Immobilized cells

IMMOBILIZED CELLS:Basics and Applications

Progress in Biotechnology Volume 1 New Approaches to Research on Cereal Carbohydrates (Hill and Munck, Editors) Volume 2 Biology of Anaerobic Bacteria (Dubourguier et al.. Editors) Volume 3 Modifications and Applications of Industrial Polysaccharides (Yalpani, Editor) Volume 4 Interbiotech '87. Enzyme Technologies (Blazej and Zemek, Editors) Volume 5 In Vitro Immunization in Hybridoma Technology (Borrebaeck, Editor) Volume 6 Interbiotech '89. Mathematical Modelling in Biotechnology (Blazej and Ottova, Editors) Volume 7 Xylans and Xylanases (Visser et al.. Editors) Volume 8 Biocatalysis in Non-Conventional Media (Tramper et al.. Editors) Volume 9 ECB6: Proceedings of the 6th European Congress on Biotechnology (Alberghina et al., Editors) Volume 10 Carbohydrate Bioengineering (Petersen et al., Editors) Volume 11 Immobilized Cells: Basics and Applications (Wijffels et al.. Editors)

Progress in Biotechnology 11

IMMOBILIZED CELLS: Basics and ApplicationsProceedings o f an International Symposium organized under auspices o f The Working Party on Applied Biocatalysis o f the European Federation o f Biotechnology Noordwijkerhout, The Netherlands November 26-29, 1995

Edited byR.H. WijffelsFood and Bioprocess Engineering Group, Wageningen Agricultural University, Wageningen, The Netherlands

R. M. BuitelaarAgrotechnological Research Institute, ATO-DLO, Wageningen, The Netherlands

C. BuckeUniversity of Westminster, School of Biological and Health Sciences, London, United Kingdom

J. TramperFood and Bioprocess Engineering Group, Wageningen Agricultural University, Wageningen, The Netherlands

ELSEVIER Amsterdam - Lausanne - New York - Oxford - Shannon - Tokyo 1996

Published by: Elsevier Science B.V. P.O. Box 211 1000 AE Amsterdam The Netherlands

ISBN 0-444-81984-3 1996 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher, Elsevier Science B.V., Permissions Department,, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, the Publisher recommends that independent verification of diagnoses and drug dosages should be made. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01293, USA. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. Al other copyright questions, including photocopying outside the USA should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. This book is printed on acid-free paper. Printed in the Netherlands.

PrefaceThis publication contains full papers of both oral and poster presentations of the symposium "Immobilized Cells: Basics and Applications" that was held in Noordwijkerhout, The Netherlands, 26-29 November 1995. Industrial processes with micro-organisms are generally based on the exploitation of free, growing cells. Alternatively, immobilized cells can be used. In that case cells are attached to or entrapped in an inert support. In a continuously operated bioreactor medium containing the substrate will be supplied. Substrate will be converted into a product by the immobilized cells and product and remaining substrate disappear with the outflowing medium. The immobilized cells are retained easily in the bioreactor and as such utilized continuously. In this way the capacity of the process is independent of the growth rate of the micro-organisms involved. Especially in cases where cells grow slowly immobilized-cell processes are more advantageous than processes with suspended cells. Also for very specific situations where the presence of biomass in the product should be prevented (e.g. champagne), drugs should be dosed gradually to the medium (e.g. islets of Langerhans), phage infections in starter cultures should be prevented (e.g. lactic-acid bacteria for cheese production) and plasmids should be stabilized (for application of genetically engineered cells in a continuous mode), immobilized cells perform better than suspended cells. The physiology of immobilized cells has been studied widely last decade. In 1990, during the symposium "Physiology of Immobilized Cells" in Wageningen, an extensive overview of this field was given. Until now, physiological knowledge did not result in wide use of such processes in practice. Since research in the field of immobilized cells started, an enormous quantity of papers have been published. For many processes the complex physiology in a heterogeneous environment is becoming clear now. In addition it is shown that in many processes it is more efficient to use immobilized cells than suspended cells. Nevertheless, applications of immobilized cells in industrial processes are limited. For this reason it was considered essential by the organizers of the symposium to cover the path from basic physiological research to applications and bring together scientists from different disciplines from academia, industry and research institutes. For applications, physiology needs to be integrated with engineering. The goal of the symposium "Immobilized Cells: Basics and Applications" was to relate basic research to applications. Another aim was to extract guidelines for characterization of immobilized cells in view of process design and application from the contributions. Reviews and recent developments of basic research methods essential for applications, bioreactors developed for applications and applications of immobilized-cell processes were discussed in 4 sessions (oral and poster presentations). Both academic and industrial research was presented in these sessions: Basics 1: physics Physical aspects of support materials in relation to fermentations or release of components are key factors in the path from basic research to applications.

Basic physical characteristics of immobilized cells and support materials such as immobilization techniques (including large scale), stability of support materials and interactions of support and media (e.g. diffusion coefficients) were discussed. This session was directed to methods for determination of these physical characteristics and to the key factors for application. Basics 2: physiology, mass transfer and dynamic modelling The physiology of immobilized cells and transfer of components through support materials have been studied widely. These basic aspects have been integrated in several dynamic models. This session was directed to the essential physiological and mass transfer aspects that are important for application of immobilized cells and to their incorporation in dynamic models, including methods for validation of these models. In addition to that, physiology of immobilized and co-immobilized bacteria, fungi, yeast, lichen, micro-algae and recombinant micro-organisms was presented. Immobilized-cell reactors Bioreactors for cultivation of immobilized cells are different in some respects from reactors for suspended cells. Essential characteristics of such immobilized-cell reactors were discussed. In this respect scale-up of the bioreactors is an essential requisite in the trail from basic research to applications. Applications Some processes with immobilized cells have been applied in practice. In this session a few of these processes in food technology, environmental technology, clinical applications, production of enzymes and amino acids have been discussed in more detail. Important phases in the development of the processes were high-lighted. Additionally, promising future applications were identified. The manuscripts presented in these proceedings give an extensive and recent overview of the research and applications of immobilized-cell technology. We hope the manuscripts will also stimulate researchers from the biological disciplines to implement in their research strategy questions that will be raised if the processes are scaled up (think big!), and industry and researchers studying scale up discuss scale-up aspects in an early stage of the research with researchers from the biological disciplines (scale-up by scaling down!). In addition the organizers hope that the integrated approach presented at this conference will stimulate universities to implement this strategy in their educational programmes as well. Helpful tools are formed by the guidelines for the characterization of immobilized cells!

The Editors, Wageningen and London, December 1995.

AcknowledgementsThe organization committee of the international symposium "ImmobiUzed Cells: Basics and Applications" acknowledges with gratitude the following organizations, that generously contributed to this symposium. Agrotechnological Research Institute, ATO-DLO, The Netherlands Applikon Dependable Instruments, The Netherlands Bavaria, The Netherlands DSM Research, The Netherlands Commission of the European communities Foundation for Biotechnology in the Netherlands Gist-brocades, The Netherlands Hitachi Plant Engineering & Construction, Japan Moet & Chandon, France Pharmacia, Sweden Solvay Duphar, The Netherlands Unilever, The Netherlands Wageningen Agricultural University, Division of Food Science and Nutrition, The Netherlands

This Page Intentiona Left Blank

ContentsPreface Acknowledgements v vii

Keynote lectureWhy immobilize? C. Wandrey 3

Basics 1: physicsAlginate gels - Some structure-function correlations relevant to their use as immobilization matrix for cells B. Thu, O. Smidsr^d, G. Skjak-Braek Determination of biofilm diffusion coefficients using micro-electrodes E.E. Beuling, J.C. van den Heuvel, S.P.P. Ottengraf Complexity and heterogeneity of microenvironments in immobilized systems J.E. Nava Saucedo, C. Roisin, J.-N. Barbotin Fundamentals of dispersion in encapsulation technology D. Poncelet, R.J. Neufeld Stable support materials for immobilization of viable cells A. Muscat, U. PriiBe, K.-D. Vorlop Evaluation of density function of pore size distribution of calcium pectate hydrogel M. Polakovic Effects of formation conditions on size distribution of thermogel beads for cell immobilization Cs. Sisak, T. Blickle, Zs. Ulbert, B. Szajani Calcium pectate gel could be a better alternative to calcium alginate gel in multiple applications of immobilized cells P. Gemeiner, J. Nahalka, A. Vikartovska, J. Nahalkova, M. Tomaska, E. Sturdik, O. Markovic, A. Malovikova, I. Zatkova, M. Ilavsky Adhesion of Lactococcus lactis diacetylactis to surfaces S. Bourassa, J.-C. Vuillemard, P. Rouxhet

19 31 39 47 55 62

70

76

84

Immobilization of Arthrobacter globiformis 193 cells into PVA cryogel. Dehydrogenation of steroid substrates V. Fokina, N. Suzina, A. Arinbasarova, A. Zubov, V.I. Lozinsky, K. Koshcheyenko Screening of immobilization materials for anaerobic wastewater treatment C.-S. Hwu, S.-K. Tseng Ecologically pure process of acetate synthesis on diverse gaseous substrates by homoacetogenic bacteria, entrapped in poly(vinyl alcohol) cryogel A.M. Ryabokon, M.V. Kevbrina, M.A. Pusheva, A.L. Zubov, V.I. Lozinsky, E.I. Rainina Entrapment of Zymomonas mobilis cells into PVA-cryogel carrier in the presence of polyol cryoprotectants V.I. Lozinsky, A.L. Zubov, T.A. Makhlis Some microorganisms during their entrapment in PAAG act as "biological accelerators" in how they affect the gel-formation rate V.I. Lozinsky, A.S. Savvichev, B.L. Tumansky, D.I. Nikitin Rhodococcus sp. immobilized by adsorption on chitin G. Bianchi, L. Setti, G. Spagna, P.G. Pifferi Immobilization of Aspergillus niger and Phanerochaete chrysosporium on polyurethane foam A. Sanroman, G. Feijoo, J.M. Lema Effect of C/N ratio and cellulose type on the cellulolytic activity of free and immobilized Trichoderma reesei V. Jirku Analysis of fungal B-glucanase system produced under conditions of mycelial immobilization V. Jirku Immobilization as a contact stimulation of yeast cell wall alterations V. Jirku Modelling the immobilization of cells in a packed bed of porous carriers R. Willaert, L. De Backer, G.V. Baron Porous silicone rubber as an immobilization matrix for microbial and mammalian cells: natural immobilization of a mass transfer limited culture environment A.J. Knights

90 98

106

112

118 126

132

136

140 148

154

162

Continuous fermentation by conventional and recombinant Saccharomyces cerevisiae immobilized in Ca-alginate beads hardened with trivalent ion E. Roca, N. Meinander, M J . Nunez, B. Hahn-Hagerdal, J.M. Lema Encapsulation by polyelectrolyte complex formation - a way to make hepatocyte cultures safe, efficient and on-line available H. Dautzenberg, J. Stange, S. Mitzner, B. Lukanoff Influence of the microenvironment on immobilized Gibberella fujikuroi C. Roisin, C. Bienaime, J.E. Nava Saucedo, J.-N. Barbotin Local mass transfer coefficients in bacterial biofilms using fluorescence recovery after photobleaching (FRAP) J.D. Bryers, F. Drummond Characteristics and selection criteria of support materials for immobilization of nitrifying bacteria EJ.T.M. Leenen, V.A.P.M. dos Santos, J. Tramper, R.H. Wijffels Stability of alginate gels applied for cell entrapment in open systems C. Vogelsang, K. 0stgaard Studies of cell viability of immobilized Chlamydomonas reinhardtii and glycerol photoproduction R. Leon, J.A. Pizzano, F. Galvan Effect of drying medium on the viability of dried Lactobacillus helveticus CNRZ 303 immobilized in calcium-alginate beads E. Selmer-Olsen, R. Pehrson, T. S^rhaug, S.-E. Birkeland Basics 2: physiology, mass transfer and dynamic modelling Comparison of retention and expression of recombinant plasmids between suspended and biofilm-bound bacteria degrading TCE J.D. Bryers, R.R. Sharp Immobilized-cell growth: Diffusion limitation in expanding micro-colonies R.H. Wijffels, C D . de Gooijer, A.W. Schepers, J. Tramper Competition and cooperation of microorganisms in a coimmobilized aerobic/anaerobic mixed culture G. John, Cs. Sisak, P. Komaromi, L. Hellendoorn, K. Schiigerl Viability of immobilized cells: Use of specific ATP levels and oxygen uptake rates P. Gikas, A.G. Livingston

173

181 189

196

205 213

221

229

239 249

257 264

Gelatin immobilized growing yeast cells: Changes in the glycosylation level of external invertase and cell wall composition E. de Alteriis, J. Zueco, R. Sentandreu, P. Parascandola Nitrification activity of immobilized activated sludge evaluated by respiration rate H. Nakamura, S. Miyabayashi, K. Noto, T. Sumino Growth of immobilized cells: Results and predictions for membrane-attached biofilms using a novel in situ biofilm thicloiess measurement technique L.M. Freitas dos Santos, P. Pavasant, E.N. Pistikopoulos, A.G. Livingston Cryo-electron microscopy of polymer particles in a high cell density synthetic biofilm V. Thiagarajan, Y. Ming, L.E. Scriven, M.C. Flickinger Investigation of oxygen consumption by E. coli immobilized in a synthetic biofilm using a thin film plug reactor (TFPR) V. Thiagarajan, K.L. Swope, M.C. Flickinger Investigation of gene expression in synthetic biofilms to extend the activity of immobilized whole cell catalysts K.L. Swope, J. Liu, L.E. Scriven, J.L. Schottel, M.C. Flickinger Screening and design of immobilized biocatalysts through the kinetic characterization by flow microcalorimetry P. Gemeiner, V. Stefuca, A. Welwardova-Vikartovska Pitfalls of initial reaction rate measurements M. Polakovic, G. Handrikova, P. Acai, V. Stefuca, V. Bales Growth and eruption of gel-entrapped microcolonies L.E. Hiisken, J. Tramper, R.H. Wijffels Quantitative characterization of viability and growth dynamics of immobilized nitrifying cells E.J.T.M. Leenen, A.A. Boogert, A.A.M. van Lammeren, J. Tramper, R.H. Wijffels Modelling of the biotransformation from geraniol to nerol by freely suspended and immobilized grape {Vitis vinifera) cells J. Guardiola, M. Canovas, J.L. Iborra Modelling and experimental validation of cell and substrate evolution in an immobilized system C. Quiros, M. Rendueles, L.A. Garcia, M. Diaz Mass transfer limitations in a bioartificial pancreas R. Willaert, G.V. Baron

272 280

290

298

304

313

320 328 336

341

349

355

362

Effect of dissolved oxygen concentration on pH-controlled fed-batch gluconate production by immobilized Aspergillus niger M. Moresi, E. Parente, A. Ricciardi, M. Lanorte Alginate-immobilized thermotolerant yeast for conversion of cellulose to ethanol N. Barron, D. Brady, G. Lx)ve, R. Marchant, P.Nigam, L. McHale, A.P. McHale Further research on polysaccharide production by immobilized cells of the lichen Pseudevemia furfuracea (L.) Zopf. in polyacrylamide T. Pereyra, C. Vicente Study of parietin production by immobilized cells of Xanthoria parietina in calcium-alginate matrix T. Pereyra, M.C. Molina, M. Segovia, J.L. Mateos, C. Vicente Effects of immobilization on polyol production by Pichia farinosa B. Bisping, U. Baumann, R. Simmering Bioprocess engineering considerations in cyclosporin A fermentation by immobilized fungus Tolypocladium inftatum T.H. Lee, G.-T. Chun, Y.K. Chang, J. Lee, S.N. Agathos Nitrate uptake by immobilized growing Chlamydomonas reinhardtii L Garbayo, C. Braban, M.V. Lobato, C. Vilchez Influence of the immobilization methodology in the stability and activity of P. putida UV4 immobilized whole cells J.V. Sinisterra, H. Dalton Immobilization of the extremely thermophylic archaeon Pyrococcus furiosus in macro-porous carriers R. Portner, H. Markl Immobilization of Mycobacterium sp. cells for sitosterol side chain cleavage in organic solvents P. Femandes, J.M.S. Cabral, H.M. Pinheiro Oxytetracycline production by free and immobilized cells of Streptomyces rimosus in batch and repeated batch cultures H.A. El-Enshasy, M.A. Farid, A.I. El-Diwany Bacterial conjugation within k-carrageenan gel beads: biotic and abiotic factors affecting plasmid transfer D.D.G. Mater, M. Craynest, J.-N. Barbotin, N. Truffaut, D. Thomas Plasmid stability in immobilized Bacillus subtilis continuous cultures M. Craynest, D. Mater, J.-N. Barbotin, N. Truffaut, D. Thomas

370

379

384

390 395

402 410

416

424

431

437

444

452

Improved stability of a naturally occurring TOL plasmid in Pseudomonas putida by immobilization M. Karbasi, E. Asilonu, T. Keshavarz Strategies in the design of an enzymatic process for the synthesis of ampicillin: A whole cell E. coli recombinant penicillin amidase biocatalyst S. Ospina, E. Barzana, O.T. Ramirez, A. Lopez-Munguia Immobilization of recombinant E. coli cells of with phenol-lyase activity S. Bielecki, R. Bolek Kinetics of a co-immobilized system L. Hellendoorn, J.C. van den Heuvel, S.P.P. Ottengraf Dynamic modelling of an integrated nitrogen removal system using co-immobilized microorganisms V.A.P.M. dos Santos, J. Tramper, R.H. Wijffels

458

464 472 479

486

Immobilized-cell reactorsScale up aspects of immobilized cell reactors J.J. Heijnen Problems in scale-up of immobilized cell cultures T. Keshavarz, C. Bucke, M.D. Lilly Performance of a liquid-impelled loop reactor with immobilized cells M.M.R. da Fonseca, D.M.R. Mateus, S.S. Alves Effect of pulsation on morphology of Aspergillus niger and Phanerochaete chrysosporium in a fluidized-bed reactor M.T. Moreira, G. Feijoo, A. Sanroman, J.M. Lema The estimation of the optimum amount of solid support in an immobilized cell bioreactor S.D. Goldfarb, H.A. Chase Double-chambered bioreactors based on plane immobilized-cell membrane structures T. Lebeau, T. Jouenne, L. Mignot, G.-A. Junter Continuous production and in situ extraction of isovaleraldehyde in a membrane bioreactor F. Molinari, F. Aragozzini, J.M.S. Cabral, D.M.F. Prazeres PEGASUS: Innovative biological nitrogen removal process using entrapped nitrifiers H. Emori, K. Mikawa, M. Hamaya, T. Yamaguchi, K. Tanaka, T. Takeshima 497 505 511

518

524

532

538

546

Treatment of nitrogenous wastewaters by immobilized cyanobacteria in an airlift-fluidized photo-bioreactor C M . Lee, C. Lu, Y.-H. Yin, P.-C. Chen Photosynthetic bio-fuel cells using immobilized cyanobacterium Anabaena variabilis M-3 T. Yagishita, S. Sawayama, K.-I. Tsukahara, T. Ogi Lactic acid fermentation using immobilized Lactobacillus casei cells A. Senthuran, V. Senthuran, R. Kaul, B. Mattiasson Lay-out of fixed bed reactor systems for effective production of biologicals with immobilized animal cells R. Portner, L Liidemann, H. Markl

556

563 570

576

ApplicationsIndustrial application of immobilized biocatalysts in Japan T. Shibatani Enzyme production with immobilized filamentous fungi S. Linko, R. Haapala, Y.-H. Zhu Microbiologial stability of an immobilized cell bioreactor with mixed lactic acid bacteria during continuous fermentation of milk C. Lacroix, L Sodini, G. Corrieu Characterization of monoclonal IgA production and activity in hollow-fiber and fluidized-bed reactors T.S. Stoll, P.-A. Ruffieux, E. Lullau, U. von Stockar, LW. Marison Organoleptic profiles of different ciders after continuous fermentation (encapsulated living cells) versus batch fermentation (free cells) J.-P. Simon, A. Durieux, V. Pinnel, V. Garre, J. Vandegans, P. Rosseels, N. Godan, A.M. Plaisant, J.-P. Defroyennes, G. Foroni Application of nitrification by cells immobilized in polyethylene glycol K. Tanaka, T. Sumino, H. Nakamura, T. Ogasawara, H. Emori Immobilized cell technology in food processing C.P. Champagne Bioencapsulation of carrot somatic embryos J.-N. Barbotin, R. Timbert, C. Bazinet, D. Thomas Alginate-polycation microcapsules for cell transplantation - Long time stability B. Thu, O. Smidsr0d, G. Skjak-Brask 585 592

600

608

615 622 633

641 649

Fumaric acid production by Rhizopus arrhizus immobilized on polyurethane sponge F. Federici, M. Petruccioli Application of immobilized yeast cells in the brewing industry N.A. Mensour, A. Margaritis, C.L. Briens, H. Pilkington, I. Russell Design and application of an immobilized loop bioreactor for continuous beer fermentation M. Andries, P.C. van Beveren, O. Goffm, C.A. Masschelein Leuconostoc oenos entrapment: Application to continuous malo-lactic fermentation A. Durieux, V. Garre, J. Mukamana, J.-M. Jourdain, D. Silva, A-M. Plaisant, J.-P. Defroyennes, G. Foroni, J.-P. Simon Practical use of an immobilized cell bioreactor for continuous prefermentation of milk I. Sodini, G. Corrieu, C. Lacroix Behavior of immobilized Nitrosomonas europaea and Paracoccus denitrificans in tubular gel for nitrogen removal in wastewater H. Uemoto, H. Saiki Treatment of high strength ammonia wastewaters using immobilized biomass W.M. Rostron, D.C. Stuckey, A.A. Young Comparison of a system using immobilized microorganisms with a conventional activated sludge process for wastewater treatment P. Chudoba, R. Pujol, H. Emori, J.C. Bourdelot, J.M. Rovel Nitrification in PVAL beads: Influence of pH and temperature on nitrite oxidation Th. Willke, K.-D. Vorlop The concept of combined phosphorus and nitrogen removal by immobilized biomass I. Wojnowska-Baryla, D. Stachowiak, E. Klimiuk Anion-active surfactants degradation by immobilized cells J. Huska, I. Zavadska, M. Dobrotova, D. Toth, P. Gemeiner, A. Vrbanova, J. Augustin Entrapped microbial cell process for treatment of chlorophenolic compounds C M . Lee, C.-J. Lu, C.-Z. Huang, C.-C. Wang Treatment of pentachlorophenol (PCP)-mineral salts medium by naturally immobilized cells of Arthrobacter strain ATCC 33790 R.U. Edgehill

655 661

672

679

687

695 703

710

718

725

731

739

745

Aerobic degradation and dechlorination of low-chlorinated biphenyls by immobilized cells of a PCB degrading bacterial co-culture F. Fava, D. Di Gioia, L. Marchetti Biodegradation of phenol by a mixed culture entrapped in Si02 films T. Branyik, G. Kuncova, J. Paca, K. Jurek, F. Kastanek Propachlor and alachlor degradation by immobilized and suspended Pseudomonas cells E. Ferrer, J. Blanco, R. Alonso, M. Martin Morpholine degradation by strain Mycobacterium aurum MOI: Improvement of cells growth and morpholine degradation rate by cells immobilization P. Poupin, N. Mazure, N. Truffaut Immobilized cells for applications in non-conventional systems L. Setti, G. Lanzarini, P.G. Pifferi Characterization of biofilm structure formed on nickel alloy fibre for ferrous sulphate oxidation by Thiobacillus ferrooxidans cells J.M. Gomez, I. Caro, D. Cantero Biological sulphate reduction with synthesis gas: Microbiology and technology R.T. Van Houten, G. Lettinga Development of mercury biotransformation process in fluidized bed reactor with immobilized microorganisms S. Ledakowicz, U. Becker, W.-D. Deckwer Cadmium removal in an airlift reactor I. Wojnowska-Baryla, R. Krzysik, A. Babuchowski Gastrointestinal protection of cellular component DNA within an artificial cell system for environmental carcinogen biomonitoring D. Quong, I.K. O'Neill, D. Poncelet, R J . Neufeld Immobilized transport mutants of bacterial cells in biosensor arrays. Improved selectivity for the simultaneous determination of glucose and lactose M. Fritzen, W. Schuhmann, J.W. Lengeler, H.-L. Schmidt

750 757

762

770 777

785 793

800 808

814

821

GuidelinesGuidelines for the characterization of immobilized cells R.H. Wijffels, R.M. Buitelaar, C. Bucke, P.B. Poulsen, M.D. Lilly, P.SJ. Cheetham, J. Tramper Index of authors Keyword index

831 835 839

This Page Intentiona Left Blank

Keynote lecture

This Page Intentiona Left Blank

R.H. Wijffels, R.M. Buitelaar, C. Bucke and J. Tramper (Eds) Immobilized Cells: Basics and Applications 1996 Elsevier Science B.V. All rights reserved.

Why immobilize?C. Wandrey

Institute of Bioteclinology, Research Centre Julicli, D-52425 Jiilich, Gennany

Introduction Why iiTunobihze? - The simplest answer may be: because nature does it! While scientists are requested to be mobile it is obviously advantageous for many biocatalytic systems to be immobile. A good example of this is Yellowstone Park (USA) where one can see thick mats of microorganisms and algae at the edge of hot springs. With a continuous supply of substrates, there is obviously selective pressure to become immobile. In nature different biocatalytic systems adapt remarkably well to reaction conditions, e.g. temperature, pH and substrate concentration in the effluent of the above mentioned hot springs. Also in the laboratory one can see - sometimes unwanted - spontaneous immobilization. In chemostate cultures, which are operated close to washout conditions, microorganisms often develop a tendency to grow on the reactor wall. If nature shows such a clear tendency towards immobilization it is obvious that one also can exploit such phenomena in teclinical systems. While the use of immobilized whole cells has already been known for a long time (e.g. production of acetic acid with microorganisms immobilized on pieces of wood) teclinical systems using immobilized isolated enzymes were applied much later. In the fifties, the first laboratory systems were described in the literature 11-2]. The first full-scale system using iminobilized enzymes went into operation in 1969. The Tanabe Company, Japan, used immobilized acylase for the kinetic racemic resolution to obtain L-methionine [3]. The Japanese were also pioneers in the field of immobilized living and dead cells (production of ethanol from glucose, production of acrylamide from acrylnitrile) [4]. Full-scale systems using immobilized higher cells came into use in the eighties. A well known example is the plant of Bayer, USA, for the production of factor VIII. Here, the cells are not immobilized on a carrier but retained by a membrane in a continuously operated system [5].

In the following, some reasons, problems, and solutions are given which may occur, if enzymes, microorganisms, or higher cells are to be immobilized. The examples used are mostly from our own developments. It is not possible here to give a comprehensive overview, since there are now dozens of systems on an industrial scale - which are seldom described in detail - and thousands of systems in laboratory scale - which are better documented.

Immobilized enzymes Enzymes should be understood here as enzymes in dead cells, partially purified enzymes, or purified enzymes. Immobilization can be achieved by carrier fixation, encapsulation, cross linking and/or the use of membranes for catalyst retention in a continuously operated system. In recent years there has been a trend towards using ''macroporous'' microcarriers. Such carriers can be produced from a mixture of glass powder and salt. After moulding and sintering the inert salt fraction is washed out [6]. The pore size and the porosity can be varied witliin a wide range by the appropriate selection of the particle size and the volume fraction of the salt so that even the centre of the particle can be reached by (reduced) enforced flow. In our experience, homogeneous biocatalysts immobilized in membrane reactors often have advantages in comparison to heterogeneous biocatalysts. Catalytic systems become complex if a coenzyme is needed. Since a coenzyme is a transport metabolite it does not make sense to immobilize the coenzyme on a carrier. For readily water-soluble reactants one can exploit the fact that in the case of NAD very low coenzyme concentrations are needed to saturate the enzymes involved. The enzymes are retained in the membrane reactor, while the coenzyme is continuously dosed into the system. For substrates with intermediate water solubility, the coenzyme can be covalently bound to soluble polymer in order to be retained by a membrane together with the enzymes. If the water solubility of the reactants is very low, one can exploit this to deliver the substrate to the water phase from an organic solvent via a membrane. Mter the product is fonned, it leaves the water phase and is obtained from the organic phase. In this case the entire catalytic system is ''iirauobilized'' in the water phase, while the organic solvent is a continuous phase [7].

Total Turnover Number (TTN) readily soluble^

s+

SHo COo NAD(H) TTN: 600 000

HCOOH NAD I soluble substrate:

HCOOH I sparingly solut S + HCOOH

c>TTN: org, solv.

SH2 +COr

5 000 SHc + CO-i

Figure 1. Continuous cofactor regeneration for reactants of different water solubility. Microorganisms Microorganisms are immobilized (in a dead fonn), if no enzyme purification is required (e.g. production of acrylamide from acrylnitrile). In other cases living, non-reproducing cells, retained by a membrane or fixed to a carrier, are used. The microbial glucose oxidase catalyzes the production of gluconic acid from glucose while the simultaneously present catalase decomposes the hydrogen peroxide fonned. Another example is the use of immobilized yeast for coenzyme-dependent reductions. Glucose is used as the hydrogen source, taking advantage of the intracellular cofactor regeneration system. The microbial system is limited with respect to other nutrients so that the glucose is mainly used for cofactor regeneradon and not for cell growth. Anaerobic microorganisms are good examples of immobilization by means of membranes or carriers. One need not be afraid of substrate limitation with respect to oxygen. Furthermore, most of the carbon source is converted to the desired product (e.g. ethanol, acetic acid, biogas). hicreasing space-time yield by

increasing substrate concentration is limited by product inliibition or as in anaerobic waste water treatment, by the substrate concentrations available in practical waste waters. All these arguments strongly support the use of immobilization techniques. Nevertheless, one has to take into account the fact that in the mentioned systems cells reproduce and form carbon dioxide. Both of these ''byproducts'' must readily leave the reactor in order to avoid accumulation. For this purpose macroporous carriers or membranes can be used. With membranes a bleed teclinique has to be apphed, where a minor part of the effluent leaves the reactor in an unfikered state in order to avoid uncontrolled accumulation. A first example of such systems is the production of ethanol from glucose by means of Zymomonas mobilis . Only 2.3 % (at most) of the glucose consumed by Zymomonas mobilis is needed for cell mass fonuation. That means that nongrowth-coupled ethanol production has a negligible effect on ethanol selectivity.Growth (and thus ethanol fonnation) is strongly inhibited by ethanol. Cell growth becomes zero at ethanol concentrations above 70 g/1. With the given ethanol selectivity of ~ 0.5 g ethanol/g glucose it does not make much sense to use glucose concentrations higher than 140 g/1. One has to make a compromise between maximal product concentration and maximal space-time yield. Here, cell immobihzation helps a lot, since for sensible ethanol concentrations the cellspecific ethanol production rate is comparatively low. One can nevertheless reach acceptable space-time yields by increasing the cell concentration. Both by carrier fixation as well as by the use of membranes, an at least 10-fold increase of catalyst concentration can be reached in comparison to chemostate conditions [89] (Figure 2). In order to get rid of daughter cells and carbon dioxide and at least partially overcome product inhibition a two-stage fluidized-bed reactor using "macroporous" microcarriers was employed. Using an initial substrate concentration of 120 g glucose/1, an ethanol concentration of 50 g/1 was reached (80 % conversion in the first stage, 99 % conversion in the second stage). Spacetime yield (at 99 % conversion) was 12 g/(lxh). hi comparison to chemostate the space-time yield could be increased by the factor of 2.5. Since, the biomass concentration was increased 10-fold, the effectiveness factor was only about 0.25. An important additional advantage of the two-stage fludized-bed cascade in comparison to a chemostate was the fact that this system could be operated with a non-sterilized glucose feed stock, which is a byproduct of starch hydrolysis. When this feed stock was used in a chemostate, Lactobacilli became dominant after some time (shift from ethanol production to lactic acid production), hi the fluidized-bed cascade contaminating Lactobacilh were continuously washed out.

due to a monoseptic precolonization of the carrier with Zymomonas mobilis and the short residence time.

Figure 2. Scanning electron micrograph of a pore inside the macroporous glass carrier (Siran^^) fiilly colonized by Zymomonas mobilis.100-

UvW^U#^

I 25o 8 12

process time, d

(120-135 g/l glucose; pH = 5.0; T = 30 C; V tiuidized bed reactor = 55 I: residence time 4 h)

Figure 3. Comparison of ethanol fennentation from hydrolized B starch in a chemostate and in a two-stage fluidized-bed cascade.

For a quite different purpose, the concentration of Zymomonas mobilis was increased in a membrane reactor. The reactor was placed into the core of the magnet of nuclear magnetic resonance machine for in-vivo NMR measurements of intracellular metabolites.

acid, alkali substrate air(0 : ) = ^ micro fihration

c={> filtrate bleed H=i>air(C02)

measuring^ ichamber 0

Figure 4. Chemostate with cell retention (by means of a microfiltration membrane) to increase cell density. By decoupling the substrate and catalyst residence time in a chemostate with partial cell retention it was possible to measure intracellular metabolites under steady-state conditions with better signal/noise-ratio.

Acetogenium kivui is a homoacetate fermenting bacterium, the acetate yield is very liigh, but there is strong product inhibition.

inhibitionI I I

Sicetogenium l(ivui 1 CeH.sOe pH 6.4, 66 C

i 3 CH3COOH

1 Mo! glucose

3 Mo! acetic acid

Figure 5. Reaction scheme for the homoacetate fermenting Acetogenium kivui.

bacterium

It is advisable to use integrated product removal to prevent acetic acid from accumulating. For this purpose electrodialysis is useful. Simultaneously the acetic acid concentration can be increased in the effluent of the electrodialysis unit.

Surplusbiomass

Microfiltration

H2O

ASubstrate (Glucose)c:(fc:;:J>|

NaOH -J^

N^

Permeate (NaAcGlucose) H2

niOH-

A

C

O2

pH6.4 64X Retentate | I |^g+.

t

u

Ac*

''Glucose

If'(HAc)

H+^

NaOH Aceticacid

Fluidized bed reactor

Electrodialysis

Product

Figure 6. hitegrated electrodialysis for the removal of acetic acid fennented by Acetogenium kivui from glucose.

10

A fluidized-bed reactor with carrier-fixed Acetogeniiim kivui is used. Product removal is effected from the bypass [10]. From biological waste water treament we can get a clear answer to the question "Why immobilize?", hi most cases the substrate concentration is so low that a continuous fermentation, following the chemostate method, would result in a very low space-time yield. By means of sludge sedimentation and recycling the biocatalyst is immobilized within the reactor system in order to increase the catalyst concentration. While in aerobic waste water treatment sludge recycling only makes sense until the oxygen transfer becomes limiting, in the anaerobic case mass transfer with respect to the biogas produced becomes limiting only at extremely high biocatalyst concentrations. In such cases the release of biogas can

1

Fixed Bed Reactor

Fixed Bed Loop Reactor

Fluidized Bed Reactor

Figure 7. Fixed-bed loop reactors or fluidized-bed reactors with an external recycling loop are used in the Jtilich High Perfomance Biogas Process.

11

be enhanced by using a fixed bed loop reactor or a fluidized-bed reactor with carrier-fixed microorganisms. Biogas is readily released due to the recycle flow. So appropriate values with respect to substrate, product and pH can be established [11]. It was possible to reduce the chemical oxygen demand of the waste water up to 250 kg per m^ total reactor volume and day. This is achieved by a very high (readily accessible) biomass concentration within a ''macroporous'' microcarrier.

Figure 8. Scamiing electron micrograph of a glass carrier colonized with Methanosarcina barken for anaerobic waste water treatment.

The advantages of the process are: no energy-intensive oxygen supply, biogas fomiation, which can replace natural gas and small amounts of exit sludge. The process has been commerialized by several licensees. Glass carriers are used in reactors with volumes up to 170 m \ volcanic stone is used as a carrier in reactors up to 1200 m\

Mammalian cells With respect to the question ''Why iimnobilize?", higher cells are of special interest because some of them (anchorage-dependent cells) can only grow if they

12

are attached to surfaces. A special tecliiiique has been developed for such cells (T-flasks, roller bottles, non-porous microcarriers). Such systems are characterized by the formation of a confluent monolayer of cells. Other higher cells (e.g. hybridoma cells) have mostly been cultivated up to now in suspension culture. Nevertheless, the basic reasons for iimnobilization are the same although higher cells show some special features. They can produce cytocines and adhesion factors. To keep them "happy", sometimes one has to supplement the medium with such compounds (being one of the reasons for the use of semm in the medium composition). The cell-specific productivity of such factors already reaches a maximum at comparatively low substrate concentrations. On the other hand, it is well known that high substrate concentrations can lead to inliibitory or even toxic product levels (ammonia, lactic acid). Under these circumstances immobilization is especially useftil, since a high flow of a medium with low concentration can be used avoiding cell washout. High cell density is good for a high concentration of supporting factors (adhesion factors, growth factors), low medium concentration is good for low concentrations of inliibitory or toxic products. Since higher cells have a diameter of about 10 ^im and sediment rapidly, they can be kept in a reactor system not only by carrier fixation but also by sedimentation (centrifugation) or by means of appropriate filters (e.g. hollow fibre membranes, spin filters). Most of the products obtained from higher cells (monoclonal antibodies, enzymes, pharmaproteins) are of high molecular weight. So there is always a danger of product retention in continuously operated systems if too "tight" filters/membranes/carriers are used. This is one of the reasons why we think that macroporous microcarriers are especially useful in this field. Using this teclinique release of product and daughter cells is easily achieved. Macroporous microcairiers also have the advantage of providing a large internal surface, while at the same time concentration gradients within the carrier can be minimized. A high density of such carriers can best be achieved in fluidized-bed reactors. But here a scale-up problem arises. A medium rising through a fluidized-bed will soon be depleted with respect to oxygen. A solution to this problem is bubble-free oxygen supply via reactor-integrated silicon tubes (cross flow mode of oxygen supply). The diffiision of carbon dioxide in silicon is even higher than for oxygen so that carbon dioxide release is not a problem. Bubble-free aeration has the additional advantage that sensitive cells in suspension are not damaged and that flotation of proteins is avoided [12].

13

^2 COo/Oo '2'"2

t f

Figure 9. Comparison of bubble-free aeration with oxygen supplementation in a recycle loop and in a reactor-integrated mode. The cross-flow mode of oxygen supply within the fluidized bed also allows an upward stream, which is just sufficient to fluidize the particles (high particle/cell density, no washout of small particles, low shear stress) The surface of the carrier must be "bycompatible". hi our own experiments we mostly used Siran'^^-beads with a diameter of about 0.5 nun (Schott AG, Mainz, Gennany). These carriers are made from borosilicate glass following a method as described above. After washing them properly with nitric acid, quite a number of mammalian cells colonize these carriers spontaneously. In other cases, a surface modification is useftil. For this purpose, coating the carriers with gelatine and afterwards ''bathing" these coated carriers in serum proved to be effective [13]. Probably, fibronectins are taken up from the semm and function as adhesion factors for mammahan cells. The following scamiing electron micrographs show beads densely colonized by hybridoma cells, Chinese hamster ovary cells, and baby hamster kidney cells. Cell densities up to 10^ cells/ml have been reached (Figure 10).

14

Figure 10. Scamiing electron micrographs of macroporous Siran^M-beads colonized by hybridoma cells (top), Chinese hamster ovary cells (middle) and baby hamster kidney cells (bottom).

15

The beads were used for the production of monoclonal antibodies (hybridoma cells), glycosyltransferases (CHO cells) and antithrombin III (BHK cells). In fluidized beds with recycle loop residence times as low as 2 h were adequate. Insitu measurement of cell density was possible using the fact that the capacity of a capacitor is influenced by living cells present between the plates of a capacitor. Thus it could also be shown that between 90 % and 99 % of the cells are within the carrier. The carrier material itself does not disturb the capacity measurement, since its capacity is constant. All cells were cultivated under steady-state conditions (constant levels of substrates including oxygen, products, pH etc.). This feature might prove in future to be very useful for improving reproducibility with respect to product fonnation. Batch and fed-batch techniques always have the disadvantage that the cells ''see" changing physiological conditions during a batch.

Outlook Immobilization of higher cells will become more and more important in the field of medical bioteclmology. It has been shown that the behaviour of suspension cultures can be quite different in comparison to ''organlike'' cultures (with high cell density). We cultivated cancer cells in a fluidized-bed reactor under steadystate conditions (for cytotoxicity testing) and stroma and stem cells (for in-vivo cell propagation). Other possible candidates are hepathocytes, skin cells, cartilage cells and maybe T cells. During evolution immobilization was perfected more and more. If there was a continuous supply of nutrients, even simple cells developed a tendency to immobilize on surfaces. Another teclmique is attachment to each other (pelletisation). "Immobilized" higher systems reached higher and higher volumespecific physiological activities until cell differentiation could be developed. Finally this cell differenfiation even allowed life outside water. Obviously, this is only possible if a lot of cells attach to each other. One can also say that the gathering and thus immobilizing of a lot of cells is not only needed for cell differentiafion but also that this iminobihzation - to each other - is the secret of becoming mobile.

References 1 Manecke G, Singer S, Makromol. Chem. 1959: 36: 119.

16

2 3 4 5

6 7 8 9 10 11 12

13

Levin Y, Pecht M, Goldstein L, Katchalski E, Biochemistry 1964: 3: 1905. Tosa T, Mori T, Fuse N, Chibata I, et al. Enzymologia 1966: 31: 214. Yamada H, CHIMIA 1993: 47: 5. Bodeker BGD, Potere E, Dove G, In: Spier RE, Griffiths JB, Berthold W, eds. 12th Meeting European Socienty for Animal Cell Technology. Oxford: Butterworth-Heinemann, 1994; 584. Kiefer W, Sura M, Schott Glaswerke, Mainz/Germany, European patent 1984: EP 117 484. Kmse W, Kragl U, Wandrey C, Forschunszentnim Jlilich GmbH, Julicli/Gemiany, patent application 1994: DE P 44 36 149.1. Weuster-Botz D, Appl Microbiol Bioteclinol 1993: 39: 679. Weuster-Botz D, Aivasidis A, Wandrey C, Appl Microbiol Bioteclinol 1993:39:685. von Eysmondt J, Breuer B, Aivasidis A, Wandrey C, BioEngineering 1989: 5:20. Aivasidis A, Wandrey C, Forschungszentrum Jiilich GmbH, Jiihcli/Gemiany, patent 1986: DE P 33 45 691. Rolef G, Biselli M, Dunker R, Wandrey C, In: Spier RE, Griffiths JB, Berthold W, eds. 12th Meeting European Society for Animal Cell Teclmology. Oxford: Butterworth-Heinemann, 1994; 481. Llillau E, Dreisbach C, Grogg A, Biselh M, Wandrey C, In: Animal Cell Teclmology Developments Processes and Products, Oxford. ButterworthHeinemann, 1992, 469.

Basics 1: physics

This Page Intentiona Left Blank

R.H. Wijffels, R.M. Buitelaar, C. Bucke and J. Tramper (Eds) Immobilized Cells: Basics and Applications 1996 Elsevier Science B.V. All rights reserved.

19

Alginate gels - Some structure-function correlations relevant to their use as immobilization matrix for cells.B. Thu, O. Smidsr0d and G. Skjak-Braek The Norwegian Biopolymer Laboratory, Department of Biotechnology, University of Trondheim, Sem Saelandsv 6/8, N-7034 Trondheim, Norway

Introduction In the last two decades advanced use of polysaccharides for gel-entrapment or encapsulation of cells has become a challenging method for the biotechnologist. Of the many proposed techniques the use of alginate gel beads stands out so far as the most promising and versatile method. This immobilization procedure can be carried out in a single step process under very mild conditions and is therefore compatible with most viable cells. The possible uses for such systems in industry, medicine, and agriculture are numerous, ranging from production of bulk chemicals to cell transplantation [1]. Large scale applications have, however, been hampered by mechanical, chemical a well as biological instability. For improving the functionality of the alginate gels it is essential to recognize that alginate is a collective term for a family of polymers with a wide range in chemical composition, sequential structure, molecular size and, hence, in their functional properties. In the present paper we will report on how features such as, charge, porosity, swelling behaviour, long time stability and gel strength depend on the chemical structure and the molecular size of the alginate molecule, and that these properties further can be modified by controlling the kinetics of the gel-formation.

Immobilization Immobilization of cells by entrapping them in a hydrogel is generally carried out by mixing the cells with a water-soluble Na-alginate cells polymer, and subsequent gelling of the polymer by adding cross-linking agents. For alginate, gelling is induced by adding cations such as calcium or strontium. By dripping the alginate-cell mixture into a solution containing multivalent cations, the calcium chloride droplets will instantaneously form gelspheres by ionotropic gelation, entrapping the cells within a threedimentional lattice of ionically crosslinked polymer (figure 1). Figure 1 Immobilization of cell in alginate Entrapment of cells in alginate gels was gel beads. first described by Hackel et al. [2], who used aluminium ions, and by Kierstan and Bucke [3], who used calcium ions. The technique had, in principle, been applied for many years by the food industry for making such products as artificial berries and caviar.

20

Limitations Gel-entrapment in alginate, as the technique is known today, has some limitations due both to the inherent nature of the alginate molecule itself, as a biodegradable polydisperse material, and to the nature of the gel as a reversible ionic network. As a consequence of the latter, substances with high affinity for calcium ions such as phosphate or citrate will sequester the crosslinking calcium ions and consequently destabilize the gel. Since the calcium ions can be exchanged with other cations, the gel will also be destabilized by high concentrations of non-gelling ions, such as sodium and magnesium. In some specific applications toxicological and immunological aspects are of significance. Although alginatefiilfilsthe requirements for additives in food and pharmaceuticals, some alginates contain small amounts of poly-phenols which might be harmful to sensitive cells. When gels are used for cell-transplantation, the alginate must also be biocompatible andfi-eefi-ompyrogen and immunogenic materials such as proteins and complex carbohydrates.

Which type of alginate should be chosen? There is of course no such thing as an ideal alginate bead to meet the requirements of all immobilized cell systems. However, alginate gel beads should ideally have high mechanical and chemical stability, controllable swelling properties, a defined pore size, and a narrow pore-size distribution and a low content of toxic, pyrogenic and immunogenic contaminants. These criteria may be met by selecting the alginates according to composition, sequential structure, molecular weight and purity, and by controlling the kinetics of the gel-formation.

Alginate chemistry Alginate was first described by the coo" British chemist E. E. C. Stanford in 1881. Its occurrence in nature is mainly limited to the marine brown algae (Phaeophyta), although exocellular polymeric material resembling alginates from brown algae are also produced by soil bacteria such as Azotobacter vmelandii and several species IVl O of Pseudomonas. Alginate exists in the brown algae, as the most abundant polypjg^re 2a) The monomer composition in saccharide comprismg up to 40% of dry ^i^j^^^^ y^. p_D-mannuronate, G: a-Lmatter. It is located in the intercellular euluronate matrix as a gel containing sodium, calcium, magnesium, strontium and barium ions and its main function is believed to be skeletal, giving both strength andflexibilityto the algal tissue. Because of its ability to retain water, and its gelling, viscosifying and stabilizing properties, alginate is widely used industrially and the

21'OOC

OH

'OOC HO -OOC HO

"OOC

J-^OH^OH "OOC

G

G

M

M

Figure 2b) The alginate chain. Chair conformation.

technical application of alginate forms the basis of the exploitation of brown seaweeds in the western hemisphere. In molecular terms alginate is a family of unbranched binary copolymers of M linked B-D-mannuronic acid (M) and a-L-guluronic acid (G) (illustrated in Figure 2a) and b)), of widely varying composition and sequential structure depending on the organism and tissue it is isolatedfrom.The monomers are arranged in a block-wise pattern along the chain with homopolymeric regions of M and G termed M- and G-blocks respectively, interspaced with regions of alternating structure (MG-block). It is well established that alginates do not have any regular repeating unit and that, except in some bacterially derived polymers, the distribution of the monomers along the polymer chain cannot be described by BemouUian statistics. Accordingly, the sequential structure is not determined by the monomer composition (monad frequencies) alone, but by measurements of diad, triad and higher orderfrequencies.The four diad (nearest Table 1 Composition and sequence parameters of algal alginates Source Laminaria japonica L digitata L. hyperborea Blade Stipe Outer cortex Lessonia nigrescens Ecklonia maxima Macrocystis pyrifera Durvillea antarctica Ascophyllum nodosum Fruiting body Old tissue Source: ref [4]0.35 0.41 0.55 0.68 0.75 0.38 0.45 0.39 0.29 0.10 0.36 0.65 0.59 0.45 0.32 0.25 0.62 0.55 0.61 0.71 0.90 0.64 0.18 0.25 0.38 0.56 0.66 0.19 0.22 0.16 0.15 0.04 0.16 0.48 0.43 0.28 0.20 0.16 0.43 0.32 0.38 0.57 0.84 0.44^GM,M

0.17 0.16 0.17 0.12 0.09 0.19 0.32 0.23 0.14 0.06 0.20

22

neighbour)frequenciesFQG, FGM FMG> and F^M, and the 8 possible triadfrequenciesFQQQ, QQ^, FMGG, FMGM, FMMM, FMMG. FGMM and FQMG can be measured by n.m.r. techniques [5][6]_Knowledge

of the diad and triadfrequenciesthen allows calculation of average block-length; N G = FQ/FJ^Q, andNj^ = FJ^F^G. For blocks consisting of at least two contiguous units, NG>I=(FG-FMGM)/FMGG. andN^i= (FM-FCu'^>Cd2^>Ba2^>Sr2^> Ca'^>Co'", Ni'", Zn'^>Mn^\ A common approach for stabilizing alginate gels are covalent crosslinking. Various techniques have been applied, including direct cross-linking of the carboxyl groups, or covalent grafting of alginate with synthetic polymers. This gives gels with improved stability and mechanical strength, but in most cases the coupling conditions are to harsh for fragile cells.Porosity

Difiusion characteristics are essential for the use of alginate gels as immobilization matrix. It is therefore important to know the pore sizes and the pore size distribution. Self-diffiision of small molecules seems to be very little affected by the alginate gel matrix whereas transport of molecules by convection is restricted by the gel network. The self-diffiision of small molecules such as glucose, ethanol and lactate has been reported to be as high as about 90 % of the difiusion rate in water [28]. The diffiision rate depends, however, on the cell load of the beads. For larger molecules such as proteins diffusional resistance occurs, although even large proteins with MW>310^ will leak out of the gel beads with rate dependent on their molecular size [22][29]. The highest diffusion rates of proteins, indicating the most open pore structure, are found in beads madefromhigh-G alginates. This may partly be related to the lower shrinkage of these types of gels, but even when this is corrected the highest diffusion rates are found in high-G alginate. A tentative model is given infigure9. The porosity of alginate gels has been studied by various techniques including electronmicrograph, gel permeation chromatography and diffusion studies. These studies reveal that the alginate gels core is macro-porous with pores ranging from 5-200nm, while the network on the bead surface is more narrow 5-16nm [30]. Evidence for a non-uniform distribution of polymer in the gel beads will be discussed below. Since the alginate matrix is negatively charged, electrostatic interaction between proteins or any other charged species and alginate must be taken into consideration. Most proteins are negatively charged at pH 7 and will not easily diffiise into the gel matrix. On the other hand, when immobilized in the gel, they tend to leak out more rapidly than should be expected from their free molecular difiusion. Several of the procedures for stabilizing the alginate gels mentioned above will also have some influence on the porosity. Exchange of calcium with strontium or barium reduces the porosity HIGH-G ALGINATE HIGH-M ALGINATE slightly, but their main effect on porosity is due to stabilization against swelling. Figure 9 Tentative model of alginate gel Larger effects are achieved by letting the network in gels made from alginates with alginate beads react with a polycation. different composition and sequense. Formation of polyanion-polycation membranes with polypeptides or

28 chitosan has been used for generating certain cut off values for proteins. By controlling the molecular weight of the polycations certain pore-sizes below certain cut-ofF values have been obtained [31]. The influence of the polycation on the pore-size distribution is, however, not clear. The porosity of calcium alginate gel may significantly be reduced by partially drying of the beads. Provided the beads are made out of an alginate rich in guluronic acid, the beads will reswell only slightly in water, and the increased alginate concentration will reduce the average pore size. Gelling kinetics Alginate gels are often regarded as non-equilibrium gels. After a cation has induced random dimerization of G-blocks, the resulting egg-box structure does not necessarily correspond to the most stable conformation. Because of the high activation energy of reopening, the structure is kinetically trapped in the initial conformation. As a result, one can see quite a dramatic difference in gel strength between Nature's own gels in the algae and the artificially prepared gels [13]. Most probably, a large degree of the G-blocks do not overlap to form the energetically most favourable and strongest network structure in ordinary laboratory or industrially prepared gels. Also laboratory produced gels made from the same alginate material, varying only the gelling kinetics can give dramatic differences in functional properties. A good example is the formation of inhomogeneous gels. Gel homogeneity Alginate gels prepared by the dialysis method often exhibit a concentration inhomogeneity in that the polymer concentration is considerably lower in the centre of the gel than at the edges. When divalent metal ions difiuse into an alginate solution, the rapid ion-binding and formation of network produce an inwardly moving gelling zone. Alginate will diffuse from the centre of the gel towards this gelling zone, leading to a depletion of alginate in the centre. A theory providing a qualitative explanation of experimental data may be found in the literature [32]. The polymer gradient is essentially governed by the relative diffusion rate between the soluble alginate

4 Slice number

8

4

6 8 Slice number

Figure 10a Concentration of alginate as function of slice number for alginate gel sylinders made from high-G alginate gelled in CaClj of various concentrations, x: 0.02M, o: 0.05M and D: O.IM.

Figure 10b Concentration of alginate as fiinction of slice number for alginate gel sylinders made from high-G alginate gelled in a mixture of 0.45% NaCl and CaClj of various concentrations; x. 0.02M, o: 0.05M and Q O.IM.

29 molecules and calcium ions. The homogeneity of the gel can thus be controlled by a careful selection of molecular weight and concentration of alginate together with the concentration of gelling and absence non-gelling ions. In general, low molecular weight alginate, low concentration of gelling ions and non-gelling ions give the highest inhomogeneity. This is illustrated in figure 10 a) and b). Homogeneous alginate gels can also be made by internal release of calcium from Ca:EDTA orfromCa:Citrate or Ca (03)2 in the presence of a slow acidifier like glucono-6-lactone (GDL) [33].

Conclusions Cells entrapped in phycocoUoid gels have many potential applications in Biotechnology, rangingfrombiocatalysts in fermentation to artificial seeds in agriculture and carrier materials for transplantation of living tissue. However, since the gelling material is a heterogeneous group of polymers, with diverse functional properties, their success as immobilization matrices will depend on a proper choice of material and methodology for each application. In addition to seaweeds as raw material for alginates of widely different chemical composition, one should bear in mind that alginates are also microbial polysaccharides with a well understood biosynthetic pathway. Both future production of special alginate qualities by fermentation is possible, as well as enzymatic modification of seaweed alginates to give tailor-made alginates for certain applications. The genes coding for the mannuronan C-5 epimerases in Azotobacter vinelandii have recently been sequenced and cloned in our laboratory [34][35]. These enzymes convert M to G in the polymer chain. This opens up the possibilities for large production of the enzyme and future use in modulating the functional properties of alginates.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Smidsr0d O, Skjak-Br^k G. TIBTECH 1990: 8: 71-78 Hackel U, Klein J, Megenet R, Wagner, F. Eur J Appl Microbiol Biotechnol 1975; 1: 291-293 Kierstan M, Bucke C. Biotechnol Bioeng 1977; 19: 387-397 Gates CG, Ledward DA. Food Hydrocoll 1990; 4(3): 215-220 Grasdalen H, Larsen B, Smidsr0d O. Carbohydr Res 1979; 68: 23-31 Grasdalen H. Carbohydr Res 1983; 118: 255-260 Haug A. Composition and Properties of Alginates, Thesis, Norwegian Institute of Technology, Trondheim 1964 Smidsrod O, Haug A. Acta Chem Scand 1968; 22: 1989-1997 Haug A, Smidsrod O. Acta Chem Scand 1970; 24: 843-854 Smidsr0d O. Some physical properties of alginates in solution and in the gel state. Report no. 34, Norwegian Institute of Seaweed Research, NTH Trykk 1973 Smidsrod O. Farad Disc Chem Soc 1974; 57(1): 263-274

30

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

Grant GT, Morris ER, Rees DA, Smith PJC, Thorn D. FEES letters 1973; 32(1): 195-198 Andresen EL, Skipnes O, Smidsr0d O, 0stgaard K, Hemmer PC. ACS Symp Ser 1977; 48: 361-381 Smidsred O, Haug A. Acta Chem Scand 1972; 26: 79-88 Preiss J, Ashwell G. J Biol Chem 1962; 237: 309-316 Gacesa P, Caswell RC, Kille P. In H0iby N, Pedersen SS, Doring G, Holder lA, eds. Antibiot Chemoter. Basel: Karger: 42: 67-71 Smidsr0d O, Haug A, Larsen B. Carbohydr Res 1967; 5: 482-485 Haug A, Laresen B, Smidsred O. Acta Chem Scand 1963; 17: 1466-1468 Skjak-Braek G, Murano E, Paoletti S. Biotechnol Bioeng 1989; 33(1): 90-94 Leo WJ, McLoughlin AJ, Malone DM. Biotechnol Prog 1990; 6: 51-53 Draget KI, Myhre S, Skjak-Braek G, 0stgaard K. J Plant Physiol 1988; 132: 552-556 Martinsen A, Skjak-Braek G, Smidsr0d O. Biotechnol Bioeng 1989; 33: 79-89 Otterlei M, 0stgaard K, Skjak-Braek G, Smidsr0d O, Soon-Shiong P, Espevik T, J Immunotherapy 1991: 10: 286-291 Klein J, Wagner F. Dechema Monogr 1978; 82: 142 Heinzen C, Pluess R, Widmer F. Commercial product from INOTEC AG, Switzerland. Goosen MFA, O'Shea GM, Sun AM. US Patent 1987; 4673566 Skjermo J, Defoort T, Dehasque M, Espevik T, Olsen Y, Skjak-Braek G, Sorgeloos P, Vadstein O. Fish & Shellfish Immunology 1995; 5: 531-534. Axelson A, Persson B. In: Neijssel OM, van der Meer RR, Luyben KCAM , eds. 4th European Congress on Biotechnology. Amsterdam: Elsevier, 1987; 1: Martinsen A, Stor01, Skjak-Braek G. Biotechnol Bioeng 1992; 39: 186-194 Martinsen A, Alginate as immobilization materials - a study of some molecular and functional properties, Thesis, 1989, Norwegian institute of Technology, Trondheim Goosen MFA, O'Shea GM, Gharapetian HM, Chou S, Sun, AM. Biotechnol Bioeng 1984; 27: 146-150 Skjak-Braek G, Grasdalen H, Smidsr0d O. Carbohydr Polym 1989; 10: 31-54 Draget KI, Ostgaard K, Smidsr0d O. Carbohydr Polym 1991; 14: 159-178 Ertesvag H, Doseth B, Larsen B, Skjak-Braek G, Valla S. J Bacteriol 1993; 176: 28462853 Ertesvag H, H0idal HK, Hals IK, Rian A, Doseth B, Valla S. Molecular Microbiology 1995; 16(4): 719-731

R.H. Wijffels, R.M. Buitelaar, C. Bucke and J. Tramper (Eds) Immobilized Cells: Basics and Applications 1996 Elsevier Science B.V. All rights reserved.

31

Determination of biofilm diffusion coefficients using micro-electrodesEvelien E. Beuling, Johannes C. Van den Heuvel and Simon P.P. Ottengraf Department of Chemical Engineering, University of Amsterdam, Nwe. Achtergracht 166, 1018 WV Amsterdam, The Netherlands

Introduction

The structure of a biofilm imposes a diffusive resistance for the transport of metabolites. As a consequence, concentration profiles will develop which affect the local microbial reaction rate. Eventually, this leads to severe mass transfer limitations and results in less effective, partially penetrated biofilms. Therefore, characterization of the mass transfer properties of biofilms is considered essential for modelling and scale-up of microbial conversions [1]. Except for biofilms with a very open and loose structure, the mass transfer inside a biofilm is predominantly diffusive and can be characterized with a single parameter: an effective diffusion coefficient. Various experimental approaches may be used to obtain a biofilm diffusion coefficient. If the conversion rate is known at different substrate concentrations, the diffusion coefficient can be calculated using a diffusion-reaction model. The results of such experiments display large differences, ascribed to biofilm inhomogenities and unknown local kinetics [2]. Knowledge of the local reaction rate is not required if the flux or the transient uptake rate of a non-reacting compound is measured. To this end the biofilm must be inactivated or a non-consumable tracer has to be used. Both methods exhibit drawbacks; the inactivation procedure might alter the biofilm properties, while the resemblance of a non-consumable tracer with relevant metabolites is often disputable. Literature values show considerable variation, connected to these experimental difficulties [3]. The aim of this experimental work was to characterize the influence of the bacterial fraction on the biofilm diffusion coefficient. Experiments with oxygen and glucose were performed in well defined model systems to circumvent the experimental problems mentioned above. An artificial biofilm consisting of agar, containing inert polystyrene particles of the same size as bacteria, was investigated and the obstruction effect of these particles was compared to that exerted by immobilized bacteria. Diffusion experiments were performed with a micro-electrode positioned in the middle of a spherical biofilm and the transient response on a concentration step in the well mixed bulk was monitored. This technique also enabled the measurement of diffusion in active biofilms. Steadystate experiments were performed in a classical diffusion cell. From the results obtained, the permeability of the bacteria was estimated and influences of the inactivation method applied were quantified.

32 Theory

In a biofilm, the cells are connected to a polymer matrix and surrounded by a continuous liquid phase with the same properties as the bulk liquid. If the bacteria and the polymers are considered impermeable they form solid obstacles (volume fraction ^) which affect diffusive transport of metabolites in two ways. The available liquid volume is reduced to a fraction (1-^) by exclusion. Furthermore, the pathlength of the diffusive molecules is increased by steric hindrance, referred to as obstruction or tortuosity. In the literature there is some confusion about the diffusion coefficient measured by different methods. Using a classical diffusion cell, the (pseudo) steady-state flux through a flat film is measured which is determined by both the tortuosity and the porosity. Consequently, the effective diffusion coefficient (Deff) is measured. In case of non steady-state methods like, for example, the transient uptake rate of a quantified amount of gelbeads, the response on a stepwise concentration change in the bulk is monitored. The diffusion coefficient is calculated from the pace at which the new equilibrium is reached. As this depends on the tortuosity only, a transient diffusion coefficient (D) is obtained which equals Deff/(1-(|)). If the solid phase is permeable, the difference between the two methods is more complex [4]. Several models have been developed to relate the effective diffusion coefficient to the structural parameters of the heterogeneous medium. For a solid phase consisting of a suspension of impermeable spheres, homogeneously distributed in a continuous liquid phase (diffusion coefficient Daq) and steadystate conditions. Maxwell deduced [5]: (l-(^) De.=777T7^-Da 2 pdj

and

50 drop diameter values (Figure 4), the production rate remains quite small. The limited production rate constitute is the largest limitation for the use of the drop technique in a large scale process. Moreover, quasy monodispersion is reached only for large beads (>1 mm).

0.5-

Air jet system Emulsification

I 20"S 15" c 10' s

Drop method Capillary jet breakup

0-

0-

0 OiO ' ^

400 800 Dropplet diameter (jim) Figure 3. Bead size distribution

1200

0

500

1000 d ()im)

1500

2000

Figure 4. Maximum production rate

Breakup of Capillary Jet When capillary jets are formed at a tip, they show instablility and break easily forming small droplets. Rayleigh [14] vibrated the jet with sonic waves, which at an optimum frequency, f, lead to a stream of uniform droplets. The optimum condition for jet breakup may be written as : >. = 4.058 d with X = ^ f and d = 1.89 d; {4a,b,c}

The diameter, d, of the resulting drop is determined by assuming that the volimie of the drop equals the cylinder having a diameter equal to the jet diameter, dj, and a height equal to the wavelength, X, The drop size is thus directly correlated with the jet diameter. The jet velocity may be freely selected until the vibration frequency is adapted to maintain the wavelength at the optimum value. However, the jet velocity must be lower than the terminal velocity of the droplets, to avoid collision and coalescence between drops. The maximum flow rate in function of the drop diameter is then obtained by equating the jet velocity with drop terminal velocity (figure 4). For large beads, the production rate may be significantly larger (up to 30 1/h) than with the drop technique. Large scale production may be obtained with the capillary jet breakup by multiplying the number of nozzles (to reach hundred

51 liters per hours for a few nozzles). For small beads (under 800 jxm), the number of nozzles to reach large production rates would be too large to constitute a simple solution. Moreover, the pressure required to insure the jet minimum velocity becomes high (several atmospheres) when the orifice diameter decreases and constitute a limitation for firagile cell encapsulation. Satellite beads are also difficult to avoid. Spinning vibrating disk

When a liquid flows on a spinning disk, the liquid may leave the disk as small jets or ligaments (Figure 5). These ligaments have a behaviour similar to jets escaping a classical orifice [15]. By applying a well designed wave on the liquid flowing on the spinning disk, the jets are broken in small and very uniform droplets (standard deviation lower than 5 %). Under certain conditions (correct wave amplitude), formation of satellite particles is avoided. Equations directing the process are more complex than with classical jet rupture. Physcal properties of the liquid (density, surface tension, viscosity), design of the rotating disk (size, rotating speed) and wave (frequency and amplitude) must be adjusted to reach correct particle size and optimum conditions for low dispersion [16]. T3rpically, on a disk of one centimeter turning at 2000 rpm, aroimd 60 ligaments are formed. Drops of 300 to 400 jim are produced at a flow rate of 6 to 8 1/h. In optimum condition, the droplet diameter is proportional to the spinning disk diameter and the production rate islinked to 5/3 power of the droplet diameter. Such a droplet generator would offer a solution for large scale production of microdroplets (less 800 |j.m). In this condition, the production rate is increased by two orders of magnitudes with regard to the simple vibrating orifice technique. In addition, high pressure is not required to cause the liquid to flow on the disk. Emulsification process If hydrogel beads are generally formed by the drop method, membrane encapsulation procedures are usually based on emulsification of the core material in a non-miscible phase. Membrane is formed by either interfacial coacervation or pol5niierisation. Due to scale-up problems with the drop methods, several authors have also considered emulsification as a potential technique for bead formation [17]

52 Liquid flow vibration

ligaments Figure 3. spinning rotating disk device Figure 4. Static mixer (Kenics)

In emulsification, droplets are not formed one by one, but rather in terms of millions by millions. The equations are thus built on statistical basis, partially from mechanistic models and partially from empirical correlations. The fundamental analyses are principally based on the work of Kolmorogov [18] and Hinze [19]. These authors stated that the energy dissipated in a turbulent flow creates a viscous stress or a djniamic pressure which tends to break the drop. The surface tension force and the internal drop viscosity counteract these deformations .

5-

We-0-^5Re-0-2

MP-

0.5 with We:

Dpu^

and

Re =^ P ^ {5a,b,c) u

where D is the hydraulic diameter of the dispersion device (impeller diameter for the turbine reactor, internal diameter for the static mixer, see below), k, a constant fimction of the design, jid, the dispersed phase viscosity. We and Re are the Weber and Reynold's numbers. Equation 5 assumes that the parameters defining the final drop size are essentially the rotational speed of the impeller, the viscosity of both continuous and dispersed phases and the interfacial tension. However, fitting of data on this equation has not been successful. Equation 5 was designed for low viscous phases and Newtonian fluids. Gel and pre-encapsulating solutions may behave in a nonideal manner. The swelling or shrinkage of droplets during gelification and/or polymerization of the membrane constituents may lead to smaller or larger diameters than expected.

53 The theoritical size dispersion with emulsification is a log normal distribution with a standard deviation of around 35 %. In real cases, the standard deviation may be larger (up to 50 %) and a satellite peak appears that may represent up to 10 % of the microcapsules volimie. Ideally, gelification should take place rapidly, but only after emulsion equilibrium is reached. In other cases, the main peak is divided in several peaks. The mean size is also no more correlated with Equation 5. The reactors used for emulsification are usually cylindrical vessels, mixed by means of various impellers (turbine, marine-style impeller or grid device) [2]. In such devices, shear, energy dissipation, and dynamic pressure are not homogeneously distributed. Dead volumes or stagnant zones may be present, as the vessel volume increases or when mixing viscous fluids such as gels. One alternative to minimize these problems may be provided by static mixers. These devices consist of a series of stationary elements mounted lengthwise in a pipe (Figure 6). The elements form intersecting channels t h a t split, rearrange and recombine component streams into smaller and smaller layers. Mean diameters of the dispersed phase ranging from a few microns to 1 nrni may be produced and the droplet size defined by an equation similar to that of Equation 5 derived from mechanistic models [20]. As homogeneous shear is applied to the whole liquid, dispersion will then lead to narrower size distribution. Moreover, as high shear is not applied near the impeller, the static mixer would be more suitable to encapsulate fragile cells. A prelimary study of static mixers to produce carrageenan beads [21] show that beads of 500 - m may be produced at 10 1/h with a 13 mm static mixer. Mean size may be easily controled by the linear velocity in the tube. The scale-up is simply realized by increasing the static mixer diameter. The production r a t e is correlated with the square of the mixer diameter (150 1/h for 5 cm static mixer). Emulsion technology has several drawbacks. The size dispersion is generally larger than with the extrusion techniques. The resulting beads or microcapsules need to be transfered from organic solvents (generally vegetal oil) and washed. Prediction of the size is more complex and experiments are needed to design the device. However, when large production is required, such as cubic meters, emulsification appears as the best or presently the one solution. Conclusions The need for a dispersion system may be divided into four categories : in laboratory scale, capillary jet breakup allows production of small but very

54 uniform dropplet batchs, at larger scale (up to hundred liters) and for large beads (1 to 3 mm), the capillary jet breakup constitue also a simple solution, for similar production but lower size, the spinning rotating disk represent a more promising solution, finally, if largest production is required, or involving an interfacial process, the engineer may consider emulsification process ~ith static mixer technology. References 1 Cheetham PS, Blunt KW and Bucke C. Biotechnol. Bioeng. 1979 21: 2155-2168 2 Poncelet D, De Smet B, et al. Appl -licrobiol Biotechnol 1995 43:644-650 3 Dautzenberg H, Loth F, et al. Makromol Chem Suppl 1985 9:203-210 4 Larich BC, Poncelet D, et al. J. Microencapsul 1994 11:189-195 5 Poncelet D, De Smet B, et al. J. Membr. Sci.1990 50: 249-267 6 Poncelet D, Neufeld RJ. Biotechnol. Bioeng.1989 33: 95-103 7 Harkins WD, Brown FE. J. Amer. Chem. Soc.1919 41: 499,1919 8 Lane W. Rev. Scient. Instrum.1947 24: 98-101 9 Poncelet D, et al. in: Doosen ~IFA, ed. Fundamental of animal Cell Encapsulation and Lnmobilization, CRC Press. 113-142 10 Miyawaki 0. Nakamura K, et al. Agric. Biol. Chem.1979 43: 1133-1138 11 Su H, Bajpai, et al. Appl. Biochem. Biotechnol. 1989 20/21: 561-569 12 Burgaski B, Li Q, et al. Mater Interf Electrochem Phenom 1994 40:1026-1031 13 Poncelet D, Burgarski B, et al. Appl Microbiol Biotechnol 1994 42:251-255 14 Rayleigh JWS. Math. Soc.1878 10: 1878. 15 Chicheportiche JM, Renaudeaux JP. pat n90 029 14 PC/ML (14 mars 1991) 16 Chicheportiche JM. PhD thesis. University Paris VI, France, 1993 17 Audet P, Lacroi, C. Proc. Biochem. 1989 12: 217-225 18 Kolmogorov AN. Dok. Akad. Nauk.1949 66: 825-831 19 Hinze JO. AIChE J.1955 1: 289-293 20 Middleman S. 1974 13:1, 78-83 21 Decamps C. Engineer Training Report, ENSBANA, Dijon, France, 1994

R.H. Wijffels, R.M. Buitelaar, C. Bucke and J. Tramper (Eds) Immobilized Cells: Basils and Applications 1996 Elsevier Science B.V.

55

Stable support materials for the immobilization of viable cellsA. Muscat, U. PriiBe and K.-D. Vorlop

Institute of Technology, FAL, Bundesallee 50, 38116 Braunschweig, Germany

Introduction Conventional natural materials for entrapment of cells (e.g. alginate, carrageenan) possess no toxicity against cells but have a low mechanical stability, are very sensitive against abrasion in stirred reactors and are biodegradable under non-sterile conditions. On the other hand, stable synthetic hydrogels for entrapment of viable cells could possess the disadvantages of a high toxicity during the immobilization process and of a very short handling time. Due to this fact it is nearly impossible to produce synthetic hydrogel beads. Conventional support materials for the adsorption of aerobic cells possess only a low oxygen permeation rate. We developed a spherical porous polymer support material for the adsorption of aerobic cells with the advantages of high elasticity and a high oxygen permeation rate through this material. This paper describes stable support materials for entrapment and adsorption of viable cells and different techniques for producing spherical support materials. Furthermore, these materials characterized with respect to their mechanical stabilities. Stable support materials for the entrapment of viable cells Natural polymers for cell entrapment possess only low mechanical and chemical stability in contrast to polyurethane (PUR) or polyvinyl alcohol (PVAL) hydrogels. Poly(carbamoyl sulfonate) PCS hydrogels PUR hydrogels have many applications especially in medicine, because of their good mechanical and chemical stability. But the conventional raw material (isocyanate prepolymer) is toxic [1] and has a short handling time (seconds). Poly(carbamoyl sulfonate) (PCS) hydrogels have the same optimal mechanical properties like conventional PUR hydrogels but they possess only a low toxicity and a handling time of up to 14 hours (adjustable by the pH) [2]. Additionally, two methods for the preparation of PCS hydrogel beads are presented (Figure 1):

7. Dropping method (Figure 1, left): A PCS solution is mixed with a solution of CaCb and adjusted to pH 4 - 6.5. Wet cell mass is added and this suspension is dropped from an apparatus for immobilization into an alginate solution. Immediately, a Ca-alginate layer is formed by ionotropic gelation (migration of Ca^^ from the core to the alginate layer). The cross-linking reaction (or gelation) of the PCS-

56 hydrogel core takes a very fast course at pH 8.5. After a while (< 1 h), the Ca-alginate layer can be dissolved in a phosphate buffer. Usually the last step is not necessary because of the simple biodegradability of Ca-alginate under non-sterile conditions. 2. Suspension method (Figure 1, right): A PCS solution is adjusted to pH up to 7.5. Wet cell mass is added immediately. The resulting suspension is poured into vegetable oil at 37C (Figure 1, left). At this temperature the cross-linking reactions from the PCS solution to the hydrogel occur very fast. After a while, the PCS hydrogel beads are separated from the oil by sieving and they are washed. There is no consumption of alginate and CaCb using the suspension method. Furthermore, it is possible to produce more and smaller (< 1 mm) beads in one batch at a faster rate. For special use it is necessary to classify the beads and to eliminate oil residues on the surface of the PCS hydrogel beads.

OS solutic

solution of CaCl2

uu-

-,1PCS solution Ca-alginate layer

PCS solution 1 mixed with cell mass

u-

^,adjustmerIt off )H up to 7.5

J

" * *' ' 1 i/egetable o 1 T= 37 C separation washing

simultaneous formation of a ff ^+ Ca-alginate layer and gelation \ \ 9 f \ of the core ^sr^ . OH OH

alginate solution pH8.5

use

Figure 1. Production of spherical PCS hydrogel beads by the dropping method (left) and by the suspension method (right).

Furthermore, PCS hydrogel membranes could be prepared [3]. They can be used in biosensors as an immobilization matrix for cells and enzymes [4], especially yeast, nitrifiers and denitrifying bacteria were immobilized in PCS hydrogels [2,3].

Polyvinyl alcohol (PVAL) hydrogels PVAL hydrogels are often used (no toxicity) for the immobilization of viable cells, but it is difficult to prepare PVAL hydrogel beads. Our entrapment process was based on the gelation of an aqueous PVAL solution (7-15 % PVAL with a high saponification rate of 99-100 %) which was frozen to a temperature lower than -5C [5]. Microorganisms were suspended in the PVAL solution and dropped into liquid nitrogen or cooled (-72C) plant oil. Afterwards, the beads were separated and thawed slowly in an insulated container.

57 contrast to the natural polymers which possess a high modulus (lower elasticity!) and a low elongation at break.

Table 1. Modulus of elasticity and elongation at break of hydrogels Hydrogel (w/w) PCS (15%) PCS (10%) PVAL (7 %) Agar (5%) [7] K-Carrageenan (4%) [7] Ca-alginate (1 %) [7] Modulus of elasticity [MPa] 0.039 0.023 0.03-0.04 0.20 0.28 0.19 Elongation at break [%] 80-120 80-120 100-240 10 27-29 55-70

Additionally, PCS and PVAL hydrogels show a higher flexibility in contrast to the natural polymers, which show a plastic behavior when compressed. The mechanical properties of the SI matrix with different porosities are estimated with a standard tensile-stress machine. SI membranes with different porosities are stressed and the force is recorded on the plotter. Figure 5 shows the influence of the elongation on the tensile stress and the elongation at break of SI membranes with different porosities. There is a decrease in the values for the elongation at break and the tensile stress (and therefore a decrease of the modulus of elasticity) by increasing porosities. The more the porosity increases in the SI material, the more the elasticity increases (lower modulus). The porous SI material shows a high elasticity and no abrasion, therefore it can be used in stirred tank and fluidized bed reactors in contrast to porous glass beads.

58 Hydrogel beads are produced by different apparatuses at our institute. Figure 1 (left) shows the conventional dropping process, in which a polymer solution is dropped into a cross-linking solution. Nevertheless, it is difficult to produce small (< 1.5 mm) PVAL hydrogel beads by this method. For this reason we developed a new special equipment [6] (Figure 2): The apparatus produces beads from a PVAL solution (dyn. viscosity = 200 mPa-s) by cutting the liquid jet coming out of a nozzle with the help of rotating wires. The beads are produced in a quantity of 10 kg beads/(h and nozzle) with a uniform diameter of 1 mm. Denitrifying bacteria were immobilized in PVAL hydrogels [5]

PVAL solution

top view

liquid nitrogen or cooled plant oilI . . ' s .

Figure 2. Production of small spherical PVAL hydrogel beads by cutting a jet of PVAL solution with the help of rotating wires (M = motor).

59 Stable support materials for the adsorption of viable cells Support materials for the adsorption of aerobic cells possess only a low oxygen permeation rate in general (e.g. porous glass beads). We developed a spherical porous silicone rubber (SI) support material for the adsorption of aerobic cells. The oxygen solubility in silicone and the oxygen permeation through silicone were measured with different apparatuses: We found a high solubility of oxygen in silicone and a high permeation rate through silicone. The permeability coefficient is more than 10 times higher than the coefficient of a PVAL or PCS hydrogel. Porous SI beads (diameter between 0.8 and 4 mm, porosity up to 65 %) for adsorption of microorganisms were produced by a suspension method (Figure 3): A silicone prepolymer is mixed with NaCl and poured into a hot (70C) and thoroughly stirred solution of glycerol. SI prepolymer/NaCl beads are formed during this process. After 1 h the SI prepolymer polymerizes to SI rubber. The beads are separated by a sieve and NaCl is washed out. Due to the size and the quantity of the NaCl spherical SI beads are obtained with a size distribution between 0.8 to 3 mm and a porosity up to 65 %.silicone prepolymer mixed with NaCl

/NaCl

#i

1m

. * Jsilicone prepolymer silicone prepolymer mixed with NaCl glycerol T= 70 X

separation washing out the NaCl use

Figure 3. Production

of porous

spherical

SI beads by a suspension

method.

60 Figure 4 shows a scanning electron micrograph of the porous surface of a SI bead (porosity 65 %). The pore size lays between 50 and 200 |j,m.

Figure 4. Scanning electron micrograph of the porous surface of a SI bead (porosity 65 %). The advantages of silicone beads are the following: They possess a high diffusivity, permeability and solubility of oxygen. SI beads are optimally sui