IsoDesign: a software for optimizing the design of 13C-metabolic flux analysis experiments

Model validation and selection in metabolic flux analysis and flux balance analysis

13C-Metabolic Flux Analysis (13C-MFA) and Flux Balance Analysis (FBA) are widely used to investigate the operation of biochemical networks in both biological and biotechnological research. Both methods use metabolic reaction network models of metabolism operating at steady state so that reaction rates (fluxes) and the levels of metabolic intermediates are constrained to be invariant. They provide estimated (MFA) or predicted (FBA) values of the fluxes through the network in vivo, which cannot be measured directly. These fluxes can shed light on basic biology and have been successfully used to inform metabolic engineering strategies. Several approaches have been taken to test the reliability of estimates and predictions from constraint-based methods and to compare alternative model architectures. Despite advances in other areas of the statistical evaluation of metabolic models, such as the quantification of flux estimate uncertainty, validation and model selection methods have been underappreciated and underexplored. We review the history and state-of-the-art in constraint-based metabolic model validation and model selection. Applications and limitations of the χ2 -test of goodness-of-fit, the most widely used quantitative validation and selection approach in 13C-MFA, are discussed, and complementary and alternative forms of validation and selection are proposed. A combined model validation and selection framework for 13C-MFA incorporating metabolite pool size information that leverages new developments in the field is presented and advocated for. Finally, we discuss how adopting robust validation and selection procedures can enhance confidence in constraint-based modeling as a whole and ultimately facilitate more widespread use of FBA in biotechnology.

mfapy: An open-source Python package for 13C-based metabolic flux analysis

¹³C-based metabolic flux analysis (¹³C-MFA) is an essential tool for estimating intracellular metabolic flux levels in metabolic engineering and biology. In ¹³C-MFA, a metabolic flux distribution that explains the observed isotope labeling data was computationally estimated using a non-linear optimization method. Herein, we report the development of mfapy, an open-source Python package developed for more flexibility and extensibility for ¹³C-MFA. mfapy compels users to write a customized Python code by describing each step in the data analysis procedures of the isotope labeling experiments. The flexibility and extensibility provided by mfapy can support trial-and-error performance in the routine estimation of metabolic flux distributions, experimental design by computer simulations of ¹³C-MFA experiments, and development of new data analysis techniques for stable isotope labeling experiments. mfapy is available to the public from the Github repository (https://github.com/fumiomatsuda/mfapy).

•

An open-source Python package, mfapy, is developed for ¹³C-MFA.

•

mfapy enables users to write Python codes for data analysis procedures of ¹³C-MFA.

•

mfapy has a flexibility and extensibility to support various data analysis procedures.

•

Computer simulations of ¹³C-MFA experiments is supported for experimental design.

Exploring the Glucose Fluxotype of the E. coli y-ome Using High-Resolution Fluxomics

We have developed a robust workflow to measure high-resolution fluxotypes (metabolic flux phenotypes) for large strain libraries under fully controlled growth conditions. This was achieved by optimizing and automating the whole high-throughput fluxomics process and integrating all relevant software tools. This workflow allowed us to obtain highly detailed maps of carbon fluxes in the central carbon metabolism in a fully automated manner. It was applied to investigate the glucose fluxotypes of 180 Escherichia coli strains deleted for y-genes. Since the products of these y-genes potentially play a role in a variety of metabolic processes, the experiments were designed to be agnostic as to their potential metabolic impact. The obtained data highlight the robustness of E. coli’s central metabolism to y-gene deletion. For two y-genes, deletion resulted in significant changes in carbon and energy fluxes, demonstrating the involvement of the corresponding y-gene products in metabolic function or regulation. This work also introduces novel metrics to measure the actual scope and quality of high-throughput fluxomics investigations.

Metabolic Flux Analysis-Linking Isotope Labeling and Metabolic Fluxes

Metabolic flux analysis (MFA) is an increasingly important tool to study metabolism quantitatively. Unlike the concentrations of metabolites, the fluxes, which are the rates at which intracellular metabolites interconvert, are not directly measurable. MFA uses stable isotope labeled tracers to reveal information related to the fluxes. The conceptual idea of MFA is that in tracer experiments the isotope labeling patterns of intracellular metabolites are determined by the fluxes, therefore by measuring the labeling patterns we can infer the fluxes in the network. In this review, we will discuss the basic concept of MFA using a simplified upper glycolysis network as an example. We will show how the fluxes are reflected in the isotope labeling patterns. The central idea we wish to deliver is that under metabolic and isotopic steady-state the labeling pattern of a metabolite is the flux-weighted average of the substrates’ labeling patterns. As a result, MFA can tell the relative contributions of converging metabolic pathways only when these pathways make substrates in different labeling patterns for the shared product. This is the fundamental principle guiding the design of isotope labeling experiment for MFA including tracer selection. In addition, we will also discuss the basic biochemical assumptions of MFA, and we will show the flux-solving procedure and result evaluation. Finally, we will highlight the link between isotopically stationary and nonstationary flux analysis.

Charting the Metabolic Landscape of the Facultative Methylotroph Bacillus methanolicus

Methanol is inexpensive, is easy to transport, and can be produced both from renewable and from fossil resources without mobilizing arable lands. As such, it is regarded as a potential carbon source to transition toward a greener industrial chemistry. Metabolic engineering of bacteria and yeast able to efficiently consume methanol is expected to provide cell factories that will transform methanol into higher-value chemicals in the so-called methanol economy. Toward that goal, the study of natural methylotrophs such as Bacillus methanolicus is critical to understand the origin of their efficient methylotrophy. This knowledge will then be leveraged to transform such natural strains into new cell factories or to design methylotrophic capability in other strains already used by the industry.

Bacillus methanolicus MGA3 is a thermotolerant and relatively fast-growing methylotroph able to secrete large quantities of glutamate and lysine. These natural characteristics make B. methanolicus a good candidate to become a new industrial chassis organism, especially in a methanol-based economy. Intriguingly, the only substrates known to support B. methanolicus growth as sole sources of carbon and energy are methanol, mannitol, and, to a lesser extent, glucose and arabitol. Because fluxomics provides the most direct readout of the cellular phenotype, we hypothesized that comparing methylotrophic and nonmethylotrophic metabolic states at the flux level would yield new insights into MGA3 metabolism. In this study, we designed and performed a ¹³C metabolic flux analysis (¹³C-MFA) of the facultative methylotroph B. methanolicus MGA3 growing on methanol, mannitol, and arabitol to compare the associated metabolic states. On methanol, results showed a greater flux in the ribulose monophosphate (RuMP) pathway than in the tricarboxylic acid (TCA) cycle, thus validating previous findings on the methylotrophy of B. methanolicus. New insights related to the utilization of cyclic RuMP versus linear dissimilation pathways and between the RuMP variants were generated. Importantly, we demonstrated that the linear detoxification pathways and the malic enzyme shared with the pentose phosphate pathway have an important role in cofactor regeneration. Finally, we identified, for the first time, the metabolic pathway used to assimilate arabitol. Overall, those data provide a better understanding of this strain under various environmental conditions.

IMPORTANCE Methanol is inexpensive, is easy to transport, and can be produced both from renewable and from fossil resources without mobilizing arable lands. As such, it is regarded as a potential carbon source to transition toward a greener industrial chemistry. Metabolic engineering of bacteria and yeast able to efficiently consume methanol is expected to provide cell factories that will transform methanol into higher-value chemicals in the so-called methanol economy. Toward that goal, the study of natural methylotrophs such as Bacillus methanolicus is critical to understand the origin of their efficient methylotrophy. This knowledge will then be leveraged to transform such natural strains into new cell factories or to design methylotrophic capability in other strains already used by the industry.

ScalaFlux: A scalable approach to quantify fluxes in metabolic subnetworks

13C-metabolic flux analysis (13C-MFA) allows metabolic fluxes to be quantified in living organisms and is a major tool in biotechnology and systems biology. Current 13C-MFA approaches model label propagation starting from the extracellular 13C-labeled nutrient(s), which limits their applicability to the analysis of pathways close to this metabolic entry point. Here, we propose a new approach to quantify fluxes through any metabolic subnetwork of interest by modeling label propagation directly from the metabolic precursor(s) of this subnetwork. The flux calculations are thus purely based on information from within the subnetwork of interest, and no additional knowledge about the surrounding network (such as atom transitions in upstream reactions or the labeling of the extracellular nutrient) is required. This approach, termed ScalaFlux for SCALAble metabolic FLUX analysis, can be scaled up from individual reactions to pathways to sets of pathways. ScalaFlux has several benefits compared with current 13C-MFA approaches: greater network coverage, lower data requirements, independence from cell physiology, robustness to gaps in data and network information, better computational efficiency, applicability to rich media, and enhanced flux identifiability. We validated ScalaFlux using a theoretical network and simulated data. We also used the approach to quantify fluxes through the prenyl pyrophosphate pathway of Saccharomyces cerevisiae mutants engineered to produce phytoene, using a dataset for which fluxes could not be calculated using existing approaches. A broad range of metabolic systems can be targeted with minimal cost and effort, making ScalaFlux a valuable tool for the analysis of metabolic fluxes.

The Design of FluxML: A Universal Modeling Language for 13C Metabolic Flux Analysis

¹³C metabolic flux analysis (MFA) is the method of choice when a detailed inference of intracellular metabolic fluxes in living organisms under metabolic quasi-steady state conditions is desired. Being continuously developed since two decades, the technology made major contributions to the quantitative characterization of organisms in all fields of biotechnology and health-related research. ¹³C MFA, however, stands out from other "-omics sciences," in that it requires not only experimental-analytical data, but also mathematical models and a computational toolset to infer the quantities of interest, i.e., the metabolic fluxes. At present, these models cannot be conveniently exchanged between different labs. Here, we present the implementation-independent model description language FluxML for specifying ¹³C MFA models. The core of FluxML captures the metabolic reaction network together with atom mappings, constraints on the model parameters, and the wealth of data configurations. In particular, we describe the governing design processes that shaped the FluxML language. We demonstrate the utility of FluxML to represent many contemporary experimental-analytical requirements in the field of ¹³C MFA. The major aim of FluxML is to offer a sound, open, and future-proof language to unambiguously express and conserve all the necessary information for model re-use, exchange, and comparison. Along with FluxML, several powerful computational tools are supplied for easy handling, but also to maintain a maximum of flexibility. Altogether, the FluxML collection is an "all-around carefree package" for ¹³C MFA modelers. We believe that FluxML improves scientific productivity as well as transparency and therewith contributes to the efficiency and reproducibility of computational modeling efforts in the field of ¹³C MFA.

Understanding metabolism with flux analysis: From theory to application

Quantitative and qualitative knowledge of metabolic rates (i.e. fluxes) over a metabolic network and in specific cellular compartments gives insights into the regulation of metabolism and helps to understand the contribution of metabolic alterations to pathology. In this review we introduce methodology to resolve metabolic fluxes from stable isotope labeling and relevant techniques in model development, model simplification, flux uncertainty analysis and experimental design that together is termed metabolic flux analysis. Finally we discuss applications using metabolic flux analysis to elucidate mechanisms pertinent to tumor cell metabolism. We hope that this review gives the readers a brief introduction of how flux analysis is conducted, how technical issues related to it are addressed, and how its application has contributed to our knowledge of tumor cell metabolism.

Multilevel interaction of the DnaK/DnaJ(HSP70/HSP40) stress-responsive chaperone machine with the central metabolism

Networks of molecular chaperones maintain cellular protein homeostasis by acting at nearly every step in the biogenesis of proteins and protein complexes. Herein, we demonstrate that the major chaperone DnaK/HSP70 of the model bacterium Escherichia coli is critical for the proper functioning of the central metabolism and for the cellular response to carbon nutrition changes, either directly or indirectly via the control of the heat-shock response. We identified carbon sources whose utilization was positively or negatively affected by DnaK and isolated several central metabolism genes (among other genes identified in this work) that compensate for the lack of DnaK and/or DnaK/Trigger Factor chaperone functions in vivo. Using carbon sources with specific entry points coupled to NMR analyses of real-time carbon assimilation, metabolic coproducts production and flux rearrangements, we demonstrate that DnaK significantly impacts the hierarchical order of carbon sources utilization, the excretion of main coproducts and the distribution of metabolic fluxes, thus revealing a multilevel interaction of DnaK with the central metabolism.

Quantification of Microbial Phenotypes

Metabolite profiling technologies have improved to generate close to quantitative metabolomics data, which can be employed to quantitatively describe the metabolic phenotype of an organism. Here, we review the current technologies available for quantitative metabolomics, present their advantages and drawbacks, and the current challenges to generate fully quantitative metabolomics data. Metabolomics data can be integrated into metabolic networks using thermodynamic principles to constrain the directionality of reactions. Here we explain how to estimate Gibbs energy under physiological conditions, including examples of the estimations, and the different methods for thermodynamics-based network analysis. The fundamentals of the methods and how to perform the analyses are described. Finally, an example applying quantitative metabolomics to a yeast model by 13C fluxomics and thermodynamics-based network analysis is presented. The example shows that (1) these two methods are complementary to each other; and (2) there is a need to take into account Gibbs energy errors. Better estimations of metabolic phenotypes will be obtained when further constraints are included in the analysis.

SUMOFLUX: A Generalized Method for Targeted 13C Metabolic Flux Ratio Analysis

Metabolic fluxes are a cornerstone of cellular physiology that emerge from a complex interplay of enzymes, carriers, and nutrients. The experimental assessment of in vivo intracellular fluxes using stable isotopic tracers is essential if we are to understand metabolic function and regulation. Flux estimation based on 13C or 2H labeling relies on complex simulation and iterative fitting; processes that necessitate a level of expertise that ordinarily preclude the non-expert user. To overcome this, we have developed SUMOFLUX, a methodology that is broadly applicable to the targeted analysis of 13C-metabolic fluxes. By combining surrogate modeling and machine learning, we trained a predictor to specialize in estimating flux ratios from measurable 13C-data. SUMOFLUX targets specific flux features individually, which makes it fast, user-friendly, applicable to experimental design and robust in terms of experimental noise and exchange flux magnitude. Collectively, we predict that SUMOFLUX's properties realistically pave the way to high-throughput flux analyses.

Investigation of useful carbon tracers for 13C-metabolic flux analysis of Escherichia coli by considering five experimentally determined flux distributions

The ¹³C-MFA experiments require an optimal design since the precision or confidence intervals of the estimated flux levels depends on factors such as the composition of ¹³C-labeled carbon sources, as well as the metabolic flux distribution of interest. In this study, useful compositions of ¹³C-labeled glucose for ¹³C-metabolic flux analysis (¹³C-MFA) of Escherichia coli are investigated using a computer simulation of the stable isotope labeling experiment. Following the generation of artificial mass spectra datasets of amino acid fragments using five literature-reported flux distributions of E. coli, the best fitted flux distribution and the 95% confidence interval were estimated by the ¹³C-MFA procedure. A comparison of the precision scores showed that [1, 2-¹³C]glucose and a mixture of [1-¹³C] and [U-¹³C]glucose at 8:2 are one of the best carbon sources for a precise estimation of flux levels of the pentose phosphate pathway, glycolysis and the TCA cycle. Although the precision scores of the anaplerotic and glyoxylate pathway reactions were affected by both the carbon source and flux distribution, it was also shown that the mixture of non-labeled, [1-¹³C], and [U-¹³C]glucose at 4:1:5 was specifically effective for the flux estimation of the glyoxylate pathway reaction. These findings were confirmed by wet ¹³C-MFA experiments.

•

Useful compositions of ¹³C-labeled glucose are investigated for ¹³C-MFA of E. coli.

•

Computer simulations revealed that [1,2-¹³C] was one of the best first choices.

•

Mixture of non-labeled, [1-¹³C] and [U-¹³C] at 0:8:2 was also suitable for ¹³C-MFA.

•

Mixture at 4:1:5 was specifically effective for estimation of glyoxylate pathway.

•

The wet ¹³C-MFA experiments of E. coli confirmed the findings.

Impact of kinetic isotope effects in isotopic studies of metabolic systems

Background

Isotope labeling experiments (ILEs) are increasingly used to investigate the functioning of metabolic systems. Some enzymes are subject to kinetic isotope effects (KIEs) which modulate reaction rates depending on the isotopic composition of their substrate(s). KIEs may therefore affect both the propagation of isotopes through metabolic networks and their operation, and ultimately jeopardize the biological value of ILEs. However, the actual impact of KIEs on metabolism has never been investigated at the system level.

Results

First, we developed a framework which integrates KIEs into kinetic and isotopic models of metabolism, thereby accounting for their system-wide effects on metabolite concentrations, metabolic fluxes, and isotopic patterns. Then, we applied this framework to assess the impact of KIEs on the central carbon metabolism of Escherichia coli in the context of ¹³C-ILEs, under different situations commonly encountered in laboratories. Results showed that the impact of KIEs strongly depends on the label input and on the variable considered but is significantly lower than expected intuitively from measurements on isolated enzymes. The global robustness of both the metabolic operation and isotopic patterns largely emerge from intrinsic properties of metabolic networks, such as the distribution of control across the network and bidirectional isotope exchange.

Conclusions

These results demonstrate the necessity of investigating the impact of KIEs at the level of the entire system, contradict previous hypotheses that KIEs would have a strong effect on isotopic distributions and on flux determination, and strengthen the biological value of ¹³C-ILEs. The proposed modeling framework is generic and can be used to investigate the impact of all the isotopic tracers (²H, ¹³C, ¹⁵N, ¹⁸O, etc.) on different isotopic datasets and metabolic systems. By allowing the integration of isotopic and metabolomics data collected under stationary and/or non-stationary conditions, it may also assist interpretations of ILEs and facilitate the development of more accurate kinetic models with improved explicative and predictive capabilities.

Electronic supplementary material

The online version of this article (doi:10.1186/s12918-015-0213-8) contains supplementary material, which is available to authorized users.

White biotechnology: State of the art strategies for the development of biocatalysts for biorefining

White biotechnology is a term that is now often used to describe the implementation of biotechnology in the industrial sphere. Biocatalysts (enzymes and microorganisms) are the key tools of white biotechnology, which is considered to be one of the key technological drivers for the growing bioeconomy. Biocatalysts are already present in sectors such as the chemical and agro-food industries, and are used to manufacture products as diverse as antibiotics, paper pulp, bread or advanced polymers. This review proposes an original and global overview of highly complementary fields of biotechnology at both enzyme and microorganism level. A certain number of state of the art approaches that are now being used to improve the industrial fitness of biocatalysts particularly focused on the biorefinery sector are presented. The first part deals with the technologies that underpin the development of industrial biocatalysts, notably the discovery of new enzymes and enzyme improvement using directed evolution techniques. The second part describes the toolbox available by the cell engineer to shape the metabolism of microorganisms. And finally the last part focuses on the 'omic' technologies that are vital for understanding and guide microbial engineering toward more efficient microbial biocatalysts. Altogether, these techniques and strategies will undoubtedly help to achieve the challenging task of developing consolidated bioprocessing (i.e. CBP) readily available for industrial purpose.

Tracking the metabolic pulse of plant lipid production with isotopic labeling and flux analyses: Past, present and future

Metabolism is comprised of networks of chemical transformations, organized into integrated biochemical pathways that are the basis of cellular operation, and function to sustain life. Metabolism, and thus life, is not static. The rate of metabolites transitioning through biochemical pathways (i.e., flux) determines cellular phenotypes, and is constantly changing in response to genetic or environmental perturbations. Each change evokes a response in metabolic pathway flow, and the quantification of fluxes under varied conditions helps to elucidate major and minor routes, and regulatory aspects of metabolism. To measure fluxes requires experimental methods that assess the movements and transformations of metabolites without creating artifacts. Isotopic labeling fills this role and is a long-standing experimental approach to identify pathways and quantify their metabolic relevance in different tissues or under different conditions. The application of labeling techniques to plant science is however far from reaching it potential. In light of advances in genetics and molecular biology that provide a means to alter metabolism, and given recent improvements in instrumentation, computational tools and available isotopes, the use of isotopic labeling to probe metabolism is becoming more and more powerful. We review the principal analytical methods for isotopic labeling with a focus on seminal studies of pathways and fluxes in lipid metabolism and carbon partitioning through central metabolism. Central carbon metabolic steps are directly linked to lipid production by serving to generate the precursors for fatty acid biosynthesis and lipid assembly. Additionally some of the ideas for labeling techniques that may be most applicable for lipid metabolism in the future were originally developed to investigate other aspects of central metabolism. We conclude by describing recent advances that will play an important future role in quantifying flux and metabolic operation in plant tissues.

OpenFLUX2: (13)C-MFA modeling software package adjusted for the comprehensive analysis of single and parallel labeling experiments

BACKGROUND

Steady-state (13)C-based metabolic flux analysis ((13)C-MFA) is the most powerful method available for the quantification of intracellular fluxes. These analyses include concertedly linked experimental and computational stages: (i) assuming the metabolic model and optimizing the experimental design; (ii) feeding the investigated organism using a chosen (13)C-labeled substrate (tracer); (iii) measuring the extracellular effluxes and detecting the (13)C-patterns of intracellular metabolites; and (iv) computing flux parameters that minimize the differences between observed and simulated measurements, followed by evaluating flux statistics. In its early stages, (13)C-MFA was performed on the basis of data obtained in a single labeling experiment (SLE) followed by exploiting the developed high-performance computational software. Recently, the advantages of parallel labeling experiments (PLEs), where several LEs are conducted under the conditions differing only by the tracer(s) choice, were demonstrated, particularly with regard to improving flux precision due to the synergy of complementary information. The availability of an open-source software adjusted for PLE-based (13)C-MFA is an important factor for PLE implementation.

RESULTS

The open-source software OpenFLUX, initially developed for the analysis of SLEs, was extended for the computation of PLE data. Using the OpenFLUX2, in silico simulation confirmed that flux precision is improved when (13)C-MFA is implemented by fitting PLE data to the common model compared with SLE-based analysis. Efficient flux resolution could be achieved in the PLE-mediated analysis when the choice of tracer was based on an experimental design computed to minimize the flux variances from different parts of the metabolic network. The analysis provided by OpenFLUX2 mainly includes (i) the optimization of the experimental design, (ii) the computation of the flux parameters from LEs data, (iii) goodness-of-fit testing of the model's adequacy, (iv) drawing conclusions concerning the identifiability of fluxes and construction of a contribution matrix reflecting the relative contribution of the measurement variances to the flux variances, and (v) precise determination of flux confidence intervals using a fine-tunable and convergence-controlled Monte Carlo-based method.

CONCLUSIONS

The developed open-source OpenFLUX2 provides a friendly software environment that facilitates beginners and existing OpenFLUX users to implement LEs for steady-state (13)C-MFA including experimental design, quantitative evaluation of flux parameters and statistics.

(13)C metabolic flux analysis of recombinant expression hosts

Identifying host cell metabolic phenotypes that promote high recombinant protein titer is a major goal of the biotech industry. (13)C metabolic flux analysis (MFA) provides a rigorous approach to quantify these metabolic phenotypes by applying isotope tracers to map the flow of carbon through intracellular metabolic pathways. Recent advances in tracer theory and measurements are enabling more information to be extracted from (13)C labeling experiments. Sustained development of publicly available software tools and standardization of experimental workflows is simultaneously encouraging increased adoption of (13)C MFA within the biotech research community. A number of recent (13)C MFA studies have identified increased citric acid cycle and pentose phosphate pathway fluxes as consistent markers of high recombinant protein expression, both in mammalian and microbial hosts. Further work is needed to determine whether redirecting flux into these pathways can effectively enhance protein titers while maintaining acceptable glycan profiles.

Sampling of intracellular metabolites for stationary and non-stationary (13)C metabolic flux analysis in Escherichia coli

The analysis of metabolic intermediates is a rich source of isotopic information for (13)C metabolic flux analysis ((13)C-MFA) and extends the range of its applications. The sampling of labeled metabolic intermediates is particularly important to obtain reliable isotopic information. The assessment of the different sampling procedures commonly used to generate such data, therefore, is crucial. In this work, we thoroughly evaluated several sampling procedures for stationary and non-stationary (13)C-MFA using Escherichia coli. We first analyzed the efficiency of these procedures for quenching metabolism and found that procedures based on cold or boiling solvents are reliable, in contrast to fast filtration, which is not. We also showed that separating the cells from the broth is not necessary in isotopic stationary state conditions. On the other hand, we demonstrated that the presence of metabolic intermediates outside the cells strongly affects the transient isotopic data monitored during non-stationary (13)C-labeling experiments. Meaningful isotopic data can be obtained by recovering intracellular labeled metabolites from pellets of cells centrifuged in cold solvent. We showed that if the intracellular pools are not separated from the extracellular ones, accurate flux maps can be established provided that the contribution of exogenous compounds is taken into account in the metabolic flux model.