Parallel labeling experiments: a novel approach for validating metabolic network models

Jennifer Au, Maciek R. Antoniewicz
Department of Chemical and Biomolecular Engineering, University of Delaware
150 Academy Street, Colburn Laboratory, Newark, DE 19716

Abstract

Metabolic network models play a central role in metabolic engineering applications. However,
these models are often not rigorously validated, but rather based on automated annotations with limited experimental supporting data. Here, we present a novel comprehensive experimental and computational framework for validating metabolic network models using parallel labeling experiments, in combination with 13C metabolic flux analysis (13C-MFA) and statistical analysis. We have applied the methodology to elucidate core metabolism of Clostridium acetobutylicum. After multiple rounds of model testing and updating we present a new validated metabolic network model for C. acetobutylicum that can be used for unbiased metabolic flux measurements using 13C-MFA. The model was used to study the metabolic stress response of C. acetobutylicum under butanol and butyrate stresses.
Although the biochemistry of C. acetobutylicum has been extensively reviewed, the central metabolic pathways have remained only partially resolved. Two recent reconstructions of genome-scale models have proposed different mechanisms for the biosynthesis of α- ketoglutarate, the precursor for glutamate, glutamine and proline. Initial stable-isotope labeling experiments and qualitative 13C-isotopomer analysis have supported the idea of an incomplete TCA cycle and suggested a Re-stereospecificity for the citrate synthase reaction. However, quantitative analysis of metabolic fluxes had not been performed. In this work, we have rigorously validated, for the first time, the metabolic model of C. acetobutylicum. Using the novel parallel labeling experiments approach we quantitatively elucidated core metabolism (including amino acid metabolism) of C. acetobutylicum. Contrary to previously proposed hypotheses, we found that while the TCA cycle runs in the oxidative direction, there is no notable flux between α-ketoglutarate and succinyl-CoA or succinate and fumarate, and that the conversion of succinyl-CoA to succinate proceeds independently. Using multiple 13C-labeled amino acid tracers, we additionally showed that there is no flux between malate and oxaloacetate, and that there exists a previously unknown metabolic cycle where carbon flows from aspartate to threonine, serine, pyruvate, oxaloacetate and back to aspartate. Finally, we identified a putative citramalate synthase gene that is in fact the first step in isoleucine biosynthesis in C. acetobutylicum.
The validated metabolic network model was used to measure metabolic fluxes in C. acetobutylicum under butanol and butyrate stress. The metabolic flux distributions resulting from this analysis provide the basis for ongoing work on improving the tolerance of C. acetobutylicum to these fermentation products. Results from this ongoing effort will be presented. We note that the systematic approach employed in this study can be easily applied to elucidate metabolism of other poorly characterized organisms that are of importance to metabolic engineering.