Pyrimidine homeostasis is accomplished by directed overflow metabolism

Abstract

Cellular metabolism converts available nutrients into usable energy and biomass precursors. The process is regulated to facilitate efficient nutrient use and metabolic homeostasis. Feedback inhibition of the first committed step of a pathway by its final product is a classical means of controlling biosynthesis. In a canonical example, the first committed enzyme in the pyrimidine pathway in Escherichia coli is allosterically inhibited by cytidine triphosphate. The physiological consequences of disrupting this regulation, however, have not been previously explored. Here we identify an alternative regulatory strategy that enables precise control of pyrimidine pathway end-product levels, even in the presence of dysregulated biosynthetic flux. The mechanism involves cooperative feedback regulation of the near-terminal pathway enzyme uridine monophosphate kinase. Such feedback leads to build-up of the pathway intermediate uridine monophosphate, which is in turn degraded by a conserved phosphatase, here termed UmpH, with previously unknown physiological function. Such directed overflow metabolism allows homeostasis of uridine triphosphate and cytidine triphosphate levels at the expense of uracil excretion and slower growth during energy limitation. Disruption of the directed overflow regulatory mechanism impairs growth in pyrimidine-rich environments. Thus, pyrimidine homeostasis involves dual regulatory strategies, with classical feedback inhibition enhancing metabolic efficiency and directed overflow metabolism ensuring end-product homeostasis.

Figure 1. Increased pyrimidine flux triggers overflow to uracil

a, Canonical pyrimidine regulatory schematic. Carbamoyl phosphate is an intermediate in both pyrimidine and arginine synthesis. Carbamoyl aspartate is committed to pyrimidine synthesis. Carbamoyl phosphate synthetase (carAB) is feedback inhibited by UMP. Aspartate transcarbamoylase (pyrBI) is feedback inhibited by UTP and CTP and activated by ATP. b, c, Extracted ion chromatograms showing uracil (b) and CTP (c) levels in wild-type (black) and feedback-defective (ΔpyrI carB*; red) strains. d, Metabolite fold changes relative to wild type in ΔpyrI, carB* and ΔpyrI carB* strains. Error bars denote ± standard error (n = 6). e, Metabolite fold changes at 5min after addition of orotate. Fold changes were computed relative to un-supplemented controls. Error bars denote ± standard deviation (n = 2–3). Time course data appear in Supplementary Fig. 2. f, Fraction of UMP and UTP derived from endogenous sources and from exogenously added ¹⁵N-orotate. Error bars denote ± standard deviation (n = 2). The ΔpyrI (blue) and wild-type (black) lines lie under the carB* (green) line. g, Competitive growth advantage of wild-type versus ΔpyrI carB* appears selectively under energy limitation. Competitions were performed in indicated media with lacZ⁻ marker in wild-type and in feedback-dysregulated strain (wild-type lacZ⁺) to control for effect of marker on growth. Calculations and experimental details are described in Methods. In brief, media were glucose-ammonia minimal media unless otherwise indicated. Alternative carbon and nitrogen sources were used in conditions K, L and O. + indicates supplements to the minimal media; +/− indicates alternating supplementation/removal of indicated nutrient every 8 h. Grey ellipses mark 95% confidence interval (n = 6–10).

Figure 2. Pyrimidine overflow pathway is initiated by catabolism of UMP by UmpH

a, Pathway schematic. b, Uracil excretion does not depend on the canonical pyrimidine interconversion enzymes Udk and Upp, but does require Udp. Excreted uracil accounts for approximately half of total uracil. Error bars mark standard deviation (n = 2–3). c, Uracil is produced by UMP degradation following orotate addition. UMP is degraded to uridine by UmpH (also known as NagD) and UmpG (also known as SurE), and uridine to uracil by Udp. Isotopic tracing from orotate to UMP, uridine and uracil appears in Supplementary Fig. 3. Residual uracil in a Δudp ΔcodA strain indicates at least one unannotated activity also makes minor contributions to uracil production. End-product levels are not perturbed following orotate addition by deletion of either ΔumpH or ΔumpG individually, but double deletion leads to increased UTP. Error bars mark ± standard deviation (n =3). d, Inability to degrade UMP to uracil causes a growth defect upon orotate addition but increases final culture density (reproducible in two biologically independent experiments each of 6–8 technical replicates). D, attenuance.

Figure 3. Cooperative inhibition of UMP kinase by UTP maintains end-product homeostasis

Variant pyrH alleles were expressed from its native promoter on low-copy plasmid (pACYC) with genomic pyrH removed. a, Schematic of downstream regulatory events in pyrimidine metabolism. Wild-type UMP kinase (PyrH) is feedback inhibited by UTP in a switch-like manner (high degree of cooperativity). Allosteric parameters of pyrH alleles appear in shaded box, with D93A lacking the switch-like behaviour. nH, Hill coefficient. b, Altered expression or regulation of UMP kinase following orotate addition impairs growth. Defects also occur in pyrH^WT/pyrH diploid strains (Supplementary Fig. 6). c, Metabolite fold changes upon orotate addition in strains with altered UMP kinase expression or allosteric regulation reveal defects in both pyrimidine and purine homeostasis. Error bars mark standard deviation (n = 3).

Figure 4. Directed overflow metabolism in biosynthesis is analogous to central carbon overflow metabolism

a, Schematic of directed overflow metabolism as a biosynthetic regulatory mechanism. b, Schematic of overflow in central carbon metabolism, using the Warburg effect as a canonical example. PDH, pyruvate dehydrogenase; PDHK, pyruvate dehydrogenase kinase (which catalyses inhibitory phosphorylation of PDH); TCA, tricarboxylic acid.