Recent Progress of Biochemical and Biomedical Engineering in Japan I

90 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

Recent Progress of Biochemical and Biomedical Engineering in Japan I Volume Editor : Takeshi Kobayashi

With contributions by S. Iijima · Y. Iwasaki · Y. Kawarasaki · K. Miyake · H. Nakano · Y. Nakashimada · N. Nishio · T. Ohshima · E.Y. Park · M. Sato · K.-I. Suehara · T. Tanaka · M. Taniguchi · T. Yamane · T. Yano

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 www.springeronline.com

ISSN 0724-6145 ISBN 3-540-20408-3 DOI 10.1007/b14089 Springer-Verlag Berlin Heidelberg New York Library of Congress Control Number 2004104352 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. Takeshi Kobayashi Department of Biological Chemistry Chubu University 1200 Matsumoto-cho, Kasugai Aichi 487-8501, Japan [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: www.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 www.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 www.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

During World War II, industrial production of penicillin was started. This was the first example of aerobic microbial cultivation on an industrial scale, and many new techniques were developed to cultivate Penicillium chrysogenum in large fermenters. Since demands for penicillin production were urgent, microbiologists, biochemists and chemical engineers were in a great hurry to start production of penicillin and as a result, most early techniques were acquired through empirical procedures. However, this was the start of biochemical engineering, and the contents of almost all chapters of the book “Biochemical Engineering” written by Shuichi Aiba,Arthur E. Humphrey and Nancy F. Millis (University of Tokyo Press, 1965) dealt with this subject and its later development under academic conditions. In 1957, glutamic acid production was started by Kyowa Hakko Co. Other amino acids were also produced industrially and industrial microbial cultivation was rapidly developed to an advanced state. The organisms were not limited to microorganisms, and mammalian cells and plant cells were then also applied to production of glycosylated proteins and complex secondary metabolites. In 1972, genetic engineering technology was developed and this technique had a drastic influence not only on basic biosciences but also on biochemical engineering. The wealth of information that has been accumulated on genetic engineering technology and as well as hybridoma technology has made it possible to produce various metabolites and proteins in microorganisms, mammalian cells and plant cells. Species barriers between microorganisms, animals and plants have been, in principle, eliminated. The areas dealt with in biochemical engineering have been expanded to many organisms. In 2003, the human genome project has been completed, and complete DNA sequences have been announced. The areas dealt with in biochemical engineering have been expanded to humans; bioinformatics and biomedical engineering are now parts of biochemical engineering. A major objective in editing this book has been to gather together the information of the expanding areas in biochemical engineering. Publication was motivated by the retirement of the editor after working at Nagoya University from April 1968 to March 2004. The editor asked his esteemed friends and colleagues doing active research in biochemical engineering in Japan to sum

X

Preface

up the information of these rapidly expanding areas in biochemical engineering. This book is not intended to be an introduction to biochemical engineering, but to serve as a reference that looks at the expanded field of biochemical engineering in Japan and also looks forward to future prospects. This book (two volumes) is composed of 15 chapters dealing with microbial cultivation techniques in biomedical engineering including tissue engineering and cancer therapy. Hopefully, this book will give readers a glimpse of the past and also a view of what may happen in biochemical engineering in Japan. Finally, I would like to thank all my friends and colleagues for their cooperation in publishing this book. I express my deepest appreciation to my wife, Noriko Kobayashi who endured the long evenings and weekends I devoted to working at Nagoya University. Nagoya, May 2004

Takeshi Kobayashi

Contents

Recent Progress in Microbial Cultivation Techniques E.Y. Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

Clarification of Interactions Among Microorganisms and Development of Co-culture System for Production of Useful Substances M. Taniguchi · T. Tanaka . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 High Rate Production of Hydrogen/Methane from Various Substrates and Wastes N. Nishio · Y. Nakashimada . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Bacterial Capsular Polysaccharide and Sugar Transferases K. Miyake · S. Iijima . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field T. Ohshima · M. Sato . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Cell-free Protein Synthesis Systems: Increasing their Performance and Applications H. Nakano · Y. Kawarasaki · T. Yamane . . . . . . . . . . . . . . . . . . . . 135 Enzymatic Synthesis of Structured Lipids Y. Iwasaki · T. Yamane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Bioprocess Monitoring Using Near-Infrared Spectroscopy K. Suehara · T. Yano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Author Index Volumes 51–90 . . . . . . . . . . . . . . . . . . . . . . . . 199 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

Contents of Volume 91 Recent Progress of Biochemical and Biomedical Engineering in Japan II Volume Editor: Takeshi Kobayashi

Metabolic Flux Analysis Based on 13C-Labeling Experiments and Integration of the Information with Gene and Protein Expression Patterns K. Shimizu Application of Knowledge Information Processing Methods to Biochemical Engineering, Biomedical and Bioinformatics Field T. Hanai · H. Honda Large-Scale Production of Hairy Root N. Uozumi Large-Scale Micropropagation System of Plant Cells H. Honda · T. Kobayashi Development of Culture Techniques of Keratinocytes for Skin Graft Production M. Kino-oka · M. Taya Transgenic Birds for the Production of Recombinant Proteins M. Kamihira · K. Nishijima · S. Iijima Functional Magnetic Particles for Medical Application M. Shinkai · A.Ito

Adv Biochem Engin/Biotechnol (2004) 90: 1– 33 DOI 10.1007/b94190 © Springer-Verlag Berlin Heidelberg 2004

Recent Progress in Microbial Cultivation Techniques Enoch Y. Park (✉) Laboratory of Biotechnology, Department of Applied Biological Chemistry, Faculty of Agriculture, Shizuoka University, 836 Ohya, Shizuoka 422-8529, Japan [email protected]

1

Development of Microbial Cultivation in Biotechnology

. . . . . . . . . .

High Cell Density Cultivation with Process Control . . . . . Cell Recycling System . . . . . . . . . . . . . . . . . . . . . . Fed-batch Culture with Various Control Techniques . . . . . Moving Identification Combined with Optimal Control . . . Fuzzy Control . . . . . . . . . . . . . . . . . . . . . . . . . . Application of Control Technique to Culture of Recombinant Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Recombinant B. subtilis . . . . . . . . . . . . . . . . . . . . . 2.4.2 Recombinant Yeast . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1 2.2 2.2.1 2.3 2.4

. . . . .

. . . . .

. . . . .

. . . . .

2 2 4 5 7

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

14 14 17

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

2

3 3.1 3.2 3.3

Image Analysis for Characterization of Mycelium . . . . . . . . Morphological Classification . . . . . . . . . . . . . . . . . . . . Mycelial Physiology . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring Mycelial Formation Using a Flow-Through Chamber

. . . .

. . . .

19 19 23 28

4

Future Prospects for Microbial Cultivation . . . . . . . . . . . . . . . . . .

32

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

Abstract Recent advances in the improvement of microbial cultivation are reviewed, with emphasis on biochemical engineering techniques as a means of obtaining high production rate of bioproduct. Possible uses of high cell density culture include their use in food industry as well as in the production of new medicines and in biotechnology. Concentration of microorganisms using a hollow fiber membrane or centrifuge, and increase in cell density by controlling the pH, dissolved oxygen, or carbon source concentrations of the culture broth with control algorithms are discussed. In a culture of filamentous microorganisms the mycelial morphology is hard to define and it is difficult to quantify its amount, and this is one of the bottlenecks hampering the improvement of production rate. Specific features of mycelial cultivation in the presence of highly pulpy mycelia and entangled-pellets are scrutinized by visual inspection through a microscope that is linked to a computer, and using software that can characterize the mycelial morphology. Image analysis technology for analyzing the mycelial image captured by a digital camera is a potential tool for morphological analysis, including analysis of the morphological development of filamentous microorganisms. Keywords Fed-batch culture · Glucose control · Image analysis · Fungal morphology

2

E.Y. Park

1 Development of Microbial Cultivation in Biotechnology World War II provided an impetus to shift from chemical synthesis to biological processes for the antibody production process. Many companies and government laboratories, assisted by many universities, took up the challenge. However, the new fermentation process gave very low production levels of valuable products, so the first fermentation efforts were modest, and many companies were at first reluctant to commit to the fermentation process for mass production of antibiotics. The low rate of production per unit volume required very large and inefficient reactors, and the low concentration made product recovery and purification very difficult. This problem was a significant constraint on the early development of fermentation technology. In the early days of the fermentation process, surface fermentation using a variety of containers was used, but this required a long growth period and was labor-intensive. Soon tank-based processes were developed, and surface fermentation was exchanged for submerged fermentation, which allowed the control of pH and dissolved oxygen (DO) concentration in the reactor. This technology made it possible to produce products on a large scale with high performance, and led to the commercial production of penicillin via submerged fermentation [1]. Since then, many cultivation techniques have been developed and many useful bioproducts, including antibiotics, enzymes, and biological active compounds have been produced in large-scale fermentors. Various advances in life science techniques, including in the areas of mutation, gene manipulation, and media, have also aided progress in fermentation technology.

2 High Cell Density Cultivation with Process Control Increasing cell mass or product concentration is one of the most effective methods for improving productivity in the fermentation process. Two concepts have been developed to achieve this goal, as shown in Fig. 1. 2.1 Cell Recycling System To maintain high cell density in the reactor, a cell recycling bioreactor has been devised (Fig. 1a). The maintenance of high cell density in the bioreactor was obtained by using a cell separator with flocculating yeast [2], centrifugal separation of cells [3], or recycling membrane-filtered cells [4]. Cells were recycled by the fermentor and a small amount of effluent was exchanged for fresh medium to avoid nutrient depletion. The system of membrane filtering of anaerobic bacteria that do not need oxygen was very effective in achieving a high concentration of cells and high productivity. Taniguchi et al. [5] report-

3

Fig. 1 Concept for high cell-density culture

Recent Progress in Microbial Cultivation Techniques

4

E.Y. Park

Table 1 Comparison of acetic acid productivity [6]

Method

Dilution rate (h–1)

Production rate (g/L/h)

Batch culture Continuous culture Immobilized on Carrageenan Ceramic monolith Hydrous titanium oxide High cell-density culture (cell recycle)

– 0.14 0.3 0.5 0.07 4.0

2.6 2.8 4.0 10.4 5.0 120

ed that lactic acid productivity by Streptococcus cremoris increased 19-fold compared with the conventional batch cultures. In the case of strictly aerobic microorganisms significant progress has also been reported. When Acetobacter aceti cells were recycled through a hollow fiber filter module, a high acetic acid production rate (120 g/L/h) was obtained [6]. The acetic acid productivities attained in various types of bioreactors and processes are compared in Table 1. Acetic acid production rate was enhanced 46-fold compared with conventional batch cultures. However, when dilution rate was increased to 4.0 h–1 the DO concentration was lower than the critical concentration of 1 ppm, even though oxygen-rich air (at a partial pressure of 0.92 atm) was supplied. Another problem with high cell density cultures was cell deactivation due to product inhibition. To overcome this, the following methods are considered: 1) the product can be removed from the fermentor; 2) acetic acid is dialyzed through anion exchange resin by electrodialysis. The process is modified into repeated batch or fed-batch systems combined with cell filtration apparatus. However, these systems present problems in scale-up and use in large-scale production of bioproducts from microbial cultivation. 2.2 Fed-batch Culture with Various Control Techniques Theoretically, obtaining high productivity high cell density culture is a useful method, but it is at a disadvantage for obtaining high concentration of the product. To overcome this, a fed-batch culture process equipped with controlling culture variables has been developed (Fig. 1b). The fed-batch culture is defined as a type of operation for microbial reactions in which one or more nutrients is supplied continuously or intermittently to the fermentor, but no culture broth is discarded from the fermentor until the end of the operation. Conventionally there are two types of nutrient-feeding mode: without feedback control and with feedback control. In the former case the feed rate of nutrient is kept constant or changes are controlled in a predetermined manner, while the latter case is feedback-controlled, utilizing a measurable parameter as control index, such as pH, DO concentration [7, 8], RQ, and nutrient concentration.

Recent Progress in Microbial Cultivation Techniques

5

Fig. 2 Schematic figure showing the fuzzy control strategy

This on- or off-line analyzed index is connected to a computer that calculates the feed rate of nutrient according to a given control strategy. Schematic features of this control strategy are shown in Fig. 2. The control strategy consists of a feedforward and a feedback control. To maintain the exponential growth phase in a fed-batch culture of microorganisms, glucose must be fed exponentially while the other nutrients should be maintained at a sufficient level. To achieve this, the mass balance equation between feed and consumption of glucose should be considered. The feedforward glucose feed rate (normal glucose feed rate, F*) can be determined in accordance with increase of the cell concentration as follows:

mX F* = 9 Yx/s S0

(1)

where m, X, S0, and Yx/s denote the specific growth rate, concentrations of cell and feed glucose, and cellular yield. Although significant progress in fed-batch culture has made, there are still many problems with microbial cultivation. Microbial processes are regulated by various biochemical enzyme reactions and are affected by many environmental conditions. Until now, only a few environmental conditions, such as pH, temperature and DO concentration could be controlled in the fermentation process during cultivation. More information is needed in order to control fermentation processes more accurately. Many researchers have developed biosensors to measure concentrations of carbon and nitrogen sources, and metabolites on- or off-line. Here two control algorithms that make use of measurable control variables are reviewed. 2.2.1 Moving Identification Combined with Optimal Control Because the process model is unclear, but time series data of glucose concentrations are available on-line, a moving identification combined with optimal control (MICOC) [9] is available as a control strategy to maintain glucose concentration in culture broth.

6

E.Y. Park

Let the process model be expressed by y(k) = a1(k – 1)y(k – 1) + b1(k – 1)u(k – 1)

(2)

where u and y are the input and output variables. a1 and b1 are the model parameters, and k is the sampling instant. Let the difference between the output y and the set point ys be e. Namely, e(k) = ys(k) – y(k)

(3)

Using Eq. 2, we have e(k + 1) = ys(k + 1) – y(k + 1) = ys(k + 1) – a1(k)y(k) – b1(k)u(k) = a1(k)e(k) – v(k) where v(k) = ys(k + 1) – a1(k)ys(k) – b1(k)u(k)

(4)

(5)

Consider the objective function to be minimized such as 1 J = 2 [e(k + 1)2 + wv(k)2] 2

(6)

where w is the weighting factor. Now the problem is to minimize J with respect to v(k) under the constraint of Eq. 4. Introducing the Lagrange multiplier l, 1 J* = 2 [e(k + 1)2 + wv(k)2] + l {e(k + 1) – a1(k)e(k) – v(k)} 2

(7)

From the necessary condition for optimality (∂J*/∂n = 0), Eq. 7 is – a1(k)e(k) v*(k) = 08 (1 + w)

(8)

where n* denotes the optimal value of n. The optimal value of u, u* can be also obtained from Eqs. 3, 5 and 8: a1(k){ys(k)–a1(k–1)y(k–1)– b1(k–1)u(k–1)} ys(k+1)–a1(k)ys(k)+ 08000006 (1+ w) u*(k) = 000000000001 b1(k) (9) Since model parameter a1 is simple pole, its absolute value must be less than unity. The assignment of a1 is rather arbitrary and was set at 0.5 in our experiments, and the assignment of a1 had little effect on the determining moving model. The model parameter b1 is adjustable by on-line data using the recursive least squares method, letting

Recent Progress in Microbial Cultivation Techniques

q = (a1, b1)

7

(parameter vector)

y = {–y(k), –y(k + 1),…, –y(k – n + 1), u(k),…, u(k – m + 1)} (data vector) Since the parameter b1is changeable with sampling time, b1 is substituted as bi . Substituting glucose concentration (S) and glucose feed flow rate (F) for y and u in the model, respectively, the parameter bi can be obtained using: k+m

 F(j){S(j + 1) – aS(j)} j=k–1

bi = 00002 k+m

(10)

 F(j)2 j=k–1

Consequently, the optimal glucose feed rate F* is calculated via: a{Ss –aS(k–1) – bi F(k–1)} Ss –aSs + 08005 (1+ w) F* = 000009 bi

(11)

where Ss denotes the set concentration of glucose. This MICOC provides the optimal glucose feed rate in Eq. 1 that can be used to maintain the glucose concentration much more accurately. 2.3 Fuzzy Control Computers coupled with sensors or analytical instruments are useful for controlling the fermentation processes. Consequently, various control strategies have been developed to control fermentation variables more accurately, but in many cases there are still numerous difficulties in constructing an accurate mathematical model for the growth of microorganisms. Another difficulty in mathematical modeling is accumulation of the inhibitory metabolites, such as ethanol or acetic acid, in culture broth during aerobic fermentation of yeast or Escherichia coli, respectively. The production of these metabolites is due to the Crabtree effect or the bacterial Crabtree effect [10], and they inhibit cell growth and product formation. Therefore, it is necessary to keep these inhibitory metabolites at a low concentration by feeding glucose in at an optimal value so as to prevent inhibitory metabolites forming while maintaining a relatively high growth rate. Konstantinov et al. [11] formulated a glucose feeding policy, called a balanced DO-stat, that kept acetic acid concentration at a low level during cultivation of recombinant E. coli, and obtained concentrations of cell and phenylalanine of 36 g/L and 24 g/L, respectively. With the aid of a glucose analyzer, glucose concentration was controlled at 0.15 g/L during fermentation of recombinant Saccharomyces cerevisiae cells and obtained 20-fold higher gene product than when glucose concentration was controlled at 10 g/L [12]. However, even if glucose concentration is maintained at a low level, it is very

8

E.Y. Park

difficult to maintain ethanol at low concentration. High ethanol concentration severely affects gene expression during recombinant S. cerevisiae culture. In this case, to maintain both ethanol and glucose concentrations simultaneously at defined values, a new control strategy is required. Many operations during the culture of microorganisms are handled only by experienced experts. These operations seem to be appropriate kinds of processes for applying fuzzy control theory. Fuzzy reasoning as a control theory has been applied for production processes of glutamic acid [13]. To obtain a precise feedforward glucose feed rate, Eq. 1 has to be compensated as follows: F = F* + DF

(12)

The glucose feed rate F* of Eq. 1 is just a rough approximation. Because of changes in cell activity and environmental conditions, the feed rate has to be corrected. To obtain X, m, and Yx/s in Eq. 1, on-line data from the turbidimeter or cell measuring apparatus, and on-line data of serial cell concentration, and glucose consumption within 20 min are used. The initial cellular yield from glucose was assumed to be 0.5. This correction was noted as DF, which was determined by the fuzzy control algorithm.As a result, the real glucose feed rate was adjusted as Eq. 12. Table 2 Production rules for fuzzy control

Production rule I DO concentration: B

DO concentration: M

Glucose

Ethanol

S B

Glucose

S

M

B

PB PM

PM PS

ZE NS

Ethanol

S B

S

M

B

PM PS

ZE NS

NM NB

Production rule II DO concentration: B

DO concentration: M

Glucose

Ethanol

S M B

Glucose

S

M

B

PB PM ZE

PM ZE NS

ZE NS NB

Ethanol

S M B

S

M

B

PM PS NE

PS ZE NM

NS NM NB

Abbreviations: B: big; M: medium; S: small; PB: positivebig; PM: positive medium; PS: positive small; ZE: zero; NS: negative small; NM:negative medium; NB: negative big

Recent Progress in Microbial Cultivation Techniques

9

Fig. 3 Membership functions of DO (a-1), glucose (a-2), and ethanol (a-3) and output (b). B: big; M: medium; S: small; PB: positive big; PM: positive medium; PS: positive small; ZE: zero; NS: negative small; NM: negative medium; NB: negative big

Table 2 shows two kinds of production rule for fuzzy control. Production rule II was obtained from modification of production rule I. Fuzzy inference is carried as follows (for example). IF {DO concentration is “B” and glucose concentration is “S” and ethanol concentration is “S”} THEN DF be “PB”. Membership functions for inputs and outputs are established as in Fig. 3. “S”,“M”, and “B” concentrations were set as 0.1, 0.2, and 0.3 g/L for glucose and 1, 2, and 3 g/L for ethanol, respectively. Concerning DO concentration, only two members, as “M” and “B”, are set since it was not varied much throughout fermentation. The inference procedure is shown in Fig. 4, which follows Mamdani’s minmax algorithm [14]. The grade of output’s fuzzy set is determined as the minimum value among the grades of input’s fuzzy set by each selected rule. Using production rules, the corresponding fuzzy set and grades are determined. Defuzzification of each variable is then carried out using a simplified centerof-gravity method (see Fig. 4). The value of DF is bound by –F*(DFmin) and +3F*(DFmax) (Fig. 3b). If DF is equal to the DFmin then the feed rate is zero. If DF is equal to the DFmax , in that case the feed rate is a maximum rate (4F*). The determination of the lower and upper bound (proportional constants such as –1 and +3) was based on experience.

10

E.Y. Park

Fig. 4 Fuzzy inference for the determination of DF. DO, G and E in the horizontal axis indicate on-line measured concentrations of dissolved oxygen, glucose, and ethanol, respectively

Recent Progress in Microbial Cultivation Techniques

11

Fig. 5 Cell growth a, concentrations of glucose b and ethanol c, glucose feed flow rate d in fed-batch cultures of baker’s yeast with control of both glucose and ethanol concentrations by production rule I (I) and production rule II (II). The dotted lines in Figs. 5I-b and 5II-b indicate “M” glucose concentration (0.2 g/L)

Cell growth yield (Yx/s) and ethanol yield (Yp/s) were defined by the produced amounts of dry cells and ethanol divided by total consumed amount of glucose, as follows: (XfVf – XiVi ) Yx/s = 000001 ˆ ˆ tf

tf

(13)

S0 Ú Fdt – Ú SFglcdt + (SiVi –SfVf ) tiˆ

tiˆ

tfˆ

Ú PFglcdt + (PfVf – PiVi )

tiˆ

Yp/s = 000002 ˆ ˆ tf

tf

(14)

S0 Ú Fdt – Ú SFglcdt + (SiVi – SfVf ) tiˆ

tiˆ

where F, Fglc, P, S, X, V and t denote glucose feed flow rate, filtration rate for on-line glucose measurement, concentrations of ethanol, glucose and cell, culture volume and culture time, respectively. Subscripts i and f denote initial and final, respectively. Fed-batch culture was carried out by controlling both glucose and ethanol concentrations via a fuzzy controller. In this experiment, two kinds of production rule (I and II) were applied. In the case of production rule I, ethanol concentration increased gradually and finally reached 12.2 g/L (Fig. 5I). The specific growth rate was maintained at 0.24 h–1 over a culture time of 5 h, but

12

E.Y. Park

it gradually decreased when ethanol accumulated to above 7 g/L, and finally became 0.14 h–1. The maximum cell concentration was 74 as OD660. A total of 46.6 g of glucose was consumed. Yx/s and Yp/s were calculated as 0.42 g cell/g glucose and 0.28 g ethanol/g glucose, respectively. This showed that it is very difficult to strictly control ethanol concentration, which seems due to the metabolic time delay from glucose consumption to ethanol production, and the production rule I not being suitable for restraining ethanol production. For example, if glucose, ethanol, and DO concentrations are all “B”, then DF is decided at “NS” by the production rule I. Since the real glucose feed rate F=DF+F*, glucose is still fed in proportion to the specific growth rate until DF=–1F*. In general, increasing the fineness of the grid of production rules gives better control performance [15]. Therefore, the production rule I was modified into the production rule II. Ethanol concentration was subdivided into three member functions and DF levels for feedback control were lowered from “PM” to “ZE”, from “PS” to “NS” and from “NS” to “NB” or “NM”. The modified production rule II was applied to baker’s yeast culture for the same culture conditions [16]. However, ethanol concentration increased gradually and finally reached 6.1 g/L, which was half of the value obtained in Fig. 5I. Therefore, production rule II is not enough to keep the ethanol concentration low. To improve ethanol concentration control, an on-off feedback controller of ethanol concentration was added to the fuzzy controller, and a critical factor was introduced to the fuzzy controller for stopping the glucose feed pump. This prevents ethanol concentration from increasing to a critical level. The critical ethanol concentration was defined as the “B” ethanol concentration multiplied by the critical factor. In this experiment, the critical factor and the critical ethanol concentration were set as 1.5 and 4.5 g/L, respectively. The results are shown in Fig. 5II. Glucose and ethanol concentrations were controlled to be in the range of 0.27±0.25 g/L and 4.2±0.8 g/L, respectively. A sudden increase in glucose might be due to a sudden increase of glucose feed rate. The specific growth rate decreased from 0.26 h–1 to 0.2 h–1 over a culture time of 4 h because of a depletion of glucose in culture broth. Glucose feed rate was stopped from the culture time of 4 h to 5.5 h, because of “B” ethanol concentration. During this culture time, the carbon source was changed from glucose to ethanol. The specific growth rate was maintained at 0.16 h–1 after a cultivation time of 8 h. A total of 39.7 g of glucose was consumed during the controlled time of 11 h. The final OD660 was 92.3, and Yx/s and Yp/s were calculated as 0.62 g dry cell/g glucose and 0.19 g ethanol/g glucose, respectively. This controller’s performance for a lower ethanol concentration than 4.5 g/L ethanol control was also tested using the production rule II. The critical ethanol concentration was set to “B” ethanol concentration (=3.0 g ethanol/L). Other experimental conditions were unchanged. The results are shown in Fig. 6. Ethanol concentration was controlled successfully at ±0.35 g/L; however, glucose concentration fluctuated in the range from 0 g/L to 2 g/L because of a sudden increase of glucose feed rate.At a culture time of 7 h, since both glucose

Recent Progress in Microbial Cultivation Techniques

13

Fig. 6 Cell growth a, concentrations of glucose b and ethanol c, glucose feed rate d in fed-batch culture of baker’s yeast with control of both “M” glucose and “M” ethanol concentrations at 0.2 g/L and 2.0 g/L by fuzzy controller. The dotted lines in Fig. 6b and c indicate “M” glucose and ethanol concentrations

and ethanol concentrations were “B”, DF was “NB”. Consequently, F=0 and glucose was depleted after a while. The determined glucose feed rate was still low until the ethanol concentration was decreased to the “M” level. Over this culture time, since ethanol was used as the carbon source by the yeast, the concentration decreased gradually. At a culture time of 9.5 h, concentrations of glucose and ethanol decreased to nearly zero (“S”) and 1.6 g/L (between “S” and “M”), respectively. Then glucose feed rate was changed to “PB” (=4F*) for “B” DO concentration (because of glucose depletion). This resulted in an abrupt increase in glucose concentration. This glucose depletion and consumption of ethanol resulted in poor cell growth. The specific growth rate was maintained at about 0.09 h–1 after a culture time of 8 h. The final OD660 was 81.4, and amount of consumed glucose was 26.8 g, during a controlled time of 18 h. Yx/s and Yp/s were calculated as 0.63 g/g and 0.14 g/g, respectively.

14

E.Y. Park

2.4 Application of Control Technique to Culture of Recombinant Microorganisms Considerable progress has been made in producing heterologous proteins by recombinant Bacillus subtilis, or yeasts by recombinant E. coli. The production rate of a gene product in recombinant microorganisms depends on various factors, among which are the interactions between host cells and the plasmid. At the cellular level, an important consideration is the inference of gene expression on host growth. High expression is desired; however, the synthesis of a cloned gene product places additional stress on the cells. This can result in a low growth rate, lowered cellular yield, and plasmid instability [17]. To lessen this negative effect of cloned gene expression, plasmids with inducible promoters may be effective in maximizing gene products, because the timing and level of cloned gene product synthesis can be controlled [18]. Inducible promoters such as trp, pho, pL, pR, tac, lac, and a runaway plasmid are wellknown for E. coli, whereas in Saccharomyces cerevisiae systems SUC2, PHO5, ADH1, GAL1, GAL7, and GAL10 have been developed. On the other hand, at the process level, to maximize gene expression with some of these inducible promoters, several processes have been proposed: the use of a fermentor equipped with a crossflow filtration system [19], an on-line glucose monitoring and controlling system [12], and exchange of the carbon source [20]. Here, application of MICOC and fuzzy control strategy to the fedculture of recombinant B. subtilis and yeasts respectively are reviewed. 2.4.1 Recombinant B. subtilis With advances in recombinant DNA technology, gene engineering systems have also been developed for Bacillus. Bacteria belonging to the genus Bacillus have attractive properties for the production of industrially important enzymes, since they are safe, non-infectious to humans, do not produce toxic substances, and can secrete gene products in the culture broth. However, as compared with E. coli and S. cerevisiae, there have been very few reports on an efficient cultivation of recombinant Bacillus in fermentor scale cultivation and the development of suitable conditions for the production of recombinant protein, because the high-density cultivation of B. subtilis was hampered by inhibitory metabolites. Most studies aimed at improving the performance of fermentation have investigated the reduction of inhibitory metabolites. In these approaches, however, glucose concentration in the culture broth was not strictly maintained at the set value, since it was estimated indirectly. To improve the control of glucose concentration directly, an on-line glucose measurement system was used. Glucose concentration was controlled at 10 g/L, 1 g/L, and 0.2 g/L, respectively, using modified MICOC [21].When glucose concentration was controlled at 1 g/L after culture time of 6 h, cell concentration increased to 184 g/L, and the specific activity and activity of b-galactosidase increased to 5.7 U/mg protein

Recent Progress in Microbial Cultivation Techniques

15

Fig. 7 Fed-batch culture of recombinant B. subtilis with glucose concentration controlled at 1 g/L. Filled circles: dry cell concentration (g/L); filled squares: viable cell concentration (¥109 cells/mL); empty squares: glucose concentration (g/L); filled triangles: enzyme activity (U/mL); empty triangles: specific enzyme activity (U/mg protein)

and 129 U/mL, respectively (Fig. 7). However,at 10 g/L of glucose, cell concentration increased to 44 g/L, while the maximum specific enzyme activity decreased to about 2.9 U/mg protein. From these results, control of glucose concentration at low levels throughout the cultivation was thought to be efficient for obtaining high cell concentration and for improving enzyme activity. When glucose concentration was controlled at 10 g/L, the accumulated inhibitory metabolites were identified to be propionic, acetic, and lactic acids, whose concentrations were 18.5 g/L, 5.4 g/L, and 0.28 g/L, respectively (Fig. 8). When glucose concentration was controlled at 0.2 g/L, the maximum concentration of propionic acid was about 3.0 g/L, one sixth of the propionic acid concentration in the culture with glucose controlled at 10 g/L. The maximum concentrations of acetic and lactic acids were 1.9 and 0.06 g/L, respectively. After 8 h of cultivation, concentrations for lactic and acetic acids sometimes decreased, suggesting consumption of these organic acids by cells. These results indicate that the productions of these organic acids were favored by high

16

E.Y. Park

Fig. 8 Profiles of lactic acid a, acetic acid b and propionic acid c concentration in fed-batch cultures, with controlling glucose concentration at 10 g/L (filled square), 1 g/L (filled triangle), and 0.2 g/L (filled circle), respectively

glucose concentrations. Low glucose concentration in culture broth is favorable for high cell-density culture and enzyme production. However, it was practically difficult to control glucose concentration at 0.2 g/L without depletion of glucose. In the case of recombinant B. subtilis 1A96, containing insecticide gene in chromosome, the culture was performed by controlling glucose concentration at 2 g/L. The cell concentration and insecticidal potency were 104 g/L and 394 kU/mL, which were 2.2 and 1.6 times, respectively, as high as those of the fed-batch culture without glucose control [22]. Aside from glucose, excess supply of the nitrogen source also causes accumulation of inhibitory metabolites. However, since it was difficult to measure the nitrogen source directly, the concentration of L-amino acid contained in the nitrogen source was monitored by an automatic biotech analyzer, instead of the nitrogen source itself [23]. To improve production of the recombinant gene product in Bacillus brevis, fed-batch cultures with controlled L-amino acid

Recent Progress in Microbial Cultivation Techniques

17

concentration were initiated. Cell growth increased with an excess of nitrogen source, but the cloned gene production was drastically decreased. When L-amino acid was controlled at 2 or 6 mM, the accumulation of ammonium ion in the culture broth was 20% lower than that in a conventional fed-batch culture. By controlling it at 5 mM, with additional feeding of Ile- and Asnenriched nitrogen source, the maximum a-amylase activity was increased from 5.14 kU/mL to 12.05 kU/mL; the specific enzyme activity, from 0.77 to 2.63 kU/mg dry cell, in comparison to those values obtained with a conventional fed-batch culture [23]. These results indicate that glucose or nitrogen source control is a very important factor for the cultivation of Bacilli, and may be applied to the cultivation of other B. subtilis strains or other Bacilli. 2.4.2 Recombinant Yeast Researche [9, 18] on cloned gene products from recombinant yeasts have revealed that an inhibitory metabolite, ethanol, accumulates in the culture broth during the aerobic fermentation of yeast. It is therefore necessary to keep the concentration of the inhibitory metabolite at a low level by feeding glucose at the optimal rate that restricts metabolite formation while maintaining a relatively high growth rate. Under such circumstances, fuzzy reasoning seems to be most appropriate for the cultivation of recombinant yeasts. Measured culture parameters such as the cell, DO, ethanol and glucose concentrations on-line were used [16].When the MICOC was used, even when the glucose concentration was restricted to a low level (0.15 g/L), during the fermentation of recombinant S. cerevisiae cells monitored with a glucose analyzer about 20 g/L of ethanol was found to accumulate in the culture broth [9], indicating the difficulty in maintaining ethanol at a low concentration. By using this strategy, a high ethanol concentration may severely affect gene expression [24]. The decrease in cell growth of S. cerevisiae 20B-12 at an ethanol concentration of 20 g/L was about 10%. Gene product (a-amylase) activity was half that of the culture with ethanol controlled at 2 g/L. The fuzzy controller used to regulate the concentrations of both ethanol and glucose simultaneously was used for the fed-batch culture of recombinant S. cerevisiae 20B-12/pNA7 [24]. The concentrations of glucose and ethanol were controlled at 0.1 and 1.9 g/L, respectively, by the fuzzy controller, and the results are shown in Fig. 9 [25]. For a culture time of 4 h, 5 g/L of glucose was consumed, at which point the control of both glucose and ethanol began. Cell growth decreased at 5 h and was then maintained until 30 h. The maximum a-amylase activity was 392 U/mL, and the specific enzyme activity 12.4 U/mg dry cells. Cloned gene expression was compared between cultures with both glucose and ethanol control (fuzzy control) and with only glucose concentration control (adaptive control).With the SUC2 promoter, the enzyme and specific enzyme activities using the fuzzy controller increased by 1.9 and 3.3 times compared with those cultures not controlling

18

E.Y. Park

Fig. 9 Expression of a-amylase gene of S. cerevisiae 20B-12 harboring plasmid pNA7 by controlling both glucose and ethanol concentrations using a fuzzy controller. Filled circles: cell concentration (g/L); filled triangles: a-amylase activity (U/mL); empty triangles: specific a-amylase activity (U/mg dry cells)

ethanol concentration, even though the cell growth was similar, while with the PGK promoter these activities were three and four times as high, respectively (Table 3). The difference in the results between the two methods of control lies in the ethanol concentration. The ethanol concentration was controlled at 2 g/L in the culture using the fuzzy controller, whereas it increased to 17 or 19.8 g/L when ethanol concentration was not controlled. The fuzzy controller was therefore shown to be very useful for controlling the concentrations of both glucose and ethanol. Table 3 Comparison of mouse a-amylase production in fed-batch cultures of recombinant S. cerevisiae 20B-12 with controlof both glucose and ethanol (fuzzy control), and of glucose only (adaptive control)

Promoter

SUC2

Control method Glucose concentration (g/L) Ethanol concentration (g/L) Cells (g/L) a-amylase activity (U/mL) Specific activity (U/mg cell) Reference

Adaptive 0.26 17.0 42.9 90.7 2.3 [21]

PGK Fuzzy 0.48 2.0 40.0 175.0 7.7 [24]

Adaptive 0.58 19.8 40.3 128.6 3.2 [21]

Fuzzy 0.30 1.9 55.0 392.4 12.4 [24]

Recent Progress in Microbial Cultivation Techniques

19

3 Image Analysis for Characterization of Mycelium Although substantial advancements had been made in the study of molecular biology in recent years, progress in mold-culturing techniques has proceeded at a slow pace, despite the importance of mold microorganisms in the production of many industrial products. In most of the cases, these microorganisms are used for producing chemical stocks, pharmaceutical products, and enzymes, but relatively little is known about the ability of mold microorganisms to change mycelial morphology, or the correlation between this morphology and productivity. Since molds show many morphological variations in submerged fermentation, it is difficult to control the morphology and so optimize the reactor performance. There are many kinds of morphology known in the cultivation of mycelia. However, mycelial morphology was difficult to characterize because there were no tools to visualize it. Since the geometrical interpretation of mycelial morphology is hard to define, and moreover is very hard to quantify, so-called concentration, cultivations of fungi, or actinomycetes are focused only on production rate with regards to a reactor operation including microbial kinetics. However, although some products were produced by cultivation of mycelium using these ideas, it is still difficult to elucidate the effect of mycelial morphology on the production of useful microbial products. Recently, analysis of the mycelial morphology has been carried out, not by visual inspection, but by a microscope that is linked to a computer, which can then characterize the mycelial morphology via Software specifically designed for such a purpose. This technology gives us the possibility of improving cultivations of fungi or actinomycetes. Thomas [26] revealed that filamentous microorganisms could be characterized by passing information on their morphology to a computer for analysis. Developments in image analysis techniques for analyzing the mycelial image captured by a camera have been quite remarkable in recent years. This image analysis technology has been used to investigate characterization, mycelial growth kinetics, and the physiology of mycelium. In this review, the applied fields of image analysis are introduced. 3.1 Morphological Classification Image analysis extracts information from a picture (image) via a computer and software. First, the image is captured from a microscope on which a CCD camera is mounted. Usually, resolution of the image is 512¥512 pixels with 256 gray values. The image is made ready for analysis by segmenting it by two thresholds and converting it into bits (binarization) using computer. To obtain a clear image for measuring, the image is treated with a low-pass filter.

20

E.Y. Park

Table 4 Morphological parameters for image analysis

Morphology

Parameter

Description

Dispersed form (filamentous form)

Hyphal length Branch length Hyphal width Branching frequency Number of tips Hyphal growth unit

Length of hypha in a mycelium Length of individual hyphal branch Width of hypha Numbers of branches in a mycelium Numbers of branches plus two Hyphal length divided by the number of tips

Clumps (entangled filamentous form)

Convex perimeter Convex area Compactness

Length of the perimeter Area inside the convex perimeter Actual area of the hyphae in a clump divided by the convex area A measure of the irregularity of the perimeter of a clump

Roughness

Pellets

Convex perimeter Convex area ratio

Clumps Ratio of the convex area of a pellet core to that of the whole pellet

Mycelial morphologies are classified by defining morphological parameters of mycelial images. Thomas defined morphological parameters [26] as shown in Table 4. Morphology is split into three groups: dispersed (filamentous) form; clumped (entangled) form; pellet form. In the case of tylosin production in a culture of Streptomyces fradiae, it is difficult to classify into these three morphologies because there are no guidelines for judging dispersed and clumped forms. Figure 10 is an image selected from a culture broth sample. The resolution of this monochrome image is 512¥512 pixels with 256 gray values (Fig. 10a). The image was then segmented by two thresholds to obtain an image in which the objects are white (gray value 256) and the background is black (gray value 0, Fig. 10b). After binarization, the image was separated out by using a different color for it to the others (Fig. 10c). To obtain a clear image for measuring, the image was treated with low-pass filter. Pellet (Fig. 10d) and mycelial images were extracted from the binary image by an opening process. Finally, the areas of the pellet and mycelia, the mycelial convex perimeter, and the pellet length were measured. If the captured image has a pellet core, then the morphology is classified as a pellet. Otherwise the morphology is filamentous or entangled form [27]. How to classify into these two forms is by using parameters of mycelial area (Ma) and convex area (Mp). If (Ma /M p2 ) ≥ 0.014 and Mp ≥ 420, the captured image is classified as an entangled filamentous form. If (Ma /M p2 ) ≥ 0.014 and Mp ≥ 420, the captured image is classified as an entangled or Mp <420, then the captured image is judged to be

Recent Progress in Microbial Cultivation Techniques

21

Fig. 10 Morphology of S. fradiae at a culture time of 84 h. a Original image on a TV monitor (512¥512 pixels, 256 gray values); b and c binary and color images, respectively; d pellet separated by an opening process. One centimeter in the photographs represents 60 mm

a filamentous form. The values 0.014 and 420 were obtained from a statistical evaluation of 250 images, as shown in Fig. 11. The parameters (Ma /M p2 ) and Mp clearly indicate two distributions with boundaries of 0.014 and 420, respectively [27]. Using these criteria, the morphological change in S. fradiae culture was investigated [27]. The area ratios of the three types of morphology were compared in cultures under high, medium, and low shear conditions, for which the conditions selected were agitation at 900 and 400 rpm, and culture in the airlift reactor, respectively. The results are shown in Fig. 12. In the case of agitation at 900 rpm, the mycelial morphology was filament and entangled filament. In the early culture phase, the mycelia were almost all filaments, but in the midculture phase this filamentous morphology was modified into entangled filaments. Finally, the entangled filaments that formed collapsed and the morphology became filamentous again due to the mechanical agitation.At 400 rpm, the filamentous morphology decreased until it comprised only 13% of the whole, but entangled filament morphology increased with culture time to reach 47%. Pellets increased and made up an average of 40% throughout the cultivation. On the other hand, in the airlift reactor, filaments increased until the

22

E.Y. Park

Fig. 11 Distribution map of Ma/Mp2 and Mp. The boundary values 0.014 and 420 were used to judge the morphology as filament or entangled filament, respectively. Images of pellets labeled with PI a and FITC b from flask culture of S. fradiae at 5 d

Fig. 12 Area ratios of filament a, entangled filament b, and pellet c during cultures with agitation at 900 rpm (gray), 400 rpm (black), and in an air-lift reactor (hatched)

mid-culture phase, but decreased to 18% by the end of the culture at 160 h. At 20 h, filaments and entangled filaments occupied the same ratio, but the entangled filaments had disappeared and formed pellets by 50 h. The pellet ratio in the reactor increased to 80% by 160 h. This time course indicates that pellets were formed from both filaments and entangled filaments.

Recent Progress in Microbial Cultivation Techniques

23

3.2 Mycelial Physiology Most studies concentrate on analyzing mycelial morphology, but there are a few reports regarding the mycelial pellet intrastructure and prediction of its physiological state. The most important thing is to get information about mycelial physiology from image analysis. Intrastructure of the mycelial pellet was visualized to clarify whether the mycelia inside the pellet are biologically active or not using a confocal scanning laser microscope [28]. In order to distinguish viable mycelia among mycelial pellets, fluorescein isothiocyanate (FITC) and propidium iodide (PI) were used. By analyzing the fluorescence of the image using image analysis, the inactive fraction of pellet present inside the pellet was visualized in the culture of Streptomyces fradiae, which produces tylosin. Exciting and emission waves for detection of the FITC fluorescence were 495 nm and 520 nm, respectively; 493 nm and 630 nm, respectively for the PI fluorescence. FITC, which binds nonspecifically to cell surface proteins, was used as a dye to label microorganisms by the procedure of Block and Srienc [29]. If the FITC binds to the mycelial proteins, it fluoresces, and the fluorescence is detected at a wavelength of 520 nm. On the other hand, PI did not penetrate into intact cells but crossed the membranes of dead cells, which caused dead cells to fluoresce [30]. This fluorescence was detected at a wavelength of 630 nm. Captured images were analyzed using an IBAS image analysis system (Kontron Co., Germany) and a NIH image analysis system. Figure 13a and b show the image intensities of a pellet labeled with PI and FITC, respectively, in a flask culture of S. fradiae at 5 d, useful for visualizing the intrastructure of the pellet for which the diameter was 340 mm. Both images showed a high intensity region in the center of the pellet. The PI fluorescence intensity suggests that there are viable mycelia in the cortex of the pellet. However, in the center, the mycelia might be dead because of the strong intensity compared with that in the cortex of the pellet. On the other hand, the fluores-

Fig. 13 Color image of the pellet of S. fradiae labeled with PI a and FITC b

24

E.Y. Park

cence intensity of FITC was strongest in the center of the pellet, decreasing gradually towards the outside. This indicates that the mycelia are more concentrated in the center of the pellet. If the total pellet area is atotal, the inactive fraction of the pellet (finactive) can be obtained by dividing the area of the inactive region inside the pellet ainactive by atotal : ainactive finactive = 01 (15) atotal The total area obtained from the image was 7.69¥104 pixels and the area of the inactive fraction was 7.93¥103 pixels. Therefore, the inactive fraction of the pellet constituted 10.3% of the total area in this case. The area of the inactive fraction was calculated from Eq. 15 based on the results of image analysis. In the case of the airlift reactor culture, the pellet area increased up to 3¥104 mm2 on average and the inactive fraction of the pellet fluctuated from 4% to 11.8% of the total area. On the other hand, in the case of the fermentor culture, the pellet area was 1.3¥104 mm2 on average and the inactive fraction of the pellet remained at 0.5% of the total area throughout the culture. The reason why the inactive fraction was a high percentage initially of the total pellet area might have been due to the presence of aged mycelial pellets in the inoculum. If a circular section is made through the pellet, the pellet diameter (d) is calculated as follows: 9 atotal d=2 8 p

f

(16)

Since the diameter of the section is assumed to be similar to that of its own pellet, the pellet diameter was calculated from Eq. 16 to be 195 mm in the airlift reactor and 129 mm in the fermentor. If the thickness of the active mycelial layer from the surface of the pellet is lactive, the inactive fraction of the pellet is represented as follows:

p (d –2lactive)2 001 4 d – 2lactive finactive = 001 = 06 pd2 d 6 4

$

%

2

Therefore, the thickness of the active mycelial layer is: d(1 – d01 finactive ) lactive = 003 2

(17)

(18)

The active mycelial layer of the pellet, as calculated from Eq. 18, was 110 mm in the case of the pellet in Fig. 13. Since the inactive fractions of pellets in the fermentor and in the airlift reactor were 0.5% and 9.4% on average, respectively, of the total pellet area, the active mycelial layers were calculated to be 60 mm thick in the fermentor and 68 mm thick in the airlift reactor. This indicates that

Recent Progress in Microbial Cultivation Techniques

25

the biologically active mycelia are present about 64 mm from the surface of the pellet. In the case of Aspergillus oryzae, the active mycelial layer was reported to be 145 mm thick [31]. This means that the thickness of the active mycelial layer depends on the species used, the pellet structure and operating conditions. If the pellet density is g, total amount of the mycelia (X) is calculated as follows: 4p X = 5 d 3 · g · n, 3

(19)

where n denotes the number of pellets in a reactor. If the differences in pellet density between fermentor and airlift reactor are negligible, the ratio of pellet number in the fermentor culture to that in the airlift reactor culture is rewritten as follows:

T Y T Y

X 63 (n)fermentor (d) fermentor r = 06 = 002 . (n)air – lift X 63 (d) air – lift

(20)

Since intracellular nucleic acid concentrations of mycelia after 2 d culture in the fermentor and air-lift reactor were 1.0 g/L and 0.5 g/L respectively, and remained at similar values until 148 h of culture (data not shown), the ratio of the pellet number in the fermentor to that in the air-lift reactor, r, was calculated to be 6.9. If the differences in mycelial density between the two reactors are negligible, the active mycelial volume per pellet, volactive, is:

p 3 volactive = 3 {d 3 – (d – 2lactive) }. 6

(21)

Therefore, the active mycelial volume per pellet in airlift reactor and fermentor cultures were calculated to be 3.8¥106 mm3 and 1.1¥106 mm3, respectively. The ratio of amount of the active mycelia in fermentor culture to that in airlift reactor culture (Ractive) is defined as follows: xfermentor (volactive · g)fermentor · nfermentor Ractive = 03 = 00006 xair – lift (volactive · g)air – lift · nair – lift (volactive)fermentor = 0302 · r (volactive)air – lift

(22)

where x indicates amount of active mycelia. From Eqs. 21 and 22, Ractive was predicted to be 2.1. The ratio of tylosin production rate (Rpr) between two reactors can be represented as follows:

Çfermentor xfermentor · Çfermentor /t Rpr = 0309 = Ractive · 04 Çair – lift xair – lift · Çair – lift /t

(23)

26

E.Y. Park

where Ç indicates the specific tylosin production rate. The maximum Çfermentor and Çair-lift were calculated to be 4.8 and 3.8 g tylosin/g dry cell weight, respectively (data not shown). Therefore, the Rpr was predicted to be 2.7. The final tylosin production rate in the fermentor culture reached 0.78 g/L/d, while it reached 0.31 g/L/d in the airlift reactor (data not shown), which indicates that the production rate in the fermentor was 2.5 times higher than that in the airlift reactor. This coincides with the predicted Rpr. This explains why the tylosin production capacity in the fermentor was 2.5 times higher than that in the airlift reactor. The difference in the active mycelial fraction between the fermentor and the airlift reactor cultures resulted in the difference in the tylosin production rate between them. Contrary to Streptomyces fradiae, which produces tylosin [27], to investigate the intracellular product distribution, Mortierella, a producer of the polyunsaturated fatty acid arachidonic acid (5, 8, 11, 14-cis-eicosatetraenoic acid,AA) was chosen. Mortierella fungi have been known to accumulate AA to approximately 70% of the total fatty acids [32]. The AA, which is a C-20 polyunsaturated fatty acid, and is known to be a precursor of prostaglandins and leukotrienes [33], has recently attracted great interest due to its unique biological activity. Using a fluorescence microscope, the intrastructure of pellet was visualized, and the accumulation of lipid bodies and their distribution were clarified. In order to distinguish mycelia and lipid bodies inside mycelial pellets, FITC and Nile red were used for staining mycelia and lipid body, respectively. Images of the intact pellet, and cross-sections stained with the Nile red and FITC are shown in Fig. 14. It was hard to observe the distribution of mycelia inside the pellet with the intact image (b), but the distribution of mycelia and lipid are clearly shown by the fluorescence intensities of the FITC (c) and Nile red (d), respectively [34]. Towards the edge of the pellet the mycelial density was high, but closer to the center the mycelial density decreased. The oily materials inside the pellet are accumulated towards the edge of the pellet judging by the Nile red image, which is similar to the results shown in the FITC image. This indicates that distributions of mycelia and lipid are not uniform inside the pellet. Profiles of FITC and Nile red fluorescent intensities in the cross-section of the pellet are shown in Figs. 14e and f, respectively. In the middle of the pellet the intensities of FITC and Nile red are low, which is evidence for the presence of a cavity inside pellet. From both images we can deduce that the fatty acid production occurred towards the edge of the pellets, the area of dense mycelia. This profile was used for the calculation of the cavity ratio in the pellet. The pellet intrastructure is constructed as a three-dimensional image in Fig. 15. The intrastructure of the pellet was doughnut-shaped, and the center is shown to be hollow. From these two images the pellet intrastructure was unveiled: the pellet was hollow inside, and mycelia and lipids accumulated on the verge on the pellet. The reason for the difference in pellet intrastructure between S. fradiae and M. alpina is not clear right now. It may be a phenomenon specific to the pellet of M. alpina.

Recent Progress in Microbial Cultivation Techniques

27

Fig. 14 Images of stained cross-sections of a pellet with FITC c and Nile red d from culture at 162 h of M. alpina. a and b are the schematic diagram of the pellet cross-section and the intact pellet cross-section respectively. e and f denote profiles of FITC and Nile red fluorescence intensities along the direction of the arrow shown in c and d, respectively. Dotted lines indicate the presence of bottom or peak in both images

Fig. 15 Three-dimensional images of stained cross-section of pellet with FITE a and Nile red b of Fig. 14. Color being close to red indicates high fluorescence intensity

28

E.Y. Park

From the results presented in this study, it is evident that M. alpina pellets are elliptical, but their intrastructure have a doughnut shape with a cave inside the pellet, which was shown using image analysis with staining by two different fluorescent dyes. This intrastructure of the fungal pellet is a very interesting phenomenon that has not otherwise been revealed. Application of the technique used here opens up the possibility of analyzing the intrastructure of fungal pellets and understanding the new variations in morphological fungal image or the fungal genetic variation related to this morphology. 3.3 Monitoring Mycelial Formation Using a Flow-Through Chamber Features of fungal culture morphology, such as the shape and size of the pellets, can be determined macroscopically. However, fungal morphology might be predetermined from the beginning of germination of spores. Microscopes have been widely used to study early developmental morphology, and the microscopic morphology, such as the form of individual hyphal elements, total hyphal length and the number of tips on the hyphal element, can be determined. In order to quantify the microscopic morphology of fungal microorganisms from fermentation samples, image analysis techniques are normally used. However, a detailed analysis of the true growth kinetics of mycelia is complicated due to the great variations between growth of different hyphae. Therefore, a flow-through cell was designed to study the growth of individually identified hyphae, and to aid with quantitative studies of the morphologies of fungi under defined conditions [35]. A schematic diagram of the flow-through chamber is shown in Fig. 16. The flow-through chamber (Fig. 16a) is located between two glass slides (76 mm¥ 26 mm¥1 mm in depth) separated by a 40 µm Teflon sheet as a spacer. The lower glass slide has two slits through which the medium is introduced and withdrawn. The flow-through chamber is fastened onto a stainless steel (SUS-304) frame with eight bolts. The dimension of the frame is 38 mm¥ 101 mm¥5 mm. A silicone sheet of 0.2 mm depth is used as a spacer between the flow-through chamber and the stainless steel frame. The lower part of the stainless steel frame is equipped with 1 mm (i.d.) tubes at the medium inlet and outlet. The assembled flow-through chamber is steam-sterilized before use. Using the flow-through chamber, hyphal length and tip numbers formed were measured (Fig. 17), and the specific rates of hyphal growth (ml) and tip formation (mtip) are calculated via the following kinetics: 1 dl ml = 2 · 4 l dt

(24)

1 dn mtip = 2 · 5 l dt

(25)

Fig. 16 Experimental system for on-line image analysis of hyphal elements (A) and picture sequence of a single hyphal element (B). The system consists of a CCD camera, a microscope, and a flow-through chamber mounted on the microscope. Temperature of the flow-through chamber was controlled at 28 °C using a Thermo Plate during the cultivation. The micro-pump with flow rate control was used to pump fresh medium through the chamber.All experiments were carried out at a flow rate of 5.2 mL/min, of which the space velocity corresponded to 10. The pictures in B were taken at 12 h, 14 h, 16 h, 18 h, 20 h, and 24 h after spore germination, as indicated by numbers in the Figure

Recent Progress in Microbial Cultivation Techniques 29

30

E.Y. Park A

for

B

Fig. 17 Development of total length of hyphal (A) and tips (B). Total concentrations of carbon and nitrogen sources were 2 g/L. Tested C/N ratios are from 0.18 to 30. Filled squares: C/N=0.18; empty squares: C/N=8.88; filled triangles: C/N=13.3; empty triangles: C/N=20; empty circle: C/N=30

where the l and n denote total hyphal length and total tip numbers, respectively. Experimental data were fitted well by Eqs. 24 and 25 (the correlation factors were higher than 0.96). Hyphae grow and form a branch at a point, and the branches grow and make a tip. Therefore, a connection between tip and branch is inevitable. To understand the relationship between tip extension and branch formation, the tip extension rate of branches (qtip) and the branch formation rate of tips (qb) are defined as follows: 1 dl qtip = 2 · 4 n dt

(26)

1 dn qb = 2 · 5 l dt

(27)

Equations 26 and 27 are determined using the data in Fig. 17. The tip formation rate and the tip extension rate are important factors because they play a critical role in a fungal morphology. When one tip is formed from a branch, the tip extension rate may be related to both the hyphal growth rate and tip formation rate, which might be a strain-specific feature. From experimental results, the tip extension rate of one tip increased exponentially with the increase in the specific hyphal growth rate (data not shown) and the following empirical equation was derived: qtip = a · ebml

(28)

where a and b are empirical constants, 3.30 (mm/tip/h) and 3.41 (h), respectively, with a correlation coefficient of 0.97.

Recent Progress in Microbial Cultivation Techniques

31

Fig. 18 Correlation between tip extension rate and specific tip formation rate. Symbols are the same as those in Fig. 17

On the other hand, the specific hyphal growth rate showed a linear relation to the specific tip formation rate, and these were empirically correlated as follows:

ml = k1 + k2 mtip

(29)

where k1 and k2 also denote empirical constants and determined to be 0.08 (h–1) and 1.55 (–), respectively, with a correlation coefficient of 0.74. From Eqs. 28 and 29, the tip extension rate can be expressed as follows: qtip = a · eb (k1 + k2 mtip ) = a · ebmtip

(30)

where a and b are constants which were determined as 4.34 (mm/tip/h) and 5.29 (h), respectively. From experimental results (Fig. 18), empirical constants a and b were determined as 5.21 (mm/tip/h) and 4.75 (h), respectively, with a correlation coefficient of 0.61, which deviated from the constants in Eq. 30 by 13%. Therefore, Eq. 30 may not exactly explain the correlation between tip extension rate and specific tip formation rate because of low correlation coefficient, but develops a tendency to correlate both rates and may be a good enough to explain the correlation between specific rates of tip formation and tip extension. This research demonstrates that the hyphal growth rate is proportioned to the tip formation rate of mycelia, and the higher hyphal growth is the more tip are formed. Morphological data of Aspergillus oryzae and M. alpina are compared in Table 5. Sophr et al. [36] investigated the morphological parameters in glucose concentration lower than 500 mg/L, that corresponded to 200 mg/L of carbon concentration. The hyphal growth and branching formation rates of M. alpina

32

E.Y. Park

Table 5 Comparison of morphological parameters between Aspergillus oryzae and M. alpina

Morphological parameter

Aspergillus oryzae

M. alpina

Specific hyphal growth rate (h–1) Tip extension rate (mm/tip/h) Branching formation rate (tip/mm/h) Reference

0.35a 75a 1.6¥10–3, a [36]

0.84b 65b 6.5¥10–3, b [35]

a

Estimated maximum value; b average value for carbon concentration lower than 10 g/L.

were 2.4- and 4-fold as high as that of Aspergillus oryzae, respectively. But the tip extension rate of both fungi was similar. This means that M. alpina shows high hyphal growth and high frequency of branch formation, but hyphal growth rate from one tip does not differ much from A. oryzae. No study on the clear relationship between microscopic morphology and pellet formation has been reported. However, in the case of M. alpina, when rates of both hyphal growth and tip formation were high, the macroscopic morphology showed a tendency to form pellets covered by filamentous mycelia; on the other hand, when hyphal growth and tip formation rates were low, the morphology in the submerged culture was inclined towards filamentous morphology (data not shown). The flow-through chamber was a useful tool for investigating the fungal morphology in a very limited period of time.

4 Future Prospects for Microbial Cultivation Various technologies and operations have been developed which have made the maximization of bioproducts from microbial cultivation possible. If we can predict the microscopic morphology, it will be very useful for the culture of mycelial microorganisms. Using the flow-through chamber, if the morphological parameters of many types of fungi are accumulated and then stored in a data-bank of fungal morphology, the morphological information should be very useful for optimizing cultivation, and should lead to a better understanding of the cultivation of mycelial microorganisms than at present. We have demonstrated that this flow-through chamber system can be used for evaluations of both morphological development and morphological parameters, and will contribute to the elucidatation of fungal morphology. However, new bioproducts from organella or microorganisms that are difficult to cultivate may be expected. In order to produce these new bioproducts, more research into new types of reactor or operation are required. Nanoreactors will be used for production of very specific bioproducts, which will open up a new era in cultivation.

Recent Progress in Microbial Cultivation Techniques

33

References 1. Hobby GL (1985) Penicillin. Meeting the Challenge.Yale University Press, New Haven, CN 2. Kuriyama H, Seiko Y, Murakami T, Kobayashi H, Sonoda Y (1985) J Ferment Technol 63:159 3. Taniguchi M, Kotani N, Kobayashi T (1987) Appl Microbiol Biotechnol 25:438 4. Lee CW, Chang HN (1987) Biotechnol Bioeng 29:1105 5. Taniguchi M, Kotani N, Kobayashi T (1987) J Ferment Technol 65:179 6. Park YS, Ohtake H, Fukaya M, Okumura H, Kawamura Y, Toda K (1989) J Ferment Technol 68:315 7. Yano T, Mori H, Kobatashi T, Shimizu S (1981) J Ferment Technol 59:295 8. Mori H, Yamane T, Kobayashi T, Shimizu S (1983) J Ferment Technol 61:391 9. Shimizu K, Morikawa M, Mizutani S, Iijima S, Matsubara M, Kobayashi T (1988) J Chem Eng Jpn 21:113 10. De Deken (1966) J Gen Microbiol 44:157 11. Konstantinov K, Kishimoto M, Seki T, Yoshida T (1990) Biotechnol Bioeng 36:750 12. Lin KH, Iijima S, Hishinuma F, Kobayashi T (1989) Appl Microbiol Biotechnol 32:313 13. Kishimoto M, Moo-Young M, Allsop P (1991) Bioprocess Eng 72:110 14. Mamdani EH (1976) Int J Man-Machine Studies 8:669 15. MacVicar-Whelan PJ (1976) Int J Man-Machine Studies 8:687 16. Park YS, Shi ZP, Shiba S, Chantal C, Iijima S, Kobayashi T (1993) Appl Microbiol Biotechnol 38:649 17. Stouthamer AH, van Verseveld EJ (1987) Trends Biotechnol 5:149 18. Georgiou G (1988) AIChE J 34:1233 19. Iijima S, Kawai S, Mizutani S, Taniguchi M, Kobayashi T (1987) Appl Microbiol Biotechnol 26:542 20. Ohshima T, Zhang XL, Iijima S, Kobayashi T (1992) Kagaku Kougaku Ronbunsyu 18:693 21. Park YS, Kai K, Iijima S, Kobayashi T (1992) Biotechnol Bioeng 40:686 22. Cayuela C, Kai K, Park YS, Iijima S, Kobayashi T (1993) J Ferment Bioeng 75:383 23. Park YS, Dojima T, Okabe M (1996) Biotechnol Bioeng 49:36 24. Shiba S, Nishida Y, Park YS, Iijima S, Kobayashi T (1994) Biotechnol Bioeng 44:1055 25. Park YS, Shiba S, Iijima S, Hishinuma F, Kobayashi T (1993) Biotechnol Bioeng 41:845 26. Thomas CR (1992) Trends in Biotechnol 10:343 27. Tamura S, Park YS, Toriyama M, Okabe M (1997) J Ferment Bioeng 83:523 28. Park YS, Tamura S, Toriyama M, Okabe M (1997) J Ferment Bioeng 84:483 29. Block DE, Srienc F (1991) Biotechnol Tech 5:95 30. Swope KL, Flickinger MC (1996) Biotechnol Bioeng 52:340 31. Carlsen M, Spohr AB, Nielsen J, Villadsen J (1996) Biotechnol Bioeng 49:266 32. Shinmen Y, Shimizu S, Akimoto K, Kawashima H, Yamada H (1989) Appl Microbiol Biotechnol 31:11 33. Gill I, Valivety R (1997) Trends Biotechnol 15:401 34. Hamanaka T, Higashiyama K, Fujikawa S, Park EY (2001) Appl Microbiol Biotechnol 56:233 35. Park EY, Hamanaka T, Higashiyama K, Fujikawa S (2002) Appl Microbiol Biotechnol 59:706 36. Spohr A, Dam-Mikkelsen C, Carsen M, Nielsen J, Villadsen J (1998) Biotechnol Bioeng 58:541

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 35– 62 DOI 10.1007/b94191 © Springer-Verlag Berlin Heidelberg 2004

Clarification of Interactions among Microorganisms and Development of Co-culture System for Production of Useful Substances Masayuki Taniguchi (✉) · Takaaki Tanaka Department of Materials Science and Technology, Niigata University, Ikarashi 2-8050, Niigata 950-2181, Japan [email protected]

1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.2.4

Introduction . . . . . . . . . . . . . . . . . . Interactions of Two Microbial Populations . . Production of Useful Substances by Co-culture Hydrogen and Methane . . . . . . . . . . . . Acids and Ethanol . . . . . . . . . . . . . . . Bioactive Compounds . . . . . . . . . . . . . Flocculant and PHB . . . . . . . . . . . . . .

2 2.1 2.2 2.3

Ethanol Production from a Mixture of Glucose and Xylose by Co-culture Bioconversion of Biomass Hydrolysate to Ethanol . . . . . . . . . . . . . Effect of Oxygen Supply on Ethanol Production from Glucose or Xylose . Ethanol Production by Co-culture of P. stipitis and a Respiratory-deficient Mutant of S. cerevisiae . . . . . . . . . . . . . . . . . . . . . . . . . . . . Properties of the Co-culture System with Membrane Bioreactor . . . . . . Ethanol Production by P. stipitis and S. cerevisiae in the Co-culture System with Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.4 2.5

3 3.1 3.2 3.3 4 4.1 4.2 4.3 4.4

5

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

36 36 38 38 40 41 41

. . . . . .

42 42 43

. . . .

44 47

. .

49

. . . .

. . . .

51 51 52 52

. . . . . .

55 55 55

. .

56

. .

58

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

59

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

60

Simultaneous Production of Bifidobacterial and Propionibacterial Cells Using a Co-culture System with Membranes . . . . . . . . . . . . . . . Production of Bifidobacterial Cells as a Probiotic . . . . . . . . . . . . Cooperative Interactions between B. adolescentis and P. freudenreichii . Cell Production Using Our Co-culture System with Membranes . . . . .

. . . . . . .

. . . .

Kefiran Production by L. kefiranofaciens under Culture Conditions Mimicking the Natural Co-culture System . . . . . . . . . . . . . . . . . Properties and Production of Kefiran . . . . . . . . . . . . . . . . . . . . Kefiran Production by a Combination of L. kefiranofaciens and Yeast Cells Kefiran Production by L. kefiranofaciens under the Conditions Established by Mimicking the Existence of Yeast Cells . . . . . . . . . . . . . . . . . . Kefiran Production by a Single Culture of L. kefiranofaciens Mimicking a Natural Co-culture System . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions and Prospects

References

. . . . . . .

36

M. Taniguchi · T. Tanaka

Abstract Co-culture systems containing two microorganisms for the production of useful substances are described. We developed a novel co-culture system composed of two fermentors and two microfiltration modules. The proposed co-culture system allowed regulation of the dissolved oxygen concentration at a level suitable for an individual microorganism in each fermentor, as well as the successful exchange of culture medium between two fermentors. By co-culture, using a combination of Pichia stipitis and Saccharomyces cerevisiae, ethanol was produced efficiently from a mixture of glucose and xylose. Moreover, the useful probiotic cells were simultaneously produced with a high productivity by our co-culture using a combination of Bifidobacterium and Propionibacterium. Kefiran production by Lactobacillus kefiranofaciens alone under the culture conditions, established by mimicking the presence and activities of yeast cells in kefir grains, was also investigated. The results obtained showed that under the culture conditions established by mimicking the actions of yeast cells on L. kefiranofaciens in kefir grains, the amount of kefiran produced was enhanced, even when the lactic acid bacterium alone was used. Keywords Co-culture of microorganisms · Cooperative interactions · Commensal relationship · Synergism · Production of useful substances Symbols and Abbreviations DO Dissolved oxygen, ppm OTR Oxygen transfer rate, mg L–1 h–1 qO2 Specific oxygen uptake rate, mg (g cell)–1 h–1 (qp)max Maximum specific ethanol production rate, g (g cell)–1 h–1

1 Introduction 1.1 Interactions of Two Microbial Populations Recent bioindustries utilizing microorganisms have concentrated on systems dominated by a single type of microorganism [1–4]. Many useful substances such as amino acids, nucleotides, antibiotics, and enzymes have been produced successfully by using a single powerful microorganism [1–4]. On the other hand, among commercial processes, alcohol brewing, biological waste-water treatment (activated sludge, methane fermentation, and so on), and manufacture of dairy and conventional fermenting products (miso, soy sauce, pickles) require multiple microbial species [5, 6]. Moreover, mixed populations of microorganisms are the rule rather than the exception in natural systems. To produce high-quality beverages and foods and to understand the natural cycles of numerous elements on our planet, we must concentrate on the analysis of microbial interactions. In nature, distinct microbial populations frequently interact with each other. The categories used to explain these interactions represent a conceptual classification system. It is difficult to accurately classify many specific cases ob-

Clarification of Interactions

37

Table 1 Interactions of two microorganisms

Interaction

Mutualism Protocooperation Commensalism Neutralism Amensalism Competition Parasitism Predation

Microorganism A

a

+ + + 0 – – – –

+ + 0 0 0 – + +

0: no effect; +: positive effect; –: negative effect.

served in nature. Possible interactions between two microbial populations can be recognized as positive interactions (commensalisms, protocooperation and mutualism), negative interactions (amensalism and competition), and interactions (parasitism and predation) which are positive for one but negative for the other population. Table 1 shows a variety of interactions between two different microbial populations [7–9]. Neutralism means that there is no change in the growth rate of one population due to the presence of another one. The populations do not have the opportunity to interact. Therefore, the populations have extremely different metabolic capabilities and do not play the same or overlapping functional roles within a community. In commensalisms, one population benefits while the other remains unaffected. There are a number of physical and chemical bases for relationships of commensalisms: production of growth factors or substrates, removal of inhibitory substances, and cometabolism of natural products form the basis for many commensal relationships between two microbial populations. Protocooperation (synergism) benefits both interacting populations, but unlike the mutualism described below, the association is not obligatory. Protocooperation provides for new or accelerated activities by microbial populations acting together, permitting microorganisms to combine their metabolic activities to perform transformation of substrates that could not be carried out by the individual population. Mutualism is an obligatory relationship between two microbial populations that benefits both populations. Close mutualistic relationships, such that the partnership is essential for the survival of one or both species, are often termed symbiosis. A mutualistic relationship is highly specific: one member of the association cannot be replaced by another related species. For example, the relationship between a phenylalanine-requiring strain of Lactobacillus and a folic acid-requiring strain of Streptococcus is a representative case for the exchange of growth factors [7]. In contrast to positive interactions, amensalism and competition represent a negative relationship between two microbial populations. When one micro-

38

M. Taniguchi · T. Tanaka

bial population produces a substance that is inhibitory to the other population, the relationship is called amensalism. The first population may be unaffected by the inhibitory substance. Competition occurs when two populations use the same resource, whether space or a limiting nutrient. Parasitism exerts a negative influence on susceptible host populations and benefits the parasite. The parasite population may be entirely dependent on the host for its nutritional requirement or may have an alternative mechanism for meeting its nutritional needs. Predation typically occurs when one organism, the predator, engulfs and digests another, the prey. Normally, predator-prey interactions are of short duration and the predator is larger than the prey. For efficient production of useful substances and microbial cells, and rapid conversion to substances with functions, utilization of positive interactions (commensalisms, protocooperation, and mutualism) between two microorganisms as described above is considered to be promising. 1.2 Production of Useful Substances by Co-culture 1.2.1 Hydrogen and Methane A number of successful results on the production of useful materials by coculture have been reported by Japanese researchers. The examples of production of useful substances by co-culture are summarized in Table 2. The conversion of organic molecules by one population into substrates for other population is the basis for a commensal relationship. For example, Clostridium butyricum evolves hydrogen from carbohydrate with limited yields because it produces organic acids besides hydrogen, while photosynthetic bacteria produce hydrogen from organic acids with high yields under illumination. To directly produce hydrogen from glucose or starch, a co-culture of Clostridium butyricum and a photosynthetic bacterium, Rhodopseudomonas sp. (Rhodobacter sphaeroides) [10, 11] or Rhodobacter sp. [12] was investigated. Photoproduction of hydrogen from raw starch in a microalgal biomass by combinations of halophilic or halotolerant bacterial species, Rhodobium marinum and Vibrio fluvialis [13], and of R. marinum and an amylase-producing lactic acid bacterium, Lactobacillus amylovorus [14] was also studied. R. marinum in pure culture could not produce hydrogen from a synthetic medium containing acetic acid and ethanol. This suggests that V. fluvialis supplied not only substrates but also some unknown factors capable of inducing hydrogen production from these substrates by R. marinum [13]. In a similar manner, it was found that in the co-culture of R. marinum and L. amylovorus, the latter produces unknown factors that promote hydrogen production by the former [14]. Just as for the photoproduction of hydrogen, methanogenesis of glucose by a defined thermophilic co-culture of Clostridium thermoaceticum and

Clarification of Interactions

39

Table 2 Examples of production of useful substances by co-culture

Target compound

Microorganism

Ref.

A

a

Hydrogen

Rhodopseudomonas sp. (Rhodobacter sphaeroides)

Clostridium butyricum

[10, 11]

Hydrogen

Rhodobacter sp.

Clostridium butyricum

[12]

Hydrogen

Rhodobium marinum

Vibrio fluvialis

[13]

Hydrogen

Rhodobium marinum

Lactobacillus amylovorus

[14]

Methane

Methanosarcina sp.

Clostridium thermoaceticum

[15]

Methane

Methanobacterium thermoautotrophicum

Desulfotomaculum sp.

[16]

Lactic acid

Streptococcus lactis

Aspergillus awamori

[17]

Acetic acid

Acetobacter sp.

Zymmonas mobilis

[18]

Mixed acids

Bifidobacterium longum

Propionibacterium freudenreichii

[19]

Ethanol

Saccharomyces cerevisiae (a respiratory deficient mutant)

Pichia stipitis

[20]

Ethanol

Saccharomyces cerevisiae

Pichia stipitis

[21]

Vitamin B12

Propionibacterium freudenreichii

Ralstona eutropha

[22]

Nisin

Lactococcus lactis subsp. lactis

Kluyveromyces marxianus

[23]

Kefiran

Lactobacillus kefiranofaciens

Saccharomyces cerevisiae

[24]

Kefiran

Lactobacillus kefiranofaciens

Torulaspora delbrueckii

[25]

Polysaccharide

Agrobacterium tumefaciens

Cellulomonas cellulans

[26]

b-Hydrooxybutyrate

Ralstona eutropha

Lactobacillus delbrueckii

[27]

Methanosarcina sp. was carried out [15]. In the co-culture, glucose was first converted to acetate (an inhibitory product for the former) by C. thermoaceticum and the acetate was subsequently converted to methane and carbon dioxide by Methanosarcina sp. The removal of the inhibitory or toxic compound is another basis for a commensal relationship. The ability to destroy or remove toxic factors is widespread in microbial communities. For example, methanogens play an important role in the degradation of organics, but this group is generally sensitive to heavy metals. Therefore, a methanogen Methanobacterium thermoautotrophicum isolated from the waste disposal site

40

M. Taniguchi · T. Tanaka

could not grow at 0.5 mM Cd2+ or 1.0 mM Cu2+. On the other hand, a sulfatereducing bacterium, Desulfotomaculum sp. isolated from the same site was able to insolubilize 3.0 mM Cd2+ or 2.0 mM Cu2+ by production of hydrogen sulfide. When both were cultured together in the presence of the heavy metal, the activity of the metal-sensitive methanogen could be maintained and methane production was observed [16]. 1.2.2 Acids and Ethanol It is very difficult to establish optimum conditions for a co-culture in a single bioreactor, because the two strains used in co-cultures do not always have similar optimum culture conditions for, say, pH, temperature, nutrients, and oxygen demand. In one attempt to solve these difficulties, a coimmobilized co-culture system of an anaerobe and an aerobe that have different oxygen demands was proposed [17]. Therefore, the coimmobilized co-culture system of an anaerobic bacterium, Streptococcus lactis and an aerobic mold, Aspergillus awamori has been developed for L-lactic acid production from starch. In the coimmobilized system, the aerobic mycelia grew on and near the oxygen-rich surface of the gel beads, while the anaerobic bacterium grew mainly in the oxygen-deficient central part of the gel beads. By using the coimmobilized co-culture system, L-lactic acid was produced successfully from starch with high yields [17]. Enhanced production of acetic acid from glucose was also investigated in a co-culture of Acetobacter sp. and Zymmonas mobilis [18]. The co-culture was initially incubated under low aeration and low agitation conditions to allow ethanol fermentation. When the glucose concentration was dropped below 5 g/L, a shift from ethanol fermentation to acetic acid fermentation was achieved by enhancing the aeration rate and/or agitation speed. In the co-culture system, the acetic acid yield from glucose was 95.5% of the theoretical value under the optimum conditions [18]. Moreover, to increase the total antimicrobial activity of culture supernatant, we have reported on a sequential conversion of lactose using the two food microorganisms [19]. That is, a co-culture of Bifidobacterium longum and Propionibacterium freudenreichii, in which lactic acid produced from B. longum was converted to acetic and propionic acids by P. freudenreichii. The antimicrobial activity of the culture supernatant containing a mixture of acetic and propionic acids without lactic acid was found to be higher than that obtained in the cultivation of B. longum alone. We have also reported on ethanol production from a mixture of glucose and xylose, not only by a co-culture of a respiratory deficient mutant of Saccharomyces cerevisiae and Pichia stipitis using a single fermentor [20], but also by using a co-culture system with two fermentors and two microfiltration modules [21]. The results will be described below in detail.

Clarification of Interactions

41

1.2.3 Bioactive Compounds The major problems in vitamin B12 production using Propionibacterium freudenreichii is the growth inhibition of the cell due to the accumulation of inhibitory metabolites such as propionic and acetic acids. As one of several approaches for controlling the propionic acid concentration at low level, a co-culture of the vitamin B12 producing bacterium and Ralstona eutropha where the latter microorganism assimilates the propionic acid produced by the former was attempted [22]. When the performance was evaluated on the basis of the amount of vitamin B12 produced per medium used, the co-culture system gave a much higher value than a cell recycle system using a hollow fiber module and periodic cultivation where dissolved oxygen (DO) concentration was alternatively changed between 0 and 1 ppm. The nisin (an antibiotic polypeptide) production by a co-culture of Lactococcus lactis subsp. lactis and Kluyveromyces marxianus [23] as well as the kefiran (a functional polysaccharide) production by a co-culture of Lactobacillus kefiranofaciens and S. cerevisiae [24] is also given as a good example of a commensal relationship based on removal of a growth-inhibitory substance. The two strains of yeast, K. marxianus and S. cerevisiae, are capable of consuming lactic acid produced by the lactic acid bacteria (producers of useful compounds) under the aerobic conditions. The amounts of nisin and kefiran produced in the co-cultures were higher than those in pure cultures of corresponding lactic acid bacteria.Very interesting results were obtained for kefiran production by an anaerobic co-culture of L. kefiranofaciens and Torulaspora delbrueckii in which the latter, at a concentration of 2¥103 cells/mL, was added to the culture containing the former (7¥107 cells/mL) preincubated for 30 h [25]. The positive effects of addition of the yeast cells on kefiran production by the lactic acid bacterium seem to be attributable to some unknown factors aside from removal of lactic acid. The high level of productivity obtained in the co-culture suggests that it is feasible to produce kefiran industrially. On the basis of this result from the co-culture described above, we have investigated kefiran production by a single culture of L. kefiranofaciens under the conditions established by mimicking the presence of yeast cells. Our experimental results will be described below in detail. 1.2.4 Flocculant and PHB Cellulomonas cellulans grows well on starch but Agrobacterium tumefaciens does not degrade starch. However, A. tumefaciens could produce acidic polysaccharide, which has high flocculation activity in a kaolin suspension, from starch metabolites produced by C. cellulans [26]. Furthermore, coexistence of the strains was suitable for production of polysaccharide. Most importantly, a large amount of polysaccharide was produced only when A. tumefaciens was

42

M. Taniguchi · T. Tanaka

used to inoculate a culture of C. cellulans simultaneously or after one day [26]. There may be some unknown interactions between the two strains that are indispensable for production of much polysaccharide. Unfortunately, the reason why the two strains, when co-cultured, produce so much polysaccharide is not clear. b-Hydroxybutyrate (PHB) is a useful biodegradable polymer which can be used as a thermoplastic. A wild type of Ralstona eutropha cannot utilize certain sugars such as glucose, fructose, and xylose, but it can produce PHB from organic acids such as lactic acid, acetic acid and butyric acid. It is considered that these sugars are first converted to organic acids using one microorganism and then such organic acids are converted to PHB by R. eutropha. Conversion of sugars directly to PHB by using mutants or gene-engineered microorganisms may also be considered. However, several problems concerning plasmid stability and low yields have been encountered in previous studies.A co-culture system where sugars were converted to lactic acid by Lactobacillus delbrueckii and the lactic acid was converted to PHB by R. eutropha in one fermentor was developed [27]. Based on the experimental studies on the effect of lactic acid concentration on the cell growth of both microorganisms, as well as the fact that L. delbrueckii prefers anaerobic conditions while R. eutropha grows well under the aerobic conditions, a scheme to control lactic acid concentration at a low level and DO at different levels in a sophisticated co-culture system was considered. It was experimentally shown that the periodic fermentation resulted in superior PHB yield with relatively high productivity as compared with the cases where DO concentration was controlled at constant values [27].

2 Ethanol Production from a Mixture of Glucose and Xylose by Co-culture 2.1 Bioconversion of Biomass Hydrolysate to Ethanol Lignocellulosic biomass represents one of the most abundant renewable energy sources and contains hemicellulose as well as cellulose. The main components of cellulose and hemicellulose are glucose and xylose, respectively. The conversion of xylose to ethanol allows potentially high ethanol yields from lignocellulose. Depending on the pretreatment and hydrolysis processes for lignocellulose, a mixture of glucose and xylose, or both xylose-rich hydrolysate and glucose-rich hydrolysate, is produced [28, 29]. The latter is obtained by hydrolysis of the hemicellulose fraction of lignocelluloses such as hardwood and wheat straw with dilute acid at low temperature, followed by cellulose hydrolysis with cellulase or concentrated acid [28]. Ethanol production from xyloserich hydrolysate and glucose-rich hydrolysate is achieved by separate two-stage fermentation processes using microorganisms specific for each sugar. However,

Clarification of Interactions

43

it is relatively difficult to produce pure pentose (xylose) in high yield from lignocellulose [29]. Taking into account the overall utilization of the sugar portions of lignocellulose, the production of a mixture of glucose and xylose from lignocellulose by acid and/or enzymatic hydrolysis appears more promising. In the case of using such a sugar mixture, a co-culture system, in which a potential glucose-fermenting yeast is cultivated with a xylose-fermenting yeast is useful. 2.2 Effect of Oxygen Supply on Ethanol Production from Glucose or Xylose Whatever the fermentation schemes, the conversion of glucose and xylose to ethanol requires use of microorganisms suitable for each sugar. Pichia stipitis is one of the better xylose-fermenting yeasts [30, 31]. However, the rate of fermentation of glucose by the yeast is lower than that of the common glucose-fermenting microorganisms such as S. cerevisiae and Z. mobilis [32]. Moreover, the fermentation performance of the xylose-fermenting yeast depends significantly upon the oxygenation level of the culture broth [32–40]. For improved conversion of both glucose and xylose to ethanol, we quantitatively estimated the influence of the specific oxygen uptake rate per cell concentration (qO2) on the fermentation performance of P. stipitis under oxygenlimited conditions [20]. Based on the optimum values of qO2 obtained for each sugar, ethanol production from a mixed solution of glucose and xylose by P. stipitis was investigated. In addition, we indicated the effect of controlling qO2 on ethanol production by a co-culture of P. stipitis and a respiratory-deficient mutant of S. cerevisiae. The respiratory-deficient mutant, No. 701RD, was obtained from S. cerevisiae No. 7. The influence of the oxygen transfer rate (OTR) on the fermentation performance of P. stipitis cultivated in the media with glucose and xylose was examined first. Then the qO2 value was estimated for each sugar consumed to determine the optimum oxygenation conditions in the batch culture in which the cell concentration was varied. Figure 1 shows the relationships between qO2 and (qp)max when glucose or xylose was used. Except for low values of qO2, the values of (qp)max for the culture with xylose were lower than those for the culture with glucose. Moreover, (qp)max decreased steeply with increasing qO2 above 15 mg (g cell)–1 h–1. The maximum values of (qp)max were attained at a qO2 value of 14.3 mg (g cell)–1 h–1 for xylose and a value of 66.7 mg (g cell)–1 h–1 for glucose. From these results, it can be concluded that it is essential to control qO2 in accordance with cell concentration for improvement of the ethanol productivity in glucose and xylose fermentation by P. stipitis, and that the optimum value of qO2 is 14.3 or 66.7 mg (g cell)–1 h–1 when xylose or glucose is used as the carbon source, respectively.

44

M. Taniguchi · T. Tanaka

Fig. 1 Relationship between specific oxygen uptake rate, qO2, and specific ethanol production rate, (qp)max. Empty squares: glucose; empty circles: xylose

2.3 Ethanol Production by Co-culture of P. stipitis and a Respiratory-deficient Mutant of S. cerevisiae The rate of fermentation of glucose to ethanol by S. cerevisiae, the yeast most widely used for ethanol production, is much higher than that by P. stipitis. We used the technique of co-culture with simultaneous inoculation of P. stipitis and S. cerevisiae to increase the level of ethanol production from a mixture of glucose and xylose [20]. Co-culturing of S. cerevisiae No. 7 with P. stipitis made it impossible to control qO2 at an optimum value for xylose fermentation by P. stipitis because of oxygen consumption by S. cerevisiae No. 7. Therefore, we considered using a respiratory-deficient mutant of S. cerevisiae No. 7. In fact, the maximum qO2 when S. cerevisiae No. 7 was cultivated in CCY medium with glucose was 95.5 mg (g cell)–1 h–1. On the other hand, the maximum qO2 of a respiratory-deficient mutant (S. cerevisiae No. 701RD) whose fermentation performance was as good as that of S. cerevisiae No. 7, was 6.7 mg (g cell)–1 h–1. Therefore, the co-culture was carried out by simultaneous inoculation of S. cerevisiae No. 701RD and P. stipitis into CCY medium containing 50 g/L of glucose and 25 g/L of xylose and subsequent incubation. Figure 2 shows the ethanol production from a glucose and xylose mixture in a co-culture of P. stipitis and S. cerevisiae No. 701RD. Glucose was almost completely consumed by 12 h and at this time about 23 g/L of ethanol had been produced. Thereafter, xylose was gradually converted to ethanol, with a high yield of about 0.5 g/g. The final ethanol concentration reached a maximum value of 37.5 g/L at 40 h

Clarification of Interactions

45

Fig. 2 Ethanol production from a mixture of glucose and xylose by co-culture of P. stipitis and a respiratory-deficient mutant of S. cerevisiae. qO2 was adjusted from 66.7 mg (g cell)–1 h–1 to 14.3 mg (g cell)–1 h–1 at 16 h after the start of cultivation as indicated by the dotted line. Empty circles: glucose; filled circles: xylose

although 0.5 g/L of xylitol was also produced. In the co-culture of P. stipitis and S. cerevisiae No. 7 (data not shown), the concentration of ethanol produced during xylose consumption by P. stipitis was 5 g/L, much less than that in the case of the co-culture of P. stipitis and S. cerevisiae No. 701RD. Moreover, the concentration of xylitol produced in the co-culture with S. cerevisiae No. 7 was more than two-fold that for the case of co-culture with the mutant strain. Table 3 shows a comparison of the fermentation performances for the different culture methods. Excess oxygen supply resulted in a low yield of ethanol from a mixture of glucose and xylose, whereas the fermentation time was shorter than that of the cultures with low oxygen supply. The difference in yield and productivity of ethanol between the single cultures (No. 1 and No. 2 in Table 3) was due to the effect of controlling qO2 at an optimum value during xylose fermentation.When S. cerevisiae No. 7 was simultaneously inoculated with P. stipitis into CCY medium containing a mixture of glucose and xylose (No. 3), the ethanol yield was far below the maximum due to the shortage of oxygen available to P. stipitis caused by the oxygen utilization by S. cerevisiae No. 7. From these results, P. stipitis was concluded to possess a strictly limited qO2 for

P. stipitis

Single culturec

Co-culture

Co-culture

2

3

4

c

b

a

P. stipitis

Single cultureb

1

40

40

40

36

Culture time [h]

37.5

29.4

35.1

26.4

Concentration [g/L]

Ethanol

CCY medium with a mixture of 50 g/L glucose and 25 g/L xylose was used; fermentation with qO2 controlled at a constant 66.7 mg (g cell)–1 h–1; fermentation with qO2 adjusted to an optimum value for each sugar consumed.

S. cerevisiae No. 701RD +P. stipitis

S. cerevisiae No. 7 +P. stipitis

Strain used

Culture technique [g/L]

No.

Culture conditionsa

0.50

0.39

0.47

0.35

Yield [g/g substrate]

Table 3 Ethanol yields and productivities achieved under various sets of culture conditions

0.94

0.74

0.88

0.73

Productivity [g/L/h]

0.50

1.13

trace

trace

Concentration [g/L]

Xylitol

46 M. Taniguchi · T. Tanaka

Clarification of Interactions

47

ethanol production from xylose with a maximum yield. Furthermore, gradual increase in ethanol concentration during the early stage of the cultivations was considered to acclimate P. stipitis cells to ethanol before ethanol production from xylose commenced. The yield and productivity of ethanol obtained in the co-culture (No. 4) of the respiratory-deficient yeast strain and P. stipitis were higher than those in a single culture (No. 2) with qO2 controlled at an optimum value for each sugar. The increase in ethanol yield was based not only on the better fermentation performance of S. cerevisiae No. 701RD as compared with that of P. stipitis, but also on the control of qO2 at an optimum value for xylose fermentation by P. stipitis due to the low oxygen consumption of the mutant yeast strain. Consequently, the maximum ethanol yield and productivity were obtained in the coculture (No. 4) of the respiratory-deficient mutant of S. cerevisiae No. 7 and P. stipitis. Control of qO2 at an optimum value for each sugar consumed throughout the fermentation was necessary for enhancement of the yield and productivity of ethanol from a mixture of glucose and xylose. If an on-line measurement system of cell concentration were used, qO2 could be controlled at a predetermined value in accordance with cell concentration in such a microaerobic fermentation by coupling of this system with an OTR control system. It is necessary to develop a fermentation system capable of both automatic regulation of the oxygenation level of the medium and in situ determination of cell concentration for control of qO2 at a desired level. 2.4 Properties of the Co-culture System with Membrane Bioreactor As described above, we experimentally investigated the effect of controlling qO2 at the optimum value for each sugar consumed on the yield and productivity of ethanol in both single cultures of P. stipitis and conventional co-cultures of P. stipitis and S. cerevisiae using a single fermentor. Moreover, we developed a novel co-culture system for efficient conversion of a mixture of glucose and xylose to ethanol by P. stipitis and S. cerevisiae No. 7 [21], as an alternative to the conventional co-culture system using P. stipitis and S. cerevisiae No. 701RD described above. We have constructed the new co-culture system by combining two sets of a bioreactor with a microfiltration module reported previously [41–47]. Therefore, the co-culture system consists of two fermentors and two microfiltration modules, which allowed regulation of qO2 at an optimum level for ethanol production from xylose by P. stipitis in co-culture with S. cerevisiae No. 7. A schematic diagram of the reactor system for co-culture is shown in Fig. 3 [21]. This reactor system, with a total working volume of 2 L, is composed of two main fermentors (A and B) and two microfiltration modules (C and D) with a nominal pore size of 0.25 mm and an effective filtration area of 0.1 m2. The microfiltration modules were used after they were sterilized with a sodium hypochlorite solution of 100 ppm and washed with sterilized water. A roller

48

M. Taniguchi · T. Tanaka

Fig. 3 Schematic diagram of new co-culture system with two fermentors and two microfiltration modules. A, B: Fermentors; C, D: microfiltration modules; E, F: roller pumps; G, H peristaltic pumps for adjusting filtration rate; I: level controller; J, K, L, M: peristaltic pumps for adding HCl or NaOH solution; N, O: control units

pump (E) was used to withdraw the culture broth containing P. stipitis cells from fermentor A and to circulate it between fermentor A and module C at a rate of 21 L/h. Since module C can be used to filtrate the broth at a rate above the desired level at the circulation rate used, the filtration rate was usually adjusted to 0.5 L/h by a peristaltic pump (G). The filtrate obtained from fermentor A was supplied to fermentor B by a peristaltic pump (G) connected to a level controller (I), which was used to maintain the a constant working volume in fermentor B. Another roller pump (F) was used to withdraw the culture broth containing S. cerevisiae cells from fermentor B and circulate it in a manner similar to that described above. The filtrate obtained from fermentor B was supplied to fermentor A by a peristaltic pump (H) at almost the same rate as the filtration rate of culture broth in fermentor A. The pH of the culture broth in both fermentors was kept constant at 5.0 by addition of 2 N NaOH and 2 N HCl by use of peristaltic pumps (J, K, L, and M) connected to pH controllers (N and O). Filtered air was sparged at a constant flow rate (0.2 vvm) into fermentor A. DO concentration was measured by use of a polarographic oxygen electrode. Except for a short time after inoculation, the DO concentration in the culture broth was negligible throughout the fermentations. Under oxygen-limited conditions, qO2 was controlled at a desired value for each sugar consumed by adjustment of the agitation rate during the fermentation as described previously [20]. On the other hand, nitrogen gas was sparged at 0.2 vvm into fermentor B throughout the fermentation in order to maintain anaerobic conditions.

Clarification of Interactions

49

2.5 Ethanol Production by P. stipitis and S. cerevisiae in the Co-culture System with Membranes The new co-culture system enables not only confinement of microbial cells to each fermentor, but also exchange of culture medium between the two fermentors, as described above. Moreover, it is possible to maintain the oxygen supply conditions suitable for an individual microorganism in each fermentor. P. stipitis and S. cerevisiae were inoculated into fermentors A and B, respectively. We expected that qO2 could be maintained at an optimum value for ethanol production by P. stipitis in fermentor A, and that concomitantly anaerobic conditions could be maintained for efficient ethanol production from glucose by S. cerevisiae in fermentor B. However, it was found to be quite difficult to control qO2 at an optimum level (14.3 mg (g cell)–1 h–1) for xylose fermentation by P. stipitis because of the adhesion of a fraction of the P. stipitis cells to the microfiltration membrane. In fact, the final ethanol concentration was about 20 g/L and hardly any ethanol was produced from xylose by P. stipitis, but 1.6 g/L of xylitol was produced by 72 h of culture. The adhesion of the cells resulted in inaccurate measurement of the total cell concentration in fermentor A and microfiltration module C, and subsequent unfavorable regulation of qO2. Therefore, after the co-culture experiment, the P. stipitis cells adhering to the membrane were recovered by the circulation of a saline solution (about 5 L) between the module and a reservoir, and the concentration of cells recovered was determined from the measured turbidity of the saline suspension in the reservoir. The ratio of the number of P. stipitis cells adhering to the membrane to the total number of cells was around 0.33, whereas for S. cerevisiae cells the ratio was 0.25. Therefore, the total cell concentration of P. stipitis was assumed to be 1.5-fold higher than that in the broth circulating between fermentor A and microfiltration module C. Figure 4 shows the ethanol production from a glucose (50 g/L) and xylose (25 g/L) mixture by the new co-culture system using P. stipitis and S. cerevisiae. Taking into account the concentration of P. stipitis cells adhering to the membrane, qO2 was controlled at an optimum level (14.3 mg (g cell)–1 h–1) for xylose fermentation by P. stipitis. That is, by adjustment of the agitation rate, the oxygen supply was increased in accordance with the total cell concentration estimated as stated above. The culture broths were separately removed from the two fermentors. The concentrations of ethanol, glucose, and xylose in the two fermentors changed in the same manner due to almost complete exchange of culture medium between the two fermentors. All of the glucose was converted to ethanol within 16 h by S. cerevisiae. At this time, the ethanol concentrations reached about 22.4 and 23.9 g/L in fermentors A and B, respectively, which are as high as that achieved using the conventional co-culture and higher than that in the single culture of P. stipitis. Then the ethanol concentrations increased up to 33.1 g/L and 33.7 g/L for fermentors A and B, respectively, during xylose fermentation as shown in Fig. 4a and b. The final xylitol concentration in fer-

M. Taniguchi · T. Tanaka Fig. 4 Ethanol production from a mixture of glucose and xylose in the new co-culture system. a Results for fermentor A with P. stipitis; b results for fermentor B with S. cerevisiae. In fermentor A, qO2 was changed stepwise from 66.7 mg (g cell)–1 h–1 to 14.3 mg (g cell)–1 h–1 at 16 h depending upon the type of sugar consumed as indicated by the dotted line. Empty circles: glucose; filled circles: xylose

50

mentor A was 1.0 g/L at 56 h, which is as high as that in fermentor B. By increase of the oxygen supply with the cells adhering to the membrane taken into consideration, a high ethanol yield was obtained in our co-culture system. We developed a new co-culture system with microfiltration membranes in which two microorganisms could be cultivated separately in two fermentors while culture medium could be simultaneously almost completely exchanged between the two fermentors. Using the new co-culture system with microfiltration membranes, it was possible to individually establish the desired con-

Clarification of Interactions

51

ditions of oxygen supply for each microorganism. Such a co-culture system is expected to be widely used for the efficient production of useful substances, rapid conversion to bioactive compounds by two microorganisms, and elucidation of the interactions between two microorganisms.

3 Simultaneous Production of Bifidobacterial and Propionibacterial Cells Using a Co-culture System with Membranes 3.1 Production of Bifidobacterial Cells as a Probiotic Bifidobacterium species are among the most important bacteria of the large intestinal tracts of mammals, and are considered to have several physiological effects including repression of growth of putrid bacteria, protection against infection and promotion of the immune reaction [48–50]. Bifidobacterial cells have been used not only as starter cultures for production of a large variety of fermented foods such as dairy and meat products, but also as so-called probiotics with function for improving the health of humans and farm animals [51–53]. Therefore, to use the bifidobacterial cells widely as useful microorganisms, it is necessary to increase productivity of the cell mass. In our earlier investigation, a fermentor with continuous separation of inhibitory metabolites by the method of cross-flow filtration using a ceramic filter was developed to improve productivity of bifidobacterial cells [54]. In 1994, Kaneko et al. found that P. freudenreichii, which is used as a starter for making Swiss-type cheese, produces a bifidogenic growth stimulator (BGS) and the chemical structure of BGS is 2-amino-3-carboxy-1,4-naphthoquinone [55]. Thereafter they reported on the properties and functions of BGS in detail [56–59]. We described previously that lactic acid produced from glucose by B. longum was converted to acetic and propionic acids by P. freudenreichii when B. longum and P. freudenreichii were co-cultivated in a fermentor with medium containing lactose as a carbon source [19]. Generally, P. freudenreichii prefers lactic acid as a carbon source over sugars such as glucose and lactose. Therefore, the cooperative relationship for cell growth seems to exist between the two food microorganisms as shown in Fig. 5. We used our co-culture system (see Fig. 3) with two microfiltration modules and two fermentors for efficient production of useful cells by use of the cooperative interactions between two microorganisms.As mentioned above, our coculture system allowed the cultivation of the individual microorganism in each fermentor as well as the successful exchange of culture medium between two fermentors. We investigated the cooperative interactions for cell growth between the two food microorganisms. In addition, we showed the simultaneous production of the two food microorganisms, especially production of bifidobacterial cells, using the co-culture system.

52

M. Taniguchi · T. Tanaka

Fig. 5 Cooperative interaction between Bifidobacterium and Propionibacterium

3.2 Cooperative Interactions between B. adolescentis and P. freudenreichii The effect of fed-batch addition of BGS on growth of B. adolescentis in a single fermentor was first examined. The supernatant, which was obtained previously from culture broth of P. freudenreichii, was used as a BGS sample. The supernatant was supplied to the culture of B. adolescentis with another peristaltic pump connected to the same pH controller used for adding NaOH solution. By gradual addition of the supernatant, the turbidity increased and the number of living cells was maintained at a higher level during the late-growth period compared to that in the control culture without the supernatant. Figure 6 shows the effect of kind of carbon source on growth of P. freudenreichii in a single fermentor. Glucose and lactic acid were used as a carbon source. This strain of P. freudenreichii consumed little or no lactose, as described previously [19], but 10 g/L of glucose was entirely consumed in 48 h and it produced propionic and acetic acids. On the other hand, 10 g/L of lactic acid was rapidly consumed and the yields of propionic and acetic acids per gram of consumed carbon source were higher than those with the culture using glucose as a carbon source. The maximum specific growth rate in the initial growth phase of the culture using lactic acid as a carbon source was about 0.20 h–1, which is 1.3 times that obtained in the culture using glucose as a carbon source. On the basis of these results, we experimentally showed that P. freudenreichii prefers lactic acid to glucose as a carbon source. Consequently, in the co-culture of B. adolescentis and P. freudenreichii, the conversion of glucose to lactic acid by B. adolescentis seems to be favorable for rapid growth of P. freudenreichii. 3.3 Cell Production Using Our Co-culture System with Membranes Simultaneous cell production was achieved by the co-culture system. B. adolescentis, B. longum, or Bifidobacterium breve and P. freudenreichii were inoculated separately into fermentors A and B (see Fig. 3), respectively, each containing the same volume (1.0 L) of TPY medium with 50 g/L of glucose. In other experiments for production of bifidobacterial cells, the volume (1.5 L) of

53 Fig. 6 Effect of kind of carbon source on growth of P. freudenreichii. Glucose a and lactic acid b were used as a carbon source. Empty circles: turbidity; empty squares: glucose; filled triangles: lactic acid; empty diamonds: propionic acid; empty triangles: acetic acid

Clarification of Interactions

medium in fermentor A for B. adolescentis, B. longum, or B. breve, was tentatively adjusted to be larger than that (0.5 L) in fermentor B for P. freudenreichii. A mixed gas of N2 and CO2 at a volume ratio of 9:1 was sparged into both fermentors (A and B) at 0.3 vvm throughout the fermentation to maintain anaerobic conditions. B. adolescentis and P. freudenreichii were inoculated into fermentor A and fermentor B, respectively. The culture broths were separately removed from the

54

M. Taniguchi · T. Tanaka

Table 4 Comparison of total cell productivity of B. adolescentis and P. freudenreichii among different culture systems

Culture methods

Strains used

Culture time [h]

Cell centration [g/L]

Single culture

B. adolescentis

12

3.0

Single culture

P. freudenreichii

96

5.0

Co-culture with membrane (Volume ratio=1:1)

B. adolescentis P. freudenreichii

56 56

3.2 8.6

Co-culture with membrane (Volume ratio=3:1)

B. adolescentis P. freudenreichii

64 64

5.5(8.3)a 9.8(4.9)a

a

These data in the parentheses show total cell mass obtained per 2 L of medium in the co-culture with volume ratio of 3 (1.5 L: fermentor A) to 1 (0.5 L: fermentor B).

two fermentors. The concentrations of glucose, acetic acid, and propionic acid in the two fermentors were changed in a similar manner due to almost complete exchange of culture medium between the two fermentors. However, a slight amount of lactic acid was detected once in fermentor A and little or no lactic acid was detected in fermentor B, because lactic acid produced by B. adolescentis was consumed only by P. freudenreichii in fermentor B. In the co-culture, the amount of bifidobacterial cells was slightly higher than that with a conventional batch culture, and the growth rate of P. freudenreichii was higher than that with a corresponding batch culture. As shown in Table 4, the sum of the amount (11.8 g) of both cells obtained per 2 L of the medium in the co-culture was about 1.5-fold higher than the total amount (8.0 g) of cells obtained in each single culture. Under similar conditions in the co-culture system, the increase in volume of medium for B. adolescentis made it possible to improve the productivity of bifidobacterial cells per volume of medium used. That is, in such a co-culture the amounts of bifidobacterial and propionibacterial cells were 8.3 g and 4.9 g, respectively. The total amount of both cells obtained per 2 L of the medium was 13.1 g, which is much higher than that (8.0 g) of cells obtained in each single culture. Enhanced total cell productivity per volume of medium used was also observed in the co-culture with a combination of B. longum and P. freudenreichii or B. breve and P. freudenreichii. These interesting results will be reported elsewhere. It was found that our co-culture system with two microfiltration modules and two fermentors allowed not only the cultivation of the individual microorganism in each fermentor, but also the successful exchange of culture medium between two fermentors. We successfully applied the co-culture system to the efficient production of bifidobacterial and propionibacterial cells by use of the cooperative interactions between the two food microorganisms.

Clarification of Interactions

55

4 Kefiran Production by L. kefiranofaciens under Culture Conditions Mimicking the Natural Co-culture System 4.1 Properties and Production of Kefiran Polysaccharides produced by lactic acid bacteria also provide a source of stabilizing, viscosifying, emulsifying, gelling, and water binding reagents for use as natural food additives [60, 61], which may be an alternative to texturizing agents of plant and animal origins. Kefiran, a water-soluble polysaccharide, is produced in kefir grains that consist of a complex population of firmly embedded lactic acid bacteria and yeasts [62]. The presence of yeasts results in the production of ethanol and also of carbon dioxide (CO2), which gives an effervescent character to these products. Kefiran, which contains approximately equal amounts of glucose and galactose, is produced by L. kefiranofaciens [63]. The chemical structure and properties of kefiran extracted from kefir grains were investigated by Kooiman [64], Mukai et al. [65], and Micheli et al. [66]. The antitumor activity of kefiran was reported by Shiomi et al. [67]. To utilize kefiran in many applied fields, such as the food, cosmetic and pharmaceutical industries, it may be necessary to produce a large quantity of the polysaccharide by L. kefiranofaciens because extraction of kefiran from kefir grains cultured in milk is complicated and the yield is fairly low. A homofermentative bacterium, L. kefiranofaciens which was isolated from kefir grains by Toba et al. [63] produced kefiran only in a medium containing expensive wine, and kefiran was obtained at a low concentration (80 mg/L) in the culture. Yokoi et al. [68] isolated a new kefiran-producing homofermentative bacterium, Lactobacillus sp. KPB-167B from kefir grains with a newly-developed milk whey medium without wine, and examined the optimum culture conditions for production of kefiran by the lactic acid bacterium using MRS medium containing lactose [69]. Recently, Mitsue et al. [70] obtained a high kefiran-producing strain (L. kefiranofaciens KF-75) suitable for industrial production of kefiran by repeated UV irradiation. They found that using the strain KF-75, higher kefiran productivity was successfully obtained by optimizing the medium and culture conditions in a 50 L jar fermentor [25]. Most notably, they reported that the highest kefiran productivity could be achieved in a co-culture of the strain KF-75 and Torulaspora delbrueckii, one of the yeast strains existing in kefir grains as described above (see Table 1) [25]. 4.2 Kefiran Production by a Combination of L. kefiranofaciens and Yeast Cells L. kefiranofaciens JCM 6985 was selected as one of the most suitable strains for kefiran production, and cultivated using lactose as a carbon source under the conditions of initial pH of 5.5 and temperature of 30 °C. Using 50 g/L of lac-

56

M. Taniguchi · T. Tanaka

tose as a carbon source, the highest kefiran yield (about 500 mg/L) was obtained at a cultivation time of 9 days. Based on the results reported by Mitsue et al. [25], the effect of addition of yeast cells on kefiran production by L. kefiranofaciens JCM 6985 was investigated. In the co-culture, L. kefiranofaciens and yeast cells (S. cerevisiae, Candida kefyr, or T. delbrueckii) at a concentration of 5¥105 cells/mL and 1¥105 cells/mL, respectively, were inoculated simultaneously. Unfortunately, little or no positive effect of the addition of yeast cells on kefiran production by L. kefiranofaciens was observed. Similar results were obtained even in the co-cultures where the yeast cells at different concentrations (1¥103~1¥105 cells/mL) were added to the culture containing 5¥105 cells/mL of L. kefiranofaciens and/or were added at the midpoint of cultivation. 4.3 Kefiran Production by L. kefiranofaciens under the Conditions Established by Mimicking the Existence of Yeast Cells We attempted to separate the positive effects of the presence and activities of yeast cells on kefiran production by L. kefiranofaciens JCM 6985 into factors of both the supplementation of nutrients to the lactic acid bacteria and the provision of conditions suitable for cell growth and kefiran production by formation of CO2 and/or ethanol [71]. The effect of addition of yeast extract on kefiran production by L. kefiranofaciens JCM 6985 was studied. The pH of the culture broth was maintained at 5.5 throughout cultivation by adding 4 N NaOH. In the culture without yeast extract, poor growth was observed and hardly any kefiran was produced. The higher the concentrations of yeast extract added, the higher the yield of kefiran. The addition of yeast extract at a concentration of 5 g/L allowed a drastic increase in kefiran production. The final kefiran concentration was 650 mg/L in the culture with gentle agitation (100 rpm), but without aeration.Yokoi et al. [72] also pointed out that the essential nutrient for Lactobacillus sp. KPB-167B was yeast extract, and that the lack of tryptone or meat extract achieved about 85% of the cell mass obtained in MRS medium. The enhanced cell growth and kefiran production appear to be mainly due to abundant vitamin groups in the yeast extract. It is necessary to examine the medium ingredients affecting the kefiran production; this point is especially important for the industrial production of kefiran. Although L. kefiranofaciens JCM 6985 grew and produced kefiran in the cultivation without aeration, no cell growth and no kefiran production by the bacterium were observed in the cultivation with aeration of N2 alone at 0.3 vvm. Then, a mixed gas of N2 and CO2 at a volume ratio of 9:1 or 5:5 was sparged into the culture broth at 0.3 vvm throughout the cultivation. When a mixed gas of 9:1 (N2:CO2) was sparged, a slight increase in the amount of kefiran produced was observed as compared with that in the cultivation without aeration. The supply of a slight amount of CO2 was found to be important for cell growth and kefiran production [71].

57 Fig. 7 Effects of ethanol addition on kefiran production by L. kefiranofaciens. When the medium containing 50 g/L of lactose was used as a carbon source, 10 g/L a or 20 g/L b of ethanol was added. When the medium containing 75 g/L of lactose was used as a carbon source, 10 g/L c of ethanol was added. Empty circles: turbidity; empty squares: lactose; filled squares: galactose; empty diamonds: glucose; empty triangles: lactic acid

Clarification of Interactions

The kefiran made by the traditional procedure contains 0.5–1.5% of ethanol and a small amount of acetic acid, as well as 0.6–1.0% of lactic acid. By initially adding 10 or 20 g/L of ethanol to the medium containing 50 g/l or 75 g/L of lactose, the effect of the addition on kefiran production by L. kefiranofaciens JCM 6985 was examined. Figure 7 shows the results of cultivations with the pH controlled at 5.5 and simultaneously with sparging of a mixed gas of N2 and CO2 (9:1). When 10 or 20 g/L of ethanol was initially added to the medium containing 50 g/L of lactose, the time when galactose could be detected in the culture broth decreased as compared to that without ethanol, although negligible

58

M. Taniguchi · T. Tanaka

changes in the profiles of lactose consumption and lactic acid production were observed. As shown in Fig. 7a and b, kefiran continued to be produced gradually in the culture broth, even after all sugars were consumed. The increased concentration of kefiran was obtained at 8 days when 10 g/L or 20 g/L of ethanol was added to the medium. These values were significantly higher than that of the cultivation without ethanol described above.When 10 g/L of ethanol was added to the medium containing 75 g/L of lactose, the maximum concentration of kefiran obtained was about 1040 mg/L at 10 days, as shown in Fig. 7c. However, in this cultivation, the added lactose was not utilized efficiently to produce kefiran, because both 5.9 g/L of glucose and 33 g/L of galactose remained in the culture broth. The reduced sugar consumption seems to be due to the accumulation of more than 50 g/L of lactic acid, in the same manner as the results of cultivation using 100 g/L of lactose as a carbon source. 4.4 Kefiran Production by a Single Culture of L. kefiranofaciens Mimicking a Natural Co-culture System Figure 8 shows a comparison of the amounts of extracellular kefiran obtained under the different culture conditions [71]. By sparging a mixed gas of N2 and CO2 at a volume ratio of 9:1 into the culture broth at 0.3 vvm throughout the cultivation, a slight increase in the amount (670 mg/L) of kefiran produced was observed as compared with that (650 mg/L) in the cultivation without aeration. However, further increase in the CO2 fraction to one-half of total gas volume caused a decrease in the amount of kefiran produced. When 10 g/L or 20 g/L of ethanol was added to the medium containing 50 g/L of lactose, the kefiran concentrations were about 930 mg/L or 840 mg/L at 8 days, respectively. The max-

Fig. 8 Comparison of the amounts of kefiran produced under different culture conditions

Clarification of Interactions

59

imum kefiran concentration was about 1040 mg/L at 10 days, when 10 g/L of ethanol was added to the medium containing 75 g/L of lactose. Although the time required to reach the maximum concentration of kefiran was longer, the addition of ethanol to the medium made it possible to increase the amount of kefiran produced. Based on the results described above, we found that factors resulting from the presence of yeast strains in kefir grains (that is, addition of yeast extract and ethanol, aeration of gas containing CO2, and their combinations) allowed promotion of kefiran production by L. kefiranofaciens. In general, to stimulate production of useful substances by cooperative actions between two microorganisms in a co-culture, it seems very difficult to optimize the culture conditions of pH, temperature, composition of aeration gas, ratio of inoculum size, time and amount of inoculation of each microorganism, and composition of nutrients in the medium used. The results obtained in this study may show one good example of production of useful compounds using a single microorganism under culture conditions established by mimicking the actions of yeast cells on L. kefiranofaciens in kefir grain as a typically natural co-culture system.

5 Conclusions and Prospects The optimum oxygenation level for ethanol production from a sugar mixture was achieved by using both a respiratory deficient mutant in a single fermentor and the co-culture system with two fermentors, as described above. It is generally difficult to establish optimum conditions for a co-culture in a single bioreactor, because the two strains used in co-cultures do not always have similar optimum culture conditions for, say, pH, temperature, nutrients, and oxygen demand. Therefore, conversion of substrates directly to useful substances by using mutants or gene-engineered microorganisms may be considered as an alternative to co-culture. However, several problems concerning plasmid stability and low yields have been encountered in previous studies. Use of mutants and/or gene-engineered microorganisms may be promising when the microorganisms lack specific enzymes for conversion of toxic or inhibitory compounds or for degradation of biopolymers such as starch and cellulose, as reported previously [73, 74]. The sequential conversion of sugars and starch to useful materials by two microorganisms is successfully utilized in several co-cultures as described above. In some simple cases, the activities of one species make a compound available for another species without actually transforming the particular compound. In other cases, removal of inhibitory or toxic substances by one population allows growth of another population. In the co-cultures where the products produced by one microorganism are inhibitory for its growth and they are consumed as substrates by another microorganism, cooperative interactions are established. The production of nisin and kefiran by lactic acid bacteria in

60

M. Taniguchi · T. Tanaka

a combination of yeast strains is a good example of protocooperation (synergism), indicating that both microorganisms benefit from the relationship.Also, the simultaneous production of bifidobacterial and propionibacterial cells is another good example of co-culture with the cooperative interactions between the two microorganisms. As mentioned above, however, there are still many unknown factors stimulating the production of the useful substances in each co-culture system. In particular, there may be little or no production of useful substances by utilizing a mutualistic relationship (symbiosis) of microbial community except for methane fermentation [75, 76]. It seems to be difficult to clarify even cooperative interactions between two populations, but it may not be impossible to do it. Further intensive studies are required for elucidation of multiple interactions in microbial communities producing useful compounds, including clarification of synergistic interactions between two populations. Moreover, for enhanced production of useful substances by utilizing cooperative actions between two populations, properties of molecules for positive interactions and their action mechanisms should be investigated. If the positive interactions between two microorganisms are fully clarified, positive culture surroundings provided by a non-producer strain could be established in a bioreactor when the producer of useful substance is a single strain, just as in kefiran production by L. kefiranofaciens alone [71]. Consequently, new bioprocesses in which similar performance to co-culture systems is obtained even when using a single microorganism are expected to be developed for improved biochemical reaction systems.

References 1. Fiechter A (2000) (ed) History of modern biotechnology I: Adv biochem bioeng/biotechnol, vol 69. Springer, Berlin Heidelberg New York 2. Fiechter A (2000) (ed) History of modern biotechnology II: Adv biochem bioeng/biotechnol, vol 70. Springer, Berlin Heidelberg New York 3. Faurie R, Thommel (2003) (eds) Microbial production of L-amino acids: Adv biochem bioeng/biotechnol, vol 79. Springer, Berlin Heidelberg New York 4. Fiechter A (1994) (ed) Biotechnics/wastewater: Adv biochem bioeng/biotechnol, vol 51. Springer, Berlin Heidelberg New York 5. Maier RM, Pepper LI, Gerba CP (2000) Environmental microbiology. Academic Press, San Diego 6. Hui YH, Khachatourians GG (1995) Food biotechnology: microorganisms. VCH, New York 7. Bailey JE, Ollis DF (1986) Biochemical engineering fundamentals, 2nd edn. McGraw-Hill, New York 8. Atlas RM, Bartha R (1998) Microbial ecology: fundamentals and applications, 4th edn. Benjamin/Cummings, California 9. Yanagita T (1992) Microbial ecology. Gakkai Shuppan Center, Tokyo (in Japanese) 10. Miyake J, Mao X-Y, Kawamura S (1984) J Ferment Technol 62:531 11. Asada Y, Miyake J (1999) J Biosci Bioeng 88:1

Clarification of Interactions 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

61

Yokoi H, Mori S, Hirose J, Hayashi S, Takasaki Y (1998) Biotechnol Lett 20:895 Ike A, Murakawa T, Kawaguchi H, Hirata K, Miyamoto K (1999) J Biosci Bioeng 88:72 Kawaguchi H, Hashimoto K, Hirata K, Miyamoto K (2001) J Biosci Bioeng 91:277 Koesnandar, Nishio N, Kuroda K, Nagai S (1990) J Ferment Bioeng 70:398 Mori K, Hatsu M, Kimura R, Takamizawa K (2000) J Biosci Bioeng 90:260 Kurosawa H, Ishikawa H, Tanaka H (1988) Biotechnol Bioeng 31:183 Kondo T, Kondo M (1996) J Ferment Bioeng 81:4298 Taniguchi M, Nakazawa H, Takeda O, Kaneko T, Hoshino K, Tanaka T (1998) Biosci Biotech Bioch 62:1522 Taniguchi M, Tohma T, Itaya T, Fujii M (1997) J Ferment Bioeng 83:364 Taniguchi M, Itaya T, Tohma T, Fujii M (1997) J Ferment Bioeng 84:59 Miyano K, Ye K, Shimizu K (2000) Biochem Eng J 6:207 Shimizu H, Mizoguchi T, Tanaka E, Shioya S (1999) Appl Environ Microbiol 65:3134 Mitsue T, Tachibana K, Fijio Y (1999) Nippon Seibutu-Kogaku Kaishi 77:99 (in Japanese) Cheirsilp B, Shimizu H, Shioya S (2003) J Biotechnol 100:43 Kurata S, Yamada K, Takatsu K, Hanada S, Koyama O, Yokomaku T, Komagata Y, Kanagawa T, Kurane R (2003) Biosci Biotech Biochem 67:8 Tohyama M, Takagi S, Shimizu K (2000) J Biosci Bioeng 89:323 Gregg DJ, Saddler J (1996) Biotechnol Bioeng 51:375 Prior BA, Kilian SG, du Preez JC (1989) Process Biochem Feb 21 du Preez JC, Mosch M, Prior BA (1986) Appl Microbiol Biotechnol 23:228 Delgenes JP, Moletta R, Navarro JM (1986) Biotechnol Lett 8:879 Delgenes JP, Moletta R, Navarro JM (1988) J Ferment Technol 66:417 Lighthelm ME, Prior BA, du Preez JC (1988) Appl Microbiol Biotechnol 28:63 Slininger PJ, Branstrator LE, Bothast RJ, Okos MR, Ladisch MR (1991) Biotechnol Bioeng 37:973 Delgenes JP, Moletta R, Navarro JM (1989) Biotechnol Bioeng 34:398 Laplace JM, Delgenes JP, Moletta R, Navarro JM (1991) Appl Microbiol Biotechnol 36:158 du Preez JC, van Driessel B, Prior BA (1989) Yeast 5:S129 Grootjen DRJ, Meijlink HHM, van der Lans RGJM, Luyben KChAM (1990) Enzyme Microb Technol 12:860 Laplace JM, Delgenes JP, Moletta R, Navarro JM (1993) J Ferment Bioeng 75:207 Chung IS, Lee YY (1986) Enzyme Microb Technol 8:503 Taniguchi M, Kotani N, Kobayashi T (1987) J Ferment Technol 65:179–184 Iijima S, Kawai S, Mizutani M, Taniguchi M, Kobayashi T (1987) Appl Microbiol Biotechnol 26:542 Hatanaka H,Wang E, Taniguchi M, Iijima S, Kobayashi T (1988) Appl Microbiol Biotechnol 27:470 Taniguchi M, Hoshino K, Shimizu K, Nakagawa I, Takahashi Y, Fujii M (1988) J Ferment Bioeng 66:633 Taniguchi M, Nakagawa I, Hoshino K, Itoh T, Ohno K, Fujii M (1989) Agric Biol Chem 53:2447 Taniguchi M, Hoshino K, Itoh T, Kumakura H, Fujii M (1992) Biotechnol Bioeng 39:886 Taniguchi M, Hoshino K, Urasaki H, Fujii M (1994) J Ferment Bioeng 77:704 Mitsuoka T (1992) The human gastrointestinal tract. In: Wood BJB (ed) Lactic acid bacteria, vol 1. Elsevier, London, p 69 Ballongue J (1993) Bifidobacteria and probiotic action. In: Salminen S, Wright A (eds) Lactic acid bacteria. Marcel Dekker, New York, p 357 Fujiwara S (2002) Biosci Microflora 21:255 Reuter G (1997) Biosci Microflora 16:43

62

Clarification of Interactions

52. Hosono A, Lee J, Ametani A, Natsume M, Hirayama M, Adachi T, Kaminogawa S (1997) Biosci Biotech Bioch 61:312 53. Shah N (2000) Biosci Microflora 19:99 54. Taniguchi M, Kotani N, Kobayashi T (1987) Appl Microbiol Biotechnol 25:438 55. Kaneko T, Mori H, Iwata M, Meguro S (1994) J Dairy Sci 77:393 56. Kaneko T (1999) Biosci Microflora 18:73 57. Mori H, Sato Y, Taketomo N, Kamiyama T, Yoshiyama Y, Meguro S, Sato H, Kaneko T (1997) J Dairy Sci 80:1959 58. Yamazaki S, Kaneko T, Taketomo N, Kano K, Ikeda T (2002) Appl Microbiol Biotechnol 59:72 59. Iwata K, Hojo K, Yoda N, Kamiyama T, Makino S, Sato M, Sugano H, Mizoguchi C, Kurama S, Shibasaki M, Endo N, Sato Y (2002) Biosci Biotech Biochem 66:679 60. Crescenzi V (1995) Biotechnol Progress 11:251 61. Mukai T, Watanabe N, Toba T, Itoh T, Adachi S (1991) J Food Sci 56:1017 62. Tamime A, Robinson R (1988) J Dairy Res 55:281 63. Toba T, Abe S, Arihara K, Adachi S (1986) Agric Biol Chem 50:2673 64. Kooiman P (1968) Carbohyd Res 7:200 65. Mukai T, Toba T, Itoh T, Adachi S (1990) Carbohyd Res 204:227 66. Micheli L, Uccelletti D, Palleschi C, Crescenzi V (1999) Appl Microbiol Biotechnol 53:69 67. Shiomi M, Sakai K, Murofushi M, Aibara K (1982) Jpn J Med Sci Biol 35:75 68. Yokoi H, Watanabe T, Fujii Y, Toba T, Adachi S (1990) J Dairy Sci 73:1684 69. Yokoi H, Watanabe T (1992) J Ferment Bioeng 74:327 70. Mitsue T, Tachibana K, Fujio Y (1998) Seibutu-Kogaku Kaishi 76:447 (in Japanese) 71. Taniguchi M, Nomura M, Itaya T, Tanaka T (2001) Food Sci Technol Res 7:333 72. Yokoi H,Watanabe T, Fujii Y, Mukai T, Toba T,Adachi S (1991) Int J Food Microbiol 13:257 73. Nakamura Y, Kobayashi F, Ohnaga M, Sawada T (1997) Biotechnol Bioeng 53:21 74. Murai T,Yoshino T, Ueda M, Haranoya I,Ashikari T,Yoshizumi H, Tanaka A (1998) J Ferment Bioeng 86:569 75. Imachi H, Sekiguchi Y, Kamagata Y, Ohashi A, Harada H (2000) Appl Environ Microbiol 66:3608 76. Sekiguchi Y, Takahashi H, Kamagata Y, Ohashi A, Harada H (2001) Appl Environ Microbiol 67:5740

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 63– 87 DOI 10.1007/b94192 © Springer-Verlag Berlin Heidelberg 2004

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes Naomichi Nishio (✉) · Yutaka Nakashimada Department of Molecular Biotechnology, Graduate School of Advanced Sciences of Matter, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8530, Japan [email protected]

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

64

1

Introduction

2 2.1 2.2 2.3

Methane Production from Soluble Materials . . . . . . Application of the UASB Reactor to Various Substrates . Application of UASB Reactor for Unfeasible Conditions Advanced UASB Reactor . . . . . . . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

65 65 67 69

3 3.1 3.2 3.3 3.4

Methane Production from Solid Materials . . Methane Production from Cellulosic Materials Municipal Sewage Sludge . . . . . . . . . . . . Coastal Sea Sediments . . . . . . . . . . . . . Solid Wastes from Traditional Japanese Foods

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

70 70 70 72 73

4 4.1 4.2 4.2.1 4.2.2 4.3

Fermentative Hydrogen Production . . . . . . . Strictly Anaerobes . . . . . . . . . . . . . . . . Facultative Anaerobes . . . . . . . . . . . . . . Hydrogen Production by Enterobacter spp. . . . Enhancement of Hydrogen Yield of E. Aerogenes High Rate Hydrogen Production . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

74 75 76 76 77 80

5

Hydrogen-Methane Production from Organic Wastes . . . . . . . . . . . . .

82

6

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84

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

85

References

Abstract To treat soluble and solid wastes and recover energy from them, high rate methane fermentation, especially using the UASB (upflow anaerobic sludge blanket) reactor, and hydrogen fermentation using various microorganisms and microbial consortia have been investigated intensively in Japan. In this chapter, recent works on high rate methane fermentation in Japan are reviewed, focusing on: 1) basic studies into the applicability of the UASB reactor for various substrates such as propionate, lactate, ethanol, glucose and phenol; 2) its applications to unfeasible conditions, such as lipid and protein containing wastes, low temperature and high salt-containing wastes; 3) progress made in the field of advanced UASB reactors, and; 4) research into methane fermentation from solid wastes, such as from cellulosic materials, municipal sewage sludge, and mud sediments. Following this, although hydrogen fermentation with photosynthetic microorganisms or anaerobic bacteria was

64

N. Nishio · Y. Nakashimada

researched, for this review we have focused on fermentative hydrogen production using strictly or facultative anaerobes and microbial consortia in Japan, since high rate production of hydrogen-methane via a two-stage process was judged to be more attractive for biological hydrogen production and wastewater treatments. Keywords Methane · Hydrogen · Fermentation · Wastewater · Solid wastes

1 Introduction Anaerobic digestion and methane fermentation has been commonly used for the biological treatment of high-strength wastewater and for energy recovery as methane. The process has been used in Japan for wastewaters in factories processing food such as beer, sugars, drinking, potatoes, and chemicals and so on. However, to expand the use of the process, it is necessary to judge its applicability to other wastewaters, including recalcitrant and toxic compounds, and for many other substrates including alcohols, volatile fatty acids, and carbohydrates. Anaerobic digestion is also useful for reducing solid organic wastes.Among them, cellulose materials are the most abundant biomass resources, although its rate of degradation and methane production are much lower than other biomass. Therefore, an artificial microbial consortium consisting of an anaerobic fungus as hydrolyzing and fermentative microbes, a hydrogenotrophic methanogen, and an aceticlastic methanogen was set up to enhance the rates. Furthermore, some waste solids sometimes include high amounts of salts. For example, soy sauce refuse is always produced as a by-product from soy sauce manufacture [1] and contains 10% NaCl; some inland sea sediments in Japan are polluted by organic matter which contains 3% NaCl. Therefore, it is necessary to examine the possibility/potential for methane production using this high salt-containing organic matter.

Fig. 1 Schematic diagram of methane fermentation

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

65

Hydrogen is a clean energy source produced by photosynthetic and fermentative microorganisms under anaerobic conditions in pure culture [2]. The methanogenic ecosystem consists of hydrolysis (first stage), acidogenesis (second stage), and methanogenesis (third stage) (Fig. 1).Although it is well-known that hydrogen can be produced in the hydrolysis and acidogenesis stages, there are only a few reports so far that have focused on this hydrogen production. If high rate and high yield production of hydrogen can be achieved, the hydrogen produced could be connected directly to a “fuel cell” without reforming. Here we will describe research by Japanese researchers into high rate hydrogen production by strictly anaerobes, by isolated facultative anaerobes, and by microbial consortium.

2 Methane Production from Soluble Materials Anaerobic digestion leads to the overall gasification of organic wastewaters and solid wastes into methane and carbon dioxide. Although these digestion processes have been used for decades, interest in the economical recovery of fuel gas from industrial and agricultural surpluses has recently increased due to the changing socio-economical situation in the world. To achieve rapid and effective anaerobic digestion, some processes are being developed, such as the upflow anaerobic filter process (UAFP) [3], the upflow anaerobic sludge blanket (UASB) [4], the anaerobic attached film expanded-bed reactor (AAFEB) [5], and the anaerobic fluidized-bed reactor (AFBR) [6] to increase the cells retained, as well as the two-phase digestion process to optimize both the acidogenesis and methanogenesis. 2.1 Application of the UASB Reactor to Various Substrates In Japan, among the processes mentioned above, the UASB process has been expanded for use in anaerobic wastewater treatments in the food processing, pulp, and chemical industries. To expand the use of this process even more, the reactors have been operated at laboratory scale for several different soluble substrates and operating conditions. The results obtained for various defined substrates under mesophilic conditions with a UASB reactor are summarized in Table 1. For alcohols, especially methanol, to achieve a high space velocity (SV) of 9.4 d–1 and a high loading rate of 75 g L–1 d–1, it was necessary to add acetate. Phenol was also converted to methane at a relatively high SV but the loading rate (2 g L-reactor–1d–1) was much lower due to its toxicity. For carbohydrates, methane yields were lower: ca. 80% than for ethanol or fatty acids. This was probably due to high cell mass yields in these substrates. At a high loading rate, the pH became acidic due to the accumulation of meta-

b

a

8.6 3.3 2 1.6 8.1

5.4 3.4 3.4

8 3 0.5

Concentration (g L–1)

6.8 2.2 7.7 7.7 7.7

3.4 3.4 2.2

9.4 5.6 4.4

Space velocity (d–1)

58 7 15 12 62

18 11 8

75 18 2

Loading rate (g L-reactor–1d–1)

97 25 93 93 88

95 86 84

93 88 100

Substrate removala (%)

98 84 93 93 98

80 81 78

77 89 100

Methane yieldb (%)

[11] [12] [8] [8] [13]

[10] [8] [8]

[7] [8] [9]

Reference

Removal = (l-effluent concentration/influent concentration) ¥ 100; expressed as (CH4 production rate/theoretical CH4 production rate) ¥ 100, using the methanogenic stoichiometry of each substrate.

Propionate Propionate (with 8.6 g L–1 methanol) Propionate (with 1.6 g L–1 butyrate) Butyrate (with 2 g L–1 propionate) Lactate

VFAs

Glucose Sucrose (with 1 g L–1 acetate) Starch (with 1 g L–1 acetate)

Carbohydrates

Methanol (with 1.5 g L–1 acetate) Ethanol (with 0.5 g L–1 acetate) Phenol

Alcohols

Substrate

Table 1 High rate methane production in a lab-scale UASB reactor fed with various defined substrates

66 N. Nishio · Y. Nakashimada

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

67

bolic fatty acids, which decrease the removal efficiency. To overcome this problem acetate addition is useful, as shown in the Table 1. For glucose, addition of bicarbonate was found to be useful for controlling pH, in which case 30 mM (=5,400 mg L–1) at 3.4 d–1 of SV corresponding to 18 g l–1 d–1 may be the maximum amount of product achievable in a single UASB reactor. 2.2 Application of UASB Reactor for Unfeasible Conditions Mesophilic and thermophilic methane fermentations with high rate have been developed and widely-used as commercial processes, and much research has also been done on using laboratory-scale UASB reactors for wastes such as alcohol distillery wastewater in Japan [14, 15]. However, it was found to be difficult to use the anaerobic process for some wastewaters, so application of UASB reactors to these wastewaters has been investigated for the last decade. Tagawa et al. [16] investigated the performance of an on-site pilot-scale multi-staged UASB (MS-UASB) reactor by feeding it with a food processing wastewater containing high concentrations of lipid and protein at thermophilic conditions (55 °C). The reactor finally achieved 50 kg-COD m–3 d–1 with a soluble COD removal of 90% (based on the influent total COD versus the effluent filtered COD), while the overall COD removal (based on the effluent total COD) was considerably to be unsatisfactory, at around only 60–70%. At present, sewage is mainly treated with aerobic processes, such as the activated sludge process.Anaerobic treatment is still difficult due to the low temperature of wastes and the lower organic matter removal efficiency. However, since the estimated running costs revealed that anaerobic treatment of sewage at less than 20 °C was more economical than aerobic treatment, high rate methane fermentation was evaluated at low temperature using AFBR and UASB for the anaerobic treatment of sewage at 13–20 °C [17]. In both the AFBR and UASB, temperatures above 15 °C did not affect the TOC removal efficiency, but at 13 °C the TOC concentration in the effluent increased. The quality of the effluent at 6 h HRT and for temperatures above 15 °C was almost the same as that by the activated sludge process at 21 °C and HRT 6 h. Uemura and Harada [18] also studied the feasibility of sewage treatment by UASB reactor using actual sewage at 4.7 h HRT, at temperatures in the range of 13–25 °C. They revealed that the average total COD removal and solid COD removal achieved were 70% and 80%, respectively, and total COD removal rate depended on influent strength, especially solid COD concentration, rather than operational temperature. The methane fermentation of wastes with high salt content seems to be difficult. Salt tolerance in methane fermentation is usually low; less than 3% NaCl in conventional methane fermentation systems [19] and less than 1% for Methanosaeta sp. [20], which is the most dominant aceticlastic methanogen in the UASB system. Tanemura et al. [21] investigated anaerobic treatment of wastewater with high salt content generated during a pickled-plum manufacturing process (TOC, 14 g L–1; ash, 150 g L–1; pH 2.7, hereafter called pickled-plum

68

N. Nishio · Y. Nakashimada

Fig. 2 The methanogenic fermentation of post-acidogenic fermentation liquor (PAF liquor) in the UASB reactor treated at 37 °C [22]. Symbols: (–) dilution rate; (---) pH in influent; (●) pH in effluent; (▲) acetic acid in effluent; (●) methane production rate

effluent) using AFBR. Five-fold-diluted pickled-plum effluent (ash ca. 30 g L–1) was treated in AFBR at a maximum volumetric TOC loading rate of 3.0 g L–1 d–1 with a TOC removal efficiency of 71%. Ten-fold-diluted pickled-plum effluent could be treated in the AFBR at a maximum volumetric TOC loading rate of 11.1 g L–1 d–1 with a TOC removal efficiency of 84.6%. Takeno et al. observed methane production during the acidogenic fermentation of mud sediment in artificial seawater [22, 23]. When the UASB reactor, including methanogenic sludge acclimated from the mud sediment, was used to treat the acidogenic fermentation liquor containing 40 mM acetate and 3% NaCl at 37 °C, the methane production rate increased with culture time at the dilution rate of 1.0 d–1 and with the increase of the dilution rate, keeping acetic acid in the effluent at a low level. A methane production rate of 90 mmol L–1 d–1 was obtained by decreasing the dilution rate to 2.4 d–1 with a low level of acetic acid concentration in the effluent (Fig. 2). The acetic acid and TOC removal efficiencies were 87.5% and 88.7%, respectively. Both the methane production rate and the acetate removal rate obtained at 3% NaCl (the salinity of seawater) were quite high, since in the UASB reactor it was reported that acetate removal rates were

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

69

63.4 mmol acetate L–1 d–1 at 2.46% NaCl and 19.1 mmol acetate L–1 d–1 at 3.49% NaCl, respectively [20]. 2.3 Advanced UASB Reactor The UASB method has been developed as an efficient anaerobic wastewater treatment process. However, the performance of this process in the removal of nitrogenous and phosphorus compounds is not high compared to aerobic processes. Furthermore, it is difficult to remove organic carbon completely, unlike the aerobic process. To overcome these problems, some other approaches have tried. Chang and Nishio investigated the removal of PO4 and NH4 from synthetic VFA wastewater by stimulating the formation of precipitates containing PO4 and NH4 in granulated sludge during UASB methanogenic fermentation. This gave removal efficiencies of PO4 and NH4 of 85 (influent=406 mg L–1) and 60% (influent=100 mg L–1), respectively, when 8 mM Ca and 8 mM Mg were added to the influent [24]. In a heat-treated liquor of sewage sludge, a PO4 removal efficiency of 70% (influent=53 mg L–1) was achieved by the addition of 2 mM Ca and 3.5 mM Mg; removal of NH4, however, was not observed during the experimental period. A lighted upflow anaerobic sludge blanket (LUASB) reactor was proposed by Sawayama et al. for removing nitrogenous and phosphorus compounds [25].A population of phototrophic bacteria was induced from UASB granules under light conditions (100 mE m–2 s–1). The ammonium and phosphate ion removal efficiencies of the LUASB reactor were higher than those of a UASB reactor. The difference in the results from runs in light and dark conditions suggested that the efficiencies of ammonium and phosphate ion removal were improved by an increase in the phototrophic bacteria in the LUASB reactor. From the effluent of the LUASB reactor, two strains of phototrophic bacteria, Rhodopseudomonas palustris and Blastochloris sulfoviridis strains were isolated [26, 27]. To remove organic carbon in the effluent from an anaerobic treatment reactor, a novel system was proposed, which consists of a UASB anaerobic pretreatment unit and a following DHS (downflow hanging sponge-cubes) aerobic post-treatment unit, as a low-cost and easy-maintenance process for developing countries [28] because the whole system requires neither external aeration input nor withdrawal of excess sludge. Although this system was targeted at sewage treatment, it would be applicable to wastewaters containing high strength organic matter. Their proposed system achieved 94% of totalCOD removal, 81% of soluble-COD removal, and near-perfect SS and totalBOD removal at an overall HRT of 8.3 h (7 h in UASB and 1.3 h in DHS unit) over six months of being fed sewage. Moreover, the DHS reactor was capable of performing high (73–78%) nitrification. At a municipal sewage treatment site, furthermore, a 550 d continuous experiment demonstrated that the whole combined system successfully achieved 94–97% of unfiltered-BOD removal,

70

N. Nishio · Y. Nakashimada

81–84% of unfiltered-COD removal, and 63–79% of SS removal at an overall HRT of 8 h (6 h for UASB and 2 h for DHS units) [29, 30]. The combined system resulted in excellent organic removal as well as fairly efficient nitrification (52–61% of ammonia-nitrogen removal).

3 Methane Production from Solid Materials Methane fermentation has been commonly used for anaerobic wastewater treatment and energy recovery. In particular, the UASB system has been developed for high-rate methane fermentation in high-strength wastewater [31], where microbial granule formation including acidogens and methanogens is the key factor. However, this process is currently applicable only for wastewater containing a small amount of suspended solid materials. It is, therefore, difficult to apply a UASB system for solid biomass directly; pretreatment to liquefy the suspended solid is required. 3.1 Methane Production from Cellulosic Materials Although cellulose and hemi-cellulose, such as xylan, are the main components in biomass, it is difficult to treat them directly using a high rate methane fermentation process. Nakashimada et al. [32] investigated direct conversion of cellulose to methane using defined methanogens and the anaerobic fungus Neocallimastix frontalis, which is found in ruminants that produce cellulolytic and hemicellulolytic enzymes such as cellulase and xylanase [33, 34]. Therefore, the fungi can degrade biomass such as cellulose and hemicellulose, and produce acetate, formate, H2/CO2, lactate, and small amounts of ethanol, all of which are suitable substrates for methanogens and methanogenic consortia [8, 13, 35]. Co-cultures of N. frontalis with Methanobacterium formicicum and with Methanosaeta concilii were successfully performed in order to produce methane. When cellulose solution was added five times to the bioreactor (24 g L–1 of cellulose was in total), formate and hydrogen accumulations were not observed. After 24 d, 150 mM CH4 was produced and 57 mM acetate remained in the medium. Then, the accumulated acetate was consumed together with methane formation, and after 55 d, 178 mM CH4 was produced (Fig. 3). 3.2 Municipal Sewage Sludge In Japan, municipal sewage sludge produced from aerobic plants for sewage treatment is currently treated by incineration and dumping, but a significant problem is that incineration can result in toxic compounds such as dioxin, and there is limited space for dumping.

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

71

Fig. 3 Fed-batch culture for methane production with a tri-culture of N. frontalis, M. formicicum, and M. concilii in the cellulose medium at 39 °C [32]. Symbols: (■) hydrogen; (●) methane; (■) lactate; (●) acetate; (▲) ethanol; (▲) formate; (◆) pH

Liquefaction of dewatered sewage sludge by a thermal process was intensively investigated. The digested sludge was thermochemically liquefied at 175 °C for 1 min [36–39] or 130 °C for 5 min [40, 41]. The liquefied sludge and its supernatant were successfully anaerobically digested [42]. To carry out anaerobic digestion of sludge without pretreatment, unlike the above, the effect of moisture content on anaerobic digestion of dewatered sewage sludge was tested under mesophilic conditions [43]; here the VS removal efficiency changed from 45.6% to 33.8%, as the moisture content of sludge fed to digester decreased from 97.0% to 89.0%, and the decrease in numbers of glucose consuming acidogens and methanogens was found to correspond to the decrease in the carbohydrate removal efficiency and the accumulation of propionic acid. Since recycling of municipal solid waste is vigorously promoted in Japan, and the necessity of energy recovery from organic waste is increasing, an anaerobic digestion demonstration plant for organic waste in Kyoto City, Japan has been operating for about two years [44]. Three kinds of waste (garbage and leftovers from hotels, yard waste, and used paper) mixed in various ratios are used. The plant has maintained stable operation with each mixture, generating biogas by the decomposition of VS at the rate of about 820 mN3 ton–1 of VS.

72

N. Nishio · Y. Nakashimada

3.3 Coastal Sea Sediments The recent environmental pollution of coastal seawater has resulted in the accumulation of large amounts of organic sediment in mud in some areas at the bottom of the Horoshima Bay in Japan. In particular, mud layers 2–3 m thick have been frequently observed in the oyster farm area due to large amounts of organic matter continuously excreted by oysters. Some purification techniques to remove organic matter from mud sediments have been reported, including the chemical oxidation of mud sediments by ozone gas [45] and the conversion of mud sediments to porous ceramic materials by incineration [46]. However, considering treatment costs, biological treatment (bioremediation) seems to be a preferred technology. It was reported that mud sediment could be treated by anaerobic acidogenic fermentation followed by the cultivation of photosynthetic bacteria [47]. In the acidogenic fermentation, almost all of the fatty acids produced were acetate. The combined addition of vitamins such as nicotinic acid, thiamine and biotin

Fig. 4 The profiles of methane production, acetic acid and pH during methane fermentation of mud sediment in a two-stage UASB reactor system at 37 °C [22]. Symbols: (▲) pH in acidogenic reactor; (■) pH in methanogenic reactor effluent; (▲) acetic acid in acidogenic reactor effluent; (■) acetic acid in methanogenic reactor effluent; (●) methane production in methanogenic reactor

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

73

dramatically influenced production of acetate from sea mud sediment. Furthermore, by using a UASB reactor, and including methanogenic sludge acclimated from the mud sediment, the post-acidogenic fermentation liquor could be treated as mentioned above [22].When a two-stage UASB reactor system was used for the high-rate anaerobic treatment of mud sediment, in each run, acetic acid from the mud sediment was present in the effluent of the acidogenic reactor, yielding approximately 110 mM methane from 278 g wet wt. mud L–1 per 4 d run (Fig. 4). 3.4 Solid Wastes from Traditional Japanese Foods Solid wastes discharges from the processing of traditional Japanese foods and liquor such as tofu and soy sauce could be a target for treatment by methane fermentation. The methane fermentation of bean curd refuse, which is a by-product of producing tofu (soy bean cake) was first characterized by Nakamura et al. [48]. They showed that the COD removal efficiency became about 80% after 25 d from the initiation of the operation. Muroyama et al. [49] found that pretreatment using supersonic irradiation was effective at enhancing the methane yield. Furthermore, they studied the methane fermentation of bean curd refuse using a 1 L reactor with a draft tube that was operated in a fed-batch mode with a once-a-day feeding cycle using two kinds of methanogens [50], and found that the maximum methane yield was as high as 53.7% (very close to the theoretical yield of 55%) although there was a critical substrate loading rate beyond which the operation becomes impossible due to excessive accumulation of unconverted solids. Soy sauce refuse (SSR) is a highly nutritious biomass like bean curd refuse, but it is difficult to process due to its high NaCl content (10% w/w). Some SSR is currently used as animal feed for cattle, but the demand for this may not remain stable into the future. In Japan, the amount of SSR produced is over 100,000 tonnes per year, so treatment of this solid waste is an important and serious problem.Although anaerobic microbial digestion often succeeds in treating wastewater from the food industry, no such method has been reported that can be applied to both solid and high salt waste such as SSR. To reduce MLSS in SSR and produce methane successfully, Nagai et al. [51] reported that thermophilic methanogenic sludge obtained from a municipal wastewater treatment plant could be used as seed. At 25 g-wet wt. L–1 of SSR, 120 mM CH4 production and 50% (w/v) MLSS reduction were observed after 35 d. Acclimatization of the sludge to the waste was effective in increasing the methane production rate in the pH-controlled fed-batch culture using the stirred tank reactor (Fig. 5) [52].

74

N. Nishio · Y. Nakashimada

Fig. 5 Methane fermentation of soy source refuse in fed-batch culture at 50 °C [52]

4 Fermentative Hydrogen Production Hydrogen gas, which is expected to become an important clean alternative energy, can be generated by several microorganisms. In Japan, biological hydrogen production has been studied intensively, using either photosynthetic microorganisms or fermentative anaerobes. Photosynthetic hydrogen production has been researched systematically and particularly thoroughly with the aid of the Japanese government for the last decade, and this work has been reviewed [53–55]. On the other hand, fermentative hydrogen production has also been investigated energetically in Japan because it has some advantages compared to photosynthetic hydrogen production, such as: 1) availability of conventional bioreactor; 2) compactness of process; 3) availability of many kinds of sugar, and; 4) potential high rates of production. However, there are also at least two disadvantages for fermentative hydrogen production compared to photosynthetic production. Theoretically, the maximum yield by fermentative hydrogen production is 4 moles from one mole of glucose. This value is much lower than that possible by photosynthetic production, and actual hydrogen yield is usually lower even than theoretical values. Furthermore, formation of by-products, such as fatty acids and alcohols, are inevitable, so effective utilization of these by-products is necessary for practical fermentative hydrogen production. In this section, we review the progress in fermentative hydrogen production

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

75

with strictly or facultative anaerobes, and discuss some ways of overcoming the above problems in Japan. 4.1 Strictly Anaerobes Strictly anaerobes isolated and characterized for hydrogen production were mainly Clostridium spp., which were applied to treat various kinds of carbohydrates. Clostridium beijerinckii strain AM21B isolated from termites[56] produced 10.9 mmol-H2 g-glucose–1 at uncontrolled pH. It could produce hydrogen from starch, yielding 101.6 mmol of hydrogen from 10 g of starch at pH 6.0, since the strain AM21B produced amylase [57]. Strain AM21B also produced hydrogen from arabinose, cellobiose, fructose, galactose, lactose, sucrose, and xylose at conversion efficiencies ranging from 15.7 to 19.0 mmol g-substrate–1 for 24 h. Clostridium sp. No. 2, isolated from termites, also converted arabinose and xylose to hydrogen with yields of 14.6 and 13.7 mmol g–1 substrate consumed, respectively, which were greater than the maximum value of 11.1 mmol g-glucose–1 consumed in 1 L of batch culture with 10 g of each substrate [58]. Continuous fermentation of xylose and glucose to hydrogen using Clostridium sp. strain No. 2 was carried out in the culture containing 0.3% substrate with the pH controlled at 6.0 [59]. The maximal hydrogen production rates of 21.0 and 20.4 mmol L–1 h–1 were obtained from xylose and glucose with dilution rates of 0.96 and 1.16 h-1, respectively. When enzymatic hydrolysate of Avicel or xylan was used at 0.3% in peptone-yeast extract medium, strain No. 2 produced 19.6 mmol or 18.6 mmol of H2 per gram of substrate consumed, respectively [60]. Furthermore, in continuous culture with Avicel hydrolysate in an aqueous two-phase system, strain No. 2 consumed 3.1 mmol L–1 h–1 of glucose and produced 13.7 mmol L–1 h–1 of hydrogen at a dilution rate of 0.17 h–1 during an 81 h period of stationary culture [61]. To produce hydrogen from xylan, Taguchi et al. isolated Clostridium sp. strain X53 from termites. Strain X53 produced xylanase and hydrogen in a batch culture with xylan [62]. The most efficient production of hydrogen from oat xylan was coincidently achieved under the same conditions as xylanase production, at 40 °C and pH 6.0. The maximum hydrogen evolution rate and the total hydrogen yield were 10.7 mmol L–1 h–1 after 8 h cultivation and 56 mmol L–1 from 10 g L–1 xylan added for 24 h, respectively. C. paraputrificum M-21 isolated from a soil sample collected from Mie University campus utilized chitin and N-acetyl-d-glucosamine (GlcNAc), a constituent monosaccharide of chitin, to produce a large amount of gas along with acetic acid and propionic acid as major fermentation products [63]. The bacterium grew rapidly on GlcNAc with a doubling time of around 30 min, and produced hydrogen, yielding 1.9 mol-H2 mol–1 GlcNAc at initial medium pH 6.5 and 45 °C. Strain M-21 produced 1.5 mol H2 from ball-milled chitin, equivalent to 1 mol of GlcNAc at pH 6.0 [64]. In addition, strain M-21 efficiently degraded

76

N. Nishio · Y. Nakashimada

and fermented ball-milled raw shrimp and lobster shells to produce hydrogen: 11.4 mmol H2 from 2.6 g of the former and 7.8 mmol H2 from 1.5 g of the latter. Characteristics of continuous hydrogen production and fatty acid formation by Clostridium butyricum strain SC-El were examined for vacuum and nonvacuum culture systems [65]. The non-vacuum cultures showed 2.0 to 2.3 mol-H2 mol–1 of glucose and 1.4 to 2.0 mol-H2 mol–1 glucose at 0.5 and 1.0% substrate concentrations, respectively. The vacuum cultures conducted at 0.28 atoms gave 1.8 to 2.3 mol-H2 mol-glucose–1 and 1.3 to 2.2 mol-H2 mol-glucose–1 at the same substrate concentrations, respectively. In addition, the total hydrogen production rate from a two-stage bioreactor consisting of a 1 L anaerobic fermenter (HRT 10 h) and a 4 L photobioreactor (HRT 36 h) feeding at 2.4 L of 1.0% glucose per day was estimated at 1.4 to 5.6 mol-H2 mol-glucose–1, which is 12–47% of the theoretical value. Yokoi et al. reported that a continuously mixed culture of C. butyricum and E. aerogenes removed oxygen in a reactor and produced hydrogen from starch with a yield of more than 2 mol-H2 mol-glucose–1 without any reducing agents in the medium [66], and the repeated batch culture using the same mixed culture produced hydrogen with a yield of 2.4 mol-H2 mol-glucose–1 under a controlled culture pH of 5.25 in a medium consisting of sweet potato starch residue and 0.1% polypepton, without addition of any reducing agents [67]. Furthermore, Rhodobacter sp. M-19 produced hydrogen from the supernatant of the culture broth which was obtained in the repeated batch culture of the mixed cells of C. butyricum and E. aerogenes under a controlled culture pH of 7.5, and the total yield of hydrogen reached 7.0 mol-H2 mol-glucose–1. 4.2 Facultative Anaerobes Japan has been interested in hydrogen production by facultative anaerobes, especially Enterobacter spp., because it has a high growth rate and several advantageous properties similar to those of other members of the Enterobacteriaceae, such as the ability to utilize a wide range of carbon sources, facultative anaerobicity, lack of an inhibitory effect on hydrogen generation under high hydrogen pressure (unlike Clostridia), and a hydrogen evolution rate higher than that of photosynthetic microorganisms. 4.2.1 Hydrogen Production by Enterobacter spp. Tanisho et al. first reported hydrogen production by E. aerogenes E.82005 isolated from the leaves of a plant [68], and investigated the effects of pH and temperature [69], usability of various substrates [70] and the effect of CO2 removal [71] on its hydrogen production. Continuous hydrogen production from molasses was also tested [72, 73].

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

77

Table 2 Yields of products from various carbon sources by E. aerogenes [75]

Substrate

Gluconate Glucose Fructose Galactose Sorbitol Mannitol Glycerol

Cave

3.67 4.00 4.00 4.00 4.33 4.33 4.67

Yield (mmol g-substrate–1) H2

CO2

Ethanol

Acetate

Butanediol

Lactate

1.44 1.97 2.17 1.90 4.96 5.20 6.69

4.85 8.25 8.02 7.87 8.29 7.68 7.59

0.86 2.59 2.73 2.65 5.80 5.30 7.05

2.69 0.81 1.32 1.02 0.74 0.37 0.17

1.59 2.66 2.46 2.61 1.27 1.43 0.15

1.35 1.99 1.53 1.28 1.08 2.15 1.95

Culture conditions: substrate, 10 g L–1; culture time, 14 h.

Rachman et al. also isolated E. aerogenes HU-101 (a hydrogen producer with high growth rate and hydrogen production rate) from anaerobic methanogenic sludge, and tried mutation to enhance the hydrogen yield from glucose as described below [74]. Using E. aerogenes HU-101, Nakashimada et al. [75] clarified the relationship between hydrogen production and carbon source as being a function of the available electrons per carbon for each carbon source (Cave). Since Cave implies the redox state of carbon in each compound, it could be used as the parameter to compare the redox state of each compound, even if the numbers of carbon atoms are different. Hydrogen and ethanol yields increased linearly with Cave, while the acetate yield decreased with Cave (Table 2). This clearly indicated that the redox state of carbon sources directly affects hydrogen production by E. aerogenes. Yokoi et al. isolated aciduric Enterobacter aerogenes HO-39 that has hydrogen-producing ability [76]. The strain was able to grow at an acidic pH of 3.3 aerobically, and at 4.0 anaerobically. Although the optimum pH for hydrogen production was 6.0 to 7.0, hydrogen could be produced at the acidic pH of 4.0, where the yield was 1.0 mol-H2 mol-glucose–1. Hydrogen and a bioflocculant could be produced simultaneously by an anaerobic culture of Enterobacter sp. BY-29 [77, 78]. For production of hydrogen and the bioflocculant in batch cultures, cultivation at 37 °C in a medium containing glucose as a carbon source and polypepton as a nitrogen source was found to be suitable. 4.2.2 Enhancement of Hydrogen Yield of E. Aerogenes The yield of hydrogen from glucose for this bacterium is less than 1 mol molglucose–1, which is lower than that for other bacteria, including Rhodobacter sp. [79] and Clostridium sp. [58]. For industrial-scale hydrogen production by E. aerogenes, enhancement of its hydrogen production ability is needed. In this

78

N. Nishio · Y. Nakashimada

Fig. 6 Schematic diagram of metabolic pathway of various substrates by E. aerogenes

context, mutation has been attempted in order to get a strain with higher hydrogen yield than the wild strain. E. aerogenes produces 2,3-butanediol (BD), ethanol and organic acids (lactate, acetate and formate) as well as hydrogen from various substrates (Fig. 6) [80, 81]. Of these metabolites, ethanol, BD and/or lactate are formed for reoxidation of reducing equivalents such as NADH in order to maintain intracellular redox balance. However, since hydrogen is produced from formate, which is formed via metabolism to acetyl-CoA from pyruvate via pyruvate formate-lyase, it would be possible to produce mostly 2 mol-H2 mol-glucose–1 by stopping the BD and lactate production that hampers hydrogen production. Mutants could be selected by the allyl alcohol (AA) method [74]. In this method, since AA is oxidized by ADH and/or BDDH to a toxic aldehyde (acrolein), mutants deficient in these enzymes can survive [82]. The yield of

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

79

Table 3 Stoichiometric analyses of end products of fermentation by E. aerogenes HU-101 and three mutants grown on glucose minimal medium [74]

Strain

HU-101 A-1 HZ-3 AY-2

Yield (mol mol-glucose–1) H2

Ethanol

Butanediol

Lactate

Acetate

Pyruvate

CO2

0.56 0.84 0.83 1.17

0.49 0.32 0.54 0.34

0.37 0.0 0.46 0.04

0.29 0.55 0.16 0.31

0.17 0.48 0.13 0.14

0.02 0.08 0.09 0.14

1.08 0.88 1.07 1.22

hydrogen from AA-resistant mutant A-1 was 0.84 mol mol-glucose–1, compared to 0.56 mol mol-glucose–1 for HU-101 (Table 3). The yields of lactate and acetate of A-1 were also higher than those of HU-101, although yields of alcoholic metabolites and CO2 were lower, indicating that pyruvate was metabolized to lactate and acetate instead of alcohols with improved hydrogen yield. Alternatively, to isolate non- or low-acid producers, a proton suicide method can be used [74]. This method is based on the lethal effects of bromine and bromite produced from a mixture of NaBr and NaBrO3 during production of acids such as lactate and acetate [83, 84]. The mutant HZ-3 produced hydrogen at almost the same yield, 0.83 mol mol-glucose–1, but less acetate and lactate, and more ethanol, BD and pyruvate than the A-1 mutant (Table 3). Double mutation by the AA and proton suicide methods to block production of both alcoholic and acidic metabolites enhanced hydrogen production compared to single mutation [74]. The hydrogen yield of AY-2, obtained by double mutation with the AA and proton suicide methods, reached 1.17 mol molglucose–1, which is 2.1-fold higher than that of HU-101 (Table 3). The yields of alcoholic and acidic metabolites could be reduced independently by the AA method and the proton suicide method, compared to those in the case of HU-101. In E. aerogenes, hydrogen is usually produced from formate, which is formed by the way of metabolism to acetyl-CoA from pyruvate via pyruvate formatelyase. However, speculating from thermodynamic calibration research, Tanisho et al. proposed that E. aerogenes possessed a hydrogen-producing pathway via NADH as the electron donor [85]. Provided that the hydrogen generation is derived from the pyruvate metabolism to ethanol and acetate during glucose fermentation, the maximum molar hydrogen yield should not exceed the sum of the ethanol and acetate produced. In the experiments by Rachman et al., however, even if acetate and ethanol production was reduced by mutation, hydrogen production was much enhanced [74]. This suggested that hydrogen should be generated from excess NADH. When in vitro enzymatic evolution of hydrogen from NADH was tested, both NADH and NADPH actually supported hydrogen formation from an extract of E. aerogenes after a reaction time of 48 h, as shown in Fig. 7 [75]. Furthermore, it was found that NAD(P)H-depen-

80

N. Nishio · Y. Nakashimada

Fig. 7 NADH- and NADPH-dependent H2 evolution in cell-free extract from E. aerogenes AY-2 [75]. Experiments were carried out in vials without pH control (initial pH was 6.3), at 37 °C, anaerobically. H2 evolution was determined after 48 h of reaction. Symbols: (●) H2 evolution using NADH; (●) H2 evolution using NADPH; (■) without reductant as a substrate

dent hydrogenase was localized in the cell membrane [75], like the membranebound hydrogenase that was reported for Klebsiella pneumoniae, a close relative of E. aerogenes [86]. If all NADH is converted to hydrogen, the theoretical maximum hydrogen yield is 4 mol mol-glucose–1, like Clostridia. The breeding of such a mutant will make hydrogen production by E. aerogenes more attractive. 4.3 High Rate Hydrogen Production In order to continuously produce hydrogen from biological resources for industrial use, the hydrogen evolution rate is as important as its overall yield. Therefore, for the practical operation of a bioreactor, a high cell density is required. To achieve this end, a variety of reactor systems with immobilized cells using several microbial support carriers have been reported. These systems include a polyurethane form for E. aerogenes E 82005, giving a hydrogen evolution rate of 13 mmol L–1 h–1 using molasses as a substrate [73]; agar gel or porous glass beads for aciduric E. aerogenes HO-39, giving 850 mL L–1 h–1 from glucose [87]; porous glass beads for Clostridium butyricum in a column reactor giving 715 and 1,150 mL L–1 h–1 at retention times of 2.0 and 1.0 h, respectively [88]. Because any matrix of support carriers occupies considerable space in the reactor, microbial cell density is limited by the size and the porosity of the carrier. Flocculation of E. aerogenes was firs reported by Tanisho et al. [72] although they did not effectively use the flocculant ability in their study. Moreover, it has

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

81

Fig. 8 Flocculated cells of E. aerogenes HU-101 in packed-bed reactor

been reported that Enterobacter sp. BY-29 produces a new biopolymer flocculant consisting of polysaccharides [77]. Using the flocculation ability of E. aerogenes HU-101 and AY-2 [74], continuous hydrogen production without any carrier in the packed-bed reactor was investigated [89]. In the culture using glucose, the flocculated cells continued to accumulate during the continuous culture, as shown in Fig. 8. In the case of AY-2, cell density in the packed-bed was estimated at 17 mg cm–3. For the HU-101 strain, the hydrogen evolution rate was 30 mmol L–1 h–1 at 0.67 h–1 of dilution rate. In contrast, for the AY-2 strain, the hydrogen evolution rate was 58 mmol L–1 h–1 at 0.67 h–1 (Fig. 9). The molar hydrogen yield of AY-2 was found to be more than 1.1 at dilution rates from 0.13 to 0.55 h–1 when the pH in the effluent was kept above 6.0.

Fig. 9 Fermentation performance of a mutant AY-2 in packed-bed reactor [89]. Symbols; (●) H2 evolution rate; (■) residual glucose; (●) optical density in the effluent; (▲) height of bed of flocculated cells in the reactor from the bottom; (■) dilution rate

82

N. Nishio · Y. Nakashimada

5 Hydrogen-Methane Production from Organic Wastes To treat wastewaters and wastes, use of microbial consortia might be feasible, because it is impossible to keep a pure culture without contamination. Also, as mentioned above, it is well-known that many kinds of hydrogen-producing bacteria work in methane fermentation, especially at the hydrolysis and acidogenesis stages. If high rate and high yield of hydrogen fermentation are achieved, the produced hydrogen gas might be connected directly to a “fuel cell” without reforming. Furthermore, methane can be produced from fatty acids such as acetate, propionate and butyrate that remain in the liquid after hydrogen fermentation. We have proposed this process as a H2-CH4 two-stage (Hy-Met) process, as illustrated in Fig. 10. There are some reports that focus on hydrogen production at the hydrolysis and acidogenesis stages. The acidogenic phase of a continuous stirred tank reactor inoculated with well acclimated sludge produced a substantial amount of hydrogen from glucose [90]. Hydrogen content of more than 65% was observed at high dilution rate and 71% at low dilution rate. A gas production rate of more than 101 d–1 was also observed from the chemostat reactors. The effect of solid retention time (SRT) on hydrogen gas production, glucose degradation, and anaerobic bacteria in anaerobic treatment processes were investigated using a 11.7 g L–1 glucose solution as a substrate [91]. Counts of general anaerobic bacteria and acid-forming bacteria in each reactor increased with increasing SRT, and counts of genus Clostridium and sulfate-reducing bacteria in the reactor decreased with increasing SRT of the reactor.When SRT increased from 2 to 10 h, the hydrogen content in the gas decreased from 12 to 9%. The capability of natural anaerobic microflora to produce hydrogen was examined by Ueno et al. with artificial wastewater containing cellulose [92]. They reported that thermophilic anaerobic microflora enriched from sludge compost produced a significant amount of hydrogen (2.4 mol mol-hexose–1). The microbial community in the microflora was investigated through isolation of

Fig. 10 Schematic diagram of Hy-Met (H2-CH4 two-stage) process

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

83

the microorganisms by both plating and denaturing gradient gel electrophoresis (DGGE) of the PCR-amplified V3 region of 16S rDNA [93]. Most of the isolates belonged to the cluster of the thermophilic Clostridium/Bacillus subphylum of low G+C gram-positive bacteria. Thermoanaerobacterium thermosaccharolyticum was isolated in the enrichment culture, and was detected with strong intensity by PCR-DGGE. Two other thermophilic cellulolytic microorganisms, Clostridium thermocellum and Clostridium cellulosi, were also detected by PCR-DGGE, although they could not be isolated. To evaluate the performance of hydrogen production from actual wastes, continuous production of hydrogen from sugary wastewater by anaerobic microflora in chemostat culture was examined as a function of HRT in the reactor [94]. The measured volumes of the evolved gas at each HRT were almost constant (avg. 3590 mL L-feed–1). Steady states on gas evolution were observed for 190 d at HRTs from 0.5 to 3 d, giving hydrogen production rates from 198 to 34 mmol L–1 d–1. A maximum production yield of hydrogen of 14 mmol g–1 carbohydrate removed was obtained at an HRT of 0.5 d. To assess the feasibility of hydrogen production from organic fraction municipal solid waste (OFMSW), two seed microorganisms, namely heat-pretreated digested sludge and hydrogen-producing bacteria enriched from soybean-meal silo, were utilized [95]. When experiments were carried out according to a full factorial central composite experimental design, a simple model developed from the Gompertz equation was found to be suitable for estimating the hydrogen production potential and rate. Through response surface methodology, it was shown that high hydrogen production potentials of 140 and 180 mL-H2 g-TVS–1 occurred when the pretreated digested sludge and the hydrogenproducing bacteria consumed OFMSW, respectively. Further experiments confirmed that the results of this study were highly reliable and that the OFMSW has considerable potential for biological hydrogen production. Hydrogen production from OFMSW was also investigated in detail using sucrose as the substrate. The effect of the iron concentration in the external environment on hydrogen production was studied at 37 °C [96]. The maximum specific hydrogen production rate was found to be 24.0 mL g–1-VSS h–1 at 4000 mg FeCl2 L–1. The maximum hydrogen production yield of 131.9 mL g–1 sucrose was obtained at an iron concentration of 800 mg FeCl2 L–1. The influence of initial pH of the culture medium on hydrogen production was also studied; hydrogen production was not observed at pH values of 3.0, 11.0 and 12.0, but low production was observed at pH values of 5.0 and 5.5. The maximum specific hydrogen production rate was 37 cm3 g-VSS–1 h–1 [97]. Looking into the inhibitory factors on hydrogen production with microbial consortia, Noike et al. reported the inhibitory effects of lactic acid bacteria on hydrogen fermentation of organic waste [98]. They suggested that the inhibitory effect of lactic acid bacteria on hydrogen production was caused by bacteriocins excreted from the bacteria that had a deleterious effect on other bacteria, and found that the inhibition on hydrogen production was reduced by heat treatment for 30 min at temperatures ranging from 50 °C to 90 °C.

84

N. Nishio · Y. Nakashimada

To elucidate the Hy-Met process, H2-CH4 two-stage fermentation by microbial consortium for solid waste from bread manufacture was tested [99]. 100 gwet wt. L–1 of the bread waste was fermented by microbial consortium under pH-uncontrolled (initial pH was 7.0) and pH-controlled conditions at pH of 7.0 or 5.0. Although pH-uncontrolled conditions gave only 70 mM H2 with 80% SS reduction, 91% SS was reduced after 24 h under the pH 7 controlled condition, and 240 mM of H2 was produced. The culture broth contained 150 mM each of acetate and butyrate, and the total organic carbon was ca. 20,000 ppm. On the other hand, under pH 5.0 controlled conditions, only 100 mM of H2 was produced, and lactate was mainly produced (ca. 220 mM) in the culture broth. The culture broth for hydrogen fermentation of the bread waste usually contained about 20,000 ppm of TOC as mentioned above. This culture broth was diluted to give 2000–5000 ppm TOC, and was fed continuously into the UASB reactor where acclimatized methanogenic granules were inoculated.When the organic carbon loading rate was increased by the stepwise increase of the dilution rate, the optimum loading was 9.5 g-TOC L–1 d–1, giving 80% TOC removal, 400 mmol L–1 d–1 of methane production rate, and ca. 0.6 of methane yield as carbon base. These results show that when the reactor volume of hydrogen and methane fermentation are set up in the ratio of 1:2.1, 91% of SS should be reduced at 29 g-wet wt. L–1 d–1 of loading rate, and hydrogen and methane yields should be 2.4 mol and 8.6 mol kg-wet wt.–1, respectively.

6 Concluding Remarks Since there are few natural resources such as petroleum and natural gas in Japan, the demand for saving energy, and the recovery of energy from unused matter is traditionally high, especially since the “oil shock”. This tendency has been intensified by the fears of depletion of fossil fuels and global warming. Therefore, in Japan high rate methane fermentation, typified by the UASB reactor, has become widely-used, and the improvement and modification of the high rate methane fermentation process has been carried out over the last decade in order to expand the practical applicability of it to various other organic matter, as reviewed above. Furthermore, to avert global warming, hydrogen has become the focus of attention, since it is seen as a potential clean future energy source. To this end, the biological production of hydrogen from biomass has been investigated using various microorganisms and microbial consortia. Since biologicallyproduced hydrogen from biomass is renewable and clean, if a cost-effective process for biological hydrogen production is developed in the future, it will contribute greatly to the improvement of the global environment. A combined hydrogen/methane production process, “Hy-Met”, might be a candidate for improving the yield and production rate of hydrogen still further. Such challenging research will be continued by many researchers all over the world in years to come.

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes

85

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

Yokotsuka T (1986) Adv Food Res 30:195 Alvin IK (1979) Enzyme Microb Tech 1:165 Young JC, McCarty PL (1969) J Wat Pollut Control Fed 41:160 Lettinga G, Roersma R, Grin P (1983) Biotechnol Bioeng 25:1701 Jowell WJ, Switzenbaum MS, Morris JW (1981) J Wat Pollut Control Fed 53:482 Sutton PM, Li A (1981) Proc 36th Industrial Waste Conf, May 1981, West Lafayette, IN, p 665 Nishio N, Silveira RG, Hamato K, Nagai S (1993) J Ferment Bioeng 75:309 Fukuzaki S, Nishio N, Nagai S (1995) J Ferment Bioeng 79:354 Chang Y-J, Nishio N, Nagai S (1995) J Ferment Bioeng 79:348 Chang Y-J, Nishio N, Maruta H, Nagai S (1993) J Ferment Bioeng 75:430 Fukuzaki S, Nishio N, Sakurai H, Nagai S (1991) J Ferment Bioeng 71:50 Fukuzaki S, Nishio N (1997) J Ferment Bioeng 84:382 Fukuzaki S, Chang Y-J, Nishio N, Nagai S (1991) J Ferment Bioeng 72:465 Harada H, Uemura S, Chen AC, Jayadevan J (1996) Biores Technol 55:215 Kida K, Tanemura K, Sonoda Y, Hikami S (1994) J Ferment Bioeng 77:90 Tagawa T, Takahashi H, Sekiguchi Y, Ohashi A, Harada H (2002) Wat Sci Technol 45:225 Kida K, Tanemura K, Sonoda Y (1993) J Ferment Bioeng 76:510 Uemura S, Harada H (2000) Biores Technol 72:275 Sonoda Y, Seiko Y (1977) Hakkokogaku Kaishi 55:22 Rinzema A, van Lier JB, Lettinga G (1988) Granular anaerobic sludge – microbiology and technology: Proceedings of the GASMAT-Workshop, Oct 1987, Lunteren, Netherlands. Pudoc, Wageningen, Netherlands, p 216 Tanemura K, Kida K, Ikbal, Matsumoto J, Sonoda Y (1994) J Ferment Bioeng 77:188 Takeno K, Nakashimada Y, Kakizono T, Nishio N (2001) Appl Microbiol Biot 56:280 Takeno K, Nakashimada Y, Kakizono T, Nishio N (2000) Proc 5th Int Symp Environmental Biotechnology, 9–13 July 2000, Kyoto, Japan, p 746 Chang YJ, Nishio N (1994) J Ferment Bioeng 77:450 Sawayama S, Yagishita T, Tsukahara K (1999) J Biosci Bioeng 87:258 Sawayama S, Tsukahara K, Yagishita T, Hanada S (2001) J Biosci Bioeng 91:195 Sawayama S, Hanada S, Kamagata Y (2000) J Biosci Bioeng 89:396 Machdar I, Harada H, Ohashi A, Sekiguchi Y, Okui H, Ueki K (1997) Wat Sci Technol 36:189 Machdar I, Sekiguchi Y, Sumino H, Ohashi A, Harada H (2000) Wat Sci Technol 42:83 Uemura S, Takahashi K, Takaishi A, Machdar I, Ohashi A, Harada H (2002) Wat Sci Technol 46:303 Pette KC,Versprille AI (1982) In: Wentworth RL (ed) Anaerobic Digestion 1981: Proc 2nd Int Symp Anaerobic Digestion, Travemunde, Germany. Elsevier, Amsterdam, p 121 Nakashimada Y, Kartikeyan S, Murakami M, Nishio N (2000) Biotechnol Lett 22:223 Teunissen MJ, Baerends RJS, Knelissen RAG, Op den Camp HJM,Vogels GD (1992) Appl Microbiol Biot 38:28 Kartikeyan S, Murakami M, Nakashimada Y, Nishio N (2001) J Biosci Bioeng 91:153 Nishio N, Kayawake E, Nagai S (1985) J Ferment Technol 63:205 Sawayama S, Inoue S, Tsukahara K, Yagashita T, Minowa T, Ogi T (1999) Renew Energ 16:1094 Sawayama S, Inoue S, Yagishita T, Ogi T, Yokoyama SY (1995) J Ferment Bioeng 79:300 Sawayama S, Inoue S, Tsukahara K, Ogi T (1996) Biores Technol 55:141 Inoue S, Sawayama S, Ogi T, Yokoyama SY (1996) Biomass Bioener 10:37

86

N. Nishio · Y. Nakashimada

40. 41. 42. 43. 44.

Tanaka S, Kobayashi T, Kamiyama K, Bildan M (1997) Wat Sci Technol 35:209 Tanaka S, Kamiyama K (2002) Wat Sci Technol 46:173 Nagai S, Kawasugi T, Mazumder TK, N N (1986) Chem Eng Commun 45:83 Fujishima S, Miyahara T, Noike T (2000) Wat Sci Technol 41:119 Komatsu T, Kimura T, Kuriyama Y, Isshiki Y, Kawano T, Hirao T, Masuda M, Yokohama K, Matsumoto T, Takeda M (2002) Water Sci Technol 45:113 Imaoka T, Osumi H (1997) J Water Waste (in Japanese) 39:587 Kai T, Sera T, Sato K, Kajitani J (1997) Haikan-Gijyutu (in Japanese) 39:26 Takeno K, Sasaki K, Watanabe M, Kaneyasu T, Nishio N (1999) J Biosci Bioeng 88:410 Nakamura K, Kameyama T, Ohtaki S, Tada T (1989) Mizushorigijutu (in Japanese) 30:33–39 Muroyama K, Ohta T, Yamade K (1994) Environ Conserv Eng (in Japanese) 23:621–629 Muroyama K, Mochizuki T, Wakamura T (2001) J Biosci Bioeng 91:208 Nagai H, Kobayashi M, Tsuji Y, Takata K, Nakashimada Y, Kakizono T, Nishio N (2000) J Jpn Soy Sauce Res Inst 26:295 Nakashimada Y, Kobayashi M, Nomura N (2002) Bioscience and Industry (in Japanese) 60:807 Miyake J, Miyake M, Asada Y (1999) J Biotechnol 70:89 Asada Y (1998) Stud Surf Sci Catal 114:321 Asada Y, Miyake J (1999) J Biosci Bioeng 88:1 Taguchi F, Chang JD, Mizukami N, Saitotaki T, Hasegawa K, Morimoto M (1993) Can J Microbiol 39:726 Taguchi F, Mizukami N, Hasegawa K, Saitotaki T, Morimoto M (1994) J Ferment Bioeng 77:565 Taguchi F, Mizukami N, Hasegawa K, Saito-Taki T (1994) Can J Microbiol 40:228 Taguchi F, Mizukami N, Taki TS, Hasegawa K (1995) Can J Microbiol 41:536 Taguchi F, Mizukami N, Yamada K, Hasegawa K, Saitotaki T (1995) Enzyme Microbial Tech 17:147 Taguchi F, Yamada K, Hasegawa K, TakiSaito T, Hara K (1996) J Ferment Bioeng 82:80 Taguchi F, Hasegawa K, SaitoTaki T, Hara K (1996) J Ferment Bioeng 81:178 Evvyernie D, Yamazaki S, Morimoto K, Karita S, Kimura T, Sakka K, Ohmiya K (2000) J Biosci Bioeng 89:596 Evvyernie D, Morimoto K, Karita S, Kimura T, Sakka K, Ohmiya K (2001) J Biosci Bioeng 91:339 Kataoka N, Miya A, Kiriyama K (1997) Wat Sci Technol 36:41 Yokoi H, Tokushige T, Hirose J, Hayashi S, Takasaki Y (1998) Biotechnol Lett 20:143 Yokoi H, Saitsu A, Uchida H, Hirose J, Hayashi S, Takasaki Y (2001) J Biosci Bioeng 91:58 Tanisho S, Wakao N, Kosako Y (1983) J Chem Eng Japan 16:529 Tanisho S, Suzuki Y, Wakao N (1987) Int J Hydrogen Energ 12:623 Tanisho S, Tu H, Wakao N (1989) Hakkokogaku Kaishi 67:29 Tanisho S, Kuromoto M, Kadokura N (1998) Int J Hydrogen Energ 23:559 Tanisho S, Ishiwata Y (1994) Int J Hydrogen Energ 19:807 Tanisho S, Ishiwata Y (1995) Int J Hydrogen Energ 20:541 Rachman MA, Furutani Y, Nakashimada Y, Kakizono T, Nishio N (1997) J Ferment Bioeng 83:358 Nakashimada Y, Rachman MA, Kakizono T, Nishio N (2002) Int J Hydrogen Energ 27:1399 Yokoi H, Ohkawara T, Hirose J, Hayashi S, Takasaki Y (1995) J Ferment Bioeng 80:571 Yokoi H, Yoshida, Mori S, Hirose J, Hayashi S, Takasaki Y (1997) Biotechnol Lett 19:569 Yokoi H, Aratake T, Hirose J, Hayashi S, Takasaki Y (2001) World J Microb Biot 17:609

45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

High Rate Production of Hydrogen/Methane from Various Substrates and Wastes 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99.

87

Miyake Y, Mao XY, Kawamura S (1984) J Ferment Technol 62:531 Magee RJ, Kosaric N (1987) Adv Appl Microbiol 32:89 Johansen L, Bryn K, Stormer FC (1975) J Bacteriol 123:1124 Dürre P, Kuhn A, Gottschalk G (1986) FEMS Microbiol Lett 36:77 Winkelman JW, Clark D (1984) J Bacteriol 160:687 Pablo HC, Méndez BS (1990) Appl Environ Microbiol 56:578 Tanisho S, Kamiya N, Wakao N (1989) Biochim Biophys Acta 973:1 Steuber J, Krebs W, Bott M, Dimroth P (1999) J Bacteriol 181:241 Yokoi H, Tokushige T, Hirose J, Hayashi S, Takasaki Y (1997) J Ferment Bioeng 83:481 Yokoi H, Maeda Y, Hirose J, Hayashi S, Takasaki Y (1997) Biotechnol Tech 11:431 Rachman MA, Furutani Y, Nakashimada Y, Kakizono T, Nishio N (1998) Appl Microbiol Biot 49:450 Akashah M, Yoshida M, Watanabe M, Nakamura M, Matsumoto Ji (1997) Wat Sci Technol 36:279 Nakamura M, Kanbe H, Matsumoto JI (1993) Wat Sci Technol 28:81 Ueno Y, Kawai T, Sato S, Otsuka S, Morimoto M (1995) J Ferment Bioeng 79:395 Ueno Y, Haruta S, Ishii M, Igarashi Y (2001) Appl Microbiol Biot 57:555 Ueno Y, Otsuka S, Morimoto M (1996) J Ferment Bioeng 82:194 Lay JJ, Lee YJ, Noike T (1999) Water Res 33:2579 Lee YJ, Miyahara T, Noike T (2001) Biores Technol 80:227 Lee YJ, Miyahara T, Noike T (2002) J Chem Technol Biot 77:694 Noike T, Takabatake H, Mizuno O, Ohba M (2002) Int J Hydrogen Energ 27:1367 Nishio N, Nakashimada Y (2003) In: Karim MIA (ed) Biotechnology for sustainable utilization of biological resources in the tropics, vol 16. IC Biotech (in press)

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 89 –111 DOI 10.1007/b94193 © Springer-Verlag Berlin Heidelberg 2004

Bacterial Capsular Polysaccharide and Sugar Transferases Katsuhide Miyake1 (✉) · Shinji Iijima2 1

2

Research Center for Advanced Waste and Emission Management, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan [email protected] Department of Biotechnology, School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

90

2 Physiological Effects of the Bacterial CPs . . . . . . . . . . . . . . . . . . . . .

92

3 Capsular Polysaccharide Synthesis Genes . . . . . . . . . . . . . . . . . . . . .

96

4 Function of Streptococcal Glycosyltransferases . . . . . . . . . . . . . . . . . . 103 5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Abstract Capsular polysaccharides (CPs) of several pathogenic bacteria are thought to be good materials for the development of new therapeutic reagents. These polysaccharides can be used as vaccines against infection of pathogenic bacteria and are also useful as inhibitors for disease caused by aberrant and abnormal cell-cell interaction, such as cancer metastasis and inflammation. Since bacterial CPs are diverse in structure and these bacteria have a variety of sugar transferases responsible for the synthesis of CPs, bacterial CP synthesis (cps) genes have attracted much interest as a source of glycosyltransferases useful for glycoengineering. In this review, we describe physiological effects of the bacterial CPs on mammalian cells, and the structure and function of the cps genes, by focusing on group B streptococci, Streptococcus agalactiae type Ia and Ib, that produce high-molecular weight polysaccharides consisting of the following pentasaccharide repeating units: Æ4)-[a-DNeupNAc-(2Æ3)-b-D-Galp-(1Æ4)-b-D-GlcpNAc-(1Æ3)]-b-D-Galp-(1Æ4)-b-D-Glcp–(1Æ and Æ4)-[a-D-NeupNAc-(2Æ3)-b-D-Galp-(1Æ3)-b-D-GlcpNAc-(1Æ3)]-b-D-Galp-(1Æ4)b-D-Glcp–(1Æ, respectively. Keywords Capsular polysaccharide · cps gene · Sialyl Lewis oligosaccharide · Streptococcus agalactiae Abbreviations CP Capsular polysaccharide cps Capsular polysaccharide synthesis GBS Group B streptococci HUVECs Human umbilical cord vein endothelial cells ELAM-1 Endothelial leukocyte adhesion molecule-1

90

K. Miyake · S. Iijima

1 Introduction Encapsulated bacteria are frequently associated with serious diseases in both humans and animals. Bacterial capsular polysaccharides (CPs) are generally composed of repeating oligosaccharides, consisting of two to ten monosaccharides, and are sometimes complemented with other components. The CPs of pathogenic bacteria confer resistance to complement-mediated opsonophagocytosis [1]. In addition, some bacteria have CPs that mimic host molecules to avoid the specific immune response of the host [2]. Lancefield’s group B streptococci (GBS), Streptococcus agalactiae, is well adapted to asymptomatic colonization of adult humans. It is frequently found in the gastrointestinal and the genitourinary tract, but it also predominantly causes invasive bacterial disease in the neonate. S. agalactiae is one of the human pathogens that causes invasive diseases such as sepsis, meningitis and pneumonia in infants, which is observed in two to three cases per 1000 live births [3]. Infection in the rectum and vagina of pregnant women with GBS, which causes infection of the amniotic cavity, is correlated with GBS sepsis in infants. These bacteria are gram-positive microorganisms that have two distinct polysaccharide antigens. One of these, group antigen (C-substance) composed of a number of rhamnose units, is common to all strains. The others are typespecific CPs that distinguish S. agalactiae into nine serotypes. Of the nine serotypes described so far, the types Ia, Ib, II, III, and V are responsible for the majority of invasive human GBS diseases. The chemical structures of these polysaccharides have already been determined [4–11] (Fig. 1). Streptococcus pneumoniae (pneumococcus) and Streptococcus pyogenes (group A streptococci) are also human pathogens causing serious diseases such as meningitis, pneumonia, bacteremia and otitis. These bacteria also produce type-specific CPs, as is the case with GBS. The CPs of these streptococci have been recognized as the major virulence factor, since there are several reports that unencapsulated streptococcal and pneumococcal mutants often lost their pathogenicity. We are particularly interested in the type-specific CPs of S. agalactiae type Ia and Ib. These polysaccharides have a linear backbone of 4)-b-D-Glcp-(1–4)-b-D-Galp-(one repeating unit with trisaccharide side chains of a-NeupNAc-(2–3)-b-D-Galp-(1–4 or 1–3)b-D-GlcpNAc-(1 linked to C3 of each b-D-galactose residue of the backbone (Fig. 1). The type specific polysaccharides from S. agalactiae Ia and Ib have very similar structures to those of sialyl Lewis carbohydrates. These polysaccharides consist of a NeuNAc-Gal-GlcNAc trisaccharide unit, which is a structure specific to sialyl Lewis carbohydrates, but does not contain a branched fucose residue observed in the cancer-specific carbohydrates. Therefore, these polysaccharides seem to be useful materials for therapeutic reagents. Structures of other streptococcal CPs are less complicated than those of GBS. These saccharides are now thought to be useful as materials for vaccines to prevent infection of these bacteria.

Bacterial Capsular Polysaccharide and Sugar Transferases

91

Fig. 1 Subunit structures of CPs from S. agalactiae types Ia, Ib, III, IV,V, and VI. Glc: glucose; Gal: galactose; GlcNAc: N-acetylglucosamine; NeuNAc: N-acetylneuraminic acid

92

K. Miyake · S. Iijima

Recently, the genes involved in CP synthesis (cps) and the mechanisms of biosynthesis have been reported in many bacteria [12–16]. The biosynthesis of CPs is a complex enzymatic pathway starting with the uptake or synthesis of the monosaccharides and their activation to nucleotide derivatives. Membranebound transferase complexes then catalyze the successive coupling of the monosaccharides to a membrane-bound lipid carrier, followed by polymerization of the sugar subunits and subsequent export and attachment of the complete CP to the cell surface [13, 14]. Until now, thirteen pneumococcal cps gene clusters and five cps clusters of GBS have been identified and analyzed. Furthermore, complete genome sequences have been determined with several strains of GBS, S. pneumoniae and S. pyogenes. We have also determined cps sequences of GBS type Ia and type Ib. These approaches to CP synthesis are very important for fighting against streptococcal diseases by developing drugs capable of inhibiting the synthesis of the CPs. A rational approach to this aim requires a deep knowledge of the genes and proteins involved in the biosynthetic process. Furthermore, this knowledge seems to also be helpful for glycoengineering using enzymes of the cps gene cluster. In this review, physiological effects of the CPs on mammalian cells, and structure, expression and functions of the cps gene clusters required for synthesis of CPs are also described.

2 Physiological Effects of the Bacterial CPs CPs of pathogenic bacteria often mimic carbohydrates of mammalian cells to evade the host immune system. Therefore, the polysaccharides could affect some important steps of the reaction of a living body and be potentially useful as materials for development of therapeutic reagents. In this section, we describe the physiological effects of the CPs of GBS on the interaction between cancer cells and endothelial cells. The most fearsome aspect of cancer is metastasis and the invasion of cancer cells into surrounding tissues and various organs, which often results in death. Therefore, it is very important to develop effective cancer metastasis inhibitors. The process of cancer metastasis consists of several steps: detachment of cancer cells from the original tumor; invasion into a blood vessel; spreading to various and distant places through the blood stream; and invasion of nearby organs from blood vessel. In the last step of metastasis, sialyl Lewis carbohydrates (sialyl Lea and Lex) of cancer cells interact with ELAM-1 of endothelial cells (rolling), then tight binding of cells is formed through other attachment factors [17]. Therefore, sialyl Lea and Lex carbohydrates play a key role in metastasis by mediating cell-cell interaction between cancer and endothelial cells. Type-specific polysaccharides from group B streptococci, S. agalactiae Ia and Ib, have very similar structures to those of sialyl Lewis carbohydrates (Fig. 2). Therefore, we investigated inhibitory effects of these bacterial poly-

Fig. 2 Comparison of cell surface polysaccharides from S. agalactiae with mammalian carbohydrates

Bacterial Capsular Polysaccharide and Sugar Transferases 93

94

K. Miyake · S. Iijima

Fig. 3 Gel filtration of polysaccharides from S. agalactiae. Total polysaccharides (empty circles) were analyzed by phenol-sulfuric acid method and type-specific polysaccharides (filled circles) were detected by ELISA using type-specific antisera

saccharides on the adhesion between endothelial cells and cancer cells, by using normal human endothelial cells (HUVEC) which express ELAM-1 to confirm the usefulness of the polysaccharides as cancer metastasis inhibitors [18]. GBS, S. agalactiae usually produces two classes of CPs. One is a type-specific polysaccharide which is similar to sialyl Lewis carbohydrates, and the other is a group B specific polysaccharide produced by any type of group B streptococci. To purify these polysaccharides, type specific and group specific antisera were separated by gel filtration chromatography. On the Sepharose 6B column, type-specific polysaccharides were eluted in high molecular weight fractions (Fig. 3) and these fractions showed little reactivity with group B specific antiserum.According to elution profiles from the gel filtration, the molecular weight of the type-specific polysaccharides ranged from about 400 to 700 kDa, which means at least 400 NeuNAc-Gal-GlcNc repeats are included in one polysaccharide molecule. The final yield was about 1–3 mg of purified type-specific polysaccharides per 500 mL culture. These polysaccharides were used in the cell-cell adhesion assay to demonstrate the inhibitory effects of metastasis. As shown in Fig. 4, the type-specific polysaccharides exhibited inhibitory effects on cell adhesion. The adhesion of Colo201 human cancer cells to IL-1bstimulated normal HUVECs was clearly inhibited by both type Ia and Ib specific polysaccharides. We were therefore able to observe the same inhibitory effects when HL60 cells were used (Fig. 4). The adhesion was reduced to 10–40% of the control and the type-specific Ia and Ib polysaccharides showed the inhibitory effects at 10 mg mL–1. This inhibition was probably due to specific structures of the type-specific polysaccharides, because the other types of polysaccharide – for instance colominic acid (polysialic acid) – showed only a little inhibitory effect at the same concentration (Fig. 4). We were able to detect the ability of type-specific polysaccharides from S. agalactiae Ia and Ib to in-

Fig. 4 Inhibitory effects of type-specific polysaccharides from S. agalactiae on adhesion of Colo201 and HL60 human cancer cells to interleukin-1b activated normal and immortalized HUVECs

Bacterial Capsular Polysaccharide and Sugar Transferases 95

96

K. Miyake · S. Iijima

hibit adhesion of cancer cells to human endothelial cells. In the experiment shown in Fig. 4, multivalent interaction between the sugar molecules and cell surface ELAM-1 may enhance the inhibitory effect. Therefore, the inhibition of cell adhesion seems to be due to multivalent sialyllactosamine structure. For instance, it is also expected that the inhibitory effect of the CPs can be increased by the introduction of fucose residues to CPs. Although we also need to investigate these effects in vivo, these results suggest that the streptococcal polysaccharides may be very useful for development of new cancer metastasis inhibitors.

3 Capsular Polysaccharide Synthesis Genes Recently, genes involved in CP synthesis (cps) and the mechanisms of biosynthesis have been elucidated in many bacteria [12–16]. Several groups including ours have analyzed cps gene clusters in many S. pneumoniae strains [19–23] and several S. agalactiae strains [24–28]. Since bacterial CPs are diverse in structure, and bacteria have a variety of sugar transferases responsible for the synthesis of CP, bacterial cps genes have attracted much interest as a source of glycosyltransferases useful for glycoengineering. Until now, the cps gene cluster of 13 out of the 90 known pneumococcal types and 5 out of 9 GBS has been sequenced. However, group A streptococci does not have cps gene cluster for CP synthesis. In these strains, genes for CP synthesis are probably dispersed on the chromosome. Structures of the cps gene clusters of GBS and pneumococci were very similar between each strain (Fig. 5). These cps gene clusters contain regulator of CP synthesis, sugar transferases, and transporter genes. In this chapter, we describe cps cluster of S. agalactiae [26–28]. The DNA sequence of 27,140 nucleotides of S. agalactiae type Ia cps locus was completely determined on both strands with overlapping clones covering the cps gene locus. Sequence analysis showed 23 complete open reading frames (ORFs), designated as cpsIaS to cpsIaL, neuB, C, D, A, orf1, ung, parE, C and bcat.All ORFs were in the same orientation except for cpsIaR, and were spaced one behind the other at short distances (Fig. 5). Four potential –35 and –10 promoter sequences were identified upstream of cpsIaR, cpsIaA, cpsIaE and orf1, respectively. Two putative Rho-independent transcription terminator sequences were found downstream of orf1 (DG=–24.5 kcal/mol) and ung (DG= –32.5 kcal/mol). These observations suggested that at least 17 ORFs from cpsIaA to orf1 may constitute one polycistronic operon and that transcription may start from cpsIaA, cpsIaE and orf1. The average G+C content of the sequenced area was 31.7%. The percent G+C content of the cps cluster agreed well with that of chromosomal DNA of S. agalactiae (34.0%) [29]. The amino acid sequence of each ORF was deduced and an overview of all cpsIa genes with their properties and translation products is shown in Table 1. CpsIaR shows 44% homology with Klebsiella aerogenes NAC protein and 41%

Fig. 5 Structure of cps locus of S. agalactiae type Ia and Ib compared with that of the corresponding region of S. pyogenes. Figures in the box show the percentage of homology with translational products

Bacterial Capsular Polysaccharide and Sugar Transferases 97

98

K. Miyake · S. Iijima

Table 1 Properties of the ORFs in the cps locus of S. agalactiae type Ia and homologies with gene products of other bacteria

ORF

Number of amino acids

Proposed function of gene product

Similar gene products (% identity)

CpsIaR

307

Regulation

Streptococcus pyogenes M1CpsY (82.4%) Klebsiella aerogenes NAC (44%) Pseudomonas sp. TcbR (41%)

CpsIaA

485

Regulation

Streptococcus pneumoniae serotype 14 Cps14A (50.2%) Streptococcus pneumoniae serotype 19 Cps19fA (50.6%) Bacillus subtilis LytR (32.9%)

CpsIaB

243

Unknown

Streptococcus pneumoniae serotype 14 Cps14B (62.6%) Streptococcus pneumoniae serotype 19 Cps19fB (64.2%)

CpsIaC

230

Chain length regulator

Streptococcus pneumoniae serotype 14 Cps14C (44.2%) Streptococcus pneumoniae serotype 19 Cps19fC (45.6%) Rhizobium melilotiExoP (22.5%)

CpsIaD

229

Chain length regulator/Export

Streptococcus pneumoniae serotype 14 Cps14D (55.9%) Streptococcus pneumoniae serotype 19 Cps19fD (56.8%) Rhizobium meliloti ExoP(29.6%)

CpsIaE

391

Glucosyltransferase

Streptococcus pneumoniae serotype 14 Cps14E (49.1%) Streptococcus pneumoniae serotype 19 Cps19fE (49.1%)

CpsIaF

149

Unknown

Streptococcus pneumoniae serotype 14 Cps14F (83.9%) Sphingomonas S88 SpsK (33.1%)

CpsIaG

180

b-1,4-galactosyltransferase

Streptococcus pneumoniae serotype 14 Cps14G (53.2%) Sphingomonas S88 SpsK (23.6%)

CpsIaH

379

Putative CP polymerase

Streptococcus pneumoniae serotype 14 Cps14H (23.6%) Salmonella typhimurium Rfc (19.5%) Shigella flexneriRfc (16.1%)

CpsIaI

334

b-1,3-N-acetylglucosaminyl transferase

Streptococcus pneumoniae serotype 14 Cps14I (26.4%) Rhizobium meliloti ExoU (23.7%)

Bacterial Capsular Polysaccharide and Sugar Transferases

99

Table 1 (continued)

ORF

Number of amino acids

Proposed function of gene product

Similar gene products (% identity)

CpsIaJ

315

b-1,4-galactosyltransferase

Streptococcus pneumoniae serotype 14 Cps14 J (37.0%) Rhizobium meliloti ExoO (26.5%)

CpsIaK

318

Sialyltransferase

Haemophilus influenzae type b Orf4 (20.8%) Shigella dysenteriae RfbX (19.8%)

CpsIaL

466

Repeat unit transporter

Streptococcus pneumoniae serotype 14 Cps14L (16.4%)

NeuB

341

Sialic acid synthesis

Escherichia coli NeuB (56.0%)

NeuC

384

Sialic acid synthesis

Escherichia coli NeuC (43.6%)

NeuD

209

Acetyltransferase

Escherichia coli NeuD (33.0%)

NeuA

413

CMP-sialic acid synthetase

Escherichia coli NeuA (33.0%)

orf1

161

Unknown

Escherichia coli EvgS (28.2%)

ung

217

Uracil DNA glycosylase

Streptococcus pneumoniae Ung (75.8%) Bacillus subtilis Ung (52.6%)

parE

619

DNA topoisomerase

Streptococcus pyogenes M1 ParE (100%)

parC

819

DNA topoisomerase

Streptococcus pyogenes M1ParC (77.3%)

bcat

324

Branched-chain amino acid aminotransferase

Streptococcus pyogenes M1Bcat (82.1%)

with Pseudomonas sp. TcbR protein [30]. These proteins are known to belong to LysR-like regulatory proteins. In the N-terminal region of LysR proteins, the helix-turn-helix structure is conserved that confers DNA binding activity. The orientation of transcription of cpsIaR was opposite to the cps gene cluster as observed with the majority of LysR-like regulation systems [30]. As shown in Fig. 5, the cpsR gene was detected only in S. agalactiae strains, but not in pneumococci. The cpsIaA gene product showed a high degree of similarity to Cps14A (50.2% identity) and Cps19fA (50.6% identity) of S. pneumoniae serotype 14 [23] and 19F [20]. Furthermore, CpsIaA exhibited similarity to LytR (32.9% identity) of Bacillus subtilis [31], as was the case with Cps14A. The pneumococcal CpsA proteins are thought to be involved in the regulation of CP expression, and the LytR protein is a transcription attenuator for the autolysin (lytABC) operon. Therefore, the CpsIaA protein may also have a specific role in the transcriptional regulation of cpsIa genes of S. agalactiae. The hydropathy

100

K. Miyake · S. Iijima

profile of CpsIaA demonstrated the presence of three N-terminal hydrophobic segments. This suggested that the putative regulatory protein may bind to the cell membrane, similarly to LytR. However, the mechanism by which these membrane proteins regulate transcription is still unclear. This gene is usually observed in all cps gene clusters. The cpsIaB to cpsIaD gene products also showed high degrees of similarity to the corresponding gene products of S. pneumoniae serotype 14 and 19F (Table 1). Among these previously-reported gene products, those showing homology to CpsIaC and CpsIaD have been suggested to play important roles in chain length determination and export of CPs. The ExoP protein of Rhizobium meliloti is composed of two domains, and CpsC and CpsD proteins of streptococcal bacteria were similar to its N-terminal and C-terminal domains of ExoP, respectively. The N-terminal domain of ExoP is expected to have a role in chain length determination based on its homology to other bacterial proteins, and the C-terminal domain is supposed to have a regulatory function [32]. As CpsIaC and IaD also showed similarity to the respective domains of ExoP protein as suggested with Cps14C and 14D, these proteins may function in chain length determination and export of type Ia CP. Recently, it is reported that CpsD is a tyrosine kinase and CpsIaC is a substrate for CpsD kinase [33]. The cpsIaE gene product showed a close similarity to Cps14E (49.1% identity) of S. pneumoniae serotype 14 [21]. Cps14E of S. pneumoniae serotype 14 appeared to be a glucosyltransferase [21]. Polypeptides that show homology to CpsIaE of S. agalactiae type Ia are known to catalyze linkage of the first sugar to the lipid carrier (Table 1), suggesting that the gene product may transfer glucose to the lipid carrier. The gene products encoded by cpsIaF and cpsIaG were homologous to the proteins Cps14F (83.9% identity) and Cps14G (53.2% identity) of S. pneumoniae serotype 14, respectively (Table 1). Since Cps14G protein is a b-1,4-galactosyltransferase [22], CpsIaG of S. agalactiae type Ia was expected to have the same enzyme activity. CpsIaF of S. agalactiae type Ia contained a hydrophobic region in the center of the molecule, as also reported in Cps14F of S. pneumoniae serotype 14 [22]. This hydrophobic region of Cps14F is thought to be anchored in the membrane, as reported with SpsK of Sphingomonas S88 [22]. S. agalactiae type Ia CpsIaH showed a degree of homology (23.6% identity) to Cps14H of S. pneumoniae serotype 14. Although these two proteins did not show close homology over the entire region, the twelve membrane spanning domains observed in Cps14H were also detected in S. agalactiae type Ia CpsIaH. In S. pneumoniae serotype 14, Cps14H is supposed to be a CP polymerase based on its homology to O-antigen polymerase (Rfc) of Shigella flexneri [34] and Salmonella typhimurium [35]. Since CpsIaH of S. agalactiae type Ia showed homology to the Rfc proteins, CpsIaH might also be a CP polymerase. Indeed, exchange of this gene between type Ia and III were recently reported to cause exchange of CPs [36]. The gene products encoded by cpsIaI and cpsIaJ showed similarity to several putative glycosyltransferases involved in the biosynthesis of CPs, lipopoly-

Bacterial Capsular Polysaccharide and Sugar Transferases

101

saccharides, and exopolysaccharides in numerous bacterial species (Table 1). Especially, CpsIaI and CpsIaJ of S. agalactiae type Ia showed moderate similarity to Cps14I and Cps14 J of S. pneumoniae serotype 14, respectively. In S. pneumoniae serotype 14, Cps14I and Cps14 J have b-1,3-N-acetylglucosaminyltransferase and b-1,4-galactosyltransferase activity, respectively [23]. This suggested that cpsIaI and cpsIaJ encode an N-acetylglucosaminyltransferase and a galactosyltransferase required for addition of the third and the fourth sugar residues in the oligosaccharide side chain of type Ia CP, respectively. DXD, DXS and ED sequences have been found in many glycosyltransferases involved in biosynthesis of CPs, lipopolysaccharides and exopolysaccharides [23, 37]. It was previously reported that a-glycosyltransferases contain two sets of DXD which are essential for the enzyme activity [38, 39]. On the other hand, b-glycosyltransferases contain one set of DXD, DXS and sometimes ED sequences aligned at suitable distances [23, 37].All of these sequences were found in CpsIaI and IaJ, but only DXD and DXS were found in CpsIaE. The lack of ED in CpsIaE was in accordance with the previous observation by Keenleyside and Whitfield that glycosyltransferases to lipid carriers often do not contain this sequence [37]. Neither of these conserved sequences was found in CpsIaG galactosyltransferase of S. agalactiae type Ia. The CpsIaG may belong to a different family of glycosyltransferase from CpsIaE, CpsIaI and CpsIaJ. cpsIaK gene products of both S. agalactiae types Ia showed a certain homology to Lst sialyltransferase of Haemophilus ducreyi (GenBank accession no. AF101047). The cpsIaL gene of S. agalactiae type Ia encoded a protein, which showed similarities to RfbX of Shigella dysenteriae and CapF of Staphylococcus aureus (Table 1). Furthermore, the hydropathy profile of CpsIaL was similar to those of RfbX-related proteins. The RfbX of S. dysenteriae is thought to be involved in one of the later CP synthesis steps, such as transfer of the repeating unit to the cell surface [21]. Therefore, CpsIaL seems to be involved in a later stage of CP synthesis. It is noteworthy that the levels of homology between CpsIaA to CpsIaG of S. agalactiae type Ia and corresponding gene products of S. pneumoniae serotype 14 were high (44–85%). However, their downstream gene products (CpsIaH to CpsIaL) showed less homology (15–40%) to those of S. pneumoniae serotype 14. In addition to these ORFs, nine ORFs were identified downstream of cpsIaL (Fig. 5).Among these, the first four ORFs were found to be related to sialic acid synthesis. These ORFs were designated as neuB, C, D and A based on their similarity to gene products required for polysialic acid synthesis in E. coli K1. In E. coli K1 strain, neuB and neuC gene products are involved in sialic acid synthesis (40–42), and it is likely that the corresponding gene products of S. agalactiae type Ia have the same function. NeuD of S. agalactiae type Ia showed similarity (33.0% identity) to NeuD of E. coli K1. The NeuD protein of E. coli shows homology to several bacterial acetyltransferases, but its target is unknown. The neuA gene product of S. agalactiae type Ia showed homology (33.0%) to NeuA of E. coli K1 [40]. These well-studied proteins have been con-

102

K. Miyake · S. Iijima

firmed to be CMP-sialic acid synthetases. These genes are detected in cps gene clusters of GBS strains that produce CPs containing sialic acid. Additional ORFs were identified downstream of neuA and designated as orf1, ung, parE, parA and bcat. ORF1 showed similarity (28.2% identity) to the N-terminal cytoplasmic domain of EvgS which seems to be a sensor protein of a two component regulatory system responding to environmental stimuli [43]. However, the function and involvement of the ORF1 protein in CP synthesis are not yet clear.An ORF showing high similarity to ung gene products of S. pneumoniae [44] and Bacillus subtilis [45] was found downstream of the orf1 gene. The Ung proteins function in DNA mismatch repair. The role of the gene in the cpsIa cluster is also still obscure. Furthermore, ParE and ParC showed close similarity to DNA topoisomerase of S. pyogenes. The last ORF was identical to the branchedchain amino acid aminotransferase (bcat) gene of S. pyogenes [46]. As shown in Fig. 5, upstream and downstream regions of cps and neu operons of the type Ia strain showed close similarity to a portion of S. pyogenes genome. Instead of cps and neu operons, S. pyogenes genome contains pyr gene cluster between cpsY (cpsR in S. agalactiae) and par genes [46]. This suggests that the cps operon was evolutionally inserted into the genome of S. pyogenes. In contrast to the cps gene cluster of GBS, gene clusters of S. pneumoniae strains are always surrounded by dexB gene and aliA gene. This fact probably means that cps gene clusters of GBS and S. pneumoniae exist at different positions of their chromosomes. We also examined the structure of the cps gene cluster of S. agalactiae type Ib and compared with cpsIa gene products. The DNA sequence of 9,987 nucleotides was determined and nine complete ORFs, designated cpsIbE to cpsIbL, and neuB were identified (Fig. 5). All ORFs were in the same orientation, and were spaced one behind the other at short distances.A possible Shine-Dalgarno sequence was identified just upstream of the potential initiation codon of each ORF. All ORFs except for cpsIbK were preceded by ATG codons. Only cpsIbK was preceded by GTG. These gene products exhibited high homology to cps genes of several other streptococci. Notably, the cpsIb gene products showed very high sequence identity with the corresponding cpsIa gene products (92%<) except for CpsH, CpsJ and CpsK (Fig. 6). As is the case with cpsIa genes, the cpsIbE, cpsIbG and cpsIbI gene products seem to have glucosyltransferase, b-1,4-galactosyltransferase, and b-1,3-N-acetylglucosaminyltransferase activity, respectively. On the other hand, CpsIbJ showed very low sequence identity with CpsIaJ (15.7%). CpsIbK, which is assumed to be a sialyltransferase also showed relatively low homology to CpsIaK (50.5%) (Fig. 6). Despite the low homology between type Ia and Ib, CpsJ and K of type Ia and III had almost identical amino acid sequences.Among CPs of these three strains, only type Ib CP has a b-1,3-linked galactose as the fourth saccharide (Fig. 1). These differences of homology between S. agalactiae strains for the respective genes may relate to the CP structure of each strain. CpsH was unique in the type III strain although type Ia and Ib CpsH shared high homology. The unique structure of CpsIIIH seems to reflect a difference in sugar subunit polymerization, since CpsH was reported to be a CP polymerase.

Bacterial Capsular Polysaccharide and Sugar Transferases

103

Fig. 6 Similarity of cps gene products between S. agalactiae type Ia, Ib, III, and S. pneumoniae type 14. Percentages shown in squares are homology to CpsIb proteins, and those shown in circles are homology between cps products of type Ia and III strains. GenBank accession numbers of these cps gene clusters are AB028896, AB050723, AF163833, and X85787, respectively

4 Function of Streptococcal Glycosyltransferases Repeating units of bacterial CPs are known to be synthesized on lipid carriers on the cell surface [13, 14]. Since CpsIaE, CpsIaG, CpsIaI and CpsIaJ showed homology to several glycosyltransferases of S. pneumoniae serotype14, the functions of these molecules were examined by analysis of the intermediates in synthesis of the oligosaccharide subunit formed by membrane fractions of E. coli harboring expression plasmids of these cps genes. Membrane fractions were used as sources of enzymes and acceptors, and 14C-labeled oligosaccharide intermediates added to lipid carriers were extracted in lipid fractions. After releasing lipid carriers by treatment with trifluoroacetic acid, these intermediates were analyzed by thin layer chromatography (TLC) [22]. In this chapter, we focus on enzymatic activity of the cps gene products of S. agalactiae type Ia and Ib.

104

K. Miyake · S. Iijima

Fig. 7 Thin-layer chromatogram of 14C-labeled sugar intermediates of CP synthesis using isolated membranes of various E. coli strains expressing type Ia cps genes (pBAPE, pBAPF, pBAPG, pBAPI and pBAPJ). TLC plates were developed twice in the case of the clone pBAPJ. Added UDP-monosaccharides are shown below the chromatograms

Bacterial Capsular Polysaccharide and Sugar Transferases

105

Membranes of the E. coli clone carrying the plasmid containing cpsIaE (pBAPE) showed incorporation of [14C] glucose and [14C] galactose. However, [14C]-glucose was the only labeled sugar detected by TLC analysis (Fig. 7, lanes 1 and 2). Furthermore, the incorporation of [14C] glucose or [14C] galactose into the lipid carrier was almost completely inhibited by excess cold glucose, but galactose showed only partial inhibition (Fig. 7, lanes 3 and 4). These results suggested that CpsIaE had glucosyltransferase activity, but not galactosyltransferase activity. Membranes of the E. coli clone carrying the plasmid containing cpsIaE-cpsIaG (pBAPG) showed incorporation of radioactivity on incubation with UDP-[14C] glucose and/or UDP-[14C] galactose. TLC analyses indicated synthesis of lactose and glucose intermediates (Fig. 7, lanes 6, 7). Only lactose was detected when UDP- [14C] galactose and cold glucose were added (Fig. 7, lane 8). Furthermore, cold galactose inhibited lactose intermediate formation (Fig. 7, lane 9). On the other hand, lactose was not detected with plasmid containing cpsIaE-cpsIaF (pBAPF) (Fig. 7, lane 5). Taken together, these results showed that cpsIaG encoded a galactosyltransferase that catalyzed the transfer of galactose as the second monosaccharide. The E. coli clone carrying plasmid containing cpsIaE-cpsIaI (pBAPI) showed incorporation of the radioactivity into the lipid carriers by reaction using 14C-labeled UDP-Nacetylglucosamine, UDP-glucose and UDP-galactose. Following incubation with UDP-[14C] N-acetylglucosamine, cold UDP-glucose and UDP-galactose, GlcpNAc-Lac trisaccharide was detected by TLC (Fig. 7, lane 11). Excess cold Nacetylglucosamine inhibited the trisaccharide intermediate formation (Fig. 7, lane 12). These results indicated that CpsIaI had N-acetylglucosaminyltransferase activity. The membrane fraction of the E. coli clone carrying plasmid containing cpsIaE-cpsIaJ (pBAPJ) showed an activity, which produced the tetrasaccharide Galp-GlcpNAc-Lac (LNnT) (Fig. 7, lane 13). Since the clone carrying pBAPI did not show the tetrasaccharide-forming activity and galactose inhibited the tetrasaccharide formation (Fig. 7, lane 14), cpsIaJ seemed to encode a galactosyltransferase, which catalyzed the transfer of galactose as the fourth monosaccharide. As judged from these glycosyltransferase assays and the results of homology analysis, it is likely that CpsIaE, CpsIaG, CpsIaI and CpsIaJ are glucosyltransferase, b-1,4-galactosyltransferase, b-1,3-N-acetylglucosaminyltransferase, and b-1,4-galactosyltransferase, respectively. CpsIaJ of S. agalactiae type Ia is a b-1,4-galactosyltransferase, which transfers galactose and produces LNnT (b-Gal-1, 4-b-GlcNAc-1, 3-b-Gal-1, 4-b-Glc). Although CpsIbJ did not show similarity to the CpsIaJ protein, it was expected to have galactosyltransferase activity from the arrangement of the cpsIb gene cluster. To examine the enzymatic activity of CpsIbJ, an expression plasmid containing cpsIbJ was constructed (pBBPJ) and introduced into E. coli. The membrane fraction of E. coli harboring the expression plasmid was used as a source of enzyme and FCHASE-labeled LNT2 (bGlcNAc-1, 3-bGal-1, 4-bGlc) as an artificial acceptor.As shown in Fig. 8a, the membrane fraction of E. coli harboring pBBPJ, which contained cpsIbJ downstream of the lac promoter, showed

106

K. Miyake · S. Iijima

Bacterial Capsular Polysaccharide and Sugar Transferases

107

clear galactosyltransferase activity. As a positive control, E. coli harboring pBAPJ that contained cpsIaJ was used. This result suggested that the CpsIbJ protein had galactosyltransferase activity, as observed with CpsIaJ. To investigate the type of linkage formed by the enzyme reaction, products were digested with b1,4- and b1,3-specific galactosidases. The reaction product of CpsIbJ could not be digested by b1,4-specific galactosidase, while that of E. coli containing pBAPJ was cut by this enzyme (Fig. 8a). On the other hand, b1,3specific galactosidase released galactose from the reaction product of pBBPJ. Together, these results suggested that CpsIbJ had b1,3-galactosyltransferase activity. The natural acceptor of CpsJ is LNT2 linked to a lipid carrier. To study whether CpsIbJ transfers galactose to the oligosaccharide acceptor without a lipid carrier, LNT2 was reacted with UDP-galactose and membrane fractions of E. coli harboring pBBPJ or pBAPJ, and reaction products were analyzed by high-performance anion-exchange chromatography. As shown in Fig. 8b, membrane fractions from E. coli harboring pBBPJ and pBAPJ could produce LNT (b-Gal-1, 3-b-GlcNAc-1, 3-b-Gal-1, 4-b-Glc) and LNnT, respectively. This result indicated again that the cpsIbJ gene codes for b1,3-galactosyltransferase, and both CpsIaJ and CpsIbJ, could transfer galactose to the sugar acceptor without a lipid carrier. As mentioned above, cpsK gene products of both S. agalactiae type Ia and Ib show some homology to Lst sialyltransferase of Haemophilus ducreyi (23.4% and 26.7%, respectively). To detect sialyltransferase activity of cpsK gene products, crude extracts of E. coli harboring cpsK expression plasmids (pETIaK and pETIbK) were used as source of enzyme. As shown in Fig. 9a, only CpsIaK protein showed sialyltransferase activity when FCHASE-labeled LNnT was used as substrate, and the crude extracts from E. coli harboring pETIbK and pETBlue2 did not show any enzymatic activity. Since FCHASE-labeled LNT was not available as an acceptor substrate for CpsIbK, enzyme reactions were performed using non-labeled oligosaccharide acceptors and analyzed, as is

Fig. 8 Detection of b1,3-galactosyltransferase activity of cpsIbJ gene product. a Thin-layer chromatogram of FCHASE-labeled oligosaccharide produced by CpsIaJ or CpsIbJ enzyme and the degradation products produced by galactosidases. Lane 1, FCHASE-LNT2. The reaction product of the membrane fraction of E. coli harboring pBAPJ (lane 2), that digested with b1,3-galactosidase (lane 3) and that digested with b1,4-galactosidase (lane 4). The reaction product of the membrane fraction of E. coli harboring pBBPJ (lane 5), that digested with b1,3-galactosidase (lane 6) and that digested with b1,4-galactosidase (lane 7). The reaction product of the membrane fraction of E. coli harboring pBluescript II SK+ alone (lane 8); FCHASE-LNnT (lane 9). b HPAEC-PAD chromatograms of products of the galactosyltransferase reaction with the membrane fraction of E. coli harboring pBAPJ or pBBPJ. The membrane fraction of E. coli harboring pBluescript II SK+ alone was used as a negative control. Arrows 1, 2, and 3 indicate the elution positions of sugar standards: 1, LNT2; 2, LNnT; 3, LNT

108

K. Miyake · S. Iijima

Fig. 9 Detection of sialyltransferase activity of cpsK gene products. a Thin-layer chromatogram of FCHASE-labeled oligosaccharide that reacted with the crude extracts of E. coli harboring pETIaK (lane 1), pETIbK (lane 2), and pETBlue2 (lane 3). FCHASE-labeled LNT2 (lane 4), LNnT (lane 5), and sialyl-LNnT (lane 6) were used as standards.Arrows indicate the nonspecific bands. The origin is also indicated. b HPAEC-PAD chromatograms of sialyltransferase reaction with crude extracts of E. coli harboring pETIaK, pETIbK, and pETBlue2. The elution positions of 3¢sialyl-LNnT and 3¢sialyl-LNT are indicated by arrows.Arrowheads indicate background peaks

the case with the galactosyltransferase assays.As shown in Fig. 9b, crude extract of E. coli harboring pETIbK showed sialyltransferase activity only for LNT. On the other hand, crude extract from E. coli harboring pETIaK converted only LNnT to 3¢sialyl-LNnT. These results suggested that cpsK genes of both strains code for sialyltransferases which transfer sialic acid to a galactose residue of LNnT or LNT as the fifth saccharide, and that they have a stringent accepter specificity.

Bacterial Capsular Polysaccharide and Sugar Transferases

109

In these studies, we could detect respective glycosyltransferase activity. Among the cps genes analyzed, cpsI, cpsJ and cpsK are more useful and show more potential for glycoengineering than cpsE and cpsG, since the former genes can use oligosaccharides without lipid carrier as substrates and the latter require that for glycosyltransferase reactions. This tendency is also reported for cps genes of S. pneumoniae.

5 Conclusions In our studies, we have shown the usefulness of the CPs and sugar transferases of streptococci. The CPs of S. agalactiae type Ia and Ib are potential materials for development of new therapeutic reagents because of their similarity to sialyl Lewis carbohydrates. Since these CPs are naturally multivalent, higher effects are expected than for monomeric sialyl Lewis carbohydrates. However, these polysaccharides lack fucose residue, and the fucose residue is known to be essential for biological activity of sialyl Lewis carbohydrates. We are now trying to use bacterial fucosyltransferases for addition of fucose to the streptococcal CPs to produce multivalent sialyl Lewis carbohydrates. The modified polysaccharides might be good therapeutic reagents for diseases mediated by cell-cell interaction such as cancer metastasis and inflammation. The streptococcal cps gene products also show potential, because the enzymes, especially the glycosyltransferases, seem to be useful for glycoengineering. The cps genes can be expressed and the gene products can show the enzymatic activity in E. coli. Therefore, cps genes of S. agalactiae type Ia and Ib could be useful for production of oligosaccharide containing sialyllactosamine structure in E. coli. CP structures of streptococci are very diverse. Several streptococcal strains were reported to react to anti-sialyl Lewis carbohydrate antibody [47]. Therefore, cps genes from these strains will also be useful for the production of various oligosaccharides in E. coli. Furthermore, we expect that S. agalactiae strains themselves are also useful as a host for the production of diverse polysaccharides, since host-vector systems are available and mutagenesis by transposon is possible in several streptococci. Therefore, various modified CPs will be produced in genetically engineered streptococci so that they can be used as therapeutic reagents in the near future. Acknowledgements We are grateful to Dr. Michio Ohta, Department of Bacteriology, Nagoya University, School of Medicine, for providing us with S. agalactiae type Ia OI1 strain and type Ib OI2 strain. We also thank Dr. Satoshi Koizumi, Dr. Tetsuo Endo, and Dr. Akio Ozaki, for high-performance anion-exchange chromatography and providing FCHASE-labeled substrates.

110

K. Miyake · S. Iijima

References 1. Moxon ER, Kroll JS (1990) Curr Top Microbiol 150:65 2. Cross AS (1990) Curr Top Microbiol 150:87 3. Boyer KM, Gadzala CA, Burd LI, Fisher DE, Paton JB, Gotoff SP (1983) J Infect Dis 148:795 4. DiFabio JL, Michon F, Brisson J-R, Jennings HJ (1989) Can J Chem 67:877 5. Jennings HJ (1990) Curr Top Microbiol 150:97 6. Jennings HJ, Katzenellenbogen E, Lugowski C, Kasper DL (1983) Biochemistry 22:1258 7. Jennings HJ, Lugowski C, Kasper DL (1981) Biochemistry 20:4511 8. Kogan G, Brisson J-R, Kasper DL, von Hunolstein C, Orefici G, Jennings HJ (1995) Carbohydr Res 277:1 9. von Hunolstein C, Ascenzi SD, Wagner B, Jelinkova J, Alfarone G, Recchia S, Waner M, Orefici G (1993) Infect Immun 61:1272 10. Wessels MR, DiFabio JL, Benedi V-J, Kasper DL, Michon F, Brisson J-R, Jelikova J, Jennings HJ (1991) J Biol Chem 266:6714 11. Wessels MR, Pozsgay V, Kasper DL, Jennings HJ (1987) J Biol Chem 262:8262 12. Arakawa Y, Wacharotayankun R, Nagatsuka T, Ito H, Kato N, Ohta M (1995) J Bacteriol 177:1788 13. Boulnois GJ, Jann K (1989) Mol Microbiol 3:1819 14. Boulnois GJ, Roberts IS (1990) Curr Top Microbiol 150:1 15. Glucksmann MA, Reuber TL, Walker GC (1993) J Bacteriol 175:7033 16. Glucksmann MA, Reuber TL, Walker GC (1993) J Bacteriol 175:7045 17. Takada A, Ohmori K, Yoneda T, Tsuyuoka K, Hasegawa A, Kiso M, Kannagi R (1993) Cancer Res 53:354 18. Miyake K, Yamamoto S, Iijima S (1996) Cytotechnology 22:205 19. Arrecubieta C, García E, López R (1995) Gene 167:1 20. Guidolin A, Morona JK, Morona R, Hansman D, Paton JC (1994) Infect Immun 62:5384 21. Kolkman MAB, Morrison DA, van der Zeijst BAM, Nuijten PJM (1996) J Bacteriol 178:3736 22. Kolkman MAB, van der Zeijst BAM, Nuijten PJM (1997) J Biol Chem 272:19502 23. Kolkman MAB, Wakarchuk W, Nuijten PJM, van der Zeijst BAM (1997) Mol Microbiol 26:197 24. Rubens CE, Heggen LM, Haft RF, Wessels MR (1993) Mol Microbiol 8:843 25. Chaffin DO, Beres SB, Yim HH, Rubens CE (2000) J Bacteriol 182:4466 26. Yamamoto S, Miyake K, Koike Y, Watanabe M, Machida Y, Ohta M, Iijima S (1999) J Bacteriol 181:5176 27. Watanabe M, Miyake K, Yanae K, Kataoka Y, Koizumi S, Endo T, Ozaki A, Iijima S (2002) J Biochem 131:183 28. Watanabe M, Miyake K, Yamamoto S, Kataoka Y, Koizumi S, Endo T, Ozaki A, Iijima S (2002) J Biosci Bioeng 93:610 29. Rotta J (1986) Pyrogenic hemolytic Streptococci. In: Sneath PHA, Mair NS, Sharpe ME, Holt JG (eds) Bergey’s manual of systematic bacteriology, vol 2. Williams and Wilkins, Baltimore 30. Schell MA (1993) Annu Rev Microbiol 47:597 31. Lazarevic V, Margot P, Soldo B, Karamata D (1992) J Gen Microbiol 138:1949 32. Becker A, Niehaus K, Pühler A (1995) Mol Microbiol 16:191 33. Morona JK, Morona R, Miller DC, Paton JC (2003) J Bacteriol 176:733 34. Morona R, Mavris M, Fallarino A, Manning PA (1994) J Bacteriol 176:733 35. Collins LV, Hackett J (1991) J Bacteriol 173:2521

Bacterial Capsular Polysaccharide and Sugar Transferases 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

46.

47.

111

Chaffin DO, Beres SB, Yim HH, Rubens CE (2000) J Bacteriol 182:4466 Keenleyside WJ, Whitfield C (1996) J Biol Chem 271:28581 Saxena IM, Brown JR M, Fevre M, Geremia RA, Henrissat (1995) J Bacteriol 177:1419 Shibayama K, Ohsuka S, Tanaka T, Arakawa Y, Ohta M (1998) J Bacteriol 180:5313 Ganguli S, Zapata G,Wallis T, Reid C, Boulnois G,Vann WF, Roberts IS (1994) J Bacteriol 176:4583 Zapata G, Crowley JM, Vann WF (1992) J Bacteriol 174:315 Zapata G, Vann WF, Aaronson W, Lewis MS, Moos M (1989) J Biol Chem 264:14769 Utsumi R, Katayama S, Taniguchi M, Horie T, Ikeda M, Igaki S, Nakagawa H, Miwa A, Tanabe H, Noda M (1994) Gene 140:73 Méjean V, Rives I, Claverys J-P (1990) Nucleic Acids Res 18:6693 Glaser P, Kunst F,Arnaud M, Coudart MP, Gonzales W, Hullo MF, Ionescu M, Lubochinsky B, Marcelino L, Moszer I, Presencan E, Santana M, Schneider E, Schweizer J, Vertes A, Rapoport G, Danchin A (1993) Mol Microbiol 10:371 Ferretti JJ, McShan WM,Ajdic D, Savic DJ, Savic G, Lyon K, Primeaux C, Sezate S, Suvorov N, Kenton S, Lai HS, Lin SP, Qian Y, Jia HG, Najar FZ, Ren Q, Zhu H, Song L, White J, Yuan X, Clifton SW, Roe BA, McLaughlin R (2001) Microbiol 98:4658 Hirota K, Kanitani H, Nemoto K, Ono T, Miyake Y (1995) FEMS Immun Med Mic 12:159

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 113– 133 DOI 10.1007/b94194 © Springer-Verlag Berlin Heidelberg 2004

Bacterial Sterilization and Intracellular Protein Release by a Pulsed Electric Field Takayuki Ohshima (✉) · Masayuki Sato Department of Biological and Chemical Engineering, Faculty of Engineering, Gunma University, 1-5-1 Tenjin-cho, Kiryu, Gunma 376-8515, Japan [email protected]; [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

2 2.1 2.2 2.3

PEF Sterilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of Temperature on PEF Sterilization . . . . . . . . . . . . . . Synergistic Effect of PEF and Bactericide Treatment on Sterilization Improvement of Chamber Structure for PEF Sterilization . . . . . .

. . . .

. . . .

. . . .

116 116 118 120

3 3.1 3.2 3.3 3.4

Recovery of Intracellular Proteins Using PEF . . . . . . . . . . . . . . . Release of the Protein under PEF . . . . . . . . . . . . . . . . . . . . . . Recovery of Recombinant Protein from E. coli by PEF Treatment . . . . . Selective Release of Recombinant Proteins from E. coli by PEF Treatment Secretion of Recombinant Protein from E. coli by PEF Treatment . . . . .

. . . . .

. . . . .

123 123 125 127 129

4

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

References

. . . .

. . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

Abstract Several biotechnological applications of high-voltage pulsed electric field (PEF) are introduced. Electrical breakdown or disruption of a biological membrane by PEF is understood to occur by electromechanical compression, which results in the formation of transmembrane pores. If the total area of induced pores is small in relation to the total surface area of the membrane, the pores are able to close again mainly through the diffusion of the lipid molecules and rearrangement of the proteins (reversible disruption). If the total area of the pores becomes unfavorably large, the membrane is no longer able to repair these perturbations (irreversible disruption), and that results in sterilization.We have investigated effective sterilization by using PEF-induced irreversible disruption of biological membranes. The treatment temperature or growth temperature was found to have a great effect on PEF sterilization. The shape of the treatment chamber also proved important for effective PEF sterilization. Therefore, a number of reactors having novel structures were developed. We have also verified that this PEF-induced reversible disruption could be utilized for the selective release of intracellular proteins from yeast and certain gene-engineered Escherichia coli. The secretion of periplasmic protein from E. coli was achieved during cultivation. Keywords Pulsed electric field (PEF) · Sterilization · Intracellular protein

114

T. Ohshima · M. Sato

1 Introduction The electrical breakdown or disruption of biological membranes in a pulsed electric field (PEF) is a well known phenomenon, which can be explained relatively easily by electromechanical compression [1]. Naturally, the charges on the capacitor plates of the membrane attract each other. This causes a thinning of the membrane provided that the membrane is compressible, and indeed there is experimental evidence to suggest that it is. The membrane thickness attained at a given membrane potential (in other words, at a given charge density on the plates of the membrane capacitor) is determined by the equilibrium between the electric compressive forces and the resulting elastic restoring forces. With increasing membrane potential, a critical membrane thickness is reached at which the electric compressive forces change more rapidly than the generated electric restoring forces. The membrane becomes unstable and breaks through. The emerging pores fill up with internal and external solution, both of which are usually highly conducting. The resulting increase in the electrical permeability of the membrane, which can be as high as eight orders of magnitude in the case of pure lipid bilayer membranes, leads to a very rapid discharge of the membrane capacitor. The mechanism for this is shown in Fig. 1. An increase in the intensity of the external field will lead first to membrane breakdown at the poles of the cell. The required field strengths are in the range of 1 to 20 kV/cm (depending on cell radius). The breakdown voltage itself is of the order of 1 V (depending on temperature, field duration, and so on) [2, 3].At higher (supercritical) field strengths, the breakdown voltage is reached for other membrane sites. This phenomenon resulted in the formation of transmembrane pores. The size or number of these pores can be adjusted by varying the electric field conditions. If the total area of induced pores is small in relation to the total surface area of the membrane, the pores are able to close again mainly through diffusion of lipid molecules, and rearrangement of the proteins (reversible disruption). On the other hand, if more and more pores are formed or if the diameter of individual pores increases as a result of secondary processes at higher field strengths, a limit value may be exceeded where the ratio of total pore area to total membrane area becomes so unfavorable that the membrane is no longer able to repair these perturbations (irreversible disruption). Based on these phenomena, biotechnological applications of PEF have been investigated. When disruption of the biological membrane is reversible, electroporation [4, 5] or electrofusion [6, 7] is possible, while irreversible disruption comes into play in the non-thermal PEF sterilization technique [8–10]. We have studied the biotechnological applications of the PEF technique, and significantly improved PEF sterilization. In addition, we have proposed a novel method for the recovery of intracellular proteins using PEF. In what follows, we shall dwell on advanced PEF sterilization and the technique for the recovery of intracellular proteins using PEF, focusing on our research results.

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

Fig. 1 Schematic diagram of cell membrane destruction in PEF

115

116

T. Ohshima · M. Sato

2 PEF Sterilization 2.1 Effect of Temperature on PEF Sterilization PEF sterilization was found to depend on the treatment temperature. The lethal effect is related to the temperature of the medium in which the bacteria are suspended. Jayaram et al. reported that the death rate of Lactobacillus brevis is significantly influenced by the liquid medium temperature and that the use of an electric field (25 kV/cm) at 60 °C results in very high destruction levels (survival ratio is about 10–9) of L. brevis cells [11]. Ohshima et al. [12] studied the treatment temperature dependency of PEF sterilization using three microorganisms, Escherichia coli, Saccharomyces cerevisiae, and Salmonella typhimurium. PEF energy (U) was calculated using the following equation based on a charge-discharge capacitor: Np 1 fp T 1 U = 3 CV 2 7 = 3 CV 2 5 n 2 n 2

(1)

where, C [F] is the capacitance of the capacitor, V [V] is the charging voltage, fp [Hz] is the frequency of the pulse (50 Hz in the present study), T [s] is the treatment time, Np [–] is the number of pulses, and v [mL] is the sample volume. Figure 2 shows the survival ratio of S. cerevisiae without (A) and with (B) PEF sterilization (16 kV/cm of PEF) at various temperatures. The sterilization efficiency increased with temperature. The minimum survival ratio was about 10–6 when 300 J/mL of energy was applied at 50 °C, while the survival ratio at 10 °C was about 10–1 (Fig. 2b). Under these experimental conditions, heat sterilization of S. cerevisiae was not observed below 45 °C without PEF treatment (Fig. 2a). Therefore, the temperature effect on sterilization below 45 °C could signify the temperature dependence of the PEF treatment. Sterilization above 50 °C, however, occurs via the synergistic effect of both PEF and heat. Ohshima et al. [12] also studied the effect of temperature on PEF sterilization of E. coli and S. typhimurium at a PEF strength of 32 kV/cm (twice that used for S. cerevisiae because PEF sterilization depends on cell size). The temperature trends observed were similar to those observed in the case of S. cerevisiae (data not shown). Therefore, the temperature dependence of PEF sterilization appears to be independent of the type of microorganism. Figure 3 shows an Arrhenius’ plot for the three microorganisms S. cerevisiae, E. coli, and S. typhimurium, where k is the death rate, regarded as a reaction rate, and T is the absolute temperature. The death rate is obtained as the slope of the best fit to survivability curves. The three microorganisms exhibited different sensitivities to temperature. In addition, PEF sterilization for all microorganisms was sensitive to temperature at relatively high temperatures, while this was not observed under lower temperature conditions. The temperature of the medium in which cells are suspended has a significant influence in

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

a

117

b

Fig. 2 Survival ratio of S. cerevisiae without a and with b PEF treatment at various treatment temperatures

Fig. 3 Arrhenius’ plots for three bacteria S. cerevisiae, E. coli, and S. typhimurium, where k is the death rate, regarded as a reaction rate, and T is the absolute temperature

determining the membrane fluidity properties. In general, a lipid bilayer has many phases and the phospholipids are closely packed in a rigid “gel structure” at low temperatures, while at high temperatures, they are less ordered and the membrane has a “liquid-crystalline” structure [13]. The phase transition temperature is about 10 °C lower than the culture temperature because bacteria in-

118

T. Ohshima · M. Sato

corporate increasing proportions of saturated and long-chain fatty acids into phospholipids as the growth temperature is increased [14]. An inflection point for each microorganism (at about 30 °C for E. coli and S. typhimurium, and 25 °C for S. cerevisiae), as shown in Fig. 3, seems to indicate that the temperature of the phase transition of each cell membrane consisted of a lipid bilayer given that the culture temperature in this study was 37 °C (E. coli and S. typhimurium) and 30 °C (S. cerevisiae). Therefore, PEF treatment causes a larger reduction in viable cells when the biological membrane has a liquid-crystalline structure, while PEF sterilization is inefficient below the phase transition temperature. In addition to the temperature of PEF treatment, the effect of growth temperature of bacteria on PEF sterilization has been studied. Ohshima et al. demonstrated that the efficiency of PEF sterilization depends on the growth temperature of the target bacteria by using E. coli [15]. E. coli cells were cultivated at 20 °C, 30 °C, 37 °C, or 42 °C, and subjected to PEF sterilization. The sterilization characteristics varied significantly with the growth temperature of E. coli. It was found that E. coli cells cultivated at around the optimal temperature (30, 37 °C) were relatively resistant to PEF, while those cultivated at 20 °C or 42 °C were easily sterilized by the same PEF treatment. Cell membranes of E. coli cells cultivated at 20 °C have a relatively high content of unsaturated fatty acid contained in phospholipid molecules, which renders the cell membranes fragile [14, 16], and so seemingly more sensitive to the PEF sterilization. E. coli cells cultivated at 20 °C were easily sterilized by the PEF treatment, especially at higher temperatures. The higher efficiency of PEF sterilization can presumably be attributed to these structural changes that occur in the cell membranes. However, E coli cells cultivated at 42 °C were also sterilized easily. The membranes of these particular cells have a high concentration of saturated fatty acids contained in phospholipids, which renders them resistant to PEF sterilization. Therefore, in the case of E. coli cells cultivated at 42 °C, factors other than structural changes in the cell membrane must be responsible for the ease of PEF sterilization. This temperature is close to that at which E. coli cells simply die due to heat; therefore, 42 °C induces heat shock proteins to become expressed [17]. It was presumably the formation of these heat shock proteins that affected the efficiency of PEF sterilization when the E. coli cells were cultivated at 42 °C. Although PEF sterilization can be performed at around room temperature, sterilization efficiency would be very different in this temperature range. If PEF sterilization is to be employed in the food industry, the processing temperature would have to be regulated in order to achieve effective sterilization. 2.2 Synergistic Effect of PEF and Bactericide Treatment on Sterilization Ozone and H2O2 are among the most commonly used bactericides in bacterial sterilization. The synergistic effect of these bactericides and PEF has been studied [12]. Figure 4 shows the survival ratio of E. coli suspended in ozonized

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

119

Fig. 4 Survival ratios of E. coli cells suspended in ozonized water, and combinations with PEF treatment

water, prepared by dissolving ozone in distilled water under a PEF (8 kV/cm). In general, ozone molecules rapidly react with bacteria, reducing the initial survival ratio obtained when bacteria are suspended in ozonized water. The survival ratio after 200 J/mL of PEF treatment with 2 ppm of ozonized water was about 10–4, while the survival ratio after 200 J/mL of PEF treatment alone was about 10–1. In addition, the death rate with PEF and ozone treatment was higher than that with PEF sterilization alone. In the latter experiment, ozone could not be detected during the PEF treatment, so the higher death rate was not caused by an ozone reaction. Evidently, bacteria damaged by an ozone reaction could be more easily destroyed by PEF, leading to the higher death rate and lower survival ratio. The same effects were observed in the case of simultaneous treatment with PEF and H2O2. At an H2O2 concentration of 1.125 mg/mL or less, no sterilization of E. coli could be achieved without the use of PEF. However, a significant synergistic effect was observed between PEF and H2O2. These results revealed that cells whose membranes have been partially damaged, which, in itself, does not lead to cell death, are nevertheless rendered more sensitive to PEF as a result of the membrane damage, and can be easily destroyed by PEF treatment. Recently, some combinational sterilization techniques have been reported with PEF. Liang et al. [18] and Terebiznik et al. [19] reported that combinational sterilization involving PEF and a bacteriocin nisin, a 3.5 kDa peptide produced by Lactococcus lactis, was effective against Salmonella sp., and E. coli. Khadre et al. [20] have tried the combinations of ozone, high pressure, and PEF for efficient sterilization. Picart et al. [21] demonstrated the variation in PEF steril-

120

T. Ohshima · M. Sato

izing efficiency with treatment media. PEF sterilizing promises to become a key technique for non-thermal sterilization. Other combinations of substances and techniques are likely to emerge in the near future. In addition, it will be necessary to tailor the combination according to the target fluid food. 2.3 Improvement of Chamber Structure for PEF Sterilization The fundamentals of PEF sterilization discussed above have been obtained using plate-plate electrode systems, in which a uniform and homogeneous PEF is generated between plate-plate electrodes. Researchers have proposed other, more efficient electrode systems for PEF sterilization [22]. Ohshima et al. [12] has reported that a partially concentrated PEF yields efficient PEF sterilization. Figure 5 shows the three types of electrode systems used for PEF sterilization. At any given voltage and energy, the highest efficiency of PEF sterilization was achieved using the needle-plate electrode. The non-uniform electric field generated between the needle and plate electrodes and the higher PEF around the needle electrode tip enhanced cell destruction efficiency. In addition, the survival ratio for the plate-plate electrode with edges was lower than that for the plate-plate electrode without edges. The PEF was also concentrated around the edges, which increased the sterilization efficiency. The study of the three types of electrodes indicated that concentration of the PEF leads to a higher lethal effect. An electric field-concentrating chamber, which consisted of plate-plate electrodes and an insulating (Plexiglas) plate with holes, as illustrated in Fig. 6, was constructed to study the survival ratio of S. cerevisiae. Insulating plate type (a) had holes with diameters the size of the electrode, which would result in a uniform electric field. Insulating plate types (b), (c), and (d) had almost the same total area of holes, each type with a different diameter. The electric field-concentrating chamber ((b), (c), and (d)) exhibited a higher efficiency than type (a). Type (d), with the smallest diameter

a

b

c

Fig. 5 Schematic diagrams of various treatment chambers: a plate-plate electrodes without edges, b plate-plate electrodes with edges, and c needle-plate electrodes

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

a

b

121

c

d Fig. 6 Schematic diagram of treatment chamber with four types of insulating plates

holes, showed the strongest lethal effect. These results indicated that concentration of the PEF enhances sterilization efficiency under relatively low PEF strengths. Smaller holes were expected to help raise the energy efficiency for killing bacteria. Using a plate-plate electrode system, a plastic (polyethylene) film having many small holes was pasted on the plate electrode surface. The hole diameter was 0.5 mm; the hole density was 2.5¥105 holes/m2, and the total exposed area of the electrode was 35 mm2. The total area of the non-insulated plate electrode was 705 mm2. The inactivation rate was greatly improved by the plastic film insulation on the electrode surface. The efficiency of PEF sterilization may have also changed due to the capacitance of the capacitor in the pulse source because the impedance of the treatment chamber was determined by the exposure area of the electrode surface (area of contact with liquid), and was also affected by the output impedance of the pulse source. However, the thin plastic film was easily damaged during PEF treatment, making extended use difficult.A porous ceramic-coated electrode, rather than the plastic film, would presumably be better suited for food industry applications. To be able to operate continuously for an extended period of time without electrode damage, the new electrode system shown in Fig. 7 was developed [23, 24]. Figure 7a and b are plate-plate and needle-plate electrode systems, respectively. Figure 7c shows stainless steel rings in a cylinder-type electrode system set at the center of a concentric stainless steel cylinder. Instead of rings, coiled stainless steel wire is used for the center electrode in Fig. 7d. Distilled water, apple juice and NaCl solution (1.1 g NaCl was dissolved in 1 liter of distilled water, resulting in a conductivity of 2.5 mS/cm; the same conductivity as the apple juice) were used. The number of rings affected the survival ratio. The optimal number was found to be between five and eight rings. Figure 8 compares

122

T. Ohshima · M. Sato

a

b

c

d

Fig. 7 Four types of electrode system: a plate-to-plate, b needle-to-plate, c rings-to-concentric cylinder, d coiled wire-to-cylinder

the energy efficiency of several kinds of electrode systems: plate-to-plate; insulating plate-to-plate; needle-to-plate; ring-to-cylinder; and coiled wire-to-cylinder, with a high conductivity liquid (2.5 mS/cm). The ring-to-cylinder and coiled wire-to-cylinder system were more effective than the others, their survival ratio reduction about four orders of magnitude higher than that of the plate-plate system.A slight reduction of the survival ratio was observed in the case of plateto-plate system (Fig. 8) as compared to the results shown in previous figures (such as Fig. 2). This is because the electrical conductivity of the liquid media is much higher (2.5 mS/cm) than that of distilled water. Recently, we proposed a new electrode configuration that consisted of two parallel wire electrodes, where the high-voltage and grounded wires are placed between two concentric Plexiglas cylinders. This new system outperformed other systems in terms of energy efficiency [25], and could be used in industrial liquid food sterilizers. The experimental results discussed above suggest the following points: (i) concentrated PEF using a plastic plate or a film with small holes yields a higher survival ratio reduction than the plate-to-plate electrode system; (ii) the ringsor coiled wire-to-cylinder type electrode system has the highest energy efficiency among all the systems studied; (iii) the use of pinholes on the surface of the plate electrode or ring edges with proper insulating materials generated

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

123

Fig. 8 Comparison of energy efficiency for sterilization between five kinds of electrode systems, where pulse voltage is 14 kV; baker’s yeast suspended in NaCl solution (2.5 mS/cm)

intense electric fields that enabled effective inactivation of microorganisms. However, it was observed that, in the case of a highly concentrated electric field, corona or streamer discharge occurs in the liquid phase. The electrical discharge leads to the formation of certain kinds of active species, such as OH radicals or hydrogen peroxide, through processes like recombination [26]. These species could kill bacteria effectively, however, at the same time, they could oxidize the contents of liquid foods. Therefore, PEF treatment should be performed so as to avoid the generation of electrical discharges.

3 Recovery of Intracellular Proteins Using PEF 3.1 Release of the Protein under PEF Although PEF has been used in the field of biotechnology, only a few reports suggested an electrical release of intracellular compounds [27]. During the PEF destruction process (Fig. 1), intracellular proteins seemed to be released.

124

T. Ohshima · M. Sato

Fig. 9 Effect of electric field strength on the release of proteins and survival ratio

Ohshima et al. reported the release of proteins from S. cerevisiae cells under PEF [28]. The relationship between cell sterilization and protein release was studied, and is shown in Fig. 9.While the amount of protein increased with the electric field strength, there was a region in which the increase of protein concentration was higher than the decrease of survival cell concentration (below 6 kV/cm). Because reversible pores seemed to form in the cell membrane in this region, the control of PEF strength could cause the release of the intracellular protein while the cell was still alive. This phenomenon (Fig. 9) resembles the biological secretion process. However, it is also independent of the biological process. Therefore, it could be applied to any bacteria culture, whether the bacteria have a biological secretion system or not. The maximal protein concentration was about 40 mg/mL, about 30% of that of glass bead homogenization, although the PEF treatment enabled the release of proteins in a much shorter time (up to 10 s). The release of intracellular protein indicated the formation of pores in the cell membrane. Benz et al. reported that the size and nature of the pores depended on the applied voltage: high voltage formed large pores, which caused cell death, while low voltage formed small and reversible pores [29, 30]. This variation in pore size suggested the selective release of intracellular protein. Enzyme activities of invertase and ADH in the supernatants from S. cerevisiae were reported for various PEF strengths. The variation in release properties with PEF strength (invertase and ADH activity increased at around 6 kV/cm and 12 kV/cm, respectively) indicated that the selective release of intracellular protein could be controlled by adjusting the applied PEF strength. Since invertase is found around the cell membrane and ADH in the cytoplasm near the center of the cell, selective protein release could be achieved through varying the PEF strength. Although protein release by PEF treatment was less efficient

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

125

compared to conventional methods, PEF treatment has the great advantage of allowing selective release of specific enzymes, such as target proteins. Biotechnology, in particular genetic engineering, has enabled the production of certain important heterologous proteins in such microorganisms as E. coli and S. cerevisiae. If we are to make use of these products as medical materials or subject them to molecular biological analysis, we need to extract them from the host strain and purify the target molecule. Thorough purification is typically a rather time-consuming and technically challenging process since all compounds contain unknown contaminates that are released as a result of traditional processes, such as glass bead homogenization. 3.2 Recovery of Recombinant Protein from E. coli by PEF Treatment Genetic engineering has led to the development of a new method for producing heterologous proteins in bacteria; there are currently numerous recombinant E. coli. In addition, using the high cell density cultivation technique [31–34], overproduction of heterologous proteins has been possible in E. coli cells. However, these products utilized in the medical field should be free of other host proteins, in particular, pyrogen. In general, cell disruption for recovery of intracellular proteins is carried out by ultrasonication or homogenization, which causes the complete disruption of the cell, and all of the intracellular proteins and materials are released. Much effort must be expended to extract and purify even a small quantity of the target molecule. By contrast, PEF appears to have enabled selective recovery of intracellular proteins [28]. Ohshima et al. [35] applied the PEF technique to recovery of protein from recombinant E. coli species. Figure 10 shows the three kinds of recombinant E. coli species and the positions where the recombinant proteins accumulate in the cell. Recombinant b-glucosidase produced by E. coli/pNC1 is accumulated in the cytoplasm [36], and recombinant a-amylase produced by E. coli/ pHI301A is accumulated in the periplasmic space [37]. E. coli/pNB6 [38] produces recombinant cellobiohydrolase and accumulates this enzyme in both the periplasmic space and cytoplasm (Fig. 10). Effect of the electrode shape on release of recombinant proteins was studied under the same electric field conditions (7.5 kV/cm, 100 J/ml) as were used in the three types of electrode systems previously mentioned in Fig. 5. Uniform PEF is generated in the plate-plate (P-P) chamber (Fig. 5a), a nonuniform PEF, concentrated around the needle electrode, is generated by using the needle-plate (N-P) chamber (Fig. 5c). Plate-plate electrodes with edge (P-PE) would generate a PEF of intermediate strength (Fig. 5b). Since recombinant b-glucosidase produced by E. coli/pNC1 is accumulated in the cytoplasm, it could only be released by the N-P chamber, while recombinant a-amylase produced by E. coli/pHI301A, which accumulated in the periplasmic space, could be released by all treatment chambers used in the present study. In the case of E. coli/pNB6, which produces recombinant cellobiohydrolase and

126

T. Ohshima · M. Sato

Fig. 10 Three types of recombinant E. coli and gene products with accumulation positions

accumulates this enzyme in both the periplasmic space and cytoplasm, the recombinant protein could be released by all chambers. However, the amount of cellobiohydrolase released in the N-P chamber was higher than those in the P-P and P-PE chambers. The amount of enzyme released, however, was very small: 10 times (a-amylase), 20 times (cellobiohydrolase), or 60 times (b-glucosidase) lower compared to the amounts achievable through ultrasonication. The effect of varying the electric field during the release of b-glucosidase from E. coli/pNC1 was studied. Increasing the electric field strength or input energy yielded a higher amount of b-glucosidase in the case of the N-P chamber, while no significant improvement was observed in the case of the P-P and P-PE chambers. An electric strength of 12 kV/cm or higher causes a spark discharge between the needle and plate electrode in the N-P chamber. The electric discharge produces certain kinds of radical species [26], which have a negative effect on the enzyme by virtue of their high oxidizing capacity. E. coli has an outer membrane in addition to its cytoplasmic membrane. The release of cytoplasmic protein, therefore, is possible only when both the outer and cytoplasmic membrane is disrupted. A uniform PEF was found to disrupt only the outer membrane, enabling the release of periplasmic enzymes, while a nonuniform PEF, highly concentrated around the needle electrode, would be necessary for the release of cytoplasmic proteins.

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

127

3.3 Selective Release of Recombinant Proteins from E. coli by PEF Treatment As described previously, gene products could not be recovered efficiently from recombinant E. coli suspended in distilled water. To improve the recovery of recombinant protein, the effect of the solution in which recombinant E. coli was suspended was studied. Glycine [39, 40], NaCl, polyethylene glycol (PEG), and combinations thereof were investigated as possible solutions. The most efficient release conditions and the release results for each recombinant E. coli are summarized in Table 1.When E. coli/pNC1 was suspended in a solution containing 5% glycine and 15% of PEG under a PEF of 10 kV/cm and 280 J/mL in an N-P chamber, 5.2 U/mL of b-glucosidase was released. The amount of b-glucosidase released increased with the input energy, which is proportional to the treatment time. However, the sample temperature rose above 50 °C when a 280 J/mL PEF was applied to the N-P chamber. Although the maximum amount of b-glucosidase possible was released, the activity and specific activity of the b-glucosidase released corresponded to only 26% and 50%, respectively, of the values obtained using ultrasonication. It was confirmed by SEM that yeast cells sterilized by PEF treatment undergo no apparent morphological changes [28]. Therefore, PEF treatment would result in a partial disruption of the cell membrane even under higher PEF values, which reduces the release efficiency of bglucosidase in the cytoplasm. In addition, the decrease of specific activity suggested that the PEF-induced pore diameter was too small for the release of large molecules like b-glucosidase, but large enough for smaller proteins. Sota et al. reported that this cloned b-glucosidase is estimated to be 50 kDa [36]. These results suggested that more than 50 kDa of the target protein in the cytoplasm could not be released efficiently by PEF treatment. However, the PEF technique would be useful for the release of smaller molecules from the cytoplasm, allowing selectivity of molecular size. On the other hand, cellobiohydrolase (pNB6) and a-amylase (pHI301A) could be released with higher specific activity than is achievable by ultrasonication: when E. coli/pNB6 was suspended in 15% of PEG solution and 10 kV/cm and 200 J/mL in an N-P chamber, 70% cellobiohydrolase with twice the specific activity was released.Although the amount of cellobiohydrolase released in the N-P chamber was higher than in the P-PE chamber, the total amount of protein released also increased, which resulted in an almost equivalent specific activity. Suspending E. coli/pHI301A in a solution of 0.9% NaCl and 10% PEG under a PEF of 10 kV/cm and 200 J/mL in a P-P chamber yielded the selective release of 89% a-amylase with more than nine times higher specific activity.Although the amounts of a-amylase released in the P-P and N-P chambers were almost equal, the total amount of protein released in the N-P chamber was higher than that released in the P-P chamber. The increase in protein release in the N-P chamber seemed to have a cytoplasmic origin. These results indicated that a uniform PEF generated in the P-P chamber could disrupt only the outer mem-

P-P (10)

HB101/pHI301A (a-amylase)



amount of released enzyme Recovery = 00007 , amount of enzyme




N-P (10)

HB101/pNB6 (cellobiohydrolase)

a

N-P (10)

Electric field strength [kV/cm]

Treatment chamber

Release condition

HB101/pNC1 (b-glucosidase)

E. coli (product)

b




0.9% NaCl +10% PEG

15% PEG

5% Glycine +15% PEG

Suspended solution

133 (149)

0.20 (0.28)

5.2 (19.8)



amount of released enzyme Purificity = 00007 . amount of enzyme

200 (76)

200 (76)

280 (106)

Input energy [J/mL] (treatment time [s])

Activity with PEF [U/mL] (activity with ultrasonication [U/mL])

215 (24)

0.096 (0.050)

1.8 (3.6)

0.89

0.70

0.26

Specific activity Recoverya with PEF [U/mL] (specific activity with ultrasonication [U/mg])

Table 1 Summary of the conditions for the most effective release of cloned proteins from three kinds of recombinant E. coli

9.0

1.9

0.5

Purificityb

128 T. Ohshima · M. Sato

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

129

brane, while a nonuniform PEF generated in the N-P chamber could disrupt both the outer and cytoplasmic membrane, which caused the increase in total released protein concentration. 3.4 Secretion of Recombinant Protein from E. coli by PEF Treatment A wide variety of technologies concerning intracellular expression have been studied and established for the high-level production of heterologous proteins. Since intracellular production methods are relatively straightforward, intracellular expression has been preferred. However, the potential advantages of protein secretion over intracellular expression should not be overlooked. Using secretory production of recombinant proteins, it is possible to maintain the natural N-terminal residue, to take advantage of the enzymatic system in the periplasm that allows disulfide bond formation [41], or to facilitate purification by physical separation of the recombinant protein from the bulk of endogenous contaminants. Despite these potential advantages, the secretion of heterologous proteins has not been extensively used as a means for high-level production because of the often low or undetectable secretion levels and incomplete processing of the precursor to the mature protein in many cases. In genetic engineering, the bacterium E. coli has been widely used as a host cell. E. coli, a Gram-negative bacterium, has an outer membrane in addition to an inner membrane. Although a recombinant protein could be transported to the periplasmic space by using a signal peptide technique, it has been difficult to secrete to the culture medium. Recently, some of the recombinant protein was successfully produced extracellularly [42–44]. However, it was unclear how these recombinant proteins passed through the outer membrane, and overproduction could cause cell lysis [44]. The study of the release profiles of intracellular proteins from microorganisms using PEF revealed that the PEF technique enabled protein release during cultivation of E. coli without cell death, through a secretion-like process. As illustrated in Fig. 11, PEF treatment could cause a partial and reversible destruction of the outer membrane, and the recombinant protein accumulated in the periplasmic space could be released to the medium (A), the outer membrane could be reconstructed spontaneously (B), and recombinant proteins may be re-produced and accumulated in the periplasmic space (C). Ohshima et al. demonstrated a novel extracellular production of recombinant protein by using the PEF technique [45]. When E. coli/pHI301A, in which recombinant a-amylase is produced in the periplasmic space, was cultivated without PEF treatment, the released a-amylase was not detected during cultivation. However, when a PEF of 12 kV and 1 Hz was applied to this culture during cultivation, a-amylase was detected in the culture medium. Figure 12 shows the cultivations of E. coli/pHI301A obtained when the PEF treatment was started after 2 h of cultivation. In this cultivation, the PEF treatment was started at the late logarithmic growth phase (OD660=0.4). The time course

130

T. Ohshima · M. Sato

Fig. 11 Schematics of “secretion” of periplasmic recombinant protein using PEF treatment. PEF treatment could cause a partial and reversible destruction of outer membrane, and recombinant protein accumulated in periplasmic space could be released to the medium a, outer membrane could be reconstructed spontaneously b, and recombinant proteins might be re-produced and accumulated in periplasmic space c

of cell growth was almost identical to that obtained in the absence of PEF, and no cell death was observed during cultivation. Though a significant increase in the total protein concentration in the medium was not observed, the a-amylase activity was increased during cultivation with the PEF treatment. These results suggested that extracellular secretion of recombinant proteins from E. coli could be successfully achieved by using the PEF technique.

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

131

Fig. 12 Cultivation profile of E. coli HB101/pHI301A with PEF treatment. Shadow area indicates the period of PEF treatment

4 Summary PEF sterilization is a method for destroying cell membranes physically, without resorting to heat or other processes that affect protein structure and function. Many recent studies have focused on applying PEF sterilization technology to the production of common liquid foods, such as orange juice, apple juice, and milk. We have investigated PEF sterilization from a biological and electrical viewpoint, and our research and development activities geared towards applications will undoubtedly continue. Dairy or fresh juice products whose safety was ensured by PEF sterilization may soon become commercially available. The present authors have proposed a novel method for PEF-induced intracellular protein release. The method is very different from the conventional cell crushing method: it involves collecting the protein through holes

132

T. Ohshima · M. Sato

generated in the cell membrane.An effective and selective recovery of periplasmic protein from Gram-negative bacteria such as E. coli can be achieved using this method. As PEF treatment is a rapid process, further advances using PEF are to be expected. PEF, with its interesting features, lies at the forefront of research in several fields, in particular, biotechnology.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

Zimmermann U,Vienken J, Halfmann J, Emeis CC (1985) Adv Biotechnol Processes 4:79 Zimmermann U, Scheurich P, Pilwat G, Benz R (1981) Angew Chem Int Ed 20:325 Coster HGL, Steudle E, Zimmermann U (1976) Plant Physiol 58:636 Andreason GL, Evans GA (1988) Biotechniques 6:650 Neumann E, Sowers AE, Jordan CA (1989) Electroporation and electrofusion in cell biology. Plenum, New York Zimmermann U (1982) Biochim Biophys Acta 694:227 Zimmermann U (1983) Trends Biotechnol 1:149 Coster HGL, Zimmermann U (1975) J Membrane Biol 22:73 Sato M, Tokita K, Sadakata M, Sakai T, Nakanishi K (1990) Int Chem Eng 30:695 Sato M, Kimura K, Ikeda K, Ohgiyama T, Hata K (1994) Sterilization of beverages under normal temperature by a high-voltage, pulsed discharge. In: Yano T, Matsumo R, Nakamura K (eds) Developments in food engineering (vol 2). Blackie, London pp 736–738 Jayaram S, Castle GSP (1992) Biotech Bioeng 40:1412 Ohshima T, Sato K, Terauchi H, Sato M (1997) J Electrostat 42:159 Mitchell WJ, Slaughter JC (1989) Biology and biochemistry for chemists and chemical engineers. Wiley, New York McElhaney RN, Souza KA (1976) Biochim Biophys Acta 443:348 Ohshima T, Okuyama K, Sato M (2002) J Electrostat 55:227 Sinensky M (1974) Proc Natl Acad Sci USA 71:522 Hockney RC (1994) Trends Biotech 12:456 Liang Z, Mittal GS, Griffiths MW (2002) J Food Protect 65:1081 Terebiznik M, Jagus R, Cerrutti P, Huergo MS, Pilosof AMR (2002) J Food Protect 65:1253 Khadre MA, Yousef AE (2002) J Food Protect 65:1441 Picart L, Dumay E, Cheftel JC (2002) Innov Food Sci Emerg Technol 3:357 Mizuno A, Hori Y (1988) IEEE T Ind App 24:387 Sato M, Ishida NM, Sugiarto AT, Ohshima T, Taniguchi H (2001) IEEE T Ind Appl 37:1646 Sato M, Ishida NM, Sugiarto AT, Ohshima T (2001) Proc 11th World Congress of Food Science and Technology, W05-5, 22–27 April 2001, Seoul, Korea, p 51 Ishida NM, Sugiarto AT, Ohshima T, Sato M (2003) Japan J Food Eng (in Japanese) 4:47 Sato M, Ohgiyama T, Clements JS (1996) IEEE T Ind App 32:106 Sale H, Hamilton WA (1967) Biochim Biophys Acta 148:781 Ohshima T, Sato M, Saito M (1995) J Electrostat 35:103 Benz R, Zimmermann U (1980) Biochim Biophys Acta 597:637 Zimmermann U, Benz R (1980) J Membrane Biol 53:33 Mori H, Yano T, Kobayashi T, Shimizu S (1979) J Chem Eng Jpn 12:313 Iijima S, Kai K, Mizutani S, Kobayashi T (1986) J Chem Tech 36:539 Kawabe T, Ohshima T, Uozumi N, Iijima S, Kobayashi T (1992) J Chem Eng Jpn 25:702 Das A (1990) Method Enzymol 182:93

Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field

133

35. Ohshima T, Hama Y, Sato M (2000) Biochem Eng J 5:149 36. Sota H, Arunwanich P, Kurita O, Uozumi N, Honda H, Iijima S, Kobayashi T (1994) J Ferment Bioeng 77:199 37. Tsukagoshi N, Ihara N, Yamagata H, Udaka S (1984) Mol Genet Genomics 193:58 38. Honda H, Naito H, Taya M, Iijima S, Kobayashi T (1987) Appl Microbiol Biot 25:480 39. Ariga O, Watari T, Ando Y, Fijishita Y, Sano Y (1989) J Ferment Bioeng 72:243 40. Ariga O, Miyakawa I, Aota T, Sano Y (1994) J Ferment Bioeng 77:71 41. Bardwell JCA, Lee J, Jander G, Martin N, Bellin D, Beckwith J (1993) Proc Natl Acad Sci USA 90:1038 42. Loo T, Patchett ML, Norris GE, Lott JS (2002) Protein Expres Purif 24:90 43. Xu R, Du P, Fan JJ, Zhang Q, Li TP, Gan RB (2002) Protein Expres Purif 24:453 44. Lee J, Saraswat V, Koh I, Song KB, Park YH, Rhee SK (2001) FEMS Microbiol Lett 195:127 45. Ohshima T, Masao N, Sasazawa Y, Sato M (1996) In: Proc YABEC ’96 Symp, 14–16 Sept 1996, Kyoto, Japan, p 86

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 135 –149 DOI 10.1007/b94195 © Springer-Verlag Berlin Heidelberg 2004

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications Hideo Nakano · Yasuaki Kawarasaki · Tsuneo Yamane (✉) Laboratory of Molecular Biotechnology, Graduate School of Biological and Agricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan [email protected]

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

1

Introduction

2

Improvement of the Escherichia coli Cell-Free Protein Synthesis System and its Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . High-Throughput Construction of a Protein Library by SIMPLEX . . . Development of SIMPLEX . . . . . . . . . . . . . . . . . . . . . . . . . Quality of the SIMPLEX-based Protein Library . . . . . . . . . . . . . . Expansion of the SIMPLEX-based Library . . . . . . . . . . . . . . . . Application of SIMPLEX for Protein Engineering . . . . . . . . . . . .

2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 3 3.1 3.2 3.3 3.4 3.5

. . . . . . .

137 137 138 138 139 140 140

Highly Efficient Eukaryotic Cell-Free Translation System: A Useful Tool for Functional Genome/Proteome Analysis . . . . . . . . . . . Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trimming of Viral CITE Sequence . . . . . . . . . . . . . . . . . . . . . . . . Characterization of CITE Activity of the TE(37–65) Sequence . . . . . . . . . Rapid Gene Expression Using the TE(37–65) Sequence . . . . . . . . . . . . An Advanced Framework for Rapid Eukaryotic Gene Expression and Analysis

141 141 143 143 146 147

References

. . . . . . .

. . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Abstract The Escherichia coli cell-free protein synthesis system can now be used for various proteins that need special requirements, such as disulfide bond formation between intraand inter-molecules, hetero-dimerization, and specific chaperons. In addition, a novel protein library construction method termed “SIMPLEX” has been developed. Some applications with SIMPLEX are described. A highly efficient eukaryotic cell-free translation system using wheat germ extract has also been developed. An advanced framework for a rapid eukaryotic gene expression and analysis is shown. Keywords Cell-free protein synthesis · E. coli S30 extract · Wheat-germ extract · Single-molecule PCR · Protein library

136

H. Nakano et al.

1 Introduction Incorporation of amino acids into proteins in cell homogenates, in cell extracts, and in cell-free fractions containing microsomes or ribosomes was demonstrated as early as the 1960s, and such a reaction system in vitro has been called a “cell-free translation” or a “cell-free protein synthesis system”. Today, cell-free protein-synthesizing systems can be reconstituted from well-characterized, highly purified components, including ribosomes, template polynucleotides, all-soluble proteineous translation factors, GTP, and a set of aminoacyl-tRNAs or a system of tRNA amynoacylation, namely, tRNA, amino acids, ATP, and aminoacyl tRNA synthetases. Using these purified components, the molecular mechanism of gene translation in/on the ribosome is now being studied extensively. Usually, however, crude cell extracts comprising all of these endogenous components and factors are used in routine laboratory practices. As materials containing all the necessary components and containing ribosomes as the main constituent, there are three cell-free translation systems that utilize organelles originated from the following different living organisms: 1. S30 extract of Escherichia coli (as a prokaryotic system) 2. Wheat germ extract (as a phytogenic system) 3. Rabbit reticulocyte lysate (as an animal system)

Nowadays all of these are commercially available as kits and are routinely used in laboratories, but their preparation procedures remain essentially unimproved since they were originally used. For many years we have worked with the situation where their amount of translation products are only detectable by incorporation of labeled amino acid into the synthesized protein (amounts as little as pico- to sub nano-gram per mL per hour). In other words, there has been one principal shortcoming of all cell-free translation and coupled transcription/translation systems: in contrast to in vivo protein synthesis, they have short lifetimes, and as a consequence, give a low yield of the protein synthesized. The poor productivity only allows the system to be applied to some biochemical analyses of gene products or analyses of translation mechanisms. Therefore, cell-free protein synthesis systems have not been regarded as an alternative method for preparative syntheses of foreign polypeptides and proteins. However, many scientists, including our group, have devoted their efforts over the last few decades to considerably increasing the performance of the cell-free protein synthesis system, and their achievements have widened its applicability to various areas of both bioscience and biotechnology. Notably, Spirin et al. reported a novel reaction system called CFCF (Continuous-Flow Cell-Free) system where the cell-free translation system was retained by a membrane bioreactor, resulting in a significant increase in the amount of protein, possibly due to the continuous supply of substrates and removal of wastes [1]. Following this milestone paper, several important papers have appeared,

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

137

such as those on long-life or robust wheat-germ cell-free systems [2, 3], and on high-yield E. coli systems [4, 5]. In addition, a reconstituted protein synthesis system containing only purified translation machinery components, which is called PURE SYSTEM, has been recently reported [6]. Together with improvements in the productivity of the cell-free expression systems, recent advances in peripheral techniques, such as template preparation using PCR and miniaturization of reaction/detection systems, allow us to apply the cell-free systems to unprecedented areas including high-throughput expression and screening of a mutant protein library or cDNA library. We will review this expanding field shortly, focusing on our recent improvements and applications of E. coli and wheat-germ systems.

2 Improvement of the Escherichia coli Cell-Free Protein Synthesis System and its Application 2.1 Overview One of the advantages of the cell-free protein synthesis systems is its high adjustability in regard to various conditions for both translation and folding, such as temperature, oxido-redox conditions, and chaperons for the individual protein. Originally, the cell-free protein synthesis reaction was done under rather reductive conditions, because cytoplasm is normally reductive in nature. However, E. coli cell-free protein synthesis systems were shown to be relatively stable under oxidative conditions. Under such oxidative conditions, active single-chain Fv was synthesized successfully [7]. In addition, some industriallyimportant enzymes, such as Burkholderia cepacia KWI-56 lipase with a sole disulfide bond and Streptomyces antibioticus Phospholipase D with four disulfide bonds were also synthesized as active form [8, 9]. In the case of B. cepacia lipase, the co-expression of its specific chaperon or the addition of the purified one, both of which are not easy in the E. coli recombinant expression system, was required for the active production of the lipase. A Fab fragment of a catalytic antibody was also successfully produced as an enzymatically active form with an intermolecular disulfide bond, suggesting that the system can produce multimer proteins as well as monomer ones (see Fig. 1) [10]. Moreover, Phanerochaete chrysosporium manganese peroxidase (MnP), which cannot be produced as active form in recombinant systems, possibly due to the complicated structure of MnP, includes five disulfide bonds, Mn2+, Ca2+ and heme as the catalytic center. Addition of heme helped the production of active form of the enzyme [11]. These examples demonstrate the high adjustability and flexibility of the cell-free protein synthesis system for various types of proteins.

138

H. Nakano et al.

Fig. 1 Cell-free protein synthesis of a Fab fragment of antibody. Autoradiography of the non-reducing SDS-PAGE of 6D9 Fab synthesized in an E. coli transcription/translation system. The genes encoding the mature light chain and the Fd of the heavy chain of the 6D9 antibody were independently cloned under T7 RNA polymerase promoter. L and H indicate the use of templates of light chain and Fd of heavy chain, respectively. The figure is slightly modified from the original [3]

2.2 High-Throughput Construction of a Protein Library by SIMPLEX 2.2.1 Development of SIMPLEX Since the cell-free protein synthesis system can use the PCR product directly as a template for transcription/translation, and is compatible with a multi-well plate format, it can be said to be quite an appropriate system for high-throughput screening. Our group developed a novel protein library construction system using both amplification of single DNA molecules and the cell-free protein synthesis system. The template DNA molecules are diluted typically to one molecule per well, amplified by PCR, and transcribed and translated by the cellfree system as illustrated in Fig. 2. The reduction of accumulation of primer dimers is the most important point for the successful amplification of a single DNA molecule that is large enough to encode fully functional proteins, typically more than 1 kbp. To amplify a single DNA molecule specifically, the nested PCR method was first employed [12]. However, the two-step PCR protocol was not suitable for high-throughput PCR to obtain a larger size of protein library. Then one step amplification of a single DNA molecule was established using single-primer amplification and hot-startable DNA polymerase [13]. Single primer PCR against a homo-tailed template drastically reduces the

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

139

Fig. 2 A schematic drawing of a novel protein library construction system named SIMPLEX

accumulation of primer dimers, because the primer dimers made of a single primer have a panhandle structure, which effectively represses the entry of a free primer to the dimer, preventing the further accumulation of the dimers. Eventually, even a single DNA molecule could be amplified, and the resulting DNA library was transformed to a protein library ready for functional screening by the cell-free protein synthesis system or in vitro expression. This novel system is named SIMPLEX: single-molecule-PCR-linked in vitro expression, and has been proved to give a highly uniform protein library even though the library comes from each DNA molecule after extensive amplification of approximately 1012. 2.2.2 Quality of the SIMPLEX-based Protein Library “Quality” of the protein library is as important as the variety of molecules that can be handled at one time. If the amount of protein produced in each clone varies one by one, for example, it would be very difficult to search and find a real positive candidate from the sea of noise. Therefore the uniformity of the

140

H. Nakano et al.

amount of proteins obtained by SIMPLEX was examined [14]. The amounts of PCR product amplified from a single green fluorescent protein (GFP) gene were almost equal to each other, and the amount of GFP proteins produced in each well were also highly uniform. The relative standard deviation (RSD) was about 8%, which is pretty similar to that of the liquid volume of micropipettes. Since the RSD index was more than 25% using a conventional E. coli expression system, SIMPLEX can provide a much more highly uniform protein library than the conventional cell-based system. 2.2.3 Expansion of the SIMPLEX-based Library Although the size of a protein library generated by SIMPLEX would be ultimately unlimited, this is not currently practical due to its cost. A miniaturized PCR system that could empower the SIMPLEX is not commercially available at present. Therefore, a practically achievable library size is limited by the capacity of existing thermocyclers and the need for costly DNA polymerase. In order to increase the achievable size of the SIMPLEX-based library, a multiple molecules PCR (such as for five molecules), was tried for the initial construction of the library. Interestingly, each five molecules in a tube were amplified equally even after extensive amplification [15]. The result means that the searchable number of variations can be easily multiplied by increasing the average number of molecules in a PCR well using a double-screening method (Fig. 3). For example, amplification of five molecule PCR instead of single molecule PCR in a 384-well plate actually can make approximately 1900 variations of proteins at one time. The system has the following features compared to the conventional systems: 1) the library size is theoretically unlimited; 2) the total time necessary to obtain a protein library from each DNA molecule is very short, only about 5 h; 3) it is easy to automate; 4) cytotoxic proteins can be expressed; 5) various assay methods can be applied because the library is constructed on a multi-well plate. Since the system needs no biological cloning or cultivation of recombinant cells, and is performed exclusively on a multi-plate format, its throughput is drastically higher than a conventional library method using agar plates. 2.2.4 Application of SIMPLEX for Protein Engineering Using the system, our group has already succeeded in making some mutant proteins with improved or modified characters. A mutant manganese peroxidase of Phanerochaete chrysosporium which has an improved resistance against hydrogen peroxide has been successfully screened from a regional specific random library constructed by SIMPLEX [11]. In addition, the enantioselectivity of the Burkholderia cepacia KWI-56 lipase has been completely reversed by high-throughput screening of a combinatorially mutated library constructed by

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

141

Fig. 3 Schematic representation of expanded SIMPLEX-based library construction and screening. 1: DNA templates in the gene pool are diluted to a specified number of molecules per well and amplified by PCR, namely multiplex-PCR, yielding a primary PCR library. 2: The PCR library is converted to the primary protein library by means of cell-free protein synthesis. 3: The protein library is screened for clones with desired properties. 4: DNA encodings for positive clones are collected, and prepared for the second round of screening where single-molecule PCR (SM-PCR) is performed to isolate single genes

SIMPLEX [17]. Since we have developed a fully-automatic system that enables the construction of the protein library by SIMPLEX on 1536-well plates, the system will be able to make novel proteins in a high-throughput manner.

3 Highly Efficient Eukaryotic Cell-Free Translation System: A Useful Tool for Functional Genome/Proteome Analysis 3.1 Overview Cell-free expression of eukaryotic genes using a eukaryotic expression platform allows us to analyze the intrinsic structures and/or functions of their products. Although a highly productive expression system based on bacterial cell-free

142

H. Nakano et al.

extract has been used for large-scale preparation of defined proteins, the bacterial system is not amenable to rapid preparation of a number of anonymous eukaryotic proteins from their genes, which is one of the essential properties for the expression platform for eukaryotic genes. The lower compatibility of the bacterial expression system for eukaryotic gene expression is due to intrinsic differences in translation between a eukaryote and a prokaryote: most eukaryotic mRNA has a unique structure at its 5¢ terminal. The structure is called 5¢ cap, and plays an essential role in ribosome recruitment onto mRNA. Besides the 5¢ cap structure, eukaryotic mRNA lacks an obvious ribosome binding sequence in the 5¢ untranslated region (UTR), which prevents itself from being translated by bacterial translation machinery. The yield of synthesized protein in the eukaryotic expression system using wheat germ extract has been increased by improvements in reaction rate [18] as well as reaction longevity [19]. During these studies, we also identified factors that caused the cessation of translation reaction. The factors are wheat endogenous nucleases and acid phosphatases [20, 21]. We reported that the addition of a copper ion, an inhibitor to both of the endogenous enzymes, to the translation mixture resulted in a significant increase in protein production [18, 21]. These efforts allowed us to synthesize various eukaryotic proteins in sufficient amounts required for the functional analysis using biochemical or immunochemical methods. The throughput of protein production in the conventional eukaryotic system was, however, not yet amenable to the high-throughput expression and screening because of the cap-dependent translation manner; although the capped mRNA can be synthesized by in vitro transcription with the existence of a cap analog (7mGpppG), the capped transcript cannot be added to the translation reaction mixture without purification. This is because residual cap analog unincorporated into transcripts seriously inhibits translation initiation by capturing the cap-binding translation initiation factor that is essential for ribosome recruitment. The purification of the transcript is, therefore, an unavoidable step for efficient protein synthesis when the capped mRNA is used as a template, despite the fact that the step increases risk of cross-contamination among samples and significantly retards the throughput of the gene expression. The use of a transcript equipped with a certain kind of viral 5¢ UTR (>140 bases), that involves a cap-independent translation enhancing element (CITE), can circumvent the tedious purification step. Nevertheless, the target gene must still be joined with the viral sequence by an overlapping extension or subcloning into the expression vector, meaning that another cumbersome process is required. In this section, we show that the sequence responsible for the viral CITE was mapped, and it could be trimmed to a length sufficient to be involved in the PCR primer. Moreover, based on this trimmed sequence, we established a framework for the rapid eukaryotic gene expression system that was amenable to the high-throughput analysis of eukaryotic proteins [22].

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

143

3.2 Trimming of Viral CITE Sequence Tobacco etch virus (TEV) is a plant RNA virus whose genes are translated in cap-independent manner. 5¢ UTR of TEV is 144 bases in length and shows high CITE activity in vitro [19, 23] and in intact cells [23, 24] when it is located upstream of a reporter gene. The sequence identity of 5¢ UTR among the viruses closely related to TEV is, however, very low (Fig. 4a). First, we tried to map the sequence responsible for CITE activity empirically on the TEV 5¢ UTR. We prepared dihydrofolate reductase (dhfr) genes with a series of fragments of the TEV 5¢ UTR, deleted according to a predicted secondary structure and a sequence similarity that was conserved between several potyvirus 5¢ UTRs (Fig. 4a). After in vitro transcription of each construct with or without 5¢ capping, each transcript was subjected to in vitro translation using wheat-germ extract. We compared the translational efficiencies of these transcripts with that of the control mRNA transcribed from pGEM’DHFR [19]. The amount of synthesized DHFR protein from each mRNA was also summarized in Fig. 4b. Among them, the TE(37–65) sequence was the most promising candidate to be a minimum unit for the cap-independent translation, due to several reasons. The TE(37–65) sequence is relatively conserved between several potyviral 5¢ UTRs (Fig. 4a). In spite of the low sequence similarities between the 5¢ UTRs of potyviruses, AAKC (where K is G or U) quartets and oligopyrimidine tracts emerge in most potyvirus 5¢ UTRs in unusual numbers (Fig. 4a). This may suggest that the two kinds of reiterative motifs are involved in the regulatory elements for the viral proliferation. In fact, the oligopyrimidine tract in picornavirus RNA was revealed to interact with p57/PTB and facilitate translation initiation [25]. Although there has been no information about biological implications of the AAKC quartets, the reiterative 4AAKC quartets in the TE(37–65) sequence consequently lead to the formation of two sets of the “B box” (UCAAGCA [26]) which is the well-conserved seven residue sequence between several potyvirus 5¢ UTRs. For these reasons, the TE(37–65) sequence probably includes one of the regulatory elements responsible for the viral capindependent translation. 3.3 Characterization of CITE Activity of the TE(37–65) Sequence The CITE activity of the TE(37–65) sequence was not influenced by the reporter gene sequence. Figure 5a and b show the time course analyses of the translation reaction of chloramphenicol acetyl transferase (cat) and green fluorescent protein (gfp). In both cases, the translational rates of the uncapped mRNAs containing the TE(37–65) sequence were comparable to that of capped versions or the mRNAs containing the entire TEV 5¢ UTR, whereas the uncapped mRNAs without the viral sequence were poorly translated. The GFP

144

H. Nakano et al.

A

B

Fig. 4 a Comparison of the TEV 5¢ UTR sequence with those of some other potyviruses. The sequence of each potyvirus alignment was carried out using the BLASTN program, followed by manual modification. The boxed sequence indicates the AAKC quartet, and underlined residues are oligo pyrimidine tracts. The asterisks below the sequences represent the matched residues between the listed potyviral 5¢ UTRs. Abbreviations: TEV, tobacco etch virus; TVMV, tobacco vein mottling virus; PVA, potato virus A; PPV, plum pox virus. b Schematic drawing of prepared TEV 5¢ UTR deletions. All constructs were prepared as a 5¢ UTR of the dhfr gene. The deleted sequences from TEV 5¢ UTR are represented by broken lines. The translation efficiency of each capped or uncapped mRNA was determined by the enzymatic activities of the synthesized DHFR. The average amount of synthesized DHFR in three independent experiments was tabulated at the left side of each construct

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

145

Fig. 5 Time-course analysis of translation reactions using capped or uncapped mRNAs. gfp a, or cat b mRNA bearing TE(1–114), TE(37-65), or the control 5¢ leader sequence derived from plasmid pRSETB (60 bases) was translated. Filled squares, filled circles, empty circles, and filled triangles represent uncapped TE(1–144) mRNA, uncapped TE(37–65) mRNA, capped TE(37–65) mRNA, and uncapped mRNA carrying the control 5¢ leader sequence, respectively. The optimum concentrations of uncapped and capped constructs were 50 and 20 mg/mL, respectively. c Immunoblot analysis of synthesized His6 tagged CAT protein. The CAT protein (11 units) synthesized from uncapped TE(37–65) cat mRNA which encoded the His6-tag peptide just next to the initiation codon (lane 1) was subjected to a SDS-polyacrylamide gel, followed by blotting on a nitrocellulose membrane.As controls, the same amount of the His6-tagged CAT protein (lane 2) or authentic CAT protein (lane 3) synthesized in the E. coli S30 in vitro transcription/translation system under the canonical SD sequence were also loaded. The CAT proteins were detected by anti-CAT rabbit antibody (upper panel) or anti His6-tag monoclonal antibody (lower panel). Closed and open arrowheads indicate the His-tagged CAT protein and authentic CAT protein, respectively. Arrows indicate proteolytically degraded CAT protein

protein synthesis using uncapped TE(37–65)-gfp mRNA lasted for 10 h, like that of the capped counterpart. The CITE activity of the TE(37–65) and its translation longevity were also confirmed by cat mRNA (Fig. 5b), suggesting that the CITE activity was independent of the reporter gene sequences. The correct recognition of the defined initiation codon is very important, especially when this rapid expression system is employed for molecular dissection, where the truncated polypeptide should be synthesized precisely. The CAT gene construct used in Fig. 2b contained a His6-tag at its N-terminal. After 6 h of translation from the uncapped TE(37–65)-cat mRNA, the reaction mixture including synthesized CAT protein (11 unit/µL) was subjected to SDS-polyacrylamide gel electrophoresis. As controls, the same amount of an authentic- and the His6-tagged CAT protein synthesized in an E. coli S30 transcription/translation system under the canonical SD sequence were also

146

H. Nakano et al.

loaded. The blotted CAT proteins were then detected with the anti-CAT antibody or the anti His6-tag monoclonal antibody (Fig. 5c). The CAT protein synthesized in wheat-germ extract (lane 1) migrated to the expected point, and it was detected using both antibodies. It indicates that the TE(37–65) sequence guided the entered ribosome to the first and defined AUG codon, and the AUG codon worked effectively as the initiation codon. The band of His-tagged CAT protein synthesized in E. coli S30 extract (lane 2) was also detected by both antibodies. However, it was found that another smaller protein was detected with the anti-CAT antibody in lane 2. The intensity of this band was increased by further incubation (data not shown). Since this protein was not detected with the anti His6-antibody, this was a partially degraded His-tagged CAT protein that lacked its N-terminal region during the translation reaction in the E. coli S30 extract. 3.4 Rapid Gene Expression Using the TE(37–65) Sequence We then tried to demonstrate a rapid gene expression from rice cDNA clones. For practical use, the gene expression system requires the efficient translation of genes of which expression levels are strictly regulated in intact cells. The rice

Fig. 6 Rapid expression of rice cDNAs. The rice cDNA encoding a small subunit of ribulose bis phosphate decarboxylase (osss1139), or homeobox-gene osh15, was amplified by two rounds of PCR to attach the T7 promoter, TE(37–65) sequence, and T7 terminator. After in vitro transcription, the uncapped transcript was added to the translation mixture without purification. The translation reaction was carried out for 2 h in the presence of 14C lecine. The translated protein was then subjected to a SDS-PAGE, followed by fluorography. An arrow and an arrowhead represent the migration points of translated OSSS1139 and OSH15, respectively. The numbers under the fluorogram indicate the amount of synthesized protein (ng/mL) after 2 h of translation reaction

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

147

cDNA clones used here were osss1139 [27] and osh15 [28], which encoded a small subunit of ribulose bis-phosphate decarboxylase and one of rice homeobox protein, respectively. The former represents a highly translatable gene and the latter represents a poorly expressed one. They were amplified by two rounds of PCRs to attach the TE(37–65) sequence and T7 promoter to their 5¢ terminals. Simultaneously, a T7 terminator sequence was also integrated to the 3¢ ends of the genes. The amplified cDNAs were transcribed, then translated without purification of transcripts. The translation reaction was carried out for 2 h at 27 °C with 14C-leucine. The whole process for gene expression was completed within 10 h. The molecular weights of the products were found to be the same as the values calculated from their gene structures (Fig. 6). The amount of each synthesized protein was about 0.5 mg/mL of reaction mixture, suggesting that the homeobox gene osh15 was synthesized as much as the housekeeping gene osss1139. 3.5 An Advanced Framework for Rapid Eukaryotic Gene Expression and Analysis A graphical overview of our current expression framework established on the basis of this study is illustrated in Fig. 7. First, a target gene or target region is amplified by pairs of the target-specific inner primers and the inner-primerspecific outer primers (shown as thin lines and bold lines in Fig. 7, respectively). By reducing the concentration of inner primers to 1/25 of the outer primers, the fragment generated by inner primers is eventually amplified by the outer primers. As a result, the amplified fragment consists of a full set of DNA structures necessary for the cell-free expression. After a brief purification to remove unreacted dNTP and primers, the amplified fragment is subjected to in vitro transcription reaction, then in vitro translation reaction mixture is added directly to the transcription mixture to terminate the transcription reaction and to allow the synthesized mRNA to be translated. This gene expression framework allows us to prepare the gene product within a day. The target region to be expressed can be arbitrarily defined by the inner primers, since the trimmed translation enhancer guides the recruited ribosome to the first and defined AUG codon (Fig. 5c). This property allows us to carry out “molecular dissection” of genes of interest using the expression platform. In addition to the molecular dissection, this expression framework can also be applied to simultaneous expression of the cDNA pool, to discover the gene product responsible for interesting biological functions such as protein-protein interactions, catalytic activity, and specific modifications including phosphorylation, processing, and so on. Although the cDNA library should be constructed on an appropriate plasmid vector in a directional insertion and the inner primers are designed to anneal on the plasmid sequence, the amplification of the cDNA pool using the four primers allows instant conversion from the cloned inert cDNAs into expression-feasible templates. The obtained template pool is then subjected to the expression platform.

148

H. Nakano et al.

Fig. 7 Schematic overview of the cell-free expression of eukaryotic genes. The novel framework for the expression and analysis of eukaryotic genes is illustrated. (i) One-step conversion of expression-inert sequence into expression-feasible template. Target region to be expressed (closed box) is amplified by a combination of four primers (represented by tailed arrows) in order to attach the transcription promoter (PT7), the terminator (T7), and the capindependent translation enhancing sequence (TE(37–65)) to the ends of the target sequence. The thin and bold arrows above the target represent target-specific inner, and inner-primerspecific outer primers, respectively. The widths of the primers represent their concentration (see text). (ii) in vitro transcription to synthesize uncapped mRNA. (iii) in vitro translation of the synthesized mRNA. The translation reaction mixture is directly added to the transcription reaction mixture. Isotope-labeled amino acid (for example, [35S]-Met) can be used for the specific labeling of translation products. (iv) Functional analysis of synthesized polypeptides

Cell-free Protein Synthesis Systems: Increasing their Performance and Applications

149

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Spirin AS, Baranov VI, Ryabova LA, Ovodov SY, Alakhov YB (1988) Science 242:1162 Kawarasaki Y, Kawai T, Nakano H, Yamane T (1995) Anal Biochem 226: 320 Madin K, Sawasaki T, Ogasawara T, Endo Y (2000) Proc Natl Acad Sci USA 97:559 Kigawa T, Yabuki T,Yoshida Y, Tsutsui M, Ito Y, Shibata T Yokoyama Y (1999) FEBS Lett 442:15 Kim D, Swartz JR (2001) Biotechnol Bioeng 12:309 Shimizu Y, Inoue A, Tomari Y, Suzuki T,Yokogawa T, Nishikawa K, Ueda T (2001) Nature Biotech 19:751 Ryabova LA, Desplancq D, Spirin AS, Plukthun A (1997) Nature Biotech 15:79 Yang J, Kobayashi K, Iwasaki Y, Nakano H, Yamane T (2000) J Bacteriol 182:295 Iwasaki Y, Nishiyama T, Kawarasaki Y, Nakano H, Yamane T (2000) J Biosci Bioeng 89:506 Jiang X, Ookubo Y, Fujii I, Nakano H, Yamane T (2002) FEBS Lett 514:290 Miyazaki-Imamura C, Oohira K, Kitagagwa R, Nakano H, Yamane T (2003) Protein Eng 16:423 Ohuchi S, Nakano H, Yamane T (1998) Nucleic Acids Res 26:4339 Nakano H, Kobayashi K, Ohuchi S, Sekiguchi S, Yamane T (2000) J Biosci Bioeng 90: 456 Rungpragayphan S, Kawarasaki Y, Imaeda T, Kohda K, Nakano H,Yamane T (2002) J Mol Biol 318:395 Rungpragayphan S, Nakano H, Yamane T (2003) FEBS Lett 540:147 Koga Y, Kobayashi K, Yang J, Nakano H, Yamane T (2002) J Biosci Bioeng 94:84 Koga Y, Kato K, Nakano H, Yamane T (2003) J Mol Biol 331:585 Nakano H, Tanaka T, Kawarasaki Y, Yamane T (1996) J Biotechnol 46:275 Kawarasaki Y, Kawai T, Nakano H, Yamane T (1995) Anal Biochem 226:320 Kawarasaki Y, Nakano H, Yamane T (1996) Plant Sci 119:67 Kawarasaki Y, Nakano H, Yamane T (1998) J Biotechnol 61:199 Kawarasaki Y, Kasahara S, Kodera N, Shinbata T, Sekiguchi S, Nakano H,Yamane T (2000) Biotechnol Progr 16:517 Carrington JC, Freed DD (2000) J Virol 60:1590 Gallie DR, Tanguay RL, Leathers V (1995) Gene 165:233 Hellen CUT, Wimmer E (1995) Curr Top Microbiol 203:31 Turpen T (1989) J Gen Virol 70:1951 Matsuoka M, Kano-Murakami Y, Tanaka Y, Ozeki Y, Yamamoto N (1998) Plant Cell Physiol 29:1015 Sato Y, Sentoku N, Nagato Y, Matsuoka M (1998) Plant Mol Biol 38:983

Received: September 2003

Adv Biochem Engin/Biotechnol (2004) 90: 151–171 DOI 10.1007/b94196 © Springer-Verlag Berlin Heidelberg 2004

Enzymatic Synthesis of Structured Lipids Yugo Iwasaki · Tsuneo Yamane (✉) Laboratory of Molecular Biotechnology, Graduate School of Bio- and Agro-Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

2 2.1 2.2 2.3 2.4 2.5

Enzymatic Synthesis of Structured Triacylglycerols . . . . Importance of MLM-type STGs . . . . . . . . . . . . . . . . Synthesis of MLM-type STGs from Natural Oils and Fats . Conversion of Oils Containing Fungal Lipase-Resistant FAs Pure MLM-type STG Containing EPA . . . . . . . . . . . . Pure MLM-type STG Containing DHA . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

154 154 154 155 156 159

3 3.1 3.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4

Enzymatic Synthesis of Structured Phospholipids . . . . . . . . . Fundamentals of Phospholipids . . . . . . . . . . . . . . . . . . . Modification of Acyl Groups . . . . . . . . . . . . . . . . . . . . . Modification of Head Groups . . . . . . . . . . . . . . . . . . . . . Transphosphatidylation by PLD . . . . . . . . . . . . . . . . . . . Syntheses of Natural Type PLs . . . . . . . . . . . . . . . . . . . . Synthesis of PLs Containing Bio-Active Compounds . . . . . . . . Synthesis of Phosphatidylserine in a Non-Organic Solvent System

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

160 160 160 163 163 164 165 167

4

Conclusions

. . . . . .

. . . . . .

. . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Abstract Structured lipids (SLs) are defined as lipids that are modified chemically or enzymatically in order to change their structure. This review deals with structured triacylglycerols (STGs) and structured phospholipids (SPLs). The most typical STGs are MLM-type STGs, having medium chain fatty acids (FAs) at the 1- and 3-positions and a long chain fatty acid at the 2- position. MLM-type STGs are synthesized by: 1) 1,3-position-specific lipasecatalyzed acyl exchange of TG with FA or with FA ethylester (FAEt); 2) 1,3-position-specific lipase-catalyzed acylation of glycerol with FA, giving symmetric 1,3-diacyl-sn-glycerol, followed by chemical acylation at the sn-2 position, and; 3) 1,3-position-specific lipase-catalyzed deacylation of TG, giving 2-monoacylglycerol, followed by reacylation at the 1- and 3-positions with FA or with (FAEt). Enzymatic preparation of SPLs requires: 1) acyl group modification, and 2) head group modification of phospholipids. Acyl group modification is performed using lipases or phospholipase A2-mediated transesterification or ester synthesis to introduce arbitrary fatty acid to phospholipids. Head group modification is carried out by phospholipase d-catalyzed transphosphatidylation. A wide range of compounds can be introduced into the polar head of phospholipids, making it possible to prepare various SPLs. Keywords Structured lipid · Triacylglycerol · Phospholipid · Lipase · Phospholipase

152 Abbreviations and Symbols CA Caprylic acid CAEt Caprylic acid ethyl ester CEC 1,3-dicapryloyl-2-eicosapentaenoylglycerol CDC 1,3-dicapryloyl-2-docosahexaenoylglycerol COhC 1,3-dicapryloylglycerol DDD Tridocosahexaenoylglycerol DHA Docosahesaenoic acid DPA Docosapentaenoic acid EEE Trieicosapentenoylglycerol EFA Essential fatty acid EPA Eicosapentaenoic acid EPAEt Eicosapentaenoic acid ethyl ester FA Fatty acid FAEt Fatty acid ethyl ester GPC Glycerophosphorylcholine LCFA Long chain fatty acid LCT Long chain triacylglycerol MCFA Medium chain fatty acid MCFAEt Medium chain fatty acid ethyl ester MCT Medium chain triacylglycerol 2-MG 2-monoacylglycerol lysoPC Lysophosphatidylcholine NeuAc N-acetylneuraminic acid OhDOh 2-monodocosahexaenoylglycerol PA Phosphatidic acid PAsA Phosphatidylascorbic acid PC Phosphatidylcholine PCh Phosphatidylchromanol PE Phosphatidylethanolamine PG Phosphatidylglycerol PI Phosphatidylinositol PL Phospholipid PLase Phospholipase PLA1 Phospholipase A1 PLA2 Phospholipase A2 PLB Phospholipase B PLC Phospholipase C PLD Phospholipase D PS Phosphatidylserine PUFA Polyunsaturated fatty acid SCFA Short chain fatty acid SL Structured lipid SPL Structured phospholipid STG Structured triacylglycerol TG Triacylglycerol

Y. Iwasaki · T. Yamane

Enzymatic Synthesis of Structured Lipids

153

1 Introduction Structured lipids (SLs) are broadly defined as any lipids which are restructured to change the positions of their fatty acids (FAs) and modified to change the FA compositions from the natural state [1]. Natural edible fats, oils and lecithin are simply mixtures of a number of molecular species of triacylglycerols (TGs) or phospholipids (PLs) that are different in terms of FA species, their positional distribution on the glycerol backbone and type of polar head groups. In contrast to such natural lipids, SLs, which may include structured triacylglycerols (STGs) and structured phospholipids (SPLs), are TGs or PLs modified chemically or enzymatically to change their chemical structure (such as type of FAs, position of FA, and type of polar head groups). The molecular structure of a lipid influences its metabolic fate in organisms (digestion and absorption), physiological activities, as well as its physical characteristics (like its melting point). Consequently, designing SLs with particular chemical structure, it is possible to control the behavior of lipids, thereby improving the nutritional and pharmaceutical properties of lipids. Based on this perspective, much attention has been directed to the syntheses of SLs. One interesting example of STGs is the family of so-called low-calorie fats. Generally, saturated long chain fatty acids (LCFAs) are poorly absorbed and so impart less energy [2]. However, TGs consisting of only saturated LCFAs are not favorable as low-calorie fats, because their use is limited due to very high melting points. Therefore, reduced-calorie fats with appropriate physical properties were developed by attaching both saturated LCFA and short chain fatty acids (SCFAs), or by adding saturated LCFA and medium chain fatty acids (MCFAs) into one TG molecule. The best-known representatives of this low-calorie fat group are Salatrim and Caprenin. Salatrim, formed by chemical interesterification between TGs having LCFA (usually hydrogenated vegetable oil) and TGs having SCFA (C2:0–C4:0), has fewer calories (4.5–5.5 kcal/g) than conventional fats (9 kcal/g) [3]. Caprenin is a randomized TG comprising MCFA (C6:0–C8:0) and behenic acid (C22:0). Caprenin has 4.3 kcal/g which is the same calorie level as Salatrim [4]. In contrast to these chemically-prepared STGs having random structure, there are other groups of STGs (sometimes called “specific STGs”). Examples of such “specific STGs” include highly absorbable oil (MCFA-LCFA-MCFA, MLM) [5–7], human milk fat substitute (oleoyl-palmitoyl-oleoyl, OPO) [8] and cocoa butter equivalent (stearoyl-oleoyl-stearoyl, SOS) [9]. These “specific STGs” are different from the chemically prepared STGs with respect to the structural randomness. Therefore, in a stricter sense, STGs can be defined as TGs having particular FAs at particular positions of glycerol. Applying a similarly strict definition, SPLs are phospholipids having particular FAs and particular polar head groups at specific positions. The syntheses of these specific SLs of particular structure require specific modifications at the desired positions in the glycerol backbone. Although

154

Y. Iwasaki · T. Yamane

chemical syntheses are sometimes simple and inexpensive, they are not capable of modifying specific positions due to the random nature of the reactions. In contrast, the reactions catalyzed by enzymes such as lipases and phospholipases are more promising for positional-specific modification of lipids. In this review article, the recent advances in enzyme-catalyzed syntheses of SLs including STGs and SPLs are described.

2 Enzymatic Synthesis of Structured Triacylglycerols 2.1 Importance of MLM-type STGs Among reports on lipase-catalyzed syntheses of STGs, most deal with MLMtype STGs, in which MCFAs (with 6~10 carbons) are attached to sn-1,3 positions and LCFAs (with more than 12 carbons) are attached at the sn-2 position. The interest in MLM-type STGs might be related to so-called medium chain triacylglycerols (MCT) comprising MCFA, which is useful for clinical purposes as a rapid energy source for patients suffering from malabsorption of lipids [10–12]. Since MCT itself does not contain essential fatty acids (EFAs), supplementary use of EFAs to MCT is necessary.A simple mixture of MCT and long-chain triacylglycerols (LCT) containing EFAs, however, does not provide enough absorbable EFAs. Mammalian pancreatic lipase hydrolyzes the ester linkages at the sn-1 and 3-positions with a preference for MCFAs over LCFAs [13, 14]. The resulting sn-2-monoacylglycerols (2-MGs) are better absorbable forms of FAs through the intestinal mucosa [5]. Therefore, LCFAs located at sn-2position of MLM type SLs are expected to be well absorbed, prompting the idea to “build-in EFAs into a MCT molecule” [5–7]. This idea was extended to an alternative concept: to use MLM-type STGs as effective carriers of LCFAs, especially bio-active FAs such as polyunsaturated FAs (PUFAs). This concept might be related to drug delivery technology, including so-called “DG prodrugs”, which are 1,3-diacyl-sn-glycerols with various drugs attached to glycerol’s sn-2 position [15–18]. 2.2 Synthesis of MLM-type STGs from Natural Oils and Fats Figure 1 shows a typical synthetic scheme for MLM-type STGs. The method is acyl exchange of oils with excess MCFA (for acidolysis) [19–21] or its ethylester (MCFAEt, for interesterification) [22]. The strategy is to substitute the FA residues specifically at the sn-1- and 3-positions of the oils with desired residues of a 1,3-specific lipase (especially of fungal origin such as Rhizomucor miehei and Rhizopus delemar), leaving the FA residues at the sn-2-position unchanged.

Enzymatic Synthesis of Structured Lipids

155

Fig. 1 Synthesis of MLM-type STGs. LCT is reacted with MCFA (acidolysis) or with MCFAEt (interesterification) in the presence of 1,3-specific lipase. TGs are schematically represented. “M” and “L” indicate MCFA and LCFA residues, respectively. Possible by-products (and impurities) are shown as small formulas

Intensive studies on acidolysis reactions were performed by Shimada et al. [19–21]. R. delemar lipase immobilized on a ceramic carrier was employed for acidolysis of various oils with caprylic acid (CA). Hydrolysis of the substrate is a side reaction that should be minimized. They found that, following recovery and reuse, the enzyme, which was used first in the presence of a certain amount of water for activation, did not hydrolyze triacylglycerol (TG) further in the subsequent reactions. With this “activated enzyme”, modification of oils containing EFAs [19], g-linoleic acid [20], arachidonic acid [21], eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) [20, 21] were successfully achieved without formation of partial glycerides. In addition, to enhance the incorporation of CA, the reaction was repeated for three cycles [20].After each cycle, the TG fraction was recovered and reacted with fresh CA in the subsequent cycle. Consequently, the FA residues at sn-1 and 3 positions could be replaced completely, while the ones at the sn-2 position remained unchanged. The final products were quite pure with respect to heterogeneity of constitutive molecular species (they comprised only a few TG species). Therefore, depending on the choice of the oils as starting material, various kinds of MLMtype SLs can be obtained by 1,3-specific-lipase-mediated acidolysis. 2.3 Conversion of Oils Containing Fungal Lipase-Resistant FAs In spite of the versatility of the lipase-catalyzed acidolysis, DHA-containing oils such as tuna oil are exceptional [20, 21]. Since fungal lipases scarcely act on DHA residues of TG [23, 24], the DHA residues at the sn-1 and 3-positions of the starting material remained [20]. We also pointed out a similar problem in acidolysis of a single cell oil rich in DHA and docosapentaenoic acid (DPA), both of which were resistant to the common fungal lipases [25]. When immobilized Rhizomucor miehei lipase (Lipozyme RM-IM) was used, the degree of acidolysis was very low (only 23%), leaving a large amount of DHA and DPA residues unexchanged. Irimescu et al. reported an alternative strategy for synthesis of DHA-containing STGs [26]. The new method includes 1,3-position-specific ethanolysis

156

Y. Iwasaki · T. Yamane

Fig. 2 Confirmation of the purity of DHA-containing STG prepared from bonito oil. Gas chromatograms of bonito oil TG a, and STG derived from it b are shown

of DHA-containing TG by immobilized Candida antarctica lipase B (Novozym) in ethanol to form DHA-containing 2-MG, followed by reesterification of 2-MG by Lipozyme RM-IM. This strategy enabled the effective preparation of DHAcontaining MLM-type STGs from bonito oil (Fig. 2). The Novozym-catalyzed ethanolysis is mentioned in more detail in the later part of this review. 2.4 Pure MLM-type STG Containing EPA In the authors’ laboratory, 1,3-dicapryloyl-2-eicosapentaenoylglycerol (CEC) was synthesized from glycerol, CA or its ethylester (CAEt), and EPA or its ethylester (EPAEt).

Enzymatic Synthesis of Structured Lipids

157

Fig. 3 Three different synthetic routes for pure MLM-type STGs. a COhC is prepared by Lipozyme RM IM-mediated esterification of glycerol with CA. Then COhC is chemically esterified with EPA. b EPAEt is converted to EPA by hydrolysis. Then, EEE is prepared by Novozym-catalyzed esterification of glycerol with EPA. Afterwards, EEE is transesterified with CAEt by lipozyme RM IM. c DDD is prepared by Novozym-catalyzed esterification of glycerol with DHA. Then DDD is converted to OhDOh by Novozym-mediated ethanolysis. Finally, OhDOh is reesterified with CAEt by Lipozyme RM IM

158

Y. Iwasaki · T. Yamane

Fig. 4 Confirmation of the purity of CEC synthesized by three-step enzymatic reaction. a High-temperature gas chromatography showing the purity of the product in the “glyceride fraction”. Note that main impurities are FAs or FAEts, but not glycerides other than the desired product. b (insert): Silver ion HPLC proving the absence of the positional isomer, 1,2-dicapryroyl-3-eicosapentaenoylglycerol (CCE) [28]

One method included preparation of 1,3-dicapryloylglycerol (COhC) from glycerol and CA by Lipozyme RM-IM-catalyzed esterification, followed by acylation with EPA at sn-2-position (Fig. 3a) [27]. Since the synthesis of the COhC is a condensation reaction, it is important to remove water generated during the reaction course. The reaction proceeded successfully by removing the water under vacuum or nitrogen blowing. The targeted COhC could be obtained with more than 80% purity even from a stoichiometric mixture of the substrates (CA: glycerol=2:1). After purification by a silica gel column, the COhC was chemically reacted with EPA in the presence of dicyclohexylcarbodiimide as a condensation agent and dimethyaminopyridine as a catalyst, yielding the desired CEC with 90% purity.

Enzymatic Synthesis of Structured Lipids

159

Alternatively, we tried a three-step method which did not require isolation of intermediates (Fig. 3b) [28]. The first step was hydrolysis of EPAEt to free EPA by Novozym under controlled vacuum, in order to remove the formed ethanol and promote faster hydrolysis. The residual water was removed completely at the end of the reaction. In the second step, EPA was esterified with the stoichiometric amount of glycerol added directly to the reaction vessel. A high yield of 1,2,3-trieicosapentaenoylglycerol (EEE) (over 90%) was obtained by removal of water under vacuum. After the reaction, the non-regio-specific enzyme (Novozym) was removed by filtration. The final step was transesterification of EEE with an excess amount of CAEt by 1,3-specific Lipozyme RM-IM, resulting in the desired CEC. The excess of CAEt and the by-products (CA, EPAEt and EPA) can be removed fractionally by molecular distillation from the glyceride fraction, which contained more than 90% CEC (Fig. 4). EPAEt and EPA can be reused in the first step and CAEt in the third step. The main advantages of this procedure are: 1) no organic solvent is employed; 2) isolation and purification of intermediates is not necessary, and; 3) reuse of the excess of reactants (CAEt) and main by-products (EPAEt and EPA) is feasible. When CA was used instead of CAEt at the final step, the reaction rate was slower than in the case of CAEt. However, since CAEt is ten times more expensive than CA, it is better to use CA for economical reasons. This problem was overcome by enzymatically synthesizing CAEt from CA and ethanol in a separate reaction, and using it for the final reaction step [29]. 2.5 Pure MLM-type STG Containing DHA The general scheme for the enzymatic synthesis of pure MLM-type STGs includes: 1) preparation of mono-acid TG (having the same acyl groups at all of the positions of the glycerol backbone, such as EEE), and then 2) replacement of FA residues, specifically at the 1,3-positions, leaving the one at the 2-position unchanged (Fig. 3b). However, this strategy cannot be applied for the synthesis of MLM-type STG with DHA residue at the 2-position, because DHA residue is resistant against fungal lipase. To solve this problem, an alternative strategy using Novozym-catalyzed ethanolysis was developed. Novozym was believed to act on DHA residues of TG in a position-non-specific manner. However, when subjected to ethanolysis of tridocosahexaenoylglycerol (DDD) in ethanol, Novozym reacted only at the 1,3-position, giving 2monodocosahexaenoylglycerol (OhDOh) and DHA ethyl ester (Fig. 3c) [26]. In other words, Novozym changed its positional specificity in ethanol [30]. Using this “ethanol effect”, DDD was converted to OhDOh, which was then reesterified with CAEt by Lipozyme RM-IM to afford 1,3-dicapryroyl-2-docosahexaenoyl-sn-glycerol (CDC) (Fig. 3c) [31].

160

Y. Iwasaki · T. Yamane

Fig. 5 Modes of action of phospholipases on phosphatidylcholine

3 Enzymatic Synthesis of Structured Phospholipids 3.1 Fundamentals of Phospholipids Phospholipids (PLs) are simply lipids which contain phosphorus. Naturally-occurring PLs are ubiquitous in all organisms. One of the most important characteristics of PLs is their amphiphilicity, which arises from the hydrophobic acyl or alkyl groups and hydrophilic polar head groups. Their amphiphilic nature makes it possible for them to form several aggregates with water, such as micelles, reverse micelles and bilayer vesicles. Physiologically, PLs are major components of bio-membranes. Physical properties, biocompatibility, and nutritional functions of PLs make them useful in industrial fields such as food, cosmetics, and pharmaceuticals. PLs can be used as emulsifiers, components of cosmetics, medical formulations, and for liposome preparations. PLases are enzymes which hydrolyze PLs. Based on their mode of action, PLases are classified into several classes; A1, A2, B, C and D (PLA1, PLA2, PLB, PLC and PLD). Figure 5 shows the mode of action of each PLase on phosphatidylcholine (PC). Enzymatic modification of PLs are divided into two categories: acyl group modification and head group modification. 3.2 Modification of Acyl Groups Natural PLs are heterogeneous mixtures of various molecular species with different FAs and polar head groups, but it is often necessary to use pure PL with

Enzymatic Synthesis of Structured Lipids

161

a defined chemical structure. In liposome technology, for example, such pure PLs should be used to control the properties of the vesicles. PC with identical FA residues at both sn-1 and 2-positions (symmetric PC) can be synthesized chemically (Fig. 6a). Natural PC is deacylated chemically to obtain GPC, followed by acylation with an appropriate FA-derivative such as FA-chloride or FA-anhydride [32]. Preparation of PC by the above method often handles GPC as an adduct with cadmium chloride to facilitate its recrystallization and to improve its dispersity in solvent [33]. However, use of such a toxic heavy metal should be avoided. A Japanese company developed and industrialized an alternative production process where GPC absorbed on an inorganic support (such as celite) is used instead of the cadmium adduct [34]. PC species with different FA residues at the sn-1- and 2-positions (asymmetric PC) are prepared by hydrolysis of the corresponding symmetric PC by PLA2, and further chemical acylation at the sn-2-position [35] (Fig. 6a). The chemical and chemoenzymatic methods described above are very useful.Yet there might be several drawbacks in the chemical methods: they require an activated acyl donor such as FA anhydride or chloride, or condensation agents such as carbodiimides; depending on the PL species, certain functional groups, such as amino groups of phosphatidylethanolamine (PE) or phosphatidylserine (PS), or hydroxyl groups of phosphatidylinositol (PI) or phosphatidylglycerol (PG), should be protected prior to deacylation/ reacylation and deprotected thereafter [36, 37]. To simplify the synthesis, several attempts were made to enzymatically introduce particular FAs into PLs using lipase and PLA2. 1,3-specific-lipase-catalyzed transesterification is employed to introduce particular fatty acids into the sn-1 position of PLs (Fig. 6b). With these enzymes, transesterification occurs exclusively at the sn-1 position [38, 39].A major problem in the lipase-catalyzed reaction is formation of deacylated PLs (such as lysoPLs) via hydrolysis of the substrate. Svenson et al. performed the transesterification of PC and heptadecanoic acid with Rhizopus arrhizus lipase [39]. By optimizing the water activity in the reaction system, almost 100% incorporation in the sn-1 position with 60% of PC recovery was achieved. Introduction of the FA into the sn-2-position is done with PLA2. Since PLA2 scarcely catalyzes transesterification, unlike lipases, the condensation reaction (the reverse reaction of hydrolysis) of the particular FA and lysophospholipid (lysoPL) is used (Fig. 6b). This reverse reaction was first reported by Pernas et al. with a very low yield (6%) [40].Afterwards, several researchers improved the yield. So far, the highest yield achieved, reported by Egger et al. [41] and Hosokawa et al. [42], was approximately 60%. PUFAs such as EPA and DHA are known to be bio-active compounds. Since PUFAs are very unstable, enzymatic conversion of PUFAs under mild conditions are preferable.A good example was reported by Hosokawa et al. [42]. PLs containing PUFA at the sn-1 position were prepared by Lipozyme RM-IMcatalyzed acidolysis of PC with a mixture of EPA and DHA. Use of a combination of water and propylene glycol as a water mimic gave 80% of the

Fig. 6 Preparation of PLs containing particular FAs at specific positions. a Chemical methods, b enzymatic methods

B

A

162 Y. Iwasaki · T. Yamane

Enzymatic Synthesis of Structured Lipids

163

theoretical maximum incorporation with a PC recovery of 80%. In addition, PC containing PUFA at the sn-2-position was prepared by PLA2-mediated condensation of lysoPC and PUFA in glycerol as a reaction media. Addition of a small amount of formamide to the reaction mixture gave the best results, with a 60% yield. 3.3 Modification of Head Groups 3.3.1 Transphosphatidylation by PLD Besides hydrolysis, PLD catalyzes a reaction in which polar head groups of PLs are replaced by other hydroxyl compounds [43, 44]. This reaction (called transphosphatidylation) can be used to prepare a PL with a particular polar head group (Fig. 7a). Various PLs can be synthesized from naturally abundant PLs such as PC or lecithin and corresponding hydroxyl compounds. Important parameters for evaluating a reaction are selectivity and yield, defined as: Selectivity (%) = [PX]/([PA] + [PX]) ¥ 100 Yield (%) = [PX]/[PC]0 ¥ 100 where PX is the desired product, PA is phosphatidic acid (in this case a byproduct caused by hydrolysis of the substrate), and [ ] denotes concentration of each component. The reaction is typically carried out in a bi-phasic system consisting of a water-immiscible organic solvent (for example diethylether or ethylacetate) containing lipids and an aqueous solution of enzyme and hydroxyl compounds (in other words, acceptor compounds such as ethanolamine, glycerol or serine). The transphosphatidylation takes place under the optimized conditions even though the system is so water-rich, although PLD is intrinsically a hydrolytic enzyme. Therefore, there is no need to control water content in the reaction system or dehydrate the enzyme prior to the reaction, unlike for lipase-catalyzed transesterification reactions. An advantage of the bi-phasic system is that the desired product is soluble in the organic phase and can be separated easily from the aqueous phase by a simple phase separation. It is also possible to introduce particular hydroxyl compounds into the polar head group by a chemical method where the hydroxyl compound and PA is condensed in the presence of 2, 4, 6-triisopropylbenzenesufonyl chloride as a condensing agent [45] (Fig. 7b). However, the most convenient way to prepare PA from lecithin as a starting material is, in fact, hydrolysis by PLD. Therefore, the enzymatic transphosphatidylation is preferable over the chemical reaction, unless other special purposes are intended (such as synthesis of D-PL, which is the stereoisomer of naturally occurring L-PL, using an appropriate chiral build-

164

Y. Iwasaki · T. Yamane

A

B

Fig. 7 Introduction of particular hydroxyl compounds into the polar heads. a Transphosphatidylation catalyzed by PLD. PC is converted into phosphatidyl X in the presence of hydroxyl compounds (X-OH). When X is H, the reaction is hydrolysis. b Chemical method. Phosphatidic acid and hydroxyl compound (X-OH) is condensed using 2, 4, 6-triisopropylbenzenesulphonyl chloride (TPS) as a condensation agent

ing block [46]). To achieve high selectivity a reasonably high concentration of the acceptor is necessary. This might be a drawback of the enzymatic method as compared to the chemical one, especially when the acceptor compound is expensive. However, the use of the bi-phasic system might overcome this problem, since the separated aqueous phase (containing the enzyme and the residual acceptor compound) can be reused several times in other batches of organic phase containing fresh lipid substrate [47]. 3.3.2 Syntheses of Natural Type PLs Juneja et al. synthesized several natural PLs such as PG [48, 49], PE [50] or PS [51, 52] from lecithin or PC using PLD-catalyzed transphosphatidylation. Under the optimized conditions, the syntheses were very successful with yields of almost 100%. Transphosphatidylation with L- and D-serine gave phosphatidyl-L- and D-serine, respectively. Interestingly, bacterial PLD showed a twofold higher reaction rate with D-serine than with L-serine, while cabbage PLD reacted only with L-serine. In addition, the reaction of egg lecithin containing 75% PC and 25% PE with choline gave PC with almost 100% purity [53]. Generally, volumetric productivity as high as possible is favorable. When a

Enzymatic Synthesis of Structured Lipids

165

higher concentration of PC was used, the efficiency of the reaction dropped due to feedback inhibition by the released choline. Removal of the choline by coexisting choline oxidase and catalase enabled the reaction with high concentration of the starting substrate [51]. As shown above, most natural-type SPLs can be synthesized enzymatically. In addition, starting from PC with defined FA residues as described in Sect. 3.2, most of natural PLs with completely defined structures can be obtained. An exception is the synthesis of PI from PC and inositol with PLD, which has not been successful yet. 3.3.3 Synthesis of PLs Containing Bio-Active Compounds The PLD-mediated transphosphatidylation can be extended to syntheses of PLs with unnatural polar heads. Until now, a large variety of artificial lipids have been synthesized. The ability of PLD to transphosphatidylate a large number of compounds can be applied for syntheses of novel PLs containing functional polar head groups. Since the phosphatidyl moiety is itself bio-compatible, it is considered to be a non-toxic carrier of bio-active compounds. The purpose of the syntheses of such functional PLs is to add physical and biological properties of PLs to the acceptor compounds. Examples intended for use in food, cosmetics and medical applications are described below. Nagao et al. synthesized phosphatidylascorbic acid (PAsA) (Fig. 8, compound 1) [54]. It was revealed that PAsA suppressed oxidation of PC in multilamelar liposomes [55]. The result indicated that the ascorbic acid moiety of PAsA was localized on the water-lipid surface of the liposomes, enhancing the effective concentration of the ascorbic acid moiety for scavenging aqueous peroxy radicals. Similarly, Koga et al. introduced 2,5,7,8-tetramethyl-6-hydroxy2-(hydroxyethyl)-chroman into the polar head, resulting in phosphatidylchromanol (PCh) (Fig. 8, compound 2) [56]. The chroman ring, which is a portion of a-tocopherol (vitamin E), is involved in the anti-oxidant activity of a-tocopherol. PCh suppressed autooxidation of lard more effectively than the original compound. The explanation for this phenomenon is that PCh formed reverse micelles in the oil, and trapped the residual water that dissolved a trace of metal ions such as iron that initiated the oxidation [57]. Both arbutin and kojic acid prevent overproduction of melanin in epidermal cells. Arbutin is a competitive inhibitor of tyrosinase, whereas kojic acid inhibits the enzyme as a chelator and an antioxidant. Since these compounds are water-soluble, their instability was a problem for their use in cosmetic products. Takami et al. converted these compounds into PLs in order to improve their physical characteristics (Fig. 8, compound 3 and 4) [58]. The resultant phosphatidylarbutin and phosphatidylkojic acid exhibited similar inhibitory activity in vitro towards tyrosinase as the original compound. The phosphatidyl moiety has a high affinity to cell membrane, so that phosphatidyl derivatives can easily penetrate into cells. Shuto et al. synthesized

Fig. 8 Structure of several functional phospholipids. Phosphatidyl derivatives of ascorbic acid (1), 2,5,7,8-tetramethyl-6hydroxy-2-(hydroxyethyl)chroman (2), arbutin (3), kojic acid (4), 5-fluorouridine (5) and N-acetylneuraminic acid (connected with octanediol linker) (6) are shown

166 Y. Iwasaki · T. Yamane

Enzymatic Synthesis of Structured Lipids

167

various PL-derivatives of nucleoside analogues, including anti-cancer agents such as 5-fluorouridine (Fig. 8, compound 5) [59]. Some of them showed superior anti-tumor activities to the parent compounds [60]. A sialic acid-containing PL was synthesized from N-acetylneuraminic acid (NeuAc) and PC by a combination of chemical and enzymatic methods (Fig. 8, compound 6) [61]. Derivatives of NeuAC are expected to be potent anti-viral agents. The inhibitory effect of the resultant liposomes for rotavirus infection was 103–104 fold higher than that for NeuAc. It has been speculated that its enhanced anti-viral activity is due to the bilayers with multivalent NeuAc displayed on the surface that the synthesized lipid forms, which then interact with the virus in a multivalent manner. Besides these, phosphatidyl derivatives of a peptide-based inhibitor (against fibronectin adhesion to integrin) [62] and dihydroxyacetone (a tanning agent for cosmetics) [63] are also interesting examples. 3.3.4 Synthesis of Phosphatidylserine in a Non-Organic Solvent System The PLD-mediated transphosphatidylation is a very versatile reaction by which various natural and unnatural PLs can be synthesized.As described previously, transphosphatidylation is generally carried out in a bi-phasic system consisting of a water-immiscible organic solvent phase (for example diethyl ether or ethyl acetate) containing lecithin, and an aqueous buffer phase containing the enzyme and an acceptor compound (Fig. 9a). Considering PLs as food for human consumption, however, the use of such toxic organic solvents should be avoided. In this aspect, we have recently developed a novel reaction system for transphosphatidylation without using toxic organic solvents [64]. The target product was PS, which has attracted much attention as it shows therapeutic effects for several memory-related disorders [65–68]. An attempt to react soybean lecithin (simply dispersed in an aqueous buffer) with an aqueous solution of L-serine and PLD was unsuccessful, giving only 20% of PS (Fig. 9b). By contrast, a suspension of lecithin, previously adsorbed on fine powders such as silica, was effectively converted into PS in an aqueous solution of L-serine and PLD (Fig. 9c). After screening of various powders for the lecithin adsorbent, calcium sulfate was revealed to be the best with respect to lecithin conversion. In addition, the use of calcium sulfate was more advantageous as it did not require prior adsorption of lecithin (in other words the reaction proceeded effectively just by adding the powder into an aqueous mixture of lecithin, L-serine and PLD). Using this “aqueous suspension system” with calcium sulfate, up to 180 mg/mL lecithin was completely converted, resulting in more than 80% PS in 24 h. The synthesized PS was easily recovered from the powder by extracting with a mixture of n-hexane, ethanol and diluted HCl. The PS content in the final product was approximately 90%, and the overall yield of PS from the soybean lecithin was calculated to be approximately 89% (Fig. 10).

Fig. 9 Comparison of several reaction systems for PS synthesis [64]. PS synthesis by PLD-mediated transphosphatidylation was performed in three different reaction systems. a ethyl acetate-buffer bi-phasic system; b aqueous system where lecithin was simply dispersed in the aqueous buffer, and; c aqueous suspension system where silica-adsorbed lecithin was suspended in the aqueous buffer. Closed circles: PS; open circles: PC; open squares: PA; open triangles: PE

168 Y. Iwasaki · T. Yamane

Enzymatic Synthesis of Structured Lipids

169

Fig. 10 Confirmation of the purity of PS prepared from soybean lecithin by PLD in aqueous suspension system [64]. Thin layer chromatogram of initial soybean lecithin a and the synthesized PS derived from it b are shown. “O” and “F” represent the origin and the front of the development, respectively

4 Conclusions Lipases and PLases have become powerful tools for enzymatic synthesis of STGs and SPLs. Many reactions described here are, in principle, very simple and well-established. Reducing production costs is very important for practical implementation of those methods. Production costs are influenced by many factors, such as choice of the substrates, catalysts, volumetric productivity, downstream purification of products, and formation of by-products.Among them, costs of the enzymes are usually predominant. Immobilized enzyme would facilitate continuous operation or repeated use of the catalysts, thereby lowering the costs. Effective production of enzymes themselves by recombinant DNA techniques is also promising. In addition, obtaining novel enzymes with enhanced stability by screening or by mutagenesis techniques is worthwhile. More importantly, what to synthesize should be considered as well as how to synthesize, because these reactions would surely become beneficial when they are applied to syntheses of useful compounds for certain purposes. Collaborative works among people who study nutritional chemistry, pharmacology and synthetic chemistry might help in the design of novel useful target compounds.

170

Y. Iwasaki · T. Yamane

References 1. Akoh CC (1995) INFORM (International News on Fats, Oils and Related Materials) 6:1055 2. Finley JW, Klemann LP, Leveille GA, Otterburn MS, Walchak CG (1994) J Agric Food Chem 42:495 3. Smith RE, Finley JW, Leveille GA (1994) J Agric Food Chem 42:432 4. Ranhorta GS, Gelroth JA, Glaser BK (1994) Cereal Chem 71:159 5. Jandacek RJ, Whiteside JA, Holcombe BN, Volpenhein RA, Taulbee JD (1987) Am J Clin Nutr 45:940 6. Ikeda I, Tomari Y, Sugano M, Watanabe S, Nagata J (1991) Lipids 26:369 7. Christensen MS, Høy CE, Becker CC, Redgrave TG (1995) Am J Clin Nutr 61:56 8. Quinlan P, Moore S (1993) INFORM (International News on Fats, Oils and Related Materials) 4:580 9. Macrae AR (1983) J Am Oil Chem Soc 62:421 10. Kaunitz H, Salnetz CA, Johnson RE, Babayan VK, Barsky G (1958) J Am Oil Chem Soc 35:10 11. Bach AC, Babayan VK (1962) Am J Clin Nutr 36:950 12. Babayan VK (1987) Lipids 22:417 13. Bottino NR, Vandenberg GA, Reiser R (1967) Lipids 2:489 14. Yang LY, Kuksis A, Myher JJ (1990) J Lipid Res 31:137 15. Kumar K, Billimoria JD (1978) J Pharm Pharmacol 30:754 16. Deverre JR, Loiseau P, Puisieux F, Gayral P, Letourneux Y, Couvreur P, Benoit JP (1992) Arzneimittel-Forsch (Drug Res) 42:1153 17. Delie F, Couvreur P, Nisato D, Michael JB, Puisieux F, Letourneux Y (1994) Pharm Res 11:1082 18. Kihel LE, Bourass J, Richomme P, Petit JY, Letourneux Y (1996) Arzneimittel-Forsch (Drug Res) 46:1040 19. Shimada Y, Sugiura A, Nakano H, Yokota T, Nagao T, Komemushi S, Tominaga Y (1996) J Am Oil Chem Soc 73:1415 20. Shimada Y, Sugiura A, Nakano H, Nagao T, Suenaga M, Nakai S, Tominaga Y (1997) J Ferment Bioeng 83:321 21. Shimada Y, Sugiura A, Maruyama K, Nagao T, Nakamura S, Nakano H, Tominaga Y (1996) J Ferment Bioeng 81:299 22. Huang KH, Akoh CC (1996) J Am Oil Chem Soc 73:245 23. Langholz P, Andersen P, Forskov T, Schmidtsdorff W (1989) J Am Oil Chem Soc 66: 1120 24. Pedersen SB, Hølmer G (1995) J Am Oil Chem Soc 72:239 25. Iwasaki Y, Han JJ, Narita M, Rosu R, Yamane T (1999) J Am Oil Chem Soc 76:563 26. Irimescu R, Furihata K, Hata K, Iwasaki Y, Yamane T (2001) J Am Oil Chem Soc 78:743 27. Rosu R, Yasui M, Iwasaki Y Yamane T (1999) J Am Oil Chem Soc 76:839 28. Irimescu R, Yasui M, Iwasaki Y, Yamane T (2000) J Am Oil Chem Soc 77:501 29. Irimescu R, Hata K, Iwasaki Y, Yamane T (2001) J Am Oil Chem Soc 78:65 30. Irimescu R, Iwasaki Y, Hou CT (2002) J Am Oil Chem Soc 79: 879 31. Irimescu R, Furihata K, Hata K, Iwasaki Y, Yamane T (2001) J Am Oil Chem Soc 78:285 32. Gupta CM, Radhakrishnan R, Khorana HG (1977) P Natl Acad Sci USA 74:4315 33. Chadha JS (1970) Chem Phys Lipids 4:104 34. US patent 4,690,784 35. Mason JT, Broccoli AV, Hung CH (1981) Anal Biochem 113:96 36. Aneja R, Chadha JS, Robels EC, Daal RV (1969) Biochim Biophys Acta 187:439

Enzymatic Synthesis of Structured Lipids 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68.

171

Japan patent S63–157993 Svensson I, Adlercreutz P, Mattiasson B (1990) Appl Microbiol Biot 33:255 Svensson I, Adlercreutz P, Mattiasson B (1992) J Am Oil Chem Soc 69:986 Pernas P, Olivier JL, Legoy MD, Bereziat G (1990) Biochem Bioph Res Co 168:644 Egger D, Wehtje E, Adlercreutz P (1997) Biochim Biophys Acta 1343:76 Hosokawa M, Takahashi K, Kikuchi Y, Hatano M (1995) J Am Oil Chem Soc 72:1287 Yang SF, Freer S, Benson AA (1967) J Biol Chem 242:477 Dawson RMC (1967) Biochem J 102:205 Aneja R, Chadha JS (1971) Biochim Biophys Acta 248:455 Kumar R, Gardner MF, Richman DD, Hostetler KY (1992) J Biol Chem 267:20288 Nakajima J, Nakashima T, Shima Y, Fukuda H, Yamane T (1994) Biotechnol Bioeng 44:1193 Juneja LR, Hibi N, Inagaki N, Yamane T, Shimizu S (1987) Enzyme Microb Tech 9:350 Juneja LR, Hibi N, Yamane T, Shimizu S (1987) Appl Microbiol Biot 27:146 Juneja LR, Kazuoka T, Yamane T, Shimizu S (1988) Biochim Biophys Acta 960:334 Juneja LR, Kazuoka T, Goto N, Yamane T, Shimizu S (1989) Biochim Biophys Acta 1003:277 Juneja LR, Taniguchi E, Shimizu S, Yamane T (1992) J Ferment Bioeng 73:357 Juneja LR, Yamane T, Shimizu S (1989) J Am Oil Chem Soc 66:714 Nagao A, Ishida N, Terao J (1991) Lipids 26:390 Nagao A, Terao J (1990) Biochem Bioph Res Co 172:385 Koga T, Nagao A, Terao J, Sawada K, Mukai K (1994) Lipids 29:83 Koga T, Terao J (1994) J Agric Food Chem 42:1291 Takami M, Hidaka N, Miki S, Suzuki Y (1994) Biosci Biotech Biochem 58:1716 Shuto S, Itoh H, Ueda S, Imamura S, Fukawa K, Matsuda A, Ueda T (1987) Tetrahedron Lett 28:199 Shuto S, Itoh H, Ueda S, Imamura S, Fukawa K, Tsujino M, Matsuda A, Ueda T (1988) Chem Pharm Bull 36:209 Koketsu M, Nitoda T, Sugino H, Juneja LR, Kim M,Yamamoto T,Abe N, Kajimoto T,Wong CH (1997) J Med Chem 40:3332 Wang P, Schuster M, Wang YF, Wong CH (1993) J Am Chem Soc 115:10487 Takami M, Suzuki Y (1994) Biosci Biotech Biochem 58:2136 Iwasaki Y, Mizumoto Y, Okada T,Yamamoto T, Tsutsumi K,Yamane (2003) J Am Oil Chem Soc 80:653 Delwaide PJ, Gyselynck-Mambourg AM, Hurlet A, Ylieff M (1986) Acta Neurol Scand 73:136 Crook TH, Tinklenberg J,Yesavage J, Petrie W, Nunzi MG, Massari DC (1991) Neurology 41:644 Nunzi MG, Milan F, Guidolin D, Toffano G (1987) Neurobiol Aging 8:501 Monteleone P, Beinat L, Tanzillo C, Maj M, Kemali D (1990) Neuroendocrinology 52:243

Received: October 2003

Adv Biochem Engin/Biotechnol (2004) 90: 173 –198 DOI 10.1007/b94197 © Springer-Verlag Berlin Heidelberg 2004

Bioprocess Monitoring Using Near-Infrared Spectroscopy Ken-ichiro Suehara (✉) · Takuo Yano Department of Information Machines and Interfaces, Faculty of Information Sciences, Hiroshima City University, 3-4-1 Ozukahigashi, Asaminami-ku, Hiroshima 731-3194, Japan [email protected], [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

2 2.1 2.2 2.3 2.4

Procedure for Near-Infrared Spectroscopy NIR Spectrum . . . . . . . . . . . . . . . Calibration Equation for NIR . . . . . . . Second Derivative for the Spectrum . . . Validation of Calibration Equation . . . .

. . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

174 175 176 176 177

3 Application of NIR to a Liquid Sample . . 3.1 Rice Vinegar Fermentation . . . . . . . . . 3.2 L-glutamic Acid Fermentation . . . . . . . 3.3 Lactic Acid Fermentation . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

177 178 184 185

4 Application of NIR to Solid Samples . . . . . . . . . . . . . . . . . . . . . . . 185 4.1 Compost Fermentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 4.2 Koji Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 5 5.1 5.2 5.3

Other Applications . . . Mushroom Cultivation . Medical Products . . . . Glycolipid Fermentation

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

193 194 194 195

6

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

Abstract Near-infrared spectroscopy (NIR) is a nondestructive analytical technique that has been used for simultaneous prediction of the concentrations of several substrates, products and constructs in mixtures sampled from fermentation processes. In this chapter, we discuss applications of NIR for the monitoring of bioprocesses involving rice vinegar, compost, glycolipid, L-glutamic acid, lactic acid fermentation, mushroom cultivation, and Koji production. This includes detailed discussion of applications of NIR to process management of rice vinegar fermentation and compost fermentation. In the present study, absorbance at wavelengths between 400 and 2500 nm was measured at 2 nm intervals. To obtain calibration equations, multiple linear regression (MLR) was performed on NIR spectral data and conventional analysis values of a calibration sample set. To validate these calibration equations, they were used to calculate concentrations of a prediction sample set, which were then compared with concentrations measured by conventional methods. There was excellent agreement between the results of the conventional method and those of the NIR method,

174

K. Suehara · T. Yano

when both were used to analyze culture broth of rice vinegar fermentation and solid-state fermented compost. These results indicate that NIR is a useful method for monitoring and control of bioprocesses. Keywords Near-infrared spectroscopy (NIR) · Nondestructive measurement · Rice vinegar · Compost · Process management

1 Introduction To operate a bioprocess with high efficiency, it is very important that nutrient concentration and environmental factors are carefully controlled. In addition, concentrations of products that inhibit growth of microorganisms should be measured and kept at low levels. Ideally, a monitoring system for bioprocess management should be easy, fast, aseptic and non-destructive. Several monitoring systems for measuring concentrations of nutrients and products have been developed, including gas chromatography, liquid chromatography and biosensors. Near-infrared spectroscopy (NIR) is also used to measure concentrations of important substances in bioprocesses. NIR is widely used for rapid and nondestructive analysis in industries such as agriculture, food, pharmaceuticals, textiles, cosmetics and polymer production [1]. NIR has been used to measure concentrations of constituents in bioprocesses involving wine, beer, sake, miso, soy sauce, vinegar, alcohol fermentation, lactic acid fermentation, mushroom production, animal cell culture, compost fermentation and enzymatic saccharification [2–11]. NIR has several advantages in such applications: several components can be assayed simultaneously; a sample can be analyzed with a photometer without pretreatment; both dry and wet samples can be analyzed. Therefore, both aqueous and solidstate fermented samples can be analyzed. NIR is a high-precision assay technique, and on-line measurement is available using fiber optics. In this chapter, use of NIR for simultaneous measurement of concentrations of constituents in rice vinegar fermentation broth is discussed, as an example of the application of NIR to the monitoring of liquid fermentation [8]. Application of NIR to the monitoring of compost fermentation is also discussed, as an example of the monitoring of solid-state fermentation processes [9–11]. Several other applications to bioprocess monitoring are also outlined.

2 Procedure for Near-Infrared Spectroscopy A schematic of the NIR procedure is shown in Fig. 1. First, large number of samples are prepared, because the NIR method involves use of statistics on a computer. To obtain a calibration equation, relationships between absorption

Bioprocess Monitoring Using Near-Infrared Spectroscopy

175

Fig. 1 Schematic of near-infrared spectroscopy procedure

value and conventional analysis values of the sample are examined using multiple linear regression (MLR) analysis. The calibration equation is validated before it is used. Sometimes, a differential spectrum may be used in NIR analysis to obtain the calibration equation. Details of the NIR analysis procedure are described below. 2.1 NIR Spectrum NIR spectra are produced when light is absorbed by organic molecule bonds such as C–N, C–H, N–H, C=O, S–H and O–H. Most of the absorption bands are caused by overtones or combinations of overtones of the bonds originating in the infrared region of the spectrum. The overtones are usually 10–1000 times weaker than the original bands in the infrared region.When near-infrared light irradiates a sample, there is little damage to the sample, because water in the sample absorbs very little near-infrared light, so very little of the near-infrared light is converted into heat. However, analysis of the NIR spectrum to obtain a calibration equation is very difficult. Assignment of absorption for NIR spec-

176

K. Suehara · T. Yano

tra is not straightforward because there can be a great deal of accumulation of weak absorption in a single spectrum. Therefore, statistical methods are used to examine the relationship between spectral data values and concentration values obtained by conventional analytical methods. 2.2 Calibration Equation for NIR In the mathematical treatment of raw spectra, second derivatives must be obtained for the calibration set of spectra [7, 12]. Multiple linear regression (MLR) using least-squares methods [13] is performed on the NIR spectral data and the measurement values obtained by conventional analytical methods, Cact, for the calibration sample set, in order to obtain the following calibration equation: Cpre = a0 + Â ai · Ai .

(1)

Here, Cpre is the predicted value calculated from the NIR spectral data, and a is the regression coefficient. The subscript i (i=1, 2, …, m) represents the wavelength used in the regression analysis. When the calculation is performed using the raw optical data, A is represented by log(1/Tr) or log(1/Ref). Tr and Ref are the NIR spectral values of transmittance and reflectance of the sample, respectively. 2.3 Second Derivative for the Spectrum When the calculation is performed using the second derivative optical data, A at wavelength j (nm) is expressed as d2log(1/Tr) or d2log(1/Ref), and is represented by the following equation: A = A1 – 2 · A2 + A3 ,

(2)

where

T Â log (1/Tr

j – 1.5 · seg – gap + 2k)

Y/(seg/2 + 1)

T Â log (1/Tr

j – 0.5 · seg + 2k)

T Â log (1/Tr

j + 0.5 · seg + gap + 2k)

seg/2

A1 =

k=0

seg/2

A2 =

k=0

seg/2

A3 =

k=0

Y/(seg/2 + 1) Y/(seg/2 + 1)

(3)

(4)

(5)

In these equations, seg represents the segment size (nm) used to average the values of several data points in the neighborhood of the wavelength, and gap represents the gap size (nm), which is the distance from the nearest data point used in Eq. 4.

Bioprocess Monitoring Using Near-Infrared Spectroscopy

177

For rice vinegar and compost fermentation, 20 and 4 nm were used as the values of seg and gap, respectively. The second derivatization, calibration and validation were performed using the NSAS program (Near-Infrared Spectral Analysis Software; Nireco Co.). 2.4 Validation of Calibration Equation To evaluate the performance of the calibration equations obtained, they are validated using the prediction sample set, which was not used for calibration. The concentrations of a prediction sample set are calculated using the calibration equations, and compared with the concentrations measured by conventional methods. If there is good correspondence between these values, the calibration equation can be used for daily analysis.

3 Application of NIR to a Liquid Sample An advantage of NIR is that it can be used to analyze liquid samples, as shown in Table 1. The constituents of various liquid samples can be analyzed using NIR, including glucose, maltose, maltotriose and fructose in aqueous sugar solution [14]; milk [15]; glucose, citric acid and lactic acid in blood anticoTable 1 Application of NIR for liquid samples

Sample

Constituent

Food

Milk Beer Wine Sake Fruit juice Soybean milk Tea Soy sauce

Moisture, lipid, protein, lactose, casein Ethanol Ethanol, sucrose, acidity Ethanol, acidity, amino acid, total sugar Glucose, fructose, sucrose Protein, moisture Caffeine, theanine, amino acid NaCl, total N, ethanol, lactic acid, glutamic acid

Culture broth

Animal cell culture Lactic acid fermentation

Glucose, L-glutamine, lactic acid, ammonia Lactic acid, glucose, biomass

Rice vinegar fermentation

Ethanol, acetic acid, OD (biomass), gluconic acid, lactic acid, and so on

L-glutamic

acid fermentation

Sugar, cell, L-glutamic acid, NH4+

Glycolipid fermentation

Soybean oil, mannosyl erythritol lipid (glycolipid)

178

K. Suehara · T. Yano

agulant and peritoneal dialysis solutions [16, 17]; soybean oil (quality evaluation [oxidization]) [18]; cholinesterase in human blood [19, 20]; and others [21, 22]. Therefore, NIR may be useful for management of cultivation. However, there have been few applications of NIR to constituent analysis of culture broth. Apparently, the high number of constituents in culture broth makes it difficult to analyze using NIR. In this section, we discuss applications of NIR to the analysis of liquid samples from rice vinegar and L-glutamic and lactic acid fermentation, including a detailed description of the procedure for the NIR analysis of liquid samples from rice vinegar fermentation. 3.1 Rice Vinegar Fermentation Yano et al. applied NIR to the management of rice vinegar fermentation [8]. In vinegar fermentation, gas chromatography is used to measure concentrations of ethanol (substrate) and acetic acid (product) in the culture broth. Organic acids such as gluconic and 2-ketoglutaric acids, which affect the taste of the vinegar, are measured by liquid chromatography. The cell density of the culture broth is calculated on the basis of the optical density, measured using a spectrophotometer. For a long time, there has been demand for the development of a simple method for measuring the concentrations of these constituents simultaneously. To determine the potential of NIR for use in the management of rice vinegar fermentation, we used NIR to predict concentrations of ethanol and acetic acid in culture broth sampled from several runs of rice vinegar fermentation. Samples of culture broth from rice vinegar fermentation (number of samples n=63) were removed from a 23 m3 fermentor (working volume, 17 m3) used in a vinegar brewery. The samples were drawn from several runs of fermentation, and contained ethanol and acetic acid at concentrations ranging from 5.7 to 34.8 g/L and from 66.9 to 109 g/L, respectively. A set of 42 samples was used as the calibration sample set, and the remaining 21 samples were used as the prediction sample set. In order to select the wavelength for calibration equations, 40 aqueous samples of an authentic ethanol, 60 aqueous samples of mixed authentic ethanol and acetic acid, and 40 aqueous samples of an authentic acetic acid were also prepared. The culture broth was incubated in a water bath to heat it to the required temperature, and was then placed in a cuvette with a light path length of 2 mm. After putting the cuvette in the cell holder of the near-infrared spectrophotometer (NIRS6500SPL, Nireco Co., Tokyo), transmittance values at wavelengths ranging from 400–2500 nm were measured at 2 nm intervals. Concentrations of ethanol and acetic acid in the samples of culture broth and authentic compounds were measured using a gas chromatograph. The optical density of the culture broth was measured at a wavelength of 660 nm using a spectrophotometer. Concentrations of organic acids (such as gluconic, lactic, 2-ketogulutaric, succinic and propionic acids) in the culture broth were measured

Bioprocess Monitoring Using Near-Infrared Spectroscopy

179

Fig. 2 Raw near-infrared spectra of authentic ethanol, authentic acetic acid and the culture broth of the rice vinegar fermentation, and a second-derivative near-infrared spectrum of the culture broth of the rice vinegar fermentation

using a high-performance liquid chromatograph. Concentrations of glucose, reducing sugar and total sugar in the broth were measured using the enzymatic method, the Somogyi-Nelson method, and the phenol-sulfuric acid method, respectively. In Fig. 2, a raw NIR spectrum and a second derivative spectrum of the culture broth from the rice vinegar fermentation are plotted between 800 and 2500 nm. The two main peaks at around 1454 and 1940 nm on the raw spectrum may be due mainly to the absorption of water [23]. Raw NIR spectra of authentic ethanol (purity, 99.5%) and acetic acid (99.7%) are also shown in Fig. 2. The spectrum of authentic ethanol has main peaks at 1696, 1732, 2084, 2278, 2310 and 2354 nm. On the spectrum of authentic acetic acid, there are three main peaks at 1678, 1720 and 2254 nm. Peaks at around 1700 nm may be due to absorption of the carbon-hydrogen (C–H) stretch first overtone [23]. The peak at 2084 nm in the ethanol spectrum and at 2254 nm on the acetic acid spectrum may be caused by a combination of O–H stretching and O–H deformation. In the ethanol spectrum, the peak at 2278 nm may be caused by a combination of O–H stretching and C–C stretching, and the peaks at 2310 and 2354 nm may be caused by the C–H group [23]. Selection of the wavelength for the ethanol and acetic acid calibration equations was examined. As shown in Fig. 3a, all spectra of aqueous solutions of ethanol had a peak due to ethanol at 1688 nm. Spectra of aqueous mixed solutions of ethanol and acetic acid are shown in Fig. 3c. All of these mixed solutions had almost the same ethanol concentration as measured by a gas chromatograph, Cact. In Fig. 3c, it is clear that d2log(1/Tr) at 1654, 1686, 1706 and

Fig. 3 Second-derivative near-infrared spectra of aqueous samples of authentic ethanol a, acetic acid d and a mixture of authentic ethanol and acetic acid b, c

180 K. Suehara · T. Yano

Bioprocess Monitoring Using Near-Infrared Spectroscopy

181

1738 nm is not affected by the concentration of acetic acid. The negative peak at 1688 nm in Fig. 3a was shifted to a shorter wavelength when acetic acid was added to the aqueous ethanol solution. Therefore, 1686 nm should be selected as the first wavelength for ethanol calibration. As shown in Fig. 3d, there were peaks due to acetic acid at 1672 and 1718 nm in all spectra of aqueous solutions of acetic acid. Spectra of aqueous mixed solutions of ethanol and acetic acid are shown in Fig. 3b. All of these mixed solutions had almost the same acetic acid concentration, Cact.According to Fig. 3b, d2log(1/Tr) at 1674, 1706 and 1718 nm was not affected by the addition of ethanol. The negative peaks at 1672 and 1718 nm in Fig. 3d were shifted to longer wavelengths upon addition of ethanol to the acetic acid solution. Therefore, 1674 nm should be selected as the first wavelength for acetic acid calibration. MLR was performed on second derivative spectra and Cact for ethanol or acetic acid in the culture broth. As a result, the following calibration equations for measuring ethanol and acetic acid were obtained: Cpre, ethanol = 13.04 – 105.2 · A1686 – 257.2 · A1738

(6)

Cpre, acetic acid = –45.1 – 3515 · A1674 + 3179 · A1718

(7)

In the equation for ethanol, Eq. 6, the multiple correlation coefficient (R) and standard error of calibration (SEC) were 0.999 and 0.374 g/L, respectively. In the equation for acetic acid, Eq. 7, the values of R and SEC were 0.940 and 0.387 g/L, respectively. The validation results are shown in Fig. 4. Ethanol and acetic acid concentrations in the prediction sample set (n=21), which was not used for calibration, were predicted using the calibration equations and compared with the actual values of Cact. In Fig. 4, Cact is plotted on the horizontal axis, and Cpre is plotted on the vertical axis. There was good agreement between Cact and Cpre. Time courses of ethanol and acetic acid concentrations in the culture broth are shown in Fig. 5. The predicted concentrations for ethanol and acetic acid are represented by closed symbols, while open symbols denote concentrations obtained by the conventional method. At 4.5 h, some of the culture broth was drawn out and fresh medium was added to the fermentor. For both ethanol and acetic acid, the values predicted using NIR were in good agreement with those obtained by the conventional method. The correlation coefficients for ethanol and acetic acid were 0.999 and 0.989, respectively. Comparison of culture broth optical density values measured using the near-infrared spectrophotometer with those measured at 660 nm using the spectrophotometer in conventional analysis revealed simple correlation coefficients (r) of >0.99 at each wavelength between 488 and 800 nm (data not shown here). Therefore, it is possible to simultaneously measure the optical density of the culture broth and the concentrations of ethanol and acetic acid. The culture broth of the rice vinegar fermentation also contains several other constituents, including organic acids. If NIR can be used to simultaneously measure concentrations of organic acids in the culture broth, it may be

Fig. 4 Correlation between ethanol concentration, Cact, obtained using the conventional method, and that predicted using NIR with the calibration equations for ethanol a and acetic acid b

182 K. Suehara · T. Yano

Bioprocess Monitoring Using Near-Infrared Spectroscopy

183

Fig. 5 Time courses of ethanol and acetic acid concentrations in the rice vinegar fermentation. Some of the culture broth was removed and fresh medium was added to the fermentor at 4.5 h. Symbols: (empty circle), Cact of ethanol; (filled circle), Cpre of ethanol; (empty triangle), Cact of acetic acid; (filled triangle), Cpre of acetic acid

Table 2 Calibration and validation results of NIR for rice vinegar fermentation broth

Constituent

Spectra

Wavelength (nm)

Sample range (g/L)

Calibration

Validation

R

SEC

r

SEP

0.997 0.976 0.999

0.387 0.248 0.002

0.947

0.400

0.983

0.012

Ethanol Acetic acid OD Glucose

2nd 2nd raw raw

1686, 1738 1674, 1718 660 1274

5.7–34.8 66.9–109 0.151–0.327 9.73–14.72

0.999 0.940 0.999 0.924

0.374 0.387 0.005 0.483

Reducing sugar Total sugar Gluconic acid

raw

1206

16.4–21.2

0.809

0.775

2nd raw 1st 2nd raw 1st 2nd raw 1st 2nd raw 1st 2nd raw 1st 2nd

1744, 472 1848, 2084 2256, 1982 1658, 506 1844, 1556 2258, 1680 1188, 1162 1858, 1998 2252, 2484 2234, 498 1844, 2066 1680, 490 1188, 1768 1790, 1000 1240, 544 1198, 1962

30.3–34.2 7.59–12.0

0.812 0.947 0.947 0.948 0.983 0.982 0.983 0.625 0.654 0.657 0.901 0.909 0.905 0.774 0.797 0.769

0.781 0.400 0.399 0.394 0.012 0.012 0.012 0.022 0.021 0.021 0.014 0.013 0.013 0.015 0.014 0.015

Lactic acid

2-ketoglutaric acid Succinic acid

Propionic acid

0.11–0.30

0.44–0.57

0.28–0.38

0.09–0.18

184

K. Suehara · T. Yano

highly useful in the management of vinegar fermentation. Accordingly, we performed MLR on NIR spectra and Cact for several organic acids in the culture broth, including gluconic, lactic, 2-ketogulutaric, succinic and propionic acids. We also performed MLR for glucose, reducing sugar and total sugar. The calibration and validation results for all constituents are summarized in Table 2. Good calibration and validation results were obtained for gluconic and lactic acids. However, the calibration results for other organic acids were not good. Assignment of wavelengths in calibration for gluconic, lactic, 2-ketogulutaric, succinic and propionic acids was not clear, but NIR can be used to predict concentrations of some organic acids in the culture broth. The optical density and the concentrations of ethanol and acetic acid in the culture broth can be simultaneously analyzed using NIR. With NIR, it is possible to return the culture broth to the fermentor after the analysis, because NIR is a nondestructive method. The present results suggest that NIR is a useful method for the monitoring and control of rice vinegar fermentation. 3.2 L-glutamic Acid Fermentation

Okayasu et al. introduced an on-line NIR analyzing system for L-glutamic acid fermentation [24]. They used this on-line sampling system with a deforming apparatus as pretreatment for NIR analysis. They attempted simultaneous determination of the main constituents of the L-glutamic acid fermentation broth, including concentration of sugar, cells, L-glutamic acid and NH +4 . However, there are problems that must be overcome before this fermentation process can be successfully monitored on-line using NIR. Bubbles in the culture broth and adhesion of the scale in the cuvette affect measurement of NIR absorption. These are common problems in fermentation processes. In the present NIR system, a cyclone-type deforming apparatus and a U-shaped pipe were attached to the outlet stream from the fermentor and the inlet stream to a NIR cuvette. To avoid adhesion of the NIR cuvette scale, the cuvette was washed after analysis using a detergent with enzymes. Also, the NIR analyzer was maintained at 25 °C, because NIR spectra are often affected by temperature. After these measures were taken, the calibration equation was formulated. A calibration equation for sugar in the culture broth of L-glutamic fermentation was formulated with the raw NIR spectral data at 1445, 1722, 2100 and 2180 nm, and the value of R was 0.992. Calibration equations for concentrations of cells, L-glutamic acid and NH +4 were formulated, and the values of R were 0.999, 0.996 and 0.966, respectively. The time courses of concentrations of sugar, cells and L-glutamic acid were successfully measured using the NIR system. This system is occasionally used in industrial fermentation processes, but wider application requires improvement of accuracy in analyzing changes in raw materials in each lot.

Bioprocess Monitoring Using Near-Infrared Spectroscopy

185

3.3 Lactic Acid Fermentation Vaccari et al. applied NIR to the control of lactic acid fermentation [25]. To obtain chemical parameters such as concentrations of substrate, nutrients and biomass, an on-line NIR system has been introduced that allows quantitative determination of glucose, lactic acid and biomass in real time during the fermentation. They used an InfraAlyzer 450 spectrophotometer (Bran-Luebbe Co., Germany) equipped with a cuvette for liquid. They measured concentrations of lactic acid and glucose in the broth, using HPLC as the conventional method. The cell (biomass) concentration was measured conventionally by a drying method. A set of 45 samples of broth was used as the calibration sample set to create the calibration equation. Lactobacillus casei was cultivated in a fermentor with a working volume of three liters. The calibration equations for lactic acid, glucose and biomass in the culture broth were formulated. Validation revealed that the values of r were 0.9988, 0.9971 and 0.9870 for lactic acid, glucose and biomass, respectively.

4 Application of NIR to Solid Samples NIR can also be used to analyze solid samples. Applications of NIR to analysis of solid samples are summarized in Table 3. There are applications of NIR in agriculture, including analysis of soil [26–29], fruit [30–32], wood [33] and other materials [20, 21]. This suggests that NIR may be useful for process control and management of solid-state fermentation. However, there have been few applications of NIR to analysis of fermented samples in solid-state fermentation. This may be due to difficulties with analysis of NIR spectra as a result of changes in properties of the fermented sample during development of fermentation. High moisture content in the fermented sample may also complicate analysis of the NIR reflectance spectrum. In Japan, where traditional fermented foods are popular, there are examples of application of NIR to solid-state fermentation. In this section, we discuss applications of NIR for analysis of solid-state fermented samples in compost fermentation and Koji production.We discuss in detail the NIR analysis procedure for solid-state fermented samples in compost fermentation. 4.1 Compost Fermentation Suehara et al. applied NIR to management of compost fermentation [9–11]. Compost fermentation is a key technology in waste treatment and recycling of residue from food processing. In general, compost fermentation is classified into three stages (Fig. 6). Easily decomposable organic compounds such as pro-

186

K. Suehara · T. Yano

Table 3 Application of NIR for solid samples

Sample Food

Constituent Rice

Starch, moisture, protein, amino acid, ash, taste

Soybean

Protein, moisture, lipid, 7S and 11S-protein, germination

Tea leaf

Protein, moisture, total N, caffeine, theanine, amino acid

Cheese

Lipid, protein, solid, moisture, ash, chirosine, pH

Fruit (apple, orange, peach, water melon, and so on)

Brix, acidity, hardness, growth period, growing location, and so on

Agriculture

Soil

Moisture, density, total C, total N, total phosphorus, Cu, Zn, Fe, Mn and so on

Solid-state fermentation

Koji production

Mycelium, protease and amylase activity

Miso production

Moisture, salt, alcohol, total and reducing sugar, nitrogen

Mushroom production

Mycelium, moisture, rice bran

Compost fermentation

Moisture, carbon, nitrogen, lipid

Fig. 6 Outline view of the development of the compost fermentation process

Bioprocess Monitoring Using Near-Infrared Spectroscopy

187

tein, carbohydrate and lipid in the compost material are decomposed rapidly by thermophilic bacteria at high temperature (>60 °C) under aerobic conditions. This process is called thermophilic composting (primary fermentation), and lasts about a week. At this stage, the material is not suitable for compost because less easily decomposable organic compounds such as cellulose, hemicellulose and lignin in the compost material have not been decomposed. Therefore, the product of the thermophilic composting process is often subjected to secondary fermentation and maturation. For rapid treatment of organic waste, monitoring and control at optimal fermentation conditions (for instance temperature, aeration rate, pH and moisture content of the compost) are very important for management of thermophilic composting. In particular, control of the moisture content of the compost is very important for good thermophilic compost fermentation. The carbon, nitrogen and lipid contents and C/N ratio of the compost are also important factors in management of compost fermentation. Because compost fermentation is a biodegradation process, organic compounds in the compost material are converted to CO2 and NH3 gas by microorganisms. Therefore, compost fermentation can be characterized according to the time courses of carbon and nitrogen content and C/N ratio [10]. Furthermore, changes in lipid content of compost can be indicators of the end point of the fermentation [11]. In the present study, we used NIR to determine the moisture, carbon, nitrogen and lipid contents, and C/N ratio of a compost sample during tofu (soybean curd) refuse compost fermentation.We also attempted to control the moisture content during compost fermentation and simultaneously measure the carbon, nitrogen and lipid contents and C/N ratio of the compost. During the fermentation, compost samples were drawn to measure the nearinfrared spectra. To formulate a calibration equation for moisture, carbon, nitrogen and lipid content of compost, 85 samples (calibration, n=50; prediction, n=35), 108 samples (calibration, n=60; prediction, n=48) and 95 samples (calibration, n=60; validation, n=35) were prepared, respectively. Moisture, carbon and nitrogen content, and lipid content of compost samples were measured by the oven drying method, the Pregie-Dumas combustion method using a CN analyzer, and the Soxhlet extraction method, respectively. The reaction rate of the compost fermentation was defined as the total reduction in weight of the dry compost in the composter per unit time. The compost sampled from the composter was put into a polyethylene bag, which was placed in a sample holder. The reflectance values were measured with a NIR spectrophotometer (NIRS6500SPL, Nireco Co., Tokyo), and the second derivative spectra were obtained numerically from the raw spectra. To evaluate the performance of the calibration equations obtained, validation was performed using the prediction sample set, which was not used for calibration. We also performed real-time monitoring and control of the moisture content and monitoring of the time course variations of the carbon, nitrogen and lipid contents and C/N ratio on a practical fermentation. Figure 7a shows raw NIR spectra of fresh tofu refuse and the compost cultured for 0.96, 2.1 and 6.5 d. The moisture content in the fresh tofu refuse and

K. Suehara · T. Yano Fig. 7 Raw a and second derivative b NIR spectra of fresh tofu refuse and the compost cultured for 0.96, 2.1 and 6.5 d. The moisture content was about 70%. Raw c and second derivative d NIR spectra of the compost (degradation ratios of the various organic materials were almost the same); moisture content was 51.3%, 70.7% and 24.0%, respectively

188

the compost cultured for 0.96, 2.1 and 6.5 d was 72.0, 69.9, 70.2 and 70.5%, respectively. The three peaks at 970, 1450 and 1940 nm in all spectra may be due mainly to absorption of water. The baseline of the spectrum shifted upwards as the fermentation developed. This shift may be caused by a change in the properties (grain, constituent, color and so on) of the compost. It was difficult to measure the constituents using the values from the raw spectrum. Figure 7b shows the second derivative NIR spectra, in which the shift of the baseline of

Bioprocess Monitoring Using Near-Infrared Spectroscopy

189

Table 4 Assignments and simple linear regression results between carbon content and second derivative spectral values

Wavelength [nm] 922 1024 1360 1584 1718 1830 2182

r

–0.895 –0.922 –0.899 –0.974 –0.942 –0.915 –0.941

SEC [%]

Assignments

Structure

Ref.

1.98 1.72 1.95 0.998 1.50 1.79 1.50

C-H str. third overtone N-H+amide I C-H str.+C-H def. C-H str. first overtone C-H str. first overtone O-H str.+C-O def. amide I+amide III

-CH2Protein -CH3 Cellulose -CH2- or -CH3 Cellulose Protein

23 34 23 23 23 23 34

Table 5 Assignments and simple linear regression results between nitrogen content and second derivative spectral values

Wavelength [nm] 900 1060 1360 1570 1820 2174

r

–0.828 –0.945 –0.941 –0.932 –0.959 –0.895

SEC [%]

Assignments

Structure

Ref.

0.192 0.112 0.116 0.125 0.097 0.153

C-H str. third overtone N-H str. second overtone C-H str.+C-H def. N-H str. first overtone O-H str.+C-O def. amide I+amide III

Protein R-NH2 -CH3 -CONHCellulose -CONH-R, Protein

23 23 23 23 34 23

the spectrum of the compost observed in Fig. 7a was corrected and the baselines overlapped each other. The effect of the properties of the compost could be made negligible using the second derivative spectrum. Fig. 7c shows raw NIR spectra of the compost, with moisture contents of 24.0, 51.3 and 70.7%, respectively. The baseline of the spectrum shifted upwards as the moisture content increased. The moisture content affected not only absorption at the wavelength assigned to water but also absorption at all wavelengths measured. The shift of the baseline of the spectrum of the compost observed in Fig. 7c was corrected in the second derivative spectrum, as shown in Fig. 7d. Using the second derivative values at the wavelength of 960 nm, the best calibration equation for moisture was obtained. In Fig. 7d, there are prominent negative peaks at 1210, 1360, 1584, 1730, 1820 and 2174 nm. Tables 4 and 5 show the wavelength assignments and correlations between the carbon and nitrogen contents obtained by the conventional method and the values of the second derivative spectra at the assigned wavelengths, respectively. To produce a suitable calibration equation, it is very important to select a wavelength at which the absorption can be assigned to the target compound. The peaks at 1360 and 1584 nm may be related to carbon

190

K. Suehara · T. Yano

compounds in the compost. These absorptions may be caused by C–H stretching and C–H deformation in the carbon compounds of the compost, such as lipid, protein and cellulose of soybean. It should be noted, however, that most wavelengths were common to both carbon and nitrogen. The first wavelength used to formulate a calibration equation of carbon content should be either 922, 1360, 1584, 1718 or 1830 nm. These absorptions may be caused by the structure of carbon compounds in the compost [13, 14, 36]. The first wavelength used to formulate a calibration equation of nitrogen content should be either 900, 1060, 1570 or 2174 nm. Because of the absorptions in the NIR spectrum caused by nitrogen-containing structure, four wavelengths were chosen. A calibration equation for carbon and nitrogen content was developed from the various first and second wavelength combinations, and the best combination (best values of R and SEC) was chosen. To determine the lipid content of compost, 1208 and 1712 nm were selected as the first and second wavelength. As a result, the following calibration equations for measuring moisture, carbon, nitrogen and lipid contents were obtained: Cpre, moisture = 41.85 – 693.8 · A960

(8)

Cpre, carbon = 21.7 – 764 · A1584 – 1451 · A1024

(9)

Cpre, nitrogen = 2.46 – 61.5 · A2174 – 84.2 · A900

(10)

Cpre, lipid = –1.11 – 129 · A1208 – 497 · A1712 .

(11)

The values of R and SEC and the validation results of the equations are summarized in Table 6 and Fig. 8. Time courses of moisture content (a), dry weight and reaction rate of compost in the composter (b), C/N ratio (c), and carbon and nitrogen contents (d) of the compost during the compost fermentation are shown in Fig. 9. In that figure, real-time monitoring and control of moisture content during compost fermentation was performed using calibration Eq. 8. Time courses of the carbon and nitrogen contents of the compost were predicted using calibration Eqs. 9 and 10. The fermentation was performed using Table 6 Calibration and validation results for prediction of moisture, carbon, nitrogen and lipid content of compost

Constituent

Moisture Carbon Nitrogen Lipid

Wavelength [nm]

Calibration

Validation

l1

l2

R [–]

SEC [%]

r [–]

SEP [%]

960 1584 2174 1208

– 1024 900 1712

0.987 0.988 0.984 0.975

1.33 0.68 0.066 0.698

0.979 0.986 0.972 0.964

1.85 0.70 0.082 0.815

Bioprocess Monitoring Using Near-Infrared Spectroscopy

191

Fig. 8 Correlation between moisture a, carbon b, nitrogen c and lipid d contents obtained by the conventional methods, Cact, and that predicted by NIR, Cpre, with the calibration equations. The solid line represents Cpre=Cact

Fig. 9 Time courses of moisture content a, dry weight and reaction rate of compost in the composter b, C/N ratio c and carbon and nitrogen content d during the experiment in which the moisture content was controlled using the prediction results of the NIR method. Carbon and nitrogen contents were also measured using the NIR method

192

K. Suehara · T. Yano

50% as the target value of the moisture content of the compost. In Fig. 9, the values predicted using the NIR method are represented by open symbols, while closed symbols denote those obtained by the conventional method. The moisture content in compost can be measured and controlled by the NIR method, as shown in Fig. 9a. The values of the moisture content of the compost measured and controlled by the NIR method were in good agreement with those obtained by the conventional method. Maximum reaction rate of the compost fermentation was 32.0 g/h (Fig. 9b). When the moisture content was not controlled during the compost fermentation, the maximum reaction rate was only 16.7 g/h (data not shown). These results suggest that the moisture content in the compost had a large effect on the reaction rate of the compost fermentation, and that development of compost fermentation was smooth when the NIR method was used during the fermentation process and the moisture content was kept at a suitable level. In Fig. 9b, the total dry weight of the compost in the composter decreased because compost fermentation is a biodegradation process in which organic compounds in the compost material are converted to carbon dioxide and ammonia by microorganisms. The predicted carbon and nitrogen contents are represented by open symbols, while closed symbols denote those obtained by the conventional method (Fig. 9d). The C/N ratio of the compost was calculated based on the values of carbon and nitrogen content of the compost obtained using Eqs. 9 and 10. This predicted C/N ratio is represented by open symbols, while closed symbols denote those calculated by the conventional method (Fig. 9c). Because the prediction error of nitrogen content was large, the prediction precision of the C/N ratio decreased during the latter half of fermentation. However, compost fermentation can be characterized using the time courses of carbon and nitrogen content and C/N ratio. In Fig. 9d, the carbon content remained at a constant value, but nitrogen content decreased. As a result, C/N ratio increased (Fig. 9d). Furthermore, the C/N ratio of the compost was close to the C/N ratio of the microorganism (about 13), because of the growth of the microorganisms in the compost. Figure 10 shows the time course of lipid content of the compost during compost fermentation. Masui et al. reported that the time course of the lipid content of compost during compost fermentation is coupled to the decrease in total dry weight of the compost in the composter, and that the change in lipid content in the compost may be an indicator of the end point of thermophilic compost fermentation [11]. In Fig. 10, the predicted lipid content is represented by open symbols, while closed symbols denote the lipid content obtained by the conventional method. During the compost fermentation, there was good agreement between the values obtained by the conventional method and those obtained by NIR. The results suggest that the calibration equation we formulated, Eq. 11, had a high practicability. We examined rapid measurement and control of the moisture content of the compost using the NIR method.We also examined simultaneous measurement of carbon, nitrogen and lipid content and C/N ratio of the compost using the

Bioprocess Monitoring Using Near-Infrared Spectroscopy

193

Fig. 10 Monitoring of the lipid content during compost fermentation using the NIR method

NIR method, and good results were obtained. The present results suggest that NIR is a useful method for management of compost fermentation, such as control of moisture, monitoring of organic matter in the material, and detection of the end point of compost fermentation. 4.2 Koji Production Kojima et al. used NIR to determine the mycelial weight of Aspergillus oryzae in Koji [35]. Koji is a raw material for the production of Japanese sake, and consists of steamed rice inoculated with A. oryzae. The mycelial weight in Koji is the most important factor in Japanese sake processing, because the rate at which starch is converted to sugar depends on the content of mycelia produced by A. oryzae. Therefore, NIR has been applied to Koji production because it allows easier measurement of the mycelial weight of A. oryzae in Koji rice. NIR measurements of ground Koji rice were performed over a range of 1100 to 2500 nm. A calibration equation for mycelial weight in Koji was formulated using NIR spectrum data at 1730, 1738, 2348 and 2360 nm, and the values of R and SEC were 0.98 and 0.56 mg/g, respectively. A good calibration equation was obtained by MLR based on the NIR spectra data and the mycelial weight was measured by the enzymatic method. Aramaki et al. also formulated a calibration equation for mycelial weight in Koji [36], using NIR data at 2348 nm. The assignment of absorbance at 2348 nm may be due to lipids in A. oryzae.

5 Other Applications In this section, other applications of NIR in bioprocess monitoring are discussed. These applications were researched by Yano, Suehara and coworkers.

194

K. Suehara · T. Yano

5.1 Mushroom Cultivation Mushrooms are important materials for food and medical products. Solid cultivation is used for commercial mushroom production, but measurement of cell mass, growth rate of mycelium, and constituents in the solid media is very difficult, because the analysis is time consuming and laborious. NIR has been applied to the prediction of cell growth rate of the mushroom Ganoderma lucidum in solid culture [37]. Simultaneous determination of water and rice bran content in solid media material has also been attempted [38]. Cell mass is conventionally measured by analyzing the concentration of glucosamine, a component of the cell wall. We predicted glucosamine concentration in culture materials using six-wavelength regression analysis. We used the values of the first derivative spectra at 1203, 1635, 1751, 2103, 2375 and 2431 nm to obtain a calibration equation. The values of R and SEC of the calibration sample set (n=30) were 0.969 and 0.622 mg/g, respectively. The value predicted by NIR was in fairly good agreement with that obtained by the conventional method. The r and standard error of prediction (SEP) of the prediction sample set (n=11) were 0.992 and 0.346 mg/g, respectively. The specific growth rate of the cell was calculated using NIR. To obtain a calibration equation for water content in the mushroom media material, we performed simple linear regression on the NIR spectral data at 1450 nm and on the water content of the calibration sample set (n=113) obtained using a drying method. The values of R and SEC were 0.995 and 1.33%, respectively. Based on the content of rice bran in the solid media determined by NIR, a calibration equation using the second derivative reflectance data at 672 and 2100 nm was obtained with values of R and SEC of 0.978 and 1.73%, respectively. To validate the calibration equations obtained, water and rice bran content in the prediction sample set (n=56) were calculated using the calibration equations. For both water and rice bran content, there was excellent agreement between the results of the conventional method and those of the NIR method. The r and SEP were 0.997 and 1.33% for water content and 0.975 and 1.84% for rice bran content, respectively. The NIR method is a useful method for rapid measurement of cell mass, growth rate of mycelium and constituent content in solid media of mushroom cultivation. 5.2 Medical Products Precise quality control is often required for the manufacture of medical products such as blood anticoagulant and peritoneal dialysis solution. The blood collected from a donor is stored after mixing with a blood anticoagulant. The peritoneal dialysis solution is introduced into the peritoneal cavity of renal failure, and waste materials in the blood are dialyzed into the solution through

Bioprocess Monitoring Using Near-Infrared Spectroscopy

195

the peritoneum. For management of production of blood anticoagulant and peritoneal dialysis solutions, it is very important to monitor concentrations of constituents such as glucose, citric acid and lactic acid. Generally, concentrations of glucose, citric acid and lactic acid in the product are measured using liquid chromatography. There is a demand for a simple method of measuring the concentrations of these constituents simultaneously. NIR has been used for simultaneous prediction of glucose and citric acid concentration in blood anticoagulant solutions [16], and for simultaneous prediction of glucose and lactic acid concentration in peritoneal dialysis solutions [17]. We used MLR to obtain calibration equations relating the NIR spectral data and the glucose, citric acid and lactic acid concentrations of a calibration sample set obtained by enzymatic methods. A calibration equation for glucose in blood anticoagulant solutions was formulated with the second derivative NIR spectral data at 2274 and 1674 nm; the values of R and SEC were 0.993 and 0.169 g/L, respectively. A calibration equation for citric acid in blood anticoagulant solutions was formulated with the second derivative NIR spectral data at 1690 nm; the values of R and SEC were 0.993 and 0.159 g/L, respectively. A calibration equation for glucose in peritoneal dialysis solution was formulated with the second derivative NIR spectral data at 2270 nm; the values of R and SEC were 0.996 and 2.03 g/L, respectively.A calibration equation for lactic acid in peritoneal dialysis solution was formulated with the second derivative NIR spectral data at 1688 and 1268 nm; the values of R and SEC were 0.997 and 0.178 g/L, respectively. In the validation results of the calibration equations, there was excellent agreement between the results from the enzymatic method and the NIR method for these constituents. The values of r for glucose and citric acid in the blood anticoagulant and for glucose and lactic acid in the peritoneal dialysis solution were 0.994, 0.997, 0.996 and 0.996, respectively. These results suggest that NIR is a useful method for management of production of blood anticoagulant and peritoneal dialysis solutions. NIR is probably also applicable to the manufacture of other medical products. 5.3 Glycolipid Fermentation The yeast Kurtzumanomyces sp. I-11 produces mannosyl erythritol lipid (MEL) from soybean oil. MEL is a biosurfactant, and is classified as a glycolipid. MEL is a typical amphiphilic compound, containing both lipophilic and hydrophilic moieties, and is composed of mannose, erythritol and fatty acids. Biosurfactants such as MEL have been shown to have special properties over their chemically synthesized counterparts. These properties include low toxicity, biodegradability, biological activity, a wide variety of possible structures and ease of synthesis from inexpensive, renewable resources. Consequently, biosurfactants have a wide range of possible industrial applications, especially in the production of food, cosmetics, pharmaceuticals and chemicals for biotechnology.

196

K. Suehara · T. Yano

MEL is produced by glycolipid fermentation from soybean oil added to a medium as a carbon source. A system for measuring concentrations of MEL and soybean oil during fermentation has been developed using NIR [39]. In the present study, MEL and soybean oil in the culture broth were extracted with ethyl acetate. NIR transmittance spectra of the ethyl acetate extract were measured.Absorption caused by MEL was observed at 1436, 1920 and 2052 nm. To obtain a calibration equation, MLR was performed between the second derivative NIR spectral data at 2040 and 1312 nm and MEL concentrations were obtained using a method of thin-layer chromatography with a flame-ionization detector (TLC/FID). The values of R and SEC were 0.994 and 0.48 g/L, respectively. Absorption caused by soybean oil was observed at 1208, 1716, 1766, 2182 and 2302 nm. A calibration equation for soybean oil was formulated with the second derivative NIR spectral data at 2178 and 2090 nm; the values of R and SEC were 0.974 and 0.77 g/L, respectively. The results from validation of the calibration equation showed good agreement between the results of the TLC/FID method and those of the NIR method for both constituents. The values of r and SEP for MEL were 0.994 and 0.45 g/L, respectively. The values of r and SEP for soybean oil were 0.979 and 0.56 g/L, respectively. Therefore, good results were obtained when the NIR method was applied to measurement of concentrations of MEL and soybean oil in practical fermentation. The present results indicate that NIR is a useful method for measurement of raw material and product in glycolipid fermentation.

6 Conclusions Rapid measurement of the main constituents of the fermentation process using the NIR method was examined. NIR proved useful for management of various bioprocesses such as rice vinegar fermentation, compost fermentation, mushroom cultivation, glycolipid fermentation and manufacture of medical products. In rice vinegar fermentation, we were able to simultaneously analyze optical density and concentrations of ethanol and acetic acid in the culture broth. In compost fermentation, we examined simultaneous measurement of moisture, carbon, nitrogen and lipid contents, and C/N ratio of the compost using NIR, and obtained good results.We were able to control moisture content of the compost using NIR. The operational procedure of NIR was very simple, and the time required for measurement (full range scanning; 32 times) was only 1.5 min. The time required can be dramatically shortened if only absorption in the wavelength used in the calibration equation is measured. In addition, it is possible to return the fermented sample to the fermentor after analysis, because NIR is a nondestructive method. On-line monitoring using an optical fiber probe is also possible, making it unnecessary to pack the fermented sample into a cuvette or

Bioprocess Monitoring Using Near-Infrared Spectroscopy

197

a polyethylene bag before measurement, because the fermented sample is not pretreated, and constituents of the sample can be measured using transmitted or reflected rays. These results suggest that NIR is a useful method for monitoring and control of bioprocesses. We plan to study computerized control systems for process management using NIR, for use in developing efficient bioprocesses.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Stark E, Luchter K, Margoshes M (1986) Appl Spectrosc Rev 22:335 Dumoulin ED, Azais BP, Guerain JT (1987) J Food Sci 52:626 Kaffka KJ, Norris KH, (1976) Acta Aliment Hung 5:267 Coventry AG, Hunston MJ (1984) Cereal Food World 29:715 Halsey SA (1985) J I Brewing 91:306 Yano T, Harata M (1994) J Ferment Bioeng 77:659 Iwamoto M (1980) Nippon Shokuhin Kogyo Gakkaishi (J Jpn Soc Food Sci) (in Japanese) 27:464 Yano T, Aimi T, Nakano Y, Tamai M (1997) J Ferment Bioeng 84:461 Suehara K, Ohta Y, Nakano Y, Yano T (1999) J Biosci Bioeng 87:769 Suehara K, Nakano Y, Yano T (2001) J Near Infrared Spec 9:35 Masui D, Suehara K, Nakano Y, Yano T (2001) Near Infrared Analysis 2:37 Mark H (1993) In: Patonay G (ed) Advances in near-infrared measurements. JAI, London, p 55 Burns DA, Ciurczak EW (1992) (eds) Handbook of near-infrared analysis. Marcel Dekker, New York, p 383 Brunt K (1992) In: Hideum KI, Isaksson T, Naes T, Tandberg A (eds) Near-infrared spectroscopy. Ellis Horwood, London, UK, p 327 Lizzano R, Barzaghi S, Cattaneo TMP, Giangiacomo R (1999): In: Davies AMC, Giangiacomo R (eds) Near-infrared spectroscopy. NIR Publications, Chichester, UK, p 339 Yano T, Funatsu T, Suehara K, Nakano Y (2001) J Near Infrared Spec 9:43 Yano T, Matsushige H, Suehara K, Nakano Y (2000) J Biosci Bioeng 90:540 Cho SY, Kim JY, Rhee (1998) J Near Infrared Spec 6:A349 Domjan G, Kako J, Valyi-Nagy I (1998) J Near Infrared Spectrosc 6:A279 Heise HM, Marbach R and Bittner A (1998) J Near Infrared Spec 6:361 Iwamoto M, Kawano S, Ozaki Y (1995) J Near Infrared Spec 3:179 Osborne BG, Fearn T (1986) Near infrared spectroscopy in food analysis. Wiley, New York, p 144 Osborne BG, Fearn T (1986) Near infrared spectroscopy in food analysis. Wiley, New York, p 86 Okayasu S, Katayama M, Miyashiro S (1995) Kagaku Kougaku Ronbun (in Japanese) 42:436 Vaccari G, Dosi E, Campi AL, Gonzales-Vara A, Matteuzzi D, Mantovani G (1994) Biotech Bioeng 43:913 Ryu KS, Park JS, Kim BJ (2001) In: Davies AMC, Cyo RK (eds) Near-infrared spectroscopy. NIR Publications, Chichester, UK, p 399 Salgo A, Nagy J, Tarnoy J, Marth P, Palmai O, Szabo-Kele G (1998) J Near Infrared Spec 6:199 Pietikainen J, Fritze H (1995) Soil Biol Biochem 27:101

198

K. Suehara · T. Yano

29. Moron A, Cozzolino D (2003) J Near Infrared Spec 11:145 30. Kawano S, Fujiwara T, Iwamoto M (1993) Engei Gakkai Zasshi (J Jpn Soc Hortic Sci) (in Japanese) 62:645 31. Sohn MR, Cho RK (1999) In: Davies AMC, Giangiacomo R (eds) Near-infrared spectroscopy. NIR publications, Chichester UK, p 797 32. Sohn MR, Kwon YK, Cho RK (2001) Near Infrared Analysis 2:55 33. Tsuchikawa S, Tsutsumi S (2001) In: Davies AMC, Cyo RK (eds) Near-infrared spectroscopy. NIR publications, Chichester, UK, p 423 34. Osborne BG, Fearn T, Hindle PH (1986) Practical NIR spectroscopy with applications in food and beverage analysis. Longman, Harlow, UK, p 13 35. Kojima Y, Asai Y, Hata Y, Ichikawa E (1994) Nippon Nogeik Kaishi (in Japanese) 68:801 36. Aramaki I, Fukuda K, Hashimoto T, Ishikawa T, Kizaki Y, Okazaki N (1995) SeibutsuKougaku Kais (in Japanese) 73:33 37. Suehara K, Nakano Y, Yano T (1998) J Near Infrared Spec 6:273 38. Yano T, Suehara K, Nakano Y (1998) J Ferment Bioeng 86:472 39. Nakamichi K, Suehara K, Nakano Y, Kakugawa K, Tamai M,Yano T (2002) J Near Infrared Spec 10:53

Received: September 2003

Author Index Volumes 51 – 90 Author Index Volumes 1–50 see Volume 50

Ackermann, J.-U. see Babel, W.: Vol. 71, p. 125 Adam, W., Lazarus, M., Saha-Möller, C. R., Weichhold, O., Hoch, U., Häring, D., Schreier, Ü.: Biotransformations with Peroxidases. Vol. 63, p. 73 Ahring, B. K.: Perspectives for Anaerobic Digestion. Vol. 81, p. 1 Ahring, B. K. see Angelidaki, I.: Vol. 82, p. 1 Ahring, B. K. see Gavala, H. N.: Vol. 81, p. 57 Ahring, B. K. see Hofman-Bang, J.: Vol. 81, p. 151 Ahring, B. K. see Mogensen, A. S.: Vol. 82, p. 69 Ahring, B. K. see Pind, P. F.: Vol. 82, p. 135 Ahring, B. K. see Skiadas, I. V.: Vol. 82, p. 35 Akhtar, M., Blanchette, R. A., Kirk, T. K.: Fungal Delignification and Biochemical Pulping of Wood. Vol. 57, p. 159 Allan, J. V., Roberts, S. M., Williamson, N. M.: Polyamino Acids as Man-Made Catalysts. Vol. 63, p. 125 Allington, R. W. see Xie, S.: Vol. 76, p. 87 Al-Rubeai, M.: Apoptosis and Cell Culture Technology. Vol. 59, p. 225 Al-Rubeai, M. see Singh, R. P.: Vol. 62, p. 167 Alsberg, B. K. see Shaw, A. D.: Vol. 66, p. 83 Angelidaki, I., Ellegaard, L., Ahring, B. K.: Applications of the Anaerobic Digestion Process. Vol. 82, p. 1 Angelidaki, I. see Gavala, H. N.: Vol. 81, p. 57 Angelidaki, I. see Pind, P. F.: Vol. 82, p. 135 Antranikian, G. see Ladenstein, R.: Vol. 61, p. 37 Antranikian, G. see Müller, R.: Vol. 61, p. 155 Archelas, A. see Orru, R. V. A.: Vol. 63, p. 145 Argyropoulos, D. S.: Lignin. Vol. 57, p. 127 Arnold, F. H., Moore, J. C.: Optimizing Industrial Enzymes by Directed Evolution. Vol. 58, p. 1 Autuori, F., Farrace, M. G., Oliverio, S., Piredda, L., Piacentini, G.: “Tissie” Transglutaminase and Apoptosis. Vol. 62, p. 129 Azerad, R.: Microbial Models for Drug Metabolism. Vol. 63, p. 169 Babel, W., Ackermann, J.-U., Breuer, U.: Physiology, Regulation and Limits of the Synthesis of Poly(3HB). Vol. 71, p. 125 Bajpai, P., Bajpai, P. K.: Realities and Trends in Emzymatic Prebleaching of Kraft Pulp. Vol. 56, p. 1 Bajpai, P., Bajpai, P. K.: Reduction of Organochlorine Compounds in Bleach Plant Effluents. Vol. 57, p. 213 Bajpai, P. K. see Bajpai, P.: Vol. 56, p. 1 Bajpai, P. K. see Bajpai, P.: Vol. 57, p. 213 Banks, M. K., Schwab, P., Liu, B., Kulakow, P.A., Smith, J. S., Kim, R.: The Effect of Plants on the Degradation and Toxicity of Petroleum Contaminants in Soil: A Field Assessment.Vol. 78, p. 75 Barut, M. see Strancar, A.: Vol. 76, p. 49

200

Author Index Volumes 51–90

Bárzana, E.: Gas Phase Biosensors. Vol. 53, p. 1 Basu, S. K. see Mukhopadhyay, A.: Vol. 84, p. 183 Bathe, B. see Pfefferle, W.: Vol. 79, p. 59 Bazin, M. J. see Markov, S. A.: Vol. 52, p. 59 Bellgardt, K.-H.: Process Models for Production of b-Lactam Antibiotics. Vol. 60, p. 153 Beppu, T.: Development of Applied Microbiology to Modern Biotechnology in Japan.Vol.69, p. 41 Berovic, M. see Mitchell, D.A.: Vol. 68, p. 61 Beyeler, W., DaPra, E., Schneider, K.: Automation of Industrial Bioprocesses. Vol. 70, p. 139 Beyer, M. see Seidel, G.: Vol. 66, p. 115 Beyer, M. see Tollnick, C.: Vol. 86, p. 1 Bhardwaj, D. see Chauhan, V.S.: Vol. 84, p. 143 Bhatia, P.K., Mukhopadhyay, A.: Protein Glycosylation: Implications for in vivo Functions and Thereapeutic Applications. Vol. 64, p. 155 Bisaria, V.S. see Ghose, T.K.: Vol. 69, p. 87 Blanchette R. A. see Akhtar, M.: Vol. 57, p. 159 Bocker, H., Knorre, W.A.: Antibiotica Research in Jena from Penicillin and Nourseothricin to Interferon. Vol. 70, p. 35 de Bont, J.A.M. see van der Werf, M. J.: Vol. 55, p. 147 van den Boom, D. see Jurinke, C.: Vol. 77, p. 57 Borah, M. M. see Dutta, M.: Vol. 86, p. 255 Brainard, A. P. see Ho, N. W. Y.: Vol. 65, p. 163 Brazma, A., Sarkans, U., Robinson, A., Vilo, J., Vingron, M., Hoheisel, J., Fellenberg, K.: Microarray Data Representation, Annotation and Storage. Vol. 77, p. 113 Breuer, U. see Babel, W.: Vol. 71, p. 125 Broadhurst, D. see Shaw, A. D.: Vol. 66, p. 83 Bruckheimer, E. M., Cho, S. H., Sarkiss, M., Herrmann, J., McDonell, T. J.: The Bcl-2 Gene Family and Apoptosis. Vol 62, p. 75 Brüggemann, O.: Molecularly Imprinted Materials – Receptors More Durable than Nature Can Provide. Vol. 76, p. 127 Bruggink, A., Straathof, A. J. J., van der Wielen, L. A. M.: A ‘Fine’ Chemical Industry for Life Science Products: Green Solutions to Chemical Challenges. Vol. 80, p. 69 Buchert, J. see Suurnäkki, A.: Vol. 57, p. 261 Bungay, H. R. see Mühlemann, H. M.: Vol. 65, p. 193 Bungay, H.R., Isermann, H.P.: Computer Applications in Bioprocessin. Vol. 70, p. 109 Büssow, K. see Eickhoff, H.: Vol. 77, p. 103 Byun, S.Y. see Choi, J.W.: Vol. 72, p. 63 Cabral, J. M. S. see Fernandes, P.: Vol. 80, p. 115 Cahill, D. J., Nordhoff, E.: Protein Arrays and Their Role in Proteomics. Vol. 83, p. 177 Cantor, C. R. see Jurinke, C.: Vol. 77, p. 57 Cao, N. J. see Gong, C. S.: Vol. 65, p. 207 Cao, N. J. see Tsao, G. T.: Vol. 65, p. 243 Carnell, A. J.: Stereoinversions Using Microbial Redox-Reactions. Vol. 63, p. 57 Cash, P.: Proteomics of Bacterial Pathogens. Vol. 83, p. 93 Cen, P., Xia, L.: Production of Cellulase by Solid-State Fermentation. Vol. 65, p. 69 Chand, S., Mishra, P.: Research and Application of Microbial Enzymes – India’s Contribution. Vol. 85, p. 95 Chang, H. N. see Lee, S. Y.: Vol. 52, p. 27 Chauhan, V. S., Bhardwaj, D.: Current Status of Malaria Vaccine Development. Vol. 84, p. 143 Cheetham, P. S. J.: Combining the Technical Push and the Business Pull for Natural Flavours.Vol. 55, p. 1 Cheetham, P. S. J.: Bioprocesses for the Manufacture of Ingredients for Foods and Cosmetics. Vol. 86, p. 83 Chen, C. see Yang, S.-T.: Vol. 87, p. 61 Chen, Z. see Ho, N. W. Y.: Vol. 65, p. 163

Author Index Volumes 51–90

201

Chenchik, A. see Zhumabayeva, B.: Vol. 86, p. 191 Cho, S. H. see Bruckheimer, E. M.: Vol. 62, p. 75 Cho, G.H. see Choi, J.W.: Vol 72, p. 63 Choi, J. see Lee, S.Y.: Vol. 71, p. 183 Choi, J.W., Cho, G.H., Byun, S.Y., Kim, D.-I.: Integrated Bioprocessing for Plant Cultures. Vol. 72, p. 63 Christensen, B., Nielsen, J.: Metabolic Network Analysis – A Powerful Tool in Metabolic Engineering. Vol. 66, p. 209 Christians, F. C. see McGall, G.H.: Vol. 77, p. 21 Chu, J. see Zhang, S.: Vol. 87, p. 97 Chui, G. see Drmanac, R.: Vol. 77, p. 75 Ciaramella, M. see van der Oost, J.: Vol. 61, p. 87 Contreras, B. see Sablon, E.: Vol. 68, p. 21 Conway de Macario, E., Macario, A. J. L.: Molecular Biology of Stress Genes in Methanogens: Potential for Bioreactor Technology. Vol. 81, p. 95 Cordero Otero, R.R. see Hahn-Hägerdal, B.: Vol. 73, p. 53 Cordwell S. J. see Nouwens, A.S.: Vol. 83, p. 117 Cornet, J.-F., Dussap, C. G., Gros, J.-B.: Kinetics and Energetics of Photosynthetic MicroOrganisms in Photobioreactors. Vol. 59, p. 153 da Costa, M. S., Santos, H., Galinski, E. A.: An Overview of the Role and Diversity of Compatible Solutes in Bacteria and Archaea. Vol. 61, p. 117 Cotter, T. G. see McKenna, S. L.: Vol. 62, p. 1 Croteau, R. see McCaskill, D.: Vol. 55, p. 107 Danielsson, B. see Xie, B.: Vol. 64, p. 1 DaPra, E. see Beyeler, W.: Vol. 70, p. 139 Darzynkiewicz, Z., Traganos, F.: Measurement of Apoptosis. Vol. 62, p. 33 Davey, H. M. see Shaw, A. D.: Vol. 66, p. 83 Dean, J. F. D., LaFayette, P. R., Eriksson, K.-E. L., Merkle, S. A.: Forest Tree Biotechnolgy. Vol. 57, p. 1 Debabov, V. G.: The Threonine Story. Vol. 79, p. 113 Demain, A.L., Fang, A.: The Natural Functions of Secondary Metabolites. Vol. 69, p. 1 Dhar, N. see Tyagi, A. K.: Vol. 84, p. 211 Diaz, R. see Drmanac, R.: Vol. 77, p. 75 Dochain, D., Perrier, M.: Dynamical Modelling, Analysis, Monitoring and Control Design for Nonlinear Bioprocesses. Vol. 56, p. 147 Dolfing, J. see Mogensen, A. S.: Vol. 82, p. 69 Drmanac, R., Drmanac, S., Chui, G., Diaz, R., Hou, A., Jin, H., Jin, P., Kwon, S., Lacy, S., Moeur, B., Shafto, J., Swanson, D., Ukrainczyk, T., Xu, C., Little, D.: Sequencing by Hybridization (SBH): Advantages, Achievements, and Opportunities. Vol. 77, p. 75 Drmanac, S. see Drmanac, R.: Vol. 77, p. 75 Du, J. see Gong, C. S: Vol. 65, p. 207 Du, J. see Tsao, G. T.: Vol. 65, p. 243 Dueser, M. see Raghavarao, K.S.M.S.: Vol. 68, p. 139 Dussap, C. G. see Cornet J.-F.: Vol. 59, p. 153 Dutta, M., Borah, M. M., Dutta, N. N.: Adsorptive Separation of b-Lactam Antibiotics: Technological Perspectives. Vol. 86, p. 255 Dutta, N. N. see Ghosh, A. C.: Vol. 56, p. 111 Dutta, N. N. see Sahoo, G. C.: Vol. 75, p. 209 Dutta, N. N. see Dutta, M.: Vol. 86, p. 255 Dynesen, J. see McIntyre, M.: Vol. 73, p. 103 Eggeling, L., Sahm, H., de Graaf, A. A.: Quantifying and Directing Metabolite Flux: Application to Amino Acid Overproduction. Vol. 54, p. 1 Eggeling, L. see de Graaf, A.A.: Vol. 73, p. 9

202

Author Index Volumes 51–90

Eggink, G., see Kessler, B.: Vol. 71, p. 159 Eggink, G., see van der Walle, G. J. M.: Vol. 71, p. 263 Egli, T. see Wick, L. M.: Vol. 89, p. 1 Ehrlich, H. L. see Rusin, P.: Vol. 52, p. 1 Eickhoff, H., Konthur, Z., Lueking, A., Lehrach, H., Walter, G., Nordhoff, E., Nyarsik, L., Büssow, K.: Protein Array Technology: The Tool to Bridge Genomics and Proteomics. Vol. 77, p. 103 Elias, C. B., Joshi, J. B.: Role of Hydrodynamic Shear on Activity and Structure of Proteins. Vol. 59, p. 47 Ellegaard, L. see Angelidaki, I.: Vol. 82, p. 1 Elling, L.: Glycobiotechnology: Enzymes for the Synthesis of Nucleotide Sugars. Vol. 58, p. 89 Enfors, S.-O. see Rozkov, A.: Vol. 89, p. 163 Eriksson, K.-E. L. see Kuhad, R. C.: Vol. 57, p. 45 Eriksson, K.-E. L. see Dean, J. F. D.: Vol. 57, p. 1 Faber, K. see Orru, R. V. A.: Vol. 63, p. 145 Fahnert, B., Lilie, H., Neubauer, P.: Inclusion Bodies: Formation and Utilisation. Vol. 89, p. 93 Fang, A. see Demain, A.L.: Vol. 69, p. 1 Farrace, M. G. see Autuori, F.: Vol. 62, p. 129 Farrell, R. L., Hata, K., Wall, M. B.: Solving Pitch Problems in Pulp and Paper Processes. Vol. 57, p. 197 Fellenberg, K. see Brazma, A.: Vol. 77, p. 113 Fernandes, P., Prazeres, D. M. F., Cabral, J. M. S.: Membrane-Assisted Extractive Bioconversions. Vol. 80, p. 115 Ferro, A., Gefell, M., Kjelgren, R., Lipson, D. S., Zollinger, N., Jackson, S.: Maintaining Hydraulic Control Using Deep Rooted Tree Systems. Vol. 78, p. 125 Fiechter, A.: Biotechnology in Switzerland and a Glance at Germany. Vol. 69, p. 175 Fiechter, A. see Ochsner, U. A.: Vol. 53, p. 89 Flechas, F. W., Latady, M.: Regulatory Evaluation and Acceptance Issues for Phytotechnology Projects. Vol. 78, p. 171 Foody, B. see Tolan, J. S.: Vol. 65, p. 41 Fréchet, J. M. J. see Xie, S.: Vol. 76, p. 87 Freitag, R., Hórvath, C.: Chromatography in the Downstream Processing of Biotechnological Products. Vol. 53, p. 17 Friehs, K.: Plasmid Copy Number and Plasmid Stability. Vol. 86, p. 47 Furstoss, R. see Orru, R. V. A.: Vol. 63, p. 145 Galinski, E.A. see da Costa, M.S.: Vol. 61, p. 117 Gàrdonyi, M. see Hahn-Hägerdal, B.: Vol. 73, p. 53 Gatfield, I.L.: Biotechnological Production of Flavour-Active Lactones. Vol. 55, p. 221 Gavala, H. N., Angelidaki, I., Ahring, B. K.: Kinetics and Modeling of Anaerobic Digestion Process. Vol. 81, p. 57 Gavala, H. N. see Skiadas, I. V.: Vol. 82, p. 35 Gefell, M. see Ferro, A.: Vol. 78, p. 125 Gemeiner, P. see Stefuca, V.: Vol. 64, p. 69 Gerlach, S. R. see Schügerl, K.: Vol. 60, p. 195 Ghose, T. K., Bisaria, V.S.: Development of Biotechnology in India. Vol. 69, p. 71 Ghose, T. K. see Ghosh, P.: Vol. 85, p. 1 Ghosh, A. C., Mathur, R. K., Dutta, N. N.: Extraction and Purification of Cephalosporin Antibiotics. Vol. 56, p. 111 Ghosh, P., Ghose, T. K.: Bioethanol in India: Recent Past and Emerging Future. Vol. 85, p. 1 Ghosh, P. see Singh, A.: Vol. 51, p. 47 Gilbert, R. J. see Shaw, A. D.: Vol. 66, p. 83 Gill, R.T. see Stephanopoulos, G.: Vol. 73, p. 1 Gomes, J., Menawat, A. S.: Fed-Batch Bioproduction of Spectinomycin. Vol. 59, p. 1

Author Index Volumes 51–90

203

Gong, C. S., Cao, N. J., Du, J., Tsao, G. T.: Ethanol Production from Renewable Resources. Vol. 65, p. 207 Gong, C. S. see Tsao, G. T.: Vol. 65, p. 243 Goodacre, R. see Shaw, A. D.: Vol. 66, p. 83 de Graaf, A. A., Eggeling, L., Sahm, H.: Metabolic Engineering for L-Lysine Production by Corynebacterium glutamicum. Vol. 73, p. 9 de Graaf, A. A. see Eggeling, L.: Vol. 54, p. 1 de Graaf, A. A. see Weuster-Botz, D.: Vol. 54, p. 75 de Graaf, A. A. see Wiechert, W.: Vol. 54, p. 109 Grabley, S., Thiericke, R.: Bioactive Agents from Natural Sources: Trends in Discovery and Application. Vol. 64, p. 101 Griengl, H. see Johnson, D. V.: Vol. 63, p. 31 Gros, J.-B. see Larroche, C.: Vol. 55, p. 179 Gros, J.-B. see Cornet, J. F.: Vol. 59, p. 153 Gu, M. B., Mitchell, R. J., Kim, B. C.: Whole-Cell-Based Biosensors for Environmental Biomonitoring and Application.Vol. 87, p. 269 Guenette M. see Tolan, J. S.: Vol. 57, p. 289 Gupta, M. N. see Roy, I.: Vol. 86, p. 159 Gupta, S. K.: Status of Immunodiagnosis and Immunocontraceptive Vaccines in India. Vol. 85, p. 181 Gutman, A. L., Shapira, M.: Synthetic Applications of Enzymatic Reactions in Organic Solvents. Vol. 52, p. 87 Haagensen, F. see Mogensen, A. S.: Vol. 82, p. 69 Hahn-Hägerdal, B., Wahlbom, C.F., Gárdonyi, M., van Zyl, W.H., Cordero Otero, R.R., Jönsson, L.J.: Metabolic Engineering of Saccharomyces cerevisiae for Xylose Utilization.Vol. 73, p.53 Haigh, J.R. see Linden, J.C.: Vol. 72, p. 27 Hall, D. O. see Markov, S. A.: Vol. 52, p. 59 Hall, P. see Mosier, N. S.: Vol. 65, p. 23 Hammar, F.: History of Modern Genetics in Germany. Vol. 75, p. 1 Hannenhalli, S., Hubbell, E., Lipshutz, R., Pevzner, P. A.: Combinatorial Algorithms for Design of DNA Arrays. Vol. 77, p. 1 Haralampidis, D., Trojanowska, M., Osbourn, A. E.: Biosynthesis of Triterpenoid Saponins in Plants. Vol. 75, p. 31 Häring, D. see Adam, E.: Vol. 63, p. 73 Harvey, N. L., Kumar, S.: The Role of Caspases in Apoptosis. Vol. 62, p. 107 Hasegawa, S., Shimizu, K.: Noninferior Periodic Operation of Bioreactor Systems. Vol. 51, p. 91 Hata, K. see Farrell, R. L.: Vol. 57, p. 197 Hecker, M.: A Proteomic View of Cell Physiology of Bacillus subtilis – Bringing the Genome Sequence to Life. Vol. 83, p. 57 Hecker, M. see Schweder, T.: Vol. 89, p. 47 van der Heijden, R. see Memelink, J.: Vol. 72, p. 103 Hein, S. see Steinbüchel, A.: Vol. 71, p. 81 Hembach, T. see Ochsner, U. A.: Vol. 53, p. 89 Henzler, H.-J.: Particle Stress in Bioreactor. Vol. 67, p. 35 Herrler, M. see Zhumabayeva, B.: Vol. 86, p. 191 Herrmann, J. see Bruckheimer, E. M.: Vol. 62, p. 75 Hewitt, C. J., Nebe-Von-Caron, G.: The Application of Multi-Parameter Flow Cytometry to Monitor Individual Microbial Cell Physiological State. Vol. 89, p. 197 Hill, D. C., Wrigley, S. K., Nisbet, L. J.: Novel Screen Methodologies for Identification of New Microbial Metabolites with Pharmacological Activity. Vol. 59, p. 73 Hiroto, M. see Inada, Y.: Vol. 52, p. 129 Ho, N. W. Y., Chen, Z., Brainard, A. P. Sedlak, M.: Successful Design and Development of Genetically Engineering Saccharomyces Yeasts for Effective Cofermentation of Glucose and Xylose from Cellulosic Biomass to Fuel Ethanol. Vol. 65, p. 163

204

Author Index Volumes 51–90

Hoch, U. see Adam, W.: Vol. 63, p. 73 Hoffmann, F., Rinas, U.: Stress Induced by Recombinant Protein Production in Escherichia coli. Vol. 89, p. 73 Hoffmann, F., Rinas, U.: Roles of Heat-Shock Chaperones in the Production of Recombinant Proteins in Escherichia coli. Vol. 89, p. 143 Hofman-Bang, J., Zheng, D., Westermann, P., Ahring, B. K., Raskin, L.: Molecular Ecology of Anaerobic Reactor Systems. Vol. 81, p. 151 Hoheisel, J. see Brazma, A.: Vol. 77, p. 113 Holló, J., Kralovánsky, U.P.: Biotechnology in Hungary. Vol. 69, p. 151 Honda, H., Kobayashi, T.: Industrial Application of Fuzzy Control in Bioprocesses. Vol. 87, p. 151 Honda, H., Liu, C., Kobayashi, T.: Large-Scale Plant Micropropagation. Vol. 72, p. 157 Hórvath, C. see Freitag, R.: Vol. 53, p. 17 Hou, A. see Drmanac, R.: Vol. 77, p. 75 Hubbell, E. see Hannenhalli, S.: Vol. 77, p. 1 Huebner, S. see Mueller, U.: Vol. 79, p. 137 Hummel, W.: New Alcohol Dehydrogenases for the Synthesis of Chiral Compounds. Vol. 58, p. 145 Ikeda, M.: Amino Acid Production Processes. Vol. 79, p. 1 Imamoglu, S.: Simulated Moving Bed Chromatography (SMB) for Application in Bioseparation. Vol. 76, p. 211 Inada, Y., Matsushima, A., Hiroto, M., Nishimura, H., Kodera, Y.: Chemical Modifications of Proteins with Polyethylen Glycols. Vol. 52, p. 129 Iijima, S. see Miyake, K.: Vol. 90, p. 89 Irwin, D. C. see Wilson, D. B.: Vol. 65, p. 1 Isermann, H. P. see Bungay, H. R.: Vol. 70, p. 109 Iwasaki, Y., Yamane, T.: Enzymatic Synthesis of Structured Lipids. Vol. 90, p. 151 Iyer, P. see Lee, Y. Y.: Vol. 65, p. 93 Jackson, S. see Ferro, A.: Vol. 78, p. 125 James, E., Lee, J. M.: The Production of Foreign Proteins from Genetically Modified Plant Cells. Vol. 72, p. 127 Jeffries, T. W., Shi, N.-Q.: Genetic Engineering for Improved Xylose Fementation by Yeasts. Vol. 65, p. 117 Jendrossek, D.: Microbial Degradation of Polyesters. Vol. 71, p. 293 Jenne, M. see Schmalzriedt, S.: Vol. 80, p. 19 Jin, H. see Drmanac, R.: Vol. 77, p. 75 Jin, P. see Drmanac, R.: Vol. 77, p. 75 Johnson, D.V., Griengl, H.: Biocatalytic Applications of Hydroxynitrile. Vol. 63, p. 31 Johnson, E. A., Schroeder, W. A.: Microbial Carotenoids. Vol. 53, p. 119 Johnsurd, S. C.: Biotechnolgy for Solving Slime Problems in the Pulp and Paper Industry. Vol. 57, p. 311 Johri, B. N., Sharma, A., Virdi, J. S.: Rhizobacterial Diversity in India and its Influence on Soil and Plant Health. Vol. 84, p. 49 Jönsson, L. J. see Hahn-Hägerdal, B.: Vol. 73, p. 53 Joshi, J. B. see Elias, C. B.: Vol. 59, p. 47 Jurinke, C., van den Boom, D., Cantor, C. R., Köster, H.: The Use of MassARRAY Technology for High Throughput Genotyping. Vol. 77, p. 57 Kaderbhai, N. see Shaw, A. D.: Vol. 66, p. 83 Karanth, N. G. see Krishna, S. H.: Vol. 75, p. 119 Karthikeyan, R., Kulakow, P. A.: Soil Plant Microbe Interactions in Phytoremediation. Vol. 78, p. 51 Kataoka, M. see Shimizu, S.: Vol. 58, p. 45

Author Index Volumes 51–90

205

Kataoka, M. see Shimizu, S.: Vol. 63, p. 109 Katzen, R., Tsao, G.T.: A View of the History of Biochemical Engineering. Vol. 70, p. 77 Kawai, F.: Breakdown of Plastics and Polymers by Microorganisms. Vol. 52, p. 151 Kawarasaki, Y. see Nakano, H.: Vol. 90, p. 135 Kell, D. B. see Shaw, A. D.: Vol. 66, p. 83 Kessler, B., Weusthuis, R., Witholt, B., Eggink, G.: Production of Microbial Polyesters: Fermentation and Downstream Processes. Vol. 71, p. 159 Khosla, C. see McDaniel, R.: Vol. 73, p. 31 Khurana, J. P. see Tyagi, A. K.: Vol. 84, p. 91 Kieran, P. M., Malone, D. M., MacLoughlin, P. F.: Effects of Hydrodynamic and Interfacial Forces on Plant Cell Suspension Systems. Vol. 67, p. 139 Kijne, J.W. see Memelink, J.: Vol. 72, p. 103 Kim, B. C. see Gu, M. B.: Vol. 87, p. 269 Kim, D.-I. see Choi, J.W.: Vol. 72, p. 63 Kim, R. see Banks, M. K.: Vol. 78, p. 75 Kim, Y.B., Lenz, R.W.: Polyesters from Microorganisms. Vol. 71, p. 51 Kimura, E.: Metabolic Engineering of Glutamate Production. Vol. 79, p. 37 King, R.: Mathematical Modelling of the Morphology of Streptomyces Species. Vol. 60, p. 95 Kino-oka, M., Nagatome, H., Taya, M.: Characterization and Application of Plant Hairy Roots Endowed with Photosynthetic Functions. Vol. 72, p. 183 Kirk, T. K. see Akhtar, M.: Vol. 57, p. 159 Kjelgren, R. see Ferro, A.: Vol. 78, p. 125 Knorre, W.A. see Bocker, H.: Vol. 70, p. 35 Kobayashi, M. see Shimizu, S.: Vol. 58, p. 45 Kobayashi, S., Uyama, H.: In vitro Biosynthesis of Polyesters. Vol. 71, p. 241 Kobayashi, T. see Honda, H.: Vol. 72, p. 157 Kobayashi, T. see Honda, H.: Vol. 87, p. 151 Kodera, F. see Inada, Y.: Vol. 52, p. 129 Kolattukudy, P. E.: Polyesters in Higher Plants. Vol. 71, p. 1 König, A. see Riedel, K: Vol. 75, p. 81 de Koning, G. J. M. see van der Walle, G. A. M.: Vol. 71, p. 263 Konthur, Z. see Eickhoff, H.: Vol. 77, p. 103 Koo, Y.-M. see Lee, S.-M.: Vol. 87, p. 173 Kossen, N.W.F.: The Morphology of Filamentous Fungi. Vol. 70, p. 1 Köster, H. see Jurinke, C.: Vol. 77, p. 57 Koutinas, A. A. see Webb, C.: Vol. 87, p. 195 Krabben, P., Nielsen, J.: Modeling the Mycelium Morphology of Penicilium Species in Submerged Cultures. Vol. 60, p. 125 Kralovánszky, U.P. see Holló, J.: Vol. 69, p. 151 Krämer, R.: Analysis and Modeling of Substrate Uptake and Product Release by Procaryotic and Eucaryotik Cells. Vol. 54, p. 31 Kretzmer, G.: Influence of Stress on Adherent Cells. Vol. 67, p. 123 Krieger, N. see Mitchell, D.A.: Vol. 68, p. 61 Krishna, S. H., Srinivas, N. D., Raghavarao, K. S. M. S., Karanth, N. G.: Reverse Micellar Extraction for Downstream Processeing of Proteins/Enzymes. Vol. 75, p. 119 Kuhad, R. C., Singh, A., Eriksson, K.-E. L.: Microorganisms and Enzymes Involved in the Degradation of Plant Cell Walls. Vol. 57, p. 45 Kuhad, R. Ch. see Singh, A.: Vol. 51, p. 47 Kulakow, P. A. see Karthikeyan, R.: Vol. 78, p. 51 Kulakow, P. A. see Banks, M. K.: Vol. 78, p. 75 Kumagai, H.: Microbial Production of Amino Acids in Japan. Vol. 69, p. 71 Kumar, R. see Mukhopadhyay, A.: Vol. 86, p. 215 Kumar, S. see Harvey, N. L.: Vol. 62, p. 107 Kunze, G. see Riedel, K.: Vol. 75, p. 81 Kwon, S. see Drmanac, R.: Vol. 77, p. 75

206

Author Index Volumes 51–90

Lacy, S. see Drmanac, R.: Vol. 77, p. 75 Ladenstein, R., Antranikian, G.: Proteins from Hyperthermophiles: Stability and Enzamatic Catalysis Close to the Boiling Point of Water. Vol. 61, p. 37 Ladisch, C. M. see Mosier, N. S.: Vol. 65, p. 23 Ladisch, M. R. see Mosier, N. S.: Vol. 65, p. 23 LaFayette, P. R. see Dean, J. F. D.: Vol. 57, p. 1 Lammers, F., Scheper, T.: Thermal Biosensors in Biotechnology. Vol. 64, p. 35 Larroche, C., Gros, J.-B.: Special Transformation Processes Using Fungal Spares and Immobilized Cells. Vol. 55, p. 179 Latady, M. see Flechas, F. W.: Vol. 78, p. 171 Lazarus, M. see Adam, W.: Vol. 63, p. 73 Leak, D. J. see van der Werf, M. J.: Vol. 55, p. 147 Lee, J.M. see James, E.: Vol. 72, p. 127 Lee, S.-M., Lin, J., Koo, Y.-M.: Production of Lactic Acid from Paper Sludge by Simultaneous Saccharification and Fermentation. Vol. 87, p. 173 Lee, S. Y., Chang, H. N.: Production of Poly(hydroxyalkanoic Acid). Vol. 52, p. 27 Lee, S. Y., Choi, J.: Production of Microbial Polyester by Fermentation of Recombinant Microorganisms. Vol. 71, p. 183 Lee, Y.Y., Iyer, P., Torget, R.W.: Dilute-Acid Hydrolysis of Lignocellulosic Biomass.Vol. 65, p. 93 Lehrach, H. see Eickhoff, H.: Vol. 77, p. 103 Lenz, R. W. see Kim, Y. B.: Vol. 71, p. 51 Licari, P. see McDaniel, R.: Vol. 73, p. 31 Lievense, L. C., van’t Riet, K.: Convective Drying of Bacteria II. Factors Influencing Survival. Vol. 51, p. 71 Lilie, H. see Fahnert, B.: Vol. 89, p. 93 Lin, J. see Lee, S.-M.: Vol. 87, p. 173 Linden, J. C., Haigh, J. R., Mirjalili, N., Phisaphalong, M.: Gas Concentration Effects on Secondary Metabolite Production by Plant Cell Cultures. Vol. 72, p. 27 Lipshutz, R. see Hannenhalli, S.: Vol. 77, p. 1 Lipson, D. S. see Ferro, A.: Vol. 78, p. 125 Little, D. see Drmanac, R.: Vol. 77, p. 75 Liu, B. see Banks, M. K.: Vol. 78, p. 75 Liu, C. see Honda, H.: Vol. 72, p. 157 Lohray, B. B.: Medical Biotechnology in India. Vol. 85, p. 215 Lueking, A. see Eickhoff, H.: Vol. 77, p. 103 Luo, J. see Yang, S.-T.: Vol. 87, p. 61 Lyberatos, G. see Pind, P. F.: Vol. 82, p. 135 MacLoughlin, P.F. see Kieran, P. M.: Vol. 67, p. 139 Macario, A. J. L. see Conway de Macario, E.: Vol. 81, p. 95 Madhusudhan, T. see Mukhopadhyay, A.: Vol. 86, p. 215 Malone, D. M. see Kieran, P. M.: Vol. 67, p. 139 Maloney, S. see Müller, R.: Vol. 61, p. 155 Mandenius, C.-F.: Electronic Noses for Bioreactor Monitoring. Vol. 66, p. 65 Markov, S. A., Bazin, M. J., Hall, D. O.: The Potential of Using Cyanobacteria in Photobioreactors for Hydrogen Production. Vol. 52, p. 59 Marteinsson, V.T. see Prieur, D.: Vol. 61, p. 23 Marx, A. see Pfefferle, W.: Vol. 79, p. 59 Mathur, R. K. see Ghosh, A. C.: Vol. 56, p. 111 Matsushima, A. see Inada, Y.: Vol. 52, p. 129 Mauch, K. see Schmalzriedt, S.: Vol. 80, p. 19 Mazumdar-Shaw, K., Suryanarayan, S.: Commercialization of a Novel Fermentation Concept. Vol. 85, p. 29 McCaskill, D., Croteau, R.: Prospects for the Bioengineering of Isoprenoid Biosynthesis. Vol. 55, p. 107

Author Index Volumes 51–90

207

McDaniel, R., Licari, P., Khosla, C.: Process Development and Metabolic Engineering for the Overproduction of Natural and Unnatural Polyketides. Vol. 73, p. 31 McDonell, T. J. see Bruckheimer, E. M.: Vol. 62, p. 75 McGall, G.H., Christians, F.C.: High-Density GeneChip Oligonucleotide Probe Arrays. Vol. 77, p. 21 McGovern, A. see Shaw, A. D.: Vol. 66, p. 83 McGowan, A. J. see McKenna, S. L.: Vol. 62, p. 1 McIntyre, M., Müller, C., Dynesen, J., Nielsen, J.: Metabolic Engineering of the Aspergillus. Vol. 73, p. 103 McIntyre, T.: Phytoremediation of Heavy Metals from Soils. Vol. 78, p. 97 McKenna, S. L., McGowan, A. J., Cotter, T. G.: Molecular Mechanisms of Programmed Cell Death. Vol. 62, p. 1 McLoughlin, A. J.: Controlled Release of Immobilized Cells as a Strategy to Regulate Ecological Competence of Inocula. Vol. 51, p. 1 Memelink, J., Kijne, J.W., van der Heijden, R., Verpoorte, R.: Genetic Modification of Plant Secondary Metabolite Pathways Using Transcriptional Regulators. Vol. 72, p. 103 Menachem, S. B. see Argyropoulos, D. S. : Vol. 57, p. 127 Menawat, A. S. see Gomes J.: Vol. 59, p. 1 Menge, M. see Mukerjee, J.: Vol. 68, p. 1 Merkle, S. A. see Dean, J. F. D.: Vol. 57, p. 1 Meyer, H. E. see Sickmann, A.: Vol. 83, p. 141 Mirjalili, N. see Linden, J.C.: Vol. 72, p. 27 Mishra, P. see Chand, S.: Vol. 85, p. 95 Mitchell, D.A., Berovic, M., Krieger, N.: Biochemical Engineering Aspects of Solid State Bioprocessing. Vol. 68, p. 61 Mitchell, R. J. see Gu, M. B.: Vol. 87, p. 269 Miyake, K., Iijima, S.: Bacterial Capsular Polysaccharide and Sugar Transferases. Vol. 90, p. 89 Möckel, B. see Pfefferle, W.: Vol. 79, p. 59 Moeur, B. see Drmanac, R.: Vol. 77, p. 75 Mogensen, A. S., Dolfing, J., Haagensen, F., Ahring, B. K.: Potential for Anaerobic Conversion of Xenobiotics. Vol. 82, p. 69 Moore, J.C. see Arnold, F. H.: Vol. 58, p. 1 Moracci, M. see van der Oost, J.: Vol. 61, p. 87 Mosier, N.S., Hall, P., Ladisch, C.M., Ladisch, M.R.: Reaction Kinetics, Molecular Action, and Mechanisms of Cellulolytic Proteins. Vol. 65, p. 23 Mreyen, M. see Sickmann, A.: Vol. 83, p. 141 Mueller, U., Huebner, S.: Economic Aspects of Amino Acids Production. Vol. 79, p. 137 Mühlemann, H.M., Bungay, H.R.: Research Perspectives for Bioconversion of Scrap Paper. Vol. 65, p. 193 Mukherjee, J., Menge, M.: Progress and Prospects of Ergot Alkaloid Research. Vol. 68, p. 1 Mukhopadhyay, A.: Inclusion Bodies and Purification of Proteins in Biologically Active Forms. Vol. 56, p. 61 Mukhopadhyay, A. see Bhatia, P.K.: Vol. 64, p. 155 Mukhopadhyay, A., Basu, S. K.: Intracellular Delivery of Drugs to Macrophages. Vol. 84, p. 183 Mukhopadhyay, A., Madhusudhan, T., Kumar, R.: Hematopoietic Stem Cells: Clinical Requirements and Developments in Ex-Vivo Culture. Vol. 86, p. 215 Müller, C. see McIntyre, M.: Vol. 73, p. 103 Müller, R., Antranikian, G., Maloney, S., Sharp, R.: Thermophilic Degradation of Environmental Pollutants. Vol. 61, p. 155 Müllner, S.: The Impact of Proteomics on Products and Processes. Vol. 83, p. 1 Nagatome, H. see Kino-oka, M.: Vol. 72, p. 183 Nagy, E.: Three-Phase Oxygen Absorption and its Effect on Fermentation. Vol. 75, p. 51 Nakano, H., Kawarasaki, Y., Yamane, T.: Cell-free Protein Synthesis Systems: Increasing their Performance and Applications. Vol. 90, p. 135

208

Author Index Volumes 51–90

Nakashimada, Y. see Nishio, N.: Vol. 90, p. 63 Nath, S.: Molecular Mechanisms of Energy Transduction in Cells: Engineering Applications and Biological Implications. Vol. 85, p. 125 Nebe-Von-Caron, G. see Hewitt, C. J.: Vol. 89, p. 197 Necina, R. see Strancar, A.: Vol. 76, p. 49 Neubauer, P. see Fahnert, B.: Vol. 89, p. 93 Nielsen, J. see Christensen, B.: Vol. 66, p. 209 Nielsen, J. see Krabben, P.: Vol. 60, p. 125 Nielsen, J. see McIntyre, M.: Vol. 73, p. 103 Nisbet, L.J. see Hill, D.C.: Vol. 59, p. 73 Nishimura, H. see Inada, Y.: Vol. 52, p. 123 Nishio, N., Nakashimada, Y.: High Rate Production of Hydrogen/Methane from Various Substrates and Wastes. Vol. 90, p. 63 Nordhoff, E. see Cahill, D.J.: Vol. 83, p. 177 Nordhoff, E. see Eickhoff, H.: Vol. 77, p. 103 Nouwens, A. S., Walsh, B. J., Cordwell S. J.: Application of Proteomics to Pseudomonas aeruginosa. Vol. 83, p. 117 Nyarsik, L. see Eickhoff, H.: Vol. 77, p. 103 Ochsner, U. A., Hembach, T., Fiechter, A.: Produktion of Rhamnolipid Biosurfactants.Vol. 53, p. 89 O’Connor, R.: Survival Factors and Apoptosis: Vol. 62, p. 137 Ogawa, J. see Shimizu, S.: Vol. 58, p. 45 Ohshima, T., Sato, M.: Bacterial Sterilization and Intracellular Protein Release by Pulsed Electric Field. Vol. 90, p. 113 Ohta, H.: Biocatalytic Asymmetric Decarboxylation. Vol. 63, p. 1 Oliverio, S. see Autuori, F.: Vol. 62, p. 129 van der Oost, J., Ciaramella, M., Moracci, M., Pisani, F.M., Rossi, M., de Vos, W.M.: Molecular Biology of Hyperthermophilic Archaea. Vol. 61, p. 87 Orlich, B., Schomäcker, R.: Enzyme Catalysis in Reverse Micelles. Vol. 75, p. 185 Orru, R.V.A., Archelas, A., Furstoss, R., Faber, K.: Epoxide Hydrolases and Their Synthetic Applications. Vol. 63, p. 145 Osbourn, A. E. see Haralampidis, D.: Vol. 75, p. 31 Oude Elferink, S. J. W. H. see Stams, A. J. M.: Vol. 81, p. 31 Padmanaban, G.: Drug Targets in Malaria Parasites. Vol. 84, p. 123 Panda, A.K.: Bioprocessing of Therapeutic Proteins from the Inclusion Bodies of Escherichia coli. Vol. 85, p. 43 Park, E.Y.: Recent Progress in Microbial Cultivation Techniques. Vol. 90, p. 1 Paul, G.C., Thomas, C.R.: Characterisation of Mycelial Morphology Using Image Analysis. Vol. 60, p. 1 Perrier, M. see Dochain, D.: Vol. 56, p. 147 Pevzner, P. A. see Hannenhalli, S.: Vol. 77, p. 1 Pfefferle, W., Möckel, B., Bathe, B., Marx, A.: Biotechnological Manufacture of Lysine. Vol. 79, p. 59 Phisaphalong, M. see Linden, J.C.: Vol. 72, p. 27 Piacentini, G. see Autuori, F.: Vol. 62, p. 129 Pind, P. F., Angelidaki, I., Ahring, B. K., Stamatelatou, K., Lyberatos, G.: Monitoring and Control of Anaerobic Reactors. Vol. 82, p. 135 Piredda, L. see Autuori, F.: Vol. 62, p. 129 Pisani, F.M. see van der Oost, J.: Vol. 61, p. 87 Podgornik, A. see Strancar, A.: Vol. 76, p. 49 Podgornik, A., Tennikova, T. B.: Chromatographic Reactors Based on Biological Activity. Vol. 76, p. 165 Pohl, M.: Protein Design on Pyruvate Decarboxylase (PDC) by Site-Directed Mutagenesis. Vol. 58, p. 15

Author Index Volumes 51–90

209

Poirier, Y.: Production of Polyesters in Transgenic Plants. Vol. 71, p. 209 Pons, M.-N., Vivier, H.: Beyond Filamentous Species. Vol. 60, p. 61 Pons, M.-N., Vivier, H.: Biomass Quantification by Image Analysis. Vol. 66, p. 133 Prazeres, D. M. F. see Fernandes, P.: Vol. 80, p. 115 Prieur, D., Marteinsson, V.T.: Prokaryotes Living Under Elevated Hydrostatic Pressure. Vol. 61, p. 23 Prior, A. see Wolfgang, J.: Vol. 76, p. 233 Pulz, O., Scheibenbogen, K.: Photobioreactors: Design and Performance with Respect to Light Energy Input. Vol. 59, p. 123 Raghavarao, K. S. M. S., Dueser, M., Todd, P.: Multistage Magnetic and Electrophoretic Extraction of Cells, Particles and Macromolecules. Vol. 68, p. 139 Raghavarao, K. S. M. S. see Krishna, S. H.: Vol. 75, p. 119 Ramanathan, K. see Xie, B.: Vol. 64, p. 1 Raskin, L. see Hofman-Bang, J.: Vol. 81, p. 151 Reuss, M. see Schmalzriedt, S.: Vol. 80, p. 19 Riedel, K., Kunze, G., König, A.: Microbial Sensor on a Respiratory Basis for Wastewater Monitoring. Vol. 75, p. 81 van’t Riet, K. see Lievense, L. C.: Vol. 51, p. 71 Rinas, U. see Hoffmann, F.: Vol. 89, p. 73 Rinas, U. see Hoffmann, F.: Vol. 89, p. 143 Roberts, S. M. see Allan, J. V.: Vol. 63, p. 125 Robinson, A. see Brazma, A.: Vol. 77, p. 113 Rock, S. A.: Vegetative Covers for Waste Containment. Vol. 78, p. 157 Roehr, M.: History of Biotechnology in Austria. Vol. 69, p. 125 Rogers, P. L., Shin, H. S., Wang, B.: Biotransformation for L-Ephedrine Production. Vol. 56, p. 33 Rossi, M. see van der Oost, J.: Vol. 61, p. 87 Rowland, J. J. see Shaw, A. D.: Vol. 66, p. 83 Roy, I., Sharma, S., Gupta, M. N.: Smart Biocatalysts: Design and Applications. Vol. 86, p. 159 Roychoudhury, P. K., Srivastava, A., Sahai, V.: Extractive Bioconversion of Lactic Acid. Vol. 53, p. 61 Rozkov, A., Enfors, S.-O.: Analysis and Control of Proteolysis of Recombinant Proteins in Escherichia coli. Vol. 89, p. 163 Rusin, P., Ehrlich, H. L.: Developments in Microbial Leaching – Mechanisms of Manganese Solubilization. Vol. 52, p. 1 Russell, N.J.: Molecular Adaptations in Psychrophilic Bacteria: Potential for Biotechnological Applications. Vol. 61, p. 1 Sablon, E., Contreras, B., Vandamme, E.: Antimicrobial Peptides of Lactic Acid Bacteria: Mode of Action, Genetics and Biosynthesis. Vol. 68, p. 21 Sahai, V. see Singh, A.: Vol. 51, p. 47 Sahai, V. see Roychoudhury, P. K.: Vol. 53, p. 61 Saha-Möller, C. R. see Adam, W.: Vol. 63, p. 73 Sahm, H. see Eggeling, L.: Vol. 54, p. 1 Sahm, H. see de Graaf, A.A.: Vol. 73, p. 9 Sahoo, G. C., Dutta, N. N.: Perspectives in Liquid Membrane Extraction of Cephalosporin Antibiotics: Vol. 75, p. 209 Saleemuddin, M.: Bioaffinity Based Immobilization of Enzymes. Vol. 64, p. 203 Santos, H. see da Costa, M.S.: Vol. 61, p. 117 Sarkans, U. see Brazma, A.: Vol. 77, p. 113 Sarkiss, M. see Bruckheimer, E. M.: Vol. 62, p. 75 Sato, M. see Ohshima, T.: Vol. 90, p. 113 Sauer, U.: Evolutionary Engineering of Industrially Important Microbial Phenotypes. Vol. 73, p. 129 Scheibenbogen, K. see Pulz, O.: Vol. 59, p. 123

210

Author Index Volumes 51–90

Scheper, T. see Lammers, F.: Vol. 64, p. 35 Schmalzriedt, S., Jenne, M., Mauch, K., Reuss, M.: Integration of Physiology and Fluid Dynamics. Vol. 80, p. 19 Schmidt, J. E. see Skiadas, I. V.: Vol. 82, p. 35 Schneider, K. see Beyeler, W.: Vol. 70, p. 139 Schomäcker, R. see Orlich, B.: Vol. 75, p. 185 Schreier, P.: Enzymes and Flavour Biotechnology. Vol. 55, p. 51 Schreier, P. see Adam, W.: Vol. 63, p. 73 Schroeder, W. A. see Johnson, E. A.: Vol. 53, p. 119 Schubert, W.: Topological Proteomics, Toponomics, MELK-Technology. Vol. 83, p. 189 Schügerl, K., Gerlach, S. R., Siedenberg, D.: Influence of the Process Parameters on the Morphology and Enzyme Production of Aspergilli. Vol. 60, p. 195 Schügerl, K. see Seidel, G.: Vol. 66, p. 115 Schügerl, K.: Recovery of Proteins and Microorganisms from Cultivation Media by Foam Flotation. Vol. 68, p. 191 Schügerl, K.: Development of Bioreaction Engineering. Vol. 70, p. 41 Schügerl, K. see Tollnick, C.: Vol. 86, p. 1 Schumann, W.: Function and Regulation of Temperature-Inducible Bacterial Proteins on the Cellular Metabolism. Vol. 67, p. 1 Schuster, K. C.: Monitoring the Physiological Status in Bioprocesses on the Cellular Level. Vol. 66, p. 185 Schwab, P. see Banks, M. K.: Vol. 78, p. 75 Schweder, T., Hecker, M.: Monitoring of Stress Responses. Vol. 89, p. 47 Scouroumounis, G. K. see Winterhalter, P.: Vol. 55, p. 73 Scragg, A. H.: The Production of Aromas by Plant Cell Cultures. Vol. 55, p. 239 Sedlak, M. see Ho, N. W. Y.: Vol. 65, p. 163 Seidel, G., Tollnick, C., Beyer, M., Schügerl, K.: On-line and Off-line Monitoring of the Production of Cephalosporin C by Acremonium Chrysogenum. Vol. 66, p. 115 Seidel, G. see Tollnick, C.: Vol. 86, p. 1 Shafto, J. see Drmanac, R.: Vol. 77, p. 75 Sharma, A. see Johri, B. N: Vol. 84, p. 49 Sharma, M., Swarup, R.: The Way Ahead – The New Technology in an Old Society. Vol. 84, p. 1 Sharma, S. see Roy, I.: Vol. 86, p. 159 Shamlou, P. A. see Yim, S.S.: Vol. 67, p. 83 Shapira, M. see Gutman, A. L.: Vol. 52, p. 87 Sharp, R. see Müller, R.: Vol. 61, p. 155 Shaw, A. D., Winson, M. K., Woodward, A. M., McGovern, A., Davey, H. M., Kaderbhai, N., Broadhurst, D., Gilbert, R. J., Taylor, J., Timmins, E. M., Alsberg, B. K., Rowland, J. J., Goodacre, R., Kell, D. B.: Rapid Analysis of High-Dimensional Bioprocesses Using Multivariate Spectroscopies and Advanced Chemometrics. Vol. 66, p. 83 Shi, N.-Q. see Jeffries, T. W.: Vol. 65, p. 117 Shimizu, K. see Hasegawa, S.: Vol. 51, p. 91 Shimizu, S., Ogawa, J., Kataoka, M., Kobayashi, M.: Screening of Novel Microbial for the Enzymes Production of Biologically and Chemically Useful Compounds. Vol. 58, p. 45 Shimizu, S., Kataoka, M.: Production of Chiral C3- and C4-Units by Microbial Enzymes. Vol. 63, p. 109 Shin, H. S. see Rogers, P. L.: Vol. 56, p. 33 Sickmann, A., Mreyen, M., Meyer, H. E.: Mass Spectrometry – a Key Technology in Proteome Research. Vol. 83, p. 141 Siebert, P. D. see Zhumabayeva, B.: Vol. 86, p. 191 Siedenberg, D. see Schügerl, K.: Vol. 60, p. 195 Singh, A., Kuhad, R. Ch., Sahai, V., Ghosh, P.: Evaluation of Biomass. Vol. 51, p. 47 Singh, A. see Kuhad, R. C.: Vol. 57, p. 45 Singh, R. P., Al-Rubeai, M.: Apoptosis and Bioprocess Technology. Vol. 62, p. 167

Author Index Volumes 51–90

211

Skiadas, I. V., Gavala, H. N., Schmidt, J. E., Ahring, B. K.: Anaerobic Granular Sludge and Biofilm Reactors. Vol. 82, p. 35 Smith, J. S. see Banks, M. K.: Vol. 78, p. 75 Sohail, M., Southern, E. M.: Oligonucleotide Scanning Arrays: Application to High-Throughput Screening for Effective Antisense Reagents and the Study of Nucleic Acid Interactions. Vol. 77, p. 43 Sonnleitner, B.: New Concepts for Quantitative Bioprocess Research and Development.Vol. 54, p. 155 Sonnleitner, B.: Instrumentation of Biotechnological Processes. Vol. 66, p. 1 Southern, E. M. see Sohail, M.: Vol. 77, p. 43 Srinivas, N. D. see Krishna, S. H.: Vol. 75, p. 119 Srivastava, A. see Roychoudhury, P. K.: Vol. 53, p. 61 Stafford, D.E., Yanagimachi, K.S., Stephanopoulos, G.: Metabolic Engineering of Indene Bioconversion in Rhodococcus sp. Vol. 73, p. 85 Stamatelatou, K. see Pind, P. F.: Vol. 82, p. 135 Stams, A. J. M., Oude Elferink, S. J. W. H., Westermann, P.: Metabolic Interactions Between Methanogenic Consortia and Anaerobic Respiring Bacteria. Vol. 81, p. 31 Stark, D., von Stockar, U.: In Situ Product Removal (ISPR) in Whole Cell Biotechnology During the Last Twenty Years. Vol. 80, p. 149 Stefuca, V., Gemeiner, P.: Investigation of Catalytic Properties of Immobilized Enzymes and Cells by Flow Microcalorimetry. Vol. 64, p. 69 Steinbüchel, A., Hein, S.: Biochemical and Molecular Basis of Microbial Synthesis of Polyhydroxyalkanoates in Microorganisms. Vol. 71, p. 81 Stephanopoulos, G., Gill, R.T.: After a Decade of Progress, an Expanded Role for Metabolic Engineering. Vol. 73, p. 1 Stephanopoulos, G. see Stafford, D. E.: Vol. 73, p. 85 von Stockar, U., van der Wielen, L. A. M.: Back to Basics: Thermodynamics in Biochemical Engineering. Vol. 80, p. 1 von Stockar, U. see Stark, D.: Vol. 80, p. 149 Straathof, A. J. J. see Bruggink, A.: Vol. 80, p. 69 Strancar, A., Podgornik, A., Barut, M., Necina, R.: Short Monolithic Columns as Stationary Phases for Biochromatography. Vol. 76, p. 49 Suehara, K., Yano, T.: Bioprocess Monitoring Using Near-Infrared Spectroscopy. Vol. 90, p. 173 Suryanarayan, S. see Mazumdar-Shaw, K.: Vol. 85, p. 29 Suurnäkki, A., Tenkanen, M., Buchert, J., Viikari, L.: Hemicellulases in the Bleaching of Chemical Pulp. Vol. 57, p. 261 Svec, F.: Capillary Electrochromatography: a Rapidly Emerging Separation Method. Vol. 76, p. 1 Svec, F. see Xie, S.: Vol. 76, p. 87 Swanson, D. see Drmanac, R.: Vol. 77, p. 75 Swarup, R. see Sharma, M.: Vol. 84, p. 1 Tabata, H.: Paclitaxel Production by Plant-Cell-Culture Technology. Vol. 87, p. 1 Tanaka, T. see Taniguchi, M.: Vol. 90, p. 35 Tang, Y.-J. see Zhong, J.-J.: Vol. 87, p. 25 Taniguchi, M., Tanaka, T.: Clarification of Interactions Among Microorganisms and Development of Co-culture System for Production of Useful Substances. Vol. 90, p. 35 Taya, M. see Kino-oka, M.: Vol. 72, p. 183 Taylor, J. see Shaw, A. D.: Vol. 66, p. 83 Tenkanen, M. see Suurnäkki, A.: Vol. 57, p. 261 Tennikova, T. B. see Podgornik, A.: Vol. 76, p. 165 Thiericke, R. see Grabely, S.: Vol. 64, p. 101 Thomas, C. R. see Paul, G. C.: Vol. 60, p. 1 Thömmes, J.: Fluidized Bed Adsorption as a Primary Recovery Step in Protein Purification. Vol. 58, p. 185

212

Author Index Volumes 51–90

Timmens, E. M. see Shaw, A. D.: Vol. 66, p. 83 Todd, P. see Raghavarao, K. S. M. S.: Vol. 68, p. 139 Tolan, J. S., Guenette, M.: Using Enzymes in Pulp Bleaching: Mill Applications.Vol. 57, p. 289 Tolan, J. S., Foody, B.: Cellulase from Submerged Fermentation. Vol. 65, p. 41 Tollnick, C. see Seidel, G.: Vol. 66, p. 115 Tollnick, C., Seidel, G., Beyer, M., Schügerl, K.: Investigations of the Production of Cephalosporin C by Acremonium chrysogenum. Vol. 86, p. 1 Torget, R. W. see Lee, Y. Y.: Vol. 65, p. 93 Traganos, F. see Darzynkiewicz, Z.: Vol. 62, p. 33 Trojanowska, M. see Haralampidis, D.: Vol. 75, p. 31 Tsao, D. T.: Overview of Phytotechnologies. Vol. 78, p. 1 Tsao, G. T., Cao, N. J., Du, J., Gong, C. S.: Production of Multifunctional Organic Acids from Renewable Resources. Vol. 65, p. 243 Tsao, G. T. see Gong, C. S.: Vol. 65, p. 207 Tsao, G.T. see Katzen, R.: Vol. 70, p. 77 Tyagi, A. K., Dhar, N.: Recent Advances in Tuberculosis Research in India. Vol. 84, p. 211 Tyagi, A. K., Khurana, J. P.: Plant Molecular Biology and Biotechnology Research in the PostRecombinant DNA Era. Vol. 84, p. 91 Ukrainczyk, T. see Drmanac, R.: Vol. 77, p. 75 Uyama, H. see Kobayashi, S.: Vol. 71, p. 241 VanBogelen, R. A.: Probing the Molecular Physiology of the Microbial Organism, Escherichia coli using Proteomics. Vol. 83, p. 27 Vandamme, E. see Sablon, E.: Vol. 68, p. 21 Verpoorte, R. see Memelink, J.: Vol. 72, p. 103 Viikari, L. see Suurnäkki, A.: Vol. 57, p. 261 Vilo, J. see Brazma, A.: Vol. 77, p. 113 Vingron, M. see Brazma, A.: Vol. 77, p. 113 Virdi, J. S. see Johri, B. N: Vol. 84, p. 49 Vivier, H. see Pons, M.-N.: Vol. 60, p. 61 Vivier, H. see Pons, M.-N.: Vol. 66, p. 133 de Vos, W.M. see van der Oost, J.: Vol. 61, p. 87 Wahlbom, C.F. see Hahn-Hägerdal, B.: Vol. 73, p. 53 Wall, M. B. see Farrell, R. L.: Vol. 57, p. 197 van der Walle, G. A. M., de Koning, G. J. M., Weusthuis, R. A., Eggink, G.: Properties, Modifications and Applications of Biopolyester. Vol. 71, p. 263 Walsh, B. J. see Nouwens, A.S.: Vol. 83, p. 117 Walter, G. see Eickhoff, H.: Vol. 77, p. 103 Wang, B. see Rogers, P. L.: Vol. 56, p. 33 Wang, R. see Webb, C.: Vol. 87, p. 195 Webb, C., Koutinas, A. A., Wang, R.: Developing a Sustainable Bioprocessing Strategy Based on a Generic Feedstock. Vol. 87, p. 195 Weichold, O. see Adam, W.: Vol. 63, p. 73 van der Werf, M. J., de Bont, J. A. M. Leak, D. J.: Opportunities in Microbial Biotransformation of Monoterpenes. Vol. 55, p. 147 Westermann, P. see Hofman-Bang, J.: Vol. 81, p. 151 Westermann, P. see Stams, A. J. M.: Vol. 81, p. 31 Weuster-Botz, D., de Graaf, A. A.: Reaction Engineering Methods to Study Intracellular Metabolite Concentrations. Vol. 54, p. 75 Weusthuis, R. see Kessler, B.: Vol. 71, p. 159 Weusthuis, R. A. see van der Walle, G. J. M.: Vol. 71, p. 263 Wick, L. M., Egli, T.: Molecular Components of Physiological Stress Responses in Escherichia coli. Vol. 89, p. 1

Author Index Volumes 51–90

213

Wiechert, W., de Graaf, A. A.: In Vivo Stationary Flux Analysis by 13C-Labeling Experiments. Vol. 54, p. 109 van der Wielen, L. A. M. see Bruggink, A.: Vol. 80, p. 69 van der Wielen, L. A. M. see von Stockar, U.: Vol. 80, p. 1 Wiesmann, U.: Biological Nitrogen Removal from Wastewater. Vol. 51, p. 113 Williamson, N. M. see Allan, J. V.: Vol. 63, p. 125 Wilson, D. B., Irwin, D. C.: Genetics and Properties of Cellulases. Vol. 65, p. 1 Winson, M. K. see Shaw, A. D.: Vol. 66, p. 83 Winterhalter, P., Skouroumounis, G. K.: Glycoconjugated Aroma Compounds: Occurence, Role and Biotechnological Transformation. Vol. 55, p. 73 Witholt, B. see Kessler, B.: Vol. 71, p. 159 Wolfgang, J., Prior, A.: Continuous Annular Chromatography. Vol. 76, p. 233 Woodley, J. M.: Advances in Enzyme Technology – UK Contributions. Vol. 70, p. 93 Woodward, A. M. see Shaw, A. D.: Vol. 66, p. 83 Wrigley, S. K. see Hill, D. C.: Vol. 59, p. 73 Xia, L. see Cen, P.: Vol. 65, p. 69 Xie, B., Ramanathan, K., Danielsson, B.: Principles of Enzyme Thermistor Systems: Applications to Biomedical and Other Measurements. Vol. 64, p. 1 Xie, S., Allington, R. W., Fréchet, J. M. J., Svec, F.: Porous Polymer Monoliths: An Alternative to Classical Beads. Vol. 76, p. 87 Xu, C. see Drmanac, R.: Vol. 77, p. 75 Yamane, T. see Iwasaki, Y.: Vol. 90, p. 135 Yamane, T. see Nakano, H.: Vol. 90, p. 89 Yanagimachi, K.S. see Stafford, D.E.: Vol. 73, p. 85 Yang, S.-T., Luo, J., Chen, C.: A Fibrous-Bed Bioreactor for Continuous Production of Monoclonal Antibody by Hybridoma. Vol. 87, p. 61 Yano, T. see Suehara, K.: Vol. 90, p. 173 Yim, S. S., Shamlou, P. A.: The Engineering Effects of Fluids Flow and Freely Suspended Biological Macro-Materials and Macromolecules. Vol. 67, p. 83 Zhang, S., Chu, J., Zhuang, Y.: A Multi-Scale Study on Industrial Fermentation Processes and Their Optimization. Vol. 87, p. 97 Zheng, D. see Hofman-Bang, J.: Vol. 81, p. 151 Zhong, J.-J.: Biochemical Engineering of the Production of Plant-Specific Secondary Metabolites by Cell Suspension Cultures. Vol. 72, p. 1 Zhong, J.-J., Tang, Y.-J.: Submerged Cultivation of Medicinal Mushrooms for Production of Valuable Bioactive Metabolites. Vol. 87, p. 25 Zhuang, Y. see Zhang, S.: Vol. 87, p. 97 Zhumabayeva, B., Chenchik, A., Siebert, P. D., Herrler, M.: Disease Profiling Arrays: Reverse Format cDNA Arrays Complimentary to Microarrays. Vol. 86, p. 191 Zollinger, N. see Ferro, A.: Vol. 78, p. 125 van Zyl, W. H. see Hahn-Hägerdal, B.: Vol. 73, p. 53

Subject Index

Acetate 65, 68–74, 78, 79 Acetic acid 40, 54, 178 Aceticlastic methanogen 64, 67 Acetobacter aceti 4 Acetogen 64 b-1,3-N-Acetylglucosaminyltransferase 101, 102, 105 N-Acetylneuramic acid 166 Acid phosphatase, wheat germ 142 Acidogenesis 64, 65, 82 Acidolysis 155 Activated sludge process 67 Active species 123 Acylation 158, 161 ADH 124 Alcohol distillery wastewater 67 Amensalism 37 a-Amylase 125 Anaerobic attached film expanded-bed reactor (AAFEB) 65 Anaerobic fluidized-bed reactor (AFBR) 65, 67, 68 Antigen, O- 100 Aqueous suspension system 167 Arachidonic acid 155 Bacillus subtilis 14 Bean curd refuse 73 Bifidobacterium adolescensis 52 Bifidobacterium longum 40, 51 Bifidogenic growth stimulator (BGS) 51 Biodegradable polymer 42 Bioprocess management/monitoring 174 Bioremediation 72 Biosurfactants 195 Biosynthesis 92, 96 Biotech analyzer, automatic 16 Biphasic system 163, 164 Blastochloris sulfoviridis 69 Bonito oil 156 Branch formation rate 30 Bread waste 84

Breakdown voltage 114 Butyrate 66, 74 C/N ratio 187 Calcium sulfate 167 5¢-Cap 142 Cap-independent translation enhancer (CITE) 142, 143, 145 Caprylic acid 155 Carbon dioxide 65, 76, 78, 79 Cell recycling bioreactor 2 Cell-free expression 139 Cell-free protein synthesis system 136–139 Cell-free translation 136 Cellobiohydrolase 125 Cellulose 42, 64, 70, 71 Center-of-gravity method 9 Chloramphenicol acetyltransferase (CAT) 143–146 Clostridium 76, 81 CMP-sialic acid 102 Co-culture system 47, 51 COD removal 67, 70, 73 Coiled wire-to-cylinder 122 Commensalism 37 Competition 37, 38 Compost fermentation 185 Condensation reaction 161 Continuous-Flow Cell Free (CFCF) system 136 Cooperative interactions 51, 54, 59 Coupled transcription/translation system 136, 138 CP polymerase 102 Crabtree effect 7 Digestion, anaerobic 64, 73 Diglyceride prodrug (DG prodrug) 154 Dihydrofolate reductase (DHFR) 143 Dihydroxyacetone 166 Dilution rate 68, 75, 81, 84

216 Disulfide bond 137 DNA library 139 DNA polymerase, hot-startable 138 DNA topoisomerase 102 DO 2 Docosahexaenoic acid (DHA) 155, 161 Docosapentaenoic acid (DPA) 155 DO-stat, balanced 7 Downflow hanging sponge-cube (DHS) 69 Eicosapentaenoic acid (EPA) 156, 161 ELAM-1 92, 94, 96 Electrical breakdown 114 Electromechanical compression 114 Endothelial cells 92, 94, 96 Energy recovery 64 Escherichia coli 7 – –, cell-free protein synthesis system 137 – –, S30 extract 136, 145, 146 Essential fatty acid 154 Esterification 158 Ethanol 40, 65, 66, 70, 71, 78, 79, 178 Ethanolysis 155, 159 Eukaryotic cell-free translation system 141, 142 Exopolysaccharides 101 Fab fragment 137, 138 Facultative anaerobes 64, 65, 75, 76 Fatty acid 153, 161 – –, polyunsaturated (PUFA) 154, 161, 163 Fed-batch culture 4 Feedforward/feedback control 5 Feedforward glucose feed rate 5 Flow-through chamber 28 Fluorescin isothiocyanate (FITC) 23 5-Fluorouridine 166 Fuel cell 65 Fuzzy control theory 8

b-Galactosidase 14 b-1,3-Galactosyltransferase 107 b-1,4-Galactosyltransferase 101, 102, 105 Ganoderma lucidum 194 Garbage 71 GBS (Group B stretococci) 90, 92, 94, 96, 102 Gel filtration chromatography 94 Gene expression 138, 146, 147 Glucose 66, 71, 75–78, 81 – / xylose 42, 43 Glucose control 1 Glucose measurement system, on-line 14 b-Glucosidase 125 Glucosyltransferase 100, 102, 105

Subject Index Glutamic acid fermentation 184 Green Fluorescent Protein (GFP) 140, 143, 145 Hemicellulose 42, 70 High cell density 2 High-throughput analysis/screening 137, 141, 142 His-tag 145, 146 Hot-startable DNA polymerase 138 Hydraulic retention time (HRT) 67, 70, 76, 80 Hydrogen 38 Hydrogen production 64, 76, 77, 81, 83 – –, biological 64, 74 – –, fermentative 64, 74 – –, photosynthetic 74 Hydrogen-methane two-stage process 64, 84 Hydrolysis 64, 65, 159, 161, 163 Hydroxybutyrate (PHB) 42 Hyphal growth rate, specific 28 Image analysis technique 19 Immobilized lipase 155, 156, 169 Invertase 124 Irreversible disruption 114 Kefiran 41, 55 2-Ketoglutaric acids 178 Kojic acid 165 Kurtzumanomyces sp. 195 Lactic acid 40, 51 – – fermentation 185 Lactobacillus kefiranofaciens 41, 55 Lactose 51, 55 Lecithin 163, 164, 167 Lighted upflow anaerobic sludge blanket (LUASB) 69 Lignocellulosic biomass 42 g-Linoleic acid 155 Lipase 137, 140, 154, 155, 161, 169 Lipid carrier 92, 107, 109 Lipopolysaccharides 100, 101 Loading rate 65, 66, 73 Long chain fatty acid (LCFA) 153 Low-calorie fat 153 Lysophosphatidylcholine (lysoPC) 161, 163 Lysophospholipid (lysoPL) 161, 163 LysR/LytR 99, 100 Mamdani’s min-max algorithm 9 Manganese peroxidase 137, 140 Mannosyl erythritol lipid (MEL) 195

Subject Index Medium chain fatty acid (MCFA) 153 Medium chain triacylglycerol (MCTG) 154 Membership function 9 Membrane, cytoplasmic 126 –, outer 126 Membrane potential 114 Mesophilic 65, 67 Metastasis 92, 94, 96 Methane 39, 64 – fermentation 64, 70, 73, 74 – yield 66, 73 Methanobacterium formicum 70 Methanogen, hydrogenotrophic 64 Methanogenesis 64, 65 Methanogenic ecosystem 65 Methanol 65, 66 Methanosaeta sp. 67, 70 MICOC 5 Microbial consortia 64, 65 Microscope 28 MLM 153, 154, 159 MLR 173, 175 Molecular dissection 147 Monoacylglycerol (MG) 154 Mortierella 26 mRNA, eukaryotic 142 Mud sediment 68 Municipal sewage sludge 70 Mutant protein 140 Mutualism 37 Mycelial morphology 19 Near-infrared spectroscopy (NIR) 174–176 Needle-plate electrode 120 Neocallimastix frontalis 70 Neutralism 37 Nisin 41, 119 Nitrification 69 Nondestructive analysis 174 Open reading frame (ORFs) 96, 101, 102 OPO 153 Optimal control 5 Oxygen transfer rate (OTR) 43 Oxygen uptake rate, specific 43 Ozonized water 118 Parasitism 38 Pathogens 90 PCR product 137 –, multiple molecules 140 –, single molecule PCR 138, 140 –, single primer PCR 138

217 PEF (pulsed electric field) 114 –, concentrated 122 PEF energy 116 PEF sterilization 114 Pellet intrastructure 23 PGK 18 Phase transition temperature 117 Phenol 65, 66 Phosphatidic acid (PA) 163 Phosphatidylarbutin 165 Phosphatidylascorbic acid 165 Phosphatidylcholine (PC) 161, 164, 165, 167 Phosphatidylchromanol 165 Phosphatidylethanolamine (PE) 161, 164 Phosphatidylglycerol (PG) 161, 164 Phosphatidylinositol (PI) 161, 165 Phosphatidylkojic acid 165 Phosphatidylserine (PS) 161, 164, 167 Phospholipase (PL) 153, 160, 161, 169 Phospholipase A2 (PLA2) 161, 163 Phospholipase D (PLD) 137, 163, 164, 167 Phototrophic bacteria 69 Pichia stipitis 40, 43 Pickled-plum effluent 67, 68 Picornavirus 143 Plate-plate electrode 120 Polysaccharide 41, 55 Polysialic acid 94, 101 Polyunsaturated fatty acid (PUFA) 154, 161, 163 Polyvirus 143 Predation 38 Primer dimer 138, 139 Prionibacterium freudenreichii 40, 41, 51 Probiotics 51 Promoter, inducible 14 Propidium iodide (PI) 23 Propionate 66, 71, 74 Propionic acid 40, 54 Protein library 138, 139, 141 Protocooperation (synergism) 37, 60 Pulsed electric field (PEF) 114, 122 PURE SYSTEM 137 Respiratory deficient mutants 40, 44 Reticulocyte lysate, rabbit 136 Reversible disruption 114 Rhodopseudomonas palustris 69 Ribosome 136 Ribulose bis-phosphate decarboxylase 147 Rice homeobox protein 147 Rice vinegar fermentation 178 Ring-to-cylinder 122

218 Saccharomyces cerevisiae 7, 40, 43 scFv 137 Sea mud sediment 64, 73 Secretion 129 Selective release 124 Selectivity 163, 164 Sequential conversion 40, 59 Sewage treatment 67, 69 Short chain fatty acid (SCFA) 153 Sialyl Lewis carbohydrates 90, 92, 109 Sialyllactosamine 96, 109 Sialyltransferase 101, 107, 108 SIMPLEX-based protein library 139, 140 SIMPLX (single molecule PCR linked in vitro expression) 138–141 Solid organic waste 64, 65, 73 SOS 153 Soy sauce refuse 64, 73 Space velocity (SV) 65–67 Starch 41, 66, 76 STGs, specific 153 Streptococcus agalactiae 90, 92, 94, 96, 99–109 Streptococcus cremoris 4 Streptococcus pneumoniae 90, 92, 96, 100–103, 109 Streptococcus pyogenes 90, 102 Streptomyces fradiae 20 Structured lipid (SL) 153 Structured phospholipid (SPL) 153, 165 Structured triacylglycerol (STG) 153, 154 –, specific 153 SUC2 17 Sucrose 66 Survival ratio 116 Symbiosis 37, 60 Synergism 37, 60

Subject Index TE(37-65) 143, 145, 146, 147 Thermophiles 67, 73 Thermophilic composting 187 Thin layer chromatography (TLC) 103, 105 Tip extension rate 30 Tip formation rate, specfic 28 Tobacco etch virus 143, 144 TOC removal 67, 68 Transcription 96, 99, 100 – terminator, rho-independent 96 Translation, in vitro 143 Transmembrane pores 114 Transphosphatidylation 163–167 Triacylglycerol (TG) 153 Tridocosahexaenoylglycerol 159 Tuna oil 155 Two-phase digestion 65 Two-stage process 64, 73 Tylosin 20 UDP-galactose 105, 107 UDP-glucose 105 Upflow anaerobic filter process (UAFP) 65 Upflow anaerobic sludge blanket (UASB) 65, 67–70, 73, 84 5¢-UTR (5¢-untranslated region) 142, 143, 145 Vinegar fermentation, rice 178 Vitamin B12 41 Volumetric loading rate 67 Wastewater 64 Wheat germ acid phosphatase 142 Wheat germ extract 136, 142, 146 Xylan 70, 75 Yard waste 71 Yield 163