Metabolic flux and metabolic network analysis of Penicillium chrysogenum using 2D [13C, 1H] COSY NMR measurements and cumulative bondomer simulation

At present two alternative methods are available for analyzing the fluxes in a metabolic network: (1) combining measurements of net conversion rates with a set of metabolite balances including the cofactor balances, or (2) leaving out the cofactor balances and fitting the resulting free fluxes to measured (13)C-labeling data. In this study these two approaches are applied to the fluxes in the glycolysis and pentose phosphate pathway of Penicillium chrysogenum growing on either ammonia or nitrate as the nitrogen source, which is expected to give different pentose phosphate pathway fluxes. The presented flux analyses are based on extensive sets of 2D [(13)C, (1)H] COSY data. A new concept is applied for simulation of this type of (13)C-labeling data: cumulative bondomer modeling. The outcomes of the (13)C-labeling based flux analysis substantially differ from those of the pure metabolite balancing approach. The fluxes that are determined using (13)C-labeling data are shown to be highly dependent on the chosen metabolic network. Extending the traditional nonoxidative pentose phosphate pathway with additional transketolase and transaldolase reactions, extending the glycolysis with a fructose 6-phosphate aldolase/dihydroxyacetone kinase reaction sequence or adding a phosphoenolpyruvate carboxykinase reaction to the model considerably improves the fit of the measured and the simulated NMR data. The results obtained using the extended version of the nonoxidative pentose phosphate pathway model show that the transketolase and transaldolase reactions need not be assumed reversible to get a good fit of the (13)C-labeling data. Strict statistical testing of the outcomes of (13)C-labeling based flux analysis using realistic measurement errors is demonstrated to be of prime importance for verifying the assumed metabolic model.     

Cumulative bondomers: a new concept in flux analysis from 2D [13C,1H] COSY NMR data

A well-established way of determining metabolic fluxes is to measure 2D [(13)C,(1)H] COSY NMR spectra of components of biomass grown on uniformly (13)C-labeled carbon sources. When using the entire set of measured data to simultaneously determine all fluxes in a proposed metabolic network model, the (13)C-labeling distribution in all measured compounds has to be simulated. This requires very large sets of isotopomer or cumomer balances. This article introduces the new concept of bondomers; entities that only vary in the numbers and positions of C-C bonds that have remained intact since the medium substrate molecule entered the metabolism. Bondomers are shown to have many analogies to isotopomers. One of these is that bondomers can be transformed to cumulative bondomers, just like isotopomers can be transformed to cumomers. Similarly to cumomers, cumulative bondomers allow an analytical solution of the entire set of balances describing a metabolic network. The main difference is that cumulative bondomer models are considerably smaller than corresponding cumomer models. This saves computational time, allows easier identifiability analysis, and yields new insights in the information content of 2D [(13)C,(1)H] COSY NMR data. We illustrate the theoretical concepts by means of a realistic example of the glycolytic and pentose phosphate pathways. The combinations of 2D [(13)C,(1)H] COSY NMR data that allow identification of all metabolic fluxes in these pathways are analyzed, and it is found that the NMR data contain less information than was previously expected.     

Metabolic flux analysis of CHO cells in perfusion culture by metabolite balancing and 2D [13C, 1H] COSY NMR spectroscopy

The physiological state of CHO cells in perfusion culture was quantified by determining fluxes through the bioreaction network using ¹³C glucose and 2D-NMR spectroscopy. CHO cells were cultivated in a 2.5 L perfusion bioreactor with glucose and glutamine as the primary carbon and energy sources. The reactor was inoculated at a cell density of 8×10⁶ cells/mL and operated at ～10×10⁶ cells/mL using unlabeled glucose for the first 13 days. The second phase lasted 12 days and the medium consisted of 10% [U-¹³C]glucose, 40% labeled [1-¹³C]glucose with the balance unlabeled. After the culture attained isotopic steady state, biomass samples from the last 3 days of cultivation were considered representative and used for flux estimation. They were hydrolyzed and analyzed by 2D [¹³C, ¹H] COSY measurements using the heteronuclear single quantum correlation sequence with gradients for artifacts suppression. Metabolic fluxes were determined using the 13C-Flux software package by minimizing the residuals between the experimental and the simulated NMR data. Normalized residuals exhibited a Gaussian distribution indicating good model fit to experimental data. The glucose consumption rate was 5-fold higher than that of glutamine with 41% of glucose channeled through the pentose phosphate pathway. The fluxes at the pyruvate branch point were almost equally distributed between lactate and the TCA cycle (55% and 45%, respectively). The anaplerotic conversion of pyruvate to oxaloacetate by pyruvate carboxylase accounted for 10% of the pyruvate flux with the remaining 90% entering the TCA cycle through acetyl-CoA. The conversion of malate to pyruvate catalyzed by the malic enzyme was 70% higher than that for the anaplerotic reaction catalyzed by pyruvate carboxylase. Most amino acid catabolic and biosynthetic fluxes were significantly lower than the glycolytic and TCA cycle fluxes. Metabolic flux data from NMR analysis validated a simplified model where metabolite balancing was used for flux estimation. In this reduced flux space, estimates from these two methods were in good agreement. This simplified model can routinely be used in bioprocess development experiments to estimate metabolic fluxes with much reduced analytical investment. The high resolution flux information from 2D-NMR spectroscopy coupled with the capability to validate a simplified metabolite balancing based model for routine use make ¹³C-isotopomer analysis an attractive bioprocess development tool for mammalian cell cultures.     

Innovations in Generation and Analysis of 2D [C,H] COSY NMR Spectra for Metabolic Flux Analysis Purposes

2D [¹³C,¹H] COSY NMR is used by the metabolic engineering community for determining ¹³C–¹³C connectivities in intracellular compounds that contain information regarding the steady-state fluxes in cellular metabolism. This paper proposes innovations in the generation and analysis of these specific NMR spectra. These include a computer tool that allows accurate determination of the relative peak areas and their complete covariance matrices even in very complex spectra. Additionally, a method is introduced for correcting the results for isotopic non-steady-state conditions. The proposed methods were applied to measured 2D [¹³C,¹H] COSY NMR spectra. Peak intensities in a one-dimensional section of the spectrum are frequently not representative for relative peak volumes in the two-dimensional spectrum. It is shown that for some spectra a significant amount of additional information can be gained from long-range ¹³C–¹³C scalar couplings in 2D [¹³C,¹H] COSY NMR spectra. Finally, the NMR resolution enhancement by dissolving amino acid derivatives in a nonpolar solvent is demonstrated.     

Kinetic analysis of dynamic 13C NMR spectra: metabolic flux, regulation, and compartmentation in hearts.

Control of oxidative metabolism was studied using 13C NMR spectroscopy to detect rate-limiting steps in 13C labeling of glutamate. 13C NMR spectra were acquired every 1 or 2 min from isolated rabbit hearts perfused with either 2.5 mM [2-13C]acetate or 2.5 mM [2-13C]butyrate with or without KCl arrest. Tricarboxylic acid cycle flux (VTCA) and the exchange rate between alpha-ketoglutarate and glutamate (F1) were determined by least-square fitting of a kinetic model to NMR data. Rates were compared to measured kinetics of the cardiac glutamate-oxaloacetate transaminase (GOT). Despite similar oxygen use, hearts oxidizing butyrate instead of acetate showed delayed incorporation of 13C label into glutamate and lower VTCA, because of the influence of beta-oxidation: butyrate = 7.1 +/- 0.2 mumol/min/g dry wt; acetate = 10.1 +/- 0.2; butyrate + KCl = 1.8 +/- 0.1; acetate + KCl = 3.1 +/- 0.1 (mean +/- SD). F1 ranged from a low of 4.4 +/- 1.0 mumol/min/g (butyrate + KCl) to 9.3 +/- 0.6 (acetate), at least 20-fold slower than GOT flux, and proved to be rate limiting for isotope turnover in the glutamate pool. Therefore, dynamic 13C NMR observations were sensitive not only to TCA cycle flux but also to the interconversion between TCA cycle intermediates and glutamate.     

A universal framework for 13C metabolic flux analysis.

A general methodology is presented for the modeling, simulation, design, evaluation, and statistical analysis of (13)C-labeling experiments for metabolic flux analysis. The universal software framework 13C-FLUX was implemented to support all steps of this process. Guided by the example of anaplerotic flux determination in Corynebacterium glutamicum, the technical details of the model setup, experimental design, and data evaluation are discussed. It is shown how the network structure, the input substrate composition, the assumptions about fluxes, and the measurement configuration are specified within 13C-FLUX. Based on the network model, different experimental designs are computed depending on the goal of the investigations. Finally, a specific experiment is evaluated and the various statistical methods used to analyze the results are briefly explained. The appendix gives some details about the software implementation and availability.     

Parallel labeling experiments with [U-(13)C]glucose validate E. coli metabolic network model for (13)C metabolic flux analysis

(13)C-metabolic flux analysis (MFA) is a widely used method for measuring intracellular metabolic fluxes in living cells. (13)C MFA relies on several key assumptions: (1) the assumed metabolic network model is complete, in that it accounts for all significant enzymatic and transport reactions; (2) (13)C-labeling measurements are accurate and precise; and (3) enzymes and transporters do not discriminate between (12)C- and (13)C-labeled metabolites. In this study, we tested these inherent assumptions of (13)C MFA for wild-type E. coli by parallel labeling experiments with [U-(13)C]glucose as tracer. Cells were grown in six parallel cultures in custom-constructed mini-bioreactors, starting from the same inoculum, on medium containing different mixtures of natural glucose and fully labeled [U-(13)C]glucose, ranging from 0% to 100% [U-(13)C]glucose. Macroscopic growth characteristics of E. coli showed no observable kinetic isotope effect. The cells grew equally well on natural glucose, 100% [U-(13)C]glucose, and mixtures thereof. (13)C MFA was then used to determine intracellular metabolic fluxes for several metabolic network models: an initial network model from literature; and extended network models that accounted for potential dilution effects of isotopic labeling. The initial network model did not give statistically acceptable fits and produced inconsistent flux results for the parallel labeling experiments. In contrast, an extended network model that accounted for dilution of intracellular CO(2) by exchange with extracellular CO(2) produced statistically acceptable fits, and the estimated metabolic fluxes were consistent for the parallel cultures. This study illustrates the importance of model validation for (13)C MFA. We show that an incomplete network model can produce statistically unacceptable fits, as determined by a chi-square test for goodness-of-fit, and return biased metabolic fluxes. The validated metabolic network model for E. coli from this study can be used in future investigations for unbiased metabolic flux measurements.     

1H and 13C NMR assignments for the glycans in glycoproteins by using 2H/13C-labeled glucose as a metabolic precursor

In order to understand the role of the glycans in glycoproteins in solution, structural information obtained by NMR spectroscopy is obviously required. However, the assignment of the NMR signals from the glycans in larger glycoproteins is still difficult, mainly due to the lack of appropriate methods for the assignment of the resonances originating from the glycans. By using [U-¹³C₆,²H₇]glucose as a metabolic precursor, we have successfully prepared a glycoprotein whose glycan is uniformly labeled with ¹³C and partially with D at the sugar residues. The D to H exchange ratios at the C1-C6 positions of the sugar residues have been proven to provide useful information for the spectral assignments of the glycan in the glycoprotein. This is the first report on the residue-specific assignment of the anomeric resonances originating from a glycan attached to a glycoprotein by using the metabolic incorporation of hydrogen from the medium into a glycan labeled with [U-¹³C₆,²H₇]glucose.     

Combining metabolic flux analysis tools and 13C NMR to estimate intracellular fluxes of cultured astrocytes

In this work, brain cell metabolism was investigated by (13)C NMR spectroscopy and metabolic flux analysis (MFA). Monotypic cultures of astrocytes were incubated with labeled glucose for 38 h, and the distribution of the label was analyzed by (13)C NMR spectroscopy. The analysis of the spectra reveals two distinct physiological states characterized by different ratios of pyruvate carboxylase to pyruvate dehydrogenase activities (PC/PDH). Intracellular flux distributions for both metabolic states were estimated by MFA using the isotopic information and extracellular rate measurements as constraints. The model was subsequently checked with the consistency index method. From a biological point of view, the occurrence of the two physiological states appears to be correlated with the presence or absence of extracellular glutamate. Concerning the model, it can be stated that the metabolic network and the set of constraints adopted provide a consistent and robust characterization of the astrocytic metabolism, allowing for the calculation of central intracellular fluxes such as pyruvate recycling, the anaplerotic flux mediated by pyruvate carboxylase, and the glutamine formation through glutamine synthetase.     

In vivo 2H/13C flux analysis in metabolism research

1. Download : Download high-res image (230KB)
2. Download : Download full-size image

     

Parallel labeling experiments for pathway elucidation and 13C metabolic flux analysis

1. Download : Download high-res image (170KB)
2. Download : Download full-size image

     

An integrated (2)H and (13)C NMR study of gluconeogenesis and TCA cycle flux in humans

Hepatic glucose synthesis from glycogen, glycerol, and the tricarboxylic acid (TCA) cycle was measured in five overnight-fasted subjects by (1)H, (2)H, and (13)C NMR analysis of blood glucose, urinary acetaminophen glucuronide, and urinary phenylacetylglutamine after administration of [1,6-(13)C(2)]glucose, (2)H(2)O, and [U-(13)C(3)]propionate. This combination of tracers allows three separate elements of hepatic glucose production (GP) to be probed simultaneously in a single study: 1) endogenous GP, 2) the contribution of glycogen, phosphoenolpyruvate (PEP), and glycerol to GP, and 3) flux through PEP carboxykinase, pyruvate recycling, and the TCA cycle. Isotope-dilution measurements of [1,6-(13)C(2)] glucose by (1)H and (13)C NMR indicated that GP in 16-h-fasted humans was 10.7 +/- 0.9 micromol.kg(-1).min(-1). (2)H NMR spectra of monoacetone glucose (derived from plasma glucose) provided the relative (2)H enrichment at glucose H-2, H-5, and H-6S, which, in turn, reflects the contribution of glycogen, PEP, and glycerol to total GP (5.5 +/- 0.7, 4.8 +/- 1.0, and 0.4 +/- 0.3 micromol.kg(-1).min(-1), respectively). Interestingly, (13)C NMR isotopomer analysis of phenylacetylglutamine and acetaminophen glucuronide reported different values for PEP carboxykinase flux (68.8 +/- 9.8 vs. 37.5 +/- 7.9 micromol.kg(-1).min(-1)), PEP recycling flux (59.1 +/- 9.8 vs. 27.8 +/- 6.8 micromol.kg(-1).min(-1)), and TCA cycle flux (10.9 +/- 1.4 vs. 5.4 +/- 1.4 micromol.kg(-1).min(-1)). These differences may reflect zonation of propionate metabolism in the liver.     

Determination of full 13C isotopomer distributions for metabolic flux analysis using heteronuclear spin echo difference NMR spectroscopy

13C-isotopomer labeling experiments play an increasingly important role in the analysis of intracellular metabolic fluxes for genetic engineering purposes. 13C NMR spectroscopy is a key technique in the experimental determination of isotopomer distributions. However, only subsets of isotopomers can be quantitated using this technique due to redundancies in the scalar coupling patterns and due to invisibility of the 12C isotope in NMR. Therefore, we developed and describe in this paper a 1H NMR spectroscopy method that allows to determine the complete isotopomer distribution in metabolites having a backbone consisting of up to at least four carbons. The proposed pulse sequences employ up to three alternately applied frequency-selective inversion pulses in the 13C channel. In a first application study, the complete isotopomer distribution of aspartate isolated from [1-13C]ethanol-grown Ashbya gossypii was determined. A tentative model of the central metabolism of this organism was constructed and used for metabolic flux analysis. The aspartate isotopomer NMR data played a key role in the successful determination of the flux through the peroxisomal glyoxylate pathway. The new NMR method can be highly instrumental in generating the data upon which isotopomer labeling experiments for flux analysis, that are becoming increasingly important, are based.     

Gas chromatography-mass spectrometry analysis of 13C labeling in sugars for metabolic flux analysis

Metabolic flux analysis, using 13C labeled substrates, has become a powerful methodology for quantifying intracellular fluxes. Most often, analysis is restricted to nuclear magnetic resonance or mass spectrometry measurement of 13C label incorporation into protein amino acids. However, amino acid isotopomer distribution insufficiently covers the entire network of central metabolism, especially in plant cells with highly compartmented metabolism, and analysis of other metabolites is required. Analysis of label in saccharides provides complementary data to better define fluxes around hexose, pentose, and triose phosphate pools. Here, we propose a gas chromatography-mass spectrometry (GC-MS) method to analyze 13C labeling in glucose and fructose moieties of sucrose, free glucose, fructose, maltose, inositol, and starch. Our results show that saccharide labeling for isotopomer quantification is better analyzed by chemical ionization than by electron ionization. The structure of the generated fragments was simulated and validated using labeled standards. The method is illustrated by analysis of saccharides extracted from developing rapeseed (Brassica napus L.) embryos. It is shown that glucose 6-phosphate isomerase and plastidial glucose 6-phosphate transport reactions are not at equilibrium, and light is shed on the pathways leading to fructose, maltose, and inositol synthesis.     

Metabolism of (1-(13)C) glucose and (2-(13)C, 2-(2)H(3)) acetate in the neuronal and glial compartments of the adult rat brain as detected by [(13)C, (2)H] NMR spectroscopy

Ex vivo ?(13)C, (2)H? NMR spectroscopy allowed to estimate the relative sizes of neuronal and glial glutamate pools and the relative contributions of (1-(13)C) glucose and (2-(13)C, 2-(2)H(3)) acetate to the neuronal and glial tricarboxylic acid cycles of the adult rat brain. Rats were infused during 60 min in the right jugular vein with solutions containing (2-(13)C, 2-(2)H(3)) acetate and (1-(13)C) glucose or (2-(13)C, 2-(2)H(3)) acetate only. At the end of the infusion the brains were frozen in situ and perchloric acid extracts were prepared and analyzed by high resolution (13)C NMR spectroscopy (90.5 MHz). The relative sizes of the neuronal and glial glutamate pools and the contributions of acetyl-CoA molecules derived from (2-(13)C, (2)H(3)) acetate or (1-(13)C) glucose entering the tricarboxylic acid cycles of both compartments, could be determined by the analysis of (2)H-(13)C multiplets and (2)H induced isotopic shifts observed in the C4 carbon resonances of glutamate and glutamine. During the infusions with (2-(13)C, 2-(2)H(3)) acetate and (1-(13)C) glucose, the glial glutamate pool contributed 9% of total cerebral glutamate being derived from (2-(13)C, 2-(2)H(3)) acetyl-CoA (4%), (2-(13)C) acetyl-CoA (3%) and recycled (2-(13)C, 2-(2)H) acetyl-CoA (2%). The neuronal glutamate pool accounted for 91% of the total cerebral glutamate being mainly originated from (2-(13)C) acetyl-CoA (86%) and (2-(13)C, 2-(2)H) acetyl-CoA (5%). During the infusions of (2-(13)C, 2-(2)H(3)) acetate only, the glial glutamate pool contributed 73% of the cerebral glutamate, being derived from (2-(13)C, 2-(2)H(3)) acetyl-CoA (36%), (2-(13)C, 2-(2)H) acetyl-CoA (27%) and (2-(13)C) acetyl-CoA (10%). The neuronal pool contributed 27% of cerebral glutamate being formed from (2-(13)C) acetyl-CoA (11%) and recycled (2-(13)C, 2-(2)H) acetyl-CoA (16%). These results illustrate the potential of ?(13)C, (2)H? NMR spectroscopy as a novel approach to investigate substrate selection and metabolic compartmentation in the adult mammalian brain.     

Biosynthesis of porphyrins and corrins. 1. 1H and 13C NMR spectra of (hydroxymethyl)bilane and uroporphyrinogens I and III

High-field NMR spectroscopic methods have been applied to study the reactions catalyzed by porphobilinogen (PBG) deaminase and uroporphyrinogen III (uro'gen III) cosynthase, which are the enzymes responsible for the formation of the porphyrin macrocycle. The action of these enzymes in the conversion of PBG, [2,11-13C]PBG, and [3,5-13C]PBG to uro'gens I and III has been followed by 1H and 13C NMR, and assignments are presented. The principal intermediate that accumulated was the correspondingly labeled (hydroxymethyl)bilane (HMB), the assignments for which are also presented.     

Compartment-specific labeling information in 13C metabolic flux analysis of plants

Metabolic engineering of plants has great potential for the low cost production of chemical feedstocks and novel compounds, but to take full advantage of this potential a better understanding of plant central carbon metabolism is needed. Flux studies define the cellular phenotype of living systems and can facilitate rational metabolic engineering. However the measurements usually made in these analyses are often not sufficient to reliably determine many fluxes that are distributed between different subcellular compartments of eukaryotic cells. We have begun to address this shortcoming by increasing the number and quality of measurements that provide (13)C labeling information from specific compartments within the plant cell. The analysis of fatty acid groups, cell wall components, protein glycans, and starch, using both gas chromatography/mass spectrometry and nuclear magnetic resonance spectroscopy are presented here. Fatty acid labeling determinations are sometimes highly convoluted. Derivatization to butyl amides reduces the errors in isotopomer resolution and quantification, resulting in better determination of fluxes into seed lipid reserves, including both plastidic and cytosolic reactions. While cell walls can account for a third or more of biomass in many seeds, no quantitative cell wall labeling measurements have been reported for plant flux analysis. Hydrolyzing cell wall and derivatizing sugars to the alditol acetates, provides novel labeling information and thereby can improve identification of flux through upper glycolytic intermediates of the cytosol. These strategies improve the quantification of key carbon fluxes in the compartmentalized flux network of plant cells.     

2D COSY 1H NMR: a new tool for studying in situ brain metabolism in the living animal

2D COSY 1H NMR with surface coil has been used to resolve and assign cerebral metabolites which had previously been detected but could not be resolved or assigned in situ in the living animal by conventional 1D 1H NMR. A wide range of cerebral metabolites, including alanine, N-acetyl aspartate, aspartate, choline derivatives, creatine/phosphocreatine pool, GABA, glucose, glutamate/glutamine pool, inositol, lactate and taurine were simultaneously resolved and assigned in situ in the whole animal using the 2D COSY correlation graphs. Global irreversible ischemia caused the appearance and the disappearance of cross-peaks in the 2D COSY 1H NMR map, corresponding to increases in alanine, GABA and lactate and glucose depletion.     

Optimization of 13C isotopic tracers for metabolic flux analysis in mammalian cells

Mammalian cells consume and metabolize various substrates from their surroundings for energy generation and biomass synthesis. Glucose and glutamine, in particular, are the primary carbon sources for proliferating cancer cells. While this combination of substrates generates static labeling patterns for use in (13)C metabolic flux analysis (MFA), the inability of single tracers to effectively label all pathways poses an obstacle for comprehensive flux determination within a given experiment. To address this issue we applied a genetic algorithm to optimize mixtures of (13)C-labeled glucose and glutamine for use in MFA. We identified tracer combinations that minimized confidence intervals in an experimentally determined flux network describing central carbon metabolism in tumor cells. Additional simulations were used to determine the robustness of the [1,2-(13)C(2)]glucose/[U-(13)C(5)]glutamine tracer combination with respect to perturbations in the network. Finally, we experimentally validated the improved performance of this tracer set relative to glucose tracers alone in a cancer cell line. This versatile method allows researchers to determine the optimal tracer combination to use for a specific metabolic network, and our findings applied to cancer cells significantly enhance the ability of MFA experiments to precisely quantify fluxes in higher organisms.     

Bioreaction network topology and metabolic flux ratio analysis by biosynthetic fractional 13C labeling and two-dimensional NMR spectroscopy

Biosynthetically directed fractional 13C labeling of the proteinogenic amino acids is achieved by feeding a mixture of uniformly 13C-labeled and unlabeled carbon source compounds into a bioreaction network. Analysis of the resulting labeling pattern enables both a comprehensive characterization of the network topology and the determination of metabolic flux ratios. Attractive features with regard to routine applications are (i) an inherently small demand for 13C-labeled source compounds and (ii) the high sensitivity of two-dimensional [13C,1H]-correlation nuclear magnetic resonance spectroscopy for analysis of 13C-labeling patterns. A user-friendly program, FCAL, is available to allow rapid data analysis. This novel approach, which recently also has been employed in conjunction with metabolic flux balancing to obtain reliable estimates of in vivo fluxes, enables efficient support of metabolic engineering and biotechnology process design.