Growth energetics and metabolic fluxes in continuous cultures of Penicillium chrysogenum

Journal of

Biotechnology45 (1996) 149-164

Growth energetics and metabolic fluxes in continuous cultures of Penicillium chrysogenum C.M. Henriksen, L.H. Christensen ‘, J. Nielsen *, J. Villadsen Center for Process Biotechnology, Department of Biotechnology, Technical University of Denmark, DK-2800 Lyngby, Denmark

Received 7 July 1995;revised 10 October 1995;accepted 17 October 1995

Abstract Continuous cultures of the penicillin producing fungus Penicillium chrysogenum have been analyzed with respect to the macromolecular composition of the mycelium. All cultivations were carried out using a chemically defined medium with glucose as the growth limiting component. Biomass was harvested at steady state and analyzed for proteins, lipids, RNA, DNA, and carbohydrates. Carbohydrates present in the cell wall, i.e., glucans and chitin, and carbohydrates serving as storage materials, i.e., glycogen, were measured. It was observed that the levels of DNA and lipids are relative constant, whereas the proteins and stable RNA levels increase with the specific growth rate and the total amount of carbohydrates decreases with the specific growth rate. Glycogen is only present in small amounts, decreasing with the specific growth rate. As an average the measured macromolecules account for 77 f 2% (w/w> of the biomass. On the basis of estimations of the metabolic costs for biosynthesis and polymerization of the different macromolecules the total ATP and NADPH requirements for cell biosynthesis from glucose and inorganic salts, i.e., YxATP.srowth and YXNAopH, have been quantified. The biosynthesis of 1 g biomass was calculated to require 39.9 mmol of ATP and 7.5 mmol of NADPH when cytosolic acetyl-CoA is formed from citrate by citrate lyase and oxaloacetate is recycled back into the TCA cycle. Other pathways of acetyl-CoA biosynthesis have been considered. The calculations show that the different biosynthetic routes for generating cytosolic acetyl-CoA have a significant influence on the theoretical value of ATP and NADPH requirements for cell biosynthesis. Combining a detailed stoichiometric model for growth and product formation of P. chrysogenum with experimental data on the macromolecular composition of P. chrysogenum and precise measurements of substrate uptake and product formation the intracellular flux distribution was calculated for different cultivation conditions. Keywords: Penicillium chrysogenam; Continuous tion

cultivation;

Macromolecular

composition;

Energetic requirement;

Metabolic

flux distribu-

1. Introduction

’ Corresponding author. ’ Present address: Novo Nordisk A/S, Bagsvaxd,

NOVO All&

DK-2880

Denmark.

0168-1656/%/$15.00 8 1996 Elsevier Science B.V. All rights reserved SD1 0168-1656(95)00164-6

Growth energetics describes the relationship between generation and consumption of Gibbs free energy - generally in the form of A’lT. According to Stouthamer and Bettenhaussen (1973) the ATP gen-

’

150

CM.

Hrrwiksen

et al./

Journal

eration rATP (mmol ATP per g DW per h) is balanced by the growth- and the non-growth- (or maintenance) associated ATP consumption, Y,,,, (mmol ATP per g DW) and mATP (mmol ATP per g DW per h), respectively, as described by Eq. 1: rATP

YxATPII.

=

+

(1)

mATP

growth-associated ATP requirement three terms (Benthin et al., 1994):

The

YxATP

=

YxATP,growth

’

‘xATP,lysis

+

YxATP,leaks

consists

YxATP,lysis

md

45 (19961 149-164

NADH was determined. Furthermore, using a detailed metabolic model (Jorgensen et al., 1995a) the metabolic flux distributions were calculated at different specific growth rates.

2. Materials and methods

of

(2)

includes ATP consumption for where ‘xATP growth membrane transport processes, biosynthesis and polymerization, YxATP,,ysisincludes ATP consumption for repolymerization of degraded macromolecules, and YxATP,,eaksincludes ATP consumption for leaks and futile cycles. Theoretical calculation of the ATP requirement for biosynthesis of a microbial (mmol ATP per g DW), is possible cell* YxATP,~rowth on the basis of experimental data on the macromolecular composition of the microorganism concerned, whereas little is known concerning the values Of

oj’Biotec.hrwlogy

YxATP,leaks.

The macromolecular biosynthesis can be divided into the biosynthesis of (1) RNA, (2) DNA, (3) proteins, (4) carbohydrates, (5) amino carbohydrates and (6) lipids. The macromolecules are synthesized by polymerization of a number of building blocks which all are synthesized from 12 precursor metabolites (Ingraham et al., 1983). Since the energetic costs are quite different for synthesis of the various macromolecular pools it is important to know the macromolecular composition when YxATP,arowthis to be calculated. In the literature a large number of experiments concerning measurements of macromolecules in bacteria and yeasts have been described. But with respect to filamentous fungi the amount of data on macromolecular composition is limited. The few attempts to describe the macromolecular composition of filamentous fungi have focused either on some of the macromolecules making up the mycelium or on a part of the mycelium, e.g., the cell wall. In this work the size of all the major macromolecular pools were measured at different specific growth rates of a steady-state chemostat. From the measurements YxATP,growth was calculated, as well as the biosynthetic requirement for NADPH and synthesis of

2.1.1. Strain The applied high yielding strain P. chrysogenum was a donation from Novo Nordisk A/S known to give about 20-25 g I-’ penicillin V after 200 h of fed-batch cultivation (Jorgensen et al., 1995b). 2.1.2. Media Both the batch medium and the feed during continuous operation were defined. The batch medium contained 25 g 1-l sucrose, 1.6 g 1-l KH,PO,, 7.0 g 1-l (NH,),SO,, 0.5 g 1-l KCl, 0.04 g I-’ FeSO, .7H,O, 0.10 g I-’ MgSO,.7H,O, 0.05 g I-’ CaCl, . 2H,O, 0.5 ml 1-l Pluronic F-68 (Fluka), and 5 ml I-’ trace metal solution. The feed contained 15 g 1-l glucose (sterilized separately), 1.6 g l- ’ KH*PO,, 7.0 gl-’ (NH,),SO,, 0.5 g l-’ KCl, 0.04 g 1-l FeSO,. 7H,O, 0.10 g I-’ MgSO,. 7H,O, 0.05 g 1-l CaCl, . 2H,O, 0.3 ml 1-l Pluronic F-68, 5 ml 1-l trace metal solution, and 6.5 g I-’ phenoxyacetic acid. The trace metal solution used for both media contained 1.0 g l- ’ CuSO, . 5H,O, 4.0 g 1-l ZnSO, .7H,O, and 4.0 g I-’ MnSO, . H,O. 2.1.3. Cultivation conditions All cultivations were carried out in high performance Chemap bioreactors (Nielsen and Villadsen, 1993) with a volume of 7 and 9 I, respectively. Both were placed on a load cell with an accuracy of + 20 g. The temperature, pH, and head space pressure were kept constant at 25.O”C, 6.50, and 1.5 bar, respectively, and the aeration rate was 1 vvm. The agitation rate was kept constant at a level sufficiently high to maintain a dissolved oxygen tension above 60% of saturation. The batch cultivations were inoculated with spores from rice cultures. Continuous cultivation was initiated after 55-63 h of batch cultivation. At this time the carbohydrates were exhausted, leaving gluconic acid as the main carbon source. This gave rise to an initial drop in the carbon dioxide evolution rate (Nielsen et al., 1994).

C.M. Henriksen et d/Journal

2.1.4. Sampling

During the cultivations samples of the biomass were taken manually twice a day. At the same time cell free samples were collected automatically by means of an in-situ membrane module (Christensen et al., 1991). The cell free samples were collected for later analysis in a fraction collector positioned in a refrigerator. The samples taken manually were used to determine both the concentration of biomass and the macromolecular composition of the biomass. For biomass measurements the samples were filtered, washed with water, and dried at 105°C until constant weight. When determining the macromolecular composition the samples were filtered, washed with 0.9% NaCl, and freeze-dried. After freeze-drying the samples were kept dry in a desiccator. 2.1.5. Steady state Attainment of steady state was based on measurements of the biomass concentration, the oxygen consumption rate, the carbon dioxide evolution rate, and the concentration of penicillin V in the cell free samples. As illustrated in Christensen et al. (1995) the applied strain of P. chrysogenum is genetically unstable resulting in a decrease in the penicillin productivity after more than six to seven residence times. Hence, all steady-state data were obtained during the first six residence times after the steady state had been reached. 2.1.6. Analysis of carbohydrates lo-12 mg of freeze-dried biomass was suspended in 15 ml of water and partly homogenized by treatment with ultrasound for 30 min. The relative amount of carbohydrates was subsequently determined by analyzing 1.0 ml of the suspension by the method developed by Dubois et al. (1956) (later discussed by Herbert et al., 1971). The reported values for the carbohydrate content of the mycelium have been corrected for the contribution from RNA and DNA. 2.1.7. Analysis of glycogen Glycogen was determined by the method described by Schulze et al. (1995) using freeze-dried biomass samples. The method is based on degradation of glycogen to glucose by amyloglycosidase and measurement of the formed glucose by an enzyme kit based on glucose kinase and glucose-6-phosphate

of Biotechnology 45 (1996) 149-164

dehydrogenase (Boehringer-Mannheim, 716 251).

151

product No.

2.1.8. Analysis of amino carbohydrates and amino acids The analysis for amino carbohydrates was carried

out by the same HPLC method which was used to determine the amino acid composition of the proteins (Barkholt and Jensen, 1989) using 6 and 24 h of acid hydrolysis, respectively. The intracellular pool of free amino acids was determined by boiling freezedried biomass in water for 15 min. The extract was centrifuged for 10 min at 10000 X g and the supernatant was subsequently analyzed without any further hydrolysis (Stewart, 1975; Malaney et al., 1989). 2.1.9. Analysis of lipids Lipids were analyzed by means of a method based on the work of Folch et al. (1957). Approx. 100 mg of freeze-dried biomass was treated 20 min with a two phase system of 200 ~1 water and 3.75 ml chloroform/methanol 2:l (v/v>. The extraction of the lipids was facilitated by simultaneous treatment with ultrasound. The extracts were filtered (Whatman 1PS 2200-0701, followed by washing of the filtrates by addition of 1.25 ml 0.9% NaCl and subsequent mixing. The filtrates were centrifuged until two distinct phases reappeared, and the upper phase, including any possible floating material in the boundary zone, was discarded. Finally, the amount of lipids extracted from the biomass was determined gravimetrically after evaporation of the lower phase using a gentle stream of nitrogen. 2.1.10. Analysis of RNA The RNA content has been determined by the Schmidt-Thannhauser method as described by Benthin et al. (1991). 2.1.11. Analysis of DNA DNA was measured by means of the diphenylamine reagent method as described by Herbert et al. (1971). 2.1.12. Analysis of proteins Total nitrogen was measured by Kjeldahl digestion. The protein content was found by subtracting the other N-containing compound, i.e., free amino

152

C.M. Henrikxn

acids, amino carbohydrates, were determined separately.

et ~l./Journal

oj’Biotechnobgy

45 (1996) 149-164

RNA and DNA, which

3. Results and discussion 3.1. Macromolecular composition Samples of biomass glucose-limited steady-state

were harvested from continuous cultures and 8.00 +

Table 1 Macromolecular composition cific growth rates a

0.04

0.06

0.08

I

0.10

IT 0.15

Specific growth rate [h-l]

of P.

chysogenumat different spe-

Specific growth rate (h-‘)

0.022

0.039

0.066

0.075

0.101

RNA b DNA’ Proteins d Carbohydrates ghlcans chitin glycogen ’ Lipids s Rest h

4.3 0.9 36.4 27.5 18.8 7.4 1.3 5.5 25.4

4.6 1.0 41.6 28.2 23.0 4.6 0.6 3.4 21.2

6.1 1.1 42.4 22.7 17.6 4.7 0.4 3.2 24.5

6.3 1.0 42.9 22.8 18.2 4.4 0.2 3.8 23.2

7.2 0.8 44.0 21.9 13.0 8.9 0.0 5.3 20.8

e

0.02 ,

a All macromolecular pools arc in % (w/w). b The pool of RNA only includes stable RNA, i.e., rRNA and tRNA. ’ Similar results are reported for Trichodermu aureoviridr by Pitt and Bull (1982). but the constant level is lower (0.5% (w/w)). Righelato et al. (1968) reported a much lower level (0.17% (w/w)) for P. chrysogenum. d Pitt and Bull (1982) reported an increase in the protein content for Trichoderma aureooiride from 40% to 50% (w/w) with increasing specific growth rates. Mason and Righelato (1976) reported values for the protein content for P. chrysogenum of 35-38% (w/w) at different specific growth rates, and Righelato et al. (1968) obtained 30% (w/w) at a specific growth rate of 0.051 h-‘. ’ The pool of carbohydrates includes: (1) structural carbohydrates which are mainly associated with the cell walls such as chitin and different glucans (Stagg and Feather, 1973; Hamilton and Knight, 1962); (2) storage carbohydrates represented by glycogen. f According to Blumenthal (1976) glycogen acounts for 5% (w/w) of the cell dry weight, but the level depends strongly on the environmental conditions. g Meisgeier et al. (1990) reported that the fatty acid content of P. chrysogenum is between 5 and 20% (w/w), whereas the total lipid content in S. cerevisine is between 4 and 8% (w/w) (Ratledge and Evans, 1989). ’ The rest of the biomass can partly be accounted for by ash, which is about 8% of the biomass (Christensen et al., 1995) and pools of different building blocks and metabolites (Nielsen, 1995).

Fig. 1. Pools of stable RNA, proteins and carbohydrates in P. chrysogenum as functions of the specific growth rate. The carbohydrates represent both cell wall components (glucans and chitin) and storage carbohydrate (glycogen).

subsequently analyzed with respect to the macromolecular composition, i.e., the relative content of RNA, DNA, proteins, carbohydrates, amino carbohydrates and lipids. The macromolecular composition of the cells normally entered a steady state after approx. six to seven residence times. Table 1 summarizes the macromolecular composition at different specific growth rates in the range 0.022-0.101 h- ’ . The measurement of the different macromolecular pools can account for only 77 + 2% (w/w) of the biomass as an average. The remaining part can partly be accounted for by ash, which is about 8% of the biomass (Christensen et al., 19951, and by pools of different building blocks and metabolites (Nielsen, 1995). Finally, the difference may be due to degradation of some of the macromolecular components during sample preparation. As can be seen from Table 1 the pools of lipids and DNA were relatively constant at 3-6% (w/w) and 1% (w/w), respectively. Fig. 1 shows the pools of stable RNA, proteins and carbohydrates as functions of the specific growth rate. The pools of stable RNA and proteins both increase linearly with the specific growth rate, whereas the pool of carbohydrates decreases linearly with the specific growth rate. The linear increase of stable RNA with the specific growth rate has also been observed in other filamentous fungi (Pitt and Bull, 1982; Carlsen, 1994) and in a number of bacteria and yeasts (Nielsen and Villadsen, 1994). The total pool of cellular RNA

C.M. Henriksen et al./Journal

of Biotechnology 45 (1996) 149-164

consists of the highly unstable messenger RNA (mRNA) and the stable RNA, i.e., ribosomal RNA (rRNA) and transfer RNA (tRNA). There are no reports on the relative content of these three pools of RNA in filamentous fungi, but in Escherichia coli the relative content is 5% mRNA, 18% tRNA and 77% rRNA at a specific growth rate of 1.0 h- ’ (Ingraham et al., 1983). At lower specific growth rates the relative tRNA content increases at the expense of the rRNA content, and a similar observation has been made in Neurosporu crassa where the number of tRNA molecules per ribosome decreases with increasing specific growth rate (Alberghina et al., 1979). However, even at low specific growth rates rRNA accounts for more than 75% of the stable RNA, i.e., the sum of rRNA and tRNA. The pool of stable RNA is therefore a good measure of the ribosome levels in the cell - in Aspergillus niger the ribosomes consist of 53% RNA and 47% protein (Berry and Berry, 1976). Ribosomes play an important role in protein synthesis, and together with a few enzymes they make up the so-called protein synthesizing system (PSS; Ingraham et al., 1983). Since protein synthesis is energetically expensive for the cell (see Table 3) there is a tight control of the PSS in the cells. Thus the ribosomal level is adjusted to the requirements and this explains the linear relation

Table 2 Amino acid composition Ammo acid cr-Aminoadipic acid b Alanine Arginine Asparagine/aspartate ’ Cysteine/cystine d Glutamate/glutamine ’ Glycine Histidine Isoleucine Leucine

of the proteins and the free intracellular

153

between the pool of stable RNA and the specific growth rate observed in Fig. 1. A similar relation between the content of stable RNA and the specific growth rate is observed for lactic bacteria during dynamic changes in the environmental conditions (Benthin et al., 1991). During shift down experiments in the specific growth rate the synthesis of rRNA is observed to halt until the level of the ribosomes has adjusted to the new growth conditions (Sturani et al., 1973). At low specific growth rates the ribosome level of N. C~QSSU is constant, whereas the ribosome efficiency decreases with the specific growth rate (Alberghina et al., 1979). A similar observation is made for Aspergillus nidulans (Bushel1 and Bull, 1976). Thus, at low specific growth rates the cells have some unused capacity which may be rapidly activated when the environmental conditions change. The exact control mechanisms for synthesis of the ribosomes in filamentous fungi are not known, but there must be some similarities to the mechanisms found in bacteria, where the depletion of just one amino acid results in repression of the rRNA synthesis (Ingraham et al., 1983). As a consequence of the tight control of the PSS and the requirements for an increasing enzyme machinery with increasing specific growth rate the protein content also increases as a function of the

pool in P. chrysogenum a

Proteins (%)

Free pool (%)

Amino acid

Proteins (o/o)

Free pool (%)

10.0 k 0.4 4.8 f 0. I % 9.6 f 0.2 1.4kO.3 14.9 f 0.6 9.2 + 0.2 2.4kO.i 4.3 f 0.1 7.5 f 0.1

0.3 f 0. I (0.7 * 0.2) 19.8 * 3.5 (50.5 rt 9.5) 1.6 f 0.4c4.1 Ifr 1.1) 9.2 f 1.7 (23.4 f 4.1) I .O + 0.5 (2.5 k I .2) 38.5 f 3.6 (98.1 f 11.3) 3.8 * 0.7 (9.7 i 2.0) 0.8 k 0.3 (2.0 + 0.8) 0.6+0.3(1.7+ 1.0) I .6 +- 1.O (4.2 f 2.9)

Lysine Methionine Phenylalanine Proline Serine Threonine Tryptophan ’ Tyrosine Valine Total

5.6 1.7 3.4 4.7 6.1 5.3 _

1.1 k0.4(2.8f 1.3) 0.4+0.3(1.1 *0.9) 0.8 f 0.5 (2.1 f 1.6) 1.7 f 0.2 (4.5 f 0.8) 3.9 f 0.4(10.0 * 1.8) 12.6 k 2.3 (32.2 * 6.8) 0.3 rt 0.2 (0.9 f 0.61 0.5 f 0.3 (1.2 * 0.9) 1.9 f 0.5 (4.8 f I .5) 100.4 (256.5)

f 0.2 f 0.1 f 0.1 f 0.2 +O.l xto.1

2.6k0.1 6.4 f 0.2 99.9

a The data represent average values of 12 steady states at specific growth rates between 0.022 h- ’ and 0.108 h- ’ The figures show the relative content on a molar basis and the figures in paranthesis are the concentrations in units of pmol per g DW. b a-aminoadipic acid is a non-protein amino acid which is used as precursor in the penicillin biosynthetic pathway and it is an intermediate in the biosynthesis of lysine in fungi. ’ The figure represents the combined pool of asparagine and aspartate. d Cysteine and cystine are both derivatized with 3,3’-dithiopropionic acid and subsequently quantified as one component. ’ The figure represents the combined pool of glutamate and gtutamine. f Tryptophan is almost completely destroyed during the acid hydrolysis.

C.M. Hrnriksen

154

n

I

I’ -

04 000

,,,

002

I

I

0.04

n

.

11

0.06

Specific

Fig. 2. Total intracellular specific growth rates.

er al./Journul

pool

.

I

11

006

growth

“““I’

0.10

0.12

0.14

016

rate [h’]

of free amino acids at different

specific growth rate (Fig. 1). The increasing pool of proteins with the specific growth rate corresponds well with findings for other fungi (Pitt and Bull, 1982; Verduyn et al., 1990). In Table 2 the average amino acid composition is shown for both the proteins and for the pool of free amino acids. Neither the amino acid composition of the proteins nor the composition of the pools of free intracellular amino acids exhibit any significant change in the applied range of specific growth rates. The amino acid composition of the proteins corresponds well with findings reported by Jaklitsch et al. (1986), whereas the composition of the pools of some free amino acids deviate significantly, which may be explained by the different cultivation media applied. The total pool of free amino acids remains almost constant at a level of 230-290 pmol per g DW corresponding to 2.93.7% of the cell dry weight (Fig. 2) and this is a much higher value than reported by Jaklitsch et al. (1986), who used a complex medium. In Succharomyces cerevisiae grown in batch cultures with ammonia as sole nitrogen source the total intracellular amino acid pool is about 275 Fmol per g DW (Watson, 1976). The carbohydrates of P. chrysogenum are mainly associated with the cell walls in form of chitin and different glucans (Stagg and Feather, 1973; Hamilton and Knight, 1962), which are believed to be the most important structural polysaccharides in fungi (Barnicki-Garcia, 1968). In fungi the cell walls not only function as an osmotic barrier but also play a crucial

ofBiotechnology

45 (1996) 149-164

role for growth and morphogenesis. The structural polysaccharides make up 22-28% (w/w) of P. chrysogenunz with a distribution of 75% being different glucans and 25% being chitin (Table 1). Righelato et al. (1968) reported a carbohydrate pool (chitin not included) of approx. 18% (w/w) at a specific growth rate of 0.05 1 h- ‘, and according to Hamilton and Knight (1962) the sugar composition of the carbohydrates present in the cell wall is 49.4% glucose, 15.2% galactose, 5.6% mannose, 2.7% xylose, 2.4% rhamnose and 24.7% glucosamine, which corresponds well with the figures in Table 1. Besides the structural polysaccharides a minor part of the carbohydrates is present as storage carbohydrates, i.e., glycogen, trehalose and other polyols. As seen in Table 1 the glycogen content of P. chrysogenum increases at low specific growth rates, an observation also made for S. cerevisiae when grown in glucose limited continuous cultures (Kiienzi and Fiechter, 1972). Lipid is a heterologous group of compounds which in fungi not only serve as a major constituent of membrane systems and cell walls but also appear as lipid bodies which serve as storage material. In some cases lipids may also be found as extracellular products (Brennan and Lijsel, 1978). A major part of the lipids can be pooled into either: (1) acylglycerides, (2) phospholipids, and (3) sterols/sterolesters. The lipid content and lipid composition of fungi depend strongly on both the cultivation conditions and on the developmental stage of the fungus. Several attempts have been made to correlate the lipid composition, and especially the fatty acid composition, of a fungus to its taxonomic classification. The lipids in P. chrysogenum consist of about 76% phospholipids, 16% sterolesters and 8% triacylglycerides (Woodin and Wang, 1987). The major constituent of these lipids is fatty acids. In P. chrysogenum the overall fatty acid composition is: 8.2% palmitic acid (C 16:O). 6.6% stearic acid (C18:0), 23.8% oleic acid (Cl8: 1). 59.1% linoleic acid (C18:2) and 2.3% linolenic acid (Cl8:3) (Meisgeier et al., 1990). The phospholipid composition in P. chrysogenum is not known, but in S. cerevisiae it is approx. 50% phosphatidyl cholin (PC), 20% phosphatidyl inositol (PI) and 30% phosphatidyl ethanol (PE); (Ratledge and Evans, 1989). The sterolesters in fungi are mainly esters of ergosterol.

C.M. Henrikven et ul./ Journul of Biotechnology 45 (1996) 149-164

3.2. Growth energetics The catabolic and the anabolic pathways together with the penicillin biosynthesis for P. chrysogenum has recently been reviewed by Nielsen (1995). Here, only the biosynthetic routes for the formation of cytosolic acetyl-CoA will be discussed since these have a significant influence on the ATP and NADPH requirement for cell biosynthesis (Fig. 3). Acetyl-CoA is one of the 12 precursor metabolites from which all the building blocks for the biosynthesis of a cell can be synthesized (Ingraham et al., 1983). The major production of acetyl-CoA takes place in the mitochondria where pyruvate is irreversibly decarboxylated and converted into acetylCoA by the pyruvate dehydrogenase complex with a simultaneous reduction of NAD+. The acetyl-CoA may enter the TCA cycle by reaction with oxaloacetate to form citrate. Since acetyl-CoA can not traverse the inner mitochondrial membrane and it is needed in the cytosol for the biosynthesis of fatty acids and certain amino acids, i.e., L-lysine, L-leucine

155

and L-a-aminoadipic acid, it must in some other form be transported across the mitochondrial membrane or alternatively it has to be produced by another biosynthetic route in the cytosol. One possible way of synthesizing cytosolic acetyl-CoA is from citrate. Unlike acetyl-CoA citrate is readily transported from the mitochondria into the cytosol where ATP:citrate lyase may catalyze the formation of acetyl-CoA and oxaloacetate at the expense of one ATP. Normally the requirement for acetyl-CoA is larger than that for oxaloacetate and the oxaloacetate not used for biosynthesis in the cytosol is reduced by malate dehydrogenase to malate with NADH as electron donor. Malate may either be transported back into the mitochondrial matrix where it is reoxidized to oxaloacetate or it may become oxidatively decarboxylated to pyruvate (catalyzed by the malic enzyme) which readily enters the mitochondria. If the pyruvate is carboxylated into oxaloacetate by pyruvate carboxylase at the expense of one ATP, the entire biosynthetic route to cytosolic acetyl-CoA from glucose requires 1 mol ATP and produces 1 mol

% glucose

co2d-AIT wJ + NADH-

rG;;\r

’

_

w

i oxaloacetate

citr at

! &t

0V

eh

-

co*

I_ NADH

malate

Fig. 3. Different biosynthetic routes to cytosolic acetyl-CoA from glucose. The enzymes involved are: 1, Pyruvate dehydrogenase complex; II, citrate synthetase; III, AlPcitrate lyase; IV, malate dehydrogenase (cytosolic); V, malate dehydrogenase (mitochondrial); VI, malic enzyme; VII, pyruvate carboxylase; VIII, camitine carrier system; IX, pyruvate decarboxylase followed by acetaldehyde dehydrogenase; and X, acetyl-CoA synthetase.

C.M. Henrikxn

156

er ul./Journul

oj’Biutechn&gy

45 119961 149-164

tion of cytosolic acetyl-CoA is from pyruvate via acetate. First, pyruvate is decarboxylated to acetaldehyde (catalyzed by pyruvate decarboxylase) which is converted to acetate (catalyzed by acetaldehyde dehydrogenase). Finally, acetate is converted into acetyl-CoA by acetyl-CoA synthetase at the expense of one ATP (which is cleaved into AMP and pyrophosphate), and the total energetic costs for synthesizing cytosolic acetyl-CoA from glucose, therefore, adds up to one ATP, with two NADH being simultanously produced. Similar to calculations for S. cerevisiue (Verduyn et al., 1990) LTP,srowth was found based on the estimates of the metabolic costs for membrane transport processes, biosynthesis and polymerization reported by Nielsen (1995). In addition, the NADPH requirement for biomass synthesis and the NADH production was calculated. The results are summarized in Table 3. The calculations are based on the cell composition of P. chrysogenum at balanced growth at a specific growth rate of about 0.1 hh ‘,

NADPH and 1 mol NADH per mol acetyl-CoA formed. There are no reports on citrate lyase activity being present in P. chrysogenum, although the active enzyme has been purified from a number of yeasts (Boulton and Ratledge, 1981), from the filamentous fungi P. spiculisporum (M5hlen, 1973) and from citric acid producing strains of Aspergillus niger (Jernejc et al., 1991; Pfizner et al., 1987). Cytosolic acetyl-CoA may also be supplied by transport of mitochondrial acetyl-CoA across the inner mitochondrial membrane as a carnitine derivative by the carnitine carrier system. This transport system normally transports acyl-groups from the cytosol to the mitochondrial matrix, but activity of the transport system in the oppositive direction has been detected in some microorganisms (Kohlhaw and Tan-Wilson, 1977; Holdsworth et al., 1988). If the biosynthetic route to cytosolic acetyl-CoA is via the camitine carrier system the cell gains 1 mol of ATP and 2 mol of NADH for each mol of cytosolic acetyl-CoA being formed from glucose. The last route for formaTable 3 ATP. NADH

and NADPH

reauirements for cell biosynthesis of P. chrv.wvnum

A

Content

ATP (mmol

NADPH

(g per g DW

per g DW)

per g

Protein

0.450

2 I.097

RNA

0.080

2.856

-0.318

- 0.650

DNA

0.010

0.402

-0.001

- 0.08

Lipids

0.050 - 0.966

Macromolecule

(mmol

NADH

DW)

(mmol

per g DW)

7.009

- 11.141

I

phospholipids

0.035

I .649

0.654

sterolesters

0.010

0.803

0.029

- 0.432

triacylglycerides

0.005

0.294

0.119

-0.173 - 0.355

0.250

Carbohydrates cell wall

0.220

2.889

- 0.355

glycogen

0.030

0.370

0.000

0.000

0.080

Soluble pool amino acids

0.040

0.082

0.452

nucleotides

0.020

0.580

- 0.054

metabolites, etc.

0.020

- 0.805 -0.174

0.080

Ash Transport b ammonia

7.796

0.000

sulphate

0.131

0.000

O.OW

phosphate

0.944

0.000

0.000

I .ooo

Total

a The calculations are based on the review by Nielsen (1995)

39.894

O.OCO

7.535

on the metabolic pathways in P. chry.wgenum

- 14.777 assuming that cytosotic

acetyl-CoA is synthesized from citrate by citrate lyase with subsequent recycling of oxaloacetate via malate/pyruvate. composition of P. chrysogenum

b Glucose

is at a specific growth rate of 0.1 h-

is assumed to be transported by facilitated diffusion,

phosphate is symported with two protons (Nielsen,

1995).

The macromolecular

’

whereas ammonia and sulphate are symported with one proton and

C.M. Hem&en

et d./Journul

of Biotechnology 45 (1996) 149-164

and synthesis of all macromolecules from glucose and inorganic salts. Furthermore, cytosolic acetylCoA is assumed to be synthesized from citrate by ATP:citrate lyase with recycling of oxaloacetate via malate/pyruvate. Under these conditions the ATP and NADPH requirements for cell biosynthesis sum to 39.9 and 7.5 mmol per g DW, respectively, whereas 14.8 mmol of NADH is produced per g dry weight. 67% of the total ATP requirement is needed for the biosynthesis of proteins, i.e., 26.8 mmol per g DW (including transport of ammonia used for protein synthesis). Two-thirds of this is required for polymerization of the amino acids and the remainder is used for the uptake of ammonia and biosynthesis of the individual amino acids. Likewise, as much as 93% of the total NADPH requirement and 75% of the NADH production is for the synthesis of amino acids destined for protein biosynthesis. If the cytosolit acetyl-CoA is assumed to be synthesized from pyruvate via acetate, YxATP,rrowlh remains at 39.9 mmol per g DW, whereas YxNADPHand YxNADH become 9.8 and 17.0 mmoi per g DW, respectively (Table 4). Finally, if cytosolic acetyl-CoA is transported across the mitochondrial membranes as a carnitine derivative, i.e., compartmentation is ignored *en ‘xATPgrowth is lowered significantly to 35.4 mmol per g DW with Y,,,,,, and YxNADH still being 9.8 and 17.0 mmol per g DW, respectively. This last figure for YxATP,growrh corresponds well with findings for other microorganisms: 34.7 mmol per g DW for E. co/i (Stouthamer, 1979) and 35.3 mmol per g DW for 5. cerevisiae (Verduyn et al., 1990). Table 4 summarizes the results. With the majority of the co-factor requirement being for protein synthesis it is obvious that the yield

1 8.00 0.02 I

I

157

I

0.04

I

0.06

0.06

0.10

Specific growlh rate [h”]

Fig. 4. YxATP,srow,,,. Y,,,,,, and YxNAoH as functions of the specific growth rate. The yields are calculated on the basis of the macromolecular composition listed in Table I. Lines represent best linear fits.

cOt?fficients

YxATP,growth~

yx,A,P,

and

YxNADH

3.3. Metabolic flux analysis On the basis of the recent review of Nielsen (1995) on the metabolic pathways of P. chrysogenum a detailed stoichiometric model has been set up for growth and product formation on a chemically defined medium with glucose as the growth limiting component. The model is listed in the Appendix and it resembles the model presented by Jorgensen et al. (1995a) for growth and product formation on a com-

ATP ( YxATP)and NADPH (Yx’,,,o,,)

requirements

Biosynthetic route to cytosolic acetyl-CoA

YrATP.growlh(mmot per g DW)

YxNADPH(mmot per g DW)

YtNAoH (mmol per g DW>

From citrate catalyzed by citrate lyase, recycling of oxalo-acetate via malate/pyntvate

39.9

7.5

14.8

From pyruvate via acetate

39.9

9.8

17.0

Transported to cytosol via the camitine carrier system

35.4

9.8

17.0

were carried out in the same way as for Table 3.

depend

on the protein content of the cell. Since the protein content of P. chrysogenun increases linearly with the specific growth rate (Fig. l), the three yield cofficients also increase linearly as functions of the specific growth rate (Fig. 4) an observation also of S. cerevisiae (Verduyn et made for YxATP,growth al., 1990).

Table 4 Influence of different biosynthetic routes to cytosolic acetyl-CoA on the theoretical NADH (Y xNADH) production for cell biosynthesis of P. chrysogenum ’

a The calculations

c

0.8

and

158

C.M. Henrikxn

et uI./Jourd

of Biotechnology 45 (1996) 149-164

plex medium. Simplifications have been made with respect to the uptake of substrates available in the cultivation medium. Furthermore, minor improvements have been introduced in the stoichiometric description of the macromolecular biosynthesis in order to make it more consistent with the review of Nielsen (1995). The model considers 72 biochemical fluxes and 77 metabolites. For glucose-limited growth on a chemically defined medium and in the applied range of specific growth rates no products besides biomass, penicillin and carbon dioxide are formed in significant amounts. Thus 64 metabolites can be assumed to be in pseudo-steady state. The degree of freedom is therefore 8, and the system can be made observable by measuring/calculating the net specific formation rates (mmol per g DW per h) of glucose, penicillin V and the six macromolecular components of the biomass, i.e., RNA, DNA, proteins, carbohydrates, amino carbohydrates and lipids.

9.0 Proteins, RNA / DNA

60.9 PP-pathway f-

=F 100.0 glucose-6-P

35.3

29.4 f fructose-fi-P

16.5 k

Christensen et al. (1995) characterized the applied high yielding strain of P. chrysogenum in continuous cultures with respect to its growth on glucose and its penicillin production. When continuous cultivation was initiated by feeding with medium containing the sidechain precursor phenoxyacetic acid a steady state was obtained after approx. five residence times. At this steady state a maximum specific penicillin productivity of about 24 pmol penicillin V per g DW per h could be maintained almost independently of the specific growth rate in the range 0.020.11 h- ’ (see also Fig. 7). After another two to five residence times the penicillin productivity dropped rapidly due to the formation of at least two non- or low-producing mutants. Finally, a second steady state was obtained at which a mixed culture of the mutants and the high yielding strain could be maintained. The growth of the fungus was furthermore characterized in terms of different yield and mainte-

9.7 ) carbohydrates 2.6 k amino carbohydrates

+ 62. I glyceraldebyde-3-P

t 140.3 3-pbosphoglycerrte t 125.0 phosphoenolpyruvate

15.2 __)

5.8 serine

, +

proteins

8. I

Fig. 5. Metabolic flux distribution in P. chrysogenum

during balanced growth at a specific growth rate of 0.

I

h

’

Calculations were based

on the stoichiometric model listed in the Appendix, the macromolecular

composition presented in Fig. I, and the specific rates of glucose

uptake and penicillin V production reported by Christensen et al. (1995).

All biochemical fluxes have been scaled with respect to the uptake

rate of glucose and refer to moles of the reactants disappearing. The specific uptake rate of glucose was

I .24

mmol per g DW per h.

CM.

Henriben

et nl./Journal

trance coefficients. The yield coefficient for biomass with respect to glucose (Y,, > was 0.51 g g- ’ , and the maintenance demand for glucose (m,) was 0.028 g g-’ h-l. From these data the steady-state biomass concentration can be calculated when the residual glucose concentration is known, and subsequently it can be used for calculating the net specific rates of glucose consumption and penicillin V production. The net specific formation rates of the macromolecular components are calculated from the measurements of stable RNA, proteins and carbohydrates presented in Fig. 1 and by assuming pools of lipids and DNA of 5% (w/w) and 1% (w/w), respectively. Furthermore, the pool of carbohydrates is assumed to consist of 75% glucan and 25% chitin as discussed previously. Finally, since the macromolecular pools can not account for the entire biomass, they have all subsequently been scaled by the same factor such that they sum to 92.2% (w/w> at all specific growth rates with the remaining 7.8% (w/w) being present as ash (Christensen et al., 1995). Fig. 5 displays the steady-state metabolic flux distribution in P. chrysogenum at a specific growth rate of 0.1 h-‘. The metabolic flux calculations show that a major part of the glucose enters the pentose phosphate (PP> pathway. The flux fpp through the PP pathway is as high as 61% of the total glycolytic flux which mainly reflects the high demand for NADPH used for biosynthesis of cysteine, a precursor for penicillin biosynthesis. As a consequence of the drain of precursor metabolites from the Embden-Meyerhof-Parnas (EMP) and the PP pathway only 60% of the glucose carbon ends up

Table 5 The relative flux distribution (Nielsen. 1995)

between the EMP and the PP pathway

Species

EMP (%)

Aspergillus niger

78

Penicillium

Penicillium

chrysogenum

digitalum

ofBiotechnology

45 (1996)

,c*

149-164

159

Specific growth rate [b’]

Fig. 6. Percentage of glucose entering the PP pathway (fpp) as a function of the specific growth rate. The calculations were carried out both at maximum specific penicillin productivity and at non-penicillin-producing conditions.

in pyruvate. 67% of the pyruvate enters the TCA cycle via acetyl-CoA, 16% is used as a precursor metabolite for the biosynthesis of amino acids ending up in either proteins or penicillin, and the remaining 17% is carboxylated into oxaloacetate in the mitochondria by pyruvate carboxylase (see also Fig. 31, thus demonstrating the importance of this anaplerotic reaction. TCA cycle intermediates are heavily drained due to the high requirement for precursor metabolites in the cytosol in the form of acetyl-CoA, oxaloacetate and cr-ketoglutarate. The requirement for cytosolic acetyl-CoA is only slightly higher than the requirement for oxaloacetate which leads to almost no recycling of cytosolic oxaloacetate back into the TCA cycle.

in batch cultivations

of P. chrysogenum and some closely related fungi

PP (%)

Conditions

Ref.

77

23

51 56-70

49

batch culture batch culture batch culture a

Shu et al. (1954) Wang et al. (1958) Heath and Koffler ( 1956) Lewis et al. (1954) Wang et al. (1958) Reed and Wang ( 1959)

83

17

77

23

batch culture batch culture

a The flux distribution was determined at different periods during a batch cultivation, and the relative flux through the PP pathway increased throughout the experiment. The listed figures am average values for the batch cultivation. The high relative flux through the PP pathway at the end of the batch cultivation could be due to conidiophore formation. Thus, both Ng et al. (1972) and Pitt and Mosley (1985) observed an increased activity of the PP pathway prior to conidiophore formation in, respectively, A. niger and P. noratum.

Ifi0

C.M. Henriksen

Ed d/Journal

Fig. 6 shows f,, as a function of the specific growth rate both at maximum specific penicillin productivity of the high yielding strain and at nonpenicillin-producing conditions. When no penicillin is produced fpp increases as a function of the specific growth rate which reflects the increasing biosynthetic demand for NADPH as depicted in Fig. 4 and the increasing need for precursor metabolites withdrawn from the PP pathway for the biosynthesis of nucleotides and certain amino acids. A similar increase m fpp with increasing specific growth rate has been determined experimentally in A. niduluns by the respirometric method (Carter and Bull, 1969). The calculated values for fpp at non-penicillin-producing conditions are comparable with values reported previously for P. chrysogenum and some closely related fungi (Table 5). At conditions with maximum specific penicillin production f,, becomes significantly higher than at non-penicillin-producing conditions (Fig. 61, and this is caused by the high demand for NADPH in reduction of sulphate for the biosynthesis of cysteine which eventually ends up in penicillin. Furthermore, f,, is highest at low specific growth rates, whereas it becomes almost constant at around 60% for higher specific growth rates, an effect caused by the growth rate independent specific productivity of penicillin, which has a more pronounced effect on the flux distribution at low specific growth rates. During a fed-batch cultivation fpp has been calculated to be about 45% (Jorgensen et al., 1995a), which is slightly lower than at low specific growth rates in the chemostat. However, the specific productivity of penicillin is also lower, about 13 pmol per g DW per h. From the metabolic flux analysis based on data from continuous cultures (Christensen et al., 1995) and on a fed-batch cultivation (Jorgensen et al., 1995a, b) it is possible to draw some conclusions concerning the potential bottlenecks for penicillin production by P. chrysogenum. Fig. 7 shows the specific productivity of penicillin (T& and the yield of penicillin on glucose (Y, ) as functions of the specific glucose uptake rate Pr,). For rs above 0.35 mmol per g DW per h rp is approx. constant, whereas Ysp decreases for increasing values of rs. At lower values of rr the specific productivity decreases, whereas the yield decreases slowly. A decrease in rp at low values of rS (or the specific

oj’Biotechdogy

45 (1996)

0.00-l l 0.00

0.25

149-164

0.50 rs [mmoles

0.75 per g DW

1.00

1.25

co.00 1.50

per h]

Fig. 7. Experimental data on the specific production of penicillin (rP) as a function of the specific uptake rate of glucose (rS). On the basis of a linear tit of rP the yield coefficient of penicillin on glucose (Y&j has been calculated. (B) Data from continuous cultures (taken from Christensen et al. (1995) supplemented with a few extra data points). (A) Data point from fed-batch cultivation (taken from J5rgensen et al. (1995a)).

growth rate in a steady-state chemostat) has also been observed by Ryu and Hospodka (1980). When rS approaches the maintenance requirements (approx. 0.15 mmol per g DW per h) the specific productivity drops to zero and hence the yield also becomes zero. From the flux analysis it is found that fpp increases for decreasing rS (see Fig. 6). This shows that there is a correlation between Ys,, and fpp. Thus a thermodynamic constraint at the glucose-6-phosphate node, which results in an upper limit for the relative flux through the PP pathway, may have a significant effect on the yield of penicillin on glucose. Since glucose accounts for about 25% of the costs for production of penicillin Ysp is an important process variable, and the process should therefore be operated at a maximum of Ysp. Another important process variable is the specific productivity (rp), which should be maximum to ensure an efficient utilization of the capital investment, i.e., the bioreactors. With the applied high yielding strain the optimum process conditions are therefore obtained with a specific glucose uptake rate of 0.35 mmol g DW _ ’ h _ ‘. However, this will result in slow growth of the biomass ( p - 0.02 h- ’ ) which may result in a high biomass concentration in a fed-batch process. This may cause limitations in the oxygen supply - an aspect that also has to be considered in design of the optimal process.

CM. Henriksen et al./Journal

of Biotechnology 45 (1996) 149-164

Appendix A

A.1. Stoichiometric

model for growth and product formation

of Penicillium

chrysogenum

The subscripts ‘cyt’ and ‘mit’ refer to the cytosolic and mitochondrial compartments, respectively.

Uptake reactions: 1. glucose-6-P - glucose - ATP = 0 2. NH;- NH: (ex.) - ATP = 0 3. H,S - SOi- (ex.) - 4 ATP - 4 NADPH = 0 Embden-Meyerhof-Pamas pathway: 4. fructose-6-P - glucose-6-P = 0 5. 2 glyceraldehyde-3-P - fructose-6-P - ATP = 0 6. 3-phosphoglycerate + ATP + NADH,,, - glyceraldehyde-3-P = 0 7. phosphoenolpyruvate - 3-phosphoglycerate = 0 8. pyruvate + ATP - phosphoenolpyruvate = 0 Pentose phosphate pathway: 9. ribose-5-P + 2 NADPH + CO, - glucose-6-P = 0 10. erythrose-4-P + fructose-6-P - 2 ribose-5-P = 0 11. glyceraldehyde-3-P + fructose-6-P - ribose-5-P - erythrose-4-P = 0 TCA cycle: 12. citrate + NADH,i, + CO, - pyruvate - oxaloacetate,it = 0 13. succinyl-CoA + 2 NADH,i, + 2 CO, - citrate = 0 14. malate + FADH, + ATP - succinyl-CoA = 0 15. oxaloacetate,i, + NADH,,, - malate = 0 Anaplerotic reactions: 16. oxaloacetate,i, - pyruvate - ATP - CO, = 0 17. acetyl-CoA + oxaloacetate,,, - citrate - ATP = 0 18. pyruvate + NADPH + CO, - oxaloacetate,yt - NADH,,, = 0 Electron transport/oxidative phosphorylation: 19. FADH, - NADH,,, = 0 20. 2.6 ATP - NADH,i, - 0.5 0, = 0 21. 1.6 ATP-FADH,-0.5 O,=O Biosynthesis of amino acids: 22. a-ketoglutarate + NADH,,, + CO, - citrate = 0 23. glutamate - a-ketoglutarate - NH: - NADPH = 0 24. glutamine - glutamate - NH: - ATP = 0 25. proline - glutamate - ATP - 2 NADPH = 0 26. cu-aminoadipate + NADH,,, + CO, - acetyl-CoA - glutamate = 0 27. lysine + a-ketoglutarate + NADH,,, - cz-aminoadipate - glutamate - 2 ATP - 2 NADPH = 0 28. arginine + malate + cr-ketoglutarate - 2 glutamate - aspartate - CO, - NH: - 7 ATP - NADPH = 0 29. serine + cY-ketoglutarate + NADH,,, - 3-phosphoglycerate - glutamate = 0 30. cysteine + a-ketobutyrate + NH: - homocysteine - serine = 0 31. glycine + N5, N lo-methylene THF - serine = 0

161

162

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

C.M. Henriksen et ul./Jnurnul

oj’Biotechmlogy

45 (1996) 149-164

aspartate + a-ketoglutarate - oxaloacetate,,, - glutamate = 0 asparagine + glutamate - aspartate - glutamine - 2 ATP = 0 homoserine - aspartate - ATP - 2 NADPH = 0 threonine - homoserine - ATP = 0 homocysteine - homoserine - H,S - 2 ATP = 0 methionine - homocysteine - N5-methyl THF = 0 cz-ketoisovalerate + CO, - 2 pyruvate - NADPH = 0 valine + cr-ketoglutarate - cr-ketoisovalerate - glutamate = 0 leucine + cr-ketoglutarate + NADH,,, + CO, - cu-ketoisovalerate - acetyl-CoA - glutamate = 0 alanine + cr-ketoglutarate - pyruvate - glutamate = 0 isoleucine + wketoglutarate + NH: + CO, - threonine - glutamate - pyruvate - NADPH = 0 chorismate - 2 phosphoenolpyruvate - erythrose-4-P - ATP - NADPH = 0 tryptophan + glutamate + pyruvate + glyceraldehyde-3-P + CO, - chorismate - glutamine - serine PRPP=O 45. prephenate - chorismate = 0 46. tyrosine + a-ketoglutarate + NADH,,, + CO, - prephenate - glutamate = 0 47. phenylalanine + cu-ketoglutarate + CO, - prephenate - glutamate = 0 48. histidine + cru-ketoglutarate + AICAR + 2 NADH,,, - PRPP - adenosine-triphosphate - glutamine = 0 Biosynthesis of nucleotides: 49. AICAR + malate + 2 glutamate - PRPP - glycine - aspartate - 2 glutamine - N”-formyl THF - CO, 4ATP=O 50. IMP - AICAR - N “-formy THF = r) 5 1. adenosine-triphosphate + malate - ‘MP - aspartate - 3 ATP = 0 52. GTP + glutamate + NADH,,, - IMP - glutamine - 4 ATP = 0 53. UTP + glutamate + NADH,,, - PRPP - aspartate - glutamine - 4 ATP = 0 54. CTP + glutamate - UTP - glutamine - ATP = 0 Biosynthesis of macromolecules: 55. RNA - 0.256 adenosine-triphosphate - 0.286 GTP - 0.262 UTP - 0.196 CTP - 0.4 ATP = 0 56. DNA - 0.24 adenosine-triphosphate - 0.24 UTP - 0.26 GTP - 0.26 CTP - 0.24 N’, N “-methylene THF -0.4ATPl.?4NADPH=O 57. lipid + 1.I16 homocysteine + 0.595 CO, - 20.156 acetyl-CoA - 0.149 glucose-6-P - 0.814 glyceraldehyde-3-P - 1.116 methionine - 0.595 serine - 1.403 0, - 23.106 ATP - 32.260 NADPH - 3.472 NADH,,, = 0 58. protein - 0.100 alanine - 0.048 arginine - 0.048 asparagine - 0.048 aspartate - 0.014 cysteine - 0.075 glutamate - 0.075 glutamine - 0.092 glycine - 0.024 histidine - 0.043 isoleucine - 0.075 leucine - 0.056 lysine - 0.017 methionine - 0.034 phenylalanine - 0.047 proline - 0.061 serine - 0.053 threonine - 0.026 tyrosine - 0.064avaline - 4.3 ATP = 0 59. chitin + glutamate - fructose-6-P - glutamine - acetyl-CoA - ATP = 0 60. glucan - glucose-6-P - ATP = 0 Biosynthesis of penicillin: 61. ACV - valine - cysteine - cy-aminoadipate - 6 ATP = 0 62. isopenicillin N - ACV - 0, = 0 63. penicillin V + cu-aminoadipate - isopenicillin N - phenoxyacetic acid - 2 ATP = 0 64. 6-APA + a-aminoadipate - isopenicillin N = 0 65. 8-HPA - 6-APA - CO, = 0 66. OPC - a-aminoadipate = 0 Miscellaneous reactions: 67. N “-methyl THF - N *, N lo-methylene THF - NADH cy, = 0

CM. Henriksen et al./Journal

68. 69. 70. 71. 72.

of Biotechnology 45 (1996) 149-164

163

sN”-formyl THF + NADPH - N5,N lo-methylene THF = 0 N5, N lo-methylene THF + NH: + CO, + NADH,,, - glycine = 0 PRPP - ribose-5-P - 2 ATP = 0 succinyl-CoA + NADH,,, - a-ketobutyrate - ATP = 0 -ATP=O

References Alberghina, L., Sturani, E.. Costantini, M.G., Martegani, E. and Zippel, R. (1979) Regulation of macromolecular composition during growth of Neurospora crassa. In: Burnett, J.H. and Trinci, A.P.J. @is.), Fungal Walls and Hyphal Growth. Cambridge University Press, Cambridge, UK, pp. 295-318. Barkholt, V. and Jensen, A.L. (1989) Amino acid analysis: Determination of cysteine plus half-cystine in proteins after hydrochloric acid hydrolysis with a disulfide compound as additive. Anal. Biochem. 177, 318-322. Bartnicki-Garcia, S. (1968) Cell wall chemistry, morphogenesis and taxonomy of fungi. Annu. Rev. Microbial. 22, 87- 108. Benthin, S., Nielsen, J. and Villadsen, J. (1991) A simple and reliable method for the determination of cellular RNA content. Biotechnol. Technol. 5(l), 39-42. Benthin, S., Schulze, U., Nielsen, J. and Villadsen, J. (1994) Growth energetics of L.ucrococcus cremoris FDI during energy-, carbon-, and nitrogen-limitation in steady state and transient cultutes. Chem. Eng. Sci. 49, 589-609. Berry, D.R. and Berry, E.A. (1976) Nucleic acid and protein synthesis in tilamentous fungi. In: Smith, J.E. and Berry, D.R. (Eds.), The Filamentous Fungi, Vol. II. Edward Arnold, London, UK, pp. 238-291. Blumenthal, H.J. (1976) Reserve carbohydrates in fungi. In: Smith, J.E. and Berry, D.R. (Eds.), The Filamentous Fungi, Vol. II. Edward Arnold, London, UK, pp. 292-307. Boulton, C.A. and Ratledge, C. (1981) Correlation of lipid accumulation in yeasts with possesion of ATPcitrate lyase. J. Gen. Microbial. 127, 169-176. BR.M~, P.J. and L&l, D.M. (1978) Physiology of fungal lipids: selected topics. In: Rose, A.H. and Morris, J.G. (Eds.), Advances in Microbial Physiology, Vol. 17. Academic Press, London, UK, pp. 47-179. Bushell, M.E. and Bull, A.T. (1976) Growth rate. dependent ribosomal efficiency of protein synthesis in the fungus Aspergillus nidulans. J. Appl. Chem. Bioteehnol. 26, 339-340. Carlsen, M. (I 994) a-Amylase production by Aspergillus oryzae. PhD thesis, The Technical University of Denmark, Lyngby, Denmark. Carter, B.L.A. and Bull, A.T. (1%9) Studies of fungal growth and intermediary carbon metabolism under steady state and nonsteady state conditions. Biotechnol. Bioeng. 11, 785-804. Christensen, L.H., Nielsen, J. and Villadsen, J. (1991) Delay and dispersion in an in-situ membrane probe for bioreactors. Chem. Eng. Sci. 46, 3304-3307. Christensen, L.H., Henriksen, CM., Nielsen, J., Villadsen, J. and Egel-Mitani, M. (1995) Continuous cultivation of P. chryso-

genus. Growth on glucose and penicilhn production. J. Biotechnol. 42, 95-107. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A. and Smith, F. ( 1956) Calorimetric method for determination of sugars and related substances. Anal. Chem. 28(3), 350-356. Folch, J.. Lees, M. and Stanley, G.H.S. (1957) A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226, 497-509. Hamilton, P.B. and Knight, S.G. (1%2) An analysis of the cell walls of Penicillium chrysogenum. Arch. Biochem. Biophys. 99, 282-287. Heath, E.C. and Koffler, H. (1956) Biochemistry of filamentous fungi. II. The quantitative significance of an “oxidative pathway” during the growth of Penicillium chrysogenum. J. Bacteriol. 71, 174-181. Herbert, D., Phipps, PJ. and Strange, R.E. (1971) Chemical analysis of microbial cells. In: Norris, J.R. and Ribbons, D.W. @Is.), Methods in Microbiology, Vol. 5B. Academic Press, London, UK, pp. 209-344. Holdsworth, J.E., Veenhuis, M. and Ratledge, C. (1988) Enzyme activities in oleaginous yeast accumulating and utilizing exogenous or endogenous lipids. J. Gen. Microbial. 134, 29072915. Ingraham, J.L., Maaloe, 0. and Neidhardt, F.C. (1983) Growth of the Bacterial Cell. Sinnauer Ass., Sunderland. Jakhtsch, W.M., Hampel, W., Riihr, M., Kubicek, C.P. and Gamerith, G. (1986) a-Aminoadipate pool concentration and penicillin biosynthesis in strains of Penicillium chrysogenum. Can. J. Microbial. 32, 473-480. Jemejc, K., Perdih, A. and Cimerman, A. (1991) ATP:citrate lyase and camitine acetyltransferase activity in a citric acid producing Aspergillus niger strain. Appl. Microbial. Biotechnol. 36, 92-95. Jorgensen, H., Nielsen, J. and Villadsen, J. (1995a) Metabolic flux distribution in Penicillium chtysogenum during fed-batch cultivations. Biotechnol. Bioeng. 46, 117- 13 1. Jorgensen, H., Nielsen, J., Villadsen, J. and Mollgarvd, H. (1995b) Analysis of penicillin V biosynthesis during fed-batch cultivations with a high-yielding strain of Penicillium chrysogenum. Appl. Microbial. Biotechnol. 43, 123- 130. Kohlhaw, G.B. and Tan-Wilson, A. (1977) Camitine acetyltransferase: Candidate for the transfer of acetyl groups through the mitochondrial membrane of yeast. J. Bacterial. 129, I1591161. Kiienzi, M.T. and Fiechter, A. (1972) Regulation of carbohydrate composition of Succharomyces cereuisiae under growth limitation. Arch. Microbial. 84, 254-265. Lewis, K.F., Blumenthal, H.J., Wenner, C.E. and Weinhouse, S.

164

C.M. Henriksen er ul./Journul

(1954) Estimation of glucose catabolism pathways. Fed. Proc. 13, 252. Malaney, G.W., Tanner, R.D., Shipp. E.A. and Slonim, A.E. (1989) The effects of high sodium chloride in the growth medium on the content of the essential amino acids in the intracellular pool in Baker’s yeast during aerobic fermentation of glucose. Food Microbial. 6, l-6. Mason, H.R.S. and Righelato, R.C. (1976) Energetics of fungal growth: the effect of growth-limiting substrate on respiration of Penicillium chrysogenum. J. Appl. Chem. Biotechnol. 26, 145- 152. Meisgeier, G., Miiller, H., Ruhland, G. and Christner, A. (1990) Qualitative und quantitative Zusammensetzung der Fetts&nespektren von Selektanten des Penicillium chrysogenum. Zentralbl. Mikrobiol. 145, 183-186. M%hlBn, A. ( 1973) Purification and some properties of ATP citrate lyase from Penicillium spiculisporum. Eur. J. B&hem. 36, 342-346. Ng, W.S., Smith, J.E. and Anderson, J.G. (1972) Changes in carbon catabolic pathways during synchronous development of conidiophores of Aspergillus niger. J. Gen. Microbial. 71, 495-504. Nielsen, J. and Villadsen, J. (1993) Bioreactors: Description and modelling. In: Stephanopoulos, G. @d.), Biotechnology, Vol. 3 (2nd edn.). VCH, Weinheim, Germany, pp. 77- 104. Nielsen, J.. Johansen, C.L. and Villadsen, J. (1994) Culture fluorescence measurements during batch and fed-batch cultivations with Penicillium chrysogenum. J. Biotechnol. 38,5 l-62. Nielsen, J. and Villadsen, J. (1994) Biomaction Engineering Principles. Plenum Press, New York. Nielsen, J. (1995) Physiological Engineering Aspects of Penicillum chrysogenum. Dissertation. The Technical University of Denmark, Lyngby. Pfizner, A., Kubicek, C.P. and Riihr, M. (1987) Presence and regulation of ATPeitrate lyase from the citric acid producing fungus Aspergillus niger. Arch. Microbial. 147, 88-91. Pitt, D.E. and Bull, A.T. (1982) Influence of culture conditions on the physiology and composition of Trichodermu aureooiride. J. Gen. Microbial. 128, 1517-1527. Pitt, D. and Mosley, M.J. (1985) Pathways of glucose catabolism and the origin and metabolism of pyruvate during calcium-induced conidiation of Penicillium notatum. Antonie van Leeuwenhoek 5 1, 365-384. Ratledge, C. and Evans, C.T. (1989) Lipids and their metabolism. In: Rose, A.H. and Harrison, J.S. (Eds.), The Yeasts, Vol. 3. Academic Press, London, UK, pp. 367-455. Reed, D.J. and Wang, C.H. (1959) Glucose metabolism in Penicillium digifatum. Can. J. Microbial. 5, 59-66.

oj’Biotechnology

45 (1996) 149-164

Righelato, R.C., Trinci, A.P.J. and Pitt, S.J. (1968) The influence of maintenance energy and growth rate on the metabolic activity, morphology and conidiation of Penicillium chrysogenum. J. Gen. Microbial. 50, 399-412. Ryu, D.D.Y. and Hospodka, J. (1980) Quantitative physiology of Penicillium chrysogenum in penicillin fermentation. Biotechnol. Bioeng. 22, 289-298. Schulze, U., Larsen, M.E. and Villadsen, J. (1995) Determination of intracellular trehalose and glycogen in Sacchuromyces cereuisiur. Anal. B&hem. 228, 143- 149. Shu, P., Funk, A. and Neish, A.C. (1954) Mechanism of citric acid formation from glucose by Aspergillus niger. Can. J. B&hem. Physiol. 32, 68-80. Stagg, C.M. and Feather, M.S. (1973) The characterization of a chitin-associated D-glucan from the cell walls of Aspergilfus niger. B&him. Biophys. Acta 320, 64-72. Stewart, P.R. (1975) Analytical methods for yeast. In: Prescott, D.M. (Ed.), Methods in Cell Biology, Vol. 12. Academic Press, New York, pp. 111-147. Stouthamer, A.H. (1979) The search for correlation between theoretical and experimental growth yields. In: Quayle, J.R. @d.), International Review of Biochemistry, Vol. 21, Microbial biochemistry. University Park Press, Baltimore, MD, pp. l-47. Stouthamer, A.H. and Bettenhaussen, C. (1973) Utilization of energy for growth and maintenance in continuous and batch cultures of microorganisms. Biochim. Biophys. Acta 301, 53-70. Sturani, E., Magnani, F. and Alberghina, F.A.M. (1973) Inhibition of ribosomal RNA synthesis during a shift down transition of growth in Neurospora crasser. B&hem. Biophys. Acta 319, 153-164. Verduyn, C., Postma, E., Scheffers, W.A. and van Dijken, J.P. ( 1990)Energetics of Sacchoromyces cereuisiue in anaerobic glucose-limited chemostat cultures. J. Gen. Microbial. 136, 405-412. Wang, C.H., Stem, I., Gilmour, C.M., Klungsoyr, S., Reed, D.J., Bialy, J.J., Christensen, B.E. and Cheldelin, V.H. (1958) Comparitive study of glucose catabolism by the respirometric method. J. Bacterial. 76, 207-216. Watson, T.G. (1976) Amino-acid pool composition of Succhuromyces cereuisiae as a function of growth rate and amino-acid nitrogen source. J. Gen. Microbial. 96, 263-268. Woodin, T.S. and Wang, J.L. (1987) Effect of temperature on sulphate transport in two Penicillia. J. Membr. Sci. 33, 221240.