A universal metabolite repair enzyme removes a strong inhibitor of the TCA cycle

Abstract

A prevalent side-reaction of succinate dehydrogenase oxidizes malate to enol-oxaloacetate (OAA), a metabolically inactive form of OAA that is a strong inhibitor of succinate dehydrogenase. We purified from cow heart mitochondria an enzyme (OAT1) with OAA tautomerase (OAT) activity that converts enol-OAA to the physiological keto-OAA form, and determined that it belongs to the highly conserved and previously uncharacterized Fumarylacetoacetate_hydrolase_domain-containing protein family. From all three domains of life, heterologously expressed proteins were shown to have strong OAT activity, and ablating the OAT1 homolog caused significant growth defects. In Escherichia coli, expression of succinate dehydrogenase was necessary for OAT1-associated growth defects to occur, and ablating OAT1 caused a significant increase in acetate and other metabolites associated with anaerobic respiration. OAT1 increased the succinate dehydrogenase reaction rate by 35% in in vitro assays with physiological concentrations of both succinate and malate. Our results suggest that OAT1 is a universal metabolite repair enzyme that is required to maximize aerobic respiration efficiency by preventing succinate dehydrogenase inhibition.

Fig. 1. Formation and removal of an inhibitory side-product of the TCA cycle.

Canonical and promiscuous activities of SDH are indicated with black and red arrows, respectively. Enol-OAA, a strong inhibitor of SDH, is converted to keto-OAA spontaneously (⇌) or by OAT1 (blue arrow). All values shown were reported at pH 7.0 and 25 °C, and those associated with SDH or OAT1 were reported for Bos taurus enzymes^,,. Q quinone, QH₂ quinol, eff. catalytic efficiency (k_cat/K_M), k_off dissociation rate constant, K_d equilibrium dissociation constant, k_cat turnover number, k rate constant.

Fig. 2. Phylogenetic relationships, domain architectures, and reactions catalyzed by FAA hydrolase family enzymes.

A Phylogram of representative FAH family enzymes. Nodes with bootstrap values ≥0.5 are shown. Typical architectures of each subfamily of enzymes is shown with conserved domains colored as indicated. Reactions catalyzed by each subfamily of enzymes is shown with enol-keto tautomers highlighted. BS Bacillus subtilis, Mm Methanococcus maripaludis, Dr Danio rerio, Hs Homo sapiens, Sc Saccharomyces cerevisiae, Ec Escherichia coli, Dm Drosophila melanogaster, Ce Caenorhabditis elegans, At Arabidopsis thaliana, Ss Saccharolobus solfataricus. B Predicted FAHD genes (colored blue) cluster with predicted L2HGDH (colored orange) and/or D2HGDH (colored purple) genes in various configurations in several diverse prokaryotic genomes. Apparently unrelated genes are colored white.

Fig. 3. Representative OAT assays.

(A) OAA in diethyl ether or (B) aqueous OAA (pH~2) was added to pH 9 buffered solution. C OAA in acetone was added to pH 7.5 buffered solution containing MDH and NADH. One μg E. coli YcgM (OAT1) was added as indicated. Open circles show progression of the assays without added enzyme.

Fig. 4. FAHD increases SDH activity at physiological substrate concentrations.

Bovine SDH activity at 2 mM succinate (A), 2 mM malate (B), or 2 mM each succinate and malate (C). Assays contained 1 mM MOPS-NaOH, pH 7.1, 5 μg (1 U) GOT, 1 mM glutamate, 0.1 mM DCIP, 0.1 mM decylubiquinone, detergent-solubilized SDH (11–33 μg total protein), and substrate(s) without (black circles) or with addition of 10 μg H. sapiens FAHD1 (red squares) or 10 μg M2-FAHD1 (open blue circles). DCIP reduction was monitored at 600 nm. Data represents the mean and SEM, n = 3 independent experiments; individual data points are represented with gray circles. In assays with malate and succinate, the difference in mean steady-state rates determined without and with FAHD1 was 35.5% and this difference is significant (2-sided t test; P = 0.0021, t = 7.065, df = 4, 95% CI = −0.001403 to −0.0006114).

Fig. 5. Phenotypic consequences of FAHD ablation.

A Serially diluted ΔycgM (ΔFAHD) or wild type E. coli containing either ycgM-pUC19 (FAHD) or empty vector (EV) were spotted on the indicated solid medium plates and pictures were taken after an incubation period. These experiments were repeated three times with similar results. B, D, F Growth of the wild-type strain with empty vector or ΔFAHD strain with empty vector or complemented with FAHD-vector for the indicated organism grown in the indicated liquid medium. Data represents mean and SEM, n = 3 biologically independent cell cultures. C, E Colony size of the indicated E. coli or S. cerevisiae strains on the indicated solid medium. Data represents mean and SD, n = 95 (E. coli) or the indicated (S. cerevisiae) biologically independent cell colonies. 2-sided t tests were used to determine whether differences in means are significant, with P values shown for key comparisons (other statistical information is listed in the source data). G The relative abundance of respiratory intermediates in ΔycgM (red slices) or wild-type (blue slices) E. coli determined by isotope ratio HILIC-MS analysis. Pie charts outlined in gray show relative abundance determined by comparing peak heights. Metabolite ID ambiguities are enclosed by dotted black lined polygons with chromatographic retention times listed where multiple peaks were detected for different isomers. Data represents mean and SD, n = 4 biologically independent cell cultures given two treatments. Statistical significance was determined with one-way ANOVA (P values are listed in Supplementary Data file 1). EPM Embden–Meyerhoff–Parnas pathway, blue, PPP pentose phosphate pathway, orange, ED Entner–Doudoroff pathway, magenta, TCA tricarboxylic acid cycle, green.