Improving formaldehyde consumption drives methanol assimilation in engineered E. coli

Abstract

Due to volatile sugar prices, the food vs fuel debate, and recent increases in the supply of natural gas, methanol has emerged as a promising feedstock for the bio-based economy. However, attempts to engineer Escherichia coli to metabolize methanol have achieved limited success. Here, we provide a rigorous systematic analysis of several potential pathway bottlenecks. We show that regeneration of ribulose 5-phosphate in E. coli is insufficient to sustain methanol assimilation, and overcome this by activating the sedoheptulose bisphosphatase variant of the ribulose monophosphate pathway. By leveraging the kinetic isotope effect associated with deuterated methanol as a chemical probe, we further demonstrate that under these conditions overall pathway flux is kinetically limited by methanol dehydrogenase. Finally, we identify NADH as a potent kinetic inhibitor of this enzyme. These results provide direction for future engineering strategies to improve methanol utilization, and underscore the value of chemical biology methodologies in metabolic engineering.

Fig. 1

The pathway of methanol assimilation in E. coli. Heterologous enzymes required for methanol assimilation in E. coli are shown in blue. Methanol dehydrogenase (Mdh) catalyzes the thermodynamically uphill NAD-dependent oxidation of methanol to formaldehyde (ΔG^0′ = +34.2 kJ mol⁻¹), which is ligated to ribulose 5-phosphate (Ru5P) in an aldol reaction catalyzed by hexulose phosphate synthase (Hps). The resulting hexulose (H6P) is isomerized by phosphohexulose isomerase (Phi) to fructose 6-phosphate (F6P), which is metabolized through native E. coli central metabolism. A portion of the F6P must be used to regenerate Ru5P through the ribulose monophosphate (RuMP) cycle to enable further formaldehyde assimilation. Two major variants of this cycle are known, the transaldolase (Tal) and sedoheptulose bisphosphatase (SBPase), which are named for the enzyme that produces sedoheptulose 7-phosphate, and differ in their ATP requirements. The red moiety denotes the position of assimilated methanol. H6P: D-arabino-3-hexulo-6-phosphate, G3P: glyceraldehyde 3-phosphate

Fig. 2

Insufficient Ru5P concentration in carbon-starved cells limits formaldehyde assimilation. a Detailed enzymatic reactions and metabolic pathways involved in methanol metabolism, showing the entry points of glycogen (the primary carbon source during starvation) through glycolysis, and xylose through the pentose phosphate pathway (PPP). Heterologous reactions unique to methanol assimilation are shown in blue. Glycolysis reactions are shown in purple, and PPP reactions in green. Dotted and dashed green lines added for clarity. b Formaldehyde levels over time after addition of 250 mM methanol to starved cells of E. coli MG1655(DE3) ΔfrmA with either no additional substrate (blue) or 6 g L⁻¹ xylose (orange). Solid lines represent cells expressing only Mdh, and dashed lines denote cells expressing the full methanol assimilation pathway (Mdh, Hps, and Phi). c Relative ribulose 5-phosphate (Ru5P) concentration in cells with no substrate or xylose. Error bars represent s.d. of n = 3 biological replicates (three individual colonies). H6P: D-arabino-3-hexulo-6-phosphate, F6P: fructose 6-phosphate, G6P: glucose 6-phosphate, FBP: fructose 1,6-bisphosphate, DHAP: dihydroxyacetone phosphate, G3P: glyceraldehyde 3-phosphate, PEP: phosphoenolpyruvate, E4P: erythrose 4-phosphate, S7P: sedoheptulose 7-phosphate, R5P: ribose 5-phosphate, Ru5P: ribulose 5-phosphate, Xu5P: xylulose 5-phosphate

Fig. 3

Chemical inhibition of glycolysis with iodoacetate is insufficient to improve formaldehyde assimilation. a Metabolic pathways in methanol metabolism, additionally depicting chemical inhibition of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) by the addition of 1 mM iodoacetate (IA). Dot sizes indicate relative pool sizes after IA addition compared to untreated cells. b Relative internal concentrations of various metabolites with or without IA treatment, quantified by LC-MS/MS. Metabolites immediately upstream of GAPDH (DHAP and FBP) increased dramatically, while metabolites downstream of the blockage (e.g., PEP and 3PG) became virtually undetectable. c Formaldehyde levels after addition of 250 mM methanol to starved cells of E. coli MG1655(DE3) ΔfrmA. Solid lines denote cells expressing only Mdh, whereas dashed lines indicate cells expressing the full pathway (Mdh, Hps, and Phi). Triangles indicate cells treated with IA. Blue lines symbolize cells supplemented with no additional substrate beyond methanol, whereas orange lines represent cells supplemented with 6 g L⁻¹ xylose (XYL). Error bars represent s.d. of n = 3 biological replicates (three individual colonies)

Fig. 4

Iodoacetate coupled with glpX overexpression increases Ru5P concentration and improves cyclic formaldehyde assimilation. a Metabolic pathways in methanol metabolism, showing the blockage of glycolysis by iodoacetate (IA), and the minor gluconeogenic FBPase encoded by glpX. Dot sizes indicate relative pool sizes after IA addition in cells expressing glpX compared to untreated cells without glpX overexpression. b FBPase catalyzes the irreversible hydrolysis of FBP to F6P (ΔG^0′ = −28.5 kJ mol⁻¹), potentially allowing the conversion of the large FBP pool resulting from IA-mediated inhibition into F6P, the entry point for the transaldolase and transketolase reactions that replenish the PPP intermediates. c Relative concentrations of F6P and Ru5P upon treatment with IA in cells with (pink) or without (blue) glpX overexpression. d Formaldehyde levels after addition of 250 mM methanol to starved cells of E. coli MG1655(DE3) ΔfrmA. Solid lines denote cells expressing only Mdh, whereas dashed lines indicate cells expressing the full pathway (Mdh, Hps, and Phi). Pink lines denote cells expressing glpX. Triangles indicate cells treated with IA, squares indicate untreated cells. e Isotopic analysis of fructose 6-phosphate extracted from IA-treated cells with (pink) or without (blue) glpX overexpression that were treated with 250 mM ¹³CH₃OH. Error bars represent s.d. of n = 3 biological replicates (three individual colonies)

Fig. 5

Ru5P regeneration is mediated by activation of the SBPase variant of the RuMP pathway. a Pathways involved in methanol metabolism, additionally showing the sedoheptulose bisphosphatase (SBPase) variant of the ribulose monophosphate (RuMP) pathway, where SBP is produced from DHAP and E4P by FbaA (orange line), and dephosphorylated to S7P by SBPase (pink line), compared to the transaldolase (TAL) pathway (gray line), where S7P is produced from F6P and E4P. Dot sizes indicate relative metabolite concentrations in cells treated with IA and overexpressing glpX, compared to untreated cells not overexpressing glpX. b LC-MS/MS analysis of assays with purified E. coli FbaA and GlpX, demonstrating that GlpX has SBPase activity. Black = DHAP, Red = E4P, Blue = SBP, Pink = S7P. c Relative SBP (left) and S7P (right) levels in cells treated with or without iodoacetate and expressing glpX (pink) or not (blue). d Formaldehyde timecourse in transaldolase-deficient E. coli cells (MG1655(DE3) ΔfrmA talA(K131A) ΔtalB) treated with 250 mM methanol, either expressing (pink) or not expressing (blue) glpX. Triangles denote cells treated with IA, and squares denote untreated cells. Error bars represent s.d. of n = 3 biological replicates (three individual colonies)

Fig. 6

Thermodynamic and kinetic isotope effect analysis reveal Mdh kinetics limit methanol assimilation flux. a Upper bounds on in vivo reversibility of Mdh derived from steady-state measurements of formaldehyde in cells with Mdh or the full pathway (Mdh, Hps, and Phi), and NAD/NADH ratios. Light blue bars represent minimum fraction of Mdh flux in forward direction, and dark blue bars represent maximum fractional reverse flux. Hypothetical “Equilibrium” and “Fully Irreversible” scenarios are plotted for comparison to experimental results under various conditions. b Dynamic simulations of formaldehyde concentration for a hypothetical assimilation pathway where formaldehyde concentration is close to equilibrium (blue) and far from equilibrium (orange), showing the effect of increasing the V_max of Mdh by a factor of 4 in each case. c Michaelis plot of Mdh activity with CH₃OH (black) and CD₃OD (red), highlighting the kinetic isotope effect (KIE) associated with deuterated methanol. d Formaldehyde concentrations over time after treatment with 250 mM ¹³CH₃OH (black) or 250 mM ¹³CD₃OD (red) of E. coli MG1655(DE3) ΔfrmA starved cells containing various plasmids: Left, Mdh-only; Middle, full pathway (Mdh, Hps, and Phi); Right, full pathway + glpX with iodoacetate (IA) treatment. e The fraction of total Mdh flux in the forward direction is plotted as a function of the ratio of steady-state formaldehyde concentrations with protonated or deuterated methanol (blue line, Eq. (2)), with experimentally measured values indicated (orange circles). The minimum and maximum formaldehyde ratios are defined by the equilibrium isotope effect (EIE) and kinetic isotope effect (KIE), respectively. Error bars represent s.d. of n = 3 biological replicates (three individual colonies), except for c in which case they represent s.d. of n = 2 technical replicates