Utilization of 5'-deoxy-nucleosides as Growth Substrates by Extraintestinal Pathogenic E. coli via the Dihydroxyacetone Phosphate Shunt

Abstract

All organisms utilize S-adenosyl-l-methionine (SAM) as a key co-substrate for methylation of biological molecules, synthesis of polyamines, and radical SAM reactions. When these processes occur, 5'-deoxy-nucleosides are formed as byproducts such as S-adenosyl-l-homocysteine (SAH), 5'-methylthioadenosine (MTA), and 5'-deoxyadenosine (5dAdo). One of the most prevalent pathways found in bacteria for the metabolism of MTA and 5dAdo is the DHAP shunt, which converts these compounds into dihydroxyacetone phosphate (DHAP) and 2-methylthioacetaldehyde or acetaldehyde, respectively. Previous work has shown that the DHAP shunt can enable methionine synthesis from MTA or serve as an MTA and 5dAdo detoxification pathway. Here we show that in Extraintestinal Pathogenic E. coil (ExPEC), the DHAP shunt serves none of these roles in any significant capacity, but rather physiologically functions as an assimilation pathway for use of MTA and 5dAdo as growth substrates. This is further supported by the observation that when MTA is the substrate for the ExPEC DHAP shunt, the sulfur components is not significantly recycled back to methionine, but rather accumulates as 2-methylthioethanol, which is slowly oxidized non-enzymatically under aerobic conditions. While the pathway is active both aerobically and anaerobically, it only supports aerobic ExPEC growth, suggesting that it primarily functions in oxygenic extraintestinal environments like blood and urine versus the predominantly anoxic gut. This reveals a heretofore overlooked role of the DHAP shunt in carbon assimilation and energy metabolism from ubiquitous SAM utilization byproducts and suggests a similar role may occur in other pathogenic and non-pathogenic bacteria with the DHAP shunt.

Fig. 1.

Salvage of SAM utilization byproducts, 5-deoxy-pentose sugars, and 6-deoxy-hexose sugars in E. coli. A) S-adenosyl-l-homocysteine (SAH), produced by methyltransferases is recycled by the Methionine Cycle (a.k.a. Active Methyl Cycle) [64, 65]. B-C) The E. coli variation of the DHAP shunt. B) For adenine salvage and detoxification of 5’-methylthioadenosine (MTA) and 5’-deoxyadenosine (5dAdo), all E. coli possess the multifunctional SAH/MTA/5dAdo nucleosidase (Pfs; a.k.a. MtnN) [23]. C) ExPEC strains possess the Pfs nucleosides (B) and remainder of the DHAP shunt genes which together compose a dual-purpose pathway for conversion of 5’-deoxy-nucleosides in the form of MTA and 5dAdo or corresponding 5-deoxy-pentoses in the form of 5-deoxy-d-ribose and 5-methylthioribose, respectively, into the central carbon metabolite dihydroxyacetone phosphate (DHAP) and acetaldehyde or (2-methylthio)acetaldehyde [13]. D) E. coli analogously metabolizes 6-deoxy-hexose sugars in the form of l-fucose and l-rhamnose to DHAP and l-lactaldehyde. During anaerobic growth l-lactaldehyde is primarily reduced to (S)-1,2-propanediol as a terminal product, whereas during aerobic growth it is primarily oxidized to l-lactate for carbon assimilation as pyruvate. Enzymes: Pfs (MtnN), SAH/MTA/5dAdo nucleosidase, E.C. 3.2.2.9; LuxS, S-ribosyl-l-homocysteine lyase, E.C. 4.4.1.21; MetH, methionine synthase, E.C. 2.1.1.13; MetE, methionine synthase, E.C. 2.1.1.4; MtnK, 5-methylthioribose/5-deoxyribose kinase, E.C. 2.7.1.100; MtnA, 5-methylthioribose-1-phosphate/5-deoxyribose-1-phosphate isomerase, E.C. 5.3.1.23; Ald2, 5-methylthioribulose-1-phosphate/5-deoxyribulose-1-phosphate aldolase, E.C. 4.1.2.62; FucI, l-fucose isomerase, E.C. 5.3.1.25; FucK, l-fuculose kinase, E.C. 2.7.1.51; FucA, l-fuculose-1-phosphate aldolase, E.C. 4.1.2.17; RhaA, l-rhamnose isomerase, E.C. 5.3.1.14; RhaK, l-rhamnulose kinase, E.C. 2.7.1.5; RhaD, l-rhamnulose-1-phosphate aldolase, E.C. 4.1.2.19; FucO, (S)-1,2-propanediol oxidoreductase, E.C. 1.1.1.77; AldA, l-lactaldehyde dehydrogenase, E.C. 1.2.1.22.

Fig. 2.. Sulfur from MTA cannot be salvaged by E. coli for growth.

A) ATCC 25922 maximum growth achieved after 24 hours measured by optical density at 600 nm when cultured with 1 mM of the indicated sulfur compound as the sole sulfur source. No further growth was observed after 24 hours. Average and standard deviation error bars are for n=3 independent replicates. B) Fold-difference in the abundance of DHAP shunt-associated metabolites when ATCC 25922 was grown aerobically in the presence of 1 mM sulfate and 1 mM MTA versus grown aerobically in the presence of 1 mM sulfate only. Metabolites were resolved by LC-MS/MS from three independent biological replicates for each growth condition. Values are the average for the three replicates and significance of fold-change was analyzed by ANOVA with P < 0.05. C) Reverse Phase HPLC quantification of E. coli ATCC 25922 metabolites when fed aerobically with [¹⁴C-methyl]-5’-methylthioadenosine. Unk – unknown (identified as 2-methylsulfinylethanol by LC-MS/MS, m/z = 109.0319), R_T = 6.7 min; Met – methionine, R_T = 8.0 min; MTR – methylthioribose, R_T = 16.7 min; MT-EtOH – 2-methylthioethanol, R_T = 21.7 min; IS - internal standard, R_T = 24.0 min; MTA – 5’-methylthioadenosine, R_T = 31.8 min.

Fig. 3.. Prevention of growth inhibition by SAM utilization byproducts requires Pfs but not the DHAP shunt.

Final culture density after 18 hours of the ATCC 25922 wild type, Δpfs, and ΔK2 (ΔmtnK ΔmtnA Δald2) strains and the K12 BW25113 wild type and Δpfs deletion strains grown aerobically with glucose and in the presence of the indicated concentration of either A) 5’-methylthioadenosine (MTA), B) 5’-deoxyadenosine (5dAdo), or C) 5-deoxy-d-ribose (5dR). Averages and standard deviation error bars are for n=3 independent replicates. Curves are the non-linear least squares weighted fit to the Hill equation. D) ID₅₀ for MTA and 5dAdo in ExPEC ATCC 25922 and commensal K12 BW25113 strains. ID₅₀ values are from the fit curves in A-B to the Hill Equation and the error bars are the parameter 95% confidence interval of the weighted fit. * Statistically significant difference, P < 0.05.

Fig. 4.. Assimilation of 5’-deoxy-nucleosides and 5-deoxy-pentoses for carbon and energy metabolism by ExPEC with the DHAP shunt.

A) Growth of either wild type ATCC 25922, ATCC 25922 ΔK2 (ΔmtnK ΔmtnA Δald2), ΔK2 + pEV (pTETTET empty vector), or ΔK2 + pK2 (pTETTET vector with DHAP shunt mtnK, mtnA, and ald2) on either 25 mM glucose (Glc) or 5 mM 5-deoxy-d-ribose (5dR) as the carbon source. B-E) Growth of ATCC 25922, ATCC 25922 Δpfs, ATCC 25922 ΔK2, K12 commensal strain BW25113, or BW25113 Δpfs on 1 mM of (B) glucose, (C) 5dAdo, (D) MTA, (E) or no carbon source. F) Growth of K12 commensal strain BW25113 pEV (pTETTET empty vector) and K12 complemented with pK2 (pTETTET with DHAP shunt) using either 5 mM glucose (Glc) or 5 mM 5-deoxy-d-ribose (5dR) as the carbon source, or no carbon source. Average and standard deviation error bars are for n=3 independent replicates.

Fig. 5.. The ExPEC DHAP shunt is an efficient carbon assimilation pathway for growth, but only under aerobic conditions.

A) Aerobic ATCC 25922 growth efficiency for various carbon growth substrates measured as dry cell weight of cells generated per gram of substrate. B) Ion exclusion HPLC quantification of ³H-labeled metabolites produced upon feeding E. coli ATCC 25922 with [³H-methyl]-5’-deoxyadenosine under anaerobic conditions. After feeding, metabolism was quenched by rapid freezing in liquid nitrogen at the indicated time in minutes. C-D) ExPEC ATCC 25922 wild type and DHAP shunt deletion (ΔK2) strains were grown under varying oxygen concentrations with either C) 25 mM glucose (Glc) or D) 5 mM 5-deoxy-d-ribose (5dR) as the sole carbon source. In addition, experiments were performed with no carbon source (NC). Final optical density measurements at 600 nm (OD₆₀₀) were taken after 24 hours of growth. Averages and standard deviation error bars in A,C-D are for n=3 independent replicates.