To reliably determine intracellular metabolite concentrations in microorganisms, accurate sampling and sample inactivation strategies are crucial. Here, we present a method for automated fast filtration and on-filter quenching of microbial samples to overcome metabolite leakage induced by cold shock and significantly reduce the sampling and treatment time compared to manual filtration methods. The whole process of sampling, sample filtration, filter wash, and quenching of the filter with liquid nitrogen was finished in less than 6-15 s, depending on the experimental setup. By integration into an automated fast sampling device, we compared our method to the conventional methanol quenching method and showed that intracellular amino acid contents in Escherichia coli were significantly increased (≥75%) with our fast filtration and on-filter quenching method. Furthermore, we investigated different filter types for the fast filtration and the efficiency of metabolite extraction from cells held on filters. Additionally, we found that the fast filtration behaves considerably different during exponential and nonexponential growth, probably due to variations of cell morphologies. Overall, we demonstrated that the automation of the fast filtration method significantly reduces the time for filtration and quenching and hence enlarge the number of metabolites that can be quantified with this leakagefree sampling method.

Engineering of enzymes and pathways is generally required for the development of efficient strains for bioproduction processes. To this end, quantitative and reliable data of intracellular metabolites are highly desired, but often not available, especially for conditions more close to industrial applications, i.e. at high cell density and product concentration. Here, we investigated the intracellular metabolite profiles of an engineered l‐lysine‐producing Corynebacterium glutamicum strain and the corresponding wild‐type strain to assess the impacts of deregulation of product inhibition of the key enzymes aspartate kinase and phosphoenolpyruvate carboxylase and to identify potentials for their further improvement. A bioreactor system with automated fast‐sampling, filtration and on‐filter quenching of the metabolism was used for a more reliable determination of intracellular metabolites in batch cultures with optical cell density (OD660) up to 40. The l‐lysine‐producing strain showed substantially different metabolite profiles in the amino acid metabolism, including increased intracellular pool sizes in the l‐lysine‐, l‐homoserine‐ and l‐threonine pathways and decreased intracellular pool sizes for all other determined amino acids. By comparing data of in vitro inhibition of the engineered enzymes and determined intracellular concentrations of the inhibitors it was found that the inferred in vivo activities of these enzymes are still significantly below their in vitro maximums. This work demonstrates the usefulness of metabolic analysis for assessing the impact of engineered enzymes and identifying targets for further strain development.