Fig 1.
Lifecycle of B. bacteriovorus HD100.
1) Prey location: B. bacteriovorus moves towards prey-rich regions. 2) Attachment: the predator anchors to the host cell, which leads to the infection. 3) Invasion: B. bacteriovorus enters the periplasm of the prey cell. 4 and 5) Growth in bdelloplast and development: the prey has a rounded appearance due to cell wall modification and B. bacteriovorus grows in the periplasm and replicates its DNA. B. bacteriovorus uses the prey cytoplasm as a source of nutrients. 6 and 7) Septation and development: the predator septates when resources become limited and it matures into individual attack phase cells. 8) Lysis: mature attack-phase cells lyse the cell wall of the bdelloplast, initiating the search for fresh prey. The complete cycle takes about 4 h.
Fig 2.
iCH457 metabolic model pipeline.
The draft of metabolic reconstruction was based on available metabolic models (iJN1411 and iJO1366), the genome sequence of B. bacteriovorus HD100 and the automatic model Seed. Manual curation is required to accurately fine-tune the information contained in the metabolic model and several steps of network validation and analysis are required to finally obtain the metabolic model iCH457. The general model iCH457 was constrained based on nutrient availability (minimal and rich in silico media), biological role (ATP production or biomass generation) and transcriptomic available data [41] for the generation of condition-specific models: iCHAP and iCHGP. GIM3E algorithm was used to construct the condition-specific models Rxns: reactions.
Fig 3.
Distribution of the reactions of the system in 12 global functional categories.
The metabolites inside the rectangles correspond with the auxotrophies in the cofactor metabolism (dark blue fraction) and the amino acid metabolism (brown fraction), respectively.
Table 1.
Comparison of the metabolic properties of iCH457 compared with other δ-proteobacteria (Geobacter spp. and Desulfovibrio vulgaris) metabolic models and with the well-establish metabolic reconstruction of P. putida (iJN1411) and E. coli (iJO1366).
Fig 4.
Evaluation of the metabolic capabilities of iCH457. A) Comparison of the growth performance of the in silico iCH457 strain and a derivative strain of B. bacteriovorus 109 Davis on different carbon sources.
Experimental values of growth rate were calculated using the mass doubling time previously compiled in Ishiguro et al. [67]. The in silico growth rate was calculated with the minimal medium defined in S1 Text supplemented with the tested carbon source. B) Comparison of the biomass production predicted in silico with the available experimental data performed with the prey-independent B. bacteriovorus 109 Davis [67]. In vivo and in silico biomass data were expressed as Kendall’s rank correlation coefficient (τ = 0,88) for iCH457. GLU: glutamate, GLN: glutamine, PYR: pyruvate, LACT: lactate, SUCC: succinate. The statistically significant differences were calculated using two-way ANOVA followed by Bonferroni test. All comparisons were found non-significant (P<0.05).
Table 2.
Reaction essentiality of the metabolic models of different bacteria.
Fig 5.
Comparison of the reaction essentiality of intracellular lifestyle and free-living bacteria.
Numbers of reactions and the corresponded percentage are represented. Only reaction included within iCH457 were considered for comparison.
Fig 6.
Prediction of the carbon flux distribution in iCH457 metabolic network.
Graphical representation of the metabolic carbon fluxes during the life cycle of B. bacteriovorus HD100. The numbers below each reaction represent the more probable flux in each phase (GP flux/AP flux) as determined by Monte Carlo sampling analysis. The thick arrows highlight the carbon flux distribution in GP compared with AP. The thin arrows highlight the reactions that are active in AP compared with GP. As the major carbon sources are amino acids, alanine and glutamate come directly from the breakdown of the dipeptides or from the single amino acids. Eritrose 4 phosphate and glyceraldehyde 3 phosphate come from the degradation pathways of serine and threonine.