Fig 1.
Pipeline and statistics for building the StressME (iZY1689) model.
① EcoliME (iJL1678b) to FoldME, OxidizeME and AcidifyME in python 3.6 ② FoldME to single StressME (°C) ③ Single StressME integrated with AcidifyME to dual StressME (°C and pH) ④ Dual StressME integrated with OxidizeME to Triple StressME (°C, pH and ROS) ⑤ Triple StressME optimization to final iZY1689.
Fig 2.
Investigating sensitivity of predicted phenotypes and proteome to the kinetome (keff, effective rate constants).
(A) and (B): Distribution of Keffs for metabolic reactions used in previous single-stress ME models; (C) and (D): Effect of Keffs on growth rates simulated by StressME under different stress conditions; (E) and (F) Effect of Keffs on simulated protein mass fractions by StressME under different stress conditions.
Fig 3.
Stress-evolved Keffs for StressME (A) Sensitivity of simulated growth rates to keff choice for each reaction. Two reactions are affected the most: reaction DXPRli (1-deoxy-D-xylulose 5-phosphate reductoisomerase, dxr); reaction RHCCE (S-ribosylhomocysteine lyase, luxS) (B) Effect of keff for the reaction DXPRli on proteome allocation (C) Effect of protein unfolding Keq for dxr on growth rates (D) Effect of keff for the reaction RHCCE on proteome allocation (E) Effect of Keff and protein properties (unfolding Keq and aggregation propensities) for the reaction RHCCE on growth rates.
Fig 4.
StressME accurately simulates individual stress responses.
(A) thermal stress (B) acid stress (C) and (D) oxidative stress with different supply of amino acids (AAs).
Fig 5.
Computing alternative optimal solutions from StressME by Flux variability analysis (FVA).
FVA for the heat-evolved strain (A)-(D): (A) glucose uptake rates (GUR), (B) oxygen uptake rates (OUR), (C) CO2 production/uptake rates (CO2) and (D) acetate production rates (APR). FVA for the wild-type strain (E)-(H). Pink zones represent the maximal boundedness of the optimal solution space to support 95%-100% of maximal biomass production rates. Negative values indicate the reverse direction of the exchange reaction.
Fig 6.
Effect of heat (Stress 1) adaptation on system-level proteome re-allocation as tradeoffs when exposed to Stress 2 (thermal or acid or oxidative or combination).
(A) wild type strain–proteome mass fraction overview from 0.0 to 1.0 (B) wild type strain–proteome zoom-in view from 0.000 to 0.015 (C) heat-evolved strain–proteome mass fraction overview from 0.0 to 1.0 (D) heat-evolved strain–proteome zoom-in view from 0.000 to 0.015.
Fig 7.
Proteome reallocation under thermal and oxidative dual stress.
Heat-evolved strain (A)-(B): (A) proteome mass fractions for COG functional groups (B) proteome mass fractions for key oxidative-response groups. Wild-type strain (C)-(D). ROS genes: ROS-activated proteome (19 ROS stress-exclusive and stress-intensified proteins). Damage genes: ROS-vulnerable proteome (31 [Fe-S] binding proteins).
Fig 8.
Proteome reallocation under thermal-acid dual stress.
Heat-evolved strain (A)-(D): (A) pH 5.0 and pH 7.0 at 26°C; (B) pH 5.0 and pH 7.0 at 32°C; (C) pH 5.0 and pH 7.0 at 40°C and (D) pH 5.0 and pH 7.0 at 42°C. Wild-type strain: (E)-(H).
Fig 9.
Metabolic flux balance analysis (MFBA) showing a switch from fully respiratory to respiro-fermentative metabolism and acetate overflow under thermal-acid stress: Heat-evolved strain (A) and (B); Wild-type strain (C) and (D). (A) and (C): Fluxome (mmol g-1 DCW h-1) at pH 7.0, temperature 26°C, 32°C and 40°C. (B) and (D): Fluxome at pH 5.0, temperature 26°C, 32°C and 40°C.
Fig 10.
Proteome reallocation under oxidative and acid dual stress.
Heat-evolved strain (A)-(B): (A) proteome mass fractions for COG functional groups (B) proteome mass fractions for key oxidative-response groups. Wild-type strain (C)-(D). ROS genes: ROS-activated proteome (19 ROS stress-exclusive and stress-intensified proteins). Damage genes: ROS-vulnerable proteome (31 [Fe-S] binding proteins).
Fig 11.
Proteome allocation under thermal-oxidative-acid triple stress.
(A) proteome mass fractions for COG functional groups in heat-adapted strain (B) proteome mass fractions for COG functional groups in heat non-adapted strain (C) ROS-activated proteome (19 ROS stress-exclusive and stress-intensified proteins) and ROS-vulnerable proteome (31 [Fe-S] binding proteins) for heat-adapted and heat non-adapted strains.
Table 1.
Eleven stress-response genes added to FoldME.
Fig 12.
Integrated StressME model to combine mechanisms of three single stress models (thermal, oxidative, acid).
User-friendly I/O platform to run simulations with a simple array of temperature, pH and ROS levels; CSV output for phenotypes, proteome and fluxome for further visualization.