INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion Appendix References

In the production of industrial enzymes, filamentous fungi are attractive as cell factories because they can secrete large amounts of protein into the medium. Presently, many important industrial enzymes such as lipases, cellulase, and amylases are produced by fermentation of filamentous fungi, and there also has been a report on the use of Aspergillus oryzae for production of mammalian proteins (14). Despite the industrial importance of these organisms, relatively little is known about their basic physiology. For example, for A. oryzae, information about which pathways are present and active under various growth conditions is limited. Such information is quite important for the optimization of a fermentation process, especially if the fluxes in wild-type cells and overproducing cells are compared, since it quantifies the drain of precursors from the central carbon metabolism to protein biosynthesis and the requirements for cofactors like NADPH for protein synthesis.

Metabolic fluxes may be calculated from measurements of fluxes into and out of the cell and metabolite mass balances (18, 28, 36, 43, 47). A more robust estimate of the metabolic fluxes is obtained from experiments with ¹³C-enriched glucose in which either the fractional enrichment of ¹³C in cellular metabolites or the isotopomer distributions of the metabolites are quantified by using nuclear magnetic resonance (31). However, this technique is technically complex, and the simple approach based on metabolite balancing is often preferred.

Our first objective was to compare a wild-type strain and an -amylase-overproducing strain of A. oryzae with respect to metabolic fluxes and enzyme activities. In order to calculate the metabolic fluxes, we developed a metabolic flux model. Our second objective was to investigate the influence of different nitrogen sources on the metabolic fluxes and enzyme activities. The working hypothesis tested was whether the metabolic fluxes in the two strains are the same and whether the NADPH supply influences the -amylase productivity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion Appendix References

Symbols. The following symbols are used for the metabolic flux analysis: G (J by Q), a stoichiometric matrix; F (= 10), degrees of freedom; g_ji (no dimension), stoichiometric coefficient for compound i in reaction j; J (= 69), number of reactions; Q (= 59), number of intracellular metabolites; X_i (moles gram [dry weight]¹), concentration of the intracellular compound i; µ (hour¹), specific growth rate; and v_j (moles gram [dry weight]¹ hour¹), rate of reaction j.

Microorganism. Two strains of A. oryzae were used. A1560 is a wild-type strain, and CF 1.1 is a recombinant strain containing additional copies of the -amylase gene (11, 46). Both strains were provided by Novo Nordisk A/S.

Media. The chemically defined medium for continuous cultivation with ammonia as a nitrogen source contained (per liter) 4 g of glucose · H₂O, 2.5 g of (NH₄)₂SO₄, 0.75 g of KH₂PO₄, 1 g of MgSO₄ · 7H₂O, 1 g of NaCl, 0.1 g of CaCl₂ · 2H₂O, 0.25 ml of antifoam (Struktol SB2121; Schill & Seilacher, Hamburg, Germany), 7.2 mg of ZnSO₄ · 7H₂O, 1.3 mg of CuSO₄ · 5H₂O, 0.3 mg of NiCl₂ · 6H₂O, and 6.9 mg of FeSO₄ · 7H₂O. When nitrate was used as the nitrogen source, (NH₄)₂SO₄ was replaced with 3.2 g of NaNO₃ per liter and 2.7 g of Na₂SO₄ per liter.

Cultivation conditions. The cultivations were carried out as constant-mass chemostats in an MBR bioreactor (11). The temperature was 30°C, and the pH was kept at 6.0 by adding 4 M NaOH or 2 M H₂SO₄. The headspace pressure was 1.5 × 10⁵ Pa, and the aeration was 0.4 to 0.9 liter liter¹ min¹, depending on the specific growth rate. The cultures were considered to be in physiological steady state when the biomass concentration, the CO₂ evolution rate, and the extracellular -amylase activity were stable for at least three residence times.

Sampling. For biomass measurements, the samples were filtered, washed with 0.9% (wt/vol) NaCl, and dried at 105°C for 24 to 48 h. For determination of the macromolecular composition of the biomass, the sample was filtered, washed with 0.9% (wt/vol) NaCl, and freeze-dried. After the freeze-drying, the cells were crushed in a mortar and kept dry in a desiccator. For determination of in vitro enzyme activities, the sample was filtered, washed with cold 0.9% (wt/vol) NaCl, and kept at 80°C until assayed (not longer than 3 months).

Analysis of extracellular -amylase. Extracellular -amylase activity was measured by a flow injection analysis system based on monitoring of the decolorization of an iodine-starch complex (10).

Analysis of carbohydrates. Freeze-dried biomass was disrupted in distilled water with a mixer mill (Retsch MM2) for 2 h at 4°C with 0.75- to 1.0-mm-diameter glass beads. The sample was analyzed by the phenol method (19). The reported values for carbohydrate content have been corrected for the contributions from RNA and DNA.

Analysis of DNA. DNA was measured by the diphenylamine reagent method with salmon testis DNA as a standard (19). The biomass used had been freeze-dried.

Analysis of RNA. RNA was determined by the Schmidt-Tannhauser method (3). An extinction coefficient of 10,800 M¹ cm¹ and an average molecular mass of nucleotides of 340 g mol¹ was used. The biomass used had been freeze-dried.

Analysis of protein, free amino acids, and chitin. Freeze-dried biomass was disrupted in distilled water in a mixer mill (Retsch MM2) for 45 min at 4°C with 0.75- to 1.0-mm-diameter glass beads. Samples were hydrolyzed with HCl (6 M) for 6, 24, and 72 h at 110°C. For analysis of free amino acids, the biomass was boiled in water for 15 min. The extract was centrifuged at 10,000 × g, and the supernatant was analyzed by high-pressure liquid chromatography (HPLC) for amino acid composition (2). Glutamine could not be distinguished from glutamic acid, and aspargine could not be distinguished from aspartic acid; therefore, the molar fractions of glutamine/glutamic acid and aspargine/aspartic acid were assumed to be 1. The nitrogen content originating from protein and chitin was calculated by subtracting the nitrogen content in RNA, DNA, and free amino acids from the total nitrogen content measured by the Kjeldahl method. The ratio of the amounts of chitin and protein was determined by HPLC analysis.

Analysis of glycerol and mannitol. Freeze-dried biomass was suspended in 0.5 mM H₂SO₄ and boiled in water for 15 min. The cells were removed by centrifugation for 10 min at 5,000 × g, and the supernatant was analyzed by HPLC with an HPX-87H Aminex ion-exclusion column (36). The column was eluted at 60°C with 5 mM H₂SO₄ at a flow rate of 0.6 ml min¹. The detector was a Waters 410 refractive index detector.

Analysis of lipids. Lipids were extracted with chloroform and methanol and determined gravimetrically (16).

Analysis of ash. Freeze-dried biomass was heated at 600°C for 15 h, and the ash content was measured by weighing the remains.

Preparation of cell extracts for measurement of total enzyme activities. A frozen sample was crushed in a mortar under liquid nitrogen. Thereafter, 0.1 g (wet weight) of mycelium was suspended in 1.5 ml of 0.1 M sodium phosphate (pH 7.4)-5 mM MgCl₂-1 mM EDTA. To the suspension was added 0.15 ml of 26 mM dithiothreitol, and the suspension was crushed in a mixer mill (Retsch MM2) with 0.75- to 1.0-mm-diameter glass beads for 15 min at 4°C.

Preparation of cell extracts for separation of mitochondria and cytosol. We prepared cell extracts by nitrogen cavitation (26) from cells grown on glucose and ammonia. The mycelia were filtered and washed with 0.9% (wt/vol) NaCl. Subsequently, 0.6 g (dry weight) of mycelia was suspended in 50 ml of 0.2 M sodium phosphate (pH 6.4) containing 0.6 M mannitol and 50 mg of Novozyme 234 (Sigma, St. Louis, Mo.). The suspension was incubated at room temperature (22 to 25°C) for 30 min with stirring. The mycelia were filtered through gauze and washed with 50 ml of ice-cold sodium phosphate-mannitol buffer. From this point on, the mycelium-cell extract was kept at 4°C. The mycelium was suspended in 50 ml of 0.1 M phosphate (pH 7.4) containing 0.25 M sucrose, 1 mM EDTA, 1 mM MgCl₂, 1 mM CaCl₂, 0.1 mM dithiothreitol, 8 µM phenylmethylsulfonyl fluoride, and 0.02 to 0.2% (wt/vol) bovine serum albumin (the cavitation buffer). Nitrogen pressure (10⁶ Pa) was applied to the suspension for 30 min and then quickly released. The broken mycelium was filtered through two layers of gauze, and the filtrate was centrifuged at 700 × g for 10 min to remove cell debris. A mitochondrial pellet was obtained by centrifuging the supernatant at 45,000 × g for 45 min. The pellet was washed with the cavitation buffer and then resuspended in 4 ml of the cavitation buffer. The 45,000 × g supernatant represented the cytosolic fraction. Since the amount of extracted protein was rather low, two cavitation experiments were carried out in parallel, one experiment with 0.02% (wt/vol) bovine serum albumin in the cavitation buffer and the other with 0.2% (wt/vol) bovine serum albumin in the cavitation buffer. These experiments were used for measurement of the amount of extracted protein and the enzyme activities, respectively. Measured protein concentrations were corrected for the bovine serum albumin in the cavitation buffer.

Analysis of in vitro enzyme activities. Enzyme assays were performed at 30°C with an HP8453 UV-visible light spectrophotometer. In all cases enzyme activities were proportional to the amount of cell extract added. Enzyme activities are expressed as micromoles minute¹ milligram¹. Protein was determined by the Bradford method (7) with bovine serum albumin as the standard. All specific enzyme activities were reduced by a factor of 2, since bovine serum albumin gives twice as high a signal as most other proteins in the Bradford assay. The enzyme activities were measured by using the reaction mixtures listed in Table 1.

Isoelectric focusing and gel staining. Isoelectric focusing was carried out on polyacrylamide gels (Ampholine PAGplate pI 4-6.5 and pI 3-10; Pharmacia Biotech, Uppsala, Sweden) under the conditions recommended by the manufacturer. The gels were washed with distilled water and stained for isocitrate dehydrogenase (NADP) by overlaying with a 0.8% (wt/vol) agarose gel containing 0.3 mg of thiazolyl blue tetrazolium bromide ml¹, 0.05 mg of N-methylphenazonium methyl sulfate ml¹, 0.5 mg of isocitrate ml¹, and 0.5 mg of NADP⁺ ml¹. For staining for total protein, the gels were fixed in trichloroacetic acid and stained with Coomassie brilliant blue.

Metabolic flux analysis. Calculations of metabolic fluxes by metabolite balancing are based on a metabolic model normally represented as a stoichiometric matrix, G. In this matrix the rows represent reactions and the columns represents compounds. Thus, the stoichiometric coefficient for the compound i participating in reaction j is positioned as element g_ji in the stoichiometric matrix. With the rates (or fluxes) of reaction j given by v_j, the mass balance for the intracellular compound i becomes:

(1)

where X_i is the intracellular concentration of compound i and µ is the specific growth rate of the biomass. The last term accounts for dilution due to the cell growth and is generally negligible for intracellular metabolites. Furthermore, a pseudo-steady-state assumption can often be used for intracellular metabolites, and equation 1 then reduces to:

(2)

Equation 2 represents a set of algebraic equations (one for each intracellular metabolite) that relate the intracellular fluxes. If some of the fluxes can be measured, e.g., the flux of glucose into the cell or the flux of precursors incorporated into the biomass, then it may be possible to calculate the nonmeasured fluxes by using the algebraic equations represented by equation 2. If the total number of reactions included in the stoichiometric model is J and the number of intracellular metabolites in the pseudo-steady-state is Q, then the number of degrees of freedom is F = J  Q; i.e., F fluxes must be measured.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion Appendix References

Enzyme activity measurements. We analyzed several key enzymes to determine if these reactions were active and what the cofactor requirements were.

Intracellular localization of enzymes. Since intracellular compartmentation plays an important role in how the metabolic fluxes are distributed, eight enzymes were localized to the mitochondrial and cytosolic fractions (Table 2). The mitochondrial matrix citrate synthase is localized almost entirely in the mitochondrial fraction (80% of total activity). The activity found in the cytosolic fraction is probably from the breakage of some mitochondria during the separation.

Biomass composition and -amylase productivities. The contents of the main biomass components were determined at different specific growth rates (Table 3). The measured biomass components could account for approximately 85% (wt/wt) of the biomass. The biomass contains approximately 10% (wt/wt) ash. The remaining 5% (wt/wt) can be accounted for by metabolites. For the flux analysis the entire biomass must be accounted for, so we normalized the measured composition to 100% (wt/wt).

1

1

View this table: [in this window] [in a new window]   TABLE 4.   Specific glucose uptake rates, specific -amylase productivities, and -amylase yields in the cultivations

Metabolic model. We developed a stoichiometric model based on literature information (4, 5, 8, 9, 12, 13, 17, 20, 21, 25, 27, 29, 30, 32, 35, 38-42, 45, 48-50), the enzymatic activities measured in this study, and the measured localizations of various enzymes. The metabolic model contains 69 reactions and 59 intracellular metabolites. Since each intracellular metabolite gives a mass balance, 10 fluxes must be specified to calculate the reaction rates. These 10 fluxes are determined from measurements of the eight biomass components, glucose, and -amylase. When the stoichiometry for production of a biomass component is known, the flux to this component is calculated by using the productivity (millimoles gram [dry weight]¹ hour¹) of the component, i.e., the specific growth rate (hour¹) times the fraction (grams gram [dry weight]¹) of the component in the biomass divided by the molar weight (grams millimole¹). To determine the stoichiometry for macromolecular synthesis, the compositions of the compounds must be known. In our analysis the composition of the proteins was taken as the average of eight measured protein compositions from the recombinant strain grown at specific growth rates of 0.05 to 0.15 h¹. The protein composition and content includes intracellular -amylase, which accounts for less than 2% of the total cellular protein in the wild-type strain. In eucaryotes most of the central carbon metabolism takes place either in the cytosol or in the mitochondria, and the metabolic model includes the major metabolic reactions occurring in these two compartments. This compartmentalization is necessary to obtain reliable balances of all components that do not pass the mitochondrial membrane. NADH and NADPH cannot pass the mitochondrial membrane, and separate balances are needed for these cofactors in each compartment. Since several biosynthetic reactions in the mitochondria require NADPH, this cofactor must be synthesized in this compartment, presumably by the mitochondrial NADP⁺-linked isocitrate dehydrogenase. Due to the compartmentalization of isocitrate dehydrogenase (NADP), malate dehydrogenase and glutamate oxaloacetate transaminase are included twice in the model, since these enzymes are present in both the cytosolic and the mitochondrial compartments. The reactions catalyzed by glucose-6-phosphate dehydrogenase and glutamate dehydrogenase (NADP) are included only once, since they are present only in the cytosol, as are citrate synthase and isocitrate dehydrogenase (NAD), which are present only in the mitochondria.

Flux calculations. Fluxes were calculated for four different scenarios (Table 5). We used the flux through the pentose phosphate pathway as an indicator for the change in fluxes. The flux through the pentose phosphate pathway is approximately constant in all four cases. We interpreted this result to mean that these pathways do not significantly alter the requirement for cytosolic NADPH. Therefore, we concluded that the assumptions regarding the biochemistry specified in Table 5 do not have much influence on the calculated flux distribution through the pentose phosphate pathway and the Embden-Meyerhof-Parnas pathway, and all further calculations were based on case a (see Table 5 and Appendix). If nitrate is the nitrogen source instead of ammonia, the calculated flux through the pentose phosphate pathway increases due to the large requirement for cytosolic NADPH in the reduction of nitrate to ammonia.

View this table: [in this window] [in a new window]   TABLE 5.   Relative fluxes through the pentose phosphate pathway with different assumptions regarding the biochemistry for growth of the recombinant strain at a specific growth rate of 0.094 h1 with ammonia or nitrate as the nitrogen source

View larger version (34K): [in this window] [in a new window]   FIG. 1.   Flux distributions of A. oryzae wild-type and recombinant strains during growth on glucose and ammonia at specific growth rates of 0.10 and 0.094 h1, respectively. For each reaction, the numbers to the left are for the wild-type strain and the numbers to the right (italic) are for the recombinant strain. All fluxes are normalized with respect to the glucose uptake rate. Solid lines represent chemical reactions, and dashed lines represent transport across the mitochondrial membrane. It was assumed that cytosolic isocitrate dehydrogenase (NADP) is active, that -ketoglutarate does not pass the mitochondrial membrane, and that oxaloacetate is transported from the cytosol to the mitochondria as malate. P, phosphate; acetyl-CoA, acetyl coenzyme A.

View larger version (13K): [in this window] [in a new window]   FIG. 2.   Relative fluxes through the pentose phosphate pathway at various specific growth rates for the wild-type strain () and the recombinant strain (). The standard deviations at 0.05 and 0.10 h1 are 6 and 3%, respectively, for the wild-type strain. All other data points are based on one chemostat cultivation. P, phosphate.

Comparison of in vitro enzyme activities with calculated fluxes. We compared the measured in vitro activities with the calculated fluxes (Table 6). Since the measured enzyme activities are maximum activities in vitro, the determined activities are measures of only the amount of enzyme present in the cell and not the in vivo activity.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion Appendix References

Enzyme measurements. We found the maximum specific activities of glutamate dehydrogenase (NADP) to be similar in the two strains and to increase with increasing flux catalyzed by this enzyme when ammonia was used as the nitrogen source. This is similar to results for Saccharomyces cerevisiae (36). When we used nitrate as the nitrogen source, the maximum specific activity was approximately 32 to 100 times higher than the calculated flux (it was 4 to 8 times higher than the calculated flux for ammonia as the nitrogen source). This effect might be due to a lower intracellular ammonia concentration during growth on nitrate compared to that on ammonia, which would require a high level of glutamate dehydrogenase (NADP) in the cell in order to maintain a certain flux to glutamate.

Biomass composition. We found that the protein and RNA contents increased with specific growth rate. Both the wild type and the recombinant strain have approximately the same levels of intracellular protein and RNA, even though the recombinant strain produces approximately 24% more total protein (intracellular protein plus extracellular -amylase) than the wild-type strain.

1

Flux analysis. The major branch points in the central metabolism where the fluxes are affected by changes in nitrogen source, specific growth rate, and -amylase productivity are at the glucose-6-phosphate node and the flux catalyzed by isocitrate dehydrogenase (NAD). We calculated that the fraction of the glucose uptake diverted through the pentose phosphate pathway increased with increasing specific growth rate and -amylase productivity. This effect is caused by the high NADPH demand for synthesis of proteins. Furthermore, at the pyruvate node we calculated an increased flux catalyzed by pyruvate carboxylase and a decreased tricarboxylic acid cycle flux in the recombinant strain compared with the wild-type strain. These changes are caused by the increased requirement for amino acid biosynthesis in the recombinant strain compared with the wild-type strain.

1

1

1

1

1