Bayesian multi-model-based ¹³C¹⁵N-metabolic flux analysis quantifies carbon-nitrogen metabolism in mycobacteria

Abstract

Metabolic flux is the final output of cellular regulation and has been extensively studied for carbon but much less is known about nitrogen, which is another important building block for living organisms. For the pathogen Mycobacterium tuberculosis (Mtb) this is particularly important in informing the development of effective drugs targeting metabolism of the pathogen. Here we performed 13C15N dual isotopic labelling of mycobacterial steady state cultures and quantified intracellular carbon-nitrogen (CN) and nitrogen (N) fluxes in addition to carbon (C) fluxes and inferred their reaction bidirectionalities. The combination of 13C15N-MFA with a Bayesian multi-model approach allowed us to resolve C and N fluxes simultaneously which was not possible with classical 13C-MFA. We quantified CN fluxes for amino acid and for the first time nucleotide biosynthesis. Our analysis identified glutamate as the central CN and N node in mycobacteria and improved resolution of the anaplerotic node. Our study describes a powerful platform to measure carbon and nitrogen metabolism in any biological system with statistical rigor.

Elucidating dynamic anaerobe metabolism with Live Cell HRMAS 13C NMR and genome-scale metabolic modeling

“…Metabolic modeling systems link experimentally-obtained substrate and metabolite fluxes to cellular pathways and genes. By design, metabolic flux analyses (MFA) informed by 13 C-NMR have remained limited to pathways carrying 13 C flux 7-10 . In contrast, dynamic flux balance analysis (dFBA) simulates time-dependent recruitment of metabolic pathways on an organismal scale given a set of nutrient exchange constraints and a biological objective such as biomass or ATP production 11 .…”

SIMPEL: using stable isotopes to elucidate dynamics of context specific metabolism

“…Because HRMS can measure many m/z values quickly, the metabolic pathways that are actively incorporating isotopes can be identified through an unsupervised, unbiased approach with untargeted data acquisition that is blind to the metabolic network. Enumerating and quantifying all metabolomic features detected in samples enriched with multiple stable isotopes remains an analytical bottleneck 5,6 , since the number of multivariate isotopologues 3 (I) increases exponentially with the number of labeled atoms (N) for each element (e) of the molecular formula, following equation ( 1).…”

“…Cellular metabolic phenotypes are defined by fluxes that is evaluated by tracking the movement of isotopes over time through biochemical pathways, most frequently involving 13 C. However, the complexity of metabolism often requires more than one labeling experiment and different 13 C substrates or alternatively 2 H, 15 N or 18 O to resolve pathway use [1][2][3] . The time-consuming manual identification and integration of isotopologues in labeling experiments is not compatible with the extensive number of replicates required for statistically meaningful results (e.g., genetic studies) and is prone to error.…”