A dynamic model-based preparation of uniformly-C-labeled internal standards facilitates quantitative metabolomics analysis of Penicillium chrysogenum.

Targeted, quantitative metabolomics can, in principle, provide precise information on intracellular metabolite levels, which can be applied to accurate modeling of intracellular processes required in systems biology and metabolic engineering. However, quantitative metabolite profiling is often hampered by biased mass spectrometry-based analyses caused by matrix effects, the degradation of metabolites and metabolite leakage during sample preparation, and unexpected variation in instrument responses. Isotope Dilution Mass Spectrometry (IDMS) has been proven as the most accurate method for high-throughput detection of intracellular metabolite concentrations, and the key has been the acquisition of the corresponding fully uniformly (U) -C-labeled metabolites to be measured. Here, we have prepared U-C-labeled cell extracts by cultivating P. chrysogenum in a fed-batch fermentation with fully U-C-labeled substrates. Towards this goal, a dynamic fed-batch model describing P. chrysogenum growth and penicillin production was used to simulate the fermentation process and design the fed-batch fermentation media. Further, a case study with extensive intracellular metabolomics data from glucose-limited cultivation of Penicillium chrysogenum under both single and repetitive glucose pulses was illustrated by using the IDMS methods with the prepared U-C-labeled cell extracts as internal standards. In conclusion, the IDMS method can be incorporated into well-established fast sampling and quenching protocols to obtain dynamic quantitative in vivo metabolome data at the timescales of (tens of) seconds and elucidate the underlying regulatory architecture. The case study revealed gross differences between single and repeated pulses, which suggests that single pulse studies have limited value for understanding of metabolic responses in large-scale bioreactors. Instead, intermittent feeding should be favored.