Microbial Reactions [Reprint 2022 ed.] 9783112620748, 9783112620731

Citation preview

Microbial Reactions H. W. Doelle, K. N. Ewings, N. W. Hollywood

Regulation of Glucose Metabolism in Bacterial Systems P. L. Rogers, K. J. Lee, M. L. Skotnicke, D. E. Tribe

Ethanol Production by Zynomonas mobilis S. Aiba

Growth Kinetics of Photo synthetic Microorganisms L. T. Fan, Y.-H. Lee, M. M. Gharpuray

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis

AKADEMIE-VERLAG BERLIN

Microbial Reactions

Microbial Reactions Managing Editor: A. Fiechter

with 63 Figures and 47 Tables

Akademie-Verlag • Berlin 1983

Die Originalausgabe erscheint im Springer-Verlag Berlin—Heidelberg—New York als Volume 23 in der Schriftenreihe Advances in Biochemical Engineering

Vertrieb ausschließlich für die DDR und die sozialistischen Länder Alle Rechte vorbehalten © Springer-Verlag Berlin, Heidelberg 1982 Erschienen im Akademie-Verlag, DDR-1086 Berlin, Leipziger Straße 3—4 Lizenznummer: 202 • 100/554/82 Gesamtherstellung: VEB Druckerei „Thomas Müntzer", 5820 Bad Langensalza Umschlaggestaltung: Karl Salzbrunn Bestellnummer: 763 225 9 (6761) • LSV 1345, 1315 Printed in GDR DDR 88,— M

Managing Editor Professor Dr. A. Fiechter Eidgenössische Technische Hochschule, Hönggerberg, CH-8093 Zürich

Editorial Board Prof. Dr. S.Aiba

Department of Fermentation Technology, Faculty of Engineering, Osaka University, Yamada-Kami, SuitaShi, Osaka S65, Japan

Prof. Dr. B. Ätkinson

University of Manchester, Dept. Chemical Engineering, Manchester/England

Prof. Dr. J. Boing

Röhm GmbH, Chem. Fabrik, Postf. 4166, D-6100 Darmstadt

Prof. Dr. E. Bylinkina

Head of Technology Dept., National Institute of Antibiotika. 3a Nagatinska Str., Moscow M-10S/USSR

Prof. Dr. H. Dellweg

Techn. Universität Berlin, Lehrstuhl für Biotechnologie, Seestraße 13, D-1000 Berlin 65

Prof. Dr. A. L. Demain

Massachusetts Institute of Technology, Dept. of Nutrition & Food Sc., Room 56-125, Cambridge, Mass. 02139/USA

Prof. Dr. R. Firn

School of Chemical Engineering, Olin Hall, Ithaca, NY 14853/USA

Prof. Dr. S. Fukui

Dept. of Industrial Chemistry, Faculty of Engineering, Sakyo-Ku, Kyoto 606, Japan Wissenschaftl. Direktor, Ges. für Biotechnolog. Forschung mbH, Mascheroder Weg 1, D-3300 Braunschweig

Prof. Dr. K. Kieslich

Prof. Dr. R. M. Lafferty

Techn. Hochschule Graz, Institut für Biochem. Technol., Schlögelgasse 9, A-8010 Graz

Prof. Dr. K. Mosbach

Biochemical Div., Chemical Center, University of Lund, S-22007 Lund/Sweden Westf. Wilhelms Universität, Institut für Mikrobiologie, Tibusstraße 7—15, D-4400 Münster

Prof. Dr. H. J. Rehm Prof. Dr. P. L. Rogers

School of Biological Technology, The University of New South Wales. PO Box 1, Kensington, New South Wales, Australia 2033

Prof. Dr. H. Sahm

Institut für Biotechnologie, Kernforschungsanlage Jülich, D-5170 Jülich

Prof. Dr. K. Schügerl

Institut für Technische Chemie, Universität Hannover, Callinstraße 3, D-3000 Hannover

Prof. Dr. H. Suomalainen

Director, The Finnish State Alcohol Monopoly, Alko, P.O.B. 350, 00101 Helsinki 10/Finland

Prof. Dr. G. T. Tsao

Director, Lab. of Renewable Resources Eng., A. A. Potter Eng. Center, Purdue University, West Lafayette, IN 47907/USA

Table of Contents

Regulation of Glucose Metabolism in Bacterial Systems H. W. Doelle, K. N. Ewings, N. W. Hollywood

1

Ethanol Production by Zymomonas mobilis P. L. Rogers, K. J. Lee, M. L. Skotnicki, D. E. Tribe . . . .

37

Growth Kinetics of Photosynthetic Microorganisms S. Aiba . . . "

85

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis L. T. Fan, Y.-H. Lee, M. M. Gharpuray

155

Regulation of Glucose Metabolism in Bacterial Systems Horst W. Doelle, Ken N. Ewings, Neil W. Hollywood Department of Microbiology, University of Queensland, St. Lucia, Qld. 4067, Australia

1 Introduction 1.1 Glucose Transport 1.2 Electron Transport Systems 2 Effect of Oxygen on Glucose Metabolism (Pasteur Effect) 2.1 Catalytic Reactions of Phosphofructokinase 2.2 Allosteric Phosphofructokinase 2.2.1 Action of Effectors on Allostery 2.2.2 Structure of the Enzyme 2.2.3 Kinetic Modelling of the Reaction Mechanism 2.3 Non-allosteric Phosphofructokinase 2.3.1 Physical and Kinetic Properties 2.3.2 Structural Modifications 2.3.3 Role in Glucose Metabolism 2.4 Model for the Mechanism of the Pasteur Effect 2.5 Summary 3 Effect of Glucose on Glucose Metabolism (Crabtree Effect) 3.1 Observation on a Crabtree Effect in Bacterial Systems 3.2 Effect of Specific Growth Rate 3.3 Effect of Glucose Concentration 3.4 Bioenergetic Aspects 3.5 Models for the Mechanism of the Crabtree Effect 3.6 Summary 4 Conclusions 5 Acknowledgements 6 Nomenclature 7 References

2 2 3 5 5 7 7 8 10 12 12 13 14 15 22 23 23 24 25 26 28 29 29 30 30 31

In the past 10 years there have been rapid developments in the elucidation of the mechanisms of the Pasteur (or oxygen) and the Crabtree (or glucose repression of the respiratory chain) Effects in bacterial systems, which convincingly exhibit the difference between the regulatory mechanisms in yeast and bacteria. The presented review will demonstrate that the enzyme phosphofructokinase plays no role in the mechanism of the Pasteur effect and that there exists no glucose repression on biomass formation in bacteria under aerobic conditions. Endproduct formation is caused aerobically and anaerobically by an oversupply of NADH 2 , whereas biomass correlates to energy supply. This development indicates very strongly that the mechanism of the Pasteur effect may be reflected solely in the change of glucose uptake rates and must therefore be sought at the cell membrane. In regard to the Crabtree Effect, the question arises whether there exists such a mechanism in bacteria. The variability of the bacterial electron transport systems, the lack of cytochrome a as terminal oxidase together with bioenergetic investigations indicate that the Crabtree effect may give cause for an alteration but not for a cessation of respiratory activity.

2

H. W. Doelle, K. N. Ewings, N. W. Hollywood

1 Introduction The regulation of glucose metabolism in facultative anaerobic microorganisms is dominated by two entirely different phenomena, the Pasteur Effect and the Crabtree Effect. The first phenomenon relates to Pasteur's observation 11 that the addition of oxygen reduces the rate of product formation and glucose utilization. Over the past century (see 2 '), it has been found that anaerobic growth proceeds with a much smaller energy yield compared with that obtained during respiratory metabolism of glucose to C02 and water. In association with the difference in energy formation per mole of glucose degraded, the cell utilizes glucose at a slower rate aerobically in comparison to that consumed during anaerobic growth. This interplay between energy availability and glucose uptake resulted in a number of theoretical concepts of metabolic regulation, but not one has yet been able to account for the Pasteur Effect in its entirety. Most studies have been concerned with the role qf the enzyme phosphofructokinase as the regulator of glycolysis, which is reflected in the extensive reviews given of this enzyme in past 3 _ 1 0 ) . The second phenomenon relates to Crabtree's observation 111 that respiration rates in tumours grown on glucose were consistently lower than for those grown on xylose. He deduced from his experiments that glycolytic activity must exert a controlling effect on respiration. This phenomenon has since also been found to occur in facultatively anaerobic yeast and bacteria. It is distinct from the Pasteur Effect in that product formation proceeds in the presence of oxygen. Two main theories, both originating in studies using Saccharomyces cerevisiae have been proposed as the basis of the Crabtree Effect. The first theory reflects the observations 1 2 ' 1 3 ) that glucose inhibits the formation of certain enzymes and is a typical example of catabolite repression 14), whereas the second theory 15) suggests that the Crabtree Effect occurs due to a relationship between specific growth rate and metabolic patterns rather than the molecular nature of the substrate.

1.1 Glucose Transport Glucose added to the medium must cross the cell membrane to be metabolized. Facultative anaerobic bacteria have two dominant systems of glucose transport: a) the active transport or permease system 182) and b) the group-translocating phosphoeno/pyruvate-phosphotransferase system (PTS), whereas strictly aerobic bacteria appear to have solely or predominantly the permease system 1 6 , 1 8 1 ) Investigations into the transport of galactoside 25_27 > ) lactose 2 8 , 2 9 ) and alanine 30) and the role of electrical and membrane potentials 31 ~ 33 ' have produced convincing evidence for the existence of the hypothetical chemiosmotic energy-coupling mechan i s m 3 4 ' 3 5 S u b s t r a t e transport and energy production by the bacterial cell are therefore considered in terms of this particular mechanism. According to Mitchell's theory of chemiosmotic energy-coupling, the permease system 1 7 - 2 0 ' is regarded as an active transport system which moves molecules (substrate) across the cell membrane against a concentration gradient (for further reading see Ref. 21 _24) ). The energy required for such active transport system comes

Regulation of Glucose Metabolism in Bacterial Systems

3

from the membrane-bound ATPase activity, which itself is driven by the protonmotive force created by the aerobic electron transport system (see 1.2). The occurrence of this ATPase activity in the particulate fraction together with the effect of ATP or related energy-rich compounds on this activity led to the suggestion that the permease enzyme may be a membrane-bound galactoside ATPase controlled by the availability of ATP or related compounds such as cyclic AMP. The group-translocating PTS system, on the other hand, does not require ATPase activity and takes its energy for active transport predominantly from the hydrolysis of phosphoeno/pyruvate or other substrate phosphorylation reactions. The standard free energy of hydrolysis of phosphoeno/pyruvate is about —56.5 kJ ( = —13.5 kcal) per mol. The liberated phosphate group is transferred via the PTS components to the substrate, causing a chemical transformation of the substrate molecule (for detailed reading see Ref. 22) ), which concomitantly leads to its transport into the cell. Since the approximate values obtainable for the standard free energies of hydrolysis of the other components of the system P-HPr and P-enzyme III are —50.2 kJ ( = —12 kcal) per mol, the total free energy available from this system easily substitutes for ATP. This substrate phosphorylation system therefore replaces the oxidative phosphorylation system. Facultative anaerobes could therefore replace the predominant permease system under aerobic conditions with a predominant PTS system under anaerobic conditions. Apart from its sugar transport, the PTS system also plays a significant physiological role as it regulates the uptake of other carbohydrates 3 9 , 4 0 ) , flagellar synthesis, transmembrane sugar transport, adenylate cyclase (cyclic-AMP synthesizing enzyme) and catabolic enzyme synthesis 4 1 ~43). Of special significance is the effect on adenylate cyclase, since cyclic AMP takes part in the regulation of the synthesis of a number of electron transport chain components, which in turn are responsible for the protonmotive force and thus ATP formation during oxidative phosphorylation. This regulatory system of PTS is itself subject to regulation 44 ' since the addition of exogenous cyclic AMP reverses the inhibition by PTS on other carbohydrate uptakes provided the exogenous inducers of the particular transport systems are present. It would appear that cyclic AMP regulates the availability of ATP via oxidative phosphorylation for the active transport or permease system, whereas PTS uses its substrate-level phosphorylation for the group-translocating system. This corresponds with the observations that bacteria utilizing glucose aerobically via the Entner-Doudoroff pathway do not in general posses a PTS system, whereas the latter dominates in those microorganisms, which use exclusively the Embden-Meyerhof-Parnas pathway 44) . It can be envisaged that microorganisms, which switch from one to the other glucose utilizing pathway or using combinations of different pathways could or may have to change from one to the other glucose transport systems depending upon the environmental conditions and particularly the energy status of the cell. Details of the mechanisms responsible for the regulation of carbohydrate uptake in bacteria has been reviewed recently 44) .

1.2 Electron Transport Systems Escherichia coli is able to synthesize a variety of redox carriers depending upon the growth phase, carbon source, strain and growth conditions 3 4 , 4 5 ~ 48) . Apart from

4

H. W. Doelle, K. N. Ewings, N. W. Hollywood

NADH dehydrog

QH2>ie H Q H

Out

H+-«

'

F

e

QH'Fe J

0

' 1/2 S + H +

-H+

< C o t e dehydrog

cyt b ^ c y t b ^

- c

• 1/2 SH 2

In

e

>

, Cytochrome o e - cyt d cyt h

,1/4 0j+ H 1/2 H 2 0

Fig. 1. Electron transport chain of E. coli arranged in the form of the Mitchell proton motive ubiquinone cycle S5) . Abbreviations see Nomenclature. (Reprinted with the permission of the authors and the American Society of Microbiology)

ubiquinone-8, benzoquinone and menaquinone-8, at least nine different cytochromes have been characterized 4 6 , 4 8 ) . The membrane-bound redox carriers include ubiquinone-8, cytochrome b 556 , b 562 , o and d. The latter two serve as terminal oxidases 4 9 - 5 1 ) . Respiration-driven proton translocation is coupled to the oxidation of NAD(P)linked or flavin-linked substrates in aerobically grown intact cells of Escherichia coli. Stoichiometric measurements of -» H + / 0 ratios provide levels of 4 for L-malate and 2 for the oxidation of succinate and D-lactate 52) . Assuming that 2 H + are required per ATP molecule synthesized by ATPase 33) , the electron transport chain is considered to be organized into two equivalent conservation segments 53,54 *, one of which is specifically associated with the NADH dehydrogenase (EC 1.6.99.3) region of the respiratory chain and the other with the cytochrome region ^K Since each of these segments translocates 2 protons per pair of transferred reducing equivalents, these results indirectly confirm that ATP synthesis via the reversible, proton-translocating ATPase occurs with an H + / A T P ratio of approximately two. These results are 55) illustrated in Fig. 1 and take into account both the proposed electron transport chain system 56) and Mitchell's proposed protonmotive ubiquinone cycle 57 • 58) . The relationship between respiratory composition and the efficiency of respirationlinked proton translocation of nine bacterial species of widely differing taxonomic and ecological status has been recently investigated 54) . Efficiencies were found ranging from 4 to 8 mol H + per g-atom of oxygen consumed. Clearly, the different patterns of respiratory chain composition 5 4 , 5 9 ) , which are exhibited by those organisms are responsible for the wide variations in proton-translocating efficiency. In Escherichia coli, which is a facultative anaerobe, not less than five proton-translocating redox segments have been identified 34) . If each of these segments is considered as a distinct building block, then, dependent on growth conditions, each block can function either separately or sequentially allowing alterations and parallel pathways to exist in the membrane for the reoxidation of reduced coenzymes. In energy production terms it appears that, under aerobic conditions, the protonmotive force generated by the respiratory chain reverses the ATPase catalysis from ATP hydrolysis to ATP synthesis, which in turn is coupled to inward proton

Regulation of Glucose Metabolism in Bacterial Systems

5

translocation. Under anaerobic conditions, ATP is gained from substrate-level phosphorylation and a second system functions in the formation of a protonmotive force at the expense of ATP hydrolysis. A third system exists in the form of the grouptranslocating system, where energy is derived directly by phosphoryl-group transfer from phosphoeno/pyruvate.

2 Effect of Oxygen on Glucose Metabolism (Pasteur Effect) 2.1 Catalytic Reactions of Phosphofructokinase Phosphofructokinase (ATP: D-fructose 6-phosphate 1-phosphotransferase, EC 2.7.1.11) catalyzes the transfer of« the y-phosphoryl group of ATP to D(—)-fructose 6-phosphate and produces D(—)-fructose 1,6-bisphosphate and ADP. The enzyme requires Mg2"1" for the reaction, because MgATP 2 ~ is the true substrate 6 0 ~62>. The reaction is essentially irreversible and characterized by a large negative value of free enthalpy 63) . Phosphofructokinase (PFK) appears to exhibit little substrate specificity in relation to the phosphoryl donor, with ATP being replaced by all of the nucleotide triphosphates, although greater affinity is shown for the purine nucleotides in bacteria 6 0 , 65) . Unlike the mammalian enzyme, information concerning the substitution of other phosphoryl acceptors has been recorded only in one other instance. In extracts of Streptococcus faecalis 66) , the PFK-mediated phosphorylation of sedoheptulose 7-phosphate to sedoheptulose 1,7-bisphosphate was recorded. ATP and citrate are the two major metabolic effectors, which cause inhibition of the mammalian and yeast enzymes 67 ', but not of the phosphofructokinase from procaryotes. An examination of Table 1 indicates that phosphofructokinase from most bacteria show no response to ATP or citrate. However, there have been some reports that have indicated a response to ATP in the form of inhibition, e.g. E. coliM'68'69), 70) Enterobacter aerogenes , Clostridium perfringens and Staphylococcus aureus71). There is a strong possibility that most of these latter examples are, in fact, allosteric enzymes, which can exhibit an inhibitory effect on the enzyme velocity substrate curves with ATP at low levels of fructose 6-phosphate. Reports of ATP inhibition were made with Mg 2 *: A T P 4 - ratios often less^than 2:1, which can indicate either the binding of free A T P 4 - or a lack of the substrate MgATP 2 ". It is highly improbable that ATP could exist as A T P 4 - in the intracellular environment, due to the strong chelating ability of the triphosphate group. Blangy et al. 60) made reference to this problem during their detailed kinetic studies of phosphofructokinase in E. coli by stating that M g 2 + : A T P 4 - ratios should be at least 10:1 to avoid complex inhibition. As pointed out by Hofmann 4 ', sigmoidality of the fructose 6-phosphate isotherm is not necessarily the result of the inhibitory action of ATP. Most of the allosteric bacterial phosphofructokinases are inhibited by phosphoeno/pyruvate and activated by ADP. There are two exceptions to this rule arising from studies on the allosteric enzymes from Clostridium pasteurianum 65) and Lactobacillus plantarum23). Neither phosphofructokinase from these sources are affected by phosphoeno/pyruvate. The clostridial enzyme was subject only to levels of fructose

H. W. Doelle, K. N. Ewings, N. W. Hollywood

6

h N Cl c4fc «

P? Z

< < I < Z I -4

X s« s S £5 S

Regulation of Glucose Metabolism in Bacterial Systems

7

6-phosphate and the relative concentrations of ATP and ADP. The latter produced a strong activation. Non-allosteric phosphofructokinases have been reported from a number of bacteria, such as Lactobacillus casei72), Lactobacillus plantarum 23-72>; Arthrobacter crystallopoietes13), Streptococcus faecalis 74) , and E. coli75 ' 7 6 ) . Since the enzymes from these sources do not have a sigmoidal rate dependence on fructose 6-phosphate, they may not have a regulatory role in glucose metabolism. As an example, it was shown 77) that L. casei has an allosteric lactate dehydrogenase subject to fructose 1,6bisphosphate activation and ATP inhibition, which displaces the pivotal role of phosphofructokinase in this organism to lactate dehydrogenase.

2.2 Allosteric Phosphofructokinase 2.2.1 Action of Effectors an Allostery Most bacterial phosphofructokinases are allosteric due to the cooperative subunit interactions that occur in the presence of fructose 6-phosphate. An important aspect of allosteric enzymes is their response to a number of metabolic effectors. In order to analyzç this response in a quantitative manner, investigators make use of the Hill equation as applied to the allosteric concept of Monod et al. 78) . From this approach a useful measure of subunit interactions, or cooperativity is given by the interaction coefficient, n H (Fig. 2). The interaction coefficient is affected differently by various effectors. The action of the allosteric effectors ADP and phosphoeno/pyruvate on the E. coli enzyme weakens considerably the kinetic interaction of fructose 6-phosphate 60) . ADP is a positive effector, shifting the fructose 6-phosphate velocity curve to the left, decreasing the nH value from 3.8 to a level of about 1.0. In fact, ADP not only influences the cooperativity of several bacterial phosphofructokinases in the same manner, it also appears to act as a competitive inhibitor of ATP for the enzyme 6 0 , 6 5 ' 7 0 , 7 9 ) . Phosphoewo/pyruvate, on the other hand, acts in a converse manner by shifting the fructose 6-phosphate velocity curve to the right, and increasing the nH value to 4.0. This is an example of a negative effector, and results in a low affinity of phosphofructokinase for its substrate fructose 6-phosphate. Amongst the lactic acid bacteria only the phosphofructokinase of Lactobacillus acidophilus shows allostery. This enzyme is activated by fructose 1,6-bisphosphate 23) . ADP does not influence the value of the interaction coefficient, but fructose 1,6-

Fig. 2. Diagrammatic Hill plot

8

H. W. Doelle, K. N. Ewings, N. W. Hollywood

bisphosphate has an effect by decreasing the nH value to 0.8 and shifting the fructose 6-phosphate isotherm to the left. NH^ ion activation also appears to change the interaction coefficient of the L. acidophilus enzyme, with a reduction from 3.3 to 2.2. The extreme thermophile, Thermus sp. X-l possesses a phosphofructokinase, which is analogous to the E. coli enzyme in its response to metabolite effectors 80) . The pH of the medium plays an important role in the ATP inhibition of phosphofructokinase and implicates the hydrogen ion as a metabolic effector. Results with the yeast enzyme show that an increase of the pH value from 6 to 8 shifts the sigmoidal fructose 6-phosphate velocity curve to the right, with a corresponding increase in the inhibitory action of ATP. At pH 6.0, ATP is less inhibitory and fructose 6-phosphate possesses a higher affinity for the enzyme than at higher pH values 8 1 _ 8 3 ) . Amongst the bacterial sources of phosphofructokinase, only the allosteric enzyme from L. acidophilus shows any pH effect 23) . At pH 8.5, the saturation curve for fructose 6-phosphate is markedly sigmoidal in the presence of ATP, with an nH value of 2.5. Sigmoidality is reduced at low pH values to an nH value of 1.1 at pH 6.0. Since the grade of cooperativity of fructose 6-phosphate is affected by the pH of the assay medium, and increases at alkaline pH, it is interesting to reflect on the versatility of the enzyme in L. acidophilus. The normal environment for the lactic acid group of bacteria is acidic, and phosphofructokinase is non-allosteric at pH 6.0, so the intracellular pH may have a regulatory effect on metabolism. L. acidophilus is also found in the alkaline intestinal environment, and could be expected to behave similarly to the allosteric enzyme from E. coli, which does not show any sigmoidal change below pH 8.0 60) . 2.2.2 Structure of the Enzyme In contrast to the phosphofructokinase from yeast 8 4 _ 9 2 ) , the prokaryotic enzyme appears to be considerably smaller with an oligomeric molecular weight of about 140000, consisting of four identical subunits of about 35000 23•. Clostridium pasteurianum phosphofructokinase has an oligomeric molecular weight of 144000 and is dissociated to subunits of 35000 in 7 M guanidine or 8 M urea 65) . . The amino acid analysis shows that the enzyme contains 34 arginine and lysine residues

Regulation of Glucose Metabolism in Bacterial Systems

9

Table 2. The amino acid composition of Escherichia coli, Clostridium pasteurianum and Thermus sp. X-l phosphofructokinases 80>. (Reprinted with permission of the authors and Academic Press Inc.) Category

Basic residues Arginine Lysine Acidic residues Aspartate Glutamate Ambivalent residues Alanine Half-cysine Histidine Serine Threonine Tryptophan Tyrosine Hydrophobic residues Isoleucine Leucine Methionine Phenylalanine Proline Valine Glycine Hydrophobicity (cal/residue) Hydrogen bonding residues (%)

Number of residues Thermus sp. X-l

E. coli

C. pasteurianum

145 74 71 227 109 118 324 108 12 40 52 76 4 32 394 109 90 20 30 35 110 169 1043 33

140 84 56 254 133 121 346 120 26 27 61 63 4 45 449 116 101 44 42 40 106 157 1091 35

133 70 63 220 108 112 269 92 17 10 58 70 0 22 405 102 103 54 31 24 91 151 1057 34

per 35000. Tryptic peptide mapping revealed 35-85 peptides, and 15 arginine-containing peptides were observed, which is in close agreement with 18 arginine residues per 35000. These results suggest that the clostridial phosphofructokinase is a tetramer consisting of four identical subunits. Homogenous Thermus sp. X-l phosphofructokinase has a molecular weight of 132000 and this molecular weight decreased to 34000 in 6 M guanidine 80) . In a comparative study of the enzyme from the extreme thermophile with phosphofructokinase from the mesophile E. coli and CI. pasteurianum, Cass and Stellwagen 80> use the amino acid composition of the respective enzymes to determine the reason for the extreme heat stability of the Thermus sp. X-l phosphofructokinase. The amino acid composition (Table 2) indicates that the hydrophobicity and the contents of potential hydrogen bonding residues do not differ to any significant degree. The amino acid composition grouped as acidic, basic and hydrophobic, or ambivalent classes of residues are very similar, varying by no more than 3 mol % for a given class among all the enzymes. These results predict a high degree of sequential and structural homology, and suggest that a small number of amino acid differences can account for substantial differences in structural stability. Clearly hydrogen binding alone cannot account for the additional stability in the phosphofructokinase from

10

H. W. Doelle, K. N. Ewings, N. W. Hollywood B A

Z

A B

Fig. 3. Schematic drawing of Bacillus stearothermophilus phosphofructokinase Each subunit is shown divided into two domains, numbered 1 and 2. The four subunits are related by three (X, Y, Z) orthogonal dyad axes (D 2 symmetry). The positions of some of the sites A, B and C are marked, showing that sites A and C lie between subunits. Dark shade'd bands are regions of P-sheeting. (Reprinted with the permission of the authors and from Macmillan Journals Ltd.)

the extreme thermophile. Although it has been calculated 51) that only 5-10 kJ per mol of additional free energy is required to stabilize an enzyme structure from 37 to 60 °C, which could be provided by hydrogen bonds or salt bridges. The structure of phosphofructokinase from Bacillus stearothermophilus originally investigated by Hengartner and Harris 94) and Hudson et al. 67) , was resolved into a tetrameric protein of molecular weight 136000 with a subunit molecular weight of 33900. A very recent and elegant study on the orthorhombic crystals of the purified enzyme makes use of X-ray diffraction and electron density mapping procedures to obtain a complete structural determination 99) . The enzyme consists of two domains (Fig. 3), numbered 1 and 2, each of which has a central p-pleated sheet sandwiched between a-helices. Each subunit forms contacts with only two of the other subunits in the tetramer. All of the subunits are related by three orthogonal dyad axes (I222 symmetry), which coincides with the crystallographic symmetry axes. Apparently there is a solvent filled hole of approximately 7 A diamter through the center of the tetramer along the Y-axis, indicated by a cylinder in Fig. 3. The positions of some of the sites A and B (active sites) and C (effector site) are marked. Sites A and C were found to lie between subunits and Table 3 indicates the compounds found to bind these sites. An interesting finding is that each subunit in the tetramer has an interface with two other subunits, one of these being bridged by fructose 6-phosphate, the other by the effectors. In addition, the region between the active site and the effector site contained amino acid side chains pointing in one direction to the fructose 6-phosphate site and in the opposite direction to the effector site. Binding of a ligand to either side, therefore, could affect the ligand binding affinity in the other site, which does indicate that both catalysis and control require the whole tetramer.

2.2.3 Kinetic Modelling of the Reaction Mechanism A reaction mechanism is extremely difficult to examine and to explain if some metabolites have the ability to act as activators or inhibitors. It was therefore necessary to establish an adequate model to explain the mechanism of sigmoidal kinetics. The first model was a concerted symmetry model 7 8 ) , which, assumes that the enzyme phosphofructokinase exists in two different conformational states, R and T, considered to be in equilibrium (Fig. 4). The kinetic data fit a model in which two states have equal affinity for ATP but differ with respect to their affinity for fructose 6-phosphate, ADP and phosphoerco/pyruvate. Although this concerted symmetry model turned

Regulation of Glucose Metabolism in Bacterial Systems F 6P

R

11

ADP

F6P

T

RT

Ti

PEP Fig. 4. Two state symmetry model of the reaction mechanism of E. coli phosphofructokinase 7 8 )

ATP, E>

L AMP

R2

ATP, S)

AMP r

ATP | E )

T2

Fig. 5. Four state kinetic model of the reaction mechanism of yeast phosphofructokinase printed with permission of the authors and Elsevier (North Holland Biomedical Press))

100)

. (Re-

out to be appropriate for the phosphofructokinases of E. coli60\ Clostridium pasteurianum 65), Thermus sp. X-l 80> and Bacillus stearothermophilus 67) , it could not explain the allosteric kinetics of the enzyme from Lactobacillus acidophilus 23) . The main problem was the dual action of fructose 1,6-bisphosphate as an activator and as a negative effector. Since the phosphofructokinase of E. coli exhibits an analogous problem with respect to the binding of ADP, it is not clear why these phenomena cannot be incorporated into Monod's model. The other alternative to the concerted symmetry model is the four state kinetic model of yeast phosphofructokinase 100), which involves conformational adjustments of the enzyme. In this model (Fig. 5) the basic conformational states, R and T, are degenerated into two isomeric subconformations, R 2 and T 2 . Fructose 6-phosphate binds only and Tj with different affinities, but not at all to R 2 and T 2 . ATP as a substrate binds to all four states with equal affinity. The inhibitor ATP E , however, binds only to R 2 and T 2 with the same affinity. Inversely, the activator or positive effector AMP has the same affinity for only the R! and T 2 states. In this model then, cooperativity from fructose 6-phosphate is generated by shifting the equilibrium between T, and Rj. Therefore, the action of ATP E merely shifts the saturation curve for fructose 6-phosphate to the right, and that of AMP to the left without any effect on the equilibrium between R t and T 2 . Consequently, the cooperative interaction of fructose 6-phosphate is not affected by these effectors. This model does not include, however, the complex actions of ADP, which not only inhibits the enzyme in competition with ATP, but also activates it synergistically with AMP 6 8 , 8 2 ) . A modification of this model has been suggested by Laurent et al. 1 0 1 ) after demonstrating that the inhibitory action of ATP occurred over a much larger range of ATP concentrations (10~6 to 10~3 M) than predicted by a concerted transition. Their results, utilizing fluorescent stopped flow and titrimetric measurements, was consistent with a concerted transition of the enzyme in the presence of fructose 6-phosphate, but the regulation by the adenylates appeared to involve a more complex mechanism involving non-concerted, sequential and conformational adjustments of the enzyme. At the moment there are clear indications, that the bacterial phosphofructokinase mechanism, although not as complex as that of the yeast enzyme, can be explained by the Monod model.

12

H. W. Doelle, K. N. Ewings, N. W. Hollywood

2.3 Non-Allosteric Phosphofructokinase 2.3.1 Physical and Kinetic Properties Since the initial report appeared 601 on the presence of a phosphofructokinase activity in crude extracts of E. coli, which eluted earlier than the allosteric form on DEAEcellulose, contrasting evidence has arisen concerning the physical and kinetic properties of this enzyme from E. coli. Polyacrylamide gel studies on crude extracts showed the enzyme to be a dimer of molecular weight 180000 M ) . Kotlarz and Buc 95) , who first succeeded in the isolation and purification of the enzyme from the original mutant strain of Morrissey and Fraenkel 102) , reported the presence of a dimeric molecule with a molecular weight of 71000. A dimeric protein of 65-67000 molecular weight has also been isolated from the wild-type E. coli K-12 103 ' 104) . In contrast to these results, Babul 105) isolated a tetrameric protein with a molecular weight of 140000 from the original mutant strain. Recent work in our laboratory 106) verified the presence of a tetrameric non-allosteric phosphofructokinase in crude extracts of the wild-type E. coli K-12. It became apparent that the discrepancies were due to the properties of a facile protein and that molecular size studies utilizing polyacrylamide gels and crude enzyme extracts with associated protein aggregation can lead to erroneous conclusions 98) . The non-allosteric phosphofructokinase from L. plantarum is a tetramer possessing identical subunits each of 38000 molecular weight 23) . In its dimeric form, the E. coli enzyme has an isoelectric point (pi) of 5.1 at 4 °C and a calculated half-time (t 1/2 ) of inactivation at 50 °C of 2.49 minutes, in comparison to a pi of 5.2 and a t 1/2 of 1.7 minutes at 75 °C for the allosteric isozyme. ATP inhibition was originally considered to be a regulatory phenomenon with all phosphofructokinases, including that of E. coli68'69). Early attempts characterized the two isozymes in E. coli as an ATP-insensitive' dimer and an ATP-sensitive' tetramer 64 • 107) . Detailed kinetic studies have shown that ATP does not inhibit the E. coli enzymes 6 0 , 7 6 , 9 5 ' 1 0 4 ) . Therefore the two phosphofructokinase isozymes would be more correctly labelled as allosteric and non-allosteric with regards to the second substrate fructose 6-phosphate. The non-allosteric isozyme can utilize tagatose 6-phosphate at approximately 40 % of the normal forward reaction rate obtainable with fructose 6-phosphate 105) . There is also a tendency to utilize a variety of nucleotide triphosphates in substitution for ATP with little specificity between the purine or pyrimidine base 104). This property is particularly evident with Mn 2 + and to a lesser degree in the presence of Mg 2 *. The Vmax values for ATP in the presence of Mn 2 + are 2-fold greater in magnitude than in the presence of Mg2"1", indicating a Mn 2 + -mediated activation of the enzyme. Free manganese also decreases the K m for fructose 6-phosphate. In addition, the rate of catalysis is dependent on the MnATP 2 - complex and not on free Mn 2 + . An interpretation of the kinetic data indicated that manganese complexed as MnATP 2 " was a prerequisite for enhanced catalysis, and by comparison a more efficient substrate than MgATP 2 ~. Although this is the first evidence of a manganese-mediated activation for a bacterial phosphofructokinase, it will not be possible to conclude a positive binding of free Mn 2 + to the enzyme without an accurate assessment of possible enzyme-Mn 2+ ternary binding complexes. A recent study has demonstrated an activation of yeast phosphofructokinase by MnATP 2 - 108).

Regulation of Glucose Metabolism in Bacterial Systems

13

The non-allosteric phosphofructokinase from E. coli has two pH optima at pH 6.8 and 8.4 95) , in contrast to the single optimum at pH 8.4 for the allosteric isozyme. Subtrate velocity curves at either pH, in the presence of either Mn 2 + or Mg2 + , do not deviate from the normal Michaelis-Menten kinetics. No further metabolic effector has been found that can modify the rate of catalysis for the enzyme in E. coli. The L. plantarum enzyme is activated by ammonium ions 23) and potassium ions 72) , but no inhibitor has yet been found. 2.3.2 Structural Modifications The non-allosteric phosphofructokinase from E. coli has been purified to homogeneity from both a mutant 1 0 5 , 1 0 9 ) and a wild-type strain 104) . Babul 105) utilized the increased affinity of the enzyme for the Blue Dextran chromophore, Cibracon Blue 185) , following pretreatment with hydroxylapatite to obtain a tetramer. In each of the other reports, which describe more prolonged periods of isolation procedures, a dimeric protein was produced. This suggested that some form of dissociation could occur during purification. Of particular interest is the affinity of the enzyme for Cibracon Blue in Babul's investigations. Although it was previously found that the enzyme does not bind to this ligand 95) , earlier studies revealed a partial binding to Blue Dextranlinked sepharose 103) using crude cell free extracts of E. coli. The resultant ATP elution of this enzyme produced a particularly facile, dimeric molecule of about 65000 molecular weight. Activity losses are high during purification procedures and only the substitution of the metabolites ADP, fructose 6-phosphate and ATP at a fixed concentration ratio of 1:5:10 mM is capable of restoring the activity to the original level. This effect is fully reversible upon dialysis in the presence of mercaptoethanol 103) . This proposed interconversion system is given in Fig. 6. The complete mechanism for this interconversion is still not clear. However, ultracentrifugation studies indicate that no association/dissociation occurs during the change from an inactive to an active dimer. Reactivation is totally dependent on the enzyme concentration, within the limits of 0.056 to 0.006mg m l - 1 protein 98 '. Unlike the interconversion process involving the urea-denatured yeast phosphofructokinase 110) , the non-allosteric bacterial enzyme can not be reactivated following urea dissociation. In crude cell free extracts, the enzyme is a tetrameric molécule which can be modified to bind to Cibracon Blue, either by the removal of phosphate as was suggested by Babul 105) , or by the combined action of heat and non-specific proteolysis 106) (Table 3). rapid

A

active

^ + ADP+F6-P g -ADP-F6-P i n a c t i v e

-ME

- B

+ ME +ME

inactive

slow

-AT

active

Fig. 6. Interconversion model of the non-allosteric E. coli phosphofructokinase 103) . (Reprinted with permission of the authors and from European Journal of Biochemistry)

14

H. W. Doelle, K. N. Ewings, N. W. Hollywood

Table 3. Apparent modification after selected treatments on the crude form of the non-allosteric phosphofructokinase from Escherichia coli K-12. The molecular weights were determined by gel filtration using Sephacryl S-200 unless otherwise indicated, n = number of tests; * = Cibracon Blue; ** = minor peak; (a) = sucrose density gradient centrifugation; N.D. = not determined 106). (Reprinted with permission of the authors and CSIRO Editorial and Publication Service) Fraction

Molecular weight x 103

Amount bound to CB*-Sepharose 6B

(%) Cell-free extract DEAE-cellulose DEAE-cellulose plus 50 °C at pH 8.4 DEAE-cellulose plu pH 5.0 DEAE-cellulose plus 50 °C at pH 5.0 CB*-Sepharose 6B plus 50 °C DEAE-cellulose, peak 2 (allosteric protein)

140 ± 2 120 ± 0.8 (a) 140 ± 1.0

4.5; n = 5 13.3; n = 3

67 ± 0.4 (140)** N.D.

87.1; n = 2 10.4; n = 2

140 ± 0.4 67 + 0.5 65 + 0.8 (a)

5.0; n = 2 92.4; n = 4

120 ± 0.9(a) 140 ± 1.0

98.5; n = 3

Incubation at 50 °C and pH 8.2 produces a dimeric molecule, which binds strongly to the dye ligand. The bound enzyme can be non-specifically eluted with high ionic strength sail,Solutions. An increase in temperature (50 °C) at pH 5.0 not only destroys proteolysis, but also results in neither a dissociation nor an increase in affinity for Cibracon Blue. When subjected to 3 M urea gel filtration, the crude tetramer dissociates to the dimer, indicating the electrostatic nature of the intermolecular forces binding pairs of tetramers. There is no doubt that the E. coli non-allosteric phosphofructokinase is susceptible to dissociation under periods of prolonged purification and modification following accelerated proteolysis in crude cell free extracts. However, neither influence on the quarternary structure of this protein is considered to be of a specific nature, but instead reflects an inherent weakness in the subunit assembly mechanism. 2.3.3 Role in Glucose Metabolism In the past it has been difficult to assign a functional role in glucose metabolism for the non-allosteric phosphofructokinase from E. coli. This has been due mainly to a low ievel of synthesis and an apparent lack of response to metabolic effectors. Based on early chromosomal mapping experiments for phosphofructokinase 102,111 ~ 1 1 3 ' 1 8 3 , 184) it is apparent that E. coli does possess two distinct genetic loci for synthesis of either of the two isoenzymes, a feature which appears to be unique to E. coli and to several species of enterobacteria 9 7 ' 1 1 4 ) amongst procaryotes. It is tempting to suggest that the second non-allosteric protein could be genetic remnant. However, there are characteristics that do suggest a useful purpose in glucose metabolism for this enzyme. This conclusion has been derived from the following: a) there is little discrimination between phosphoryl donors; b) the enzyme is active at a pH less than

Regulation of Glucose Metabolism in Bacterial Systems

15

7.0; c) the enzyme is activated by MnATP 2 - and possibly Mn 2 + ; and d) the enzyme is not inhibited by phosphoeno/pyruvate. What advantage can these properties confer on the non-allosteric enzyme in terms of a functional role in glycolysis? Manganese is an essential trace element required for the growth of E. coli in minimal media and therefore intracellular concentrations of this element are low in comparison to that of Mg 2 *. Experimental data would suggest that competition for the ATP 4 " species by both M n 2 + and Mg2"1" would be in favor of the MnATP 2 - complex formation, having an association constant of 10" 5 M 115) compared to 10" 3 M for the MgATP 2 " species 116) . If one takes into account the defined experimental conditions required to obtain these magnitudes, it is still feasible to predict a positive role for manganese. The definite lack of specificity observed towards each of the phosphoryl donors in the presence of Mn 2 + l 0 4 ) confers a decisive advantage to the cell despite the low intracellular levels of this cofactor. The ability to function as a catalyst during periods of low or neutral pH, a situation most likely to occur during glycolysis, confers a distinct advantage upon E. coli. Certainly the protein is not regulated in terms of allosteric feedback inhibition normally associated with the other isoenzyme, but a lack of response towards metabolic effectors such as phosphoeno/pyruvate and ADP should be viewed as being economically expedient for an organism. Large quantities of the constitutive nonallosteric enzyme are not synthesized at any stage because a more defined role is required. This role can only be one of a safeguard to ensure that carbon flow can continue via the Embden-Meyerhof-Parnas pathway during periods of metabolic change. The process of interconversion also suggests that this protein is capable of surviving periods of low intracellular metabolite concentration.

2.4 Model for the Mechanism of the Pasteur Effect The observation made by Pasteur, that in the presence of oxygen the cell yield from glucose was greater than that obtained under anaerobic conditions and that the presence of oxygen suppresses alcohol formation u is generally referred to as the Pasteur Effect. Oxygen controls the life of these facultative microorganisms by either stimulating or inhibiting their metabolic functions. One of the main functions of oxygen under aerobic conditions is to act as a final hydrogen acceptor in the electron transport chain mediating in the supply of energy required for growth. A reduction in oxygen supply can affect the electron transport chain, since the withdrawal of the final electron acceptor results in a smaller difference in the redoxpotential between the dehydrogenated couple and the final acceptor couple. In the absence of oxygen, the hydrogen from the metabolized couple is transferred to another organic compound of the same or similar redoxpotential, using N A D + / N A D H + H + as mediator. Such shortening of the electron transport chain means, of course, a reduction in ATP production 117) , which is reflected in a loss of biomass formation. The relationship between energy and growth 118) can be determined using the adenylate energy charge formular 119) : ATP + |

2

ADP

ATP + ADP + AMP

H. W. Doelle, K. N. Ewings, N. W. Hollywood

16 12

20

50

30 40 INPUT p0 2

60

Fig. 7. Effect of input p 0 2 on the respiratory activity of E. coli K-12. O , N A D H dehydrogenase; cytochrome a 2 (d); cytochrome b,; succinate dehydrogenase; A , isocitrate dehydrogenase; A , 2-ketoglutarate dehydrogenase 121) . (Reprinted with permission of the authors and the American Society for Microbiology)

70

During the transition from aerobiosis to anaerobiosis, oxidative phosphorylation and a significant amount of biomass is lost to the facultative anaerobe. The enzyme 2-ketoglutarate dehydrogenase 1 2 0 ) is repressed, leading to a branched biosynthetic pathway instead of the amphibolic tricarboxylic acid cycle. Investigations into the effects of dissolved oxygen tension on the respiratory enzymes in E. coli demonstrated the gradual collapse of the electron transport chain 121) . Cytochrome d and cytochrome bj were the first to show an increase to maximal value soon after the increase in oxygen affinity (Fig. 7) occurring at 5 mm Hg O z 1 2 2 _ 1 2 5 ) . As the partial pressure of oxygen approaches zero, the respiratory enzymes decreased and obtained a lower value than the corresponding

0

10

20

30 INPUT

40 p0 2

50

60

70

Fig. 8. The effect of input partial pressure of oxygen on the specific PFK activity ( O , mlJ per mg protein); cell yield ( • , g l - 1 ) , dissolved oxygen level ( # , DOT) and specific acid production rate ( • , N NaOH per biomass per time, ml g" 1 h" 1 ) in a glucose-limited chemostat culture of Escherichia coli K-12 with a dilution rate of 0.2 h _ 1 and pH 7.0 at 37 °C 130) . (Reprinted with permission of the authors and the Publishers Swets and Zeitlinger)

17

Regulation of Glucose Metabolism in Bacterial Systems

aerobic levels, with the exception of cytochrome bj. None of the respiratory enzymes, however, were completely repressed indicating that no stringent control system exists in the respiratory system. These results together with the observation that 2-ketoglutarate dehydrogenase synthesis becomes zero just prior to the increase in respiratory activity (Fig. 8), are in accordance with earlier observations of Amarasingham and Davis 120> , who suggested that the functioning of a branched tricarboxylic acid cycle occurs under oxygen limitation. Due to the location of the allosteric phosphofructokinase in the upper part of the EMP pathway, the irreversibility of the reaction, complex kinetics, and susceptibility towards a large variety of effectors, it is highly adapted to be rate-controlling for the substrate flux through the glycolytic or EMP pathway 126) . The underlying problem, concerning the inhibition of glycolysis by respiration is whether allostery can account for the observed decrease in glucose utilization during aeration. It is perhaps fortuitous, that the allosteric concept, i.e. that the specifically induced conformational changes which enable certain enzymes to act as chemical transducers, has enabled enzymologists to interpret metabolic regulation and in particular the 'Pasteur Effect' as a consequence of allosteric enzyme regulation. In Saccharomyces cerevisiae, a series of sequential feed-back control inhibitors on a number of allosteric enzymes has been postulated by Sols 127) , to account for the regulation of glycolysis. Sols postulates that there are metabolic pathways, two in anaerobiosis, and three in aerobiosis, crucial to the degradation of glucose. These routes, outlined in Fig. 9 are: 1) from extracellular glucose to glucose 6-phosphate 2) from glucose 6-phosphate to ATP and a carbon by-product in anaerobiosis, or to ATP and citrate in aerobiosis 3) from citrate to more ATP and C0 2 plus water. Each of these pathways contains an enzyme known to be controlled by allosteric feedback inhibition. The specific regulatory mechanisms are as follows: a) feed-back control of the NAD-dependent isocitrate dehydrogenase by its dependence on allosteric activation by AMP. When the energy level in the cell is high with a very low AMP/ATP ratio, the slowing down of isocitrate dehydrogenase activity allows the citrate concentration to increase, which increases the feedback inhibition of the preceding pathway; b) feedback inhibition of phosphofructokinase by ATP in anaerobiosis, and by ATP and citrate in aerobiosis. The allosteric control by ATP is reinforced by a strong

Glucose

I

ADP-AMP1

Fig. 9. An hypothetical regulatory mechanism for the Pasteur Effect in yeast systems with permission of the author and Pergamon Press Ltd.)

127)

. (Reprinted

18

H. W. Doelle, K. N. Ewings, N. W. Hollywood

allosteric activation by AMP. Feedback inhibition of phosphofructokinase raises the concentration of glucose 6-phosphate, which is feedback inhibiting the preceding pathway. c) allosteric feedback inhibition of hexokinase or of the catalyzed transport of glucose across the cell membrane by glucose 6-phosphate. In support of this mechanism is the finding by Banuelos et al. 1 7 6 ) that inorganic phosphate is an allosteric activator of phosphofructokinase from Saccharomyces cerevisiae. Although not included in the scheme, phosphate could be an important metabolite in situations where changes in the concentrations of phosphate occur. For example, inorganic phosphate levels do change during the shift from aerobiosis to anaerobiosis 128). Sequential feedback inhibition of allosteric enzymes completely ignores the possible role of enzyme synthesis in the regulation of the 'Pasteur Effect'. Model and Rittenberg 129) found that about 25 % of the glucose metabolized by E. coli during active growth in a minimal medium passes through the HMP pathway, and that this value remains constant for a range of glucose concentrations. The EMP pathway is responsible for the rest of the glucose metabolism. Anaerobic growth leads to lower HMP participation (7%) in glucose metabolism. It has been shown in continuous cultures of E. coli152), that the amounts of phosphofructokinase 107 ' 130) , aldolase 121) and pyruvate kinase 131) increase when the tension of dissolved oxygen in the medium falls below 5 mm Hg 0 2 . Similarly, in continuous culture of Saccharomyces cerevisiae, the specific activities of hexokinase, phosphofructokinase and alcohol dehydrogenase increase with the diminishing availability of oxygen 132) . On the other hand, amounts of the HMP pathway enzymes glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase remain constant during changes in the dissolved oxygen tension in cultures of E. coli121,133), and in Saccharomyces cerevisiae 132) . These results suggest that the synthesis of the enzymes of the EMP pathway is enhanced by anaerobiosis. In addition to changes in enzyme synthesis during the aerobic-anaerobic transition, it should be noted that facultative anaerobes can produce, for the reoxidation of reduced coenzymes, different alternative and parallel electron transport pathways, with different abilities to translocate protons and, thereby, to synthesize ATP. Extensive reviews have been compiled on this subject of bacterial respiration 34) , the regulation of its rate 134), electron transfer 135) and anaerobic energy conservation 136) . It is only necessary to emphasize that protein synthesis is also involved in the regulation of electron transfer depending on the oxygen concentration. From the preceding information it is clear that a change from aerobiosis to anaerobiosis can influence the synthesis of a variety of glycolytic and oxidative enzymes in microorganisms. The question is, can oxygen function as a corepressor or as a repressor? According to the work of Lilius 137) , it can do neither, since he demonstrated aerobically, growth of E. coli on glucose subjected to an anaerobic shock by the addition of the reductant sodium dithionite, produced an anoxic state that lasted for only 15 min in comparison to the increase in enzyme synthesis which lasted for longer than 1 h. Conversely, the addition of the oxidant persulphate, acting as an uncoupler of oxidative phosphorylation, to the actively growing culture, caused an increase in the utilization of glucose and a high respiration rate despite the presence of oxygen. Other regulators have been suggested. The redoxpotential of the

Regulation of Glucose Metabolism in Bacterial Systems

19

medium 138) , the effective intracellular redoxpotential 1 3 9 ) , the redox state of the components of the regulatory chain 140) and the response of the N A D + / N A D H + H + couple to the extracellular redox state 137) are all factors proposed to regulate enzyme synthesis. During anaerobiosis, organic compounds have to serve as electron acceptors and accordingly, E. coli responds by increasing the synthesis of lactate dehydrogenase, alcohol dehydrogenase and a variety of reductases. In a similar manner, the levels of the glycolytic enzymes are increased in order to satisfy anaerobic energy demands. These investigations suggest that the mechanism of the 'Pasteur Effect' cannot be fully explained by the catalytic and kinetic behaviour of phosphofructokinase alone, but the synthesis of the enzyme must play a vital role. Any description of the mechanism of the 'Pasteur Effect' in E. coli should include a number of experimentally proven observations, such as the presence of two phosphofructokinase isozymes, the allosteric and non-allosteric nature of each, the change in enzyme synthesis of the allosteric isozyme, the response to extracellular redoxpotential and a possible change in glucose transport across the cellular membrane. In gram negative bacteria, glucose is transported across the cell membrane by a group translocation process known as the phosphoeno/pyruvate: sugar phosphotransferase system (PTS). The chemical reaction catalyzed by the PTS system involves transfer of the phosphoryl moiety of phosphoeno/pyruvate to glucose in the presence of the enzyme components of the PTS and Mg2"1", to form the sugar phosphate and pyruvate. Phosphoryl transfer from phosphoeno/pyruvate to glucose requires the intermediate participation of several catalytic enzymes. Pertinent information on these enzymes has been adequately covered in a review by Saier 141 It is of interest to note that the PTS system appears to have the capacity to affect the rate of glucose utilization 1 4 2 ' 1 4 3 ) . Working with cultures of E. coli, Herbert and R o m berg 142) observed that, over a 5-fold range of growth rate, 14 C-glucose was transported at rates, which, above a threshold value, increased linearily with growth rate, and which corresponded to the rates at which glucose was utilized by the growing cultures. Since external glucose enters the metabolic pathways of E. coli predominantly via the PTS system 1 4 4 , 1 4 5 ) , it was concluded that it is the activity of the PTS system which sets the pace of overall glucose utilization. This has also been confirmed by Carter and Dean 146) on studies with strains of Klebsiella aerogenes, where it was shown that the activities of the PTS system were linearily dependent on growth rate, whereas hexokinase activities varied over a wide range. However, Neijssel and Tempest 1 4 7 ) found that the sudden addition of glucose to a glucose-limited culture of K. aerogenes or E. coli148) resulted in the immediate stimulation of respiration. This implies that the PTS system was not saturated before further provision of glucose was made, and hence, that it was not the pacemaker reaction for regulating the rate of cell synthesis, particularly at low dilution rates. Recently, Hunter and Kornberg 143) , utilizing mutant derivatives of E. coli impaired in one or the other PTS enzyme II component, showed that the increase in respiration was due to the increased PTS activity and paralleled the ability of the cell to take up glucose. They concluded by stating that the capacity of the PTS system is not only rate limiting for glucose utilization in the glucose-limiting chemostat, but is also adequate to account for the stimulation of respiration previously observed by Neijssel et al. 148) , when the glucose concentration in the medium is suddenly raised.

H. W. Doelle, K. N. Ewings, N. W. Hollywood

20

The glucose phosphotransferase system requires phosphoeno/pyruvate for activity, and therefore the intracellular level of phosphoerco/pyruvate must have a controlling effect on the transport of glucose, and the concomitant phosphorylation of glucose to form glucose 6-phosphate. Kornberg and Smith 149) used a stoichiometric argument as an explanation for phosphofructokinase mutants of E. coli which failed to grow on glucose, but grew well on glucose 6-phosphate. These mutants were impaired in the generation of phosphoeno/pyruvate via the EMP pathway but were capable of metabolizing glucose via the H M P pathway, where the phosphoeno/pyruvate yield is one molecule per molecule of glucose. Glucose 6-phosphate is transported independently of the PTS system 16) . Phosphoeno/pyruvate is the only potent allosteric inhibitor of phosphofructokinase in E. coli60). This inhibition of the allosteric isozyme is reversed by high levels of the substrate fructose 6-phosphate. Using mutants of E. coli lacking the allosteric isozyme and containing only the non-allosteric phosphofructokinase, Robinson and Fraenkel 1 5 0 ) indicated that the absence of the allosteric type of phosphofructokinase has little effect on the growth rate or the yield of this organism (Y ) when grown on glucose. Additional evidence that the allosteric phosphofructokinase plays no role in the mechanism of the Pasteur Effect comes from chemostat culture experiments using the same mutant 1 5 1 ) . This study (Fig. 10) demonstrated that the Pasteur Effect can occur with or without the allosteric phosphofructokinase and that a change in pathway utilization from H M P to EMP pathway was not necessary for product formation to occur. This regulation is brought about by any interference to the electron transport system which produces

0

20

40 INPUT

60 p02

100

200

Fig. 10. The effect of different steady states of partial pressure of oxygen on the growth ( • , g l - 1 ) and specific acid production rate (O, N NaOH per biomass per time, m l - 1 g"1 h" 1 ) of E. coli K-12 mutant DF 1000 ( ) and AM 1 ( )151). (Reprinted with permission of the authors and FEMS-Microbiology Letters)

Regulation of Glucose Metabolism in Bacterial Systems

21

Glucose

Q

I -G 6 - P

I

F 6-P" HMP

EMPc

if

F BP \

Biosynthesis + NADP Endproducts +NADH 2

• i I I I PEP'

V

Biosynthesis electron transport

Non-cyclic TCA ( Anaerobiosis) CyclicTCA (Aerobiosis)

Fig. 11. A hypothetical regulatory mechanism for the Pasteur Effect in bacterial systems 9 8 )

an oversupply or accumulation of N A D H 2 151) , as will also be demonstrated later in the effect of glucose on glucose metabolism. With all the evidence available it should be possible to postulate a mechanism for the 'Pasteur Effect' in E. coli. Such a mechanism is outlined in Fig. 11 and includes three stages: 1) the transient stage during which phosphoenolpyruvate is rapidly depleted to a lower level than that reached during anaerobiosis 2) the stage where the PTS system is decreased and glucose utilization is via the H M P pathway 3) the stage of complete oxidative metabolism and biosynthesis. The hypothesis which explains this model is as follows: Stage 1: (i) Oxygen causes a change in the extracellular redoxpotential, which induces a transient cessation of cellular growth but not metabolism. Cyclic A M P levels increase and trigger the initiation of enzyme synthesis. An unknown signal causes a repression of the glycolytic enzymes in the EMP pathway (ii) Fructose 6-phosphate is rapidly depleted, due to the remaining phosphofructokinase, which is then inhibited by phosphoeno/pyruvate. Stage 2: (i) The PTS system operates at a decreased level due to the H M P pathway reducing the level of phosphoeno/pyruvate via glucose 6-phosphate (ii) The E M P pathway is at its lowest activity.

22

H. W. Doelle, K. N. Ewings, N. W. Hollywood

Stage 3: (i) The oxidative enzymes of the tricarboxylic acid cycle and nascent electron transport chains reach a sufficient level for the production of increased levels of energy and protein synthesis (ii) Cyclic A M P levels drop. Implicit in this scheme is the role of the H M P pathway which is postulated to act in a dual role, a) as a constant supply of carbon and N A D P H + H + for cellular oxidative biosynthesis of nucleic acid and amino acids, and b) as a means of removing excess glucose 6-phosphate in order to reduce the level of phosphoeno/pyruvate. It is clear from this mechanism that three basic assumptions are made, 1) that the level of phosphoeno/pyruvate is the major factor affecting the rate of glycolysis, 2) the allosteric phosphofructokinase is transiently inhibited prior to repression and 3) de novo glycolytic enzyme synthesis will occur at a decreased level governed by the increase in the already available energy from oxidative phosphorylation. Cyclic A M P is implicated as a trigger for the inducement of enzyme synthesis 152 " 1 5 5 ) due to its known involvement in relieving glucose catabolite repression. However, the still unanswered questions are, in what form is the signal that actually initiated enzyme repression and derepression? Where in the order of cellular events does it occur, at the transcriptional or translational level? The occurrence of a 'Pasteur Effect' in the absence of the allosteric phosphofructokinase can also be explained by the hypothetical model. Since no additional phosphofructokinase synthesis occurs during the aerobic-anaerobic transition, EMP pathway activity remains low and carbon flow via the H M P pathway predominates. Any additional phosphoeno/pyruvate required for the PTS system of glucose transport under anaerobic conditions, is obtained by a redirection of the carbon flow from a pentose shunt system to the energy-yielding part of the EMP pathway leading to pyruvate. To do this, the synthesis of glyceraldehyde 3-phosphate dehydrogenase is increased 100-fold 15 6-phosphogluconate dehydrogenase 5-fold and AMP-activated pyruvate kinase synthesis is repressed in favour of the FBP-activated enzyme. Therefore, the mechanism of the Pasteur Effect can be sought for in a change in cellular energy arising as a consequence of the cessation of electron transport activity, together with a change in the rate of glucose transport across the cell membrane. Since the cessation of the electron transport activity and the rearrangement of glucose carbon flow towards repression can occur without invoking a biosynthetic control, it is proposed that regulation of glucose metabolism must occur at the cell membrane producing a change in the glucose uptake rate. The change from oxidative to substratelevel phosphorylation together with the increase in phosphoeno/pyruvate formation would favour the PTS system of glucose transport, which represses the formation of cyclic A M P and other glucose uptake systems. Such a change takes place at constant glucose concentration and constant specific growth rate and appears to be the only unique feature of the effect of oxygen on bacterial glucose metabolism.

2.5 Summary Perturbations in both synthesis and kinetics of the glycolytic enzyme phosphofructokinase have in the past implicated this protein with a key pivotal role in the Pasteur Effect. The latter is now known to encompass a number of mechanisms each being

Regulation of Glucose Metabolism in Bacterial Systems

23

interwoven in a series of complex reactions ; none of which can account for the process as a whole. The functional roles of both phosphofructokinase isozymes in E. coli are still important. Control by phosphoeno/pyruvate on the major allosteric protein allows an optimal flow of carbon during repression in direct response to the reduction of available ATP. In contrast the non-allosteric protein ensures a minor but necessary continuous carbon flow under diverse metabolic conditions. Further studies of the mechanism of the Pasteur Effect in microorganisms will need to consider all of the metabolic phenomena reviewed in this article with special attention to the changes in glucose uptake rates and glucose transport, before proceeding to the final and obvious unsolved question : What is the source and form of the initial signal that triggers such a complex mechanism?

3 Effect of Glucose on Glucose Metabolism (Crabtree Effect) 3.1 Observation on a Crabtree Effect in Bacterial Systems Investigations into the glucose effect in bacterial systems have not been well documented and were initially reported when Strasters and Winkler 156> noted that glucose enhanced glycolysis, suppressed the tricarboxylic acid cycle, decreased the activity of the pentose cycle (HMP pathway) and the cytochrome content in Staphylococcus aureus. Detailed investigations into the regulation of 2-ketoglutarate dehydrogenase synthesis in E. coli120) revealed that the synthesis of this enzyme was not only repressed under anaerobic conditions but also during the logarithmic phase of growth on glucose under aerobic conditions. This observation led to the conclusion that respiration was not a constant feature of aerobic metabolism. It was also noted 1 0 9 ' 1 5 7 ) , that glucose repressed the formation of TCA cycle enzymes in complex media more actively than in simple media. These results were explained by assuming that with an adequate glucose supply, enough ATP is available from the EMP pathway to minimize the role of the TCA cycle in producing energy. Under these conditions, the biosynthetic role of the cycle dominates. As this function is not required to the same extent in the presence of complex media, the full repressive effect of glucose would be visible. The effect of glucose on bacterial electron transport systems varies widely in its nature and extent and depends very much on the genus or species. A comparison between cultures of Salmonella typhimurium grown on glucose and on TCA cycle intermediates 158) revealed a repression of cytochrome synthesis in the glucose grown cultures. The distribution of quinones between large and small respiratory particles in E. coli was found to be different 1 5 9 ) depending on whether growth occurred on glucose or other substrates. Escherichia coli B harvested in the stationary phase of growth is capable of oxidative phosphorylation with an apparent efficiency of 3.5 + 0.3 moles phosphate esterified per mol of N A D H oxidized 160>. If the cells were harvested during the logarithmic phase, a considerably lower phosphorylation efficiency was observed. It was suggested that part of the enzymatic or structural apparatus of oxidative phosphorylation is subjected to catabolite repression. Detailed investigations into the effect of glucose on glucose metabolism started mainly with the development of the continuous culture system. It was this technique,

24

H. W. Doelle, K. N. Ewings, N. W. Hollywood

together with the introduction of the oxygen electrode, which enabled controlled experiments to exclude any possible interference with the Pasteur Effect. Using a turbidostat-controlled continuous culture technique 161) with a constant biomass at 300 mg dry weight per 1, a switch-over from respiration to repression occurred at a glucose input concentration between 0.15 and 0.2%. The start of repression coincided with an inhibition of TCA cycle activity owing to the repression of 2-ketoglutarate dehydrogenase and isocitrate dehydrogenase synthesis. Of the glycolytic enzymes assayed, only lactate dehydrogenase activity increased, whereas phosphofructokinase, fructose 1,6-bisphosphate aldolase, glucose 6-phosphate and 6-phosphogluconate dehydrogenases remained at a constant level of activity. These results led to the conclusion that the glucose effect may be mediated through regulation at the pyruvate level and does not cause a major shift in the glucose utilizing pathways. Since turbidostat-controlled continuous culture techniques involve the maintenance of a constant biomass by varying dilution rates, these results could be caused by either increasing specific growth rate 15) or by catabolite repression 14).

3.2 Effect of Specific Growth Rate At low specific growth rate (0.1 h" 1 ), the aerobic growth of E. coli is characterized by a fully functional tricarboxylic acid cycle, and a fully respiratory mode of metabolism with high levels of cytochrome d as the terminal oxidase 162). Increasing the dilution rate from 0.1 to 0.5 h _ 1 in a chemostat culture at three different glucose concentrations ranging from 0.1 to 0.5%, revealed 162 ' a proportional increase in the specific glucose uptake rate (Qg) and specific acid production rate. Product formation under conditions of high specific growth rate is therefore apparently brought about in a similar manner as described in yeast 15) , i.e. by a relationship between the specific growth rate, the rate of glucose consumption and subsequent equilibrium rates of metabolic routes rather than the molecular nature of the substrate. The increase in Q g values are associated with increases of lactate dehydrogenase and in particular glyceraldehyde 3-phosphate dehydrogenase synthesis. The latter increased approximately proportional to the specific acid production rate. All other enzymes of the EMP and HMP pathways tested e.g. phosphofructokinase, aldolase, glucose 6-phosphate and 6-phosphogluconate dehydrogenases were not affected. The increase of these two enzymes could indicate an increasing use of the lower part of the EMP pathway and possibly the methylglyoxal bypass 163) to pyruvate. These results conform with the observations that the PEP-phosphotransferase activity sets the pace of overall glucose utilization 142) , and that glucose utilization may be regulated by the initial uptake step rather than by the level of glycolytic enzymes 164). Associated with increasing specific growth rate and Q g values is a repression of the respiratory pathways 162) . The ratio of moles C 0 2 produced per mole of glucose utilized, the synthesis of 2-ketoglutarate dehydrogenase, succinate dehydrogenase, NADH dehydrogenase and cytochrome d decreased significantly, whereas the oxygen uptake rate (Qo2) increased and the glucose yield values (Yg) stayed constant (Table 4). The constancy of the Yg values indicate that ATP production is maintained despite the fact that a greater proportion of NADH 2 channels its reducing equivalents to the

25

Regulation of Glucose Metabolism in Bacterial Systems

Table 4. The effect of specific growth rate and glucose concentration on the molar growth yields during aerobic growth of E. coli K-12 162) . (Reprinted with permission of the authors and The Faculty Press) D h"1

Glucose input concentration ( %) 0.1

Y g (g dry weight per mole glucose utilized) 0.1 94 0.2 95 0.3 95 0.4 93 0.5 94

0.3

0.5

62 87 85 97 97

71 87 97 97 97

reduction of organic substrates rather than to the electron transport chain. The logical requirement of a greater efficiency of oxidative phosphorylation and the reduction in cytochromed synthesis at high specific growth rates indicates that electron transport occurs to a greater extent through the cytochrome o limb. The use of this limb causes the translocation of 4 H + per mol of NADH 2 oxidized and hence a doubling in the efficiency of oxidative phosphorylation. The observation of constant Yg values during aerobic growth forms the key difference between Escherichia coli and Saccharomyces cerevisiae. Unlike E. coli, S. cerevisiae possesses only one main cytochrome oxidase, the cytochrome a—a 3 complex 165). Growth at high specific growth rates causes both repression of oxygen uptake 15) and of cytochrome oxidase synthesis 166) , which leads to a decrease in energy production and consequently to a decrease in Y g values 167). These experimental observations indicate that the glucose effect in bacteria may be completely different from that existing in Saccharomyces cerevisiae.

3.3 Effect of Glucose Concentration Attempts to distinguish a glucose effect from a specific growth rate effect consisted of a chemostat technique, in which the influence of increased glucose input concentration at a constant specific growth rate on glucose metabolism was studied 162) . Using this technique, it was noted that high glucose concentrations tend also to repress parts of the electron transport chain through a mechanism distinct from that of the specific growth rate. At specific growth rates of 0.1 and 0.2 h - 1 , increases in glucose concentration lead to increases in specific acid production but do not affect Q g values. At higher specific growth rates of 0.3-0.5 h - 1 , a maximal acid production rate of approximately 10 ml g - 1 h _ 1 (expressed as NaOH and dry matter resp.) appears to exist, which is also obtainable at low specific growth rates and high glucose concentrations. Presumably, the effect on specific production values occurs through an effect of glucose on the proportion of glucose utilized to titratable endproducts. Similarly to the specific growth rate effect, no change in Y g values occurs and enzymatic analyses clearly established that no change in the glucose utilization pathways takes place. Although

26

H. W. Doelle, K. N. Ewings, N. W. Hollywood

increases in glucose concentration result in higher levels of acid production, a reduction in moles CO z produced per mole glucose utilized, repression of NADH dehydrogenase and succinate dehydrogenase, the levels of 2-ketoglutarate dehydrogenase, cytochrome d and the oxygen uptake rate are not affected. In contrast to the effect of specific growth rates, increased glucose concentrations did not affect the glucose uptake rate, the functioning of the tricarboxylic acid cycle and the cytochrome d. Similarities exist in the constancy of the Yg values, acid production and the repression of NADH and succinate dehydrogenase. If a greater proportion of the glucose metabolized forms glycolytic endproducts, then one must expect that a decreased proportion of glucose carbon would be used for cellular biosynthesis. The relative constant Y g values, however, do not indicate such change. This inverse relationship between moles C 0 2 produced per moles glucose utilized and specific acid production could indicate that the increased acid production could come from a shift in endproduct formation from carbon dioxide to acetate. Such a shift is possible as the repression of succinate dehydrogenase would slow down tricarboxylic acid cycle activity. Such a balance in endproduct formation would maintain the biosynthesis of cellular components and thus keep the Y g value constant. Despite this endproduct rearrangement, the same question as for the effect of specific growth rates exist: how can the electron transport via the cytochromes produce the oxidative phosphorylation or ATP required to maintain cellular biosynthesis despite the repression of NADH dehydrogenase and thus phosphorylation siteT.

3.4 Bioenergetic Aspects High glucose concentrations as well as specific growth rates repress NADH dehydrogenase and thus phosphorylation site I of the electron transport chain. Repression, therefore, occurs due to an accumulation of NADH 2 180). The constant Yg values obtained indicate that ATP production must be maintained despite the repression of phosphorylation site I. This observation with E. coli is in contradiction to the results obtained in yeast. Saccharomyces cerevisiae possesses only one main cytochrome oxidase, the cytochrome a—a 3 complex. Associated with electron transport via this oxidase is a translocation of 6 H + ions, and a resultant synthesis of 3 moles of ATP per mole of NADH oxidized 165) . Growth at high specific growth rates causes both repression of oxygen uptake 1 5 , 1 6 7 ) and a repression of mitochondrial and cytochrome synthesis 166 ' 168) , which results in a severe repression of oxidative phosphorylation. This decrease in energy production from oxidative phosphorylation leads to the reduction in Yg values noted during the growth of S. cerevisiae at high specific growth rates 167) . In E. coli, the constant Y g values obtained during aerobiosis under both conditions of high glucose concentration or high specific growth rates indicates that oxidative phosphorylation still proceeds. Evidence for the existence of oxidative phosphorylation comes from investigations, which involved the use of the uncoupler 2,4-dinitrophenol 180) . Under high glucose concentrations, the presence of 2,4-dinitrophenol increases Q g values suggesting that glucose transport can not be the rate-limiting step in aerobic cells as was also found to be the case in Klebsiella aerogenes 147 • 148,169) . This activation of Q g is also reflected in the oxygen uptake rate confirming reports on 2,4-dinitrophenol activation in cell respiration 170) using Pseudomonas stutzeri.

27

Regulation of Glucose Metabolism in Bacterial Systems

Table 5. The effect of 2,4-dinitrophenol on biomass formation and molar growth yields at different specific growth rates and glucose input concentrations during aerobic glucose metabolism of E. coli K-12 180) . (Reprinted with permission of the authors and The Faculty Press) D h"1

Glucose input concentration ( %) 0.1 + DNP

Dry weight (g 1 ') 0.1 0.48 0.2 0.51 0.3 0.48 0.4 0.49 0.5 0.49

— DNP

0.3 + DNP

— DNP

0.5 + DNP

— DNP

0.52 0.53 0.53 0.51 0.52

1.32 1.17 1.32 1.32 1.00

1.03 1.45 1.41 1.35 1.58

1.87 1.35 1.32 1.05 1.00

1.98 2.41 2.70 2.30 1.00

80 71 79 79 71

62 87 85 97 97

67 49 43 49 71

71 87 97 97 97

Y g (g dry weight per mole glucose utilized) 0.1 88 94 0.2 92 95 0.3 86 95 0.4 89 93 94 0.5 87

Since a correlation must exist between energy production and growth, the presence of 2,4-dinitrophenol should exhibit an effect on either YG or Y 0 values. In Escherichia coli180), the Y 0 values are found to be dependent upon the specific growth rate, whereas the maintenance requirement (qS 2 ) independent of the specific growth rate but dependent on glucose concentration. This increase in q™2 values with increasing substrate concentration is similar to the response observed with different substrates 1 7 1 ) . The effect of 2,4-dinitrophenol on biomass formation is insignificant at low glucose input concentrations irrespective of specific growth rate. If the glucose concentration was raised to 0.5%, a pronounced effect occurs with increasing specific growth rates. This uncoupler therefore exhibits a significant effect only at high glucose concentrations (Table 5). A similar effect occurs in relation to the oxygen uptake rate. The presence of 2,4-dinitrophenol therefore reduces the 'true yield' values (YSAX), the YG values and Qq 2 , but not qS2 with increasing glucose concentrations (Table 6 and Fig. 12). Table 6. The effect of 2,4-dinitrophenol on qg 2 and YS ,X at different glucose input concentrations during aerobic glucose metabolism of E. coli K-12 180) . (Reprinted with permission of the authors and The Faculty Press) Glucose input

%

0.1 0.3 0.5

vYm a x

qS2 1.2 3.1 3.3

o

DNP

+ DNP

-

DNP

1.0 3.2 4.4

38.4 71.4 166.6

+ DNP 21.2 55.5 100.0

28

H. W. Doelle, K. N. Ewings, N. W. Hollywood

12

6

10

5

8

L

6 . -4

_P

3 »"J ~ _P 2

u

0.1

0.2

0.3

0.4

0.5

0.1

D-

0.2 D

0.3

0.4

0.5

Fig. 12. Influence of specific growth rate and different glucose input concentration on the specific oxygen uptake rate in the absence (a) and presence (b) of 2,4-dinitrophenol 1 8 0 1 ( # , 0.1 %; O , 0.2% and • , 0.5% glucose). (Reprinted with permission of the authors and The Faculty Press)

3.5 Models for the Mechanisms of the Crabtree Effect The Crabtree Effect in Escherichia coli is manifested in a repression of phosphorylation site I of the electron transport chain, namely NADH dehydrogenase, and a repression of succinate dehydrogenase. This repression can either be caused by the metabolic rate of glucose utilization or increase of glucose concentration and does not affect oxidative phosphorylation. As glucose uptake rates increase proportionally to the specific growth rate, but remain constant with increasing glucose concentrations at constant specific growth rates, and Yg values remain constant in both cases, transport limitation can be excluded from the hypothetical model. The following mechanism for the Crabtree Effect in Escherichia coli has ben proposed 172) : a) Electron transport chain under the effect of specific growth rate: SDH

D-LDH I

NADH

FP 1 I FP —» Fe —> cyt. b 5 5 6 —• cyt. O I

organic endproducts

Q

cyt. b 5 5 8 —> cyt. d

> 02 02

b) Electron transport chain under glucose repression: SDH

D-LDH I

FP 4

NADH

FP —• Fe —• cyt. b 5 5 6 —> cyt. O I

organic endproducts

Q

cyt. b 5 5 8 —• cyt. d

> 02 • Oz

Regulation of Glucose Metabolism in Bacterial Systems

29

There is no doubt that the production of organic endproducts is caused by the accumulation of NADH 2 . The repression of phosphorylation site I, succinate dehydrogenase and in the case of increased specific growth rates, also of cytochrome d does not prevent oxidative phosphorylation to take place. The appearance and increase of D-lactate dehydrogenase with increased glucose concentration could be an indication for a replacement of succinate dehydrogenase by D-lactate dehydrogenase as additional electron entry point into the electron transport chain. During aerobiosis, the direction of carbon flow through the predominant HMP pathway does not change, but the increased glyceraldehyde 3-phosphate dehydrogenase suggests a pull towards pyruvate. This increased flow through the energyyielding part of the EMP pathway may supply additional phosphoeno/pyruvate for the PTS system. The increased accumulation of NADH 2 restricts carbon flow through the TCA cycle and switches pyruvate utilization towards acetate and ethanol. Escherichia coli therefore exhibits oxidative phosphorylation with the concomitant production of organic endproducts. This phenomenon raises doubts as to the use of the term fermentation 173) . It is finally the glucose concentration and not the metabolic rate, which terminates or limits biomass formation. This limitation of growth occurs at high glucose input concentrations and low specific growth rates, after a maximal value for the specific acid production rate is obtained 180). Further increases in glucose concentration or specific growth rate leads to non-utilized glucose in the medium. It is very unlikely that this is caused by glucose transport limitations, since the addition of 2,4dinitrophenol increases Q g and the oxygen uptake rate. It is rather thought that at this stage a carbon plus energy dual limitation 147 • 169) may exist with optimal overall growth efficiency under the conditions used. Anabolism and catabolism are evenly matched, which is supported by the earlier mentioned reports on the effect of 2,4dinitrophenol on YJ5ax values demonstrating the uncoupling of ATP synthesis and non-availability of energy for further growth.

3.6 Summary The term Crabtree Effect originally referred to a decrease in the respiratory activity of cells grown on glycolysable substrates compared to that found in cells grown on non-glycolyzable substrates. The present indications are that such a system does not exist in bacterial systems. However, some components of the electron transport chain are repressed by both specific growth rate and glucose concentration via separate mechanisms. This would indicate that neither the term Crabtree Effect nor the term catabolite repression would describe both the effect of specific growth rate and glucose concentration on the electron transport chain.

4 Conclusions The majority of reports are those concerning the facultative anaerobic bacterium Escherichia coli and one should be aware of the danger of generalization with respect to the basis of regulation. The present state of knowledge indicates, that the mechanism of regulation of glucose utilization is quite different from that reported for yeasts.

30

H. W. Doelle, K. N. Ewings, N. W. Hollywood

Studies on the effect of oxygen and glucose concentration on glucose metabolism demonstrate that the formation of organic endproducts, is not unique to the anaerobic state. Such formation can be caused by interferences with the electron transport chain, which lead to an oversupply or accumulation of NADH 2 . This phenomenon can occur under aerobic and anaerobic conditions. In contrast to anaerobic endproduct formation, the aerobic one is not necessarily accompanied by a decrease in ATP production. This means, of course, that microorganisms are able to produce organic endproducts from glucose under fully respiratory conditions in the presence of oxidative phosphorylation. Since it is the energy status of the cell, which causes a rearrangment of the carbon flow via the Embden-Meyerhof-Parnas pathway, organic endproduct formation occurs irrespective of the pathway used for glucose utilization to pyruvate. In withdrawing the final electron acceptor oxygen from the medium, a complete cessation of the electron transport system leads to decreased ATP production. This change in the energy status occurs in association with some unknown regulatory control system for the increased uptake of substrate. The evidence available appears to favour the involvement of cyclic AMP in this regulation. The molecular basis for such regulation is still unknown 186). Recent developments indicate that the isoenzymes of phosphofructokinase are not essential to the occurrence of the Pasteur Effect, nor is the EMP pathway of glucose utilization. Studies with phosphofructokinase mutants of E. coli have shown, that the mechanism of the Pasteur Effect can be sought for at the cell membrane level. The Crabtree Effect can be thought of as a two-fold mechanism involving both an effect of specific growth rates and a direct glucose concentration effect. The latter can be considered a specialized case of catabolite repression, whereas the former is one involving a regulation of metabolism by the respective rates of energy producing pathways. High metabolic rates are accompanied by a repression of TCA cycle activity although respiration continuous through-out by the use of alternate electron transport limbs.

5 Acknowledgements We wish to thank Professor Fiechter for his advice and his prepublished manuscript as well as all those which helped in the critical assessment of this manuscript. For that part of work, which has been carried out in our laboratory, we wish to thank most sincerely Professor Fraenkel for kindly sending us his mutants, the Australian Research Grant Commission, University Research Grant Commission and the Postgraduate Award Scheme for their continued financial support. We also acknowledge the work of all those postgraduate students, who were involved in various stages.

Regulation of Glucose Metabolism in Bacterial Systems

31

6 Nomenclature AMP ADP ATP Pi cAMP FBP PFK A A* I NE — PEP F 6-P G 6-P Yg Qg QO2 Y0 Qo2

adenosine monophosphate adenosine diphosphate adenosine triphosphate inorganic phosphate cyclic 3',5'-adenosine monophosphate fructose 1,6-bisphosphate phosphofructokinase activator (at inhibitory or non-inhibitory levels of ATP) absolute necessary for enzyme activity inhibitor no effect not determined phosphoeno/pyruvate fructose 6-phosphate glucose 6-phosphate biomass (gM-1) molar growth yield glucose glucose ( m M g - 1 h" 1 ) specific glucose uptake rate biomass x time oxygen ( m M g - 1 h" 1 ) specific oxygen uptake rate biomass x time biomass molar growth yield (gg _ 1 ) g atom oxygen I^YQ1

For Fig. 1: cyt.: cytochrome; SH 2 : reduced substrate; S: oxidized substrate; QH 2 : ubiquinol; QH*: ubisemiquinone; Q: ubiquinone; Q: ubiquinone reaction center; e: electron; NAD/NADH: nicotinamide adenine dinucleotide oxidized/ reduced form

7 References 1. Pasteur, L.: Compt. rend. Acad. Sei. 52, 1260 (1861) 2. Doelle, H. W.: Microbial Metabolism. In: Benchmarks in Microbiology, Bd. 5. Stroudsburg: Dowden, Hutchinson and Ross 1974 3. Bloxham, D. P., Lardy, H. A.: Phosphofructokinase, In: The Enzymes. Boyer, P. D. (ed.), Vol. 8, 239, New York: Academic Press Inc. 19733 4. Hofmann, E.: Rev. Physiol. Biochem. 75, 1 (1976) 5. Krebs, H. A.: Essays Biochem. 8, 1 (1972) 6. Mansour, T. E.: Curr. Topics Cell Regul. 5, 1 (1972) 7. Ramaiah, A.: Curr. Topics Cell Regul. 8, 297 (1976) 8. Stadtman, E. R.: Adv. Enzymol. 28, 41 (1966) 9. Tejwani, G. A.: Trends Biochem. Sei. 3, 30 (1978) 10. Uyeda, K.: Adv. Enzymol. 48, 193 (1979) 11. Crabtree, H. G.: Biochem. J. 23, 536 (1929) 12. Epps, H. M. R., Gale, E. F.: Biochem. J. 36, 619 (1942)

32 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.

H. W. Doelle, K. N. Ewings, N. W. Hollywood Monod, J.: Growth II., p. 223 (1947) Magansanik, B.: Cold Spring Harbor Symp. Quant. Biol. 26, 249 (1961) Beck, C., von Meyenburg, H. K.: J. Bacteriol. 96, 479 (1968) Romano, A. H. et al.: J. Bacteriol. 104, 808 (1970) Burstein, C., Kepes, A.: Biochim. Biophys. Acta 230, 52 (1971) Craven, G. R. et al.: J. Biol. Chem. 240, 246B (1965) Marr, A. G. et al.: J. Bacteriol. 87, 356 (1964) Zipser, D.: J. Mol. Biol. 7, 739 (1963) Harold, F. M.: Curr. Topics Bioenerg. 6, 83 (1977) Rosen, B. P. (ed.): Bacterial Transport, Marcel Dekker Inc. 1978 Simon, W. A., Hofer, H. W.: Biochim. Biophys. Acta 481, 450 (1977) Wilson, D. B.: Ann. Rev. Biochem. 47, 933 (1978) Cecchini, G., Koch, A. L.: J. Bacteriol. 123, 187 (1975) Purdy, D. R., Koch, A. L.: J. Bacteriol. 127, 1188 (1976) West, I. C., Mitchell, P.: J. Bioenerg. 3, 445 (1972) West, I. C.: J. Bacteriol. 122, 1162 (1970) West, I. C., Mitchell, P.: Biochem. J. 132, 587 (1973) Collins, S. H. et al.: J. Bacteriol, 126, 1232 (1976) Collins, S. H., Hamilton, W. A.: J. Bacteriol. 126, 1224 (1976) Hirata, H. et al.: Proc. Natl. Acad. Sci. U.S. 70, 1804 (1973) West, I. C., Mitchell, P.: FEBS-Lettr. 40, 1 (1974) Haddock, B. A., Jones, C. W.: Bacteriol. Revs. 41, 47 (1977) Mitchell, P.: Biol. Revs. 41, 445 (1966) Knopfel, H. P.: Ph. D. Thesis, Diss. No. 4906 E.T.H. Zurich (1972) Dietzler, D. N. et al.: J. Biol. Chem. 250, 7188 (1975) Dietzler, D. N. et al.: J. Biol. Chem. 250, 7194 (1975) Saier, M. H. Jr., Roseman, S.: J. Biol. Chem. 251, 6606 (1976) Winkler, H. H., Wilson, T. H.: Biochim. Biophys. Acta 135, 1030 (1967) Morris, D. M., Lawson, J. W.: Can. J. Microbiol. 25, 235 (1979) Mugharbil, U., Cirillo, V. P.: J. Bacteriol. 133, 203 (1978) Saier, M. H. Jr., Staley, J. T.: J. Bacteriol. 131, 716 (1977) Dills, S. S. et al.: Microbiol. Revs. 44, 385 (1980) Ashcroft, J. R., Haddock, B. A.: Biochem. J. 148, 349 (1975) Haddock, B. A., Schairer, H. W.: Eur. J. Biochem. 35, 34 (1973) Jones, C. W.: Symp. Soc. Gen. Microbiol. 27, 23 (1977) Shipp, W. S.: Arch. Biochem. Biophys. 150, 454 (1972) Castor, L. N., Chance, B.: J. Biol. Chem. 234, 1587 (1959) Haddock, B. A. et al.: Biochem. J. 154, 285 (1976) Pudek, M. R„ Bragg, P. D.: FEBS-Lettrs. 50, 111 (1975) Lawford, H. G., Haddock, B. A.: Biochem. J. 136, 217 (1973) Brice, J. M. et al.: Biochem. Soc. Trans. 2, 523 (1974) Jones, C. W. et al.: Eur. J. Biochem. 52, 265 (1975) Downie, A., Cox, G. B.: J. Bacteriol. 133, 477 (1978) Cox, G. B. et al.: Biochem. J. 117, 551 (1970) Mitchell, P.: Fed. Proc. 26, 1370(1967) Mitchell, P.: FEBS-Lettrs. 59, 137 (1975) Jurtshuk, P. et al.: C. R. C. Crit. Rev. Microbiol. 3, 399 (1975) Blangy, D. et al.: J. Mol. Biol. 31, 13 (1968) Mavis, R. D., Stellwagen, E.: J. Biol. Chem. 245, 647 (1970) Nissler, K. et al.: Acta Biol. Med. Germ. 36, 1027 (1977) Bohme, H. J. et al.: Acta Biol. Med. Germ. 34, 15 (1975) Doelle, H. W.: Eur. J. Biochem. 50, 335 (1975) Uyeda, K., Kurooka, S.: J. Biol. Chem. 245, 3315 (1970) Sokatch, J. R.: Arch. Biochem. Biophys. 99, 401 (1962) Hudson, P. J. et al.: Biochem. Soc. Trans. 5, 725 (1977) Atkinson, D. E., Walton, G. M.: J. Biol. Chem. 240, 757 (1965) Griffin, C. C. et al.: Biochem. Biophys. Res. Commun. 27, 287 (1967)

Regulation of Glucose Metabolism in Bacterial Systems

33

70. Sapico, V., Anderson, R. L.: J. Biol. Chem. 244, 6280 (1969) 71. Lowry, O. H., Passoneau, J. V. : Naunyn-Schmiedeberg's Arch. Exp. Pathol. Pharmakol. 248,185 (1964) 72. Doelle, H. W.: Biochim. Biophys. Acta 258, 404 (1972) 73. Ferdinandus, J., Clark, J. B.: Biochem. J. 113, 735 (1969) 74. Grillo, J. F., Vandermark, P. J . : Am. Soc. Microbiol. Abstr. No. P214, 176 (1973) 75. Doelle, H. W.: Proc. Austral. Biochem. Soc. 5, 30 (1972) 76. Fraenkel, D. G. et al.: J. Biol. Chem. 248, 4865 (1973) 77. Gordon, G. L., Doelle, H. W.: Eur. J. Biochem.

Characteristics

Z. mobilis

Z. anaerobia

Dimensions in microns

1.4-2.0x4.0 -5.0

1.0-1.5 x 2.0 -3.0

Occurrence

Fermenting plant juices

Tainted ciders, perries & beers

Ability to utilize sucrose

Utilizes sucrose and produces levan

Does not utilize sucrose

Cytochromes in aerobically

a 2 and c

b and c

grown cells In anaerobically grown cells

a 2 , b and c

b and c

Not required

Essential

Requirements for lipoic acid and biotin

(2) if motile, the motility is by means of \-A lophotrichous flagella (as a tuft or group, either at one or both ends of a cell) but motility is not an essential feature (3) cells occur singly or in pairs (4) cells are Gram-negative (5) no spores, capsules, intracellular lipids or glycogen; no resting stages known (6) anaerobic but can tolerate some oxygen (7) cells can convert glucose or fructose to nearly equimolar quantities of ethanol and C 0 2 by the Entner-Doudoroff pathway (8) sucrose is utilized as a carbon source by many strains and this is often accompanied by levan formation (9) a wide variety of other sugars and fatty acids are not utilized as carbon sources (10) the G + C content of the cellular DNA is about 47.5-49.5%. Carbohydrate Metabolism Glucose and Fructose Metabolism The ability of Zymomonas to metabolize glucose and fructose was reported first by Barker and Hillier 6) and then by Lindner 18). A molar equation for the conversion of glucose to ethanol was established by Kluyver and Hoppenbrouwers 27) as follows: 1 glucose 1.8 ethanol + 1.9 C 0 2 + 0.15 lactic acid. At that time, it was postulated that the mechanism of glucose and fructose metabolism in Zymomonas was similar to that of Saccharomyces. Gibbs and De Moss 2 8 ' 2 9 ' however found that Zymomonas did not follow the glycolytic pathway but utilized the Entner-Doudoroff pathway anaerobically in association with a pyruvate decarboxylase. This finding was tested and confirmed later by Stern et al. 3 0 ) using radiorespirometric techniques. Additional enzymatic confirmation of the utilization of this pathway has been provided by other authors 3 1 ~33). It has been suggested further that Zymomonas is the only species of bacteria able to utilize the Entner-Doudoroff pathway anaerobically 34) . Analysis of the energetics of the Entner-Doudoroff pathway shows that only 1 mole of ATP is produced per mole of glucose or fructose utilized.

Ethanol Production by Zymomonas mobilis

41

It has been found that only 2-2.6% of glucose or fructose utilized by Zymomonas is converted to biomass, the remainder being converted almost quantitatively into carbon dioxide and ethanol 3 5 _ 3 7 ) . However, the molar conversion of glucose to ethanol was found to be variable. With some strains of Z. mobilis, the amount of ethanol produced per mole of glucose or fructose varied from 1.5 to 1.9 mole depending on the culture conditions 29 • 3 2 ' 3 8 ) . With a further strain, originally classified as Z. anaerobia NCIB 8777, the ethanol yield from fructose was found to be much lower than from glucose indicating the additional formation of glycerol and dihydroxyacetone 3 6 , 3 7 ) . Sucrose Metabolism The ability to utilize sucrose as a carbon source was the main characteristic distinguishing Z. mobilis from Z. anaerobia in the original classification 22) . It was demonstrated, however, that some strains of Zymomonas could lose and regain the ability to use sucrose 20) . In later work by Dadds et al. 3 9 ) , and Richards and Corbey 40) , it was considered that the ability to metabolize sucrose was an inducible and strain specific phenomenon. The hydrolysis of sucrose to glucose and fructose with concomitant formation of levan was reported to be the first step in sucrose metabolism 4 1 , 4 2 ) . The results of Dawes et al. 4 2 ) with Z. mobilis NCIB 8938 established that levan, a polymer of fructose sub-units, was formed in sucrose media but not from a mixture of glucose and fructose. Both growing and non-growing cells were able to form levan. Furthermore, it was found that levan was synthesized readily by crude extracts of Z. mobilis. Approximately 10% of the sucrose utilized by these extracts was converted into levan, the remainder of the sucrose being hydrolyzed to glucose and fructose. Dawes et al. 4 2 ) did not establish whether or not an invertase was present and suggested tentatively that a levansucrase alone could have accounted for both levan formation and sucrose hydrolysis according to the reaction: n C 1 2 H 2 2 O u + ROH - R (C 6 H 10 O 5 ) n OH + n C 6 H 1 2 0 6 sucrose acceptor levan glucose and fructose They postulated that if the acceptor was fructose, then levan would be formed; however if the acceptor was water, no levan would be formed and the products would be fructose and glucose. No evidence for sucrose phosphorylase activity in Z. mobilis was found 42) . Swings and De Ley 2 ) reiterated that two important questions with sucrose metabolism remain to be answered: (1) is the hydrolysis of sucrose effected by levansucrase or by a separate invertase or both? (2) is the acquisition of sucrose-utilizing ability due to the induction of specific enzymes, which also give rise to levan production, or is it due to the selection from the population of certain cells possessing constitutive enzymes? Other Carbohydrate Sources It has been found that raffinose can support the growth of some strains of Z. mobilis although no gas is produced 8). Dadds et al. 3 9 ) confirmed also that

42

P. L. Rogers et al.

strains of Zymomonas could be induced to utilize raffinose. Sorbitol could permit the growth of some strains of Zymomonas without gas formation 8 ) . However, Z. mobilis was unable to utilize other sugars such as lactose, maltose or cellobiose. It was also unable to use pentose sugars or fatty acids. Other Nutritional Requirements Sources of Nitrogen Ammonium salts, a-amino acids and peptides may be utilized as sources of nitrogen, but nitrates and nitrites are not assimilated. Belaich and Senez 38> reported that a mixture of 20 amino acids (in synthetics medium) or even NH 4 C1 (in minimal medium) could replace yeast extract and support the growth of Z. mobilis NCIB 8938 in the presence of glucose and pantothenate. Belaich et al. 4 3 ) reported that the growth rate and the growth yield of Z. mobilis in both synthetic and minimal media were dependent on the concentration of calcium pantothenate; however, the values obtained in minimal media were lower than those obtained from synthetic media with the same concentrations of calcium pantothenate. Bexon and Dawes 44) reported data on the nutritional requirements of Z. anaerobia NCIB 8227. It was found that Z. anaerobia was more exacting in its requirements for amino acids than Z. mobilis. The maximum growth of Z. anaerobia in the defined medium was obtained with the complete mixture of 21 amino acids; however, the addition of arginine, tryptophan, cystine and glutamic acid was sufficient to support the growth of Z. anaerobia. It has been found also that the effects of amino acid addition on the growth of strains of Zymomonas varied significantly from strain to strain 4 0 ' 4 5 ) . Sources of Sulphur Compounds Sulphate, sulphite, methionine, thiamine and to a lesser extent cysteine can be used as sources of sulphur in synthetic media *6). Anderson and Howard 4 7 > reported that 50% of the cell sulphur was derived from sulphate and methionine; the origin of the rest was unknown. Hydrogen sulphide was the main volatile sulphur compound formed, with traces of dimethyl sulphide and dimethyl disulphide. In synthetic medium, excessive amounts of hydrogen sulphide were liberated by Z. anaerobia when calcium pantothenate was deficient. Sulphate and zinc ions were found to stimulate the formation of hydrogen sulphide 47) . Millis 8) found that Z. anaerobia var. pomaceae did not produce hydrogen sulphide in cider or beer. Richards and Corbey 40) observed that hydrogen sulphide production was variable and sensitive to environmental conditions. De Ley and Swings M) reported that only 4 strains among their 46 strains of Zymomonas tested formed hydrogen sulphide. However, it was considered that the ability to form hydrogen sulphide may be a stable feature in some of the strains of Zymomonas e.g. Z. mobilis NCIB 8938 and Z. anaerobia NCIB 8227 46) . . Effect of Phosphate Ions Phosphate ions were considered to be important in facilitating the catabolic activity of the cells. Senez and Belaich 48) reported that when the growth of Z. mobilis

Ethanol Production by

Zymomonas mobilis

43

was limited by phosphate in the presence of excess glucose, catabolic activity was reduced by about 20 %. However the decreased catabolic activity was fully and immediately restored by the addition of inorganic phosphate. These results suggested that the catabolism of glucose was controlled by coupled phosphorylation, and that the catabolic activity of the phosphate-limited cells depended on hydrolysis of ATP by an adenosine triphosphatase system. Effect of Various Metallic Ions Evidence has been provided that magnesium ions play an important role in preventing RNA degradation and prolonging the survival of strains of Zymomonas. Dawes and Large 49) reported that the degradation of RNA under substrate-limited conditions was suppressed by the presence of 33 mM MgCl 2 . It was found that the addition of 33 mM MgCl2 to the starvation medium did not affect the ATP content, but brought about an increase in viability. Sly and Doelle 33) reported that the addition of 10 mM MgCl2 resulted in optimal activity of glucose-6-phosphate dehydrogenase but that higher concentrations inhibited enzyme activity. The effect of various trace elements has not been thoroughly investigated although it has been observed that molybdenum stimulated growth 46) . Vitamin and Growth Factor Requirements Vitamin requirements appear to be variable. Most strains of Zymomonas exhibit a definite requirement for pantothenate 3 8 , 4 6 ) . Inexplicably, a strain of Z. anaerobia NCIB 8227 previously reported to require pantothenate, biotin and lipoate 441 was shown to require only pantothenate on re-examination after a period of time had elapsed 50) . Dadds et al. 3 9 ) reported that a strain of Z. mobilis isolated from beer showed a requirement for p-aminobenzoic acid, folic acid, cyanocobalamin and biotin. Five months later, however, one of the subcultures from the original isolates did not require these vitamins and exhibited only a requirement for pantothenate. Van Pee et al. 4 5 > tested 38 different strains of Zymomonas for their requirements and concluded that (1) most or all strains required pantothenate and biotin and occasionally some other growth factors (2) results with the same strain differed from one author to another (3) upon ageing of the subculture some strains appeared to require only pantothenate and (4) differences in growth factor requirements had no particular taxonomic meaning for Zymomonas. Effect of Medium Composition on Kinetics Studies by Bauchop and Elsden 51) and Belaich et al. 4 3 ) have illustrated that the medium composition can have a significant effect on the kinetics of growth and ethanol production by Z. mobilis. From the data shown in Table 3, it is interesting to note that while limiting concentrations of pantothenate resulted in a reduction in the specific growth rate (|i), the specific glucose uptake rate (qs) remained relatively constant. The significance of the uncoupling between growth and uptake of sugars will be discussed later in more detail.

44

P- L- Rogers et al.

Table 3. Effect of medium composition on kinetic parameters for Z. mobilis A T C C 10988 M e d i u m composition Components

Kinetic parameters Concentration (gl"1)

Glucose Peptone Yeast extract

10.0 10.0

Glucose Peptone Yeast extract

2.0 2.0 2.0

Glucose A m i n o acids 3 excess-Pantothenate limiting-Pantothenate Glucose NH 4 C1 excess-Pantothenate limiting-Pantothenate

Ref.

Yx/s

6.5

Bauchop and Elsden 5 1 )

0.046

Belaich et al. 0.40

0.046

8.67

5 x 10~ 3 5 x 10" 7

0.39 0.20

0.036 0.016

10.82 12.17

2.0 2.0 5x10" 5x10"

0.30 0.15

0.026 0.014

11.48 11.02

431

2.0 •

Alanine, arginine, aspartic acid, cysteine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, hydroxyproline, phenylalanine, proline, serine, tryptophan, tyrosine, and valine were added as final concentration of 60 m g l " 1

Effect of Environmental Conditions Influence of Temperature Millis 8) found that the optimum temperature for the growth of various strains of Zymomonas was between 25 and 31 °C. De Ley and Swings 24) tested the growth of 40 different strains in liquid medium at different temperatures and the results are summarized in Table 4. Forrest 52) reported that Z. mobilis ATCC 10988 growing below 33 °C had a constant growth yield of 8.3 g biomass per mole of glucose utilized (Y x/s = 0.046). The activation energy for both growth and glucose metabolism was estimated as 11.1 Kcal per mole. Above this temperature the growth yield decreased and the activation energy

Table 4. G r o w t h of Zymomonas different temperatures

strains in liquid medium at

Incubation temperature °C

% of strains growing

30 34 36 38 40

100 97 97 74 5

Ethanol Production by Zymomonas

mobilis

45

for growth changed to —6.9 Kcal per mole. However the activation energy for the specific uptake rate of glucose did not change, and it was apparent that the degree of coupling between growth and glucose uptake decreased at temperatures above 33 °C. The effect of increasing temperature on the degree of coupling between growth and ethanol production was apparent also from the data of Viikari et al. 5 3 ) who reported that all the strains of Zymomonas species tested ceased to grow above 37 °C. However, some of their strains still possessed the ability to produce ethanol at temperatures as high as 45 °C. Selection of Optimum pH The wide pH range for growth, from 3.5 to 7.5, and the relative acid tolerance are both typical features of strains of Zymomonas8'16). Swings and De Ley 2) reported the growth response at different pH values as shown in Table 5.

Table 5. Growth response of Zymomonas to different p H Initial pH

% of strains growing

3.05 3.50 3.70 3.85 5-7 7.50 8.00

0 43 71 90 100 87 0

strains

Effect of Dissolved Oxygen Concentration Kluyver and Hoppenbrouwers 27) reported that the ethanol yield in semi-aerobic conditions decreased to one-third of the value for strict anaerobic conditions. Shimwell 17) found that the decrease in ethanol production under conditions of aerobiosis resulted from an accumulation of acetaldehyde. Belaich and Senez 38 ' however reported that aeration of glucose cultures did not result in the accumulation of acetaldehyde; it did result, however, in further oxidation of ethanol to acetate, with consumption of 1 mole of 0 2 per mole of acetate formed. The aerobic metabolism of glucose by Z. mobilis could be described by: 1 Glucose + 1 0 2 -> 1 Ethanol + 1.7 C 0 2 + 1 Acetate + 0.2 Lactate . Belaich and Senez 38) reported also that cells grown both under aerobic and anaerobic conditions possessed a degree of respiratory activity indicating that the respiratory enzymes were constitutive. It was also considered by other authors that some strains of Zymomonas were more aerotolerant than others 2 , 3 6 ) .

46

P. L. Rogers et al.

Effect of High Sugar Concentrations Kluyver and Hoppenbrouwers 271 observed that the pulque strain (Z. mobilis) was able to grow with "alcoholic fermentation" in a medium containing 250 g 1 _ 1 glucose. Swings and De Ley 2 ' reported that all strains of Zymomonas tested grew well in 200 g 1 _ 1 glucose medium and 54% of strains tested could grow in 400 g l" 1 glucose medium. Effect of High Salt Concentrations Susceptibility to relatively high concentrations of NaCl was reported by Swings and De Ley 2) . No strains of Zymomonas were found which were able to grow in the presence of 2 0 g l _ 1 NaCl, only 71 % of strains being able to grow in medium containing 10 g l" 1 NaCl. Ethanol Tolerance Strains of Zymomonas have been isolated historically from alcoholic beverages containing 20 to 100 g 1 _ 1 ethanol. Swings and De Ley 21 reported that all strains of Zymomonas tested grew readily in the presence of 55 g 1 _ 1 ethanol; 73% of strains could grow in 77 g 1 _ 1 of ethanol while only 47% of strains were able to grow in 100 g 1 _ 1 of ethanol. However, no detailed studies of the kinetics of ethanol inhibition in Z. mobilis have been reported until recently.

3 Genetic Studies 3.1 Strain Selection In a review of the biology of Zymomonas by Swings and De Ley 2) forty different isolates from various parts of the world were compared for a number of properties, including ethanol tolerance, temperature tolerance, growth in high sugar concentrations, use of different sugars and flocculation. Considerable variation between the strains was observed. Because of this wide variation, it seemed likely that there would be better strains of Zymomonas available than those originally chosen for kinetic studies. Therefore strains isolated from a number of different locations were compared for their suitability for ethanol production. Viikari et al. 5 3 ) compared strain ATCC 10988 with seven brewery isolates, and found that several of these isolates produced more ethanol and were more temperature tolerant than strain ATCC 10988. In this laboratory, twelve strains from different locations have been compared. The strains tested are listed together with their sources in Table 6. In initial studies in test tubes, it was immediately apparent that some strains were unable to use sucrose 54) and would be unsuitable therefore for industrial production of ethanol from sucrose-based substrates such as molasses or sugar beet. Of the remaining strains, several were observed to have shorter lag periods and to grow faster than the other strains 54) . The best strains including ZM1 were chosen for more detailed studies.

Ethanol Production by Zymomonas mobilis

47

Table 6. Strains of Zymomonas used for strain selection studies Strain

Source

Z. mobilis

ZM1 (ATCC 10988) Agli 3TH Delft ZM6 (ATCC 29191) ZM4 (CP4) B70 ZAbi ZalO

Fermenting Agave juice Fermenting Agave juice Fermenting Arenga sap Fermenting Elaeis sap Fermenting sugarcane juice Infected British ale Beer Beer

Z. mobilis subsp. pomaceae

ATCC 29192 238 S30.A S30.2

Sick cider Cider Apple pulp Cider

W h e n the four strains were compared in liquid medium with 200 g l " 1 glucose, strain Z M 4 was f o u n d to produce ethanol at a considerably faster rate than the other three strains (Table 7). The final ethanol concentration w a s also higher with this strain (Table 7), suggesting that Z M 4 m a y be more ethanol-tolerant than the other strains. This w a s confirmed in media with added ethanol; only Z M 4 could grow with m o r e than 60 g 1 _ 1 e t h a n o l 5 4 ) . In media with 200 g 1 _ 1 sucrose, strains Z M 1 and Z M 4 gave similar results, both producing ethanol faster than the other t w o strains (Table 7). In a o n e litre

Table 7. Ethanol production by Z. mobilis strains growing on 200 g 1 1 glucose or sucrose in test tubes Strain

ZM1 ZM4 ZM6 Agli

Ethanol production from: 200 g l " 1 glucose

200 g 1

Rate ( g l " 1 h" 1 )

Final conc. (gl ')

Rate ( g l - 1 h" 1 )

Final conc. (g 1 ')

1.43 2.00 1.25 1.40

60 81 77 55

1.22 1.21 0.80 1.00

56 56 51 39

Table 8. Final ethanol concentrations produced by Z. mobilis strains from 200 g l" 1 glucose in test tubes at different temperatures Strain

ZM1 ZM4 ZM6 Agli

Final ethanol concentration (g 1 ') 30 °C

37 °C

42 °C

60 81 77 55

40 52 0 0

0 30 0 0

1

sucrose

48

P. L. Rogers et al.

fermentor, the ethanol yields followed the same general trends as in test tubes, with strain ZM4 producing 117 g l " 1 ethanol from 250 g l - 1 glucose medium and 89 g 1 _ 1 ethanol from 250 g 1 _ 1 sucrose medium 55) . The reduced yield on sucrose was partly due to the production of levan 54) . At temperatures above 30 °C, all strains grew and produced ethanol more slowly, but once again strain ZM4 appeared the best (Table 8).

3.2 Strain Improvement Since strain ZM4 was clearly better than ZM1, although there was still room for further improvement of, for example, ethanol tolerance, attempts were made to genetically manipulate ZM4. Mutagenesis Techniques Since no attempts to mutate Zymomonas had been reported previously, it was first necessary to devise a suitable method of mutagenesis for this organism. Initially, ultraviolet light was tested as a mutagenic agent, selecting for an increased level of colonies resistant to the antibiotic rifampicin. Strains ZM1 and ZM4 normally have a low level of spontaneous mutation to rifampicin resistance of about 10~8, and no increase in this level could be detected after exposure to UV light. However, with nitrosoguanidine (NTG) as the mutagen, a tenfold increase in the number of rifampicin-resistant colonies was obtained. There was considerable strain variation in the level of sensitivity to this mutagen, with some strains only requiring 2 5 m g l _ 1 NTG for 60 min to give the best levels of mutagenesis, while others needed up to four times this amount. The best rates of mutagenesis were obtained when NTG was added directly to growing cultures in rich medium rather than in buffer. Selection of an Ethanol-Tolerant Mutant Having determined the optimum time of exposure to NTG, strain ZM4 was treated with this mutagen and ethanol-tolerant mutants were selected. Strain ZM4 could grow on agar plates with up to 8 % v/v ethanol, but after mutagenesis a few cells were able to form colonies on plates with 12% v/v ethanol. These colonies were purified and tested for ethanol production in media with high levels of glucose. The best mutant, ZM48, appeared more ethanol-tolerant in test-tube culture (Table 9) but did not show any real improvement in a one litre fermentor. However, after a second round of mutagenesis on strain ZM48, several colonies appeared on 15% v/v ethanol plates. Some of these produced higher final ethanol concentrations but grew slower than the parent strain (Table 9). The best isolate, strain ZM481, showed more promise in test tubes, particularly with high levels of glucose where it produced a higher final ethanol concentration and grew faster than the parent strain (Table 9). In the one litre fermentor it was confirmed that ZM481 was a more promising strain than ZM4 at high ethanol concentrations. In long term continuous culture with cell recycle, ZM4 could sustain 65 g l " 1 ethanol, but it was difficult to maintain a stable recycle operation due to a loss of

Ethanol Production by

Zymomonas mobilis

49

Table 9. Final ethanol concentrations of strain Z M 4 and ethanoltolerant mutants growing on 300 g 1 _ 1 glucose medium Strain

Final ethanol conc. (g l - 1 )

Time taken to reach max. ethanol conc. (h)

ZM4 ZM444 ZM48 ZM485 ZM488 ZM481

93 98 101 108 112 110

108 110 101 110 132 84

viability at ethanol levels above 75 g 1 _ 1 5 6 , 5 7 ) . The mutant strain ZM481, however, retained a higher viability in high ethanol concentrations so that ethanol concentrations could be maintained at 85 g 1 _ 1 in the cell recycle system — a significant increase over the parent strain ZM4. Selection of a Flocculent Mutant For cell recycle in many of the continuous culture experiments carried out in this laboratory, a cross-flow membrane unit has been used. However, scale-up of such a system for industrial use may be impractical due to membrane clogging by molasses or starch hydrolysates, and therefore other cell recycling systems have been examined. With yeasts, spontaneously flocculating strains are available which can be recycled on the basis of this property. Although many strains of Zymomonas are reported to be flocculent 2) their flocculation is insufficient for utilization in cell recycling. Strain ZM4 does not flucculate spontaneously, but after NTG mutagenesis a highly flocculent mutant was isolated. This mutant forms small granular floes in the fermentor which settle out as floes up to several millimetres in diameter within a minute or so after agitation is stopped. This mutant, ZM401, shows promise in a continuous system with cell recycling by means of an external settling tank, and also in semibatch culture where productivités of 50 g l - 1 h _ 1 have been obtained (J. H. Lee, unpublished observations). Selection for Temperature Tolerance Temperature is likely to play an important role in the optimal control of ethanol production, particularly where cell recycle, vacuum operation or simultaneous saccharification and fermentation are contemplated. Therefore, if the optimum temperature for Zymomonas could be increased from 30 °C, the process would be more economical overall. Viikari et al. 5 3 ) found that their strains did not grow above 37 °C, and ethanol production decreased markedly above this temperature although it did not completely cease even at 45 °C. The strains studied in this laboratory also show reduced levels of growth at or above 37 °C, although ZM4 proved to be the most temperature tolerant with limited growth and ethanol production at 42 °C. However, it was fairly easy to select mutant colonies which grew on plates at 40 °C and above. These mutants grew more readily in liquid media at the elevated temperatures, although increases of only a few degrees have so far been obtained.

50

P. L. Rogers et al.

3.3 Development of other Genetic Techniques in Zymomonas Isolation of Auxotrophs In order to facilitate strain construction and genetic mapping, a bank of auxotrophic mutants is currently being isolated in this laboratory. It was first necessary to develop a minimal medium for isolation of such mutants, since no suitable medium was available (most experiments reported on Zymomonas have used rich media containing yeast extract). Mutants have been isolated after NTG mutagenesis, and already a number of single and double auxotrophs with various combinations of spontaneous antibiotic resistance markers have been characterized. Conjugation The transfer of several conjugable drug resistance plasmids of the IncPI and IncFII groups into Z. mobilis has been tested in this laboratory using the membrane filter mating technique 58) . All four of the plasmids which were tested were transferred from

Table 10. Transfer of plasmids from E. coli and Ps. aeruginosa to Z. mobilis Donor"

Recipient"

Approx. frequency of plasmid transfer

E. coli (Rl-í/rrf-19)

ZM1 ZM6 ZM1 ZM4 ZM6 ZM1 ZM4 ZM6 ZM1

5 x 10" 2 lxlO'2 4 x 10" 4 6 x 10~4 9 x 10" 3 1 x 10~2 2xl(T3 4 x 10" 3 4 x 10 _ 1

E. coli (pJB4JI)

E. coli (pRDl) Ps. aeruginosa (R68.45)

" Donor strains were easily distinguishable from the recipient strains morphologically, and usually doubly marked antibiotic-resistant recipient strains were also used

Table 11. Retransfer of R plasmids between Z. mobilis strains Donor"

Recipient"

Approx. frequency of plasmid transfer

ZM1 ZM4 ZM6 ZM4 ZM1 ZM1

ZM6 ZM6 ZM6 ZM1 ZM6 ZM6

8 x 10 _1 9 x 10 - 1 5 x 10" 1 8x10"' 8 x 10 - 1 8 x 10" 1

(pRDl) (pRDl) (pRDl) (pJB4Jl) (Rl-rfrd-19) (R68.45)

" Donor and recipient strains used were usually doubly marked antibioticresistant strains so that they could be easily distinguished from each other

Ethanol Production by Zymomonas

mobilis

51

E. coli and Ps. aeruginosa to strains ZM1, ZM4 and ZM6 at the high frequencies of 10~4 to 10" 1 (Table 10). Plasmid R68.45 was transferred most readily, with about 30 % of the recipient cells gaining the plasmid (Table 10). When Z. mobilis strains carrying any of the four plasmids were used as donors in conjugation experiments with other marked strains of Z. mobilis, these plasmids could be transferred at very high frequencies (Table 11). As would be expected for such plasmids 58) , the transfer between Z. mobilis strains was considerably higher than for the intergeneric matings. However, transfer between homogenic strains was no higher than between strains of completely different origins (Table 11). Chromosome Mobilization When one spontaneous rifampicin-resistant strain of Z. mobilis containing plasmid R68.45 was used as the donor in a conjugation experiment with another marked strain of Z. mobilis, it was possible to test for mobilization of the chromosomal rifampicin resistance. A low level of transfer of this gene was detected in a filter mating, at a frequency of almost 10 - 6 which is 100 times higher than the background level of spontaneous mutation 58) . Some of the auxotrophic mutants isolated in the background of ZM4 were also used in tests for chromosome mobilization by plasmid R68.45. Depending on the particular mutant used, levels of transfer ranged from 10" 5 to more than 10~3. Since chromosome transfer by this method has now been established for Z. mobilis, plasmid R68.45 is currently being used to construct strains more useful for ethanol production, such as a flocculent, highly ethanol-tolerant strain from donor and recipient strains ZM401 and ZM481. Plasmid Isolation The utilization of sucrose by Zymomonas has been reported by some authors to be an inducible, strain-specific phenomenon 2 ' 3 9 ' 4 l , 4 2 , 5 9 ) . A product formed by most strains which do utilize sucrose is levan, a fructose polymer 39 > 42) . However, levan production disappears with increasing temperature 54) and one strain, Agl 1, does not produce any detectable levan when growing on sucrose. To test whether the inducibility of growth on sucrose and levan production might be correlated with plasmid-borne genes, and their possible amplification in the presence of sucrose, a method was developed for the isolation of natural Zymomonas plasmids. The method used is based on the alkaline extraction method of Birnboim and Doly 60) with several modifications to allow isolation of good plasmid preparations with as little chromosomal DNA contamination as possible. Other methods involving lysis with Triton X-100 or phenol/chloroform extractions sheared all the DNA very easily, as did the unmodified alkaline extraction method. However, with the modified method, good preparations were obtained with plasmids which were clearly visible on 1 % agarose gels. A range of different strains of Zymomonas were grown in glucose medium and tested for the presence of natural plasmids. Interestingly, the most widely used strain (ZM1), which has been maintained in various laboratories for a longer time than the other strains, was the only one with no detectable plasmids (Fig. 2). This

52

P. L. Rogers et al.

m Ü m

.

r pw 9» « ** f* IS

m ...

-

=r.

^

—

Fig. 2. Plasmids of Z. mobilis on a 1 % agarose gel. From left, plasmid standards from E. coli V517; B70; A g l l ; ATCC 29192; Z M 6 ; Z M 4 ; Z M 1 ; ATCC 10988; E. coliV 517 61)

Fig. 3. Plasmids of Z. mobilis on a 1 % agarose gel. From left, E. coli V517; S30.2; S30.A; 238; ZAbi; ZalO; 3TH Delft; E. coli V517

strain is a derivative of ATCC 10988 which was obtained as NCIB 8938 by the University of Queensland Culture Collection and kept there for many years. Strain ATCC 10988 obtained from the American Type Culture Collection was found to have both large and small plasmids (Fig. 3). All strains other than ZM1 had at least one plasmid, and usually several (Figs. 2 and 3). Plasmids ranged in size from 1 megadalton to about 40 megadaltons, and plasmids at both ends of the size range could be found in the same strains (for example, strains 238, ZalO and B70). Another interesting point is that plasmids of apparently the same sizes-occurred in strains from widely differing origins (for example strains ZM4 from sugarcane juice and ATCC 29192 from sick cider both had plasmids of about 9.4 megadaltons). Whether these plasmids are, in fact, identical is currently being investigated using restriction endonucleases as well as by more accurate sizing. Several of the very small plasmids (1 to 4 megadaltons in size) are also being developed as cloning vehicles in this organism. It was not possible to demonstrate any correlation between growth on sucrose and an alteration in the plasmid profiles of any of the levan-producing strains examined. Also, strain ZM1 which had no detectable plasmids, was able to grow

53

Ethanol Production by Zymomonas mobilis

on sucrose just as well as the other strains. Therefore, at this stage there is no evidence for any connection between the inducibility of growth on sucrose and the presence or amplification of plasmids.

3.4 Further Improvement of Zymomonas Strain ZM481 was selected as the most ethanol-tolerant mutant from two sequential mutagenesis experiments on strains ZM4 and ZM48. No success was obtained in attempts to further improve this tolerance by a third round of mutation; no colonies appeared on plates with ethanol concentrations higher than 15% v/v. It may, however, be possible to select better mutants of ZM481 on the basis of greater viability in liquid media with high ethanol concentrations, since this appears to be the main advantage ZM481 has over ZM4. Similarly, strains with better viability at 42 °C and above, may give higher productivities under these particular conditions. Another factor which should help to increase the yield on sucrose is the abolition of levan formation. Although it has not been possible to isolate mutants of strains ZM1 or ZM4 which do not produce levan from sucrose, the availability of the non-levan producing strain A g l l should facilitate genetic construction of a ZM4 derivative to give higher ethanol yields on sucrose-based substrates. In order to help with locating the genes for such characteristics as levan production, as well as for the other mutagenesis experiments, as method of transposon mutagenesis has been developed in Z. mobilis. Although the Tn5-containing plasmid pJB4JI, which has been used for transposon mutagenesis in Rhizobium 61) is stable in most strains of Z. mobilis58), the subsequent introduction of plasmid pJB3JI ( kanamycinsensitive derivative of R68.45) promotes "jumping" of the transposon and loss of pJB4JI in derivatives of strain ZM4. Such a method should also prove useful for putting markers on some of the natural Z. mobilis plasmids. Z. mobilis does have one disadvantage if it is to be used industrially, due to the fact that it can only use glucose, fructose and sucrose, and cannot utilize other carbon sources such as maltose and starch 27) . Since some of the substrates proposed for industrial ethanol production are basically starch (e.g. corn, cassava and wheat), it would be very useful to have a strain of Z. mobilis which could use starch. This would then allow direct conversion of starch to ethanol by Z. mobilis, rather than a more costly hydrolysis followed by ethanol production. In high productivity systems, where high rates of starch hydrolysis will have to be achieved, the cost of using a commercial glucoamylase enzyme would be a significant factor in the process.

4 Kinetic Studies on Zymomonas

mobilis

Following detailed batch and continuous culture evaluation of Z M 1 3 , 4 ) , strain ZM4 was selected for a more comprehensive study under similar conditions. From the strain selection program and the preliminary genetic studies, the advantages of ZM4 when compared to ZM1 can be summarized as follows:

54

P. L. Rogers et al.

(1) higher rates of ethanol production (2) shorter lag period and faster growth on high concentration glucose and sucrose media (3) enhanced ethanol tolerance (4) enhanced temperature tolerance (5) amenability to genetic manipulation as shown by a relatively high frequency of antibiotic-resistance plasmid transfer both from strains of Escherichia coli and Pseudomonas to Z. mobilis and between various strains of Z. mobilis.

4.1 Kinetic Studies on Glucose Media Batch Culture Z. mobilis strain ZM4 was grown in batch culture with initial glucose concentrations in the range of 100-300 g l " 1 at 30 °C and pH = 5.0. Data for the increase in biomass, utilization of glucose and the production of ethanol have been reported previously 63) . The kinetic parameters calculated from these batch culture data are shown in Table 12. From the results it is evident that the specific growth rate of Z. mobilis ZM4 was significantly affected by an increase in the concentration of glucose while the specific glucose uptake rate and the specific ethanol production rate were largely unaffected. Some decrease in the biomass yield was observed with an increase in the initial glucose concentration, while the ethanol yields were relatively constant.

Table 12. Kinetic parameters of Z. mobilis Z M 4 in batch culture at various glucose concentrations Kinetic parameters

Specific growth rate |i Specific ethanol productivity q p Specific glucose uptake rate q s Biomass yield Yxls Ethanol yield Y p/5 % of theoretical yield Time period of calculation (h)

Initial glucose concentraion (g l" 1 ) 100

150

200

250

300

0.35 5.2 10.9 0.032 0.48 94.0 0-10

0.27 4.2 8.9 0.030 0.47 92.0 0-13.5

0.22 5.1 10.5 0.018 0.49 96.0 0-18.5

0.18 5.4 11.3 0.015 0.48 94.0 0-19

0.13 4.3 8.7 0.015 0.49 96.0 0-20

Continuous Culture Continuous culture studies with ZM4 were carried out with 60, 100, 135 and 170 g l" 1 glucose media 63) . Steady-state values for the biomass, ethanol and glucose concentrations are shown in Figs. 4, 5, 6 and 7 respectively. It is evident that for the media with the lower concentrations of glucose, for example 60, 100, 135 g l - 1 , the cultures behaved as glucose-limited chemostats at low dilution rates. However, for an inflow of 170 g 1 _ 1 glucose medium, the glucose was in excess at all

55

Ethanol Production by Zymomonas mobilis

DILUTION

RATE

( h"1 )

Fig. 4. Steady state continuous culture data and selected kinetic parameters for Z. mobilis ZM4 on 60 g l" 1 glucose medium (pH = 5.0, T = 30 °C)

dilution rates. This was taken to indicate that the growth was ethanol-limited. The maximum steady-state values of the kinetic parameters for ZM4 with 60,100,135 and 170 g l" 1 glucose media are shown in Table 13. The maximum values for the specific rates found in continuous culture tend to confirm those obtained from batch culture. Substained oscillations were observed at a dilution rate of D = 0.1 h _ 1 with 200 g 1 _ 1 glucose medium. These oscillations were sustained for at least 400 h at this dilution rate and are characteristic of the behaviour of a product-limited continuous culture system 4 , 6 3 ) .

56

P. L. Rogers et al.

0

0.1

0-2

DILUTION

0.3

RATE

0-4

0.5

1

( h" )

Fig. 5. Steady state continuous culture data and selected kinetic parameters for Z. mobilis Z M 4 on 100 g 1 ~1 glucose medium

Comparative Kinetics of ZM1 and ZM4 When the results obtained from batch and continuous culture with Z M 4 were compared with those obtained with Z M 1 3 ' 4 ) , it was evident that Z M 4 was significantly better than ZM1. The differences were clear in batch culture at high glucose concentrations with Z M 4 being less inhibited by the increased ethanol concentrations. The maximum ethanol concentration attained with Z M 4 was 127 g l - 1 while 104 g 1 _ 1 was attained with ZM1. Comparison of the kinetic parameters revealed that Z M 4 had q s and q p values approximately double those of ZM1 as well as a considerably higher specific growth rate (Table 1 and 12).

Ethanol Production by Zymomonas mobilis

0

0.1

DILUTION

57

0.2

( h~1)

RATE

0.3

Fig. 6. Steady state continuous culture data and selected kinetic parameters for Z. mobilis ZM4 on 135 g 1 _ 1 glucose medium Table 13. Comparison of kinetic parameters of Z. mobilis Z M 4 in continuous cultures Maximum steady state kinetic parameters

Specific growth rate n Specific ethanol production rate q p Specific glucose uptake rate q s Biomass yield Yx/a Ethanol yield Y p/S % of theoretical yield Maintenance energy coefficient m

Concentration of glucose (g 1 ')

,

60

100

135

170

0.45 5.2 10.4 0.041 0.50 98.0 3.8

0.375 5.4 10.8 0.037 0.50 98.0 4.1

0.275 5.40 10.80 0.025 0.50 98.0 5.1

0.24 5.1 10.5 0.022 0.48 95.0 5.1

58

P. L. Rogers et al.

003 — • o> 0.02 o,

0.01 „ 0 >*

4 i~ 3

Z

2 W C o «

'sucrose

0 CT

q . glucose q fructose-

20

a

. a *— -Cb-

60

40 1

ETHANOL ( 9 I" ) Fig. 11. The effect of ethanol on the specific growth rate, specific catabolic activities and specific rate of levan production for Z. mobilis ZM4. (Data derived from continuous culture experiments)

Discussion

of the Kinetics on Fructose and Sucrose

Media

Other authors have studied the characteristics of growth of various strains of Z. mobilis on fructose and sucrose media. Relatively low biomass yields and slow growth rates on fructose medium have been noted by Dawes et al. 4 2 ' and McGill and Dawes 3 7 ) . On sucrose media, a reduction in ethanol yield has been attributed to levan formation 4 1 , 4 2 ) . The mechanism of sucrose hydrolysis and levan formation has not been elucidated fully. Furthermore, the mechanism of synthesis of the enzyme(s), i.e. whether the enzymes are constitutive or inducible, and the question as to which enzymes are involved, still remains to be investigated. Two distinct enzymes are considered to be involved in sucrose hydrolysis by Z. mobilis 2> 4 1 • 42 namely levansucrase (EC. 2.4.1.10) and an invertase (EC. 3.2.1.26), and the presence and activity of these enzymes may

P. L. Rogers et al.

64

well be strain-dependent. With Z. mobilis ZM4, the formation of levan and rapid accumulation of both fructose and glucose during exponential growth may indicate that both enzymes are present with appreciable activity. The rapid increase in fructose levels also appears to initiate the formation of levan in batch culture and it is interesting to note that no levan was detected in continuous culture when the fructose concentration was at low levels. The effect of ethanol inhibition on the enzyme(s) involved in sucrose hydrolysis and levan production is of considerable significance in the present work and necessitates further detailed investigation.

4.3 Effects of Temperature on the Kinetics of Ethanol Production Temperature is considered to play an important role in the optimal control of ethanol production. For this reason, it was decided to evaluate the effects of temperature on the growth and ethanol production of ZM4 over the temperature range 30-40 °C. Batch Culture Batch cultures were carried out on 250 g 1" 1 glucose medium for the temperature range 30-40 °C. The experimental results for the increase of biomass, glucose utilization and ethanol production have been reported previously M ) . In Fig. 12, data are presented for the effect of elevated temperature and ethanol concentration on cell viability. Kinetic parameters calculated from the batch culture data are given in Table 15. From the results, it is clear that the kinetic parameters (x, q p and q s are relatively constant over the temperature range 30—40 °C. However, significant effects of temperature on ethanol production were observed. These included a reduction of ethanol yield and a lower final ethanol concentration at the end of batch culture. It was apparent that the loss of viability due to ethanol accumulation was greater at higher temperatures as shown in Fig. 12. Continuous Culture The effect of higher temperatures on ethanol production by ZM4 was clearly shown in a series of continuous culture experiments. Data illustrating sustained oscillations in biomass, ethanol and glucose concentrations at a dilution rate D = 0.1 h _ 1 with 150 g glucose medium at 37 °C have also been reported 64) . The maximal ethanol concentration was about 50-55 g l - 1 . However, at 30 °C with the same culture conditions, the steady-state ethanol concentration was 73 g l" 1 . Discussion of Temperature

Effects

Previous studies have been concerned with the effect of increased temperature in controlling Z. mobilis infections in the making of cider and p e r r y 3 , 9 ) and with the role of temperature in uncoupling metabolism from growth 52) . Millis 8 ) found that the optimum temperature for Z. mobilis strains was 25-31 °C. Swings and De Ley 2) reported that ZM4 grew at 38 °C but not 40 °C. Dadds and Martin 4 6 ) , however, found that a strain isolated from beer grew at 40 °C.

Ethanol Production by Zymomonas

mobilis

65

n 10 9 8 _

G

7

s

~d 6 Z a> o

ri

5

=1 4 u ILI -

3

CD


2 1 0

Fig. 12. Effect of temperature and ethanol concentration on viable cell count for Z. mobilis Z M 4

Investigations by F o r r e s t 5 2 ' with Z. mobilis A T C C 10988 on 2—10g 1~1 glucose medium indicated that growth and ethanol production were closely coupled between 24 and 33 °C. At higher temperatures (up to 39 °C) decreasing biomass yields were obtained suggesting an uncoupling of growth and ethanol production. In the studies with ZM4, some decrease in biomass yield occurred with increasing temperature. However, the specific growth rate and the specific rates of glucose uptake and ethanol production were largely unaffected. The m a j o r effects were a decrease in ethanol yield at higher temperatures (perhaps indicating an increased production of aldehydes and lactic acid) and an increased sensitivity to ethanol inhibition. Similar observations of the effect of temperature on ethanol inhibition have been reported by various authors with strains of S. cerevisiae6S~6S). Detailed kinetic studies have been carried out in our laboratory with a strain of S. uvarum which provided further supportive evidence of increasing inhibition at higher temperatures in the range 2 5 ^ 3 °C 6 9 ) . By comparison with yeasts, it appeared that Z. mobilis Z M 4 responded to increased temperature in a similar way although Z M 4 did not have the ability to produce ethanol above 42 °C. However, other strains of Z. mobilis have been reported which are capable of ethanol production but not growth at 45 ° C 5 3 ) .

66

P. L. Rogers et al.

Table 15. The effect of temperature on the kinetic parameters of Z. mobilis ZM4 growing on 250 g 1 glucose medium Kinetic parameters

1

Temperature (°C)

Specific growth rate n Specific ethanol production rate q p Specific glucose uptake rate q s Biomass yield Y l/S Ethanol yield Yp/s % of theoretical yield Time period of calculation (h) Maximum ethanol concentration ( g l - 1 )

30

34

37

40

0.18 5.4 11.3 0.015 0.48 94 0-19 117

0.17 4.9 11.1 0.015 0.44 86 0-19 103

0.18 5.1 12.8 0.014 0.40 78 0-18 91

0.17 5.3 13.2 0.013 0.40 78 0-16 80

4.4 Effects of Medium Composition on the Kinetics of Ethanol Production The medium composition was evaluated in order to minimize media costs and maximize ethanol productivity. It was decided to evaluate the effect of the concentration of yeast extract, K H 2 P 0 4 , M g S 0 4 . 7 H 2 0 and ( N H ^ S C ^ on the kinetics of ZM4 in 250 g l - 1 glucose medium. Yeast Extract

(Oxoid)

The effect of yeast extract concentration on biomass formation, glucose utilization and ethanol production was studied in batch culture. Biomass formation in the medium containing 2.5 g l - 1 yeast extract was found to be about one-half that of the culture with 5-10 g l " 1 yeast extract (viz. 3.7 g 1 _ 1 ). A lower concentration of ethanol was also produced with 2.5 g 1 _ 1 yeast extract indicating the role of a component in the yeast extract in minimizing the effects of ethanol inhibition. The kinetic parameters calculated from the batch culture results are shown in Table 16. It is evident that the kinetics were not affected greatly by yeast extract Table 16. The effect of yeast extract on the kinetic parameters of Z. mobilis ZM4 growing on 250 g 1 ~1 glucose medium Kinetic parameters

Specific growth rate n Specific ethanol production rate q p Specific glucose uptake rate q s Biomass yield Yx/s Ethanol yield Yp/S % of theoretical yield Time period of calculation (h) Maximal ethanol concentration at the end of batch culture ( g l - 1 )

Concentration of yeast extract (gl ') 2.5

5.0

10.0

0.12 4.1 8.3 0.015 0.49 96.0 0-18.5

0.17 4.5 9.6 0.016 0.47 92.0 0-18.5

0.18 5.4 11.3 0.015 0.48 94.0 0-19.0

69

119

117

Ethanol Production by Zymomonas mobilis

67

concentration in the range of 5-10 g l - 1 although the values of q p and q s showed some increase at the higher concentration. However, it is apparent that cells grown with 2.5 g l - 1 yeast extract medium were limited by the restricted availability of one or more essential nutrients. As pointed out by Senez and Belaich 48) , components of yeast extract are utilized as building blocks for cell biosynthesis in Zymomonas and not as energy sources for growth. In their study, a significant proportion of the cellular carbon was found to be derived from organic compounds present in yeast extract or peptone. In the absence of yeast extract, it was shown that calcium pantothenate was essential to the growth of Z. mobilis 3 8 , 4 3 ) . The growth yields and growth rates in synthetic medium (amino acids as nitrogen source and growth factors) and in minimal medium (NH4C1 supplemented with calcium pantothenate) corresponded to approximately one-half of the values obtained with yeast extract medium. The metabolic activity, nevertheless, measured in terms of the rates of glucose uptake and ethanol production was found to be independent of the culture conditions and relatively constant. It was noted also that glucose utilization and ethanol production continued at the same rate although the growth rate was significantly changed, thereby providing evidence of appreciable uncoupling between growth and metabobolism under conditions of nutritional deficiency. Effect of

KH2P04

The effect of KH 2 P0 4 concentration over the range of 1-5 g 1 ~1 was evaluated using continuous culture techniques. The steady-state data for biomass, ethanol and glucose concentrations obtained from the continuous cultures were subjected to kinetic analysis and the results are given in Table 17. From the results it is evident that the overall kinetics were largely unaffected by changes in the concentration of K H 2 P 0 4 over the above range. The effect of high concentrations of potassium ions has been investigated also with a view to understanding the reason(s) behind the significant inhibition of Z. mobilis when grown on molasses. Analysis of undiluted sugarcane molasses revealed a potassium ion level in excess of 50 g l - 1 .

Table 17. Effect of KH 2 P0 4 concentration on kinetic parameters of Z. mobilis ZM4 in continuous culture with 100 g l" 1 glucose medium Kinetic parameters

Specific growth rate n Specific ethanol production rate q p Specific glucose uptake rate q8 Biomass yield Y,/s Ethanol yield Yp/s % of theoretical

Concentration of KH 2 P0 4 (g 1 1

2

5

0.35 5.4 10.8 0.032 0.49 97.0

0.35 5.3 10.7 0.032 0.49 97.0

0.35 5.1 10.2 0.032 0.50 98.0

68

P. L. Rogers et al.

Kinetic studies with 10 and 20 g 1 ~ 1 KC1 added to 250 g l " 1 glucose medium have shown that the overall kinetics and the total ethanol produced are reduced appreciably at the higher concentration (unpublished observations). Effect ofMgSOf

7 H20

To study the effect of magnesium ions on the growth of ZM4, the concentration of M g S 0 4 . 7 H 2 0 was changed from 0.25 to 5 g l" 1 in 250 g 1 _ 1 glucose medium. The kinetic parameters calculated from batch culture data are shown in Table 18. The effect of magnesium on the growth of Z. mobilis ZM4 was investigated as it has been reported that magnesium ions play an important role in preventing R N A degradation and prolonging survival of cells of Z. mobilis under substrate starvation conditions 49) . It has been reported also in previous studies on Z. mobilis that a concentration of 10" 2 M MgCl 2 produced optimal activity of glucose-6-phosphate dehydrogenase in cell free extracts of Z. mobilis, but higher concentrations inhibited the activity of this enzyme 33) . With ZM4 it was observed that the overall kinetic profiles in batch culture were not affected by the MgS0 4 . 7 H 2 0 concentration in the range from 0.5 to 1.0 g l - 1 . However, the concentrations of biomass and the biomass yield were significantly higher with 5 g 1 _ 1 M g S 0 4 • 7 H 2 0 medium, although the specific growth rate was relatively constant. The higher biomass concentrations gave rise to reduced specific rates of glucose uptake and ethanol production. The results indicated that biomass synthesis was prolonged, with a longer exponential growth phase, at the higher concentrations of magnesium ions.

Table 18. The effects of M g S 0 4 • 7 H 2 0 on the kinetic parameters of Z. mobilis ZM4 growing on 250 g 1 _ 1 glucose medium Kinetic parameters

Specific growth rate |i Specific ethanol production rate q p Specific glucose uptake rate q s Biomass yield Y x/s Ethanol yield Y p/S % of theoretical yield Time period of calculation (h) Maximal ethanol concentration ( g l - 1 ) Maximal biomass concentration (g I - 1 )

Effect of

Concentration of M g S 0 4 • 7 H 2 0 (g 1 ') 0.25

0.5

1.0

5.0

0.17 5.1 10.6 0.016 0.48 94.7 0-19.5 103.0 3.1

0.17 5.4 11.1 0.016 0.48 95.0 0-19.5 116.0 3.2

0.18 5.4 11.3 0.015 0.48 94.0 0-19.0 117.0 3.6

0.17 3.5 7.1 0.024 0.49 96.0 0-18.0 119.0 4.4

(NHJ2S04

The effect of (NH 4 ) 2 S0 4 concentration was evaluated over the range of 0.5 to 5 g 1 _ 1 using 250 g l - 1 glucose medium. The kinetic parameters calculated from the batch cultures are shown in Table 19. From the results it is evident that overall kinetics were unaffected by changes in the concentration of (NH 4 ) 2 S0 4 over the

Ethanol Production by Zymomonas

69

mobilis

above range. It was concluded that in the presence of excess yeast extract (viz. 10 g l _ 1 ) t h e concentration of (NH 4 ) 2 S0 4 did not greatly affect the growth and ethanol production of Z. mobilis. Table 19. The effects of (NH 4 ) 2 S0 4 on the kinetic parameters of Z. mobilis growing on 250 g l" 1 glucose medium Kinetic parameters

Concentration of (NH 4 ) 2 S0 4 (g 1 ')

Specific growth rate n Specific ethanol production rate q p Specific glucose uptake rate q s Biomass yield Yx/S Ethanol yield Yp/S % of theoretical yield Time period of calculation (h)

0.5

1.0

2.0

5.0

0.17 5.3 11.3 0.015 0.49 96.0 0-19

0.18 5.4 11.3 0.015 0.48 94.0 0-19

0.18 5.6 11.5 0.015 0.49 96.0 0-19

0.18 5.8 11.9 0.015 0.49 96.0 0-19

5 Development of High Productivity Systems with Z. mobilis 5.1 Cell Recycle Studies One of the advantages which Z. mobilis offers over the various yeasts is the considerably higher specific rate of ethanol production. This property means that selected strains of Z. mobilis are able to produce ethanol appreciably faster than comparable concentrations of yeast. To exploit this characteristic in achieving a high productivity process, a continuous culture system with cell recycle was developed (Fig. 13). Cells of Z. mobilis have high motility with average dimensions of 1-1.4 microns diameter and 2-6 microns length. In the development of a high productivity system, cell recycle was achieved by using various cross-flow microfiltration units together

MEDIUM

T A N G E N T I A L FLOW MICROFILTRATION FERMENTATION FILTRATE

Fig. 13. Continuous culture system with cell recycle using tangential flow microfiltration with various membranes

70

P. L. Rogers et al.

with selected membranes 56) . A polyamide membrane 5 7 ) was found to be most suitable for sustaining a relatively constant permeate rate at high cell concentrations (20-40 g T 1 ) . Cell Recycle with ZM1

The steady-state results with strain ZM1 using 100 g 1 _ 1 glucose medium are shown in Fig.. 14, together with the values for the specific ethanol production rate and the volumetric productivity. The dotted line for q p indicates its maximum value in earlier continuous culture studies without cell recycle while the dotted line for cell concentration shows the minimum concentration required for complete conversion of glucose to ethanol based on this value q p (it was assumed that no loss of cell viability occurred at the high cell concentrations). There is obviously close agreement between the experimental results and these projections. From Fig. 14 it is clear that no loss of activity occurred at high biomass concentrations and that very

DILUTION

RATE

( h"1 )

Fig. 14. Effect of dilution rate on the steady state operation of a continuous cell recycle system with Z. mobilis on 100 g 1 _ 1 glucose medium

Ethanol Production by Zymomonas mobilis

71

high volumetric ethanol productivities (up to 120 g 1 Z. mobilis ZM1 in a continuous cell recycle system. Cell Recycle

with

1

h *) could be achieved with

ZM4

Strain ZM4 was evaluated in a continuous culture cell recycle system with a view to developing a process with both high productivity and a relatively high level of ethanol. As ZM4 was more ethanol-tolerant than ZM1, medium containing 150 g I - 1 glucose concentration was used. To illustrate the improved performance with ZM4, a comparison of data with the two strains is presented in Fig. 15 for dilution rates up to 1.0 h - 1 . Higher ethanol

DILUTION

RATE

( h"')

Fig. 15. Comparative studies with Z. mobilis ZM1 on 100 g 1 1 glucose medium, and Z. mobilis ZM4 on 150 g1 _ 1 glucose medium, in a continuous cell recycle system

P. L. Rogers et al.

72

concentrations (viz. above 70 g l " 1 ) were achieved with Z M 4 and no residual glucose was detected. Other published data with Z M 4 with polyamide membranes in a crossflow microfiltration unit indicated that volumetric productivities as high as 200 g l " 1 h " 1 could be reached 57) . Cell Recycle with Ethanol-Tolerant

ZM481

Using the ethanol-tolerant strain ZM481 continuous cell recycle experiments were carried out with 180 and 200 g l " 1 glucose media. Steady-state conditions were sustained at 180 g l " 1 glucose concentration. The results at this concentration are compared with the maximum steady-state values achieved with Z M 4 using 150 g 1 _ 1

DILUTION

RATE

( h"')

Fig. 16. Comparative studies with Z. mobilis Z M 4 on 150 g 1 1 glucose medium, and Z. mobilisZM4HÌ on 180 g l 1 glucose medium, in a continuous cell recycle system

Ethanol Production by Zymomonas

73

mobilis

glucose medium as shown in Fig. 16. The higher ethanol concentrations attained with ZM481 are most likely a consequence of its enhanced viability at higher ethanol levels.

5.2 Vacuum Fermentation with Zymomonas mobilis The theory behind vacuum fermentation is that it provides a means of removing inhibitory ethanol directly from the culture vessel, thereby allowing higher sugar concentrations in the medium with consequent higher volumetric productivities. With a strain of S. cerevisiae for example, Cysewski and Wilke 7 0 ) reported a productivity of 82 g l" 1 h _ 1 with vacuum fermentation using a feed stream containing 334 g 1 _ 1 glucose. In a comparative study, strains of Z. mobilis have been used in a vacuum fermentation system 71) . One of the problems associated with the use of yeasts in such a system — that of supplying oxygen to the culture vessel to maintain cell viability, 180 CONDENSATE ETHANOL

140 „

PRODUCTIVITY

100 ° 4 Z

60

ETHANOL

BIOMASS

—a 200

ffl-

250 GLUCOSE

GLUCOSE

-ai

300

800

Immobilized yeast (20% glucose) Chibata & Tosa 7 8 )

40

100

2.5

>2100

Immobilized yeast (Molasses - 1 9 . 7 % R.S.) Ghose & Bandyopadhyay 79)

25

71

2.9

>1800

5.4 Comparison of Various Systems To gain some perspective on the improvements achieved with different strains of Z. mobilis and with various modes of fermentation, a comparison is provided in Table 22. Clearly a combination of a relatively high ethanol level together with a high productivity is desirable in minimizing processing costs. High concentrations of ethanol can be obtained in batch cultures, which offer advantages in lowering product recovery costs. However, it is apparent from Table 22 that ethanol productivities are far too low to be competitive with other systems. The problem of low productivity can be overcome in semi-batch culture using a flocculent mutant of ZM4 (viz. ZM401). Providing the problems associated

Ethanol Production by Zymomonas mobilis

79

Table 22. Comparison of ethanol productivities achieved in various culture systems with different strains of Z. mobilis on glucose media System used

Strain used

Glucose input (gl- 1 )

Cone, of ethanol (gl" 1 )

Ethanol productivity (gl-'h"1)

Ref.

Batch

ZM1 ZM4 ZM481

200 300 300

104.0 127.0 124.0

3.3 5.1 2.5

Continuous

ZMI ZM4 ZM481

150 170 150

55.0 60.0 62.6

11.0 12.5 12.0

Recycle

ZMI ZM4 ZM481

100 140 180

45.0 70.0 85.0

120.0 120-200 85*

Immobilized cell reactor

ZM4

150

62.5

53.0

Grote et al. 80)

Vacuum

ZM4

200

60.0

85.0

Lee, J. H. et al. 71>

Semi-batch

ZM401 (flocculent strain)

175

82.0

46.0

Lee, J. H. et al. unpublished

This work

* This productivity was sustained at a dilution rate of 1.0 h 1 . It is anticipated that productivities of 120-200 g 1 _1 h" 1 could be achieved readily with this strain at higher dilution rates

with scale-up and the use of commercial raw materials can be overcome, semibatch culture with Z. mobilis may well offer the best potential for future commercialization of the process, at least in the initial stages. In continuous culture the optimal conditions for the growth of Z. mobilis can be maintained at close to the maximal specific rate of ethanol production. Higher ethanol productivities are obtained when compared to batch cultures, however, the steady-state ethanol concentrations are significantly lower due to the effect of ethanol inhibition on growth. High ethanol productivities can be sustained in a cell recycle system (viz. 120 to 2 0 0 g l _ 1 h - 1 ) and relatively high ethanol concentrations (viz. 85-90 g l " 1 ) can be achieved by using improved strains of Z. mobilis (viz. ZM481). At the optimal conditions the specific rates of glucose uptake and ethanol production can be maintained close to their maximal rates. It would seem that the effective commercialization of such a system will depend on the development of suitable stable and low cost techniques for cell recycle, perhaps by using highly productive strains which can also flocculate. Immobilized cell reactors have been considered as potential systems for commercial ethanol production due to their simple operation, low cost installation and their facility for minimizing ethanol inhibition effects due to their plug flow character. However, the ethanol productivities are found to be much lower than for the cell recycle system due to mass transfer and diffusion limitations.

80

P. L. Rogers et al.

The technology of vacuum fermentation was evaluated also with Z. mobilis71 * and from the data summarized in Table 22, it is clear that some increase in productivity can be achieved. Furthermore the vacuum system allows relatively high sugar solutions (e.g. 350 g 1 _ 1 and higher) to be processed. However, the additional capital and operating costs for a relatively small increase in productivity (approx. 20 %) detract from the commercial potential of the system. The ethanol productivities published for yeasts and Z. mobilis in various systems are compared in Table 23 and it is clear from the data that Z. mobilis is superior to yeast for high productivity processes. From the viewpoint of the commercial potential of the various systems with Z. mobilis it would appear that the semi-batch process in association with a flocculent strain offers the greatest immediate promise. In the longer term the continuous cell recycle system has advantages despite the obvious scale-up difficulties. Essentially such a process would operate with a highly productive strain of Z. mobilis which could maintain cell viability at relatively high ethanol concentrations (e.g. strain ZM481). By operating at high ethanol concentrations (e.g. 85-90 g l" 1 ) there would be little, if any, biomass production as growth would be fully inhibited and ethanol would be produced entirely as a product of the uncoupled metabolism and maintenance energy requirements. This minimization of growth would facilitate cell recycle operation (as there would be little need to remove excess biomass) and would give some additional economic advantage in terms of an increased ethanol yield.

Table 23. Comparison of ethanol productivities achieved in various culture system with Z . mobilis and strains of yeast on glucose medium System used

Microorganism

Glucose input (gl"1)

Cone. ethanol (gl"1)

Ethanol productivity (gl"1 h"1)

Ref.

Batch

Z. mobilis S. uvarum

250 250

119 109

5.9 2.7

This work del Rosario et al. u

Continuous

Z. mobilis S. cerevisiae

135 100

60.0 41.0

12.5 7.0

This work Cysewski & Wilke 8 2 )

Recycle

Z. mobilis Z. mobilis S. cerevisiae

140 180 150

70.0 85.0 60.5

120-200 85.0 32.0

This work This work Ghose &

S. uvarum

200

60.0

36.0

Z. mobilis

200

180.0'

85

S. cerevisiae

334

160.0°

82

T y a g i

Vacuum with cell recycle

" Ethanol concentration in vapour stream from vacuum fermentor

83,84)

del Rosario et al. 11 Lee, J. H. et al. 71) Cysewski & Wilke 7 0 )

Ethanol Production by Zymomonas

mobilis

81

6 Conclusions Given the obvious advantages

o f Z . mobilis

f o r industrial e t h a n o l

production

( T a b l e 2 4 ) it is w o r t h w h i l e t o c o n s i d e r t h e o v e r a l l r e s e a r c h strategy w h i c h

is

n e e d e d t o realise this p o t e n t i a l (Fig. 21). C l e a r l y t h e i s o l a t i o n o f n e w w i l d - t y p e strains o f Z . mobilis

is o f great i m p o r t a n c e in o r d e r t o i n c r e a s e t h e diversity o f

g e n e t i c m a t e r i a l a v a i l a b l e f o r further m a n i p u l a t i o n . T h e c l o n i n g o f g e n e s Z . mobilis

into

w h i c h w o u l d p e r m i t t h e direct c o n v e r s i o n o f c e l l u l o s e , h e m i c e l l u l o s e a n d

s t a r c h - b a s e d r a w m a t e r i a l s t o e t h a n o l is a n e x c i t i n g p r o s p e c t . A s w e l l a s t h e n e e d f o r d e t a i l e d kinetic s t u d i e s o n t h e n e w strains w h i c h are i s o l a t e d or p r o d u c e d b y g e n e t i c m a n i p u l a t i o n , t h e c o n v e r s i o n o f v a r i o u s currentlyutilized r a w m a t e r i a l s b y Z . mobilis

should be assessed. Preliminary experiments with

the j u i c e s f r o m s u g a r - b a s e d s u b s t r a t e s (sugar c a n e , s w e e t s o r g h u m , s u g a r a n d f o d d e r b e e t ) a n d t h e h y d r o l y s a t e s o f c o r n , w h e a t a n d c a s s a v a h a v e r e v e a l e d n o particular d i f -

Tables 24. Advantages of Zymomonas

process for ethanol production

1. Considerably faster specific rates of sugar uptake and ethanol production (rates of 3-4 times faster than yeasts) 2. Higher ethanol yields and lower biomass yields compared to yeasts due to different carbohydrate metabolism 3. Higher productivities (120-200 g l " 1 h " 1 ) in continuous processes with cell recycle (maximum reported values for yeasts are 30-40 g l " 1 h " 1 ) 4. Simpler growth conditions. Zymomonas grow anaerobically and do not require the controlled addition of oxygen to maintain viability at high cell concentrations 5. Ethanol tolerance is comparable if not better than yeasts. Concentrations of 85 g l " 1 (11 % v/v) can be sustained in continuous culture and up to 127 g l " 1 (16% v/v) in batch culture 6. Studies with strains of Z. mobilis over a period of several years have revealed no significant contamination or bacteriophage infection problems 7. The wide range of in vivo and in vitro techniques developed for the genetic manipulation of bacteria can be applied to the important problem of broadening the range of substrates utilized by Z. mobilis to include starch and cellulosic raw materials

1. Strain Selection A Mutation 2. Improve Ethanol Tolerance, Temperature Sensitivity 3. Broaden Substrate Range

1. Studies on e n z y m e s of Entner — Doudoroff Pathway 2. Ethanol Inhibition Kinetics

1. Batch Culture 2. Continuous Culture 3. Micro— calorimetry

1. Techniques to overcome any inhibition problems ¡eg. M o l a s s e s l 2. Evaluation of Sugars,Starch Hydrolysates

1. C o n t i n u o u s Cell Recycle 2. Vacuum Fermentation 3. Cell Immobilization 4. Computer Control & Optimization 5. Scale Up

Fig. 21. Multidisciplinary research strategy developed to evaluate the potential of various strains of Z. mobilis for industrial ethanol production

82

P. L. Rogers et al.

ficulties. The only commercial raw material for which the process was inhibited was sugarcane molasses. Much of this inhibition was due to the high salt levels — an effect which has been partially overcome by membrane "desalting" of the molasses and the selection of salt-tolerant strains of Z. mobilis. In terms of the choice of a suitable high productivity fermentor design, various systems have been evaluated at a laboratory scale. Parallel research has been directed also at the development of suitable on-line instrumentation 8 1 ) and the interaction of a microprocessor with various production systems 7 1 A mathematical model which predicts both batch and continuous culture data has been verified (unpublished results) and will be the basis for future optimization studies. The challenge, however, after more than four years laboratory research lies at the industrial pilot scale level. For example, what problems will be posed by scale-up, particularly of the cell recycle process? Will the flocculation characteristics of strains of Z. mobilis be influenced by various commercial raw materials? And what of the stability of the highly-productive strains during extended semi-batch or continuous operation? Will the higher protein levels of Z. mobilis 65% crude protein) render it more useful than yeast as a by-product for animal feed supplementation? What potential does Z. mobilis offer for significant genetic manipulation to allow direct conversion of cellulose, hemicellulose and starch-based raw materials? These and other questions point the direction for future research. They also emphasize the considerable need for collaborative studies between industry and academia to ensure the full exploitation of a potentially promising new process. Note added in proof: The considerable research by Dr. H. W. Doelle at the University of Queensland on sugar cane juice and other aspecto of Z. mobilis kinetics 8 5 ~ 9 0 ) supports this conclusion.

7 Acknowledgements The authors wish to acknowledge that support for this research was provided by the University of New South Wales, the Colombo Plan and the National Energy Research, Development and Demonstration Program administered by the Australian Commonwealth Department of National Development and Energy.

8 List of Symbols H qs qp qglu qsuc qfru qlev

specific specific specific specific specific specific specific

growth rate sugar uptake rate ethanol production rate glucose uptake rate sucrose uptake rate fructose uptake rate levan production rate

h_1 g g-1 g g-1 g g" 1 g g-1 g g-1 g g-1

h-1 h-1 h_1 h-1 h-1 h-1

Ethanol Production by Zymomonas

Y I/S Yp/S m

mobilis

biomass yield ethanol yield maintenance energy coefficient

83 gg1 gg-1 g g" 1 h - 1

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

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

del Rosario, E. J., Lee, K. J., Rogers, P. L.: Biotech. Bioeng. 21, 1477 (1979) Swings, J., De Ley, J.: Bacteriol. Rev. 41, 1 (1977) Rogers, P. L., Lee, K. J., Tribe, D. E.: Biotechnol. Lett. 1, 165 (1979) Lee, K. J., Tribe, D. E„ Rogers, P. L.: ibid, 1, 421 (1979) Lindner, P.: Atlas der Mikroskopischen Grundlagen der Gaerungskunde, Tafel 68, 3rd Edn. Berlin 1928 Barker, B. T. P., Hillier, V. F . : J. Agrie. Sei. 5, 67 (1912) Barker, B. T. P.: Ann. Rep. Agrie. Hort. Res. Stn. Univ. of Bristol. 174 (1948) Millis, N. F . : Some bacterial fermentations of cider P h D Thesis, Univ. of Bristol, Bristol, U.K. 1951 Millis, N. F . : J. Gen. Microbiol. 15, 521 (1956) Roelofsen, P. A.: Natuurwet. Tijdschr. Ned. Indie 101, 274 (1941) Van Pee, W., Swings, J.: E. Afr. Agr. For. J. 30, 311 (1971) Faparusi, S. E.: J. Food Sei. 39, 755 (1974) Okafor, N . : J. Appl. Bacteriol. 38, 81 (1975) de Lima, G. O., Schumacher, D. I. E., da Silva, E. C.: Rev. Inst. Antibiot. Univ. Recife 10, 3 (1970) Ruiz-Argueso, T., Rodriguez-Navarro, A . : Appl. Microbiol. 30, 893 (1975) Shimwell, J. L.: J. Inst. Brew. 43, 507 (1937) Shimwell, J. L.: ibid, 56, 179 (1950) Lindner, P.: Z. Ver. Dsch. Zuckerind. 81, 25 (1931) Kluyver, A. J., Van Niel, K.: Zentralb. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. II, 94, 369 (1936) Kluyver, A. J.: in: Bergey's Manual of Determinative Bacteriology, 7th Edn. (Breed, R. S., Murray, E. G . D., Smith, N. R. (Eds.)), p. 199. Baltimore: The Williams and Wilkins Company, 1957 Carr, J. G., Passmore, S. M.: J. Inst. Brew. 77, 462 (1971) Carr, J. G. in: Bergey's Manual of Determinative Bacteriology, 8th Edn., (Buchanan, R. E., Gibbons, N. E. (Eds.)), p. 352. Baltimore: The Williams and Wilkins Company, 1974 Swings, J., De Ley, J.: Inst. J. Syst. Bacteriol. 25, 324 (1975) De Ley, J., Swings, J.: ibid. 26, 146 (1976) Swings, J., Kersters, K., De Ley, J.: J. Gen. Microbiol. 93, 266 (1976) Swings, J., Kersters, K., De Ley, J.: Inst. J. Syst. Bacteriol. 27, 271 (1977) Kluyver, A. J., Hoppenbrouwers, W. J.: Arch. Mikrobiol. 2, 245 (1931) Gibbs, M., De Moss, R. J.: Arch. Biochem. Biophys. 34, 478 (1951) Gibbs, M., D e Moss, R. J.: J. Biol. Chem. 207, 689 (1954) Stern, I. J., Wang, C. H „ Gilmour, C. M.: J. Bacteriol. 79, 601 (1960) Ribbons, D. W., Dawes, E. A.: Biochem. J. 81, 3 (1961) Dawes, E. A., Ribbons, D. W., Large, P. J.: ibid. 98, 795 (1966) Sly, L. I., Doelle, H. W.: Arch. Microbiol. 63, 197 (1968) Kersters, K., De Ley, J.: Antonie van Leeuwenhoek J. Micro. Serol. 34, 393 (1968) Schreder, K., Bunner, R., Hampe, R.: Biochem. Z. 273, 223 (1934) McGill, D. J., Dawes, E. A., Ribbons, D. W.: Biochem. J. 97, 44 (1965) McGill, D. J., Dawes, E. A.: ibid. 125, 1059 (1971) Belaich, J. P., Senez, J. C.: J. Bacteriol. 89, 1195 (1965) Dadds, M. J. S., Martin, P. A., Carr, J. G . : J. Appl. Bacteriol. 36, 531 (1973) Richards, M., Corbey, D. A.: J. Inst. Brew. 80, 241 (1974) Ribbons, D. W., Dawes, E. A., Rees, D. A.: Biochem. J. 82, 45P (1962)

84 42. 43. 44. 45. •46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90.

P. L. Rogers et al. Dawes, E. A., Ribbons, D. W „ Rees, D. A.: ibid. 98, 804 (1966) Beiaich, J. P., Belaich, A., Simonpietri, P.: J. Gen. Microbiol. 70, 179 (1972) Bexon, J., Dawes, E. A.: ibid. 60, 421 (1970) Van Pee, W., Vanlaar, M., Swings, J.: Acad. R. Sei. Outre-Mer (Brussels) Bull Science (2), 206 (1974) Dadds, M. J. S., Martin, P. A.: J. Inst. Brew. 79, 386 (1973) Anderson, R. J., Howard, G. A.: ibid. 80, 245 (1974) Senez, J. C., Belaich, J. P.: Colloq. Int. C.N.R.S. 124, 357 (1965) Dawes, E. A., Large, P. J.: J. Gen. Microbiol. 60, 31 (1970) Stephenson, M. P. et al.: ibid. 76, 247 (1973) Bauchop, T., Elsden, S. R.: ibid. 23, 457 (1960) Forrest, W. W.: J. Bacteriol. 94, 1459 (1967) Viikari, L., Nybergh, P., Linko, M.: In: Abstracts of 6th Int. Fermentation Symp., p. 80, London, Ontario 1980 Skotnicki, M. L. et al.: Appl. Env. Microbiol. 41, 889 (1981) Lee, K. J. et al.: Biotechnol. Lett. 3, 207 (1981) Rogers, P. L., Lee, K. J., Tribe, D. E.: Process Biochem. 15 (6), 7 (1980) Lee, K. J. et al.: Biotechnol. Lett. 2, 487 (1980) Skotnicki, M. L., Tribe, D. E., Rogers, P. L.: Appl. Env. Microbiol. 40, 1 (1980) Lavers, B. H. et al.: In: Abstracts of 6th Int. Fermentation Symp., p. 81, London, Ontario 1980 Birnboim, H. C., Doly, J.: Nucleic Acids Res. 7, 1513 (1979) Beringer, J. E. et al.: Nature 276, 633 (1978) Macrina, F. L. et al.: Plasmid 1, 417 (1978) Lee, K. J. et al.: Biotechnol. Lett. 2, 339 (1980) Lee, K. J. et al.: ibid. 3, 291 (1981) Nagodawithana, T. W., Steinkraus, K. H . : Appl. Env. Microbiol. 3, 158 (1976) Van Unden, N.: In: The Yeasts, Vol. 2, p. 75. (Rose, A. H., Harrison, J. S. (Eds.)), New York: Academic Press 1971 Krouwel, P. G., Braber, L.: Biotechnol. Lett. 1, 403 (1979) Navarro, J. M., Durand, G . : Ann. Microbiol. 128 B, 215 (1978) Lee, J. H., Williamson, D., Rogers, P. L.: Biotechnol. Lett. 2, 141 (1980) Cysewski, G. R., Wilke, C. R.: Biotech. Bioeng. 19, 1125 (1977) Lee, J. H. et al.: Biotechnol. Lett. 3, 177 (1981) Kierstan, M., Bucke, C.: Biotech. Bioeng. 19, 387 (1977) Grinbergs, M., Hildebrand, R. P., Clarke, B. J.: J. Inst. Brew. 83, 25 (1977) Divies, C.: French Pat. 844, 766 Wick, E., Popper, K.: Biotech. Bioeng. 19, 235 (1977) White, F. H., Portno, A. D . : J. Inst. Brew. 84, 228 (1978) Wada, M., Kato, J., Chibata, I.: Eur. J. Appl. Microbiol. Biotechnol. 11, 67 (1981) Chibata, I., Tosa, T.: Trends in Biochemical Science, April, 88 (1980) Ghose, T. K., Bandyopadhyay, K. K.: Biotech. Bioeng. 22, 1489 (1980) Grote, W., Lee, K. J., Rogers, P. L.: Biotechnol. Lett. 2, 481 (1980) Lee, J. H. et al.: ibid. 3, 251 (1981) Cysewski, G. R., Wilke, C. R.: Biotech. Bioeng. 18, 1297 (1976) Ghose, T. K., Tyagi, R. D . : ibid. 21, 1387 (1979) Ghose, T. K., Tyagi, R. D . : ibid. 21, 1401 (1979) Cromie, S., Doelle, H. W.: Biotechnol. Lett. 2, 357 (1980) Lyness, E., Doelle, H. W . : ibid. 2, 549 (1980) Cromie, S., Doelle, H. W.: Eur. J. Appl. Micro. Biotechnol. 11, 116 (1981) Lyness, E., Doelle, H. W.: Biotechnol. Lett. 3, 257 (1981) Lyness, E., Doelle, H. W.: Biotech. Bioeng. 23, 1449 (1981) Cromie, S., Doelle, H. W.: Eur. J. Appl. Micro. Biotechnol. 14, 69 (1982)

Growth Kinetics of Photosynthetic Microorganisms Shuichi Aiba Department of Fermentation Technology, Faculty of Engineering, Osaka University, Suita-shi, Osaka, Japan 565

1 Introduction 2 Growth Yield (I), Basic Concept 2.1 Quantum Requirement 2.2 Effect of C 0 2 and/or 0 2 on Photosynthesis 2.3 Culture Conditions of Photosynthetic Microorganisms 3 Growth Yield (II), Practical Aspect 3.1 Opalescent Plate Method 3.2 Integrating Sphere Photometer 3.3 Chemical Actinometry 3.4 Monte Carlo Method 3.5 Determination of Growth Yield 4 Growth Kinetics 4.1 Background 4.2 Specific Growth Rate in Light-limited Environment 4.3 Specific Rate of Light Absorption and Specific Growth Rate 4.4 Phosphorus Content of Algae 4.5 Estimation of Specific Growth Rate in Algae 4.6 Decomposition of Periphytic Algae 5 Algal Growth vs. Environmental Control 5.1 Construction of one-dimensional Model 5.2 Estimation of Parameters 5.3 Estimation of Parameters (cont'd) 5.4 Simulation 5.5 Strategy 6 Summary 7 Appendix 8 Acknowledgement 9 Nomenclature 10 References

86 90 90 91 93 95 95 97 98 100 101 110 110 114 117 119 125 128 131 133 135 137 140 141 142 144 148 148 154

Needless to say, any studies on photosynthetic microorganisms should be accompanied by a proper assessment of microbial absorption of light energy whatever the purposes of these works might be — biomass production, analysis of excessive growth of blue-green algae in waters, evaluation of lightenergy conversion efficiency, etc. In this context, this article begins with a review on the experimental procedures of how to avoid the multiple-scatterings of light before light energy absorbed by microorganisms suspended in liquid could be assessed in situ. This discussion will be augmented by a theoretical consideration in Appendix that has a potential significance in an optimal design of photo-reactor system, if required. However, here in this review paper, the excessive algal growth in still waters attributable to eutrophication would make the point. In order to simulate the emergence of water-bloom in lakes and/or in ponds, quite a few laboratory data on algal growth characteristics, efficiencies of light

86

S. Aiba

energy-conversion, and so forth must be made available in coordination with observations in the field. The idea of how to correlate laboratory data with field observations could be termed "the study on scale-up" which is, indeed, the heart of the Biochemical Engineering.

1 Introduction Photosynthetic microorganisms — photosynthetic bacteria, blue-green and green algae, etc. — have been playing a significant role in the material cycle in lake and marine ecosystems during the past millenniums. Their activities have been hidden from the interests of engineering and science at large except that microbiologists, biochemists, ecologists, etc. have been revealing various characteristics of photosynthetic microorganisms; for instance, biochemists have been interested in the mechanism of light energy conversion to chemical energy, especially in the study on enzymes and biochemical pathways for the photochemical reaction. Rather recently, some people who are engaged in the study of photosynthetic microorganisms have turned their eyes to the application of those microbes for the sake of general public, although a blue-green alga, Spirulina maxima, that grows on the surface of shallow ponds, rich in bicarbonate, is said to have been used, since ancient times, as a source of raw protein by native Africans living nearby in the Republic of Chad 1). It was in the late 40's and/or early 50's here in Japan that microbiologists and plant physiologists in cooperation with chemical engineers began to study mass outdoor cultivation of green alga, Chlorella2). The main interest then was to clarify the algal growth kinetics that is affected by light intensity, temperature, p H of culture medium, etc., and to resolve the problem of how to design and operate the outdoor plant in order to maximize the algal product yield. The outdoor plant, whose efficiency of operation depends principally on the intensity of sunlight and outdoor temperature, has not always been successful in the suburbs of Tokyo; the cell yield of 1 6 ~ 1 7 g m _ 2 d _ 1 experienced in the plant even during a short period of summer season cannot keep up with that of 15 ~ 25 g m ~ 2 d _ 1 for longer periods in many other places in the world 3) . According to Endo et al. who isolated a fast-growing species of Chlorella regularis S-50, the autotrophic growth (maximum value of specific growth rate = 0.30 h - 1 ; gassing with C0 2 -enriched air (volumetric percentage of C 0 2 = 5 %)) was accelerated remarkably when added by glucose or acetate into the culture medium by around 0.1 % (w/v) (mixotrophic culture); maximum values of specific growth rate in mixotroph were enhanced by around 50 % compared to autotroph 4) . Indoor and submerged cultures are feasible and free from any changes in weather conditions. Annual production of Chlorella in Japan (1976) is estimated as about 350 tons (dry matter), coming mostly from mixotrophic cultures 5>. Other Asian countries (The Republic of China (Taiwan) 3 ) , Philippines and Malaysia), however, cultivate Chlorella almost in open ponds; their annual production in 1976 is estimated as about 940 tons (dry matter) 5 ) . Nearly all of these algae produced in Asian regions are processed in Japan as a health food or a source of pigments. A fraction, less than 5% of the total amount (1,290 tons) is used to feed goldfish and carp 5 ) .

Growth Kinetics of Photosynthetic Microorganisms

87

The largest outdoor mass culture of Spirulina sp. is in Mexico; the yield is of the order of 10 g cells m - 2 d - 1 . In 1976, the daily output, 2 tons of dried cells, was exported to Japan for further processing 3) . The mass culture technique of algae throughout the world, including U.S.A., Germany, Israel, etc. has been developed at a snail's pace. Indeed, no outstanding progress in both the culture technique per se and the processing as food has been made during the past 30 years. Unless ways and means to utilize the algal product more extensively not only for feeding chicken, fish, silkworm, etc., but also for direct human consumption as food are developed, the mass algal culture would remain limited, local and sporadic. In this context, it would be worthwhile mentioning the necessity of developing least expensive and most efficient means of cell separation from culture medium. The subject on whether extended applications of algal product as food, source of chemicals, etc. should precede technical improvements of higher cell yield, more efficient separation of cells, etc. or the latter technical renovation should be realized first resembles an argument — which should come first, a chicken or an egg —. Whatever the future aspect of algae as feeding materials to animals or food for human consumption might be, nutritional and technical problems in connection with the use of algae as food or source of chemicals will be put outside the scope of this paper. For separation of cells that needs to be improved, the review article that appeared in 1971 in this series may be consulted 6) . Figure 1 cited from the recent paper of Goldman 3) summarizes an idea of how to utilize algal biomass; it is noted hereby that algal separation must be efficient before biomass can be used for various purposes such as feed, raw materials, from which precious substances are to be extracted, and so forth. Benemann et al. exploited a small-scale strainer 7 ) that could be used to concentrate algal cells most efficiently, but ample room still remains for improving the structure and operation for separation of algal cells on a large scale. Some species of blue-green algae (such as Anabaena, Aphanizomenon, and Gloeotrichia) are characterized by its ability to fix atmospheric nitrogen, and have been used

Fig. 1. Possible applications of algal mass culture 3>

88

S. Aiba

as "natural" fertilizer in farming fields in southeastern Asian countries. Separation of algal cells from suspending media is not needed, and the fields, in which blue-green algae have flourished, can be reclaimed alternately. As was mentioned earlier, extraction of commercial chemicals from algal biomass, the use of algae as human protein supplement, etc., described in Fig. 1 are not so straightforward. In addition, anaerobic digestion as illustrated in the figure to produce methane as "energy" or to extract commercial chemicals are still in the stage of "idea", though "energy production" sounds stimulating in the day of "energy crisis" 8). As a corollary, this review article does not comprehend topics that serve to resolve technical problems on energy "extraction" out of algal biomass. As apparent from Fig. 1, algal growth is closely related to the removal of nutrients — phosphorus, nitrogenous as well as carbonaceous materials from wastewaters. Herein lies an argument that algae can be used to abate the nutrient loading in polluted waters, but commendable means to harvest the algal crop must be made available 91 before the argument becomes fruitful. In fact, many papers have been presented from various circles to utilize algae to remove inorganic phosphorus from effluent in the secondary treatment plants of sewage and industrial wastes 1 0 ' U i 12)_

It is well known that enrichment of rivers and/or streams gives rise to "eutrophication", — over-nourishment — of sea and/or lake waters 13>14) . Eutrophication sometimes triggers the emergence of red tide in coastal areas and water-bloom on the surface of inland waters. Once the red tide and/or water-bloom occurs, a great damage will be inflicted upon fish-farms, coastal fisheries, etc. because of suffocation of fishes, oysters, etc., while a huge capital investment is required to remove offensive smells, if the inland waters are used for drinking. Algal growth that is ubiquitous in terrestrial and littoral waters does have a significant role either to aggravate the environment or to benefit the society when an adequate management is warranted. The growth of alga can be designated as an index to judge the progress of eutrophication when too frail to permit the chemical analysis of phosphorus in the water. AGP (algal growth potential) is advocated as a means to evaluate eutrophication in waters 15 • 16 • 17) . Blue-green algae such as Microcystis aeruginosa, Microcystis sp., Anabaena sp., etc. are known to become prosperous — water-bloom thriving — when the environment of inland waters such as pH, temperature, nutrient concentration, etc. turns out to be favorable for the algal growth 18). On the other hand, Flagellata such as Heterosigma sp., Exuviaella sp., Skeletonema costatum, etc., are isolated from seawater when red tide occurs 19). It is not reasonable to attribute simply any specific factors involved in eutrophication to the cause of water-bloom 20) and/or red tide. The real cause of this undesirable phenomenon is too teemed with various factors to facilitate a straightforward analysis. In spite of the complicated picture on the cause of water-bloom and/or red tide, it might be said that occurrence of red tide and/or water-bloom should be closely related to an anomaly of algal growth behavior in the ecosystem. Accordingly, it might not be wide of the mark to refer to a subject on growth yield of algae, particularly in terms of light-energy-conversion efficiency later on in this article. Algal species reviewed in this paper will be restricted only to a few of photosynthetic microorganisms in inland waters. Whatever the objective of algal cultivation either in

Growth Kinetics of Photosynthetic Microorganisms

89

a closed vessel under light irradiation or in an open and outdoor pond might be — biomass production, waste treatment, and so forth —, a question of what is the efficiency of conversion of light energy to cell material is stimulating. Data on algal harvest, 15 to 25 g m~ 2 d" 1 , etc. 3 ' are reported, while those on quantum requirement, 11~16 quanta needed to reduce one molecule of C 0 2 in Chlorella cells are available elsewhere 21 \ However, the former comes from field data and hence, it is difficult to assess from the data the efficiency of algal cells synthesized per light energy absorbed. By the same token, the latter data on quantum yield (reciprocal of the quantum requirement) per se are not necessarily equal to the conversion efficiency in the algal growth. Consequently, it is worthwhile reviewing some procedures of how to evaluate light energy absorbed by photosynthetic microorganisms in a bio-photoreactor. Instead of the quantum yield, Y kJ (g cells harvested per light energy (kJ) absorbed, g k J - 1 ) and/or Y (chemical energy retained in cells per light energy absorbed, %) values will then be discussed with respect to some species of blue-green and green algae, and in addition, a photosynthetic bacterium. However, prior to the discussion on Y kJ (or Y) values, not a few references on physiological and biochemical aspects of photosynthetic microbes will be briefly reviewed to make clear a background of the algal growth yield. Lambert-Beer's law is usually used to estimate light intensity within a culture medium; in fact, a bio-photoreactor is conventionally designed from the law on the attenuation of light energy along the depth of liquid column. It is important to reassess here the effect of light-scattering caused by the presence of photosynthetic microorganisms on the light-intensity distribution within the reactor. Even though experimental methods to estimate light energy scattered by the cells are available (see later), the approach only by experimentation is not always satisfactory from the viewpoint of designing bio-photoreactor. The latter sophisticated analysis on the distribution of light intensity in a reactor will be referred to in Appendix. The purpose of this article is to present a basis of how to simulate the algal growth in an eutrophic environment. Even though the algal species will change from one to another, depending on the topics in this paper, the growth kinetics to be presented will become the basis, from which simulations of water-bloom and of whatever else in the algal growth could be made. Innumerable workers of limnology, ecology, plant physiology, etc. have been continuing studies on photosynthetic microorganisms during the past 100 years, especially during more than 30 years in this century 22) . Yet, there has been accomplished no and/or little progress 23) on the important problem of how to control the emergence and/or disappearance of a specific of algae in nature, albeit one cannot overlook splendid achievements by various workers on the mechanism of light energy as converted to chemical energy in those photosynthetic microorganisms. To be modest, it is important to "synthesize" (or "simulate") microbial phenomenon (an anomaly of algal growth, for instance) from biochemical and physiological studies of algae. A substantial aim of "simulation", even though unsophisticated, is considered to be in finding the possibility to extract factor(s) responsible for a particular event such as deterioration of water quality due to algal growth by comparing calculation with observation, and to unveil clue(s) to control the environment.

90

S. Aiba

2 Growth Yield (I), Basic Concept 2.1 Quantum Requirement It is well known that pathways through which electrons are transferred to reach ultimately to NADP by the stimulus of light quantum (photon) in algae are not existent in photosynthetic bacteria. Photosystem II in algae that deals with oxidation of water to gain proton and electron, and to release oxygen is absent from the bacteria. Species of pigment to catch up light energy are different among those photosynthetic microorganisms. Chlorophylls, carotenoids, and phycobilins are the principal types of pigment in algae, whereas in photosynthetic bacteria there exist, in addition to carotenoids, bacteriochlorophylls, whose absorption peaks extend nearly from X = 750 to 950 nm depending on the bacterial species 2 4 , 2 5 ) . Absorption maxima of algal pigments are in the visible range; for example, chlorophyll a at X = 420, and 660 nm; P-carotene at X = 425, 450 and 480. Contents and species of those pigments are subjected to change depending on culture conditions. Taking for granted the difference in electron transfer mode between algae and photosynthetic bacteria, the two-light reaction scheme of photosynthesis with respect _ to algae will be referred to here, i.e. light energy, hv, of one photon, where h = Planck's constant and X = c/X = frequency of light, c = velocity of light in vacuum space, X = wavelength, is doubled to have e~ transported to NADP. According to Calvin cycle, 2 molecules of NADPH must be oxidized to have one molecule of C 0 2 reduced photosynthetically 26) . Then, quantum requirement defined by the number of photons required to reduce one molecule of C 0 2 (or the reciprocal = quantum yield) 271 turns out to be (2 x 2 x 2 = 8), independent of the wavelength of light (quantum yield = 0.125), since 2e~ must participate in yielding NADPH from NADP + + 2e~ + H + . The classical work of Warburg and Negelein 28) claimed from their experimental work on Chlorella that quantum requirement ( = 4 ) remains unchanged irrespective of the wavelength of light (X = 436 ~ 660 nm). Since photon has energy of 12.0/A. J m o l - 1 , the quantum requirement ( = 4 ) corresponds to the efficiency of conversion from light to chemical energies as 46% (X = 436 x 10" 7 cm) and 69% (X = 660 x 10" 7 cm), assuming that one mol of C 0 2 assimilated photosynthetically corresponds to 24 g cells, whose enthalpy change is taken as 21 kJ g _ 1 (see later) 29) . The quantum requirement of 4 for Chlorella is doubtful, but no definite answer to this question has been presented. Emerson and Lewis observed that quantum requirement of Chlorella depends on the wavelength of light used and claimed the value as 16 (X = 480 nm) to 11 (X = 680 nm) per molecule of C 0 2 reduced 21) . Although ample room is still left open for careful discussion on the reason to have caused a marked disagreement of the quantum requirement observed in these previous works, the fact that fairly concentrated suspensions of Chlorella were used in both cases is worthy of attention. In the former work of Warburg and Negelein, no transmitted light through the concentrate algal suspension was assumed to facilitate the assessment of light absorbed by the cells 28) . The concentrated suspension of Chlorella used in the work of Emerson and Lewis allowed only a fraction (about 5 %)

Growth Kinetics of Photosynthetic Microorganisms

91

of the incident light energy to pass through the suspension 21) . Under these circumstances, the algal cells might not have been exposed equally to the irradiation of light due to the shading by neighboring cells, and this could have distorted algal physiological activities as a whole. In this context, the quantum requirement remains to be reassessed. To ward off the shading effect on the algal physiology, the use of a dilute suspension of photosynthetic microorganisms is required. However, energy of light scattered necessarily by the cells must be accounted for before dependable measurements of light energy absorbed by the cells are made possible. The experimental procedure to measure light energy absorbed by a dilute suspension of algal and/or bacterial cells will be described in the next chapter.

2.2 Effect of C 0 2 and/or 0 2 on Photosynthesis Here again, photosynthetic bacteria and algae cannot be treated in the same category so far as the discussion on C 0 2 and/or 0 2 effect on photosynthesis goes. Photosynthetic bacteria have to be exluded from the following argument primarily because of the essential difference in photosynthetic mechanism especially relevant to C 0 2 and 0 2 . In spite of the exclusion of photosynthetic bacteria from this discussion, it is, nevertheless, of significance to refer to the effect of C 0 2 and/or 0 2 on algal growth as another basis of the discussion on growth yield in this chapter. Ever since the discovery of photorespiration in a leaf of hybrid tobacco by Dekker 30) who observed a rapid deceleration of C 0 2 evolution immediately following extinction of light that has irradiated the leaf, not a few workers have studied the existence and/or non-existence of photorespiration in microalgae 3 1 , 3 2 ' 3 3 ) . In fact, this topic of photorespiration in microalgae seems to have been admixed with the adverse effect of O z on photosynthesis in algae (Warburg effect). The argument is significant from a viewpoint that the algal growth could have been controlled by establishing an environment to minimize photorespiration, if the photorespiration were to play an appreciable role in C 0 2 reduction in photosynthetic mechanism of these algal cells. It might be inferred that pros and cons of photorespiration in microalgae have not reached yet the final settlement, and the argument per se seems to be outside of this brief review on growth yield of photosynthetic microorganisms. However, the above argument is considered to be related closely with the discussion on competitive inhibition of Ribulose-l,5-bisphosphate (RuP 2 ) carboxylase and oxygenase by 0 2 and/or C 0 2 — the key and bi-functional enzyme at the threshold of C 0 2 fixation in photosynthesis 3 4 , 3 5 ) . If partial pressures of 0 2 and C 0 2 in a culture medium of algae were controlled in such a way as to permit dominant performance of RuP 2 oxygenase rather than carboxylase, in another term, oxygenation overwhelming carboxylation at the entrance of photosynthetic fixation of C0 2 , phosphoglycolate resulted from the oxygenation might lead to glycolate excretion (a specific feature of photorespiration), a n d vice

versa.

The following equation will be used to discuss the effect of C 0 2 and/or 0 2 on photosynthesis; this equation deals with competitive inhibition of enzyme kinetics to

92

S. Aiba

assess a relative magnitude of oxygenation reaction of RuP 2 carboxylase i - J L . i

^ C ^ + K J l + C^/K;)

35 36)

'

. (i)

provided: v = rate of carboxylation, M s - 1 vmax = maximum rate of carboxylation, M s " 1 CCQ2 = C 0 2 concentration in liquid, M CQ2 = 0 2 concentration in liquid, M K m = Michaelis constant with respect to C0 2 , M K; = equilibrium constant between enzyme (RuP 2 carboxylase) and inhibitor (0 2 ), M It is possible to assess the relative magnitude of oxygenation from Eq. (1), once both values of K m and K, are given. It is assumed in this assessment that partial pressure of 0 2 and C0 2 , i.e. p 0 2 and p cc , 2 are in equilibrium with 0 2 and C 0 2 in water (30 °C), respectively. Taking K m = 3.4 x 10" 5 M, K; = 3.7 x 10" 4 M , 3 4 ) and Bunsen coefficient, a C o 2 = 0.665 ml (ml H 2 0 ) _ 1 , lamps, but even the incident light intensity is not shown in consistent terms as apparent from the table. Obviously, it is hard to extract from Table 1 any idea on growth kinetics, i.e. specific growth rate as affected by incident light intensity, temperature, pH, etc., and the subject will be discussed separately later on in this paper. The table gives partly a sectional view on the microbial growth despite the essential difficulty to handle consistently different divisions, genera and species of the photosynthetic microorganisms.

Genus

(sn -SflOOdO'lft) s'iq.sfîovuy

'pi

§

•ti

Si

B3XqdoupXQ

sCM O

s-s

.-H

m 1 /-s

1 oo

v^i^domo

1 ° vT

•a s

•ti

CN rH m t C M X m t

O

o

1

r-.

O

CM

0 cn

(0.7 - 4)

rH

iH en x 0 \0 w 00 rH

\o

sphaeroides S

0.22 (max. value)

0.10 (max. value)

o s© m «a- rH C M rH CM i-H CN 0 r-«- r-» r>. r-. tn u-> m m r-«. oo r-

O en oo o r-- oo rCM C MCM ~ CM en en cn

en rH m

5

-a

•a u

00 UJ* >~) M

00 o >h

e

i

^UID

6-S

=i

SB ft

i—1 >-) 1 •i! >->

U

= 1 ~

L

=

t j - J i

o

$

60 CO

J3 S ft iLO

en

S

10 is the rate-limiting step in the supply of C 0 2 for algal growth, whereas dehydration of carbonic acid into C 0 2 for pH < 8 is quite rapid and most unlikely to limit the supply rate of C0 2 . Designating M' to the rate of carbon supply from the dissociation of bicarbonate ion in water, M ' = k H C O -C T C (l/2~)

(27)

Eq. (27) assumed tacitly that total inorganic carbon composed of bicarbonate by one-half or less 4 5 , 7 9 ' limits the rate of carbon supply under a particular range of pH ( > 10) in water 45) . On the other hand, the amount of carbon, M", required for sustaining the specific growth rate, |i, of algae is shown by: M " = a'nX provided : M' M" kHCOC TC X a' |x

= = = = = = =

(28)

carbon supply rate, g carbon 1 _ 1 d " 1 carbon consumption rate, g carbon 1 _ 1 d - 1 dissociation rate constant for HC0 3 " C 0 2 (aq.) + O H , s total inorganic carbon concentration, g 1 _ 1 cell concentration, g 1 _ 1 carbon content in algal cells, g carbon g - 1 specific growth rate, d " 1

1

In Eqs. (27) and (28), the following values were employed to estimate M' and M " values : k H ™- = 2 x 10" 4 s" 1 at 25 °C45> C TC

= 10xl0-3gl"18O).

The above-mentioned value of C x c was cited from field observations of a eutrophic lake 80) . In addition, values taken for a', X and n were: a' = 0.5 g carbon g " 1 , 4 0 ' X = 50 x 10~3 g l - 1 , 8 0 ) and H = 1 d " 1 (cf. Table 1).

Growth Kinetics of Photosynthetic Microorganisms

113

Then M' = = M" = =

(2x 10~4) (3.6 x 103 x 24) (10 x 10~3) (1/2 0.086 ~ g carbon l" 1 d " 1 (0.5) (1) (50 x 10" 3 ) 0.025 g carbon l" 1 d " 1

Consequently, M'

M" .

Accordingly, carbon supply in polluted waters would not limit algal growth. Chemostat data on both Scenedesmus quadricauda and Selenastrum capricornutum at 27 °C, pH = 7.3 showed that saturation constants in a Monod equation (|i vs. C TC ) were of the order of 0.4~0.6 mg P 1 (total inorganic carbon) 4 5 ) , respectively. Those values are considerably smaller than C TC value ( = 10 mg l" 1 ) used in the earlier calculation. This might also support that carbon is unlikely to limit the growth of algae 81) . In spite of the conservative assessment mentioned above, it is interesting to find out a reference that claims carbon as the substrate to limit algal growth 82). This controversial paper emphasizes the symbiotic role of bacteria to supply C 0 2 to algae, albeit the reference is still open for further rebuttal. As a consequence, the nutrient to limit algal growth in natural environment will be assumed as phosphorus 83) (and light intensity) in this paper. Herein lies a motive to refer later on in this chapter to phosphorus uptake rate as viewed from growth kinetics and also, to simulate and control phosphorus level in a shallow river in the next chapter. Now, it would be worthwhile to refer to heterotrophy of carbon in the growth of photosynthetic microorganisms, because organic carbon such as glucose, acetate, etc. is fairly rich in some wastewaters and in fact, the growth of these microbes could be utilized to reduce the pollution. As was mentioned previously, photosynthetic bacteria require hydrogen donors such as lactate, succinate or acetate for their growth. Some green algae and diatoms can grow in dark by metabolizing both glucose and acetate (Chlorella ellipsoidea, Chlorella pyrenoidosa, Nitzschia alba, etc.), while bluegreen algae such as Anabaena sp., Oscillatoria spp., etc. exhibit dark growth on glucose but not on acetate. The picture of metabolic capabilities of these algae on a specific organic substrate is crisscrossing 84) . According to Droop, the value of saturation constant in the Monod equation of growth for heterotrophic bacteria that can metabolize mannitol, glycerol and lactose ranges from 2 to 20 mg 1 _ 1 , whereas the value for algal growth in the dark is somewhat higher; for instance, 7 5 m g l _ 1 for glucose in Navícula pelliculosa (diatom) and 57 m g - 1 for lactate in Cylindrotheca fusiformis (blue-green alga) 84 '. Incidentally, the saturation constant for acetate in Rhodopseudomonas spheroides S (of light- and substrate-limited anaerobic culture) was 69 mg l - 1 5 9 ) . Whatever the objective of growth kinetics in photosynthetic microorganisms might be — assessment of saturation constant in a Monod equation, determination of growth-limiting substrate, etc. —, the decay and succession of species in natural environment is deemed to be a significant subject. The subject on decay and succession

114

S. Aiba

Fig. 8. A diurnal variation of DO (dissolved oxygen) in a shallow stream (Tama-gawa, running through the metropolitan area of Tokyo) (example taken on Aug. 15, 1972) 85>

Chronological time Aug.15, 1972 will be finally mentioned in this chapter. Fig. 8 demonstrates a diurnal variation of DO (dissolved oxygen) in a shallow stream (Aug. 15,1972) that runs through the metropolitan area of Tokyo 85) (water temperature = 27 °C, saturation of DO = 7.47 mg • l - 1 37>). A marked super-saturation around noon (nearly 250% of saturation) observed could apparently be attributed to the photosynthetic activity of algae. The super-saturation of DO observed diurnally comes from periphytic as well as planktonic algae (the order of magnitude in contribution by the former periphytic algae far exceeds the latter in this instance). The periphytic algae will be released after growth, and also they will be subjected to decay and succession in a long run. A reasonable assessment on the decay and/or succession of algae is particularly significant before one could establish a model that simulates the behavior of algae in a shallow stream. The marked variation of DO in running waters (cf. Fig. 8) has been pointed out and simulated by previous workers, but they have not necessarily paid attention onto the physiological characteristics of algae involved 8 6 , 8 7 ) .

4.2 Specific Growth Rate in Light-limited Environment Ryther studied extensively the effect of solar light intensity on photosynthetic activities of 14 species of planktonic marine algae during the summer of 195 5 88). He measured the photosynthetic activity of each species of Chlorophyta, Diatoms, and Dinoflagellates by using the technique of 14COf " uptake. Before the addition of 14C into a 20 ml glass-stoppered bottle, containing seawater suspension of each alga, the bottle was exposed nearly for 2 h to sunlight to minimize the error due to come from adaptation. The ratio of photosynthetic activity at a given value of solar light intensity to the maximum activity ever observed in each run was defined as relative photosynthesis, and this ratio was plotted against solar light intensity.

115

Growth Kinetics of Photosynthetic Microorganisms 1.0

Fig. 9. Relative photosynthesis vs. solar light 88) intensity 88)

00

20 Solar

light

40 intensity

60

80

100

( k lux)

Figure 9 is a diagram (relative photosynthesis vs. solar light intensity) that has resulted from averaging the data on these marine planktonic algae 88) . In view of the less importance of Chlorophyta in marine algae, Ryther secured the diagram (Fig. 9), weighting the data on Chlorophyta by 50%, while 100% for Diatoms and Dinoflagellates. The temperature ranged from 18 to 23 °C, but the effect of temperature on relative photosynthesis remained to be disclosed 88) . It is evident from Fig. 9 that relative photosynthesis is enhanced up to the solar light intensity of nearly 30 klux, beyond which the photosynthetic activity deteriorates; the latter range of light intensity might correspond to inhibition due to intensive light in analogy to inhibition of high substrate concentration in enzyme kinetics 89) . Assuming that photosynthetic activity, P, expressed either by 0 2 evolution rate per unit cell (or unit amount of chlorophyll a) or by specific growth rate, etc. is shown by the same function as in enzyme kinetics, the pattern in Fig. 9 will be shown as follows: P =

(29)

a' + b'l + c'l 2

where P = photosynthetic activity, g 0 2 (g cell) phyll a ) ' 1 h" 1 I = light intensity, klux or kJ c m - 2 h " 1 a', b', c' = empirical constants

1

h

1

or g 0 2 (g chloro-

The value of c' in Eq. (29) might be taken as zero in laboratory experiments. A hyperbolic function of I to affect photosynthetic activity of microalgae has been demonstrated previously by many workers, for instance, by Tamiya et al. on Chlorella ellipsoidea 2). According to Ikushima, photosynthetic activities of aquatic plants are also expressed by Eq. (29), wherein c' = 0 90) . The effect of temperature on photosynthetic activity of microalgae is shown in Fig. 10, in which specific growth rate, n h - 1 , of Microcystic aeruginosa is plotted against light intensity, I 0 , of a tungsten lamp through an infrared cut-filter 91 \

S. Aiba

116 025

3 6 "C

n-O-

33"C

•

0

Fig. 10. Specific growth rate, n, of Microcystis aeruginosa vs. light intensity, I0 (tungsten lamp with infrared cut-filter); parameter is temperature. 91) Symbols: O 20 °C; A 24 °C; • 29 °C; 33 °C; * 36 °C. (pH = 7.5 ~ 8.5) 10

15

Io Cklux)

20

25

Parameter in Fig. 10 is temperature in each determination of n, provided: specific rate of 0 2 evolution, Qq 2 , observed at given temperatures and light intensities was converted to specific growth rate (see Eq. (34) that appears later). Algal cells used for the observation of Qq 2 were sampled from a turbidostat culture at 33 °C and I 0 = 2 klux, and the algal suspension was transferred to a specific device that has been kept at a desired temperature to measure Qq 2 within 10 min at various values of I 0 99>. In another word, the ordinate reading of p. in Fig. 10 does not correspond to the actual value of specific growth rate of this alga exposed indefinitely to the irradiation. Q 0 l values were simply translated into the value of \i. The effect of temperature on (i is not clear in Fig. 10 when I 0 is less than 4 klux, while that effect is most evident when light intensity exceeds 5 klux. The specific growth rate, (j. is apparently governed by a hyperbolic function of I 0 (cf. Eq. (29), where c' = 0) for each temperature. Myers and Kratz 9 2 ) reported the same pattern as in Fig. 10 for Anacystis nidulans (Qo2 vs• light intensity, parameter being temperatures, 25 and 39 °C), whereas exactly the same feature of temperature effect, 2, 8 and 20 °C, on photosynthetic activity, mg carbon (mg chlorophyll a ) - 1 h " 1 was presented by Nielsen and Jorgensen for Skeletonema costatum (diatom) 93) . Irrespective of the difference in algal species used, the photosynthetic activity expressed either by specific rate of 0 2 evolution or by specific growth rate is linearly correlated with light intensity and is insensitive to temperature, when the light is weak. This is the so-called "light-limited" environment. When irradiation light is intensified exceeding the "light-limiting" condition, the effect of temperature on (x becomes revealed, (x values that levelled off at each temperature are apparently controlled by enzymatic reaction(s), whose rate(s) being enhanced by the increase in temperature, as demonstrated in Fig. 10. Assuming that the slope of a linear (lightlimited) region in the figure depends on contents of pigments 9 2 ) , i.e. on prehistoric picture of algal cells used for Q 0 2 determination, Bader presented a structured model of photosynthesis in microalgae 94) .

Growth Kinetics of Photosynthetic Microorganisms

117

4.3 Specific Rate of Light Absorption and Specific Growth Rate It is most evident that light intensity is an important parameter to define the environmental condition for algal growth. If a sophisticated analysis of light energy conversion to cell material is required, light absorbed by the cell rather than the incident light needs to be appreciated (cf. previous chapter). Data in Tables 2 and 3 list £ values, kJ g ~1 h " 1 . Equation (17) and Fig. 6 will be discussed further in this context. Rhodopseudomonas spheroides S 59) was cultivated batchwise in a bio-photoreactor (the same one as used in Table 3). The medium composition using acetate as hydrogen donor, temperature at 30 °C, pH = 7.0 ~ 7.2 and anaerobic cultures, blowing nitrogen gas were also the same, respectively. However, in this batch culture, incident light intensity, I 0 kJ cm~ 2 h - 1 , was kept constant at nearly 5 x l O " 3 k J c m " 2 • h - 1 , whereas in the previous study on continuous culture, I 0 values ranged nearly from (1.4 ~ 5.6) x 10" 3 k J cm" 2 h " 1 (see Table 3). Specific growth rate was estimated from batph culture and by using Fig. 6, it was possible to plot specific rate of light absorption against specific growth rate as shown in Fig. 11 59) (cf. Eq. (17)). Solid and open circles in the figure correspond to those in Fig. 6, respectively. Despite a marked deviation from the linear correlation as indicated by Eq. (17) when the previous data in Table 3 were used (not shown here, and this point was referred to earlier), both circles in Fig. 11 are deemed to justify Eq. (17). Solid lines passing through these data points in Fig. 11 were taken arbitrarily. However, the line through solid circles intersects with the ordinate at negative value. This implies that the maintenance coefficient, m, in Eq. (17) is negative. Clearly, the negative value of m is unacceptable. Accordingly, solid circles that were estimated from Lambert-Beer's law, disregarding the scattering of light by the bacterial cells in the reactor will be discarded 59) . Another line through open circles yielded: (cf. Eq. (18))

0.01

0.02

Specific growth rate I IT' )

Fig. 11. Specific rate of light-energy absorption vs. specific growth rate (light-anaerobic and batch culture of Rhodopseudomonas spheroides S at 30 °C, pH = 7.0 ~ 7.2) 59) . Symbols : O light absorption estimated from Monte Carlo method; • light absorption assessed from LambertBeer's law, disregarding light scattering

S. Aiba

118

Fig. 12. Experimental data of light-anaerobic and batch culture of Rhodopseudomonas spheroides S at 30 °C, p H = 7.0 ~ 7.2, see open circles. Eq. (30) is shown by curved surface, where (Y k ,) G = 7.9 x 10" 3 g k j - 1 , m = 6.7 x 10~ 2 kJ g " 1 h " 1 , and K = 0.51 x 10" 3 M. Short bars between open circles (observation) and cross symbols (calculation) in the figure represent the deviation 591

If data for Run Nos. 26 and 27 (I0 = (4.8 ~ 5.6) x 10"3 kJ cm" 2 h _ 1 ) in Table 3 (continuous culture of Rhodopseudomonas spheroides S) were used, 441 and if the linear relationship between E, and |i ( = D) were assumed, m = 0.2 kJ g _ 1 h - 1 , (YkJ)G ^ 8.8 x 10"3 g k J - 1 could be obtained. However, other data (Run Nos. 24, 25 and 28 in the table) for smaller values of I0 did not yield any reasonable data of m and (YkJ)G due to unknown reasons. This controversial point remains to be dissolved. Apparently, open circles in Fig. 11 represent the "light-limited", residual concentration, S, of acetate being far from limiting. Consequently, an overall picture, in which acetate concentration becomes another factor to limit the bacterial growth rate will be mentioned. Hence, data for open circles, added by 3 more data points for smaller values of S and (i, were rearranged as shown in Fig. 12. The equation to represent a curved surface of S) in Fig. 12 is:

H=

1

K + S

(30)

(Yk provided: K = saturation constant, M In connection with Eq. (30), a question on the use of additivity of the specific growth rate originating from light energy absorbed and that coming from substrate (acetate) energy expended may arise (cf. Eqs. (21) and (22)), i.e. H = (Y u ), v + ( Y k J ) 2 ( $ - m )

(30)'

provided: v = specific rate of substrate (acetate) consumption, kJ g _ 1 h ' 1 (YkJ)i = growth yield from acetate, g k J - 1 (YkJ)2 = growth yield from light, g kJ" 1 However, the following facts that experimental data in Fig. 12 dealt with lightanaerobic culture of Rhodopseudomonas spheroides S, using acetate as hydrogen donor

Growth Kinetics of Photosynthetic Microorganisms

119

and no growth could be observed in the absence of acetate could justify the use of Eq. (30). Parameters in Eq. (30), i.e. m, (Y k ) G and K were estimated, different from the preceding estimation of m and (Y kJ ) G , by a least-squares method for nonlinear functions, SALS (Statistical Analysis with Least-Squares Fitting Program, Computer Center, Univ. of Tokyo) 59). The result of this estimation was: m = 6.7 x 10~2 kJ g _ 1 h - 1 , (YkJ)G = 7.9 x 10" 3 g kJ" 1 , and K = 0.51 x 10" 3 M. Cross symbols in Fig. 12 represent calculated values of (i, while open circles are experimental data, most of which appeared already as open circles in Fig. 11. Short bars in Fig. 12 designate deviation between observation and calculation. The 3 dimensional diagram was prepared by using a perspective program (Computer Center, Univ. of Tokyo) and an X-Y plotter subroutine package (Hitachi Works, Tokyo) 5 9 ) . With respect to Rhodopseudomonas spheroides S, the fact that nearly the same value °f (YkJ)G = (8 ~ 9) x 10~3 g k J - 1 were obtained despite the difference both in culture conditions (batch or continuous) and in estimation procedures is interesting. However, m values differ considerably from 0.2 to 0.07 kJ g _ 1 h - 1 . It is also interesting that (YkJ)G and m values of Rhodopseudomonas spheroides S were within the range for those of blue-green algae, respectively (cf. Eq. (18)). Ragonese and Williams 95} presented a kinetic model for algal growth and attested to a rationale, referring to published data on growth of Chlorella ellipsoidea 2). They integrated the rate equation (proportional to light absorption rate) by using LambertBeer's law. A constraint of this model that was derived from Lambert-Beer's law is self-explanatory (cf. Fig. 6). The proper estimate of light-scattering from photosynthetic microorganisms in a culture medium is not always feasible even if Monte Carlo method is available. It might be urged that Lambert-Beer's law could be used to estimate the light-intensity distribution in an open field such as lake and/or running waters, because the cell concentration is usually in a region near the origin of Fig. 6 for instance, where no appreciable difference between the use of LambertBeer's law and that of Monte Carlo method is noted.

4.4 Phosphorus Content of Algae It has been demonstrated previously in this paper that phosphorus is most likely to be a substance to limit the growth of photosynthetic microorganisms, microalgae in particular when eutrophication and water-bloom in inland waters are discussed. Unless otherwise noted, orthophosphate (or phosphate) will be simply stated hereafter as phosphorus, and its concentration in liquid medium will be expressed as mg P ml" 1 throughout. Needless to say, phosphorus is an indispensable substance to sustain the living of microorganisms whatever their genera and species might be. Inorganic phosphorus, condensed inorganic phosphorus (polyphosphate) in both acid-soluble and acidinsoluble fractions, and phosphorus bound with nucleic acids and organic substances (carbohydrate, lipid and/or protein) are the principal phosphorus compounds in microbial cells. Polyphosphates that are found ubiquitously in bacteria, blue-green and other algae, fungi and also, in protozoa are the "storage" materials to supply

120

•rg _Ë "3 o

S. Aiba

o

5.0

/

o»

o

§ 4.0

o "o

ora

/

9%

o

o„ o„

E

o 5

3.0

c f 2.0

o.

o

o13

o o

P o , o

1.5

o3

1.0

°0

6

10

fP

15

1 20

o„

Fig. 13. Photosynthetic activity, Q 0 j nmol 0 2 m g - 1 m i n - 1 , of Chlorella ellipsoidea vs. phosphorus content, f P ng P mg " 1 ; the vicissitude of f P is demonstrated. Symbols: 0> Chlorella ellipsoidea cultured in P-free medium at 30 °C, pH = 6.0 (initial). O , Chlorella ellipsoidea, which had been starved with phosphorus, was transferred into basal medium (phosphorus concentration = 2 8 5 m g P l - 1 ) after rinsing the algal cells with P-free medium. Subscript numbers indicate the lapse of time, t, h, in the cultivation (after the transfer) at 30 °C, pH = 6.0 (initial). Arrows refer to the direction of each "trajectory" of f P

(^gPcng "W)

phosphorous to the cells when needed 96) . It is well known that phosphatase activity is enhanced remarkably when an algal culture becomes "phosphorus-limiting" 97) . Despite the fact that the role and function of phosphorus and/or polyphosphate in microorganisms have been studied extensively by many workers 96) , there have been published only a few works on kinetic behavior of phosphorus. For instance, a subject of how to correlate specific uptake rate of phosphorus of microorganisms in phosphorus-limiting environment to its specific growth rate still abounds with uncertainties 98) that are remaining to be disclosed. Attention will be paid in this paper onto a kinetic feature on phosphorus in some algae that appear commonly in the study of environmental pollution and its control. An example of the vicissitude of phosphorus content, f P n g P m g - 1 of Chlorella ellipsoidea is shown in Fig. 13. The content of phosphorus here implies that of total phosphorus compounds, judging from the procedure of hydrolysis of algal cells with a mixture of sulfuric acid and hydrogen peroxide, followed by the measurement of phosphate with the method of Fiske and SubbaRow 99) . Square symbols in the figure demonstrate the decrease of f P when the algal cells were put into a P-free medium, while open circles indicate clearly a trajectory of how P-starved cells recover its phosphorus pools intracellularly when the cells were transferred into a basal medium, rich in phosphorus (285 m g P l _ 1 ) . During these cultivations at 30 °C, C0 2 -enriched air (5 % v/v) was charged at a rate of 0.5 w m into a suspension stirred magnetically in a conical flask that was irradiated by a fluorescent lamp (light intensity at the flask surface = 5 klux; this irradiation was apparently of "light-limiting", because saturation level of light was around 20 klux at 30 °C). Algal cells in question, after rinse with P-free medium, were transferred to a closed and small glass vessel (5.6 ml) to measure the photosynthetic activity, Q 0 z , in P-free medium at 30 °C, 6.5 klux (from halogen lamp). Before the transfer, C0 2 -enriched

Growth Kinetics of Photosynthetic Microorganisms

121

air had been supplied enough into the P-free medium. It is interesting to remark from Fig. 13 that the decrease of f P during the process of P-starvation was accompanied by a sharp decrease of QQ2 when f P decreased below 5 |ig P m g - 1 ( = 0.5% w/w), whereas the recovery of f P in the basal medium entailed that of QQ 2 99) . This kind of vicissitude of f P values in algal cells has already been pointed out by other workers 100 ' 97 >. in this demonstration, it might be urged that f P c r i t , implying phosphorus content, below which photosynthetic activity deteriorates, was of around 0.5 %, while the maximum value of f P , f Pmax was about 2.3 % in this example (Fig. 13). These values might be subjected to change depending on algal species used, and also on the culture conditions including temperature, pH, light intensity, etc. Accordingly, it is not worthwhile to state precisely both values of f P and f P of algae. It is more significant to extract from an example of Fig. 13 that specific growth rate, p. h" 1 , of algae is correlated apparently with f P values rather than extracellular concentration, P e . mg P I - 1 , because QQ2 values in the figure should correspond to those of n (for details, see next section). It might be argued that the vicissitude of f P was "magnified" to a certain extent, because the cells used to measure the recovery of f P had been "P-starved." In another run when algal cells (Chlorella ellipsoidea) whose f P = 20 ng P m g - 1 , max

QO2 = 2.5 x 10" 2 nmol 0 2 m g - 1 m i n - 1 were transferred to basal (phosphorus-rich) medium (initial pH = 6.0), QQ2 values increased to 6.0 x 10~ 2 |xmol 0 2 m g - 1 m i n - 1 during 50 h of growth at 30 °C by consuming intracellular phosphorus, and hence f P values decreased to f P = 12 ng P m g - 1 at 50th h. During a period of 100 h that ensued, QO2 values deteriorated to about 5.0 x 10~ 2 nmol 0 2 m g - 1 m i n - 1 without any change of f P = 12 |ig P m g - 1 . This latter period corresponded to the "linear growth" of cells 9 9 ) . It might be urged that some nutrient other than phosphorus might have limited the algal growth during the period of 100 h when f P = 12 ng P mg" 1 99) . In these specific examples of Fig. 13 and the run mentioned above, the "luxury uptake" of phosphorus might be commensurable to Af P = f P — f P = (23 ~ 20) — 12 = 11 ~ 8 ng P m g - 1 ( = 1.1 ~ 0.8%), provided: f P = phosphorus content, below which some nutrient other than phosphorus limits algal growth; in this example, taken as 12 ng P m g - 1 . The "overplus", on the other hand, might correspond to Af P = / P m a x - fp cru = (23 ~ 20) - 5 = (18 ~ 15) ng P mg" 1 ( = 1.8 ~ 1.5%), albeit references have not necessarily clarified the distinction between "luxury uptake" and "overplus" of phosphorus in algal growth 9 7 ' 9 9 ) . In connection with phosphorus content in algae that play a significant role in water pollution and its control, a concept of A G P (Algal Growth Potential) is introduced. A G P of water indicates the potentiality of the water to proliferate algae that are found ubiquitously in a polluted environment and/or grow a specific green alga, Selenastrum capricornutum that withstands severe and malnutritional environment without appreciable morphological change. A G P is indicative of the progress and/or degree of eutrophication of water in terms of its phosphorus availability. A G P of effluent from the secondary treatment plants of domestic sewage in Tokyo area is shown in Table 4 101) . Periphytic algae isolated from a river running through the metropolitan area of Tokyo or from a settling basin of the treatment plant were

122

S. Aiba

Table 4. A G P of effluent f r o m Secondary Treatment Plant of municipal wastewaters 1 0 1 ' Run Sampl- Date No. ing

% % mgPI"'

Algal sp.

mgPr'

June 30. 1972

1 M

2

plant

0.98

0.08

250

0.36

0.39

1.08

0.83

0.09

264

0.28

0.38

1.37

1.42

0.09

340

0.39

0.38

1.06

0.72

0.06

213

0.31

0.47

1.28

1.22

0.04

220

0.54

0.56

July 22, Chlamvdo 0.65 monas so. 1972

0.66

0.03

72

0.88

0.86

May 24, 1972

Chlorella sp.

Oct. 11, 1972 lune 30, 1972

M plant

5

SH plant

6

May 24, 1972

Stiqeoclonlum

tenue

=

Concentration

of

T - P in effluent at the

-

Concentration

of

T - P in a mixture of

=

Concentration

of T - P in filtrate at the end of

A G P

=

Algal

' >-100 -no AGP % m g l ' ' AGP%

V

1.06

3 4

AGP

mgPI

start of algal c u l t i v a t i o n , mg P I * '

algae and effluent after the

cultivation, mgP I' 1 .

algal cultivation , m g P I " 1 ,

Growth Potential, mg I"'.

inoculated by about 0.2 m g l - 1 into a L-type glass vessel containing the effluent (about 500 ml) that had been filtered through a Millipore filter (pore size = 1.2 |im). The vessel was subjected to a see-saw motion (30 times min - 1 ) at 20 °C, 4 klux from a fluorescent lamp (14 h for light and 10 h for dark) for about 20 days. AGP shown in the table is the algal concentration in each vessel after the cultivation. Generally, the level of AGP is larger than that in oligotrophic water by one order of magnitude 101 • 102) . The last two columns in the table are equivalent to phosphorus content, f P % that appeared earlier. The column before the last pertains to the measurement of algal content of phosphorus by digesting the cells after the cultivation, whereas the last deals with a convenient procedure that merely measures phosphorus concentration in liquid before and after the cultivation. As is evident from Fig. 14, the conveninent method to assess the value of f P is considered to be plausible, if deviations from the solid line as noted from the figure, i.e. the overestimation of f P values is not too prohibitive. A comparison of phosphorus content (upper rows for Chlorella sp. in Table 4) with f P values in Fig. 13 (Chlorella ellipsoidea) may lead to a surmise that algae in the AGP test might have been "P-starved" considerably. The relationship between specific growth rate, ( x h - 1 and phosphorus content, f P mg P m g _ 1 is assumed to be hyperbolic as shown below: fP - fP . H

=

Umax

K

f r

f r

, r

m l n

provided: f Pmin. = minimum value of phosphorus content to sustain the living of cells, mgPmg"1

^

123

Growth Kinetics of Photosynthetic Microorganisms

Fig. 14. P - P e / A G P x 100 vs. P eo - P ^ / A G P x l O O (cf. Table 4). A convenient procedure to estimate phosphorus content, f P mg P g - 1 , from the determination of A G P 101>

AGP

x 100

where f ca = conversion factor from 0 2 evolved to algal cell material synthesized, g cell g O f 1 f ch = chlorophyll a content in algae, g chlorophyll a(g cell)" 1 ( = 0.0035) 113) The following stoichiometric equation of algal photosynthesis might be consulted to estimate the value of f ca , i.e. f ca = 0.645 in this case. Sunlight NH 3 + 5.7 C 0 2 + 12.5 H 2 0 C 5 7 H 9 . 8 0 2 . 3 N + 6.25 0 2 +

9.1

H

2

0

(39)

85,72,113)

128

S. Aiba

4.6 Decomposition of Periphytic Algae Decomposition of algae, a consequence of periphytic algal growth in polluted waters will be referred to briefly. There have been published not many works on decomposition of periphytic algae 7 0 , 1 1 5 _ 1 1 7 ) , and indeed, it is not feasible to discuss the decomposition in general terms, because decomposition rate is governed not only by pH and temperature of water, but also by constituents of decomposer, micro-flora and -fauna that vary from case to case. The rate of decomposition and that of release of total phosphorus (T-P) at 20 °C in dark and aerobic decomposition of periphytic algae are illustrated in a) and b) of Fig. 17, respectively 70) . These algae isolated either from a sedimentation basin of the secondary treatment plant of domestic sewage in Tokyo area or from periphyton in a polluted river were precultivated for 20 days at 20 °C in L-shaped glass vessel (500 ml in working volume) that contained a filtrate of effluent from the secondary treatment of wastewaters through a Millipore filter (pore size = 1 . 2 |im). The vessels were exposed to fluorescent light (I0 = 4.5 klux). These algal cells harvested were transferred into a glass vessel (nominal volume = 51) containing an effluent from the plant (4 1), and air was charged through a

9 3 Or

1 I

7 ;

i

3 0J,

o

l O

20

30

40

Period for decomposition , t

not— 0

1

10

1

20

1

30

1

40

Period for decomposition . t

50

100

(days)

1

50

I days )

ll

.

100

Fig. 17. Dark and aerobic decomposition of periphytic algae in an effluent from secondary treatment of wastewater (20 °C). a Decomposition, b Regeneration of phosphorous 7 0 )

Growth Kinetics of Photosynthetic Microorganisms

129

sintered glass plate, shielding the vessel from ambient light. It is noted from Fig. 17 a) that the rapid rate of decrease in cell mass concentration was followed by its deceleration and finally, by levelling-off. Regeneration of phosphorus (b) in the figure) continued regardless of the levelling-off of the decomposition in a). Nitrogenous compounds were also regenerated (not shown in Fig. 17) in parallel with the regeneration of phosphorus as shown in the figure. Apparently, these regenerations cause the secondary loading in water. For detailed discussion on the secondary loading see the Ref. 7 0 , 1 1 5 ) . It is clear from Fig. 17 a) that there exists "refractory" portion that remains resistant to the decomposition even after 50 to 100 days of exposure to the decomposer. The fraction, f, is defined here as the ratio of refractory components to cell concentration, XQ at t = 0. Then, fXo refers to the concentration of refractory materials. If (X — fXo) is plotted semi-logarithmically against t using data shown in Fig. 17 a), linear relationships could be assumed, respectively for t = 0 to t = 20 days (not shown here). This implies that decomposition rate for an early period could be represented by the 1st order reaction, i.e. dX — = - k ( X - fXo)

(40)

where k = decomposition rate constant, d " 1 XQ — algal cell concentration, nig l" 1 at t = 0 f = refractory fraction of algal cells The refractory fraction, f, in this example of Fig. 17 a) extended nearly from 0.20 (Chlamydomonas sp.) to 0.50 (Chlorella sp.) at 20 °C. The value o f f reported by Foree who used various media including effluent from activated sludge plant, seawater and freshwater at 20 °C in the decomposition of green algae and diatoms ranged from 6 to 100% in dark and aerobic decomposition tests 117). Clearly, f value depends on ecological and decomposer species besides environmental conditions of whether aerobic or anaerobic at a specified temperature, k values in Eq. (40) assessed from data in Fig. 17 a) were 70) : k = 0.1 d " 1 for Chlorella sp., k = 0.08 d _ 1 for Stigeoclonium tentie, and k = 0.2 d _ 1 for Chlamydomonas sp. at 20 °C.

According to Jewell and McCarty, the value of k under dark and aerobic conditions ranged from 0.01 to 0.06 d _ 1 (at 20 °C) when mixed cultures of algae, bacteria, and zooplankton were subjected to the decomposition (cf. Eq. (40)). Axenic algal cultures, when seeded with decomposers, followed the 2nd order kinetics during initial stages of decomposition, but exhibited later on the 1st order reaction 116). Further, Jewell studied dark and aerobic decompositions of aquatic weeds. He showed that the decomposition rate also followed the 1st order kinetics, k values in Eq. (40) being from 0.05 to 0.19 d " 1 at 18 °C in deionized water; when domestic wastewater was used, these values were doubled 115).

S. Aiba

130

Table 6. Micro-flora and -fauna of periphyton community (outdoor artificial stream, July 25 to September 2, 1975) 118) Sampling

Time (days)

1

4

S,M

S,M

8

S

A,S,M

11

2

S,F,Sp S,M,0

13

S,F

S,P,F F,S

17

P.F

20

P, F

25

Sp,F

28

location 3

4 E,Av

S,M

Av

F.RAv

Av,M

F.RAw Av,M,E Av,E P,M,Av

E

Sp

Sp,F

Av,E

Sp

Sp

Sp

Sp

32

Sp

34

Sp

Sp

Av

39

Sp

Sp

Av

Sp

A

= Achnantes

Av = Arcella

lanceolala vulgaris

E

= Euqlypha

spp.

F

= Fragilaria

constatum

M

= Melosira

O

= Oscillatoria

P

= Pinnularia

S

= Synedra

Sp = Spirogyra

(protozoa)

varians

(diatom) (diatom)

spp.

( blue-green a l g a )

gibba ulna

(diatom)

(protozoa)

(diatom) (diatom)

distenta

(green alga)

In spite of inextricable pictures that should have underlain the decomposition of algae in various media, i.e. effluent from secondary treatment plant, seawater, deionized water, etc., whose spectra of decomposers extending perhaps widely, it is interesting to confirm from these publications that k values were of the order of 10~2 to 10" 1 d " 1 so far as the 1st order of decomposition kinetics was applicable 1 1 5 , 1 1 6 , 7 0 ' It is conceivable that periphytic algae in polluted and running waters would experience succession of its genera and/or species during long runs of proliferation and/or decomposition. It is also envisaged that once the specific layer of periphytic algal population reaches a critical thickness, these algae cannot tolerate shear stress of running waters and they are detached from their habitat. This phenomenon must also be appreciated to analyze the secondary loading in natural environment. No systematic work has yet been done along this line. A spectrum of periphytic algae and protozoa that was observed in an artificial stream of effluent from a plant for the secondary treatment of domestic sewage is shown in Table 6 118). The artificial stream that was constructed near the plant in Tokyo area consisted of a number of troughs (60 cm long, 24 cm deep, 24 cm wide and 9 cm thick) fixed on a steel frame rack. The total length of this stream was about 90 m. Effluent from the plant was charged appropriately into this channel to realize an artificial stream, whose linear velocity was controlled not only by the amount of effluent discharged, but also by placing a trapezoidal block of acrylic resin onto particular troughs, thus elevating the bottom of these portions. Thus far, linear velocity of effluent at Locations 1 to 4 in Table 6 was 16.3, 45.0,14.1 and 17.1 cm s" 1 , respectively. The depth of water at Locations 1 to 4 was: 9.0, 8.0, 18.5 and 16.0 cm, respectively. Location 4 was covered with plates, thus shielding ambient light (dark). The spectrum of micro-flora and -fauna was observed during about 40 days (water temperature: 24~26 °C, pH: 6.9~7.2) by removing square slabs intermittently one by one from each location; the unglazed slabs (total number = 80; each surface area = 90.25 cm2) were fixed, in juxtaposition, initially at the channel-bed. Peri-

Growth Kinetics of Photosynthetic Microorganisms

131

phytic microorganisms grown on the slab surface were scraped off with a toothbrush. It is noted from Table 6 that in the dark area protozoa were dominant throughout, whereas in other areas a succession of periphytic algae from diatoms to green algae took place at around 20th day after the start of this run. Although data on biomass of periphyton, mg c m - 2 and chlorophyll a |ig cm" 2 are not shown here, both items increased gradually, exhibiting peaks at around 13th to 15th day. Minimal concentrations of biomass were observed at about 18th day 1 1 8 '. Right after this detachment, the succession of algal species might have taken place, and another peak of algal population density was noted at 33rd to 34th day of operation 118>. Mclntire also studied this sort of periphytic algal spectrum in his laboratory stream, and claimed that some of diatoms he observed were insensitive to linear velocity of stream, while Nitzschia linearis, Achnantes lanceolata, etc. were more abundant in a region where the current velocity was relatively large. Melosira varians was abundant in still waters 119). However, in Table 6 no such distinctive picture on the algal habitat was recognized.

5 Algal Growth vs. Environmental Control A diurnal and conspicuous fluctuation of DO in a polluted and shallow river as shown earlier in Fig. 8 is apparently caused by growth of periphytic algae. The algal growth should have reduced nutrient levels in the water, especially on phosphorus and nitrogenous compounds. However, decay or decomposition that accompanies inevitably the algal growth and in addition, detachment of these algae downstream due to flood that could seasonably happen should cause the secondary loading. The secondary loading is far-reaching to an extent of another cause of red tide in seawater. Once the red tide should occur, reflecting dominance of specific species of marine planktons, coastal fisheries as well as fish farms are heavily damaged because of death of edible fishes. In fact, the consequence of algal growth due to enrichment in running waters is serious enough to deserve grave concern. Another example of eutrophication in terrestrial waters, i.e. lakes and ponds may be cited. A representative and ominous symptom of eutrophication in these waters is in the emergence of water-bloom. Water-blooms that have been recognized as early as in the 15th century in Switzerland 50 ' are attributed to dominant growth of bluegreen algae such as Microcystis aeruginosa, Anabaena flos-aquae, etc. at the water surface. Whenever water-bloom flourishes especially in summer seasons, not only scenic beauty is impaired, but also an offensive odor coming from the bloom requires huge capital investment to strip the water of this odor, particularly if the water is to be processed for drinking. Needless to say, when a massive bloom decomposes and settles onto benthic sediment, the secondary loading originating from the benthic mud further aggravates water quality. It is actually constituting a vicious cycle of damaging and threatening the environmental beauty and indigenous life. There have been published so many works on simulation of the aforementioned marked DO variation in running waters 8 6 ' 1 2 0 ) . Although these works incorporated into their differential equations proper terms to represent activities of photosynthetic

132

S. Aiba

microorganisms in order to simulate the pattern concerned, no attention has been paid on microbial physiology and on the after-effect of decomposition of algae on the deterioration of water quality. No simulations have been presented yet on appearance and/or disappearance of water-bloom 50) at the surface of still waters, albeit gas vacuoles in these blue-green algae that should trigger the algal ascent and/or descent in water columns depending on photosynthetic activities of these algae have been studied extensively by Walsby and his associates 5 0 , 1 2 1 ) . A solution to the problem of how to maintain environmental serenity in populated areas is simple in one sense. The answer is to minimize the primary production of biomass, composed principally of algae. However, this solution must be discussed in concrete terms. The process of how to put the answer into practice is teemed with various difficulties and/or unknowns. Herein lies the significance of simulation in the sense of environmental assessment. Not a few works on simulation of rivers, estuaries, lakes, etc. have been presented from various circles. For example, Schofield and Krutchkoff published a deterministic model of dynamic and eutrophic estuary, in which chemical and biological systems are shown by 12 simultaneous differential equations including 40 unknown parameters 122) . They solved numerically these equations, parameters having been estimated by the least-squares method. As a particular model becomes sophisticated, the number of unknown parameters increases necessarily. A compromise should be made between minimizing the number of unknown parameters and maximizing the reliability of simulation from the viewpoint of practice. In another word, the compromise might suggest that environmental assessment be made by reconstruction of unit processes comprising a given problem, each process — whatever biological, physical or chemical it might be — being pursued either on the spot or in laboratories to minimize the number of unknown parameters. If a model to simulate reliably an environmental problem becomes available, one could foresee the effect of changing the value of any control variable on the given objective. For instance, one could assess the effect of increased flow rate in a stream on periphytic algae and its secondary loading. According to Oglesby and Edmondson 23) , addition of nutrient-poor city water, ranging from about 380 to 22,710 m 3 d " 1 periodically all year round (2 periods of several months, each, of pause) into Green Lake, Seattle has been confirmed to rehabilitate the eutrophic lake (normal water volume = 4 x 106 m3). However, this method of a partial control of the lake's eutrophication has not necessarily been established from a rationale of simulation. The principal aim of this chapter is to demonstrate simulation of a shallow and polluted river (Tama-gawa) running through the metropolitan area of Tokyo. Assuming that P0 4 -P is a key nutrient to destine the progress of enrichment, 2 3 ' 1 2 3 ) the fate of P0 4 -P in the middle region extending from 18th to 40th km from the estuary will be simulated. In this simulation, not only algal phenomena, but also physical as well as hydraulic characteristics of the river must be taken into account. Likewise, BOD and DO profiles along the river will be simulated. Since the procedure of solving numerically partial differential equations and in addition, that of estimating parameters involved in these equations (see later) are published elsewhere 85 ' 112 ' 124 ', these descriptions will be minimized in this paper. Only the salient point to have made an approach to the problem of environment, in

Growth Kinetics of Photosynthetic Microorganisms

133

which the growth of photosynthetic microorganisms should have played a significant role, will be made clear. Indeed, there has been found no panacea to environmental control. Although it looks like naive (see next), the approach, together with accumulation of information to be accrued therefrom, would only be eligible for yielding the most effective means that can be applied practically to the control of environment. Simulation of water-bloom and its control will be presented separately.

5.1 Construction of one-dimensional Model Figure 18 shows a map of Tama-gawa, a polluted river flowing down through heavily populated areas of Tokyo. Locations of field stations along the river are also shown in the figure. Several tributaries come into the main stream which is characterized by its shallowness (average depth of water: less than 100 cm). Accordingly, models to describe profiles of P0 4 -P, DO and BOD concentrations are assumed to be onedimensional as shown below.

^ r = I (ADl I ) - IT (AUP) + M P A " k*PWA - n M f p S - k 7 PS + P T - P s

9ac 0 2

0 /

= -

0c O2 \

0AB

0 /

= -

a

I A D l - ^ J - — (AUC 02 ) - k,BA + k 2 (CS 2 - Co2) A

- r„MS +

—

(41)

0B\

, u.Ib - j MS + (CG )T - (C 02 ) s a + b I b + CIB

(42)

8

( A D l — J - QHAUB) - k t B A - k 3 B A + k 4 M S + B T - BS (43)

For symbols, see the list of Nomenclature. A brief explanation of Eqs. (41) to (43) is as follows: The left-hand side of each equation represents the rate of change in either phosphorus, DO or BOD concentration with time per unit length of the river, g P m " 1 d " 1 for instance. The 1 st and 2nd terms on the right-hand side of these equations denote input due to dispersion (backmixing) and output owing to current convection, respectively. The term before last and the last terms in Eqs. (41) to (43) are the (net) loading from tributaries and seepage due to groundwater (disappearance and/or emergence), respectively. The 3rd and 4th terms on the right-hand side of Eq. (41) pertain to phosphorus emergence due to hydrolysis of condensed phosphates and adsorption (disappearance) of phosphorus onto suspended solid in the river water, whereas the 5th and 6th terms deal with phosphorus disappearance from the water because of algal uptake of phosphorus and the settling of phosphorus onto the river bottom (see Fig. 19). It is feasible to discern the terms relevant to decrease in DO due to bio-oxidation of BOD, increase of DO by surface reaeration, DO decrease and increase because of algal

134

S. Aiba

Densely

populated

Water

reservoir

TOKYO

bay

TOKYO International Observation

area

Airport

station

Boundaries of Metropolitan (TOKYO) area

Fig. 18. Locations of Tamagawa and field stations U 3 )

respiration and photosynthetic activities, respectively (cf. Eq. (29), provided: I in Eq. (29) is replaced by light intensity, I B at the river bottom.). Although samplings of water and its chemical analyses were made at regular time-intervals (Aug., 1972) in each station (see Fig. 18) from St. 1 to St. 9, the simulation which appears later on handled a region extending from St. 3 to St. 7 (distance: about 22 km). Especially, water at St. 3 was sampled and analyzed every hour for a whole day of Aug. 18, 1972 to secure data of boundary condition. Items examined on water were: temp., pH, SS (suspended solid, W g m~ 3 ), transparency (m), BOD (B, g BOD irT 3 ), DO (CQ2, g 0 2 m" 3 ), P 0 4 - P (P, g P m" 3 ). NH 3 -N, N 0 3 - N and N 0 2 - N were also measured, but those data will not appear in the simulation here. Each item was measured by the Standard Method125). In order to estimate the amount of water, Q s m 3 d - 1 , that disappears and/or emerges at the river bottom, water balance in the region from St. 3 to St. 9 was attempted in Aug., 1972. Data on daily flow rate of water in the main stream at each station and those of tributaries were provided by Research Institute of Environment, Tokyo Metropolitan Government 8 5 ' 1 2 4 > 1 1 2 ) . Q s values thus obtained were used to estimate the last terms on the right-hand side in Eqs. (41) to (43). P s , for instance, was equated to PQS/Ax, where P is phosphorus concentration in the main stream for a particular region (Ax), where the value of Q s was predetermined. Data on hydraulics of Tama-gawa that were used in the simulation came also from the Research Institute. Hydraulics in Aug., 1972 was: Q = 100 x 104 m 3 d _ 1 H = 0.4~0.6m § = 50-80 m U = (1.3 — 2.1) x 103 m h _ 1 .

135

Growth Kinetics of Photosynthetic Microorganisms

Inflow

/// / Current

}/

from

tributaries

Outflow

to

tributaries

Water Surface

Hydrolysis of condensed phosphate (dissolved)

/ / / / / / /

Adsorption

by

J///// Current

suspended solid

P04-P

Dispersion

Uptake by Benthic Organisms (Algae)

iL

Adsorption ( R e lease) to (from) River-bed (dead) Organisms (Algae)

River- bed

Dispersion

4

Seepage

Fig. 19. Schematic representation of unit processes responsible for P0 4 -P balance in a shallow and polluted river U 3 ) . Regarding "sink" of P0 4 -P on the river-bed, "adsorption" was used in this figure instead of the use of "settling" in text. N o concrete difference between "adsorption" and "settling" was conceivable in this context

The observation of algal population density, M(x) (g chlorophyll a) m " 2 in Aug., 1972 revealed the dominance of Stigeoclonium tenue (green alga) and Synedra ulna (diatom) n 2 ' 124'85>. Specific rate of 0 2 evolution and specific rate of respiration of algae, r p g 0 2 (g chlorophyll a ) - 1 h " 1 were measured by the light- and darkbottle procedure 112) . This procedure was extended to estimate from a stoichiometric equation (cf. Eq. (39)) the value of specific growth rate of periphytic algae as elaborated from Eqs. (35) to (38) in previous chapter. In addition, the light- and dark-bottle procedure permitted an assessment of coefficients, a', b' and c' in Eq. (42) as was mentioned earlier (cf. Fig. 16).

5.2 Estimation of Parameters Not a few coefficients and parameters are involved in Eqs. (41) to (43). It is necessary to estimate directly on the spot or separately in laboratories these parameters or coefficients. The following is a rough sketch of the procedure used for estimation of some of the coefficients and parameters. i) Dispersion coefficient, D L m 2 h - 1 appearing throughout Eqs. (41) to (43) was assessed by a formula presented by Thackston and Krenkel who examined various formulae on D L ever published by other workers 126) . Although Thackston and Krenkel also surveyed widely reaeration rate constant, k 2 h - 1 , in Eq. (42) 127) , the value depended most probably on hydraulics of river. Consequently, k 2 values as "unknowns" were estimated from another approach (see next). ii) The rate coefficient of bio-oxidation, kj, that appears both in Eqs. (42) and (43)

136

S. Aiba

was estimated from a formula by Krenkel et al. 128) , taking temperature effect into account 85 • 112) . k 5 that concerns with hydrolysis of condensed phosphates in water was assessed referring to the work by Shannon and Lee 129) . iii) k4 in Eq. (43) could be assessed from daily growth of periphytic algae per mg chlorophyll a, multiplied by a factor, f BOD , where f B O D is a conversion factor from algal cells to BOD 85) . The daily growth of algae was estimated from 0 2 evolution rate (cf. Eq. (38)). However, since the secondary loading of BOD was assumed to be equal to that amount caused by the daily growth of algae and in addition, because of another assumption of steady state in P, C 0 2 and B on daily basis in the simulation later on, a separate assessment of k4 in Eq. (43) was not required. iv) a in Eq. (41) was evaluated experimentally by measuring total phosphate of water (Tama-gawa) that was filtered beforehand through a Millipore filter (pore size = 1.2 (im) 113) . Namely, the difference between total phosphate and phosphate in water divided by the latter value was equated to a. Both values of kg and f P in the equation were determined experimentally. For the former experiment, water sampled from the river was used to observe the decrease of P due to adsorption of phosphate to SS during more than one week at room temperatures 113) . v) Values of parameters that were assessed and used in the simulation are summarized below: a' = 8.95 x 102, b' = - 4 . 1 6 x 1(T 2 , c' = 2.05 x 10" 6 r p = 1.76 g 0 2 (g chlorophyll a)" 1 h " 1 dominant alga: Synedra ulna (loc. cit.) a' = 7.33x 103, b' = —6.94x 10" 1 , c' = 3.40x 1(T 5 r p = 0.87 g 0 2 (g chlorophyll a)" 1 h " 1 dominant alga: Stigeoclonium tenue . v-i) M(x) = 0.30 g chlorophyll a m~ 2 for Synedra ulna from St. 3 to St. 9, Aug., 1972, whereas M(x) = 0.50 g chlorophyll a m " 2 at S. 3, M(x) = 0.40 g chlorophyll a m " 2 at St. 6 and M(x) = 0.0 at Stns. 5 and 7, respectively for Stigeoclonium tenue; the latter distribution was expressed by polygonal lines 1 1 3 , U 2 ) . The fractional area of river bottom occupied by Stigeoclonium tenue was estimated as 50% at Stns. 3 and 6, and 0% at Stns. 5 and 7 (Aug., 1972), while the rest was assumed to be the occupation by Synedra ulna. The distribution of this fractional area was also approximated by polygonal lines 1 1 3 , 1 1 2 ) . v-ii)

a f BOD fp kt k5 kg

= = = = = =

0.125, f ch = 0.0035 g chlorophyll a (g cell)" 1 , 0.41 g BOD (g cell)" 1 , 13°.131> 2.4 g P (g chlorophyll a ) ' 1 = 0.0084 g P (g cell)" 1 , 0.012-0.014 h " 1 , 1.43 x 10" 2 h " 1 , 6.02xlO"7m3(gSS)-1h-1 .

Growth Kinetics of Photosynthetic Microorganisms

5.3 Eestimation of Parameters (cont'd)

137 85,124)

.

Evaluation of k 3 and k 7 , both of which are relevant to the decrease of BOD and P due to settling, respectively are apparently beyond the scope of experimental determination. k 2 in Eq. (42) was also treated in this category (loc. cit.). In each case, steady state of either P, Cq 2 or B (phosphorus, D O or BOD concentration) in Eqs. (41) to (43) was assumed, respectively on diurnal basis, i.e. the left-hand side of each equation was equated to zero to facilitate estimation of these unknown parameters. The assumption of steady state of each nutrient concentration on daily basis has already been applied to an assessment of the secondary loading of BOD (the 3rd term from right on the right-hand side of Eq. (43)), and actually, as far as the simulation in this work pertains to hourly rather than weekly or monthly, the assumption per se does not present any serious problem in this simulation. The entire region of simulation from St. 3 to St. 7 (distance = 22 km) was subdivided into the following three sections for daily balance of nutrient to permit eventually an assessment of each unknown parameter. Section

St. to St.

I II III

Distance

x (km) 5.5 6.5 10.0

3-5 5-6 6-7

Since balance equations with respect to nutrients are exactly similar, only that of phosphorus will be shown next.

P

P

B

0 = A B D LB

A + PB

j —x 2

2

PA+PB

x 24 - A a D l a

(diffusion: P J

5

TV ~ A J —x 2

x 24

(diffusion: P 2 )

- A b U b P b x 24 + A a U a P a x 24 (outflow: P 3 )

(inflow: P 4 )

^ , Pa + PB AA + AB _ . . , P A + P B WA + WB A a + A „ _ + k5a x x 24 - k 6 x x 24 2 2 2 2 2 (hydrolysis: P 5 )

(adsorption to SS: P 6 )

- - SA -I- S B _ P A P B S A -I- SB -uMfpp x x 24 - k 7 x 2 2 2 (algal uptake: P 7 )

(settling: P 8 )

138

S. Aiba

+ Q]PT

-

Pa+PB Q o ^ r ^

-

(inflow from (outflow to tributaries: P 9 ) tributaries: P 1 0 ).

^ PA^PB Q s ^ r ^ 2

(seepage: P n )

It is easy to establish Eq. (44) from Eq. (41) when an average quantity of variable is taken with respect to upstream and downstream ends in each section; note that P T in Eq. (41) was divided into inflow and outflow regarding tributaries as shown in Eq. (44), provided: PT = av. phosphorus concentration in tributaries flowing into this section of the main stream, and subscripts, A and B are for upstream and downstream boundaries of the section, whereas short-bar superscript implies daily average, unless otherwise noted. From Eq. (44), P 8 = Pj - P 2 - P3 + P 4 + P 5 - P 6 ~ P 7 + P 9 - Pio - Pn •

(45)

k7 = — Pa

(46)

Then

+

_ . ' . . PBVSA+Sb^

x 2 4

Likewise, -D7 k

" "

k3 =

c

, _ ( g o >

'B A + BB\

,

2

+ ^ > . ) ( ^ )

i x

24

(47>

(48)

AA + A B \ _

1 x x 24

Since each term on the right-hand side of Eq. (45) could be assessed from field data (Aug., 1972), daily amount of phosphorus settled onto the river bottom could be determined, followed by estimation of k7 as shown in Eq. (46). Similarly, numerators in Eqs. (47) and (48) could be assessed from another balance of DO and BOD on daily basis (cf. Eqs. (42) and (43), respectively) to allow the assessment of unknown parameters, k 2 and k 3 . The result of these estimations was 130 ' 124-8S>. Section

k2 (h- 1 )

k3 (h" 1 )

k7 (mh"1)

I II III

0.19 0.43 0.29

0.080 0.092 0.028

0.035 0.045 0.011

G r o w t h Kinetics of Photosynthetic Microorganisms

139

Table 7. P 0 4 - P balance (kg d " 1 ) 1 1 3 )

\ p Sectiwi

I I

M

\

P,

P,

P,

P.

P5

P6

p,

P,

P,

1

1

368

150

2

0.1

25

65

262

23

-67

-1

1

268

368

4

0.1

32

82

185

32

143

1

-1

295

268

3

0.1

49

24

107

32

-22

P,. P„

unit : kg day' 1

Average values of linear velocity of water, U ( = 0.47 m s " 1 ) , water depth, H ( = 0.5 m) (loc. cit.), and average slope of the river-bed throughout Sections I to III ( = 2.7 x 10" 3 m/m; data from Tokyo Metropolitan Government) were substituted into an empirical formula on k2 (Thackston and Krenkel) 127) . The result was: k 2 = 0.15 h - 1 , i.e. k 2 values estimated from Eq. (47) were nearly two ~ threefold of that value calculated from the formula. This inconsistency might be partly due to the difference in scale between the shallow river here and deep streams on the Continent (U.S.A.). Table 7 illustrates the daily balance of phosphorus that was needed to estimate the value of k7 as shown above. The fact that Eq. (45) does not exactly hold in each row of the table comes from the procedure that counted fractions over 0.5 as unity, disregarding the rest. An overall picture in daily phosphorus balance for the region from Stns. 3 to 7 was constructed from the data in Table 7 as shown in Fig. 20. It is remarked from Fig. 20 that contributions of diffusion, adsorption of phosphorus to SS and hydrolysis of condensed phosphates to an overall balance of phosphorus are negligible and/or relatively small, whereas those of algal uptake and settling are con-

unit

^ k g day"'

Fig. 20. Material balance of P 0 4 - P between Stns. 3 and 7

U3)

140

S. Aiba

siderably large. In this example, daily input from tributaries is also remarkable. Though not shown here, the same pattern was obtained in the balance of BOD and DO, respectively 85) . In both cases, the effect of diffusion on the balance was also negligible. DO increase due to photosynthesis by periphytic algae was almost of the same order of magnitude as that originating from reaeration, while DO decrease due to bio-oxidation was fairly small, the absolute value being about one-tenth of that coming from photosynthetic activities of algae 85) . Regarding BOD balance, the disappearance due to settling, emergence of BOD because of the secondary loading and inflow from tributaries were nearly of the same order of magnitude; the decrease due to bio-oxidation was also nearly of one-tenth of that value in the preceding items 85) .

5.4 Simulation 85

124

*

Equations (41) and (43) are solved numerically by a difference model, respectively. For simplicity, the formulation of this difference model is omitted here; the 1st point of the numerical computation used is in that the value of L ; k , where L is either P, CQ2 or B, and i, k are segment number of having divided distance and time for the calculation, is equated to L i + 1 k + 1 appearing once due to diffusion in the difference formula to solve for L; k + 1 from the value of L; k . This convenient procedure reduced considerably the time of computation compared to the multi-step method of Bella and Dobbins 1321 who repeated computations with respect to diffusion, convection and biochemical reaction terms in Eq. (41) for example, consecutively 85) . The procedure used here was also superior to the conventional Crank-Nicolson (Lassonen) method 132) . The 2nd point is in that both values of D L and U were assumed to be dependent only on location, i.e., time-independent. The region from St. 3 to St. 7 (see Fig. 18) (distance = 22 km) was subdivided into segments by the following procedure to avoid pseudodispersion. In this segmentation, inflows from tributaries must be taken into account. ( A ^ U . - i + Qh) At = AjAx; = ( A A + QoO At.

(49)

In fact, the following values of At and Ax; were taken: At = 0.5 h Axj = 200 ~ 1,300 m i = 0 ~ 34 . This segmentation must satisfy that G values ( = D, At/Axf) g 1/2 to warrant the stability and convergence of numerical solution. Field data taken at St. 3 every hour throughout the day of Aug. 18, 1972 (loc. cit.) were used as boundary condition, while those at 0:00 h on the same day atStns. 3, 5, 6 and 7 were employed as initial conditions, i.e.

141

Growth Kinetics of Photosynthetic Microorganisms t = ( k + l ) At

L(0,t) dt

(50)

L.,o = L(i- 0) . Irrespective of BOD, DO and P0 4 -P, good agreement was confirmed between computation (HITAC 8700/8800, Computer Center, Univ. of Tokyo) and observation. For details, see Ref. 8 5 • 124) .

5.5 Strategy Once computation program to simulate the concentration profile of nutrient in a shallow and polluted river is made available, it is feasible to predict the pattern of nutrient concentration downstream if any one of the control variables were changed from the current level to another. This approach to the environment is defined tentatively as "strategy" for the control of environment. P0 4 -P profile in Tama-gawa is shown as an example of this discussion on "strategy". Solid line in Fig. 21 simulates the profile of P0 4 -P concentration observed at 12.00 noon on Aug. 18,1972. Irregular peaks observed (and/or simulated) around 39th and 23rd km from the estuary correspond obviously to the effect of PO4-P inflows from tributaries (cf. Fig. 18). If the region were shielded from sunlight and the river-bed were free from periphytic algae, the pattern is shown above the solid line, implying the progress of enrichment because of the absence of phosphorus uptake by these algae. However, the quality of water in terms of BOD and DO could be improved (BOD decrease and increase in DO compared to the current level), since the secondary loading from the decomposition and/or detachment of periphytic algae was absent (BOD and DO profiles are not shown here). For details, the references may be consulted 8 5 , 1 2 4 , 1 3 0 ) . If wetted periphery and water depth of the river were enlarged ( x 2) and reduced ( x 1/2), respectively without changing flow rate of the river, the concentration of

Distance (St.3)

4ao

(St.5) 34.5

from

I

(St.6) 28.0

Estuary

(km) (St. 7) 18.0

Fig. 21. PO4.-P concentration profiles predicted (chronological time = 12:00). , actual data (Aug. 21, 1972), ; absence of algae on river-bed, , wetted periphery enlarged ( x 2 ) , and water depth reduced ( x 1/2) (wetted periphery and water depth in Aug., 1972 were taken as unity, respectively). , absence of PO4-P inflow from tributaries 8 5 ' 1 1 3 ' 1 2 4 )

142

S. Aiba

P0 4 -P is reduced next to the actual data (Aug. 18, 1972; see solid line). It is surmised that the reduction of water depth enhances periphytic algal growth, thus resulting in the reduction of P0 4 -P concentration in water due to algal uptake of phosphorus. Contrary to the previous condition assigned, the quality of water becomes deteriorated in the sense that concentration of BOD is enhanced, whereas that of DO decreases. This is apparently attributable to the increase of secondary loading from periphytic algae, whose growth is accelerated by the reduction of water depth 85.124,130) It was pointed out previously (cf. Fig. 20) that inflows of phosphorus (and BOD) from tributaries are serious in this particular example. The profile of P0 4 -P extremely improved (decreased in concentration) in Fig. 21 was the simulation with an assumption that all tributaries be cut off. The improvement of water quality not only in P0 4 -P, but also in BOD (reduction) and DO (rather constant; 0.5 m g l - 1 throughout) was expected from the absence of tributaries 8 5 ' 1 2 4 ' 1 3 0 ) . Although not shown in Fig. 21, the increase of water quantity flowing down the river was effective to decrease the nutrient concentration, just in between solid line (actual data) and broken line (simulation when water depth were reduced ( x 1/2) and wetted periphery increased ( x 2)), if flow rate of water in the main stream in Aug., 1972 were doubled. Improvement of water quality (BOD and DO) ensued 85 ' 124,130) ¡ s ^ e effect of "dilution". Thus far, it is interesting and of significance to discuss ways and means to improve the water quality in a shallow and polluted river. This approach could obviously be applicable to the control of environment other than this specific example, if an environmental problem concerned were disintegrated into unit processes, whatever biological, chemical or physical they might be, and if rate equations, i.e. quantitative expression of phenomena involved in those unit processes were integrated conversely to synthesize the problem. The simulation of water pollution discussed so far in this paper might have demonstrated an example of the above-mentioned procedure of "analysis into unit processes" and "synthesis into the problem", particularly centering around a mutual relationship between algal growth and one of its nutrients, phosphorus.

6 Summary 1) Representative means to measure light energy absorbed by photosynthetic microorganisms suspended in liquid medium were discussed, i.e. working principles of opalescent plate method and integrating sphere photometer, both of which have been used by many workers, were presented. In addition, chemical actinometry using Reinecke's salt solution, as applied to the measurement of light absorption by photosynthetic microbes, and Monte Carlo method to assess statistically light energy absorbed by the cells were elaborated. The latter two methods have hardly been used as far as the assessment of growth yield is concerned. 2) The use of Lambert-Beer's law in a bio-photoreactor overestimates lightenergy absorption especially in a concentrate suspension in comparison with the esti-

Growth Kinetics of Photosynthetic Microorganisms

143

mate from Monte Carlo method that does not overlook light scattered by the cells. 3) Values of Y kJ (defined by amount (g) of dried cells harvested per unit energy of light (kJ) absorbed by the cells) for autotrophic growth of green and blue-green algae were of (3~5) x 1 0 ' 3 g k J - 1 . These values were smaller than those of heterotrophs (either aerobic or anaerobic, in general) nearly by one order of magnitude. 4) Values of Y (efficiency of light energy converted to chemical energy as cell material) for blue-green and green algae in autotrophic growth were of (6 ~ 11)%; in fact, quantum requirements of those cells ranged most probably from 18 to 25 (much larger than 8 that has been maintained theoretically by previous workers). 5) Y kJ values of Chlorella vulgaris in mixotrophic growth (organic carbon: glucose) were enhanced 3 to 4 times of those in autotrophic growth. 6) Y kJ values of Rhodopseudomonas spheroides S (light-anaerobic culture using acetate as hydrogen donor) were in between Y kJ values of 3) and 5) mentioned above. 7) By analogy to heterotrophs, specific rate of light absorption, was correlated with specific growth rate, by : t = — + ^ (Y kJ ) G

m .

(17)

It was difficult to assume the linear relationship between and |i (due to unknown reasons) in some cases of autotrophic and/or in photo-organotrophic growth of photosynthetic microorganisms studied in this work. 8) If the difference of genus, species as well as culture conditions, either auto-, photo-organo- or mixo-trophic, were disregarded, (Y kJ ) G and m values were: ( Y k J ) G = 1 0 " 2 ~ 10~3 g k J - 1 m = l O ^ - l O k J g " 1 h"1 (cf. Eqs. (18) and (25)). 9) A convenient means to assess |i value of periphytic algae in lotic environment was demonstrated. In this assessment the algal growth was assumed to be limited by light intensity as shown in Eq. (29). 10) As far as eutrophication (and/or enrichment) of terrestrial waters is concerned, phosphorus (P0 4 -P) was most likely to limit algal growth. Hence, a possibility exists to retard eutrophication (and/or enrichment) by controlling phosphorus concentration in the water. 11) The concept of A G P was useful not only for measuring potentiality of water to proliferate algae in water, but also to estimate phosphorus content in algae (Fig. 14). 12) If |i values of algae were limited solely by phosphorus, n was shown as a function of intracellular content, f P , rather than extracellular concentration, P e ., of phosphorus (Eqs. (31) and (32)). 13)' There are found two extreme values of frP , i.e. f Pmax and f Pmm. . The value of fpmax could be observed when phosphorus-starved cells are transferred to phosphorusrich medium, whereas f P . was the minimum value, below which algal cells could not

144

S. Aiba

sustain its living. In addition, there has been confirmed a critical value of f P , below which photosynthetic activity of algae is apparently limited by intracellular supply of phosphorus, provided the cells are exposed to phosphorus-free medium. A bare orderestimation of f Pm a y and f Pm i.n values was: 1

4

f Pmin = 0 . 2 - 0 . 3 % .

(Fig. 13)

14) Although the relationship between specific rate of phosphorus uptake, Q P and remains to be revealed, data on Q P obtained here suggested apparently two systems of "carrier-mediated" transport of phosphate in algal cells. Saturation constant, K P , in a Monod type of equation was: K P = 1 0 _ 4 ~ 10~5 mg P m l - 1 . Most probably, this range of K P corresponds to that of permeability of phosphate, P a , in Chlorella ellipsoidea, i.e. P a = 10~ 3 ~ 10" 4 ml mg" 1 m i n - 1 ( = 1 0 " 7 ~ 1 0 - 8 cm s" 1 ). Ample room is left open for further studies on phosphate transport that would give a molecular basis of growth kinetics. 15) Decomposition rate of periphytic algae and its detachment were observed. In an early period of decomposition, the rate was expressed approximately by the 1st order reaction, *if refractory fraction of cells was deleted. Succession of algal species was observed after the algae were detached (cf. Table 6). Obviously, decomposition and/or detachment of periphytic algae in lotic environment causes the secondary loading. 16) Nutrient balance of BOD, DO and phosphorus on daily basis in a shallow and polluted river revealed a significant role of periphytic algae in aggravating water quality through the secondary loading. 17) Concentration of BOD, DO and phosphorus in the river could be simulated by one-dimensional model, in which items relevant to algal growth and decay were incorporated. From this simulation, a "strategy" of how to improve water quality by controlling algal growth was extracted. For instance, dilution of water by increasing flow rate or cutoff of tributaries into the main stream were considerably effective for the improvement. 18) This paper as an example of "numerical physiology" on photosynthetic microorganisms is considered to have exhibited a new facet of Biochemical Engineering.

7 Appendix Monte Carlo method to assess light-energy distribution in a bio-photoreactor 57) . It is well known that Lambert-Beer's law is used to assess light intensity distribution in a photosystem, wherein scattering of light can be disregarded 95) . However, when the scattering of light cannot be overlooked (cf. Fig. 6), the use of this law would result in a wrong interpretation on growth kinetics of a photosynthetic microorganism (for instance, see Fig. 11). Monte Carlo method permits one to estimate through the statistical procedure both absorption and scattering of light in a given suspension of

145

Growth Kinetics of Photosynthetic Microorganisms

photosynthetic microorganisms without solving an integro-differential equation (Boltzmann's equation) that follows: uV/ ( Z , X , u) = - {£a(X) +

I ( X , X , u)

+ JJZ S (X) /(x, X, u ) 4n

P)l (u',u)dco'

(51)

provided: I = local light intensity, kJ c m - 2 h _ 1 u = unit vector X = wavelength, nm = sectional area per unit volume as function of X , cm" 1 X = position vector to = solid angle subscripts: a = absorption s = scattering The local light intensity, I(%, X , u) can be assessed by solving Eq. (51) under boundary condition of 7°(x°, X°, u°), where supercript 0 implies incident light. The 1st term on the right-hand side of Eq. (51) indicates the decrease of light intensity due to absorption and scattering when the light passes along u-direction. The 2nd term represents the increase of light intensity along u-direction, because light along u'-direction other than u is assumed to scatter and then pass along the udirection. A phase function, p^(u', u), for u' -*• u is assumed. The sum of both increase and decrease of light intensities referred to above is equated to the gradient of 7 ( x , X , u) along u as shown on the left-hand side of the equation. Light energy, 1°, incident onto the reactor and that of absorption, I a , in the reactor are given by: 10

= / / i

0 sj 2u

Ia = f J" J" 0 Vj 2n

«) d c o

ds

/(X, K u) dco dv dk

(52) (53)

It is hardly possible to solve for Eq. (51) analytically in a 6-dimensional Euclidean space and to integrate (see Eqs. (52) and (53)). Monte Carlo method is used to assess 7(x, X , u) and I a not analytically, but statistically. This procedure is to use often probability density function and/or variable, and first of all, cross sectional area of absorption and scattering, and £S(A,), respectively as functions of wavelength of light must be presented. Supposing that the semi-integral attenuance, — log (It/I0), is measured with respect to a specific microbial suspension by a recording spectrophotometer (light-path length, L, of cuvette = 1 cm), and assuming that no absorption of infrared light occurs, the value of I' 0 is: = i; + 1 ;

(cf. Eq. (3))

146

S. Aiba

i.e. 1 = T' + R'

(54) (cf. (Eq. (5))

where T' = It'/Io = transmittance R' = VJVQ = reflectance ' implies infrared light (cf. Chapter 3). Taking for granted the experimental fact 5 2 ) that the ratio of transmittance to reflectance is independent of wavelength, X, T(X)/R(X)

= T'/R'

(55)

If T(/.) values are measured spectrophotometrically, it is possible to estimate R(X) values from Eq. (55). Substituting T(A) and R(X) values into the following equations 5 8 ) that pertain to the solution of Eq. (51) in one-dimensional case, Xa(A.) and ZS(A) values could be assessed by an appropriate program, for example Program Library, C7 PO Wl, Computer Center, Univ. of Tokyo 59) .

T(X)

25

= (y + 6) e

R(A.) = —

: 5L

- (y - 5) e" 6L

_ 5 2 ) 1 ' 2 (e 6t - e " 5 t )

(56)

(57)

(7 + 5) eSL - (7 - 5) e" 5L 6 = {Z a M (I a (X) + X s a))} 1/2

(58)

7 = S a W + (1/2)

(59)

The point of this statistical procedure is to trace each photon trajectory within the reactor vessel (total number of photons traced = N), and to equate an expected value of random variable which defines whether absorbed photon is within volume element or not, for instance, to absorption, Ta/I° (cf. Eqs. (52) and (53); see Eqs. (65) to (68) later). Steps of computation for one of N photons are as follows 5 7 ) : Step 1. Select an appropriate wavelength, A,, that is in between and The value of k should satisfy: (k - l)/n ^

< k/n

(60)

147

Growth Kinetics of Photosynthetic Microorganisms

provided: E, = uniform random number between 0 and 1 n = natural number M 1/n = / f(X.) d l M-i

(61)

f(X.) = probability density function of wavelength with respect to light source f(/.) is approximated to as a sum of equi-area n rectangles, whose bottom and height being (X; — ^.¡-i) and f; (— ). The probability density is then V n(^-^i-i)/ uniform between _ i and so far as f(X) is concerned. Generating again uniform random number, i;, between 0 and 1, 1 =

(62)

Step 2. In accordance with specifications of light source, determine u° on the lightincident surface of the vessel. No reflection at the surface is assumed. Step 3. Following the geometrical configuration between light source and vessel, and in addition, from u°, determine x°. Step 4. Free-path length, /, of this photon is estimated from generating uniform random number, that satisfies the following equation: I=

In (1 - t) — W +

(63)

It is assumed that distribution of £a(/.) + £S(X.) within the reactor is uniform and then, the distribution of free-path length of photons is exponential. Step 5. Determine Xi from / and (/¡_,, u ^ , ) . If Xi is outside the vessel, this sequence of random walk terminates. If not, determine whether the photon is absorbed or scattered by using random number, with reference to the following criteria: absorption: scattering:

£ ^ Xa(/.)/{Ia(X) + XS(A)} > Za(>,)/{Xa(>0 + IS(X)}

(64)

Step 6. If absorbed, the sequence of random walk terminates. If scattered, the new direction, ui; that is different from U;_i, is determined by generating two uniform random numbers, ¡^, that define the directional cosine by: cos 27i^2, 2

si" 2t^ 2 , 1 - 2 ^ )

Return to Step 4. The above-mentioned series of steps could yield the following sequence of random walk for one trial of the photon {(X°, A

(Xi, «i),

,0ci. "i> k=Xj}

148

S. Aiba

N sets of such sequence of random walk are provided. Designating Sj and Vj to surface and volume elements, where subscript j is number of each element in view of geometry of the reactor (and light source), the following random variables, Xs ( X i - i > i d and %v.(xd are introduced next. x5.(Xi-i> xd is concerned with surface element and defines whether the photon, whose trajectory is from Xi-i to passes through Sj plane or not fl =

(pass) (no pass)

(65>

Xv.(Xi), dealing with volume element defines whether the photon, whose absorption is at Xb exists within Vj or not fl Xv:(}Ci)=-UJ i/J >S >

•S S s •S

S ^ .r¡ i « 43 u ^ •S c • 1 à «eM -s «j. s • ? ca .. o M ai M

a;

—- ** .y ? 5C •o «o-a

2 a E -X D, g D, u " O. M 2 jí ® S « "•OS Ë 5 Pi > H

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis

oo ri

— r- tj- o so m so "«t Os

U,

r j m oo ^

cd O

«>>

SO SO so so

TT tt m

00 so

3 o C/3

c o a 6 o o

os cn oo © vo oo oo f i r^ oo r i ò j i — "o C« c o c

I

I

o,

V)Tti/-ir-- ON . 8 9 > 102>. Acid hydrolysis of cellulose on industrial scale was started during the first world war, and it is still practiced extensively in the USSR 127) .

176

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

Studies on the use of acid pretreatments prior to enzymatic hydrolysis have been initiated only recently. Acids which have been proposed for use in pretreatments are sulfuric acid 151) , hydrochloric acid 5 7 ) , and phosphoric acid 1 2 9 , 1 4 8 \ Sulfuric acid. Han and Callihan 5 7 ) treated sugarcane bagasse and rice straw with sulfuric acid. Sugar production from sugarcane bagasse was maximized by treating it with 50% H 2 S 0 4 at 121 °C for 15 min, followed by dilution of the acid to 1 % and heating for 15 min at 121 °C, the sugar yield was 23%. Rice straw was treated with 3 % H 2 S 0 4 at 130 °C for 4 h; however, little improvement in growth of Cellulomonas sp. was observed. Sulfuric acid was also used by Tsao et al. 141) to pretreat lignocellulosics. They used a mild acid to separate the pentose fraction; the remaining residue was dried and mixed with concentrated acid (70-80 %) to dissolve the cellulose, which was eventually precipitated out of solution by the addition of methanol. This reprecipitated cellulose was amorphous and could be easily saccharified by acid or enzymatic hydrolysis. This method was further investigated by Wilke 153) , who obtained 35-49 % overall sugar yield from corn stover. Another sulfuric acid pretreatment was developed by Sasaki et al. 1 2 3 ) , who employed concentrated acid. In their method, cellulose was dissolved in a short period of time, followed by adding acetone for precipitating the dissolved cellulose. This process was applied to rice hulls, and 95 % of the cellulose was converted to glucose by a commercial enzyme system in 24 h. Knappert et al. 7 6 ) studied partial acid hydrolysis as a pretreatment for enhancing enzymatic hydrolysis of oak, corn stover, newsprint, and Solka Floe. They employed a continuous flow reactor, with temperature varying from 160° to 220 °C, and acid concentration from 0 to 1.2% at a fixed treatment time of 0.22 min. For all substrates, except Solka Floe, glucose yields were increased. In several cases, pretreatment resulted in 100% conversion of the potential glucose content of the substrate after 24 h of enzymatic hydrolysis. Dunning and Lathrop 3 4 ) carried out studies to establish an integrated process for utilizing agricultural residues. They pretreated a variety of agricultural residues by varying the acid concentration, temperature, and time. It has been found that pretreatment with 4.4%, at 100 °C for 55 min gives significantly improved results upon acid hydrolysis. Hydrochloric acid. Rice straw was pretreated with 3 % HC1 at 130 °C for 4 h by Han and Callihan 57) . They obtained decreased cell yield of Cellulomonas sp. grown on the treated substrate compared to the control. Phosphoric acid. Walseth 1481 investigated the possibility of pretreating cellulose with phosphoric acid. The pretreatment was carried out using 85% phosphoric acid and storing the samples at 2 °C for two different periods of time, namely, 10 min and 2 h. The substrate thus obtained was swollen and amorphous and hence, was highly reactive to enzymatic degradation. Walseth obtained about 80% digestion in 100 h under appropriate conditions of hydrolysis. Ghose and Kostick 4 9 ) have discussed the possibility of using 85 % phosphoric acid to swell cellulose as a means of pretreatment to enhance the saccharification of cellulose. Stone et al. 1 2 9 ) observed increased swelling with an increasing phosphoric acid concentration. Furthermore, they observed a linear relationship between the initial hydrolysis rate and accessible surface area of cellulose swollen by phosphoric acid.

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis

177

4.2.3 Gases Pretreatment with gases has the advantage in that it facilitates uniform penetration throughout the substrate, thereby giving uniform coverage. On the other hand, a gaseous medium is harder to work with than a liquid medium, and the recovery of the former for reuse poses more problems than that of the latter 153). Chlorine dioxide. Chlorine dioxide is an active agent of the chlorite pulping technique, which utilizes sodium chlorite in an acetic acid solution to solubilize lignin. Saarinen et al. 1 2 0 1 used the chlorite technique to improve the digestibility of birch wood for ruminants. Sullivan and Hershberger 131) passed C102 through a bed of dried and ground wheat straw. The maximum increase in the in vitro digestibility of treated wheat straw was 43 % over that of untreated wheat straw. They postulated that the increase was attributable to decomposition of lignin by chlorine and its oxides. Nitrogen oxides. Brink et al. 1 7 , 1 8 ) investigated a pulping scheme using nitrogen oxides. It gave a rapid rate of delignification and a higher overall sugar yield than conventional pulping processes at comparable lignin contents. These promising results indicate that nitrogen oxides may provide an effective pretreatment for enzymatic hydrolysis. Borrevik et al. 1 6 ) and Wilke 153> used nitrogen oxides as pretreatment agents for lignocellulosics. Wilke pretreated 100 g of wheat straw in a 22 1 vessel with 5 g of NO, followed by an addition of 6 g of 0 2 . The treatment lasted for 24 h at 25 °C. The best result obtained was a xylose yield of 69 % based on the original xylan content of straw. By applying NO first to a suspension of cellulose and then leaving it to stand a few minutes before feeding oxygen, NO was allowed to permeate uniformly throughout the substrate prior to the reaction of NO with 0 2 to form N 0 2 . Nitrogen dioxide then reacts with water to form nitric acid, which oxidizes cellulose and degrades lignin. Sulfur dioxide. Sulfur dioxide pretreatment is relatively unexplored as evident from the few publications. According to Dunlap et al. 32) sulfur dioxide pretreatment may be effective for lignocellulosics. Their process reacts gaseous S0 2 with moist lignocellulosics at 120° for 2 to 3 h. This treatment apparently disrupts the lignin-carbohydrate complex and depolymerizes the lignin. Millett et al. 97) studied the reactions of various species of wood with gaseous sulfur dioxide for a period of 2 to 3 h at room temperature and gas pressure of 30 psi without free water. Impressive results were obtained for both hardwoords and softwoods. Most of the hardwood carbohydrates were converted to sugars after being pretreated with S0 2 for 2 h, and 70-85 % of the softwood carbohydrates were converted to sugars after being pretreated for 3 h. Ozone. Schurz 124) has commented that ozone may be effective in degrading lignin without producing excessive amounts of pollutants. Recently Binder et al. 1 6 3 1 carried out ozone pretreatment of wheat straw, and obtained a drastic increase in the biodegradability of cellulose. They have observed that ozone attacks both lignin and carbohydrates, though the rate of reaction with latter is slower and that a 50% reduction in the original lignin content is optimal. A similar and independent observation has also been made by Fan et al. 42) .

178

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

4.2.4 Oxidizing Agents Oxidizing agents that have been used to pretreat lignocellulosics are listed in Table 5 57) . These agents cause structural modification of cellulose by penetrating into the cellulose and then oxidizing. Han and Callihan 57 ' have noted that some oxidants penetrate and react with both the crystalline and amorphous regions of cellulose, while other oxidants only attack the amorphous regions. Some oxidizing agents exist in gaseous form, and these have already been considered under the gas pretreatment section. Table 5. Oxidizing agents for cellulose571 Sodium chlorite (NaC102) Potassium bromate (KBr0 3 ) Potassium iodate (KI0 3 ) Potassium permanganate (KMn0 4 ) Potassium peroxydisulfate (K 2 S 2 0 8 ) Potassium Perchlorate (KC104)

Sodium hypochlorite (NaOCl) Hydrogen peroxide (H 2 0 2 ) Nitrogen dioxide (N0 2 ) Chlorine dioxide (C102) Sulfur dioxide (S0 2 ) Ozone (0 3 )

Hydrogen peroxide. The action of H 2 0 2 with F e + + as a catalyst causes cotton cellulose to oxidize and decompose into C 0 2 5 6 , 7 8 ) . The reaction is thought to be similar to what takes place when brown-rot fungi degrades lignocellulosics. Elmund et al. 35) treated feedlot waste at a slurry concentration of 5% with 150 ml H 2 0 2 and 10 ml ferrous sulfate at 22 "C for 6-8 h. The resultant substrate gave 46% cellulose conversion upon 24 h of enzymatic hydrolysis. Peracetic acid. Toyama and Ogawa 1 3 6 , 1 3 7 ) used 20% peracetic acid (consisting of acetic anhydride and 35 % hydrogen peroxide, 1:1 by volume) for delignifying corn stalks, sawdust from broad leaved trees, and sawdust from coniferous trees. A significant increase in the rate of enzymatic hydrolysis was observed. Recently Fan et al. 4 2 ' obtained a drastic increase in the digestibility of wheat straw upon peracetic acid pretreatment. This increase was attributed to the extensive delignification achieved by this pretreatment. 4.2.5 Cellulose Solvents Cellulose solvents, such as cadoxen and CMCS, are able to swell and transform solid cellulose into a soluble state 1 1 4 ' 1 3 2 '. This ability to dissolve cellulose has been exploited as a means of lignocellulosic pretreatment. Once cellulose is dissolved, the major factors that deter its degradation, high crystallinity and the presence of lignin, can be reduced. Turbak et al. 143> reported on a vast array of cellulose solvents, shown in Table 6. The cost of many of these solvents along with their toxicity may prohibit their industrial application. Cadoxen. Cadoxen is an alkaline solution, containing ethylene diamine and water. The solution can readily dissolve cellulose, which can be reprecipitated into a soft floe by adding excess water. Tsao et al. 6 5 , 8 1 , 1 4 0 - 1 4 2 1 have shown that the dissolved cellulose can be hydrolyzed easily before the floe re-crystallizes, and that up to 10%

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis Table 6. Solvents for cellulose

179

1431

Calcium thiocyanate Strontium thiocyanate Quaternary pyridinium salts Hydrazine hydrate Benzyl trimethylammonium hydrate Methylamine in DMSO Triethylamine oxide Cuprammonium hydroxide

bis (P-y-dihydroxy propyl) disulfide CMCS Cadoxen Phosphoric acid Nitric acid Hydrochloric acid Sulfuric acid

of the cellulose can be dissolved in cadoxen at room temperature, and that approximately 90% glucose yield can be attained based on the total amount of cellulose reprecipitated from the solution. This pretreatment, however, has disadvantages of solvent toxicity and high cost. CMCS. CMCS solvent is composed of sodium tartrate, ferric chloride, sodium sulfite, and sodium hydroxide solution. The CMCS solvent dissolves cellulose and the cellulose can be reprecipitated easily by adding excess water. Up to 4 % cellulose can be dissolved in CMCS at room temperature. Tsao et al. 1 4 2 ) reported approximately 95 % glucose yield based on the total amount of cellulose reprecipated from the solution. CMCS dissolves less cellulose than cadoxen; however, it is nontoxic. 4.2.6 Solvent Extraction Extraction of lignin by means of organic solvents has been studied extensively in systems using mineral acids as catalysts and in uncatalyzed systems 61) . In both the catalyzed and uncatalyzed systems, organic acids liberated during the process have been shown to accelerate delignification. Solvent extraction has been used mainly for pulping purposes. Kleinert 75) extracted lignin from spruce and poplar sawdust by a mixture of ethanol and water. Wilke 152> used a similar approach; wheat straw was pretreated with a mixture of 50% ethanol-water at pH 6 at 180 °C and 230 °C for approximately 1 h. Overall reducing sugar yields of 39.6% and 37.1 % were obtained at 180° and 230 °C, respectively, compared to a reducing sugar yield of 24% for untreated wheat straw. Wilke also studied the effect of a benzene-ethanol extraction upon 2 mm Wiley milled wheat straw and attained an overall reducing sugar yield of 38.2%. Sarkanen et al. 122> carried out delignification of western cottonwood using an ethanol-water system and a variety of catalysts. They have concluded that in the acidcatalyzed systems, hardwoods and straws are readily delignified, providing pulps in generally higher yields than the conventional kraft or soda processes. Softwoods require high temperatures for delignification by the acid-catalyzed process, and consequently, the pulp yields tend to be low. In base-catalyzed systems, ammonium and sodium sulfides are appropriate catalysts and the method is applicable to a wide variety of species. Selvam and Ghose 108) used ethylene glycol to extract lignin from rice husks. Their results have indicated that an optimum treatment involves heating the rice husks at 170 °C in a substrate to ethylene glycol ratio of 8 g per 100 ml with an acid catalyst

180

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

concentration of 0.8 % (conc. HC1) for 60 min; after 48 h of enzymatic hydrolysis, 46 mg of reducing sugar per ml was obtained. According to them, ethylene glycol pretreatment yields better results than milling, heating, and alkali pretreatment of rice husks. Organosolv delignification of southern yellow pine was investigated by April et al. 7 ) and Bowers and April 1 3 ) . Aqueous phenol was found to be more effective than aqueous n-butanol in an uncatalyzed system. Aqueous phenol led to a 90% delignification in 2 h at 205 °C. The yield of residual solids was 36%. Four hours of cooking with aqueous n-butanol at 205 °C resulted in a 40 % delignification. Solvent recovery was 90-95 % for n-butanol and 70-78 % for phenol. April et al. 7) also investigated the use of aqueous n-butanol for extracting lignin from sweet gum. The maximum delignification (92 %) was obtained at the reaction time of 2 h at 200 °C. Recently, organosolv delignification studies have been initiated by Nolan et al. 1 0 6 ) to investigate the use of solvent delignification as a pretreatment for enzymatic hydrolysis. An integrated process scheme has been presented involving n-butanol pulping; preliminary experimental results have also been presented. Economic evaluation of organosolv pulping was investigated by Nguyen et al. 1 0 5 ) ; they concluded that the recovery efficiency of ethanol used in lignin extraction is the most important step in determining the process economics. 4.2.7 Swelling Agents Swelling agents, primarily strong electrolytic solvents, are used for pretreatment of cellulose. Two kinds of swellings are known, intercrystalline swelling and intracrystalline swelling. For example, water can penetrate and loosen only the amorphous region of cellulose, this is considered as an intercrystalline swelling agent. On the other hand, swelling agents such as certain salts and alkali solutions (Table 7), affect

Table 7. Swelling agents for cellulose

1501

Sodium hydroxide Sulfuric acid Nitrogen dioxide in dimethylsulfoxide Zinc chloride Ruthenium red Phosphoric acid Trimethylbenzylammonium hydroxide Iron tartrate complex Methacrylate embedding Sodium zincate

both the amorphous and crystalline regions of cellulose; they are called intracrystalline swelling agents 124) . In other words, intracrystalline swelling agents are effective in loosening up the crystalline region of cellulose. As an illustration, the action of swelling agents untwists the outer skin of cotton fibers and causes it to split and form collars; the inner cellulose layers swell rapidly between the collars 150) .

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis

181

4.3 Biological Pretreatments Biological pretreatment utilizes wood attacking microorganisms that can degrade lignin. They can be classified into three categories, brown rots, white rots, and red rots. Brown rots attack mainly cellulose; white rots and red rots attack both lignin and cellulose. Present research is aimed towards finding those organisms which can degrade lignin. Table 8 lists the microorganisms which are under investigation 124) . Ander and Eriksson 1} classified four types of microorganisms that degrade wood components: bacteria, soft-rot fungi, brown-rot fungi, and white-rot fungi. According to them, the white-rot fungi is the most promising for lignocellulosic pretreatment. Table 8. Some microorganisms that attack wood

124)

Brown rots (attack cellulose) Piptoporus betulinus Laetiporus sulphureus Trametes quercina Fomitopsis pinicola Gloephyllum saepiarium White rots (attack both lignin and cellulose) Fomes fomentarius Phellinus igniarius Ganoderma appalanatum Armillaria mellea Pleurotus ostreatus Red rot (attacks both lignin and cellulose) Fomitopsis annosa

The white-rot fungi are basidiomycetes which have over 1,000 species; only about 25 species have been examined for lignin decomposition. Kirk and Harkin 73) used a white-rot fungus to remove 42% of the lignin, 3% of the glucan (including cellulose), and 30% of the hemicellulose of birch wood. Degradation of lignin by white-rot fungi is co-oxidative process consequently accompanying carbon source is necessary (e.g., cellulose and/or hemicellulose). The fungus ligninases appear to attack the phenolic residues with demethylation and ring cleavage. Eriksson and Goodell 3 6 , 3 7 ) achieved almost specific lignin degradation by using cellulase-less mutants of whiterot fungi. By exposing Polyporus adustus to UV light, a regulatory gene, which controls the synthesis of cellulase, mannanase, and xylanase, was destroyed. Ander and Eriksson u reduced the lignin content of birch rods by 17 % and stiffness by 10 % by exposing them to a cellulase-less mutant white-rot fungus for six weeks. Recently, Detroy et al. 3 0 , 1 6 1 ) degraded lignin in wheat straw using P. ostreatus. They observed that losses of lignin and cellulose were 22 and 14%, respectively, after 30 days, and were 40 and 32 %, respectively, after 70 days. The organism appeared to selectively degrade lignin during the first six days. A study using Cyathus stercoreus also resulted in preferential degradation of lignin; the experiment with pressure cooked wheat straw degraded 45 % of the lignin, but consumed only 20 % of the cellulose 161

182

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

Biological deligniflcation appears to be a promising technique, but its low rate has prevented its usage in large scale industrial processes. The possibility, however, exists to accelerate its rate through genetic modification of lignin degrading microorganisms or through partial physical and/or chemical processing of a substrate prior to biological pretreatment.

4.4 Economic Analysis of Pretreatments Recently, Fan and coworkers 42 ' identified several effective pretreatments, both physical and chemical, for the enzymatic hydrolysis of wheat straw. Preliminary cost analyses were conducted, the results of which are enlisted in Table 9 and 10. The cost of a chemical pretreatment included only the cost of chemicals used and that of a physical pretreatment included only the cost of energy consumed by laboratory scale apparatus. Therefore, the cost of physical pretreatments would have been further reduced for an industrial scale of operation. The costs of chemical pretreatments varied from 4 $ per kg of wheat straw for caustic pretreatment to $ 11.25 per kg of wheat straw for ethylene glycol treatment. The costs of physical pretreatments varied from 1 if per kg of wheat straw for Fitzmilling to $ 2.24 per kg of wheat straw for rolling-milling. Table 9. Cost analysis of chemical pretreatment methods Type of Pretreatment

Yield of sugar per kg of wheat straw, g kg" 1

Extent of hydrolysis after 8h, gl"1

Cost of chemicals per kg of wheat straw, $ per kg" 1

Pretreatment cost based on sugar, $ per kg" 1

Caustic-AC Sulfite-AC Hypochlorite-AC Peracetic Acid Butanol Ethylene glycol-AC Sulfuric acid Standard

341.0 252.9 239.9 279.9 72.4 241.2 140.5 70.0

20.5 16.0 13.3 20.9 4.7 18.2 9.4 2.1

0.04 0.32 0.94 7.51 2.86 11.25

0.12 1.26 3.92 26.84 39.49 46.53 0.78

0.11 —

—

Table 10. Cost analysis of physical pretreatment methods Type of . pretreatment

Yield of sugar per kg of wheat straw, g kg" 1

Extent of hydrolysis after 8 h , g I" 1

Cost of energy consumed per kg of wheat straw, $ per kg

Pretreatment cost based on sugar, $ per kg

Ball-milling, 8 h Roller-milling, */4 h Fitz-milling, fine Extrusion, with pressure Y-Irradiation 50 Mrads

255.98 160.0 100.0

9.1 6.48 3.0

1.48 2.24 0.01

5.82 14.0 0.1

64.29

2.5

0.01

0.16

207.02

5.6

0.1

0.48

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis

183

Among the physical pretreatments, ball-milling gave the most promising results in terms of the hydrolysis rate and sugar yield. The pretreatment is clean and easy to operate, but the pretreatment time of 8 h may make it impractical on a large scale operation. Among the chemical pretreatments, caustic and sulfite pretreatments appeared most promising. Caustic pretreatment was identified as a potential candidate for large scale process development based on pretreatment cost, hydrolysis rate, and sugar yield it rendered. In addition, ethylene glycol pretreatment was notable because of its effectiveness and possibility of recovery through an appropriate scheme, which would significantly reduce its cost. The chemical pretreatments, however, have disadvantages which must not be ignored. These include use of specialized corrosion resistant equipment, need of extensive washing, and disposal of chemical wastes 42) .

5 Acknowledgement This work was supported by the Department of Energy grant no. DE-FG02-79ET00080 37.

6 References 1. Ander, P., Eriksson, K. E.: In: Progress in Ind. Microbiol. Vol. 14, M. J. Bull (Ed.), p. 1. New York: Elsevier 1978 2. Anderson, A. W., Han, Y. W. : NSF-RA-770278, Non Conventional Protein Food Processing Conference, PB-283965, p. 202, 1977 3. Anderson, D. C., Ralston, A. T.: J. Anim. Sci. 37 (1), 148 (1973) 4. Andren, R. K.., Erickson, R. J., Medeiros, J. E.: In: Biotech. Bioeng. Symp. No. 6, E. L. Gaden Jr. et al. (Eds.), p. 177. New York: Interscience 1976 5. Andren, R. K., Mandels, M. H., Medeiros, J. E. : In : Appi. Poly. Symp. No: 28, T. E. Timell (Ed.), p. 205. New York: Interscience 1975 6. Andren, R. K., Nystrom, J. M.: AIChE Symp. Ser. No. 158, Vol. 72, 91 (1976) 7. April, G. C. et al.: Presented at the 89th AIChE Nat. Meet. Portland, Oregon, Aug. 17—20, 1980 8. Arthur, Jr., J. C.: In: Energetics and Mechanisms in Radiation Biology, G. O. Phillips (Ed.), New York: Academic Press 1968 9. Autrey, K. M., McCaskey, T. A., Little, J. A.: J. Dairy Sci. 58 (1), 67 (1975) 10. Baker, A. J.: J. Anim. Sci. 36 (4), 768 (1973) 11. Batsawde, K. B. et al.: In: Bioconversion of Cellulosic Substances into Energy Chemicals and Microbial Protein Symp. Proc., T. K. Ghose (Ed.), p. 387. New Delhi: IIT Delhi 1978 12. Bender, F., Heaney, D. P., Bowden, A. : Forest Prod. J. 20, 36 (1970) 13. Bowers, G. H., April, G. C.: TAPPI 60 (8), 102 (1977) 14. Blouin, F. A., Arthur, Jr., J. C.: J. Chem. Eng. Data 5, 470 (1960) 15. Brenner, W. et al.: In: New Approaches for the Acid Hydrolysis of Cellulose, Dept. of Appi. Sci., New York Univ., NYU/DAS-77-30, New York 1977 16. Borrevik, R. K., Wilke, C. R., Brink, D. L.: In: Effect of Nitrogen Oxide Pretreatment on Enzymatic Hydrolysis of Cellulose, Lawrence Berkeley Lab., Univ. of Calif., Berkeley, Ca. LBL 7879, Sept. 1978 17. Brink, D. L.: U.S. Pat. 4,076,579, Feb. 1978 18. Brink, D. L. et al.: In: Abstracts of Papers of the 167th Amer. Chem. Soc. Meet., Los Angeles, Ca. March 31—April 5, 1974

184

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

19. Browning, B. L. : In : Handbook of Pulp and Paper Technology, 2nd Ed. K. W. Britt (ed.), p. 7. New York: Van Nostrand Reinhold 1970 20. Casey, J. P.: In: Pulp and Paper, 2nd Ed., Vol. 1, New York: Interscience 1960 21. Casey, J. P.: In: Pulp and Paper, 2nd Ed., Vol. 1, p. 399. New York: Interscience 1960 22. Chandra, S., Jackson, M. G . : J. Agr. Sci. Camb. 77, 11 (1971) 23. Chang, M.: J. Polym. Sci.: Part C, No. 36, 343 (1971) 24. Clarke, S. D., Dyer, I. A.: J. Anim. Sci. 37, 1022 (1973) 25. Cowling, E. B., Kirk, T. K. : Biotech. Bioeng. Symp. No. 6, E. L. Gaden Jr. et al. (Eds.), p. 163. New York: Interscience 1976 26. Cowling, E. B.: In: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 163. New York: Interscience 1975 27. Cysewski, G. R., Wilke, C. R . : Biotech. Bioeng. 18, 1297 (1976) 28. Datta, R . : Presented at AIChE/73rd Annual Meet., Chicago, Nov. 16—20, 1980 29. Dehority, B. A., Johnson, J. J . : J. Dairy Sci. 44, 2242 (1961) 30. Detroy, R. W. et al.: In: Biotech. Bioeng. Symp. No. 10, C. D. Scott (Ed.), p. 135. New York: Interscience 1980 31. Donefer, E., Adeleye, I. O. A., Jones, T. A. O. C. : In: Adv. Chem. Ser., G. J. Hajny, E. T. Reese (Eds.), 95, p. 328, Amer. Chem. Soc., Wash. D.C. 1969 32. Diinlap, C. E., Thomson, J., Chiang, L. C.: AIChE Symp. Ser. No. 158, Vol. 72, p. 58, 1976 33. Dunlap, C. E., Chiang, L. C.: In: Cellulose Degradation — A Common Link, M. C. Shuler (Ec.), C.R.C. Press 1980 34. Dunning, J. W., Lathrop, E. C.: Ind. Eng. Chem. 57,-24 (1945) 35. Elmund, G. K., Grant, D. W., Morrison, S. M.: In: Management of Livestock Wastes, Proc. 3rd Int. Symp. on Livestock Wastes, p. 275, A.S.A.E. Publication, 1975 36. Eriksson, K. E., Goodell, E. W.: Can. J. Microbiol. 20, 371 (1974) 37. Eriksson, K. E.: In: Bioconversion of Cellulosic Substrates into Energy Chemicals and Microbial Protein Symp. Proc., T. K. Ghose (Ed.), p. 195. New Delhi: IIT Delhi 1978 38. Fan, L. T., Lee, Y. H., Beardmore, D . H . : Presented at Und Int. Symp. on Bioconv. and Biochem. Eng., IIT, Delhi, India, March 3 - 6 , 1980 39. Fan, L. T., Lee, Y. H., Beardmore, D. H.: Biotech. Bioeng. 22, 177 (1980) 40. Fan, L. T., Lee, Y. H., Beardmore, D. H.: In: Adv. Biochem. Eng. Vol. 14, A. Fiechter (Ed.), p. 101, Berlin: Springer 1980 41. Fan, L. T., Lee, Y. H., Beardmore, D. H.: Biotech. Bioeng. 23, 419 (1981) 42. Fan, L. T., Gharpuray, M. M., Lee, Y. H.: Presented at 3rd Symp. on Biotech, in Energy Production and Conservation, Gatlinburg, Tennessee, May 12—15, 1981 43. Feist, W. C., Baker, A. J., Tarkow, H.: J . Ani. Sci. 30, 832 (1970) 44. Ferguson, W. S.: Biochem. J. 36, 786 (1942) 45. Ferguson, W. S.: J. Agr. Sci. 33, 174 (1943) 46. Forss, K., Fremer, K. E.: In: Symp. on Enzymatic Hydrolysis of Cellulose, M. Bailey et al. (Eds.), p. 41. Helsinki : SITRA 1975 47. Ghose, T. K . : Biotech. Bioeng. 11. 239 (1969) 48. Ghose, T. K. : In : Adv. Biochem. Eng. Vol. 6, T. K. Ghose, A. Fiechter, N. Blakebrough (Eds.), p. 39. Berlin: Springer 1977 49. Ghose, T. K., Kostick, J. A.: In: Adv. Chem. Ser., G. J. Hajny, E. T. Reese (Eds.), 95, p. 415, Amer. Chem. Soc., Wash. D.C. 1969 50. Glegg, R. E., Kertesz, Z. I.: J. Poly. Sci. 26, 289 (1957) 51. Godden, W.: J. Agr. Sci. 10, 437 (1920) 52. Guggolz, J. R., Köhler, G. O., Klopfenstein, T. J . : J. Anim. Sci. 33, 151 (1971) 53. Guggolz, J. R. et al.: J. Anim. Sci. 33, 167 (1971) 54. Hale, J. D.: In: The Pulping of Wood, 2nd Ed. R. G. MacDonald (Ed.), Vol. 1, p. 14. New York: McGraw Hill 1969 55. Hall, J. A., Saeman, J. F., Harris, J. F . : Unasylva 10, 7 (1956) 56. Halliwell, G. : In : Bioconversion of Cellulosic Substrates into Energy Chemicals and Microbial Protein, T. K. Ghose (Ed.), p. 81. New Delhi: IIT Delhi 1977 57. Han, Y. W., Callihan, C. D.: Appi. Microbiol. 27, 159 (1974) 58. Han, Y. W., Chen, W. P.: Biotech. Bioeng. 20, 567 (1978)

The Nature of Lignocellulosics and Their Pretreatments for Enzymatic Hydrolysis 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100.

185

Han, Y. W., Pence, J. W., Anderson, A. W.: Appl. Microbiol. 29, 708 (1975) Han, Y. W., Lee, J. S., Anderson, A. W.: J. Agr. Food Chem. 23, 928 (1975) Hansen, S. M., April, G. C.: Biosources Digest 3 (1), Jan. 1981 Harris, E. E. et al.: U.S. Dept. Agr. Forest Serv. Rep. 1618, Wash. D.C. 1969 Heaney, D. C., Bender, F.: Forest Prod. J. 20, 98 (1970) Howsmon, J. A., Marchessault, R. H.: J. Appl. Poly. Sei. 1, 313 (1959) Hsu, T. A., Ladisch, M. R., Tsao, G. T.: Chemtech, May 1980 Huffman, J. G., Kitts, W. D., Krishnamurthi, C. R.: Can. J. Anim. Sei. 51, 457 (1971) Israilides, C. J., Grant, G. A., Han, Y. W.: Develop. Ind. Microbiol., 20, 603 (1979) Janes, R. L.: In: The Pulping of Wood, R. G. MacDonald (Ed.), 2nd'Ed. Vol. 1, p. 34. New York: McGraw Hill 1969 Jones, R. P., Greenfield, P. F., Nickiin, D. J.: Cellulose Hydrolysis, State of the Art and Near Term Potential, a report submitted to the Queensland Dept. of Commercial and Industrial Development, Jan. 1980 Katz, M., Reese, E. T.: Appl. Microbiol. 16 (2), 419 (1968) Kelly, J. A.: Radiolytic Hydrolysis of Cellulose, U.S. Environ. Protec. Agency Report EPA-670/2-73-030,1973 Kelsey, R. G., Shafizadeh, F.: Biotech. Bioeng. 22, 1025 (1980) Kirk, T. K., Harkin, J. M.: AIChE Symp. Ser. No. 133, 124 (1973) Klee, A. J., Rogers, C. J.: In: Proceedings of the Second Pacific Chemical Engineering Congress, Vol. 2, p. 759. New York: AIChE 1977 Kleinen, T. N.i.TAPPI, 57, 99 (1974) Knappert, D., Grethlein, H., Converse, A.: Biotech. Bioeng. 22, 1449 (1980) Kobayashi, T.: Proc. Biochem. 19 (Feb. 1971) Koenigs, J. W.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 151. New York: Interscience 1975 Kumakura, M. Kaetsu, I.: Biotech. Bioeng. 20, 1309 (1979) Kunz, N. D., Gainer, J. L„ Kelly, J. L.: Nuclear Technol. 16, 556 (1972) Ladisch, M. R., Flickinger, M. C., Tsao, G. T.: Energy 4, 263 (1978) Lawton, E. J. et al.: Science 113, 380 (1951) Lee, Y. H., Fan, L. T.: Presented at Vlth Int. Ferment. Symp. London, Ontario, Canada, July 2 0 - 2 5 , 1980 Lee, Y. Y. et al.: Biotech. Bioeng. Symp. No. 8, C. D. Scott (Ed.), p. 75. New York: Interscience 1978 Lehmann, F.: Ger. Pat. No. 169,800, March 26, 1905 Linden, J. C., Murphy, V. G., Moreira, A. R.: Presented at the Vlth Int. Ferment. Symp., London, Ontario, Canada, July 20—25, 1980 Linko, M.: In: Adv. Biochem. Eng. Vol. 5, T. K. Ghose, A. Fiechter, N. Blakebrough (Eds.), p. 27. Berlin: Springer 1977 Lipinsky, E.: In: Adv. Chem. Ser., R. D. Brown, Jr., L. Jurasek (Eds.), 181, p. 1. Amer. Chem. Soc., Wash. D.C. 1979 Lloyd, R. A., Harris, J. R.: U.S. Dept. Agr. Forest Serv. Report, 2029, Wash. D.C. 1963 MacDonald, D. G., Mathews, J. F.: Biotech. Bioeng. 21, 1091 (1979) Mandels, M„ Hontz, L., Nystrom, J.: Biotech. Bioeng. 16, 1471 (1974) Mandels, M., Sternberg, D.: J. Ferment. Technol. 54, 267 (1976) Mellenberger, R. W. et al.: J. Anim. Sei. 32, 756 (1971) Mellenberger, R. W. et al.: J. Anim. Sei. 30, 1005 (1970) Millett, M. A. et al.: J. Anim. Sei. 31, 781 (1970) Millett, M. A., Baker, A. J., Satter, L. D.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 193. New York: Interscience 1975 Millett, M. A., Baker, A. J., Satter, L. D.: Biotech. Bioeng. Symp. No. 6, E. L. Gaden Jr. et al. (Eds.), p. 125. New York: Interscience 1976 Millett, M. A., Effland, M. J., Caulfield, D. F.: In: Adv. Chem. Ser., R. D. Brown Jr., L. Jurasek (Eds.), 181, p. 71. Amer. Chem. Soc., Wash. D.C. 1979 Millett, M. A.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 319. New York: Interscience 1975 Millett, M. A., Goedken, V. L.: TAPPI 48, 367 (1965)

186 101. 102. 103. 104. 105.

L. T. Fan, Y.-H. Lee, M. M. Gharpuray

Moore, W. E., Eflland, M. J., Millett, M. A.: J. Agr. Food Chem. 20, 1173 (1972) Moo-Young, M., Chahal, D. S„ Vlach, D.: Biotech. Bioeng. 20, 107 (1978) Nesse, N., Wallick, J., Harper, J. M.: Biotech. Bioeng. 19, 323 (1977) Nisizawa, K.: J. Ferment. Technol. 51, 267 (1973) Nguyen, X. N. et al.: Presented at the 89th AIChE Nat. Meet., Portland, Oregon, August 17-20, 1980 106. Nolan, E. J. et al.: Presented at the Vlth Int. Ferment. Symp., London, Ontario, Canada, July 2 0 - 2 5 , 1980 107. Nystrom, J.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 221. New York: Interscience 1975 108. Selvam, P. V., Ghose, T. K.: Presented at the Und Int. Symp. on Bioconv. and Biochem. Eng., IIT Delhi, India, March 3 - 6 , 1980 109. Peitersen, N., Andersen, B.: AIChE Symp. Ser. No. 172, Vol. 74, 100 (1978) 110. Peitersen, N. : In : Symp. on Enzymatic Hydrolysis of Cellulose, M. Bailey, T. M. Enari, M. Linko (Eds.), p. 407. Helsinki: SITRA 1975 111. Pew, J. C.: TAPPI 40, 553 (1957) 112. Pew, J. C., Weyna, P.: TAPPI 45, 247 (1962) 113. Phillip, B. et al.: In: Adv. Chem. Ser., R. D. Brown Jr., L. Jurasek (Eds.), 181, p. 127. Amer. Chem. Soc., Wash. D.C. 1979 114. Phillip, B., Schleicher, H., Wagenknecht, W.: Chemtech, Nov. 1977, p. 702. 115. Pritchard, G. I., Pigden, W. J., Minson, D. J.: Can. J. Anim. Sci. 42, 215 (1962) 116. Ranby, B. : In: Adv. Chem. Ser., G. J. Hajny, E. T. Reese (Eds.), 95, p. 139. Amer. Chem. Soc., Wash. D.C. 1969 117. Rogers, C. J. et al.: Environ. Sci. Technol. 6, 715 (1972) 118. Rydholm, S. A.: Pulping Processes, p. 218. New York: Interscience 1965 119. Ryu, D. D. Y: Presented at 6th Int. Ferment. Symp., London, Ontario, Canada, July 20—25, 1980 120. Saarinen, P., Jenson, W. J., Alhojarvi, J.: Acta Agral. Fennkja. 94, 41 (1959) 121. Saeman, J. F., Millett, M. A., Lawton, E. J.: Ind. Eng. Chem. 44, 2848 (1952) 122. Sarkanen, K. V., Assiz, S., Chiang, V.: Presented at the AIChE 89th Nat. Meet., Portland, Oregon, Aug. 17—20, 1980. 123. Sasaki, T., Nakagawa, S., Kainuma, K.: Presented at Amer. Asso, of Cereal Chemists Annual Meet., San Francisco, Oct. 23—28, 1977 124. Schurz, J.: In: Bioconversion of Cellulosic Substances into Energy Chemicals and Microbial Protein Symp. Proc., T. K. Ghose (Ed.), p. 37. New Delhi: IIT, Delhi 1978 125. Scott, R. W„ Millett, M. A., Hajny, G. J.: Forest Prod. J. 19 (4), 14 (1969) 126. Shafizadeh, F. : TAPPI Forest Biology Wood Chemistry Conf., Madison, June 20—22, 1977 127. Sharkov, V. I., Levanova, V. P.: Zh. Priklad. Khim. 33, 2563 (1960) 128. Spano, L. A. et al. : In: Interagency-Environment Research and Development Program Report, EPA-600/7-77/038,1977 129. Stone, J. E. et al. : In : Adv. Chem. Ser., G. J. Hajny, E. T. Reese (Eds.) 95, p. 223. Amer. Chem. Soc., Wash. D.C. 1969 130. Stranks, D. W.: Forest Prod. J. 9, 228 (1959) 131. Sullivan, J. T„ Hershberger, T. V.: Science 130, 1252 (1959) 132. Tanaka, M. et al.: J. Ferment. Technol. 57, 186 (1979) 133. Tarkow, H., Feist, W. C.: In: Adv. Chem. Series, G. J. Hajny, E. T. Reese (Eds.) 95, p. 197. Amer. Chem. Soc., Wash. D.C. 1969 134. Tassinari, T., Macy, C.: Biotech. Bioeng. 19, 1321 (1977) 135. Taylor, P. C.: In: Machinery and Equipment for Rubber and Plastics, Vol. 1, Seaman, R. G., Merrill, A. M. (Eds.), p. 13. New York: Bill Brothers Pub. Corp. 1952 136. Toyama, N., Ogawa, K.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 225. New York: Interscience 1975 137. Toyama, N.: Biotech. Bioeng. Symp. No. 6, E. L. Gaden Jr. et al. (Eds.), p. 207. New York: Interscience 1976 138. Toyama, N., Ogawa, K. : In: Symp. on Enzymatic Hydrolysis of Cellulose, M. Bailey, T. M. Enari, M. Linko (Eds.), p. 375, 1975

The Nature of Lignoceliulosics and Their Pretreatments for Enzymatic Hydrolysis

187

139. Toyama, N., Ogawa, K.: In: Bioconversion of Cellulosic Substances into Energy Chemicals and Microbial Protein Symp. Proc., T. K. Ghose (Ed.), p. 373. New Delhi: IIT Delhi 1978 140. Tsao, G. T.: J. Chinese Inst. Chem. Eng. 9, 1 (1978) 141. Tsao, G. T. et al.: In: Annual Reports on Fermentation Processes, Vol. 2, D. Perlman (Ed.), p. 1. New York: Academic Press 1978 142. Tsao, G. T.: Presented at U.S.-R.O.C. Joint Seminary Ferment. Eng., Philadelphia, May 30— June. 1,1978 143. Turbak, A. F. et al.: Chemtech, Jan. 1980 144. Tyagi, R. D., Ghose, T. K.: In: Bioconversion of Cellulosic Substances into Energy Chemicals and Microbial Protein Symp. Proc., T. K. Ghose (Ed.), p. 585. New Delhi: IIT Delhi 1978 145. Van Soest, P. J.: In: Adv. Chem. Ser., G. J. Hajny, E. T. Reese (Eds.) 95, p. 267. Amer. Chem. Soc., Wash. D.C. 1969 146. Waiss, Jr. A. C. et al.: J. Anim. Sci. 35, 109 (1972) 147. Walker, H. G. et al.: Cereal Chem. 47, 513 (1970) 148. Walseth, C. S.: TAPPI 35, 228 (1952) 149. Wardrop, A. B.: In: Lignins, K. V. Sarkanen, C. H. Ludwig (Eds.), p. 19. New York: Wiley Interscience 1971 150. Warwicker, J. O. et al.: In: Shirley Institute Pamph. No. 93, 1966 151. Wilke, C. R. : In : Pilot Plant Studies of the Bioconversion of Cellulose and Production of Ethanol, Lawrence Berkeley Lab., Univ. of Calif., Berkeley, Ca., LBL 6860, June 1977 152. Wilke, C. R. : In : Process Development Studies of the Bioconversion of Cellulose and Production of Ethanol, Lawrence Berkeley Lab., Univ. of Calif., Berkeley, Ca., LBL 6881, Feb. 1978 153. Wilke, C. R.: In: Process Development of the Bioconversion of Cellulose and Production of Ethanol, Lawrence Berkeley Lab., Univ. of Calif., Berkeley, Ca., LBL 7880, Sept. 1978 154. Wilke, C. R., Mitra, G.: Biotech. Bioeng. Symp. No. 5, C. R. Wilke (Ed.), p. 253. New York: Interscience 1975 155. Wilke, C. R., Yang, R. D. : In: Proc. of the VHIth Cellulose Conf. 1. Wood Chemicals-A Future Challenge, T. E. Timell (Ed.), p. 175. New York: Interscience 1975 156. Wilke, C. R., Yang, R. D., Stockar, U. v.: Biotech. Bioeng. Symp. No. 6, E. L. Gaden Jr. et al. (Eds.), p. 155. New York: Interscience 1976 157. Williams, M. A., Baer, S.: Feedstuffs 36, 48 (1964) 158. Wilson, R. K., Pigden, W. J.: Can. J. Anim. Sci. 44, 122 (1964) 159. Wilson, P. N., Brigstocke, T.: Process Biochem. 12, 17 (1977) 160. Woodman, H. E„ Evans, R. E.: J. Agr. Sci. 37, 202 (1947) 161. Anonymous: A Fungus to Digest Lignin, Gasohol U.S.A. 2 (10), 26 (1980) 162. Anonymous: Continuous Cellulose to Glucose Process, Chem. Eng. News, Oct. 8, 1979 163. Binder, A., Pelloni, L., Fiechter, A.: Eur. J. Appi. Microbiol. Biotech. 11, 1 (1980)