynthesis genes ipnA [107] and very likely of acvA [108]. The intergenic region between acvA and ipnA was found to contain 4 in vitro

64


PACC binding sites designated ipnA1, ipnA2, ipnA3 and ipnA4AB, recognized by a GST::PACC(amino acids 31–195) fusion protein (Fig. 4). The fusion protein was demonstrated to bind to the core consensus GCCARG [107]. A mutation analysis of each of these sites using ipnA-lacZ gene fusions revealed that in vivo the binding site ipnA3 was most important for PACC dependent ipnA expression, whereas sites ipnA2 and ipnA4AB were less important, although site ipnA2 was bound with highest affinity by PACC in vitro. Binding site ipnA1 apparently was not required for PACC dependent ipnA expression [109]. As observed for expression of an ipnA-lacZ gene fusion [101], expression of an acvA-uidA gene fusion was increased in a PacC5 mutant strain [108]. pacC5 is an allele which is active irrespective of the ambient pH [107]. Furthermore, addition of amino acids histidine and valine to the culture medium led to acidification of ambient pH and to reduced acvA-uidA expression. This effect was not observed in a deletion strain (D183–312) carrying a deletion spanning PACC binding site ipnA3 (at nt 265–270) nor in the PacC5 mutant strain with a constitutively active PACC protein. Taken together, these data suggest that PACC also regulates acvA expression of A. nidulans predominantly from binding site ipnA3 [108]. At alkaline ambient pH, PACC prevents transcription of acid-expressed genes [107, 110, 111]. PACC must be specifically proteolysed to yield the functional (for both positive and negative roles) version containing the N-terminal 40% of the protein (Fig. 4). The processed form is functional as both activator and repressor. PACC proteolysis occurs in response to a signal provided by the six regulatory pal gene products in alkaline environments [107, 110–112]. In wild-type strains, the pal pathway is thought to introduce a modification of PACC at alkaline pH, disrupting intramolecular interactions to allow activating proteolysis [112] which leads to the removal of a negative-acting C-terminal domain. The mechanism how PACC can avoid its proteolytic activation in the absence of signal transduction has been studied [113]. The activation of PACC requires two sequential proteolytic steps. First, the ‘closed’ translation product is converted to an accessible, committed intermediate by proteolytic elimination of the C-terminus (Fig. 4). This ambient pH-regulated cleavage is required for the final, pH-independent processing reaction and is mediated by a distinct signalling protease (possibly PalB) [114]. Interestingly, ambient pH signalling also regulates nuclear localisation of PACC [115]. The P. chrysogenum pacC gene of strain NRRL1951 is encoded by 1979 bp interrupted by a single intron of 56 bp. The predicted 641-residue protein (Mr 68,681) exhibits most of the features described for the A. nidulans PACC protein, including three zinc fingers of the Cys2His2 class. A fusion protein of glutathione thiotransferase with amino acids 46–154 of P. chrysogenum PACC (GST::PACC (46–154)) overexpressed in E. coli and purified, bound in vitro to the intergenic region between P. chrysogenum acvA and ipnA. By computer analysis seven PACC binding consensus sites (5¢-GCCARG-3¢) were found in the intergenic region. This is consistent with the finding that steady state ipnA mRNA levels were increased at alkaline pH [86]. The upstream region of the P.


65

Fig. 4 Regulatory genes involved in the regulation of the penicillin biosynthesis genes of A. nidulans. The effects of the indicated amino acids are mediated by the ambient pH most likely via pal genes and the central regulatory protein PACC. The four PACC binding sites bound in vitro by PACC are marked by triangles in the intergenic region between acvA and ipnA. Site 3, which seems to be of major importance for both ipnA-lacZ [109] and acvA-uidA expression [108], is marked by a filled triangle. The model of nuclear import of PACC is adapted from Mingot et al. [71]. The two identified AnCF binding sites are marked with stippled boxes and with Roman numbers (I and II). The AnBH1 binding site is indicated by a dark grey box

66


chrysogenum aatA (penDE) gene contains eight binding sites for PACC, whereas that of A. nidulans has just one such sequence, suggesting that these genes might be regulated by PACC as well [86]. In addition, Chu et al. [116] and Feng et al. [117] reported independently of each other that partially purified crude extracts of P. chrysogenum bound in vitro to the sequences TGCCAAG and GCCAAGCC, respectively. These binding sites identified almost certainly correspond to PACC binding sites [86]. Furthermore, the A. chrysogenum PACC homologue was found to activate transcription of the ipnA gene (see chapter Schmitt et al.). PacC genes are not confined to b-lactam producing fungi, as the cloning of the A. niger pacC gene showed [118]. Hence, PACC represents a wide-domain regulator which is involved in the regulation of expression of b-lactam biosynthesis genes. Because the use of glucose or sucrose as the C-source leads to acidification of the medium [83, 93, 101] it was conceivable that the glucose/sucrose effect was due to pH regulation [101]. This was further supported by the observation that external alkaline pH could bypass sucrose repression of steady state ipnA transcript levels and penicillin titres.Additional experiments confirmed that alkaline pH is the factor derepressing penicillin production in 3% sucrose broth [101]. Furthermore, the analysis of PacC mutants revealed that mutations in pacC bypassed C-source regulation of ipnA transcript levels, i.e. pacC mutations caused derepression of steady state levels of the ipnA mRNA in sucrose broth despite external acidic pH resulting from sucrose utilisation [101]. However, neither acidic external pH nor mutations palA1, palB7 and palF5 mimicking the effects of growth at acidic pH, prevented C-source derepression [101]. Furthermore, the PACC binding sites determined in vitro by the use of a fusion polypeptide containing the PACC DNA-binding domain are not located in the cis-acting region which was shown to mediate C-source repression of ipnA-lacZ expression [101, 107]. Taken together, these data support the model of independent regulatory mechanisms, one mediating C-source regulation and another mediating pH regulation through the pacC-encoded transcriptional regulator [101, 107]. Since alkaline pH values per se seem to derepress ipnA transcription, Espeso et al. [101] proposed that alkalinity represents a physiological signal which triggers penicillin biosynthesis. The authors concluded that carbon limitation, either by using less favourable C-sources or by reducing the concentration of favourable C-sources, results in external alkalinisation, whereas sufficient availability of a favourable C-source causes external acidification. Thus, carbon and pH regulation normally act in concert, although through different mechanisms [101]. In contrast to the situation in A. nidulans, alkaline ambient pH did not seem to override the negative effect of repressing C-source on ipnA transcription in P. chrysogenum because full ipnA expression was dependent on C-source derepression irrespective of the ambient pH value [86]. The reason for the pH mediated regulation of penicillin biosynthesis is unclear. It might be connected to the observation that b-lactams exhibit increased


67

toxicity on at least some bacterial species at alkaline pH. Furthermore, bacterial competition with fungi may be more intense at alkaline pH [110]. 3.5.5 Nitrogen Regulation The effect of the availability of nitrogen source on the penicillin biosynthesis has been discussed for a long time. Sanchez et al. [119] reported the inhibition of the penicillin biosynthesis in P. chrysogenum by high levels of ammonium. In A. chrysogenum, it was found that ammonium concentrations ((NH4)2SO4) higher than 100 mmol/L strongly interfered with cephalosporin C production [120]. It was demonstrated that in P. chrysogenum, ammonium directly influenced the expression of penicillin biosynthesis genes. By using gene fusions of both penicillin biosynthesis genes ipnA and acvA with the E. coli reporter gene uidA, it was shown that the expression of both genes was repressed by addition of 40 mmol/L (NH4)Cl to lactose-grown mycelia [87]. In A. nidulans and N. crassa, global nitrogen repression/derepression is mediated by the major positive control genes areA and nit-2, respectively [121, 122]. The homologous gene of P. chrysogenum, nre, was shown to complement Nit-2 mutants of N. crassa [123]. Each of these three genes encodes regulatory factors with a single Cys-X2-Cys-X17-Cys-X2-Cys-type zinc finger that in combination with an immediate downstream basic region constitutes a DNA-binding domain. The overall amino acid sequences of these three regulatory proteins show only 30% identity, but they have 98% identity in their DNA-binding domains. These transcription factors recognise the consensus sequence GATA and can be grouped together into a GATA protein family [124]. The optimal binding sites for NIT-2 were found to consist of at least two GATA elements, which can face in the same or opposite directions, with a spacing which can vary from 3 to 30 bp [125]. A protein consisting of 181 amino-acid residues of (the 835-residue) P. chrysogenum NRE [126], containing its zinc-finger domain, fused to the N-terminus of E. coli b-GAL bound with high affinity to a DNA fragment derived from the intergenic region between acvA and ipnA of P. chrysogenum.Although there are six GATA sequences found in the intergenic region, missing contact experiments using the b-GAL-NRE fusion protein revealed that NRE strongly interacts with a site that contains two of these GATA sequences [126]. In this binding site, the two GATA core sequences are arranged in a head-to-head fashion and separated by 27 bp. Therefore, it appears very likely that nitrogen metabolite regulation of penicillin biosynthesis genes is mediated through NRE, although in vivo studies are clearly needed to relate NRE binding to potential regulatory functions [126]. This suggests that the availability of favoured nitrogen sources and thus good growth conditions, leads to reduced penicillin synthesis by the fungus. In A. nidulans, however, no evidence for nitrogen dependent regulation of the penicillin biosynthesis has been reported so far. This is also consistent with

68


the observation that the intergenic region between acvA and ipnA of A. nidulans only contains a single GATA motif, whereas six GATA sequences are found in the corresponding P. chrysogenum region [126]. Interestingly, the P. chrysogenum promoter was shown to respond to nitrogen control when transformed in A. nidulans indicating that the nitrogen repressing system of A. nidulans acted on the heterologous promoter [127]. Furthermore, it is worth to note that in the intergenic region of the corresponding A. chrysogenum genes [91], there are even 15 GATA motifs present. It is thus conceivable that these genes are also regulated by a GATA factor. 3.5.6 Amino Acids as Mediators of Regulation Because penicillin and cephalosporin are synthesised from the amino acid precursors L-a-AAA, L-cysteine and L-valine, it was conceivable that amino acids play a role in the regulation of their biosyntheses. This was supported by the observation that in both P. chrysogenum and A. nidulans the addition of L-lysine to fermentation medium led to reduced penicillin titres [83, 128]. Since La-AAA is a branch point between L-lysine and penicillin/cephalosporin biosynthesis pathways, L-lysine inhibition of penicillin biosynthesis was suggested to operate at one or more steps of the L-lysine pathway. This was based on the notion that L-lysine feedback inhibited several enzymes of the lysine biosynthesis pathway which might result in a reduced L-a-AAA pool available for penicillin production [13]. However, amino acids also directly affect the expression of b-lactam biosynthesis genes. In A. chrysogenum, it was reported that the addition of D,L-methionine to the medium led to a three- to fourfold increase in production of cephalosporin C. The increased production was paralleled by increased steady state levels of mRNAs of cephalosporin biosynthesis genes acvA, ipnA, cefEF and, to a slight extent cefG [129]. In A. nidulans, differential effects due to various amino acids in the medium on the expression of penicillin biosynthesis genes acvA and ipnA, and penicillin production were measured. L-Amino acids with a major negative effect on the expression of acvA-uidA and ipnA-lacZ gene fusions, i.e. histidine, valine, lysine and methionine (only at concentrations greater than 10 mmol/L), led to decreased penicillin titres and a decreased ambient pH during cultivation of the fungus. An analysis of deletion clones lacking binding sites of the pH dependent transcriptional factor PACC (see above) in the intergenic region between acvA-uidA and ipnA-lacZ gene fusions and in a PacC5 mutant strain suggested that the negative effects of L-histidine and L-valine on acvA-uidA expression were due to reduced activation by PACC under acidic ambient conditions caused by these amino acids (Fig. 4). The repressing effect caused by Llysine and L-methionine on acvA, however, was even enhanced in one of the deletion clones and the pacC5 mutant strain, suggesting that these amino acids act independently of PACC by so far unknown mechanisms on the gene expression [108].


69

A specific effect of the cross-pathway control on the penicillin biosynthesis was excluded. However, a secondary effect was found. It was shown that amino acid limitation led to significantly increased transcription of lysA but not of lysF. The lysF-encoded homoaconitase acts upstream of the a-aminoadipate branch point, whereas the lysA gene product, saccharopine dehydrogenase, catalyses the ultimate step of the lysine-specific branch. Starvation-dependent changes in transcription levels of lysA were dependent on the presence of the central transcriptional activator of the cross-pathway control (CPCA) which is the homologue of the Saccharomyces cerevisiae GCN4. Overproduction of CPCA decreased expression of ipnA and acvA reporter gene fusions and even more drastically reduced penicillin production. This data suggests that, upon amino acid starvation, the cross-pathway control overrules secondary metabolite biosynthesis and favours the metabolic flux towards amino acids instead of penicillin in A. nidulans [130]. 3.5.7 Influence of Oxygen The availability of oxygen is important for penicillin production. Good aeration of mycelia with oxygen is a prerequisite for high b-lactam titres [95, 131]. Since several enzymes require oxygen for their activity, like IPNS and DAOC synthetase/DAC hydroxylase, it is conceivable that this is the reason for oxygen requirement. The importance of oxygen is also supported by the possibility of increasing cephalosporin production genetically by introducing a bacterial oxygen binding protein in A. chrysogenum [132]. However, there is a contradictory report which shows that reduction of oxygen led to increased acvA and ipnA expression of P. chrysogenum, possibly as part of a stress response [84]. 3.5.8 The CCAAT-Box Binding Protein Complex AnCF Based on results with a moving window analysis of the acvA-ipnA intergenic region of A. nidulans and together with band shift and methyl interference assays, a CCAAT-containing DNA motif (box I) located 409 bp upstream of the ATG initiation codon of the acvA gene was identified which is bound by a protein complex designated AnCF (syn. PENR1), for Aspergillus nidulans CCAAT binding factor [89, 133] (Fig. 4). This CCAAT box I is of major importance for the regulation of both genes, since a 4 bp-deletion within this site (DCCA-G) led to an eightfold increase of acvA expression and simultaneously, to a reduction of ipnA expression to about 30% [89]. Furthermore, the A. nidulans aatA promoter region also contains a functional CCAAT element (box II), located about 250 bp upstream of the transcriptional start sites of aatA. It was specifically bound by the same AnCF regulatory protein complex. Substitution of the CCAAT core sequence by GATCC led to a fourfold reduction of ex-

70


pression of an aatA-lacZ gene fusion [134] indicating that the identified binding site was functional in vivo and positively influenced aatA expression (Fig. 4). The first CCAAT-box binding factor characterised in detail, was the S. cerevisiae HAP complex, which consists of at least four subunits: HAP2, HAP3 and HAP5 form a heterotrimeric complex that is essential for DNA binding, HAP4 is an acidic protein which acts as the transcriptional activation domain [135, 136]. AnCF of A. nidulans was found to consist at least of three subunits designated HapB, HapC and HapE, all of which are necessary and sufficient for binding of AnCF to the DNA [137, 138]. Deletion mutants of AnCF subunit-encoding genes in A. nidulans, i.e. hapB, hapC and hapE, show an identical phenotype of slow growth and poor conidiation. Band shift experiments and deletion analysis showed that AnCF is also involved in the regulation of the acetamidase gene amdS, which is required for the use of acetamide as the nitrogen and Csource [133, 139]. Consequently, hap mutant strains hardly grew on acetamide as the sole nitrogen and C-source indicating that AnCF plays a role in regulating amdS expression [137, 138]. In addition, the intergenic region of the bidirectionally transcribed genes lamA and lamB (needed for utilisation of lactams) as well as the promoter region of gatA (g-amino butyric acid transaminase) contain CCAAT boxes which were bound by DNA binding factors [140]. An additional CCAAT binding factor (AnCP) of A. nidulans had been proposed [141, 142] which bound in vitro to the Taka-amylase gene promoter of A. oryzae. It turned out that this complex is in fact AnCF [143, 144] although it is unclear whether AnCF always consists of the same subunits, apart from those essential for DNA binding. For AnCF, its binding consensus motif was determined by band shift assays as RRCCAATC/ARCR [13]. Deletion or mutagenesis of the AnCF binding sites in the promoters of the penicillin biosynthesis genes had opposite effects; the expression of acvA was increased eightfold, while the expression of ipnA [89] and aatA [134] was reduced. Consistent with data obtained by deletion of the AnCF binding site, in a DhapC background, expression of both an ipnA-lacZ gene fusion and an aatA-lacZ gene fusion was reduced to 10% compared to their expression in the wild type. Hence, AnCF is a positively acting factor of ipnA and aatA expression. Penicillin titres were reduced in a DhapC background as well, but only by about 30%. In addition, expression of an acvA-uidA gene fusion was hardly affected by the DhapC mutation [145]. The minor effect of lack of AnCF on penicillin production is consistent with the view that acvA expression is rate-limiting in A. nidulans wild-type strains [92]. Consistently, decreased expression of ipnA and aatA even by a factor of five as observed in the DhapC strain, only results in a reduction of penicillin production of about 30% when the expression of the acvA gene is only marginally affected. The observation that acvA expression was less affected in a DhapC strain was unexpected because it was shown that specific deletion of four nucleotides of the AnCF binding site between acvA and ipnA (box I) resulted in a strong in-


71

crease of acvA expression [89]. Thus it appears likely that, in addition to AnCF, a repressor protein binds closely to or overlaps the AnCF binding site which would explain that the AnCF binding site exhibits a repressing effect on acvA expression in the wild type. Consistent with this view, lack of AnCF binding in the DhapC mutant did not prevent binding of this putative repressor protein and hence, acvA-uidA expression was not increased. Deletion of the AnCF binding site, however, prevented binding of both AnCF and the repressor causing the phenotype of increased acvA expression. However, the existence of such a putative repressor remains to be shown experimentally. Recently, negative regulation could be assigned to AnCF as well. The lysine biosynthesis gene lysF of A. nidulans is negatively regulated by AnCF [146]. Moreover, it was shown that AnCF is negatively autoregulated by repression of the hapB gene [147]. The physiological function of HAP-like regulatory factors in lower eukaryotes remains obscure. In yeast, the HAP complex activates the expression of genes whose products are required for respiration. Hence, HAP mutants are not able to grow on non-fermentable carbon sources [135, 136]. Because A. nidulans, however, is an aerobic growing fungus, S. cerevisiae may not be a good model for the role of HAP-like complexes in aerobic growing eukaryotes. Furthermore, lack of a functional HAP complex (Dhap strains) is not lethal for A. nidulans [137, 138]. In addition, in A. nidulans AnCF regulates secondary metabolism genes (penicillin biosynthesis genes). Hence, it will be interesting to elucidate whether this particular function of a HAP-like complex requires so far for HAP-complexes unknown accessory proteins, or, alternatively, which mechanism is involved allowing a HAP-like complex to regulate certain sets of genes (see next section). It seems very likely that AnCF is conserved among the industrially important b-lactam producing fungi. This assumption is supported by the observation that DNA fragments spanning the corresponding intergenic regions between acvA and ipnA of P. chrysogenum and A. chrysogenum, and the promoter region of the P. chrysogenum aatA gene were able to dilute the complexes of the corresponding A. nidulans probes and AnCF protein [89, 134]. Computer analysis showed that DNA elements with a high degree of sequence identity to the A. nidulans AnCF site reside within the intergenic regions of both P. chrysogenum and A. chrysogenum and the aatA promoter region of P. chrysogenum. These sites could be potential targets of homologous AnCF complexes in P. chrysogenum and A. chrysogenum. 3.5.9 The A. nidulans bHLH Protein AnBH1 AnCF was shown to bind to a single CCAAT box (box II) present in the promoter of the A. nidulans aatA gene.Attempts at purifying components of AnCF by DNA affinity chromatography using a DNA fragment encoding the region of the CCAAT box II of aatA led to the identification of several protein bands

72


in an SDS-PAGE [145]. Some of these did not correspond to the known components of AnCF, i.e. HapB, HapC and HapE [138, 145]. This finding suggested that there might be additional proteins binding to the aatA promoter adjacent or overlapping to the CCAAT box. Using affinity chromatography and standard protein purification, a novel transcription factor designated AnBH1 was isolated. The corresponding anbH1 gene was cloned and found to be located on chromosome IV. The deduced AnBH1 protein belongs to the family of basic-region helix-loop-helix (bHLH) transcription factors. AnBH1 binds in vitro as a homodimer to a not previously described asymmetric E-box within the aatA promoter which overlaps with the AnCF binding site. Since deletion of anbH1 appeared to be lethal the anbH1 gene was replaced by a regulatable alcApanbH1 gene fusion. The analysis of aatAp-lacZ expression in such a strain indicated that AnBH1 acts as a repressor of aatA gene expression and therefore counteracts the positive action of AnCF [148] (Fig. 4). 3.5.10 Velvet A (veA) The velvet gene veA gene was previously shown to mediate a developmental light response [149]. In A. nidulans strains containing a wild-type allele of the velvet gene (veA+), light reduces and delays cleistothecial formation and the fungus develops asexually, whereas in the dark, fungal development is directed toward the sexual stage, forming cleistothecia. Under conditions inducing the sexual development, the veA deletion (DveA) strain is unable to develop sexual structures [150], indicating that veA is required for cleistothecium and ascospore formation. Kato et al. [151] demonstrated that veA regulates the expression of genes implicated in the synthesis of the mycotoxin sterigmatocystin and penicillin. In a DveA strain ipnA transcripts were abundant. However, surprisingly the veA deletion mutant produced less penicillin than the wild type. This contradiction might be explained by the finding that transcript of acvA analysed by RT-PCR was only detected in the veA+ strain, in both light and dark cultures [151]. Hence the authors concluded VeA repressed the transcription of the ipnA gene, and was necessary for the expression of the acvA gene. However, by contrast, previously it was shown that acvA transcript was present in veA mutant strains [24]. Consistently, acvA expression when measured via a gene fusion was clearly detectable in veA mutant strains because most of the laboratory strains of A. nidulans contain a veA mutation [83].


73

3.5.11 The Cephalosporin C Regulator CPCR1 Identified in A. chrysogenum Is very Likely also Present in Both A. nidulans and P. chrysogenum CPCR1 was identified by Schmitt and Kück [152] as binding to a region in the promoter of the cephalosporin biosynthesis gene (ipnA) pcbC, located 418 nucleotides upstream of the translational start codon (see chapter Schmitt et al.). By using degenerate oligonucleotides and PCR a putative homologous gene was isolated from P. chrysogenum (PcRFX1) which consists of 855 amino acids. PcRFX1 and CPCR1 share an overall similarity of 29% identical amino acids. However, the similarity in the DNA binding domain to CPCR1 is 60% of identical amino acid residues. Also, a putative homologue was identified in the genome of A. nidulans (see chapter Schmitt et al.). 3.5.12 Recessive Trans-Acting Mutations Affecting the Expression of Penicillin Biosynthesis Genes In A. nidulans, a mutagenesis approach led to the identification of mutants carrying recessive mutations, designated prg (for penicillin regulation) [153] and npeE1 (impaired in penicillin biosynthesis) [154]. Segregation analysis led to the identification of two different complementation groups designated prgA1 and prgB1. For npeE1, genetic analysis showed that the gene is located on linkage group IV [154]. To date, it has not been clarified whether npeE1 differs from prgA1 and prgB1. The mutants exhibited both reduced ipnA-lacZ expression and reduced penicillin titres compared with the wild-type strain. For mutants PrgA1 and PrgB1, it was demonstrated that they also differed in acvA-uidA expression levels from the wild type and that these mutants contained reduced intracellular amounts of IPNS [153]. The results obtained by genetic and biochemical analyses indicated that the mutants most likely carry mutations in positively acting regulatory genes (Fig. 4). A gene was isolated designated suAprgA1 that complemented the prgA1 phenotype to the wild type, i.e. the expression of both gene fusions and the penicillin production nearly reached wild-type levels. Analysis of suAprgA1 in the prgA1 mutant strain did not reveal any mutation in the suAprgA1 gene or unusual transcription of the gene. This suggested that the gene is a suppressor of the prgA1 mutation. The suAprgA1 gene has a size of 1245 bp. Its five exons encode a deduced protein of 303 amino acids. The putative SUAPRGA1 protein showed similarity to both the human p32 protein and the Mam33p of S. cerevisiae. The suAprgA1 gene is located on chromosome VI. Deletion of the suAprgA1 gene led to a reduction of ipnA-lacZ expression to about 50% and to a slight reduction of the acvA-uidA expression. The DsuAprgA1 strain produced about 60% of the amount of penicillin compared with the wild-type strain [155]. SUAPRGA1 was localised in the mitochondria. It appears to bind Ca2+ ions (Gehrke A, Van den Brulle J, Bielen H, Read N, Brakhage AA, unpublished

74


results). It is likely that SUAPRGA1 is involved in the generation of a physiological signal which is required for the full expression of the penicillin biosynthesis genes and thus the penicillin production [155]. This assumption is also supported by the observation that overexpression of the suAprgA1 gene in A. nidulans using the alcA promoter of A. nidulans did not result in an increase of penicillin production or expression of penicillin biosynthesis genes beyond the levels observed in wild-type strains (Van den Brulle et al. unpublished data). Using nitrosoguanidine, Cantoral et al. [156] isolated nine mutants of the P. chrysogenum Wis54-1255 strain impaired in penicillin production. Biochemical and genetic analyses suggested that two of these mutants (Npe2 and Npe3) carry mutations in regulatory genes affecting the expression of the entire penicillin biosynthesis gene cluster. 3.5.13 G Protein Mediated Signal Transduction Until now, only little information is available on the signal transduction cascades involved in the b-lactam biosynthesis. Previous work revealed that synthesis of the carcinogenic mycotoxins sterigmatocystin (ST) and aflatoxin (AT) in Aspergillus species is negatively controlled by FADA, the a-subunit of a heterotrimeric G-protein. FADA negatively regulated both asexual reproduction (conidiation) and AF/ST synthesis in these aspergilli [157]. In an A. nidulans strain containing a constitutively activated FADA (fadA G42R), both conidiation and sterigmatocystin production are repressed. Furthermore, the dominant activating fadA allele, fadAG42R, also led to an increased steady state mRNA level of the ipnA gene and concomitantly increased penicillin titres. Taken together, FADA appears to be a member of a signal transduction cascade activating the penicillin biosynthesis and, interestingly, has opposite roles in regulating the biosynthesis of penicillin and the mycotoxin ST in A. nidulans [158]. 3.5.14 Post-Transcriptional Regulation Discrepancies observed between expression of structural genes and enzyme specific activities of the corresponding proteins suggested that besides transcriptional regulation, post-transcriptional regulation of penicillin biosynthesis genes occurs [13]. It was proposed that the glucose effect on IAT specific activity was post-transcriptionally mediated [85]. Furthermore, some discrepancies between ipnA expression and detectable IPNS specific activity were observed by comparing a wild-type strain of A. nidulans with a strain carrying a disrupted acvA gene [159, 160].


75

3.6 Regulation of Penicillin Biosynthesis in Fungal Production Strains Apart from the academic interest in elucidating the molecular regulation of biosynthesis of secondary metabolites in lower eukaryotes, there is a strong interest from an industrial point of view because b-lactam compounds are still among the most sold antibiotics in the world’s antibiotic market (see chapter Barber et al.). Hence, it is desirable to analyse high producing production strains which are highly mutated and have been derived from several different strain development programmes. This will help to elucidate both the molecular basis of deregulation and thus high production and also any remaining bottlenecks. Nowadays, industrial penicillin and cephalosporin production is mainly carried out with P. chrysogenum and A. chrysogenum, respectively. Most of these strains have been produced by mutagenesis followed by screening or selection. In 1972, the initial Panlabs Inc. P. chrysogenum strain made 20,000 units of penicillin per mL in seven days (an activity equivalent to 12 mg pure penicillin G, Na salt per mL [95]). In 1990, the improved strain made 70,000 units per mL in seven days. Penicillin titres in industry in 1993 were as high as 100,000 units per mL [19, 30]. Two important genetic features of P. chrysogenum production strains have been identified: (i) amplification of structural genes and (ii) their massively increased steady state mRNA levels. Between 8 and 16 copies are present in the high producer strain P. chrysogenum BW1890 [161]. The P. chrysogenum high titre producing strains P-2 and AS-P-78 (old production strain) carry approximately nine and six copies, respectively, of penicillin biosynthesis genes [162]. Fierro et al. [163] showed that in the high titre P. chrysogenum strains E1 and AS-P-78 the amplifications are organised in tandem repeats.A conserved TTTACA hexanucleotide sequence may be involved in their generation. This TTTACA sequence borders the 106.5-kb long penicillin biosynthesis gene cluster in the wild-type strain NRRL 1951 and also the P. notatum strain ATCC 9478 (Fleming’s isolate). In P. chrysogenum mutants independently isolated, it was shown that in all three mutants deletion of the penicillin biosynthesis gene cluster had occurred at a specific site within the conserved hexanucleotide sequence [164]. It was suggested that this site may represent a hot-spot for site specific recombination after mutation with nitrosoguanidine, the process possibly being part of a fungal SOS system similar to that found in E. coli [163, 164]. In other members of a strain improvement series, the length of the amplicon was found to be 57.5 kb. Furthermore, cDNA screening has failed to identify any further transcribed elements within the coamplified region apart from those derived from the structural penicillin biosynthesis genes [165]. Taken together, these data indicated the presence of recombinogenic regions flanking the penicillin biosynthesis gene cluster [163, 165]. Sequence analysis has shown that no mutations have been generated within the promoter regions of the penicillin biosynthesis structural genes [165]. In addition, data obtained from several production strains indicated that penicillin

76


titres were not proportionally increased with copy number. Northern blot analysis established that the ipnA mRNA steady state level of strain BW 1890 was 32to 64-fold that of NRRL1951, an increase too great to be due to the amplification alone [161]. These findings suggest that the increased penicillin production in amplified strains may be due to altered regulation of the biosynthesis pathway through changes in trans-acting regulatory factors [165]. Therefore, it will be of considerable interest to compare regulatory genes already found in both A. nidulans and P. chrysogenum between P. chrysogenum wild-type and production strains. Several studies revealed interesting differences [166]. In contrast to the gene amplification of structural genes reported in P. chrysogenum production strains, in the cephalosporin C production strain A. chrysogenum LU4-79-6, the b-lactam biosynthesis genes seem to be present in single copy (see chapter Schmitt et al.). There are certainly numerous other mutations involved which lead to a high-producing phenotype. These also include factors such as stability of biosynthesis enzymes and deregulation of enzymes involved in amino acid biosynthesis pathways, and hence the amount of precursor amino acids produced [13]. Furthermore, the transport of intermediates of the penicillin biosynthesis between organelles, and the number of organelles can be predicted to be important for penicillin production strains. 3.7 Evolution of b -Lactam Biosynthesis Genes in Fungi b-Lactam biosynthesis genes were found in both some bacterial species and some fungi. Based on several observations, a horizontal transfer of b-lactam biosynthesis genes from bacteria to fungi during evolution was proposed by several authors [51, 167–170]. The arguments in favour of a horizontal gene transfer are as follows. (i) ipnA genes of fungi and bacteria show high sequence similarities. More than 60% of the nucleotide bases and 50% of the deduced amino acids are identical. (ii) Bacterial as well as fungal b-lactam genes are organized in clusters. In bacteria, the b-lactam biosynthesis genes are organized into a single cluster, as are the penicillin biosynthesis genes in fungi. The cephalosporin biosynthesis genes in A. chrysogenum are organized into two clusters located on different chromosomes (Fig. 3). This finding led to the assumption that the b-lactam biosynthesis genes were transferred as a single cluster from an ancestral prokaryote to a common ancestor of the b-lactam synthesising fungi. In the eukaryotic ancestor, the biosynthesis genes were split onto two chromosomes. One part encodes the early genes of b-lactam biosynthesis, the other the late genes. Later in the lineage an ancestor of A. nidulans and P. chrysogenum diverged from A. chrysogenum and has presumably lost the second cluster with the genes for the late stage of cephalosporin biosynthesis [7] (Fig. 3). (iii) The GC content in the third position of codons encoding the ipnA gene of A. nidulans and P. chrysogenum is unusually high and could indicate an evolutionary origin from streptomycetes which show GC contents of


77

greater than 70% [51]. (iv) Fungal acvA and ipnA genes do not contain introns indicating a bacterial origin of the genes [13]. Based on the DNA sequences of ipnA genes from Gram-positive streptomycetes and fungi and a rate of nucleotide substitution of 10–9 nucleotide changes per site per year [171],Weigel et al. [170] proposed that the transfer occurred 370 million years ago. The cloning and sequencing of an ipnA gene from a Gram-negative bacterium, Flavobacterium sp., however, led to an extension/modification of the hypothesis of horizontal gene transfer. The ipnA gene of Flavobacterium sp. shares 69% sequence identity with the streptomycetes gene and 64–65% with the fungal genes (A. chrysogenum, P. chrysogenum) [172]. A recent reevaluation of the divergence times of organisms using a protein clock suggested that Gram-positive and Gram-negative bacteria split about 2 billion years ago, prokaryotes and a eukaryotic ancestor split about 3.2–3.8 billion years ago [173]. If the gene transfer had occurred only 370 million years ago from streptomycetes to fungi as proposed by Weigel et al. [170], it could be expected that the fungal and streptomycetes genes show a greater homology than the Gram-positive (streptomycetes) and Gram-negative genes (Flavobacterium sp.).As outlined above, this is not the case [172]. Hence,Aharonowitz et al. [51] suggested that multiple gene transfer events might have occurred from bacteria to fungi. It is difficult to imagine, however, why these multiple gene transfers then happened at about the same time what would be expected from the degree of similarity between the proteins of the various organisms. In addition, Smith et al. [174] argued against a horizontal transfer. The authors criticised that the hypothesis of a horizontal gene transfer, e.g. of the ipnA gene, was made with a very limited data set and was based solely on assumptions about rates of change. They rooted the tree with two distantly related b-lactam biosynthesis enzymes. They compared the similarity of both IPNS of A. nidulans, P. chrysogenum, A. chrysogenum, S. clavuligerus, S. anulatus and Flavobacterium sp., and DAOC synthase/synthetase of S. clavuligerus and A. chrysogenum. Based on these similarities, a tree arose with conventional evolutionary descent. The authors argued that the simplest interpretation is that the genes for the two enzymes are the result of a duplication that occurred before the prokaryote/eukaryote divergence. However, if the genes appeared very early in the evolution why have most of the eukaryotes and fungi lost the gene cluster? This question cannot be seriously answered at the moment. Thus, the evolutionary origin of b-lactam biosynthesis remains speculative.

4 Applied Implications The increasing knowledge of the molecular genetics of b-lactam biosynthesis has opened up new possibilities to rationally improve b-lactam production strains and to engineer new biosynthesis pathways. This leads to the question, however, whether an improvement of productivity is still possible. Cephalo-

78


sporin C production with A. chrysogenum is well below the productivity reached with P. chrysogenum for penicillin. Therefore, there is much effort needed to increase cephalosporin C production. For penicillin, several theoretical models have been established based on the available experimental data [175]. By using detailed stoichiometric models, the theoretical yield was calculated to be 0.47–0.50 moles penicillin per mole glucose [176, 177]. Until today, the maximum theoretical yields calculated are eight to ten times higher than the overall yields observed in fed-batch cultures, and there is therefore a considerable potential for further improvement of the process. However, it seems unlikely that the maximum theoretical yields can be reached in a real process since the penicillin biosynthesis is indirectly coupled to other cellular reactions. Therefore, taking this into account it was estimated that it should still be possible to improve the current yields by a factor of four to five [175]. Several molecular strategies have been followed to improve or alter b-lactam production. (i) Introducing additional copies or overexpression using strong promoters of b-lactam biosynthesis genes. (ii) Metabolic engineering of b-lactam biosynthesis pathways by expression of heterologous genes, e.g. production of cephalosporin precursors in P. chrysogenum. (iii) Use of the increasing knowledge of peptide synthetase genes such as those for ACVS enzymes to produce novel compounds by genetic engineering [178]. (iv) Manipulation of regulatory genes which has not been reported yet because results on the identification of regulatory genes are just being accumulated. 4.1 Increase of Expression of Penicillin Biosynthesis Genes In an A. nidulans wild-type strain acvA expression is rate-limiting for penicillin production. The acvA gene promoter was replaced by the strong inducible ethanol dehydrogenase promoter (alcAp). The expression level of alcAp was determined using a strain in which the reporter gene, lacZ, is under the control of alcAp, and was found to be up to 100 times greater than that from the acvA promoter when induced in fermentation conditions with the artificial inducer cyclopentanone. Penicillin yields were found to be increased by as much as 30fold when the acvA gene was overexpressed by induction of the alcA promoter. Glucose, which strongly represses transcription from alcAp, also repressed penicillin biosynthesis in the overproducing strain [92]. Overexpression of both the ipnA and aatA gene of A. nidulans using the alcA promoter resulted in tenfold higher levels of ipnA or aatA transcripts than those resulting from transcription of the corresponding endogenous genes. This increase caused a 40-fold rise in IPNS activity or an 8-fold rise in IAT activity. Despite this rise in enzyme levels, forced expression of the ipnA gene resulted only in a modest increase in levels of exported penicillin (increase by about 25%), whereas forced expression of the aatA gene even reduced penicillin production (decrease by about 10–30%), showing that neither of these enzymes is rate-limiting for penicillin biosynthesis [75].


79

Consistent with data obtained with A. nidulans, only transformants of the low-producing, single gene copy strain Wis54-1255 containing extra copies of the whole biosynthesis gene cluster produced more penicillin, whereas transformants carrying only extra copies of individual genes did not [179]. 4.2 Genetic Engineering of b -Lactam Biosynthesis Pathways Processes based on genetic engineering to produce novel cephalosporin derivatives biosynthetically have been introduced (see chapters of Schmitt et al. and Evers et al.).

5 Future Prospects Although considerable progress has been made in the understanding of the molecular regulation of penicillin/cephalosporin biosynthesis in fungi, our picture is far from being complete. Research on the regulation of biosyntheses of b-lactam antibiotics is heading towards the elucidation of (i) further regulatory circuits involved, (ii) inducing/repressing signals, (iii) signal transduction pathways (missing links between regulatory circuits and regulatory genes), (iv) additional regulators (transcriptional factors, co-activators/co-repressors) and (v) the mode of action of these regulatory proteins. Understanding these aspects will also help to explain the possible physiological and ecological functions of b-lactams for the producing fungi, the evolution of the pathways and also the recruitment of trans-acting factors to regulate the biosynthesis genes. The application of this knowledge will contribute not only to a further increase of blactam production and to the production of novel related compounds, but also to the identification of new b-lactam producing organisms by genetic means. Acknowledgements We gratefully acknowledge the former and current members of the laboratory for their dedicated work. Research in the authors’ laboratory was supported by the Deutsche Forschungsgemeinschaft (Priority Programme SPP1152) and the European Union (EUROFUNGII).

References 1. Lazzarini A, Cavaletti L, Toppo G, Marinelli F (2000) Rare genera of actinomycetes as potential producers of new antibiotics. Antonie Van Leeuwenhoek 78:399–405 2. Bennett JW, Bentley R (1989) What’s in a name? – Microbial secondary metabolism. In: Neidleman SL (ed) Advances in applied microbiology, vol 34.Academic Press, pp 1–28 3. Bhatnagar D, Yu J, Ehrlich KC (2002) Toxins of filamentous fungi. Chem Immunol 81:167–206 4. Heatley NG (1990) Early work at Oxford on penicillin. Biochemist 12:4–7 5. Fleming A (1929) On the antibacterial action of cultures of a Penicillium, with special reference to their use in the isolation of B. influenza. Br J Exp Pathol 10:226–236

80


6. Brotzu G (1948) Ricerche su di un nuovo antibiotico. Lavori dell’Instituto d’Igiene di Cagliari 1948:1–11 7. Skatrud PL (1991) Molecular biology of the b-lactam-producing fungi. In: Bennett JW, Lasure LL (eds) More gene manipulations in fungi. Academic Press, New York, NY, pp 364–395 8. Elander RP (2003) Industrial production of b-lactam antibiotics. Appl Microbiol Biotechnol 61:385–392 9. Pontecorvo G, Roper JA, Hemmons LM, MacDonald KD, Bufton AWJ (1953) The genetics of Aspergillus nidulans. Adv Genet 5:141–238 10. Clutterbuck AJ (1993) Aspergillus nidulans, nuclear genes. In: O’Brien SJ (ed) Genetic maps, 6th edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, pp 371–384 11. Macdonald KD, Holt G (1976) Genetics of biosynthesis and overproduction of penicillin. Sci Prog 63:547–573 12. Arst HN Jr, Scazzocchio C (1985) Formal genetic methodology of Aspergillus nidulans as applied to the study of control systems. In: Bennett JW, Lasure LL (eds) Gene manipulations in fungi. Academic Press, pp 309–343 13. Brakhage AA (1998) Molecular regulation of b-lactam biosynthesis in filamentous fungi. Microbiol Mol Biol Rev 62:547–585 14. Brakhage AA, Caruso ML (2004) Biotechnical genetics of antibiotic biosynthesis. In: Esser K, Lemke P-A (eds) The Mycota. Kück U (ed) Genetics and biotechnology, 2nd edn. Springer, Berlin Heidelberg New York, pp 317–353 15. Jensen SE, Demain AL (1995) b-Lactams. In: Vining LC, Stuttard C (eds) Genetics and biochemistry of antibiotic production. Butterworth-Heinemann, Newton, Mass, pp 239–268 16. Brakhage AA (1997) Molecular regulation of the penicillin biosynthesis in Aspergillus nidulans. FEMS Microbiol Lett 148:1–10 17. Brakhage AA, Andrianopoulos A, Kato M, Steidl S, Davis MA, Tsukagoshi N, Hynes MJ (1999) HAP-like CCAAT-binding complexes in filamentous fungi: implications for biotechnology. Fungal Genet Biol 27:243–252 18. Martin JF (2000) Molecular control of expression of penicillin biosynthesis genes in fungi: regulatory proteins interact with a bidirectional promoter region. J Bacteriol 182:2355–2362 19. Martin JF, Gutiérrez S, Demain AL (1997) b-Lactams. In: Anke T (ed) Fungal biotechnology. Antibiotics. Chapman and Hall, Weinheim, pp 91–127 20. van de Kamp M, Driessen AJ, Konings WN (1999) Compartmentalization and transport in b-lactam antibiotic biosynthesis by filamentous fungi. Antonie Van Leeuwenhoek 75:41–78 21. Ingolia TD, Queener SW (1989) b-Lactam biosynthetic genes. Med Res Rev 9:245–264 22. Nüesch J, Heim J, Treichler H-J (1987) The biosynthesis of sulfur-containing b-lactam antibiotics. Ann Rev Microbiol 41:51–75 23. MacCabe AP, Riach MBR, Unkles SE, Kinghorn JR (1990) The Aspergillus nidulans npeA locus consists of three contiguous genes required for penicillin biosynthesis. EMBO J 9:279–287 24. MacCabe AP, van Liempt H, Palissa H, Unkles SE, Riach MBR, Pfeifer E, von Döhren H, Kinghorn JR (1991) d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine synthetase from Aspergillus nidulans – molecular characterization of the acvA gene encoding the first enzyme of the penicillin biosynthetic pathway. J Biol Chem 266:12646–12654 25. Smith DJ, Burnham MRK, Bull JH, Hodgson JE, Ward JM, Browne P, Brown J, Barton B, Earl AJ, Turner G (1990a) b-Lactam antibiotic biosynthetic genes have been conserved in clusters in prokaryotes and eukaryotes. EMBO J 9:741–747


81

26. Smith DJ, Earl AJ, Turner G (1990b) The multifunctional peptide synthetase performing the first step of penicillin biosynthesis is a 421 073 dalton protein similar to Bacillus brevis peptide antibiotic synthetases. EMBO J 9:2743–2750 27. Diez BS, Gutierrez S, Barredo JL, van Solingen P, van der Voort LHM, Martin JF (1990) The cluster of penicillin biosynthetic genes. Identification and characterization of the pcbAB gene encoding the a-aminoadipyl-cysteinyl-valine synthetase and linkage to the pcbC and penDE genes. J Biol Chem 265:16358–16365 28. Gutiérrez S, Diez B, Montenegro E, Martin JF (1991) Characterization of the Cephalosporium acremonium pcbAB gene encoding a-aminoadipyl-cysteinyl-valine synthetase, a large multidomain peptide synthetase: linkage to the pcbC gene as a cluster of early cephalosporin biosynthetic genes and evidence of multiple functional domains. J Bacteriol 173:2354–2365 29. Gutiérrez S,Velasco J, Fernandez FJ, Martin JF (1992) The cefG gene of Cephalosporium acremonium is linked to the cefEF gene and encodes a deacetylcephalosporin C acetyltransferase closely related to homoserine O-acetyltransferase. J Bacteriol 174:3056–3064 30. Mathison L, Soliday C, Stepan T, Aldrich T, Rambosek J (1993) Cloning, characterization, and use in strain improvement of the Cephalosporium acremonium gene cefG encoding acetyl transferase. Curr Genet 23:33–41 31. Matsuda A, Sugiura H, Matsuyama K, Matsumoto H, Ichikawa S, Komatsu K-I (1992) Molecular cloning of acetyl coenzyme A: deacetylcephalosporin C O-acetyltransferase cDNA from Acremonium chrysogenum: sequence and expression of catalytic activity in yeast. Biochem Biophys Res Comm 182:995–1001 32. Martin JF, Liras P (1989) Organization and expression of genes involved in the biosynthesis of antibiotics and other secondary metabolites. Annu Rev Microbiol 43: 173–206 33. Seno ET, Baltz RH (1989) Structural organization and regulation of antibiotic biosynthesis and resistance genes in Actinomycetes. In: Shapiro S (ed) Regulation of secondary metabolism in Actinomycetes. CRC Press, Boca Raton, Fla, pp 1–48 34. Malpartida F, Hopwood DA (1986) Physical and genetic characterization of the gene cluster for the antibiotic actinorhodin in Streptomyces coelicolor A3(2). Mol Gen Genet 205:66–73 35. Aharonowitz Y, Bergmeyer J, Cantoral JM, Cohen G, Demain A, Fink U, Kinghorn J, Kleinkauf H, MacCabe A, Palissa H, Pfeifer E, Schwecke T, van Liempt H, von Döhren H, Wolfe S, Zhang J (1993) d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine synthetase, the multienzyme integrating the four primary reactions in b-lactam biosynthesis, as a model peptide synthetase. Biotechnology (NY) 11:807–810 36. von Döhren H, Keller U, Vater J, Zocher R (1997) Multifunctional peptide synthetases. Chem Rev 97:2675–2705 37. Zhang J, Demain AL (1992) ACV Synthetase. Crit Rev Biotechnol 12:245–260 38. van Liempt H, von Döhren H, Kleinkauf H (1989) d-(L-a-Aminoadipyl)-L-cysteinyl-Dvaline synthetase from Aspergillus nidulans. J Biol Chem 264:3680–3684 39. Theilgaard HB, Kristiansen KN, Henriksen CM, Nielsen J (1997) Purification and characterization of d-(L-a-aminoadipyl)-L-cysteinyl-D-valine synthetase from Penicillium chrysogenum. Biochem J 327:185–191 40. Schwecke T,Aharonowitz Y, Palissa H, von Döhren H, Kleinkauf H, van Liempt H (1992). Enzymatic characterisation of the multifunctional enzyme d-(L-a-aminoadipyl)-L-cysteinyl-D-valine synthetase from Streptomyces clavuligerus. Eur J Biochem 205:687–694 41. Baldwin JE, Bird JW, Field RA, O’Callaghan NM, Schofield CJ,Willis AC (1991) Isolation and partial characterization of ACV synthetase from Cephalosporium acremonium and Streptomyces clavuligerus. Evidence for the presence of phosphopantothenate in ACV synthetase. J Antibiot 44:241–248

82


42. Kimura H, Miyashita H, Sumino Y (1996) Organization and expression in Pseudomonas putida of the gene cluster involved in cephalosporin biosynthesis from Lysobacter lactamgenus YK90. Appl Microbiol Biotechnol 45:490–501 43. Schlumbohm W, Stein T, Ullrich C,Vater J, Krause M, Marahiel MA, Kruft V,WittmannLiebold B (1991) An active serine is involved in covalent substrate amino acid binding at each reaction center of gramicidin S synthetase. J Biol Chem 266:23135–23141 44. Stein T, Vater J, Kruft V, Wittmann-Liebold B, Franke P, Panico M, McDowell R, Morris HR (1994) Detection of 4¢-phosphopantetheine at the thioester binding site for L-valine of gramicidin S synthetase 2. FEBS Lett 340:39–44 45. Stein T, Vater J, Kruft V, Otto A, Wittmann-Liebold B, Franke P, Panico M, McDowell R, Morris HR (1996) The multiple carrier model of nonribosomal peptide biosynthesis at modular multienzymatic templates. J Biol Chem 271:15428–15435 46. Mootz HD, Schorgendorfer, Marahiel MA (2002) Functional characterisation of 4-phosphopantetheinyl transferase genes of bacterial and fungal origin by complementation of Saccharomyces cerevisiae lys5. FEMS Microbiol Lett 213:51–57 47. Keszenman-Pereyra D, Lawrence S, Twfieg M-E, Price J, Turner G (2003) The npgA/cfwA gene encodes a putative 4¢-phosphopantetheinyl transferase which is essential for penicillin biosynthesis in Aspergillus nidulans. Curr Genet 43:186–190 48. Baldwin JE, Shiau C-Y, Byford MF, Schofield CJ (1994) Substrate specificity of d-(L-aaminoadipyl)-L-cysteinyl-D-valine synthetase from Cephalosporium acremonium: demonstration of the structure of several unnatural tripeptide products. Biochem J 301:367–372 49. Mootz HD, Marahiel MA (1999) Design and application of multimodular peptide synthetases. Curr Opin Biotechnol 10:341–348 50. Roach PL, Clifton IJ, Fülöp V, Harlos K, Barton GJ, Hajdu J, Andersson I, Schofield CJ, Baldwin JE (1995) Crystal structure of isopenicillin N synthase is the first from a new structural family of enzymes. Nature 375:700–704 51. Aharonowitz Y, Cohen G, Martin JF (1992) Penicillin and cephalosporin biosynthetic genes: structure, organization, regulation, and evolution. Annu Rev Microbiol 46:461–495 52. Baldwin JE, Gagnon J, Ting H-H (1985) N-terminal amino acid sequence and some properties of isopenicillin-N synthetase from Cephalosporium acremonium. FEBS Lett 188:253–256 53. Hollander IJ, Shen VQ, Heim J, Demain AL, Wolfe S (1984) A pure enzyme catalyzing penicillin biosynthesis. Science 224:610–612 54. Pang CP, Chakravarti B, Adlington RM, Ting H-H, White RL, Jayatilake GS, Baldwin JE, Abraham EP (1984) Purification of isopenicillin N synthetase. Biochem J 222:789–795 55. Perry D,Abraham EP, Baldwin JE (1988) Factors affecting the isopenicillin N synthetase reaction. Biochem J 255:345–351 56. Cohen G,Argaman A, Schreiber R, Mislovati M,Aharonowitz Y (1994) The thioredoxin system of Penicillium chrysogenum and its possible role in penicillin biosynthesis. J Bacteriol 176:973–984 57. Roach PL, Clifton IJ, Hensgens CMH, Shibata N, Schofield CJ, Hajdu J, Baldwin JE (1997) Structure of isopenicillin N synthase complexed with substrate and the mechanism of penicillin formation. Nature 387:827–830 58. Borovok I, Landman O, Kreisberg-Zakarin R, Aharonowitz Y, Cohen G (1996) Ferrous active site of isopenicillin N synthase: genetic and sequence analysis of the endogenous ligands. Biochem 35:1981–1987 59. Baldwin JE, Adlington RM, Moroney SE, Field LD, Ting H-H (1984) Stepwise ring closure in penicillin biosynthesis. Initial b-lactam formation. J Chem Soc Chem Commun 1984:984–986


83

60. Baldwin JE,Abraham EP (1988) The biosynthesis of penicillins and cephalosporins. Nat Prod Rep 5:129–145 61. Wolfe S, Demain AL, Jensen SE, Westlake DWS (1984) Enzymatic approach to synthesis of unnatural b-lactams. Science 226:1386–1392 62. Coque J-JR, Martin JF, Calzada JG, Liras P (1991) The cephamycin biosynthetic genes pcbAB, encoding a large multidomain peptide synthetase, and pcbC of Nocardia lactamdurans are clustered together in an organization different from the same genes in Acremonium chrysogenum and Penicillium chrysogenum. Mol Microbiol 5:1125– 1133 63. Samson SM, Belagaje R, Blankenship DT, Chapman JL, Perry D, Skatrud PL, van Frank RM, Abraham EP, Baldwin JE, Queener SW, Ingolia TD (1985) Isolation, sequence determination and expression in Escherichia coli of the isopenicillin N synthetase gene from Cephalosporium acremonium. Nature 318:191–194 64. Alvarez E, Meesschaert B, Montenegro E, Gutiérrez S, Diez B, Barredo JL, Martin JF (1993) The isopenicillin N acyltransferase of Penicillium chrysogenum has isopenicillin N amidohydrolase, 6-aminopenicillanic acid acyltransferase and penicillin amidase activities, all of which are encoded by the single penDE gene. Eur J Biochem 215:323–332 65. Luengo JM (1995) Enzymatic synthesis of hydrophobic penicillins. J Antibiot (Tokyo) 648:1195–1212 66. Brunner R, Röhr M (1975) Phenylacyl:coenzyme A ligase. Methods Enzymol 43:476–481 67. Kogekar R, Deshpande VD (1982) Biosynthesis of penicillin in vitro. Purification and properties of phenyl/phenoxyacetic acid activating enzyme. Indian J Biochem Biophys 19:257–261 68. Martinez-Blanco H, Reglero A, Fernández-Valverde M, Ferrero MA, Moreno MA, Peñalva MA, Luengo JM (1992) Isolation and characterization of the acetyl-CoA synthetase from Penicillium chrysogenum. Involvement of this enzyme in the biosynthesis of penicillins. J Biol Chem 267:5474–5481 69. Gouka RJ, van Hartingsveldt W, Bovenberg RA, van Zeijl CM, van den Hondel CA, van Gorcom RF (1993) Development of a new transformant selection system for Penicillium chrysogenum: isolation and characterization of the P. chrysogenum acetyl-coenzyme A synthetase gene (facA) and its use as a homologous selection marker. Appl Microbiol Biotechnol 38:514–519 70. Minambres B, Martinez-Blanco H, Olivera ER, Garcia B, Diez B, Barredo JL, Moreno MA, Schleissner C, Salto F, Luengo JM (1996) Molecular cloning and expression in different microbes of the DNA encoding Pseudomonas putida U phenylacetyl-CoA ligase. Use of this gene to improve the rate of benzylpenicillin biosynthesis in Penicillium chrysogenum. J Biol Chem 271:33531–33538 71. Mingot JM, Peñalva MA, Fernández-Cañón JM (1999) Disruption of phacA, an Aspergillus nidulans gene encoding a novel cytochrome P450 monooxygenase catalyzing phenylacetate 2-hydroxylation, results in penicillin overproduction. J Biol Chem 274:14545–14550 72. Rodriguez-Saiz M, Barredo JL, Moreno MA, Fernández-Cañón JM, Peñalva MA, Diez B (2001) Reduced function of a phenylacetate-oxidizing cytochrome P450 caused strong genetic improvement in early phylogeny of penicillin-producing strains. J Bacteriol 183:5465–5471 73. Queener SW, Neuss N (1982) Biosynthesis of b-lactam antibiotics. In: Morin RB, Gorman M (eds) Chemistry and biology of b-lactam antibiotics, vol 3.Academic Press, London, pp 1–81 74. Barredo JL, van Solingen P, Diez B, Alvarez E, Cantoral JM, Kattevilder A, Smaal EB, Groenen MAM,Veenstra AE, Martin JF (1989) Cloning and characterization of the acyl-

84

75. 76.

77.

78. 79.

80.

81.

82.

83.

84. 85.

86.

87.

88.

89.

90.

A. A. Brakhage et al. coenzyme A: 6-amino-penicillanic-acid-acyltransferase gene of Penicillium chrysogenum. Gene 83:291–300 Fernández-Cañón JM, Peñalva MA (1995) Overexpression of two penicillin structural genes in Aspergillus nidulans. Mol Gen Genet 246:110–118 Tobin MB, Fleming MD, Skatrud PL, Miller JR (1990) Molecular characterization of the acyl-coenzyme A:isopenicillin N acyltransferase gene (penDE) from Penicillium chrysogenum and Aspergillus nidulans and activity of recombinant enzyme in Escherichia coli. J Bacteriol 172:5908–5914 Alvarez E, Cantoral JM, Barredo JL, Diez B, Martin JF (1987) Purification to homogeneity and characterization of acylcoenzyme A:6-amino penicillanic acid acyltransferase of Penicillium chrysogenum. Antimicrob Agents Chemother 31:1675– 1682 Queener SW (1990) Molecular biology of penicillin and cephalosporin biosynthesis. Antimicrob Agents Chemother 34:943–948 Tobin MB, Baldwin JE, Cole SCJ, Miller JR, Skatrud PL, Sutherland JD (1993) The requirement for subunit interaction in the production of Penicillium chrysogenum acylcoenzyme A: isopenicillin N acyltransferase in Escherichia coli. Gene 132:199–206 Aplin RT, Baldwin JE, Cole SC, Sutherland JD, Tobin MB (1993) On the production of alpha, beta-heterodimeric acyl-coenzyme A:isopenicillin N-acyltransferase of Penicillium chrysogenum. Studies using a recombinant source. FEBS Lett 319:166–170 Tobin MB, Cole SCJ, Miller JR, Baldwin JE, Sutherland JD (1995) Amino-acid substitutions in the cleavage site of acyl-coenzyme A:isopenicillin N acyltransferase from Penicillium chrysogenum: effect on proenzyme cleavage and activity. Gene 162:29–35 Aplin RT, Baldwin JE, Roach PL, Robinson CV, Schofield CJ (1993) Investigations into the post-translational modification and mechanism of isopenicillin N:acyl-CoA acyltransferase using electrospray mass spectrometry. Biochem J 294:357–363 Brakhage AA, Browne P, Turner G (1992) Regulation of Aspergillus nidulans penicillin biosynthesis and penicillin biosynthesis genes acvA and ipnA by glucose. J Bacteriol 174:3789–3799 Renno DV, Saunders G, Bull AT, Holt G (1992) Transcript analysis of penicillin genes from Penicillium chrysogenum. Curr Genet 21:49–54 Litzka O, Then Bergh K, Brakhage AA (1995) Analysis of the regulation of Aspergillus nidulans penicillin biosynthesis gene aat (penDE) encoding acyl coenzyme A:6aminopenicillanic acid acyltransferase. Mol Gen Genet 249:557–569 Suárez T, Peñalva MA (1996) Characterisation of a Penicillium chrysogenum gene encoding a PacC transcription factor and its binding sites in the divergent pcbAB-pcbC promoter of the penicillin biosynthetic cluster. Mol Microbiol 20:529–540 Feng B, Friedlin E, Marzluf GA (1994) A reporter gene analysis of penicillin biosynthesis gene expression in Penicillium chrysogenum and its regulation by nitrogen and glucose catabolite repression. Appl Environ Microbiol 60:4432–4439 Pérez-Esteban B, Orejas M, Gómez-Pardo E, Peñalva MA (1993) Molecular characterization of a fungal secondary metabolism promoter: transcription of the Aspergillus nidulans isopenicillin N synthetase gene is modulated by upstream negative elements. Mol Microbiol 9:881–895 Then Bergh K, Litzka O, Brakhage AA (1996) Identification of a major cis-acting DNA element controlling the bidirectionally transcribed penicillin biosynthesis genes acvA (pcbAB) and ipnA (pcbC) of Aspergillus nidulans. J Bacteriol 178:3908–3916 Kosalkova K, Marcos AT, Fierro F, Hernando-Rico V, Gutiérrez S, Martin JF (2000) A novel heptameric sequence (TTAGTAA) is the binding site for a protein required for high level expression of pcbAB, the first gene of the penicillin biosynthesis in Penicillium chrysogenum. J Biol Chem 275:2423–2430


85

91. Menne S,Walz M, Kück U (1994) Expression studies with the bidirectional pcbAB-pcbC promoter region from Acremonium chrysogenum using reporter gene fusions.Appl Microbiol Biotechnol 42:57–66 92. Kennedy J, Turner G (1996) d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine synthetase is a rate limiting enzyme for penicillin production in Aspergillus nidulans. Mol Gen Genet 253:189–197 93. Espeso EA, Peñalva MA (1992) Carbon catabolite repression can account for the temporal pattern of expression of a penicillin biosynthetic gene in Aspergillus nidulans. Mol Microbiol 6:1457–1465 94. Soltero FV, Johnson MJ (1952) The effect of the carbohydrate nutrition on penicillin production by Penicillium chrysogenum Q-176. Appl Microbiol 1:52–57 95. Swartz RW (1985) Penicillins. In: Blanch HW, Drew S, Wang DIC (eds) Comprehensive biotechnology. The principles, applications and regulations of biotechnology in industry, agriculture and medicine, vol 3. The practice of biotechnology: current commodity products. Pergamon Press, Oxford, UK, pp 7–47 96. Hönlinger C, Kubicek CP (1989) Regulation of d-(L-a-aminoadipyl)-L-cysteinyl-D-valine and isopenicillin N biosynthesis in Penicillium chrysogenum by the a-aminoadipate pool size. FEMS Microbiol Lett 65:71–76 97. Revilla G, Ramos FR, Lopez-Nieto MJ, Alvarez E, Martin JF (1986) Glucose represses formation of d-(L-a-aminoadipyl)-L-cysteinyl-D-valine and isopenicillin N synthase but not penicillin acyltransferase in Penicillium chrysogenum. J Bacteriol 168:947– 952 98. Bailey C, Arst HN Jr (1975) Carbon catabolite repression in Aspergillus nidulans. Eur J Biochem 51:573–577 99. Hynes MJ, Kelly J (1977) Pleiotropic mutants of Aspergillus nidulans altered in carbon source metabolism. Mol Gen Genet 150:193–204 100. Brakhage AA, Turner G (1995) Biotechnical genetics of antibiotic biosynthesis. In: Kück U (ed) The Mycota II. Genetics and biotechnology. Springer, Berlin Heidelberg New York, pp 263–285 101. Espeso EA, Tilburn J, Arst HN Jr, Peñalva MA (1993) pH regulation is a major determinant in expression of a fungal biosynthetic gene. EMBO J 12:3947–3956 102. Espeso EA, Fernández-Cañón JM, Peñalva MA (1995) Carbon regulation of penicillin biosynthesis in Aspergillus nidulans: a minor effect of mutations in creB and creC. FEMS Microbiol Lett 126:63–68 103. Fernández-Cañón JM, Reglero A, Martinez-Blanco H, Luengo JM (1989) I. Uptake of phenylacetic acid by Penicillium chrysogenum Wis54-1255: a critical regulatory point in benzylpenicillin biosynthesis. J Antibiot 42:1389–1409 104. Hillenga DJ, Versantvoort HJM, van der Molen S, Driessen AJM, Konings WN (1995) Penicillium chrysogenum takes up the penicillin G precursor phenylacetic acid by passive diffusion. Appl Environm Microbiol 61:2589–2595 105. Shah AJ, Tilburn J, Adlard MW, Arst HN Jr (1991) pH regulation of penicillin production in Aspergillus nidulans. FEMS Microbiol Lett 77:209–212 106. Rossi A, Arst HN Jr (1990) Mutants of Aspergillus nidulans able to grow at extremely acidic pH acidify the medium less than wild type when grown at more moderate pH. FEMS Microbiol Lett 66:51–53 107. Tilburn J, Sarkar S,Widdick DA, Espeso EA, Orejas M, Mungroo J, Peñalva MA,Arst HN Jr (1995) The Aspergillus PacC zinc finger transcription factor mediates regulation of both acidic- and alkaline-expressed genes by ambient pH. EMBO J 14:779–790 108. Then Bergh K, Brakhage AA (1998) Regulation of the Aspergillus nidulans penicillin biosynthesis gene acvA (pcbAB) by amino acids: implication for involvement of transcription factor PACC. Appl Environm Microbiol 64:843–849

86


109. Espeso EA, Peñalva MA (1996) Three binding sites for the Aspergillus nidulans PacC zinc-finger transcription factor are necessary and sufficient for regulation by ambient pH of the isopenicillin N synthase gene promoter. J Biol Chem 271:28825–28830 110. Arst HN Jr (1996) Regulation of gene expression by pH. In: Brambl R, Marzluf GA (eds) The Mycota III, biochemistry and molecular biology. Springer, Berlin Heidelberg New York, pp 235–240 111. Arst HN, Peñalva MA (2003) pH regulation in Aspergillus and parallels with higher eukaryotic regulatory systems. Trends Genet 19:224–231 112. Orejas M, Espeso EA, Tilburn J, Sarkar S, Arst HN Jr, Peñalva MA (1995) Activation of the Aspergillus PacC transcription factor in response to alkaline ambient pH requires proteolysis of the carboxy-terminal moiety. Genes Dev 9:1622–1632 113. Espeso EA, Roncal T, Diez E, Rainbow L, Bignell E, Alvaro J, Suarez T, Denison SH, Tilburn J, Arst HN Jr, Peñalva MA (2000) On how a transcription factor can avoid its proteolytic activation in the absence of signal transduction. EMBO J 19:719–728 + Erratum EMBO J 719:2391 114. Diez E, Alvaro J, Espeso EA, Rainbow L, Suarez T, Tilburn J, Arst HN Jr, Peñalva MA (2002) Activation of the Aspergillus PacC zinc finger transcription factor requires two proteolytic steps. EMBO J 21:1350–1359 115. Mingot JM, Espeso EA, Diez E, Peñalva MA (2001) Ambient pH signaling regulates nuclear localization of the Aspergillus nidulans PacC transcription factor. Mol Cell Biol 21:1688–1699 116. Chu Y-W, Renno D, Saunders G (1995) Detection of a protein which binds specifically to the upstream region of the pcbAB gene in Penicillium chrysogenum. Curr Genet 27:184–189 117. Feng B, Friedlin E, Marzluf GA (1995) Nuclear DNA-binding proteins which recognize the intergenic control region of penicillin biosynthetic genes. Curr Genet 27:351– 358 118. MacCabe AP, van den Hombergh JPTW, Tilburn J, Arst HN Jr, Visser J (1996) Identification, cloning and analysis of the Aspergillus niger gene pacC, a wide domain regulatory gene responsive to ambient pH. Mol Gen Genet 250:367–374 119. Sanchez S, Flores ME, Demain AL (1988) Nitrogen regulation of penicillin and cephalosporin fermentations. In: Sanchez-Esquival S (ed) Nitrogen source control of microbial processes. CRC Press, Boca Raton, Fla, pp 121–136 120. Shen Y-Q, Heim J, Solomon NA,Wolfe S, Demain AL (1984) Repression of b-lactam production in Cephalosporium acremonium by nitrogen sources. J Antibiot (Tokyo) 37:503–511 121. Fu YH, Marzluf GA (1990) nit-2, the major positive-acting nitrogen regulatory gene of Neurospora crassa, encodes a sequence-specific DNA-binding protein. Proc Natl Acad Sci USA 87:5331–5335 122. Kudla B, Caddick MX, Langdon T, Martinez-Rossi NM, Benett CF, Silbey S, Davis RW, Arst HN Jr (1990) The regulatory gene areA mediating nitrogen metabolite repression in Aspergillus nidulans. Mutations affecting specificity of gene activation alter a loop residue of a putative zinc finger. EMBO J 9:1355–1364 123. Haas H, Bauer B, Redl B, Stöffler G, Marzluf GA (1995) Molecular cloning and analysis of nre, the major nitrogen regulatory gene of Penicillium chrysogenum. Curr Genet 27:150–158 124. Marzluf GA (1997) Genetic regulation of nitrogen metabolism in the fungi. Microbiol Mol Biol Rev 61:17–32 125. Chiang TY, Marzluf GA (1994) DNA recognition by the NIT2 nitrogen regulatory protein: importance of the number, spacing, and orientation of GATA core elements and their flanking sequences upon NIT2 binding. Biochem 33:576–582


87

126. Haas H, Marzluf GA (1995) NRE, the major nitrogen regulatory protein of Penicillium chrysogenum binds specifically to elements in the intergenic promoter regions of nitrate assimilation and penicillin biosynthetic gene clusters. Curr Genet 28:177– 183 127. Kolar M, Holzmann K,Weber G, Leitner E, Schwab H (1991) Molecular characterization and functional analysis in Aspergillus nidulans of the 5¢-region of the Penicillium chrysogenum isopenicillin N synthetase gene. J Biotechnol 17:67–80 128. Demain AL (1957) Inhibition of penicillin formation by lysine. Arch Biochem Biophys 67:244–245 129. Velasco J, Gutiérrez S, Fernandez FJ, Marcos AT, Arenos C, Martin JF (1994) Exogenous methionine increases levels of mRNAs transcribed from pcbAB, pcbC, and cefEF genes, encoding enzymes of the cephalosporin biosynthetic pathway, in Acremonium chrysogenum. J Bacteriol 176:985–991 130. Busch S, Bode HB, Brakhage AA, Braus GH (2003) Impact of the cross-pathway control on the regulation of lysine and penicillin biosynthesis in Aspergillus nidulans. Curr Genet 42:209–219 131. Hilgendorf P, Heiser V, Diekmann H, Thoma M (1987) Constant dissolved oxygen concentrations in cephalosporin C fermentation: applicability of different controllers and effect on fermentation parameters. Appl Microbiol Biotechnol 27:247–251 132. DeModena JA, Gutiérrez S,Velasco J, Fernández FJ, Fachini RA, Galazzo JL, Hughes DE, Martin JF (1993) The production of cephalosporin C by Acremonium chrysogenum is improved by the intracellular expression of a bacterial hemoglobin. Biotechnology (NY) 11:926–929 133. van Heeswijck R, Hynes MJ (1991) The amdR product and a CCAAT-binding factor bind to adjacent, possibly overlapping DNA sequences in the promoter region of the Aspergillus nidulans amdS gene. Nucleic Acids Res 19:2655–2660 134. Litzka O, Then Bergh K, Brakhage AA (1996) The Aspergillus nidulans penicillin biosynthesis gene aat (penDE) is controlled by a CCAAT containing DNA element. Eur J Biochem 238:675–682 135. McNabb DS, Xing Y, Guarente L (1995) Cloning of yeast HAP5: a novel subunit of a heterotrimeric complex required for CCAAT binding. Genes Dev 9:47–58 136. Guarente, L (1992) Messenger RNA transcription and its control in Saccharomyces cerevisiae. In: Jones EW, Pringle JR, Broach JR (eds) The molecular and cellular biology of the yeast Saccharomyces cerevisiae, vol 2. Gene expression. Cold Spring Harbor Laboratory Press, pp 49–98 137. Papagiannopoulos P, Andrianopoulos A, Sharp JA, Davis MA, Hynes MJ (1996) The hapC gene of Aspergillus nidulans is involved in the expression of CCAAT-containing promoters. Mol Gen Genet 251:412–421 138. Steidl S, Papagiannopoulos P, Litzka O, Andrianopoulos A, Davis MA, Brakhage AA, Hynes MJ (1999) AnCF, the CCAAT binding complex of Aspergillus nidulans, contains products of the hapB, hapC and hapE genes and is required for activation by the pathway-specific regulatory gene amdR. Mol Cell Biol 19:99–106 139. Littlejohn TG, Hynes MJ (1992) Analysis of the site of action of the amdR product for regulation of the amdS gene of Aspergillus nidulans. Mol Gen Genet 235:81–88 140. Richardson IB, Katz ME, Hynes MJ (1992) Molecular characterization of the lam locus and sequences involved in the regulation of the AmdR protein of Aspergillus nidulans. Mol Cell Biol 12:337–346 141. Kato M,Aoyama A, Naruse F, Kobayashi T, Tsukagoshi N (1997) An Aspergillus nidulans nuclear protein, AnCP, involved in enhancement of Taka-amylase A gene expression, binds to the CCAAT-containing taaG2, amdS, and gatA promoters. Mol Gen Genet 254:119–126

88


142. Nagata O, Takashima T, Tanaka M, Tsukagoshi N (1993) Aspergillus nidulans nuclear proteins bind to a CCAAT element and the adjacent upstream sequence in the promoter region of the starch-inducible Taka-amylase A gene. Mol Gen Genet 237:251–260 143. Kato M, Aoyama A, Naruse F, Tateyama Y, Hayashi K, Miyazaki M, Papagiannopoulos P, Davis MA, Hynes MJ, Kobayashi T, Tsukagoshi N (1998) The Aspergillus nidulans CCAAT-binding factor, AnCP/AnCF, is a heteromeric protein analogous to the HAP complex of Saccharomyces cerevisiae. Mol Gen Genet 257:404–411 144. Kato M, Naruse F, Kobayashi T, Tsukagoshi N (2001) No factors except for the hap complex increase the Taka-amylase A gene expression by binding to the CCAAT sequence in the promoter region. Biosci Biotechnol Biochem 65:2340–2342 145. Litzka O, Papagiannopoulos P, Davis MA, Hynes MJ, Brakhage AA (1998) The penicillin regulator PENR1 of Aspergillus nidulans is a HAP-like transcriptional complex. Eur J Biochem 251:758–767 146. Weidner G, Steidl S, Brakhage AA (2001) The Aspergillus nidulans homoaconitase gene lysF is negatively regulated by the multimeric CCAAT-binding complex AnCF and positively regulated by GATA sites. Arch Microbiol 175:122–132 147. Steidl S, Hynes MJ, Brakhage AA (2001) The Aspergillus nidulans multimeric CCAAT binding complex AnCF is negatively autoregulated via its hapB subunit gene. J Mol Biol 306:643–653 148. Caruso ML, Litzka O, Martic G, Lottspeich F, Brakhage AA (2002) Novel basic-region helix-loop-helix transcription factor (AnBH1) of Aspergillus nidulans counteracts the CCAAT-binding complex AnCF in the promoter of a penicillin biosynthesis gene. J Mol Biol 323:425–439 149. Yager LN (1992) Early developmental events during asexual and sexual sporulation in Aspergillus nidulans. Bio/Technology 23:19–41 150. Kim H, Han K, Kim D, Han D, Jahng K, Chae K (2002) The veA gene activates sexual development in Aspergillus nidulans. Fungal Genet Biol 37:72–80 151. Kato N, Brooks W, Calvo AM (2003) The expression of sterigmatocystin and penicillin genes in Aspergillus nidulans is controlled by veA, a gene required for sexual development. Eukaryotic Cell 2:1178–1186 152. Schmitt EK, Kück U (2000) The fungal CPCR1 protein, which binds specifically to b-lactam biosynthesis genes, is related to human regulatory factor X transcription factors. J Biol Chem 275:9348–9357 153. Brakhage AA,Van den Brulle J (1995) Use of reporter genes to identify recessive transacting mutations specifically involved in the regulation of Aspergillus nidulans penicillin biosynthesis genes. J Bacteriol 177:2781–2788 154. Pérez-Esteban B, Gómez-Pardo E, Peñalva MA (1995) A lacZ reporter fusion method for the genetic analysis of regulatory mutations in pathways of fungal secondary metabolism and its application to the Aspergillus nidulans penicillin pathway.J Bacteriol 177:6069–6076 155. Van den Brulle J, Steidl S, Brakhage AA (1999) Cloning and characterization of an Aspergillus nidulans gene involved in the regulation of penicillin biosynthesis. Appl Environ Microbiol 65:5222–5228 156. Cantoral JM, Gutiérrez S, Fierro F, Gil-Espinosa S, van Liempt H, Martin JF (1993) Biochemical characterisation and molecular genetics of nine mutants of Penicillium chrysogenum impaired in penicillin biosynthesis. J Biol Chem 268:737–744 157. Hicks JK, Yu JH, Keller NP, Adams TH (1997) Aspergillus sporulation and mycotoxin production both require inactivation of the FadA G a protein-dependent signaling pathway. EMBO J 16:4916–4923 158. Tag A, Hicks J, Garifullina G,Ake C Jr, Phillips TD, Beremand M, Keller N (2000) G-protein signalling mediates differential production of toxic secondary metabolites. Mol Microbiol 38:658–665


89

159. Brakhage AA, Browne P, Turner G (1994) Analysis of the regulation of the penicillin biosynthesis genes of Aspergillus nidulans by targeted disruption of the acvA gene. Mol Gen Genet 242:57–64 160. Turner G, Browne PE, Brakhage AA (1993) Expression of genes for the biosynthesis of penicillin. In: Maresca B, Kobayashi GS, Yamaguchi H (eds) Molecular biology and its applications to medical mycology. NATO ASI Series H; Cell Biology, vol 69. Springer, Berlin Heidelberg New York, pp 125–138 161. Smith DJ, Bull JH, Edwards J, Turner G (1989) Amplification of the isopenicillin N synthetase gene in a strain of Penicillium chrysogenum producing high levels of penicillin. Mol Gen Genet 216:492–497 162. Barredo JL, Diez B, Alvarez E, Martin JF (1989a) Large amplification of a 35-kb DNA fragment carrying two penicillin biosynthetic genes in high penicillin producing strains of P. chrysogenum. Curr Genet 16:453–459 163. Fierro F, Barredo JL, Diez B, Gutiérrez S, Fernández FJ, Martin JF (1995) The penicillin gene cluster is amplified in tandem repeats linked by conserved hexanucleotide sequences. Proc Natl Acad Sci USA 92:6200–6204 164. Fierro F, Montenegro E, Gutiérrez S, Martin JF (1996) Mutants blocked in penicillin biosynthesis show a deletion of the entire penicillin gene cluster at a specific site within a conserved hexanucleotide sequence. Appl Microbiol Biotechnol 44:597–604 165. Newbert RW, Barton B, Greaves P, Harper J, Turner G (1997) Analysis of a commercially improved Penicillium chrysogenum strain series: involvement of recombinogenic regions in amplification and deletion of the penicillin biosynthesis gene cluster. J Ind Microbiol Biotechnol 19:18–27 166. Jekosch K, Kück U (2000) Glucose dependent transcriptional expression of the cre1 gene in Acremonium chrysogenum strains showing different levels of cephalosporin C production. Curr Genet 37:388–395 167. Carr LG, Skatrud PL, Scheetz ME II, Queener SW, Ingolia TD (1986) Cloning and expression of the isopenicillin N synthetase gene from Penicillium chrysogenum. Gene 48:257–266 168. Landan G, Cohen G, Aharonowitz Y, Shuali Y, Graur D, Shiffman D (1990) Evolution of isopenicillin N synthase genes may have involved horizontal gene transfer. Mol Biol Evol 7:399–406 169. Peñalva MA, Moya A, Dopazo J, Ramon D (1990) Sequences of isopenicillin N synthetase genes suggest horizontal gene transfer from prokaryotes to eukaryotes. Proc R Soc Lond B Biol Sci 241:164–169 170. Weigel BJ, Burgett SG, Chen VJ, Skatrud PL, Frolik CA, Queener SW, Ingolia TD (1988) Cloning and expression in Escherichia coli of isopenicillin N synthetase genes from Streptomyces lipmanii and Aspergillus nidulans. J Bacteriol 170:3817–3826 171. Li W-H, Luo C-C, Wu C-I (1985) Evolution of DNA sequences. In: MacIntyre (ed) Molecular evolutionary genetics. Plenum Press, New York, pp 1–94 172. Cohen G, Shiffman D, Mevarech M,Aharonowitz Y (1990) Microbial isopenicillin N synthase genes: structure, function, diversity and evolution. Trends Biotechnol 8:105–111 173. Feng D-F, Cho G, Doolittle RF (1997) Determining divergence times with a protein clock: update and reevaluation. Proc Natl Acad Sci USA 94:13028–13033 174. Smith MW, Feng D-F, Doolittle RF (1992) Evolution by acquisition: the case for horizontal gene transfers. Trends Biochem Sci 17:489–493 175. Nielsen J (1995) Physiological engineering aspects of Penicillium chrysogenum. Polyteknisk Forlag, Denmark 176. Hersbach GJM, van der Beek CP, van Dijck PWM (1984) The penicillins: properties, biosynthesis and fermentation. In: Vandamme EJ (ed) Biotechnology of industrial antibiotics. Marcel Dekker, New York, pp 45–140

90


177. Jorgensen HS, Nielsen J,Villadsen J, Mollgaard H (1995) Metabolic flux distributions in Penicillium chrysogenum during fed-batch cultivations. Biotechnol Bioeng 46:117–131 178. Marahiel MA, Stachelhaus T, Mootz HD (1997) Modular peptide synthetases involved in nonribosomal peptide synthesis. Chem Rev 97:2651–2673 179. Theilgaard HB, van Den Berg M, Mulder C, Bovenberg R, Nielsen J (2001) Quantitative analysis of Penicillium chrysogenum Wis54–1255 transformants overexpressing the penicillin biosynthetic genes. Biotechnol Bioeng 72:379–388 180. O’Sullivan J, Sykes RB (1986) b-Lactam antibiotics. In: Pape H, Rehm H-J (eds) Biotechnology, a comprehensive treatise in 8 volumes, vol 4. VCH Verlagsgesellschaft, Weinheim, Germany, pp 247–281 181. Ullán RV, Liu G, Casqueiro J, Gutiérrez S, Bañuelos O, Martin JF (2002) The cefT gene of Acremonium chrysogenum C10 encodes a putative multidrug efflux pump protein that significantly increases cephalosporin C production. Mol Genet Genomics 267:673–683 182. Ramon D, Carramolino L, Patino C, Sanchez F, Penalva MA (1987) Cloning and characterization of the isopenicillin N synthetase gene mediating the formation of the blactam ring in Aspergillus nidulans. Gene 57:171–181 183. Peñalva MA, Vian A, Patino C, Perez-Aranda A, Ramon D (1989) Molecular biology of penicillin production in Aspergillus nidulans. In: Hershberger CL, Queener SW, Hegeman G (eds) Genetics and molecular biology of industrial microorganisms. American Society for Microbiology, Washington DC, pp 256–261 184. Montenegro E, Barredo JL, Gutierrez S, Diez B, Alvarez E, Martin JF (1990) Cloning, characterization of the acyl-CoA:6-amino penicillanic acid acyltransferase gene of Aspergillus nidulans and linkage to the isopenicillin N synthase gene. Mol Gen Genet 221:322–330 185. Whiteman PA,Abraham EP, Baldwin JE, Fleming MD, Schofield CJ, Sutherland JD,Willis AC (1990) Acyl coenzyme A:6-aminopenicillanic acid acyltransferase from Penicillium chrysogenum and Aspergillus nidulans. FEBS Lett 262:342–344 186. Veenstra AE, van Solingen P, Bovenberg RAL, van der Voort LHM (1991) Strain improvement of Penicillium chrysogenum by recombinant DNA techniques. J Biotechnol 17:81–90

Received: March 2004


Novel Genes Involved in Cephalosporin Biosynthesis: The Three-component Isopenicillin N Epimerase System Juan F. Martín 1 (✉) · Ricardo V. Ullán 2 · Javier Casqueiro 2 1

2

University of León, Area of Microbiology, Faculty of Biology and Environmental Sciences, 24071 León, Spain Institute of Biotechnology (INBIOTEC), Avda. del Real n°1, 24006 León, Spain

1

Cephalosporins and Cephamycin-Producing Organisms

. . . . . . . . . .

92

2 2.1 2.2

Biosynthesis of Cephalosporins and Cephamycins . . . . . . . . . . . . . . Genes Encoding Enzymes Involved in Cephalosporin Biosynthesis . . . . . The Missing Gene: Where Is the Gene(s) Encoding the IPN Epimerization Step? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

93 93

Isomerases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Individual Amino Acid Racemases . . . . . . . . . . . . . . . . . . . . . Epimerization Domains in Non-Ribosomal Peptide Synthetases (NRPSs) D-Alanine Racemases in the Cyclosporin and HC-Toxin Biosynthesis Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hydroxyproline 2-Epimerase of Pseudomonas putida . . . . . . . . . . Isomerization of Fatty Acids as Acyl-CoA Derivatives . . . . . . . . . .

95 96 97

3 3.1 3.2 3.3 3.4 3.5 4 4.1 4.2 4.3 4.4 4.5

. . . . . .

93

. . 99 . . 100 . . 100

The Isopenicillin N Epimerase System . . . . . . . . . . . . . . . . . . . . A Transcriptional Map of the Region Located Downstream of pcbC Revealed Additional Genes Involved in Cephalosporin Biosynthesis . . . . The IPN Epimerase System Consists of Two Proteins with High Similarity to Acyl-CoA Synthetases and Acyl-CoA Racemases, Respectively . . . . . Targeted Inactivation of ORF1 and ORF2 Results in Mutants Blocked in Cephalosporin Production . . . . . . . . . . . . . . . . . . . . . . . . . The Disrupted Transformants Lack Isopenicillin N Epimerase Activity and Accumulate Isopenicillin N . . . . . . . . . . . . . . . . . . . . . . . Complementation of Both cefD1 and cefD2 Mutations is Required for Restoration of Epimerase Activity . . . . . . . . . . . . . . . . . . . . . .

. 101 . 101 . 101 . 101 . 102 . 102

. . . . . . . 103

5

Mechanism of Action of the Fungal Isopenicillin N Epimerase

6

Bacterial Isopenicillin N Epimerases

7

A Cephalosporin Gene Cluster in Kallichroma tethys Includes an Isopenicillin N Epimerase . . . . . . . . . . . . . . . . . . . . . . . . . . 106

8

Summary and Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . . 106

References

. . . . . . . . . . . . . . . . . . . . . 104

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

92

J. F. Martín et al.

Abstract Cephalosporin is one of the best b-lactam antibiotics, widely used in the treatment of infectious diseases. It is synthesized by Acremonium chrysogenum. The levels of cephalosporin produced by the improved strains obtained by classical mutation and selection procedures are still low compared to the penicillin titers obtained from the high-producing Penicillium chrysogenum strains. Most of the genes encoding the cephalosporin biosynthesis enzymes have been cloned, and some improvement of cephalosporin production has been achieved by removing bottlenecks in the pathway. One of the poorly-known steps involved in cephalosporin biosynthesis is the conversion of isopenicillin N into penicillin N catalyzed by the isopenicillin N epimerase system. This epimerization reaction is catalyzed by a two-component protein system encoded by the cefD1 and cefD2 genes that correspond, respectively, to an isopenicillinyl-CoA ligase and an isopenicillinyl-CoA epimerase. Comparative analysis of those proteins with others in the databanks provide evidence indicating that they are related to enzymes catalyzing the catabolism of toxic metabolites in animals. There are several biochemical mechanisms, reviewed in this article, for the biosynthesis of D-amino acids in secondary metabolites. The conversion of isopenicillin N to penicillin N in cephamycin-producing bacteria is mediated by a classical pyridoxal phosphate-dependent epimerase that is clearly different from the epimerization system existing in Acremonium chrysogenum. Modification of gene expression by directed manipulation of the cefD1-cefD2 bidirectional promoter region is a promising strategy for improving cephalosporin production. Improving our knowledge of the mechanism of epimerization systems is important if we wish to understand how microorganisms synthesize the high number of rare D-amino acids that are responsible, to a large extent, for the biological activities of many different secondary metabolites. Keywords Cephalosporin biosynthesis · Isopenicillin N · Penicillin N · Epimerization system · Coenzyme A activation · Thioesterase

1 Cephalosporins and Cephamycin-Producing Organisms b-lactam antibiotics are produced by a restricted number of microorganisms although they belong to a variety of unrelated taxons, including filamentous fungi, filamentous Gram-positive bacteria with high G+C content (Streptomyces and Amycolatopsis) [1] and unicellular Gram-negative bacteria Flavobacterium, Xanthomonas and Lysobacter [2]. This distribution of b-lactam antibiotics that does not conform to established phylogenetic patterns is the basis for the proposal of a horizontal transfer of the b-lactam biosynthesis genes among soil-inhabiting microorganisms [3, 4]. Several species of filamentous fungi belonging to the genera Acremonium (syn. Cephalosporium), Anixiopsis, Arachnomyces, Spiroidium, Scopulariopsis, Diheterospora, and Paecilomyces have been reported to produce cephalosporins [5, 6]. An early study revealed the presence of non-ribosomal peptide intermediates of cephalosporin biosynthesis in Paecilomyces persicinus [5]. Later we confirmed the presence of the pcbAB, pcbC and cefEF genes in this Paecilomyces species by hybridization (J.L. Barredo, S. Gutiérrez and J.F. Martín, unpublished results) but no further information is available on the molecular ge-

Novel Genes Involved in Cephalosporin Biosynthesis

93

netics of cephalosporin biosynthesis in this strain. Recently, an incomplete cephalosporin gene cluster has been found in the marine fungus Kallichroma tethys [7]. A comparative analysis of the K. tethys and an A. chrysogenum cephalosporin gene cluster is described at the end of this article.

2 Biosynthesis of Cephalosporins and Cephamycins The biosynthesis of cephalosporins is an excellent model for the study of secondary metabolism since considerable information on the enzymology (reviewed in [8–10]), molecular genetics and gene expression mechanisms [2, 11, 12] have accumulated in the last few years. The biosynthesis of cephalosporins by Acremonium chrysogenum (syn. Acremonium strictum) and cephamycins by Amycolatopsis (Nocardia) lactamdurans and Steptomyces clavuligerus (reviewed in [13, 14]) begins with the formation of the tripeptide d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (ACV) by the ACV synthetase [15–17], followed by cyclization of ACV to isopenicillin N (IPN). IPN is later converted into penicillin N by an epimerase activity that has remained uncharacterized so far in A. chrysogenum. After the epimerization step, penicillin N is transformed by a deacetoxycephalosporin C synthase (expandase) into deacetoxycephalosporin C (DAOC), and finally, DAOC is converted into deacetylcephalosporin C (DAC) and cephalosporin C by DAC synthase (hydroxylase) and DAC acetyltransferase, respectively [9] (Fig. 1). 2.1 Genes Encoding Enzymes Involved in Cephalosporin Biosynthesis The genes pcbAB [18] and pcbC [19] encoding, respectively, ACV synthetase and IPN synthase are linked together in chromosome VII in the so-called “early cephalosporin gene cluster” [20]. There are two other genes linked to this cluster, ORF3 encoding a putative D-hydroxyacid dehydrogenase and cefT that encodes a transmembrane protein of the major facilitator superfamily (MFS) that may be involved in cephalosporin secretion [21) (Fig. 2). The genes cefEF encoding the bifunctional DAOC synthase (expandase)-hydroxylase [22] and cefG encoding the DAC acetyltransferase [23–25] are linked together in the so-called “late cephalosporin cluster” on chromosome I [20] (Fig. 2). 2.2 The Missing Gene: Where Is the Gene(s) Encoding the IPN Epimerization Step? The IPN epimerization step has remained unclear for decades. Demain and coworkers reported that isopenicillin N was converted into penicillin N by extracts of A. chrysogenum [26–28] although the epimerizing enzyme was extremely labile [29], preventing purification and, therefore, further characteri-

Fig. 1 Biosynthesis pathway of cephalosporin C in A. chrysogenum. The step(s) catalyzed by the isopenicillin N epimerase is indicated by a question mark


95

Fig. 2 Cephalosporin gene clusters in A. chrysogenum. The cluster of “early” genes located on chromosome VII includes, in addition to pcbAB and pcbC, the epimerase genes cefD1 and cefD2 and the cefT gene encoding a transmembrane protein. The cluster of “late” genes located on chromosome I includes cefEF and cefG

zation of the protein. However, the IPN epimerase has been purified from S. clavuligerus [30, 31] and Amycolatopsis lactamdurans [32], and the cefD gene encoding this protein was cloned from both microorganisms [33, 34]. The bacterial IPN epimerases are pyridoxal-phosphate-dependent enzymes and do not appear to require ancillary proteins for their epimerization activity. Repeated attempts to clone the homologous cefD gene of A. chrysogenum using the bacterial cefD gene or oligonucleotides based on conserved amino acid sequences of bacterial epimerases with the A. chrysogenum preferred codon usage as probe were unsuccessful (S. Gutiérrez and J.F. Martín, unpublished data). These results suggested that the fungal IPN epimerization system was different to the bacterial one.

3 Isomerases The term “isomerases” designates a broad group of enzymes that are able to catalyze a variety of molecular transpositions resulting in the optical inversion of an asymmetric carbon atom. These molecular transpositions extend from simple optical inversions or racemization (epimerases and racemases) to more complex interconversions by tautomerization or intramolecular group transpositions (mutases). Isomerases are classified into five groups: 1. 2. 3. 4. 5.

Racemases and epimerases Cis-trans isomerases Intramolecular oxidoreductases Intramolecular transferases Intramolecular lyases

96


We refer in this article only to racemases and epimerases. The term racemase is used for enzymes that work on substrates with a single active site, whereas the term epimerase refers to enzymes acting on substrates with more than one active center. 3.1 Individual Amino Acid Racemases With the exception of glycine, all other amino acids are optically active and contain at least one asymmetric (optically active) carbon atom. An amino acid may have 2n optical isomers where n is the number of asymmetrically substituted carbon atoms in the molecule. Most amino acid racemases are of bacterial origin, although two different alanine racemases have been reported to be involved in fungal non-ribosomal peptide synthesis [35, 36]. The molecular mechanism of action of bacterial amino acid racemases is well known. Racemization of amino acids by these enzymes requires pyridoxal phosphate (PLP) (Fig. 3). The racemization process proceeds in four steps: 1. Formation of the initial imine (Schiff base). The amino group of the amino

acid forms an imino bond by a Schiff base mechanism with the aldehyde group of PLP. As a result of the Schiff base reaction, a system of conjugated double bonds is formed between the quaternary N+ atom of the PLP and the a carbon of the amino acid that leads to a debilitation of the C-H bond at the a carbon due to the tendency of an electron to migrate towards the quaternary N+. 2. Formation of the transitional imine. The weakening of the C-H bond at the a-carbon of the amino acid gives rise to a deprotonation forming the transitional imine.After deprotonation, the residual electron at carbon 2 confers a negative charge to this carbon atom (Fig. 3). 3. Protonation. Due to the excess negative charge at carbon 2 of the transitional imine, a protonation takes place in the opposite orientation at carbon 2, giving rise to the amino acid enantiomer. Following the protonation reaction there is a rearrangement of the double bonds in the transitional imine with the recovery of the positive charge at the quaternary nitrogen. 4. Hydrolysis. In the last step the racemases hydrolyze the double bond between the PLP and the amino group of the amino acid regenerating the PLP cofactor. The released amino acid is in the D-configuration. Bacterial racemases catalyze reversible reactions and, therefore, there is a final equilibrium of the two enantiomeric forms of the amino acid. Further reactions that consume the D-isomer for biosynthetic purposes displace the equilibrium reaction towards the D-form of the amino acid.


97

Fig. 3 Proposed molecular mechanism of conversion of an L-amino acid to its D-form by PLP-dependent amino acid racemases. The L- and D-forms of the amino acids are in shaded circles. PLP is pyridoxal phosphate. B1, B2, B3 are aminoacid residues involved in proton donation and abstraction. A tyrosine residue in the protein (Tyr, shaded) has been proposed to be involved in proton abstraction from the PLP intermediate (based on the model of Watababe et al. [74])

3.2 Epimerization Domains in Non-Ribosomal Peptide Synthetases (NRPSs) Many non-ribosomally synthesized peptides contain D-amino acids [37–40]. Precursor studies revealed that the L-amino acids rather than the D-enantiomers are incorporated into the relevant D-amino acid positions. This has been shown for bacitracin, actinomycin, etamycin, and penicillins. In the actinomycins, position 2 of the cromophore-linked peptide contains D-valine (or Dalloisoleucine). In vivo studies on actinomycin D synthesis indicated that L-valine rather than D-valine is the precursor of peptide-bound D-valine and that the epimerization proceeds via loss of hydrogen at C-2 of valine [41]. The multifunctional enzyme actinomycin synthetase II (280 kDa) assembles the amino

98


acids L-threonine and L-valine in positions 1 and 2 in the precursor peptide of actinomycin by activating them as enzyme-bound thioesters via their corresponding adenylates [42]. The C-terminal region of the ACV synthetases of Penicillium chrysogenum, Aspergillus nidulans, Acremonium chrysogenum and Amycolatopsis lactamdurans located at the end of the third module shows high similarity to the homologous regions of gramicidin synthetase I (GS1) and tyrocidine synthetase I (TY1) and the third module of the surfactin synthetase I. Since all peptide antibiotics synthesized by these peptide synthetases contain a D-amino acid in its carboxyl terminal region (D-Phe in gramicidin and tyrocidine; D-leu in surfactin) it was proposed that an epimerization region of about 365 amino acids is located in this region [43, 44]. Epimerization domains of several non-ribosomal peptide synthetases are now known and all contain characteristic signature sequences. These motifs have been found in the V domain of ACV synthetases [17] and in HC-toxin synthetase domain epimerizing L-Pro to D-Pro [35] (Fig. 4). However, very little is known about the molecular mechanism of epimerization. The involvement of a basic amino acid in these motifs as a proton donor/acceptor during racemization of phenylalanine has been proposed [44]. Epimerization of the amino acids in the epimerization domain of the nonribosomal peptide synthetases occurs while the amino acid residue is attached to the phosphopantetheine arm of the corresponding module. Similarly, the methylacyl-CoA racemase catalyzes the epimerization of the methylacyl group

Fig. 4 Conserved motifs E1 to E7 in the epimerization domain (E) of the third (Val) module of the ACV synthetases of P. chrysogenum, A. chrysogenum, A. lactamdurans and A. nidulans compared to the GrsA (GSI) and the SrfA modules of B. subtilis and the HC toxin synthetase of C. carbonum (from [40])


99

while it is activated with CoA. Further biochemical analyses of these epimerization domains are required to confirm that these integrated domains of nonribosomal peptide synthetases constitute an authentic catalytic site for amino acid epimerization. 3.3 D-Alanine Racemases in the Cyclosporin and HC-Toxin Biosynthesis Pathways On the other hand, the D-alanine component of both cyclosporin and HC-toxin is provided by a distinct alanine racemase that may interact with the peptide synthetase [36]. Cyclosporin is an undecapeptide containing D-Ala produced by Tolypocladium niveum. Analysis of the primary sequence of the cyclosporin synthetase indicated that it lacks an epimerization signature motif in its D-Alaactivating domain [45, 46]. Hoffmann and coworkers [36] provided conclusive evidence showing that a separate alanine racemase catalyzes the synthesis of D-alanine. The structural properties and kinetics of the T. niveum D-alanine racemase indicate a close relationship with the procaryotic alanine-racemases involved in cell wall biosynthesis. The T. niveumD-alanine racemase has a molecular mass of 37 kDa, in the range of the bacterial alanine racemases (about 40 kDa). The activity of the T. niveum racemase also depends on PLP. This cofactor was found to be loosely bound to the enzyme and could be removed by gel filtration or dialysis. The activity of the enzyme was inhibited by hydroxylamine, an inhibitor of PLP-dependent enzymes and by L-(1-aminoethyl)-phosphonate (alo-P) [36]. These authors provided evidence that the D-alanine racemase is limiting for cyclosporin biosynthesis. The cyclic tetrapeptide HC-toxin is an essential virulence determinant for the plant pathogenic fungus Cochliobolus carbonum. The major form of HCtoxin contains the D-isomers of Ala and Pro. The source of D-Ala has been clarified by Cheng and Walton [35]. The central enzyme in HC toxin biosynthesis is HC-toxin synthetase (HTS), a 570-kDa NRPS with four amino acid-activating domains. The HTS has only one epimerase motif, in the L-Pro activating domain. The HC-toxin producer Cochliobolus carbonum contains an alanine racemase encoded by the toxG gene involved in the conversion of L- to D-alanine, separated from the HC-toxin synthetase [35]. Therefore, in this filamentous fungus there is an epimerization domain converting L- to D-proline integrated in the HC-toxin synthetase and a separate alanine racemase. The interaction of a non-ribosomal peptide synthetase with a separate low molecular weight amino acid racemase is another interesting example of the collaboration of different amino acids activating and isomerizing enzymes for the biosynthesis of secondary metabolites.

100


3.4 Hydroxyproline 2-Epimerase of Pseudomonas putida An entirely different isomerization mechanism is involved in the conversion of 4-hydroxy-L-proline to 4-hydroxy-D-proline in Pseudomonas putida catalyzed by the enzyme hydroxyproline-2-epimerase. The enzyme purified to homogeneity does not contain PLP [47] and resembles proline racemase of Clostridium sticklandii [48]. Kinetics and structural data suggested a model for hydroxyproline 2-epimerase in which two cysteine residues act as reciprocal donor/acceptor of the a-hydrogen of the substrate, thereby effecting racemization [49]. 3.5 Isomerization of Fatty Acids as Acyl-CoA Derivatives Some isomerases catalyze isomerization reactions on acyl-CoA derivatives that are required for modifications of the fatty acids that allow b-oxidation of these compounds in eukaryotic cells. The best known are the methylacyl-CoA racemase, the D3,5-D2,4 dienoyl-CoA isomerase and D3-D2-enoyl-CoA isomerase [50]. All of these isomerases act on substrates that are activated as CoA-derivatives. The formation of the CoA derivative requires ATP and is catalyzed by a different type of enzymes: the acyl-CoA synthetases that are classified into subclasses depending upon the length of the fatty acid used as substrate. These acyl-CoA synthetases are usually associated with the endoplasmic reticulum or the mitochondrial membrane [50]. In a first step, the fatty acid is activated with ATP, forming the corresponding acyladenylate and releasing pyrophosphate. In a second step of the reaction the sulfhydryl group of CoA displaces the AMP residue from the acyl-adenylate forming acyl-CoA. Methylacyl-CoA racemase. The methylacyl-CoA racemase catalyzes the reversible conversion of (2R)methyl-acyl-CoA into (2S)methylacyl-CoA. This reaction occurs in the racemization of (2R)pristanic acid into its 2S enantiomer acid that is involved in the catabolism of pristanic acid, a product formed during degradation of the chlorophyll [51, 52]. The 2S-pristanoyl (but not its 2R enantiomer) is later degraded by pristanoyl-CoA oxidase through the b-oxidation pathway [53]. ∆3,5-∆2,4 Dienoyl-CoA isomerase and ∆3-∆2 Enoyl-CoA isomerase. During boxidation of the unsaturated fatty acids, the presence of a double bond in some configurations prevents recognition and, therefore, further degradation of these fatty acids. The dienoyl-CoA isomerase has both isomerase and reductase activity and converts trans-3-cis-5 3,5-dienoyl-CoA into trans-2, trans-4 2,4dienoyl-CoA that is further degraded [54]. Similarly, in the degradation of even-numbered unsaturated fatty acids, the ∆3-∆2-enoyl-CoA isomerase isomerizes the trans-∆3-enoyl-CoA to trans-∆2enoyl-CoA that then may be degraded by hydratase activity present in the same proteins [55, 56].


101

4 The Isopenicillin N Epimerase System 4.1 A Transcriptional Map of the Region Located Downstream of pcbC Revealed Additional Genes Involved in Cephalosporin Biosynthesis Genes encoding enzymes involved in the biosynthesis of secondary metabolites are usually clustered both in bacteria [57] and in filamentous fungi [58–60]. Since the genes encoding all other proteins involved in CPC biosynthesis are clustered in two separate loci, we hypothesized that the gene encoding the protein(s) involved in the conversion of isopenicillin N into penicillin N might be located in one of the two cephalosporin gene clusters. To search for genes located in the early cluster downstream from the pcbC gene, a transcript map of this region was made [61] using RNA extracted from mycelia of A. chrysogenum grown for 48 h in MDFA medium [62] and five probes covering 9 kb downstream of the pcbC gene. Hybridization results showed the presence in this region of two new genes cefD1 and cefD2 (Fig. 2). 4.2 The IPN Epimerase System Consists of Two Proteins with High Similarity to Acyl-CoA Synthetases and Acyl-CoA Racemases, Respectively Sequence analysis of the region downstream of pcbC, corresponding to the two new transcripts, revealed the presence of two open reading frames ORF1 and ORF2 (Fig. 2). ORF1 has 2193 nucleotides and it is interrupted by five introns. This gene was also cloned from a previously constructed cDNA library [63] and the sequence confirmed the presence of the five introns. ORF1 encodes a protein of 642 amino acids with a deduced molecular mass of 71 kDa that showed similarity to long chain acyl-CoA synthetases, particularly those from Homo sapiens and other eukaryotes [64]. The ORF1-encoded protein has all the characteristic motifs of the acyl-CoA ligases involved in the activation (usually through an adenylation step) of the carboxyl group of fatty acids or amino acids [65]. ORF2 consists of 1146 nucleotides and it is interrupted by one intron that was confirmed by RT-PCR. ORF2 encoded a protein of 383 amino acids with a deduced molecular weight of 41.4 kDa. The encoded protein showed high similarity to a-methylacyl-CoA racemases from eukaryotic cells [61]. 4.3 Targeted Inactivation of ORF1 and ORF2 Results in Mutants Blocked in Cephalosporin Production Targeted inactivation of both genes was performed with the double marker technique [66] developed for targeted inactivation in A. chrysogenum [67]. Nine

102


transformants showed a drastic reduction in the cephalosporin production and seven additional transformants showed no cephalosporin production at all. Inactivation of cefD1-cefD2 in the seven non-producing transformants took place by a canonical double crossing-over at the right position, as shown by Southern blot analysis [61]. In two other mutants with drastic reduction in the cephalosporin production a non-canonical recombination process had occurred at the cefD1-cefD2 locus. By contrast, in the transformant showing no reduction in cephalosporin production, an ectopic integration of the plasmid had occurred. 4.4 The Disrupted Transformants Lack Isopenicillin N Epimerase Activity and Accumulate Isopenicillin N Enzyme analysis showed that the isopenicillin N epimerase activity was absent in the ORF1-ORF2 disrupted strains, both at 72 and 96 h, whereas the A. chrysogenum C10 showed a high level of IPN epimerase activity. These results clearly indicated that the proteins encoded by the ORF1 and ORF2 are involved in the epimerization step in cephalosporin C biosynthesis. Therefore, the corresponding genes have been named cefD1 and cefD2 according to standard b-lactam gene nomenclature since the designation cefD is used to describe the bacterial epimerase gene [68]. HPLC analysis showed that in A. chrysogenum C10 culture broth the peaks corresponding to both isopenicillin N and penicillin N were present. However, in the three disrupted strains tested, the peak corresponding to penicillin N was absent and at the same time there was an accumulation of isopenicillin N, confirming that the disrupted strains are blocked in isopenicillin N epimerase [61]. 4.5 Complementation of Both cefD1 and cefD2 Mutations is Required for Restoration of Epimerase Activity The IPN epimerase activity was measured in three representative transformants TCD1, TCD2 and TCD1+2 to study which of the two ORFs was responsible for the epimerization step. The epimerase activity was only restored in A. chrysogenum TCD1+2 transformants (in which cefD1 and cefD2 are present). In transformants TCD1 or TCD2 where just one of the genes was functional, no epimerase activity was detected. These results indicate that both cefD1 and cefD2 proteins are indeed required for the epimerization of isopenicillin N into penicillin N. Cephalosporin C production studies with three of the TCD1, TCD2 and TCD1+2 transformants showed that, in TCD1+2 transformants, the cephalosporin production was restored to levels similar to those of the parental strain A. chrysogenum C10. HPLC analysis confirmed that TCD1+2 transformants produced authentic cephalosporin C as in A. chrysogenum C10 [61].


103

5 Mechanism of Action of the Fungal Isopenicillin N Epimerase A mechanism of action of the A. chrysogenum two-component IPN epimerase system has been proposed on the basis of homology of the CefD1 and CefD2 proteins with known eukaryotic epimerases (Fig. 5). Such epimerization systems that require previous activation of the substrate as CoA-derivatives have been reported to be involved in the racemization of phytanic acid and in the inversion of ibuprofen in humans. Phytanic acid occurs naturally as a mixture of the (3R) and the (3S)-diastereomers and is converted after a-oxidation into pristanic acid. The oxidases and dehydrogenases responsible for further b-oxidation of phytanic acid act only on (2S)-2-methylacyl-CoAs [51, 52, 69, 70] and a racemization step is required for complete degradation. In summary, a-methyl-branched fatty acids are racemized as CoA thioesters by a specific a-methylacyl-CoA racemase similar to the protein encoded by cefD2 . Another similar example is the epimerization of 2-arylpropionic acids (such as ibuprofen), an important group of non-steroidal antiinflamatory drugs.

Fig. 5 Proposed model for the conversion of the L-a-aminoadipyl side-chain of isopenicillin N into the D-a-aminoadipyl side-chain of penicillin N by the fungal three-protein system. For comparison, the direct epimerization by the bacterial isopenicillin N epimerase is shown (see text for details)

104


A unique feature regarding 2-arylpropionic acid metabolism is the stereoselective conversion of the R-enantiomer to its S-diastereomer [71]. The pathway for this metabolic event begins with the activation of the 2-arylpropionic acids as an acylCoA ester by an acyl-CoA synthetase (that resembles the protein CefD1) that is later racemized by a 2-arylpropionyl-CoA epimerase (similar to CefD2), and finally the S-diastereomer is released by a thioesterase [72, 73]. The protein encoded by the cefD2 has high similarity to a-methylacyl-racemases and 2-arylpropionyl-CoA epimerases, suggesting that after activation of isopenicillin N to isopenicillinyl-CoA through an isopenicillin-adenylate, epimerization to penicillinyl-CoA occurs (Fig. 5). The product of both genes cefD1 and cefD2 are necessary for full epimerase activity and for cephalosporin biosynthesis which supports the proposed activation and epimerization model. Finally, the required hydrolysis of the CoA thioesters requires a thioesterase (CefD3 protein). It is not known if there is a specific thioesterase dedicated to the hydrolysis of penicillinyl-CoA. In animal cells, hydrolysis of the CoA thioesters has been reported to occur in a non-stereoselective manner by different thioesterases [72]. The 2-arylpropionyl-CoA epimerase has been purified from rat liver [73], showing that it is similar to a-methylacyl-CoA racemases [52]. Enzyme purification and studies on the catalysis exerted by the two CefD1 and CefD2 proteins are required to confirm this model.

6 Bacterial Isopenicillin N Epimerases The Streptomyces clavuligerus and Amycolatopsis lactamdurans IPN epimerases appear to work by an entirely different mechanism since they consist of a single protein with an estimated molecular weight of 59,000 (for the A. lactamdurans) and 63,000 (for the S. clavuligerus enzyme) that catalyze a pyridoxal phosphate-dependent removal of the proton at C2 of the a-aminoadipyl chain, followed by reintroduction in the D-configuration [30, 32].Another bacterial isopenicillin N epimerase has been reported in the Gram-negative bacterium Lysobacter lactamgenus with identities of 48.7 and 57.0% with the A. lactamdurans and S. clavuligerus enzymes, respectively. The conversion of IPN into penicillin N appears to follow the well-known mechanism of PLP-dependent bacterial amino acid racemases. The reactions catalyzed by the A. lactamdurans and S. clavuligerus enzymes have been shown to proceed reversibly, as occurs with other bacterial racemases. The generally accepted mechanism [74] proceeds through the steps of Fig. 3. PLP bound at the active-site lysyl residue reacts with the substrate to form an external Schiff base through transaldimination. The subsequent a-hydrogen abstraction results in the formation of a resonance-stable deprotonated intermediate.When reprotonation occurs at the a-carbon of the substrate moiety on the opposite face of the planar intermediate, then an antipodal aldimine is formed. The e-amino group


105

of the enzyme’s lysine residue is substituted for the isomerized substrate amino acid through transaldimination, and the internal aldimine is regenerated. This mechanism has been shown to be involved in alanine racemization by Bacillus stearothermophilus alanine racemase [75, 76], an enzyme that occurs widely in bacteria and plays a central role in the biosynthesis of D-alanine, a component of peptidoglycan. On the basis of X-ray crystallography studies Shaw et al. [75] have proposed that Lys39 and Tyr265 of the B. subtilis alanine racemase serve as the basis to abstract the a-hydrogen from the alanyl-PLP aldimine and to protonate the carbanion intermediate. Both Lys39 and Tyr265 are also conserved in both DadB and Alr alanine racemases of Salmonella typhimurium [77, 78] suggesting that these two residues play a common role in this type of racemases. cefD-like genes encoding isopenicillin N epimerase-like enzymes have been reported in Bradyrhizobium japonicum (Accession number BAC46545), Caulobacter crescentus (NP419755), Pirellula sp. (CAD75434), Sulfolobus tokodaii (BAB67245), Synechocystis sp. (NP44105), Xanthomonas campestris (N636586) and Ralstonia solanacearum (NP522257). The last two Gram-negative bacteria show putative IPN epimerases with 69.2% identity among themselves but with only 15–20% identity with the IPN epimerases of the cephamycin producers. The IPN epimerases of A. lactamdurans, S. clavuligerus and L. lactamgenus contain a putative PLP consensus binding site SXHKXL (SGHKWL in A. lac-

A

B Fig. 6 Bacterial IPN epimerases. (A) Conserved amino acids (shaded) in the IPN epimerases of Amycolatopsis lactamdurans, Lysobacter lactamgenus and Streptomyces clavuligerus. (B). Conserved PLP-binding box in IPN epimerase-like proteins from different bacteria. The lysine proposed to be involved in formation of the Schiff base is indicated by an asterisk. The consensus PLP-binding motif is boxed in the lower line

106


Fig. 7 Alignments of amino acids of the IPN-CoA epimerase (CefD2) of A. chrysogenum and the unknown protein encoded by ORF1 of K. tethys. The identical amino acids are shaded

tamdurans) internal to a long conserved amino acid stretch (boxed in Fig. 6A) that is also present in PLP-requiring lysine, histidine and alanine decarboxylases. This consensus sequence is also conserved in several IPN epimerase-like enzymes (Fig. 6B) but since this motif is common to a variety of PLP-requiring enzymes, the exact function of those IPN epimerase-like enzymes in different bacteria remains uncertain.

7 A Cephalosporin Gene Cluster in Kallichroma tethys Includes an Isopenicillin N Epimerase Kallichroma tethys is a wood-inhabiting marine fungus that occurs exclusively in tropical waters [79]. Two genes pcbAB and pcbC encoding the initial two steps of the cephalosporin biosynthesis pathway were cloned from this fungus [7]. A third uncharacterized open reading frame, located downstream of pcbC in this fungus, corresponds to the cefD2 gene encoding isopenicillinyl-CoA epimerase [61] (Fig. 7). This ORF is located in the same position with respect to the pcbAB-pcbC cluster as in A. chrysogenum, suggesting that the cluster has been conserved in K. tethys, a fungus that appears to be phylogenetically related to A. chrysogenum. However, K. tethys does not produce active antibiotic [7] and it remains unknown if it contains the cefD1 gene (isopencillinyl-CoA synthetase) that is strictly required for cephalosporin biosynthesis.

8 Summary and Future Outlook The epimerization system converting isopenicillin N into penicillin N has finally been elucidated at the genetic level. It consists of three components encoded by two linked cefD1 and cefD2 genes that activate isopenicillin N as isopenicillinylCoA and converts this compound to penicillinyl-CoA. A putative third component thioesterase, named CefD3, that later releases penicillin N has not been lo-


107

cated so far. This novel mechanism provides another good example of the adaptation of enzyme systems used in eukaryotes for performing reactions required for isomerizations during the biosynthesis of secondary metabolites. This novel epimerization system may be considered as a system for late modification of amino acids in non-ribosomal peptides or their condensed derivatives (like isopenicillin N) similar to the D-alanine racemases involved in providing D-alanine for cyclosporin or HC-toxin biosynthesis. There is a functional similarity between the CoA-mediated activation and isomerization of isopenicillin N and the phosphopantetheine-linked conversion of L to D-amino acids by non-ribosomal peptide synthetases. In both cases the L-amino acid is activated by formation of a thioester of its carboxyl group with the SH group of phosphopantetheine. However, it is unclear if the complete isomerization reaction is performed by the integrated epimerization domains of non-ribosomal peptide synthetase, or whether (as yet unknown) additional enzymes similar to the CefD2 protein are involved in the peptide isomerization. It is likely that discrete epimerases may work in combination with the non-ribosomal peptide synthetases. Further biochemical work is, therefore, required to establish the exact mechanism of isomerization by the non-ribosomal peptide synthetases. Although the complexity of these large proteins makes molecular analysis complicated, the availability of the cloned genes will facilitate this work. The existence of a dedicated isopenicillinyl-CoA thioesterase (putative CefD3 protein) is unclear at this time. Cleavage of the isopenicillinyl-CoA may be catalyzed by a dedicated thioesterase or by non-specific thioesterases that occur in the cell. The first hypothesis is attractive because a number of discrete thioesterases or integrated thioesterase domains into the non-ribosomal peptide synthetases have been found. In particular, there is a thioesterase in the carbonyl-terminal region of the A. chrysogenum ACV synthetase that may be connected to the epimerase activity involved in the L- to D-valine conversion [80]. Further biochemical work is therefore required to elucidate this hypothesis.

References 1. Shiffman D, Mevarech M, Jensen SE, Cohen G, Aharonowitz Y (1988) Mol Gen Genet 214:562–569 2. Aharonowitz Y, Cohen G, Martín JF (1992) Annu Rev Microbiol 46:461–495 3. Peñalva MA, MoyaA, Dopazo J, Ramon D (1990) Proc R Soc London Ser B 241:164–169 4. Cohen G, Shiffman D, Mevarech M, Aharonowitz Y (1990) Trends Biotechnol 8:105–111 5. Eriquez LA, Pisano MA (1979) Antimicrob Agents Chemother 16:392–397 6. Elander RP (1983) In: Demain AL, Solomon NA (eds) Antibiotics containing the b-lactam structure. Springer, Berlin Heidelberg New York, pp 97–146 7. Kim CF, Lee SK, Price J, Jack RW, Turner G, Kong RY (2003) Appl Environ Microbiol 69:1308–1314 8. Demain AL (1983) In: Demain AL, Solomon NA (eds) Antibiotics containing the b-lactam structure. Springer, Berlin Heidelberg New York, pp 189–228 9. Nüesch J, Heim J, Treichler H-J (1987) Annu Rev Microbiol 41:51–75

108


10. Martín JF, Liras P (1989) In: Fiechter A (ed) Advances in biochemical engineering/ biotechnology, vol 39. Springer, Berlin Heidelberg New York, pp 153–187 11. Martín JF, Gutiérrez S, Demain AL (1997) In: Anke T (ed) Fungal biotechnology. Chapman and Hall, Weinheim, Germany, pp 91–127 12. Martín JF (2000b) J Bacteriol 182:2355–2362 13. Martín JF (1998) Appl Microbiol Biotechnol 50:1–15 14. Liras P (1999) Antonie van Leeuwenhoek 75:109–124 15. Baldwin JE, Bird JW, Field RA, O’Calaghan NM, Schofield CJ, Willis AC (1991) J Antibiotics 44:241–248 16. Aharonowitz Y, Bergmeyer J, Cantoral JM, Cohen G, Demain AL, Fink U, Kinghorn J, Kleinkauf H, Maccabe A, Palissa H, Pfeifer E, Schwecke T, van Liempt H, von Döhren H, Wolfe S, Zhang J (1993) Biotechnology 11:807–810 17. Martín JF (2000) J Antibiot 53:1008–1021 18. Gutiérrez S, Díez B, ÁlvarezE, Martín JF (1991) J Bacteriol 173:2354–2365 19. Samson SM, Belagaje R, Blankenship DT, Chapman JL, Perry D, Skatrud PL, van Frank RM, Abraham EP, Baldwin JE, Queener SW, Ingolia TD (1985) Nature 318:191–194 20. Gutiérrez S, Fierro F, Casqueiro J, Martín JF (1999) Antonie van Leeuwenhoek 75:81– 94 21. Ullán RV, Liu G, Casqueiro J, Gutiérrez S, Bañuelos O, Martín JF (2002) Mol Genetics Genomics 267:673–683 22. Samson SM, Chapman JL, Belagaje R, Queener SW, Ingolia TD (1987) Proc Natl Acad Sci USA 84:5705–5709 23. Gutiérrez S, Velasco J, Fernández FJ, Martín JF (1992) J Bacteriol 174:3056–3064 24. Matsuda A, Sugiura H, Matsuyama K, Matsumoto H, IchikawaS, Komatsu KI (1992) Biochem Biophys Res Commun 182:995–1001 25. Velasco J, Gutiérrez S, Campoy S, Martín JF (1999) Biochem J 337:379–385 26. Konomi T, Herchen S, Baldwin JE, Yoshida M, Hunt N, Demain AL (1979) Biochem J 184:427–430 27. Sawada Y, Baldwin JE, Singh PD, Solomon NA, Demain AL (1980) Antimicrob Agents Chemother 27:380–387 28. Lübbe C, Wolfe S, Demain AL (1986) Appl Microbiol Biotechnol 23:367–368 29. Baldwin JE, Keeping JW, Singh PD, Vallejo CA (1981) Biochem J 194:649–651 30. Jensen SE, Westlake DWS, Wolfe S (1983) Can J Microbiol 29:1526–1531 31. Usui S, Yu CA (1989) Biochim Biophys Acta 999:78–85 32. Láiz L, Liras P, Castro JM, Martín JF (1990) J Gen Microbiol 136:663–671 33. Kovacevick S, Tobin MB, Miller JR (1990) J Bacteriol 172:3952–3958 34. Coque JJR, Liras P, Martín JF (1993) Mol Gen Genet 236:453–458 35. Cheng Y-Q, Walton JD (2000) J Biol Chem 275:4906–4911 36. Hoffmann K, Schneider-Scherzer E, Kleinkauf H, Zocher R (1994) J Biol Chem 269: 12710–12714 37. Kleinkauf H, von Döhren H (1996) Eur J Biochem 192:1–15 38. Martín JF (1996) In: Meyers RA (ed) Encyclopedia of molecular biology.VCH, Germany, pp 207–217 39. von Döhren H, Keller U, Vater J, Zocher R (1997) Chem Rev 97:2675–2705 40. Martín JF, Gutiérrez S, Aparicio JF (2000) In: Lederberg J (ed) Encyclopedia of microbiology, vol 4, 2nd edn. Academic, San Diego, CA, pp 213–237 41. Mason KT, Shaw GJ, Katz E (1977) Arch Biochem Biophys 180:509–513 42. Stindl A, Keller U (1994) Biochemistry 33:9358–9364 43. Fuma S, Fujishima Y, Corbell N, D’Souza C, Nakano MM, Súber P,Yamane K (1993) Nucl Acids Res 125:93–97 44. Stachelhaus T, Marahiel MA (1995) FEMS Microbiol Lett 21:93–97


109

45. Walsh CT, Zydowsky LD, McKeon FD (1991) J Biol Chem 267:13115–13118 46. Zocher R, Nihira T, Paul E, Madry N, Peeters H, Kleinkauf H, Keller U (1986) Biochemistry 25:550–553 47. Adams E, Norton IL (1964) J Biol Chem 293:1525–1535 48. Rudnick G, Abeles RH (1975) Biochemistry 14:4515–4522 49. Ramaswamy SG (1984) J Biol Chem 259:249–254 50. Hiltunen JK, Qin Y (2000) Biochim Biophys Acta 1484:117–128 51. van Veldhoven PP, Croes K, Asselberghs S, Herdewijn P, Mannaerts GP (1996) FEBS Lett 388:80–84 52. Schmitz W, Conzelmann E (1997) Eur J Biochem 244:434–440 53. Schmitz W, Helander HM, Hiltunen JK, Conzelmann E (1997) Biochem J 326:883–889 54. Chen LS, Jin SJ, Tserng KY (1994) Biochemistry 33:10527–10534 55. Gurvitz A, Mursula AM, Firzinger A, Hamilton B, Kilpelainen SH, Hartig A, Ruis H, Hiltunen JK, Rottensteiner H (1998) J Biol Chem 273:31366–31374 56. Geisbretch BV, Gould SJ (1999) J Biol Chem 273:33184–33191 57. Martín JF, Liras P (1989b) Ann Rev Microbiol 43:173-206 58. Fierro F, Barredo JL, Díez B, Gutiérrez S, Fernández FJ, Martín JF (1995) Proc Natl Acad Sci USA 92:6200–6204 59. Keller NP, Hohn TM (1996) Fungal Genet Biol 21:17–29 60. Laich F, Fierro F, Martín JF (2002) Appl Environ Microbiol 68:1211–1219 61. Ullán RV, Casqueiro J, Bañuelos O, Fernández FJ, Gutiérrez S, Martín JF (2002) J Biol Chem 277:46215–46225 62. Velasco J, Gutiérrez S, Fernández FJ, Marcos AT, Arenos C, Martín JF (1994) J Bacteriol 176:985–891 63. Velasco J, Gutiérrez S, Casqueiro J, Fierro F, Campoy S, Martín JF (2001) Appl Microbiol Biotechnol 57:300–356 64. Uchiyama A, Aoyama T, Kamijo K, Uchida Y, Kondo N, Orii T, Hashimoto T (1996) J Biol Chem 271:30360–30365 65. Turgay K, Krause M, Marahiel MA (1992) Mol Microbiol 6:529–546 66. Mansour SL, Thomas KR, Capecchi MR (1988) Nature 336:348–352 67. Liu G, Casqueiro J, Bañuelos O, Cardoza RE, Gutiérrez S, Martín JF (2001) J Bacteriol 183:1765–1772 68. Martín JF, Ingolia TD, Queener SW (1990) In: Leong SA, Berka R (eds) Molecular industrial mycology. Marcel Dekker, New York, pp 149–195 69. Schmitz W, Fingerhut R, Conzelmann E (1994) Eur J Biochem 222:313–323 70. Schmitz W, Albers C, Fingerhut R, Conzelmann E (1995) Eur J Biochem 231:815–822 71. Caldwell J, Hutt AJ, Fournel-Gigleux S (1988) Biochem Pharmacol 37:105–114 72. Knihinicki RD, Day RO, Williams KM (1991) Biochem Pharmacol 42:905–911 73. Shieh WR, Chen CS (1993) J Biol Chem 268:3487–3493 74. Watababe A, Kurokawa Y, Yoshimura T, Kurihara T, Soda K, Esaki N (1999) J Biol Chem 274:4189–4194 75. Shaw JP, Petsko GP, Ringe D (1997) Biochemistry 36:1329–1342 76. Stamper CG, Morollo AA, Ringe D (1998) Biochemistry 37:10438–10443 77. Wasserman SA, Daub E, Grisafi P, Botstein D, Walsh CT (1984) Biochemistry 23:5182–5187 78. Galakatos NG, Daub E, Grisafi P, Botstein D,Walsh CT (1986) Biochemistry 25:3255–3260 79. Kohlmeyer J, Vokmann-Kohlmeyer B (1993) Mycol Res 97:753–761 80. Kallow W, Kennedy J, Arezi B, Turner G, von Döhren H (2000) J Mol Biol 297:395–408 Received: June 2004

Adv Biochem Engin/Biotechnol (2004) 88: 111–135 DOI 10.1007/b99259 © Springer-Verlag Berlin Heidelberg 2004

Compartmentalization and Transport in b -Lactam Antibiotics Biosynthesis M. E. Evers 1 · H. Trip 1 · M. A. van den Berg 2 · R. A. L. Bovenberg 2 · A. J. M. Driessen 1 1

2

University of Groningen, Department of Molecular Microbiology & Groningen Biomolecular Sciences and Biotechnology Institute, Kerklaan 30, 9751 NN Haren, The Netherlands [email protected] · [email protected] DSM Anti-infectives, A. Fleminglaan 1, 2611 XT Delft, The Netherlands [email protected] · [email protected]

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

1

Introduction

2 2.1 2.2 2.3 2.4 2.5

Localization of Biosynthesis Enzymes of the Penicillin Biosynthesis Pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine Synthetase (ACVS) . . . . . . . . Isopenicillin N Synthase (IPNS) . . . . . . . . . . . . . . . . . . . . . . . . Acyl-Coenzyme A: Isopenicillin N Acyltransferase (IAT) . . . . . . . . . . . Side Chain Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Localization of Enzymes of Cephalosporin Biosynthesis and Other b-lactams

3

Compartmentalization of Penicillin Biosynthesis

4 4.1 4.2 4.3 4.4

Synthesis of b -lactam Precursors L-a-Aminoadipate Synthesis . . . Cyclization of L-a-Aminoadipate Cysteine Synthesis . . . . . . . . Valine Synthesis . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

118 118 120 120 122

5 5.1 5.2 5.3 5.4 5.5

Uptake of b -Lactam Precursors from the Growth Medium Uptake of Amino Acids . . . . . . . . . . . . . . . . . . . Uptake of Sulfate and Phosphate . . . . . . . . . . . . . . Uptake of Nitrogen-Containing Compounds . . . . . . . Uptake of Side Chain Precursors . . . . . . . . . . . . . . Uptake of Sugars . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

123 123 125 126 126 127

6

Transport Across the Microbody Membrane

7

Excretion of b -lactams into the Medium

8

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

References

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

113 114 115 115 116 116

. . . . . . . . . . . . . . 117 . . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . . . . . . 128

. . . . . . . . . . . . . . . . . . . 129

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

112

M. E. Evers et al.

Abstract Classical strain improvement of b-lactam producing organisms by random mutagenesis has been a powerful tool during the last century. Current insights in the biochemistry and genetics of b-lactam production, in particular in the filamentous fungus Penicillium chrysogenum, however, make a more directed and rational approach of metabolic pathway engineering possible. Besides the need for efficient genetic methods, a thorough understanding is needed of the metabolic fluxes in primary, intermediary and secondary metabolism. Controlling metabolic fluxes can be achieved by adjusting enzyme activities and metabolite levels in such a way that the main flow is directed towards the desired product. In addition, compartmentalization of specific parts of the b-lactam biosynthesis pathways provides a way to control this pathway by clustering enzymes with their substrates inside specific membrane bound structures sequestered from the cytosol. This compartmentalization also requires specific membrane transport steps of which the details are currently uncovered. Keywords b-Lactam · Biochemical engineering · Compartmentalization · Penicillium chrysogenum · Transporter proteins

List of Abbreviations AAP Amino acid permease ABC ATP-binding-cassette ACS Acetylcoenzyme A synthetase ACVS d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine synthetase ad-7-ACA Adipoyl-7-aminocephalosporanic acid ad-7-ADC A3-Aminodeacetoxycephalosporanic acid A. niger Aspergillus niger APS Adenosine-5-phosphosulfate A. nidulans Aspergillus nidulans C. acremonium Cephalosporium acremonium CoA Coenzyme A DAC Deacetylcephalosporin C DAOC Deacetoxycephalosporin C Gap General amino acid permease GFP Green fluorescent protein GST Glutathione S-transferase IAT Acyl-coenzyme A:isopenicillin N acyltransperase IPNS Isopenicillin N synthase MFS Major Facilitator Superfamily N. crassa Neurospora crassa N. lactamdurans Nocardia lactamdurans NBD Nucleotide binding domain OPC 6-Oxopiperidine-2-carboxylic acid PA Phenylacetic acid PAP S3-Phospho-adenosine-5-phosphosulfate P. chrysogenum Penicillium chrysogenum PCL Phenylacetyl-coenzyme A ligase pmf Proton motive force PMP Peroxisomal membrane protein POA Phenoxyacetic acid PTS Peroxisomal targeting signal

Compartmentalization and Transport in b-Lactam Antibiotics Biosynthesis S. cerevisiae St. clavuligerus Sut

113

Saccharomyces cerevisiae Streptomyces clavuligerus Sulfate transporter

1 Introduction Industrial penicillin and cephalosporin fermentation is performed using filamentous fungi. However, sequence analysis of one of the biosynthesis genes encoding isopenicillin N synthase (pcbC) of various organisms involved in the biosynthesis of these compounds, suggests that the origin of these genes stems from prokaryotic organisms. In bacteria apparently the production of b-lactams has evolved as a means of improving their ability to compete with other prokaryotes.Via horizontal gene transfer, biosynthetic genes may have been acquired by filamentous fungi some 370 million years ago [1–4]. The advantage for these fungi of possessing these genes is thought to be of ecological significance. Antibiotic production provides fungi with the possibility to protect released enzymes and released nutrients against bacteria competing for the same substrates. Other explanations are found in detoxification mechanisms, for instance, to prevent the accumulation of acids such as phenylacetic acid in the cell [5].After the discovery of the application of penicillin as an antimicrobial agent in humans in the early 1940s, classical strain improvement has been applied to obtain higher production yields. High production and secretion of the b-lactams, however, drains intracellular pools of primary metabolites. In addition, specific metabolic engineering of industrial strains has been applied; however, this requires extensive knowledge of control of the metabolic fluxes in order to obtain predictive models and the desired results. The importance of compartmentalization and transport processes in industrial penicillin biosynthesis has become clear and the different aspects of these topics are being studied in several laboratories. The aim of this chapter is to describe recent developments that are important towards the compartmentalization and transport in b-lactam antibiotics by filamentous fungi. The localization of biosynthesis enzymes and the compartmentalization of biosynthesis, precursors, intermediates and products will be discussed in relationship to their consequences for intra- and extracellular transport.

2 Localization of Biosynthesis Enzymes of the Penicillin Biosynthesis Pathway Figure 1 in brief depicts the major enzymatic steps involved in b-lactam biosynthesis. This section discusses the cellular localization of the key enzymes.

114

M. E. Evers et al.

Fig. 1 Localization of the penicillin biosynthesis; d-(L-a-aminoadipyl)-L-cysteinyl-D-valine synthetase (ACVS) and isopenicillin N synthase (IPNS) are present in the cytosol (C) whereas acyl-coenzyme A:isopenicillin N acyltransferase (IAT) and possibly also phenylacetyl-coenzyme A ligase (PCL) are localized in peroxisomes (P). In the mitochondria (M), part of the synthesis of precursor amino acids takes place

2.1 a -Aminoadipyl)-L-cysteinyl-D-valine Synthetase (ACVS) d -(L-a The first step in the biosynthesis of penicillins and cephalosporins is the condensation of three precursor amino acids, namely L-a-aminoadipate, L-cysteine and L-valine into the tripeptide d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (LLD-ACV). This step is catalysed by a multi-enzyme complex of 424 kDa with non-ribosomal peptide synthetase activity termed ACV synthetase (ACVS) [6–10]. ACVS is encoded by the acvA gene that is part of a cluster which includes the other two key enzymes of the penicillin biosynthesis pathway. The localization of ACVS has been a matter of debate for some time. Initially, it was described as a membrane associated protein and found to co-sediment with vesicles of either Golgi or vacuolar origin [11–13]. However, the amino acid sequence of P. chrysogenum ACV synthetase contains no recognizable targeting information for the endoplasmic reticulum or the vacuole, and although the protein is hydrophobic of nature, it does not harbour any trans-membrane regions. Localization studies by traditional fractionation experiments were ob-

Compartmentalization and Transport in b-Lactam Antibiotics Biosynthesis

115

scured by the fact that ACVS is a highly unstable enzyme and very sensitive to proteolytic degradation. For this purpose improved protocols of cell lysis were designed and used in combination with an immuno-gold electron microscopical analysis to determine the subcellular location of this protein. On the basis of these studies, ACVS turned out to be a cytosolic enzyme [11–14]. Likewise, an ACVS fusion with green fluorescent protein in Aspergillus nidulans also localizes to the cytosol [15]. The cytosolic localization is more in pair with the pH optimum of this enzyme, as the acidic vacuole would not support activity. Moreover, the vacuole is highly proteolytic which seems contradictory with the protease sensitivity of the multidomain ACVS and the release of a product tripeptide. ACVS consists of three major modules, one for each amino acid. These modules are divided into domains that are specialized for partial reactions of the total condensation reaction, hereby combining adenylation activity, peptide-bond formation, epimerization and product release by thioesterase activity, in one multi-enzyme [7]. The localization of ACVS in the cytosol bears consequences for the recruitment of the three precursor amino acids. In general, acquisition of L-a-aminoadipate, L-cysteine and L-valine can either proceed through de novo synthesis or uptake from the growth medium; see below. 2.2 Isopenicillin N Synthase (IPNS) The second step in b-lactam synthesis is the oxidative cyclisation of LLD-ACV into isopenicillin N (IPN). In this step, the bicyclic penam nucleus, consisting of the b-lactam and thiazolidine rings is generated [3, 7]. This step is mediated by IPN synthase (IPNS) a protein of 38 kDa that is encoded by the ipnA gene that is part of the penicillin biosynthesis gene cluster. From X-ray diffraction experiments using the substrate analogue d-(L-a-aminoadipoyl)-L-cysteinyl-L-Smethyl-cysteine in the crystal it was concluded that closure of the b-lactam ring precedes the closure of the five-membered thiazolidine ring [3, 16]. Based upon the results of fractionation experiments it became evident that IPNS behaves like a soluble, cytosolic enzyme [17]. This means that LLD-ACV produced by ACVS can directly be used as the substrate for IPNS. The question if these two enzymes are organized in a metabolon or large complex is not known. 2.3 Acyl-Coenzyme A: Isopenicillin N Acyltransferase (IAT) The third and final step in b-lactam synthesis is the exchange of the L-aaminoadipate for phenylacetyl- or phenoxyacetyl group by acyl-coenzyme A:isopenicillin N acyltransferase (IAT) resulting in the formation of respectively penicillin G and penicillin V [18–21]. IAT is a hetero-dimeric enzyme consisting of a 11 kDa a-subunit and a 28 kDa b-subunit. It is synthesized as a 40 kDa pre-protein from the aat gene and undergoes autocatalytic processing to form the heterodimer [22, 23]. Both subunits possess a C-terminal PTS1 signal that

116

M. E. Evers et al.

targets this enzyme to a microbody or peroxisome. Fractionation studies as well as immuno-gold labelling experiments indeed localized IAT to microbodies [17]. Consequently the substrate IPN has to enter the microbody before it can be converted. It is not known yet whether this occurs by diffusion over the membrane, or by facilitated or active transport. 2.4 Side Chain Activation Before the side chain can be used in the substitution reaction catalysed by IAT mentioned above, PA and POA, the side chain precursors, have to be activated to their CoA thioesters. Theoretically this activation can be carried out by an enzyme displaying either acetyl-coenzyme A synthetase (ACS) activity, phenylacetyl-coenzyme A ligase (PCL) activity, or alternatively via a glutathione-dependent pathway involving glutathione S-transferase (GST) activity. This phenomenon has not well been studied, but current view considers the last option unlikely. A gene encoding a cytosolic ACS of P. chrysogenum has been identified and isolated. Disruption of this gene did not result in a decrease in penicillin production, meaning that ACS cannot be solely responsible for activation of precursors [17, 24]. In another study a PCL of P. chrysogenum containing a C-terminal peroxisomal targeting signal (PTS1; SKI) was identified [25]. This suggests a peroxisomal location of the activating enzyme which would seem advantageous as it is then in the same compartment as IAT and the peroxisomal concentration of activated precursors would be higher than in the cytosol. In addition PA and POA are more likely to easily diffuse across the peroxisomal membrane than their activated counterparts thereby providing a means of retention. However, overproduction of a presumably cytosolic located heterologous PCL from Pseudomonas putida U increased the penicillin production by 100% whereas overproduction of the homologous peroxisomal PCL of P. chrysogenum did not affect penicillin production [25, 26]. Although these results give an ambiguous view on this step, a bias towards a role of the peroxisomal PCL is provided by additional observations that will be described below. 2.5 Localization of Enzymes of Cephalosporin Biosynthesis and Other b -lactams Cephalosporins and cephamycins are produced by the filamentous fungus Acremonium chrysogenum (syn. Cephalosporium acremonium) (cephalosporin C), and by Gram-positive actinomycetes such as Nocardia lactamdurans and Streptomyces clavuligerus (cephamycin C) [27, 28]. Cephalosporin and penicillin biosynthesis have the first two steps in common, i.e. the formation of ACV by ACV synthetase and the subsequent cyclization to isopenicillin N by IPN synthase. Here the penicillin and cephalosporin pathways diverge. Cephalosporin biosynthesis proceeds with the epimerisation of IPN into penicillin N by epimerase activity of the cefD1 and cefD2 gene products in C. acre-


117

monium [29] and the completely different cefD gene product in N. lactamdurans and St. clavuligerus [27, 30]. After epimerisation, penicillin N undergoes expansion of the thiazolidine ring to a dihydrothiazine ring by deacetoxycephalosporin C synthetase, yielding deacetoxycephalosporin C (DAOC). DAOC is hydroxylated by DAC synthase to give deacetylcephalosporin C (DAC). In the cephamycin producers N. lactamdurans and St. clavuligerus, DAC is carbamoylated and methoxylated to form cephamycin C [31]. In C. acremonium, DAC is acetylated by the cefG-gene encoded enzyme DAC acetyltransferase [32, 33], yielding cephalosporin C. IPN epimerase, DAOC synthetase, DAC synthase and DAC acetyltransferase behave as soluble cytosolic proteins with pH optima above 7.0, reviewed in [34]. Currently, these enzymes are believed to be localized in the cytosol, but no direct studies have addressed the compartmentalization issue. Expression of the expandase gene of St. clavuligerus (cefE) or the C. acremonium expandase-hydroxylase gene and the acetyl transferase gene in P. chrysogenum and feeding adipic acid has led to efficient production of adipoyl-7-aminodeacetoxycephalosporanic acid (ad-7-ADCA) and adipoyl-7aminocephalosporanic acid (ad-7-ACA) respectively. Removal of the adipyl side chain gives 7-ADCA and 7-ACA, respectively, which are important intermediates in the production of semi-synthetic cephalosporins [35, 36].

3 Compartmentalization of Penicillin Biosynthesis As can be concluded from the above the first two steps of penicillin biosynthesis, the condensation into the tripeptide and the conversion into isopenicillin N, take place in the cytosol of filamentous fungi. The final step, side chain exchange and most probably the side chain precursor activation take place in the microbody. Microbodies (also termed peroxisomes, glyoxysomes, glycosomes depending on the organism and function) are indispensable organelles that can be found in practically all eukaryotic cells.Although their morphology is relatively simple (a proteinaceous matrix surrounded by a single membrane) their physiological properties are remarkably complex. The organelles are involved in pathways of primary, intermediary and secondary metabolism. They may be regarded as organelles in which specific metabolic conversions take place mostly by non-membrane bound enzymes. Various peroxisomal metabolic pathways function in the cytosol of peroxisome-deficient mutants, although in some cases with lower final cell yield compared to wild-type cells [26, 37]. For other peroxisomal pathways, the membrane needs to function as an intact boundary, otherwise metabolic pathways may be severely affected even though all the enzymes of the pathway are synthesized and active in the cytosol [38]. A general major advantage of the presence of a peroxisomal permeability barrier is that it permits the cells to precisely adjust the levels of different intermediates of primary metabolism required for specific metabolic pathways (metabolic flux control by a physical barrier). It was in 1991 that the importance

118

M. E. Evers et al.

of the microbody with respect to penicillin biosynthesis (secondary metabolism) became evident when IAT was shown to be located in this organelle [17]. When the putative targeting signal was removed the enzyme was not directed to the microbody but instead localized in the vacuole and surrounding cytosol. Under these conditions, production of penicillin was halted although the enzyme was expressed in vivo and active in vitro [39]. This might be explained by the possibility that another essential enzymatic step, the precursor activation by PCL might occur only inside microbodies and that these activated precursors are now sequestered from IAT inside the microbody. The other explanation, namely that IAT is not able to perform the catalytic reaction in the cytosol seems less likely, because in a mutant of A. nidulans lacking functional peroxisomes, penicillin production still occurred with the peroxisomal enzymes mislocalized to the cytosol [40].Although this suggests that peroxisomes are not essential for penicillin production per se, a positive correlation between penicillin yield and peroxisome numbers has been implicated [17]. The exact reason for this correlation is not known, but this may relate to an increase in the amount of enzymes of the biosynthesis pathway. For detailed information on import of peroxisomal proteins and biogenesis of peroxisomes see reviews by Purdue et al. [41] and van der Klei et al. [42] and references therein.

4 Synthesis of b -lactam Precursors 4.1 a -Aminoadipate Synthesis L-a L-a-Aminoadipate is an intermediate of the L-lysine biosynthesis pathway. The

intracellular level of L-a-aminoadipate can be a limiting factor in the overall penicillin synthesis rate, as shown by the observation that addition of L-aaminoadipate to the growth medium enhances the b-lactam production [43]. Therefore, the lysine biosynthsisc pathway is extremely important for, and interconnected with, the b-lactam biosynthesis pathway. Recent insights in the lysine biosynthesis route in yeast may alter the classical view on this pathway significantly. L-Lysine biosynthesis starts with the condensation of acetyl-CoA and a-ketoglutarate into homocitrate by homocitrate synthase. This enzymatic reaction was until recently believed to take place in the mitochondria of P. chrysogenum because of insights in this route in higher eukaryotes. However, a localization study using a GFP-fusion of homocitrate synthase indicates that in P. chrysogenum this protein is located in the cytosol, although it could not be excluded that minor amounts might be present in the nucleus and mitochondria [44]. Using an in silico approach by performing a Saccharomyces cerevisiae database context sensitive motif search to identify new peroxisomal proteins, it was established that both the LYS1 and LYS4 proteins that encode a homoaconitase and a saccharopine dehydrogenase, respectively, contain a C-terminal


119

Fig. 2 Compartmentalization of lysine biosynthesis in relation to penicillin biosynthesis. aAminoadipate is a branch point intermediate at which the lysine and penicillin biosynthesis routes converge. In the last step of the penicillin biosynthesis, the a-aminoadipate moiety of isopenicillin N (IPN) is exchanged for phenylacetic acid and becomes available again for penicillin or lysine biosynthesis. Part of a-aminoadipate is lost by the cyclization into 6-oxopiperidine-2-carboxylic acid (OPC). OPC is excreted into the medium by an unknown mechanism. Main routes are depicted in bold arrows, hypothetical routes are in light grey. PM, plasma membrane

peroxisomal targeting signal (PTS1). GFP-fusions of these proteins localized to peroxisomes. This suggests that these proteins are peroxisomal localized [44, 45] (Fig. 2). Furthermore, examination of micro-array experiments to determine the role of peroxisomes under physiological conditions revealed that in a peroxisome-deficient mutant five genes of the lysine biosynthesis pathway are highly up-regulated, among them LYS1 and LYS4. The other three genes that are up-regulated are LYS20 (homocitrate synthase), LYS12 (homoisocitrate dehydrogenase) and LYS9 (another saccharopine dehydrogenase). LYS12 contains a putative PTS1 whereas LYS9 and LYS20 contain PTS2-like sequences. The observed expression pattern of the genes of a peroxisome-deficient mutant grown on rich medium surprisingly resembled a lysine starvation response even when sufficient lysine was present in the medium. The authors explain their findings by mislocalization of a-aminoadipate semialdehyde to the cytosol. When

120

M. E. Evers et al.

a-aminoadipate semialdehyde is not contained inside the peroxisome the level in the cytosol will increase and stimulate the Lys14p transcriptional activator [46]. In contrast, no peroxisomal PTS could be detected in Lys5p or in the amino acid aminotransferases that are thought to be a part of the lysine biosynthetic pathway. Therefore, it is believed that it is not very likely that all the lysine biosynthesis enzymes have an exclusively peroxisomal location, and that part of the pathway may be cytosolic [46]. The question arises if these proteins are also localized in the peroxisomes of P. chrysogenum, which is important for the question where the L-a-aminoadipate is formed? Consequently, a peroxisomal location poses some important questions about the mechanism of release of L-a-aminoadipate, as this charged amino acid is unlikely to pass the membrane passively. 4.2 a -Aminoadipate Cyclization of L-a During b-lactam biosynthesis, part of the a-aminoadipate is lost by the irreversible formation of 6-oxopiperidine-2-carboxylic acid (OPC), the cyclized d-lactam of a-aminoadipate. This compound is excreted into the medium. The extent of OPC formation ranges from 6 to 60% relative to the formation of penicillin (on a molar basis), depending on strain and cultivation conditions. The route leading to OPC is not understood [47] nor is it clear how this compound is excreted. 4.3 Cysteine Synthesis The synthesis of cysteine in P. chrysogenum is dependent on the active uptake of sulfate from the exterior of the cell. The sulfate assimilation pathway catalyses the reduction of sulfate via sulfite to sulfide and subsequently sulfide is converted into cysteine (Fig. 3). The reduction of sulfate into sulfite is catalysed by three enzymes: ATP sulfurase converts inorganic sulfate into adenosine-5phosphosulfate (APS) which is then activated into 3-phospho-adenosine-5phosphosulfate (PAPS) by APS-kinase and reduced to sulfite by PAPS reductase [48–50]. Sulfite is reduced to sulfide by sulfite reductase [51]. The location of enzymes involved in the reduction of sulfate has not been described. Sulfide is the basis for biosynthesis of L-cysteine, which occurs via two different pathways in b-lactam producing fungi: the transsulfuration and the sulfhydrylation pathway. L-Cysteine, synthesized via the transsulfuration pathway, is formed by cleavage of L-cystathionine derived from the intermediate Lhomocysteine, which is formed from L-methionine or O-acetyl-L-homoserine [52]. Otherwise, direct acetylation of L-serine yields O-acetyl-L-serine that, in the presence of sulphide, is converted to L-cysteine by action of the enzyme Oacetyl-L-serine sulfhydrylase (OASS). Theoretically, the yield of penicillin on glucose would be substantially higher when L-cysteine is synthesized exclusively via the direct sulfhydrylation pathway [53, 54]. In A. nidulans and C. acre-


121

Fig. 3 Sulfate uptake and metabolism in P. chrysogenum. PM, plasma membrane

monium both pathways are described, although A. nidulans prefers the direct sulfhydrylation, while C. acremonium utilizes the transsulfuration pathway [55, 56]. For the industrial penicillin producer P. chrysogenum only the presence of the transsulfuration pathway was demonstrated as mutants, disturbed in this pathway, were unable to grow on inorganic sulfate [57], which is the main source of sulfate during industrial fermentations. Recently, Østergaard et al. [58] reported the purification of OASS from P. chrysogenum. This enzyme is localized in the mitochondria (van den Berg MA, Westerlaken I, Hillekens R and Bovenberg RAL, unpublished results) and the analogous enzyme of the transsulfuration pathway, O-acetyl-L-homoserine sulfhydrylase (OAHS), is located in the cytosol. Moreover, a cloned cDNA encoding OASS was fused to eGFP and shown to encode active OASS enzyme located in the mitochondria (van den Berg MA., Westerlaken I., Hillekens R. and Bovenberg RAL, unpublished results). Isolated UV mutants that were unable to grow on inorganic sulfate unless OAS, or a more reduced sulfate source like cysteine or methionine,

122

M. E. Evers et al.

was added to the medium are likely to be disturbed in serine transacetylase. These findings suggest a distinctive role of the direct sulfhydrylation pathway for growth. This in contrast to the cytosolic transsulfuration pathway, which seems to be used for penicillin production in P. chrysogenum, as an increase in detectable OAHS activity correlates with the onset of penicillin G biosynthesis in shake flask experiments (van den Berg MA, Westerlaken I, Hillekens R and Bovenberg RAL, unpublished results). 4.4 Valine Synthesis Valine synthesis starts with the condensation of pyruvate with hydroxyethyl thiamine pyrophosphate into a-acetolactate by Acetohydroxy acid synthase. The conversion of a-acetolactate into L-valine is catalysed by three enzymes. Acetohydroxy acid isomeroreductase converts a-acetolactate into dihydroxyisovalerate. Dihydroxy acid dehydrase converts dihydroxyisovalerate into a-ketoisovalerate which in turn is converted to L-valine by the branched chain amino acid Glutamate transaminase. All four enzymes are thought to be located inside the mitochondrial matrix [59]. Consequently valine has to be translocated to the cytosol to become available for ACVS. The mechanism by which amino acids that are synthesized inside the mitochondrial matrix are transported into the cytosol has not been investigated. The inner membranes of mitochondria contain a family of transporter proteins (the mitochondrial carrier family) of related sequence and structure that are involved in the uptake and excretion of various metabolites, nucleotides and cofactors [60, 61] (and references therein). A number of these transporters have been biochemically characterized by overexpression and functional reconstitution into liposomes, most notably the ATP/ADP translocase and the phosphate transporters. The mitochondrial transporters operate by various mechanisms, which include uniport, symport, and antiport mechanisms [62] (and references therein). As to amino acid transport, a few transporters have been biochemically characterized: two human aspartate-glutamate transporters, citrin and aralar1, mediate the antiport of aspartate for glutamate [63]; a human glutamate transporter, GC, catalyses the uptake of glutamate in symport with protons [60], and an ornithine transporter from rat liver mitochondria that catalyses the uptake of ornithine in an antiport reaction for citrulline or protons [64]. In S. cerevisiae, an ornithine transporter,ARG11, mediates the exchange of ornithine for protons, but transports also arginine and lysine with less affinity [65]. The genomic sequence of S. cerevisiae suggests presence of 35 putative members of the mitochondrial transporter family and an increasing number can now be associated with a particular transport reaction [60, 66] (and references therein). It is unknown how valine, cysteine and aaminoadipate are transported across the mitochondrial inner membrane. Interestingly, one member of the mitochondrial carrier family in S. cerevisiae turned out to be an adenine nucleotide transporter in peroxisomes [67].


123

5 Uptake of b -Lactam Precursors from the Growth Medium 5.1 Uptake of Amino Acids Since penicillin and cephalosporin are synthesized from three amino acid precursors, amino acids might play an important role in the regulation of b-lactam synthesis. This may be either directly, e.g. precursor availability for ACV synthesis, or indirectly, e.g. by affecting expression of penicillin synthesis genes. All amino acids are synthesized by filamentous fungi, but can also be taken up from the extracellular medium and be used as nitrogen and carbon source. In industrial P. chrysogenum fermentations for penicillin production, corn steep liquor is often used as nitrogen source. This supplement is rich in amino acids, which are consumed in the exponential phase of a fed-batch cultivation, rather than ammonia.When the amino acids are depleted the cells start to utilize ammonia as the nitrogen source [68]. A number of effects of the addition of extracellular amino acids on b-lactam synthesis have been reported. The addition of the three amino acid precursors aaminoadipate, cysteine and valine to P. chrysogenum in nitrogen-less medium leads to efficient incorporation into ACV. Only a-aminoadipate increases the rate of ACV synthesis and the overall penicillin synthesis rate [43, 68]. Based on such studies, it was suggested that the intracellular a-aminoadipate concentration may be limiting for the penicillin biosynthesis [69]. The addition of lysine to P. chrysogenum and A. nidulans cultures leads to a reduction of penicillin biosynthesis [70]. L-Lysine inhibits homocitrate synthase, the first enzyme in the lysine biosynthesis pathway, thereby blocking the production of a-aminoadipate, which is the branch-point metabolite between the lysine and penicillin biosynthetic pathways [71]. Lara et al. reported a stimulating effect on penicillin synthesis by addition of L-glutamate in minimal media. Interestingly, this effect was also observed with non-metabolisable analogues of L-glutamate [72]. The effect of externally added amino acids on the expression of penicillin biosynthesis genes acvA and ipnA was investigated in A. nidulans. The negative effect of histidine and valine is due to a reduced activation of the transcriptional factor PACC under acidic conditions. The presence of these amino acids leads to a decreased ambient pH during cultivation of the fungus. The negative effect of lysine and methionine, that also cause an acidification of the medium, does not involve PACC. The mechanism by which these amino acids act is unclear [73]. In Acremonium chrysogenum, the addition of DL-methionine to the medium led to increased mRNA levels of cephalosporin biosynthesis genes pcbAB (acvA), pcbC (ipnA) and cefEF, encoding deacetylcephalosporin C synthetase/hydroxylase and a three- to fourfold increase in the production of cephalosporin C [74, 75]. Uptake of amino acids in filamentous fungi is mediated by active amino acid permeases. In general, fungi possess a multiplicity of amino acid permeases

124

M. E. Evers et al.

that are involved in the uptake of amino acids from the environment as nitrogen and/or carbon source or as building blocks for the synthesis of proteins and peptides [76]. Biochemical and genetic characterization of fungal amino acid transporters has been performed most extensively in S. cerevisiae. Most fungal amino acid permeases show significant sequence similarities and form a unique family referred to as the AAP family [77], a subfamily of the APC family [78]. The permeases have a common structural organization with 12 putative a-helical transmembrane segments and cytoplasmically located N- and Cterminal hydrophilic regions [79–81]. Uptake occurs as secondary transport, i.e. by proton symport, with the pmf as driving force in order to allow uptake against the concentration gradient [77, 78, 82]. An exception is an amino acid permease encoded by the mtr locus of Neurospora crassa [83, 84]. This permease is unrelated to the AAP family, but instead it belongs the amino acid/auxin permease (AAAP) family [85]. Recently, a functional and structural homologue of the mtr encoded permease was found in P. chrysogenum (unpublished results). Genome analysis of S. cerevisiae revealed 24 members of the AAP family, of which most have been functionally characterized [79]. Some of them are specific for one or a group of related L-amino acids, such as Dip5p (glutamate and aspartate), Put4p (proline), Can1p (arginine). Others have a broader specificity, like Agp1p, which transports most neutral amino acids [79, 86].The general amino acid permease Gap1p, transports all L-and D-amino acids and nonproteinogenic amino acids such as citrulline and ornithine [87]. In P. chrysogenum, so far three amino acid permeases have been cloned and characterized [88]; (Trip et al. unpublished results), and various other activities have been classified on the basis of transport and competition assays. Nine amino acid transport systems have been reported: system I for L-methionine [89]; II for L-cysteine [90]; III for all amino acids [91, 92], analogous to Gap1p of S. cerevisiae; IV for acidic amino acids; V for L-proline; VI for L-lysine and L-arginine, VII for L-arginine; VIII for L-lysine and IX for L-cysteine [92]. The first two systems are expressed under sulphur starvation, while systems III-V are expressed under nitrogen and carbon starvation (NCR and CCR). Systems VI-VIII appear constitutive [76]. System VI was studied by Hillenga et al., 1996 [93], using plasma membranes fused with liposomes containing cytochrome c. Factors that interfere with the analysis of the plasma membrane transport processes when performed with intact mycelium, like metabolism and compartmentalisation, were circumvented this way. Inhibition studies with analogues revealed a narrow substrate specificity for arginine and lysine and quantitative analysis of arginine uptake suggest a H+-arginine symport stoichiometry of one-to-one [93]. Uptake of a-aminoadipate, an acidic amino acid similar in structure to glutamate, might be mediated by both the acidic amino acid transport system [92], and general amino acid transport system [76]. The acidic amino acid permease gene, DipP, was cloned and biochemically characterized. This transporter is homologous to Dip5 of S. cerevisiae [79] and is capable of transporting a-aminoadipate, albeit with much lower affinity than the preferred substrates aspartate and glutamate (Km of 800 and 35 mmol/l, re-


125

spectively) (H. Trip, unpublished results). Transport studies with penicillin producing mycelium show that a-aminoadipate uptake is strongly competed by leucine which is a substrate for the general amino acid permease, and not for DipP. This suggests that the general amino acid permease provides the main route for a-aminoadipate uptake into the cell. The expression of DipP is, like Dip5 in S. cerevisiae, under nitrogen catabolite repression and is strongly induced when glutamate is the only nitrogen source in the culture medium (H. Trip, unpublished results). ARLP encodes a permease specific for aromatic amino acids and leucine [88] and MTRP encodes a permease specific for neutral aliphatic and aromatic amino acids (H. Trip, unpublished results). MtrP is a structural and functional homologue of the mtr locus encoded protein of N. crassa and therefore not related to the AAP family, but a member of the AAAP family [88]. The physiological role of these permeases is unclear. It has been postulated that the ACV precursors a-aminoadipate, cysteine and valine are sequestered in the vacuole of P. chrysogenum. Cysteine and valine are produced in the vacuole due to proteolytic degradation of proteins. The presumed vacuolar localization of ACVS would then benefit from a direct withdrawal of these amino acids from the vacuolar pools [13]. The recent observation that ACVS is located in the cytosol [14] and the fact that in S. cerevisiae the acidic amino acids glutamate and aspartate are not accumulated in the vacuole, but instead, are located almost exclusively in the cytosol [94], do not support the vacuolar storage of the acidic amino acid a-aminoadipate. In S. cerevisiae, four vacuolar amino acid transporters have been identified, one of which,AVT6, mediates the efflux of the acidic amino acids glutamate and aspartate from the vacuole. These transporters do not show homology with amino acid permeases from the cellular membrane [95]. 5.2 Uptake of Sulfate and Phosphate The uptake of sulfate is an important step in the regulation of sulphur metabolism in P. chrysogenum. This uptake has been studied with mycelium and isolated plasma membrane vesicles. These experiments showed that uptake is mediated by a electroneutral sulfate/proton symport mechanism [96]. The P. chrysogenum membrane vesicles were fused with cytochrome-c oxidase containing liposomes to provide the system with a proton motive force. Sulfate uptake was solely dependent on the transmembrane pH gradient, and occurred with high affinity (Km~30 mmol/l). Apart from sulfate, the transporter also showed affinity for analogous divalent oxyanions like thiosulfate, selenate and molybdate. The genes of two putative sulfate transporters (designated SutA and SutB), and PAPS reductase (parA) have been cloned and sequenced [28]. SutB is the major sulfate transporter, while the exact function of SutA remains to be elucidated. This protein has been implicated in thiosulfate uptake or is possibly involved in an intracellular sulfate uptake activity. Expression studies were performed to determine if there is a relationship between penicillin biosyn-

126

M. E. Evers et al.

thesis and sulfate metabolism. Under sulphur starvation conditions the expression levels of both sulfate transporters are elevated. A positive correlation was observed between the levels of sutB mRNA and the penicillin biosynthesis, but such a correlation was not apparent for sutA and parA mRNAs. The parA mRNA levels are controlled by the sulphur content of the medium. It is generally believed that SutB is the main route for sulfate uptake during b-lactam biosynthesis [97]. Phosphate transport in P. chrysogenum has hardly been studied. In fungi uptake of phosphate occurs through proton and sodium phosphate symport. In fermentation media phosphate addition does not in itself inhibit penicillin production, but it strongly enhances the effect of glucose repression of transcription of the genes of the penicillin cluster [98]. In S. cerevisiae at least 5 transporters are involved in this process namely PHO84, 87, 89, 90 and 91. Deletion of all five genes is lethal [99]. Pho90 and Pho91 have the highest phosphate transporting capacity, whereas Pho84 and Pho87 are specific phosphate sensors. Pho89 has a very low transporting capacity and is not involved in phosphate signalling [99]. Pho84 is a phosphate proton symporter belonging to the Major Facilitator Superfamily (MFS) proteins and contains 12 membrane spanning segments. 5.3 Uptake of Nitrogen-Containing Compounds Ammonium, nitrate, urea and amino acids are possible nitrogen sources in blactam synthesis. The uptake and synthesis of amino acids has been described above. Although one of the earliest reports about an active ammonium transport system concerned the uptake of methylammonium by P. chrysogenum [100], no major new insights have been obtained since then. At high concentration, methylammonium is toxic to cells, and this was used to screen for mutants of S. cerevisiae and A. nidulans that are impaired in methylammonium uptake [101, 102]. This screen lead to the identification of several genes that encode (methyl)ammonium transporters (MEP/AMT). In A nidulans, two ammonium transporters have been described. These two proteins, MeaA and MepA are also involved in the retention of ammonium as determined by crossfeeding studies [103]. Although the molecular mechanism of transport is still unclear, studies using the LeAMT1 plasma membrane ammonium transporter of tomato (Lycopersicon esculentum) that was functionally expressed in Xenopus oocytes, indicate that ammonium ions are the substrates rather than ammonia. Uptake seems to take place by means of a uniport mechanism [104]. 5.4 Uptake of Side Chain Precursors Phenylacetic acid (PA) and phenoxyacetic acid are weak acids that rapidly enter P. chrysogenum cells through passive diffusion and distribute across the


127

membrane according to the transmembrane pH gradient [105, 106]. However, various reports have implicated active transport in the acquisition of phenylacetic acid from the medium [107]. The major differences in these studies may relate to concentration of phenylacetic acid used, and eventually the type of strains (low- vs high-yielding strains). When high concentrations of PA (60–3000 mmol/l) are used, PA readily enters the cells through passive diffusion in both low- and high-yielding strains. However, at low concentrations (1.4–100 mmol/l) accumulation of PA in the low yielding strain exceeds the accumulation of PA in the high yielding by a factor 10 at the lowest concentrations, suggesting the involvement of a transporter protein [108]. However, the latter may also relate to side-chain activation. Instead of uptake, the activity of the CoA ligase may be responsible for the observed retention. However, during b-lactam biosynthesis, high concentrations (millimolar) of PA or phenoxyacetic acid are fed to the cells, which makes that passive diffusion will be the dominating route of entry into the cell. 5.5 Uptake of Sugars The supply of sugars as the major carbon in industrial fermentation of P. chrysogenum is of importance, as it accounts for more than 10% of the overall costs [109]. Moreover, sugar plays an important role in the regulation of penicillin biosynthesis. Glucose and sucrose impose a strong inhibitory effect on blactam production by repression of penicillin biosynthesis genes (acvA and ipnA in P. chrysogenum, ipnA in A. nidulans) as well as by post-transcriptional (down)regulation (IAT in A. nidulans) [110]. Lactose does not inhibit b-lactam biosynthesis, which, for P. chrysogenum, was suggested to be due to the slow hydrolysis into glucose and galactose resulting from very low b-galactosidase activity [98]. Lactose has been traditionally used for penicillin biosynthesis, but during industrial fermentation, a limiting glucose-feed is now regularly used, avoiding carbon source/catabolite regulation [68]. Glucose uptake in fungi has been best studied for S. cerevisiae. Glucose transport occurs by facilitated diffusion [111] which involves transporters that belong to the MFS family [112, 113]. A family of 20 different hexose transporters or related proteins (Hxtp) is thought to be involved in sugar transport and regulation [111, 114]. In a hxt1–7 disruption mutant strain, glucose uptake is abolished, whereas the expression of any one of the genes HXT1, 2, 3, 4, 6 or 7 can restore glucose uptake [115]. Hxt1p and Hxt3p are low affinity transporters (Km=50–100 mmol/l), Hxt2p and Hxt4p are equipped with a moderately low affinity (10 mmol/l) and Hxt6p and Hxt7p are high affinity glucose transporters (1–2 mmol/l). A galactose permease was also shown to transport glucose with high affinity (Km=1–2 mmol/l) [114]. In filamentous fungi, glucose uptake systems have been described for A. nidulans [114, 116], A. niger [117], and N. crassa [118–123]. In general, at least two systems appear to be present, a constitutive, passive, low-affinity system, and a glucose repressible, proton

128

M. E. Evers et al.

motive force-driven, high-affinity system. The high-affinity system generally has a much lower Km-value than found in S. cerevisiae. The Km for the highaffinity system in A. nidulans is 0.04–0.06 mmol/l [81, 124]; for P. chrysogenum a value of 0.2 mmol/l has been reported [125]. Like in S. cerevisiae, more than two transporter proteins might be involved in glucose transport, but since mutants disrupted in one or more glucose transporter genes are not yet available, individual characterization is complicated. Little information is available on lactose transport in fungi. The best characterized fungal lactose uptake system is the inducible LAC12 gene product of Kluyveromyces lactis [126, 127], which transports lactose in symport with protons. Proton symport seems to be a general mechanism for disaccharide transport in fungi [126–128]. In P. chrysogenum, lactose is taken up by an energy-dependent system, mostly likely proton motive force-driven system. The lactose transport activity is induced when cells are growth on lactose (van de Kamp et al., unpublished).

6 Transport Across the Microbody Membrane As mentioned previously, some of the enzymatic steps of the penicillin and lysine synthesis pathway take place inside the microbody. The exact reason why these steps are localized in this intracellular organelle is not clear, but it has been hypothesized that the microbody lumen provides an optimal environment for these enzymes for instance with respect to pH, metabolite concentration etc. The internal pH of peroxisomes in the yeast Hansenula polymorpha has been reported to be acidic (pH 5.8–6) [129]. However, the pH optimum of IAT is in the alkaline range, and the enzyme is inactive at pH values lower than 6 [130, 131]. The same has been reported for PCL which is likely localized in the microbody. The pH of P. chrysogenum microbodies has also been investigated with the enhanced yellow fluorescent protein (eYFP) that was targeted to the microbody by means of a C-terminal PTS1 signal SKL. Based on the fluorescence characteristics, it was concluded that the microbody is not acidic, but slightly alkaline (pH 7.0–7.5) [132]. This is more in accordance with the pH optimum of the abovementioned biosynthesis enzymes. Studies on microbodies in human fibroblasts even suggest that the luminal pH may be as alkaline as pH 8.2 [133]. Other possible advantages for compartmentalization of key enzymatic steps may relate to the higher concentrations of both enzymes and substrates, the prevention of draining catalytic intermediates into unwanted side reaction pathways, and/or regulation of the biosynthesis pathway. The subcellular distribution of the various enzymatic steps over different organelles poses, however, important problems towards the transport of the metabolites. For a long time it was believed that peroxisomes are permeable to small compounds. For instance, it was not possible to obtain peroxisomes while maintaining the permeability barrier of the membrane. Also, a porin-like protein has been found


129

to be associated with the peroxisomal membrane [134–137]. However, the in vivo studies on the luminal pH and identification of various transporters now suggest that the peroxisomal membrane represents a permeability barrier. NAD(H), NADP(H), acetyl-CoA, ATP and protons cannot freely pass peroxisomal membranes of different organisms [67, 133, 138, 139]. The necessity for peroxisomal membrane proteins (PMPs) with a transport function is therefore obvious. Biochemical studies, however, have suffered from the fact that the organelles are very fragile, while PMPs appear of low abundance [140]. Of the known peroxisomal transporters only one has been studied in detail with respect to substrate specificity, namely the peroxisomal adenine nucleotide transporter Ant1p of S. cerevisiae. This system is very homologous to the mitochondrial transporter family. Ant1p has been overproduced, purified from the peroxisomal membrane fractions and reconstituted into liposomes [67]. The system has been suggested to function as an ATP/AMP antiporter, supply the microbody lumen with cytosolic ATP. So far, experimental evidence is lacking for the involvement of transporters in the uptake of IPN and PA or the extrusion of a-aminoadipate and penicillins.

7 Excretion of b -lactams into the Medium The mechanism of excretion of b-lactams into the medium has been a subject of speculation for a long time.Various options need to be considered, i.e., passive diffusion, vesicular transport and the involvement of transport proteins. Passive diffusion phenomena are strongly dependent on the physicochemical characteristics of the membrane, like fluidity, degree of saturation and the acyl chain length of the lipid fatty acids but also on the intrinsic properties of the compound, like charge, size and hydrophobicity. Penicillins V and G are amphiphatic, moderately hydrophobic molecules and negatively charged at the cytosolic pH. The diffusion of these molecules has been studied in model membranes, and it was suggested that they can permeate a membrane composed of phospholipids [106]. The permeability characteristics of the membrane were, however, greatly reduced when sterols were present in the membrane. Since plasma membranes of P. chrysogenum contain 30% ergosterol, a concentration that suffices to block most of the passive permeation, passive diffusion seems very unlikely [106]. During recent years it has become increasingly evident that all living cells are equipped with multidrug transporters that are capable of expelling unrelated, mostly hydrophobic compounds across the membrane. These transporters convey multidrug resistance to cells. Due to the physiochemical characteristics of penicillins, MDR transporters are likely candidates for b-lactam secretion. MDR transporters can be subdivided into six families (for a review see [141, 142]). Two of the transporter families have already been implicated in b-lactam extrusion and will be briefly discussed here; namely the ATP-binding-

130

M. E. Evers et al.

cassettes (ABC) transporter superfamily and the MFS of proteins. The ABCtransporters form a very large family of proteins with a very broad spectrum of substrate specificity, they translocate both small and large molecules across membranes. They are characterized by the presence of two cytosolic nucleotide binding domains (NBD’s) each containing the highly conserved Walker A and Walker B motifs that specify the nucleotide binding site, and two transmembrane domains consisting of six transmembrane spanning segments [141]. The MFS is also referred to as the uniporter-symporter-antiporter family. These proteins are secondary transporters that transport small molecules in a proton motive force-dependent manner. They can be classified into 17 families. This includes the drug:H+ antiporter families that specify membrane proteins with either 12 or 14 membrane spanning segments [142]. Recently, the first experimental evidence has been obtained that secretion of b-lactam in filamentous fungi may indeed involve active transport. This concerned a study on the involvement of ABC-transporters of A. nidulans in drug resistance. After identification of a number of ABC-transporter genes, a disruption mutant for the atrD gene displayed increased sensitivity towards the chemically unrelated compounds valinomycin, nigericin and cycloheximide. Moreover, in a halo size assay, used as a measure of the amount of penicillin produced, a reduced penicillin production was detected for an atrD deletion strain [143]. This suggests a role of the ATRDp in penicillin secretion, although this needs to be verified by direct transport assays. In another study, the region downstream of the acvS gene of Acremonium chrysogenum was examined which identified a gene encoding a membrane protein (CefT) belonging to the MFS. The deduced protein sequence revealed that this protein belongs to the family of drug: H+ antiporters with 12 transmembrane segments. Disruption of the gene showed that it was not required for cephalosporin synthesis and that growth of A. chrysogenum was not affected. However, amplification of the full length gene (2 to 4 copies) resulted in a twofold increase in the cephalosporin C production [29]. Both studies, however, await direct proof that the identified transporters (AtrD and CefT) mediate antibiotic transport. Also in P. chrysogenum, a series of ABC transporters have been identified that are expressed under penicillin producing conditions [144]. Some of these MDR-like ABC transporters are induced when cells are challenged with b-lactam, suggesting a role in b-lactam excretion. In various antibiotic producing organisms, genes have been identified that confer resistance to the produced antibiotic presumably by transporting the drug out of the cells. In the gene cluster for antibiotic biosynthesis of Streptomyces argillaceus the genes mtrA and mtrB are present that encode a putative ABCtransporter and render the organism resistant to mithramycin [145]. In b-lactam producing actinomycetes like Nocardia lactamdurans genes are found that code for transporter proteins either belonging to the ABC-transporter superfamily or the MFS [27].


131

8 Concluding Remarks In recent years, major insights have been obtained in the compartmentalization of the b-lactam biosynthesis pathway in filamentous fungi. The exact molecular reasons for the localization of the last step of biosynthesis steps in a microbody are not known, although the specific pH in this organelle seems favourable for the catalytic activity of the key enzymes. Control of the cellular distribution of a-aminoadipate to direct it either into the lysine or penicillin biosynthesis pathway may be crucial now it appears that critical enzymatic steps take place in the microbody. However, many of the molecular details still need to be resolved (see chapter Brakhage et al.). Other questions concern if b-lactam synthesis is limited by transport process, as for instance, cellular secretion? Intrinsic to the approach of removing bottlenecks from a metabolic production process, new limiting factors are found one of which may related to transport, a process often ignored in metabolic pathway engineering programs. The antibiotic resistance of bacteria necessitates the discovery and production of new antibiotics. Genetic engineering enables us to intervene in metabolic and biosynthetic pathways thereby providing new opportunities of product formation. Such challenging metabolic reprogramming efforts also require insights in critical transport steps and possible limitation by exciting substrate specificities. Metabolic flux analyses of genetically altered strain, genome sequencing and transcriptome profiling, and directed evolution promise to be interesting tools for the near future. Acknowledgements This work was supported by grants from the European Union (BIOT CT 94-2100, and Eurofung cell factory RTD BIO4CT96-0535 and QLK3-CT-1999-00729), STW (Stichting Toegepaste Wetenschappen), the EET K20002 Cell Factory Project, and by DSM Anti-infectives (Delft, The Netherlands).

References 1. Carr LG, Skatrud PL, Scheetz ME, Queener SW, Ingolia TD (1986) Gene 48:257 2. Cohen G, Argaman A, Schreiber R, Mislovati M, Aharonowitz Y (1994) J Bacteriol 176:973 3. Cooper RDG (1993) Bioorg Medic Chem 1:1 4. Ingolia TD, Queener SW (1989) Med Res Rev 9:245 5. Dhar MM, Khan AW (1971) Nature 233:182 6. Aharonowitz Y, Cohen G, Martin JF (1992) Annu Rev Microbiol 46:461 7. Kleinkauf H, von Dohren H (1996) Eur J Biochem 236:335 8. Kleinkauf H, von Dohren H (1987) Annu Rev Microbiol 41:259 9. Kleinkauf H, von Dohren H (1995) Antonie Van Leeuwenhoek 67:229 10. Zhang J, Wolfe S, Demain AL (1992) Biochem J 283:691 11. Fawcett P, Abraham EP (1975) Methods Enzymol 43:471 12. Kurylowicz W, Kurzatkowski W, Kurzatkowski J (1987) Arch Immunol Ther Exp (Warsz) 35:699

132

M. E. Evers et al.

13. Lendenfeld T, Ghali D,Wolschek M, Kubicek-Pranz EM, Kubicek CP (1993) J Biol Chem 268:665 14. van der Lende TR, van deKamp M, van den Berg M, Sjollema K, Bovenberg RA, Veenhuis M, Konings WN, Driessen AJM (2002) Fungal Genet Biol 37:49 15. Vousden W, Turner G (2001) Fungal Genetics Conference Fungal Genet Newsletter 48:160 16. Burzlaff NI, Rutledge PJ, Clifton IJ, Hensgens CM, Pickford M,Adlington RM, Roach PL, Baldwin JE (1999) Nature 401:721 17. Muller WH, van der Krift TP, Krouwer AJ, Wosten HA, van der Voort LH, Smaal EB, Verkleij AJ (1991) EMBO J 10:489 18. Barredo JL, van Solingen P, Diez B, Alvarez E, Cantoral JM, Kattevilder A, Smaal EB, Groenen MA, Veenstra AE, Martin JF (1989) Gene 83:291 19. Luengo JM (1995) J Antibiot (Tokyo) 48:1195 20. Tobin MB, Fleming MD, Skatrud PL, Miller JR (1990) J Bacteriol 172:5908 21. Whiteman PA,Abraham EP, Baldwin JE, Fleming MD, Schofield CJ, Sutherland JD,Willis AC (1990) FEBS Lett 262:342 22. Aplin RT, Baldwin JE, Roach PL, Robinson CV, Schofield CJ (1993) Biochem J 294:357 23. Aplin RT, Baldwin JE, Cole SC, Sutherland JD, Tobin MB (1993) FEBS Lett 319:166 24. Gouka RJ, van Hartingsveldt W, Bovenberg RA, van Zeijl CM, van den Hondel CA, van Gorcom RF (1993) Appl Microbiol Biotechnol 38:514 25. Gledhill L, Greaves PA, Griffin JP (1997) International Patent IPN WO97/02349 26. Minambres B, Martinez-Blanco H, Olivera ER, Garcia B, Diez B, Barredo JL, Moreno MA, Schleissner C, Salto F, Luengo JM (1996) J Biol Chem 271:33531 27. Liras P (1999) Antonie van Leeuwenhoek 75:109 28. van de Kamp M, Pizzinini E,Vos A, van der Lende TR, Schuurs TA, Newbert RW, Turner G, Konings WN, Driessen AJM (1999) J Bacteriol 181:7228 29. Ullan RV, Liu G, Casqueiro J, Gutierrez S, Banuelos O, Martin JF (2002) Mol Genet Genomics 267:673 30. Martin JF (1998) Appl Microbiol Biotechnol 50:1 31. Martin JF, Liras P (1989) Adv Biochem Eng/Biotechnol 39:153 32. Mathison L, Soliday C, Stepan T, Aldrich T, Rambosek J (1993) Curr Genet 23:33 33. Matsuda A, Sugiura H, Matsuyama K, Matsumoto H, Ichikawa S, Komatsu K (1992) Biochem Biophys Res Commun 186:40 34. van de Kamp M, Driessen AJM, Konings WN (1999) Antonie van Leeuwenhoek 75:41 35. Crawford L, Stepan AM, McAda PC, Rambosek JA, Conder MJ, Vinci VA, Reeves CD (1995) Biotechnology (NY) 13:58 36. Robin J, Jakobsen M, Beyer M, Noorman H, Nielsen J (2001) Appl Microbiol Biotechnol 57:357 37. Sulter GJ, van der Klei IJ, Schanstra J, Harder W,Veenhuis M (1991) FEMS Microbiol Lett 82:297 38. van der Klei IJ, Harder W, Veenhuis M (1991) Arch Microbiol 156:15 39. Muller WH, Bovenberg RA, Groothuis MH, Kattevilder F, Smaal EB, van der Voort LH, Verkleij AJ (1992) Biochim Biophys Acta 1116:210 40. De Lucas JR, Valenciano S, Dominguez AI, Turner G, Laborda F (1997) Arch Microbiol 168:504 41. Purdue PE, Lazarow PB (2001) Annu Rev Cell Dev Biol 17:701 42. van der Klei IJ, Veenhuis M (2002) Curr Opin Cell Biol 14:500 43. López-Nieto MJ, Ramos FR, Luengo JM, Martin JF (1985) Appl Microbiol Biotechnol 22:343 44. Banuelos O, Casqueiro J, Steidl S, Gutierrez S, Brakhage A, Martin JF (2002) Mol Genet Genomics 266:711


133

45. Geraghty MT, Bassett D, Morrell JC, Gatto GJ Jr, Bai J, Geisbrecht BV, Hieter P, Gould SJ (1999) Proc Natl Acad Sci USA 96:2937 46. Breitling R, Sharif O, Hartman ML, Krisans SK (2002) Eukaryot Cell 1:978 47. Henriksen CM, Nielsen J, Villadsen J (1998) J Antibiot (Tokyo) 51:99 48. Foster BA, Thomas SM, Mahr JA, Renosto F, Patel HC, Segel IH (1994) J Biol Chem 269:19777 49. Renosto F, Seubert PA, Segel IH (1984) J Biol Chem 259:2113 50. Renosto F, Martin RL, Segel IH (1987) J Biol Chem 262:16279 51. Marzluf GA (1993) Annu Rev Microbiol 47:31 52. Nuesch J, Heim J, Treichler HJ (1987) Annu Rev Microbiol 41:51 53. Hersbach GJM, van der Beek CP, van Dijck PWM (1984) In: Vandamme EJ (ed) Biotechnology of industrial antibiotics. Marcel Dekker, NY, pp 45–140 54. Jørgensen HS, Nielsen J, Villadsen J, Møllergaard H (1995) Biotechnol Bioeng 46:117 55. Pieniazek N, Stepien PP, Paszewski A (1973) Biochim Biophys Acta 297:37 56. Treichler HJ, Liersch M, Nüesch J (1979) In: Sebek OK, Laskin AI (eds) Genetics of industrial microorganisms. AMS, Washington, pp 97–104 57. Döbeli H, Nüesch J (1980) Antimicrob Agents Chemother 18:111 58. Østergaard S, Theilgaard HBA, Nielsen J (1998) Appl Microbiol Biotechnol 50:663 59. Kubicek-Pranz EM, Kubicek CP (1991) Production and biosynthesis of amino acids by fungi. In: Arora DK, Elander RP, Mukerji KG (eds) Handbook of applied mycology, vol 4: fungal biotechnology. Marcel Dekker, New York, p 313 60. Fiermonte G, Palmieri L, Todisco S, Agrimi G, Palmieri F, Walker JE (2002) J Biol Chem 277:19289 61. Palmieri F (1994) FEBS Lett 346:48 62. Palmieri L, Runswick MJ, Fiermonte G, Walker JE, Palmieri F (2000) J Bioenerg Biomembr 32:67 63. Palmieri L, Pardo B, Lasorsa FM, del Arco A, Kobayashi K, Iijima M, Runswick MJ, Walker JE, Saheki T, Satrustegui J, Palmieri F (2001) EMBO J 20:5060 64. Indiveri C, Tonazzi A, Stipani I, Palmieri F (1999) Biochem J 341:705 65. Palmieri L, De MV, Iacobazzi V, Palmieri F, Runswick MJ, Walker JE (1997) FEBS Lett 410:447 66. Marobbio CM, Vozza A, Harding M, Bisaccia F, Palmieri F, Walker JE (2002) EMBO J 21:5653 67. Palmieri L, Rottensteiner H, Girzalsky W, Scarcia P, Palmieri F, Erdmann R (2001) EMBO J 20:5049 68. Nielsen J, Jorgensen HS (1995) Biotechnol Prog 11:299 69. Honlinger C, Kubicek CP (1989) FEMS Microbiol Lett 53:71 70. Demain AL (1957) Arch Biochem Biophys 67:244 71. Luengo JM, Revilla G, Lopez MJ, Villanueva JR, Martin JF (1980) J Bacteriol 144:869 72. Lara F, del Carmen MR, Vazquez G, Sanchez S (1982) Biochem Biophys Res Commun 105:172 73. Then Bergh K, Brakhage AA (1998) Appl Environ Microbiol 64:843 74. Drew SW, Demain AL (1973) Biotechnol Bioeng 15:743 75. Velasco J, Gutierrez S, Fernandez FJ, Marcos AT, Arenos C, Martin JF (1994) J Bacteriol 176:985 76. Horak J (1986) Biochim Biophys Acta 864:223 77. Andre B (1995) Yeast 11:1575 78. Jack DL, Paulsen IT, Saier MH (2000) Microbiology 146:1797 79. Regenberg B, During-Olsen L, Kielland-Brandt MC, Holmberg S (1999) Curr Genet 36:317 80. Horak J, Wolf DH (1997) J Bacteriol 179:1541

134

M. E. Evers et al.

81. 82. 83. 84. 85. 86.

Brown CE, Romano AH (1969) J Bacteriol 100:1198 Seaston A, Inkson C, Eddy AA (1973) Biochem J 134:1031 Koo K, Stuart WD (1991) Genome 34:644 Stuart WD, Koo K, Vollmer SJ (1988) Genome 30:198 Young GB, Jack DL, Smith DW, Saier MH Jr (1999) Biochim Biophys Acta 1415:306 Iraqui I,Vissers S, Bernard F, de Craene JO, Boles E, Urrestarazu A, Andre B (1999) Mol Cell Biol 19:989 Jauniaux JC, Grenson M (1990) Eur J Biochem 190:39 Trip H, Evers ME, Konings WN, Driessen AJM (2002) Biochim Biophys Acta 1565:73 Benko PV, Wood TC, Segel IH (1967) Arch Biochem Biophys 122:783 Skye GE, Segel IH (1970) Arch Biochem Biophys 138:306 Benko PV, Wood TC, Segel IH (1969) Arch Biochem Biophys 129:498 Hunter DR, Segel IH (1971) Arch Biochem Biophys 144:168 Hillenga DJ, Versantvoort HJ, Driessen AJM, Konings WN (1996) J Bacteriol 178:3991 Kitamoto K, Yoshizawa K, Ohsumi Y, Anraku Y (1988) J Bacteriol 170:2683 Russnak R, Konczal D, McIntire SL (2001) J Biol Chem 276:23849 Hillenga DJ, Versantvoort HJ, Driessen AJM, Konings WN (1996) J Bacteriol 178: 3953 van de Kamp M, Schuurs TA, Vos A, van der Lende TR, Konings WN, Driessen AJM (2000) Appl Environ Microbiol 66:4536 Martin JF, Casqueiro J, Kosalkova K, Marcos AT, Gutierrez S (1999) Antonie van Leeuwenhoek 75:21 Giots F, Donaton MC, Thevelein JM (2003) Mol Microbiol 47:1163 Hackette SL, Skye GE, Burton C, Segel IH (1970) J Biol Chem 245:4241 Arst HN Jr, Cove DJ (1969) J Bacteriol 98:1284 Dubois E, Grenson M (1979) Mol Gen Genet 175:67 Monahan BJ, Unkles SE, Tsing IT, Kinghorn JR, Hynes MJ, Davis MA (2002) Fungal Genet Biol 36:35 Ludewig U, von Wiren N, Frommer WB (2002) J Biol Chem 277:13548 Eriksen SH, Jensen B, Schneider I, Kaasgaard S, Olsen J (1995) Appl Microbiol Biotechnol 42:945 Hillenga DJ (1999) Thesis University of Groningen Fernandez-Canon JM, Reglero A, Martinez-Blanco H, Ferrero MA, Luengo JM (1989) J Antibiot (Tokyo) 42:1410 Eriksen SH, Soderblom TB, Jensen B, Olsen J (1998) Biotechnol Bioeng 60:310 Cooney CL, Acevedo F (1977) Biotechnol Bioeng 19:1449 Litzka O, Then Bergh K, Van den Brulle I, Steidl S, Brakhage AA (1999) Antonie van Leeuwenhoek 75:95 Kruckeberg AL (1996) Arch Microbiol 166:283 Horak J (1997) Biochim Biophys Acta 1331:41 Marger MD, Saier MH Jr (1993) Trends Biochem Sci 18:13 Reifenberger E, Boles E, Ciriacy M (1997) Eur J Biochem 245:324 Reifenberger E, Freidel K, Ciriacy M (1995) Mol Microbiol 16:157 Mark CG, Romano AH (1971) Biochim Biophys Acta 249:216 Torres NV, Riol-Cimas JM, Wolschek M, Kubicek CP (1996) Appl Microbiol Biotechnol 44:790 Madi L, McBride SA, Bailey LA, Ebbole DJ (1997) Genetics 146:499 Scarborough GA (1970) J Biol Chem 245:1694 Scarborough GA (1970) J Biol Chem 245:3985 Schneider RP, Wiley WR (1971) J Bacteriol 106:479 Schneider RP, Wiley WR (1971) J Bacteriol 106:487

87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122.


135

123. Slayman CL, Slayman CW (1974) Proc Natl Acad Sci USA 71:1935 124. Mark CG, Romano AH (1971) Biochim Biophys Acta 249:216 125. Christensen LH, Henriksen CM, Nielsen J,Villadsen J, Egel-Mitani M (1995) J Biotechnol 42:95 126. Chang YD, Dickson RC (1988) J Biol Chem 263:16696 127. Dickson RC, Barr K (1983) J Bacteriol 154:1245 128. van der Rest ME, de Vries Y, Poolman B, Konings WN (1995) J Bacteriol 177:5440 129. Nicolay K, Veenhuis M, Douma AC, Harder W (1987) Arch Microbiol 147:37 130. Alvarez E, Cantoral JM, Barredo JL, Diez B, Martin JF (1987) Antimicrob Agents Chemother 31:1675 131. Alvarez E, Meesschaert B, Montenegro E, Gutierrez S, Diez B, Barredo JL, Martin JF (1993) Eur J Biochem 215:323 132. van der Lende TR, Breeuwer P,Abee T, Konings WN, Driessen AJM (2002) Biochim Biophys Acta 1589:104 133. Dansen TB, Wirtz KW, Wanders RJ, Pap EH (2000) Nat Cell Biol 2:51 134. Lemmens M, Verheyden K, Van Veldhoven P, Vereecke J, Mannaerts GP, Carmeliet E (1989) Biochim Biophys Acta 984:351 135. Reumann S, Maier E, Benz R, Heldt HW (1996) Biochem Soc Trans 24:754 136. Reumann S, Bettermann M, Benz R, Heldt HW (1997) Plant Physiol 115:891 137. Reumann S, Maier E, Heldt HW, Benz R (1998) Eur J Biochem 251:359 138. van Roermund CW, Elgersma Y, Singh N,Wanders RJ, Tabak HF (1995) EMBO J 14:3480 139. Visser WF, van Roermund CW,Waterham HR,Wanders RJ (2002) Biochem Biophys Res Commun 299:494 140. Schafer H, Nau K, Sickmann A, Erdmann R, Meyer HE (2001) Electrophoresis 22:2955 141. Saier MH Jr, Beatty JT, Goffeau A, Harley KT, Heijne WH, Huang SC, Jack DL, Jahn PS, Lew K, Liu J, Pao SS, Paulsen IT, Tseng TT, Virk PS (1999) J Mol Microbiol Biotechnol 1:257 142. Pao SS, Paulsen IT, Saier MH Jr (1998) Microbiol Mol Biol Rev 62:1 143. Andrade AC, Van Nistelrooy JG, Peery RB, Skatrud PL, De Waard MA (2000) Mol Gen Genet 263:966 144. Bovenberg RAL, Driessen AJM, Schuurs TA, Nieboer M, van den Berg MA, Konings WN, Westerlaken I (2001) International Patent WO0132904 145. Fernandez E, Lombo F, Mendez C, Salas JA (1996) Mol Gen Genet 251:692

Received: January 2004


Control of Fluxes Towards Antibiotics and the Role of Primary Metabolism in Production of Antibiotics Nina Gunnarsson 1 · Anna Eliasson · Jens Nielsen Biocentrum-DTU, Center for Microbial Biotechnology, Building 223, Søltofts plads, 2800 Lyngby, Denmark 1 Present address: Fluxome Sciences A/S, Søltofts plads, Building 223, 2800 Lyngby, Denmark [email protected] · [email protected] · [email protected]

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

1

Introduction

2 2.1 2.2 2.3

Flux Control in Biosynthesis Pathways Towards Antibiotics . Principles of MCA . . . . . . . . . . . . . . . . . . . . . . . MCA of the Penicillin V Pathway in Penicillium Chrysogenum Overexpression of Biosynthesis Genes . . . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

140 141 142 145

3 3.1 3.2 3.3 3.4 3.5

Integrated Analysis . . . . . . Metabolite Balancing . . . . . The Use of Labelled Substrates Identification of Pathways . . Node Flexibility . . . . . . . . Maximum Theoretical Yield .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

146 147 151 152 155 156

4 4.1 4.1.1 4.1.2 4.1.3 4.2

Linking the Primary and Secondary Metabolism . Precursor Supply . . . . . . . . . . . . . . . . . . Precursor Requirements in Polyketide Production Precursor Requirements in b-Lactam Production . The Role of Synthesis of Specific Precursors . . . . Cofactor Supply . . . . . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

157 157 160 161 163 171

5

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

References

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

Abstract Yield improvements in antibiotic-producing strains have classically been obtained through random mutagenesis and screening. An attractive alternative to this strategy is the rational design of producer strains via metabolic engineering, an approach that offers the possibility to increase yields while avoiding the problems of by-product formation and altered morphological properties, which frequently arise in mutagenized strains. An important aspect in the design of strains with improved yields by metabolic engineering is the identification of rate-controlling enzymatic reactions in the metabolic network. Here we describe and discuss available methods for identification of these steps, both in antibiotic biosynthesis pathways and in the primary metabolism, which serves as the supplier of precursors and cofactors for the secondary metabolism. Finally, the importance of precursor and cofactor supply from primary metabolism in the biosynthesis of different types of antibiotics is discussed and recent developments in metabolic engineering towards increased product yields in antibiotic producing strains are reviewed.

138

N. Gunnarsson et al.

Keywords Metabolic flux analysis · Metabolic control analysis · Metabolic engineering · Polyketides · b-Lactams · Glycopeptides List of Abbreviations L-a-aa L-a-Amino adipic acid a-AAR L-a-Aminoadipic acid reductase ACP Acyl-carrier protein LLD-ACV d-(L-a-Aminoadipyl)-L-cysteinyl-D-valine ACVS ACV synthetase 7-ADC A7-Aminocephalosporanic acid 6-AP A6-Aminopenicillinic acid AT Acyl-CoA:isopenicillin N acyltransferase DAB 2,4-Diaminobutyric acid DAC Deacetylcephalosporin DAC-AT Acetyl CoA:DAC acetyltransferase DACS DAC synthase DAHP 3-Deoxy-D-arabino heptulosonate-7-phosphate DAOC Deacetoxycephalosporin DAOCS DAOC synthase ED pathway Entner-Doudoroff pathway EMP pathway Embden Meyerhof Parnas pathway E4P Erythrose-4-phosphate FCC Flux control coefficient F6P Fructose-6-phosphate G6P Glucose-6-phosphate HCS Homocitrate synthase IOA Isooctanic acid IPN Isopenicillin N IPNS Isopenicillin N synthase LAT Lysine 6-amino transferase MCA Metabolic control analysis MOA 6-Methyloctanic acid NRPS Non-ribosomal peptide synthases PenN Penicillin N PEP Phosphoenolpyruvate 3PGA 3-Phosphoglycerate 6PGA 6-Phosphogluconate PKS Polyketide synthase PP pathway Pentose phosphate pathway RIBU5P Ribulose-5-phosphate TCA cycle Tricarboxylic acid cycle tcm C Tetracenomycin C TR Thioredoxin-thioredoxin reductase system TREHAL Trehalose

Control of Fluxes Towards Antibiotics and the Role of Primary Metabolism

139

1 Introduction The penicillin production process was the first biotech process to be implemented in the pharmaceutical industry. It was introduced in the early 1940s and since its introduction there has been a continuous optimisation of the process, particularly focusing on improving the overall conversion yield of sugar to penicillin and the productivity, i.e. the amount of penicillin produced per unit time. Through classical strain improvement programs the productivity of industrial strains of Penicillium chrysogenum used for production of penicillin has been improved more than 1000-fold. Also for many other antibiotics the productivity has been improved substantially by classical mutagenesis and screening. Despite success in these strain improvement programs the antibiotic industry is, however, beginning to introduce a more rational approach to strain improvement – namely metabolic engineering where directed genetic modifications are introduced with the purpose of improving the performance of the applied strain [1, 2]. There are several reasons for introduction of metabolic engineering in the antibiotics industry: – Compared with mutagenesis and screening, directed genetic modifications generally do not result in appearance of undesirable side effects, e.g. formation of by-products and/or altered morphology. – Metabolic engineering allows for specific elimination of the formation of byproducts. – Through metabolic engineering one obtains a fundamental insight into the physiology of the applied microorganisms, and this may be applied to design improved fermentation processes. – Through metabolic engineering it is possible to apply the same microorganism (super-host or plug-bug) to produce different antibiotics. Hereby optimisation of the central carbon metabolism in the applied microorganism allows for improvement of production of two or more antibiotics. This is exemplified by the use of P. chrysogenum to produce both penicillin and 7-ADCA [3]. Improvement of antibiotic production through metabolic engineering often involves improving the flux through the biosynthesis pathway leading to the antibiotic of interest. There is therefore much focus on establishing the biosynthesis pathway and identifying the genes involved in the biosynthesis. Generally the flux through the biosynthesis pathway can be improved by overexpressing the genes encoding the biosynthesis genes, e.g. through insertion of multiple gene copies, promoter replacement and/or through overexpression of transcriptional activators. When the flux through the biosynthesis pathway has been improved flux-control will, however, often move to other parts of the metabolism. Thus, in high-yielding strains of P. chrysogenum, where there is a very high flux through the biosynthesis pathway, the supply of precursors for penicillin biosynthesis may become limiting for the overall production. Efficient strain improvement should therefore involve an evaluation of the precursor supply – particularly if the ap-

140


Fig. 1 Simplified overview of precursor requirements in the biosynthesis of various antibiotics

plied microorganism is to be applied as a super-host system for the production of more than one antibiotic.A key element in this evaluation involves identification of the precursors for biosynthesis of the antibiotic of interest and subsequently mapping of the fluxes in the central carbon metabolism at different environmental conditions. Figure 1 gives an overview of how the central carbon metabolism is linked to the production of many different antibiotics. In this review we will discuss how the fluxes towards antibiotics are controlled. Our focus will be on how the production of antibiotics is linked to the primary metabolism, and how possible limitations in the primary metabolism may be identified. We will primarily focus on a few types of antibiotics, i.e. b-lactams, polyketides and glycopeptides, in order to illustrate some general principles rather than give a detailed account of the status for production of many different antibiotics.

2 Flux Control in Biosynthesis Pathways Towards Antibiotics A requirement for application of metabolic engineering in the improvement of flux through a pathway leading to a desirable product, e.g. a specific antibiotic, is to have a fundamental understanding of the biosynthesis pathway leading to the product. Specifically, it is attractive to have information concerning possible regulation in the pathway and kinetics of the individual enzymes. This information may be used to identify which enzymatic step that exerts the main control of flux through the pathway, a potentially crucial issue in the design of strains with increased flux towards the product of interest. Metabolic Control Analysis (MCA) is a useful tool for integration of knowledge at the level of regulation and kinetics of the individual enzymes, which enables quantitative analysis of flux control in metabolic pathways [4, 5]. The principles of MCA are extensively described elsewhere [6–8], but in the following we give a short summary of the theory of MCA and its applications in antibiotic biosynthesis.


141

2.1 Principles of MCA At steady state all enzymatic reactions in a linear pathway proceed at the same rate, which is identical to the steady state flux through the pathway (Fig. 2). Thus, in order to increase the flux through the pathway, all individual enzymatic reaction rates in the pathway must be increased to the same level. Alteration of an individual enzyme activity may affect the pathway flux to various degrees, depending on the level of control this specific enzymatic step exerts on the flux through the pathway. The degree of control of a specific enzymatic step is dependent on the properties of the enzyme in relation to the properties of the remaining enzymes in the pathway, and can be determined through MCA of the pathway. We consider the pathway in Fig. 2, consisting of L reactions and K metabolites. The rate of each reaction can be described by a kinetic expression, e.g. Michaelis-Menten kinetics with feedback inhibition by the product. When the

Fig. 2 Linear metabolic pathway leading from substrate (S) to product (P). The substrate is converted to the product via K intermediate metabolites (X1,2...i...K) and L reactions, each catalysed by an enzyme (denoted E1,2...j...L).At steady state the forward rates of these reactions will be identical to the steady-state flux J through the pathway (v1=v2=...vj...vL=J). In MCA, the elasticity coefficients are measures of the sensitivity of local enzymatic reaction rates towards changes in metabolite concentrations, while the flux control coefficients are global properties that quantify the sensitivity of the total flux J through the pathway in relation to changes in individual enzymatic activities

142


enzyme kinetics is given the elasticity coefficients (Eq. 1) can be calculated for each reaction in the pathway: Xi ∂nj j eXi = 5 6 i Œ {1, 2…K}, j Œ {1, 2…L} nj ∂Xi

(1)

The elasticity coefficient quantifies the relative change in reaction rate j due to a relative change in concentration of metabolite i, and is thus a measure of the sensitivity of the reaction with respect to changes in metabolite concentrations. Elasticity coefficients can be calculated for each of the reactions in the pathway and are local properties, i.e. they are not dependent on properties of the complete pathway. Therefore, they provide information only for a single enzymatic step and not about the flux control exerted by the specific enzyme on the flux through the pathway. In contrast, the flux control coefficients (FCCs) (Eq. 2), are global properties that are dependent on properties of the complete pathway: Ej ∂J FCC = C jJ = 4 6 j Œ {1, 2…L} J ∂Ej

(2)

The FCCs quantify the relative change in the steady state flux J through the pathway, caused by a relative change in the activity of the j-th enzyme. As a consequence of the normalization, the FCCs of a pathway sum up to 1 (the so-called flux-control summation theorem). Thus, in a linear pathway, a FCC of 0 implies that no flux control is exerted, while a FCC of 1 means that the reaction in question is the only flux-controlling step and, thus, is rate-limiting for the pathway. Normally, the flux control is distributed on several reactions in the pathway. The elasticity coefficients and control coefficients are connected by the flux-control connectivity theorem, a central theorem for MCA (Eq. 3): L

j

Â C Jj eXi = 0 i Œ {1, 2…L}

(3)

j=1

The connectivity theorem allows the calculation of FCCs from in vitro obtained kinetic data of the enzymes in the pathway, in combination with suitable kinetic expressions for the individual enzymes in the pathway. Flux control coefficients can, however, also be determined by other, direct or indirect, methods [9]. 2.2 MCA of the Penicillin V Pathway in Penicillium Chrysogenum Biosynthesis pathways to secondary metabolites are often complex and in many cases only partially elucidated. As a consequence, kinetic parameters of enzymes involved in the pathways, as well as methods for analysing pathway intermediates, are often not available. However, as knowledge concerning these


143

pathways accumulates, the application of MCA is rendered more feasible. One of the most extensively studied antibiotic biosynthesis pathways is the one leading to penicillin V in Penicillium chrysogenum (Fig. 3), and this information has been used to perform MCA of this pathway. In the first model presented for this pathway [10] only the two first steps of the pathway were considered, i.e. the condensation of L-a- aminoadipic acid, Lcysteine and L-valine to LLD-ACV by ACVS and the conversion of LLD-ACV to isopenicillin N (IPN) by IPNS (Fig. 3). The last step, i.e. conversion of IPN to penicillin V catalysed by the enzyme AT, was not included in the model, since in vitro activity assays of AT revealed an order of magnitudes higher activity than required to support the in vivo flux leading to penicillin. Furthermore, the

Fig. 3 The biosynthesis pathway to penicillin V in P. chrysogenum. L-a-Aminoadipic acid, L-cysteine and L-valine are converted to d-(L-a-aminoadipyl)-L-cysteinyl-D-valine (LLDACV) by ACV synthetase (ACVS). LLD-ACV is further converted to isopenicillin N by isopenicillin N synthase (IPNS), and finally the a-aminoadipic acid side chain of IPN is replaced by phenoxyacetic acid through the action of acyl-CoA:isopenicillin N acyltransferase (AT). Formation of bisACV is a spontaneous reaction, and reduction back to LLD-ACV takes place via the thioredoxin-thioredoxin reductase system (TR)

144


intracellular concentration of IPN was lower than the Km value of AT for IPN, indicating that the enzyme was in excess. This led to the conclusion that this enzymatic step would not be flux controlling, a notion that was later supported by a model including the complete pathway [11]. The kinetics of ACVS and IPNS were assumed to be of Michaelis-Menten type with respect to the substrates, with non-competitive inhibition of ACVS by LLD-ACV and competitive inhibition of IPNS by glutathione. Moreover, the kinetics of IPNS was considered to be of first order with respect to the dissolved oxygen concentration. From the proposed kinetics for the two enzymes, the elasticity and flux control coefficients were calculated during fed-batch fermentation. The analysis revealed that there was a dramatic shift in flux control during the fed-batch fermentation, with the main flux control exerted by ACVS during the initial part of the fermentation and by IPNS during the final part of the fermentation. This was explained by accumulation of the intermediate LLD-ACV during the timecourse of the fermentation, which resulted in inhibition of the first enzyme, ACVS, and thereby a shift in flux control to the enzyme catalysing the conversion of the inhibitor. Due to the dramatic shift in the pathway a single flux controlling step could not be identified, but clearly it would be beneficial to increase the activity of IPNS in order to avoid LLD-ACV accumulation, and thereby feedback inhibition of ACVS. It was proposed that an increased activity of IPNS could be achieved by increasing the concentration of dissolved oxygen, since this enzyme is dependent on oxygen. The effect of dissolved oxygen concentration on the distribution of flux control was later investigated in continuous cultures of a high-yielding strain of P. chrysogenum [12]. It was found that IPNS exerted the largest part of the flux control at low dissolved oxygen concentrations, whereas at high dissolved oxygen concentrations the flux control was evenly distributed between ACVS and IPNS. Thus, an increased oxygen concentration could not completely shift the flux control to ACVS. The reason for this finding was shown by Theilgaard et al. [13] to be a consequence of a previously unknown inhibitory effect of bisLLD-ACV on ACVS. Formation of bis-LLD-ACV occurs via an oxygen-driven reaction where a sulfide bridge is formed between to molecules of LLD-ACV (Fig. 3). Although a specific thioredoxin system (a specific reductase) capable of converting bis-LLD-ACV back to LLD-ACV exists in P. chrysogenum [14], it is probable that an increased oxygen concentration leads to an increased level of bis-LLD-ACV and hereby strong inhibition of ACVS. This information was used to revise the kinetic model to comprise the oxidation of LLD-ACV to bisLLD-ACV, reduction of bis-LLD-ACV to LLD-ACV by the action of the thioredoxin system, and the non-competitive feedback inhibition of ACVS by LLDACV and bis-LLD-ACV [15]. From analysis of this model it was found that the main flux control of the biosynthesis pathway to penicillin V in P. chrysogenum is exerted by IPNS.


145

2.3 Overexpression of Biosynthesis Genes The genes involved in antibiotic biosynthesis, i.e. structural genes, transport genes, resistance genes and pathway-specific regulators, are often grouped together in a cluster with coordinated expression. One can imagine that a tight regulation, ensuring a coordinated activity of the individual enzymes in a secondary metabolic pathway, is beneficial for the cell in order to avoid accumulation of intermediates, particularly if these are toxic compounds. To avoid accumulation of intermediates, the enzymatic activities of the pathway need to be adjusted so that the product of the initial reaction is rapidly converted into succeeding intermediates and finally into the product. This requires that the capacity of the initial reaction is lower than the capacity of remaining enzymes in the pathway, which means that the committed step controls the flux through the pathway to a large extent. The committed step in penicillin synthesis is the condensation of valine, cysteine and L-a-aminoadipic acid to form LLD-ACV by the action of ACVS and in Aspergillus nidulans, overexpression of this enzyme resulted in a large increase in penicillin production [16]. However, the situation in a high-yielding strain that has been developed through strain-improvement programs based on random mutagenesis may be entirely different to that in low-yielding natural strains. During the many rounds of mutations, properties that lead to higher antibiotic yields have been selected for, and this most probably includes alleviation of flux control at the committed step and consequently a more even distribution of control at the various steps of the pathway. In accordance with this, MCA of the penicillin biosynthesis pathway in a high-yielding strain of P. chrysogenum indicated that the flux control was distributed between ACVS and IPNS, and that IPNS exerted the main control of flux through the pathway. In a study by Theilgaard et al. [17], a single gene copy strain of P. chrysogenum was transformed with the structural genes of the penicillin pathway and strains with improved penicillin productivity were analysed with respect to the ACVS and IPNS activities, intracellular concentration of LLD-ACV, and penicillin productivity. It was found that an increase in either IPNS or ACVS activity led to increased penicillin productivity, and that the largest effect was obtained when the activity of both enzymes was increased. Thus, even though MCA of the pathway indicated IPNS as a prime target for amplification, an even better result was achieved by the amplification of both activities. In accordance with this, overproducing strains of P. chrysogenum obtained by classical strain-improvement techniques contain amplifications of the complete gene cluster, rather than any single gene in the cluster [18–20]. The effect of overexpression of biosynthesis genes has also been investigated in the polyketide producer Streptomyces glaucescens. The carbon skeleton of polyketide antibiotics is assembled by the action of polyketide synthases (PKSs). After the assembly of the carbon backbone, the polyketide is typically modified by other enzymes, encoded by genes in the biosynthesis cluster, to

146


Fig. 4 Simplified pathway to cephalosporin C in A. chrysogenum. Abbreviations: L-a-aa, L-a-amino adipic acid; LLD-ACV, d-(L-a-aminoadipyl)-L-cysteinyl-D-valine; IPN, isopenicillin N; PenN, penicillin N; DAOC, deacetoxycephalosporin; DAC, deacetylcephalosporin; DAOCS, DAOC synthase; DACS, DAC synthase; DAC-AT, Acetyl CoA: DAC acetyltransferase

form the final product. In the biosynthesis of the aromatic polyketide tetracenomycin C (tcm C) by S. glaucescens, four PKSs are responsible for the synthesis of the initial polyketide intermediate [21]. In a study by Decker et al. [22], the genes encoding these enzymes were overexpressed and their influence on the production of tcm C and intermediates in the pathway to tcm C was examined. It was found that the introduction of a vector harbouring all four PKS genes led to an increased tcm C titre as compared to a control strain harbouring the vector without insert. Moreover, introducing the gene encoding the acyl-carrier protein (ACP) alone also resulted in increased tcm C titres. In both cases, the introduction of additional copies of PKS encoding genes resulted in pronounced accumulation of pathway intermediates, as well as increased tcm C titres. This indicates a shift of flux control towards later enzymatic steps in the pathway, and thus, it is likely that an even more increased tcm C productivity can be achieved through overexpression of all structural genes of the tcm C biosynthesis cluster. There are also examples of biosynthesis pathways where enzymatic steps late in the pathway exert a large degree of flux control on the pathway. In production of cephalosporin C by Acremonium chrysogenum, overexpression of cefEF (Fig. 4) resulted in 15% increased cephalosporin C production [23]. However, it was later found that the plasmid used for overexpression of these genes also contained the cefG gene, which encodes the acetyltransferase catalysing the last step of the pathway [24, 25]. Moreover, overexpression of cefG alone resulted in substantially increased cephalosporin C yields [25, 26].

3 Integrated Analysis Production of secondary metabolites normally occurs at a low level in naturally producing organisms, and the flux through the biosynthesis pathway is typi-


147

cally orders of magnitude lower than the fluxes through many of the pathways of the central carbon metabolism. Supply of precursors for secondary metabolite production is therefore not an issue in naturally producing strains, and even in strains with a substantially improved production level. However, as the productivity is continuously increased, the drain of precursors and cofactors from the central carbon metabolism may become an issue. Particularly in situations where building blocks are drained from anabolic reactions or specifically synthesized precursors are required for antibiotic synthesis, redirection of carbon fluxes in the central carbon metabolism may be necessary (this will be further discussed later in this chapter). In order to evaluate the influence of drain of precursors and cofactors for secondary metabolite production it is not sufficient to consider the individual pathways leading to the precursors or supplying cofactors for antibiotic synthesis, but the complete primary metabolic network needs to be taken into consideration. Metabolic flux analysis allows the quantification of intracellular reaction rates or fluxes, which is a term used to emphasize that these are rates of metabolic pathways rather than single reactions. The result of this analysis gives a snapshot of cellular metabolism. Such a snapshot may not be particularly useful on its own, but valuable information can be obtained by comparing the metabolic flux distribution for a specific situation with the distributions found for other situations. For example, analysis of flux distributions in high- and low yielding strains or in a certain strain subjected to different cultivation conditions may provide important clues to the link between the primary and secondary metabolism through the supply of precursors and cofactors for secondary metabolism. Metabolic flux analysis is also a useful tool in determining the flexibility of metabolic branch points and calculating the maximum theoretical product yields. Below we will briefly review the concepts of metabolic flux analysis and its applications for analysis of antibiotic-producing strains. A detailed description of the principles and applications of metabolic flux analysis is given by Stephanopoulus et al. [8]. 3.1 Metabolite Balancing Intracellular fluxes can be quantified by combining experimental measurements, e.g. substrate uptake rates and the secretion rates of metabolic products, with mass balances applied around intracellular metabolites, so-called metabolite balances. The mass balances are based on the stoichiometry of the intracellular reactions that are included in the model. Normally, only the major intracellular reactions are included, in order to keep the model as simple as possible. In a metabolic model with K metabolites and J reactions, the stoichiometry of the j-th intracellular reaction can be specified as K

Â gji Xmet, i = 0

i =1

(4)

148


Fig. 5a,b A metabolic model typically includes the main reactions of cellular metabolism. In the figure, some of the reactions around the glucose-6-phosphate (G6P) node are considered, i.e. the phosphorylation of glucose to form G6P, the further conversion of G6P to fructose-6-phosphate (F6P), conversion of G6P to the storage compound trehalose (TREHAL) and the oxidative branch of the PP-pathway, leading to ribulose-5-phosphate (RIBU5P) and carbon dioxide: a stoichiometry of reactions around G6P, defined according to Eq. (4); b the stoichiometric balances for all reactions can be written in matrix notation (Eq. 5). The J rows in the matrix G each represent one reaction, while each column represent one of the K metabolites. The stoichiometries of the reactions around the G6P branch point have been defined in the matrix, and the dots represent additional reactions and metabolites that may be included in the model

In Eq. (4), Xmet,i is the i-th pathway intermediate. The stoichiometric coefficient gji may be positive, negative or zero depending on if the metabolite referred to is formed, consumed or not participating in the reaction, respectively (see example in Fig. 5). The stoichiometry of all the J reactions in the metabolic model can be written in a compact form using matrix notation: GXmet = 0

(5)

The matrix G then contains the stoichiometric information of all reactions and metabolites included in the metabolic model. In flux analysis, the intracellular reaction rates are in focus. The rate of a chemical reaction is defined as the forward rate v, and the formation rate of a metabolite with the stoichiometric coefficient g is then gv. The net synthesis rate of the i-th metabolite is the sum of its formation rate in all J reactions:


149

J

rmet,i = Â gji n j

(6)

j=1

In the example in Fig. 5, the net synthesis rate of G6P is then rmet,G6P= v1–v2–2v3–v4. The net synthesis rate of all metabolites included in the model can be written using matrix notation as rmet = GTv

(7)

Inside the cell the net formation and consumption of metabolites will generally be balanced, i.e. the net formation of a given metabolite will equal its net consumption. This will hold for most conditions, except when the system is perturbed drastically. However, since the turnover of intracellular metabolites is high, the concentration of intracellular metabolites generally adjusts rapidly to new levels after environmental perturbations.A pseudo-steady state assumption [27] can therefore be made for the intracellular metabolite concentrations even after a drastic change in the environmental conditions. Balancing of all intracellular metabolites corresponds to a situation where there is no accumulation of these and their net formation rates are consequently zero: rmet = GTv = 0

(8)

A consequence of the steady state assumption is that the reactions in a linear pathway can be lumped into overall reactions, i.e. a linear pathway segment can be lumped into a single overall reaction (exemplified in Fig. 6). Accordingly, only balances over branch-point metabolites need to be included in Eq. (8). From this vector equation the pathway fluxes in the rate vector v can be determined. The equation represents K linear algebraic balances

Fig. 6 Glucose-6-phosphate is converted to ribulose-5-phosphate through a sequence of reactions in a linear pathway, i.e. the oxidative branch of the pentose phosphate pathway. During steady state, the rates of the reactions in a linear pathway are clearly identical to the flux through the pathway (r1=r2=r3=v4) and these reactions can be lumped into an overall reaction in the metabolic model. Abbreviations: G6P, glucose-6-phosphate; TREHAL, trehalose; F6P, fructose-6-phosphate; 6PGA, 6-phosphogluconate; RIBU5P, ribulose-5-phosphate

150


with J unknowns, and the degrees of freedom in this set of algebraic equations is therefore F=J–K. If F fluxes can be experimentally determined, e.g. by measuring uptake and secretion rates of substrates and products, a unique solution to the system can be obtained. In secondary metabolite producers it is often not possible to determine F fluxes experimentally, since they typically do not secrete metabolic products that are directly representative of primary metabolic fluxes. Therefore, additional constraints must be added to the model in order to determine the fluxes in the primary metabolism. These can be balances over cofactors, i.e. NADH, NADPH and FADH2, and over ATP. Alternatively, one may apply a concept of linear programming where an overall optimisation criterion, e.g. maximization of the specific growth rate, is used to find a solution [28–31]. Jørgensen et al. [32] constructed a stoichiometric model of P. chrysogenum including 61 internal fluxes and 49 intracellular metabolites. The model included the main primary metabolic pathways, i.e. the Embden Meyerhof Parnas (EMP) pathway, the pentose phosphate (PP) pathway and the tricarboxylic acid (TCA) cycle, as well as the biosynthesis pathway to penicillin V and the drain of central metabolites to biomass. Since a complex medium was used in the cultivations, the model also considered the uptake of 21 amino acids, lactate and g-aminobutyrate, in addition to glucose. In order to solve the system, balances over NADPH, NADH, FADH2 and ATP had to be included in the model. The degrees of freedom after including these cofactors were 33, and since exactly 33 fluxes could be measured experimentally, a unique solution could be obtained. The experimentally determined fluxes included uptake rates of glucose, lactate, g-aminobutyrate and 21 amino acids, formation rates of penicillin V and intermediates in the penicillin V pathway; and the formation rates of biomass components such as cellular RNA/DNA, protein, lipid, carbohydrate and amino carbohydrate. From the flux analysis it was found that there was a correlation between the flux towards penicillin biosynthesis and the flux through the PP pathway, and flux analysis hereby provided information about complex correlations between different parts of the metabolism that would have been difficult to identify by other means. The correlation between PP pathway flux and penicillin synthesis will be further discussed later. Stoichiometric models have also been developed for antibiotic producing actinomycetes, e.g. Streptomyces coelicolor and Streptomyces lividans. Daae and Ison [33] constructed a stoichiometric model for S. lividans, including 57 reactions and 53 balances over metabolites including ATP, NADH, NADPH and FADH2. The model was used for a theoretical analysis of the sensitivity of flux estimations towards perturbations in measured fluxes and biomass composition. Later, Rossa et al. [34] used a simplified version of this model, including 36 reactions and 46 metabolites, in the analysis of an actinorhodin and undecylprodigiosin producing S. lividans strain. Also here, interesting correlations between primary metabolic fluxes and antibiotic production were observed, and the information obtained in this way was subsequently used in the design of an improved producer strain (further discussed later in this chapter).


151

In the construction of a metabolic model like the ones described above, several difficulties may arise. A detailed knowledge of the primary metabolism is required in order to set up balances over cofactors, and therefore this is not possible in poorly characterized species. Even in the cases where the metabolism is well characterized, assumptions have to be made concerning the cofactor specificity of enzymes and the possible existence of transhydrogenases, i.e. enzymes capable of reversibly converting NADH to NADPH. Likewise, assumptions have to be made for the balance over ATP, e.g. concerning the P/O ratios for NADH and FADH2. In order to circumvent these problems it is necessary to identify other possible constraints on the fluxes in the system. One approach to this is to combine metabolite balancing with feeding labelled tracers to the cells and measuring the distribution of labelling in the different intracellular metabolites as described further in the following. 3.2 The Use of Labelled Substrates By combining metabolite balances with the use of 13C-labelled glucose and measurements of 13C-enrichment patterns in metabolites, a robust estimation of flux distribution can be achieved [35–37]. When [1-13C]-glucose is used as the carbon-source for growth, the 13C-labelling will be distributed in metabolites and cell constituents in a manner dependent on which metabolic pathways are active and to which extent they are active. Using this approach, it is possible to set up labelling balances over branch-point metabolites much in the same way as metabolite balances. However, whereas the output of metabolite balancing is absolute fluxes, the use of labelling balances enables estimation of relative fluxes.As is illustrated in Fig. 7, the combination of labelling and metabolite balancing may provide information that cannot be accessed through either of the approaches.

Fig. 7a,b The principle of: a metabolite; b labelling balancing (modified from Christensen and Nielsen [38]). Two reactions with fluxes x and y result in the same metabolite, which is produced with the flux z. If only the outgoing flux z is known (z=80), the fluxes x and y cannot be calculated using metabolite balancing. However, from the fractional labelling of the metabolites it can be deduced that the reactions are equally active, and by combining labelling and metabolite balancing their fluxes can be quantified (x=y=40)

152


The additional constraints provided by labelling balances eliminate the need for including cofactor and energy balances in the metabolic model and allows a more detailed analysis of the metabolic network. An iterative method combining labelling balancing and metabolite balancing can be used for estimations of the flux distribution in metabolic networks [39].Alternatively, labelling data may be used for estimation of individual fluxes of special interest. In particular, it is possible to estimate the flux through the PP pathway using a few GC-MS measurements and a simple algebraic expression [40]. Since the flux through the PP pathway is often important in antibiotic production due to its function as the main NADPH-generating pathway in the primary metabolism, a simple and accurate method for estimating this flux is very useful. Due to the low intracellular concentrations of central metabolites it is impractical to directly use these compounds for analysis of labelling patterns. However, since central metabolites are converted to amino acids through conserved biosynthesis pathways in which the carbon transitions are well known, the positional labelling of central metabolites may be obtained through labelling analysis of the amino acids. The amino acids in total cell protein can be made available for analysis by NMR or GC-MS by hydrolysis of the cell material [35, 41]. A consequence of the use of proteinogenic amino acids for analysis is that steady state cultivation is required for flux quantification through the 13C-tracer approach. However, 13C-labelling methods can be applied in batch cultivation for a qualitative assessment of flux distribution. Thus, the distribution of labelling in the amino acids in batch cultivation reflects the relative contribution of different pathways, although it is not possible to obtain the absolute value of the fluxes.A batch experiment could for example be used to give initial information concerning the relative activity of different pathways in a poorly characterized species. It is also possible to use labelling data from a batch cultivation to follow changes in the primary metabolism during the timecourse of the growth and production phases. In this case, changes in positional labelling of central metabolites will be reflected in the amino acid labelling patterns, but with a delayed effect. Quantitative changes in fluxes can therefore not be estimated during batch fermentation, but through measurement of the labelling patterns of amino acids it is possible to obtain qualitative assessment of the fluxes and perhaps even identify trends in the flux changes. This has been illustrated by analysis of nystatin production by Streptomyces noursei [42], where changes in central metabolic fluxes were observed during the timecourse of batch fermentation (see also later). 3.3 Identification of Pathways Metabolic flux analysis through metabolite balancing requires that the main reactions in the network are known. However, it is also possible to use a metabolite balancing for assessment of which pathways are likely to be active, i.e. by examining which metabolic models fit best to a given set of experimental data. A


153

more powerful approach for pathway identification is, however, the use of labelled substrates in combination with NMR or GC-MS analysis of intracellular metabolites, e.g. amino acids, as this approach provides direct information concerning which metabolic pathways are active in the cell. An illustrative example of this is the distinction between the EMP and Entner-Doudoroff (ED) pathways for glucose catabolism. Catabolism of [1–13C] glucose through the EMP pathway results in labelling in position 3 of pyruvate, while catabolism via the ED pathway leads to labelling in position 1 of pyruvate (Fig. 8), a difference which is easily detected in the labelling patterns of the pyruvate-derived amino acids alanine and valine. Recently, the ED pathway was unexpectedly identified as the main catabolic route in the glycopeptide producing actinomycete Nonomuraea ATCC 39727 using a 13C-labelling approach [43]. The presence of the pathway was later confirmed by identification, sequencing and expression analysis of the genes encoding the enzymes of the pathway. The degradation of compounds that are not commercially available in a 13Clabelled version can be investigated through a reciprocal labelling approach, where [U-13C6] is used as the carbon-source. Co-catabolism of the unlabelled compound can then be traced in the labelling patterns of central metabolites, allowing identification of the degradative pathway. This approach was used in the elucidation of adipate degradation in a 7-aminocephalosporanic acid (adipoyl-7-ADCA) producing P. chrysogenum strain [44]. Production of adipoyl-7-ADCA in this strain has been achieved by the introduction of an expandase gene from Streptomyces clavuligerus in combination with feeding of adipate [3]. Adipate is cleaved off adipoyl 7-ADCA in a post-fermentation ena)

b)

Fig. 8a,b Catabolism of [1-13C]-glucose via: a the EMP pathway; b the ED pathway lead to 13C-labelling in position 3 and 1 of pyruvate, respectively

154


zymatic process, yielding the desired product 7-aminocephalosporanic acid (7ADCA), and adipate can hereby be reused in the fermentation process (Fig. 9). Metabolism of adipate by P. chrysogenum is therefore undesirable in terms of overall process economy. In the study by Thykær et al. [44], degradation of unlabelled adipate was followed in P. chrysogenum during growth on [U-13C6]. It was found that unlabelled carbon made its way into intermediates of the TCA-cycle, supporting an earlier hypothesis that adipate is degraded to succinyl CoA and acetyl CoA by b-oxidation. Furthermore, degradation of unlabelled adipate did not affect the labelling of acetyl CoA either in the cytosol or in the mitochondria, and therefore it was likely that adipate metabolism took place in the microbodies, with further metabolism of acetyl CoA via the glyoxylate shunt resulting in formation of a C-4 compound that was subsequently transferred to the mitochondria. Presence of the key enzyme of the glyoxylate shunt was shown by an enzymatic assay, and flux analysis of chemostat cultivations with and without the addition

Fig. 9 Pathway to 7-ADCA in P. chrysogenum carrying the expandase gene of S. clavuligerus. The acyl chain of isopenicillin N is substituted by adipate, which is supplied in the growth medium, to form adipoyl-6-APA. This compound is further converted to adipoyl-7-ADCA through the action of the expandase, and to obtain the desired product 7-ADCA, adipate is enzymatically cleaved off adipoyl-7-ADCA in a post-fermentation process. The adipate thus released can be reused in the feed medium, allowing an economic production process. Abbreviations: 7-ADCA, 7-aminocephalosporanic acid, 6-APA, 6-aminopenicillinic acid


155

of adipate demonstrated that catabolism of adipate via this pathway could ensure net synthesis of C-4 compounds and hereby replace the anaplerotic flux from pyruvate to oxaloacetate via pyruvate carboxylase. 3.4 Node Flexibility In order to improve the overall yield in a given process it is obviously desirable to redirect the carbon flux towards the pathway leading to the product of interest, while reducing the flux through pathways leading to byproducts. The relative fluxes through the various branches of a metabolic network are dependent on the flux-split ratios at the metabolic branch points, and in order to modify these ratios, it is important to gain insight into the regulation of flux around the branch points. In the branched pathway in Fig. 10, the yield of the product P1 on the substrate S is a function of the fluxes through the pathways, i.e. YSP1=JP1/JS. A straightforward approach to increase the yield of P1 would be to enhance the enzyme activities in the pathway leading from the branch point to P1, e.g. by overexpression of the structural genes encoding these enzymes. However, the effectiveness of such a change depends on the regulation of the flux distribution at the branch-point, i.e. the flexibility of the node [45]. In a flexible node, the flux-split ratio will change rapidly in response to the need of the cell, and the competing enzymes around the node typically have similar substrate affinities and reaction velocities. Thus, if the branch point in the above scheme is flexible, an increase in the enzyme activities in the pathway leading to P1 will serve its purpose and increase the flux through this pathway in relation to the flux leading to P2. In contrast, in a rigid node the competing enzymes are tightly controlled by feedback regulation or enzyme transactivation by metabolites from the competing branch. In this case, it may not be possible to alter the fluxsplit ratio by changing the enzyme activities around the node. It may not be straightforward to spot which branch points in a metabolic network are crucial to product and byproduct formation. These crucial branch points, i.e. the principal nodes, can be identified experimentally by systematic variation of the product yield and analysis of the flux-split ratios at various nodes. Once the principal nodes have been identified, their flexibility can be as-

Fig. 10 Branched pathway, leading from the substrate S to an intermediate (I), which can be further converted to two different products (P1 and P2). The yield of the respective products is determined by the flux-split ratio at the branch point

156


sessed by introducing perturbations in the metabolic network and analysing the effects of these on the flux partitioning at the principal nodes [46, 47]. In a study by van Gulik et al. [48], the rigidity of the principal nodes for penicillin production in a high-yielding strain of P. chrysogenum was investigated. The principal nodes were identified through comparison of the flux distributions in the primary metabolic network during different penicillin production rates, i.e. during a fermentation where no production occurred and during high-productivity fermentations. Since significant flux partitioning changes occurred at the glucose-6-phosphate, 3-phosphoglycerate, mitochondrial pyruvate and mitochondrial isocitrate nodes when penicillin productivity increased, these were assigned as the principal nodes. The flux partitioning changes at the 3-phosphoglycerate and mitochondrial pyruvate nodes were explained by the increased synthesis of cysteine and valine, respectively, for penicillin biosynthesis.Accordingly, the changes at the glucose-6-phosphate and mitochondrial isocitrate nodes were considered to be due to the demand for NADPH in the synthesis of these precursors. In order to assess the rigidity of the principal nodes, perturbations of the flux distribution were introduced by growing the cells on alternative carbon sources, i.e. ethanol and acetate. Surprisingly, the use of ethanol or acetate as the carbon source in chemostat cultivations of P. chrysogenum resulted in approximately as high specific penicillin productivities as was the case when glucose was the carbon source.Accordingly, the fluxes towards cysteine and valine synthesis, as well as the flux through isocitrate dehydrogenase, were largely unchanged during growth on the three different carbon-sources. The flux through the oxidative branch of the PP pathway was approximately constant whether the cells were using glucose or acetate as the carbon-source. During growth on ethanol this flux was largely decreased, as NADPH was amply supplied through the action of acetaldehyde dehydrogenase, and the function of the PP pathway therefore became mainly anabolic. In conclusion, the principal nodes in penicillin production appeared to be highly flexible and thus, flux partitioning at these nodes was not expected to present a problem in a potential further increase of penicillin yield. 3.5 Maximum Theoretical Yield Using a stoichiometric model for P. chrysogenum, Jørgensen et al. [32] calculated the maximum theoretical yield of penicillin to 0.43 mol penicillin/mol glucose. At conditions of maximum yield, no cell growth occurs and all citrate is drained from the TCA cycle to form L-a-aminoadipic acid for penicillin synthesis. Moreover, the upper part of glycolysis and the PP pathway must operate in a cyclic mode in order to supply sufficient NADPH for penicillin synthesis. In filamentous fungi, synthesis of cysteine may occur either through a transsulfuration pathway where homocysteine reacts with serine to form cystathionine, which subsequently is converted to cysteine and a-ketobutyrate, or


157

through direct sulfhydrylation of serine. At the time of the study by Jørgensen et al. [32], it was believed that P. chrysogenum only used the transsulfuration pathway and this pathway was therefore used in the stoichiometric model. However, the direct sulfhydrylation pathway to cysteine is energetically more favourable than the transsulfuration pathway [49], and when this pathway was included in the model the maximal theoretical yield could be increased to 0.50 mol/mol. Thus, it was suggested that the introduction of such a pathway in P. chrysogenum could be an approach to increased penicillin yield. In a later study it was, however, shown that the direct sulfhydrylation pathway is present in P. chrysogenum along with the transsulfuration pathway [50], but so far the relative activities of the two pathways have not been resolved.

4 Linking the Primary and Secondary Metabolism Microbial production of secondary metabolites typically takes place through complex pathways, involving specialized enzymes that play no role in the central carbon metabolism of the cell. Nevertheless, secondary metabolism is intrinsically linked to the central part of the metabolism via the supply of precursors and cofactors. As mentioned earlier, the amount of precursors and cofactors required for antibiotic synthesis is usually sufficiently low to be easily accommodated by the central carbon metabolism of the cell. However, in high-yielding producer strains the requirements for precursors and cofactors eventually become limiting for the antibiotic yield. Another issue is the source of precursors for secondary metabolism, i.e. the possible role of specifically synthesized precursors in the overall flux control towards secondary metabolites. In the following sections, we will discuss interactions between the primary and secondary metabolites for two classes of antibiotics that are produced at high yields, i.e. polyketide and b-lactam antibiotics, and antibiotics that require specifically synthesized precursors, i.e. glycopeptide and polypeptide antibiotics. The precursor and cofactor requirements for synthesis of these compounds are summarized in Table 1. 4.1 Precursor Supply When considering the supply of precursors for antibiotic synthesis, it is informative to examine the source of these precursors in the microbial metabolism. Thus, precursors for secondary metabolism can be central metabolites, but they may also be cellular building blocks that are drained from the anabolic reactions of the cell or specifically synthesized precursors. These situations are depicted in a simplified manner in Fig. 11. The ability of a secondary metabolic pathway to drain metabolites from the central carbon metabolism depends on the intracellular concentration of

Anthracyclines Tetracyclines Actinorhodin

Erythromycin Rapamycin Ascomycin Rifamycins Nystatin

Vancomycin Teichoplanin Dalbavancin

Polymyxin

Aromatic polyketides (type II PKS)

Complex polyketides (type I PKS)

Glycopeptides

Polypeptides Oxaloactetate Pyruvate Acetyl CoA PEP E4P

PEP E4P Acetyl CoA Othersa

Acetyl CoA Buturyl CoA Propionyl CoA Methylmalonyl CoA

2,4-Diaminobutyric acid Phenylalanine Leucine Other amino acids Lipidsb

Tyrosine b-Hydroxytyrosine p-Hydroxyphenylglycine 3,5-Dihydroxyphenylglycine Other amino acids Sugarsa Lipidsb

NADPH

NADPH

L-a-Aminoadipic acid Valine Cysteine

NADPH

NADPH

NADPH

Cofactor requirements

Precursors (intermediate metabolism)

Glycopeptide antibiotics include sugar residues, which typically differ from one glycopeptide to another; these sugars may be natural or specifically synthesized. b Some glycopeptide and polypeptide antibiotics include a fatty acid chain. Abbreviations: PEP, phosphoenolpyruvate; E4P, erythrose-4-phosphate.

a

Acetyl CoA a-Ketoglutarate Pyruvate 3-Phosphoglycerate

Penicillins Cephalosporins

b-Lactams

Acetyl CoA Malonyl CoA

Precursors (central metabolism)

Antibiotics of this type

Type of antibiotic

Table 1 Summary of precursor and cofactor requirements in b-lactam, polyketide, glycopeptide and polypeptide antibiotics

158 N. Gunnarsson et al.


159

Fig. 11a–c Overview of the drain of precursors in the production of different types of antibiotics: a the antibiotic precursors are central metabolites; b the antibiotic is synthesized from cellular building blocks such as amino acids and amino acid precursors; c the antibiotic is synthesized from specifically synthesized precursors

the intermediate and the kinetics of the first enzyme in the pathway leading to the secondary metabolite. Since the flux through the central carbon metabolism, in the final end leading to CO2 and ATP, is typically much higher than the fluxes leading to biomass and antibiotic formation, a small increase in the flux towards the antibiotic will not alter the intracellular pool of the central metabolite to any great extent. It is only when the flux towards the antibiotic is substantially increased that it will significantly influence the intracellular pool of the central metabolite, and thereby the ability of the first enzyme of the pathway to convert the precursor. Therefore, the supply of precursors is generally not a problem when these are drained directly from the central carbon metabolism (i.e. Fig. 11a). In this case an increased flux towards the secondary metabolite may often be obtained simply by increasing the expression of the biosynthesis genes, e.g. by the manipulation of pathwayspecific regulators. When the precursors for antibiotic synthesis are cellular building blocks such as amino acids, carbon is drained from the anabolic routes rather than the central carbon metabolism (Fig. 11b). In such a situation, it may not be sufficient to increase expression of the biosynthesis genes, but also the activity of the anabolic pathway may need to be increased in order to obtain a high yield of the product. Examples of this will be discussed later in connection with penicillin biosynthesis.

160


Finally, the situation is generally more complex when there is a requirement for specific precursors that are synthesized by enzymes encoded by genes in the biosynthesis gene cluster (Fig. 11c). Possible implications of this type of complex biosynthesis pathways on the metabolic engineering strategy towards increasing the product yield will be discussed later. 4.1.1 Precursor Requirements in Polyketide Production The carbon skeletons of polyketide antibiotics are assembled from the coenzyme A esters of short chain fatty acids by the action of polyketide synthase (PKS) enzymes. After assembly by the PKS, other enzymes may modify the polyketide backbone by amination, hydroxylation, methylation, oxidation, reduction or attachment of sugars. This, together with the variations introduced by the PKS enzymes, makes the polyketides an extremely diverse group of antibiotics. Structurally, polyketide antibiotics can be divided into two classes: aromatic and complex [51]. Aromatic polyketides are in bacteria synthesized by type II polyketide synthases, i.e. mono- or bifunctional enzymes that operate in a multienzyme complex, which carries out the condensation and reduction steps in sequence [52]. The starter unit in aromatic polyketides is most often acetyl CoA, and the polyketide chain is elongated by addition of malonyl CoA units [51]. A high yield of the aromatic polyketide actinorhodin has been achieved in Streptomyces lividans through the introduction of multiple copies of the gene encoding the pathway-specific regulator act II-ORF4 [53]. The protein encoded by act II-ORF4 mediates expression of the actinorhodin biosynthesis genes in Streptomyces coelicolor [54], and heterologous expression of this activator in S. lividans resulted in expression of the otherwise inactive actinorhodin biosynthesis gene cluster of S. lividans. The high actinorhodin titres in fed-batch fermentations of S. lividans overexpressing act II-ORF4 demonstrate that precursor supply is not a problem in production of aromatic polyketides in the gram per litre range. However, a later study has shown that even higher actinorhodin productivities could be achieved by altering primary metabolic pathways of S. lividans [55]. This will be further discussed later. In contrast to the aromatic polyketides, complex polyketides like macrolides and polyethers are synthesised from various acyl-units such as acetyl CoA, buturyl CoA, propionyl CoA and methylmalonyl CoA. Complex polyketides are synthesized by the action of type I PKSs, which are multifunctional enzymes containing domains for the condensation and reduction steps and functional domains for b-carbonyl processing.While the precursors for aromatic polyketides, i.e. acetyl and malonyl CoA, are drained from glycolysis and fatty acid biosynthesis, the origin of the precursors for complex polyketides is less clear. Thus, propionyl CoA and 2-methylmalonyl CoA may originate from catabolism of odd-numbered fatty acids, reduction of acrylate, rearrangement of succinyl CoA and catabolism of methionine, threonine or valine. The latter two


161

processes are thought to be the main routes to 2-methylmalonyl CoA and propionyl CoA in polyketide synthesis [56]. Correlations have been observed between macrolide antibiotic production and the level of catabolic enzymes in the valine- and threonine degradation pathways [57, 58]. In order to study this correlation further, Tang et al. [59] disrupted the vdh gene, encoding valine dehydrogenase, of the macrolide antibiotic producers Streptomyces ambofaciens and Streptomyces fradiae. It was found that the inability to catabolize valine led to a large decrease in antibiotic production in a defined medium. Furthermore, antibiotic production could be restored either by supplying propionate in the growth medium or by reintroducing the vdh gene. 4.1.2 Precursor Requirements in b -Lactam Production The precursors in penicillin production by P. chrysogenum are the amino acids valine and cysteine, and an intermediate in the pathway to lysine, i.e. L-aaminoadipic acid (see Fig. 3). In a study by Jørgensen et al. [60] it was shown that addition of these precursors to the growth medium of a high-yielding strain of P. chrysogenum resulted in increased penicillin yields, demonstrating that precursor supply may limit penicillin production at high yields. In particular, the supply of L-a-aminoadipic acid appears to be important in penicillin production. During classical strain improvement of P. chrysogenum, properties that are beneficial for penicillin production have been promoted, including changes in the pathway to L-a-aminoadipic acid and lysine (Fig. 12).

Fig. 12 Simplified pathway to L-a-aminoadipic acid and lysine in P. chrysogenum. HCS, Homocitrate synthase, a-AAR, L-a-aminoadipic acid reductase

162


In this pathway, lysine exerts feedback inhibition on the first step of the pathway, i.e. the conversion of a-ketoglutarate to homocitrate by homocitrate synthase [61–63]. Therefore, penicillin production is normally inhibited by the presence of lysine [64–66]. However, in high penicillin-yielding strains this inhibition has been alleviated [67], implying that deregulation of the lysine pathway is important to achieve high yields of penicillin. Lysine also exerts feedback inhibition on the conversion of L-a-aminoadipic acid to d-adenyl-a-aminoadipate by L-a-aminoadipic acid reductase [68]. This inhibition is more pronounced in high-yielding strains of penicillin than in low-yielding strains [69]. Thus, high-yielding strains of P. chrysogenum, obtained by classical strain improvement, seem to have acquired mutations favouring a larger supply of L-aaminoadipic acid to the penicillin biosynthesis pathway. Another indication of the importance of L-a-aminoadipic acid supply in penicillin production is given by the properties of the first enzyme in the penicillin biosynthesis pathway,ACVS.ACVS is responsible for the condensation of L-a-aminoadipic acid, L-valine and L-cysteine to LLD-ACV (Fig. 3). In the characterization of the P. chrysogenum AVCS [13], it was found that the Km value for L-a-aminoadipic acid was almost an order of magnitude smaller than the Km values of ACVS from other species. It is possible that this high affinity for L-aaminoadipic acid, and thus the increased ability to drain L-a-aminoadipic acid from central metabolism, is part of the reason why P. chrysogenum has developed as a superior penicillin producer. Also in the cephamycin C producer Streptomyces clavuligerus, L-a-aminoadipic acid supply has proved to be important for the product yield. In the cephamycin C biosynthesis pathway, the initial step is the same as the initial step of the penicillin pathway in P. chrysogenum, i.e. the condensation of L-aaminoadipic acid, L-cysteine and L-valine by ACVS (Fig. 13). Lysine and L-a-aminoadipic acid are, however, synthesized by different pathways in S. clavuligerus and P. chrysogenum. In S. clavuligerus, lysine is formed from aspartate by a pathway that does not involve L-a-aminoadipic acid as an intermediate, and L-a-aminoadipic acid is formed from lysine and a-ketoglutarate by the action of lysine 6-amino transferase (LAT). An increased cephamycin C yield could be achieved by supplying lysine to the growth medium of S. clavuligerus [70], indicating that the supply of L-a-aminoadipic acid may be a limiting factor for cephamycin C biosynthesis. Moreover, the introduction of an additional copy of the LAT gene has resulted in an increased cephamycin C yield [71, 72]. In a study by Khetan et al. [73] LAT was shown to have high apparent Km values for lysine and a-ketoglutarate in relation to the intracellular concentrations of these compounds, which led to the suggestion that increased cephamycin C biosynthesis might be achieved by increasing the intracellular concentrations of these precursors by metabolic engineering of the primary metabolism. When precursors are drained from the anabolic pathways, additional demand is put on pathways that normally do not carry high fluxes. This is reflected in the precursor requirements for penicillin and biomass synthesis, re-


163

Fig. 13 Simplified pathway to cephamycin C in S. clavuligerus. LAT, lysine 6-amino transferase, LLD-ACV, d-(L-a-aminoadipyl)-L-cysteinyl-D-valine, ACVS, ACV synthetase

spectively, in P. chrysogenum (Table 2). In a high-yielding strain of P. chrysogenum, the valine, cysteine and L-a-aminoadipic acid requirements for antibiotic synthesis are 4, 20 and 6 times higher, respectively, than the requirements for biomass synthesis (these values would be even higher for current production strains). In contrast, the central carbon metabolite precursor requirements for antibiotic synthesis are in the range of the requirements for biomass synthesis. Thus, production of penicillin at high yields does not drain remarkably high amounts of metabolites from the central metabolism, but rather requires channelling of these metabolites into the pathways leading to precursors of the biosynthesis pathway. 4.1.3 The Role of Synthesis of Specific Precursors Many antibiotics are synthesized from specialized precursors that are not part of the normal cellular metabolism (Table 3). Especially non-ribosomally synthesized peptide antibiotics, e.g. glycopeptides and polypeptides, are known to contain non-proteinogenic amino acids, D-amino acids, hydroxy acids and other unusual building blocks. Enzymes that are encoded by genes in the biosynthesis cluster perform the synthesis of these specialized precursors. D-Amino acids are common both in glycopeptides and polypeptides. They are formed from the corresponding L-amino acids by amino acid racemases. Racemization takes place after the activation of the amino acid on an enzyme-

164


Table 2 Specific precursor requirements for production of penicillin at high yield and corresponding precursor requirements for biomass formation by P. chrysogenum

Precursors

Precursor requirements for producta mmol (g dw)–1

Precursor requirements for biomassb mmol (g dw)–1

Precursor requirements of products in relation to biomassc mmol (mmol)–1

Building blocks

Valine 1.4 Cysteine 1.4 L-a-Aminoadipic acid 1.4

0.34 0.07 0.22

4.1 20 6.4

Central metabolite precursors

3PGA Pyruvate a-Ketoglutarate Acetyl CoA

0.88 1.9 1.2 2.3

1.6 1.4 1.2 0.6

1.4 2.7 1.4 1.4

a

The specific precursor requirements are based on a yield of 480 mg penicillin (g dw)–1 [32]. Precursor requirements for biomass formation by P. chrysogenum are taken from Nielsen [49]. c The specific amount of precursor needed for penicillin production divided by the specific precursor requirement for biomass synthesis. Abbreviations: 3PGA, 3-phosphoglycerate. b

Table 3 Examples of specifically synthesized precursors in various antibiotics

Antibiotic

Specific precursors

Glycopeptide antibiotics

p-Hydroxyphenylglycine 3,5-Dihydroxyphenylglycine b-Hydroxytyrosine Vancosamine 2,4-Diaminobutyric acid p-Hydroxyphenylglycine p-Hydroxyphenylglycine

Polymyxin Ramoplanin Nocardicin

bound thioester-linked intermediate as illustrated in Fig. 14 for phenylalanine in the cyclic decapeptide gramicidin S [74–76]. A high content of the non-proteinogenic amino acid 2,4-diaminobutyric acid (DAB) is characteristic for the polypeptide antibiotics of the polymyxin group (Fig. 15). Addition of aspartate and DAB had stimulatory effects on production of polymyxin E by Bacillus polymyxa, indicating that DAB constitutes a rate-controlling factor in polymyxin biosynthesis [77]. DAB is synthesized from L-aspartic acid, possibly via aspartyl 4-phosphate and aspartate 4-semialdehyde [78]. In addition to serving as a precursor, DAB


165

Fig. 14 Racemization of phenylalanine during the formation of gramicidin S. The initiation module PheATE of the gramicidin S synthetase has three domains. The adenylation (A) domain is responsible for recognition and ATP dependent activation of L- phenylalanine. Next, the amino acyl group is transferred from the L-Phe-AMP intermediate to the phosphopantetheinyl arm of the thiolation (T) domain and after that epimerized by the epimerization (E) domain. The module involved in the downstream chain elongation has a clear preference for the D-isomer.Abbreviations: L/D-Phe-AMP, phenylalanyl-adenosine-5¢-monophosphate diester; L/D-Phe-S-Ppant-T, phenylalanyl-S-phosphopantetheine- T domain acyl thioester enzyme covalent adduct

Fig. 15 Structure of polymyxin B and E. The amino acid in position 6 may be D-phenylalanine (polymyxin B) or D-leucine (polymyxin E), and the fatty acid may be either 6-methyloctanic acid (MOA) or 6-methylheptanoic acid (IOA). For other polymyxins the amino acids in positions 3, 7 and 10 may also differ

seems to have a wider regulatory effect. The presence of DAB was found to stimulate channelling of L-amino acids into polymyxin E while the incorporation into cell protein was repressed. The antibiotics of the polymyxin group also contain a branched short-chain fatty acid. The fatty acid is typically 6-methyloctanic acid (MOA) or isooctanic acid (IOA) synthesized from L-isoleucine and L-valine, respectively [79]. Consequently, a majority of the building blocks in polymyxin synthesis are derived from the amino acids of either the pyruvate or aspartate series (Fig. 16). A strategy for strain optimisation aiming at improved polymyxin production could be overexpression of key enzymes in the pathways leading to L-valine, L-leucine and to L-amino acids in the aspartate group. However, the physiological effects of such an approach are dependent on the interplay between the enzymes, and therefore it is important to find a properly

166


Fig. 16 Simplified view of synthesis of building blocks for polymyxin. 1) PEP carboxylase, 2) pyruvate carboxylase 3) aspartate aminotransferase, 4) aspartase, 5) aspartate kinase.Abbreviations: PEP, phosphoenolpyruvate; IOA, isooctanic acid; MOA, 6-methyloctanic acid

balanced expression. Alleviation of feedback inhibition on aspartate kinase may also result in improved polymyxin production. Synthesis of specific precursors also plays an important role in the biosynthesis of glycopeptide antibiotics. The glycopeptide antibiotics vancomycin and teichoplanin are currently used as the antibiotics of last resort against infections of multi-resistant gram-positive bacteria. Biosynthesis of these antibiotics includes synthesis of specific precursors, assembly of the heptapeptide backbone through the action of non-ribosomal peptide synthases (NRPS), crosslinking of the heptapeptide, glycosylation and halogenation.A partial understanding of this complex pathway has been achieved, primarily by the identification and analysis of the biosynthesis gene clusters of chloroeremomycin [80] and balhimycin [81], both compounds differing from vancomycin only in the glycosylation patterns (Table 4). The non-proteinogenic amino acids p-hydroxyphenylglycine, 3,5-dihydroxyphenylglycine and b-hydroxytyrosine are common to the vancomycinand teichoplanin-type antibiotics. Feeding experiments have shown that p-hydroxyphenylglycine and b-hydroxytyrosine are derived from tyrosine, while 3,5-dihydroxyphenylglycine is derived from acetate in both the vancomycin [83] and the teichoplanin-producing organisms [84]. During the sequencing of the chloroeremomycin and balhimycin gene clusters, several genes encoding enzymes with a putative function in the synthesis of non-proteinogenic amino acids were identified and recently, the biosynthesis pathway to p-hydroxyphenylglycine was elucidated by examination of the function of these enzymes [85, 86]. It was verified that tyrosine is the main precursor in p-hydroxyphenylglycine synthesis, which occurs by a cyclic route in which tyrosine is

b

a

D-Glucose Vancosamine

D-Glucose Epivancosamine

N-Methyl-D-leucine L-Asparagine D-Glucose Dehydrovancosamine

Acyl chain

D-Glucosamine

D-Glucosamine

Acyl chain

D-Mannose

D-Mannose

b-Hydroxytyrosine Tyrosine

D-Chloro-b-hydroxytyrosine

L-Chloro-b-hydroxytyrosine

D-p-Hydroxyphenylglycine

A40926/ Dalbavancinb Nonomuraea sp.

L-3,5-Dihydroxyphenylglycine

Teichoplaninb Actinoplanes teichomyceticus

D-p-Hydroxyphenylglycine

Balhimycina Amycolatopsis mediterranei

L-3,5-Dihydroxyphenylglycine

Chloroeremomycina Amycolatopsis orientalis

Vancomycin, chloroeremomycin and balhimycin share the same heptapeptide backbone and only differ in the glycosylation patterns. Teichoplanin and A40926 share the same heptapeptide backbone, but differ in glycosylation and halogenation patterns. Dalbavancin is a semi-synthetic derivative of A40926 [82].

Other

Sugars

Amino acids

Vancomycina Amycolatopsis orientalis

Table 4 Building blocks of some glycopeptide antibiotics

Control of Fluxes Towards Antibiotics and the Role of Primary Metabolism 167

168


Fig. 17 Biosynthesis pathway to p-hydroxyphenylglycine [85]. a – tyrosine transaminase; b – prephenate dehydrogenase; c – 4-hydroxymandelate synthase (HmaS); d – 4-hydroxymandelate oxidase (Hmo); e – p-hydroxyphenylglycine transaminase (HpgT)

used as an amino-donor, yielding p-hydroxyphenylpyruvate, which is an intermediate of the pathway (Fig. 17). It is not clear whether tyrosine or prephenate is used as the initial substrate of the pathway.A gene encoding a putative prephenate dehydrogenase has been identified in the chloroeremomycin gene cluster, which has led to the hypothesis that prephenate is the initial substrate [85]. However, conversion of tyrosine to p-hydroxyphenylglycine through the action of tyrosine transaminase is the first step of the catabolic route of tyrosine degradation in many microorganisms [87], and it is possible that this route contributes to the pathway during certain conditions. Regardless of this, it is clear that tyrosine is the main precursor for p-hydroxyphenyl-glycine synthesis due to the cyclic mode of the pathway. The synthesis of 3,5-dihydroxyphenylglycine (Fig. 18) has recently been shown to occur via a type III polyketide synthase, using malonyl CoA as the starter and elongation units [88, 89]. The final step of the pathway is the amination of 3,5-dihydroxyphenolic acid to form 3,5-dihydroxyphenylglycine, in which tyrosine again acts as the amino donor [88, 90]. Thus, similar to p-hydroxyphenylglycine formation, the formation of 3,5-dihydroxyphenyl-glycine requires tyrosine as amino donor.


169

Fig. 18 Biosynthesis pathway to 3,5-dihydroxyphenylglycine [88, 89]. a – reactions performed by type III PKS/chalcone synthase like protein; b – post-PKS modifications; c – transamination by Pgat (balhimycin)/HpgT (chloroeremomycin)

As a consequence of the tyrosine requirements in synthesis of these amino acids, it can be imagined that the supply of tyrosine may be rate controlling during high-yield glycopeptide production. In the production of teichoplanin at a low yield, tyrosine requirements are one third of the requirements for biomass formation (Table 5) and in relation to the drain of valine, cysteine and L-a-aminoadipic acid in high-yield penicillin production (Table 2), this represents a small flux through the anabolic reactions leading to the precursor amino acid. However, the ability of P. chrysogenum to channel large fractions of carbon through the reactions leading to valine, cysteine and L-a-aminoadipic acid is most likely a result of strain development through random mutagenesis. In an attempt to over-produce teichoplanin one may rapidly run into a limitation of tyrosine supply, and increasing the flux through the tyrosine synthetic pathway should therefore be a target for rational design of glycopeptide over-producing strains. An approach towards increased tyrosine supply could be modulation of the flux-split ratios at branch points in the biosynthesis route to aromatic amino acids (Fig. 19), i.e. by overexpressing some enzyme activities (chorismate mutase, arogenate transaminase and arogenate dehydrogenase) or by decreasing the activities of some enzymes (anthranilate synthase and prephenate dehydratase) around the chorismate and prephenate branch points. The success of

170


Table 5 Specific precursor requirements for teichoplanin production and biomass formation by Actinoplanes teichomyceticus. Only synthesis of the heptapeptide backbone is included in the table

Precursors

Precursor requirements for producta mmol (g dw)–1

Precursor requirements for biomassb mmol (g dw)–1

Precursor requirements of products in relation to biomassc mmol (mmol)–1

Building blocks Tyrosine

0.04

0.12

0.33

Central metabolite precursors

0.04 0.04 0.02

0.52 0.36 3.75

0.08 0.11 0.005

PEP E4E Acetyl CoA

a

The precursor requirements are based on a yield of 10 mg teichoplanin (g dw)–1 [91]. Precursor requirements for biomass formation are based on the biomass composition of E. coli [92]. c The specific amount of precursor needed for teichoplanin production divided by the specific precursor requirement for biomass synthesis. Abbreviations: PEP, phosphoenolpyruvate, E4P, erythrose-4-phosphate b

such an approach depends on the regulation of the enzymes around the branch points. Since chorismate and tyrosine exert feedback inhibition on enzymatic steps in the pathway, it is not probable that increased intracellular concentrations of these compounds can be achieved. However, an increased flux towards tyrosine and antibiotic is possible through this approach if the capacity of the antibiotic synthetic pathway is high enough to keep the concentrations of these intermediates at a low level. Another strategy for increased supply of tyrosine for glycopeptide biosynthesis could be alleviation of the feedback inhibition on DAHP synthase and arogenate dehydrogenase (Fig. 19). The use of strains expressing feedback-resistant DAHP synthases is a common approach to increase the carbon flow to chorismate in microbial production of aromatic amino acids [94, 95]. Traditionally, random mutagenesis and screening have been applied to obtain these mutants. However, mutational changes resulting in tyrosine-feedback resistance have been described [96–99] and recently, desensitized DAHP synthases were obtained by site-directed mutagenesis [99]. Also in production of nonaromatic amino acids, reduction of the susceptibility of key enzymes to feedback inhibition has been shown to have a major impact. Thus, lysine production in a wild-type strain could be improved from 0 g/l to 25 g/l by introduction of a plasmid carrying a feedback-resistant aspartate kinase, obtained by site-directed mutagenesis [100]. Accordingly, alleviation of the feedback control in anabolic pathways may be a relevant approach for yield improvement not only in glycopeptide producing strains, but also in polypeptide and b-lactam producers.


171

Fig. 19 Simplified view of the biosynthesis route to aromatic amino acids in actinomycetes. The properties of the enzymes in the pathway may vary from species to species, but the typical feedback inhibition patterns of streptomycetes are included in the figure [93]. 1) DAHP synthase, 2) anthranilate synthase, 3) chorismate mutase, 4) prephenate dehydratase, 5) arogenate transaminase, 6) arogenate dehydrogenase. Abbreviations: PEP, phosphoenolpyruvate, E4P, erythrose-4-phosphate, DAHP, 3-deoxy-D-arabino heptulosonate-7-phosphate

4.2 Cofactor Supply Secondary metabolism is often connected to the primary metabolism via the drain of co-factors needed for biosynthesis, as well as through the drain of precursors (Table 1). Specifically, polyketide and b-lactam synthesis requires a relatively large amount of reducing power in the form of NADPH. In the central carbon metabolism, NADPH is regenerated from NADP+ mainly through the oxidative branch of the pentose-phosphate (PP) pathway. In this pathway, two NADPH units are generated from each glucose-6-phosphate by the action of glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase. However, in some streptomyces species, 6-phosphogluconate dehydrogenase uses NAD+ instead of NADP+ as the cofactor [101]. NADPH is also regenerated by isocitrate dehydrogenase in the TCA cycle. The largest amount of NADPH is typically formed in the PP pathway, and one could therefore imagine that extensive production of secondary metabolites such as b-lactams and polyketides, would require an increased flux through this pathway.

172


In penicillin production, NADPH is required for synthesis of the precursors cysteine and valine. In addition, reduction of bis-ACV by the specific thioredoxin system requires NADPH, as discussed earlier. Synthesis of cysteine takes place in the cytosol, while valine synthesis occurs in the mitochondrial compartment, and as NADPH cofactors most likely cannot be transferred across the mitochondrial barrier, the NADPH needed in these reactions must be generated in the corresponding cellular compartment. Cysteine synthesis requires 5 units of NADPH per molecule of cysteine, while valine synthesis requires 2 NADPH units in total. However, one of the NADPH units in valine synthesis is the one used in synthesis of glutamate, the amino-group donor in the transamination reaction where valine is formed from a-ketoisovalerate. Glutamate is synthesized in the cytosol and thus, only one mitochondrial NADPH unit is required in the synthesis of one molecule of penicillin, while the rest of the required NADPH is of cytosolic origin. Cytosolic NADPH is mainly generated through the PP pathway, while the main NADPH generating reaction in the mitochondria is the reduction of isocitrate to a-ketoglutarate in the TCA-cycle. During flux analysis of P. chrysogenum growing in fed-batch and continuous culture, a correlation was found between the flux-split ratio at the glucose-6phosphate node and the yield of penicillin on glucose [32, 102]. The results indicated that an increased flux through the PP pathway was required to meet the NADPH demands of penicillin biosynthesis. Later a study of the central carbon metabolism of a high- and a low-yielding strain of P. chrysogenum revealed that the high-yielding strain exhibited a slightly higher flux through the PP pathway than the low-yielding strain [103]. The growth conditions for the highyielding strain were chosen so that the cells produced penicillin in one of the fermentations, while in a second fermentation, no penicillin was produced. It was shown that the flux through the PP pathway was the same in the two fermentations, and thus, the increased flux was considered to be a feature of the high-yielding strain rather than an effect of penicillin biosynthesis. It is possible that a high PP pathway flux is a feature that has been acquired during classical strain-improvement programs in order to accommodate an increased flux through the penicillin biosynthesis pathway. In contrast to the above observation, van Gulik et al. [48] found a strong correlation between PP pathway flux and penicillin production in a high-yielding P. chrysogenum strain, grown in chemostat at conditions of either high penicillin productivity or no penicillin formation. Moreover, when the strain was grown on carbon- and nitrogen sources that result in reduced NADPH formation, i.e. xylose and NO3 in place of glucose and NH4, decreased specific penicillin productivity was observed. Taken together, these studies point to the availability of cytosolic NADPH as a critical factor in penicillin production. However, it is interesting to compare the estimations of PP pathway flux in the different studies, since varying approaches were used for this task. In the studies by Jørgensen et al. [32], Henriksen et al. [102] and Gulik et al. [48], where a strong correlation between PP pathway flux and penicillin productivity was observed, the flux estimates through the PP pathway were based on metabolite balances where the balance


173

over NADPH plays an important role in the flux estimation. In contrast, the flux analysis in the study by Christensen et al. [103] was based on metabolite balancing in combination with 13C-labelling balancing and no NADPH balance was needed to estimate the PP pathway flux. NADPH is involved in a large number of reactions in the cell, and it may therefore be difficult to account for all NADPH formed and consumed using metabolite balancing. In addition, the occurrence of futile cycles such as the ACVoxidation/bisACV reduction previously mentioned results in NADPH consumption that cannot be accounted for by metabolite balancing. As the mechanism of polyketide biosynthesis resembles the NADPH-dependent condensation/reduction cycle in fatty acid synthesis, it has been proposed that polyketide synthases also use this cofactor as the reducing power. Indeed, polyketide biosynthesis in cell-free extracts of several Streptomyces species has been shown to require NADPH [104, 105], and it is therefore generally considered that the enzymes carrying out the reduction steps in polyketide biosynthesis require NADPH as the cofactor. Biosynthesis of the polyketide antibiotic avermectin has been reported to correlate with an increased activity of the PP pathway in Streptomyces avermitilis [106]. Another case where antibiotic synthesis has been found to correlate with an increased flux through the PP pathway is the production of the cyclopentanone antibiotic methylenomycin by Streptomyces coelicolor A3(2) [107]. The authors of the study suggested that the observed increase in PP pathway flux was related to NADPH usage in the antibiotic synthesis. During actinorhodin production in Streptomyces lividans, the correlation between PP pathway flux and antibiotic production is different from the cases mentioned above.As discussed earlier, S. lividans can be forced to overproduce actinorhodin if transformed with a plasmid carrying multiple copies of the pathway-specific activator actII-ORF4, and very high yields of actinorhodin can be obtained from the recombinant strain when cultivated in fed-batch fermentation [53]. Rossa et al. [34] performed flux analysis on S. lividans strains overproducing either actinorhodin or undecylprogidin, and found that the production of both of these antibiotics were inversely related to the flux through the PP-pathway. In a theoretical analysis of actinorhodin biosynthesis based on elemental- and redox balancing in combination with analysis of the biosynthesis steps of the pathway, Bruheim et al. [108] estimated the cofactor requirements to be six units of NADPH per actinorhodin formed. These requirements were greatly exceeded by the amount of NADPH formed in the PP pathway and TCA cycle, as quantified by flux analysis during the production phase. Thus, there appeared to be no co-factor limitation in actinorhodin production by S. lividans. This notion was verified in a study where S. lividans mutants with a partly or completely disabled PP pathway were analysed for actinorhodin productivity [55]. It was found that strains with 50% reduced PP pathway activity produced substantially higher titres of actinorhodin and undecylprogidin than the strain with intact PP pathway activity. A reduced PP pathway activity, thus, did not lower the NADPH supply to an extent that would

174


limit actinorhodin biosynthesis. In contrast, complete disabling of the PP pathway led to decreased actinorhodin production, indicating that some of the NADPH produced by this pathway is needed for antibiotic biosynthesis.An explanation for the increased productivity in the strains with decreased PP pathway activities may be the reduced loss of carbon in the form of CO2, potentially leading to increased availability of the intermediates that serve as precursors for polyketide synthesis. A similar correlation between polyketide synthesis and PP pathway flux was found in a study by Jonsbu et al. [42], where the production of nystatin by Streptomyces noursei was shown to correlate with a decreased PP pathway flux. In the referred study, central carbon fluxes were analysed during batch fermentation of S. noursei, and it was found that the flux through the PP pathway decreased, while the flux through the TCA cycle increased, during the polyketide-producing phase of the fermentation.

5 Concluding Remarks In this chapter we have discussed the importance of precursor and cofactor supply in the production of different types of antibiotics, the methods available for identification of rate-controlling steps in precursor supply and antibiotic biosynthesis pathways, and strategies for improving antibiotic yield through metabolic engineering. In short, these themes can be summarized in the following points: – Is precursor supply a problem in secondary metabolite production? Precursor supply is normally not a limiting factor in secondary metabolite production at natural levels, and overexpression of the biosynthesis cluster or manipulation of pathway-specific regulatory genes may therefore often be a feasible first strategy for improved antibiotic production. However, as the product yield is increased through strain development, it is probable that the supply of precursors or cofactors from primary metabolism eventually will become rate controlling for antibiotic biosynthesis. Particularly in situations, where precursors for antibiotic production are drained from cellular pathways that normally do not carry high fluxes, e.g. pathways in amino acid anabolism, precursor supply may become a problem. This is illustrated by the deregulation of anabolic pathways of high-yielding strains of P. chrysogenum,developed through classical strain improvement programs based on random mutagenesis. – Identification of connections between primary and secondary metabolism: precursor and cofactor supply The production of secondary metabolites at high yields may require changes in the carbon-flow of the primary metabolism in order to accommodate precursor and cofactor supply for the secondary biosynthesis pathway. In order to identify required changes in the metabolic fluxes, the complete primary


175

metabolic network rather than individual pathways leading to precursors or generating cofactors, needs to be taken into consideration. Through metabolic flux analysis of high-and low yielding strains or of a specific strain subjected to different growth conditions, connections between the primary and secondary metabolism can be identified.A good example of this is the study of an actinorhodin overproducing strain of Streptomyces lividans, where the information obtained via metabolic flux analysis could be directly used for design of an improved producer strain. – Strategies for improving productivity via an increased flux towards antibiotic precursors Production of antibiotics at high yields does not necessarily drain high amounts of central metabolites from the primary carbon metabolism, but rather requires chanelling of these metabolites into the pathways leading to antibiotic precursors. An attractive approach to achieve this is the modification of flux-partitioning at relevant branch points in the metabolic network, thereby redirecting the flow of carbon towards antibiotic precursors. However, the outcome of such a strategy is highly dependent on the regulation of the enzymes around these branch points, i.e. the flexibility of the nodes. This is particularly relevant when the antibiotic precursors are amino acids or intermediates in the anabolic pathways leading to amino acids, since the enzymes in amino acid synthetic pathways are often tightly controlled via feedback inhibition. Thus, it may not be possible to increase the intracellular concentration of a precursor, but if the capacity of the biosynthesis pathway towards antibiotic is sufficiently high to keep the precursor concentration at a low level, an increased flux toward the product may still be achieved through this approach. As an alternative or in connection with the above strategy, directed evolution may be applied to obtain enzymes less sensitive to feedback inhibition. – Rate-controlling steps in the biosynthesis pathway Metabolic Control Analysis is a powerful tool for analysis of the degree of flux control exerted by the different enzymatic steps of a pathway. Since information regarding enzyme kinetics and intracellular concentrations of pathway intermediates are generally required for MCA, this tool has to our knowledge only been used for analysis of the penicillin biosynthesis pathway. However, MCA is potentially useful for analysis of many antibiotic biosynthesis pathways, as more biochemical information on the enzymatic steps in these becomes available. In wild-type antibiotic producers, the first step towards the antibiotic is often tightly controlled and overexpression of the first enzyme in the biosynthesis pathway may therefore result in increased productivity. Alleviation of flux control at the committed step, however, necessarily means that the control of flux through the pathway is shifted to succeeding steps in the biosynthesis, which may result in the accumulation of intermediates and formation of by-products. Overexpression of all structural genes in a biosynthesis cluster is therefore generally a more effective approach, illustrated by the fact that improved penicillin producers carry

176


multiple copies of the complete biosynthesis cluster rather than any single gene in the cluster. However, there are situations where a specific enzymatic step late in the pathway is controlling the formation rate of the product to a large extent, e.g. the conversion of deacetylcephalosporin to cephalosporin C in Acremonium chrysogenum.

References 1. Bailey JE (1991) Science 252:1668 2. Nielsen J (2001) Appl Microbiol Biotechnol 55:263 3. Crawford L, Stepan AM, McAda PC, Rambosek JA, Conder MJ, Vinci VA, Reeves CD (1995) Biotechnology (NY) 13:58 4. Kacser H, Burns JA (1973) Symp Soc Exp Biol 27:65 5. Fell D (1992) Biochem J 286:313 6. Fell D (1997) Understanding the control of metabolism. Portland Press, London 7. Liao JC, Delgado J (1993) Biotechnol Prog 9:221 8. Stephanopoulos GN, Aristidou AA, Nielsen J (1998) Metabolic engineering. Academic Press, San Diego 9. Nielsen J (1998) Biotechnol Bioeng 58:125 10. Nielsen J, Jørgensen HS (1995) Biotechnol Prog 11:299 11. Pissarra P de N, Nielsen J, Bazin MJ (1996) Biotechnol Bioeng 51:168 12. Henriksen CM, Nielsen J, Villadsen J (1997) Biotechnol Prog 13:776 13. Theilgaard HA, Kristiansen KN, Henriksen CM, Nielsen J (1997) Biochem J 327:185 14. Cohen G, Argaman A, Schreiber R, Mislovati M, Aharanowitz Y (1994) J Bacteriol 176:973 15. Theilgaard HA, Nielsen J (1999) Antoine van Leeuwenhoek 75:145 16. Kennedy J, Turner G (1996) Mol Gen Genet 253:189 17. Theilgaard HA, van den Berg MA, Mulder CA, Bovenberg RAL, Nielsen J (2001) Biotechnol Bioeng 72:379 18. Barredo JL, Diez B, Alvarez E, Martin JF (1989) Curr Genet 16:453 19. Fierro F, Barredo JL, Díez B, Gutierrez S, Fernández FJ, Martín JF (1995) Proc Natl Acad Sci 92:6200 20. Newbert RW, Barton B, Greaves P, Harper J, Turner J (1997) J Microbiol Biotechnol 19:18 21. Summers RG, Wendt-Pienkowski E, Motamedi H, Hutchinson CR (1992) J Bacteriol 174:1810 22. Decker H, Summers RG, Hutchinson CR (1994) J Antibiot 47:54 23. Skatrud PL, Queener SW (1989) Gene 79:331 24. Gutiérrez S, Velasco J, Fernández FJ, Martín JF (1992) J Bacteriol 174:3056 25. Mathison L, Soliday C, Stepan T, Aldrich T, Rambosek J (1993) Curr Genet 23:33 26. Gutiérrez S, Velasco J, Marcos AT, Fernández FJ, Fierro F, Barredo JL, Díez B, Martín JF (1997) Appl Microbiol Biotechnol 48:606 27. Vallino JJ, Stephanopoulus G (1993) Biotechnol Bioeng 41:633 28. Bertsimas D, Tsitsiklis JN (1997) Introduction to linear optimisation.Athena Scientific, Belmont, Massachusetts 29. Varma A, Palsson BO (1993) J Theor Biol 165:477 30. Varma A, Boesch BW, Palsson BO (1993) Biotechnolo Bioeng 42:59 31. Schilling CH, Covert MW, Famili I, Church GM, Edwards JS, Palsson BO (2002) J Bacteriol 184:1 32. Jørgensen H, Nielsen J, Villadsen J (1995) Biotechnol Bioeng 46:117


177

33. Daae EB, Ison AP (1999) Met Eng 1:153 34. Rossa CA,White J, Kuiper A, Postma PW, Bibb M, Teixera de Mattos MJ (2002) Met Eng 4:138 35. Marx A, de Graaf AA, Wiechert W, Eggeling L, Sahm H (1996) Biotechnol Bioeng 49:111 36. Sauer U, Hatzimanikatis V, Bailey JE, Hochuli M, Szyperski T, Wüthrich K (1997) Nat Biotechnol 15:448 37. Schmidt K, Carlsen M, Nielsen J, Villadsen J (1997) Biotechnol Bioeng 55:831 38. Christensen B, Nielsen J (1999) Adv Biochem Eng Biotechnol 66:209 39. Christensen B, Nielsen J (2000) Biotechnol Bioeng 68:652 40. Christensen B, Christiansen T, Gombert AK, Thykær J, Nielsen J (2001) Biotechnol Bioeng 74:517 41. Christensen B, Nielsen J (1999) Met Eng 1:282 42. Jonsbu E, Christensen B, Nielsen J (2001) Appl Microbiol Biotechnol 56:93 43. Gunnarsson N, Mortensen UH, Sosio M, Nielsen J (2004) Mol Microbiol 52:895–902 44. Thykaer J, Christensen B, Nielsen J (2002) Met Eng 4:151 45. Stephanopoulus GN, Vallino JJ (1991) Science 252:1675 46. Vallino JJ, Stephanopoulus G (1994) Biotechnol Prog 10:327 47. Vallino JJ, Stephanopoulus G (1994) Biotechnol Prog 10:320 48. van Gulik WM, de Laat WTAM, Vinke JL, Heijnen JJ (2000) Biotechnol Bioeng 68:602 49. Nielsen (1997) Physiological engineering aspects of Penicillium chrysogenum. World Scientific, Singapore 50. Østergaard S, Theilgaard HA, Nielsen J (1998) Appl Microbiol Biotechnol 50:663 51. Katz L, Donadio S (1993) Annu Rev Microbiol 47:875 52. Hopwood DA, Sherman DH (1990) Annu Rev Genet 24:37 53. Bruheim P, Sletta H, Bibb MJ, White J, Levine DW (2002) J Ind Microbiol Biotechnol 28:103 54. Gramajo HC, Takano E, Bibb MJ (1993) Mol Microbiol 7:837 55. Butler MJ, Bruheim P, Jovetic S, Marinelli F, Postma PW, Bibb MJ (2002) Appl Environ Microbiol 68:4731 56. Hutchinson CR, Decker H, Madduri K, Otten SL, Tang L (1993) Antonie van Leeuwenhoek 64:165 57. Omura S, Tsuzuki K, Tanaka Y, Sakakibara H, Mamada H, Masuma R (1983) J Antibiot 36:1792 58. Lee SH, Lee KJ (1991) J Gen Microbiol 137:2547 59. Tang L, Zhang Y-X, Hutchinson CR (1994) J Bacteriol 176:6107 60. Jørgensen HS, Nielsen J, Villadsen J, Mølgaard H (1995) Appl Microbiol Biotechnol 43:123 61. Demain AL, Masurekar PS (1974) J Gen Microbiol 82:143 62. Friedrich CG, Demain AL (1977) J Antibiot 30:760 63. Luengo JM, Revilla G, López MJ, Villanueva JR, Martín JF (1980) J Bacteriol 144:869 64. Demain AL (1957) Arch Biochem Biophys 67:244 65. Somerson NL, Demain AL, Nunheimer TD (1961) Arch Biochem Biophys 93:238 66. Masurekar PS, Demain AL (1972) Can J Microbiol 18:1045 67. Luengo JM, Revilla G, López MJ,Villanueva JR, Martín JF (1979) J Gen Micribiol 115:207 68. Affenzeller K, Jaklitsch WM, Hönlinger C, Kubicek CP (1989) FEMS Microbiol Lett 58:293 69. Lu Y, Mach RL, Affenzeller K, Kubicek CP (1992) Can J Microbiol 38:758 70. Mendelovitz S, Aharonowitz Y (1982) Antimicrob Agents Chemother 21:74 71. Malmberg LH, Hu WS, Sherman DH (1993) J Bacteriol 175:6916 72. Khetan A, Malmberg LH, Sherman DH, Hu WS (1996) Ann NY Acad Sci 782:17

178


73. Khetan A, Malmberg LH, Kyung YS, Sherman DH, Hu WS (1999) Biotechnol Prog 15:1020 74. Stachelhaus T, Marahiel MA (1995) J Biol Chem 270:6163 75. Stein T, Kluge B, Vater J, Franke P, Otto A, Wittmann-Liebold B (1995) Biochemistry 34:4633 76. Luo L, Walsh CT (2001) Biochemistry 40:5329 77. Kuratsu Y, Arai Y, Inuzuka K, Suzuki T (1983) Agric Biol Chem 47:2607 78. Ito M, Aida K, Uemura T (1970) Progress in antimicrobial and anticancer chemotherapy, vol 2. University of Tokyo Press, Tokyo, Japan, p 1128 79. Ito M, Aida K, Uemura T (1969) Agric Biol Chem 33:262 80. van Wageningen AM, Kirkpatrick PN,Williams DH, Harris BR, Kershaw JK, Lennard NJ, Jones M, Jones SJ, Solenberg PJ (1998) Chem Biol 5:155 81. Pelzer S, Süssmuth R, Heckmann D, Recktenwald J, Huber P, Jung G, Wohlleben W (1999) Antimicrob Agents Chemother 43:1565 82. Steiert M, Schmitz FJ (2002) Curr Opin Investig Drugs 3:229 83. Hammond SJ,Williamson MP,Williams DH, Boeck LD, Marconi GG (1982) J Chem Soc Chem Commun 344 84. Heydorn A, Petersen BO, Duus JØ, Bergmann S, Suhr-Jessen T, Nielsen J (2000) J Biol Chem 275:6201 85. Hubbard BK, Thomas MG, Walsh CT (2000) Chem Biol 42:1 86. Choroba OW, Williams DH, Spencer JB (2000) J Am Chem Soc 122:5389 87. Lindblad B, Lindstedt G, Lindstedt S, Rundgren M (1977) J Biol Chem 252:5073 88. Pfeifer V, Nicholson GJ, Ries J, Recktenwald J, Schefers AB, Shawky R, Schröder J, Wohlleben W, Pelzer S (2001) J Biol Chem 276:38370 89. Li TS, Choroba OW, Hong H, Williams DH, Spencer JB (2001) Chem Commun 2156 90. Sandercock AM, Charles EH, Scaife W, Kirkpatrick PN, O’Brien SW, Papageorgiou EA, Spencer JB, Williams DH (2001) Chem Commun 1252 91. Vara AG, Hochkoepple A, Nielsen J, Villadsen J (2002) Biotechnol Bioeng 77:589 92. Neidhardt FC, Ingraham JL, Schaechter M (1990) Physiology of the bacterial cell: a molecular approach. Sinauer Associates, Sunderland, Massachusetts 93. Hodgson DA (2000) Adv Microb Phys 42:47 94. Frost JW, Draths KM (1995) Annu Rev Microbiol 49:557 95. Bongaerts J, Krämer M, Müller U, Raeven L, Wubbolts M (2001) Met Eng 3:289 96. Weaver LM, Herrmann KM (1990) J Bacteriol 172:6581 97. Edwards RM, Taylor PP, Hunter MG, Fotheringham IG (1987) WO 87/00202 98. Jossek R, Bongaerts J, Sprenger GA (2001) FEMS Microbiol Lett 202:145 99. LiaoH, Lin L, Chien HR, Hsu W (2001) FEMS Microbiol Lett 194:59 100. Sugimoto M, Ogawa Y, Suzuki T, Tanaka A, Matsui H (1997) US Patent 5 688 671 101. Dekleva ML, Strohl WR (1988) Can J Microbiol 34:1235 102. Henriksen CM, Christensen LH, Nielsen J, Villadsen J (1996) J Biotechnol 45:149 103. Christensen B, Thykær J, Nielsen J (2000) Appl Microbiol Biotechnol 54:212 104. Strohl WR, Bartel PL, Connors NC, Zhu CB, Dosch DC, Beale JM, Floss HG, StutzmannEngwall K, Otten SL, Hutchinson CR (1989) Biosynthesis of natural and hybrid polyketides by anthracyclin-producing Streptomycetes. In: Hershberger CL, Queener SW, Hegeman G (eds) Genetics and molecular biology of industrial microorganisms. Am Soc Microbiol, Washington, p 68 105. Rajgarhia VB, Priestley ND, Strohl WR (2001) Met Eng 3:49 106. Ikeda H, Kotaki H, Tanaka H, Omura S (1988) Antimicrob Agents Chemother 32:282 107. Obanye AIC, Hobbs G, Gardner DCJ, Oliver SG (1996) Microbiology 142:133 108. Bruheim P, Butler M, Ellington TE (2002) Appl Microbiol Biotechnol 58:735 Received: February 2004


Industrial Enzymatic Production of Cephalosporin-Based b -Lactams Michael S. Barber 1 · Ulrich Giesecke 2 · Arno Reichert 3 ·Wolfgang Minas 3 (✉) 1

MBA, 18 Croydon Road, Caterham, SurreyCR3 6QB, UK Anbics Laboratories AG, Maria-Ward-Strasse 1a, 80638 Munich, Germany 3 Anbics Management-Services AG, Technoparkstrasse 1, 8005 Zurich, Switzerland [email protected] 2

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

1

Introduction

2 2.1 2.2

The Cephalosporin Market . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Market Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Bulk Active Ingredients and Sterile Products . . . . . . . . . . . . . . . . . 189

3 3.1 3.1.1 3.1.2 3.1.3 3.2 3.2.1 3.2.2 3.2.3 3.2.4

Production of 7-ACA . . . . . . . . . . . . Fermentation . . . . . . . . . . . . . . . . Strains . . . . . . . . . . . . . . . . . . . . Culture Conditions . . . . . . . . . . . . . CPC Purification . . . . . . . . . . . . . . Conversion of Cephalosporin C into 7-ACA Chemical Cleavage . . . . . . . . . . . . . Enzymatic Cleavage . . . . . . . . . . . . . The Enzymes DAO and GAC . . . . . . . . Deacetyl-7-ACA by CAH . . . . . . . . . .

4

Process Economics of 7-ACA Production . . . . . . . . . . . . . . . . . . . 206

5 5.1 5.2 5.2.1 5.2.2

Advanced Intermediates 3¢ Position . . . . . . . . 7¢ Position . . . . . . . . Chemical Route . . . . . Biocatalytic Route . . . .

6

APIs by 7¢¢ and 3¢¢ Modified 7-ACA . . . . . . . . . . . . . . . . . . . . . . . 211

7

Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

References

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

. . . . . . . . . .

. . . . .

190 192 192 194 198 199 199 201 202 205

207 207 208 208 209

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

Abstract Cephalosporins are chemically closely related to penicillins both work by inhibiting the cell wall synthesis of bacteria. The first generation cephalosporins entered the market in 1964. Second and third generation cephalosporins were subsequently developed that were more powerful than the original products. Fourth generation cephalosporins are now reaching the market. Each newer generation of cephalosporins has greater Gram-negative

180

M. S. Barber et al.

antimicrobial properties than the preceding generation. Conversely, the ‘older’ generations of cephalosporins have greater Gram-positive (Staphylococcus and Streptococcus) coverage than the ‘newer’ generations. Frequency of dosing decreases and palatability generally improve with increasing generations. The advent of fourth generation cephalosporins with the launch of cefepime extended the spectrum against Gram-positive organisms without a significant loss of activity towards Gram-negative bacteria. Its greater stability to b-lactamases increases its efficacy against drug-resistant bacteria. In this review we present the current situation of this mature market. In addition, we present the current state of the technologies employed for the production of cephalosporins, focusing on the new and environmentally safer ‘green’ routes to the products. Starting with the fermentation and purification of CPC, enzymatic conversion in conjunction with aqueous chemistry will lead to some key intermediates such as 7-ACA, TDA and TTA, which then can be converted into the active pharmaceutical ingredient (API), again applying biocatalytic technologies and aqueous chemistry. Examples for the costing of selected products are provided as well. Keywords Biocatalysis · Enzymation · Cephalosporin C · Fermentation · Acremonium chrysogenum · b-Lactam List of Abbreviations 7-ACA 7-Amino cephalosporanic acid ACV d-(L-a-Aminoadipoyl)-L-cysteinyl-D-valine tripeptide 7-ADCA Amino-desacetoxy cephalospranic acid 6-APA 6-Amino penicillinic acid API Active pharmaceutical ingredient CA Cephalosporin C acylase CAH Cephalosporin C acetyl hydrolase CLEA Cross-linked enzyme aggregates CLEC Cross-linked enzyme crystals CPC Cephalosporin C DA-7-ACA Deacetyl-7-ACA DAC Deacetylcephalosporin C D-Amino acid oxidase DAO DO AC Deacetoxycephalosporin C GAC Glutaryl acylase GL-7-ACA Glutaryl-7-ACA HIC Hydrophobic interaction chromatography IEX Ion exchange chromatography KA-7-ACA Ketoadipoyl-7-ACA MMTD 2-Mercapto-5-methyl-1,3,4-thiadiazole PGA Penicillin G amidase D-2-(2,3-Dioxo-4-ethyl-1-piperazin-carbonylamino)-2-(4-hydroxyPip-pHPG phenyl)acetic acid PMV Packed mycelium volume TDA 7-(Amino-3-(5-methyl-1,3,4-thiadiazole-2-yl)thiomethyl-3-cephem-4carboxylic acid TM Metric ton TTA 7-Amino-3-[(1,2,5,6-tetrahydro-2-methyl-5,6-dioxo-1,2,4-triazin-3-yl)thiomethyl]-cephalosporanic acid TZ-7-ACA Tetrazolylacetyl-7-ACA

Industrial Enzymatic Production of Cephalosporin-Based b-Lactams TZM YX/S

181

Tetrazolylacetic acid methylester Yield coefficient consumption of amount substrate to yield amount product

1 Introduction The discovery of cephalosporin C (CPC) goes back to Giuseppe Brotzu, working at an institute in Cagliari on the island of Sardinia. Upon isolation from seawater near a sewage outlet a microorganism that had antibiotic activity against Gram-positive and Gram-negative bacteria was obtained. Upon classification as Cephalosporium, this organism was sent to Oxford in 1948. Work started in 1953 with the isolation and purification resulting several years later in the published structure of CPC [1]. In 1964 cefalotin was launched as the first semisynthetic CPC antibiotic. More than 50 semi-synthetic cephalosporins are being marketed today with sales of some US$9 billion annually. The market for “antibiotics” (perhaps better defined as anti-infective agents) sold at dose form level worldwide is estimated to be close to $60 billion in 2002, taking into account materials supplied under various aid programmes, secular and religious charity aided programmes, national bilateral aid and UN assistance programmes. The market is broken down by product type in Table 1 and by region in Table 2. In the context of this chapter, antibiotic means a chemical derived by fermentation or from a raw material obtained by fermentation, and having anTable 1 World anti-infectives market at ex-manufacturer/local primary distributor level

Product area

Estd. sales US$ Billion

Cephalosporins (all) Penicillins (incl Amoxiclav) All other betalactams Quinolones incl fluoroquinolones Macrolides (Erythromycin, Spiramycin and semi-synthetic derivatives thereof) Aminoglycosides (natural and semisynthetic) Tetracyclines All other antibacterials (incl anti-TB, topicals etc.) Antivirals (excl vaccines) Anti-infective vaccines Antifungals/antiparasitics Total

10 8 3 7 6 3 3 6 7 4 4 61

Data is rounded to one significant figure and, for the smaller groups above, should be regarded as indicative, rather than definitive. Source: Michael Barber and Associates, Caterham, UK (2003).

182

M. S. Barber et al.

Table 2 World anti-infectives market at ex-manufacturer/local primary distributor level by region

Country/Region

Estd. sales % of total US$ billion

Estd. popln. of region mill

Expenditure on anti-infectives per capita

USA Europe (all other, incl Russia) Europe (Germany, France, Italy) Japan S & E Asia/Australasia China Africa Americas incl Canada India Middle East Other Indian sub-continent Total

11 8

18 13

300 550

37 15

7

12

200

35

7 7 5 4 4 4 3 1–2 61

12 12 8 6 6 6 5 2 100

150 750 1500 800 550 1000 350 350 6.5 bill

47 9 3.3 5 7 4 9 4 Ave 9.4

Data is rounded, particularly sales value data in the Developing World, where it is difficult to obtain reliably consistent sales data. Source: Michael Barber and Associates, Caterham, UK (2003).

tibacterial, antiviral or antifungal properties. By common association, totally synthetic molecules, such as the fluoroquinolones and sulfonamides, which have substantially the same effect, are included in the general term “antibiotic”. There are, however, a number of important fermentation-derived antibiotics that have no antibacterial or similar activity. These include the anthracycline anticancer agents (e.g. doxorubicin, daunomycin, asparaginase, etc.) immunosuppressants (ciclosporin, mycophenolic acid, tacrolimus, sirolimus, etc.) vitamins, (e.g. ascorbic acid, pantothenic acid, cyano- and hydroxocobalamin, etc.). Those are not included in Table 1 and Table 2. It is apparent from Table 1 that the two series of b-lactam antibiotics, the cephalosporins and the penicillins, are the largest selling antibiotics. Both of these contain the “b-lactam” structural unit. The important structural feature is the nature of any group attached to the N atom at the top left-hand corner of the diagram in Fig. 1. This atom carries the principal side chain R1 of the cephalosporin series. Semisynthetic analogues are made by replacing the natural side chain with synthetic variants. The term “cephalosporins” includes more than 50 semisynthetic antibiotics derived from cephalosporin C (CPC), a natural antibiotic with no clinically useful antibacterial activity in its own right. b-Lactams exert their activity by inhibiting the cell wall synthesis of bacteria. The cephalosporins, the first of which, cefalotin, was introduced in 1964,

Industrial Enzymatic Production of Cephalosporin-Based b-Lactams

183

Fig. 1 Semi-synthetic b-lactam synthesis. The cephalosporin nucleus (middle) is derivatized on R1 and R2 to yield the active cephalosporin. Examples for 3¢ and 7¢ side chains are shown and the respective final products (API) indicated

184

M. S. Barber et al.

collectively have the largest sales by value of any group of antibacterial agents. However, by volume, in terms of tonnes sold or the number of prescriptions written, penicillins outweigh cephalosporins by a factor of some four to one. There are two main groups of cephalosporin antibiotics; one is derived from penicillin (G or V), the second from cephalosporin C (CPC). Most products that have intrinsic oral activity are better made from penicillin, while most products made from CPC are insufficiently absorbed from the human gut to be therapeutically adequate by the oral route, unless converted to pro-drugs by esterification [2, 3]. The penicillin-derived products are mainly based on 7-ADCA (hitherto made entirely from penicillin G) or on other more complex transformation products (made originally from penicillin V but now increasingly from penicillin G). The economics of the different processes involved are difficult to compare but essentially all (except for cefaclor as made by Lilly) depend on the price of commodity potassium penicillin G, principally supplied from China. Most products are multi-source at both the dose form and bulk active ingredient levels. The first cephalosporins were only available as injections. Orally active products, cefalexin, cefradine, etc., followed some five years later. Though these were initially manufactured from cephalosporin C, Lilly realised early on that a cheaper synthesis was essential if API costs were to be reduced to a level that allowed cefalexin to be competitive, initially with ampicillin and later with amoxicillin.Accordingly, it developed a synthesis of a suitable precursor, starting from penicillin V, which enabled it both to compete on price and have sufficient material to meet the market demand it was creating. Gist Brocades made a further advance in the synthesis of the primary intermediates when it found that its “Delft” process for making 6-APA was equally applicable to the manufacture of the key cephalosporin intermediates, 7-ACA and 7-ADCA [4, 5]. The early cephalosporins had good activity against a wide range of Grampositive bacteria, including a number of strains that produce penicillinase (but which remain methicillin sensitive). In contrast, they tended to have little activity against enterococci and weak and erratic activity against Gram-negative organisms. Cefazolin and cefradine (typical first generation products, Table 3) are still widely used in China but generally less and less elsewhere. The therapeutic limitations of the first generation of products led to the development of the so-called “second generation” products (cefamandol, cefaclor and cefuroxime, Table 3). These are characterised by a slightly poorer effect on Gram-positive bacteria but a significantly improved activity against enterobacteria and better resistance towards b-lactamases, especially those from Gram negative species. The third generation products (e.g. cefotaxime, ceftriaxone and cefixime, Table 3), have even better activity against Gram-negative bacteria, especially enterobacteria species. Most of the third generation products have good activity against streptococci, which helps to compensate for the variably weaker activity against staphylococci. The newest, fourth generation, products (cefepime,

Industrial Enzymatic Production of Cephalosporin-Based b-Lactams

185

Table 3 Overview of cephalosporins 2002

Names

Market volume MT bulk free acid equiv.a

Sales revenues $mill at dose form levelb

Sum $mill

1st Generation

Cefalotin Cefatrizine (o) Cefazolin Cefapirin Cefalexin (op) Cefadroxil (op) Cefradine (op)

70–80 105–115 850–900 25 2800–2900 450–480 1500–1570 5800–6070

80 110–120 730–770

Molecular Biotechnology of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases: Vol 88 (Advances in Biochemical Engineering Biotechnology)

Food Biotechnology (Advances in Biochemical Engineering Biotechnology)

White Biotechnology (Advances in Biochemical Engineering Biotechnology)

Marine Biotechnology I (Advances in Biochemical Engineering Biotechnology Vol 96)

Marine Biotechnology II (Advances in Biochemical Engineering Biotechnology Vol 97)

Biochemical Engineering and Biotechnology

Biochemical Engineering and Biotechnology

Biochemical Engineering and Biotechnology

Advances in Biochemical Engineering Biotechnology History of

Biotechnology of Aroma Compounds (Advances in Biochemical Engineering Biotechnology)

Biotechnology in India: v. 1 (Advances in Biochemical Engineering biotechnology)

Microscopy Techniques (Advances in Biochemical Engineering Biotechnology)

Advances In Biochemical Engineering Biotechnology Biotransformations

Advances In Biochemical Engineering Biotechnology Plant Cells

Biofuels (Advances in Biochemical Engineering Biotechnology)

Advances in Biochemical Engineering. Biotechnology Modern Adva

Advances In Biochemical Engineering - Biotechnology - Modern Adva

Disposable Bioreactors (Advances in Biochemical Engineering Biotechnology)

Advances in Biochemical Engineering Biotechnology Biopolyesters

Nano Micro Biotechnology (Advances in Biochemical Engineering Biotechnology, Volume 119)

Plant Cells (Advances in Biochemical Engineering Biotechnology, Vol. 72)

Advances in Biochemical Engineering, Volume 008: Mass Transfer in Biotechnology

Tools and Applications of Biochemical Engineering Science (Advances in Biochemical Engineering Biotechnology)

Biotechnology in China II: Chemicals, Energy and Environment (Advances in Biochemical Engineering Biotechnology, Volume 122)

Biomethanation I (Advances in Biochemical Engineering & Biotechnology, Volume 81)

Biomanufacturing (Advances in Biochemical Engineering & Biotechnology, Volume 87)

Advances In Biochemical Engineering Biotechnology. Thermal Biosensors Bioactivity Bioaffinity

Disposable Bioreactors (Advances in Biochemical Engineering Biotechnology, 115)

Advances in Biochemical Engineering Biotechnology, Volume 071, Biopolyesters

Advances in Biochemical Engineering & Biotechnology, Volume 082, Biomethanation II

Optical Sensor Systems in Biotechnology (Advances in Biochemical Engineering Biotechnology, Volume 116)

Molecular Biotechnology of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases: Vol 88 (Advances in Biochemical Engineering Biotechnology)

Food Biotechnology (Advances in Biochemical Engineering Biotechnology)

White Biotechnology (Advances in Biochemical Engineering Biotechnology)

Marine Biotechnology I (Advances in Biochemical Engineering Biotechnology Vol 96)

Marine Biotechnology II (Advances in Biochemical Engineering Biotechnology Vol 97)

Biochemical Engineering and Biotechnology

Biochemical Engineering and Biotechnology

Biochemical Engineering and Biotechnology

Advances in Biochemical Engineering Biotechnology History of

Biotechnology of Aroma Compounds (Advances in Biochemical Engineering Biotechnology)

Biotechnology in India: v. 1 (Advances in Biochemical Engineering biotechnology)

Microscopy Techniques (Advances in Biochemical Engineering Biotechnology)

Advances In Biochemical Engineering Biotechnology Biotransformations

Advances In Biochemical Engineering Biotechnology Plant Cells

Biofuels (Advances in Biochemical Engineering Biotechnology)

Advances in Biochemical Engineering. Biotechnology Modern Adva

Advances In Biochemical Engineering - Biotechnology - Modern Adva

Disposable Bioreactors (Advances in Biochemical Engineering Biotechnology)

Advances in Biochemical Engineering Biotechnology Biopolyesters

Nano Micro Biotechnology (Advances in Biochemical Engineering Biotechnology, Volume 119)

Plant Cells (Advances in Biochemical Engineering Biotechnology, Vol. 72)

Advances in Biochemical Engineering, Volume 008: Mass Transfer in Biotechnology

Tools and Applications of Biochemical Engineering Science (Advances in Biochemical Engineering Biotechnology)

Biotechnology in China II: Chemicals, Energy and Environment (Advances in Biochemical Engineering Biotechnology, Volume 122)

Biomethanation I (Advances in Biochemical Engineering & Biotechnology, Volume 81)

Biomanufacturing (Advances in Biochemical Engineering & Biotechnology, Volume 87)

Advances In Biochemical Engineering Biotechnology. Thermal Biosensors Bioactivity Bioaffinity

Disposable Bioreactors (Advances in Biochemical Engineering Biotechnology, 115)

Advances in Biochemical Engineering Biotechnology, Volume 071, Biopolyesters

Advances in Biochemical Engineering & Biotechnology, Volume 082, Biomethanation II

Optical Sensor Systems in Biotechnology (Advances in Biochemical Engineering Biotechnology, Volume 116)

Recommend Documents