Computational Systems Biochemistry Group

The fundamental function of metabolic networks is to provide the biological system with energy and building blocks. Depending on the prevailing metabolic objectives (e.g. detoxification of a toxin or accumulation of biomass), the distribution of fluxes to fulfill this function in the metabolic network may vary considerably. The study of this variation, however, would be informative for the basic understanding of the metabolic principles, limitations and its robust behaviour with respect to perturbations. The fact that experimental measurements of internal fluxes may cover only a tiny fraction of the known metabolic network, results in both, the requirement of flux predictions by constraint-based optimization approaches, and the lack of comprehensive validation of these predictions by experimentally measured internal fluxes. Without this validation, the predictions remain somehow hypothetical as the optimization principles underlying metabolic conversions are unknown.

On the other hand, there is a rapidly growing wealth of modern high-throughput technologies such as RNA- or DNA micro-arrays, 2D gel electrophoresis, protein chips, and mass-spectroscopy of small organic molecules that have paved the way for taking "molecular photographs" of a cell, i.e. to asses the levels of mRNAs, proteins and metabolites at certain time points and under varying external conditions (Hsiao & Kuo, 2006; McKenzie, 2001). The question arises to what extend such molecular photographs can be used to estimate accompanying changes of metabolic fluxes and from this to conclude on changes of the metabolic objectives of the cell - the ultimate goal of cellular regulation. So far, the usability of these high-throughput data has been limited by the fact that there is no simple 1:1 relationship neither between the mRNA level and the level of corresponding expressed gene product (Gygi et al., 1999; Greenbaum et al., 2003), nor between the expression level of an enzyme and the flux rate of the catalyzed reaction (Famili et al., 2003; ter Kuile & Westerhoff, 2001; Rossell et al, 2006). The latter relationship is a focus of our work. Specifically, the following four questions have been addressed using mathematical modeling :

(i) Is it possible to predict the functional state of a cellular system based on experimentally determined gene expression profiles?

(ii) Is it possible to computationally predict profiles of enzyme activity/gene expression required for the fulfillment of predefined metabolic functions?

(iii) Is it possible to identify modules consisting of reactions that simultaneously are active or inactive in different flux distributions?

(iv) How can the robustness of a cellular reaction network be quantitatively determined with respect to missing catalytic activities (= loss of function mutation in the gene of the respective enzyme)?

In summary, we develope flexible methods that allow the exploitation of available information on changes of enzyme levels, metabolic functions, constrained reversibility of a reaction, and external substrates for predicting changes of stationary fluxes. These predictions include predictions on changes in the metabolic output that point out to the functional status of a cell or a tissue (Manuscript in press: Hoffmann & Holzhütter, 2008).

Researchers

Sabrina Hoffmann
Prof. Hermann-Georg Holzhütter

References

Famili I, Forster J, Nielsen J, Palsson BO. (2003) Saccharomyces cerevisiae phenotypes can be predicted by using constraint-based analysis of a genome-scale reconstructed metabolic network. Proc Natl Acad Sci U S A., 11;100(23):13134-9. [PubMed]

Greenbaum D, Colangelo C, Williams K, Gerstein M. (2003) Comparing protein abundance and mRNA expression levels on a genomic scale. Genome Biol., 4(9):117. [PubMed]

Gygi SP, Rochon Y, Franza BR, Aebersold R. (1999) Correlation between protein and mRNA abundance in yeast. Mol Cell Biol., 19(3):1720-30. [PubMed]

Hoffmann S and Holzhütter HG. (2008) Uncovering metabolic objectives pursued by changes of enzyme levels. Ann N Y Acad Sci., 1158:57-70. [PubMed]

Hsiao A, Kuo MD. (2006) High-throughput biology in the postgenomic era. J Vasc Interv Radiol., 17(7):1077-85. [PubMed]

MacKenzie S. (2001) High-throughput interpretation of pathways and biology. Drug News Perspect., 14(1):54-7. [PubMed]

Rossell S, van der Weijden CC, Lindenbergh A, van Tuijl A, Francke C, Bakker BM, Westerhoff HV. (2006) Unraveling the complexity of flux regulation: a new method demonstrated for nutrient starvation in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A., 14;103(7):2166-71. [PubMed]

ter Kuile BH, Westerhoff HV. (2001) Transcriptome meets metabolome: hierarchical and metabolic regulation of the glycolytic pathway. FEBS Lett., 6;500(3):169-71. [PubMed]

		Exploitation of High-Throughput Data to Increase the Reliability of Metabolic Flux Predictions
Home Research VirtualLiver People Publications Software Positions Events Contact		The fundamental function of metabolic networks is to provide the biological system with energy and building blocks. Depending on the prevailing metabolic objectives (e.g. detoxification of a toxin or accumulation of biomass), the distribution of fluxes to fulfill this function in the metabolic network may vary considerably. The study of this variation, however, would be informative for the basic understanding of the metabolic principles, limitations and its robust behaviour with respect to perturbations. The fact that experimental measurements of internal fluxes may cover only a tiny fraction of the known metabolic network, results in both, the requirement of flux predictions by constraint-based optimization approaches, and the lack of comprehensive validation of these predictions by experimentally measured internal fluxes. Without this validation, the predictions remain somehow hypothetical as the optimization principles underlying metabolic conversions are unknown. On the other hand, there is a rapidly growing wealth of modern high-throughput technologies such as RNA- or DNA micro-arrays, 2D gel electrophoresis, protein chips, and mass-spectroscopy of small organic molecules that have paved the way for taking "molecular photographs" of a cell, i.e. to asses the levels of mRNAs, proteins and metabolites at certain time points and under varying external conditions (Hsiao & Kuo, 2006; McKenzie, 2001). The question arises to what extend such molecular photographs can be used to estimate accompanying changes of metabolic fluxes and from this to conclude on changes of the metabolic objectives of the cell - the ultimate goal of cellular regulation. So far, the usability of these high-throughput data has been limited by the fact that there is no simple 1:1 relationship neither between the mRNA level and the level of corresponding expressed gene product (Gygi et al., 1999; Greenbaum et al., 2003), nor between the expression level of an enzyme and the flux rate of the catalyzed reaction (Famili et al., 2003; ter Kuile & Westerhoff, 2001; Rossell et al, 2006). The latter relationship is a focus of our work. Specifically, the following four questions have been addressed using mathematical modeling : (i) Is it possible to predict the functional state of a cellular system based on experimentally determined gene expression profiles? (ii) Is it possible to computationally predict profiles of enzyme activity/gene expression required for the fulfillment of predefined metabolic functions? (iii) Is it possible to identify modules consisting of reactions that simultaneously are active or inactive in different flux distributions? (iv) How can the robustness of a cellular reaction network be quantitatively determined with respect to missing catalytic activities (= loss of function mutation in the gene of the respective enzyme)? In summary, we develope flexible methods that allow the exploitation of available information on changes of enzyme levels, metabolic functions, constrained reversibility of a reaction, and external substrates for predicting changes of stationary fluxes. These predictions include predictions on changes in the metabolic output that point out to the functional status of a cell or a tissue (Manuscript in press: Hoffmann & Holzhütter, 2008). Researchers Sabrina Hoffmann Prof. Hermann-Georg Holzhütter References Famili I, Forster J, Nielsen J, Palsson BO. (2003) Saccharomyces cerevisiae phenotypes can be predicted by using constraint-based analysis of a genome-scale reconstructed metabolic network. Proc Natl Acad Sci U S A., 11;100(23):13134-9. [PubMed] Greenbaum D, Colangelo C, Williams K, Gerstein M. (2003) Comparing protein abundance and mRNA expression levels on a genomic scale. Genome Biol., 4(9):117. [PubMed] Gygi SP, Rochon Y, Franza BR, Aebersold R. (1999) Correlation between protein and mRNA abundance in yeast. Mol Cell Biol., 19(3):1720-30. [PubMed] Hoffmann S and Holzhütter HG. (2008) Uncovering metabolic objectives pursued by changes of enzyme levels. Ann N Y Acad Sci., 1158:57-70. [PubMed] Hsiao A, Kuo MD. (2006) High-throughput biology in the postgenomic era. J Vasc Interv Radiol., 17(7):1077-85. [PubMed] MacKenzie S. (2001) High-throughput interpretation of pathways and biology. Drug News Perspect., 14(1):54-7. [PubMed] Rossell S, van der Weijden CC, Lindenbergh A, van Tuijl A, Francke C, Bakker BM, Westerhoff HV. (2006) Unraveling the complexity of flux regulation: a new method demonstrated for nutrient starvation in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A., 14;103(7):2166-71. [PubMed] ter Kuile BH, Westerhoff HV. (2001) Transcriptome meets metabolome: hierarchical and metabolic regulation of the glycolytic pathway. FEBS Lett., 6;500(3):169-71. [PubMed]