AEM
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.
Agricola
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, May 2005, p. 2294-2302, Vol. 71, No. 5
0099-2240/05/$08.00+0     doi:10.1128/AEM.71.5.2294-2302.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Metabolic Network Analysis of Streptomyces tenebrarius, a Streptomyces Species with an Active Entner-Doudoroff Pathway

Irina Borodina,1 Charlotte Schöller,2 Anna Eliasson,1 and Jens Nielsen1*

Center for Microbial Biotechnology, BioCentrum-DTU, Building 223, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark,1 Scientific Affairs, Alpharma ApS, Dalslandsgade 11, 2300 Copenhagen, Denmark2

Received 12 July 2004/ Accepted 28 November 2004


   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Streptomyces tenebrarius is an industrially important microorganism, producing an antibiotic complex that mainly consists of the aminoglycosides apramycin, tobramycin carbamate, and kanamycin B carbamate. When S. tenebrarius is used for industrial tobramycin production, kanamycin B carbamate is an unwanted by-product. The two compounds differ only by one hydroxyl group, which is present in kanamycin carbamate but is reduced during biosynthesis of tobramycin. 13C metabolic flux analysis was used for elucidating connections between the primary carbon metabolism and the composition of the antibiotic complex. Metabolic flux maps were constructed for the cells grown on minimal medium with glucose or with a glucose-glycerol mixture as the carbon source. The addition of glycerol, which is more reduced than glucose, led to a three-times-greater reduction of the kanamycin portion of the antibiotic complex. The labeling indicated an active Entner-Doudoroff (ED) pathway, which was previously considered to be nonfunctional in Streptomyces. The activity of the pentose phosphate (PP) pathway was low (10 to 20% of the glucose uptake rate). The fluxes through Embden-Meyerhof-Parnas (EMP) and ED pathways were almost evenly distributed during the exponential growth on glucose. During the transition from growth phase to production phase, a metabolic shift was observed, characterized by a decreased flux through the ED pathway and increased fluxes through the EMP and PP pathways. Higher specific NADH and NADPH production rates were calculated in the cultivation on glucose-glycerol, which was associated with a lower percentage of nonreduced antibiotic kanamycin B carbamate.


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
The ability of actinomycetes to make secondary metabolites with different useful properties (antibacterials, antitumor agents, immunosuppressants, etc.) is widely exploited in the pharmaceutical industry. Two thirds of the antibiotics produced by microorganisms are made by actinomycetes. In particular, the Streptomyces genus is remarkable in this aspect, representing about 80% of the actinomycete antibiotics (16).

Antibiotics are formed from specific precursors that are drained from the central carbon metabolism, and overproduction of antibiotics therefore requires that the precursors are supplied in sufficient quantities. Improvement of antibiotic production has traditionally been based on random mutagenesis methods, and in the future these methods will play an important role. However, metabolic engineering enables the introduction of rational changes to the central carbon metabolism to increase fluxes of precursors and cofactors to antibiotics. Metabolic flux analysis is a valuable tool in guiding metabolic engineering strategies, as it enables rapid phenotypic characterization of different mutants; through analysis of different mutants, one may gain insight into the correlation between antibiotic production and the fluxes through specific branches of the metabolic network.

In this study, correlations between the primary and the secondary metabolism in the antibiotics producer Streptomyces tenebrarius (11, 25) were investigated. This species produces the aminoglycoside antibiotic complex nebramycin, consisting primarily of tobramycin carbamate, kanamycin B carbamate, and apramycin (17, 18, 23). Tobramycin and kanamycin B carbamates can be hydrolyzed into the active forms (tobramycin and kanamycin B) in a chemical modification step. When S. tenebrarius is used for tobramycin production, synthesis of other nebramycin complex components decreases the tobramycin yield and causes problems in downstream processing. In particular, kanamycin B is difficult to separate from tobramycin due to their chemical similarity. The compounds differ only by one hydroxyl group, which is present in kanamycin B but is reduced during biosynthesis in tobramycin (Fig. 1). Strain development and cultivation optimization at the company Alpharma (Copenhagen, Denmark) have been used in the past to increase the tobramycin carbamate yield and reduce the kanamycin B carbamate content in the complex. Further improvement of the process requires better knowledge of the cellular metabolism; it is therefore valuable to know whether changes in the antibiotic ratio are connected to changes in the primary metabolism and how this can be used for directed improvement of the strain. For analysis of the primary metabolism of S. tenebrarius, metabolic flux analysis was combined with carbon labeling experiments. The cells were grown on 13C-labeled substrate, the labeling of amino acids derived from the cell proteins was analyzed by gas chromatography-tandem mass spectrometry (GC-MS), and through computer modeling the fluxes in the central carbon metabolism were calculated (4, 6).



View larger version (17K):
[in this window]
[in a new window]
  		FIG. 1. Chemical structures of tobramycin, tobramycin carbamate, kanamycin B, and kanamycin B carbamate.

 

   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Chemicals.
All chemicals were of chemical purity degree or higher, purchased from Merck (NJ) or Sigma-Aldrich Co. (MO). The [1-13C]-labeled glucose was from Omicron Biochemicals, Inc. (IN).

Strain and culture conditions.
Streptomyces tenebrarius strain TD507 (Alpharma ApS, Copenhagen, Denmark) was used as the nebramycin complex-producing microorganism. The strain originates from S. tenebrarius ATCC 17920 and has been developed by successive mutations either by UV or N-methyl-N'-nitro-nitrosoguanidine treatments. The strain was stored in 1-ml cryotubes at –80°C at a biomass concentration of 9.3 g (dry weight) · liter–1.

Cultures were grown in batch fermentors with 300 ml growth medium. The medium contained 2 g · liter–1 (NH4)2SO4, 0.08 g · liter–1 KH2PO4, 0.5 g · liter–1 MgSO4 · 7H2O, 10 mg · liter–1 ZnSO4 · 7H2O, 10 mg · liter–1 FeSO4 · 7H2O, 0.2 mg · liter–1 MnSO4 · H2O, 0.02 mg · liter–1 CuSO4 · 5H2O, 0.02 mg · liter–1 CoCl2 · 6H2O, 2 g · liter–1 CaCl2, and 1 ml · liter–1 Pluronic as antifoam. As a carbon source, one of the following was used: 8 g · liter–1 glucose or 4 g · liter–1 glucose and 4 g · liter–1 glycerol.

For labeled substrate cultivations, one of the following carbon source combinations was used: 4 g · liter–1 [1-13C]glucose and 4 g · liter–1 naturally labeled glucose or 4 g · liter–1 [1-13C]glucose and 4 g · liter–1 naturally labeled glycerol.

The fermentors were autoclaved with water and Pluronic, and all other components were added afterwards by sterile filtration. Fermentors were inoculated with 0.3 ml (0.1% [vol/vol]) of stock culture. The aeration rate was 1 volume of air per volume of culture suspension per min (vvm). For carbon labeling experiments, the inlet air was bubbled through a 2 N NaOH solution to remove atmospheric carbon dioxide. The pH of the medium was kept at 6.8 by the automatic addition of 0.1 N NaOH, and the temperature was kept at 37°C throughout the cultivation.

Biomass dry weight.
A total of 5 ml of culture broth was filtered through a 0.45-µm-pore-size predried filter (Supor-450; Pall Corporation). The filtrate was collected and used for metabolite analysis by high-performance liquid chromatography (HPLC) and for analysis of antibiotics. The biomass on the filter was washed twice with distilled water; the filter with biomass was then dried in a microwave oven at 170 W for 20 min and cooled down in a desiccator for 15 min, and the mass gain was measured.

Glucose, glycerol, and extracellular metabolite analysis.
Glucose and several extracellular metabolites were analyzed by HPLC with an Aminex HPX-87H column (Bio-Rad Laboratories, Hercules, CA) operating at 60°C. The mobile phase was 5 mM H2SO4 at a flow rate of 0.6 ml · min–1. Glucose, ethanol, glycerol, and succinate were quantified with a differential refractometer (Waters 410; Millipore, Bedford, MA), whereas acetate and pyruvate were quantified with a tunable absorbance detector set at 210 nm (Waters 486; Millipore, Bedford, MA).

Antibiotic analysis.
Tobramycin, tobramycin carbamate, kanamycin B, kanamycin B carbamate, and apramycin concentrations were analyzed on a Waters HPLC system (Milford), equipped with a WISP 717 column and a fluorescence detector (Waters 474) set on 338 nm for excitation and 418 nm for emission measurement. The total eluent flow was 1.2 ml · min–1. One liter of the eluent solution contained 175 ml methanol, 3.2 g sodium hexanesulfonate, 12.8 g sodium sulfate, 0.8 to 1.2 g acetic acid (to obtain a pH of 3.4), and Milli-Q water (Milli-Q ultrapure water purification system; Millipore). Eluent B differed from eluent A by having twice the amount of methanol. The antibiotics were derivatized by o-phthalaldehyde and a mercaptan to give fluorescent compounds. The derivatization was performed at room temperature with a reaction coil installed after the separation column.

GC-MS analysis.
From the labeled substrate, cultivation samples of 20 to 30 ml each were taken in duplicate for the analysis of cell mass labeling. Samples were filtered through a 0.45-µm-pore-size filter (Supor-450; Pall Corporation), and the cells were washed twice with distilled water, collected from the filter, and stored at –20°C until further analysis.

A total of 15 mg of wet cell mass was hydrolyzed with 6 M HCl at 105°C for 24 h to release free amino acids from protein molecules or for 0.5 h to free glucose molecules from the cell wall. The hydrolysate was centrifuged at 15,000 x g to remove the cell debris, divided into two aliquots, and dried at 105°C. The crude hydrolysates were subjected to derivatization that rendered them volatile: amino acids were converted to N-ethoxycarbonyl amino acid ethyl esters and N-(N,N-dimethylaminomethylene) amino acid methyl esters and glucose to glucose pentaacetate (4). Analysis was made by gas chromatography coupled with mass spectrometry as previously described (4). The signal intensities were corrected for occurrence of natural isotopes in the atoms of the derivative part and in oxygen and nitrogen atoms of the amino acids. The corrected intensities were used for calculating the summed fractional labeling (SFL) of a fragment according to the following formula:

where ij is the peak intensity of mass isotopomer mj and n is the number of carbon atoms in the fragment. The SFL value shows the amount of 13C atoms per 100 fragments.

To compare the labeling states of two metabolites with different numbers of carbon atoms, carbon normalized labeling (CNL) was calculated as CNL = SFL/n, where SFL is the summed fractional labeling of all the carbons in the fragment and n is the number of carbon atoms in the fragment.

Modelling of metabolic fluxes.
A mathematical framework used for quantification of the fluxes was described previously (27, 28). The input to the program included the stoichiometric model, the transitions of carbon atoms in each reaction, the measured SFLs of metabolite fragments, the measured fluxes of substrate uptake, and the calculated fluxes towards cell mass. The primary metabolic network of S. tenebrarius was reconstructed based on data about carbon metabolism in related species (12). The fluxes towards cell mass were calculated based on precursor requirements for cell mass biosynthesis in Corynebacterium glutamicum (20).

The numeric method (described in more detail elsewhere) (5) can be summarized as follows. An arbitrary set of fluxes was chosen that fits the stoichiometric constraints; using this set of fluxes, the SFLs were calculated. The differences between the measured and the calculated SFLs gave an error; if the error was more than the set value, a new set of fluxes was made and the procedure was repeated until a good fit was obtained. The set of metabolic fluxes that resulted in the best fit was found by an iterative process of error minimization.

Identification of the edd gene.
Degenerate primers for the amplification of internal edd gene fragment were designed to match regions of high identity in multiple alignments of protein sequences from Nonomuraea sp. ATCC 39727 and other bacterial species. The conserved amino acid regions were chosen with a maximum distance from each other to amplify the largest part of the gene possible. To reduce degeneracy of the primers, the codon preference of Streptomyces coelicolor was considered (Kazusa DNA Research Intitute, Japan; http://www.kazusa.or.jp/codon/). The primers were extended with 20-nucleotide-long tags at the 5' ends to facilitate subsequent sequencing. PCRs were performed using the GC-RICH PCR system (Roche, Mannheim, Germany) and S. tenebrarius genomic DNA as a template. The coding strand was sequenced at MWG Biotech AG (Reinach, Switzerland).

Nucleotide sequence accession number.
The coding strand sequence was submitted to GenBank (accession number AY618459).


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Influence of carbon source on antibiotic production.
The biosynthesis of kanamycin and tobramycin is likely to be differentiated only by a reductive reaction, and we therefore investigated how the degree of reduction of the substrate influenced the formation of the antibiotic complex composition in S. tenebrarius cultivation. Cultivations were performed with two different carbon sources: pure glucose and an equimolar mixture of glucose and glycerol, where glycerol is a more reduced carbon source than glucose.

Cultivations were performed under controlled conditions on defined minimal medium. The concentrations of antibiotics in the medium were analyzed when the carbon source(s) was exhausted (Fig. 2). Tobramycin concentration was not significantly influenced by the nature of the carbon source, while the kanamycin concentration decreased 2.6 times and the apramycin concentration increased 2.9 times when half of the glucose was replaced by glycerol.



View larger version (21K):
[in this window]
[in a new window]
  		FIG. 2. Final concentrations of antibiotics in S. tenebrarius cultivations on glucose and on an equimolar mixture of glucose and glycerol. Tobramycin shows the summed concentrations of tobramycin and tobramycin carbamate, and kanamycin shows the summed concentrations of kanamycin B and kanamycin B carbamate. Concentrations shown are means of three (glucose) or two (glucose-glycerol) cultivations.

 
Cultivations on labeled substrate.
To estimate the fluxes in the central carbon metabolism during growth on glucose and the glucose-glycerol mixture, carbon labeling experiments were performed. Two batch cultivations were carried out under the same conditions as mentioned above. However, in the glucose cultivation, half of the glucose was labeled in the first position and the other half was naturally labeled; in the other cultivation, a mixture of [1-13C]glucose and naturally labeled glycerol was used as a carbon source. The cells grew at similar specific growth rates and had similar specific glucose uptake rates in the exponential phases of both cultivations (Fig. 3; Table 1). Although the final tobramycin concentration was lower in the fermentation with glucose as a carbon source when a labeled carbon source was used (Fig. 3) than with a nonlabeled carbon source (Fig. 2), the change in the relative production of the different components of the antibiotic complex was the same upon the shift of carbon source in the two sets of experiments. During the production phase, which starts when the cells are still growing exponentially but continues into the growth deceleration phase, the substrate uptake continued to be higher in the cultivation with glucose-glycerol as a carbon source, and the specific rate of antibiotic production was two times higher. The tendency of higher tobramycin and lower kanamycin production in cultivation on glucose-glycerol was preserved in labeled substrate fermentation. During fermentation, biomass samples were taken at the end of the exponential growth phase (47 h for glucose fermentation and 33 h for glucose-glycerol fermentation) and during the production phase (70 h for glucose fermentation and 49 and 68 h for glucose-glycerol fermentation), and the SFLs of derivatized proteinogenic amino acids were determined by GC-MS analysis (Table 2).



View larger version (23K):
[in this window]
[in a new window]
  		FIG. 3. Cultivation profile on minimal medium with [1-13C]glucose (A) and [1-13C]glucose-glycerol (B) as the carbon sources. The arrows indicate when samples for biomass labeling analysis were taken. Biomass dry weight {square}, glucose {blacktriangleup}, glycerol {diamondsuit}, tobramycin and tobramycin carbamate •, kanamycin and kanamycin carbamate {triangledown}, and apramycin {blacksquare}.

 

View this table:
[in this window]
[in a new window]
  		TABLE 1. Some kinetics parameters of growth in the labeled substrate cultivations

 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Measured summed fractional labelings

 
Labeling analysis. (i) Pyruvate labeling.
Pyruvate is a central carbon metabolite; its labeling pattern can be deduced from the labeling patterns of the amino acids alanine and valine, as these have pyruvate as a precursor. The amount and diversity of the fragments measured allowed calculation of the fractional labeling of all three carbon atoms in pyruvate (Fig. 4; Table 2).



View larger version (12K):
[in this window]
[in a new window]
  		FIG. 4. Labeling of pyruvate atoms in cultivation on glucose. The SFL of the pyruvate second and third carbon atoms was calculated as an average of SFLs of Ala116, Ala99, half-SFLs of Val127, and half-SFLs of Val144. The labeling of the first pyruvate carbon atom was found as an average between differences: Ala158 – PYR(2, 3) and Val186 – PYR(2, 3). The SFL of the pyruvate second carbon was calculated as the difference between Val143 and PYR(1). Eventually, the SFL of the third position was found as the difference between SFL of PYR(2, 3) and PYR(2).

 
When [1-13C]glucose is metabolized into pyruvate via the Embden-Meyerhof-Parnas (EMP) pathway, 50% of the pyruvate formed is labeled in position 3. If glucose is metabolized via the Entner-Doudoroff (ED) pathway, 50% of the pyruvate formed will be labeled in position 1. This pathway is less energetically favorable than the more common EMP pathway. When the pentose phosphate (PP) pathway is active, the labeled carbon atom is converted to carbon dioxide and the resulting pyruvate does not carry any 13C atoms.

For the cultivation on glucose the rate of labeling of the first pyruvate position was much higher than that of natural labeling (1.1%), clearly pointing to an active ED pathway. However, the EMP pathway was also active, as could be judged from the labeling of the third carbon atom in pyruvate. When pyruvate labeling at the exponential growth phase and after onset of the production phase was compared, an interesting change in the incorporation pattern was observed: the labeling of the first atom decreased, while the labeling of the second and third atoms increased. This can be interpreted as a metabolic shift from ED towards EMP activity when the cells passed from growth to secondary metabolite production. It is possible to judge activity of the oxidative branch of the PP pathway in the glucose cultivation from the data of glucose labeling in the medium. The SFL of glucose atoms in the medium was equal to 61.7% ± 0.1%. If glucose was metabolized solely through the EMP and ED pathways, the summed fractional labeling of pyruvate should be about 31% (one mole of glucose gives 2 mol of pyruvate). However, at 47 h the measured SFL was 26.1%, and this SFL value implies that glucose is partly metabolized through the oxidative PP pathway.

(ii) Tricarboxylic acid (TCA) cycle metabolites.
The average labeling of carbon atoms (CNL) in the two TCA cycle metabolites (oxaloacetate and {alpha}-ketoglutarate) was found from the fragments of aspartate (Asp216) and glutamate (Glu230), respectively (Fig. 5). The CNL values were equal as expected because there was a constant interconversion between oxaloacetate and {alpha}-ketoglutarate atoms in the TCA cycle. The increase of labeling with time can be explained by the increase of the labeling of the second and third carbon atoms of pyruvate, which enters the TCA cycle in the form of acetyl-coenzyme A.



View larger version (15K):
[in this window]
[in a new window]
  		FIG. 5. Labeling of the TCA cycle metabolites oxaloacetate and {alpha}-ketoglutarate in cultivation on glucose. The labelings are normalized with respect to carbon atoms.

 
Computer simulations of fluxes at the exponential phase.
Based on knowledge about the primary metabolism of streptomycetes (12), a metabolic model was constructed for S. tenebrarius. Using this model, the fluxes through the different pathways were estimated (see Materials and Methods) for cells in the exponential growth phase, where the system can be considered to be in pseudo-steady state (Fig. 6). The irreversibility of fructose-6-phosphate conversion into fructose-1,6-biphosphate (which was initially imposed in the model) resulted in a poor fit of the glucose-6-phosphate (G6P) labeling in the fermentation on the two carbon sources (104% predicted against 85.1% measured). One factor that could explain this deviation would be that the model predicts too-low flux through the oxidative PP pathway activity, but this option was excluded due to an almost complete accumulation of the label from glucose in pyruvate. Thus, there must be a route resulting in dilution of the labeling at the level of glucose-6-phosphate; the only possible option is reversibility of the reactions converting fructose-6-phosphate into glyceraldehyde-3-phosphate and dihydroxyacetone-phosphate. Dihydroxyacetone-phosphate is an intermediate product of glycerol catabolism, and part of it will therefore be only naturally labeled. Hence, activity of fructose-1,6-biphosphatase was included in the model for the simulations of growth on glucose-glycerol.



View larger version (35K):
[in this window]
[in a new window]
  		FIG. 6. The metabolic fluxes as calculated in a computer simulation. The values are averages between fluxes obtained in two simulations. Numbers in boldface type are fluxes for the glucose cultivation, and numbers in italics are fluxes for the glucose-glycerol cultivation. All fluxes are given in mmol · g (dry weight)–1 · h–1. The numbers in circles show exchange coefficients for reversible reactions, calculated as MIN (v_forward, v_reverse)/100 + MIN (v_forward, v_reverse), where v_forward is the forward flux, and v_reverse is reverse flux. The function MIN (v_forward, v_reverse) gives the smaller flux (forward or reverse) as the result.

 
(i) Glycolysis pathways.
The distribution of fluxes around the G6P branch point was quite different for the two different carbon sources. In cultivation on glucose, the ED and EMP pathways were equally important for glucose conversion, while about 20% of the glucose was shunted through the PP pathway (some G6P was channeled into biomass). Less than half of the carbon that was oxidized in the first step of the PP pathway returned to the EMP pathway, whereas the remaining glucose that entered the PP pathway was used as building blocks for nucleotides and aromatic amino acids.

In the glucose-glycerol cultivation, most of the G6P (85.1%) was directed towards the ED pathway. The remaining 12.4% went to the PP pathway, which was just sufficient to satisfy the needs for precursor metabolites required for biomass synthesis. Thus, in this cultivation the main role of the PP pathway was supplying precursors and not the formation of NADPH. Despite the presence of labeling in the third position of pyruvate, there was no net flux of G6P into the EMP pathway. This can be explained by high exchange rates in the reactions leading from G6P to glyceraldehyde-3-phosphate. The lower rate of PP pathway activity in glucose-glycerol cultivation is also reasonable from a biological point of view: the ED pathway supplied the cell with NADPH required for growth. It is therefore not necessary to use the PP pathway for supply of NADPH; this pathway is exclusively used for supply of biomass precursors.

(ii) Anaplerotic pathways.
Out of the two anaplerotic pathways that were included in the model, pyruvate carboxylation was chosen as the one that gave the best fit to the labeling patterns in the glucose cultivation. This was surprising because phosphoenolpyruvate carboxylase is regarded as the enzyme primarily responsible for TCA cycle metabolite replenishing in Streptomyces (2, 8).

NADH and NADPH production.
With respect to the availability of cofactors for synthesis of antibiotics, it is important to know how much NADH and NADPH is synthesized and consumed by the cell. In this case it was interesting to know whether more reducing equivalents were produced per g (dry weight) per hour when glucose was partly substituted by glycerol. The cells grew at approximately the same specific growth rates during both cultivations, and the cell compositions were assumed to be similar. Consequently, the specific fluxes of NADH production and NADPH consumption during biomass synthesis were considered similar for both cultivations. Therefore, the main difference was determined by the amount of cofactors produced, which could be calculated from the estimated fluxes and the stoichiometric coefficients for NADH and NADPH in each of the reactions included in the model (Fig. 7).



View larger version (12K):
[in this window]
[in a new window]
  		FIG. 7. Specific reducing cofactor production in the cultivation on glucose and on glucose-glycerol.

 
The specific net production rate of NADH in the exponential growth phase was higher with cultivation on glucose-glycerol (4.8 mmol · g–1 [dry weight] · h–1) than in the cultivation on glucose (3.1 mmol · g [dry weight]–1 · h–1). This can be explained by a higher carbon uptake rate, 5.66 carbon-millimole (c-mmol) substrate · g (dry weight)–1 · h–1 in the glucose-glycerol cultivation compared to 4.00 c-mmol substrate · g (dry weight)–1 · h–1 in the cultivation on pure glucose. Another reason is the higher degree of reduction of glycerol, which results in the formation of more NADH during catabolism.

The specific net production rate of NADPH was slightly higher in the cultivation on glucose-glycerol, 1.6 mmol · g (dry weight)–1 · h–1, compared with the cultivation on glucose, 1.2 mmol · g (dry weight)–1 · h–1. Dividing these values by the specific growth rate (approximately 0.06 h–1), one obtains the specific NADPH formation for the two cultivations, 27 and 20 mmol · g (dry weight)–1, respectively. The requirement for NADPH for biomass biosynthesis varies for different organisms; for Penicillium chrysogenum, it was calculated to be 8.5 mmol · g (dry weight)–1 (21), and for C. glutamicum the value was 14.9 mmol · g (dry weight)–1 (20). From this, it is clear that the predicted NADPH formation rate in S. tenebrarius can satisfy the biomass biosynthesis.

Judging from the substrate consumption rate (which was twice as high with glucose-glycerol cultivation), the tendency of higher specific production rates of the reduced cofactors with the glucose-glycerol cultivation should also hold for the stationary phase, when the antibiotics were produced.

Identification of the edd gene.
To further confirm the presence of an ED pathway in Streptomyces tenebrarius, a 1,272 kb-fragment of the edd gene, encoding the ED pathway specific enzyme 6-phosphogluconate dehydratase, was amplified. The amplification was performed with degenerate primers on genomic DNA of S. tenebrarius as the template. A total of 675 bp was successfully sequenced. The putative amino acid sequence showed 70% identity over 225 residues to Edd from Pseudomonas putida, Pseudomonas aeruginosa, and Escherichia coli. In a clustering analysis, the sequence grouped with Edd from other actinomycetes with proven ED pathway activity—Nonomuraea sp. ATCC 39727 and Mycobacterium smegmatis (Fig. 8).



View larger version (18K):
[in this window]
[in a new window]
  		FIG. 8. Unrooted phylogenetic tree derived by neighbor-joining analysis of Edd amino acid sequences from different bacteria species with a proven Entner-Doudoroff pathway presence. The bootstrap values on the branches indicate the number of times a given branch appeared in 1,000 bootstrap replications. The sequences were obtained from the National Center for Biotechnology Information database (http://www.ncbi.nlm.nih.gov). ClustalX 1.8 was used for making multiple sequence alignment and for construction of the tree. The tree was drawn with Njplot.

 

   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Actinomycetes are an important group of bacteria for the biotechnological industry, mainly due to their ability to produce a wide range of bioactive metabolites. The secondary metabolism of actinomycetes has been studied intensively, but the primary metabolism of most actinomycetes is, however, poorly characterized. In this study, a correlation between the primary metabolism and the aminoglycoside production in S. tenebrarius was sought.

Detection of Entner-Doudoroff pathway.
The Entner-Doudoroff pathway for carbohydrate metabolism (9) is most common in gram-negative bacteria, although it has also been encountered in gram-positive bacteria and even in some eucaryotes (7). The ED pathway has never been described in Streptomyces before and is unusual for actinomycetes altogether. To the best of our knowledge, the only actinomycetes found to possess an ED pathway are Mycobacterium smegmatis (1) and Nonomuraea spp. ATCC 39727 (10). The edd gene has not been found in the sequenced genomes of the two Streptomyces species (S. coelicolor and S. avermitilis) (10). ED pathway activity was shown in S. tenebrarius by labeling experiments, and the presence of the characteristic gene edd was confirmed by PCR and subsequent sequencing. The activity of the pathway could not be verified by the enzymatic assay of Budgen and Danson (3), because of a high nonspecific NADH oxidative activity of the cell extract (data not shown).

Labeling experiments indicated simultaneous activity of glycolysis and gluconeogenesis, which basically means that there is an inefficient energy metabolism. Possibly, the ED pathway functions in a cyclic mode as in pseudomonads, where part of the dihydroxy-acetone phosphate produced in the ED pathway is recycled back to fructose-6-phosphate and then to gluconate instead of being directly processed into pyruvate (7). An alternative explanation is that gluconeogenesis is active only during metabolism of glycerol. An analogy could again be drawn with pseudomonads, where disruption of the edd gene is deleterious for growth on glycerol (19), which shows that glycerol is first turned into glucose-6-phosphate and then metabolized via the ED pathway. Lower rates of S. tenebrarius biomass yield during growth on glucose-glycerol (0.34 g [dry weight]/g substrate versus 0.50 g [dry weight]/g glucose) support the latter hypothesis. Normally, the formation of a futile cycle due to fructose-6-phosphate and fructose-1,6-diphosphate interconversion is avoided by microorganisms. For instance, E. coli operates the ED pathway only in a linear mode, and pseudomonads are devoid of 6-phosphofructokinase and therefore cannot metabolize carbohydrates via the EMP pathway. It has been shown that when 6-phosphofructokinase was introduced into Alcaligenes eutrophus (related to pseudomonads), the growth of the organism was impaired. However, when 6-phosphofructokinase was introduced into an Edd– mutant, the organism successfully used the EMP pathway and grew normally (26).

Metabolic shift.
Another interesting discovery was that there is a metabolic shift when S. tenebrarius was grown on glucose and the cells passed from the growth phase to the production phase. A change in metabolism has previously been observed with several other streptomycetes (Table 3). It is apparent that the cells adjust their metabolism to the change in function: from growth to secondary metabolite production.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Metabolic shift in actinomycetes on transition from growth to secondary metabolism

 
Some common observations can be made for metabolic shifts in Streptomyces that use the EMP pathway and the PP pathway. For Streptomyces lividans, which produces the antibiotics actinorhodin and undecylprodigiosin, it has been observed that for decreasing specific growth rates the activity of the PP pathway decreases and vice versa (24). A similar observation of a decreased PP flux at low specific growth rates and production of nystatin in Streptomyces noursei has also been reported (14).

For two antibiotic-producing actinomycetes that use the ED pathway (Nonomuraea spp. ATCC 39727 and S. tenebrarius), a common trait is that the activity of this pathway decreases after transition from growth to production, and so a larger part of the glucose (maybe even all) is metabolized via the EMP and PP pathways.

The PP and ED pathways both produce reduction equivalents in the form of NADPH. Because NADPH is necessary for growth, it is logical that the supply pathways are mostly active during the growth period. When cell growth slows down, the need for this cofactor decreases and the energetically more favorable EMP pathway can be used.

Cofactor requirement for antibiotics synthesis.
Could cofactor supply be the limiting factor for antibiotic production? NADPH is regarded as the cofactor necessary for biosynthesis of several antibiotics: ß-lactams, polyketides, and glycopeptides. NADPH is produced in the PP and ED pathways but not in the EMP pathway. In overproducing strains of S. lividans, the flux through the PP pathway during production phase was higher than in strains with lower levels of antibiotic production (24). Increased activity of the PP pathway has also been observed in connection with higher productivity of the polyketide avermectin in S. avermitilis (13) and the cyclopentanone antibiotic methylenomycin in S. coelicolor A3(2) (22).

The levels of aminoglycosides produced by S. tenebrarius are only slightly lower than glucose, from which they are made. Tobramycin and apramycin each have one hydroxyl group reduced during biosynthesis. However, aminoglycosides need glutamate as a source of amino groups for the amination reactions. Glutamate is produced in the cell by amination of {alpha}-ketoglutarate with concurrent consumption of NADPH. If the reducing equivalents used both for the uptake of ammonia (5 mol per mol of nebramycin component) and for the reduction of the hydroxyl group (1 mol per mol of tobramycin or apramycin) are taken into consideration, the reducing equivalent requirement will be about 10.6 µmol per mg nebramycin (average molecular mass is 520 g · mol–1). The requirements of reducing cofactors for antibiotic synthesis will therefore be about 20 µmol · g (dry weight)–1 · h–1 in the cultivation with glucose as carbon source and 40 µmol · g (dry weight)–1 · h–1 in the cultivation with glucose-glycerol as carbon source. These requirements represent <2 to 3% of the total cofactor production flux, and it is therefore not likely that cofactor supply for antibiotics biosynthesis is limiting. On the other hand, an increased ratio of NADPH:NADP+ in glucose-glycerol cultivation could in principal work as a selective pressure for production of more reduced antibiotics (tobramycin and apramycin).

Suggestions for future work.
During preparation of the manuscript, the tobramycin biosynthetic cluster was isolated from S. tenebrarius (15). Two genes were identified as possible candidates for paromamine reduction, which is the decisive step directing the aminoglycoside flux toward tobramycin instead of kanamycin. Overexpression of this reductase could shift the balance from kanamycin B towards tobramycin production.


   ACKNOWLEDGMENTS
 
This work was financially supported by the Danish Ministry and Alpharma ApS through a stipend to I.B.

We are grateful to Nina Gunnarsson for scientific discussions during the course of the work. We also acknowledge the advice and help of our colleagues Søren Bendiksen, Hans Peter Smits, Michael Lynge Nielsen, Thomas Grotkjær, and Mats Åkesson.


   FOOTNOTES
 
* Corresponding author. Mailing address: Center for Microbial Biotechnology, BioCentrum-DTU, Building 223, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark. Phone: 45 4525 2696. Fax: 45 4588 4148. E-mail: jn{at}biocentrum.dtu.dk. Back


   REFERENCES
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Bai, N. J., M. R. Pai, P. S. Murthy, and T. A. Venkitasubramanian. 1976. Pathways of glucose catabolism in Mycobacterium smegmatis. Can. J. Microbiol. 22:1374-1380.[Medline]
  2. Bramwell, H., H. G. Nimmo, I. S. Hunter, and J. R. Coggins. 1993. Phosphoenolpyruvate carboxylase from Streptomyces coelicolor A3(2): purification of the enzyme, cloning of the ppc gene and over-expression of the protein in a streptomycete. Biochem. J. 293:131-136.
  3. Budgen, N., and M. J. Danson. 1986. Metabolism of glucose via a modified Entner-Doudoroff pathway in the thermoacidophilic archaebacterium Thermoplasma acidophilum. FEBS Lett. 196:207-210.[CrossRef]
  4. Christensen, B., and J. Nielsen. 1999. Isotopomer analysis using GC-MS. Metab. Eng. 1:282-290.
  5. Christensen, B., and J. Nielsen. 2000. Metabolic network analysis of Penicillium chrysogenum using 13C-labeled glucose. Biotechnol. Bioeng. 68:652-659.[CrossRef][Medline]
  6. Christensen, B., and J. Nielsen. 2000. Metabolic network analysis. A powerful tool in metabolic engineering. Adv. Biochem. Eng. Biotechnol. 66:209-231.[Medline]
  7. Conway, T. 1992. The Entner-Doudoroff pathway: history, physiology and molecular biology. FEMS Microbiol. Rev. 9:1-27.[Medline]
  8. Dekleva, M. L., and W. R. Strohl. 1988. Activity of phosphoenolpyruvate carboxylase of an anthracycline-producing streptomycete. Can. J. Microbiol. 34:1241-1246.[Medline]
  9. Entner, N., and M. Doudoroff. 1952. Glucose and gluconic acid oxidation of Pseudomonas saccharophila. J. Biol. Chem. 196:853-862.[Free Full Text]
  10. Gunnarsson, N., U. H. Mortensen, M. Sosio, and J. Nielsen. 2004. Identification of the Entner-Doudoroff pathway in an antibiotic-producing actinomycete species. Mol. Microbiol. 52:895-902.[CrossRef][Medline]
  11. Higgins, C. E., and R. E. Kastner. 1967. Nebramycin, a new broad-spectrum antibiotic complex. II. Description of Streptomyces tenebrarius. Antimicrob. Agents Chemother. 7:324-331.
  12. Hodgson, D. A. 2000. Primary metabolism and its control in Streptomycetes. Adv. Microb. Physiol. 42:47-238.[Medline]
  13. Ikeda, H., H. Kotaki, H. Tanaka, and S. Omura. 1988. Involvement of glucose catabolism in avermectin production by Streptomyces avermitilis. Antimicrob. Agents Chemother. 32:282-284.[Abstract/Free Full Text]
  14. Jonsbu, E., B. Christensen, and J. Nielsen. 2001. Changes of in vivo fluxes through central metabolic pathways during the production of nystatin by Streptomyces noursei in batch culture. Appl. Microbiol. Biotechnol. 56:93-100.[CrossRef][Medline]
  15. Kharel, M. K., D. B. Basnet, H. C. Lee, K. Liou, J. S. Woo, B. G. Kim, and J. K. Sohng. 2004. Isolation and characterization of the tobramycin biosynthetic gene cluster from Streptomyces tenebrarius. FEMS Microbiol. Lett. 230:185-190.[CrossRef][Medline]
  16. Kieser, T., M. J. Bibb, M. J. Buttner, K. F. Chater, and D. A. Hopwood. 2000. General introduction to actinomycete biology, p. 1-42. Practical Streptomyces genetics. The John Innes Foundation, Norwich, United Kingdom.
  17. Koch, K. F., F. A. Davis, and J. A. Rhoades. 1973. Nebramycin: separation of the complex and identification of factors 4, 5, and 5'. J. Antibiot. (Tokyo) 26:745-751.[Medline]
  18. Koch, K. F., and J. A. Rhoades. 1970. Structure of nebramycin factor 6, a new aminoglycosidic antibiotic. Antimicrob. Agents Chemother. 10:309-313.
  19. Lessie, T. G. and P. V. Phibbs, Jr. 1984. Alternative pathways of carbohydrate utilization in pseudomonads. Annu. Rev. Microbiol. 38:359-388.[CrossRef][Medline]
  20. Marx, A., A. A. de Graaf, W. Wiechert, L. Eggling, and H. Sahm. 1996. Determination of the fluxes in the central metabolism of Corynebacterium glutamicum by nuclear magnetic resonance spectroscopy combined with metabolite balancing. Biotechnol. Bioeng. 49:111-129.[CrossRef]
  21. Nielsen, J. 1997. Physiological engineering aspects of Penicillium chrysogenum, p. 63-138. World Scientific, London, United Kingdom.
  22. Obanye, A. I., G. Hobbs, D. C. Gardner, and S. G. Oliver. 1996. Correlation between carbon flux through the pentose phosphate pathway and production of the antibiotic methylenomycin in Streptomyces coelicolor A3(2). Microbiology (Reading). 142:133-137.
  23. O'Connor, S., L. K. Lam, N. D. Jones, and M. O. Chaney. 1976. Apramycin, a unique aminocyclitol antibiotic. J. Org. Chem. 41:2087-2092.[CrossRef][Medline]
  24. Rossa, A. C., J. White, A. Kuiper, P. W. Postma, M. Bibb, and M. J. Teixeira de Mattos. 2002. Carbon flux distribution in antibiotic-producing chemostat cultures of Streptomyces lividans. Metab. Eng. 4:138-150.[CrossRef][Medline]
  25. Stark, W. M., N. G. Knox, and R. M. Wilgus. 1971. Strains of Streptomyces tenebrarius and biosynthesis of nebramycin. Folia Microbiol. (Prague) 16:205-217.
  26. Steinbuchel, A. 1986. Expression of the Escherichia coli pfkA gene in Alcaligenes eutrophus and in other gram-negative bacteria. J. Bacteriol. 166:319-327.[Abstract/Free Full Text]
  27. Wiechert, W., and A. A. de Graaf. 1996. In vivo stationary flux analysis by 13C labeling experiments. Adv. Biochem. Eng. Biotechnol. 54:109-154.[Medline]
  28. Wiechert, W., M. Mollney, N. Isermann, M. Wurzel, and A. A. de Graaf. 1999. Bidirectional reaction steps in metabolic networks. III. Explicit solution and analysis of isotopomer labeling systems. Biotechnol. Bioeng. 66:69-85.[CrossRef][Medline]

Applied and Environmental Microbiology, May 2005, p. 2294-2302, Vol. 71, No. 5
0099-2240/05/$08.00+0     doi:10.1128/AEM.71.5.2294-2302.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:


This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.
Agricola
Right arrow Articles by Borodina, I.
Right arrow Articles by Nielsen, J.

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
J. Bacteriol. Microbiol. Mol. Biol. Rev. Eukaryot. Cell All ASM Journals