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 page 1

1.1 Citric acid from lemons 1
1.2 Synthetic citric acid 2
1.3 Microbial citric acid 2
1.4 Citric acid by the surface method 3
1.5 The submerged process for production of citric acid 4
1.6 Continuous and immobilized processes 5
1.7 Yeast based processes 6
1.8 The koji process 7
1.9 Uses of citric acid 7
1.10 Effluent disposal 8
1.11 Conclusions 8
1.12 References 9
2 Biochemistry of Citric Acid Accumulation by Aspergillus niger 11

2.1 Introduction 11
2.2 Glucose catabolism in A. niger and its regulation 12
2.3 Regulation of citric acid biosynthesis 19
2.4 Role of citrate breakdown in citrate accumulation 21
2.5 Export of citric acid from A. niger 24
2.6 References 25
3 Biochemistry of Citric Acid Production by Yeasts 33

3.1 Introduction 33
3.2 Synthesis of citric acid from n-alkanes 35
3.3 Synthesis of citric acid from glucose 46
3.4 Conclusions 50
3.5 References 50
4 Strain Improvement 55
4.1 Introduction 55
4.2 General aspects of strain improvement 55
v
vi Contents
4.3 Isolation of recombinant strains using the parasexual cycle in A. niger 60

4.4 Genetic engineering 61
4.5 Concluding remarks 64
4.6 Acknowledgements 65
4.7 References 65
5 Fungal Morphology 69
5.1 Introduction 69
5.2 Factors affecting Aspergillus niger morphology in submerged culture 69
5.3 Effect of agitation 70
5.4 Effect of nutritional factors 74
5.5 Effect of inoculum 82
5.6 Conclusions and perspectives 82
5.7 References 82
6 Redox Potential in Submerged Citric Acid Fermentation 85

Nomenclature 85
6.1 Introduction 85
6.2 Overview 86
6.3 Theory 87
6.4 Measurement of redox potential 88
6.5 Significance of redox potential 89
6.6 Redox potential in citric acid fermentation 91
6.7 Regulation of the redox potential 95
6.8 Regulation of redox potential in citric acid fermentation 95
6.9 Scale-up based on redox potential 101
6.10 Conclusions 102
6.11 References 103
7 Modelling the Fermentation Process 105

7.1 Introduction 105
7.2 Aspergillus based models 107
7.3 Yeast based models 113
7.4 Conclusion 119
7.5 References 119
8 Mass and Energy Balance 121

Nomenclature 121
8.2 Metabolic description of A. niger growth 123
8.3 Mass and energy balances 125
8.4 Kinetics of growth and citric acid production by A. niger 128
8.5 Carbon and available electron balances 130
8.6 Conclusion 131
8.7 References 132
9 Downstream Processing in Citric Acid Production 135

9.1 Pretreatment of fermentation broth 135
9.2 Precipitation 136
Contents vii
9.3 Solvent extraction 139

9.4 Adsorption, absorption and ion exchange 142
9.5 Liquid membranes 143
9.6 Electrodialysis 144
9.7 Ultrafiltration 145
9.8 Immobilization of micro-organisms 146
9.9 References 146
10 Fermentation Substrates 149

10.2 Molasses 150
10.3 Refined or raw sucrose 156
10.4 Syrups 156
10.5 Starch 157
10.6 Hydrol 157
10.7 Alkanes 157
10.8 Oils and fats 158
10.9 Cellulose 158
10.10 Other medium redients 158
10.11 Conclusion 159
10.12 References 159
11 Design of an Industrial Plant 161

Nomenclature 161
11.1 Design of an industrial plant 163
11.2 Data required 163
11.3 Design basis 165
11.4 Scope definition 167
11.5 Process package 167
11.6 Raw material 169
11.7 Substrate preparation 169
11.8 Fermentation 170
11.9 Design of a stirred tank reactor 171
11.10 Airlift and bubble column reactors 174
11.11 Product isolation 176
11.12 Cell removal 177
11.13 Purification 178
11.14 Crystallization stages 182
11.15 Product packaging 183
11.16 Effluent and by-products 183
11.17 In conclusion 183
11.18 References 184
Index 187
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 53-
345 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
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.
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
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
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 n-
alkane 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.
Table 1.1 Applications of citric acid
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
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
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.
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 2-
desoxyglucose 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).
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 glyceraldehyde-
3-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
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-2,6-
diphosphate (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 Fru2,6-
P2 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 Fru-
2,6-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 F-
2,6-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-6-
phosphate (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
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 hexokinase-
encoding 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
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,
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 NADP-
specific 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
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
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 acetyl-
CoA 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).
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.
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 3-
phosphoglycerate 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.
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,
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.
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 mannitol-
1-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
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
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
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
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.
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,
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.
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
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.
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
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.
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,
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
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
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
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
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 pre-
treatment 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 Uptake of alkanes into the cell.

2 Oxidation of alkanes into the corresponding fatty acids.
3 Conversion of the fatty acids to acyl CoA esters.
4 Metabolism of fatty acyl CoA esters to acetyl CoA, or incorporation into cellular lipid.
5 Synthesis of TCA cycle intermediates.
6 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
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 P-
450 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.
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
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.
Table 3.3 Comparison of the properties of acyl CoA synthetases for C. lipolytica
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 FAD-
dependent 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,
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
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
Figure 3.3 Compartmentation in C. lipolytica during growth on n-alkanes

Table 3.5 Adenine nucleotide levels (mM) during a batch fermentation
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
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
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
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).
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 1-
oxidoreductase, 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, ß-
oxidation. 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
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,
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 fructose-
1,6-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
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
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.
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.
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,
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
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.
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 2-
oxoglutararic acid biosynthesis from ethanol by the thiamine-auxotrophic yeast Yarrowia lipolytica,
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
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.
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 N-
alkanes 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
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
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
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-di-
methylaminobenzaldehyde, 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, g- 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
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,
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
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.
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 rate-
limiting).
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
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
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 trehalose-
6-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-6-
phosphate 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 fructose-
2,6-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 high-
yielding, further strain improvement will most certainly be attempted. At the moment the
primary strategy for strain breeding is probably still mutagenesis and selection. Quantitative
Figure 4.3 Citric acid fermentation from glucose by an A. niger N400 wild-type strain (+) and
transformants overproducing phosphofructokinase (O), pyruvate kinase (D) or
phosphofructokinase and pyruvate kinase (䊐). 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.
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
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
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
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
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 2-
deoxyglucose-resistant mutant strains of Aspergillus niger in semi-solid culture, Applied
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
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
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,
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
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
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
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
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
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
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
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, tm, can be used. The relative
mixing time is equal to the mixing time divided by the circulation time (tm = 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 tm for both reactors. As shown in this figure, by increasing tm, 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 tm
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
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
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
Figure 5.8 Time course for the mean length of filaments under different glucose levels in fed-
batch 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
Figure 5.9 Time course for the perimeter of the clumps under different glucose levels in fed-
batch 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.
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
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 l-
1
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.
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
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.
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.
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.
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
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
6
Redox Potential in Submerged Citric

Acid Fermentation
MARIN BEROVIC
Nomenclature
a,b constants (-)

ae electron activity (-)
aox activity of oxidized form (-)
ared activity of reduced form (-)
k redox reaction balance constant (-)
n number of electrons in redox reaction (-)
EO2/H2O standard potential—1223 mV (mv)
Eh potential measured in a solution, (mV)
based on standard hydrogen electrode
Eo standard redox potential of a 50 per cent reduced substance (mV)
based on standard hydrogen electrode
F Faraday constant
N stirred speed (s-1)
Qg volumetric gas flow rate (vvm)
pO2 dissolved oxygen partial pressure (atm)
pO2crit critical dissolved oxygen partial pressure (atm)
P citric acid concentration (g/l)
R gas constant (cal/°C mol)
rH negative log of partial pressure of gaseous hydrogen (-)
S sugar concentration (g/l)
T temperature (°C)
X biomass concentration (g/l)
t residence time (s)
Y yield factors (-)
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
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 Ç 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++ 2e- « H2
is a standard one. The oxidation potential is related to the electron activity in solution:
Eh = -RT/F lnae (6.2)
Eh = -RT/F ln[(1/k)1/n (ared/aox)] (6.3)

Eh = kRT/nF + RT/nF ln(ared/aox) (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 KCl-
gel 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 (6.8)
A relatively easier method is to use ascorbic acid at various pHs (Hewitt, 1950):
Eh ascorbic acid = 375 - 60 pH (6.9)
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 micro-
organisms 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
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
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
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
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
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
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
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
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
Table 6.1 Stirred tank reactor dimensions
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 micro-
organism 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.
Table 6.2 Scaling up redox profiles
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.
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

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
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,
KJÆRGAARD, L, 1976. European Journal of Applied Microbiology, 2, 215.
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,
LENGEL, Z L and NYIRI, L, 1965. An automatic aeration control system for biosynthetic processes,
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
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
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 (7.2)

Cube root of growth linear with time (7.3
Growth linear with time Xt = X0 + ktt (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
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).
Table 7.1 Calculated values for tm and k (from Röhr et al., 1981)
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 (7.11)
rXc = µcXc + ktXb - kdXc + DXc (7.12)
rXd = kdXc - DXb (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).
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)
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).
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)
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.
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.
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
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 concentration (C-mol m-3)

K ATP consumption for polymerization of
biomass precursors (mol ATP (C-mol DM)-1)
mATP specific maintenance requirements of ATP (mol ATP (C-mol DM)-1 h-1)
mS specific maintenance requirements of
substrate (C-mol (C-mol DM)-1 h-1)
mO specific maintenance requirements of oxygen (mol O2 (C-mol DM)-1 h-1)
N moles ATP generated per 1 C-mol of
substrate by substrate level phosphorylation (mol ATP (C-mol)-1)
NP moles ATP generated per 1 C-mol of
substrate before the biomass or product
formation diverges from the catabolic
pathway (mol ATP (C-mol)-1)
rate of ATP consumption (mol ATP m-3 h-1)
rate of ATP consumption in maintenance
processes (mol ATP m-3 h-1)
rate of ATP conversion of compound j (C-mol m-3 h-1)
maximum true yield of biomass on ATP (C-mol DM (mol ATP)-1)
maximum true yield of product on ATP (C-mol DM (mol ATP)-1)
yield on available electrons (C-mol DM (mol e-)-1
true yield on available electrons (C-mol DM (mol e-)-1
yield of ATP on substrate (moles ATP
generated per 1 C-mol of substrate from the
catabolic breakdown reaction) (mol ATP (C-mol)-1)
yield of ATP on substrate (moles ATP
required for maintenance) (mol ATP (C-mol)-1)
yield of oxygen on substrate (mol O2 (C-mol)-1)
yield factor for compound j on compound i (C-mol (C-mol)-1)
121
Y¢i,j true yield factor for compound j on

compound i (C-mol (C-mol)-1)
YC upper limit of Yi,j based on carbon
availability (-)
YE upper limit of Yi,j based on energy
availability (-)
YH upper limit of Yi,j based on reducing potential (-)
d P/O ratio (mol ATP (0.5 mol O2)-1)
g¢j generalized degree of reduction of
compound j (-)
sj weight fraction of C in compound j (-)
h fraction of available electrons transferred to
the biomass (-)
e fraction of available electrons transferred
to oxygen (-)
xP fraction of available electrons transferred
to products (-)
Subscripts
O oxygen
P product
X biomass
S substrate
max maximum
pre biomass precursors
r 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 macro-
chemical 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
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):
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
d = 2.17 mol ATP (0.5 mol O2) (Roels, 1983; Garret and Grisham, 1995), while a mean
value of the true biomass yield on ATP, YAmTaP,X
x
, of 0.371 C-mol DM (mol ATP)-1, and the true
max
citric acid yield YAT P,X = -6 C-mol citric acid (mol ATP)-1 (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):
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):
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 C-
mol 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.
¢
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:
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
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:
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-
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 bio-
technological 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.
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
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 tri-
calcium 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 an-
hydrate 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
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
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:
Figure 9.2 Flowsheet of the simplified non-citrate method of citric acid separation and
purification
• 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 carbon-
bonded 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.
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 non-
citrate 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, water-
insoluble 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,
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
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
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
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.
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,
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,
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.
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
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
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
Table 10.4 Content of vitamins in beet molasses (from Smirnow, 1983)
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.
Table 10.6 Composition of cane molasses (from Smirnow, 1983)
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.
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 ethylene-
diamineacetic 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
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).
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.
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 macro-
elements. 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
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.
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 specific area (m2 m-3)

A area (m2)
B permeability coefficient (m2)
Cn (n = 1 – 5) constants
Cp specific heat capacity (J kg-1 °C-1)
dhole pore diameter (m)
D diameter (m)
D liquid diffusivity (m2 s-1)
e fractional voidage (-)
g acceleration due to gravity (m s-2)
h bed height (m)
hf film heat transfer coefficient (W m-2 °C-1)
HD dispersion height (m)
J flux (m3 m-2 s-1)
k thermal conductivity (W m-1 °C)
kL liquid side mass transfer coefficient
at the gas–liquid interface (m s-1)
K fluid consistency index (Pa s)
K² Kozeny constant (-)
lp pore length (m)
L liquid height in the reactor (m)
L characteristic length in dimensionless numbers (m)
N impeller speed (s-1)
Np power number (-)
P power input (kW)
DP pressure drop (N m-2)
Q volumetric gas flow rate at NTP (m3 s-1)
r filtration resistance of the filter medium (m-1)
R universal gas constant
161
T temperature (°C)
u superficial gas velocity (m s-1)
V volume (m3)
Greek letters
a ratio of maximum hydrostatic head to the (-)

pressure at liquid surface
or; average specific cake resistance (m-1)
em membrane porosity
s interfacial tension (N m-1)

s dynamic viscosity (Pa s)
f gas hold-up (-)
y density (kg m-3)
mf fraction of filter area immersed (-)
µ viscosity (N s m-2)
w mass of dry solids in a filtrate volume V (kg m-3)
Abbreviation and dimensionless numbers
BOD biological oxygen demand (mg 1-1)

Bo Bond number gDrr/s
De Deborah number ug(1+f)l/fds
Flg aeration number Q/ND3
Fr Froude number N2D/g
Ga Galilei number gDR3/heff
HMF hydroxy-methyl-furfural
NF nanofilter
Pr Prandtl number Cpµ/k
PFD process flow diagram
RCS readily carbonizable substrate
Sc Schmidt number µ/rD
SVC standard variable cost of production
UF ultrafiltration
VVM volumetric flow of air per unit reactor volume per minute (min)
Subscripts
a apparent
d downcomer
D total area
eff effective
g gas
i impeller
L liquid
o unaerated
Design of an industrial plant 163
r riser
R reactor
T 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.
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?
Table 11.1 Technical data required
• 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
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
Table 11.2 Details to demarcate with a formal project scope definition

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
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 (aDR3), 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 DT 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
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).
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 a, 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, µw, can be taken to be equal to the
viscosity at zero biomass concentration.
• a and ß are positive and generally in the range, 0.25 < a, 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.
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 a, ß and d
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)a(ug)ß | scale 1 = (Pg/V)a(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 (11.5)
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.
11.9.7 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.
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 rate-
limiting 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 (f) and effective viscosity (heff) 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:
D Ptotal = DPsparger + DPhead (11.7)
while the energy input can be calculated from (Deckwer, 1985):

Pg = Qrg[RT ln(1 + a) + 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, a, is the ratio of the maximum hydrostatic head, to the pressure
at the liquid surface:
a = rL(1 - f)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.
f = C3ug (11.10)
while Mashelkar (1970) proposes:
f = (ug/rL)/(30 + 2ug)(72/s)-1/3 (11.11)

-1
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:
f = 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: g = 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:
hf = 1380(ug)0.22/(Pr)0.5 (11.13)
In the case of airlift reactors, Chisti (1989) quotes two equations:

hf = 8.71(Ar/Ad)0.25(ug)0.22/(Pr)0.5 (11.14)
and
hf = 13.34(1 + Ar/Ad)-0.7(ug)0.275 (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:
kLaa(ug)n (11.16)
or:
kLaa(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)
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.
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/dq, to the pressure drop across the filter, DP, 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 d: a = a’DPd, where a’ 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 d = 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: ADyf. If AD is the total filter area
and yf 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.
Table 11.3 Classification of ultrafiltration and nanofiltration
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)
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):
DP = -(u h µ)/B (11.23)
where:
B = [1/K²][e3(S2(1 - e)2)] (11.24)
B, is the bed permeability coefficient, while the Kozeny constant, K², 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.
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- < Br-
These 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).
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
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.
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.
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-jet-
bioreactor, 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.
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 scale-
down: 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,
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 cation exchange 138, 155

absorptive purification 181 cell removal 177
activated carbon 139 cellulose 158
acyl CoA synthetase 38–9 centrifugation 139
adenine nucleotides 43 chemical plant design 163
adsorption 142 citrate synthase 41–4
aerated systems 172 citrate-free recovery 139 flowsheet 140
aeration 5 citric acid applications 8 biochemistry 11
agitation effects 70 biosynthetic pathway 12–14 chemical
air-lift bioreactor 129, 131, 174 method 2 continuous production 5 koji
alcohol dehydrogenase 37 process 7, 58 microbial 2 overproduction
alcohols 37 129 regulation 19–21 solid state process 7
aldehyde dehydrogenase 37 submerged process 4, 58 surface method 3
aliphatic alcohols 141 synthetic 2 transport 24–5, 44–6, 49–50
alkali metal salts 138 uses 7, 8 yeast based processes 6 yeast
alkanes 6, 157 uptake 35 synthesis from alkanes 35–46 yeast
alkylsuphoxides 141 synthesis from glucose 46–50
alternative oxidase 123 cloned genes 16
ammonia 20, 158 coagulating agents 139
anion exchange resins 142 colour removal 179
approaches to design 174 compartmentation 42
arabinose 151 constant specific power input 172
aspergillus model 107 copper 81
available electron balance 130 counter current flow 143
axial impellers 173 crystallisation 136, 138, 182
crystallization stages 181
beet molasses 151–3 microelements 152 cube root growth model 106, 107
vitamins 153 amino acid content 152 customer information in design 164
betaine 151 customer requirements in design 163
bipolar membranes 144 cyclic AMP 21
blackstrap 153 cytochrome P-450 hydroxylase 36
bubble column reactor 174
butanol 139 date syrup 156
design of an industrial plant 163
cane molasses 153–5 vitamins 153 design basis of an industrial plant 165–7
composition 154 design constraints 173
caramel 151 di-calcium citrate 136, 137
carbon balance 130 dilution rate 111
carbon content of substrates 150 dimensional analysis 176
Carmen-Kozeny equation 179 dissolved oxygen 4, 23
188 Index
downstream processing 135–46 kinetic modelling 105

Kjærgaard equation 87
economics 1 K a 172
L
effective viscosity 175 Kozeny constant 179
effluent 8, 183
electrodialysis 138, 144, 181 scheme 145 lemon 1
electron activity 87, 88 lime 136–7
elemental composition 124 linear growth model 106–7, 118
elementary balances 122 liquid membranes 143
energy balance 121 log growth model 106–7
energy consumption 144 logistic growth equation 113
energy yield coefficients 128 Ludeking-Piret equation 107–8, 113, 119
equilibrium distribution coefficient 141
esters 141 magnesium 81
ethers 141 manganese 21, 22, 25, 59–60, 81
ethylenediamineacetic acid 155 mannose 151
evaporation 182 market 2, 9
exchange potential 180 marketing information 164
mass balance 121, 126
fats 158 mass transfer 171
fermentation aspects of design 170 mass transfer correlations 176
filter types 177 mass yield coefficient 122
fixed bed filter 142 mechanistic models 105
formaline 156 melanoidines 151
four compartment cell 144 membrane applications 178
Froude number 173 metabolic control analysis 61–2, 106
fructose-2, 6-bisphosphate 19 metabolic description of A. niger growth 123
metathesis reaction 180
gas hold-up 175 methyl citrate 40–1
genes 15–19, 44 methyl isocitrate 40–1
geometric similarity 172 microelements 158
gluconic acid 12, 58 microfiltration 178
glucose oxidase 12 microporous hollow fibres 143
glucose uptake 20 mitochondria 22
glutamate 49 mixed biomass 111
glycerol 22, 49 modelling 105–19
glycolytic pathway 12, 58 molasses 150–6 colour 151 microflora 154
glyoxylate cycle 39 non-volatile compounds 151 pH 153
growth kinetics 127 volatile compounds 151 nitrogen
compounds 151, 153
Hagen-Poiseuille equation 178 Monod equation 106, 111
heat removal 3 monopotassium phosphate 158
heat transfer 170, 175 morphology 5, 23, 60 A. niger 71 dissolved
hydrocarbons 149 oxygen 81 effect of carbon source 75 effect
hydrol 157 of inoculum 82 effect of nutritional factors
74 effect of pH 79 initial glucose
concentration 75 nitrogen limitation 78
idiophase 117, 119, 123 phosphate level 78 trace metal levels 81
immobilization 5, 145 mutagenesis 56
initial conditions model 113 mycelium formation 126
inoculum 4
ion exchange 180–1, 138–9, 142
iron 81 NADH oxidation 23, 62–3
isocitric acid 59 NADH: ubiquinone oxidoreductase 63
isocitrate lyase 41 nanofiltration 145, 178
neokestose 151
kestose 151 Nernst equation 87, 98
ketones 141 new technology in design 167
nitrogen metabolism 49
Index 189
nitrogen source 158 reverse osmosis 145

nomenclature 85, 121, 126, 161–3 Rushton impeller 173
non-aerated power input 171 Rushton turbine 173
non-ionic resins 142
non-Newtonian behaviour 175 scale down 173
scale up 101, 170, 172
oils 158 scope definition 167–8
organophosphorus compounds 141 seeding 138
oxalate biosynthesis 15 shear 70
oxalic acid 13, 58–9 removal 135 shear rate 173, 175
oxidation potential 87, 8 simple structured model for A. niger 107
oxidative phosphorylation 124 solvent extraction 139–41, 181
oxygen yield coefficient 128 starch 58, 157
oxygenation 3 sterilisation 155
stirred tank reactor 71, 170 design 171
parasexual cycle 60 stoichiometry 124
pectinolytic enzymes 145 structured models 105
pellets 5, 69 submerged method 150
pentose phosphate pathway 12 substrate level phosphorylation 128
peroxisomes 38–40 substrate preparation 169
pH 3, 6, 23, 24 substrates 149
phase related model 117 sucrose 151, 153
phosphofructokinase overexpression 64 sulphamide 156
phosphorus source 158 superficial gas velocity 176
Poiseuille equation 177 surface method 150
polyhydric alcohols 131 syrups 156
polysulphone 145
potassium ferricyanide 91 technical data in design 165
potassium ferrocyanide 155 teratogenic effect 142
power law 175 three-compartment cell 144
precipitation 136 flow sheet 137 transcriptional regulation 15–16
pre-treatment 144 submerged fermentation transport costs 150
135 surface process 135 treatment of molasses 155
process economics 149 therapies-6-phosphate 20, 64
process flowsheet 169 tricalcium citrate 136, 137
process package 167 tricarboxylate transporter 22
product drying 183 isolation 176 packaging tri-n-butylphosphate 141
183 trioctylphosphine oxide 141
purification 178 trophophase 117–18, 123
pyruvate kinase overexpression 64 tubular loop reactor 71
QSSA 125 ultrafiltration 139, 145–6, 178

unstructured models 105
raffinose 151, 153
raw materials 169 vacuum evaporation 138
redox dyes 88
redox electrodes 88 calibration 88 wastes 136
redox potential 85 in citric acid fermentation water 159
91–5 measurement 88 optimal 102 water-soluble amines 141
regulation 95 significance 89 theory 87
time course 92–3 dissolved oxygen xylose 151
relationship 90 temperature change 94
redox regulation chemical 97 physical 97–100
refined sucrose 156 yeast based models 113
regime analysis 173 yeasts 6
regulatory enzymes 18 yield coefficients 106–7, 113, 125–9
regulatory network 14
respiratory chain 62, 63 zinc 81

Citric Acid Biotechnology PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Citric Acid Biotechnology PDF

Uploaded by

Copyright:

Available Formats

Citric Acid Biotechnology

Borregaard Industries Ltd, Norway

Department of Bioscience and Biotechnology, University of Strathclyde, UK

Gluppevelen 15, 1614 Fredikstad, Norway

This edition published in the Taylor & Francis e-Library, 2002.

Copyright © Taylor & Francis 1999

British Library Cataloguing in Publication Data

Library of Congress Cataloguing-in-Publication Data are available

Cover design by Jim Wilkie

ISBN 0-203-48339-1 Master e-book ISBN

1 A Brief Introduction to Citric Acid Biotechnology page 1

2 Biochemistry of Citric Acid Accumulation by Aspergillus niger 11

3 Biochemistry of Citric Acid Production by Yeasts 33

4.3 Isolation of recombinant strains using the parasexual cycle in A. niger 60

6 Redox Potential in Submerged Citric Acid Fermentation 85

7 Modelling the Fermentation Process 105

8 Mass and Energy Balance 121

9 Downstream Processing in Citric Acid Production 135

9.3 Solvent extraction 139

10 Fermentation Substrates 149

11 Design of an Industrial Plant 161

Jacobus van der Merwe

A Brief Introduction to Citric Acid

MICHAEL MATTEY AND BJØRN KRISTIANSEN

1.1 Citric acid from lemons

Figure 1.1 Synthesis of citric acid

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

Citric acid is used in food, confectionery and beverages, in pharmaceuticals and in

Table 1.1 Applications of citric acid

1.10 Effluent disposal

Regardless of the method of production the disposal of waste is an increasing problem

Books must follow science, not science books.

Biochemistry of Citric Acid

MARKUS F.WOLSCHEK AND CHRISTIAN P.KUBICEK

2.2 Glucose catabolism in A. niger and its regulation

2.2.1 The citric acid biosynthetic pathway

2.2.2 Transcriptional regulation of the citric acid synthesizing pathway

2.2.3 Glucose metabolism and its regulation

2.3 Regulation of citric acid biosynthesis

2.4 Role of citrate breakdown in citrate accumulation

2.4.1 Role of the citric acid cycle

2.4.2 Respiratory activity and the role of NAD regeneration

2.5 Export of citric acid from A. niger

BRINDLE, P K, HOLLAND, J P, WILLET, C E, INNIS, M A and HOLLAND M J, 1990. Multiple

KIRIMURA, K, MATSUI, T, SUGANO, S and USAMI, S, 1996. Enhancement and repression of

MACHIDA, M, GONZALEZ, T V J, BOON, L K, GOMI, K and JIGAMI, Y, 1996. Molecular

PAPAGIANNI, M, MATTEY, M and KRISTIANSEN, B, 1994. Morphology and citric acid

SCHMIDT, M, WALLRATH, J, DÖRNER, A and WEISS, H, 1992. Disturbed assembly of the

WALLRATH, J, SCHMIDT, J and WEISS, H, 1991. Concomitant loss of respiratory chain

Biochemistry of Citric Acid

been advanced through developments in the understanding of protein targeting. It is rarely

3.2 Synthesis of citric acid from n-alkanes 3.2.1 Growth on alkanes

1 Uptake of alkanes into the cell.

3.2.2 Uptake of alkanes

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

3.2.3 Initial oxidation of n-alkanes