Citric Acid Biotechnology.pdf

Citric Acid Biotechnology

Citric Acid Biotechnology
BJØRN KRISTIANSEN Borregaard Industries Ltd, Norway MICHAEL MATTEY Department of Bioscience and Biotechnology, University of Strathclyde, UK JOAN LINDEN Gluppevelen 15, 1614 Fredikstad, Norway

UK Taylor & Francis Ltd, 1 Gunpowder Square, London EC4A 3DF USA Taylor & Francis Inc., 325 Chestnut Street, 8th Floor, Philadelphia, PA 19106

This edition published in the Taylor & Francis e-Library, 2002.
Copyright © Taylor & Francis 1999 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, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 0-7484-0514-3 (cased) Library of Congress Cataloguing-in-Publication Data are available Cover design by Jim Wilkie ISBN 0-203-48339-1 Master e-book ISBN ISBN 0-203-79163-0 (Glassbook Format)

Contents

1

A Brief Introduction to Citric Acid Biotechnology 1.1 Citric acid from lemons 1.2 Synthetic citric acid 1.3 Microbial citric acid 1.4 Citric acid by the surface method 1.5 The submerged process for production of citric acid 1.6 Continuous and immobilized processes 1.7 Yeast based processes 1.8 The koji process 1.9 Uses of citric acid 1.10 Effluent disposal 1.11 Conclusions 1.12 References

page 1 1 2 2 3 4 5 6 7 7 8 8 9 11 11 12 19 21 24 25 33 33 35 46 50 50 55 55 55
v

2 Biochemistry of Citric Acid Accumulation by Aspergillus niger 2.1 Introduction 2.2 Glucose catabolism in A. niger and its regulation 2.3 Regulation of citric acid biosynthesis 2.4 Role of citrate breakdown in citrate accumulation 2.5 Export of citric acid from A. niger 2.6 References 3 Biochemistry of Citric Acid Production by Yeasts 3.1 Introduction 3.2 Synthesis of citric acid from n-alkanes 3.3 Synthesis of citric acid from glucose 3.4 Conclusions 3.5 References 4 Strain Improvement 4.1 Introduction 4.2 General aspects of strain improvement

vi

Contents

4.3 4.4 4.5 4.6 4.7

Isolation of recombinant strains using the parasexual cycle in A. niger Genetic engineering Concluding remarks Acknowledgements References

60 61 64 65 65 69 69 69 70 74 82 82 82 85 85 85 86 87 88 89 91 95 95 101 102 103 105 105 107 113 119 119 121 121 122 123 125 128 130 131 132 135 135 136

5 Fungal Morphology 5.1 Introduction 5.2 Factors affecting Aspergillus niger morphology in submerged culture 5.3 Effect of agitation 5.4 Effect of nutritional factors 5.5 Effect of inoculum 5.6 Conclusions and perspectives 5.7 References 6 Redox Potential in Submerged Citric Acid Fermentation Nomenclature 6.1 Introduction 6.2 Overview 6.3 Theory 6.4 Measurement of redox potential 6.5 Significance of redox potential 6.6 Redox potential in citric acid fermentation 6.7 Regulation of the redox potential 6.8 Regulation of redox potential in citric acid fermentation 6.9 Scale-up based on redox potential 6.10 Conclusions 6.11 References 7 Modelling the Fermentation Process 7.1 Introduction 7.2 Aspergillus based models 7.3 Yeast based models 7.4 Conclusion 7.5 References 8 Mass and Energy Balance Nomenclature 8.1 Introduction 8.2 Metabolic description of A. niger growth 8.3 Mass and energy balances 8.4 Kinetics of growth and citric acid production by A. niger 8.5 Carbon and available electron balances 8.6 Conclusion 8.7 References 9 Downstream Processing in Citric Acid Production 9.1 Pretreatment of fermentation broth 9.2 Precipitation

Contents

vii

9.3 9.4 9.5 9.6 9.7 9.8 9.9

Solvent extraction Adsorption, absorption and ion exchange Liquid membranes Electrodialysis Ultrafiltration Immobilization of micro-organisms References

139 142 143 144 145 146 146 149 149 150 156 156 157 157 157 158 158 158 159 159 161 161 163 163 165 167 167 169 169 170 171 174 176 177 178 182 183 183 183 184 187

10 Fermentation Substrates 10.1 Introduction 10.2 Molasses 10.3 Refined or raw sucrose 10.4 Syrups 10.5 Starch 10.6 Hydrol 10.7 Alkanes 10.8 Oils and fats 10.9 Cellulose 10.10 Other medium redients 10.11 Conclusion 10.12 References 11 Design of an Industrial Plant Nomenclature 11.1 Design of an industrial plant 11.2 Data required 11.3 Design basis 11.4 Scope definition 11.5 Process package 11.6 Raw material 11.7 Substrate preparation 11.8 Fermentation 11.9 Design of a stirred tank reactor 11.10 Airlift and bubble column reactors 11.11 Product isolation 11.12 Cell removal 11.13 Purification 11.14 Crystallization stages 11.15 Product packaging 11.16 Effluent and by-products 11.17 In conclusion 11.18 References Index

Contributors

Ho Ai Meng Amy Blk 135 Pasir Ris Street 11, # 06-239, Singapore 510135

Marin Berovic Department of Chemistry and Biochemical Engineering, National Chemistry Laboratory for Biotechnology and Industrial Mycology, 1115 Slo, Ljubljana, Hajdrihova 19 POB 30, Slovenia

Pawel Gluszca Department of Bioprocess Engineering, Technical University of Lodz, Wolczanska 175 90-924 Lodz, Poland

Bjørn Kristiansen Borregaard Industries Ltd, PO Box 162, 1701 Sarpsborg, Norway

Liliana Krzystek Department of Bioprocess Engineering, Technical University of Lodz, Wolczanska 175 90-924 Lodz, Poland

Christian Kubicek Institute for Biochemical Technology and Microbiology, University of Technology Getreidemarkt 9/1725, A-1060 Wien, Austria

x

Contributors

Staniskaw Ledakowicz Department of Bioprocess Engineering, Technical University of Lodzul, Wolczanska 175 90-924 Lodz, Poland

Wladyslaw Lesniak Food Biotechnology Department, Academy of Economics, Komandorska 118/120 PL 53345 Wroclaw, Poland

Michael Mattey Department of Bioscience and Biotechnology, University of Strathclyde, Todd Centre 33 Taylor Street, Glasgow G4 0NR

Maria Papagianni 8 Kamvounion Street, 54 621 Thessaloniki, Greece

George Ruijter Section of Molecular Genetics of Industrial Microorganisms, Wageningen Agricultural University, Dreijentlaan 2, 6703 HA Wageningen, The Netherlands

Jacobus van der Merwe NCP, Project Engineering Division, PO Box 494, Germiston 1400, South Africa

Jaap Visser Section of Molecular Genetics of Industrial Microorganisms, Wageningen Agricultural University, Dreijentlaan 2, 6703 HA Wageningen, The Netherlands

Frank Wayman Department of Bioscience and Biotechnology, University of Strathclyde, Todd Centre 33 Taylor Street, Glasgow G4 0NR

Markus Wolschek Institute for Biochemical Technology and Microbiology, University of Technology Getreidemarkt 9/1725, A-1060 Wien, Austria

1

A Brief Introduction to Citric Acid Biotechnology
MICHAEL MATTEY AND BJØRN KRISTIANSEN

1.1 Citric acid from lemons
They are going to be squeezed, as a lemon is squeezed—until the pips squeak. My only doubt is not whether we can squeeze hard enough, but whether there is enough juice. (Sir Eric Geddes, 1918)

It is probably no more than a coincidence that Sir Eric Geddes uttered his now famous phrase at the time that the industrial production of citric acid by fungal fermentation was being developed to circumvent the high price and lack of availability of lemon juice. However, the association of taxation and squeezing lemons is appropriate, as the history of citric acid reflects the politics and economics of the era as well as the science. Indeed the production of citric acid is a ‘classical’ biotechnology phenomenon, where the science, though important, is secondary to the economics and politics of production. This book seeks to reflect that balance between practical science, fundamental understanding and economics. Citric acid derives its name from the Latin citrus, the citron tree, the fruit of which resembles a lemon. The acid was first isolated from lemon juice in 1784 by Carl Scheele, a Swedish chemist (1742–1786), who made a number of discoveries important to the advance of chemistry, amongst them hydrofluoric, tartaric, benzoic, arsenious, molybdic, lactic, citric, malic, oxalic, gallic and other acids as well as chlorine, oxygen (1772, published in English in 1780, predating the discovery by Priestly in 1774), glycerine and hydrogen sulphide. Citric acid was thus one amongst many natural organic acids. Citric acid was produced commercially from Italian lemons from about 1826 in England by John and Edmund Sturge, but with the increasing importance of citric acid as an item of commerce, production was started in Italy by the lemon growers, who established a virtual monopoly during the rest of the nineteenth century. Lemon juice remained the commercial source of citric acid until 1919 when the first industrial process using Aspergillus niger began in Belgium. Lemon juice itself remains an important product. World lemon production averages about 3.3 million metric tonnes (US Foreign Agricultural statistics); about 75 per cent comes from the United States, Italy, Spain and Argentina, with the rest from some 15 other producer countries.
1

2

Citric Acid Biotechnology

Figure 1.1 Synthesis of citric acid

Marketing of lemons is the subject of political control both in Europe and the USA. In Europe the processing of lemons to juice carries a processing subsidy which makes it attractive to process the lemons rather than sell them as fresh produce; additionally the EU intervention mechanism results in significant quantities of lemons being destroyed. In the USA marketing is controlled by the United States Department of Agriculture (USDA) Lemon Administrative Committee which determines how many lemons will be sold into the fresh market and what growing areas will be allowed to sell them. The economic result of any monopoly tends to be to make the product expensive; without the spur of competition the control of costs, the development of the process and the efficiency of production are neglected. Citric acid in the nineteenth century was no exception; the Italian monopoly resulted in high prices that tempted the entrepreneurs of the era to seek alternative sources of the increasingly useful product. Unable to find an alternative botanical source of citric acid, the nineteenth century advances in chemistry and microbiology were examined. By the turn of the century both possibilities existed.

1.2 Synthetic citric acid Citric acid had been synthesized from glycerol by Grimoux and Adams (1880) and later from symmetrical dichloroacetone (i) by treating with hydrogen cyanide and hydrochloric acid to give dichloroacetonic acid (ii), and converting this into dicyano-acetonic acid (iii) with potassium cyanide, which on hydrolysis yields citric acid (iv), as shown in Figure 1.1. Several other routes using different starting materials have since been published. All chemical methods have so far proved uncompetitive or unsuitable, mainly on economic grounds, with the starting material worth more than the end product, although poor yields due to the number of reaction steps in the synthesis and precautions necessary when handling hazardous compounds involved have contributed to the problem.

1.3 Microbial citric acid The concept of microbiological action yielding useful products followed from Pasteur’s pioneering studies on fermentation and resulted in systematic investigations of fungi and bacteria. Amongst them Wehmer, in 1893, showed that a ‘Citromyces’ (now Penicillium) accumulated citric acid in a culture medium containing sugars and inorganic salts. This work did not lead directly to a commercial process but the subsequent search for other organisms capable of this synthesis did. Many other organisms were found to accumulate citric acid including strains of Aspergillus niger, A. awamori, A. fonsecaeus, A. luchensis, A. phoenicus, A. wentii, A. saitoi, A. lanosius, A. flavus, Absidia sp., Acremonium sp., Aschochyta

A brief introduction to citric acid biotechnology

3

sp., Botrytis sp., Eupenicillium sp., Mucor piriformis, Penicillium janthinellum, P. restrictum, Talaromyces sp., Trichoderma viride and Ustulina vulgaris. Currie (1917) found strains of A. niger that produced citric acid when cultured in media with low pH values, high sugar levels and mineral salts. Prior to this A. niger was known to produce oxalic acid; the key difference was the low pH which, as we now know, suppressed both the production of oxalic acid, which would be toxic, and gluconic acid, which has a significantly higher production rate from sugar than citric acid. Currie subsequently joined Chas. Pfizer & Co. Inc. and his discovery formed the basis of the citric acid plant established in the USA by the firm in 1923. This plant and the other similar processes established first in Belgium then in England by J.E.Sturge, in Czechoslovakia and in Germany in the next few years used the ‘surface process’. The details of this process are not well documented despite its long history, due in part to the restriction of information by manufacturers. In biotechnological terms, citric acid is known as a bulk, or low value, product. The market is, and always has been, very competitive, so the profit margins are small. Improvements in productivity depend on the detail of the various processes, many of which are not easily protected by patents, so that secrecy is important and understandable.

1.4 Citric acid by the surface method The general details of the original process are straightforward. The fungal mycelium is grown as a surface mat on a liquid medium in a large number of shallow trays with a capacity of 50 to 100 litres. Each tray has a surface area of about 5 m2 and a depth of between 5 and 20 cm. The trays are manufactured from high purity aluminium or stainless steel and usually can be lifted by just two men. The trays are stacked in racks in a chamber to allow operation under relatively aseptic conditions. Various sucrose sources were used initially but cane molasses and then beet molasses soon became the norm as the sugar source. The molasses are diluted to the required concentration, usually 15 per cent and the pH adjusted to 5–7. After sterilization, the medium is pumped into the trays and inoculation carried out directly from spores, either by adding a liquid suspension or by blowing the spores in with the air stream. Aerating the chambers is important for two purposes, oxygenation and heat removal. The air requirement depends on the stage of growth. Initially sterile air at low rates is used to prevent contamination during the germination stage, which takes about 12 hours. Later, when growth is maximal, rates of up to 10 m3 per cubic metre medium per minute are needed to ensure heat dispersal. The heat generation is considerable, around 1 kJ h-1 m-3 medium and the surface and medium temperatures are ideally around 28°C to 30°C. This high volume air is not necessarily sterile, as contamination is normally not a problem once the pH has fallen, after about 24 hours growth. The pH falls to about 2, or slightly lower, and remains at that level until the end of the process, hence the need for high-grade materials for the construction of the trays. The incoming air is humidified to 40–60 per cent to prevent moisture loss from the high surface area of the medium. Cultivation continues for 8 to 15 days, with the objective of minimizing the residence time to maximize the plant productivity. The details of time, productivity and yield are closely guarded secrets, but productivity of the order of 1 kg per square metre per day can be obtained and yield is up to 75 per cent of the initial sugar level. At the end of the process, which can be monitored by total acid production or judged by experience, the mycelial mat is removed by filtration and washed, as it contains up to 15 per cent of the total citric acid. The washings and spent medium are treated with lime (calcium hydroxide) at about 90°C to precipitate the insoluble tri-calcium tetrahydrate

4

Citric Acid Biotechnology

salt of citric acid. It is not possible to crystallize the acid directly from the crude molasses medium although this can be done if pure sucrose is used as the carbon source. The precipitate of calcium citrate is washed and suspended in enough sulphuric acid to precipitate the calcium as calcium sulphate. This releases the citric acid into solution from where it can be treated further as required. The surface process, though commercially profitable for many years, is labour intensive and inefficient in its use of space; there is a limit as to how high a large tray can be lifted! The production of citric acid by surface culture was challenged at the beginning of the 1940s by the development of submerged fermentation processes. When Shu and Johnson published their work on the effect of medium ingredients and their concentrations on citric acid production in submerged culture, the fundamental technology for submerged production was ready to be exploited on an industrial scale (Shu and Johnson, 1948a, 1948b).

1.5 The submerged process for production of citric acid The submerged process has become the method of choice in the industrialized countries because it is less labour intensive, gives a higher production rate, and uses less space. Several designs of reactor have been used, particularly in pilot scale systems; the stirred tank reactor is the most common design although air-lift reactors, with a higher aspect ratio than the stirred tank reactor are also used. The reactors are constructed of high-grade stainless steel, an important requirement in view of the low pH levels developed, the ability of citric acid to solubilize metal ions and the presence of manganese in stainless steels. Inferior grades of steel have caused problems in the past, both of leaching and pitting or general corrosion. Industrial rumours suggest it may still happen though not by design! The empirical process of ‘conditioning’ a reactor, whereby a few batches are processed before optimal production levels are achieved, may be related to this problem. The other general requirement for reactors for citric acid production is the provision of aeration systems that can maintain a high dissolved oxygen level. With both tank and tower reactors sterile air is sparged from the base, although extra inputs are often used with tower reactors. The reactor may be held above atmospheric pressure to increase the rate of oxygen transfer into the fermentation broth. The influence of dissolved oxygen on citric acid formation has been examined and the dissolved oxygen levels are routinely monitored. The oxygen levels are also affected by the rheology of the broth. A typical plant will consist of four areas: medium preparation, reactor section, broth separation and product recovery. The medium preparation will involve dilution of the molasses, or other raw material, addition of nutrients and other pre-treatment such as ferrocyanide, and sterilization, either in-line or in the reactor. Where in-line sterilization is used the reactors are steam sterilized separately. It is usual to prepare an inoculum for the production reactor in a smaller reactor, in which the conditions may be modified to give rapid growth rather than product formation. Primary inoculation is by spores and the initial phase of the growth is critical. When a separate inoculum stage is used, the correct stage for transfer, characteristically between 18 and 30 hours, is judged by pH level. Production temperature, like the inoculum temperature, is about 30°C. The process is allowed to continue until the rate of citric acid production falls below a predetermined value, which is reached many hours before the production ceases altogether.

A brief introduction to citric acid biotechnology

5

Many reports suggest that the morphology of the mycelium is crucial to the ultimate yield; not only with respect to the shape of hyphae, but also their aggregation. Several studies suggest that hyphae should be abnormally short, bulbous and heavily branched. It is recognized that this condition is brought about by manganese deficiency or related to the addition of ferrocyanide, which is probably the same thing. The mycelium should also form small (less than 0.5 mm) pellets with a smooth, hard surface. Such pellets are produced when a number of factors are controlled, such as ferrocyanide levels, manganese levels, low iron (less than 1 ppm), low pH, control of aeration and agitation or the amount of spore inoculum. It is clear that this morphological appearance is not in itself necessary for a successful yield, but is a result of the correct process parameters. Pellet formation is not necessary, but does give a broth with a lower energy requirement for mixing. When a change to a filamentous growth type occurs, the dissolved oxygen level may fall by 50 per cent for a fixed input. That filamentous growth can give satisfactory yields has been demonstrated and consideration of the diffusion characteristics of pellets versus filamentous mycelium would suggest that while yields may be similar, productivity should be greater without the additional diffusional constraint of pellets. Aeration is a significant factor in the cost of the process, and although a constant aeration rate is used in many laboratory scale studies, the industrial practice is to use relatively low aeration rates initially (0.1 vvm) rising to 0.5–1 vvm as growth proceeds. Such aeration rates will lead to foaming and various devices and agents are available to minimize the problem. Although very high yields are possible, the productivity is a more important consideration on an industrial basis, and it is rare that the process is allowed to continue to the maximum yield. The processes run today owes much to the pioneering work carried out by D.S.Clark and his co-workers at the Northern Regional Research Laboratories in Canada during the 1950s and early 1960s. Here, the technology for large-scale production of citric acid with A. niger using molasses was established. After the fermentation characteristics were worked out, attention was given to the controlling mechanisms of the fermentation. Numerous reports have been published on the role of metal ions on the citric acid cycle, in particular. After decades of academic discussion, there is general agreement about the factors that regulate the fermentation and give rise to the high yields obtained in industry (Mattey, 1992).

1.6 Continuous and immobilized processes A process for continuous production of citric acid has been described (Kristiansen and Sinclair, 1979), but no commercial application of this has been made in spite of the high productivity values obtained (Kristiansen and Charley, 1981). The process does not use the carbon source as the limiting substrate so that excess sugar will pass out of the reactor. As the carbohydrate substrate is one of the major cost factors, the continuous process will be less efficient than the batch process. This might be overcome by using several reactors in series, but this offsets any advantage from the continuous process. Fed-batch processes have been used industrially so that the conversion of sugar concentrations greater than 15 per cent can be achieved, but the gain does not seem to be sufficient to allow the fed-batch method to become standard. The possibility of using the mycelium in an immobilized system has occurred to several workers and attempts on a small scale have been reported. Immobilization of

6

Citric Acid Biotechnology

mycelium in alginate beads or collagen proved possible, but with very low production rates. The difficulties of avoiding oxygen limitation when preparing beads, and preventing further growth, which reduces oxygen transfer rates, have led to the immobilization of conidia which are then grown under nitrogen limitation to the desired compact pellet. While giving a manageable system, the productivity was still too low to be of industrial interest. Other constructs for immobilization that have been more successful are the use of exchange filtration, and a rotating disc with an adhering mycelial film, reminiscent of sewage treatment techniques. These radical methods are unlikely to gain acceptance, even were they to give economic productivity gains, unless the engineering problems of scale-up can be overcome without making the capital costs too large.

1.7 Yeast based processes From about 1965 methods using yeasts were developed, first from carbohydrate sources, then from n-alkanes. At this time hydrocarbons were relatively cheap and plants were built to use the method. The economics have altered since then and plants that have been built to utilize both yeast technologies have apparently switched back to carbohydrate feedstocks. The potential advantages of using yeasts rather than filamentous fungi are the higher initial sugar concentrations that can be tolerated and the faster conversion rates possible. Further, the insensitivity to metal ions means that crude (and hence cheaper) grade molasses can be used without costly pre-treatment. Since 1968, when the patent for citric acid production from molasses by eight genera of yeasts was allowed, there have been many process modifications reported. Candida, Hansenula, Pichia, Debaromyces, Torulopsis, Kloekera, Trichosporon, Torula, Rhodotorula, Sporobolomyces, Endomyces, Nocardia, Nematospora, Saccharomyces, and Zygosaccharomyces species are known to produce citric acid from various carbon sources. Out of these genera the Candida species, including C. lipolytica, C. tropicalis, C. guillermondii, C. oleophila and C. intermedia have been used. The original process incorporated calcium carbonate into the medium to maintain a neutral pH, and generally a pH above 5.5 was used. Various additions have been proposed to reduce the isocitric acid contamination that afflicts yeasts even on carbohydrate media. Halogen substituted alkanoic mono- or di-substituted acids, n-hexadecyl citric acid or trans-aconitic acid, and even lead acetate have been patented, despite the possibility of toxic residues in the resulting citric acid. Many mutants have been selected for reduced isocitrate production. An osmophilic strain, which would convert sugar concentrations as high as 28 per cent without pre-treatment of the molasses substrate, has been patented. Tower reactors of fairly standard design are used, but with improved cooling systems as the rate of heat production is high. A continuous process has been described where the pH is maintained at 3.5 with ammonium hydroxide. The industrial production of citric acid from n-alkanes is not now economic, although a plant was built, and operated, around 1970 at Saline, Reggio Calabria, Italy (Liquichimica). This process was based on a low aconitase mutant of C. lipolytica in a batch process with stirred, aerated tank reactors of 400 m3, operating on a 72 hour cycle. The conversion from alkanes was reported to exceed 130 per cent (by weight). The theoretical yield is 250 per cent, but part of the alkanes was converted to biomass and carbon dioxide. The yeast was removed by centrifugation and the purification was traditional. The medium used was based

A brief introduction to citric acid biotechnology

7

on the process developed for the yeast strain that had a substrate concentration of 10 per cent n-decane, although n-alkanes from 9 to 20 carbons could be used. The availability and cost of Libyan n-alkanes, which lead to the development of this and other plants, including the dual substrate plants, has changed over the last three decades. One unique feature of the n-alkane process is the insolubility of the substrate. To ensure a rapid conversion the nalkane has to be thoroughly dispersed, so additives such as polyoxypropylene glycol ether, at concentrations from 20 to 200 ppm, are used to enhance this.

1.8 The koji process A third method for the production of citric acid is the koji process, using Aspergillus species. This is the solid state equivalent of the surface process described previously. It was originally developed in Japan where it uses the readily available rice bran and fruit wastes. It is confined to south-east Asia and is a relatively small-scale process. The carbohydrate source, which is principally starch and cellulose, is sterilized by steaming and the resulting semi-solid paste (about 70 per cent water), at a pH of about 5.5, is inoculated by spraying on spores of A. niger. Additions of ferrocyanide or copper may be made. The incubation temperature is 30°C and the process takes about four to five days. Yields are low because of the difficulty of controlling trace metals and the process parameters. The fungus produces sufficient cellulases and amylases to break down the substrate, though the low yields may reflect the rate limitations of this step.

1.9 Uses of citric acid Citric acid is used in food, confectionery and beverages, in pharmaceuticals and in industrial fields. Its uses depend on three properties: acidity, flavour, and salt formation. Chemically citric acid is 2-hydroxy-1,2,3-propane tricarboxylic acid (77-92-9). It has three pKa values at pH 3.1, 4.7 and 6.4. As these three values are relatively close together the second H+ is appreciably dissociated before the first is completed, and similarly with the third. Because of this overlapping the solution is well buffered throughout the titration curve and there are no breaks from about pH 2 (the approximate pH of a 0.2M solution) to pH 7. Citric acid forms a wide range of metallic salts including complexes with copper, iron, manganese, magnesium and calcium. These salts are the reason for its use as a sequestering agent in industrial processes and as an anticoagulant blood preservative. It is also the basis of its antioxidant properties in fats and oils where it reduces metal-catalysed oxidation by chelating traces of metals such as iron. There are two components to its use as a flavouring: the first is due to its acidity, which has little aftertaste; the second to its ability to enhance other flavours. A process to remove sulphur dioxide from flue gases has been developed where citric acid is used as a scrubber, forming a complex ion which then reacts with H2S to give elemental sulphur, regenerating citrate. This may become more important with increased environmental pressures. Citric acid esters of a range of alcohols are known; the triethyl, butyl and acetyltributyl esters are used as plasticizers in plastic films and monostyryl citrate is used instead of citric acid as an antioxidant in oils and fats. A summary of the uses of citric acid is given in Table 1.1.

8 Table 1.1 Applications of citric acid

Citric Acid Biotechnology

1.10 Effluent disposal Regardless of the method of production the disposal of waste is an increasing problem for manufacturers both from a cost and a regulatory viewpoint. Gypsum (calcium sulphate) is not valuable enough to purify and use in, for example, plaster. It may be disposed of to landfill sites, at a cost, and in some cases may be pumped out to sea, where tidal conditions permit. A more serious problem is the disposal of the filtrate from the precipitation where molasses has been used as a raw material; the waste is non-toxic, but has a high biological oxygen demand, so that it cannot be disposed of to rivers untreated. Anaerobic digestion, with fuel gas as a useful by-product, is probably the future method of choice, although animal feedstuff formulation in the form of condensed molasses solubles is another possibility. It can also be used as a medium for the growth of yeasts for animal feeds.

1.11 Conclusions
Books must follow science, not science books. (Francis Bacon, Propositions touching Amendment of Laws)

For the last 80 years citric acid has been produced on an industrial scale by the fermentation of carbohydrates, initially exclusively by Aspergillus niger, but in recent times by Candida yeasts as well, with the proportion derived from the Candida process increasing. The higher productivity of the yeast-based process suggests it will be the method of choice for any new plants that may be built. The intimate knowledge about the large-scale fermentation and subsequent recovery processes are still regarded as industrial property. Nevertheless, the citric acid process is

A brief introduction to citric acid biotechnology

9

one of the rare examples of industrial fermentation technology where academic discoveries have worked in tandem with industrial know-how, in spite of an apparent lack of collaboration, to give rise to a very efficient fermentation process. The current world market for citric acid and its derivatives is difficult to estimate accurately; no international statistics are collected, but industry estimates suggest that upwards of 400 000 tonnes per year may be produced. Citric acid is a ‘mature product’ but the upward trend in its use seen over many years is an annual 2–3 per cent increase. The price is such that profit margins are low, and with significant, but erratic, quantities appearing on the world market from countries such as China the situation is unlikely to improve. The lemon, which started it all, is doing well, with an estimated world production of 3 to 4 million tonnes per year. Commercial varieties such as ‘Eureka’ are all high acid lemons, with the acid content exceeding 4.5 per cent by weight, so that some 140 000 tonnes of citric acid are still produced by lemons! The various themes touched on in this introduction are dealt with in greater depth in the following chapters.

1.12 References
COOPER, W C and CHAPOT, H, 1977. Fruit Production—with special emphasis on fruit for processing. In Citrus Science and Technology, Vol. 2. Eds S Nagy, P E Shaw and M K Veldhuis (AVI Publishing Co., Westport, CT, USA). CURRIE, J N, 1917. The citric acid fermentation of A. niger, Journal of Biological Chemistry, 31, 5. GRIMOUX, E and ADAMS, P, 1880. Synthese de l’acide citrique, C.R.Hebd. Seances Acad. Sci., 90, 1252. KRISTIANSEN, B and CHARLEY, R C, 1981. The effect of medium composition on citric acid production in continuous culture, Presented at 2nd European Congress of Biotechnology, UK. KRISTIANSEN, B and SINCLAIR, C G, 1979. Production of citric acid in continuous culture, Biotechnology and Bioengineering, 21, 297. MATTEY, M, 1992. The production of organic acids, CRC Critical Reviews in Biotechnology, 12, 81. PASTEUR, L., 1875. Nouvelle observations sur la nature de la fermentation alcoolique, C.R. Acad. Sci., 80, 452. REUTHER, W, CALAVAN, E C and CARMAN, G E, 1967. The Citrus Industry, Vol. 1. History, World Distribution, Botany and Varieties. Univ. Calif. Div. Agric. Nat. Res., San Pablo, California. ROSENBAUM, J B, MCKINNEY, W A, BEARD, H L, CROCKER, L and NISSEN, W I, 1973. Sulphur Dioxide Emission Control by Hydrogen Sulphide Reaction in Aqueous Solution. The Citrate System. US Bureau of Mines, Report 1774. SCHEELE, C, 1793. Crells Ann. 2, 1 1784, from Sämmtliche Physische und Chemische Werke. Hermbstädt (Berlin). SHU, P and JOHNSON, M J, 1948a. Citric acid production submerged fermentation with Aspergillus niger, Industrial and Engineering Chemistry, 40, 1202. SHU, P and JOHNSON, M J, 1948b. The interdependence of medium constituents in citric acid production by submerged fermentation, Journal of Bacteriology, 54, 161. WEHMER, C, 1893. Note sur la fermentation Citrique, Bull. Soc. Chem. Fr, 9, 728.

2

Biochemistry of Citric Acid Accumulation by Aspergillus niger
MARKUS F.WOLSCHEK AND CHRISTIAN P.KUBICEK

2.1 Introduction The biochemical mechanism by which Aspergillus niger accumulates citric acid has attracted the interest of researchers since the late 1930s when the optimization of this accumulation to give a commercial process began. In this sense, the various theories which have been proposed to explain the accumulation of citric acid in such high yields also reflect the general biochemical knowledge at the time the respective research was done. In view of the high input into this research through more than 50 years it is therefore rather disappointing that there is still no explanation of the biochemical basis of this process which would consistently explain all the observed factors influencing this fermentation. Reasons for this are manifold. First, citric acid is only accumulated when several nutrient factors are present, either in excess (i.e. sugar concentration, H+, dissolved oxygen), or at suboptimal levels (trace metals, nitrogen and phosphate), and thus is subject to multifactorial influence. Hence it is unlikely that single biochemical events are solely responsible for citric acid overflow. Secondly, an appreciable part of the literature consists of work which has been performed using low or only moderately producing strains or by applying nutrient conditions not optimal for citric acid production, and while this may be justified for special reasons in individual cases, the respective results are not comparable to those obtained by others. Moreover, their significance for the understanding of the commercial citric acid fermentation is questionable. Thirdly, the biochemical knowledge of filamentous fungi is still significantly inferior to that of, for example, Saccharomyces cerevisiae or higher eukaryotes and, moreover, results from these sources cannot be uncritically transformed to filamentous fungi, which impedes a biochemically correct interpretation of results in several areas. Hence, although a considerable amount of basic biochemical research has been carried out with A. niger, the present state of understanding of the events relevant for citric acid accumulation (not to say production) is still a poorly resolved puzzle. This chapter attempts to draw the currently recognizable picture and to aid in the further fitting together of the other scattered bits and pieces.
11

12

Citric Acid Biotechnology

2.2 Glucose catabolism in A. niger and its regulation 2.2.1 The citric acid biosynthetic pathway It is well known, since the famous tracer studies by Cleland and Johnson (1954), and Martin and Wilson (1951), that citric acid is mainly formed via the reactions of the glycolytic pathway. Like most other fungi Aspergillus spp. utilize glucose and other carbohydrates for energy and cell synthesis by channelling glucose into the reactions of the glycolytic and the pentose phosphate pathway, respectively. The pentose phosphate pathway accounts for only a minor fraction of metabolized carbon during citric acid fermentation, and this decreases throughout prolonged cultivation (Legisa and Mattey, 1986; Kubicek, unpublished data). Legisa and Mattey (1988) speculated that this may be due to inhibition of 6-phosphogluconate dehydrogenase by citrate, but evidence for this is lacking. It should be noted that both arabitol and erythritol are accumulated as by-products until late stages of the fermentation (Roehr et al., 1987); hence a complete blockage of the pentose phosphate pathway is obviously not taking place. A. niger possesses a further pathway of glucose catabolism which is catalyzed by glucose oxidase (Hayashi and Nakamura, 1981). This enzyme is induced by high concentrations of glucose and strong aeration in the presence of low concentrations of other nutrients (Mischak et al., 1985; Rogalski et al., 1988; Dronawat et al., 1995), conditions which are also typical for citric acid fermentation; glucose oxidase will hence inevitably be formed during the starting phase of citric acid fermentation and convert a significant amount of glucose into gluconic acid. However, due to the extracellular location of the enzyme, it is directly influenced by the external pH and will be inactivated at pH <3.5 (Mischak et al., 1985; Roukas and Harvey, 1988). Because of the pKa values for citric acid, its accumulation decreases the pH of the culture filtrate to pH 1.8 thereby inactivating glucose oxidase (Mischak et al., 1985). It is not known if, and by which mechanism gluconic acid can be catabolized to citric acid during further fermentation. The catabolism of glucose via glycolytic catabolism leads to 2 moles of pyruvate, and their subsequent conversion to the precursors of citrate (i.e. oxaloacetate and pyruvate). Cleland and Johnson (1954) were the first to show that A. niger uses 1 mole of the carbon dioxide which is released during the formation of acetyl-CoA and 1 mole of pyruvate to form 1 mole of oxaloacetate (Figure 2.1a). This reaction is of utmost importance to high citric acid yields, because oxaloacetate could otherwise only be formed by one turn of the tricarboxylic acid cycle, which would be accompanied by the loss of two moles of CO2 and only two thirds of the carbon of glucose could therefore accumulate as citric acid (Figure 2.1b). The enzyme catalyzing this reaction was shown to be pyruvate carboxylase (Woronick and Johnson, 1960; Bloom and Johnson, 1962), which was characterized by Feir and Suzuki (1969) and Wongchai and Jefferson (1974). Unlike the enzyme from several other eukaryotes, the pyruvate carboxylase of A. niger is localized in the cytosol (Bercovitz et al., 1990; Jaklitsch et al., 1991). Glycolytic pyruvate will therefore be converted to oxaloacetate, and further to malate by the cytosolic malate dehydrogenase isoenzyme (Ma et al., 1981), thereby also regenerating 50 per cent of the glycolytically produced NADH (cf. Figure 2.2). It has been postulated (Kubicek, 1988) that, analogous to higher eukaryotes, the cytosolic malate may serve as the co-substrate of the mitochondrial tricarboxylic acid carrier, and that such an enhanced malate concentration may stimulate export of citrate from the mitochondrion. It should be noted that the fixation of carbon dioxide, while convincing and experimentally verified, does not seem to occur during the early phases of fermentation: Kubicek et al. (1979b), by continuously quantifying carbon

Biochemistry of citric acid accumulation by A. niger

13

Figure 2.1 Metabolic pathways from glucose to citric acid by (a) involvement of an anaplerotic carbon dioxide fixation (Cleland and Johnson, 1954), and (b) sole involvement of the citric acid cycle. Only relevant intermediates are given, and arrows may indicate more than a single enzymatic step. Note that in (b), each of the two acetyl-CoA molecules is subject to one turn of the tricarboxylic acid cycle.

dioxide and oxygen in the exit air of a pilot plant citric acid fermentation, observed that during the first 70 hours of fermentation the respiratory coefficient (i.e. CO2 released/O2 taken up) is close to 1; it starts to decrease thereafter and reaches the level predicted from the operation of the pyruvate carboxylase reaction (0.66) only at stages where citrate accumulation is already taking place at a constant rate (e.g. <120 hours). The Cleland and Johnson reaction may therefore only be important at later stages of fermentation, whereas the initial phase of citric acid accumulation takes place without anaplerotic carbon dioxide supply. Although not directly within the topic of this chapter, it should be noted that A. niger is also capable of accumulating another organic acid—oxalic acid—as a (toxic) byproduct of citric acid fermentation. The biosynthesis of this compound is controversial (Müller, 1975; Müller and Frosch, 1975; Kubicek et al., 1988), and appears to depend on whether glucose or citric acid is used as the carbon source. In the latter case, the glyoxalate cycle has been implicated in its biosynthesis (Müller, 1975). Its biosynthesis on glucose as a carbon source occurs by the hydrolysis of oxaloacetate catalyzed by oxaloacetate hydrolase, which is cytosolically located and appears to act as a valve by which the carbon overflow can be chanelled into an energetically neutral pathway (Figure 2.3) and so compete with citrate overproduction (Kubicek, 1988). Although production of oxalate is, because of its toxicity, of considerable interest to citric acid fermentation, the regulation of its biosynthesis is controversial (Kubicek et al., 1988; Strasser et al., 1994).

Figure 2.2 Metabolic and regulatory network of citric acid biosynthesis from sucrose in A. niger. For convenience, sucrose is assumed to be split into glucose and fructose by invertase extracellularly (Boddy et al., 1993; Rubio and Maldonado, 1995) and only the monosaccharides are taken up. The double line indicates the plasma membrane, the hatched double line the mitochondrial membrane. Circles inserted into the membranes indicate known or assumed transport steps (hatched: characterized in A. niger; full: assumed, but not yet characterized; empty: countertransport, to be verified). Thick lines and arrows indicate metabolic reactions; thin lines and arrows indicate regulatory interactions (*activation: // inhibition). Intermediates of regulatory importance are boxed.

Biochemistry of citric acid accumulation by A. niger

15

Figure 2.3 Pathway of oxalate biosynthesis by Aspergillus niger. Note that concentrations of acetate corresponding to those of oxalate have not been detected in culture filtrates of A. niger, and the metabolism of acetate therefore requires further study

2.2.2 Transcriptional regulation of the citric acid synthesizing pathway It is uncertain to what extent the apparently high flux through the glycolytic pathway, which is obviously necessary for citric acid accumulation, requires an activation of transcription of the genes encoding glycolytic and other enzymes (e.g. citrate synthase). The quantification of enzyme activities in cell-free extracts of A. niger mutants, which were selected according to a reduced lag in growth on high sucrose concentrations and correspondingly increased rates of citric acid accumulation, revealed enhanced hexokinase and phosphofructokinase activities (Schreferl-Kunar et al., 1989). Also, a class of A. niger mutants, resistant to 2desoxyglucose and displaying reduced hexokinase activity, exhibited decreased rates of citric acid production (Fiedurek et al., 1988; Kirimura et al., 1992; Steinböck et al., 1994). Torres et al. (1996a) showed that high glucose concentrations (>50 g/l) are a prerequisite for the formation of a low-affinity glucose transporter. However, knowledge of the transcriptional regulation of the respective genes is still lacking. Only preliminary data are as yet available to understand whether an enhancement of transcription of selected glycolytic genes would increase the rate of citric acid accumulation. Ruijter et al. (1996b) have—selectively and in combination—amplified the genes encoding phosphofructokinase 1 (pfkA) and pyruvate kinase (pkiA), but the rates of citrate accumulation by the moderately citric acid producing strain used (N400) were not increased. Torres (1994a, 1994b), using the biochemical system theory and a constrained linear optimization method, calculated that the activities (V max) of at least seven glycolytic enzymes must be simultaneously increased to obtain an effect. Clearly, such an increase can only be achieved by appropriate manipulation of the transcription factors regulating the genes encoding the enzymes for citric acid biosynthesis. Unfortunately, transcriptional regulation of glycolytic genes has not yet been studied in sufficient detail in A. niger nor in any related fungus. In Saccharomyces cerevisiae, mutations in the GCR1 gene, which encodes a DNA-binding protein (Baker, 1986, 1991), were found to exhibit strongly reduced levels of most glycolytic enzymes. Another protein, GCR2, was shown to interact physically with GCR1 (Uemura and Jigami, 1992). The authors proposed that both factors co-operate together in a transcriptional activation complex. Further factors involved in the regulation of the glycolytic genes have been described in yeast (RAP1, REB1, ABF1; Brindle et al., 1990; Chambers et al., 1990; McNeil et al., 1990; Huie et al., 1992). GCR1-binding sites are generally located near RAP1-binding sites (Huie et al., 1992). Furthermore, several glycolytic genes contain consensus binding sites for binding of ABF1 and REB1 in the vicinity of RAP1- and GCR1-binding sites (Brindle et al., 1990; Chambers et al., 1990; Chasman et al., 1990; Scott and Baker, 1993).

16

Citric Acid Biotechnology

Table 2.1 Genes encoding enzymes involved in the biosynthesis of citric acid by A. niger, which have already been cloned from A. niger or other Aspergillus spp.

It is intriguing that the above named binding sites have so far not been detected in the 5'noncoding sequences of the few glycolytic genes studied in A. niger or the close relative Aspergillus nidulans (Table 2.1, Figure 2.4). Transcriptional regulation of glyceraldehyde3-phosphate dehydrogenase (Punt et al., 1988, 1990, 1992) and of 3-phosphoglycerate kinase (Clements and Roberts, 1986; Streatfield et al., 1992) has been studied in some detail in the closely related fungus A. nidulans. Its transcription depends on positive control by several co-operating DNA-binding proteins since a truncated core promoter of the pgkA gene only containing the CAAT, TATA and CT-rich elements could not trigger transcription. Punt et al. (1988) identified a ‘glycolytic box’ as responsible for transcription. No differences in expression of gpdA were observed on 1% glucose or 0.1% fructose (Punt et al., 1990). A 24-bp region, which shares 60 per cent similarity with the ‘glycolytic box’, is also present at -638 and -488 of the pgkA promoter (Figure 2.4). However, another sequence, located between -161 and -120, in the pgkA promoter was shown to be essential for expression of the respective gene. It consists of two non-overlapping octameric sequences that match in seven out of eight nucleotides to the higher eukaryotic consensus ATGCAAAT (Falkner et al., 1986). A 17-base pair sequence was found in the 5'-regions of the A. nidulans and A. niger pkiA genes that may act as an upstream regulating sequence (de Graaff et al., 1992). This sequence was shown to be distinct from the proposed cis-acting element mediating increased transcription of pyruvate kinase on glycolytic carbon sources (de Graaff et al., 1988).

Figure 2.4 Regulatory nucleotide motifs present in the 5'-non transcribed sequences of three glycolytic genes of A. nidulans (gpdA, pgkA) and A. niger (pkiA). Only characterized nucleotide sequences are given. The boxed bar marker indicates 100 bp, and all genes are drawn to scale.

Table 2.2 Enzymes of A. niger, involved in citric acid biosynthesis and catabolism, with noteworthy regulatory properties

Biochemistry of citric acid accumulation by A. niger

19

2.2.3 Glucose metabolism and its regulation The disappointing result that amplification of selected genes did not lead to an increase in the rate of citric acid accumulation by A. niger indicates the operation of very tight fine control of at least some of the enzymes involved. In fact, this fine control was already a major target of investigation throughout the early 1980s, and as a consequence several of the enzymes involved in it have been comparably well characterized (Table 2.2). Most recently, the method to study the concentration of intracellular metabolites in A. niger has also been critically reassessed (Ruijter and Visser, 1996). Based on these data, Figure 2.2 shows those regulatory interactions between metabolites and enzymes which are believed to be of major importance to the regulation of citric acid biosynthesis in A. niger. Similar to the situation in yeast and higher eukaryotes, citrate and phosphoenolpyruvate (PEP) seem to be the major factors negatively affecting the glycolytic flux, whereas Fructose-2,6diphosphate (Fru-2,6-P2) and Fru-1,6-P2 appear to be the major activators.

2.3 Regulation of citric acid biosynthesis Citrate is one of the best known inhibitors of glycolysis, and the ability of A. niger to overproduce citrate by an active glycolytic pathway has therefore attracted biochemical interest for a long time; it is considered to be of major consequence for the fermentation rate (cf. Habison et al., 1979). However, under appropriate nutrient conditions (see below), this inhibition is more than counteracted by the accumulation of various positive effectors of PFK1 (NH4+, inorganic phosphate, AMP, Fru-2,6-P2), and hence this feedback does not occur (Habison et al., 1983; Arts et al., 1987). A series of investigations by Kubicek and co-workers favour the assumption that Fru2,6P2 may play a major role in the counteraction of citrate inhibition: Kubicek-Pranz et al. (1990) found that the triggering of citric acid accumulation by replacing A. niger in high concentrations (14% w/v) of sucrose or glucose (Shu and Johnson, 1948b; Xu et al., 1989a) is paralleled by a rise in the intracellular concentration of Fru-2,6-P2. Also, mycelia cultivated on carbon sources which allow higher yields of citric acid (i.e. those which are taken up rapidly; Hossain et al., 1984; Kubicek and Roehr, 1986; Honecker et al., 1989; Xu et al., 1989a) showed higher concentrations of Fru-2,6-P2. The concentration of Fru2,6-P2 correlates therefore positively with the rate of citrate production, and this fact may be responsible for the lack of citrate inhibition of PFK1. The reason for the increased F2,6-P2 level is not completely clear, but it appears to be due to an increased Fru-6-P supply for PFK2, since this enzyme is only poorly regulated in A. niger (Harmsen et al., 1992). The biosynthesis and regulation of Fru-2,6-P2 links regulation of PFK1 to that of earlier steps in glycolysis. Torres (1994a, 1994b) has recently concluded from theoretical calculations that a major part of the actual control of citric acid production must occur at hexose uptake and/or phosphorylation, which is in accordance with such an assumption. The biochemistry of these early steps in A. niger glycolysis is not completely clear, however Steinböck et al. (1994) found a single hexo/glucokinase only in the citric acid producing strain ATCC 11414, which was inhibited by citrate and weakly sensitive to trehalose-6phosphate (Arisan-Atac et al., 1996). The inhibition by citric acid was due to chelation of Mg2+ which is required to chelate the co-substrate ATP, and is most probably irrelevant under physiological conditions where Mg2+ is present in excess. However, the inhibition by trehalose-6-phosphate appears to be relevant to the flux towards citric acid, since a

20

Citric Acid Biotechnology

recombinant strain of A. niger, which carries a disrupted copy of the constitutively expressed trehalose-6-phosphate synthase gene tpsA (Wolschek and Kubicek, 1997), produces citric acid at increased rates (Arisan-Atac et al., 1996). Similarly, a strain bearing multiple copies of tpsA and hence overproducing trehalose-6-phosphate synthase exhibited a reduced rate of citrate production. These data indicate that the cellular level of trehalose-6-phosphate regulates the flux from glucose to citric acid and are thus in accordance with the conclusions of Torres et al. (1996a) that hexokinase most likely accounts for the major part of regulation at the early steps of glycolysis, thereby supplying an increased concentration of substrate for PFK2. However, most recently Panneman et al. (1996) reported on the isolation and characterization of a glucokinase from A. niger N400, a strain producing only low levels of citric acid, which has properties different from the hexo/glucokinase purified by Steinböck et al. (1994). They also concluded that by analogy with A. nidulans (Ruijter et al., 1996a) there may be also at least one separate hexokinase as well. The difference between the results of Steinböck et al. (1994) and Panneman et al. (1996) are currently unresolved. Hybridization of an A. niger ATCC 11414 DNA with a Kluyveromyces lactis hexokinaseencoding gene as a probe showed hybridization to a single fragment only (F.Narendja and C.P.Kubicek, unpublished data). The gene from strain ATCC 11414 has recently been cloned in our laboratory, and its characterization has to be awaited for clarification of this situation. Whatever the results of this investigation, the results by Arisan-Atac et al. (1996) clearly show that a relief from trehalose-6-phosphate inhibition positively influences the glycolytic flux at high sugar concentrations, and the hexose-phosphorylating step is therefore a major regulatory point in this fermentation. Glucose uptake by A. niger was investigated by Torres et al. (1996a). A. niger ATCC 11414 contains two transporters with different Km and Vmax. However, the high-affinity permease can only be detected during growth on low glucose concentration (1% w/v), whereas the low-affinity permease is detectable in the presence of high glucose concentrations. The latter may therefore contribute to the increased glycolytic flux during growth on high glucose concentrations. Several lines of evidence suggest that the regulation of PFK1 by Fru-2,6-P2 may not be the only parameter regulating citrate accumulation. Citrate inhibition of PFK1 also seems in vivo to be antagonized by ammonium ions (Habison et al., 1979). This antagonism is functionally linked to the well known effect of trace metal ions (particularly manganese ions) on citric acid accumulation (Shu and Johnson, 1948b; Tomlinson et al., 1950; Trumpy and Millis, 1963), as one of the effects caused by manganese deficiency is an impairment of macromolecular synthesis in A. niger (Kubicek et al., 1979a; Hockertz et al., 1987), which causes increased protein degradation (Kubicek et al., 1979a; Ma et al., 1985). As a consequence, mycelia accumulate elevated concentrations of NH4+ (Kubicek et al., 1979a). Proof for the role of manganese ions in this process has been obtained by the isolation of mutants of A. niger whose PFK1 was partially citrate-insensitive and whose citric acid accumulation was simultaneously more tolerant to the presence of Mn2+ (Schreferl et al., 1986). Furthermore, several authors have reported that the exogenous addition of NH4+ during citric acid fermentation even stimulates the rate of citrate production (Shepard, 1963; Choe and Yoo, 1991; Yigitoglu and McNeil, 1992), which is consistent with this effect of NH4+ on PFK1. The latter authors documented that both the time of addition as well as the concentration of NH4+ were important, and its addition during inappropriate fermentation phases even decreased acid accumulation. The reason for the impairment of macromolecular synthesis under manganese deficient cultivation conditions had originally been assumed to be at the translational level (Ma et

Biochemistry of citric acid accumulation by A. niger

21

al., 1985). However, Hockertz et al. (1987) have demonstrated that the absence of manganese ions from the nutrient medium of A. niger causes a reversible inhibition of DNA, but not RNA biosynthesis. This is supported by the findings that the effect of manganese deficiency can be mimicked by addition of hydroxyurea, an inhibitor of ribonucleotide reductase (Hockertz et al., 1987). They proposed that manganese deficiency may primarily impair DNA synthesis by causing a shortage of desoxyribonucleotides required for DNA replication. A further mechanism of regulation of PFK1 was proposed by Legisa and co-workers, who postulated that PFK1 is regulated by phosphorylation by cyclic-AMP dependent protein kinase A (Legisa and Bencina, 1994). They speculate that a high concentration of sucrose causes an increase in mycelial cyclic-AMP levels which trigger the phosphorylation of PFK1, thereby converting an inactive (non-phosphorylated) form into an active (phosphorylated) form (Legisa and Gradisnik-Grapulin, 1995). The support for their model is their observation that PFK1 was inactivated by treatment with alkaline phosphatase (Legisa and Bencina, 1994). However, this model, while intriguing, has to be treated cautiously until solid evidence for it has been obtained, as the molecular weight of the PFK1 purified by Legisa and Bencina (1994) and used for their studies was 48 kDa which is not that of native PFK1 (84 kDa). Moreover, the method section of their paper does not indicate whether (and how) the alkaline phosphatase had been removed or inactivated prior to the PFK1 assay. If this was not done, the ‘inactivation of PFK1’ may have been due to a removal of Fru-6-P from the assay and thus be an artefact. Proof for a regulation of PFK1 by phosphorylation is therefore still needed. A stimulation of citric acid accumulation by increased cyclic-AMP levels had also been postulated earlier (Wold and Suzuki, 1973, 1976a, 1976b). They showed that the stimulatory effect was dependent on the zinc concentration of the medium. Adenylate cyclase from A. niger has been described as Zn2+ dependent (Wold and Suzuki, 1974). A bottleneck of their investigations, however, is that they were using 1% (w/v) sucrose throughout, and hence the relevance of their findings to the effect of zinc under citric acid fermentation conditions is unclear. Xu et al. (1989b) studied the intracellular concentration of cyclic-AMP in A. niger during citric acid biosynthesis on media with and without Mn2+ ions added, and with high (14%) and low (1%) sucrose concentrations. They reported that the cyclic-AMP levels were growth rate dependent, and comparable if phases of similar growth rates were compared. Whether or not cyclic-AMP is in fact involved in the regulation of citrate overproduction remains to be assessed.

2.4 Role of citrate breakdown in citrate accumulation 2.4.1 Role of the citric acid cycle The reason why A. niger accumulates such massive amounts of citric acid has, since the early studies by Ramakrishnan et al. (1955), attracted numerous investigations. Although citrate has been considered an ‘overflow’ product (Foster, 1949), which implies that it accumulates as a result of an excessive substrate supply rather than a limited catabolism, an excessive amount of work has been concerned with the attempt to identify a bottleneck in the tricarboxylic acid cycle as the reason for its accumulation. Numerous workers claimed that inactivation of an enzyme degrading citrate (e.g. aconitase or the isocitrate dehydrogenases) would be essential for the accumulation of citric acid (for review see Smith et al., 1974; Berry et al., 1977; Roehr et al., 1983, 1996; Kubicek and Roehr,

22

Citric Acid Biotechnology

1986). While this view has an extraordinary long half-life in the review literature, solid evidence for the presence of an intact citric acid cycle during citric acid fermentation was presented 25 years ago (Ahmed et al., 1972), and explanations based on this view are therefore simply incorrect. The requirement of citric acid accumulation of a deficiency in some metal ions (e.g. Mn2+, Fe3+) has frequently been used to explain an inhibition of some enzymes of the TCA cycle (for review see Kubicek and Roehr, 1986). Thus, iron deficiency has been claimed to inhibit aconitase (Szczodrak and Ilczuk, 1985). However, the activity of this enzyme during citric acid accumulation has been demonstrated clearly by others both in vitro (La Nauze, 1966; Ahmed et al., 1972; Mattey, 1977) as well as in vivo (Kubicek and Roehr, 1985). It should be kept in mind that the enzymes of the respiratory chain, which also require iron, are highly active during citric acid accumulation (Ahmed et al., 1972; Hussain et al., 1978). By a similar rationale, the necessity for Mn2+ deficiency has been used to claim an inhibition of either of the two isocitrate dehydrogenases which require divalent metal ions for activity (cf. Gupta and Sharma, 1995). However, this requirement is for chelation of the substrate (i.e. isocitrate; cf. Bowes and Mattey, 1979; Meixner-Monori et al., 1986). In view of the fact that Mg2+ (which is present in excess) can take over the chelating role of Mn2+ efficiently, this interpretation is unlikely to explain the effect of Mn2+. Several other explanations for citric acid accumulation are based on the postulation of a metabolic inhibition of the NADP-specific isocitrate dehydrogenase by citrate (Mattey, 1977) or glycerol (Legisa and Mattey, 1986), which would create a bottleneck in the tricarboxylic acid cycle and—because of the Keq of aconitase—lead to a spilling over of citrate. Unfortunately, none of the explanations which are based on an inhibition of NAD-or NADPspecific isocitrate dehydrogenases have ever been supported by evidence from in vivo experiments. The ‘glycerol theory’ (Legisa and Mattey, 1986; Gradisnik-Grapulin and Legisa, 1996), has recently been reassessed by studying the effect of increased mycelial glycerol concentrations on the oxidation of 1,5–14C-citrate by mycelia and isolated mitochondria of A. niger (Arisan-Atac and Kubicek, 1996). The appearance of 14C-labelled CO2—which because of the labelling position applied can only be released during the metabolic conversion of citrate to a-ketoglutarate—was virtually unaffected by the glycerol concentration, thereby clearly disproving an effect of glycerol on the activity of isocitrate dehydrogenases and consequently this theory. Also, in contrast to the enzyme from crude cell-free extracts (Legisa and Mattey, 1986), the purified NADP-specific isocitrate dehydrogenase was not inhibited by citrate (Arisan-Atac and Kubicek, 1996). It is surprising that the question of whether the isocitrate dehydrogenase step of the TCA cycle is active during citric acid fermentation or not has never been viewed from a theoretical point of view: using the cellular concentration of free and protein-bound glutamic acid as an indicator of metabolic flux from glucose to a-ketoglutarate, there is no indication for a significant change in this flux unless at late stages of fermentation where the fungal growth (and also the need for glutamic acid) has stopped, and this flux is only 17 per cent lower than that occurring in a culture accumulating 78 per cent less citric acid, and hence may not be of high relevance to the mechanism of citric acid accumulation (O.Zehentgruber and C.P.Kubicek, unpublished data). With regard to the mechanisms which trigger the initial accumulation of citrate from the mitochondria, a fact completely overlooked so far is the activity of the tricarboxylate transporter. This carrier competes directly with aconitase for citrate, and if its affinity for citrate were much higher than that of aconitase, would pump citrate out of the mitochondria without any necessity for inhibition of one of the TCA cycle enzymes. As the tricarboxylate carrier of mammalian tissues and yeast occurs by countertransport with malate (Evans et

Biochemistry of citric acid accumulation by A. niger

23

al., 1983), such a situation is conceivable when its counter-ion malate accumulates in the cytosol. Malate accumulation has in fact been shown to precede citrate accumulation (Roehr and Kubicek, 1981). However, the mitochondrial citrate carrier of A. niger has not yet been investigated, and this hypothesis clearly needs thorough investigation before it can be used to explain citrate accumulation. It is also not known to what extent changes in the flux through the NAD-dependent-, NADP-dependent isocitrate dehydrogenases, a-ketoglutarate dehydrogenase and succinate dehydrogenase, contribute to a rise in the intramitochondrial citrate concentration. As these enzymes are known to be regulated by the mitochondrial NADH/NAD and NADPH/NADP ratios, as well as by AMP, cis-aconitate and oxaloacetate (Chan et al., 1965; Meixner-Monori et al., 1985, 1986), fluctuations in the level of mitochondrial TCA metabolites are likely.

2.4.2 Respiratory activity and the role of NAD regeneration Formation of citric acid is dependent on strong aeration, and dissolved oxygen tensions higher than those required for vegetative growth of A. niger stimulate citric acid fermentation (Clark and Lentz, 1961; Kubicek et al., 1980; Dawson et al., 1988a). On the other hand, sudden interruptions in the air supply cause an irreversible impairment of citric acid production without any harmful effect on mycelial growth (Kubicek et al., 1980; Dawson et al., 1988b). The biochemical basis of this observation appears to be related to the presence of an alternative respiratory pathway, which is obviously required for re-oxidation of the glycolytically produced NADH, by a continuously maintained, high oxygen tension (Kubicek et al., 1980; Zehentgruber et al., 1980; Kirimura et al., 1987, 1996), whose activity is impaired by short interruptions in the air supply (Kubicek et al., 1980). Weiss and colleagues (Wallrath et al., 1991; Schmidt et al., 1992; Prömper et al., 1993) studied the role of the standard and alternative respiratory pathways in citric acid accumulation in detail. They detected that the assembly of the proton pumping NADH:ubiquinone oxidoreductase is impaired during citric acid accumulation (Schmidt et al., 1992), which could be the reason for the importance of the activity of the alternative pathway. Interestingly, disruption of the gene encoding the NADH-binding subunit of complex I in a low producing strain of A. niger increased its catabolic overflow, yet this strain excreted much less citrate than its parent (Prömper et al., 1993). These findings stress the fact that citric acid accumulation is not a mono-causal process, and citrate accumulation in high amounts depends on a delicate balance of several factors, whose interrelationship is not yet fully understood. The requirement of a high oxygen supply is also related to another effect of Mn2+ ions on A. niger, i.e. on the morphology of the fungus: whereas A. niger grows in long and smooth filaments when supplied with optimal concentrations of Mn2+ ions, Mn2+ deficient grown mycelia are strongly vacuolated, highly branched, contain strongly enthickened cell walls and exhibit a bulbous appearance (Kisser et al., 1980; Papagianni et al., 1994). This type of morphology has been shown to provide a much better rheology (Olsvik et al., 1993) and enables a higher oxygen transfer (Fujita et al., 1994; Iwahori et al., 1995); it may thus be required for optimal citric acid yields. NAD regeneration may also be related to the effect of pH on citric acid fermentation: the almost quantitative conversion of glucose to citric acid, as occurs during the idiophase of fermentation, yields 1 ATP and 3 NADH. While part of the NADH pool can be reoxidized by the alternative, salicylhydroxamic acid (SHAM) sensitive respiratory pathway described above, this yield of ATP probably still exceeds that of the cell’s maintenance demands. Roehr et al. (1992) speculated that the ATP will be consumed by the plasma membrane

24

Citric Acid Biotechnology

bound ATPase during maintenance of the pH gradient between the cytosol and the extracellular medium. The involvement of this enzyme in the maintainance of the pH gradient in citric acid producing A. niger has been shown by Mattey et al. (1988). Hence the requirement of a low pH for citric acid accumulation may be, at least in part, related to a high turnover of the ATP formed, which otherwise would lead to a metabolic imbalance and so stop acidogenesis. However, this explanation still requires experimental verification. Most recently, single-point mutagenesis of a plant ATPase and its expression in yeast resulted in increased H+-pumping and increased growth rates at low pH (Morsomme et al., 1996).

2.5 Export of citric acid from A. niger Torres et al. (1994a) proposed that the two citric acid transport steps, i.e. that from the mitochondria to the cytosol, and that from mycelia into the culture filtrate, are among the most important regulatory points for the obtention of high yields. The mechanism of transport of mitochondrial citrate into the cytosol is still completely unknown, except for the hypothesis that it occurs by countertransport with the glycolytically overproduced malate (see above). ATP-citrate lyase, an enzyme which in other cells uses the cytosolic citrate for lipid biosynthesis, appears to be unable to manage this high efflux but its precise regulation under citric acid producing conditions is not understood (Pfitzner et al., 1987; Jernejc et al., 1991). The latter authors have also purified a cytosolic and a mitochondrial carnitine acetyltransferase from A. niger, which exhibited similar kinetic and physicochemical properties (Jernejc and Legisa, 1996). As the activity of this enzyme was in considerable excess of that of ATP-citrate lyase they concluded that transfer as a carnitine ester may be the major physiological source of acetyl-CoA for lipid biosynthesis. If this is indeed the case it would explain why the cytosolic citrate pool is rather stable. Because of their findings of a cytosolic isoenzyme of carnitine acetyltransferase, Jernejc and Legisa (1996) also speculated that this enzyme transfers acetyl-CoA to the mitochondria and thus for citrate biosynthesis. This is an intriguing speculation, but requires the identification of a cytosolic pathway from pyruvate to acetylCoA which is not yet known. Mattey and co-workers (Mattey, 1992; Kontopidis et al., 1995) explained the export of citrate through the plasma membrane in terms of the large pH gradient between the cytosol and the extracellular medium, and postulated that citrate efflux from the cells may occur by diffusion of the 2(-) citrate anion, driven by a gradient. If this assumption is correct, the low pH would be responsible for the citrate gradient necessary for transport and consequently less citrate would be secreted at higher pH values. However, recent studies in our laboratory clearly showed that citrate export requires ATP, and its Vmax is not strongly affected by the external pH (Netik et al., 1997); this renders the diffusion hypothesis rather unlikely. Netik et al. (1997) also reported that citrate export is strongly increased in mycelia grown under manganese deficiency, which is consistent with previous observations that the intracellular concentration of citrate in manganese sufficient and deficient grown mycelia is not greatly different (Kubicek and Roehr, 1985; Legisa and Kidric, 1989; Prömper et al., 1993), despite the five- to seven-fold higher extracellular levels under the latter conditions. The reason for the requirement of manganese deficiency for citrate export is not clearly understood, but may be related to an absolute requirement of citrate uptake for manganese ions, probably because of a requirement for chelated citrate as a substrate for the permease (Netik et al., 1997).

Biochemistry of citric acid accumulation by A. niger

25

The reason for the reciprocal effect of Mn2+ ions on export and import of citric acid may also be related to yet another effect of manganese deficiency, i.e. inhibition of triglyceride and phospholipid synthesis as well as a shift in the ratio of saturated to unsaturated fatty acids of whole mycelial lipids (Orthofer et al., 1979; Jernejc et al., 1989) and of isolated plasma membranes (Meixner et al., 1985). The different behaviour of the citrate export and import system of A. niger may also be seen in the light of earlier studies on the antagonism of several membrane affecting compounds on the detrimental action of manganese ions, e.g. lower alcohols (Moyer, 1953), lipids (Millis et al., 1963; Gold and Kieber, 1967), or tertiary amines (Batti, 1969). Also the technically important ability of Cu2+ ions to antagonize the deleterious effect of Mn2+ may be related to citrate excretion, as Cu2+ strongly inhibited the uptake of citric acid from the medium (Netik et al., 1997). However, the effect of Cu2+ (Schweiger, 1959) may also reside in its inhibition of the uptake of Mn2+ by A. niger (Hockertz et al., 1987), which occurs by a specific, high affinity transport system (Seehaus et al., 1990). The properties of the uptake and the export system are otherwise similar (?pH driven proton symport) and it may be speculated that they are catalyzed by the same enzyme system.

2.6 References
AHMED, S A, SMITH, J E and ANDERSON, J G, 1972. Mitochondrial activity during citric acid production by Aspergillus niger, Transactions of the British Mycological Society, 59, 51–61. ARISAN-ATAC, I and KUBICEK, C P, 1996. Glycerol is not an inhibitor of mitochondrial citrate oxidation in Aspergillus niger, Microbiology UK, 142, 2937–2942. ARISAN-ATAC, I, WOLSCHEK, M and KUBICEK, C P, 1996. Trehalose-6-phosphate synthase A affects citrate accumulation by Aspergillus niger under conditions of high glycolytic flux, FEMS Microbiology Letters, 140, 77–83. ARTS, E, KUBICEK, C and ROEHR, M, 1987. Regulation of phosphofructokinase from Aspergillus niger: effect of fructose-2,6-bisphosphate on the action of citrate, ammonium ions and AMP, Journal of General Microbiology, 133, 1195–1199. BAKER, H V, 1986. Glycolytic gene expression in Saccharomyces cerevisiae: nucleotide sequence of GCR1, null mutations, and evidence for expression, Molecular and Cellular Biology, 6, 3774–3784. BAKER, H V, 1991. GCR1 of Saccharomyces cerevisiae encodes a DNA binding protein whose binding is abolished by mutations in the CTTCC sequence motif, Proceedings of the National Academy of Sciences of the USA, 88, 9443–9447. BATTI, M A, 1969. US Patent 3,438.863. BERCOVITZ, A, PELEG, Y, BATTAT, E, ROKEM, J S and GOLDBERG, I, 1990. Localisation of pyruvate carboxylase in organic acid producing Aspergillus strains, Applied and Environmental Microbiology, 56, 1594–1597. BERRY, D R, CHMIEL, A and AL-OBAIDY, Z, 1977. In: Genetics and Physiology of Aspergillus. Eds J E SMITH and J A PATEMAN (Academic Press, London), pp. 405–423. BLOOM, S J and JOHNSON, M J, 1962. The pyruvate carboxylase of Aspergillus niger, Journal of Biological Chemistry, 237, 2718–2720. BODDY, L M, BERGES, T, BARREAU, C, VAINSTAIN, M H, JOBSON, M J, BALLANCE, D J and PEBERDY, J F, 1993. Purification and characterization of an Aspergillus niger invertase and its DNA sequence, Current Genetics, 24, 60–66. BOWES, I and MATTEY, M, 1979. A study of mitochondrial NADP +-specific isocitrate dehydrogenase from selected strains of Aspergillus niger, FEMS Microbiology Letters, 7, 323– 325.

26

Citric Acid Biotechnology

BRINDLE, P K, HOLLAND, J P, WILLET, C E, INNIS, M A and HOLLAND M J, 1990. Multiple factors bind the upstream activation sites of the yeast endolase genes ENO1 and ENO2: ABF1 protein, like repressor activator protein RAP1, binds cis-acting sequences which modulate repression or activation of transcription, Molecular and Cellular Biology, 10, 4872–4895. CHAMBERS, A, STANWAY, C, TSANG, J S H, HENRY, Y, KINGSMAN, A J and KINGSMAN, S M, 1990, ARS binding factor 1 binds adjacent to RAP1 at the UASs of the yeast glycolyticgenes PGK and PYK1, Nucleic Acids Research, 18, 5393–5399. CHAN, M F, STACHOV, C S and SANWAL, B D, 1965. The allosteric nature of NAD-specific isocitric dehydrogenase of Aspergilli, Canadian Journal of Biochemistry, 43, 111–118. CHASMAN, D I, LUE, N F, BUCHMAN, A R, LAPOINTE, J W, LORCH, Y and KORNBERG, R D, 1990. A yeast protein that influences the chromatin structure of UASG and functions as a powerful auxiliary gene activator, Genes and Development, 4, 503–514. CHOE, J, and YOO, Y J, 1991. Effect of ammonium ion concentration and application to fed-batch culture for overproduction of citric acid, Journal of Fermentation and Bioengineering, 72, 106–109. CLARK, D S and LENTZ, C P, 1961. Submerged citric acid fermentation on beet molasses: effect of pressure and recirculation of oxygen, Canadian Journal of Microbiology, 7, 447–452. CLARK, D S, ITO, K and HORITSU, H, 1966. Effect of manganese and other heavy metals on submerged citric acid fermentation of molasses, Biotechnology and Bioengineering, 8, 465– 471. CLELAND, W W and JOHNSON, M J, 1954. Tracer experiments on the mechanism of citric acid formation by Aspergillus niger, Journal of Biological Chemistry, 208, 679–692. CLEMENTS, J M and ROBERTS, C F, 1986. Transcription and processing signals in the 3phosphoglycerate kinase (PGK) gene from Aspergillus nidulans, Gene, 44, 97–105. DAWSON, M W, MADDOX, I S, BOAG, I F, and BROOKS, J D, 1988a. Application of fed-batch culture to citric acid production by Aspergillus niger: the effects of dilution rate and dissolved oxygen tension, Biotechnology and Bioengineering, 32, 220–226. DaWSON, M W, MADDOX, I S, and BROOKS, J D, 1988b. Effect of interruptions to the air supply on citric acid production by Aspergillus niger, Enzyme and Microbial Technology, 8, 37–40. DE GRAAFF, L and VISSER, J, 1988. Structure of the Aspergillus nidulans pyruvate kinase gene, Current Genetics, 14, 553–560. DE GRAAFF, L, VAN DEN BROECK, H and VISSER, J, 1988. Isolation and characterisation of the pyruvate kinase gene of Aspergillus nidulans, Current Genetics, 13, 315–321. DE GRAAFF, L, VAN DEN BROECK, H and VISSER, J, 1992. Isolation and characterisation of the Aspergillus niger pyruvate kinase gene, Current Genetics, 22, 21–27. DRONAWAT, S N, SVIHLA, C K and HANLEY, T R, 1995. The effects of agitation and aeration on the production of gluconic acid by Aspergillus niger, Applied Biochemistry and Biotechnology, 51/52, 347–354. EVANS, C T, SCRAGG, A H and RATLEDGE, C, 1983. A comparative study of citrate efflux from mitochondria of oleaginous and non-oleaginous yeasts, European Journal of Biochemistry, 130, 195–204. FALKNER, F G, MOCIKAT, R and ZACHAU, H G, 1986. Sequences closely related to an immunoglobulin gene promoter/enhancer element occur also upstream of other eukaryotic and of prokaryotic genes, Nucleic Acids Research, 14, 8819–8827. FEIR, H A and SUZUKI, I, 1969. Pyruvate carboxylase of Aspergillus niger: kinetic study of a biotin-containing enzyme, Canadian Journal of Biochemistry, 47, 697–710. FIEDUREK, J, SZCZODRAK, J and ILCZUK, Z, 1988. Citric acid synthesis by Aspergillus niger mutants resistant to 2-deoxy-D-glucose: decreased synthesis with biomass increase, Acta Microbiologica Polonica, 36, 303–307. FOSTER, J W, 1949. Chemical Activities of Fungi (Academic Press, London). FUJITA, M, IWAHORI, K, TATSUTA, S and YAMAKAWA, K, 1994. Analysis of pellet formation of Aspergillus niger based on shear stress, Journal of Fermentation and Bioengineering, 78, 368–373.

Biochemistry of citric acid accumulation by A. niger

27

GOLD, W and KIEBER, R J, 1967. Process of making citric acid by fermentation, US Patent 337,2094. GRADISNIK-GRAPULIN, M and LEGISA, M, 1996. Comparison of specific metabolic characteristics playing a role in citric acid excretion between some strains of the genus Aspergillus, Journal of Biotechnology, 45, 265–270. GUPTA, S, and SHARMA, C B, 1995. Citric acid fermentation by the mutant strain of the Aspergillus niger resistant to manganese ions inhibition, Biotechnology Letters, 17, 269–274. HABISON, A, KUBICEK, C P and ROEHR, M, 1979. Phosphofructokinase as a regulatory enzyme in citric acid accumulating Aspergillus niger, FEMS Microbiology Letters, 5, 39–42. HABISON, A, KUBICEK, C P and ROEHR, M, 1983. Partial purification and regulatory properties of phosphofructokinase from Aspergillus niger, Biochemical Journal, 209, 669–676. HARMSEN, H, KUBICEK-PRANZ, E M, VISSER, J, ROEHR, M and KUBICEK, C P, 1992. Regulation of 6-phosphofructo-2-kinase from the citric acid producing fungus Aspergillus niger, Applied Microbiology and Biotechnology, 37, 784–787. HAYASHI, S and NAKAMURA, S, 1981. Multiple forms of glucose oxidase with different carbohydrate compositions, Biochimica et Biophysica Acta, 657, 40–51. HOCKERTZ, S, PLÖNZIG, J and AULING, G, 1987. Impairment of DNA formation is an early event in Aspergillus niger under manganese starvation, Applied Microbiology and Biotechnology, 25, 590–593. HOCKERTZ, S, SCHMID, J and AULING, G, 1987. A specific transport system for manganese in the filamentous fungus Aspergillus niger, Journal of General Microbiology, 133, 3513–3519. HONECKER, S, BISPING, B, YANG, Z and REHM, H-J, 1989. Influence of sucrose concentration and phosphate limitation on citric acid production by immobilized cells of Aspergillus niger, Applied Microbiology and Biotechnology, 31, 17–24. HOSSAIN, M, BROOKS, J D and MADDOX, I S, 1984. The effect of sugar source on citric acid production by Aspergillus niger, Applied Microbiology and Biotechnology, 19, 383–391. HUIE, M A, SCOTT, E W, DRAZINIC, C M, LOPEZ, M C, HORNSTRA, I K, YANG, T P and BAKER, H V, 1992. Characterization of the DNA-binding activity of GCR1: in vivo evidence for two GCR1-binding sites in the upstream activating sequences of TPI of Saccharomyces cerevisiae, Molecular and Cellular Biology, 12, 2690–2700. HUSSAIN, M, RAHMAN, R, CHOUDHURY, N and HUSSAIN, I, 1978. Studies on mitochondrial respiration of some high citric acid-yielding mutants of Aspergillus niger, Journal of Fermentation Technology, 56, 253–256. IWAHORI, K, TATSUTA, S, FUJITA, M and YAMAKAWA, M, 1995. Substrate permeability in pellets formed by Aspergillus niger, Journal of Fermentation and Bioengineering, 79, 387– 390. JAKLITSCH, W M, KUBICEK, C P and SCRUTTON, M C, 1991. Intracellular organisation of citrate production in Aspergillus niger, Canadian Journal of Microbiology, 37, 823–827. JERNEJC, K and LEGISA, M, 1996. Purification and properties of carnitine acetyltransferase from citric acid producing Aspergillus niger, Applied Biochemistry and Biotechnology, 60, 151– 158. JERNEJC, K, VENDRAMIN, M and CIMERMAN, A, 1989. Lipid composition of Aspergillus niger in citric acid accumulating and nonaccumulating conditions, Enzyme and Microbial Technology, 11, 452–456. JERNEJC, K, PERDIH, A and CIMERMAN, A, 1991. ATP:citrate lyase and carnitine acetyltransferase activity in a citric-acid-producing Aspergillus niger strain, Applied Microbiology and Biotechnology, 36, 92–95. KIRIMURA, K, HIROWATARI, Y and USAMI, S, 1987. Alterations of respiratory systems in Aspergillus niger under the conditions of citric acid fermentation, Agricultural Biological Chemistry, 51, 1299–1303. KIRIMURA, K, SARANGBIN, S, RUGSASEEL, S and USAMI, S, 1992. Citric acid production by 2-deoxyglucose-resistant mutant strains of Aspergillus niger, Applied Microbiology and Biotechnology, 36, 573–577.

28

Citric Acid Biotechnology

KIRIMURA, K, MATSUI, T, SUGANO, S and USAMI, S, 1996. Enhancement and repression of cyanide-insensitive respiration in Aspergillus niger, FEMS Microbiology Letters, 141, 251– 254. KISER, R-C and NIEHAUS, W G JR, 1981. Purification and kinetic characterization of mannitol1-phosphate dehydrogenase from Aspergillus niger, Archives of Biochemistry and Biophysics, 211, 613–621. KISSER, M, KUBICEK, C P and ROEHR, M, 1980. Influence of manganese on morphology and cell-wall composition of Aspergillus niger during citric acid fermentation, Archives of Microbiology, 128, 26–33. KONTOPIDIS, G, MATTEY, M and KRISTIANSEN, B, 1995. Citrate transport during the citricacid fermentation by Aspergillus niger, Biotechnology Letters, 17, 1101–1106. KUBICEK, C P, 1988. The role of the citric acid cycle in fungal organic acid fermentations. Biochemical Society Symposia, 54, 113–126. KUBICEK, C P and ROEHR, M, 1980. Regulation of citrate synthase from the citric acid producing fungus Aspergillus niger, Biochimica et Biophysica Acta, 615, 449–457. KUBICEK, C P and ROEHR, M, 1985. Aconitase and citric acid accumulation in Aspergillus niger, Applied and Environmental Microbiology, 50, 1336–1338. KUBICEK, C P and ROEHR, M, 1986. Citric acid fermentation. CRC Critical Reviews in Biotechnology, 3, 331–373. KUBICEK, C P, HAMPEL, W A and ROEHR, M, 1979a. Manganese deficiency leads to elevated amino acid pools in citric acid accumulating Aspergillus niger, Archives of Microbiology, 123, 73–79. KUBICEK, C P, ZEHENTGRUBER, O and ROEHR, M, 1979b, An indirect method for studying fine control of citric acid accumulation by Aspergillus niger, Biotechnology Letters, 1, 47–52. KUBICEK, C P, ZEHENTGRUBER, O, EL-KALAK, H and ROEHR, M, 1980. Regulation of citric acid production by oxygen: effects of dissolved oxygen tension on adenylate levels and respiration in Aspergillus niger, European Journal of Applied Microbiology and Biotechnology, 9, 101–116. KUBICEK, C P, SCHREFERL-KUNAR, G, WÖHRER, W and ROEHR, M, 1988. Evidence for a cytoplasmic pathway of oxalate biosynthesis in Aspergillus niger, Applied and Environmental Microbiology, 54, 633–637. KUBICEK-PRANZ, E M, MOZELT, M, ROEHR, M and KUBICEK, C P, 1990. Changes in the concentration of fructose–2,6-bisphosphate in Aspergillus niger during stimulation of acidogenesis by elevated sucrose concentration, Biochimica et Biophysica Acta, 1033, 250– 255. LA NAUZE, J M, 1966. Aconitase and isocitric acid dehydrogenases in Aspergillus niger in relation to citric acid accumulation, Journal of General Microbiology, 44, 73–81. LEGISA, M and BENCINA, M, 1994, Evidence for the activation of 6-phosphofructo-1-kinase by cAMP-dependent protein kinase in Aspergillus niger, FEMS Microbiology Letters, 118, 327– 334. LEGISA, M and GRADISNIK-GRAPULIN, M, 1995. Sudden substrate dilution induces a higher rate of citric acid production by Aspergillus niger, Applied and Environmental Microbiology, 61, 2732–2737. LEGISA, M and KIDRIC, J, 1989. Initiation of citric acid accumulation in the early stages of Aspergillus niger growth, Applied Microbiology and Biotechnology, 31, 453–457. LEGISA, M and MATTEY, M, 1986, Glycerol as an initiator of citric acid accumulation in Aspergillus niger, Enzyme and Microbial Technology, 8, 607–609. LEGISA, M and MATTEY, M, 1988. Citrate regulation of the change in carbohydrate degradation during the initial phase of citric acid production by Aspergillus niger, Enzyme and Microbial Technology, 10, 33–36. MA, H, KUBICEK, C P and ROEHR, M. 1981, Malate dehydrogenase isoenzymes in Aspergillus niger, FEMS Microbiology Letters, 12, 147–151. MA, H, KUBICEK, C P and ROEHR, M, 1985. Metabolic effects of manganese deficiency in Aspergillus niger: evidence for increased protein degradation, Archives of Microbiology, 141, 266–268.

Biochemistry of citric acid accumulation by A. niger

29

MACHIDA, M, GONZALEZ, T V J, BOON, L K, GOMI, K and JIGAMI, Y, 1996. Molecular cloning of a genomic DNA for enolase from Aspergillus oryzae, Bioscience, Biotechnology and Biochemistry, 60, 161–163. MARTIN, S M and WILSON, P W, 1951. Uptake of 14CO2 by Aspergillus niger in the formation of citric acid, Archives of Biochemistry, 27, 150–157. MATTEY, M, 1977. Citrate regulation of citric acid production by Aspergillus niger. FEMS Microbiology Letters, 2, 71–74. MATTEY, M, 1992. The production of organic acids, CRC Critical Reviews in Biotechnology, 12, 87–132. MATTEY, M, LEGISA, M and LOWE, S, 1988. Effect of lectins and inhibitors on membrane transport in Aspergillus niger, Biochemical Society Transactions, 16, 969–970. MCKNIGHT, G L, O’HARA, P J and PARKER, M L, 1986. Nucleotide sequence of the triosephosphate isomerase gene from Aspergillus nidulans: implications for a differential loss of introns, Cell, 46, 143–147. MCNEIL, J B, DYKSHOORN, P, HUY, J N and SMALL, S, 1990. The DNA-binding protein RAP1 is required for efficient transcriptional activation of the yeast PYK glycolytic gene, Current Genetics, 18, 405–412. MEIXNER, O, MISCHAK, H, KUBICEK, C P and ROEHR, M, 1985. Effects of manganese deficiency on plasma membrane lipid composition and glucose uptake in Aspergillus niger, FEMS Microbiology Letters, 26, 271–274. MEIXNER-MONORI, B, KUBICEK, C P and ROEHR, M, 1984. Pyruvate kinase from Aspergillus niger: a regulatory enzyme in glycolysis? Canadian Journal of Microbiology, 30, 16–22. MEIXNER-MONORI, B, KUBICEK, C P, HABISON, A, KUBICEK-PRANZ, E M and ROEHR, M, 1985. Presence and regulation of the a-ketoglutarate dehydrogenase multienzyme complex in the filamentous fungus Aspergillus niger, Journal of Bacteriology, 161, 265–271. MEIXNER-MONORI, B, KUBICEK, C P, HARRER, W, SCHREFERL, G and ROEHR, M, 1986. NADP-specific isocitrate dehydrogenase from the citric acid accumulating fungus Aspergillus niger, Biochemical Journal, 236, 549–557. MILLIS, N F, TRUMPY, B H and PALMER, B M, 1963. The effect of lipids on citric acid production by an Aspergillus niger mutant, Journal of General Microbiology, 30, 365–379. MISCHAK, H, KUBICEK, C P and ROEHR, M, 1985. Formation and location of glucose oxidase in citric acid producing mycelia of Aspergillus niger, Applied Microbiology and Biotechnology, 21, 27–31. MORSOMME, P, D’EXAERDE, A D, DEMEESTER, S, THINES, D, GOFFEAU, A and BOUTRY, M, 1996. Single point mutations in various domains of a plant plasma membrane H+-ATPase expressed in Saccharomyces cerevisiae increase H+-pumping and permit yeast growth at low pH, EMBO Journal, 15, 5513–5526. MOYER, A J, 1953. Effect of alcohols on the mycological production of citric acid in surface and submerged culture, Applied Microbiology, 1, 1–13. MÜLLER, H M, 1975. Oxalate accumulation from citrate by Aspergillus niger. I. Biosynthesis of oxalate from its ultimate precursor, Archives of Microbiology, 103, 185–189. MÜLLER, H M and FROSCH, S, 1975. Oxalate accumulation from citrate by Aspergillus niger. II. Involvement of the tricarboxylic acid cycle, Archives of Microbiology, 104, 159–162. NETIK, A, TORRES, T V, RIOL, J, and KUBICEK, C P, 1997. Uptake and secretion of citric acid by Aspergillis niger is reciprocally regulated by manganese ions, Biochima et Biophysics Acta. 1326, 287–294. OLSVIK, E S, TUCKER, K G, THOMAS, C R, and KRISTIANSEN, B, 1993. Correlation of Aspergillus niger broth rheological properties with biomass concentration and the shape of mycelial aggregates, Biotechnology and Bioengineering, 42, 1046–1052. ORTHOFER, R, KUBICEK, C P and ROEHR, M, 1979. Lipid levels and manganese deficiency in citric acid producing strains of Aspergillus niger, FEMS Microbiology Letters, 5, 403–406. PANNEMAN, H, RUIJTER, G J, VAN DEN BROECK, H C, DRIEVER, E T M and VISSER, J, 1996. Cloning and biochemical characterisation of an Aspergillus niger glucokinase, European Journal of Biochemistry, 240, 518–525.

30

Citric Acid Biotechnology

PAPAGIANNI, M, MATTEY, M and KRISTIANSEN, B, 1994. Morphology and citric acid production of Aspergillus niger PM1, Biotechnology Letters, 16, 929–934. PFITZNER, A, KUBICEK, C P and ROEHR, M, 1987. Presence and regulation of ATP:citrate lyase from the citric acid producing fungus Aspergillus niger, Archives of Microbiology, 147, 88–91. PRÖMPER, C, SCHNEIDER, R and WEISS, H, 1993. The role of the proton-pumping and alternative respiratory chain NADH:ubiquinone oxidoreductases in overflow catabolism of Aspergillus niger, European Journal of Biochemistry, 216, 223–230. PUNEKAR, N S, VAIDYANATHAN, C S and APPAJI RAO, N, 1985a. Role of Mn2+ and Mg2+ in catalysis and regulation of Aspergillus niger glutamine synthase, Indian Journal of Biochemistry and Biophysics, 22, 142–151. PUNEKAR, N S, VAIDYANATHAN, C S and RAO, N A, 1985b. Role of glutamine synthetase in citric acid fermentation by Aspergillus niger, Journal of Biosciences, 7, 269–287. PUNT, P J, DINGEMANSE, M A, JACOBS-MEIJSING, B J M, POUWELS, P H and VAN DEN HONDEL, C A M J J, 1988. Isolation and characterization of the glyceraldehyde-3-phosphate dehydrogenase gene of Aspergillus nidulans, Gene, 69, 49–57. PUNT, P J, DINGEMANSE, M A, KUYVENKOFEN, A, SOEDE, R D M, POUWELS, P H and VAN DEN HONDEL, C A M J J, 1990. Functional elements in the promoter region of the Aspergillus nidulans gpdA gene encoding glyceraldehyde-3-phosphate dehydrogenase, Gene, 93, 101–109. PUNT, P J, KRAMER, C, KUYVENKOFEN, A, POUWELS, P H and VAN DEN HONDEL, C A M J J, 1992. An upstream activating sequence from the Aspergillus nidulans gpdA gene, Gene, 120, 62–73. RAMAKRISHNAN, C V, STEEL, R and LENTZ, C P (1955). Mechanism of citric acid formationand accumulation in A. niger, Arch, Biochem. Biophys., 55(1): 270–273. ROEHR, M, KUBICEK, C P, 1981. Regulatory aspects of citric acid fermentation by Aspergillus niger, Process Biochemistry, 16, 34–37. ROEHR, M, KUBICEK, C P and KOMINEK, J, 1983. Citric acid. In: Biotechnology, Vol. 3. Eds H J REHM and G REED (Verlag Chemie, Weinheim), pp. 331–373. ROEHR, M, KUBICEK, C P, ZEHENTGRUBER, O and ORTHOFER, R, 1987. Accumulation and partial re-consumption of polyols during citric acid fermentation by Aspergillus niger, Applied Microbiology and Biotechnology, 27, 235–239. ROEHR, M, KOMINEK, J and KUBICEK, C P, 1992. Industrial acids and other small molecules. In Aspergillus: Biology and Industrial Applications. Eds J W Bennett and M A Klich (Butterworth-Heinemann), pp. 91–131. ROEHR, M, KUBICEK, C P and KOMINEK, J, 1996. Citric acid. In: Biotechnology, 2nd edition. Eds H J REHM, G REED, A PÜHLER and P STADLER, Vol. 6, Ed. M ROEHR (Verlag Chemie, Weinheim), pp. 307–345. ROGALSKI, J, FIEDURECK, J, SZCZODRAK, J, KAPUSTA, K and LEONOWICZ, A, 1988. Optimization of glucose oxidase synthesis in submerged cultures of Aspergillus niger G-143 mutant, Enzyme and Microbial Technology, 10, 508–511. ROUKAS, T and HARVEY, L, 1988. The effect of pH on production of citric and gluconic acid from beet molasses using continuous culture, Biotechnology Letters, 10, 289–294. RUBIO, M C and MALDONADO, M C, 1995. Purification and characterization of invertase from Aspergillus niger, Current Microbiology, 31, 80–83. RUIJTER, G J and VISSER, J, 1996. Determination of intermediary metabolites in Aspergillus niger, Journal of Microbiological Methods, 25, 295–302. RUIJTER, G J, PANNEMAN, H, VAN DEN BROECK, H C, BENNETT, J M and VISSER, J, 1996a. Characterisation of the Aspergillus nidulans fra1 mutant: hexose phosphorylation and apparent lack of involvement of hexokinase in glucose repression, FEMS Microbiology Letters, 139, 223–228. RUIJTER, G J, PANNEMAN, H and VISSER, J, 1996b. Metabolic engineering of the glycolytic pathway in Aspergillus niger. In Advances in Citric Acid Technology. Proceedings of an International Conference, Bratislava, Slovakia Oct. 7–9. Eds B Kristiansen et al. (Scarabeus Ltd, Bratislava), pp. 28–30.

Biochemistry of citric acid accumulation by A. niger

31

SCHMIDT, M, WALLRATH, J, DÖRNER, A and WEISS, H, 1992. Disturbed assembly of the respiratory chain NADH:ubiquinone reductase (complex I) in citric acid accumulating Aspergillus niger strain B 60, Applied Microbiology and Biotechnology, 36, 667–672. SCHREFERL, G, KUBICEK, C P and ROEHR, M, 1986. Inhibition of citric acid accumulation by manganese ions in Aspergillus niger mutants with reduced citrate control of phosphofructokinase, Journal of Bacteriology, 165, 1019–1022. SCHREFERL-KUNAR, G, GROTZ, M, ROEHR, M and KUBICEK, C P, 1989. Increased citric acid production by mutants of Aspergillus niger with increased glycolytic capacity, FEMS Microbiology Letters, 59, 297–300. SCHWEIGER, L B, 1959. Method of producing citric acid by fermentation, US Patent 291,6420. SCOTT, E W and BAKER, H V, 1993. Concerted action of the transcriptional activators REB1, RAP1 and GCR1 in the high-level expression of the glycolytic gene tpi, Molecular and Cellular Biology, 13, 543–550. SEEHAUS, C, PILZ, F and AULING, G, 1990. High-affinity manganese transport systems by filamentous fungi, Zentralblatt für Bakteriologie, 272, 357–358. SHEPARD, M W, 1963. Method of producing citric acid by fermentation, US Patent 308, 3144. SHU, P and JOHNSON, M J, 1948a. Citric acid production by submerged fermentation with Aspergillus niger, Industrial and Engineering Chemistry, 40, 1202–1205. SHU, P and JOHNSON, M J, 1948b. The interdependence of medium constituents in citric acid production by submerged fermentation, Journal of Bacteriology, 54, 161–167. SMITH, J E, NOWAKOWSKA-WASZCZUK, A and ANDERSON, J G, 1974. In Industrial Aspects of Biochemistry, Vol 1. Ed. E SPENCER (Elsevier, Amsterdam), pp. 297–317. STEINBÖCK, F, HELD, I, CHOOJUN, S, ROEHR, M and KUBICEK, C P, 1994. Characterization and regulatory properties of a single hexokinase from the citric acid accumulating fungus Aspergillus niger, Biochimica et Biophysica Acta, 1200, 215–223. STRASSER, H, BURGSTALLER, W and SCHINNER, F, 1994. High-yield production of oxalic acid for metal leaching processes by Aspergillus niger, FEMS Microbiology Letters, 119, 365– 370. STREATFIELD, S J and ROBERTS, C F, 1993. Disruption of the 3’phosphoglycerate kinase in Aspergillus nidulans, Current Genetics, 23, 123–128. STREATFIELD, S J, TOEWS, S and ROBERTS, C F, 1992. Functional analysis of the expression of the 3'-phosphoglycerate kinase pgk gene in Aspergillus nidulans, Molecular and General Genetics, 233, 231–241. SZCZODRAK, J and ILCZUK, Z, 1985. Effect of iron on the activity of aconitate hydratase and synthesis of citric acid by Aspergillus niger, Zentralblatt für Mikrobiologie, 140, 567–574. TOMLINSON, N, CAMPBELL, J J R and TRUSSELL, P C, 1950. The influence of zinc, iron, copper and manganese on the production of citric acid by Aspergillus niger. II. Evidence for the essential nature of copper and manganese, Journal of Bacteriology, 61, 17–25. TORRES, N, 1994a. Modelling approach to control of carbohydrate metabolism during citric acid accumulation by Aspergillus niger. I. Model definition and stability of the steady state, Biotechnology and Bioengineering, 44, 104–111. TORRES, N, 1994b. Modelling approach to control of carbohydrate metabolism during citric acid accumulation by Aspergillus niger. II. Sensitivity analysis, Biotechnology and Bioengineering, 44, 112–118. TORRES, N, RIOL-CIMAS, J M, WOLSCHEK, M and KUBICEK, C P, 1996a. Glucose transport by Aspergillus niger: the low affinity carrier is only formed during growth on high glucose concentrations, Applied Microbiology and Biotechnology, 44, 790–794. TORRES, N V, VOIT, E and GONZALEZ-ALCON, C, 1996b. Optimization of nonlinear biotechnological processes with linear programming: application to citric acid production by Aspergillus niger, Biotechnology and Bioengineering, 49, 247–258. TRUMPY, B H and N F MILLIS, 1963. Nutritional requirements of an Aspergillus niger mutant for citric acid production, Journal of General Microbiology, 30, 381–393. UEMURA, H and JIGAMI, Y, 1992. Role of GCR2 in transcriptional regulation of yeast glycolytic genes, Molecular and Cellular Biology, 12, 3834–3842.

32

Citric Acid Biotechnology

WALLRATH, J, SCHMIDT, J and WEISS, H, 1991. Concomitant loss of respiratory chain NADH:ubiquinone reductase (complex I) and citric acid accumulation in Aspergillus niger, Applied Microbiology and Biotechnology, 36, 76–81. WOLD, W S M and SUZUKI, I, 1973. Cyclic AMP and citric acid accumulation by Aspergillus niger, Biochemical and Biophysical Research Communications, 50, 237–244. WOLD, W S M and SUZUKI, I, 1974. Demonstration in Aspergillus niger of adenylate cyclase, a cyclic-AMP binding protein, and intra- and extracellular phosphodiesterases, Canadian Journal of Microbiology, 20, 1567–1576. WOLD, W S M and SUZUKI, I, 1976a. The citric acid fermentation by Aspergillus niger: regulation by zinc of growth and acidogenesis, Canadian Journal of Microbiology, 22, 1083–1092. WOLD, W S M and SUZUKI, I, 1976b. Regulation by zinc and adenosine 3',5'-cyclic monophosphate on growth and citric acid accumulation in Aspergillus niger, Canadian Journal of Microbiology, 22, 1093–1101. WOLSCHEK, M F and KUBICEK, C P, 1997. The filamentous fungus Aspergillus niger contains two ‘differentially regulated’ trehalose-6-phosphate synthase-encoding genes, tpsA and tpsB, Journal of Biological Chemistry, 272, 2729–2735. WONGCHAI, V and JEFFERSON, W E JR, 1974. Pyruvate carboxylase from Aspergillus niger: partial purification and some properties, Federation Proceedings, 33, 1378. WORONICK, C L and JOHNSON, M J, 1960. Carbon dioxide fixation by cell-free extracts of Aspergillus niger, Journal of Biological Chemistry, 235, 9–15. XU, D-B, MADRID, C P, ROEHR, M and KUBICEK, C P, 1989a. Influence of type and concentration of the carbon source on citric acid production by Aspergillus niger, Applied Microbiology and Biotechnology, 30, 553–558. XU, D-B, ROEHR, M and KUBICEK, C P, 1989b. Aspergillus niger cyclic AMP levels are not influenced by manganese deficiency and do not correlate with citric acid accumulation, Applied Microbiology and Biotechnology, 32, 124–128. YIGITOGLU, M, and MCNEIL, B, 1992. Ammonium and citric acid supplementation in batch cultures of Aspergillus niger B60, Biotechnology Letters, 14, 831–836. ZEHENTGRUBER, O, KUBICEK, C P and ROEHR, M, 1980. Alternative respiration of Aspergillus niger, FEMS Microbiology Letters, 8, 71–74.

3

Biochemistry of Citric Acid Production by Yeasts
MICHAEL MATTEY

3.1 Introduction In terms of bulk production citric acid is widely regarded as one of the most important of the organic acids produced by microbiological methods, although reliable estimates of world production are not easily obtained. A widespread perception has been that most of the production is achieved with Aspergillus niger, in what has come to be regarded as the ‘traditional’ fermentation process, although the first indications of a microbiological process for the production of citric acid were from Wehmer (1893), who noted that ‘Citromyces’ (now Penicillium) could accumulate citric acid. Indeed until around 1970 A. niger was almost exclusively the organism used for the production of citric acid; the ability of other filamentous fungi to excrete citric acid was known and has been reviewed (Röhr and Kubicek, 1992), but they are of limited importance. A more important class of production organisms found within the Candida yeasts, and an increasing proportion of the total production of citric acid is now manufactured using strains of Yarrowia lipolytica (the asexual form is Candida lipolytica, syn. Saccaromycopsis). The Candida genus (family Cryptococcaceae, subfamily Cryptococcoideae) contains 30 species, and six varieties, many of which are pathogenic to animals, including humans. With increasing numbers of immunodeficient people, either through retroviral disease or the anti-rejection drugs used in organ transplantation, the pathogenic species such as C. albicans have assumed a new importance. Many Candida species have been isolated from fruit, seeds, soil, and similar sources. Vegetative growth consists of budding cells and pseudomycelium, or true mycelium with blastospores. C. lipolytica was isolated from margarine (hence its name, Gr. Lipos, fat; lysis, breaking). The cells are variable, long oval to almost cylindrical and short oval. A well developed pseudomycelium is frequently formed with some true mycelium. The organism, as well as hydrolysing fats, will liquify gelatine, but does not ferment sugars. As well as the industrial importance of the organism, it is being developed as a cloning vehicle for the expression of heterologous proteins. Some understanding of the pathways involved in citric acid production has resulted from the cloning of particular genes as a result of this development; in particular our knowledge of the peroxisomal pathways has
33

34

Citric Acid Biotechnology

been advanced through developments in the understanding of protein targeting. It is rarely known whether the perfect or imperfect form is being used in a particular process. Indeed, C. lipolytica is a dimorphic yeast and the morphology may vary, particularly during the course of a batch process. The impetus for the study of citric acid production from Candida yeasts appears to have had two aspects: first the availability in the 1960s of relatively cheap hydrocarbons as feedstock; and secondly the use of hydrocarbons for the production of glutamic acid by Corynebacterium (Yamada et al., 1963). Particularly in Japan the further use of this feedstock in a number of fermentation processes was explored. This phase came to a halt in 1973 when the price of crude oil was increased dramatically. This process is reflected in the number of patents granted: from one in 1967 and two in 1968 up to 15 in 1972, then dropping to three by 1975 (source: Chemical Abstracts). Possibly as a result of the extent of industrial commitment to yeast-based processes, alternative carbon sources were explored; the finding that alkane-utilizing yeasts can also use sugars to produce citric acid formed the basis of the present expansion of the industrial importance of C. lipolytica. The potential advantages of using yeasts rather than A. niger are the higher initial sugar concentrations that can be tolerated and the faster conversion rates possible. The sensitivity of A. niger to metal ions, particularly manganese, is well known, and a source of increased costs associated with the pretreatment of molasses. This is avoided with Candida yeast that is far less sensitive to metal ions. The fermentation with Candida yeasts appears to be biphasic (Marchal et al., 1977) with citric acid accumulating after the growth phase, when nitrogen is exhausted. Although nitrogen limitation, with nitrogen usually supplied in the form of ammonium salts, is the trigger for acid accumulation, other parameters influence the yield and productivity of the process. The pH is maintained above pH 5.0, unlike the situation in A. niger where a medium pH below 2 is required for a good yield. Lowering the pH with the C. lipolytica fermentation results in the production of polyols (Tabuchi and Hara, 1970), mainly erythritol and arabitol. The addition of iron salts to the medium lowers the yield of citric acid, although some iron is required for normal growth. The addition of iron increases the activity of aconitate hydratase, and this is thought to result in the conversion of citric to isocitric acid (Tabuchi et al., 1973). The influence of iron on the growth and synthesis of citric and isocitric acid in ethanol-containing media showed similar effects (Kamzolova et al., 1996). Changes in the concentration of iron caused abrupt switching between the predominant formation of either citric or isocitric acids. One noteworthy feature of the process is the requirement for thiamine. Unless this is added, oxoacids, mainly oxoglutarate, accumulate and the yield of citric acid is reduced. The reason for this requirement is not known but is likely to be related to the level of oxidative decarboxylation required, where thiamine pyrophosphate (TPP) is a co-factor. When ß-oxidation is the main assimilatory pathway the requirement for pyruvate dehydrogenase would not be significant, but the flux through oxoglutarate dehydrogenase might be elevated. The other obvious requirement for TPP is for the transketolase reaction; although the role of the pentose phosphate cycle in the metabolism of C. lipolytica during citric acid accumulation is not known, the production of erythritol and arabitol under conditions of low pH might be indicative of its activity. The activities of enzymes of the TCA cycle have been measured after thiamine-limited growth with ethanol as a substrate (Morgunov et al., 1995). This will use essentially the same pathway as growth on alkanes, and thiamine limitation is similarly accompanied by oxogluarate production. The activity of the oxoglutarate dehydrogenase complex is greatly reduced and oxidative

Biochemistry of citric acid production by yeasts

35

decarboxylation of oxoglutarate becomes the limiting reaction in the TCA cycle. This leads to oxoglutarate accumulation within the cells and secretion into the culture medium. The glyoxylate cycle is used as an alternative pathway when the TCA cycle is impaired in this way. Although growth is usually limited by nitrogen exhaustion, limitation of growth by sulphur, magnesium or phosphorus gives a similar effect (McKay et al., 1994). Citric acid levels between 50 and 220 mM were measured after 168 hours, with nitrogen and sulphur limitation giving the highest specific production rates. Potassium limitation was ineffective (6 mM), and the glucose uptake rate was only 50 per cent of that achieved when nitrogen or sulphur was limiting.

3.2 Synthesis of citric acid from n-alkanes 3.2.1 Growth on alkanes An important feature of the growth of yeast on alkanes is that the flow of carbon from the substrate to the cellular materials is significantly different to that found during conventional growth on a carbohydrate, in that during growth on carbohydrates fatty acids are synthesized while carbohydrates are degraded, whilst the opposite is true for growth on alkanes. The metabolic sequence when growth on alkanes occurs is therefore: 1 2 3 4 5 6 Uptake of alkanes into the cell. Oxidation of alkanes into the corresponding fatty acids. Conversion of the fatty acids to acyl CoA esters. Metabolism of fatty acyl CoA esters to acetyl CoA, or incorporation into cellular lipid. Synthesis of TCA cycle intermediates. Gluconeogenesis, synthesis of amino acids, nucleic acids, etc.

Since the utilization of alkanes overlaps considerably with the metabolism of fatty acids the oleaginous yeasts such as C. lipolytica were obvious targets for fermentation processes with this type of feedstock.

3.2.2 Uptake of alkanes Alkanes are of limited solubility in water so that the uptake of alkanes by cells could be of three types: by direct contact between the alkane droplets and the microbial cells; through the soluble phase; or by ‘solubilization’ by micelle formation in an emulsion with subsequent uptake. All three mechanisms are believed to occur. Once contact between a hydrophobic alkane droplet and the hydrophobic cell membrane has been made the alkane will dissolve in the lipid phase and be transported across the membrane. Despite this, particular areas of the membrane may become specialized for the rapid uptake of alkanes. Meissel et al. (1973, 1976) have observed distinctive channels in the cell wall of yeast grown on alkanes when studied by electron microscopy. Similar channels have been observed by Osumi et al. (1975), together with protrusions on the cell surface which reach the cell membrane through electron-dense channels. The hypothesis has been put forward that alkanes attach to these channels and migrate through them to the membrane and into the endoplasmic reticulum which appears to be particularly associated with the

36

Citric Acid Biotechnology

Figure 3.1 Hydroxylation of an alkane by cytochrome P-450 dependent mono-oxygenase

cytoplasmic end of these channels. The endoplasmic reticulum is the site of the initial oxidation of the alkanes. Alkane emulsions adhere to the cell wall of Candida yeast by a non-enzymatic mechanism (Einsele et al., 1975; Käppeli and Fiechter, 1976, 1977). The binding is due to a lipopolysaccharide in the cell wall that is induced by alkanes. The lipopolysaccharide, which is mannan with about 4 per cent covalently linked fatty acid, has been isolated and characterized (Käppeli et al., 1978).

3.2.3 Initial oxidation of n-alkanes Three oxidation mechanisms are known but the one most likely to be operating in Candida yeast is the mixed function oxidase (mono-oxygenase). A cytochrome P-450 hydroxylase system, dependent on NADPH+ and H+, has been described in several species of Candida (Liu and Johnson, 1971; Lebeault, 1971; Duppel et al., 1973), see Figure 3.1. The formation of P450 has been shown to be inducible by long-chain alkanes, alkenes, secondary alcohols and ketones (Gallo et al., 1973) with hexadecane increasing the specific activity by 150-fold relative to cells grown on glucose. As well as the P-450, a microsomal NADPH-cytochrome c-reductase was increased (Gallo et al., 1971). The influence of carbon and nitrogen sources on a number of NAD+- and NADP+-linked dehydrogenases was examined (Hirai et al., 1976a); no significant effects other than on the NADP+-cytochrome c-reductase were seen. The cytochrome P-450 concentration was linearly related to hexadecane uptake rates when cells were cultivated under conditions of oxygen limitation in a chemostat (Gmunder, 1979), leading to the suggestion that cytochrome P-450 is the rate-limiting step in alkane uptake and oxidation. It is unlikely however that flux control is in fact dependent on a single step in a steady state system such as this. Alkane molecules are susceptible to such oxidations at one or both of the terminal methyl groups. A monoterminal oxidation pathway appears to be operating in C. lipolytica. The fatty acids in cell lipids of active, alkane degrading cells of C. lipolytica grown on various alkanes showed a pattern corresponding to the n-alkane chain length. The alkanes are oxidized to the corresponding fatty acids that are incorporated into lipids, either directly, after chain elongation or by ß-oxidation.

Biochemistry of citric acid production by yeasts

37

Table 3.1 Ratio of odd chain fatty acids and C17 acids to total cellular fatty acids in Candida lipolytica cells grown on n-alkanes and glucose (Tanaka et al., 1976)

Table 3.1 shows the relationship for C. lipolytica grown on a variety of substrates. The correlation between odd chain length alkane substrate and odd chain length fatty acids in the cells is clear. The high activity of a mono-oxygenase system, with its oxygen radical mechanism, suggests that protection against damage by free radicals might be important when yeast grows on alkanes. Indeed a considerable increase in copper and zinc superoxide dismutase (SOD) is seen during growth on n-alkanes as compared to glucose (Kujumdzievasavova et al., 1991). A correlation between SOD and catalase was noted and resistance to oxygen free radicals observed as a result of the high levels of copper/zinc SOD, which also protected against deleterious effects of Cu2+ and Zn2+ in the medium.

3.2.4 Oxidation of higher alcohols The product of the microsomal oxidase system is a higher alcohol corresponding to the chain length of the alkane. These alcohols are oxidized to the corresponding fatty acid through the aldehyde. NAD+-linked alcohol dehydrogenase and NAD+-linked aldehyde dehydrogenase, specific to long-chain substrates, carry out these reactions (Lebault et al., 1970a, 1970b). Both enzymes are inducible by alkanes as well as long-chain alcohols or aldehydes. A soluble alcohol oxidase may also be present in some strains of Y. lipolytica (Ilchenko et al., 1994). The enzyme was purified from strain H-222 grown on n-alkanes, and showed maximum activity with carbon chain lengths ranging from 10 to 18 (see Table 3.2). It appeared that several other specific alcohol oxidases might have been present. The

38

Citric Acid Biotechnology

localization of these enzymes within the cells appears generalized. Early reports are confusing, possibly because until 1974 the occurrence of peroxisomes was not known. Osumi et al. (1974) have detected both dehydrogenases in peroxisomes, mitochondria and microsomes.

3.2.5 Peroxisomes in yeast metabolizing n-alkanes When grown on alkanes, specific organelles, less than 1 µm in diameter, infrequently seen in cells grown on carbohydrate substrates, become numerous (Osumi et al., 1974; Teranishi et al., 1974; Mishina et al., 1978). These organelles have been identified as peroxisomes by cytochemical staining for catalase. Indeed their appearance is directly related to the increased catalase activity seen during the metabolism of alkanes. These organelles contain several of the enzyme systems involved in the initial oxidation of alkanes and similar substrates, and transport between the various compartments of substrates and intermediates is a complex area. Two peroxisomal targeting signals are known (PTS1 and PTS2) and it is suggested that PTS receptors, which have been found in several subcellular locations, shuttle between the cytosol and the peroxisomal membrane. The PTS1 protein is highly conserved and the human homologue (PTS1R) has been cloned as a result. Interestingly this is mutated in a group of patients afflicted with a fatal peroxisomal disorder (Subramani, 1996). Protein unfolding is not required for the import of peroxisomal matrix proteins, which is markedly different from other mechanisms for the translocation of proteins. The gene pay5 encodes a peroxisomal integral membrane protein in Y. lipolytica, pay5p, of 380 amino acids (41.7 kDa) (Eitzen et al., 1996) homologous to the mammalian PAF-1 protein which is essential for peroxisome assembly. Pay5p is targeted to mammalian peroxisomes in an interesting example of the evolutionary conservation of targeting mechanisms. In humans, mutation of PAF-1 results in the Zellweger syndrome. Mutants of Y. lipolytica (pay5-1) also show defective peroxisome synthesis.

3.2.6 Activation of fatty acids to CoA esters Two acyl CoA synthetases have been isolated from C. lipolytica (Mishina et al., 1978) with different locations, specificity functions and regulation. Their properties are summarized in Table 3.3. Synthetase I is constitutive while synthetase II is inducible by fatty acids. The enzymes could be distinguished immunochemically (Hosaka et al., 1979). The synthetase I is widely distributed including mitochondria where glycerophophate acyltransferase is also located, while synthetase II is located in the peroxisomal compartment where ß-oxidation occurs. Evidence for ß-oxidation has been obtained from the study of peroxisomes from C. tropicalis (Kawamoto et al., 1978). The stoichiometry of the process demonstrated that the ß-oxidation system was similar to that described for castor bean (Cooper and Beevers, 1969) and liver (Lazarow and de Duve, 1976). Acyl CoA esters are oxidized by acyl CoA oxidase, a FAD-containing enzyme, to enoyl CoA, forming hydrogen peroxide from molecular oxygen. The catalase present in the peroxisome breaks down the hydrogen peroxide. The enoyl CoA is then metabolized to give acetyl CoA with CoA and NAD+ as hydrogen acceptor. Acyl CoA oxidase has been purified from C. lipolytica and C. tropicalis (Shimizu et al., 1979) from which organism it has been crystallized. Its substrate specificity is summarized in Table 3.4.

Biochemistry of citric acid production by yeasts Table 3.3 Comparison of the properties of acyl CoA synthetases for C. lipolytica

39

The peroxisome contains an NAD+-dependent glycerol-3-phosphate dehydrogenase (Kawamoto et al., 1979) which is thought to act as a shuttle hydrogen carrier with the FADdependent glycerol-3-phosphate dehydrogenase present in the mitochondria, regenerating NAD+ and generating energy.

3.2.7 Synthesis of intermediates of the tricarboxylic acid cycle While growing on alkanes it is clear that the substrate is degraded to the level of acetyl CoA, or propionyl CoA in the case of odd chain length acids, and while lipids may be incorporated from the fatty acids all other intermediates must be synthesized from the two-carbon precursor. In general, yeast growing under gluconeogenic conditions utilizes the glyoxylate cycle as an anaplerotic mechanism. The role of this cycle has been demonstrated in Candida yeast grown on alkanes (Hildebrandt and Weide, 1974). The two characteristic enzymes of the glyoxylate cycle, isocitrate lyase and malate synthase, are induced by growth on n-alkanes (Nabeshima et al., 1977). However, while the level of isocitrate lyase is considerably elevated compared to the levels in glucose grown cells, the level of NAD dependent isocitrate dehydrogenase is lower (Hirai et al., 1976a; Tanaka et al., 1977). The distribution of the flux of intermediates between the TCA cycle and the glyoxylate cycle is determined by the relative activities of these two enzymes which therefore suggests that a high level of glyoxylate cycle activity occurs during growth on n-alkanes. Much of the isocitrate lyase is present in the particulate fraction of the cells, and the enzymes of the glyoxylate cycle have been localized to the peroxisomal compartment (Hirai et al., 1976b). However, citrate synthase, aconitase and malate dehydrogenase,

40

Citric Acid Biotechnology

Figure 3.2 Methyl isocitrate cycle in C. lipolytica

the characteristic enzymes of the TCA cycle, are present in the mitochondrial compartment as might be anticipated (Kawamoto et al., 1977). Fatty acid ß-oxidation is not present in the mitochondria of C. lipolytica (Mishina et al., 1978) or C. tropicalis (Kawamoto et al., 1978) but appears confined to the peroxisome. This implies that acetyl CoA required for citrate synthesis must be transported to the mitochondria from the peroxisome, probably by the carnitine acyltransferase system (Kawamoto et al., 1978). Methyl isocitrate cycle The propionyl CoA, derived from odd chain length n-alkanes, is metabolized by a cyclic pathway analogous to the first steps in the TCA cycle (Tabuchi, 1975a, b). This pathway is based on the accumulation of pyruvate and seven carbon tricarboxylic acids in C. lipolytica grown on odd chain length alkanes. Methyl citrate, methylaconitate and methylisocitrate were detected as were the key enzymes described in the methyl citrate cycle. In this pathway propionyl CoA from the odd chain fatty acid ß-oxidation sequence reacts with oxaloacetate

Biochemistry of citric acid production by yeasts

41

via a methylcitrate-condensing enzyme, analogous to citrate synthase (Uchiyama and Tabuchi, 1976). The resulting methylcitrate is isomerized to methylisocitric acid, possibly by aconitase and the methylisocitrate is cleaved in a manner analogous to citrate lyase to give succinate and pyruvate (Tabuchi and Satoh, 1976, 1977). The two key enzymes, methylcitrate synthase and methylisocitrate lyase, clearly differ from citrate synthase and isocitrate lyase. Both the methyl tricarboxylic acid converting enzymes seem to be constitutive, like the TCA cycle enzymes, while the key enzymes of the glyoxylate cycle are inducible. The overall effect of this path is to convert propionate to pyruvate. The main factor in the level of citric acid production from alkanes is the amount of isocitrate lyase (Behrens et al., 1977; Tanaka et al., 1977; Aiba and Matsuoka, 1979; Matsuoka et al., 1980, 1984). The role and control of isocitrate lyase has been examined in C. lipolytica and C. tropicalis, and it has been suggested to be a rate limiting enzyme for the process. The characteristics of the enzyme were determined in crude extracts by Marchal et al. (1977); the main features were a K m value of 0.6 mM (at pH 7.0) with inhibition by phosphoenolpyruvate and succinate. Significantly, citrate at 5 mM was not inhibitory. When grown on glucose the level of isocitrate lyase was only 2 per cent of that found when grown on alkanes, where the level was four times that found when grown on acetate, the classic two-carbon substrate. Induction of this enzyme is clearly greater in the presence of alkanes. The role of isocitrate lyase in the production of citrate from carbohydrate substrates is unlikely to be significant and it might well be that the critical factor in the overproduction is not the details of the regulation but that maximal fluxes are obtained, that is, the absence of effective regulation! A similar situation may occur in A. niger. The control of flux through the TCA cycle versus the glyoxylate cycle is usually thought of in terms of competition between NAD+-dependent isocitrate dehydrogenase and isocitrate lyase, with adenine nucleotide regulation of isocitrate dehydrogenase and repression/de-repression of isocitrate lyase but with no metabolite level control of isocitrate lyase, which is the situation in E. coli (Nimmo, 1984). This situation cannot occur in Y. lipolytica as isocitrate lyase is apparently in the peroxisome and the NAD+-dependent isocitrate dehydrogenase is mitochondrial, with no common pool of isocitrate. The isocitrate lyase from Yarrowia lipolytica has been purified and characterized (Hones et al., 1991). The active form was obtained as a single peak from an ion exchange column, with a specific activity of 7.4 U/mg. The molecular mass was estimated to be between 200 and 210 kDa, and appears to have four subunits of about 50 kDa. The pH optimum was pH 6.0 and a Km of 0.3 mM was estimated. The enzyme was non-competitively inhibited by succinate and oxalacetate. The gene for isocitrate lyase has been cloned by complementation of a mutation (acuA3) in the structural gene of isocitrate lyase of E. coli (Barth and Scheuber, 1993). The open reading frame was 1668 bp long and had no introns in contrast to the genes sequenced from other filamentous fungi. The deduced protein was 555 amino acids with a molecular mass of 62 kDa, which is similar to that observed for the purified monomer. The enzyme has a putative glyoxosomal targeting sequence S– L–K at the carboxy-terminus and contained a partial repeat which is typical for eukaryotic isocitrate lyases, but is absent from the E. coli sequence. Deletion mutants, as expected, were unable to utilize acetate, ethanol, fatty acids or alkanes, but surprisingly the growth on glucose was also reduced. Citrate synthase from several strains of Y. lipolytica which are citrate producers have been isolated and purified to homogeneity (Morgunov et al., 1994). The enzyme was a dimer with a subunit molecular mass of 40 kDa, and exhibited a Km value of 10 and 5 µM

42

Citric Acid Biotechnology

Figure 3.3 Compartmentation in C. lipolytica during growth on n-alkanes

Biochemistry of citric acid production by yeasts Table 3.5 Adenine nucleotide levels (mM) during a batch fermentation

43

with acetyl CoA and oxalacetate respectively. The enzyme activity observed in extracts is greater than that of isocitrate lyase, aconitase or isocitrate dehydrogenase (Behrens et al., 1977; Omar and Rehm, 1980). Candida lipolytica has both an NAD+ and an NADP+dependent isocitrate dehydrogenase. The NADP+-dependent enzyme had Michaelis– Menten type kinetics with respect to isocitrate, and a Km of 80 µM for isocitrate at pH 7.0. There was inhibition by oxalacetate at 5 mM of about 40 per cent. The energy metabolism of C. lipolytica has been examined during growth on n-alkanes; in particular, the concentration of adenine nucleotides during a batch fermentation was measured (Marchal et al., 1977). After 25 hours there was a sharp drop in the concentration of ADP and AMP while the ATP level rose. The total adenine nucleotide levels fell slightly, then recovered. These changes coincided with the exhaustion of nitrogen in the medium and the effective cessation of growth. At this point citric acid excretion began, together with isocitric acid. The proportion of isocitric to citric acid was high, about 40 per cent, although the intracellular citric to isocitric ratio was close to that expected from the aconitase equilibrium at about 90 per cent citric:10 per cent isocitric acid. The adenylate energy charge (Atkinson, 1970) reflected the changes in adenine nucleotides, rising to approach 1. However, the dramatic changes are seen in the ATP:AMP ratio, and the most common allosteric effectors amongst the adenine nucleotides are AMP and ATP, so that the ATP:AMP ratio is a better indicator of regulatory changes. The enzyme that is regarded as a significant target for allosteric regulation during growth on alkanes is mitochondrial NAD+-specific isocitrate dehydrogenase (Marchal et al., 1977). The activity of this enzyme with respect to isocitrate and AMP is sigmoidal, consistent with its structure which has four co-operative binding sites. The enzyme is totally dependent on AMP for activity, with maximal activity shown at 0.1 mM AMP and 50 per cent activity at 0.05 mM. At values below 0.01 mM the enzyme is virtually inactive. Magnesium also behaved as an allosteric activator of the enzyme, apparently with two co-operative sites. Since the substrate for the enzyme is magnesium isocitrate this is perhaps surprising. There was no correlation between the rate of n-alkane uptake and nitrogen

44

Citric Acid Biotechnology

exhaustion or changes in adenine nucleotide levels. This is unexpected since the cessation of growth should greatly reduce the energy demand and hence the substrate uptake. The implication is that the coupling between electron transport and ATP synthesis becomes ‘loose’, or that energy is used in, for example, a transport process. The mitochondrial ATP synthase genes have been studied in Y. lipolytica (Matsuoka et al., 1994) and a 6.6 kilobase region sequenced. This closely resembled the human mitochondrial genome with ATP synthase subunits 8 and 6 being followed by the genes for cytochrome c oxidase subunit 3, NADH-ubiquinone oxidoreductase subunit 4 and ATP synthase subunit 9. All the genes were transcribed from the same strand of DNA into multigenic RNAs starting from a nonanucleotide sequence, 5'-ATA-TAAATA-3', similar to other yeast mitochondrial promoters. In addition to these apparently normal mitochondrial genes there is a cyanide-resistant oxidase (Medentsev and Akimenko, 1994) located in the inner membrane. Its activity is typically blocked by benzohydroxamic acid. This resembles the situation found in A. niger and the circumstances of uncoupled electron transport are also similar. The activity of the NAD+ isocitrate dehydrogenase, which is already low compared to the level found in cells grown on glucose, is almost totally inhibited by the drop in AMP, and the evolution of carbon dioxide mirrors this, being sharply reduced to very low levels when growth ceases. The metabolic production of carbon dioxide from acetate via the TCA cycle during growth is stopped by the inhibition at the level of isocitrate dehydrogenase. Isocitrate lyase was high compared to cells grown on glucose so that the entire carbon flux through the mitochondrial compartment is via the glyoxylate cycle. The activity of citrate synthase in C. lipolytica has not been extensively studied, but it is reported to show limited inhibition by ATP (40 per cent at 5 mM ATP). Since the concentration of ATP reported in the whole cell during the citric acid accumulation phase varied from 0.6 mM at the start to 0.8 mM at the end, it is unlikely to rise much above 5 mM in the vicinity of citrate synthase, even if most of the ATP is in the mitochondrial compartment. Further, the level of acetyl CoA, which will be high during growth on n-alkanes, will reduce the ATP inhibition of citrate synthase. The 3-phosphoglycerate kinase gene has been isolated from a genomic library, by probing with a PCR fragment amplified with primers deduced from two highly conserved regions of various pyruvate kinases (Ledall et al., 1996). It encodes a polypeptide of 417 residues with extensive homology to other kinases. The expression of the gene is higher on gluconeogenic substrates, such as alkanes, than on glycolytic ones. Pyruvate kinase has also been cloned as part of the development of expression/secretion systems for heterologous proteins (Buckholz and Gleeson, 1991; Strick et al., 1992). Genomic clones were selected by their specific hybridization to synthetic oligodeoxyribonucleotide probes based on conserved sequences. The gene predicts a protein that is highly homologous to the corresponding Saccharomyces cerevisiae enzyme and the gene further transforms wild type Y. lipolytica with a twofold increase in pyruvate kinase activity. The gene sequence contained an intron, the first reported in a Y. lipolytica gene.

3.2.8 Transport of citric acid Although the outline of the biochemical pathways for the over-production of citric acid from n-alkanes is clear, two problems remain: the secretion mechanism and the reason for the simultaneous production of isocitric acid. The transport mechanism(s) could involve

Biochemistry of citric acid production by yeasts

45

direct citrate excretion across the plasma membrane by some form of facilitated diffusion, active transport, or vacuolar transport, possibly by accumulation in vacuoles and exocytosis. Studies of the activities of enzymes in acetate mutants of Y. lipolytica (Rymowicz et al., 1993) indicated that the excretion of isocitric and citric acids depended more on the transport system than metabolite levels within the cell. The vacuolar transport idea appeared promising when vacuoles from Y. lipolytica isolated during exponential growth showed the ability to concentrate citric acid through a citrate uniporter. The vacuoles showed high ATPase activity (1000 mU/mg protein at six hours growth, falling to 270 mU/mg after 48 hours), which was not sensitive to orthovanadate, nor was it inhibited by azide or oligomycin. The citrate transport rate was up to 12 nmol/mg protein/min after 12 hours growth, and calcium was also transported (140 nmol/mg/min). The vacuoles generated both a proton gradient and a membrane potential. However during the stationary phase, after nitrogen exhaustion, the transport ability fell to zero for both calcium and citrate. This observation was found to be true regardless of the growth limiting substrate, the carbon source, or whether citrate was released from the cells or not. The conclusion was that the citrate transporting system of the vacuolar membrane was not involved in the citrate release into the medium, and that that process was associated with transport systems in the plasma membrane. The ratio of excreted isocitric to citric acid is higher than would be expected from the thermodynamic equilibrium of aconitase, being as high as 40 per cent in wild-type yeasts, rather than the 7 per cent expected on an equilibrium basis. Nonetheless it is apparent that the strategies used to mitigate this unwanted production of isocitric acid all have a common theme in that they inhibit or delete aconitase. By limiting the formation of isocitrate in the first place the problem is resolved. The strategies reported include: the use of iron-free medium, which resulted in impairment of aconitase activity (Kimura and Nakanishi, 1973); the addition of sodium tetraborate, which may complex with iron to give a similar outcome (Furakawa et al., 1982); and the selection of mutants with low aconitase activities (Benckiser, 1974). These have been selected by their ability to grow on n-alkanes but not citric acid and the low aconitase results in improved citrate to isocitrate ratios and decreased biomass, but the complete absence of aconitase would presumably be lethal. Other strategies are the use of inhibitors such as monofluoroacetate (Benckiser, 1974) which is metabolized to monofluorocitrate and acts as a competitive inhibitor of aconitase (Akiyama et al., 1972, 1973a, 1973b), and 2,4-dinitrophenol, an uncoupler of oxidative phosphorylation; and the addition of alcohols, up to oleyl alcohol (Kimura and Nakanishi, 1973). The first two strategies are of use in selecting mutants but would be undesirable in a commercial fermentation. The relative levels of isocitrate lyase and aconitase in determining the ratio of isocitrate to citrate was underlined by Finogenova et al. (1986) with a study of a series of C. lipolytica mutants. Mutants with a high isocitrate lyase activity and a low aconitase level synthesized citric acid almost exclusively regardless of whether the carbon source was glucose, alkane, ethanol, acetate or glycerol. The mutant low in isocitrate lyase but with a high level of aconitase produced primarily isocitric acid on alkanes, where the ratio of citrate to isocitrate was 1:3.6, while on glucose the ratio was 1.8:1. Wild-type strains with high levels of both enzymes gave intermediate results. In the wild-type strains the ratio could be shifted towards isocitrate synthesis by inhibiting isocitrate lyase with aconitate, the reverse of the industrial strategy. The explanation advanced by Marchal et al. (1980) is still valid. They suggested that the high isocitrate ratio was a result of compartmentation within the cell. Whereas citrate is mainly mitochondrial, isocitrate is in the mitochondrial, the cytoplasmic and the peroxisomal compartments. Isocitrate will be exported from the mitochondrial

46

Citric Acid Biotechnology

compartment to the cytosol and then to the peroxisome where it will be converted to glyoxylate and succinate. The absence of aconitase from the cytoplasmic compartment will result in higher isocitrate levels with lower citrate levels, and it is presumably from the cytoplasm that the acids are exported. It is further possible that citrate and isocitrate have differential transport, but no mutants have been reported to suggest that there is a separate export mechanism for each acid.

3.3 Synthesis of citric acid from glucose 3.3.1 Introduction The production of citric acid from glucose by C. lipolytica was established before the industrial production from alkanes was initiated (Tabuchi et al., 1969; Abe et al., 1970), although an inhibitor of aconitase was required to minimize isocitric acid production (Pfizer, 1972). The productivity on glucose was found to be similar to that on n-alkanes, so that the industrial process, which in some cases had started out to make citric acid from alkanes, was adapted to synthesis from glucose without major problems. Such industrial plants were designed for yeasts, but the feedstock could be altered. To over-produce citrate from glucose, the same obstacles must be overcome as with synthesis from n-alkanes: an undiminished supply of precursors for citrate synthesis in the form of oxaloacetate and acetyl CoA, a reduction in the catabolism of the citrate, an unregulated citrate synthase and a transport mechanism.

3.3.2 Pathway for citrate synthesis from glucose The pathways involved in the synthesis of citrate in Candida yeasts are similar to those of other organisms in basic properties; the over-accumulation is a result of differences in regulation rather than differences in mechanisms. The outline of the biochemical pathways is shown in Figure 3.4. The basic difference between the pathways on n-alkanes and glucose lies in the source of acetyl CoA: in the case of n-alkanes, ß-oxidation from fatty acids; in the case of glucose, by glycolysis. In both cases oxalacetate is synthesized by an anaplerotic route, either the glyoxylate cycle or in the case of glucose, pyruvate carboxylase; the immediate reaction leading to citrate is citrate synthase with both substrates. The differences lie in the direction of the pathways: with n-alkanes as a substrate, gluconeogenesis is required for the synthesis of metabolites derived from the glycolytic sequence; when glucose is the substrate, fatty acids must be synthesized from acetyl CoA. Pyruvate carboxylase was shown to be the source of oxalacetate by Aiba and Matsuoka (1978). Its relative activity in glucose-grown cells is almost ten times that in cells grown on n-alkanes (Finogenova et al., 1986), but it was only 10 per cent of the activity of citrate synthase. The incorporation of carbon dioxide into pyruvate and thence into citrate appears to involve both carbon dioxide from metabolism within the cell and from the medium. In theory, however, the amount of carbon dioxide released in the pyruvate dehydrogenase reaction to yield acetyl CoA should be enough to form the oxaloacetate needed for citrate synthesis. When grown at a medium pH of 4.5, C. lipolytica showed 20 per cent incorporation of exogenous carbon dioxide into one of the carboxyl groups of citrate, but this fell to 8 per cent at pH 6.0 (Zyakun et al., 1992).

Biochemistry of citric acid production by yeasts

47

Figure 3.4 The metabolic relationships of citrate metabolism in yeasts with n-alkanes or glucose as substrate. 1, Alkane monooxygenase; alkane, reduced rubridoxin:oxygen 1oxidoreductase, 1.14.15.3. 2, Alcohol dehydrogenase; alcohol:NAD+ oxidoreductase, 1.1.1.1. 3, Aldehyde dehydrogenase; aldehyde:NAD+ oxidoreductase, 1.2.1.3. 4, ßoxidation. 5, Hexokinase; ATP:D-hexose 6-phosphotransferase, 2.7.1.1. 6, Glycolysis. 7, Pyruvate carboxylase; pyruvate:carbon dioxide ligase (ADP), 6.4.1.1. 8, Pyruvate dehydrogenase; pyruvate:lipoate oxidoreductase (acceptor acylating), 1.2.4.1. 9, Citrate synthase; citrate:oxaloacetate lyase (CoA acylating), 4.1.3.7. 10, Aconitase; citrate (isocitrate) hydrolyase, 4.2.1.3. 11, Isocitrate dehydrogenase; threo-DS-isocitrate:NAD oxidoreductase (decarboxylating), 1.1.1.41(42). 12. Isocitrate lyase; threo-DS-isocitrate:glyoxylate-lyase, 4.1.3.1. 13, Malate synthase; 1-malate glyoxylate-lyase (CoA-acetylating), 4.1.3.2., 14, Gluconeogenesis

48

Citric Acid Biotechnology

Table 3.7 Activities of enzymes of a Candida lipolytica strain grown on glucose

The subcellular location of the various enzymes has been determined (Sokolov et al., 1995a). Pyruvate carboxylase was found in the cytoplasmic compartment in C. lipolytica, and many other yeasts. The NADP+-dependent isocitrate dehydrogenase was found to be distributed in both cytoplasmic and mitochondrial compartments, while ATP-citrate lyase was found in the cytoplasmic compartment. The presence of this latter enzyme would be required for lipid synthesis, but is presumably regulated so that it does not degrade a significant amount of the cytoplasmic citrate. The number of peroxisomes in yeasts grown on glucose is very small. The properties of the NAD+-dependent isocitrate dehydrogenase are thought to be central to citrate accumulation when glucose is used as a substrate as well as n-alkanes. The AMP requirement for activity means that the very low AMP levels found during the stationary phase induced by nitrogen depletion will result in very low activity of the isocitrate dehydrogenase (Bartels and Jensen, 1979). The enzyme was shown to be allosterically regulated by AMP (Sokolov et al., 1995b) although with excess isocitrate the rate became AMP independent. It is also inhibited by ATP which is high during citric acid accumulation. The export of isocitrate from the mitochondial compartment is presumably reduced when glucose is a substrate, as the glyoxylate cycle is non-functional because of the low level of isocitrate lyase and the absence of malate synthase (Finogenova et al., 1986). This may be the reason for the improved ratio of citrate to isocitrate produced when glucose is a substrate. The presence of the NADP+-dependent isocitrate dehydrogenase in both cytoplasm and mitochondria has been noted but its role, if any, is not known. An important factor in the over-production of citric acid is the maintenance of the flux through glycolysis when metabolite levels rise. In particular the inhibition of phosphofructokinase by citrate might be expected to regulate the precursors for citrate production when citrate levels are high. In citrate producing strains of Y. lipolytica the citrate inhibition of phosphofructokinase appears to be weak (Sokolov et al., 1996), while AMP has no effect. Ammonium suppressed the inhibitory effect of citrate and activated the enzyme, a similar mechanism to that suggested for A. niger (Habison et al., 1983). However in Y. lipolytica under conditions of nitrogen exhaustion, when growth has ceased it is less likely that there is a significant pool of intracellular ammonium. The entry of glucose into the cell is normally regulated, and under conditions of citrate accumulation there is indeed a reduction in the glucose uptake rate (Aiba and Matsuoka, 1978), suggesting that the regulation is present to some extent. The regulation of hexokinase has been shown to be sensitive to trehalose-6-phosphate, which occurs in yeasts at about 0.2 mM (Blazques et al., 1993). This is well above the apparent Ki for Y. lipolytica hexokinase and it was concluded that this compound was physiologically significant. There was, however,

Biochemistry of citric acid production by yeasts

49

no activity against glucokinase up to 5 mM so that high levels of glucose might avoid the regulatory step at hexokinase. A related substrate, glycerol, has attracted some attention, and the activities of glycerol kinase and the NAD+ and FAD-dependent glycerol-3-phosphate dehydrogenases, involved in the glycerol phosphate shuttle between cytoplasm and mitochondria, were determined (Morgunov et al., 1991). Glycerol kinase was localized in the cytoplasm but both glycerol phosphate dehydrogenases were associated with the membrane fraction of the cells. The glycerol kinase was purified and found to be inhibited by AMP, but insensitive to fructose1,6-bisphosphate.

3.3.3 Nitrogen metabolism during growth on glucose Yeasts and fungi contain both NAD + and NADP + dependent forms of glutamate dehydrogenase as well as glutamine synthetase. Glutamate dehydrogenase functions both as an anabolic and catabolic enzyme: The NADPH form acts primarily in the direction of glutamate synthesis although it is reversible; the NADH form acts as a catabolic enzyme providing a-oxoglutarate for the citric acid cycle. The activity of the NADPH enzyme is increased under the nitrogen depletion which precedes citric acid excretion in Y. lipolytica (Peskova et al., 1996), while that of glutamine synthetase is decreased as might be expected. Both the NADPH and the NADH glutamate dehydrogenases were located in the cytosolic compartment in Y. lipolytica which is consistent with a role in synthesis of glutamate rather than energy metabolism. Glutamine synthetase was also cytoplasmic. Interestingly the enzymes in the closely related C. maltosa are mitochondrial, and the organism does not produce citric acid. Aspartate aminotransferase was located in the mitochondria in Y. lipolytica. Glutamate dehydrogenases are normally allosterically regulated by inhibition by ATP or GTP and activation by ADP or GDP. It is not known whether this situation occurs in Y. lipolytica, but it would be consistent with the high level of ATP and low ADP seen during the period of nitrogen starvation. The importance of nitrogen levels to citric acid production was demonstrated by Moresi (1994) who determined kinetic constants for a Y. lipolytica strain at different initial glucose concentrations in the medium. Although increasing the glucose concentration from 40 to 108 g/l gave a negative effect on the growth rate, the yield coefficients for glucose and nitrogen were approximately constant. By using a production medium without nitrogen, a citrate lag phase was observed during which the intracellular nitrogen fraction decreased from about 8 per cent to a new low equilibrium value of less than 3 per cent. The idiophase was found to be a non-growth associated process, and the citric acid formation rate was dependent only on the cellular nitrogen concentration. The strain used in this study was capable of equalling the productivity of the best A. niger mutants (about 1.05 g/l/h), but not the selectivity as citric acid was only 85.5 per cent of the acid excreted; the majority of the rest was isocitric acid.

3.3.4 Transport of citric acid during growth on glucose The effect of various inhibitors on the excretion of citric acid has indicated that the export of citric acid is energy requiring. The addition of protein synthesis inhibitors to cultures of

50

Citric Acid Biotechnology

C. lipolytica at the time of nitrogen exhaustion inhibited the production of citric acid (Trutko et al., 1993). At the same time, dinitrophenol (an uncoupler of oxidative phosphorylation), reumycin (respiratory chain shunting agent), or arsenate (which forms ADP-arsenyl instead of ATP) all decreased the yield of citric acid in proportion to the concentration of the agent. There was no significant effect on biomass yield. Since the over-production of citric acid appears to involve some ‘uncoupling’ of the electron transport chain from ATP synthesis, with maximal levels of ATP resulting, the requirement for ATP shown here may be connected with export rather than synthesis, as may be the requirement for protein synthesis. On the other hand, Kulakovskaya et al. (1994) showed that the activity of the plasma membrane ATPase of Y. lipolytica decreased by a factor of ten during the course of nitrogen limited growth with glucose as a carbon source. Citric acid excretion was independent of glucose concentration and resistant to diethylstilboestrol, an inhibitor of the plasma membrane ATPase, for the first 30 minutes of the excretion process. They concluded that the process is independent of energy provision.

3.4 Conclusions The over-production of citric acid by yeasts from both alkane and carbohydrate sources is now well established, both commercially and scientifically. The basic pathways and some of the enzymology are understood, although many details remain to be resolved. With both substrates the overproduction appears to represent a mechanism for recycling reducing equivalents and energy produced by unbalanced growth conditions in the form of the absence of, and subsequent intracellular restriction on, a primary substrate from the growth medium. Further developments in enzymology may arise coincidentally from the use of the organism as a cloning vehicle, but one of the main unresolved problems is the mechanism of excretion, which is central to the problem of high productivity.

3.5 References
ABE, M, TABUCHI, T and TAHARA, Y, 1970. Studies on organic acid fermentation in yeasts: further investigations on production of citric and d-isocitric acid by yeasts. Journal of the Agricultural Chemistry Society of Japan, 44, 499–504. AIBA, S and MATSUOKA, M, 1978. Citrate production from n-alkane by Candida lipolytica in reference to carbon fluxes in vivo. European Journal of Applied Microbiology and Biotechnology, 5, 247–261. AIBA, S and MATSUOKA, A, 1979. Identification of metabolic model: citrate production from glucose by Candida lipolytica, Biotechnology and Bioengineering, 1, 1373–1386. AKIYAMA, S, SUZUKI, T, SUMINO, Y, NAKAO, Y and FUKUDA, H, 1972. Production of citric acid from n-paraffins by fluoroacetate-sensitive mutants of Candida lipolytica. In 4th Proceedings IFS: Fermentation Technology Today, 613. AKIYAMA, S, SUZUKI, T, SUMINO, Y, NAKAO, Y and FUKUDA, H, 1973a. Induction and citric acid productivity of fluoroacetate-sensitive mutant strains of Candida lipolytica, Agricultural and Biological Chemistry, 37, 879–884. AKIYAMA, S, SUZUKI, T, SUMINO, Y, NAKAO, Y and FUKUDA, H, 1973b. Agricultural and Biological Chemistry, 37, 885–888. ATKINSON, D E, 1970. Enzymes as control elements in metabolic regulation. In The Enzymes, Vol. 1. Ed. P D BOYER (Academic Press, London), pp. 461–489. BARTELS, P D and JENSEN, P K, 1979. Role of AMP in regulation of the citric acid cycle in mitochondria from bakers yeast, Biochemica et Biophysica Acta, 582, 246–259.

Biochemistry of citric acid production by yeasts

51

BARTH, G and SCHEUBER, T, 1993. Cloning of the isocitrate lyase gene (ICVL1) from Yarrowia lipolytica and characterisation of the deduced protein, Molecular and General Genetics, 241, 422–430. BEHRENS, U, HIRZEL, K and SCHULZE, E, 1977. Enzymatische Untersuchungen zur Citrat-Isocitrat Akkumulation bei Hefen, Nahrung, 21, 525–529. BENCKISER, J A, 1974. GesmbH, United States Patent 3,843,465. BLAZQUES, M A, LAGUNAS, R, GANCEDO, C and GANCEDO, J M, 1993. Trehalosephosphate, a new regulator of yeast glycolysis that inhibits hexokinases, Federation of European Biochemical Societies Letters, 329, 51–54. BUCKHOLZ, R G and GLEESON, M A G, 1991. Yeast systems for the commercial production of herterologous proteins, Biotechnology, 9, 1067–1072. COOPER, T G and BEEVERS, H, 1969. ß-oxidation in glyoxosomes from castor bean endosperm, Journal of Biological Chemistry, 244, 3514–3520. DUPPEL, W, LEBEAULT, J M and COON M J, 1973. Properties of a yeast cytochrome P-450 containing enzyme system which catalyses hydroxylation of fatty acids, alkanes and drugs. European Journal of Biochemistry, 36, 583–592. EINSELE, A, SCHNEIDER, H and FEICHTER, A, 1975. Journal of Fermentation Technology, 53, 241. EITZEN, G A, TITORENKO, V L, SMITH, J J, VEENHUIS, M, SZILARD, R K and RACHUBINSKI, R A, 1996. The Yarrowia lipolytica gene pay5 encodes a peroxisomal integral membrane protein homologous to the mammalian peroxisomal assembly factor, PAF1. Journal of Biological Chemistry, 271, 20300–20306. FINOGENOVA, T V, SHISHKANOVA, N V, ERMAKOVA, I T and KATAEVA, I A, 1986. Properties of Candida lipolytica mutants with the modified glyoxylate cycle and their ability to produce citric and isocitric acid, Applied Microbiology and Biotechnology, 23, 378–383. FURAKAWA, T, OGINO, T and MATSUYOSHI, T, 1982. Fermentative production of citric acid from n-paraffins by Saccharomyces lipolytica, Journal of Fermentation Technology, 60, 281– 293. GALLO, M, BERTRAND, J C and AZOULAY, E, 1971. Federation of European Biochemical Societies Letters, 19, 45. GALLO, M, BERTRAND, J C and ROCHE, B, 1973. Alkane oxidation in Candida tropicalis, Biochemica et Biophysica Acta, 296, 624–638. GMUNDER, F K, 1979. Die Assimilation von Hexadecan durch Candida tropicalis, Dissertation ETH Zurich. HABISON, A, KUBICEK, C P and RÖHR, M, 1983. Partial purification and properties of phosphofructoskinase from Aspergillus niger, Biochemical Journal, 209, 669–676. HILDEBRANDT, W and WEIDE, H, 1974. Allg. Mikrobiol, 14, 47. HIRAI, M, SHIOTANI, T, TANAKA, A and FUKUI, S, 1976a. Effect of carbon and nitrogen sources on the level of several NADP- and NAD-linked dehydrogenase activities of hydrocarbon utilisable Candida yeasts, Agricultural Biological Chemistry, 40, 1819–1827. HIRAI, M, TAKASHI, S, TANAKA, A and FUKUI, S, 1976b. Intracellular localization of several enzymes in Candida tropicalis grown on different carbon sources, Agricultural Biological Chemistry, 40, 1979–1986. HONES, I, SIMON, M and WEBER, H, 1991. Characterisation of isocitrate lyase from the yeast Yarrowia lipolytica, Journal of Basic Microbiology, 31, 251–258. HOSAKA, K, MISHINA, M, TANAKA, T, KAMIRGO, T and NUMA, S, 1979. Acyl-coenzyme-A synthetase I from Candida lipolytica, European Journal of Biochemistry, 93, 197–204. ILCHENKO, A P, MORGUNOV, I G, HONEK, H, MAUERBERGER, S, VASILKOVA, N N and MULLER, H G, 1994. Purification and properties of alcohol oxidase from the yeast Yarrowia lipolytica, Biochemistry-Moscow, 59, 969–974. KAMZOLOVA, S V, SHISHKANOVA, N V, ILCHENKO, A P, DEDYUKHINA, E G and FINOGENOVA, T V, 1996. Effects of iron ions on biosynthesis of citric and isocitric acids by mutant Yarrowia lipolytica N-1 under conditions of continuous cultivation, Applied Biochemistry and Microbiology, 32, 35–38. KÄPPELI, O and FIECHTER, A, 1976. The mode of interaction between the substrate and cell surface of the hydrocarbon-utilzing yease Candida tropicalis, Biotechnology and Bioengineering, 18, 967–974.

52

Citric Acid Biotechnology

KÄPPELI, O and FIECHTER, A, 1977. Component of the cell surface of the hydrocarbon utilising yeast Candida tropicalis with possible relationship to hydrocarbon transport, Journal of Bacteriology, 131, 917–921. KÄPPELI, O, MULLER, M and FIECHTER, A, 1978. Chemical and structural alterations at the cell surface of Candida tropicalis induced by hydrocarbon substrates, Journal of Bacteriology, 133, 952–958. KAWAMOTO, S, UEDA, M, NOZAKI, C, YAMAMURA, M, TANAKA, A and FUKUI, S, 1977. Localization of carnitine acyl transferase in peroxisomes and in mitochondria of n-alkane grown Candida tropicalis, Federation of European Biochemical Societies Letters, 96, 37–40. KAWAMOTO, S, NOZAKI, C, TANAKA, A and FUKUI, S, 1978. Fatty acid beta-oxidation system in microbodies of n-alkane grown Candida tropicalis, European Journal of Biochemistry, 83, 609–613. KAWAMOTO, S, YAMADA, T, TANAKA, A and FUKUI, S, 1979. Distinct subcellular localization of NAD-linked and FAD-linked glycerol-3-phosphate dehydrogenases in N-alkane-grown Candida tropicalis, Federation of European Biochemical Societies Letters, 97, 253–256. KIMURA, K and NAKANISHI, T, 1973. British Patent 1,332,180. KUJUMDZIEVASAVOVA, A V, SAVOV, V A and GEORGIEVA, E I, 1991. Role of superoxide dismutase in the oxidation of n-alkanes by yeasts, Free Radical Biology and Medicine, 11, 263– 268. KULAKOVSKAYA, T V, MATYASHOVA, R N, PETROV, V V and KURANOVA, E V, 1994. ATPase of the plasma membrane of the yeast Yarrowia lipolytica is not involved in citrate excretion, Microbiology, 63, 12–15. LAZAROW, P B and DE DUVE, C, 1976. A fatty acid acyl Co-A oxidizing system in rat liver peroxisomes enhancement by clofibrate, a hypolipidemic drug, Proceedings of the National Academy of Science of the USA, 73, 2043–2046. LEBAULT, J M and AZOULAY, E, 1971. Metabolism of alkane by yeast, Lipids, 6, 444–447. LEBAULT, J M, ROCHE, B and DUVNJAK, Z, 1970a. Alcool-et aldehyde deshydrogenasesparticulaires de Candida tropicalis cultivé sur hydrocarbures, Biochemistry Biophysics Acta, 220, 373–385. LEBAULT, J M, MEYER, F and ROCHE, B, 1970b. Oxidation des alcools supérieurs chez C. tropicalis cultivé hydrocarbures, Biochemistry Biophysics Acta, 220, 386–395. LEBAULT, J M, LODE, E T and COON, M J, 1971. Fatty acid and hydrocarbon hydroxylation in yeast: role of cytochrome P-450 containing enzyme system in Candida tropicalis, Biochemistry Biophysics Research Communications, 42, 413–419. LEDALL, M T, NICAUD, J M, TRETON, B Y and GAILLARDIN, C M, 1996. The 3-phosphoglycerate gene of the yeast Yarrowia lipolytica de-represses on gluconeogenic substrates, Current Genetics, 29, 446–456. LIU, C M and JOHNSON, M J, 1971. Alkane oxidation by a participate preparation of Candida, Journal of Bacteriology, 106, 830–834. MARCHAL, R, VANDECASTEELE, J-P and METCHE, M, 1977. Regulation of the central metabolism in relation to citric acid production in Saccharomycopsis lipolytica, Archives of Microbiology, 113, 99–104. MARCHAL, R, METCHE, M and VANDECASTEELE, J-P, 1980. Intracellular concentration of citric acid and isocitric acid in cultures of the citric acid excreting yeast Saccharomycopsis lipolytica grown on alkanes, Journal of General Microbiology, 116, 535–538. MATSUOKA, M, UEDA, Y and AIBA, S, 1980. Role and control of isocitrate lyase from Candida lipolytica, Journal of Bacteriology, 144, 692–697. MATSUOKA, M, HIMENO, T and AIBA, S, 1984. Characterisation of Saccharomyces lipolytica mutants that express temperature sensitive synthesis of isocitrate lyase, Journal of Bacteriology, 157, 899–908. MATSUOKA, M, MATSUBARA, M, INQUE, J, KAKEHI, M and IMANAKA, T, 1994. Organisation and transcription of the mitochondrial ATP synthase genes in the yeast Yarrowia lipolytica, Current Genetics, 26, 382–389. MCKAY, I A, MADDOX, I S and BROOKS, J D, 1994. High specific rates of glucose utilisation under conditions of restricted growth are required for citric acid accumulation by Yarrowialipolytica IMK-2, Applied Microbiology and Biotechnology, 41, 73–78. MEDENTSEV, A G and AKIMENKO, V K, 1994. Localisation of cyanide-resistant oxidase in mitochondria of the yeast Yarrowia lipolytica, Microbiology, 63, 233–236.

Biochemistry of citric acid production by yeasts

53

MEISSEL, M N, MEDVEDEVA, G A and KOZLOVA, T M, 1973. Proceedings of the Third International Specialist Symposium on Yeasts, Otaniemi, Helsinki. MEISSEL, M N, MEDVEDEVA, G A and KOZLOVA, T M, 1976. Mikrobiologiya, 45, 844. MISHINA, M, KAMIRYO, T, HAGIHARA, T, TANAKA, A, FUKUI, S, OSUMI, M and NUMA, S, 1978. Subcellular location of two long chain acyl-CoA synthetases in Candida lipolytica, European Journal of Biochemistry, 89, 321–328. MISHINA, M, KAMIRYO, T, TASHIRO, S and NUMA, S, 1979. Separation and characterisation of two long chain acyl CoA synthetases from Candida lipolytica, European Journal of Biochemistry, 82, 347–354. MORESI, M, 1994. Effect of glucose concentration on citric acid production by Yarrowia lipolytica, Journal of Chemical Technology and Biotechnology, 60, 387–395. MORGUNOV, I G, ILCHENKO, A P and SHARYSHEV, A A, 1991. The enzymes of glycerol metabolism in the yeast Yarrowia (Candida) lipolytica, Biochemistry—Russia, 56, 146– 153. MORGUNOV, I G, SHARYSHEVA, A A, MIKULINSKAYA, O V, SOKOLV, D M and FINOGENOVA, T V, 1994. Isolation, purification and properties of citrate synthase from a citrate producing strain of the yeast Yarrowia (Candida) lipolytica, Biochemistry—Moscow, 59, 975– 981. MORGUNOV, I G, CHERNYAVSKAYA, O G and FINOGENOVA, T V, 1995. Mechanism of 2oxoglutararic acid biosynthesis from ethanol by the thiamine-auxotrophic yeast Yarrowia lipolytica, Microbiology, 64, 372–374. NABESHIMA, S, TANAKA, A and FUKUI, S, 1977. Effects of carbon sources on the levels of glyoxylate enzymes in n-alkane utilizable yeasts, Agricultural and Biological Chemistry, 41, 275–285. NIMMO, H G, 1984. Control of Escherichia coli isocitrate dehydrogenase: an example of protein phosphorylation in a prokaryote, Trends in Biochemical Sciences, 9, 475–478. OMAR, S H and REHM, H J, 1980. European Journal of Applied Microbiology and Biotechnology, 11, 42. OSUMI, M, MIWA, N, TERANISHI, Y, TANAKA, A and FULUI, S, 1974. Ultrastructure of Candida yeast grown on n-alkanes: appearance of microbodies and its relationship to high catalase activity, Archives of Microbiology, 99, 181–201. OSUMI, M, FUSAKO, F, TERANISHI, Y, TANAKA, A and FUKUI, S, 1975. Development of microbodies in Candida tropicalis during incubation in a n-alkane medium, Archives of Microbiology, 103, 1–11. PESKOVA, E B, SHARYSHEV, A A and FIBGENOVA, T V, 1996. Intracellular organization of nitrogen metabolism in the yeast Yarrowia lipolytica, Applied Biochemistry and Microbiology, 32, 383–387. PFIZER INC., 1972. British Patent 1,283,786. RÖHR, M and KUBICEK, C P, 1992. Citric acid. In: Biotechnology, Vol. 3. Eds H REHM and G REED (Verlag Chemie, Mannheim), pp. 444–454. RYMOWICZ, W, WOJTATOWICZ, M, ROBAK, M and JURGIELEWICZ, W, 1993. The use of immobilized Yarrowia lipolytica cells in calcium alginate for citric acid production, Acta Microbiologica Polonica, 42, 163–170. SHIMIZU, S, YASUI, K, TANI, Y and YAMADA, H, 1979. Acyl-co-A oxidase from Candida tropicalis, Biochemical Biophysical Research Communications, 91, 108–113. SOKOLOV, D M, SHARYSHEV, A A and FINOGENOVA, T V, 1995a. Subcellular location of enzymes mediating glucose metabolism in various groups of yeasts, Biochemistry—Moscow, 60, 1325– 1331. SOKOLOV, D M, SOLODOVNIKOVA, N Y, SHARYSHEV, A A and FINOGENOVA, T V, 1995b. The role of NAD-isocitrate dehydrogenase in the biosynthesis of citric acid by yeasts. Applied Biochemistry and Microbiology, 31, 450–454. SOKOLOV, D M, SOLODOVNIKOVA, N Y, SHARYSHEV, A A and FINOGENOVA, T V, 1996. The role of phosphofructokinase in the regulation of citric acid biosynthesis by the yeast Yarrowia lipolytica, Applied Biochemistry and Microbiology, 32, 286–290. STRICK, C A, JAMES, L C, O’DONNELL, M M, GOLLAHER, M G and FRANKE, A E, 1992. The isolation and characterisation of the pyruvate kinase encoding gene from the yeast Yarrowia lipolytica, Gene, 118, 65–72. SUBRAMANI, S, 1996. Convergence of model systems for peroxisome biogenesis, Current Opinion in Cell Biology, 8, 513–518.

54

Citric Acid Biotechnology

TABUCHI, T and HARA, S, 1970. Conversion of citrate fermentation to polyol fermentation in Candida lipolytica, Journal of Agricultural Chemical Society of Japan, 47, 485–489. TABUCHI, T and SATOH, T, 1976. Distinction between isocitrate lyase and methylisocitrate lyase in Candida lipolytica, Agricultural Biological Chemistry, 40, 1863–1869. TABUCHI, T and SATOH, T, 1977. Purification and properties of methylisocitrate lyases, a key enzyme in propionate metabolism in Candida lipolytica, Agricultural Biological Chemistry, 41, 169– 174. TABUCHI, T and SERIZAWA, N, 1975a. The production of 2-methylisocitrate from odd carbon Nalkanes by a mutant of Candida lipolytica, Agricultural Biological Chemistry, 39, 1055–1062. TABUCHI, T and UCHIYAMA, H, 1975b. Methylisocitrate condensing and methylisocitrate cleaving enzymes: evidence for the pathway of oxidation of propionyl CoA to pyruvate, Agricultural Biological Chemistry, 39, 1049–1054. TABUCHI, T, TANAKA, M and ABE, M, 1969. Journal of Agricultural Chemical Society of Japan, 43, 154. TABUCHI, T, TAHARA, Y, TANAKA, M and YANAGIUCHI, S, 1973. Journal of Agricultural Chemical Society of Japan, 47, 617. TANAKA, A, HAGIHARA T, NISHIKAWA, Y, MISHINA, M and FUKUI, S, 1976. Effect of the anti-lipogenic antibiotic cerulenin on growth and fatty acid composition of the n-alkane utilizing Candida lipolytica, European Journal of Applied Microbiology, 3, 115–124. TANAKA, A, NABESHIMA, S, TOKDUA, M and FUKUI, S, 1977. Partial purification of isocitrate lyase from Candida tropicalis and some kinetic properties of the enzyme, Agricultural Biological Chemistry, 41, 795. TERANISHI, Y, KAWAMOTO, S, TANAKA, A, OSUMI, M and FUKUI, S, 1974. Agricultural and Biological Chemistry, 38, 1213–1225. TRUTKO, S M, MATYASHOVA, R N and AKIMENKO, V K, 1993. The effect of cell deenergization and malate addition on over-synthesis of citric acid by the yeast Candida lipolytica, Microbiology, 62, 603–606. UCHIYAMA, H and TABUCHI, T, 1976. Agricultural and Biological Chemistry, 40, 1411–1418. WEHMER, C, 1893. Note sur la fermentation citrique, Bullitin Societe Chemie Francaise, 9, 728. YAMADA, K, TAKAHASHI, J and VKOBAYASHI, K, 1963. Agricultural and Biological Chemistry, 27, 773. ZYAKUN, A M, MUSLAEVA, I N, ASHIN, V V, PESHENKO, V P, ADANIIN, V M, MASHKINA, L P, MATYASHOVA, R N and FINOGENOVA, T V, 1992. Heterotrophic fixation of carbon dioxide by Candida lipolytica and its role in citric acid biosynthesis. Microbiology, 61, 390–397.

4

Strain Improvement
GEORGE J.G.RUIJTER AND JAAP VISSER

4.1 Introduction Many factors need to be considered by citric acid producers to obtain the economically most favourable process. Strain breeding is one of these factors. In this chapter we will summarize ways to improve citric acid production genetically. Commercial production of citric acid is performed mainly with Aspergillus niger and to some extent with Candida (or Yarrowia) lipolytica. As the existing fermentation processes usually give high yields, the main objective of strain breeding nowadays is shortening of fermentation time. However, other factors may also be relevant for strain improvement. For example, accumulation of a high concentration of citric acid by A. niger results from quite extreme culture conditions and strain breeding may decrease the sensitivity of the process to these conditions. The number of reports considering strain improvement that have appeared in literature is limited. Röhr et al. (1983), Kubicek and Röhr (1986) and Mattey (1992) have reviewed much of the older work. However, some research and screening activities are ‘hidden’, i.e. performed by industry and not published for obvious reasons. We have structured this chapter more or less on the basis of the methodology used for strain improvement of A. niger: 1 mutagenesis and selection; 2 the use of the parasexual cycle; and 3 genetic engineering. As strain breeding involves fungal genetics, some aspects of the genetic methodology used have been included.

4.2 General aspects of strain improvement The initial A. niger production strains were isolated from their natural habitat, soil. Better strains have been derived from these isolates by various procedures. Basically, two methods can be distinguished for strain selection. In the first method, acid production is tested for
55

56

Citric Acid Biotechnology

individual colonies obtained from single spores (single spore method). Such a method requires automated screening procedures that enable testing of thousands of colonies and is therefore usually done by plate tests. A pH indicator is commonly included in the medium to estimate acid production, but since a pH indicator does not distinguish between citric acid and other acids, improved methods have been developed, e.g. using p-dimethylaminobenzaldehyde, which specifically measures citric acid (Röhr et al., 1979). Yields are evaluated by determining the ratio between acid zone and colony diameter. Obviously, statistical analysis of screening results is quite important to evaluate the significance of a difference in acid production between the parental strain and strains derived from it. Liquid cultures, such as shake flasks, are not suitable in the initial stage of screening, but can be used in later steps for a limited number of selected strains. An alternative to screening by plates would be the use of ‘high throughput’ screening procedures, making use of microtitre plate technology. This has an even higher capacity than plates as it can be automated to a large extent. Nowadays, microtitre plate technology is commonly used by industry in all kinds of screening processes, but it is not clear whether citric acid producers also employ it. The second method comprises selection of mutants with a specific trait from a large population using a suitable discriminative growth condition (passage method). Selection may be on the basis of resistance against an antimetabolite (Kirimura et al., 1992) or failure to grow on a particular carbon source (Akiyama et al., 1973a). Mutants can arise spontaneously or be produced by mutagenic treatment. A variety of methods are used for mutagenesis including exposure to chemicals, UV light, C- and X-ray radiation (see e.g. Begum et al., 1990; Hamissa et al., 1991; Avchieva and Vinarov, 1993; Gupta and Sharma, 1995). A serious drawback of mutagenic treatment is that high doses increase the chances of obtaining more than one mutation per genome at a time. Thus, in addition to a mutation that results in improved citric acid production, an isolate may have other mutations that might, for example, result in (slightly) reduced viability. To minimize the chances to introduce such unwanted mutations, mutagenic treatment should be performed in such a way that a high percentage of survival is obtained. When a better producing mutant is isolated it should be maintained in a proper way to prevent decay, i.e. lose its particular characteristics favourable for citric acid production. Decay is most pronounced during the vegetative stage and therefore storage of spores is the best way to preserve a strain. The optimal storage method depends on the organism. A. niger conidiospores are usually stored on silica beads at 4°C or suspended in a 20 to 30 per cent glycerol solution and frozen. Apart from natural variation certain mutations may be particularly unstable, i.e. losing such mutation may be advantageous for the fungus. For example a certain mutation may result in improved citric acid production, but concomitantly cause reduced vitality. This necessitates careful preservation of original strains and possibly frequent re-isolation. The biochemistry of citric acid biosynthesis has been reviewed before (Kubicek and Röhr, 1986; Mattey, 1992) and will not be treated at length here. Some aspects will however be discussed in order to understand the rationale behind some strategies. Biosynthesis of citric acid from hexoses is depicted in Figure 4.1. Following uptake, hexoses are degraded mainly via glycolysis yielding pyruvate. Part of the pyruvate is converted to acetyl CoA, part to oxaloacetate. Finally, these two compounds are condensed to citric acid, which is secreted and accumulated in the medium. Only in a few cases is the genetic basis or biochemical mechanism of the improved performance by a mutant known. Schreferl-Kunar et al. (1989) isolated several mutants that grew better than the parent

Strain improvement

57

Figure 4.1 Schematic representation of biosynthesis of organic acids and polyols with A. niger. The following steps are depicted: 1, glucose oxidase; 2, lactonase; 3, glucose transport; 4, hexokinase or glucokinase; 5, phosphoglucose isomerase; 6, fructose transport; 7, hexokinase; 8, mannitol–1-phosphate dehydrogenase; 9, mannitol-1phosphate phosphatase; 10, mannitol transport; 11, phosphofructokinase; 12, aldolase; 13, triosephosphate isomerase; 14, glyceraldehyde–3-phosphate dehydrogenase; 15, phosphoglycerate kinase; 16, phosphoglycerate mutase; 17, enolase; 18, pyruvate kinase; 19, pyruvate dehydrogenase; 20, pyruvate carboxylase; 21, oxaloacetate hydrolase; 22, oxalate transport; 23, malate dehydrogenase; 24, citrate synthase; 25, tricarboxylate carrier; 26, citrate transport. Dashed arrows are used for multiple steps in biosynthesis of erythritol and glycerol. PPP, pentose phosphate pathway

58

Citric Acid Biotechnology

on 14 per cent sucrose. The rationale of this selection procedure is the notion that a high rate of citric acid production requires the ability for fast sugar metabolism. Four mutants consumed sucrose faster and gave higher citric acid yields than the parental strain. Unfortunately, it is not clear whether the productivity or just the final yield is improved in these mutants, although faster sucrose consumption suggests increased productivity. Interestingly, all four mutants had about twofold higher activity of the glycolytic enzymes hexokinase and phosphofructokinase, suggesting that the activity of these two enzymes is important in controlling the rate of sugar consumption. In the following sections a few specific objectives for strain improvement will be discussed.

4.2.1 Improved yield on alternative substrates In most liquid fermentation processes glucose, obtained from hydrolysed starch, or sucrose, in beet or cane molasses, are used as substrates. The semi-solid ‘Koji’ process uses agricultural raw materials containing polysaccharides, such as starch and cellulose. Polysaccharides give low productivities in submerged fermentation processes (Begum et al., 1990), supposedly because their rate of hydrolysis to sugars is too slow. However, since polysaccharides are less expensive than glucose syrups or molasses, there have been some attempts to improve strains aiming at direct use, i.e. without prior hydrolysis, of polysaccharides (e.g. starch) in liquid fermentation (Rugsaseel et al., 1993; Suzuki et al., 1996). A mutant originally isolated as being 2-deoxyglucose resistant during growth on cellobiose (Sarangbin et al., 1993) showed enhanced citric acid production from soluble starch (Suzuki et al., 1996). The mutant had increased glucoamylase activity during citric acid fermentation on starch and the most probable explanation for these observations is decreased repression by glucose of both ß-glucosidase, the enzyme catalysing hydrolysis of cellobiose to glucose, and glucoamylase, one of the enzymes catalysing hydrolysis of starch. Although some mutants give improved yields on starch, these yields are still low compared to those obtained on glucose and sucrose.

4.2.2 Decreased formation of by-products During citric acid fermentation, conversion of substrate into compounds other than citric acid is undesirable for two major reasons. Firstly, by-products reduce the final yield and secondly they complicate recovery of citric acid from the broth. In addition to citric acid, A. niger readily accumulates other acids, mainly gluconic acid and oxalic acid, and polyol compounds, e.g. mannitol, erythritol and glycerol. Polyol compounds are formed from sugars, but will be reconsumed once the sugar substrate is exhausted. Therefore, polyol compounds are probably not a major problem for the final yield of citric acid as long as the sugar substrate is completely consumed. The subsequent production and reconsumption of polyols may, however, reduce the rate of citric acid production. To our knowledge, no data are available on strains with reduced polyol production, but this may be related to the functions of polyols in fungal physiology. It has been shown that conidiospores have a high mannitol content, probably functioning as carbon storage, whereas glycerol is the major solute in osmotic adjustment of the mycelium (Witteveen and Visser, 1995). Thus, polyol production is probably vital and not liable to alterations. Gluconic acid is formed from glucose with concomitant reduction of oxygen to hydrogen peroxide by the enzyme glucose oxidase. Glucose oxidase is an extracellular enzyme,

Strain improvement

59

localized mainly in the cell wall (Witteveen et al., 1992) and induced by hydrogen peroxide and high glucose concentration (Witteveen et al., 1993). The enzyme is not stable at pH values below 2 to 3 and hence not induced since no H2O2 is formed. Oxalic acid is produced by oxaloacetate hydrolase, which is a cytoplasmic enzyme (Kubicek et al., 1988). Biosynthesis of oxaloacetate hydrolase is also regulated by external pH, but in this case the mechanism is unclear. Induction of the enzyme is optimal at pH 5 to 6, whereas a very low oxaloacetate hydrolase activity is observed at pH 2 (Kubicek et al., 1988). In pure sugar fermentations, production of gluconic and oxalic acid can thus be kept to a minimum by starting the fermentation at a relatively low pH. In fermentations using molasses as a substrate, an initial pH of 5 to 6 is commonly employed, because conidiospores will not germinate at lower pH values. Therefore, in processes using molasses, production of gluconic and oxalic acid may be a problem favouring production strains lacking glucose oxidase and oxaloacetate hydrolase. In our laboratory a number of gox mutants have been isolated. One of the mutations, goxC, results in the absence of glucose oxidase activity and strains carrying goxC do not produce gluconic acid from glucose (Witteveen et al., 1990). Interestingly, a goxC mutant produces more oxalic acid from glucose than wild-type A. niger (Van de Merbel et al., 1994). The major problem with production of citric acid by C. lipolytica is the simultaneous production of considerable amounts of isocitric acid. Wild-type C lipolytica strains produce approximately equimolar amounts of citric acid and isocitric acid from n-alkanes, whereas less isocitric acid is produced from sugar substrates (Finogenova et al., 1986). Akiyama et al. (1973a) reasoned that a low activity of aconitase, the enzyme catalyzing the conversion of citric acid to isocitric acid, was essential to reduce production of isocitric acid. They selected a mutant that was more sensitive to fluoroacetate than the wild-type strain. This mutant had approximately 1 per cent of the wild-type aconitase activity and produced virtually no isocitric acid (Akiyama et al., 1973a, 1973b).

4.2.3 A. niger mutants with a decreased sensitivity towards manganese It is commonly known that the manganese concentration should be extremely low during citric acid fermentation with A. niger. Any addition of manganese results in a lower yield (Kubicek and Röhr, 1986; Mattey, 1992). Manganese deficiency has multiple effects on physiology, e.g. increased protein turnover and altered cell wall composition, which probably means that the manganese effect is not clearly related to a particular cellular function. In pure sugar fermentations, manganese is usually removed by cation exchangers, whereas in molasses, manganese is precipitated with ferrocyanide. Obviously, mutants with a higher manganese tolerance would be advantageous, as this would make removal of manganese less critical. Gupta and Sharma (1995) reported an A. niger mutant which was more tolerant to manganese; it seems however that their parental strain is already quite tolerant as addition of 0.5 ppm manganese does not decrease the yield, while usually a level below 1 ppb is recommended (Mattey, 1992). Nevertheless, in the presence of 1.5 ppm manganese, citric acid production by the mutant was threefold higher than obtained with the parental strain and similar to production in the absence of manganese. One of the effects of manganese deficiency is a relatively high intracellular NH level, which presumably is due to increased protein turnover (Kubicek et al., 1979). This high NH concentration partially counteracts inhibition of the glycolytic enzyme, phosphofructokinase, by citrate (Arts et al., 1987). A mutant isolated by Schreferl et al. (1986) contained a

60

Citric Acid Biotechnology

phosphofructokinase that was less sensitive to citrate than the one in the parental strain; this mutant accumulated approximately threefold more citric acid compared to the parent on a medium containing 20 mM manganese. However, the citric acid yield of the mutant in the presence of manganese was only half that obtained with the parental strain on manganese deficient medium, indicating that the effects of manganese cannot be attributed to phosphofructokinase alone.

4.2.4 Morphology of A. niger Characteristic for citric acid fermentation with A. niger is a rather abnormal morphology, which has been attributed to manganese deficiency, although other process conditions, such as pH, impeller speed and seeding level also affect morphology. Hyphae are abnormally short and stubby and the mycelium shows excessive branching. The aggregation of mycelium into compact pellets is also reported to be important, but this may vary between strains and with process conditions. An important benefit of such compact pellets is better rheology of the broth. A lower viscosity of the broth makes it easier to mix, requiring a lower power input for mixing and resulting in a higher dissolved oxygen tension. Efficient aeration is quite important as productivity decreases at lower dissolved oxygen tension and interruption of the oxygen supply even results in cessation of citric acid formation. For processes operating with a filamentous mycelium, mutants with altered morphology, i.e. more branching, resulting in more compact aggregates, might be beneficial. Such mutants were easily obtained in the case of Fusarium graminearum (Wiebe et al., 1989), but we are not aware of such mutants in A. niger.

4.3 Isolation of recombinant strains using the parasexual cycle in A. niger Crossing these strains might combine beneficial characteristics of different strains. A. niger does not have a sexual cycle and crossings therefore involve the so called ‘parasexual cycle’, which is not a life cycle, but a series of independent steps, i.e. fusion of hyphae resulting in heterokaryon formation, fusion of the nuclei of the different parents to form a diploid, mitotic recombination and finally haploidization of the diploid strain to yield haploid strains again (Pontecorvo et al., 1953). If crossing of strains is impossible due to heterokaryon incompatibility, fusion of protoplasts can be used to obtain heterokaryons. Protoplasts can be prepared by treatment of mycelium with cell wall lysing enzymes in an osmotically stabilized medium. Usami and coworkers (Kirimura et al., 1988a, 1988b; Sarangbin et al., 1994) have investigated the application of A. niger diploids and haploid recombinants in citric acid fermentation. They have fused protoplasts of a strain optimized for submerged fermentation and a strain optimal for semi-solid fermentation. Some of the resulting diploid strains and haploid recombinants were better producers than both parents (Kirimura et al., 1988a, 1988b), but most were without significant improvement. The reason for higher production by diploids or haploid recombinants may be combination of beneficial mutations or complementation of adverse mutations introduced in the parents during previous mutagenic treatment. In the case of diploids the presence of two copies of the genome might result in overproduction of certain enzymes.

Strain improvement

61

4.4 Genetic engineering Strain improvement by the techniques described in the previous sections is largely a trial and error process involving laborious screening procedures. Moreover, in many cases the improved performance is ‘magic’, as the underlying mechanism is not identified. Genetic engineering, on the contrary, is a rational approach as particular metabolic steps are manipulated. The use of recombinant DNA technology to improve citric acid production has been employed only recently, although transformation of A. niger was reported in 1985 (Buxton et al., 1985; Kelly and Hynes, 1985). C. lipolytica transformation is also possible (Davidow et al., 1985), but we are not aware of any reports of genetic engineering of C. lipolytica to improve citric acid production. Different protocols for transformation of A. niger exist, but the most commonly used method involves polyethyleneglycol mediated uptake of DNA by protoplasts, followed by regeneration on a suitable selective medium. Introduction of DNA fragments (either circular or linearized) into A. niger results in integration of the DNA into the genome of the recipient strain. Integration can occur either at the homologous locus or at other loci. Multiple copies (tandemly integrated or scattered over the genome) of the gene introduced can be obtained. Expression of the gene introduced depends on the copy number and on the site of integration. To date there are three cases of genetic engineering concerning citric acid production by A. niger, which will be addressed in Sections 4.4.2 and 4.4.3, but first we will discuss some aspects of metabolic modelling.

4.4.1 Quantitative analysis of metabolism For genetic engineering it is necessary to have at least some idea of which enzymatic step should be altered to increase the metabolic flux through the pathway of citric acid biosynthesis. However, to find the optimal strategy for metabolic engineering, it is necessary to analyze the metabolism involved quantitatively. For example, the simple finding that an enzyme has a low activity in vitro does not mean that it is ‘rate-limiting’ in vivo, since the activity of an enzyme in the cell also depends on the concentrations of its substrates, products and possible effectors. To understand the control properties of a metabolic pathway, two major theoretical frameworks have been developed. Metabolic control analysis (MCA) was established independently by Kacser and Burns (1973) and Heinrich and Rapoport (1974), whereas biochemical systems theory (BST) was developed by Savageau (1976). The majority of the literature concerns MCA and the formalism of MCA and its applicability in biotechnology have been reviewed extensively (Kell and Westerhoff, 1986; Fell, 1992, 1997). Only a few of the basic concepts of MCA and BST will be discussed here. Both theories use the characteristics of the metabolic pathway under study, i.e. the kinetic properties of the enzymes, to describe it quantitatively. With this description it is possible to perform a sensitivity analysis. The effect of a small variation in, for example, the activity of an enzyme on the steady-state flux through the pathway (which is the rate of conversion of the primary substrate to the final product) can be calculated. In MCA the ‘flux control coefficient’ (C) was introduced, which is defined as the fractional change in flux (J) divided by the fractional variation in enzyme activity (e): (dJ/J)/(de/e). In most cases flux control coefficients have values between 0 (the flux does not change upon an increase in enzyme activity, i.e. no flux control) and 1 (the change in flux is proportional to the change in enzyme activity, i.e. the enzyme is completely ratelimiting).

62

Citric Acid Biotechnology

An important feature of MCA is the summation theorem, which states that the sum of the flux control coefficients of the enzymes in a pathway is equal to 1. As a consequence, the flux control coefficient of any enzyme in a very long pathway is probably very small as there are many enzymes contributing to control. Moreover, when an enzyme with some flux control is overproduced, control readily shifts to another step in the pathway. In practice this means that genetic engineering is not easy in complex pathways. The benefit of control analysis in designing strategies to optimize biotechnological processes depends heavily on the availability of enzyme kinetic data and on the reliability of these data. In the case of citric acid biosynthesis by A. niger quite a few enzymes have been studied now but a few, such as the transport steps, are less well or not at all investigated, hampering a precise analysis. Recently, a few attempts have been performed to analyse flux control in citric acid biosynthesis by A. niger (Torres, 1994a, 1994b; Ruijter et al., 1996; Torres et al., 1996a). Torres performed modelling of the first part of the pathway, i.e. up to pyruvate (see Figure 4.1) using BST formalism and suggested that sugar transport and phosphorylation, which are lumped into one step in the model, form the most important step in controlling the flux through the pathway. Thus, according to this model, the cellular amount of sugar transporter and/or hexokinase should be increased to obtain a higher metabolic flux. To a certain extent these findings correlate with experimental data. As described in Section 4.2 certain mutants with improved citric acid production had increased activity of hexokinase and phosphofructokinase (Schreferl-Kunar et al., 1989) and Steinböck et al. (1994) found that some 2-deoxyglucose resistant mutants had lower hexokinase activity and produced less citric acid than the parent. From an investigation of glucose transport in A. niger, Torres et al. (1996b) concluded that hexokinase contributed more to flux control in glycolysis than glucose transport. In a subsequent study (Torres et al., 1996a) it was concluded from flux optimization calculations that simultaneous overproduction of seven enzymes was required for a significant increase in flux. For practical reasons this is not achievable at the moment. Firstly, most of the A. niger genes required for this approach are not available and secondly, simultaneous overexpression of seven enzymes in a controlled way is experimentally difficult to accomplish. Notably, this model has not incorporated the metabolism from pyruvate to extracellular citric acid and hexokinase might have flux control in the conversion of glucose to pyruvate, but the control in the complete pathway (hexose to citric acid) might be in later steps i.e. between pyruvate and citric acid. Nevertheless, a modelling approach is worthwhile. It may not produce an exact solution to improve the process, but it provides a guideline for genetic engineering of A. niger.

4.4.2 Manipulation of the respiratory chain in A. niger In addition to the normal respiratory chain, A. niger possesses alternative respiratory enzymes, including an NADH oxidase and an ubiquinol oxidase (Zehentgruber et al., 1980; Kirimura et al., 1987; see also Figure 4.2). In the course of a citric acid fermentation the activities of the normal respiratory enzymes decrease whereas the activities of the alternative oxidases increase (Kirimura et al., 1987; Wallrath et al., 1991). The alternative oxidases do not pump protons concomitantly with electron transport and their physiological function is thought to be removal of excess reducing equivalents. Such a function is in agreement with the presence of the alternative oxidases during citric acid production. Conversion of hexoses to citric acid results in net production of ATP and NADH. Since

Strain improvement

63

Figure 4.2 Schematic representation of the normal and alternative respiratory chains. The normal respiratory chain (lower part) contains three complexes: NADH:ubiquinone oxidoreductase (complex I), ubiquinol:cytochrome c oxidoreductase (complex III) and cytochrome c oxidase (complex IV). In the alternative respiratory chain (top part) electrons are tranferred directly from ubiquinol to oxygen

there is no growth in the stage of citric acid accumulation, the cells probably do not require much ATP, and a switch from normal respiration to alternative oxidases would enable the fungus to reoxidize its NADH without concomitant ATP production. A very attractive hypothesis has been put forward by the group of Weiss. They found that the proton-pumping NADH:ubiquinone oxidoreductase (complex I) is very fragile in A. niger B-60, which is a good citric acid producer, compared to a wild-type A. niger strain (Schmidt et al., 1992; Wallrath et al., 1992). The selective loss of complex I might result in an increased NADH/NAD+ ratio in the cell, because the affinity of the alternative NADH oxidase for NADH is approximately one order of magnitude lower than that of complex I. Excretion of citric acid is a possibility in order to get rid of the excess reducing equivalents. As such, the switch to alternative oxidases is not a reaction of the fungus to citric acid production, but the loss of complex I results in initiation of citric acid accumulation. To test this hypothesis one of the subunits of complex I was inactivated in a ‘wild-type’ (bad producing) A. niger strain by disruption of the corresponding gene, nuo51, by molecular genetic techniques (Prömper et al., 1993). The mutant was unable to form a functional complex I and should accordingly accumulate citric acid as B-60 does. Unexpectedly, the mutant excreted virtually no citric acid, whereas the wild-type A. niger strain produced approximately 30 per cent of the yield obtained with B-60. However, the mutant accumulated high intracellular levels of TCA cycle intermediates, including citrate. Apparently, the mutant is indeed unable to reoxidize NADH under these conditions, resulting in accumulation of TCA cycle intermediates. Prömper et al. propose that, in contrast to wild-type A. niger and strain B-60, the mutant is unable to excrete citric acid (or other TCA cycle intermediates). This postulate, i.e. the presence of a citrate carrier, may explain the differences in citric acid production between wild-type A. niger and B-60, but does not resolve the discrepancy between wild-type A. niger and the mutant lacking complex I. It would be interesting to test the effect of disruption of nuo51 in strain B-60. In addition to the effect it might have on initiation of citric acid accumulation, it might also bring about an increase in the rate of acid production. Assuming an excess of ATP during citric acid production, inactivation of complex I would be a way to decrease such an excess, since less ATP is produced per NADH.

4.4.3 Engineering of glycolysis in A. niger Obviously a high metabolic flux is necessary for fast citric acid accumulation. To date, two reports have been published in which attempts to increase metabolic flux and hence productivity are described. Arisan-Atac et al. (1996) describe an increase in the rate of

64

Citric Acid Biotechnology

citric acid accumulation by a mutant of A. niger strain B-60 in which the gene encoding a subunit of trehalose-6-phosphate synthase, ggsA, was disrupted. This mutant lacks trehalose6-phosphate synthase activity and the rationale for construction of this strain was the following. Trehalose-6-phosphate and the enzyme catalyzing its biosynthesis have recently been shown to play a role in regulation of glycolytic flux in the yeast Saccharomyces cerevisiae (Thevelein and Hohmann, 1995). Trehalose-6-phosphate inhibits hexokinase activity in S. cerevisiae in vitro (Blázquez et al., 1993) and this was found to be also the case in A. niger (Arisan-Atac et al., 1996; Panneman et al., 1996). Inactivation of trehalose-6phosphate synthase would result in the inability to synthesize trehalose-6-phosphate and if, under citric acid producing conditions, trehalose-6-phosphate inhibits glycolysis in A. niger, the absence of trehalose-6-phosphate synthase might result in an increased glycolytic flux and increased citric acid production. This was indeed found to be the case. The ggsA disruption strain produced the same final yield of citric acid as the wild-type strain, but reached this yield in a shorter fermentation time. This is the only case where genetic engineering of A. niger results in improved citric acid production. Recently we have studied in our laboratory, the effects of overproduction of two glycolytic enzymes, phosphofructokinase and pyruvate kinase (Figure 4.1) on citric acid production by A. niger (Ruijter et al., 1997). A few experimental studies have suggested that phosphofructokinase might be an important step in control of the glycolytic flux. Firstly, cultivation on a high concentration of sucrose, glucose or fructose stimulated citric acid accumulation by A. niger and these conditions also led to increased intracellular levels of fructose-2,6-bisphosphate, a potent activator of phosphofructokinase (Kubicek-Pranz et al., 1990). Secondly, as already addressed in Section 4.2, mutants selected for the ability to grow fast on high concentrations of sucrose exhibited increased citric acid production and in these strains the activities of hexokinase and phosphofructokinase were twofold higher than in the parental strain (Schreferl-Kunar et al., 1989). We have overexpressed phosphofructokinase and pyruvate kinase, both individually and simultaneously, in A. niger N400 (Ruijter et al., 1997). Unfortunately, moderate overexpression of these enzymes (three to five times the wild-type level) did not enhance citric acid production by the fungus significantly (Figure 4.3). Overexpression of pyruvate kinase even appeared to have a negative effect on citric acid production. Thus, phosphofructokinase and pyruvate kinase do not seem to contribute in a major way to flux control of the metabolism involved in the conversion of glucose to citric acid. However, it must be noted that in cells overproducing phosphofructokinase, the concentration of fructose-2,6-bisphosphate was decreased approximately twofold compared to the wild-type. Hence, the fungus appears to adapt to overexpression of phosphofructokinase by decreasing the specific activity of the enzyme through a reduction in the level of fructose-2,6-bisphosphate. From his modelling studies Torres (1994b) also concluded that phosphofructokinase and pyruvate kinase did not have flux control. In the model of Torres, however, regulation of phosphofructokinase by fructose2,6-bisphosphate was not included. Our data suggest that overproduction of phosphofructokinase, while maintaining or increasing fructose-2,6-bisphosphate levels, may still increase glycolytic flux in A. niger.

4.5 Concluding remarks Although the strains utilized for commercial production of citric acid are undoubtedly highyielding, further strain improvement will most certainly be attempted. At the moment the primary strategy for strain breeding is probably still mutagenesis and selection. Quantitative

Strain improvement

65

Figure 4.3 Citric acid fermentation from glucose by an A. niger N400 wild-type strain (+) and transformants overproducing phosphofructokinase (O), pyruvate kinase (,) or phosphofructokinase and pyruvate kinase (ᮀ). Citric acid, glucose and dry weight are indicated (data taken from Ruijter et al., 1997)

analysis of metabolism and metabolic pathway engineering are only just being implemented, but in our view this is a promising approach, not so much as an alternative to the traditional strain breeding methods, but complementary to it.

4.6 Acknowledgements GR is financially supported by the Dutch Ministry of Economic Affairs, the Ministry of Education, Culture and Science, The Ministry of Agriculture, Nature Management and Fishery in the framework of an industrial relevant research programme of The Netherlands Association of Biotechnology Centres (ABON).

4.7 Reference
AKIYAMA, S, SUZUKI, T, SUMINO, Y, NAKAO, Y and FUKUDA, H, 1973a. Induction and citric acid productivity of fluoroacetate-sensitive mutant strains of Candida lipolytica, Agricultural and Biological Chemistry, 37, 879–884. AKIYAMA, S, SUZUKI, T, SUMINO, Y, NAKAO, Y and FUKUDA, H, 1973b. Relationship between aconitate hydratase activity and citric acid productivity in fluoroacetate-sensitive mutant strain of Candida lipolytica, Agricultural and Biological Chemistry, 37, 885–888.

66

Citric Acid Biotechnology

ARISAN-ATAC, I, WOLSCHEK, M F and KUBICEK, C P, 1996. Trehalose-6-phosphate synthase A affects citrate accumulation by Aspergillus niger under conditions of high glycolytic flux, FEMS Microbiology Letters, 140, 77–83. ARTS, E, KUBICEK, C P and RÖHR, M, 1987. Regulation of phosphofructokinase from Aspergillus niger: effect of fructose-2,6-bisphosphate on the action of citrate, ammonium ions and AMP, Journal of General Microbiology, 133, 1195–1199. AVCHIEVA, P B and VINAROV, A Y, 1993. Obtaining citric acid producing mutants of the yeast Candida lipolytica, Microbiologiya, 62, 161–165. BEGUM, A A, CHOUDHARY, N and ISLAM, M S, 1990. Citric acid fermentation by gamma-ray induced mutants of Aspergillus niger in different carbohydrate media, Journal of Fermentation and Bioengineering, 70, 286–288. BLAZQUEZ, M A, LAGUNAS, R, GANCEDO, C and GANCEDO, J M, 1993. Trehalose-6-phosphate, a new regulator of yeast glycolysis that inhibits hexokinases, FEBS Letters, 329, 51–54. BUXTON, F P, GWYNNE, D I and DAVIES, R W, 1985. Transformation of Aspergillus niger using the argB gene of Aspergillus nidulans, Gene, 37, 207–214. DAVIDOW, L S, APOSTOLAKOS, D, O’DONNELL, M M, PROCTOR, A R, OGRYDZIAK, D M and WING, R A, 1985. Integrative transformation of the yeast Yarrowia lipolytica, Current Genetics, 10, 39–48. FELL, D A, 1992. Metabolic control analysis—A survey of its theoretical and experimental development, Biochemical Journal, 286, 313–330. FELL, D, 1997. Understanding the Control of Metabolism (Portland Press, London). FINOGENOVA, T V, SHISHKANOVA, N V, ERMAKOVA, I T and KATAEVA, I A, 1986. Properties of Candida lipolytica mutants with the modified glyoxylate cycle and their ability to produce citric and isocitric acid. II. Synthesis of citric and isocitric acid by C. lipolytica mutants and peculiarities of their enzyme systems, Applied Microbiology and Biotechnology, 23, 378–383. GUPTA, S and SHARMA, C B, 1995. Citric acid fermentation by the mutant strain of the Aspergillus niger resistant to manganese ions inhibition, Biotechnology Letters, 17, 269–274. HAMISSA, F A, EL-ABYAD, M S, ABDU, A and GAD, A S, 1991. Raising potent UV mutants of Aspergillus niger van Tieghem for citric acid production from beet molasses, Bioresource Technology, 39, 209–213. HEINRICH, R and RAPOPORT, T A, 1974. A linear steady-state treatment of enzymatic chains: general properties, control and effector strength, European Journal of Biochemistry, 42, 89– 95. KACSER, H and BURNS, J A, 1973. The control of flux. In Rate Control of Biological Processes. Ed. D D DAVIES (Cambridge University Press), pp. 65–104. KELL, D B and WESTERHOFF, H V, 1986. Metabolic control theory: its role in microbiology and biotechnology. FEMS Microbiology Reviews, 39, 305–320. KELLY, J M and HYNES, M J, 1985. Transformation of Aspergillus niger by the amdS gene of Aspergillus nidulans, EMBO Journal, 4, 475–479. KIRIMURA, K, HIROWATARI, Y and USAMI, S, 1987. Alterations of respiratory systems in Aspergillus niger under the conditions of citric acid fermentation, Agricultural Biology and Chemistry, 51, 1299–1303. KIRIMURA, K, LEE, S P, NAKAJIMA, I, KAWABE, S and USAMI, S, 1988a. Improvement in citric acid production by haploidization of Aspergillus niger diploid strains, Journal of Fermentation Technology, 66, 375–382. KIRIMURA, K, NAKAJIMA, I, LEE, S P, KAWABE, S and USAMI, S, 1988b. Citric acid production by the diploid strains of Aspergillus niger obtained by protoplast fusion, Applied Microbiology and Biotechnology, 27, 504–506. KIRIMURA, K, SARANGBIN, S, RUGSASEEL, S and USAMI, S, 1992. Citric acid production by 2-deoxyglucose-resistant mutant strains of Aspergillus niger, Applied Microbiology and Biotechnology, 36, 573–577. KUBICEK, C P and RÖHR, M, 1986. Citric acid fermentation, Critical Reviews in Biotechnology, 3, 331–373. KUBICEK, C P, HAMPEL, W and RÖHR, M, 1979. Manganese deficiency leads to elevated amino acid pools in citric acid producing Aspergillus niger, Archives in Microbiology, 123, 73–79. KUBICEK, C P, SCHREFERL-KUNAR, G, WÖHRER, W and RÖHR, M, 1988. Evidence for a cytoplasmic pathway of oxalate biosynthesis in Aspergillus niger, Applied Environmental Microbiology, 54, 633–637.

Strain improvement

67

KUBICEK-PRANZ, E M, MOZELT, M, RÖHR, M and KUBICEK, C P, 1990. Changes in the concentration of fructose-2,6 bisphosphate in Aspergillus niger during stimulation of acidogenesis by elevated sucrose concentrations, Biochimica et Biophysica Acta, 1033, 250–255. MATTEY, M, 1992. The production of organic acids, Critical Reviews in Biotechnology, 12, 87–132. PANNEMAN, H, RUIJTER, G J G, VAN DEN BROECK, H C, DRIEVER E T M and VISSER, J, 1996. Cloning and biochemical characterisation of an Aspergillus niger glucokinase. Evidence for the presence of separate glucokinase and hexokinase enzymes, European Journal of Biochemistry, 240, 518–525. PONTECORVO, G, ROPER, J A and FORBES, E, 1953. Genetic recombination without sexual reproduction in Aspergillus niger, Journal of General Microbiology, 8, 198–210. PRÖMPER, C, SCHNEIDER, R and WEISS, H, 1993. The role of the proton-pumping and alternative respiratory chain NADH-ubiquinone oxidoreductases in overflow catabolism of Aspergillus niger, European Journal of Biochemistry, 216, 223–230. RÖHR, M, STADLER, P J, SALZBRUNN, W O J and KUBICEK, C P, 1979. An improved method for characterisation of citrate production by conidia of Aspergillus niger, Biotechnology Letters, 1, 281–286. RÖHR, M, KUBICEK, C P and KOMINEK, J, 1983. Citric acid. In Biotechnology, Vol 3. Eds H-J REHM and G REED (Verlag Chemie), pp. 419–454. RUGSASEEL, S, KIRIMURA, K and USAMI, S, 1993. Selection of mutants of Aspergillus niger showing enhanced productivity of citric acid from starch in shaking culture, Journal of Fermentation and Bioengineering, 75, 226–228. RUIJTER, G J G, SNOEP, J L and VISSER, J, 1996. Modelling of carbohydrate metabolism in citric acid producing Aspergillus niger. In BioThermoKinetics of the Living Cell. Eds H V WESTERHOFF, J L SNOEP, F E SLUSE, J E WIJKER and B N KHOLODENKO (BioThermoKinetics Press), pp. 413–416. RUIJTER, G J G, PANNEMAN H and VISSER, J, 1997. Overexpression of phosphofructokinase and pyruvate kinase in citric acid producing Aspergillus niger, Biochimica et Biophysica Acta, 133, 317–326. SARANGBIN, S, KIRIMURA, K and USAMI, S, 1993. Citric acid production from cellobiose by 2deoxyglucose-resistant mutant strains of Aspergillus niger in semi-solid culture, Applied Microbiology and Biotechnology, 40, 206–210. SARANGBIN, S, MORIKAWA, S, KIRIMURA, K and USAMI, S, 1994. Formation of autodiploid strains in Aspergillus niger and their application to citric acid production from starch, Journal of Fermentation and Bioengineering, 77, 474–478. SAVAGEAU, M A, 1976. Biochemical System Analysis: A Study of Function and Design in Molecular Biology (Addison Wesley, Reading, MA). SCHMIDT, M, WALLRATH, J, DORNER, A and WEISS, H, 1992. Disturbed assembly of the respiratory chain NADH-ubiquinone reductase (complex I) in citric-acid-accumulating Aspergillus niger strain B-60, Applied Microbiology and Biotechnology, 36, 667–672. SCHREFERL, G, KUBICEK, C P and RÖHR, M, 1986. Inhibition of citric acid accumulation by manganese ions in Aspergillus niger mutants with reduced citrate control of phosphofructokinase, Journal of Bacteriology, 165, 1019–1022. SCHREFERL-KUNAR, G, GROTZ, M, RÖHR, M and KUBICEK, C P, 1989. Increased citric acid production by mutants of Aspergillus niger with increased glycolytic capacity, FEMS Microbiology Letters, 59, 297–300. STEINBÖCK, F, CHOOJUN, S, HELD, I, RÖHR, M and KUBICEK, C P, 1994. Characterisation and regulatory properties of a single hexokinase from the citric acid accumulating fungus Aspergillus niger, Biochimica et Biophysica Acta, 1200, 215–223. SUZUKI, A, SARANGBIN, S, KIRIMURA, K and USAMI, S, 1996. Direct production of citric acid from starch by a 2-deoxyglucose-resistant mutant strain of Aspergillus niger, Journal of Fermentation and Bioengineering, 81, 320–323. THEVELEIN, J M and HOHMANN, S, 1995. Trehalose synthase: guard to the gate of glycolysis in yeast? Trends in Biochemical Sciences, 20, 3–10. TORRES, N V, 1994a. Modelling approach to control of carbohydrate metabolism during citric acid accumulation by Aspergillus niger: I. Model definition and stability of the steady-state, Biotechnology and Bioengineering, 44, 104–111.

68

Citric Acid Biotechnology

TORRES, N V, 1994b. Modelling approach to control of carbohydrate metabolism during citric acid accumulation by Aspergillus niger: II. Sensitivity analysis, Biotechnology and Bioengineering, 44, 112–118. TORRES, N V, VOIT, E O and GONZÁLEZ-ALCÓN, C, 1996a. Optimisation of non-linear biotechnological processes with linear programming: application to citric acid production by Aspergillus niger, Biotechnology and Bioengineering, 49, 247–258. TORRES, N V, RIOL-CIMAS, J M, WOLSCHEK, M and KUBICEK, C P, 1996b. Glucose transport by Aspergillus niger: the low-affinity carrier is only formed during growth on high glucose concentrations. Applied Microbiology and Biotechnology, 44, 790–794. VAN DE MERBEL, N C, RUIJTER, G J G, LINGEMAN, H, BRINKMAN, U A TH and VISSER, J, 1994. An automated monitoring system using on-line ultrafiltration and column liquid chromatography for Aspergillus niger fermentations, Applied Microbiology and Biotechnology, 41, 658–663. WALLRATH, J, SCHMIDT, M and WEISS, H, 1991. Concomitant loss of respiratory chain NADH:ubiquinone reductase (complex I) and citric acid accumulation of Aspergillus niger, Applied Microbiology and Biotechnology, 36, 76–81. WALLRATH, J, SCHMIDT, M and WEISS, H, 1992. Correlation between manganese-deficiency, loss of respiratory chain complex I activity and citric acid production in Aspergillus niger, Archives in Microbiology, 158, 435–438. WIEBE, M G, ROBSON, G D and TRINCI, A P J, 1989. Effect of choline on the morphology, growth and phospholipid composition of Fusarium graminearum, Journal of General Microbiology, 135, 2155–2162. WITTEVEEN, C F B and VISSER, J, 1995. Polyol pools in Aspergillus niger, FEMS Microbiology Letters, 134, 57–62. WITTEVEEN, C F B, VAN DE VONDERVOORT, P J I, SWART, K and VISSER, J, 1990. Glucose oxidase overproducing and negative mutants of Aspergillus niger, Applied Microbiology and Biotechnology, 33, 683–686. WITTEVEEN, C F B, VEENHUIS, M and VISSER, J, 1992. Localisation of glucose oxidase and catalase activities in Aspergillus niger, Applied Environmental Microbiology, 58, 1190–1194. WITTEVEEN, C F B, VAN DE VONDERVOORT, P J I, VAN DEN BROECK, H C, VAN ENGELENBURG, F A C, DE GRAAFF, L H, HILLEBRAND, M H B C, et al., 1993. Induction of glucose oxidase, catalase, and lactonase in Aspergillus niger, Current Genetics, 24, 408–416. ZEHENTGRUBER, O, KUBICEK, C P and RÖHR, M, 1980. Alternative respiration of Aspergillus niger, FEMS Microbiology Letters, 8, 71–74.

5

Fungal Morphology
MARIA PAPAGIANNI

5.1 Introduction In submerged culture the morphology of filamentous micro-organisms varies between pellets and free filaments depending on culture conditions and the genotype of the strain. All the growth forms have their own characteristics concerning growth kinetics, nutrient consumption and broth rheology. Of the two extremes, pellet suspensions exhibit Newtonian rheological behaviour, while the filamentous form produces more viscous media with consequent effects of poor mixing and mass transfer. This is unfortunate, as it is very often the case that the disperse filamentous form is the productive form. Another drawback with the pelleted suspension is that cell growth occurs only at the surface of the pellets where contact with oxygen and other nutrients is adequate, and the cells growing within a pellet respond to a very different environment. Further into the pellet, mass transfer limitation will gradually occur and cells could autolyse. This chapter deals with the factors that affect Aspergillus niger morphology in submerged culture and the influence of morphology on productivity in citric acid fermentation.

5.2 Factors affecting Aspergillus niger morphology in submerged culture According to many reports, the morphology of the mycelium is crucial to the process of fermentation, not only in relation to the shape of the hyphae themselves and the aggregation into microscopic clumps (micro-morphology), but also in the pelleted form of growth (macro-morphology). In all cases reported, the mycelium of acidogenic Aspergillus niger was found to conform to the morphological type described by Snell and Schweiger (1951): short, swollen branches which may have swollen tips. The mycelial pellets should be small with a hard, smooth surface (Clark, 1962). It is known that this is brought about by adjustment of aeration and agitation (Svenska Sockerfabrik, 1964), adjustment of pH (Fried and Sandza, 1959), concentration of manganese and other trace metals (Shu and Johnson, 1948; Clark et al., 1966; Kisser et al., 1980), and inoculum level (Berry et al., 1977). However, it is not known whether the pelleted or filamentous form is more desirable for citric acid production.
69

70

Citric Acid Biotechnology

Since the morphological form can strongly influence the overall process productivity, research on various aspects of morphological development has attracted the interest of academia and industry and attempts have been made to induce a particular form of growth and to relate morphology to product synthesis. Initial investigations of mycelial morphology relied on manual measurements from photographs and little quantitative work was presented until the 1970s. Detailed morphological characterization of the free filamentous form was first presented by Metz et al. (1981). Their method, which made use of an electronic digitizer to make measurements from microphotographs, was time consuming, inaccurate and also difficult to automate. In 1988, Adams and Thomas presented the first image analysis method for morphological measurements of a filamentous fungus, using images taken directly from the microscope to an image analyzer. Since then, highly automated methods have been developed which have many applications and allow detailed characterization of growth and simple differentiation of filamentous micro-organisms (Thomas, 1992). With the large variety of products produced by filamentous organisms and their complex physiology, a method that provides accurate and reproducible quantitative morphological characterization is invaluable in studies of process optimization and modelling. In this chapter, the effects of agitation, nutritional factors (type and concentration of carbon source, nitrogen and phosphate limitation, pH, dissolved oxygen tension, trace metals levels) and inoculum size will be discussed with respect to the micro-morphology of A. niger.

5.3 Effect of agitation In submerged fermentation, agitation is important for adequate mixing, mass transfer and heat transfer. For aerobic fermentation, mixing is required to ensure sufficient oxygen transfer throughout the reactor vessel and aeration has been shown to have a critical effect on the submerged process of citric acid fermentation. Agitation creates shear forces that affect micro-organisms in several ways, causing morphological changes, variation in their growth and product formation and also damage in the cell structure. For the dispersed form, in filamentous fermentation the effects of agitation superimposed on the fermentation process are difficult to quantify. Changes in morphology of filamentous fungi as a result of intensive agitation conditions have been observed in many cases (Dion et al., 1954; Belmar-Beiny and Thomas, 1991; Papagianni et al., 1994). Under these conditions hyphae were thick, short and densely branched and this morphological type is usually associated with increased product yields. However, high impeller speeds were found to promote mycelial growth and possibly to stimulate the occurrence of metabolic pathways which resulted in low productivity of citric acid. The effect of stirrer speed on growth and productivity of three Aspergillus niger strains was reported by Ujcova et al. (1980). Higher speeds resulted in thicker and highly branched filaments. There was a drop in productivity at higher speeds although growth remained rapid. A similar effect for penicillin fermentation was reported by König et al. (1981). At higher speeds only a short period of penicillin production was maintained and a large fraction of the substrate was converted into carbon dioxide. It has been reported that increased agitation can lead to breakage of hyphae for a number of micro-organisms (Märkl and Bronnenmeier, 1985; Belmar-Beiny and Thomas, 1991). Although A. niger cultures are normally resistant to shear damage, mycelial fragmentation due to mechanical forces has been reported. The damage of hyphae and the

Fungal morphology

71

Figure 5.1 Effect of agitation on citric acid production and the relation between production and morphology parameters in the tubular loop reactor

consecutive release of intracellular material may account for the decreased productivities reported in many cases under intensive agitation conditions. This has been proven for Penicillium chrysogenum. Smith et al. (1990) and Makagiansar et al. (1993) observed that the lower rates of penicillin synthesis at high agitation speeds were due to increased damage of hyphae, since it involved a greater frequency of circulation of mycelia through the high energy dissipation zone around the impeller. Case study 1 The effect of agitation on A. niger morphology and citric acid production has been studied in a tubular loop (TLR) and a stirred tank reactor (STR), through a series of batch experiments carried out at different circulation times (4 to 18 s) in the case of the TLR and at stirrer speeds from 100 to 600 rpm for the STR (Papagianni, 1995). Both reactors were inoculated with a vegetative inoculum. The inoculum filaments started to clump within 24 hours of fermentation in all experiments. Morphological measurements using image analysis showed that by increasing the intensity of agitation, the size of clumps (P1) decreased, as did the length of the filaments (L) that arose from the cores of clumps, while the diameter of filaments (d) increased. The perimeter of the clumps was measured by joining the tips of the filaments that arose from the core of the clumps. For the estimation of the core of clumps (P2), lines were drawn around the core and their combined length was measured. For the estimation of L, the length of the filaments and branches that arose from the core of the clump was measured. In both fermenters, specific rates of citric acid formation, sp.rp, increased with agitation; the amount of citric acid produced at the end of the runs (at 168 h) was dependent on the circulation time and the stirrer speed. In the STR, as the stirrer speed increased citric acid production increased up to a point (300 rpm) beyond which it remained constant. In Figures 5.1 and 5.2 the effect of agitation on citric acid production and the relation between production and morphology is shown. In both reactors low dissolved oxygen levels were observed under conditions of low agitation intensities. If the intensity of agitation was changed, different patterns of morphological development were observed in both fermenters. The effect of intensive agitation was more pronounced in the stirred tank reactor, since the length of filaments was reduced by a factor of three, while in the loop rector the reduction was much smaller. Figure 5.3 shows the

72

Citric Acid Biotechnology

Figure 5.2 Effect of agitation on citric acid production and the relation between production and morphology in the stirred tank reactor

Figure 5.3 Time course for the clump perimeter, P1, at different stirrer speeds in the stirred tank fermentations

time course of P1 in the STR fermentations whilst the time course for the hyphal length in the TLR is shown in Figure 5.4. The mean diameters of filaments also changed during fermentation; filaments became thinner with time in all experiments performed in both fermenters. The reduction of hyphal diameter was found to be dependent on agitation: the faster the broth circulation and the higher the stirrer speed, the more rapid was the reduction. These differences in morphological development during fermentation could not be explained by the assumption of increased branching alone (newly formed branches would

Fungal morphology

73

Figure 5.4 Time course for the specific growth rate and hyphal length, L, as a function of circulation time in the tubular loop reactor

lower the mean values of L and P1) under conditions of intensive agitation. It is known that increased branching frequency is associated with increased specific growth rates (Katz et al., 1971; Morrison and Righelato, 1974). This is not the case for these experiments. For the runs in which values for the specific growth were comparable rate (e.g. at circulation times 4s and 10s in the TLR), very different time courses of L as well as of P1 and P2 were observed. This also applied to stirred tank fermentations; at 500 rpm, L decreased rapidly for the first 100 hours and for the rest of the run decrease was slower, while m values for the period 30 to 72 hours were lower than those noted at 300 rpm. Hyphal fragmentation as a result of increased agitation intensity could explain the different patterns of time courses observed in these studies. Thus, under intensive agitation conditions a cycle of fragmentation and regrowth takes place while at low agitation intensities a gradual ageing process predominates and the filaments grow long, with few branches remaining; mean values of L and P1 are higher. Mitard and Riba (1988), studying the effect of stirrer speed on A. niger growth and morphology, observed that there was a relationship between the specific growth rate of the organism and the rupture of mycelial aggregates. As the aggregates were broken, the specific growth rate reduced; it increased again as the liberated filaments went on growing. This could explain the lower specific growth rates observed at 500 rpm during the period between 30 and 72 hours and the rapid reduction of the length of filaments. At this stirrer speed P1 and P2 also decreased for 100 hours from inoculation; after this period their mean values increased again towards the end of the run. This indicates that fragmentation and regrowth took place. From Figures 5.1 and 5.2, it can be observed that beyond a stirrer speed of 300 rpm, changes in morphology were not followed by changes in citric acid production. However, the specific rates of citric acid formation were stirrer speed dependent as was the morphology. As the specific production rate increased with the stirrer speed, the parameters

74

Citric Acid Biotechnology

Figure 5.5 Plots of the final values for citric acid concentration and the morphology parameter P1 against the relative mixing time in the loop and stirred reactors

P1 and L reduced. The core perimeter, P2, appeared to be not affected by agitation and it cannot be directly linked to citric acid production. As Figures 5.1 and 5.2 show, the morphological characteristics of the broth are similar in the two reactor systems and there appear to be no fundamental differences between the results obtained from the loop and the stirred tank reactor. To compare the two different systems the circulation times in the STR were estimated and they were found to be 1.1, 1.8, 2.2, 3, 3.8 and 6 seconds, while the mixing times were 8, 12, 14, 15, 16 and 22 seconds for the stirrer speed range of 600 to 100 rpm. Compared to those of the loop reactor, both circulation and mixing times in the STR were smaller. As a means of comparing the mixing characteristics of the two ferments, the dimensionless parameter relative mixing time, Jm, can be used. The relative mixing time is equal to the mixing time divided by the circulation time (Jm = tm/tc). It was calculated and found to be within the range of 4 to 8 in both reactors. In Figure 5.5, the final values for citric acid concentration and P1 have been plotted against Jm for both reactors. As shown in this figure, by increasingJm, citric acid production increased, while the perimeter of clumps decreased, with good agreement between the two reactor systems. It appears that the amount of product and morphology is a function of Jm and for this fermentation, production and morphology are strongly linked.

5.4 Effect of nutritional factors Citric acid accumulation is strongly influenced by the composition of the nutrient medium. The medium constituents which have been found to exert an effect on citric acid fermentation are: type and concentration of the carbon source, supply of nitrogen and phosphate, pH, dissolved oxygen levels and concentration of certain trace metals. The influence on

Fungal morphology

75

morphology of A. niger or other filamentous micro-organisms in submerged culture for most of these factors has been studied.

5.4.1 Type and concentration of carbon source The carbon source for citric acid fermentation has been the focus of much study, frequently with a view to the utilization of polysaccharide sources (Gupta et al., 1976; Hossain et al., 1984; Xu et al., 1989). The nature of the source has been shown in many cases to affect citric acid production, since it exerts a strong effect on levels of enzyme activity within the TCA cycle. In general, only sugars, which are rapidly taken up by the fungus, allow a high final yield of citric acid. Polysaccharides, unless hydrolyzed, are not included in this category. Information concerning the role of the nature of sugar source on A. niger morphology in citric acid fermentation is limited. However, it has been observed that factors favouring increased growth rates, such as media rich in easily assimilated nutrients, affect morphology by reducing pellet formation in filamentous organisms (Hemmersdorfer et al., 1987). Not only the type, but also the concentration of the carbon source is critical to this fermentation, influencing the rate of production and the final yield, in addition to growth of the fungus. In the following two case studies, the influence of glucose concentration on morphology and productivity of A. niger in batch and fed-batch culture will be discussed. Case study 2 Studies in conventional batch culture confirmed that the initial glucose concentration in the fermentation medium affects the rates of citric acid fermentation and morphology of A. niger (Papagianni, 1995). In batch experiments performed in an STR at initial glucose concentrations of 150 g l-1, 100 g l-1 and 60 g l-1 it was found that the specific production rate decreased with the initial glucose concentration. For the first 48 hours, the specific growth rate increased as initial glucose concentrations decreased. Initial glucose concentration clearly affected the length of filaments, L, as Figure 5.6 shows. In early fermentation stages, L decreased with decreasing glucose levels. Towards the end of the fermentation, when the differences in the sugar level diminish, it seems that the length of the filaments converges. The parameter L could be regarded as an indication of branching, since the high degree of clumping made impossible the counting of branching points. As these fermentations were performed at the same stirrer speed and pH and the specific growth rate was found to increase for the first 48 hours of fermentation, there should be a link between specific growth rate and branching frequency for these experiments; a large value for L would indicate few branches. Increased branching frequency and reduction in the hyphal growth unit with increasing specific growth rates have been reported in the literature (Katz et al., 1972; Morrison and Righelato, 1974). Case study 3 To eliminate the effect of a constantly changing glucose concentration in the batch experiments, a series of fed-batch runs, where glucose levels were maintained constant during the citric acid production phase, were performed under otherwise identical fermentation conditions. The range of glucose levels included the following concentrations: 130 g l-1, 88 g l-1, 70 g l-1, 44 g l-1, 17 g l-1 and 5 g l-1. The specific rate of citric acid formation was found to increase with glucose levels (see Figure 5.7). In contrast, for

76

Citric Acid Biotechnology

Figure 5.6 Effect of initial glucose concentration on length of filaments in batch fermentations

Figure 5.7 Effect of glucose levels on the specific rate of citric acid production in fed-batch fermentations

Fungal morphology

77

Figure 5.8 Time course for the mean length of filaments under different glucose levels in fedbatch fermentations

the early stages of fermentation (48 hours), the highest specific growth rate values were observed when glucose was maintained at the lowest glucose concentration of 5 g l-1. Later in the fermentation, glucose level had little influence on the specific growth rate. The mean length of filaments at 24 hours in fed-batch runs with glucose levels in the range from 130 g l-1 to 17 g l-1, decreased with decreasing glucose levels (Figure 5.8). Mean values of P1 (Figure 5.9) and L were found to be smaller in fed-batch runs with glucose levels between 130 g l-1 and 17 g l-1 after 24 hours of fermentation, than those observed in the batch run with initial glucose at 150 g l-1. The values of the specific growth rate were significantly higher in fed-batch experiments for the first two days of fermentation than those obtained in the batch run at 150 g l-1 glucose. Since it was observed in both cultures that the specific growth rate values for the first 48 hours were increased with decreasing glucose levels, the reduction in L could be a result of increased branching. Fermentation rates and morphology developed in a different way when glucose was kept at 5 g l-1 throughout fermentation. The very low glucose levels in this run affected metabolism, since citric acid formation was reduced in favour of cell growth and the changes in metabolism were accompanied by changes in morphology, as shown in Figures 5.8 and 5.9. Morphological changes also included the appearance of pellets after three days of fermentation, although the bulk of the mycelium was in the form of clumps. It has been suggested that factors favouring increased growth rates may reduce pellet formation in fungi (Hemmersdorfer et al., 1987). This is in contrast to our observations

78

Citric Acid Biotechnology

Figure 5.9 Time course for the perimeter of the clumps under different glucose levels in fedbatch fermentations

of the values for the specific growth rate which were the highest noted and for the first time pellets appeared in the broth. It seems clear that the carbon source concentration affects citric acid production rates and A. niger growth and morphology. For this system morphology and product formation were closely related, since the reduction of the length of filaments and size of clumps was associated with increased specific production rates.

5.4.2 Nitrogen and phosphate limitation In order to accumulate citric acid, growth must be restricted, but it is not clear whether phosphate or nitrogen is the necessary limiting factor. According to Shu and Johnson (1948), phosphate does not have to be limiting, but when trace metal levels are not limiting, additional phosphate results in side reactions and increased growth. Kubicek and Röhr (1977) showed that citric acid accumulated whenever phosphate was limited even when nitrogen was not. In contrast, Kristiansen and Sinclair (1979), using continuous culture, concluded that nitrogen limitation was essential for citric acid production. Pellet formation in filamentous fungi has been discussed in many cases and among the factors considered to induce it, is the limitation of particular nutrients, including nitrogen (Braun and Vecht-Lifshitz, 1991). On the other hand, factors favouring increased growth rates, including excess phosphate concentrations, have been shown to reduce pellet formation. The following case study examines the effect of increased phosphate concentration on morphology of the free filamentous form.

Fungal morphology

79

Figure 5.10 Effect of phosphate level on A. niger morphology in the loop reactor at a circulation time of 18 s

Case study 4 Figure 5.6 shows the time courses of the morphological parameters P1, P2 and L, of A. niger clumps, for two experiments with KH2PO4 concentration in the fermentation medium of 0.1 g l-1 and 0.5 g l-1. Both experiments were carried out in the loop fermenter of case study 1, at the circulation time of 18 s. As Figure 5.10 shows, a small increase in phosphate level led to drastic changes in morphology. The clump perimeter became almost three times larger, while the perimeter of the core remained small. The clumps lost their compact structure and the form of an extended growth around a small core predominated. The length of filaments increased during fermentation, while at the lower phosphate level, after an early growth phase, L remained at the same levels until the end of the run. The fermentation itself was also drastically affected, since the yield of citric acid on glucose consumed fell from 70 per cent to 39 per cent on increasing the phosphate level, while biomass concentration increased from 6.1 g l-1 at the lower to 11.5 g l-1 at the higher phosphate level. The differences in specific growth rates and biomass concentrations could explain the different time course for the hyphal length L in the two runs. In contrast to the very high increase of P1, L remained comparatively small. This could be an indication of increased branching with increasing phosphate levels. Although the strain used in this work did not form macroscopic pellets but microscopic clumps, factors such as increased phosphate concentrations seem to exert similar controls on fundamentally different morphological types.

5.4.3 pH Culture pH can have a profound effect on citric acid production by A. niger, since certain enzymes within the TCA cycle are pH sensitive. The maintenance of a low pH during fermentation is vital for a good yield and it is generally considered necessary for the pH

80

Citric Acid Biotechnology

Figure 5.11 Effect of culture pH on citric acid production in the stirred tank reactor

to fall to around pH 2 within a few hours of the initiation of the process, otherwise the yields are reduced (Mattey, 1992). Information concerning the effect of pH upon the morphology of citric acid producing A. niger is very limited. Reports on the effect of pH on morphology for other fungi are contradictory; either it influences morphology greatly or it has no effect at all (Pirt and Callow, 1959; Van Suijdam and Metz, 1981; Miles and Trinci, 1983). The following case study examines the effect of culture pH on citric acid production and A. niger morphology in the stirred tank reactor. Case study 5 Fermentations were carried out at 500 rpm and uncontrolled pH (resulting in a final pH of 1.6) and controlled pH (by addition of titrants) at 2.1 and 3 (Papagianni, 1995). Citric acid production was highest when pH was maintained at 2.1, with 122 g l-1 at the end of the run (168 hours of fermentation), compared to 75 g l-1 at pH 3 and 65 g l-1 at pH 1.6, as shown in Figure 5.11. Biomass levels were slightly increased with increasing pH: 5.65 g l-1 at pH 1.6, 6.21 g l-1 at 2.1 and 7.40 g l-1 at pH 3. The highest values of specific rates of citric acid formation were obtained at pH 2.1, with a maximum value 0.35 h-1 while it reached 0.18 l1 in the other two runs. Mean values of the morphological parameters P1, P2, L and d at the end of the runs performed at 500 rpm and pH 2.1, 3 and uncontrolled are shown in Table 5.1. P1, P2 and L increased with pH whilst there was no unidirectional response for the diameter of filaments. In addition to the small size of clumps and small length of filaments at final pH 1.6, there was an unusually high number of swollen cells and tips in the mycelium in this run, as shown in Figure 5.11. A number of swollen cells were always present at stirrer speeds above 400 rpm in the STR. The low pH in this experiment seemed to aid the development of this morphological form which also gave low citric acid concentrations.

Fungal morphology

81

Table 5.1 The influence of pH on selected morphological parameters (measurements taken at the end of fermentation in a lab-scale fermenter operating at constant stirred speed of 500 rpm)

These experiments also indicated that productivity and morphology are linked. As with agitation, the conditions which promote a certain morphological type, i.e. that of small clumps and short filaments, favour citric acid production, much as observed with the relationship between macro-morphology and citric acid production.

5.4.4 Dissolved oxygen tension It has been shown that oxygen acts as a direct regulator of citric acid accumulation as it is favoured by increasing the dissolved oxygen tension of the fermentation medium (Kubicek et al., 1980). As mentioned in the agitation case study, lower dissolved oxygen levels occurred with morphologies not associated with high yields on citric acid. However, reports on the effect of dissolved oxygen tension on the macro-morphology of A. niger suggest that no direct relationship exists between the two. Gomez et al. (1988), in their work on citric acid production from A. niger, found that no difference in morphology for pellets and filaments could be ascribed to dissolved oxygen levels, although production on citric acid was enhanced, particularly from pellets, by increasing the dissolved oxygen at different fermentation stages. Similarly, Van Suijdam and Metz (1981) showed that oxygen tension in the range of 12 to 300 mg Hg had no influence on the morphology of P. chrysogenum. These reports contradict the limitation hypothesis made by Hemmersdorfer et al. (1987), which suggests that lack of any particular nutrient, including oxygen, induces pellet formation.

5.4.5 Trace metal level A number of divalent metals have been suggested as being required in limiting amounts for a successful citric acid process. These include Fe2+, Cu2+, Zn2+, Mn2+ and Mg2+ (Shu and Johnson, 1948; Mattey, 1992). Only the effect of manganese concentration has been shown to influence A. niger morphology. Manganese ions are known to be specifically involved in many cellular processes, such as cell wall synthesis, sporulation and production of secondary metabolites (Kubicek and Röhr, 1977). Cellular anabolism of A. niger is impaired under Mn deficiency and/or nitrogen and phosphate limitation. The protein breakdown under Mn deficiency results in a high intracellular concentration. This causes inhibition of the enzyme phosphofructokinase (essential enzyme in the conversion of glucose and fructose to pyruvate), leading to a flux through glycolysis and the formation of citric acid (Habison et al., 1979; Röhr and Kubicek, 1981). Kisser et al. (1980) studied morphology and cell wall composition of A. niger under conditions of Mn sufficient and deficient cultivation in an otherwise citric acid producing

82

Citric Acid Biotechnology

medium. Omission of Mn ions (less than 10-7) from the nutrient medium resulted in abnormal morphological development that was characterized by increased spore swelling and squat, bulbous hyphae. The inhibition of glucoprotein turnover caused by the presence of Mn ions led to a possible loss of hyphal polarity and increased branching and chitin synthesis. Clark et al. (1966) also discussed changes in A. niger morphology following the addition of Mn. The authors noticed an undesirable change in morphology from the pellet like form to filamentous form with the addition of 2 ppb Mn to ferrocyanide-treated molasses. Morphological changes, which included prevention of clumping, absence of swollen cells and reduced diameters of filaments, accompanied by a 20 per cent reduction in citric acid yield, following the addition of 30 mg l-1 Mn to a Mn-free medium, were also reported by Papagianni (1995).

5.5 Effect of inoculum Among the factors that determine morphology and the general course of fungal fermentations, the amount and type of inoculum is of prime importance. Early attempts have been made to standardize inocula for citric acid production in submerged culture (Martin and Waters, 1952; Steel et al., 1955). Van Suijdam et al. (1980) reported that A. niger pellets would only form at inoculum sizes below 1011 spores per m3. However, the effect of inoculum on mycelial morphology in submerged culture has been assessed mainly by the presence or absence of pellets and their characteristics (Smith and Calam, 1980; Vecht-Lifshitz et al., 1990). The reason for this was the lack of an adequate method to monitor mycelial morphology during fermentations. Morphology was quantified by an image analysis method (Tucker et al., 1992) in the work of Tucker and Thomas (1992); a sharp transition from pelleted to dispersed forms of growth for Penicillium chrysogenum was reported, as inoculum levels rose towards 5 × 105 spores per m3. This suggests that research on the inoculum in citric acid fermentation could now be more systematic, making use of the technological advances in characterization and monitoring of morphology in fungal fermentations.

5.6 Conclusions and perspectives Discussion of the factors influencing A. niger morphology in submerged culture should distinguish between macro- and micro-morphology although a number of similarities exist in relation to citric acid production and responses to the environment. This chapter has concentrated on micro-morphology. The case studies presented suggest a strong relationship between morphology and productivity in citric acid fermentation. The observations indicate that it might be possible to manipulate the morphology parameters in order to improve bioreactor performance and process yields. Image analysis provides the tools for monitoring these parameters; however, further research is required to reveal possible general trends in metabolite regulation in relation to morphology of the producer micro-organism.

5.7 References
ADAMS, H L, and THOMAS, C R, 1988. The use of image analysis for morphological measurements on filamentous micro-organisms, Biotechnology and Bioengineering, 32, 707–712.

Fungal morphology

83

BELMAR-BEINY, M T and THOMAS, C R, 1991. Morphology and clavulanic acid production of Streptomyces clavuligerous: effect of stirrer speed in batch fermentations, Biotechnology and Bioengineering, 37, 456–462. BERRY, D R, CHMIEL, A and AL OBAIDI, Z, 1977. Citric acid production by A. niger. In Genetics and Physiology of Aspergillus. eds J E SMITH and J A PATEMAN (Academic Press, London), pp. 405–426. BRAUN, S and VECHT-LIFSHITZ, S E, 1991. Mycelial morphology and metabolite production, Trends in Biotechnology, 9, 63–68. CLARK, D S, 1962. Submerged citric acid fermentation of ferrocyanide-treated molasses: morphology of pellets of A. niger, Canadian Journal of Microbiology, 8, 133–136. CLARK, D S, ITO, K and HORITSU, H, 1966. Effect of manganese and other heavy metals on submerged citric acid fermentation of molasses, Biotechnology and Bioengineering, 8, 465–471. DION, W M, CARILLI, A, SERMONTI, G and CHAIN, E B, 1954. The effect of mechanical agitation on the morphology of Penicillium chrysogenum Thom in stirred fermentors. Rend. Ist. Super. de Sanita, 17, 187–205. FRIED, J H and SANDZA, J G, 1959. Production of citric acid, US Patent 2 No. 910 409, Chemical Abstracts, 54, 7063b. GOMEZ, R, SCHNABEL, I and GARRIDO, J, 1988. Pellet growth and citric acid yield of Aspergillus niger 110, Enzyme and Microbial Technology, 10, 188–191. GUPTA, J K, HELDING, L G and JØRGENSEN, O B, 1976. Effect of sugars, hydrogen ion concentration and ammonium nitrate on the formation of citric acid by Aspergillus niger, Acta Microbiologie Academy of ScienceHungary, 23, 63–67. HABISON, A, KUBICEK, C P and RÖHR, M, 1979. Phosphofructokinase as a regulatory enzyme in citric acid producing A. niger, FEMS Microbiology Letters, 5, 39–42. HEMMERSDORFER, H, LEUCHTENBERGER, A, WARDSACK, C and RUTTLOFF, H, 1987. Journal of Basic Microbiology, 27, 309–315. HOSSAIN, M, BROOKS, J D and MADDOX, I S, 1984. The effect of the sugar source on citric acid production by Aspergillus niger, Applied Microbiology and Biotechnology, 19, 393–397. KATZ, D, GOLDSTEIN, D and ROSENBERG, R F, 1971. Model for branch initiation in Aspergillus nidulans based on measurements of growth parameters, Journal of Bacteriology, 109, 1097– 1100. KISSER, M, KUBICEK, C P and RÖHR, M, 1980. Influence of manganese on morphology and cell wall composition of A. niger during citric acid fermentation, Archives in Microbiology, 128, 26– 33. KÖNIG, B, SEEWALD, C and SCHÜGERL, K, 1981. Process engineering investigations of penicillin production, European Journal of Microbiology and Biotechnology, 12, 205–211. KRISTIANSEN, B and SINCLAIR, C G, 1979. Production of citric acid in continuous culture, Biotechnology and Bioengineering, 21, 297–315. KUBICEK, C P and RÖHR, M, 1977. Influence of manganese on enzyme synthesis and citric acid accumulation by Aspergillus niger, European Journal of Applied Microbiology, 4, 167–173. KUBICEK, C P, ZEHENTGRUBER, O, HOUSAM, E K and RÖHR, M, 1980. Regulation of citric acid production by oxygen: effect of dissolved oxygen tension on adenylate levels and respiration in Aspergillus niger, Applied Microbiology and Biotechnology, 9, 101–115. MAKAGIANSAR, H Y, AYAZI-SHAMLOU, P, THOMAS, C R and LILLY, M D, 1993. The influence of mechanical forces on the morphology and penicillin production of Penicillium chrysogenum, Bioprocess Engineering, 9, 83–90. MÄRKL, H and BRONNENMEIER, R, 1985. Mechanical stress and microbial production. In Fundamentals of Biochemical Engineering, vol 2. Ed. H BRAUER (VCH Verlagsgesellshaft, Weinheim), pp. 370–392. MARTIN, S M and WATERS, W R, 1952. Production of citric acid by submerged fermentation, Industrial and Engineering Chemistry, 44, 2229–2233. MATTEY, M, 1992. The production of organic acids, Critical Reviews in Biotechnology, 12, 87–132. METZ, B, DE BRUIJN, E W and VAN SUIJDAM, J C, 1981. Method for quantitative representation of the morphology of moulds, Biotechnology and Bioengineering, 23, 149–162. MILES, E A and TRINCI, A P J, 1983. Effect of pH and temperature on morphology of batch and chemostat cultures of Penicillium chrysogenum, Transactions of the British Mycological Society, 81, 193–200.

84

Citric Acid Biotechnology

MITARD, A and RIBA, A, 1988. Morphology and growth of Aspergillus niger ATCC 26036 cultivated at several shear rates, Biotechnology and Bioengineering, 32, 835–840. MORRISON, K B and RIGHELATO, R C, 1974. The relationship between hyphal branching, specific growth rate and colony radial growth in Penicillium chrysogenum, Journal of General Microbiology, 81, 517–520. PAPAGIANNI, M, 1995. Morphology and citric acid production of Aspergillus niger in submerged culture, PhD Thesis, University of Strathclyde, Glasgow, Scotland. PAPAGIANNI, M, MATTEY, M and KRISTIANSEN, B, 1994. Morphology and citric acid production of Aspergillus niger PM1, Biotechnology Letters, 9, 929–934. PIRT, S G and CALLOW, D S, 1959. Continuous-flow culture of the filamentous mould Penicillium chrysogenum and the control of its morphology, Nature, 184, 307–310. RÖHR, M and KUBICEK, C P, 1981. Regulatory aspects of citric acid fermentation by Aspergillus niger, Process Biochemistry, 16, 34–44. SHU, P and JOHNSON, M J, 1948. The interdependence of medium constituents in citric acid production by submerged fermentation, Journal of Bacteriology, 56, 577–585. SMITH, M G and CALAM, C T, 1980. Variations in inocula and their influence on the productivity of antibiotic fermentations, Biotechnology Letters, 2, 261–266. SMITH, J J, LILLY, M D and FOX, R I, 1990. The effect of agitation on the morphology and penicillin production of Penicillium chrysogenum, Biotechnology and Bioengineering, 35, 1011–1023. SNELL, R L and SCHWEIGER, L B, 1951. Citric acid by fermentation, British Patent No. 653 808, Chemical Abstracts, 45, 8719a. STEEL, R, MARTIN, S M and LENTZ, C P, 1955. A standard inoculum for citric acid production in submerged culture, Canadian Journal of Microbiology, 1, 150–157. SVENSKA SOCKERFABRIK, A B, 1964. A method for producing citric acid, British Patent No. 951 629, Chemical Abstracts, 60, 2304a. THOMAS, C R, 1992. Image analysis: putting the filamentous micro-organisms in the picture, Trends in Biotechnology, 10, 343–348. TUCKER, K G and THOMAS, C R, 1992. Mycelial morphology: the effect of spore inoculum level, Biotechnology Letters, 14, 1071–1074. TUCKER, K G, KELLY, T, DELGRAZIA, P and THOMAS, C R, 1992. Fully automatic measurement of mycelial morphology by image analysis, Biotechnology Progress, 8, 353–359. UJCOVA, E, FENCL, Z, MUSILCOVA, M and SEICHERT, L, 1980. Dependence of release of nucleotides from fungi on fermentor turbine speed, Biotechnology and Bioengineering, 22, 237– 241. VAN SUIJDAM, J C and METZ, B, 1981. Influence of engineering variables upon the morphology of filamentous molds, Biotechnology and Bioengineering, 23, 111–148. VAN SUIJDAM, J C, KOSSEN, N W F and PAUL, P G, 1980. An inoculum technique for the production of fungal pellets, European Journal of Applied Microbiology, 8, 353–359. VECHT-LIFSHITZ, S E, MAGDASI, S and BRAUN, S, 1990. Pellet formation and cellular aggregation in Streptomyces tendae, Biotechnology and Bioengineering, 35, 890–896. XU, D B, KUBICEK, C P and RÖHR, M, 1989. A comparison of factors influencing citric acid production by Aspergillus niger grown in submerged culture and on filter paper, Applied Microbiology and Biotechnology, 30, 444–449.

6

Redox Potential in Submerged Citric Acid Fermentation
MARIN BEROVIC

Nomenclature a,b ae aox ared k n EO2/H2O Eh Eo F N Qg pO2 pO2crit P R rH S T X
J

Y

constants electron activity activity of oxidized form activity of reduced form redox reaction balance constant number of electrons in redox reaction standard potential—1223 mV potential measured in a solution, based on standard hydrogen electrode standard redox potential of a 50 per cent reduced substance based on standard hydrogen electrode Faraday constant stirred speed volumetric gas flow rate dissolved oxygen partial pressure critical dissolved oxygen partial pressure citric acid concentration gas constant negative log of partial pressure of gaseous hydrogen sugar concentration temperature biomass concentration residence time yield factors

(-) (-) (-) (-) (-) (-) (mv) (mV) (mV)

(s-1) (vvm) (atm) (atm) (g/l) (cal/°C mol) (-) (g/l) (°C) (g/l) (s) (-)

6.1 Introduction In living organisms oxidation–reduction systems play such an intimate and essential a part, that life itself might be defined as a continuous oxidation–reduction reaction. It is not
85

86

Citric Acid Biotechnology

surprising, therefore, that theoretical speculations and experimental studies on oxidation and reduction processes in animals and plants have been actively pursued since the isolation of oxygen over 150 years ago (Hewitt, 1950). Helmholtz (1883) was the fast to describe the decolorization of litmus in a medium containing decaying protein. This was a reductive process since on passing air into solution, the original colour could be obtained again. Ehrlich (1885) injected redox dyes into living animals, killed them and investigated the redox state of the dyes in the organs. He attributed the varying state of reduction to the oxygen uptake of the organs. These dyes could therefore be used as indicators for particular reducing conditions. (Potter (1910) carried out the fast electrometric measurement of reducing conditions in bacterial cultures. He detected with a platinum electrode that the bacterial culture had a more negative potential than un-inoculated nutrient medium.) Gillespie (1920) followed the development of bacterial cultures and showed that strongly negative potentials become more positive when air is passed into the culture. Gillespie was also the first who applied the physical–chemical term ‘redox potential’ although the terms redox potential, reduction–oxidation potential, electrode potential and reduction potential were and are still used synonymously by various authors. Redox potential detectors are usually not added to standard bioreactor instrumentation for a number of reasons, most of them related to conventional thinking in bioreactor instrumentation practices. As pH measurement represents the sum of all pH influencing compounds, redox potential measurement represents the sum of all redox potential influencing compounds in fermentation broth.

6.2 Overview Redox potential is, however, a parameter that can give valuable information about metabolism taking place in various aerobic and anaerobic microbial cultures (Kjærgaard, 1977). The significance of redox potential levels for high yielding citric acid biosynthesis has been demonstrated in submerged citric acid fermentation (Berovic, 1996; Berovic and Cimerman, 1993). Although only limited attention has been paid to this phenomenon in the past, some interesting and informative research work has been presented. Some workers have advocated the use of redox potential measurements for monitoring and controlling dissolved oxygen (Shibai et al., 1975; Radjai et al., 1984). At constant pH the relation between redox potential and dissolved oxygen partial pressure can be simplified by logarithmic relation (Jacob, 1970; Memmert and Wandrey, 1987). During the last few years a great deal of the attention for redox potential measurement and it uses, has been given to anaerobic bioprocesses (Beck and Schink, 1995). The importance of redox potential measurements was referred to in articles on waste water bioprocessing, as in the case of propionate degrading Methanospirillum and Methanocorpusculum bacteria in a fluidized bed reactor, where degradation was inhibited at redox potential below –300 mV (Heppner et al., 1992), and in anaerobic digestion in methanogenic fermentation where volatile fatty acids were used as the substrate (Peck and Chynoweth, 1992). Redox potential measurements have also been found to be important in extremely thermophilic Thermotoga sp. bioprocessing, where most thermodynamic problems were associated with the relatively high redox potential (Janssen and Morgan, 1992). In various aerobic processes the importance of the redox potential has been observed. In the case of

Redox potential in submerged citric acid fermentation

87

the biochemical transformation of l-sorbose to 2-keto-l-gluconic acid by a mutant strain of Pseudomonas (Tengerdy, 1961), it was found that the redox potential indicated the oxygen demand of the culture. The importance of redox potential was also very significant in fermentations with Proteus vulgaris, Clostridium paraputrificium and Candida utilis (Jacob, 1970; Balakireva et al., 1974), Lactobacillus sanfrancisco (Stolz et al., 1993) and Lactoccocus lactis (Vonktaveesuk et al., 1994). In Acetobacterium malicum degradation of fatty acids, there were differences in redox potentials at which electrons were released during oxidative pyruvate formation (Strochaker and Schink, 1991). In acetone–butanol fermentation by Clostridium acetobutylicum, redox potential measurements were used in batch and continuous fermentation. A correlation between redox potential and switch from an acidogenic to solventogenic metabolism was reported (Penguin et al., 1994). Although the redox condition in a fermentation broth is reflected in the redox potential values measured, its characteristics cannot be generalized and the role of redox potential should be studied for each microbial process. Publications on regulation of redox potential levels are rare. In the experiments of Lengel and Nyiri (1965) and Kjærgaard (1977), on various bioprocesses, the redox potential was regulated by addition of reductants, while in Candida guilermondii fermentation by Huang and Wu (1974), the addition of n-paraffins was used.

6.3 Theory Oxidation is a process in which a substance, molecule or ion loses or gives up electrons. Reduction, on the other hand, is a process in which a substance, molecule or ion, is involved in the taking up of electrons. Whenever one substance in a system is oxidized, another substance must be reduced. The relation between reduction and oxidation may be expressed as: Reduced form Oxidized form + electron(s)

However, since free electrons never exist in any noteworthy concentration, reduction and oxidation reactions are always coupled together, so that one reaction releases just as many electrons as the other one consumes. Thus a pair of reactions always takes part in such a process. These simultaneous and complementary reduction and oxidation processes are generally known as redox reactions (Bühler and Galster, 1980). The oxidation (or reduction) capacity of a solution is characterized by the free electron activity in it. Despite the fact that the lifetime of a free electron is extremely short (10-1–10-15 seconds) there is a statistical possibility of free electron existence at the moment of transformation from electron-donor systems to electron-acceptor systems. (Balakireva et al., 1974). The thermodynamic probability of electron emergence under activated reaction-capable conditions (Inczedy, 1970) is understood as electron activity in a solution: ae = (1/k)1/n (ared/aox) (6.1) The oxidation potential is a quantitative measure of redox capacity of a solution. It is an electrical unit of charge of free energy in a redox interaction of the given system with a standard system. The system: 2H++ 2eH2 is a standard one. The oxidation potential is related to the electron activity in solution: Eh = -RT/F lnae (6.2)

88

Citric Acid Biotechnology

Eh = -RT/F ln[(1/k)1/n (ared/aox)] Eh = kRT/nF + RT/nF ln(ared/aox)

(6.3) (6.4)

In the first part of equation (6.4), kRT is equal to Eo, the standard redox potential of a 50 per cent reduced substance, based on a standard hydrogen electrode. Eh = Eo + RT/nF ln(ared/aox) (6.5)

Equation (6.5) is the well-known Nernst equation. The redox potential of the measured substance, or substrate, depending on pH, is expressed in the Kjærgaard equation (Kjærgaard, 1977): Eh = EO2/H2O + RT/4F lnaO2 - RT/4F ln2.303 pH (6.6)

The potential values measured are dependent on pH, so that in each case measurements of redox potential should be accompanied by a statement of the pH value at which they were taken. In general a pH variation of one unit causes a potential variation of 57.7 mV (Jacob, 1970).

6.4 Measurement of redox potential In principle there are two ways of measuring a redox potential: by redox dyes and by electrodes. Measurement of the redox potential by dyes is not exact and requires a number of different dyes to obtain semi-quantitative measurements; furthermore, many of these dyes may be toxic to the cells or may inhibit the enzyme activities in biological liquids (Hill, 1973). Therefore this method is not used in biochemical engineering. In bioreactors, combined sterilizable platinum as indicator and calomel or silver/silver chloride electrode as reference electrodes are employed. As electrolyte 3M KC1 solution or sometimes KClgel are used. It has been suggested that a decrease in Eh for a tenfold decrease in concentration of dissolved oxygen amounts to 14.8 mV (Ishizaki et al., 1974). Clark and Cohen (1923) introduced the concept of rH in order to eliminate pH dependence on the potential (Clark and Cohen, 1923): rH = -logaH2 (6.7)

An rH of 0 corresponds to a pO2 of 0 atm and pH = 0, and rH = 42 corresponds to a solution in which pO2 = 1 atm and pH = 0 (see Figure 6.1).

6.4.1 Calibration of redox electrodes For calibration of the redox electrodes various redox buffers are in use. In this case two saturated solutions of quinhydrone at two different pH values at 25 °C are recommended (Kjærgaard, 1977): Eh qinhydrone = 699 - 59.1 pH A relatively easier method is to use ascorbic acid at various pHs (Hewitt, 1950): Eh ascorbic acid = 375 - 60 pH (6.8) (6.9)

Redox potential in submerged citric acid fermentation

89

Figure 6.1 Electrode potential versus pH. Continuous line, theoretical curves, broken line, actual system (from Hewitt, 1950)

6.5 Significance of redox potential Redox potential in microbial cultures is caused by the existence of reversible oxido– reduction couples, irreversible reductors, and the action of free oxygen and free hydrogen (Rabotnova, 1963). It is dependent on pH value, dissolved oxygen concentration, equilibrium constant and oxido–reduction potentials in the liquid (Ishizaki et al., 1974). Mass transfer of oxygen in aerobic cultures requires a potential difference between oxygen concentration in the cell and the surrounding medium. The concentration of oxygen decreases from the solution towards the cells, and it is highly probable that the intracellular redox potential of microorganisms is always slightly more negative than the extracellular redox potential (Jacob, 1970). Several investigations have revealed that the redox potential yields more information about the oxidative status in aerobic or partially aerobic microbial cultures than concentration of dissolved oxygen (Wimpenny, 1969; Andreeva, 1964; Kjærgaard, 1976, 1977). Most commercial dissolved oxygen probes, when used in industrial conditions, are often susceptible to failure or erratic signal behaviour during the fermentation cycle, especially when dissolved oxygen is a limiting factor. In a L-leucine fermentation, the redox signal

90

Citric Acid Biotechnology

Figure 6.2 Relationship between oxygen tension and redox potential (from Shibai et al., 1975)

was useful in determining the oxygen transfer requirements when dissolved oxygen was practically zero (Shibai et al., 1974; Akashi et al., 1978). At constant pH the relation between redox potential and dissolved oxygen partial pressure can be simplified by the following equation (Shibai et al., 1975), demonstrated in Figure 6.2: logpO2 = aEh + b (6.10)

A similar relationship between pO2 and Eh has been observed in amino acid production by Corynebacterium glutamicum (Radjai et al., 1984): logpO2 = 0.0157Eh - 0.071 (6.11)

Shibai et al. (1975) carried this further for inosine production by Bacillus subtilis; pO2crit was determined by measuring the dissolved oxygen, the redox potential and cell respiration rate in pH and temperature controlled culture. When the dissolved oxygen partial pressure was above 1.10-2 atm, the redox potential had a linear relationship with the logarithm of the dissolved oxygen partial pressure. Therefore pO2 = 1.10-2 atm was estimated by determining the redox potential, on the assumption that there was a linear relationship even at the pO2 level less than 1.10-2 atm. The redox potential was markedly lowered by the physiological change in the cells, when cell respiration was inhibited at Eh = -180 mV, which corresponded to pO2 = 2.10-4 atm. pO2 in this culture was recorded as nearly zero when the cell rapidly biosynthesized the product. It went up above 1.10-2 atm at the end of fermentation, when the substrate was almost completely assimilated. The data showed that maximum production was obtained under limited oxygen supply, where cell respiration was inhibited. When cell respiration

Redox potential in submerged citric acid fermentation

91

Figure 6.3 Process parameters of high citric acid yielding fermentation

was not inhibited, as the pO2 level rose above 1.10-2 atm, the cell did not produce the maximum amount of L-leucine. The lowest values of pO2crit that have been reported were 4.10-3 atm for Saccharomyces cerevisiae, 3 × 10-4 atm for inosine and 2 × 10-4 atm for the leucine producer (Akashi et al., 1978).

6.6 Redox potential in citric acid fermentation Although citric acid production is the oldest industrial process, in addition to our own work (Berovic and Cimerman, 1982; Berovic and Roselj, 1997), there are only two other publications on redox potential measurements (Matkovicz and Kovacz, 1957; Tengerdy, 1961). In our research on submerged citric acid fermentation using beet molasses as a substrate, the relevance of redox potential levels for high product yielding biosynthesis has been demonstrated. For a high citric acid yielding fermentation there is an optimal course of the redox potential profile with two maxima of 260 and 280 mV and two minima of 180 and 80 mV of essential importance. This redox potential course has been evaluated by analysis of more than 200 fermentations (Berovic, 1996; Berovic and Cimerman, 1993). The time course for a typical batch fermentation is shown in Figure 6.3. Beet molasses contains different organic and inorganic redox couples, substances and several metal ions that could significantly influence redox potential of the whole fermentation broth. Addition of K4Fe(CN)6, a well known redox substance, to the substrate

92

Citric Acid Biotechnology

Figure 6.4 Redox potential measurements and citric acid formation by Aspergillus niger. Curve 1: aerated sterile sugar beet molasses substrate including potassium ferrocyanide; curve 2: inoculated and aerated sterile sugar beet molasses substrate excluding potassium ferrocyanide; curve 3: inoculated and aerated sterile sugar beet molasses substrate including potassium ferrocyanide

causes the formation not only of metal ion complexes, but also the Fe3+/Fe2+ redox couple, which regulates the ion balance of the substrate (Clark and Cohen, 1923). The balance of various redox couples and especially metal ion in fermentation broth is of essential importance for citric acid biosynthesis. Related to this the influence of various influent factors and substances on redox potential levels of beet molasses substrate have been studied (Figure 6.4). A reference redox potential profile was obtained for sterile medium only, with no addition of K4Fe(CN)6 (curve 1). After 24 hours of aeration, the redox potential reached a stationary phase that was unchanged until the end of the experiment. In experiments where inoculated substrate was used in absence of any addition of K4Fe(CN)6 (curve 2), and with an initial addition of this compound (curve 3), the redox potential profile exhibited a typical single peak. Only in the case where inoculated substrate with primary and secondary addition of K4Fe(CN)6 was used (results shown in Figure 6.3), was the twin

Redox potential in submerged citric acid fermentation

93

Figure 6.5 Process variables in a low yielding, abnormal citric acid fermentation

peak redox potential course observed. The different metabolic activities in the fermentation process are summarized in all the redox reactions and detected in redox potential measurement. From these experiments we concluded that only in the presence of Aspergillus niger were the relevant changes detected. Similar observations were made by Kwong and Rao (1991, 1992) in amino acid fermentation using Corynebacterium glutamicum. Redox potential measurements in citric acid fermentation might also give valuable indications of product biosynthesis in the fermentation. From evaluation of more than 200 batches, we concluded that a high yielding fermentation is directly related to levels and time course of redox potential. The yield of citric acid is reflected in the time course of the redox potential. This is shown in Figures 6.5 and 6.6, where experiments with different histories are shown. Figure 6.5 presents the characteristics of an unsuccessful fermentation with respect to

94

Citric Acid Biotechnology

Figure 6.6 Effect of temperature shift on citric acid fermentation with Aspergillus niger and suger beet molasses medium

citric acid production. Growth was diffused and the low citric acid production was therefore expected. This is also reflected in the course of the redox potential. The second peak is low and almost negligible (190 mV). The effect of temperature change is well reflected in the course of redox potential. Figure 6.6 presents the data from an experiment started at an initial temperature of 20°C. The temperature was changed after 20 hours to 30°C. The effect of this change can be clearly seen from the redox curve. In the same experiment foaming caused a loss of the substrate (89 hours), which was also indicated in a new peak of the redox potential. In a high yielding fermentation on beet molasses substrate, the redox potential course starts at a level of 0 to 20 mV, as shown in Figure 6.3. After 12 hours, the culture reaches a

Redox potential in submerged citric acid fermentation

95

level of oxygenation which significantly influences germination of conidia in the lag phase and the subsequent development of bulbous cells that appears at the first peak of the redox potential at 260 mV. After the first peak, a period of inhibition followed by the first redox minimum at 180 mV occurs. In this phase it seems that microbial activity stops. Oxygen partial pressure in fermentation broth increases and the carbon dioxide and redox potential decreases, indicating a reduced level of activity for the micro-organism. This phase is a progressive transition from glucose to fructose consumption. For this reorganization, a low redox potential level is needed, resulting in the change in morphology (Smith, 1983). After this phase, the microbial growth mode changes to spherical pellets. This was indicated by the second redox peak at 280 mV. After a decrease in redox potential to 80 mV, the second minimum, citric acid production starts. As reported by Tengerdy (1961), at the lowest redox potential level, the peak oxygen demand and initiation of rapid excretion of citric acid can be observed. The low redox potential reveals the reducing state of the complex redox system of the fermentation broth, where the respiratory enzyme system signifies strong metabolic activity. It seems that citric acid biosynthesis (Matkovicz and Kovacz, 1957), as well as some other microbiological reactions, proceeds favourably at the redox potential near the minimum of the redox curve for the particular culture involved (Hewitt, 1950; Tengerdy, 1961). This was found to be true in riboflavin fermentation. The redox potential time-course in a high citric acid yielding fermentation reaches a final level of 180 mV. Interestingly if significant amounts of oxalic acid, up to 20 mg/l, are produced, the redox potential will only reach levels of 100 to 120 mV at the end of fermentation. It has also been found that oscillations in redox potential greater than ±20 mV have a strong influence on further development of fermentation (Berovic, 1996).

6.7 Regulation of the redox potential Although measurements and observation of redox potential have been published in several articles, its regulation and process control have only rarely been discussed. In a few fermentation processes, as in Bacillus lichenoformis cultivation, a chemical method based on addition of glucose (Kjærgaard and Jørgensen, 1976, 1979) has been used. Huang and Wu (1974) added n-paraffins for regulation of redox potential in a Candida guilermondii fermentation. In a continuous process for production of xylanase by Bacillus amiloliquefaciens, a physical method based on regulation of dilution rate and agitation (Memmert and Wandrey, 1987) was used. Constant maintenance of redox potential in various bioprocesses were reported by Lengel and Nyiri (1965). Radjai et al. (1984) found that the redox potential minimum for amino acid fermentations with Corynebacterium glutamicum was directly influenced by the agitation rate. The minimum redox potential of the culture became less negative as the rate of agitation was increased. This is consistent with the increase obtained in the oxygen transfer rate and subsequently in dissolved oxygen partial pressure as agitation speed is increased.

6.8 Regulation of redox potential in citric acid fermentation The aim of using regulation of redox potential in citric acid fermentation was to establish a method for redox potential regulation that will conduct a fermentation process towards the essential redox levels needed for high citric acid production. According to this, two different methods, a chemical method, similar to those used for pH control by using

Figure 6.7 Process parameters of citric acid fermentation at chemical regulation of redox potential using 0.1% sodium sulphite as reductant and 0.1% hydrogen peroxide as oxidant

Redox potential in submerged citric acid fermentation

97

oxidants and reductants, and a physical method, based on simultaneous agitation and aeration control were tested (Berovic and Cimerman, 1993).

6.8.1 Chemical methods In the first experiment, 0.1% hydrogen peroxide was used as oxidant, and 0.1% sodium sulphite as reductant. The results of the experiment are presented in Figure 6.7. The first redox peak reached 260 mV. At this level no oxidant was added. The microbial growth form was bulbous cells. After the first redox maximum, the first minimum of 180 mV was obtained by addition of 80 ml of the sodium sulphite. After this addition microbial growth turned from bulbous to filamentous hyphae growth forms. At the second redox maximum, 280 mV, reached by addition of 20 ml of the hydrogen peroxide solution, filamentous hyphal aggregates were the dominant growth form. The second redox minimum was obtained by addition of 195 ml of sodium sulphite solution. After this the microbial form did not change. Shear in the bioreactor caused formation of hyphal fragments which were present until the end of fermentation. This period also exhibited an unchanged redox potential of 80 mV. Although this method gave a redox time-course which was similar to the time-course in high yielding fermentations, the microbial growth after addition of sodium sulphite turned completely to low citric acid producing filamentous growth forms. The addition of reductant did not stop microbial growth, but it produced an unproductive growth form. At the end of fermentation a biomass level of 5.5 g/l and 12.8 g/l citric acid were obtained. In a further experiment, a water solution of 0.1% hydrogen peroxide as oxidant and a 20% solution of glucose as reductant were used. In Figure 6.8 the results of such an experiment are presented. The first maximum of 260 mV was obtained at 25 hours. In this period, bulbous cell agglomerates appeared. Following this the regulation of redox potential levels started. The first redox minimum of 180 mV was obtained at 30 hours by addition of 275 ml of the glucose solution. The second maximum of 280 mV was reached soon after by addition of 26 ml of hydrogen peroxide. This phase was characterized by formation of small spherical pellets with short and thin peripherial hyphae. The second minimum of 200 mV appeared after the second peak soon after the regulation was stopped. This was followed by a slow, unaided drift up to a third peak of 240 mV. In this phase mycelium growth in the pellet form with thin and long peripherial hyphae appeared. The redox then fell slowly to a third minimum of 100 mV followed by a gentle increase to 180 mV at the end of fermentation. Using these agents, regulation of redox potential of the fermentation broth was possible. The redox potential time-course was close to the optimal and addition of either compound did not inhibit the development of a productive growth mode, mycelial pellets with thick and long peripherial hyphae being the typical morphology feature. At the end of the fermentation, a biomass level of 11.1 g/l and 40 g/l of citric acid were obtained (see Figure 6.8).

6.8.2 Physical methods Physical parameters such as temperature, head space pressure, agitation and aeration strongly influence the oxygen transport coefficient in the liquid phase. Therefore by lowering the temperature and increasing head space pressure, agitation and aeration, the dissolved oxygen partial pressure pO2, will increase. Increasing pO2 influences strongly the potential difference between oxygen level in the liquid phase and the oxygen level in

98

Citric Acid Biotechnology

Figure 6.8 Process parameters of citric acid fermentation at chemical regulation of redox potential using 20% glucose as reductant and 0.1% hydrogen peroxide as oxidant

the microbial cell. According to the Nernst equation (equation 6.5), increasing aox in a cell could strongly influence the metabolic and enzyme activities in micro-organisms (Harrison, 1972). In submerged citric acid fermentation with Aspergillus niger beet molasses, in addition to the maxima and minima redox levels, timing of these events is of essential importance (Berovic and Cimerman, 1982, 1993). The possibilities of regulating both redox maxima and the first redox minimum were tested, using agitation and aeration as a physical means of regulating redox potential. The first maximum of Eh = 220 mV appeared at 23 hours (Qg = 0.4 vvm speed of agitation = 400 rpm). As the level of 220 mV was too low for further process development, increasing the aeration rate Qg to 1 vvm and agitation to 600 rpm gave a redox level of 260 mV. By further reducing the aeration rate to 0.3 vvm and agitation to 200 rpm, at 30 hours the first redox minimum Eh = 180 mV was obtained. After this step, aeration was increased to 1.2 vvm and agitation to 700 rpm. The second redox maximum Eh = 280 mV at 36 hours appeared. The fermentation then proceeded at constant conditions of Qg = 1 vvm and N = 600 rpm until the end of the process. As the course of redox potential was not maintained by aeration and agitation during the last phase, it started to deviate from the optimal course with a third maximum (265 mV) occurring at 48 hours and a third minimum (120 mV) at 75 hours. This gave final biomass and citric acid concentrations of 11.4 and 68.5 g/l respectively (see Figure 6.9).

Figure 6.9 Regulation of the first redox minimum and both maxima by manipulating airflow rate and stirred speed

Figure 6.10 Regulation of both redox minima and both maxima by manipulating airflow rate and stirred speed

Redox potential in submerged citric acid fermentation Table 6.1 Stirred tank reactor dimensions

101

Finally, optimized redox level profiles were followed using simultaneous regulation of aeration and agitation during the whole course of the fermentation. This resulted in 14.7 g/ l of biomass and 95 g/l of citric acid at the end of the fermentation. The results are given in Figure 6.10. 6.9 Scale-up based on redox potential The aim of scale-up is to develop a method based on the physiological needs of the microorganism that would give high yielding and reproducible results on various scales. Scale-up is usually based on criteria such as: geometrical similarity, power input, volumetric oxygen transfer coefficient, mixing time, etc. (Nienow, 1992; Dunn et al., 1992, Dubuis et al., 1993). However, we decided on scale-up based on redox potential, being the most relevant process parameter for our process. As redox potential indicates oxygen demand of the culture, the basic idea was to use a physiological criterion of our bioprocess for scale-up. If redox potential indicates a microbial demand for oxygen, it could also reflect information on the appropriate aeration and agitation conditions needed to meet this demand. 6.9.1 Bioreactor dimensions The experiments were performed in 10 l Bioengineering AG, and 100 and 1000 l Chemap AG bioreactors. These were all equipped with Rusthon turbines, but were not geometrically similar. The reactor dimensions are given in Table 6.1. 6.9.2 Media composition The fermentation substrate consisted of diluted beet molasses with 12.5 per cent of total reducing sugars. It was treated by addition of potassium hexacyanoferrate K4[Fe(CN)6], which balanced the ratio of heavy metals ions by the formation of metal complexes (Clark et al., 1965). K4[Fe(CN)6] was added in two stages, before sterilization (primary addition) and after (secondary addition) (Cimerman et al., 1974; Berovic and Cimerman, 1982). The fermentations were carried out at T = 30°C. 6.9.3 Laboratory scale experiments Basic research for scale-up was performed in a 10 l laboratory fermentor. The best redox profile was determined from some 200 fermentations. The objective of the scale-up was to obtain a similar redox profile in the larger reactors by regulating the agitation and aeration.

102 Table 6.2 Scaling up redox profiles

Citric Acid Biotechnology

6.9.4 Pilot scale experiments Fermentations were carried out in the 100 l and 1000 l reactors; aeration and agitation were increased and decreased stepwise as outlined above to obtain the two desired redox potential maxima and minima. This was achieved as indicated in Table 6.2. The results from the experiments are summarized in Table 6.3, with similar results obtained on all scales.

6.10 Conclusions For high citric acid yielding submerged fermentation on beet molasses the optimal redox potential time course and its typical redox levels, with two maxima, 260 and 280 mV, and two minima, 180 and 80 mV, are essentially important (Berovic, 1996; Berovic and Cimerman, 1993). It is possible to influence the fermentation by changing the redox potential profile as well as the magnitude of the maxima and minima. Regulating the redox by using hydrogen peroxide as oxidant and sodium sulphite or glucose as reductant, resulted in a favourable redox profile for the whole process, but the fermentation was affected to such an extent that poor growth and reduced citric acid yields were obtained. A better method for regulating the redox potential during fermentation is through alteration in aeration and agitation. The desired redox profile is attained by respectively increasing, and decreasing the aeration and agitation to obtain the desired maximum and minimum values. It is a simple practical approach based on changing the gradient of oxygen transfer in the fermentation broth, which influences changes in intracelluar oxygen concentration and therefore the microbial physiology of the cell.

Redox potential in submerged citric acid fermentation

103

This method of regulating the redox profile was used as a scale-up criterion with the process successfully scaled up from 10 to 1000 l. Considering the results obtained, it is evident that this new scale-up method leads to very reproducible results even in geometrically non-similar bioreactors.

6.11 References
AKASHI, K, IKEDA, S, SHIBAI, H, KOBAYASHI, K and HIROSE, Y, 1978. Biotechnology and Bioengineering, 20, 27–34. ANDREEVA, E A, 1964. Mikrobiologija, 43, 780. BALAKIREVA, L M, KANTERE, V M and RABOTNOVA, I L, 1974. The redox potential in microbiological media, Biotechnology and Bioengineering Symposium, No. 4, 769–780. BECK, S O and SCHINK, B, 1995. Acetate oxidation through a modified citric acid cycle in Propionobacterium freudenreichii, Archives in Microbiology, 163, 182–187. BEROVIC, M, 1996. PhD thesis, University of Ljubljana, Ljubljana. BEROVIC, M and CIMERMAN, A, 1982. Redox potential in submerged citric acid fermentation, European Journal of Applied Microbiology, 16, 185. BEROVIC, M and CIMERMAN, A, (1993). Redox potential an effective tool for scaling-up of citric acid fermentation from laboratory to pilot scale, 3rd International Congress on Bioreactor & Bioprocess Fluid Dynamics. Ed. A NIENOW (MEP Publications), pp. 533–545. BEROVIC, M and ROSELJ, M, 1997. Possibilities of redox potential regulation in submerged citric acid fermentation on beet molasses substrate (unpublished). BÜHLER, H and GALSTER, H, 1980. Redox Measurement—Principles and Problems (Dr Ingold AG, Zurich). CIMERMAN, A, SKAFAR, S and JOHANIDES, V, 1974. YU Patent P2481/74. CLARK, W M and COHEN, B, 1923. Public Health Report Washington, 38, 666. CLARK, O S, ITO, K and TYMCHUK, P, 1965. Effects of potassium ferrocyanide addition on the chemical composition of molasses mash used in the citric acid fermentation, Biotechnology and Bioengineering, 7, 269–271. DUBUIS, B, PLÜSS, R, ROMETTE, J L, KUT, O M and BORNE, J, 1993. Physical factors affecting the design and scale-up of fluidized bed bioreactors for plant cell culture, 3rd International Congress on Bioreactor & Bioprocess Fluid Dynamics. Ed. A.NIENOW (MEP Publications), pp. 89–100. DUNN, I J, HEINZLE, E, INGHAM, J and PRENOSIL, J E, 1992. Biochemical Reaction Engineering (VCH), pp. 121–123. ERLICH, P, 1965. Recommendations 1964 of the International Union of Biochemistry (Elsevier). GILLESPIE, L J, 1920. Soil Science, 9, 199. HARRISON, D E F, 1972. Physiological effects of dissolved oxygen tension and redox potential on growing populations of micro-organisms, Journal of Applied Chemistry and Biotechnology, 22, 417–440. HELMHOLTZ, H, 1883. Archives of Anatomy and Physiology. HEWITT, L F, 1950. Oxidation–reduction potentials. In Potentials in Bacteriology and Biochemistry, 6th edition (Livingstone, Edinburgh). HEPPNER, B, ZELLNER, G and DIEKMANN, H, 1992. Start-up and operation of a propionate degrading fluidised bed reactor, Applied Microbiology and Biotechnology, 36, 810–816. HILL, R, 1973. Bioenergetics, 4, 229. HUANG, S Y and WU, C S, 1974. Redox potential in yeast cultivation broth using n-paraffins as carbon source, Journal of Fermentation Technology, 52, 818–827. INCZEDY, J, 1970. Period Polytechnic Chemical Engineering, 14, 2. ISHIZAKI, A, SNIBAI, H and HIROSE, Y, 1974. Basic aspects of electrode potential change in submerged fermentation, Agricultural and Biological Chemistry, 38, 2399. JACOB, H E, 1970. Methods in Microbiology, Vol. 2. Eds J R NORRIS and D W RIBBONS (Academic Press). JANSSEN, P H and MORGAN, H W, (1992). Heterotrophic sulphur reduction; end product inhibition, FEMS Microbiology Letters, 2, 213–218. KJÆRGAARD, L, 1976. European Journal of Applied Microbiology, 2, 215.

104

Citric Acid Biotechnology

KJÆRGAARD, L, 1977. Advances in Biochemical Engineering, 77, 131. KJÆRGAARD, L and JØRGENSEN, B B, 1976. Maintenance of a constant redox potential during fermentation by automatic addition of glucose, 5th International Fermentation Symposium, Berlin. KJÆRGAARD, L and JØRGENSEN, B B, 1979. Redox potential as state variable in fermentation systems, Biotechnology and Bioengineering Symposium, No. 9, 85–94. KWONG, S C W and RAO, G, 1991. Utility of culture redox potential for identifying metabolic state changes in amino acid fermentation, Biotechnology and Bioengineering, 22, 1034–1040. KWONG, S C W and RAO, G, 1992. Effect of reducing agents in an aerobic amino acid fermentation, Biotechnology and Bioengineering, 40, 851–857. LENGEL, Z L and NYIRI, L, 1965. An automatic aeration control system for biosynthetic processes, Biotechnology and Bioengineering, 7, 91–100. MATKOVICZ, B and KOVACZ, E, (1957). Untersuchung der zitronnensäureproduction und der potentialveränderung in oberflächen und tiefkulturen, Naturwissenschaften, 44, 447. MEMMERT, K and WANDREY, C, 1987. Proceedings 4th European Congress on Biotechnology, 3, 153. NIENOW, A E, 1992. Scale-up and scale-down of stirred tank reactors, EFB Bioreactor Engineering Course Lecture Notes. Eds M BEROVIC and T KOLOINI, pp. 209–230. PECK, M and CHYNOWETH, D P, 1992. On-line fluorescent monitoring of the metanogenic fermentation, Biotechnology and Bioengineering, 39, 1151–1160. PENGUIN, S, GOMA, G, DELORME, P and SOUCAILLE, P, 1994. Metabolic flexibility of Clostridium acetobutylicum in response to methyl viologen addition, 42, 611–616. POTTER, M C, 1910. Durham University, Philosophical Society Proceedings, 3. RABOTNOVA, I L, 1963. Die Baedeutung physikalisch-chemischer Faktoren fur die Lebenstatigkeit der Bakterien (Fischer-Verlag). RADJAI, M K, HATCH, R T and CADMAN, T W, 1984. Optimisation of amino acid production by automatic self tuning digital control of redox potential, Biotechnology and Bioengineering Symposium, No. 14, 657. SHIBAI, H, ISHIZAKI, A, KOBAYASH, K and HIROSE, Y, 1974. Studies on oxygen transfer in submerged fermentations, Agricultural and Biological Chemistry, 37, 91–97. SHIBAI, H, ISHIZAKI, A, KOBAYASH, K and HIROSE, Y, 1975. Studies on oxygen transfer in submerged fermentations, Agricultural and Biological Chemistry, 38, 2407–2410. SMITH, J E, 1983. University of Strathclyde, Glasgow, United Kingdom (private communication). STOLZ, P, BOCKER, V, VOGEL, R F and HAMMES, W P, 1993. Utilisation of maltose and glucose by Lactobacilli isolated from Sourdough, FEMS Microbiology Letters, 109, 237–242. STROCHAKER, J and SCHINK, B, 1991. Energetic aspects of malate and lactate fermentation by Acetobacterium malicum, FEMS Microbiology Letters, 90, 83–88. TENGERDY, R P, 1961. Redox potential changes in 2-keto-l-gulonic acid fermentation. Correlation between redox potential and dissolved oxygen concentration, Biotechnology and Bioengineering, 3, 255. VONKTAVEESUK, P, TONOKAWA, M and ISHIZAKI, A, 1994. Simulation of the rate of l-lactate fermentation using Lactococcocus lactis Io-1 by periodic electrodialysis, Journal of Fermentation and Bioengineering, 7, 508–512. WIMPENNY, J W T, 1969. Biotechnology and Bioengineering, 11, 623.

7

Modelling the Fermentation Process
FRANK WAYMAN, HO AI MENG AMY AND BJØRN KRISTIANSEN

7.1 Introduction This chapter will focus on the kinetic modelling of industrial citric acid production by Aspergillus niger and Candida (= Saccharomycopsis = Yarrowia) lipolytica. A good working definition of kinetic modelling is the mathematically expressed correlation between the rates and concentrations of reactants and products. When applied to appropriate mass balances, it is possible to predict the utilization of substrates and the yield of individual products. A well constructed model can be used to express the course of a whole fermentation experiment based on a small set of initial values for the fermentation variables. Such models can then be used as a basis for simulations which are essential for the optimal design and operation of a given process.

7.1.1 Different types of kinetic models Unstructured models use a single biomass component to describe the total biomass concentration in steady state conditions. They are deterministic in their approach (i.e., they are primarily used to fit a restricted set of data) and perform poorly when applied to significantly different operating conditions. Structured models are also deterministic, but are an improvement on unstructured models as they may use more than one set of equations to represent different phases of growth and production. It is also possible to represent the biomass as more than one component to describe the physiological state of the micro-organisms during a fermentation. Both types of model involve the specific growth rate (µ) as a function of the concentration of substrate (S), product (P) and biomass (X). A mechanistic model describes the conversion from substrates into products by applying the known concentrations of metabolites and properties of all the enzymes in the reaction sequence. This obviously requires a great deal of preliminary research into the physiology of the chosen micro-organism, and so far has only been possible to model simple reaction sequences with a small number of enzymatic steps. Such models are closely related to the models used for metabolic control analysis, where the activity
105

106

Citric Acid Biotechnology

of individual enzymes is related to their effect upon the overall reaction rate. This second process has been carried out for the production of citric acid through the glycolytic pathway by A. niger (Torres 1994a, 1994b) and was successful in showing that the activities of glycolytic enzymes have little influence on the rate of product formation in this system. A shortage of information currently prohibits the creation of a mechanistic model for biomass production.

7.1.2 Unstructured models based on the Monod equation and other equations The 1942 Monod equation (equation 7.1) relies on the principle that even when there are many substrates the rate of biomass production depends on the concentration of just one limiting substrate. At low concentrations of this substrate (S), µ is proportional to S, but for increasing values of S an upper value µmax for the specific growth rate is gradually reached. The maximum specific growth rate (µmax) and the saturation coefficient (Ks) must be determined experimentally. This model has been shown to correlate with fermentation data for many different micro-organisms, but fits best with well mixed unicellular systems. (7.1)

7.1.3 Other growth modelling equations Trinci (1970) measured the growth of Aspergillus nidulans colonies and pellets in terms of radius and dry weight. He showed that the growth of pellets could be described as exponential at the start of growth (equation 7.2), changing to a cube-root phase (equation 7.3) and ending in a linear phase (equation 7.4): Log of growth linear with time lnXt = lnX0 + µt Cube root of growth linear with time Growth linear with time Xt = X0 + ktt (7.2) (7.3 (7.4)

This model was used to explain that part of the pellets was either not growing, or was growing sub-optimally, because it would otherwise be expected that the pellets would continue to grow exponentially until exhaustion of the nutrients. These equations are useful when attempting to describe hyphal growth in filamentous fungi, particularly in conditions that lead to the formation of pellets. The rates of formation and depletion within the fermentation are linked to the formation of biomass by yield coefficients, and are described mathematically as follows: (7.5) (7.6)

Modelling the fermentation process

107

(7.7) where rX, rS and rN are the rates of biomass accumulation, carbon source consumption and nitrogen source consumption, respectively. The Luedeking–Piret (1959) equation (equation 7.8) is used to relate the formation of products to either the biomass concentration or the rate of biomass accumulation: rP = (a · rX) + (ß · X) (7.8)

This equation is often simplified by substituting zero for one of the product formation constants, a or ß, giving equations for growth specific and non-growth specific product formation.

7.2 Aspergillus based models Filamentous micro-organisms have growth kinetics that are quite distinct from those of unicellular micro-organisms. All cells within the mycelium may contribute towards the maintenance of internal conditions, but it has long been established that growth only occurs at the hyphal tips. In A. niger, citrate excretion is also restricted to the apical cells (Kristiansen and Sinclair, 1979). The modelling of filamentous fungi has been advanced to a stage where structured models that describe the growth, differentiation and secondary metabolite production have been developed. One example of this is the structured model developed for penicillin fermentation by Thomas and Paul (1994). However, the modelling of citric acid fermentation has yet to reach such an advanced stage.

7.2.1 A simple struct ured model for growth and citrate production in A. niger A typical growth curve for an A. niger under citric acid producing conditions has been presented by Kubicek and Röhr (1989) as shown in Figure 7.1. This shows a fast-growth phase followed by a slow-growth phase. This change in growth rate is due to a change in the physiological state of the mycelium from the normal growth form to the citrate excreting form. An examination of the kinetics of citric acid production by Aspergillus niger growing on sucrose was carried out in a pilot plant (Röhr et al., 1981). Cell growth and product formation were subdivided into several phases, each described by a simple deterministic model. The growth phases identified were the hyphal growth phase (B x), pellet growth phase (Cx), restricted growth phase (D x), transition period between trophophase and idiophase (E x) and idiophase growth (Fx). The growth in each phase was described by logarithmic, cube root and linear equations and the best fitting equation was identified by evaluating the degree of linearity within a limit of 5 per cent maximum deviation. The three phases are illustrated in Figure 7.2, where the cell growth during citric acid fermentation by Aspergillus niger B60/B3 has been plotted. Product formation was related to the growth rate by a modified Luedeking–Piret equation. However, although the same descriptions were applicable for both growth and

108

Citric Acid Biotechnology

Figure 7.1 Typical example of growth and citric acid accumulation by A. niger in submerged culture (from Kubicek and Röhr, 1989)

acid formation (Figure 7.3), the acid formation kinetics usually differentiated from the growth kinetics by a term that represented the lag time. This lag time is also known as the maturation time, when the culture has taken up all the ammonium ions but does not yet produce citric acid. The respective rate of product formation was then said to be proportional to the rate of cells entering this physiological state and was expressed as follows: rPt = krX(t–tm) (7.9)

where rP, t, X have their usual meanings and k and tm are the product formation rate constant and maturation time respectively. One problem with this equation is that k was found not to be constant but increases in value during the fermentation. It was assumed that there were at least two different types of cells within the mycelium with different productivities, and a production term for each type could be described by the Luedeking–Piret equation. Therefore, equation 7.9 was modified to become: rPt = k1rX(t–tm) + k2(X)t-tm (7.10)

where k1 is a growth-associated constant and k2 is a non-growth associated constant.

Figure 7.2 Growth during citric acid fermentation by A. niger B60/B3 (from Röhr et al., 1981).

Figure 7.3 Citric acid concentration during citric acid fermentation by A. niger B60/B3 (from Röhr et al., 1981).

Modelling the fermentation process Table 7.1 Calculated values for tm and k (from Röhr et al., 1981)

111

The constants k1 and k2 were determined by a computer-based optimization procedure following the determination of tm, and are shown in Table 7.1. Figure 7.4 shows a comparison of the experimental values for citric acid production and the line calculated by the modified equation above. It can be clearly seen that the model closely resembles the data from citric acid production (Röhr et al., 1981).

7.2.2 Expressions for mixed biomass types In the above model, the state of cells present in the fermentation was not differentiated, i.e., the description for rate of cell growth was considered applicable to describe all states of cells. Since it is known that at least two types of active cells are present, a separate expression for the rate of cell growth of different states of active cells should be more practical in the modelling of such a bioprocess. An example of a set of different rate expressions for the different types of cells present in the fermentation was established by Kristiansen and Sinclair (1979) for a single-stage ideally mixed CSTR: rXb = µbXb - ktXb - DXb rXc = µcXc + ktXb - kdXc + DXc rXd = kdXc - DXb (7.11) (7.12) (7.13)

where D is the dilution rate and subscripts b, c and d refer to basic, citric acid-producing (carbon storage) and deactivated cells respectively. The dilution rate (D) in these equations can be substituted by the overall growth rate (µ) if they are to be applied to a batch type fermentation (Sinclair et al., 1987). It was assumed that the rate constants of the above equations took the following forms: (7.14) (7.15) (7.16) where the first rate constant, µb, was the Monod expression for growth on a limiting substrate, the second constant, µc, accounted for the increase in mass of bulbous citrate producing cells and kt was the rate of transformation of basic to storage cells.

Figure 7.4 Citric acid concentration during citric acid fermentation, as determined experimentally and as calculated from biomass values and Table 7.1 (from Röhr et al., 1981).

Modelling the fermentation process

113

7.2.3 A model of the A. niger process using initial conditions The only published A. niger citric acid model to date which calculates the outputs from the initial conditions has been that written by Ho et al. (1994). This model used four different Monod type growth rate equations with different values for µmax and ks. The influence of each on the overall growth rate was altered by the calculated concentration changes of each substrate and product. A value representing available intracellular nitrogen is also calculated. The volumetric rates of all components in the system were then calculated from the different growth rates with a combination of different yield coefficients that were derived from batch data. The equations were then linked together, producing the model results that are plotted alongside the original experimental data in Figure 7.5. 7.3 Yeast based models 7.3.1 A model of the S. lipolytica process using initial conditions Klasson et al. (1991) described the citric acid process by Saccharomycopsis lipolytica NRRL Y-7576 growing on glucose using the logistic growth curve equation: rX = KX(1 - X/Xmax) (7.17) The equation has no expression for the limiting substrate, ammonia, which is not present in the medium during most of cell growth, as shown in Figure 7.6. However, the model does limit the maximum cell concentration to the initial level of ammonia with a yield coefficient. Xmax = X0 + YX/N · N0 (7.18) This simplifies the model as there is no need to calculate a value for intracellular available nitrogen. The concentration of glucose never becomes limiting because citric acid production will continue as long as it is present in the medium (Klasson et al., 1991). The formation of product was related to the physiological state of the cells. The Luedeking–Piret equation was used to model the rate of product formation of this batch fermentation and was written as follows: rP = (K · rX) + (qm · X) (7.19) where qm represents the non-growth associated constant and K denotes the growth associated constant. K was found to be negative which confirms the proposed model where production was initiated at an intermediate point in the bioprocess and that citric acid is produced mainly by resting yeast cells. The glucose substrate consumption rate was modelled generally by the following expression: -rS = (1/YX/S · rX) + (1/YP/S · rP) + mX (7.20) where the first term on the right-hand side of the equation describes the rate of substrate consumed to synthesize new cell material, the second term describes the rate of substrate consumed to synthesize excreted product and the last term describes the maintenance energy required by the cell. The model accurately predicted the levels of the broth components throughout the batch (Figure 7.7), but with a few exceptions. It did not predict the disruption of cells caused by the high shear forces present in the bioreactor. This leads to a fall in biomass concentration after growth ceases and this effect could be added to the model as a death rate term (kd). This first inaccuracy then leads to further inaccuracies in product formation and substrate removal which accumulate as the model progresses.

Figure 7.5 Comparison of A. niger model with experimental data. X, Biomass; S, Substrate; P, product (from Ho et al., 1994)

Figure 7.6 Typical batch glucose and ammonia consumption, and biomass and acid production profiles from S. lipolytica fermentation (from Klasson et al., 1989)

Figure 7.7 Comparison of S. lipolytica model with experimental data. X, (biomass); S, (substrate); P, (product); (from Klasson et al., 1991)

Modelling the fermentation process

117

Figure 7.8 Citric acid production in batch culture by Yarrowia lipolytica ATCC 20346. Concentrations of biomass (X) and citric acid (P) and weight fraction of intracellular nitrogen (ZN) as a function of the fermentation time (t). Continuous lines were calculated using the equations and yield coefficients reported (from Moresi, 1994)

7.3.2 A phase related yeast model The production of citric acid by Yarrowia lipolytica ATCC 20346 growing on glucose and the fermentation was schematically subdivided into three different phases which could be characterized by unstructured kinetic models (Moresi, 1994). The three phases were trophophase (when ammonia is removed from the medium and growth occurs), citric acid lag phase (where rapid growth from stored ammonia occurs) and idiophase (when citric acid production occurs) (Figure 7.8).

118

Citric Acid Biotechnology

Figure 7.9 Repeated batch citric acid production by Yarrowia lipolytica ATCC 20346. Concentrations of biomass (X) and citric acid (P) and weight fraction of intracellular nitrogen (ZN) as a function of the fermentation time (t) (from Moresi, 1994)

A linear type growth equation for trophophase was found to give an acceptable correlation to the experimental profiles of X, S and N. After the nitrogen source is depleted, the exhaustion does not prevent further growth but growth is accompanied by a simultaneous decrease in the intracellular nitrogen content (ZN) before the start of citric acid excretion (Figure 7.8). The cell growth rate of this citric acid lag phase was modelled by the following expression: rX = -X (d ln zN/dt) (7.21)

Modelling the fermentation process

119

The citric acid production phase or idiophase was modelled using Luedeking–Piret kinetics: rp = (YP/S/YX/S) (rX) + mPX (7.22)

However, as citric acid is formed by resting yeast cells, rX approximates to 0, so the above Luedeking–Piret equation is reduced to: rP mPX (7.23)

Confirmation of this can be seen in Figure 7.9, where the citric acid concentration increases linearly with time while the biomass concentration remains constant in a repeated-batch experiment.

7.4 Conclusion Kinetic modelling is a powerful tool in the design and optimization of all biotechnological processes, and the citric acid process is no different. The Candida lipolytica process is quite well understood, and so it is not surprising that the published models are accurate and applicable to a wide range of fermentation conditions. On the other hand, the physiological change that occurs in Aspergillus niger mycelium during the citric acid process is far from being well defined, and models which concentrate purely on the production phase are therefore less susceptible to error. The single published model which covers this transition phase is only capable of predicting the course of the fermentation within a very narrow range of starting parameters. Outside this range, errors start to accumulate after the transition phase. If Aspergillus models are to reach the same degree of predictive accuracy as the Candida models, a greater knowledge is needed of the internal metabolic changes that occur during the transition phase, when the organism appears to become stressed.

7.5 References
HO, S F, KRISTIANSEN, B and MATTEY, M, 1994. Phase-related mathematical model of the production of citric acid by Aspergillus niger, European Federation of Biotechnology International Conference on Modelling of Filamentous Fungi, Otocec, Slovenia, p. 57. KLASSON, T K, CLAUSEN, E C and GADDY, J L, 1989. Continuous fermentation for the production of citric acid from glucose, Applied Biochemistry and Biotechnology, 20, 491–509. KLASSON, T K, CLAUSEN, E C, GADDY, J L and ACKERSON, M D, 1991. Modelling lysine and citric acid production in terms of initial limiting nutrient concentrations, Journal of Biotechnology, 21, 271–282. KRISTIANSEN, B and SINCLAIR, C G, 1979. Production of citric acid in continuous culture, Biotechnology and Bioengineering, 21, 297–315. KUBICEK, C P and RÖHR, M, 1989. Citric acid fermentation, CRC Critical Reviews in Biotechnology, 4, 331–373. LUEDEKING, R and PIRET, E L, 1959. A kinetic study of the lactic acid fermentation, Journal of Biochemistry, Microbiology, Technology and Engineering, 1, 393–412. MONOD, J, 1942. Recherches sur la croissance des cultures bacteriannes (Hermann and Cie, Paris). MORESI, M, 1994. Effect of glucose concentration on citric acid production by Yarrowia lipolytica, Journal of Chemical Technology and Biotechnology, 60, 387–395. RÖHR, M, ZEHENTGRUBER, O and KUBICEK, C P, 1981. Kinetics of biomass formation and citric acid production by Aspergillus niger on pilot plant scale, Biotechnology and Bioengineering, 23, 2433–2445.

120

Citric Acid Biotechnology

SINCLAIR, C G, KRISTIANSEN, B and BU’LOCK, J D, 1987. Kinetics and Modelling (Taylor and Francis, Open University Press), p. 56. THOMAS, C R and PAUL, G C, 1994. Modelling of the penicillin fermentation, European Federation of Biotechnology International Conference on Modelling of Filamentous Fungi Abstract, Otocec, Slovenia, p. 19. TORRES, N V, 1994a. Modelling approach to control of carbohydrate metabolism during citric acid production by Aspergillus niger: I. Model definition and stability of the steady state, Biotechnology and Bioengineering, 44, 104–111. TORRES, N V, 1994b. Modelling approach to control of carbohydrate metabolism during citric acid production by Aspergillus niger: II. Sensitivity analysis, Biotechnology and Bioengineering, 44, 112–118. TRINCI, A P J, 1970. Kinetics of the growth of mycelial pellets of Aspergillus nidulans, Archiv für Mikrobiologie, 73, 353–367.

8

Mass and Energy Balance
LILIANA KRZYSTEK AND STANISKAW LEDAKOWICZ

Nomenclature C K mATP mS mO N NP concentration ATP consumption for polymerization of biomass precursors specific maintenance requirements of ATP specific maintenance requirements of substrate specific maintenance requirements of oxygen moles ATP generated per 1 C-mol of substrate by substrate level phosphorylation moles ATP generated per 1 C-mol of substrate before the biomass or product formation diverges from the catabolic pathway rate of ATP consumption rate of ATP consumption in maintenance processes rate of ATP conversion of compound j maximum true yield of biomass on ATP maximum true yield of product on ATP yield on available electrons true yield on available electrons yield of ATP on substrate (moles ATP generated per 1 C-mol of substrate from the catabolic breakdown reaction) yield of ATP on substrate (moles ATP required for maintenance) yield of oxygen on substrate yield factor for compound j on compound i (C-mol m-3) (mol ATP (C-mol DM)-1) (mol ATP (C-mol DM)-1 h-1) (C-mol (C-mol DM)-1 h-1) (mol O2 (C-mol DM)-1 h-1) (mol ATP (C-mol)-1)

(mol ATP (C-mol)-1) (mol ATP m-3 h-1) (mol ATP m-3 h-1) (C-mol m-3 h-1) (C-mol DM (mol ATP)-1) (C-mol DM (mol ATP)-1) (C-mol DM (mol e-)-1 (C-mol DM (mol e-)-1

(mol ATP (C-mol)-1) (mol ATP (C-mol)-1) (mol O2 (C-mol)-1) (C-mol (C-mol)-1)
121

122

Citric Acid Biotechnology

Y i,j YC YE YH
@ C j

Ij D

A

NP

true yield factor for compound j on compound i upper limit of Yi,j based on carbon availability upper limit of Yi,j based on energy availability upper limit of Yi,j based on reducing potential P/O ratio generalized degree of reduction of compound j weight fraction of C in compound j fraction of available electrons transferred to the biomass fraction of available electrons transferred to oxygen fraction of available electrons transferred to products

(C-mol (C-mol)-1) (-) (-) (-) (mol ATP (0.5 mol O2)-1) (-) (-) (-) (-) (-)

Subscripts O P X S max pre r oxygen product biomass substrate maximum biomass precursors real

8.1 Introduction For any bioprocess, rapid formation of an appropriate amount of biomass is of great importance, both when it is the only expected product and when the purpose is to synthesize definite chemical compounds. The simplest of the generally applied criteria of bioprocess efficiency are mass yield coefficients. Precise determination of the amount of biomass produced and substrate used is a starting point for the evaluation of all other process yields and for mass and energy balances as well as elementary balances of carbon, oxygen, nitrogen, etc. What makes the elementary balances much easier is the assumption of a constant average weight fraction of carbon in the biomass (Erickson et al., 1979). Thus, it is specified how much substrate an organism has used in the synthesis of its own mass and how much in the production of energy. Furthermore, it is also important to understand the possibilities to increase or decrease yields. This understanding is only possible by an extensive knowledge of the biochemical intracellular reactions and processes that lead to biomass or products. Such knowledge is best presented in the form of a metabolic model. The model is based on the specification of a set of intracellular chemical reactions, which are derived from the available biochemical information. This chapter deals with mass and energy balances for Aspergillus niger growth kinetics. The relations obeyed by observed and true yield coefficients resulting from balance equa tions for carbon, reduction potential and energy during intensive cell growth and citric

Mass and energy balance

123

Figure 8.1 Typical example of growth and citric acid accumulation by A. niger in submerged culture (air-life bioreactor) (reprinted from Krzystek et al. (1996) with kind permission of Elsevier Science)

acid overproduction are derived and discussed. Quantitative balances based on one macrochemical equation for checking the consistency of experimental data and evaluation of the efficiency of conversion of organic substrates by A. niger are also presented.

8.2 Metabolic description of A. niger growth The process of submerged citric acid production carried out by A. niger in sugar mineral medium is characterized by the following features (Figure 8.1): in the phase of fast mycelial growth (trophophase), formation of citric acid is observed; during the stationary phase (idiophase), product formation is maximized, but hardly any growth occurs. The substrate, being the source of carbon and energy, is converted either to biomass, CO2 and H2O or to citric acid. Citric acid accumulation results from the disruption of the tricarboxylic acid cycle, in which destruction of citric acid is blocked. NADH formed during intensive growth of A. niger is a substrate for transformations of a respiratory chain in the presence of cytochromes, with which the synthesis of ATP is coupled in the process of oxidative phosphorylation. The main pathway of electron transport during citric acid accumulation is an alternative respiratory system (Kubicek and Röhr, 1986). The functioning of this ‘alternative oxidase’ stimulates glycolysis since it permits oxidation and NADH recirculation without ATP synthesis and contributes to citric acid overproduction. A list of the most important metabolic processes accompanying citric acid accumulation with sucrose as a carbon source is shown in Table 8.1. The complex machinery of cellular processes has been arranged into fundamental reaction patterns: catabolic pathways (breakdown of substrates into energy and small molecules), anabolic pathways (synthesis of precursors for biomass), polymerization of the precursors to biomass, and the maintenance metabolism keeping the cellular machinery operative (Roels, 1983). The reactions

124

Citric Acid Biotechnology

Table 8.1 Stoichiometry of main metabolic pathways appearing in the citric acid production by A. niger on sucrose (reprinted from Krzystek et al. (1996), with kind permission of Elsevier Science)

listed in Table 8.1 such as synthesis of biomass precursors (I), production of citric acid (II), catabolism of sucrose (VII), oxidative phosphorylation (VIII), polymerization of biomass precursors (IX) and maintenance processes (X) are considered predominant. The formation of any other by-products (i.e. polyhydric alcohols (III–VI)) is assumed to be negligible, since their total concentration is at a low level in comparison to citric acid (Röhr et al., 1987). The stoichiometry determines the mass and energy balances of energy carriers (ATP, GTP) and reducing equivalents (NADH2, NADPH2, FADH2):

Mass and energy balance

125

The rate equations for substrate consumption and product formation could be derived after applying a quasi-steady state approximation (QSSA) to biomass precursors, energy carriers and reducing equivalents (Krzystek et al., 1996):

The equations obtained have the structure in which separate terms occur for substrate and oxygen consumption associated with growth, product formation and maintenance. Formally, they are identical to the most common assumption made on an a priori base, postulated by Pirt (1975):

but now the yields are related to stoichiometric coefficients resulting from metabolic reactions (Roels, 1983):

The above values were calculated taking the molecular mass MX of biomass 28.3 g DM (C-mole DM)-1 assuming an ash content of 8 per cent DM. The P:O ratio was estimated as -1 @ = 2.17 mol ATP (0.5 mol O2) (Roels, 1983; Garret and Grisham, 1995), while a mean ma x -1 value of the true biomass yield on ATP, YA T P,X, of 0.371 C-mol DM (mol ATP) , and the true max -1 citric acid yield YAT P,X = -6 C-mol citric acid (mol ATP) (Andrews, 1989; Roels, 1983). The citric acid formation pathway generates 1/6 mol ATP from 1 C-mol of sucrose (Table 8.1), and, for simple products whose synthesis does not diverge from the catabolic pathway YATP,P = . 8.3 Mass and energy balances The yield coefficients are essentially thermodynamic quantities. They result from a balance between the energy generated by the catabolic reactions and that consumed by the anabolic reactions for the production of new cell mass. The following equations can be shown to hold the mass and energy balances of ATP and reducing equivalents (Andrews, 1989):

126

Citric Acid Biotechnology

Formation of mycelium of A. niger is an energy consuming process (Table 8.1, pathways (I) and (IX)) and it implies that cell growth is energy limited. In turn, energy is generated in the reaction of citric acid formation (Table 8.1, pathway (II)), so this is clearly a carbon limited product. The yields YC, YH and YE represent their upper limits, referred to as carbon-limited, reduction-limited and energy-limited, respectively. However, the mass and energy balances on the reactions (equations (8.1)–(8.3)) give the relations (8.9)– (8.11) which are obeyed, not only by the observed yields, but also the true yields. In the case of theoretical yield all the substrate was used for production of a single product and the cell maintenance is ignored. For the biomass the theoretical yield equals the smallest of the following:

The value of YEX corresponds to the value of Yav,e = 0.11 C-mol DM (mol)-1. Similarly, the calculations of theoretical citric acid yield give: YCP = 1 YHP = 4/3 = 1.33 YEP = 1.33 and: (YSP)max = YCP = 1 C-mol citric acid (C-mol sucrose)-1. Simultaneous consideration of the mass and energy balances allows the calculation of the highest biomass and product yields to be expected in practice (real values):

Mass and energy balance

127

Figure 8.2 Possible biomass and product yields during citric acid production (reprinted from Krzystek et al. (1996) with kind permission of Elsevier Science)

The graphical interpretation is given in Figure 8.2, where the citric acid yield is shown as a function of dimensionless biomass. The highest theoretical biomass and citric acid yields correspond to the point where the energy balance line for the value of YEX/YCX = 0.5, crosses the diagonal line representing the mass balance (taking equality in relation (8.12) and ignoring cell maintenance). The maximum real yield for citric acid is 0.8 C-mol citric acid (C-mol sucrose)-1 and the corresponding maximum real yield of biomass is 0.18 Cmol DM (C-mol sucrose)-1 (i.e. 0.9 g citric acid (g sucrose)-1 and 0.18 g DM (g sucrose)1 , respectively. The yield coefficient YE (equation (8.10)) can also be determined taking into account the ATP requirement for maintenance. Calculated values for cells and product are as follows:

The formation of biomass is also an energy consuming process as well as the production of citric acid is a carbon-limited process. The calculated value of Y av,e now equals 0.16 C-mol DM (mol)-1.

128

Citric Acid Biotechnology

8.4 Kinetics of growth and citric acid production by A. niger The mass and energy equations (8.8)–(8.11) written in terms of rates are as follows:

The form of equations (8.15) and (8.16) is identical to that of (8.4, 8.5) and (8.6, 8.7) if the energy yield coefficients YE have the values of Y E (i.e. taking into account the ATP requirement for maintenance). This makes the linear growth equations represent in fact the overall energy balance where:

The energy and reduction limitation are the same in the case of formation of citric acid since YATP,P = C-mol citric acid (mol ATP)-1 and the maintenance requirement for ATP is met approximately by substrate-level phosphorylation. In addition, for carbon limited products (YSP)max = YCP = YHP, thus from (8.14) and (8.15) the following equation can be shown to hold:

Mass and energy balance

129

Citric acid is formed by A. niger during the growth phase as well as in the stationary phase, although mainly in the stationary phase when citric acid formation is maximized and hardly any growth occurs (Kubicek and Röhr, 1986). It implies that in the intensive growth phase the substrate consumption rates (sucrose and oxygen) can be described as:

On the basis of batch experimental data the yield coefficients in equations (8.37) and (8.38) were verified (Krzystek et al., 1996). The equations are as follows:

The verification has been performed using the data from submerged citric acid processes in sucrose mineral medium at the initial pH value about 2.5 carried out in a pilot plant air-lift bioreactor (Gluszcz and Michalski, 1994). An agreement between the theory and experimental results was observed, confirming the linear growth equation to be an energy

130

Citric Acid Biotechnology

Table 8.2 Yield coefficients in citric acid production by A. niger. (reprinted from Krzystek et al. (1996), with kind permission of Elsevier Science)

balance with the true yields coefficients. Experimentally obtained yield coefficients of citric acid and biomass on sucrose and oxygen are presented in Table 8.2 and compared with theoretical and true yield coefficients. The yield of citric acid on sucrose reached 83 per cent of real maximum theoretical values. Yields of biomass on sucrose and oxygen were 96 per cent and 109 per cent, respectively. On the basis of experimental data the maintenance coefficients were also estimated: mS = mO = 0.026 C-mol sucrose (mol O2) (C-mol DM)-1(h)-1. The resulting specific maintenance requirement of ATP was: mATP = 0.015 mol ATP (C-mol DM)-1(h)-1, which is of the order of common value for mATP (Solomon and Erickson, 1981).

8.5 Carbon and available electron balances Material and energy balances and their regularities based on the concept of one macroscopic equation introduced by Minkevich and Eroshin (1973) were used to evaluate sugar conversion to citric acid, mycelium, CO2 and to check the accuracy of experimental data by Röhr et al. (1983, 1987) and Nowakowska-Waszczuk and Sokolowski (1987). In submerged citric acid production by A. niger in sugar mineral medium the carbon balances may be based on the measurements of sugar consumed and biomass, citric acid and CO2 produced, according to the equations:

The energy in the organic substrate is incorporated into biomass, evolved as heat (released as a result of combustion or as electron equivalents that can be transferred to oxygen) or incorporated into extracellular products. Available electron balance may be written in the form:

Mass and energy balance

131

A balance analysis for sugar and oxygen uptake made by Röhr et al. (1983, 1987) showed that during the first phase of citric acid accumulation (up to 130 hours) more sugar is taken up than the production of biomass, CO2 and citric acid can account for. In contrast, during later phases of fermentation, more citric acid, CO2 and mycelium are formed than sugar uptake would theoretically allow. A similar pattern was also reflected in a balance for oxygen uptake, where less uptake occurs during the early phase of fermentation than needed for complete balance; the reverse was observed during the late stage of fermentation. This was caused by the intermediate accumulation and partial re-consumption of a number of polyhydric alcohols such as glycerol, arabitol, erythritol and mannitol, up to almost 9 g l-1. This finding explained earlier observations (Shu and Johnson, 1948) of accumulation of more citric acid (9.9 g l-1) during the late stages of fermentation than sugar uptake can account for (6.7 g l–1), since the polyols become degraded during the late stages of fermentation. The polyols as by-products of citric acid production account for 70 per cent of ‘lacking material’ (Röhr et al., 1983, 1987). Nowakowska-Waszczuk and Sokolowski (1987), calculating the amounts of glucose carbon utilized during the fermentation and its distribution to mycelium, citric acid and CO2, also observed that the carbon content of consumed sugar and products did not balance. During the first 24 hours of the process carried out in an air-lift bioreactor of 0.8 m3 working volume (height about 11 m) only about 76 per cent sugar carbon was found in the products (biomass, citric acid and CO2). From the second day a surplus of carbon was found in the products. When the sugar content was very low or had been completely consumed, the surplus carbon in the products was reduced to 0.9–5.3 per cent. In two different experiments the conversion of glucose carbon to citric acid, mycelium and CO2 was 76.4 and 81 per cent, 13.8 and 12.1 per cent and 13.54 and 11.8 per cent respectively. The pool of electrons transferred to oxygen in the two runs was 4.89 and 4.55, corresponding to 1.22 and 1.14 mmol O2 dm-3. The carbon content in mycelium and the degree of reduction of its carbon was accepted as 0.46 and 4.29, respectively.

8.6 Conclusion Substrate and oxygen requirements as well as biomass and product yields, which are some of the basic parameters that need to be considered in determining the feasibility of the fermentation process, may only be estimated properly if material and energy balances can be applied to the bioprocess. Available electron, ATP and carbon balances as well as the comparison of estimated values of yields and maintenance parameters can be used to test the consistency of the data in fermentations and to gain insight into the possibility of by-

132

Citric Acid Biotechnology

product formation. Material balance for carbon is in particular widely examined during the course of the process since it can be readily confirmed by actual proof on the basis of experimental data. The assumption that the micro-organisms are alike in their chemical and biochemical properties leads to formulation of a macro-chemical equation. Obviously, this is a gross simplification that only holds as a first rough approximation. However, for many biotechnological processes it is important to know accurately the yield values. This can be realized by formulating the metabolic model. A useful metabolic model should represent, in a concise way, the main biochemical properties of a specific micro-organism specified as a limited number of reactions. Such simple metabolic models give the proper formulation of the Pirt type relation linking as well as more accurate yield values and the biochemical insight how these can be manipulated. In linear relations following from restrictions out of the metabolic network the yield values are a function of the biochemical ATP and decarboxylation stoichiometry. For citric acid production by A. niger the assumption of typical mechanistic quantities such as and P:O enables the calculation of theoretical and true yields for growth and citric acid production appearing in kinetic equations based on known mechanism of the process. The analysis of the distribution of carbon and energy source for biomass growth, product synthesis and maintenance processes stresses the importance of maintenance requirements in the process. Thus, it is a useful way for process design and optimization.

8.7 References
ANDREWS, G, 1989. Estimating cell and product yields, Biotechnology and Bioengineering, 33, 256–265. ERICKSON, L E, MINKEVICH, I G and EROSHIN, V K, 1979. Utilization of mass-energy balance regularities in the analysis of continuous culture data, Biotechnology and Bioengineering, 21, 575–591. GARRET, R H and GRISHAM, Ch M, 1995. Biochemistry (Saunders). GLUSZCZ, P and MICHALSKI, H, 1994. Cultivation of A. niger in a pilot plant external loop airlift bioreactor, FEMS Microbiological Reviews, 14, 83–88. KRZYSTEK L, GLUSZCZ P and LEDAKOWICZ S, 1996. Determination of yield and maintenance coefficients by A. niger, The Chemical Engineering Journal, 62, 215–222. KUBICEK, C P and RÖHR, M, 1986. Citric acid fermentation. CRC Critical Reviews of Biotechnology, 3, 331–373. MINKEVICH, I G and EROSHIN, V K, 1973. Productivity and heat generation of fermentation under oxygen limitation, Folia Microbiologica, 18, 376–385. NOWAKOWSKA-WASZCZUK, A and SOKOLOWSKI, A, 1987. Application of carbon balance to submerged citric acid production by A. niger, Applied Microbiology and Biotechnology, 26, 363– 364. PIRT, S J, 1975. Principles of Microbe and Cell Cultivation (Blackwell). ROELS, J A, 1983. Energetics and Kinetics in Biotechnology (Elsevier). RÖHR, M, KUBICEK, C P, ZEHENTGRUBER, O and ORTHOFER, R, 1983. A balance of carbon and oxygen conversion rates during pilot plant citric acid fermentation by A. niger: identification of polyols as major by-products, International Journal of Microbiology, 1, 19–25. RÖHR, M, KUBICEK, C P, ZEHENTGRUBER, O and ORTHOFER, R, 1987. Accumulation and partial re-consumption of polyols during citric acid fermentation by A. niger, Applied Microbiology and Biotechnology, 27, 235–239. SHU, P and JOHNSON, M J, 1948. Citric acid: production by submerged fermentation by A. niger, Industrial Engineering Chemistry, 40, 1202–1204.

Mass and energy balance

133

SOBOTKA, M, MACHON, V, SEICHERT L, UJCOVA, E and MARSCHALKOVA, Z, 1985. Chemical engineering aspects of submerged production of citric acid, Folia Microbiologica, 30, 381–392. SOLOMON, B O and ERICKSON, L E, 1981. Biomass yields and maintenance requirements for growth on carbohydrates, Process Biochemistry, February–March, 44–49.

9

Downstream Processing in Citric Acid Production
PAWEL GLUSZCZ AND STANISLAW LEDAKOWICZ

9.1 Pretreatment of fermentation broth On completion of the citric acid fermentation the obtained solution contains, besides the desirable product, the mycelium and varying amounts of other impurities, e.g. mineral salts, other organic acids, proteins, etc. The method of citric acid recovery from the fermentation broth may vary depending on the technology and raw materials used for the production (Grewal and Kalra, 1995). In the surface process the fermentation fluid is drained off the trays and hot water is introduced to wash out the remaining amount of citric acid from the mycelial mats. Thorough washing at this stage is necessary, because the mycelium retains about 15 per cent of the product formed in the fermentation. After 1–1.5 hours the wash water is drained off and then added to the fermentation liquor and mycelial mats are removed from trays, disintegrated and flushed into the washing vessel using limited amounts of water. In this vessel the mycelium is heated to about 100°C by steam. The hot pulp is subsequently dewatered by pressure filtration. The solution containing 2–4 per cent of citric acid is added to the fermentation fluid, whereas the filtration cake, containing not more than 0.2 per cent of citric acid, is dried to yield a protein-rich feed-stuff. In the submerged fermentation the mycelium is far more difficult to separate from the fermentation broth. After the fermentation process is completed the mycelium containing broth is heated to a temperature of 70°C for about 15 minutes, to obtain partial coagulation of proteins, and then filtered, usually by means of the continuous filters (e.g. a rotating vacuum drum filter or a belt discharge filter). Because of the slimy consistency of mycelium forming in the submerged process, filter aids may be required. If the mycelium is to be used as a feedstuff, the filter aid must also be digestible, e.g. from cellulosic materials. If during the fermentation process oxalic acid is formed as a side product due to suboptimal control of the fermentation process, it has to be removed from the broth. This is usually achieved by increasing the pH of the fermentation fluid with the calcium hydroxide to pH = 2.7–2.9 at a temperature of 70–75°C. Calcium oxalate thus precipitated may be removed from the solution by filtration or centrifugation, and the citric acid remains in solution as the mono-calcium citrate. Oxalate removal increases the rate of filtration of the calcium
135

136

Citric Acid Biotechnology

citrate and gypsum in the subsequent steps of downstream processing and reduces the yellow hue of the citric acid solution. Recovery of citric acid from pretreated fermentation broth may be accomplished by several procedures: classical method of precipitation, solvent extraction, adsorption/ absorption on ion-exchange resins, and recently developed, more sophisticated methods such as electrodialysis, ultra- and nanofiltration or application of liquid membranes.

9.2 Precipitation The standard method of citric acid recovery has involved precipitating the insoluble tricalcium citrate by the addition of an equivalent amount of lime to the citric acid solution. Successful operation of the precipitation depends on citric acid concentration, temperature, pH and rate of lime addition. To obtain large crystals of high purity, milk of lime containing calcium oxide (180–250 kg/m3) is added gradually at a temperature of 90°C or above and pH below, but close to, 7. The concentration of citric acid in the solution should be above 15 per cent. The process of neutralization usually lasts about 120–150 minutes. The minimum loss of citric acid due to solubility of calcium citrate is 4–5 per cent. If precipitation is properly done, most impurities remain in the solution and may be removed by washing the filtered calcium citrate. Washing is performed with the smallest amount possible of hot water (approx. 10 m3 of water per tonne of acid at the temperature 90°C) until no saccharides, chlorides or coloured substances can be detected in the effluent. The calcium citrate is then filtered off and subsequently treated with concentrated sulphuric acid (60–70 per cent) to obtain citric acid and the precipitate of calcium sulphate (gypsum). After filtering off the gypsum a solution of 25–30 per cent of citric acid is obtained. The filtrate is treated with activated carbon to remove residual impurities or may be purified in ion-exchange columns. The purified solution is then concentrated in vacuum evaporators at temperature below 40°C (to avoid caramelization), crystallized, centrifuged and dried to obtain citric acid crystals. If crystallization is performed at temperatures below 36.5°C, the citric acid mono-hydrate is formed and above this transition temperature citric acid anhydrate may be obtained. The schematic flow-chart of the standard precipitation method is shown in Figure 9.1. The disadvantage of this technology is the large amount of lime required for citric acid neutralization and of sulphuric acid for calcium citrate decomposition. Moreover, it results in the formation of large amounts of liquid and solid wastes (solution after calcium citrate filtration and gypsum). For one tonne of citric acid, 579 kg of calcium hydroxide, 765 kg of sulphuric acid and 18m3 of water are consumed and approximately one tonne of waste gypsum is produced. With the aim of decreasing the amount of lime and sulphuric acid by about one third, Ayers (1957) has proposed recovery of citric acid by precipitation of di-calcium acid citrate. An additional advantage of this method is that di-calcium acid citrate has a definite crystalline structure and washes cleaner than the amorphous tri-calcium citrate. Moreover, fewer impurities are precipitated from a fermentation fluid with the di-calcium salt than with the normal salt, when the reaction mixture is completely neutralized. Di-calcium acid citrate precipitates from a citric acid solution that has been partially neutralized by the addition of calcium hydroxide, calcium oxide or calcium carbonate at an elevated temperature. It is believed that an equilibrium exists between tri-calcium citrate and citric acid on the one side, and di-calcium acid citrate on the other. At room temperature the rate of calcium hydrogen citrate formation is negligible, but if the

Downstream processing in citric acid production

137

Figure 9.1 Flowsheet of the standard precipitation method of citric acid recovery from fermentation broth

138

Citric Acid Biotechnology

temperature is elevated above 40°C the complete conversion of tri-calcium citrate mixed with aqueous solution of citric acid occurs within a reasonable length of time (about 24 hours). According to this principle a new method of citric acid recovery has been developed. The citric acid solution, obtained from the fermentation broth, is divided in two parts. The first part, about two thirds of the total volume, is completely neutralized with milk of lime, and the tri-calcium citrate is filtered off and added to the remaining part of the original citric acid solution. If the obtained mixture is heated above 40°C, a precipitate of di-calcium acid citrate will result. As an alternative method, an amount of calcium hydroxide no greater than two thirds of that required for complete neutralization may be added directly to a citric acid solution. This mixture of tri-calcium citrate and citric acid may then be converted to di-calcium acid citrate by heating above 40°C, preferably to 80–95°C (depending on the boiling point of the solution). It has been found that the results of the process may be improved, both by shortening the time and by increasing the yield, if the mixture is seeded with di-calcium acid citrate crystals (practically about 10 to 25 per cent of the expected yield). As an alternative to the classical methods of precipitation, separation and purification of citric acid from fermentation solutions, Schultz (1963) has suggested isolating the citric acid from the fermentation solution in the form of its alkali metal salts and recovery of the acid from such salts directly in one single operation. This process is based on the fact that certain alkali metal salts of citric acid crystallize from a fermentation solution after neutralization of the acid by the addition of alkaline alkali metal compounds (hydroxides, bicarbonates or carbonates) in such a manner that the mono-, di- or tri-alkali metal citrates are obtained. The impurities contained in fermentation broth influence or even inhibit crystallization of salts, so not all the theoretically possible alkali metal salts of citric acid can be produced in crystalline form according to the process. Of the sodium salts, however, all three possible salts can be recovered in the form of crystals. Before neutralization the fermentation solution may be concentrated by vacuum evaporation to a concentration of at least 40 per cent, calculated for free citric acid. After neutralizing the alkali metal salts crystallize on standing or on slowly stirring the solution; seed crystals may be added to enhance the rate of the process. Crystallization is ordinarily completed within 24 hours. Separation of the crystals from the solution is performed by the usual methods (filtration, centrifugation). After washing the crystals with a small amount of water, an almost white or slightly yellowish-brown precipitate is obtained, depending upon the type of alkali metal citrate recovered. Subsequent purification of citric acid may be performed by ion exchange on cation exchange resins or by electrodialysis. The yield of citric acid on recovering it in the form of its alkali metal salts is between 50 per cent and about 80 per cent depending on the salt used. Citric acid remaining in the fermentation broth may be recovered by the ‘classical’ method of precipitation in the form of a calcium citrate and following treatment with the sulphuric acid. According to this process considerable savings in chemicals are achieved and the amount of the spent gypsum produced is reduced. Moreover, the obtained gypsum filters more rapidly, due to the presence of alkali metal ions, than gypsum from the ‘classical’ technology, produced in the absence of the alkali metal ions. The use of purer raw materials than molasses (e.g. sucrose or glucose) in citric acid production leads to simplified methods for its recovery and purification. Crystalline or raw sugar are the best raw materials in view of the high acid yield and relatively short fermentation times attained. Crystalline sugar is also favoured by the reduced risk of infection with foreign

Downstream processing in citric acid production

139

micro-organisms due to the low initial pH value of the nutrient medium (2.5 to 3), and by the considerable reduction of the total amount of wastes and effluents. Crystalline sugar based fermentation makes it possible to use a modified, citrate-free method of citric acid recovery (Lésniak, 1989), applied in industrial practice in several citric acid manufacturing plants in Poland and the Slovak Republic. This technology consists of direct removal of impurities from the post-fermentation liquor, i.e. colloids (proteins), mycelium derived substances, coloured substances formed on heating the fermentation solution, and mineral salts introduced with the nutrient medium, substrate and water. These impurities must be removed as they interfere with the subsequent crystallization process. The first step of purification of the solution is achieved using suitably selected coagulating agents and activated carbon and then filtering off the precipitates (Adamczyk et al., 1985). Further treatment involves the removal of the remaining impurities by ultrafiltration and retention of the mineral salts using ion-exchange resins. The purified citric acid solution is concentrated, crystallized, centrifuged and dried according to the classical production process flowsheet. After the separation of citric acid crystals the supernatant liquid from the centrifuge is recycled back to the concentration section where the so-called second crop and then a third crop of crystals is obtained. The supernatant liquid obtained after removing the third crop crystals by centrifugation contains a large amount of impurities and must be purified by the classical method involving the precipitation of calcium citrate. Thus the citrate-free method can be used for purifying only up to 80 per cent of the whole amount of citric acid. This necessitates the construction of a separate process line in order to avoid plants using the above technology to manufacture merely a 50 per cent solution of citric acid, making it necessary to purify a part of the citric acid by the calcium citrate method. It is also possible to produce half of the acid amount in crystalline form and the rest in liquid form. In this case, citric acid solution purified by the citrate free method is thickened, crystallized and centrifuged to obtain the first crop. The supernatant liquid from the centrifuge (citric acid concentration of about 50 per cent) is purified by the described method so as to meet the quality standard requirements in liquid form. The advantage of this technology lies in the fact that about half of the product is obtained in crystalline form and the use of lime and sulphuric acid is eliminated as well as the formation of large amounts of effluents and solid wastes. The flowsheet of the simplified, non-citrate method of citric acid recovery is shown in Figure 9.2.

9.3 Solvent extraction An alternative method of citrate-free recovery of citric acid from a fermentation broth is extraction by means of a selective solvent which is insoluble or only sparingly soluble in the aqueous medium (Kertes and King, 1986; Hartl and Marr, 1993; Schügerl, 1994). The solvent should be chosen so as to extract the maximum amount of citric acid and the minimum amount of impurities. The citric acid can then be recovered from the extract either by distilling off the solvent or by washing the extract with the water. From the aqueous solution purified citric acid is subsequently crystallized by concentration. In the first patent concerning citric acid solvent extraction (Chemische Fabrik J A Benkiser, 1932) it has been proposed to apply n-butanol and then to wash the solution of citric acid in n-butanol with water. Since the first report a number of solvent combinations have been suggested and a great amount of information and patents have been published. In general, extraction methods may be divided into three basic groups:

140

Citric Acid Biotechnology

Figure 9.2 Flowsheet of the simplified non-citrate method of citric acid separation and purification

Downstream processing in citric acid production

141

• Extraction with organic solvents which are partly or wholly immiscible with water, such as certain aliphatic alcohols, ketones, ethers or esters (Kasprzycka Guttman et al., 1989). • Extraction with organophosphorus compounds, such as tri-n-butylphosphate (TBP) (Pagel and Schwab, 1950) and alkylsulphoxides, e.g. trioctylphosphine oxide (TOPO) (Nikitin and Egutkin, 1974; Grinstead, 1976). • Extraction with water-insoluble amines or a mixture of two or more of such amines, as a rule dissolved in a substantially water-immiscible organic solvent, and extraction with amine salts (Baniel, 1981; Bauer et al., 1988; Bizek et al., 1992; King, 1992; Prochazka et al., 1994; Juang and Huang, 1995). Each solvent used for extraction is characterized by its equilibrium distribution coefficient which is defined as the ratio of the acid concentration of the extract to the acid concentration of the aqueous phase. For low concentrations of citric acid in the raw fluid the distribution coefficient depends strongly on the type of solvent; at higher acid concentrations differences between solvents are much reduced. Extraction with organic solvents (in practice ketones and alcohols are used) may be useful in cases where the acid has a relatively high concentration in the aqueous system from which it is to be extracted. These solvents have rather low distribution coefficients (0.02–0.36), thus the extract is always more diluted than the raw liquor and multistage extraction is necessary as a rule. Moreover, solvents with relatively higher distribution coefficients (such as butanols) are too water-miscible, so they require energy-consuming steps of subsequent solvent recovery. Thus, these extraction systems are relatively inefficient for acid recovery from the dilute aqueous solutions found in most fermentation streams. Organophosphorus extractants have a significantly higher distribution ratio than carbonbonded solvents under comparable conditions, e.g. using undiluted TPB for citric acid extraction a distribution ratio of about 2 may be obtained at a 0.1 mol initial acid concentration at 25°C (Pagel and Schwab, 1950). Alkylosulphoxides have been shown to extract carboxylic acids with a distribution ratio even higher than that of TBP (Nikitin and Egutkin, 1974). The value of the distribution coefficient is influenced not only by acid concentration but also by temperature. In TBP the distribution ratio for citric acid decreases by a factor of 4 in the 0–80°C range. This property allows perfect control of the process: extraction at low temperature (10–30°C) and re-extraction with water at higher temperature (70–95°C). For the extraction by means of amines, aliphatic, araliphatic or aromatic amines, or their mixture, preferably with the average aggregate number of carbon atoms at least 20 for each amino group, may be used. These reagents have the advantage of providing a favourable coefficient of distribution of the citric acid between the aqueous and amine phases so the acid may be extracted even from highly dilute solutions. On the other hand there is a problem of decomposing the amine salt and recovering the acid and the amine separately, since the amines are too expensive to be thrown out. Usually the amine is liberated by treatment of the salt with an inorganic base (e.g. calcium hydroxide) or inorganic acid, and the salt is thus obtained instead of free citric acid. In addition to the expenditure of chemicals, this process has the disadvantage of requiring a number of processing steps. The extraction by the amine salts may be considered as a variant of the extraction with amines. In some cases the amount of acid that can be extracted with the water-immiscible amine is stoichiometrically considerably in excess of the amine present in the amine solution.

142

Citric Acid Biotechnology

The possible excess amount of extracted acid depends on several parameters, e.g. concentration of the acid in the raw liquor, the nature of the amine and its solvent. In some cases this phenomenon may be applied for extracting the acid from its concentrated aqueous solution by means of salts of amines with the same acid. From the extract the excess acid can be recovered by washing with water. The organophosphorus and aliphatic amine extractants were developed initially for the needs of inorganic extractive separation technologies. When these solvents are used for the recovery of citric acid intended for the food industry, the question concerning their toxicity should be settled. It is known that some of these compounds show teratogenic effects. On the other hand the amine extractant patented by Baniel et al. (1981) and Baniel (1982) has received approval by the US Food and Drug Administration for the use in food and drug technology (Melsom, 1987; US Food and Drug Administration, 1975). Of the great amount of patents concerning recovery of citric acid from the fermentation broth by extraction only this one has been applied in large scale production.

9.4 Adsorption, absorption and ion exchange As crystalline sugar or other pure raw materials are used more often in citric acid production, methods of its recovery and purification by adsorption and ion exchange on polymeric resins are gaining interest. One of the methods, sometimes used as a step in other noncitrate recovery technologies mentioned above, involves adsorption of contaminants onto a non-ionic resin based on polystyrene or polyacrylates and collection of the citric acid in the rejected phase. The patent literature suggests more efficient adsorption/absorption methods that make it possible to separate the citric acid from fermentation broth in a single step (Fauconnier et al., 1996). Kulprathipanja (1988, 1989), Kulprathipanja et al. (1989), Kulprathipanja and Strong (1990) and Kulprathipanja and Oroskar (1991) have proposed several methods based on a similar principle, involving polymeric adsorbents of different types. One group of such adsorbents may be neutral, non-ionogenic, macro-reticular, waterinsoluble styrene-based polymers cross-linked with di-vinylbenzene. Better selectivity and higher capacity of the adsorbent may be achieved using weakly basic anionic exchange resins, impregnated with tertiary amine or pyridine (Kulprathipanja et al., 1989; Juang and Huang, 1995; Juang and Chou, 1996), or strongly basic anionic exchange resins containing quaternary ammonium functional groups. In the simplest case the adsorbent may be applied in the form of a dense compact fixed bed which is alternatively contacted with the feed mixture and desorbent. Any of the conventional equipment employed in static bed fluid–solid contacting may be used for such a semi-continuous process. The citric acid is recovered from the adsorbent by desorption with water or dilute inorganic acid (preferably sulphuric acid of a concentration of 0.1–0.2 N). According to the patents mentioned, the complete separation of citric acid from salts and carbohydrates is achieved by adjusting the pH of the feed solution below the first ionization constant of citric acid. The pH value required to maintain adequate selectivity is inversely proportional to the concentration of citric acid in the feed mixture. Polymeric resins proposed for use in citric acid recovery are manufactured by several chemical companies and sold under different trade names, so they are commercially available. They may differ slightly in physical properties such as porosity, skeletal density, specific surface area and dipole moment. The preferred adsorbents should have a surface area of 100–1000 m2/g. The various types of polymeric adsorbents were originally designed for different chemical technologies, e.g. for decolorizing dye wastes,

Downstream processing in citric acid production

143

decolorizing pulp mill bleaching effluent or removing pesticides from waste effluent. Their effectiveness in the separation of citric acid from A. niger fermentation broth is rather unexpected. The efficiency of the ion-exchange separation process may be greatly enhanced by applying a so called simulated moving bed counter-current flow system. In this case the apparatus consists of at least two static beds, connected with appropriate valving so that the feed mixture is passed through one adsorbent bed while the desorbent material can be passed through the other. Progressive changes in the function of each ion-exchange bed simulate the counter-current movement of the adsorbent in relation to liquid flow. In such a system, the adsorption and desorption operations are continuously taking place, which allows both continuous production of an extract and a raffinate stream and the continual use of feed and desorbent streams. The simulated moving bed system applied for citric acid recovery in a pilot scale is proposed by Edlauer et al. (1990). The disadvantage of the ion-exchange method may be seen in the fact that elution of citric acid from the adsorption bed may require a large amount of desorbent, due to the tailing effect known in chromatography, causing considerable dilution of the resulting citric acid solution. The periodical regeneration of the ion-exchange resins by inorganic bases may also be a source of unwanted effluent wastes.

9.5 Liquid membranes Recently more sophisticated methods of citric acid separation with the application of liquid membranes are being developed (Basu and Sirkar, 1991; Friesen et al., 1991; Juang, 1995; Albulescu and Guzun-Stoica, 1996). Liquid membranes containing mobile carriers consist of an inert, micro-porous support impregnated with a water-immiscible, mobile ion-exchange agent. The mobile carrier, which is held in the pores of the support membrane by capillarity, acts as a shuttle, picking up ions from an aqueous solution on one side of the membrane, carrying them across the membrane and releasing them to the solution on the opposite side of the membrane (Baker et al., 1977). The flow of the complexed ion is coupled to the flow of the second ion (e.g. the hydrogen ion). This process is categorized as ‘coupled transport’, and the membranes in which it takes place are called coupled transport membranes. The coupling of the flows of the two ions permits one of the ions to be pumped ‘up-hill’ from a solution in which it is dilute to a solution in which it is more concentrated (Fyles et al., 1982). For citric acid separation by liquid membranes, the tertiary amines which give the best results in solvent extraction can also be used. In the extraction step, the basic amine reacts with hydrogen ions in the feed solution to form a tertiary alkylammonium cation. This cation then associates as an ion pair with the citrate anion to form an alkylammonium salt, which is transported across the membrane and stripped from the organic carrier solution into the aqueous product phase. This reaction regenerates the tertiary amine, which then diffuses back to the feed side of the membrane, where it recomplexes with hydrogen and citrate ions. Supported liquid membranes have not been adopted for industrial scale, primarily due to a lack of long-term stability resulting from loss of membrane by solubility, osmotic flow of water across the membrane, progressive wetting of the support pores, and pressure differential across the membrane (Danesi et al., 1987). To eliminate these problems microporous hollow fibres have been employed by Basu and Sirkar (1991). In this case the permeator consists of two sets of identical hydrophobic microporous hollow fibres. One set carries the feed solution

144

Citric Acid Biotechnology

of citric acid and the other the strip solution flowing in the lumen. The organic liquid membrane is contained in the shell side between these two sets of hollow fibres. This technique has been shown to be promising for citric acid separation even in the large scale, as the extent of citric acid recovery of up to 99 per cent was linear with the membrane area, suggesting easy scale-up. The use of liquid membranes for the recovery of citric acid from fermentation broths offers unique advantages over conventional techniques: lower energy consumption, higher separation factors in a single stage, the ability to concentrate citric acid during separation and smaller size of the complete separation apparatus. These advantages may result in a reduction in overall recovery costs and in amount of wastes.

9.6 Electrodialysis Another environmentally friendly alternative to the conventional methods of citric acid recovery may be electrodialysis. This process enables separation of salts from a solution and their simultaneous conversion into the corresponding acids and bases using electrical potential and mono- or bipolar membranes. Bipolar membranes are special ion exchange membranes which, in an electrical field, enable the splitting of water into H+ and OH- ions (Strathmann et al., 1993). By integrating bipolar membranes with anionic and cationic exchange membranes a three- or four-compartment cell may be arranged, in which electrodialytic separation of salt ions and their conversion into base and acid takes place (Voss, 1986; Sappino et al., 1996). According to Karklins et al. (1996), complete transformation of sodium tri-citrate into citric acid in a four-compartment cell may be achieved a little faster, but voltage on electrodes is higher than in a three-chamber cell. Specific electroenergy consumption of the four-compartment cell was about 40 per cent higher than that of a three-chamber apparatus. When converting organic salts, high final acid concentrations may be achieved, as opposed to mineral salts. It makes the process especially advantageous for citric acid recovery, as the evaporation step normally required can be omitted. On the other hand organic salts such as sodium citrate have a relatively large molecular weight and the solution also shows relatively low conductivity. These properties make the separation more difficult and lead to higher energy consumption, as in the case of inorganic compounds. The energy consumption (excluding pumping) for the separation of 1 kg of citric acid using bipolar membranes is in the range of 6.1 × 103 to 7.2 × 103 kWs (Novalic et al., 1995). Due to low mass transfer at low pH values it is advantageous to adjust the pH of the feed acid stream to 7.5 (Moresi and Sappino, 1996; Novalic et al., 1996). Before the fermentation solution comes to the electrodialysis some pretreatment steps are normally necessary: filtration of the broth, removal of ionogenic substances (especially Ca++ and Mg++ ions) and neutralization by means of sodium hydroxide. In the subsequent electrodialytic step the sodium citrate solution is converted into base and citric acid, which is simultaneously concentrated and for the most part purified. The produced NaOH may be reused for the neutralization (Novalic and Kulbe, 1996). Although there have been several patents published concerning recovery and purification of organic acids by electrodialysis (Gomez et al., 1991), this method is still applied only in laboratory scale and requires optimization. The economics are mainly influenced by the relatively high energy consumption, the membrane costs and the membrane life time. However, due to the wider commercial availability of bipolar membranes in the past few years and various advantages of the electrodialysis technique it is expected that

Downstream processing in citric acid production

145

Figure 9.3 Scheme of citric acid separation by means of electrodialysis with bipolar membranes (from Novalic and Kulbe, 1996)

this technology will soon be competitive with other processes (Novalic et al., 1996). Besides the elimination of environmental problems, the use of electrodialysis enables continuous separation of the citric acid from the broth during fermentation, leading to the decrease of an inhibiting influence of the product. It is also possible to apply this technique for recovery of the citric acid in continuous fermentation processes. The scheme of the proposed method (Novalic and Kulbe, 1996) for citric acid separation by means of electrodialysis with bipolar membranes is shown in Figure 9.3.

9.7 Ultrafiltration Continuous separation and concentration of citric acid may be also achieved by ultra and/ or nanofiltration. Visacky (1996) verified in a laboratory scale a two-stage membrane process for citric acid recovery from the broth obtained in A. niger cultivation on sucrose. Polysulphone membrane with cut-off 10 000 used in the first stage allowed the product to pass through to the permeate stream, while the retentate stream contained most of peptides and proteins from the broth. The rejection coefficient for the product in this step was 3 per cent, for the reducing sugars 14 per cent and for the proteins 100 per cent. Tighter nanofiltration membrane with cut-off 200 in the second stage rejected approximately 90 per cent of citric acid and 60 per cent of reducing sugars (mono-saccharides). Concentration of the product in the retentate stream was increased three times in comparison to the feed. A similar two-stage membrane technique was adapted by Bohdziewicz and Bodzek (1994) for simultaneous separation and concentration of pectinolytic enzymes and citric acid from

146

Citric Acid Biotechnology

a fermentation broth. The dilute citric acid solution obtained as a permeate in the first step of the post-fermentation fluid ultrafiltration was then concentrated up to 20 per cent using reverse osmosis. Such membrane processes may give important benefits in industrial technologies of citric acid recovery: low energy consumption, no wastes in comparison to the conventional chemical methods, possibility of use in continuous processes. However, they require practical verification and optimization in a pilot and industrial scale.

9.8 Immobilization of micro-organisms It is worth noting that some of the problems arising in the downstream processing of citric acid produced by submerged cultivation, especially in a continuous process, might be minimized by immobilization of micro-organisms in the bioreactor. In the past few years, immobilization of microbial cells has received increasing interest. The successful application of immobilized micro-organisms as living biocatalysts, involving more careful handling and often having higher production rates than free micro-organisms, has prompted a rapid development of this technique. Citric acid production by immobilized A. niger has been performed on a laboratory scale with the use of calcium alginate gel (Eikmeier and Rehm, 1984; Tsay and To, 1987), polyacrylamide gel (Gary and Sharma, 1992; Mittal et al., 1993), polyurethane foam (Lee et al., 1989; Sanroman et al., 1994; Pallares et al., 1996) and cryopolymerized acrylamide (Wang and Liu, 1996). The profitable effect of the immobilization of A. niger mycelium in view of the citric acid recovery from the fermentation broth depends on the type of the support material and process conditions. Further research is required to take full advantage of this technology, but it seems to be promising, especially in combination with other recently developing recovery techniques, such as ultrafiltration or ion-exchange.

9.9 References
ADAMCZYK, E, LESNIAK, W, PIETKIEWICZ, J, PODGÓRSKI, W, ZIOBROWSKI, J and KUTERMANKIEWICZ, M, 1985. Polish Patent 128,527. ALBULESCU, C and GUZUN-STOICA, A, 1996. Emulsion liquid membrane extraction of citric acid, Proceedings of the International Conference Advances in Citric Acid Technology, Bratislava, October, p. 32. ALTER, J E, BLUMBERG, R, 1981. US Patent 4,251,671. AYERS, R, JR, 1957. US Patent 2,810,755. BAKER, R W, TUTTLE, M E, KELLY, D J and LONSDALE, H K, 1977. Coupled-transport membranes, Journal of Membrane Science, 2, 213–221. BANIEL, A M, 1981. Eur. Patent 0049,429. BANIEL, A, 1982. US Patent 4,334,095. BANIEL, A M and GONEN, D, 1991. US Patent 4,994,609. BANIEL, A M, BLUMBERG, R and HAJDU, K, 1981. US Patent 4,275,234. BASU, R and SIRKAR, K K, 1991. Hollow fibre contained liquid membrane separation of citric acid, AIChE Journal, 37, 383–393. BASU, R and SIRKAR, K K, 1992. Citric acid extraction with microporous hollow fibres, Solvent Extraction and Ion Exchange, 10, 119–144. BAUER, U, MARR, R, RUECKL, W and SIEBENHOFER, M, 1988. Extraction of citric acid from aqueous solutions, Chemical and Biochemical Engineering, 2, 230–232. BIZEK, V, HORACEK, J, RERICHA, R and KOUSOVA, M, 1992. Amine extraction of hydroxycarboxylic acids. 1. Extraction of citric acid with l-ocyanol/n-heptane solutions of trialkalyamine, Industrial Engineering Chemistry Research, 31, 1554–1562.

Downstream processing in citric acid production

147

BOHDZIEWICZ, J and BODZEK, M, 1994. Ultrafiltration preparation of pectinolytic enzymes from citric acid fermentation broth, Process Biochemistry, 29, 99–107. CHEMISCHE FABRIK J A BENCKISER, 1932. German Patent 555,810. DANESI, P R, REICHLEY-YINGER, L and RICKERT, P G, 1987. Lifetime of supported liquid membranes: the influence of interfacial properties, chemical composition and water transport on the long-term stability of the membranes, Journal of Membrane Science, 31, 117–124. EDLAUER, R, KIRKOVITS, A E, WESTERMAYER, R and STOJAN, O, 1990. Eur. Patent 377,430. EIKMEIER, H and REHM, H J, 1984. Production of citric acid with immobilized Aspergillus niger, Applied Microbiology and Biotechnology, 20, 363–370. FAUCONNIER, N, BEE, A, ROGER, J and PONS, J N, 1996. Adsorption of gluconic and citric acids on maghemite particles in aqueous medium, Progress in Colloid & Polymer Science, 100, 212– 220. FRIESEN, D T, BABCOCK, W C, BROSE, D J and CHAMBERS, A R, 1991. Recovery of citric acid from fermentation beer using supported-liquid membranes, Journal of Membrane Science, 56, 127–141. FYLES, T M, MALIK-DIEMER, V A, MCGAVIN, CA and WHITFIELD, D M, 1982. Membrane transport systems. III. A mechanistic study of cation-proton coupled countertransport, Canadian Journal of Chemistry, 60, 2259–2266. GARY, K and SHARMA, C B, 1992. Continuous production of citric acid by immobilized whole cells of Aspergillus niger, Journal of General and Applied Microbiology, 38, 605–615. GOMEZ, O, RAMON, J, RAMON, M, LUIS, J and ZORI, D, 1991. Eur. Patent 438,369. GREWAL, H S and KALRA, K L, 1995. Fungal production of citric acid, Biotechnology Advances, 13, 209–234. GRINSTEAD, R R, 1976. US Patents 3,980,701–4. HARTL, J and MARR, J, 1993. Extraction processes for bioproduct separation, Separation Science and Technology, 28, 805–819. JUANG, R S, 1995. Recovery of citric acid from aqueous streams by supported liquid membranes containing various salts of tri-n-octylamine, presented at the AIChE Annual Meeting, Miami, paper 28f. JUANG, R S and CHANG, H L, 1995. Distribution equilibrium of citric acid between aqueous solutions and tri-n-ocytlamine-impregnated macroporous resins, Industrial Engineering Chemistry Research, 34, 1294–1301. JUANG, R S and CHOU, T C, 1996. Sorption kinetics of citric acid from aqueous solution by macroporous resins containing a tertiary amine, Journal of Chemical Engineering Japan, 29, 146–151. JUANG R S and HUANG W T, 1995, Kinetics studies of extraction of citric acid from aqueous solution with tri n-octylamine, Journal of Chemical Engineering Japan, 28, 274–281. KARKLINS, R, SKRASTINA, I and LEMBA, J, 1996. Electrodialysis method in citric acid and its salts recovery process, Proceedings of the International Conference Advances in Citric Acid Technology, Bratislava, October, p. 30. KASPRZYCKA GUTTMAN, T, JAROSZ, K, SEMENIUK, B, MYSLINSKI, A, WILCZURA, H and KURCINSKA, H, 1989. Polish Patent 160,397. KERTES, A S and KING, C J, 1986. Extraction chemistry of fermentation product carboxylic acids, Biotechnology and Bioengineering, 28, 269–282. KING, C J, 1992. Amine based system for carboxylic acids recovery, CHEMTECH, 22, 285–291. KULPRATHIPANJA, S, 1988. US Patent 4,720,579. KULPRATHIPANJA, S, 1989. US Patent 4,851,574. KULPRATHIPANJA, S and OROSKAR, A R, 1991. US Patent 5,068,419. KULPRATHIPANJA, S and STRONG, S A, 1990. US Patent 4,924,027. KULPRATHIPANJA, S, OrOSKAR, A R and PRIEGNITZ, J W, 1989. US Patent 4,851,573. LEE, Y, LEE, C W and CHANG, H N, 1989. Citric acid production by A. niger immobilised on polyurethane foam, Applied Microbiology and Biotechnology, 30, 141–143. LESNIAK, W, 1989. A modified method of citric acid production, Polish Technical Review, 5, 17–19. MILSOM, P E, 1987. Organic acids by fermentation, especially citric acid. In: R D KING and P S J CHEETHAM, eds, Food Biotechnology Vol. 1 (Elsevier), pp. 273–307. MITTAL, Y, MISHRA, I M and VARSHNEY, B S, 1993. Characterisation of metabolically active developmental stage of Aspergillus niger cells immobilized in polyacrylamide gel, BiotechnologyLetters, 15, 41–46.

148

Citric Acid Biotechnology

MORESI, M and SAPPINO, F, 1996. Effect of temperature and pH on sodium citrate recovery from aqueous solutions by electrodialysis, Proceedings of International Conference Advances in Citric Acid Technology, Bratislava, October, p. 29. NIKITIN, YU E and EGUTKIN, N L, 1974. Neftekhimiya, 14, 780–785. NOVALIC, S and KULBE, K D, 1996. Separation and concentration of citric acid by means of electrodialytic bipolar membrane technology, Proceedings of International Conference Advances in Citric Acid Technology, Bratislava, October, pp. 41–44. NOVALIC, S, JAGSCHITS, F, OKWOR, J and KULBE, K D, 1995. Behaviour of citric acid during electrodialysis, Journal of Membrane Science, 108, 201–205. NOVALIC, S, OKWOR, J, and KULBE, K D, 1996. The characteristics of citric acid separation using electrodialysis with bipolar membranes, Desalination, 105, 277–282. PAGEL, H A and SCHWAB, K D, 1950. Analytical Chemistry, 22, 1207. PALLARES, J, RODRIGUEZ, S and SANROMAN, A, 1996. Citric acid production by immobilised Aspergillus niger in a fluidised bed reactor, Biotechnology Techniques, 10, 53–57. PROCHAZKA, J, HEYBERGER, A, BIZEK, V, KOUSOVA, M and VOLAUFOVA, E, 1994. Amine extraction of hydroxy-carboxylic acids. 2. Comparison of equilibria for lactic, malic and citric acids, Industrial Engineering Chemistry Research, 33, 1565–1573. SANROMAN, A, PINTADO, J and LEMA, J M, 1994. A comparison of two techniques for the immobilisation of Aspergillus niger in polyurethane foam, Biotechnology Techniques, 8, 389– 394. SAPPINO, F, MANCINI, M and MORESI, M, 1996. Recovery of sodium citrate from aqueous solutions by electrodialysis, Italian Journal of Food Science, 8, 239. SCHÜGERL, K, 1994. Solvent Extraction in Biotechnology (Springer-Verlag). SCHULTZ, G, 1963. US Patent 3,086,928. STRATHMANN, H, RAPP, H J, BAUER, B and BELL, C H, 1993. Desalination, 90, 303–310. TSAY, S S and TO, K Y, 1987. Citric acid production using immobilized conidia of Aspergillus niger TMB 2022, Biotechnology and Bioengineering, 29, 297–304. US FOOD and DRUG ADMINISTRATION, 1975. Federal Register, 40, 49080–49082. VISACKY, V, 1996. Membrane nanofiltration for citric acid isolation, Proceedings of the International Conference Advances in Citric Acid Technology, Bratislava, October, p. 31. VOSS, H, 1986. Deacidification of citric acid solutions by electrodialysis, Journal of Membrane Science, 27, 165–172. WANG, J L and LIU, P, 1996. Comparison of citric acid production by Aspergillus niger immobilised in gels and cryogels of polyacrylamide, Journal of Industrial Microbiology, 16, 351–353.

10

Fermentation Substrates
WLADYSLAW LESNIAK

10.1 Introduction Fermentation industries have an advantage over some other manufacturing industries in that their raw materials can be altered, within limits, allowing some buffering against increasing world prices. However, the past 20 years have seen global changes in the prices of all raw materials and consequently all fermentation substrates have suffered increases to varying extents. For processes where different substrates can be used, or both synthetic and biological production routes exist, process economics is of paramount importance for survival. For processes where the product is only obtainable through fermentation, profit margins can be sustained by passing the price increases resulting from substrate cost increases on to the consumer. Production of bulk products such as citric acid and antibiotics are obvious examples. These products therefore may have had less pressure on them than the others to search for the cheapest possible substrate, but even here there is competition between rival companies and ways to lower costs and increase profits are thus continually being sought. The choice of substrate is therefore always under review (Ratledge, 1977). There is always pressure to find a cheaper or better substrate, but the new substrate may present storage problems, may be difficult to sterilize or have an unwanted variability in composition. Increased productivity is not the only yardstick to be used. The substrate may have a residue which poses product recovery and purification problems. The cheapest substrate is therefore not often the best. In addition to these problems, any change in substrate or amendment to the formulation of the medium will influence the characteristic of the fermentation process, and has to be carefully evaluated. A substrate must be readily available throughout most of the year. Seasonably produced crops from which process wastes are used as fermentation feedstock are not suitable if the harvest period is short and the material to be used is subject to contamination and spoilage. Thus the industry must have substrates that are relatively stable and can be stored reasonably easily for more than half a year. A process, for example, citric acid production, can be changed to accommodate a new substrate. The advent of cheap hydrocarbons in the 1960s led to many companies switching
149

150

Citric Acid Biotechnology

Table 10.1 Relative carbon contents of fermentation substrates (from Ratledge, 1977)

over to this substrate. Aspergillus niger, the traditional producer, cannot grow on alkanes, but a variety of yeast can and some will accumulate citric acid sufficiently enough for industrial processes to be established (Shennan and Levi, 1974). The price of the substrate is crucial. However, it is important to take into consideration the amount of available carbon. This differs according to the type of substrate being used (see Table 10.1). This suggests that if the choice of substrate is not limited, a carbohydrate could be replaced by alkanes with no loss in process productivity (an important optimization parameter for the citric acid process). However, others factors have to be taken into consideration before this is accepted—increased aeration or agitation rates may be necessary with alkanes (being a more reduced substrate) and this factor must be met by the savings from the change of substrate. Transport costs for substrates from the collection or production point to the fermentation plant have to be considered. These costs may become significant if too much water is present and will mitigate against the use of some waste materials at sites removed from their point of production. One substrate may be more attractive to use than another simply because it poses fewer problems in the processes both before and after the fermentation. Fermentation media for citric acid biosynthesis should consist of substrates necessary for growth of the producer micro-organism and its citric acid biosynthesis, primarily the carbon, nitrogen, phosphorus and microelements sources. Moreover process water and air can be included as fermentation substrates. The basic substrate for citric acid fermentation in plants using the surface method of fermentation is beet or cane molasses. Plants using submerged fermentation can use not only beet or cane molasses, but a substrate of higher purity such as hydrolysed starch, technical and pure glucose, refined or raw sugar, purified and condensed beet or cane juice. This is because use of a pure substrate may result in increases in yield, or reduction in fermentation time.

10.2 Molasses Molasses is a widely used substrate, coming in a variety of qualities. High quality molasses is usually demanded for citric acid production while poorer quality molasses is used mainly in the production of low value products such as alcohol, where the producer micro-organism has a much greater tolerance to impurities in the medium.

Fermentation substrates

151

The composition of cane and beet molasses has recently been compared and the uses of molasses as a fermentation feedstock have been discussed elsewhere (Hastings, 1971). Cane and beet molasses are not identical in composition; often one type will be preferred to the other. They are sometimes mixed to take advantage of the additional nutrients arising from the differences in composition. Besides substrate type (sugar beet, sugar cane), the chemical composition of molasses depends on many factors such as soil and climate conditions, fertilization type, crop method, time and conditions of storage, production technology, technical equipment of plant, etc.

10.2.1 Beet molasses Beet molasses consists of about 65–80 per cent dry substance and 20–25 per cent water. The main ingredient of molasses is sucrose, 44–54 per cent by weight. Other sugars (carbohydrates) which can be found in higher amounts are inverted sugar 0.4–1.5 per cent, raffinose 0.5–2.0 per cent and kestose and neokestose 0.6–1.6 per cent. Raffinose is a natural part of sugar beet, while kestose is the result of microbial action during sugar beet treatment. Other sugars in molasses are arabinose, xylose and mannose in amounts of 0.5–1.5 per cent. All sugars (except sucrose) are included in the non-nitrogen organic substances of molasses. Products of chemical and thermal sugar decomposition (melanoidines, caramel) and organic acids also belong to this group. Caramel consists of sugar anhydride and colouring matters; melanoidines are made in hot solution as the result of a reaction between reducing sugars and amino acids. In addition to the non-volatile dark coloured compounds, there are about 40 volatile compounds as aliphatic aldehyde, methylglyoxal, diacetyl, acetoin, acetone, oxymethylfurfurol and others. The non-volatile organic acids present in molasses are glutaric, malonic, succinic, aconitic, malic and lactic acid; the remainder are oxalic, citric and tartaric acid. These can all react with calcium to form insoluble salts that can influence the precipitation and recovery of the citric acid crystals. Molasses contain such volatile acids as formic, acetic, propionic, butyric and valeric acid. Almost all organic acids, volatile and non-volatile, are potassium or calcium salts. The colour of molasses ranges between 1.2 and 4.6 cm3 of 0.1 N iodine solution (to which should be added 94 cm3 of water to get the colour identical to that of 2 per cent molasses solution). Molasses containing higher amounts (over 1 per cent) of volatile acids are normally too dark to be used as feedstock for the citric acid fermentation, though the exact relationship between content of these substances and fermentation yield has not been established. Other ingredients of molasses that have a negative influence on colour and thus fermentation yield are colloidal substances. Beet molasses contains about 4–6 per cent of colloids, whose chemical constitution has only recently been documented. Mostly, they are high-molecular coloured complexes. Some of these colloids (of negative potential) can be removed from solution by acid coagulation (pH 3.2, molasses dilution 20–30 per cent, temperature 80°C) and colloids of positive potential by alkaline coagulation (pH over 8.0). Nitrogen compounds contained in molasses are mostly betaine (about 60–70 per cent of total nitrogen), amino acids (20–30 per cent of nitrogen), protein (3–4 per cent of nitrogen) and traces of nitrogen in ammonium nitrate and amide. Betaine comes from beet and is not used by micro-organisms as a nitrogen source. It is not known to influence

152

Citric Acid Biotechnology

Table 10.2 Amino acid content of beet molasses (from Smirnow, 1983)

Table 10.3 Content of microelements in beet molasses (from Smirnow, 1983)

the fermentation. The amino acids content in molasses depends on the soil and climate conditions and beet cultivation. Amino acid content of beet molasses is shown in Table 10.2. The content of mineral substances in beet molasses amounts to 8.5–14.0 per cent. The main ingredient of the mineral ash is K2O (60–70 per cent of the total), CaO (4.5–7.0 per cent) and MgO (about 1 per cent). The level of P2O5 in ash is normally very low (0.2–0.6 per cent), because over 90 per cent of phosphorus contained in beet is removed in the sugar extraction process. If the method of juice alkalization by Na3PO4 (pH 8.3–8.5) is used in the sugar production, the contents of P2O5 in molasses ash can reach 1.2–2.0 per cent. There are also many other elements, so-called microelements, which have a great effect on the citric acid fermentation process. The amount of particular microelements in different molasses can range widely as indicated in Table 10.3. Another important

Fermentation substrates Table 10.4 Content of vitamins in beet molasses (from Smirnow, 1983)

153

ingredient of molasses is vitamins, especially those that are known to stimulate microbial activity. The content of vitamins (mg/100 g) in beet and cane molasses is shown in Tables 10.4 and 10.5 respectively. The pH of molasses depends on the sugar extraction technology. It was considered that a neutral, or slightly alkaline molasses gave the best citric acid yields. However, a fermentation technology to tolerate the slightly acidic molasses produced in modern refineries has been developed. Today, it is considered that for citric acid fermentation the buffering capacity of the medium is more important than the pH value of the molasses. It is defined as the amount of 1N solution of sulphuric acid (in cm3) used to reduce pH from 5.0 to 3.0 in 100 g molasses solution diluted in 1:1 ratio with water and acidified to pH 5.0. The buffer capacity of beet molasses usually ranges from 60 to 95 cm3. Citric acid production needs molasses with low buffer ability, to make possible the required rapid fall of medium pH during fermentation.

10.2.2 Cane molasses Cane molasses differs from beet molasses in its chemical composition. It contains less sucrose and more inverted sugar, has lower content of nitrogen and raffinose, more intensive colour and lower buffer capacity. Cane molasses of raw sugar conversion also differs from beet molasses and even from blackstrap cane molasses. The composition of cane molasses is shown in Table 10.6.

154 Table 10.6 Composition of cane molasses (from Smirnow, 1983)

Citric Acid Biotechnology

Table 10.7 Alternative analysis of cane molasses sample

Beet and cane molasses can also contain other substances which appear in small amounts, but are often crucial in deciding whether the molasses are suitable for use in citric acid biosynthesis. These are pesticides, fungicides and herbicides used in beet and cane cultivation and also substances used for defoaming in sugar production process. All have mostly toxic properties and negatively affect molasses usability. It is considered that the best molasses for citric acid fermentation can be, as a rule of thumb, characterized as shown in Table 10.7. According to all cited requirements, beet molasses is more suitable for citric acid fermentation than cane molasses. It is especially relevant in submerged fermentation where the quality of the substrate is more important for productivity and fermentation yield. The microflora of molasses can be an agent of negative influence on yield and productivity of fermentation. Molasses will always contain a certain number and type of micro-organisms, sometimes the count can be higher than 10 000 per g of molasses. The most common micro-organism in molasses is sporulating rods of Bacillus species (over 90 per cent of total molasses microflora), bacteria producing acids and gases (E. coli, Pseudomonas and others), heterofermentative lactic acid bacteria ( Leuconostoc mesenteroides), sometimes yeasts of Candida species, and very rarely, moulds of Penicillium, Aspergillus and other species.

Fermentation substrates

155

Bacteria of Bacillus species appear in molasses because their spores are present in beet and are unaffected by high temperatures, even 125°C (Bacillus subtilis). They are destructive because some of them (B. megaterium, B. mesentericus) are able to reduce nitrates to nitrites. Strains of Aspergillus niger can be very sensitive to nitrites (a NO2 concentration in medium of 0.05 per cent will retard growth and cut the citric acid production by 50 per cent). The greatest antagonists of Aspergillus niger among non-sporulating bacteria are E. coli and Pseudomonas. They grow very quickly in many media over a wide temperature range, decomposing sugar in solution to unwanted acids, alcohol and gases, and are able to reduce nitrates to nitrites. Bacteria of Leuconostoc species convert sucrose to dextran. They also produce unwanted volatile acids such as formic, acetic and propionic acid. Yeasts of Candida species can propagate over a wide range of temperature (5–55°C) and pH value of medium (2–8). They can be very undesirable to Aspergillus niger strains, especially in submerged fermentation, where they can stop citric acid biosynthesis.

10.2.3 Treatment of molasses for citric acid production Due to the varying chemical composition of molasses it is always required to evaluate any new delivery in a scaled down version of the citric acid production vessels. Even very good molasses is no guarantee for high yields of citric acid biosynthesis without special pretreatment. The basic operation in molasses preparation is a treatment for heavy metal ions removal. Potassium ferrocyanide or other complex compounds are commonly used. Potassium ferrocyanide reacts with many heavy metals, mostly causing their precipitation. It was noted that for 21 microelements found in molasses, potassium ferrocyanide reacts with 18 of them (Leopold and Valtr, 1964). Potassium ferrocyanide removes not only metals of negative influence but also some of the microelements necessary for mycelium growth. Therefore its addition to molasses has to be strictly regulated. The optimum amount of ferrocyanide depends on molasses type and ranges from 200 to 1000 mg/dm3 of medium (about 300 g of molasses); of the metals 80–85 per cent of the total is complexed as precipitate, 7–14 per cent is complexed in solution and 7–10 per cent is in elemental free state. At the optimum dose level of ferrocyanide, the part in elemental state is usually constant and ranges between 50 and 100 mg/dm3, depending on strain and fermentation type. This has been used to develop a quick method of optimal ferrocyanide dosage in molasses media. (Lesniak, 1976). Ferrocyanide is normally added before sterilization. However it can also be partially added before and after sterilization or the total amount can be added after sterilization. Another compound complexing with heavy metals is the sodium salt of ethylenediamineacetic acid (EDTA). This compound reacts with metals of I and II valency at pH 7.0, with metals of III valency at pH 3–5 and with multivalency metals at pH 1. Ca and Mg ions give Trilon B soluble salts and they are not removed from solution. Other heavy metal complexing compounds can also be used, e.g. sodium polyphosphates, potassium rhodanate, 2,4-dinitrophenols and 8-oxyquinoline. Molasses media are sometimes purified by ionites, especially on cation exchanger. Not all microelements should be removed during this process, as some of them are necessary for growth of the Aspergillus niger mycelium. To protect the fermentation process from unwanted micro-organisms, the molasses must be sterilized. The most economical method is steam sterilization. For sporulating bacteria a temperature of 130°C or above for 30 minutes is recommended. However, steam sterilization

156

Citric Acid Biotechnology

of the medium may not be sufficient to ensure total sterility because some micro-organisms can enter the fermentation broth via addition ports or from the air. Because of this, other sterilizing agents such as formaline (at 0.006–0.01 per cent) (in particular for the surface fermentation) and furan derivatives are used. Sulphamide preparations do not totally destroy the bacteria, but antibiotics, though they do not have any negative influences, are too expensive (Karklinsh and Probok, 1972). Applying chemical sterilizing agents enables softening of sharp thermal sterilization conditions that have a negative effect on molasses quality. Other methods of sterilization tested are UV and gamma radiation, ultrasound, and ultrafiltration. They are not used in practice as they are cost-prohibitive compared with steam sterilization. In tropical countries where date production is considerable, date syrup is a major product. The chemical composition of this material differs from that of sugar beet molasses, but when mixed with an equal volume of beet molasses it gives the same yield of citric acid as for beet molasses based on the amount of sugar converted (Shadafza et al., 1976). Molasses from the starch industry (hydrol molasses) is also widely used in citric acid fermentation.

10.3 Refined or raw sucrose Refined sugar of beet or cane is almost pure sucrose which Aspergillus niger strains ferment very well (Lesniak, 1989). This sugar is a very good substrate for the submerged fermentation because in surface fermentation, the rate of diffusion of acid in sugar solutions is too low. Preparation of a refined sugar solution as a fermentation medium is based on its diluting with water to a concentration of 15–22 per cent, adding necessary nutrients (NH4NO3, KH2PO4, MgSO4) and acidifying with hydrochloric or sulphuric acid to pH 2.6–3.0 (Lesniak, 1972). Normally the batch medium is sterilized in the fermentation vessel. In this case, all the ingredients of the fermentation medium are added straight into the bioreactor or are prepared separately by diluting in hot water (85–95°C) and then pumped into the bioreactor. In this case, sugar is diluted to 50–60 per cent concentration and pumped into the fermenter that has had an exact amount of sterile water added, resulting in a total sugar concentration of 15–22 per cent. Sterilization in the fermenter lasts about 0.5–1 hour at 110–120°C. The solution is then cooled to 32–35°C with continuous stirring and aeration before the inoculum of Aspergillus niger spores or mycelium is added. The use of continuous sterilizers, where the sugar solution is sterilized separately from the other ingredients, is becoming more common.

10.4 Syrups Syrups of beet or cane sugar can also be used as basic substrate for the submerged citric acid fermentation. The great advantage with this substrate is its purity; however, the quality of the syrups deteriorates rapidly during storage. Because of this they can only be used during the sugar campaign season and only if the citric acid plant is not too far from the sugar factory because of the large transport costs. Preparation of the syrups for fermentation entails dilution with water to a sugar concentration of 15–20 per cent, addition of necessary nutrients (NH4NO3, KH2PO4, MgSO4, (NH4)2C2O4), acidification with hydrochloric or sulphuric acid to pH 4–5 and sterilization at 121°C for 0.5–1 hours (Kutermankiewicz et al., 1980).

Fermentation substrates

157

10.5 Starch Starch can be an attractive feed stock for many fermentation processes. It can be used directly by many micro-organisms and is frequently incorporated into fermentation media as a partial ingredient. Starch is widely used as the principal substrate for the production of amylases and amyloses in the food and brewing industries. The production of citric acid from sources of starch such as corn, wheat, tapioca and potato is widely used. The suitability of these substrates for citric acid fermentation depends on their purity and method of hydrolysis. Acid hydrolysis, enzymatic hydrolysis, or a combination of the two, are used. Preparation of starch substrates for fermentation is based on their enzymatic liquefaction and saccharification to a defined hydrolysis level. Additional nutrients are added, depending on which starch is used. The pH is adjusted to 3–4 using hydrochloric or sulphuric acid and the medium is sterilized at 121°C for 0.5–1 hour. Good citric acid yields have been obtained using pure starch (potatoes, wheat or maize), hydrolysed only to 10–15 DE with a-amylase (Bolach et al., 1985). This was possible, as the applied Aspergillus niger strain had the ability to produce its own amylolytic enzymes which helped in the saccharification of the starch to available sugars. Dextrose syrup, obtained by enzymatic hydrolysis of starch, is now employed as a basic substrate for citric acid biosynthesis in laboratory and industrial scale. In this case it is especially important to restrict the amount of heavy metals below critical levels; heavy metals should therefore be removed by ion exchange. When using an Aspergillus niger strain resistant to higher concentrations of heavy metals, practically the same yield may be obtained on decationized and non-decationized dextrose syrup (Pietkiewicz et al., 1996).

10.6 Hydrol This is a paramolasses obtained as a by-product during crystalline glucose production from starch. Because of the high glucose content (40–45 per cent) and high purity coefficient it is a very good substrate for citric acid production (Lesniak et al., 1986). Preparation of hydrol for fermentation involves dilution to a sugar concentration of 15– 18 per cent, addition of necessary nutrients and adjustment of pH with hydrochloric or sulphuric acid to 3.0–4.0. The solution is sterilized at 121°C for 0.5 hour and cooled to 32–35°C.

10.7 Alkanes The low price of alkanes, coupled with the ability of many organisms to utilize them, produced major changes in the fermentation industry during the 1960s and 1970s. Citric acid production, using Candida lipolytica, is a typical example and has been the subject of many patents (Maldonado and Charpentier, 1975; Kimura and Nakanishi, 1985). However, there are few industrial citric acid processes that are based on alkanes. There are two main reasons for this. Firstly, in these processes isocitric acid would also be produced at concentrations that would cause product recovery problems, as well as reduced citric acid yields (Wojtatowicz and Sobieszczanski, 1981). Secondly, a fourfold increase in price since 1973 no longer makes alkanes a cheap substrate.

158

Citric Acid Biotechnology

10.8 Oils and fats Oils and fats are also being increasingly used as substrates in many fermentations. The oils should be liquid at the temperature of fermentation; the concentration of the oils may be up to 10 per cent but there is no reason to believe that concentrations up to 30 per cent may not be used. The prices of oils and fats vary according to their fatty acid composition, and often are very cheap. The price of the cheapest oils is such that, because of their high carbon content, they are not much more expensive that raw sugar. For citric acid production, oils are now being used as principal carbon source in a manner analogous to the previous use of alkanes. With palm oil as carbon source, a yield of citric acid of 145 per cent using a mutant of Candida lipolytica has been reported (Ikeno et al., 1975). There are examples of oil being added in small concentrations to Aspergillus niger fermentation (Gutcho, 1973) and even being used as a sole carbon source for Aspergillus niger fermentation. It was found that citric acid could be produced on these substrates with good yield. In particular with an initial 8 per cent concentration of vegetable oil, a yield of 104 per cent was obtained (Elimer, 1994). These oils and fats may replace alkanes in several fermentations, but it is unlikely that they will remain at their current low prices.

10.9 Cellulose Cellulose is the major renewable form of carbohydrate in the world: about 1011 tonnes are synthesized annually and much of this is waste. To use it as fermentation feedstock, it must be first hydrolyzed to starch and then to sugar, either chemically or by cellulases. The technology and economics of these processes are constantly being improved, but it is still not apparent when the production of sugar syrups by this route is going to become profitable. In the long term, cellulose could become a major resource of the fermentation industry in general, including citric acid fermentation.

10.10 Other medium ingredients 10.10.1 Other nutrients Other substances are used as sources of nitrogen, phosphorus and micro and macroelements. Organic compounds (ammonia, amino acids) or non-organic compounds (ammonia salt, nitrates) can be used as nitrogen source. The most commonly use phosphorus source is phosphoric acid or its salts. Whenever high purity carbon substrates (refined sugar and starch) are used, ammonium nitrate or ammonium sulphate will be used as nitrogen source and monopotassium phosphate as phosphorus source (Lesniak and Kutermankiewicz, 1990). When using molasses, additional nitrogen is rarely required, as it will contain sufficient amounts of organic and inorganic nitrogen compounds to support the metabolic growth process. If the nitrogen level becomes too high, some of the sugar is converted into production of excess biomass and not citric acid. The most important microelements are magnesium, sulphur, zinc, iron, copper and manganese. They are very seldom added to the medium. In complex media the level of trace metals will normally be too high, and the main concern is simply to remove them. This

Fermentation substrates

159

is very different from academic research into citric acid fermentation. Here, a refined sugar is invariably used as the carbon source and much work has been done on the level of nutrients, in particular trace metals required for optimal acid yields and the role of individual metal ions.

10.10.2 Water Water used for diluting basic substrates should be at least of drinking water quality. There should not be organic compounds and products of their decomposition (NH3, and H2S) and the level of trace metals must be controlled. All the water must be sterilized to remove contaminating micro-organisms.

10.11 Conclusion Citric acid is a bulk product, with the substrate cost being a major part of the plant operating cost. In terms of bulk, the carbon source is the most important substrate. The efficiency of its conversion to citric acid will determine the profitability of the fermentation process. For this reason, the carbon source is also the most important substrate for process economics. This chapter has, therefore, concentrated on the various forms of carbon sources used. Most processes are based on molasses, although the use of cleaner sources is gaining ground. Whatever the source, its cost and preparation in order to permit optimal fermentation conditions are two important aspects of the technology in citric acid production.

10.12 References
BOLACH, E, LESNIAK, W and ZIOBROWSKI, J, 1985. Acta Aliment. Polonica, 11, 1. ELIMER, E, 1994. Studies on Use of Plant Fats for Citric Acid Production by Aspergillus niger, PhD thesis, University of Wroclaw, Poland. GUTCHO, S J, 1973. Chemicals by Fermentation (Noyes Data Corporation, Park Ridge, NY, USA). HASTING, J J H, 1971. Advances in Applied Microbiology, 14, 1. IKENO, Y, MASUDA, M, TANNO, K, OOMORI, I and TAKAHASHI, N, 1975. Journal of Fermentation Technology, 53, 752. KARKLINSH, R J and PROBOK, A K, 1972. Organic acid biosynthesis (in Russian), Zinatne, Riga. KIMURA, K and NAKANISHI, T, 1985. German Patent 2 065 206. KUTERMANKIEWICZ, M, LESNIAK, W and BOLACH, E, 1980. Przem. Ferm. i Owoc.-Warzyw, 6, 27. LEOPOLD, H and VALTR, Z, 1964. Die Nahrung 1, 37. LESNIAK, W, 1972. Studies on Submerged Citric Acid Fermentation, PhD Thesis, University of Wroclaw, Poland. LESNIAK, W, 1976. Przem.Ferm. i Rolny, 6, 22. LESNIAK, W, 1989. Polish Technical Review, 5, 185. LESNIAK, W and KUTERMANKIEWICZ, M, 1990. Citric Acid Production—Basic Review (in Polish), STC, Warsaw. LESNIAK, W, PODGORSKI, W and PIETKIEWICZ, J, 1986. Przem. Ferm. i Owoc.-Warzyw, 6, 22. MALDONADO, P and CHARPENTIER, M, 1975. German Patent 2 551 469. PIETKIEWICZ, J, PODGORSKI, W and LESNIAK, W, 1996. Proceedings of the International Conference on Advances in Citric Acid Technology (Bratislava, Slovak Republic), p. 9. RATLEDGE, C, 1977. Fermentation substrates, Annual Reports on Fermentation Processes, Vol. 1, Chapter 3.

160

Citric Acid Biotechnology

SHADAFZA, D, OGAWA, T and FAZELI, A, 1976. Journal of Fermentation Technology, 54, 67. SHENNAN, L and LEVI, J D, 1974. Progress in Industrial Microbiology, 13, 3. SMIRNOW, W A 1983. Food Acids (in Russian), Moscow, 105. WOJTATOWICZ, M and SOBIESZCZANSKI, J, 1981. Acta Microbiologica Polonica, 30, 69.

11

Design of an Industrial Plant
JACOBUS D VAN DER MERWE

Nomenclature a A B Cn Cp dhole D D e g h hf HD J k kL K K lp L L N Np P ,P Q r R specific area area permeability coefficient (n = 1 – 5) constants specific heat capacity pore diameter diameter liquid diffusivity fractional voidage acceleration due to gravity bed height film heat transfer coefficient dispersion height flux thermal conductivity liquid side mass transfer coefficient at the gas–liquid interface fluid consistency index Kozeny constant pore length liquid height in the reactor characteristic length in dimensionless numbers impeller speed power number power input pressure drop volumetric gas flow rate at NTP filtration resistance of the filter medium universal gas constant (m2 m-3) (m2) (m2) (J kg-1 °C-1) (m) (m) (m2 s-1) (-) (m s-2) (m) (W m-2 °C-1) (m) (m3 m-2 s-1) (W m-1 °C) (m s-1) (Pa s) (-) (m) (m) (m) (s-1) (-) (kW) (N m-2) (m3 s-1) (m-1)

161

162

Citric Acid Biotechnology

T u V

temperature superficial gas velocity volume

(°C) (m s-1) (m3)

Greek letters

=

ratio of maximum hydrostatic head to the pressure at liquid surface or; average specific cake resistance Am membrane porosity s interfacial tension I dynamic viscosity B gas hold-up O density f fraction of filter area immersed µ viscosity M mass of dry solids in a filtrate volume V

(-) (m-1) (N m-1) (Pa s) (-) (kg m-3) (-) (N s m-2) (kg m-3)

Abbreviation and dimensionless numbers BOD Bo De Flg Fr Ga HMF NF Pr PFD RCS Sc SVC UF VVM biological oxygen demand Bond number Deborah number aeration number Froude number Galilei number hydroxy-methyl-furfural nanofilter Prandtl number process flow diagram readily carbonizable substrate Schmidt number standard variable cost of production ultrafiltration volumetric flow of air per unit reactor volume per minute (mg 1-1) gDrH/I ug(1+B) /B@s Q/ND3 N2D/g gDR3/Deff Cpµ/k µ/HD

(min)

Subscripts a d D eff g i L o apparent downcomer total area effective gas impeller liquid unaerated

Design of an industrial plant

163

r R T

riser reactor top

11.1 Design of an industrial plant 11.1.1 The customer requirement The focus of this chapter is to outline the many facets that are integral to the design of an industrial plant. It is assumed that the request to design the plant originated with a customer requirement: either internal or external to the process engineering team. Although the exact terminology used might differ from company to company, the project cycle will more or less follow the scheme presented in Figure 11.1. All the necessary planning, project documentation, detail design and project execution aspects cannot be covered in detail: such a treatise would warrant a separate volume. Rather an approach is presented which will provide a person with a reasonable technical background with a framework from which to proceed. The methodology is generic to the design of any (fermentation) process plant. As such the designer is advised to consult the standard chemical engineering texts on the subject. For the purpose of this chapter, the process engineer is presumed to be working as a member of a multidisciplinary team, within an environment that has formal project procedures in place. Thus aspects such as tender enquiries, procurement and project documentation will not be discussed. Where applicable, the use of specialists in the field will strongly be recommended. For certain unit operations, it is cost effective to consult with a specialist vendor on the issue, where the main function of the citric acid producers’ process engineer is the accurate definition of the unit requirements.

11.1.2 Chemical plant design Although citric acid is a fermentation product, it is still a bulk commodity chemical. At the outset it would therefore be appropriate to consider some similarities in designing a citric acid facility and a chemical plant: • Isolation steps utilize unit operations standard to the chemical industry. • The standard variable cost of production is a key economic factor. • Plant capacities are increasing as new investors attempt to get maximum benefit from the economy-of-scale principle. • Transport cost of raw material or product can be a major consideration.

11.2 Data required The design process will start with some level of research and/or process development to be done (Figure 11.1). Following on the R&D phase, the process design team needs certain information, before setting the design basis. This can be grouped as marketing inputs that will affect the design and technical data.

164

Citric Acid Biotechnology

Figure 11.1 Schematic representation of the project design sequence

11.2.1 Customer and marketing information • Would there be any seasonality to the product demand? If so, this will impact on the sizing of units. • What is the required product specification, including the crystal size distribution? • Would the plant be required to produce both anhydrous and monohydrate grades? • What type and size packaging is required?

Design of an industrial plant Table 11.1 Technical data required

165

• Are there constraints imposed on the packaging material due to environmental legislation? (see Livingstone and Sparks (1994) for a discussion on the effect of new packaging laws). • Is the final product application known? For example, will it be used in dry formulations, where a specific crystal size distribution is important, or will it be dissolved and used in solution? • Are there any other specific customer requirements?

11.2.2 Technical data This includes all data that would ultimately be required for the design of the plant (see Table 11.1). It is recommended to compile a process data book during the initial phases of the design and thereafter use it as a standard. The plant location has not necessarily been determined yet, therefore aspects such as available steam pressure and other site-related issues are omitted at this point.

11.3 Design basis This stage is also referred to as the conceptual or preliminary design phase. At this point the process evaluation has been done with regard to Reisman (1988, p. 52): • • • • site selection; a comparison of yields and productivity with different strains and substrates; alternative processing routes; and plant capacity and utilization.

A process description and material balance quantifies aspects such as effluents, by-products and site storage requirements for raw materials and products. Usually such a mass balance is calculated backwards—i.e. starting with the stated customer requirement. The approach to process development has also been fixed at this stage. In other words: To

Figure 11.2 Schematic flow diagram for the production of citric acid

Design of an industrial plant

167

what extent will licensed technology be sought and which aspects might be developed fully in-house. The process flow diagram (PFD) with the mass balance will now form the baseline for more detail design.

11.3.1 New technology In deciding on a processing route, the design team would have to consider new developments in technology, which might offer a competitive advantage. The driving force to consider such technology, in spite of possible increased risk, would be due to one or more of the following factors: • • • • lower capital outlay required for the process; cost competitive standard variable cost (SVC); environmental pressure to reduce effluents; or cost or availability of substrate.

The production of citric acid is a fairly mature technology, and it is unlikely to expect a radical breakthrough. More likely would be incremental advances (see Roussel et al. (1991, p. 54) for the context of the terms radical and incremental) in technology, as major producers focus R&D efforts on staying competitive. Thus the processing of citric acid can be expected to remain within the scheme as set out in Figure 11.2. Two areas where new technology might impact, will be on fermenter design and direct crystallization routes. The former refers to the probable phasing out of mechanically stirred vessels, while the latter includes all processes aimed at recovering citric acid without a precipitation sequence. This would include: • • • • • membrane applications; novel ion exchange resins; solvent extraction; electrodialysis; and chromatography.

11.4 Scope definition The project scope definition is rather like a questionnaire or checklist that prompts the design team to ensure that nothing has been overlooked. Once a detail scope definition has been documented, the elements required for the process package fall naturally into place. Depending on the process engineering company and client, the actual format may vary, but the aspects listed in Table 11.2 will always be addressed.

11.5 Process package Having documented the project scope, the process engineer can proceed with the detailed process package(s). Ideally it would not be an iterative process (Figure 11.1), but seldom would enough cost data be available at the scope definition stage to obtain final project

168

Citric Acid Biotechnology

Table 11.2 Details to demarcate with a formal project scope definition

Design of an industrial plant

169

approval. At this stage the design concepts and capacities are frozen, so that the focus is on the detail design of the individual units. The plant battery limits have also been set, and the various interfaces are clearly defined. In parallel to the detail unit design, the drawing office can start on plant layout options.

11.5.1 Process flowsheet In terms of the process flow, two general rules apply: • Limit the number of unit operations. • Simplify the flow sheet. The above stems from the fact that for each additional piece of equipment it is not only the unit cost to be considered, but the installed cost, which includes the associated civil work, piping, valves, instrumentation and electrical requirement. Hence one additional operation, such as a ‘polishing’ filter, should carefully be scrutinized from an economical point of view. Often it is cost effective to increase the specification level on the primary operation. Following is a discussion on the design of typical unit operations for the production of citric acid. While a significant percentage of the world’s demand is still produced via surface fermentation, this is not covered. It is unlikely that further new plants utilizing this technology will be purpose-built. A typical PFD for the classical process is also not duplicated: these are available in standard references (Reisman, 1988) and for the purposes of this section Figure 11.2 will suffice.

11.6 Raw material The main design concerns with regards to raw material are the logistics involved and storage volume required. Molasses is a seasonal product and if the intention is to operate the citric acid plant throughout the year, this must be taken into account. In the worst case, the citric acid producer might have to provide several months’ storage capacity on site. At the other extreme the producer would be situated adjacent to a corn starch producer, where the substrate would be available year round, with minimal on-site storage required. Obviously such a scenario would offer significant cost advantages. The logistical issues to be addressed include: • • • • Transport by road, rail or shipping? Off-loading facilities required and metering of quantities. Site access. Will delivery be a 24-hour, seven-day-a-week operation, or can it be planned as a day shift activity, Monday to Friday?

11.7 Substrate preparation Prior to fermentation, the substrate must be sterilized and the concentration adjusted to the required sugar loading. On any large-scale operation, sterilization will be done continuously and not on a batch basis. One exception would be the seed fermenters, where sterilization

170

Citric Acid Biotechnology

can still be done in situ. One of the reasons to avoid batch sterilization of media on the production scale, is the possible formation of complexes such as hydroxymethyl-furfural (HMF). Formed from the reaction between glucose and ammonium and nitrogen compounds at elevated temperatures, HMF is a strong respiratory inhibitor. Due to the long heating and cooling cycle of large scale batch sterilization, HMF produced might inhibit the subsequent fermentation to non-optimal productivity levels. In discussing fermentation design, Söderberg (1983) lists the advantages of continuous sterilization and Wallhäuser (1985) offers a comprehensive treatment of media and vessel sterilization.

11.8 Fermentation The mechanically stirred tank reactor (STR) has been the standard in bioprocessing for at least 40 years. Although the picture is now changing, fewer scale-up studies (especially on the larger scale) have been done on airlift reactors, while scale-up and mixing in the STR has been extensively researched and several design correlations are available. The reader is referred to a review article by Berovic (1991) where he discusses advances in reactor design.

11.8.1 Scale-up and design Bioreactor design usually involves some experimentation on a scale smaller than the production scale. Various approaches have been followed in designing the production scale reactor. These include: • • • • rules of thumb; scaling according to one specific parameter; geometric similarity; and scale-down method.

Rule of thumb guidelines, based on historical data and conventions followed with previous successful designs, provide the starting point for further detail calculations. The following set (Sections 11.8.2–11.8.3) is not intended as a complete list, but as a summary of typical aspects which would be included.

11.8.2 Heat transfer • Heat production during aerobic fermentations is proportional to the oxygen consumption rate. • The heat liberated in the fermenter increases proportional to the volume (=DR3), while the available surface area for heat transfer increases proportional to the square of the tank diameter. • In order to improve the heat transfer, the engineer has three choices: – Increase the available ,T to the reactor. – Improve the heat transfer coefficient. Unfortunately, no practical approach to accomplish this has been proposed yet. The heat transfer coefficient is a function of

Design of an industrial plant

171

the mixing power input into the vessel, but only to an exponent of 0.2 to 0.3 (Oldshue, 1985). As the mechanical energy input is dissipated as thermal energy, it does not help to increase agitator power input in order to improve heat transfer. Indications are that the presence of dispersed air bubbles at the heat transfer surfaces increases the coefficient (De Maerteleire, 1982). This is possibly due to a scouring action on the surface. Effective utilization of mixing energy input to ensure complete gas dispersion, would thus be an important consideration. – Increase the available heat transfer surface.

11.8.3 Mass transfer • It is generally taken that the rate-limiting step in oxygen transfer is the gas-to-liquid interface transfer. Further it is assumed that the concentrations within the gas bubble are homogeneous and that the overall reaction is not limited by the cell oxygen uptake rate. A special case applies with pellet morphology, where diffusion effects into the biomass cluster might come into play. • The oxygen transfer coefficient (kLa) is a positive rising function of the superficial gas velocity and specific power input. Hence increasing either of these parameters will enhance kLa. • Mixing becomes less ideal as the scale increases. Even with increased specific power input, mixing times still increase with scale. This is an important factor to bear in mind with regard to the existence of local areas within the reactor of substrate limitation (Oldshue, 1989).

11.9 Design of a stirred tank reactor When designing the bioreactor, the engineer has to specify the vessel geometry, the power input and aeration requirements of the system. The oxygen uptake rate would normally be determined experimentally as a function of the specific growth rate, while the required power input is dependent on the system rheology. At the same time the vessel geometry affects the superficial gas velocity and hence the required power input for a given oxygen transfer rate. Therefore these aspects are not arbitrary, but interrelated.

11.9.1 Non-aerated power input The non-aerated power input, P0 is important in terms of correctly sizing the agitator motor for start-up operation. It also gives the maximum loading that will occur in case of a compressor failure. The relevant equation is:
(11.1)

In the laminar flow range the power number declines linearly from an initial maximum value, while it is independent of the Reynolds number (and constant) in the turbulent flow regime. The power number varies according to the impeller geometry, but shows the same profile for a wide range of impellers (Mockel and Wollechensky, 1990).

172

Citric Acid Biotechnology

11.9.2 Aerated systems Gas hold-up during aeration decreases the effective density of the medium and hence the agitator power demand. The correlation used to predict the aerated power demand is:
(11.2)

which can be applied for non-Newtonian media (Taguchi and Miyamoto, 1966). In equation (11.2), the constant, C1, is dependent on system geometry.

11.9.3 Correlation of kLa The commonly used form for correlating kLa to superficial gas velocity, ug, and power input per unit volume, P/V, is:
(11.3)

The constant C2 and exponents =, ß and d are system specific (i.e. scale dependent and function of geometry) and must be determined experimentally. However, the following generalizations can be stated: • For Newtonian fluids the value of d is usually small and in the region of 0.10–0.14. Hence errors in the viscosity term do not drastically influence the accuracy of the result. Applying the concept of an apparent viscosity, this term can also be used in correlating data for non-Newtonian rheology. The wall viscosity, µM, can be taken to be equal to the viscosity at zero biomass concentration. •
=

and ß are positive and generally in the range, 0.25 < =, ß < 0.9.

11.9.4 Scale-up according to geometric similarity With this approach, the geometry of a reference vessel on the smaller scale is used as the basis for the specification of the large scale vessel. Several ratios (such as aspect ratio; impeller to tank diameter, etc.) are determined and then held constant. In terms of the process requirements, this approach does not yield an equivalent micro-environment. This method should not be used for large scale design purposes, but can be useful when scaling at the laboratory scale.

11.9.5 Scale-up according to one specific parameter The two parameters that are commonly used with this approach are: • Constant specific power input. • Constant kLa. Scale-up with constant specific power input This method would be recommended when scaling from the laboratory scale of 20 l to a bench scale unit of, say 200–300 l. Using this approach when scaling directly to a production volume, leads to an uneconomical power consumption.

Design of an industrial plant

173

Scale-up with constant kLa When transferring scale to production volumes, this is the preferred approach. However, the direct application of equation (11.3) can lead to errors, as the parameters =, ß and @ could be scale-dependent. For example, the influence of viscosity effects are more pronounced on the larger scale. In practice the validity of parameters on the larger scale should therefore be confirmed. Setting kLa | scale 1 = kLa | scale 2 implies: • The aspect ratio usually increases with scale. Using the definition of superficial gas velocity (ug = Q/A), it follows that at constant VVM, the velocity, ug, increases. Thus, the specific power input required decreases. • If Pg/V is held constant, a lower aeration rate is required on the larger scale. • In cases where viscosity effects are negligible, it can be stated directly that: (Pg/V)=(ug)ß | scale 1 = (Pg/V)=(ug)ß | scale2 (11.4) 11.9.6 Design constraints Parameters such as P g/V and u g cannot be chosen arbitrarily; at the lower limit of power input in the STR is the required impeller speed for complete gas dispersion. Gas dispersing capacity varies between axial and radial type impellers and must be confirmed in each case. Axial impellers will handle a higher gas loading before flooding (at the same speed) than the Rushton impeller (McFarlane et al., 1995). For Rushton turbines the minimum speed required for gas dispersion was confirmed as (Hudcova et al., 1989): (FlG)F = 30(D/T)3.5(Fr)F while a suggested working agitator speed is given by: (11.6) This is between the flooding point and point of gross recirculation, with FlG the aeration number and Fr the Froude number. These correlations are applied with regard to the bottom impeller only; as long as the bottom impeller is not flooded, dispersion at the higher impellers will not be a constraint. At the upper limit, the maximum impeller speed is set by the shear rate that can be allowed in the system. In practice this constraint leads to tip speeds in the range of 4–7 m s-1 for production scale vessels. (11.5)

11.9.7 Regime analysis and scale-down The application of regime analysis together with the scale-down technique was essentially developed at the Technical University of Delft (Oosterhuis, 1984; Sweere et al., 1987). The concept is applied in four parts: • • • • Regime analysis of the process at the production scale. Simulation of certain (rate-limiting) mechanisms at laboratory scale. Process optimization and modelling at laboratory scale. Translation of optimized conditions back to the production scale.

174

Citric Acid Biotechnology

The purpose of doing the initial regime analysis is to establish which mechanisms are rate determining. This is done by comparing the order of magnitude of the different characteristic times. If the small scale study is to be representative, the relative time-constants of ratelimiting steps must remain in the same order. It should be noted that the method is only an order of magnitude comparison and not an exact procedure. Thus if the time constants for two processes (such as mixing time and oxygen transfer) are similar, the mechanisms should be investigated further.

11.10 Airlift and bubble column reactors Developments during the last decade indicate that pneumatically agitated vessels will eventually replace mechanically agitated reactors. This is due to several advantages of such a design, as pointed out by Mashelkar (1970) and Söderberg (1983): • No need to maintain sterility around an agitator shaft entry point. • No mechanical constraints due to agitator shaft length or motor and gearbox size. • Lower heat load, as the agitator power input can contribute as much as 30 per cent to the total energy input. • Lower fabrication cost for the vessel. • Lower cost in terms of structural steel. • Lower maintenance cost. • The vessel functions as a variable mixing power unit simply through controlling the aeration. Airlift reactor refers to configurations where a draft tube is employed to set up a liquid circulation pattern in the vessel. Such a draft tube can be internal or an external loop. Where the vessel does not have a draft tube, the term bubble column is used, rather than tower reactor. Scale-up and design of these reactors are done with empirical correlations established in terms of the macroscopic parameters such as pressure drop, gas hold-up, liquid velocity distribution and mixing properties.

11.10.1 Approaches to design Three methods can be recommended for the design of an airlift or bubble column reactor: • Using the scale-down technique (Choi, 1990) as discussed in Section 11.9.7. • Scale-up with constant superficial gas velocity. • Maintaining constant kLa on scaling up. Understandably, the critical parameter in these types of reactor is the superficial gas velocity: not only the power input, but also gas hold-up (B) and effective viscosity (Deff) can be correlated to ug. In the case of airlift fermenters, the important geometrical consideration is the ratio of downcomer to riser area and the corresponding liquid velocities. The bubble column pressure drop is simply the sum of the sparger pressure drop and the hydrostatic head in the reactor:
,

Ptotal = ,Psparger + ,Phead

(11.7)

Design of an industrial plant

175

while the energy input can be calculated from (Deckwer, 1985): Pg = QHg[RT ln(1 + =) + 1/2(ug)2] (11.8)

This is the sum of the gas kinetic energy and the compression energy to overcome the pressure drop. The parameter, =, is the ratio of the maximum hydrostatic head, to the pressure at the liquid surface:
=

= HL(1 - B)gL/PT.

(11.9)

Gas hold-up The gas residence time in the reactor is determined primarily by the liquid circulation velocity and the bubble swarm rise velocity. As the liquid circulation velocity increases, the degree of back-mixing in an airlift reactor and the fractional hold-up increases. This means more efficient utilization of the available oxygen than in an STR of similar geometry. Therefore airlift reactors typically employ aspect ratios as high as 10, to develop high liquid circulation velocity (Onken and Weiland, 1983). Hold-up has been correlated as being directly proportional to ug, i.e.
B

= C3ug

(11.10)

while Mashelkar (1970) proposes:
B

= (ug/HL)/(30 + 2ug)(72/I)-1/3
-1

(11.11)

which is applicable for Newtonian fluids in the range 5 < ug < 12 cm s . For non-Newtonian fluids, Barker and Worgan (1981) correlated hold-up data to the consistency index, K, according to:
B

= 3.09 + 4.5K - 4.51K2

(11.12)

Effective viscosity and shear rate Aspergillus niger fermentation broths exhibit deviation from Newtonian fluid behaviour, often correlated with the Power Law model. It has been shown that the non-Newtonian behaviour can also be observed with pellet morphology and is due to the presence of the biomass (Mitard and Riba, 1988; Allen and Robinson, 1990). Apparent viscosity is calculated from the average shear rate in the vessel, which in an STR is directly proportional to the impeller speed. For bubble columns, Popovic and Robinson (1989) suggest a direct relationship to ug: C = ßug where ß is a constant [m -1]. At the same specific power input, the shear rate in a bubble column is one order of magnitude lower than in an STR. It has also been established that at a specific power input in the range 1–5 W/kg, µa does not impact on kLa for µa < 1.0 Pa.s. At higher viscosity (and the same specific power input), kLa decreases by two orders of magnitude, as µa increases to 100 Pa.s. This is explained in terms of reduced interfacial area per unit volume available for mass transfer. Heat transfer The correlation put forward by Mashelkar (1970) for calculation of the process side film heat transfer coefficient in bubble columns is:

176

Citric Acid Biotechnology

hf = 1380(ug)0.22/(Pr)0.5 In the case of airlift reactors, Chisti (1989) quotes two equations: hf = 8.71(Ar/Ad)0.25(ug)0.22/(Pr)0.5 and hf = 13.34(1 + Ar/Ad)-0.7(ug)0.275

(11.13)

(11.14)

(11.15)

11.10.2 Mass transfer correlations Oxygen transfer as a function of ug Several researchers (Mashelkar, 1970; Barker and Worgan, 1981; Deckwer, 1985) have found a direct dependence of kLa on superficial gas velocity. These correlations are of the form: kLa=(ug)n or: kLa=(ug)n/(µa)b (11.17) with the coefficient, n, typically in the range 0.7–0.8. For airlift fermenters, Popovic and Robinson (1989) found: (11.18) which can be reduced to equation (11.16) or (11.17). In this case Al is proportionality constant, and refers to the sparger hole diameter [m]. The exponents a1, b1, d1, e1, f1, g1 and h1 must be determined experimentally. Dimensional analysis approach The general equation proposed for design purposes (Deckwer, 1985), incorporates several dimensionless groups: kLa = C4(D/DR2)Scb1 Bob2 Gab3 Fr(1 + C5Dem)-b4 (11.19) (11.16)

Again C4 and C5 are constants and b1–4 must be determined experimentally. Depending on the system under investigation, not all groups would necessarily be relevant. The Froude number, Fr, is often omitted where vortex formation is not a factor, while the Deborah, De, number accounts for the elastic properties of the broth. A number of correlations of the format of equation (11.18) and (11.19) are summarized by Chisti (1989).

11.11 Product isolation In schematic form, the isolation sequence is presented in Figure 11.2 as: biomass removal; purification; and crystallization. Irrespective of the technology employed, the first step remains the separation of cell mass from the fermentation broth. Thereafter the sequence and unit operations during purification will depend on the specific technology used. Finally, the crystallization section is again generic to bulk producers.

Design of an industrial plant

177

11.12 Cell removal Conventionally, this is a filtration operation: unlike in the baker’s yeast industry, centrifuges are not the preferred units for harvesting biomass. This is due mainly to the higher capital, operating and maintenance cost of centrifuges in comparison to filtration operations. The product of value in this case is also the filtrate and not the biomass. Suitable filter types could include: • • • • Rotary vacuum drum filters. Plate and frame filter press. Continuous belt filter. Disc filters.

The design of the filter is based on the application of the Poiseuille equation (Boss, 1983). The equation can be written in various forms (Coulson and Richardson, 1983, p. 323), but always relates the rate of filtration, dV/dG, to the pressure drop across the filter, ,P, filtration area, A, liquid viscosity, µ, resistance to filtration, r, and cake compressibility, d: (11.20) where a is the average specific cake resistance, which is a function of the pressure applied and the cake compressibility @: = = =’,P@, where =’ is a constant related to the size of the particles in the cake. A value of d = 0 corresponds to an incompressible filter cake, while @ = 1 would be a gelatinous protein sludge. From equation (11.20) it follows that for such cases, the rate of filtration is independent of the applied pressure. The resistance of the filter medium, r, is often expressed as an additional cake resistance, with a fictitious thickness, to simplify the handling of experimental data. On a (batch) laboratory apparatus, the rate of filtration is measured versus the applied pressure. The parameters of the equation are then determined through curve fitting of the data. Scaling up the filter design, is in essence the specification of a required filter area to achieve a certain rate of filtration, at the allowable pressure drop. Equation (11.20) can be extended to account for applications such as a rotary drum filter, where the total filter area is not continuously submerged in the slurry (Peters and Timmerhaus, 1968, p. 487), by defining an effective area: ADOf. If AD is the total filter area and Of the fraction immersed in the slurry, then: (11.21) where VR is the volume of filtrate per revolution of the filter. Final points to consider in selecting a suitable filter include: • Will washing of the filter cake be necessary to recover the maximum citric acid? • What is the desired or acceptable moisture content of the cake? • Can a filter aid such as diatomaceous earth be used, or is this undesirable? If so, is this from an economical point, or because it impacts on the fodder value of the biomass? • Can gypsum be used as a filter aid? • Is there an additional time constraint on the required rate of filtration? This applies in cases where the operation must be completed within a short time (usually less than 16 hours) after the end of fermentation, before complete cell lysis occurs.

178 Table 11.3 Classification of ultrafiltration and nanofiltration

Citric Acid Biotechnology

11.13 Purification 11.13.1 Membrane applications Microfiltration, ultrafiltration and nanofiltration have been investigated for application in citric acid processes. Microfiltration, as with lactic acid production, offers possibilities with regard to cell retention in continuous citric acid fermentation systems (Enzminger and Asenjo, 1986; Daniel and Brauer, 1994; Rubbico et al., 1996). Where the aim is to remove proteins or enzymes from the fermentation broth, ultrafiltration (UF) or nanofiltration (NF) is employed (Bohdziewicz and Bodzek, 1994). With NF it is also possible to remove residual sugars from the process liquor (Raman et al., 1994) at low pH. This is due to the structure of the NF membrane, where the active membrane layer typically consists of negatively charged groups. Thus salts are rejected due to electrostatic interaction between the ions and the membrane while sugars are rejected on molecular size. At low pH values, the citric acid is un-dissociated and permeates the membrane in spite of sugars (with a similar molecular weight) being rejected. Conceivably it is therefore possible to put together a process scheme which would eliminate the lime precipitation step by removing, not only proteins, but also a significant percentage of residual sugars, through NF. Industrial exploitation of this concept has been hindered by: • • • • Developing membrane material that offers long-term stability at low (±2) pH. Membrane fouling. Cost of membranes. Energy requirements due to high trans-membrane pressures.

These factors are being resolved with continued research in the field: developments such as ceramic membranes offer mechanically rigid filters resistant to chemical attack. One favourable aspect of the membrane applications is that scale-up can be done through modular duplication of pilot plant units. Hence it is possible to predict accurately the performance of a full-scale unit from a series of laboratory or pilot plant tests. As with standard filter operations, it is however of prime importance to use a representative sample in doing experiments. The application areas of UF and NF overlap to some extent, but can be grouped according to molecular cut-off point (Gyure, 1992). Mathematically, ultrafiltration can be modelled with the Hagen–Poiseuille equation (Kula, 1985). Analogous to equation (11.20), the flux, J is related to the trans-membrane pressure applied, dynamic viscosity and membrane properties: (11.22)

Design of an industrial plant

179

Equation (11.22) applies in the ideal case, where the membrane pores are of uniform distribution and size and fouling of the membrane or concentration polarization can be neglected. The latter effect occurs due to an accumulation of the retained solute at the membrane surface, resulting in a concentration higher than the bulk concentration. Such concentration of the retained species means that the component will diffuse back to the bulk flow conditions. At a sufficiently high concentration of the retained species, saturation concentrations at the membrane surface lead to the formation of a gel layer, which then offers an additional filtration resistance. Once such gel polarization is established, the flux becomes independent of the pressure: increased pressure forms a thicker gel layer, which in turn offers increased resistance and hence the flux does not increase. In practice, the UF process is better described by the mass transfer limited (i.e. diffusion limited) models. For a comprehensive treatment of the topic, the reader is referred to the Ultrafiltration Handbook (Cheryan, 1986); Reisman (1988) also presents some comparative data on capital and operating costs of membrane units.

11.13.2 Colour removal The function of the colour removal step is to yield an aesthetically acceptable (i.e. white) food grade product. While colour can be removed with resin applications, the norm is still to use activated carbon for this purpose. Specifying the required carbon loading per volume of citric acid process liquor, requires some laboratory experiments. This is done to quantify the carbon/acid ratio and give an indication of the volume of process liquor that can be treated, as well as the kinetics of colour removal. Treatment can either be in a fixed bed system, or by simply adding fine activated carbon directly to the process liquor. After a calculated residence time in contact with the activated carbon, the latter is then removed by filtration. This method offers the advantage that a simple stirred batch tank, sized for the calculated residence time, will suffice. A disadvantage is the filtration step required afterwards and the disposal cost of spent carbon. In a fixed bed system, the spent carbon can be regenerated with steam. These columns are usually installed as two parallel units: while one is in operation the other is regenerated. For laminar flow through the column, the pressure drop at a specific superficial fluid velocity can be calculated from the Carman–Kozeny equation (Coulson and Richardson, 1983):
,

P = -(u h µ)/B

(11.23)

where: B = [1/K ][e3(S2(1 - e)2)] (11.24)

B, is the bed permeability coefficient, while the Kozeny constant, K , is generally assumed to be 5. This constant is a function of particle shape and porosity. It should be noted that the equation was derived on the basis of the bed consisting of uniformly sized, spherical particles. Where significant deviation occurs from this situation, some corrections have to be taken into account (Coulson and Richardson, 1983). Equation (11.24) can also be applied to calculate the pressure drop across an ion-exchange column resin bed. In such a case it is to be expected that the particles will be of uniform size and spherical.

180

Citric Acid Biotechnology

11.13.3 Ion-exchange Ion exchange involves the interchange of any ion between the process liquor and the polymer resin. The essential points (Dechow, 1983) are that ion exchange reactions are: • • • • • stoichiometric; reversible; possible with any ionizable compound; a function of the resin selectivity and reaction kinetics; subject to the usual chemistry kinetic behaviour with regard to concentration and temperature.

Typically, resins are polymers based on the cross-linking of polystyrene with divinylbenzene. Other possibilities are the cross-linking of divinyl-benzene with an acrylate or acrylonitrile, as well as phenol-formaldehyde and polyalkylamine resins. Three types of resin exchange reaction can be found in the production of citric acid: • Demineralization. • Metathesis, which is the conversion of salts of citric acid to the acid. • Adsorptive purification. The metathesis reaction can be represented as: [resin]-H+ + Na+-[citrate][resin]-Na+ + H+-[citrate]-

with equilibrium constant, K, defined as the concentration of products divided by reagents (at equilibrium). A large K-value indicates a high affinity of the resin for sodium ions and means that an excess of strong acid will be required to regenerate the resin. The polymer structure and composition determine the resin selectivity for a specific ion, but at ambient temperatures, two generalizations apply to dilute aqueous solutions: • the exchange potential increases with increasing ion valence; and • at the same ion valence, the exchange potential increases with atomic number. Thus, in increasing order of exchange potential: Na+ < Ca++ < Al+++ and Li+ < Na+ < K+ or Mg++ < Ca++ < Sr++ < Ba++ Similarly, for anions: F- < Cl- < BrThese principles are important in monitoring the ion exchange column effluent: it follows that the monovalent ions would be expected to ‘break through’ first as the resin reaches capacity loading. As the resin reactions are stoichiometric, the quantity of resin material required to remove a certain concentration of cations or anions, can easily be calculated. The resin capacity is expressed as equivalents per kilogram (on a dry basis) or per litre on a wet basis (eq/l). The equivalents number is simply an indication of the number of active sites available for adsorption and can be obtained from the resin supplier. Analysis of the process liquor will then determine the quantity of resin to be used for the required throughput. Because the exchange process is an equilibrium reaction, the resin is utilized at a level well below the theoretical capacity, thereby shifting the equilibrium in the desired direction (Le Châtelier’s principle). This does not significantly increase costs, as the resin cost is typically only 10 per cent of the unit cost (Dechow, 1983).

Design of an industrial plant

181

Specifying the ion-exchange unit Once the required ion loading to be removed and the specific resin capacity are known, the unit can be specified. Three configurations are employed: batch stirred tank, batch column and continuous operation. The most common installation is the batch column, either as two units in parallel to allow uninterrupted operation, or with an adsorption– regeneration cycle on a single column. Because the resin cost is not a major factor, it is sound design practice to allow for some excess capacity in specifying the column size. As an example, consider specifying a column of diameter d1 and a unit with diameter d2 = 1.5d1. Assuming the bed height in both columns to be h, the increase in resin volume is: V2/V1 = (1.5)2 = 2.25 (11.25)

Thus the resin volume and hence column capacity more than doubles, but the capital cost to fabricate and install the unit does not increase by the same factor. The calculation of the pressure drop for a specific column geometry can be done with equation (11.23). Adsorptive purification Possibly a commercially viable direct crystallization route, this process is already employed for the recovery of lysine. In some cases, adsorption is done directly from the fermentation broth (Van Walsem et al., 1997) thereby simplifying the flowsheet considerably. A scheme proposed for the recovery of citric acid (Ernst and McQuigg, 1992) uses temperature swing adsorption (TSA) to purify the citric acid solution. In this case the regeneration is done by utilizing the difference in resin capacity for citric acid as a function of temperature. Thus citric acid is adsorbed at ambient temperatures and desorbed with hot water. Resin capacity in excess of 155 g citric acid per litre resin, with a 96 per cent reduction in RCS values were reported. While such a process eliminates the lime precipitation route and associated by-product disposal dilemma, the environmental focus might shift to the actual resin in this case. The adsorptive resins are structured from poly-vinylpyridine-co-divinylbenzene, the production process of which generates some environmental concerns in itself. Although the effluents (pyridine compounds) can be treated, it ultimately becomes a cost which is passed on to the end user in the pricing of the resin.

11.13.4 Electrodialysis Recovering lactic acid by electrodialysis has been researched and several processes patented during the last 20 years (Nomura et al., 1987; Siebold, et al., 1995). Although it is actively being researched (Karklins et al., 1996; Moresi and Sappino, 1996), it is this author’s opinion that employing electrodialysis for citric acid recovery is not at the point of commercial exploitation yet. The technology is proven, but the energy consumption does not yet offer a competitive advantage (Novalic and Kulbe, 1996). Membrane filtration routes, solvent extraction and adsorptive ion exchange seem more likely to succeed on a cost competitive basis.

11.13.5 Solvent extraction As a direct crystallization route, the application of solvent extraction seems to offer interesting possibilities. Similar to membrane processes, this might necessitate the use of more refined

182

Citric Acid Biotechnology

substrates such as glucose syrups, to avoid the extraction of impurities in beet and cane molasses. Once again, with increasing environmental pressure on reducing effluents, this might become an economically viable alternative to the classical lime precipitation route. Presenting a discussion of solvent extraction principles is beyond the scope of this text. Suffice it to say that processes utilizing butan-2-ol tributyl phosphate plus kerosene and tertiary amines have been published (Melsom and Meers, 1985). A further simplification of the flow sheet would be direct extraction from the fermentation broth (Stuckey, 1997), which is currently being researched.

11.14 Crystallization stages Unit operations included in this section are the evaporator, crystallizer, centrifuge and dryer. These units require specialist vendor input and will rarely be designed in-house.

11.14.1 Evaporation At the evaporation stage, the process liquor will contain 15 to 20 per cent citric acid in solution. It is an energy intensive operation and the efficient utilization of energy is an important design consideration. The norm is to specify multiple effect evaporators, where vapour from one effect is condensed in the subsequent unit re-boiler, with the process side operated at a lower pressure. Mechanical vapour recompression can also be considered and depending on the relative steam/electricity cost, is often economical at large capacities. The types of evaporators employed vary, but forced circulation and falling film types have been used successfully for a number of years. In planning the energy integration, care must be taken to ensure that the citric acid will not be discoloured through exposure to high temperature. Especially if a direct crystallization route is considered, trace amounts of residual sugar are still present at the evaporation stage. This means that the temperature in the first effect evaporator, where steam is used as the heating medium, must be limited to well below 100°C. Such a constraint necessitates the use of low-pressure steam and also dictates the vacuum required in subsequent stages. Removing colour from the concentrated citric acid solution after the evaporator stage presents practical problems and should be avoided where possible. As the solution is close to saturation, the prevention of blockages due to crystals settling in a unit such as an activated carbon column is cumbersome.

11.14.2 Crystallization Typically this is a two-stage operation, where the first stage is the final purification step. The second crystallization must yield the correct crystal size distribution, according to the specified customer requirement. Usually continuous forced circulation crystallizers, in line with pusher centrifuges will be used. The mother liquor produced from the crystallizers can be recycled through an adsorptive ion exchange unit, or utilized for the production of sodium citrate. Alternatively, the acid can be recovered with the lime precipitation route. In discussing the unit specification with a vendor, it will be necessary to stipulate if the option of anhydrous and monohydrate acid is required. As the crystallization temperature of monohydrate citric acid is lower, this impacts on the capacity of the vacuum ejectors/pumps.

Design of an industrial plant

183

11.14.3 Product drying Again a specialist vendor can consult on a suitable type of dryer. This might typically be a moving bed/fluidized bed unit. From the centrifuge, the wet cake should not contain more than 5 per cent moisture; a simple energy balance will therefore determine the dryer heat load. Where high humidity conditions exist, the unit will have to incorporate an air-drying sequence to obtain product according to specification.

11.15 Product packaging The process engineer’s responsibility here lies in correctly transferring the customer requirement into a technical specification. Before this is done, it should be decided in principle if product packaging is to be done during one, two or three shifts. Where volumes or specific market needs warrant, one grade of product might have a dedicated packaging line. These issues also set the storage hopper volume required. In planning this area, cognisance should be taken of the time required for final product quality approval. This usually implies that product is kept in an area immediately adjacent to the bagging area, while quality control analysis is being done. Therefore sufficient space should be allocated for unhindered flow of material in the area.

11.16 Effluent and by-products The main out-flows produced during the process are: • biomass from the cell removal stage; • excess water; • gypsum—if recovery is done along the classical route; and • retentate from membrane filtration steps. As a protein source, the biomass does have some value: a recent study (Szoltysek et al., 1996) reported on an investigation into using the mycelium as a component of chicken feed. Aqueous effluent from the plant can be expected to have a relatively high BOD: in the region of 12 000–14 000 mg/l, or even higher where molasses is used as substrate. While technically this does not pose a problem, it does have a cost implication to reduce these levels. Similarly, the disposal of gypsum could be a significant cost factor, if the plant is not located in a region where there is a demand from industries in the construction sector. Volumes of retentate are small relative to the other effluents and disposal does not seem to be a problem. For example, this could be mixed as a protein source with animal fodder.

11.17 In conclusion The preceding paragraphs bear out the fact that plant design is largely a discipline generic to the chemical engineering industry. However, it should also be stressed that fermentation plants require unambiguous communication between the chemical engineer and microbiologist. Provided the process engineer can correctly interpret the sometimes unusual requirements of a ‘living system’, the application of sound engineering practice will ensure a successful design.

184

Citric Acid Biotechnology

11.18 References
ALLEN, D G and ROBINSON, C W, 1990. Measurement of rheological properties of filamentous fermentation broths, Chemical Engineering Science, 45, 37–48. BARKER, T W and WORGAN, J T, 1981. The application of airlift fermenters to the cultivation of filamentous fungi, European Journal of Applied Microbiology and Biotechnology, 13, 77–83. BEROVIC, M, 1991. Advances in aerobic bioreactor design, Chemical Biochemical Engineering Quarterly, 5, 189–192. BOHDZIEWICZ, J and BODZEK, M, 1994. Ultrafiltration preparation of pectinolytic enzymes from citric acid fermentation broth, Process Biochemistry, 29, 99–107. BOSS, F C, 1983. Filtration. In Fermentation and Biochemical Engineering Handbook, ed. H C VOGEL (Noyes Publications). CHERYAN, M, 1986. Ultrafiltration Handbook (Technomic Publishing Company Inc.). CHISTI, M Y, 1989. Airlift bioreactors (Elsevier Applied Science). CHOI, P B, 1990. Designing airlift loop fermenters, Chemical Engineering Progress, December, 32– 37. COULSON, J M and RICHARDSON, J F, 1983. Chemical Engineering, Volume Two (Pergamon Press). DANIEL, ST and BRAUER, H, 1994. Continuous production of citric acid in the reciprocating-jetbioreactor, Bioprocess Engineering, 11, 123–127. DECHOW, F J, 1983. Ion exchange. In Fermentation and Biochemical Engineering Handbook, ed. H C VOGEL (Noyes Publications). DECKWER, W, 1985. Bubble column reactors. In Biotechnology, Volume 2, ed. H BRAUER, (VCH). ENZMINGER, J D and ASENJO, J A, 1986. Use of cell recycle in the aerobic fermentative production of citric acid by yeast, Biotechnology Letters, 8, 7–12. ERNST, E E and MCQUIGG, D W, 1992. Adsorptive purification of carboxylic acids, presented at AIChE meeting, Miami, November. GYURE, D C, 1992. Set realistic goals for cross-flow filtration, Chemical Engineering Progress, November. HUDCOVA, V, MACHON, V and NIENOW, A W, 1989. Gas-liquid dispersion with dual Rushton turbine impellers, Biotechnology and Bioengineering, 34, 617–628. KARKLINS, R, SKRASTINA, I and LEMBA, J, 1996. Electrodialysis method in citric acid and its salts recovery process. Presented at Advances in Citric Acid Technology, Bratislava, Slovakia. KULA, M, 1985. Recovery operations. In Biotechnology, Volume 2, ed. H BRAUER (VCH). LIVINGSTONE, S and SPARKS, L, 1994. The new German packaging laws: effects on firms exporting to Germany, International Journal of Physical Distribution & Logistics Management, 24, 15– 25. MASHELKAR, R A, 1970. Bubble columns, British Chemical Engineering, 15, 274–281. MCFARLANE, C M, ZHAO, X and NIENOW, A W, 1995. Studies of high solidity ratio hydrofoil impellers for aerated bioreactors, Biotechnology Progress, 11, 608–618. MELSOM, P E and MEERS, J L, 1985. Citric Acid. In Comprehensive Biotechnology, Vol. 3, ed. M MOO-YOUNG (Pergamon Press). MITARD, A and RIBA, J P, 1988. Morphology and growth of Aspergillus niger ATCC 26036 cultivated at several shear rates, Biotechnology and Bioengineering, 32, 835–840. MOCKEL, H O and WOLLECHENSKY, E, 1990. Modelling of the calculation of the power input for aerated single- and multistage impellers with special respect to scale-up, Acta Biotechnology, 10, 215–224. MORESI, M and SAPPINO, F, 1996. Effect of temperature and pH on sodium citrate recovery from aqueous solutions by electrodialysis, Presented at Advances in Citric Acid Technology, Bratislava, Slovakia. NOMURA, Y, IWAHARA, M and HONGO, M, 1987. Lactic acid production by electrodialysis fermentation using immobilized growing cells, Biotechnology and Bioengineering, 30, 788– 793. NOVALIC, S and KULBE, K D, 1996, Separation and concentration of citric acid by means of electrodialytic bipolar membrane technology, presented at Advances in Citric Acid Technology, Bratislava, Slovakia. OLDSHUE, J Y, 1985. Transport phenomena, reactor design and scale-up, Biotechnology Advances, 3, 219–237.

Design of an industrial plant

185

OLDSHUE, J Y, 1989, Fluid mixing, Chemical Engineering Progress, May. ONKEN, U and WEILAND, P, 1983. Airlift fermenters: construction, behaviour and uses, Advances in Biotechnological Processes, 1, 67–95. OOSTERHUIS, N M G, 1984. Scale-up of bioreactors: a scale-down approach, PhD Thesis, Technical University of Delft, Holland. PETERS, M S and TIMMERHAUS, K D, 1968. Plant Design and Economics for Chemical Engineers (McGraw-Hill). POPOVIC, M K and ROBINSON, C W, 1989. Mass transfer studies of external-loop airlifts and a bubble column, AIChE Journal, 35, March. RAMAN, L P, CHERYAN, M and RAJAGOPALAN, N, 1994. Consider nanofiltration for membrane separations, Chemical Engineering Progress, March, 68. REISMAN, H B, 1988. Economic Analysis of Fermentation Processes (CRC Press). ROUSSEL, P A, K N and ERICKSON, T J, 1991. Third generation R&D: managing the link to corporate strategy, Harvard Business School Press. RUBBICO, R, LO PRESTI, S, BRAVI, M, MORESI, M and SPINOSI, M, 1996. Repeated batch citrate production by Yarrowia lipolytica using yeast recycling by cross-flow microfiltration, Agro-Food-Industry Hi-Tech, March/April. SIEBOLD, M, VAN FRIELING, P, JOPPIEN, R, RINDFLEISCH, D, SCHUGERL, K and ROPER, H, 1995. Comparison of the production of lactic acid by three different lactobacilli and its recovery by extraction and electrodialysis, Process Biochemistry, 30, 81–95. SÖDERBERG, A C, 1983. Fermentation Design. In Fermentation and Biochemical Engineering Handbook, ed. H C VOGEL (Noyes Publications). STUCKEY, D C, 1997. Solvent extraction in biotechnology: some novel techniques, presented at Biotech South Africa ’97, Grahamstown, South Africa. SWEERE, A P J, LUYBEN, K CH A M and KOSSEN, N W F, 1987. Regime analysis and scaledown: tools to investigate the performance of bioreactors, Enzyme and Microbial Technology, 9, 386–398. SZOLTYSEK, K, GRZESIAK, E and FRITZ, Z, 1996. Possibility of using Aspergillus niger mycelium in fodder industry. Presented at Advances in Citric Acid Technology, Bratislava, Slovakia, October. TAGUCHI, H and MIYAMOTO, S, 1966. Power requirement in non-Newtonian fermentation broth, Biotechnology and Bioengineering, 8, 43–54. VAN WALSEM, H J, THOMPSON, M C and FECHTER, W L, 1997. Simulated moving bed in the production of lysine, presented at Biotechnology South Africa ’97, Grahamstown, South Africa. WALLHÄUSER, K H, 1985. Sterilization. In Biotechnology, Volume 2, ed. H J REHM and G REED (VCH).

Index

absorption 142 absorptive purification 181 activated carbon 139 acyl CoA synthetase 38–9 adenine nucleotides 43 adsorption 142 aerated systems 172 aeration 5 agitation effects 70 air-lift bioreactor 129, 131, 174 alcohol dehydrogenase 37 alcohols 37 aldehyde dehydrogenase 37 aliphatic alcohols 141 alkali metal salts 138 alkanes 6, 157 uptake 35 alkylsuphoxides 141 alternative oxidase 123 ammonia 20, 158 anion exchange resins 142 approaches to design 174 arabinose 151 aspergillus model 107 available electron balance 130 axial impellers 173 beet molasses 151–3 microelements 152 vitamins 153 amino acid content 152 betaine 151 bipolar membranes 144 blackstrap 153 bubble column reactor 174 butanol 139 cane molasses 153–5 vitamins 153 composition 154 caramel 151 carbon balance 130 carbon content of substrates 150 Carmen-Kozeny equation 179

cation exchange 138, 155 cell removal 177 cellulose 158 centrifugation 139 chemical plant design 163 citrate synthase 41–4 citrate-free recovery 139 flowsheet 140 citric acid applications 8 biochemistry 11 biosynthetic pathway 12–14 chemical method 2 continuous production 5 koji process 7, 58 microbial 2 overproduction 129 regulation 19–21 solid state process 7 submerged process 4, 58 surface method 3 synthetic 2 transport 24–5, 44–6, 49–50 uses 7, 8 yeast based processes 6 yeast synthesis from alkanes 35–46 yeast synthesis from glucose 46–50 cloned genes 16 coagulating agents 139 colour removal 179 compartmentation 42 constant specific power input 172 copper 81 counter current flow 143 crystallisation 136, 138, 182 crystallization stages 181 cube root growth model 106, 107 customer information in design 164 customer requirements in design 163 cyclic AMP 21 cytochrome P-450 hydroxylase 36 date syrup 156 design of an industrial plant 163 design basis of an industrial plant 165–7 design constraints 173 di-calcium citrate 136, 137 dilution rate 111 dimensional analysis 176 dissolved oxygen 4, 23

188 downstream processing 135–46 economics 1 effective viscosity 175 effluent 8, 183 electrodialysis 138, 144, 181 scheme 145 electron activity 87, 88 elemental composition 124 elementary balances 122 energy balance 121 energy consumption 144 energy yield coefficients 128 equilibrium distribution coefficient 141 esters 141 ethers 141 ethylenediamineacetic acid 155 evaporation 182 exchange potential 180 fats 158 fermentation aspects of design 170 filter types 177 fixed bed filter 142 formaline 156 four compartment cell 144 Froude number 173 fructose-2, 6-bisphosphate 19 gas hold-up 175 genes 15–19, 44 geometric similarity 172 gluconic acid 12, 58 glucose oxidase 12 glucose uptake 20 glutamate 49 glycerol 22, 49 glycolytic pathway 12, 58 glyoxylate cycle 39 growth kinetics 127 Hagen-Poiseuille equation 178 heat removal 3 heat transfer 170, 175 hydrocarbons 149 hydrol 157 idiophase 117, 119, 123 immobilization 5, 145 initial conditions model 113 inoculum 4 ion exchange 180–1, 138–9, 142 iron 81 isocitric acid 59 isocitrate lyase 41 kestose 151 ketones 141 kinetic modelling 105 Kjærgaard equation 87 K a 172 L Kozeny constant 179

Index

lemon 1 lime 136–7 linear growth model 106–7, 118 liquid membranes 143 log growth model 106–7 logistic growth equation 113 Ludeking-Piret equation 107–8, 113, 119 magnesium 81 manganese 21, 22, 25, 59–60, 81 mannose 151 market 2, 9 marketing information 164 mass balance 121, 126 mass transfer 171 mass transfer correlations 176 mass yield coefficient 122 mechanistic models 105 melanoidines 151 membrane applications 178 metabolic control analysis 61–2, 106 metabolic description of A. niger growth 123 metathesis reaction 180 methyl citrate 40–1 methyl isocitrate 40–1 microelements 158 microfiltration 178 microporous hollow fibres 143 mitochondria 22 mixed biomass 111 modelling 105–19 molasses 150–6 colour 151 microflora 154 non-volatile compounds 151 pH 153 volatile compounds 151 nitrogen compounds 151, 153 Monod equation 106, 111 monopotassium phosphate 158 morphology 5, 23, 60 A. niger 71 dissolved oxygen 81 effect of carbon source 75 effect of inoculum 82 effect of nutritional factors 74 effect of pH 79 initial glucose concentration 75 nitrogen limitation 78 phosphate level 78 trace metal levels 81 mutagenesis 56 mycelium formation 126 NADH oxidation 23, 62–3 NADH: ubiquinone oxidoreductase 63 nanofiltration 145, 178 neokestose 151 Nernst equation 87, 98 new technology in design 167 nitrogen metabolism 49

Index nitrogen source 158 nomenclature 85, 121, 126, 161–3 non-aerated power input 171 non-ionic resins 142 non-Newtonian behaviour 175 oils 158 organophosphorus compounds 141 oxalate biosynthesis 15 oxalic acid 13, 58–9 removal 135 oxidation potential 87, 8 oxidative phosphorylation 124 oxygen yield coefficient 128 oxygenation 3 parasexual cycle 60 pectinolytic enzymes 145 pellets 5, 69 pentose phosphate pathway 12 peroxisomes 38–40 pH 3, 6, 23, 24 phase related model 117 phosphofructokinase overexpression 64 phosphorus source 158 Poiseuille equation 177 polyhydric alcohols 131 polysulphone 145 potassium ferricyanide 91 potassium ferrocyanide 155 power law 175 precipitation 136 flow sheet 137 pre-treatment 144 submerged fermentation 135 surface process 135 process economics 149 process flowsheet 169 process package 167 product drying 183 isolation 176 packaging 183 purification 178 pyruvate kinase overexpression 64 QSSA 125 raffinose 151, 153 raw materials 169 redox dyes 88 redox electrodes 88 calibration 88 redox potential 85 in citric acid fermentation 91–5 measurement 88 optimal 102 regulation 95 significance 89 theory 87 time course 92–3 dissolved oxygen relationship 90 temperature change 94 redox regulation chemical 97 physical 97–100 refined sucrose 156 regime analysis 173 regulatory enzymes 18 regulatory network 14 respiratory chain 62, 63 reverse osmosis 145 Rushton impeller 173 Rushton turbine 173 scale down 173 scale up 101, 170, 172 scope definition 167–8 seeding 138 shear 70 shear rate 173, 175 simple structured model for A. niger 107 solvent extraction 139–41, 181 starch 58, 157 sterilisation 155 stirred tank reactor 71, 170 design 171 stoichiometry 124 structured models 105 submerged method 150 substrate level phosphorylation 128 substrate preparation 169 substrates 149 sucrose 151, 153 sulphamide 156 superficial gas velocity 176 surface method 150 syrups 156 technical data in design 165 teratogenic effect 142 three-compartment cell 144 transcriptional regulation 15–16 transport costs 150 treatment of molasses 155 therapies-6-phosphate 20, 64 tricalcium citrate 136, 137 tricarboxylate transporter 22 tri-n-butylphosphate 141 trioctylphosphine oxide 141 trophophase 117–18, 123 tubular loop reactor 71 ultrafiltration 139, 145–6, 178 unstructured models 105 vacuum evaporation 138 wastes 136 water 159 water-soluble amines 141 xylose 151 yeast based models 113 yeasts 6 yield coefficients 106–7, 113, 125–9 zinc 81

189