Genome-Scale Metabolic Model of Caldicellulosiruptor bescii Reveals Optimal Metabolic Engineering Strategies for Bio-based Chemical Production

Abstract

Metabolic modeling was used to examine potential bottlenecks that could be encountered for metabolic engineering of the cellulolytic extreme thermophile Caldicellulosiruptor bescii to produce bio-based chemicals from plant biomass. The model utilizes subsystems-based genome annotation, targeted reconstruction of carbohydrate utilization pathways, and biochemical and physiological experimental validations. Specifically, carbohydrate transport and utilization pathways involving 160 genes and their corresponding functions were incorporated, representing the utilization of C5/C6 monosaccharides, disaccharides, and polysaccharides such as cellulose and xylan. To illustrate its utility, the model predicted that optimal production from biomass-based sugars of the model product, ethanol, was driven by ATP production, redox balancing, and proton translocation, mediated through the interplay of an ATP synthase, a membrane-bound hydrogenase, a bifurcating hydrogenase, and a bifurcating NAD- and NADP-dependent oxidoreductase. These mechanistic insights guided the design and optimization of new engineering strategies for product optimization, which were subsequently tested in the C. bescii model, showing a nearly 2-fold increase in ethanol yields. The C. bescii model provides a useful platform for investigating the potential redox controls that mediate the carbon and energy flows in metabolism and sets the stage for future design of engineering strategies aiming at optimizing the production of ethanol and other bio-based chemicals. IMPORTANCE The extremely thermophilic cellulolytic bacterium, Caldicellulosiruptor bescii, degrades plant biomass at high temperatures without any pretreatments and can serve as a strategic platform for industrial applications. The metabolic engineering of C. bescii, however, faces potential bottlenecks in bio-based chemical productions. By simulating the optimal ethanol production, a complex interplay between redox balancing and the carbon and energy flow was revealed using a C. bescii genome-scale metabolic model. New engineering strategies were designed based on an improved mechanistic understanding of the C. bescii metabolism, and the new designs were modeled under different genetic backgrounds to identify optimal strategies. The C. bescii model provided useful insights into the metabolic controls of this organism thereby opening up prospects for optimizing production of a wide range of bio-based chemicals.

Keywords: Caldicellulosiruptor; bio-based chemical production; central carbon metabolism; metabolic engineering; metabolic modeling; redox balance.

FIG 1

Experiment-based model validation using experimental measurements from WT and Δldh using glucose and fructose as sole carbon sources (5, 15). (A) Growth yield validation; (B) product yield validation. A dashed line is included in each panel to mark the 1:1 comparison between experimental and computational data. Error bars represent the standard deviations among experimental replicates.

FIG 2

Pathway diagram showing metabolic reactions related to the optimization of ethanol production in C. bescii. The solid black arrows indicate essential metabolic reactions in the C. bescii model, while the dotted black arrows indicate alternative reactions in the model. The blue arrows indicate reactions that are needed to enable ethanol production. The red arrows indicate reactions that are required for the optimization of ethanol production. The red crosses indicate reactions that should be blocked for the optimization of ethanol production. Functions essential for the optimization of ethanol production are highlighted in yellow background, and engineered functions in the E1 strain of C. bescii are on a green background. Metabolic flux ranges are shown for selected functions, with gray representing the flux range under the no-ethanol simulation, blue representing half-maximum ethanol simulation, and red representing the maximum ethanol simulation. Underlined compounds indicate products of metabolic reactions when ethanol production approaches the maximum. Full names of the abbreviated compounds and enzymes can be found in Table S1 in the supplemental material.

FIG 3

Simulation of maximum ethanol productions for the evaluation of diverse engineering designs. (A) Simulation of ethanol production by the E1 strain with BF-H₂ase as either reversible (black) or H2-producing only (red). (B) Simulation of engineering designs with the BF-H₂ase constrained to the H₂-producing direction. Strain E1M was proposed over the previous engineering design of E1 as a base strain. Further engineering designs were performed with insertion or deletion of genes on the E1M. The symbol “–” indicates the removal of a function, while the symbol “+” indicates the insertion of a function into the E1M model. The red dashed line indicates the experimentally measured ethanol production by the E1 strain.

FIG 4

(A to D) Design of four potential engineering strategies predicted by the C. bescii model. Boxes with a white background indicate the native metabolic functions in C. bescii; boxes with a green background indicate insertion mutants present in existing strains of engineered C. bescii; boxes with a cyan background indicate the new engineering strategy proposed in this study. The products of each reaction are represented in boldface with underlines. Key intermediates of each engineering design are represented as ovals with a yellow background.