Experimental determination of Escherichia coli biomass composition for constraint-based metabolic modeling

PLOS ONE RESEARCH ARTICLE Experimental determination of Escherichia coli biomass composition for constraint-based metabolic modeling Vetle Simensen ID1☯, Christian Schulz ID1☯, Emil Karlsen1☯, Signe Bråtelund1‡, Idun Burgos1‡, Lilja Brekke Thorfinnsdottir1, Laura Garcı́a-Calvo1, Per Bruheim1, Eivind Almaas1,2* a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Biotechnology and Food Science, NTNU - Norwegian University of Science and Technology, Trondheim, Norway, 2 K.G. Jebsen Center for Genetic Epidemiology Department of Public Health and General Practice, NTNU - Norwegian University of Science and Technology, Trondheim, Norway ☯ These authors contributed equally to this work. ‡ SB and IB also contributed equally to this work. * Abstract OPEN ACCESS Citation: Simensen V, Schulz C, Karlsen E, Bråtelund S, Burgos I, Thorfinnsdottir LB, et al. (2022) Experimental determination of Escherichia coli biomass composition for constraint-based metabolic modeling. PLoS ONE 17(1): e0262450. https://doi.org/10.1371/journal.pone.0262450 Editor: Chen-Guang Liu, Shanghai Jiao Tong University, CHINA Received: August 20, 2021 Accepted: December 24, 2021 Published: January 27, 2022 Copyright: © 2022 Simensen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the manuscript and its Supporting information files. Funding: This work was funded by the following grants: European Research Area in Industrial Biotechnology 2 (https://era-learn.eu/networkinformation/networks/era-ib-2), ERA-IB-2, grant number 271585 (CS); The Research Council of Norway (https://prosjektbanken.forskningsradet. no/en/project/FORISS/294605), grant number 294605 (CS); The Research Council of Norway Genome-scale metabolic models (GEMs) are mathematical representations of metabolism that allow for in silico simulation of metabolic phenotypes and capabilities. A prerequisite for these predictions is an accurate representation of the biomolecular composition of the cell necessary for replication and growth, implemented in GEMs as the so-called biomass objective function (BOF). The BOF contains the metabolic precursors required for synthesis of the cellular macro- and micromolecular constituents (e.g. protein, RNA, DNA), and its composition is highly dependent on the particular organism, strain, and growth condition. Despite its critical role, the BOF is rarely constructed using specific measurements of the modeled organism, drawing the validity of this approach into question. Thus, there is a need to establish robust and reliable protocols for experimental condition-specific biomass determination. Here, we address this challenge by presenting a general pipeline for biomass quantification, evaluating its performance on Escherichia coli K-12 MG1655 sampled during balanced exponential growth under controlled conditions in a batch-fermentor set-up. We significantly improve both the coverage and molecular resolution compared to previously published workflows, quantifying 91.6% of the biomass. Our measurements display great correspondence with previously reported measurements, and we were also able to detect subtle characteristics specific to the particular E. coli strain. Using the modified E. coli GEM iML1515a, we compare the feasible flux ranges of our experimentally determined BOF with the original BOF, finding that the changes in BOF coefficients considerably affect the attainable fluxes at the genome-scale. Introduction The increasing availability of large-scale omics data has propelled the study of complex biological systems, pushing the field of systems biology to the forefront of cutting-edge biological PLOS ONE | https://doi.org/10.1371/journal.pone.0262450 January 27, 2022 1 / 17 PLOS ONE (https://prosjektbanken.forskningsradet.no/en/ project/FORISS/269084), grant number 269084 (EK); The Research Council of Norway (https:// prosjektbanken.forskningsradet.no/en/project/ FORISS/269432), grant number 269432 (VS); VS, LBT, and LGC were also supported by NTNU internal grants. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Experimental determination E. coli biomass composition research [1, 2]. Central to this development is the realization that biology is best understood not merely by considering its individual constituents, but rather investigating the emergent properties of the system as a whole. One of the predominant subfields of systems biology is the in silico study of metabolism using genome-scale metabolic models (GEMs) [3–6]. Here, the genetically encoded metabolic potential of an organism is used to construct a stoichiometric network of biochemical transformations. The steady-state fluxes of the metabolic system can subsequently be calculated using approaches such as flux balance analysis [7]. These flux phenotypes are usually computed by maximizing growth, assuming optimal biomass production to be a reasonable cellular objective [8]. Growth in these models is implemented as a pseudo-reaction called the biomass objective function (BOF), whose reactants are the metabolic precursors needed to generate the molecular constituents of the cell. By appropriately scaling the stoichiometric coefficients of these precursors using experimental biomass measurements, the flux through the BOF directly corresponds to the specific growth rate, allowing for quantitative predictions of growth phenotypes [9, 10]. While experimental data on the condition-dependent biomass compositions of some well-studied organisms are available [11, 12], this is commonly lacking for most organisms. The usual strategy has therefore been to either adopt existing organism-specific biomass compositions from different conditions or employ parts or the whole composition from another organism entirely [13, 14]. This, however, is a sub-optimal approach as the biomass composition depends on the particular organism and strain [15, 16]. The biomass composition is also not static, but is rather continually adjusted in response to changing environmental conditions [12]. In many instances, the predicted flux phenotypes of these models have been shown to be highly susceptible to variations in the biomass composition [17]. Dikicioglu et al. [18] demonstrated how the predicted flux distributions of a Saccharomyces cerevisiae GEM were sensitive to changes in the stoichiometric coefficients of the BOF within experimentally determined bounds. Lakshmanan et al. [17] observed a similar sensitivity when varying the biomass composition in GEMs of eight different yeast species, showcasing the impact on both growth rate and gene e (...truncated)