The Rut pathway for pyrimidine degradation: novel chemistry and toxicity problems

Abstract

The Rut pathway is composed of seven proteins, all of which are required by Escherichia coli K-12 to grow on uracil as the sole nitrogen source. The RutA and RutB proteins are central: no spontaneous suppressors arise in strains lacking them. RutA works in conjunction with a flavin reductase (RutF or a substitute) to catalyze a novel reaction. It directly cleaves the uracil ring between N-3 and C-4 to yield ureidoacrylate, as established by both nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry. Although ureidoacrylate appears to arise by hydrolysis, the requirements for the reaction and the incorporation of (18)O at C-4 from molecular oxygen indicate otherwise. Mass spectrometry revealed the presence of a small amount of product with the mass of ureidoacrylate peracid in reaction mixtures, and we infer that this is the direct product of RutA. In vitro RutB cleaves ureidoacrylate hydrolytically to release 2 mol of ammonium, malonic semialdehyde, and carbon dioxide. Presumably the direct products are aminoacrylate and carbamate, both of which hydrolyze spontaneously. Together with bioinformatic predictions and published crystal structures, genetic and physiological studies allow us to predict functions for RutC, -D, and -E. In vivo we postulate that RutB hydrolyzes the peracid of ureidoacrylate to yield the peracid of aminoacrylate. We speculate that RutC reduces aminoacrylate peracid to aminoacrylate and RutD increases the rate of spontaneous hydrolysis of aminoacrylate. The function of RutE appears to be the same as that of YdfG, which reduces malonic semialdehyde to 3-hydroxypropionic acid. RutG appears to be a uracil transporter.

FIG. 1.

Comparison of Rut pathway products (E. coli K-12) to those of other pyrimidine catabolic pathways. (A) The Rut pathway, which has been studied only in vivo in E. coli K-12 (31); (B) known reductive (52) and oxidative (22, 28, 48) pathways for catabolism of pyrimidine rings (upper and lower pathways, respectively). Although the enzyme that initiates the oxidative pathway was originally called uracil oxidase, it is a classical monooxygenase (28). An additional pathway (not shown) has recently been proposed in Saccharomyces kluyveri (1).

FIG. 2.

RutA acts on [¹⁴C]uracil and thymine. (A) Products from [¹⁴C-6]uracil in the presence of increasing amounts of cell extract (μl) from NCM4384 (UpBCon1) grown on uridine at 37°C. (B) Products from [¹⁴C-6]uracil (lanes 1 to 3) or [¹⁴C-2]uracil (lanes 4 to 6) in the presence of RutA (lanes 2 and 5) or RutA and RutB (lanes 3 and 6). Samples in lanes 1 and 4 are from controls with no enzyme. (C) Products from [¹⁴C-6]uracil (lanes 4 and 5) or [¹⁴CH₃]thymine (lanes 6 and 7) in the presence of RutA. Samples in lanes 1 and 2 are from controls with no enzyme, and the sample in lane 3 contained a mixture of unlabeled thymine (T), uracil (U), and barbituric acid (B), which were detected by UV absorbance and have been circled. All reactions mixtures that contained RutA also contained the flavin reductase Fre. Reactions were run for 20 min at room temperature with agitation as described in Materials and Methods, and the mixtures were frozen at −20°C before being analyzed by thin-layer chromatography. For samples in panels A and C, reactions were run at pH 8.2, and for samples in panel B, they were run at pH 7.

FIG. 3.

NMR evidence that RutA cleaves the uracil ring between N-3 and C-4 and incorporates O from O₂ at C-4. (A) 1D ¹³C-NMR spectrum of the C-4 resonances of the product of the RutA reaction (RutA product C4) and uracil (Uracil C4). Uracil was uniformly ¹³C/¹⁵N labeled. The C-4 resonance of uracil shows a large coupling of 65 Hz to C-5 and a small coupling of 11 Hz to N-3. The C-4 resonance of the product lacks the coupling to N-3, indicating that the bond between N-3 and C-4 was broken. The spectrum was recorded at 600 MHz. (B) The spectrum is as in panel A, except that [¹³C-4, C-5]uracil was used, the RutA reaction mixtures were bubbled with ¹⁸O₂ or ¹⁶O₂, and then equal volumes were combined. The uracil and product C-4 resonances are split by the ¹J_C-4-C-5 coupling of 65 Hz. The product C-4 resonance also exhibits an ¹⁸O isotope shift of −0.02 ppm, indicating that oxygen was incorporated at this position. The spectrum was recorded at 800 MHz. The different chemical shifts of the species in panels A and B result from the fact that spectrum A was recorded in H₂O, whereas spectrum B was recorded in DMSO.

FIG. 4.

Mass spectrometric evidence that RutA yields ureidoacrylate and a trace of its peracid. (A) Ureidoacrylate m/z 133.0491 corresponds to [C₄H₆N₂¹⁶O₂¹⁸O + H]⁺ (calculated value, 133.0494); m/z 139.0565 corresponds to [¹³C₄H₆¹⁵N₂¹⁶O₂¹⁸O + H]⁺ (calculated value, 139.0569). (B) Peracid of ureidoacrylate. m/z 147.0396 corresponds to [C₄H₆N₂O₄ + H]⁺ (calculated value, 147.0400); m/z 153.0471 corresponds to [¹³C₄H₆¹⁵N₂O₄ + H]⁺ (calculated value, 153.0475).

FIG. 5.

In vitro reactions catalyzed by RutA/F, RutB, and the short-chain dehydrogenase YdfG. (A) RutA/F reaction. The RutA/F reaction yields ureidoacrylate in vitro. However, there also appears to be a small amount of ureidoacrylate peracid in reaction mixtures (Fig. 4B), and we infer that this is the product of the RutA/F reaction. Mukherjee et al. (37) have evidence that the peracid is quickly reduced to ureidoacrylate by NADH spontaneously (Spont.) under conditions similar to ours (see Discussion and Fig. 6) (B) RutB reaction. The RutB reaction yields 2 mol of ammonium, HCO₃⁻, and malonic semialdehyde (3-oxopropionate) from ureidoacrylate. Carbamate and aminoacrylate, which hydrolyze spontaneously, are the presumed intermediates. (C) YdfG reaction. The known short-chain dehydrogenase YdfG (18) reduces malonic semialdehyde to 3-hydroxypropionic acid. In our case, malonic semialdehyde was generated from ureidoacrylate by the RutB reaction, which was run simultaneously.

FIG. 6.

Proposed in vivo pathway for pyrimidine ring degradation in E. coli K-12 (A) and possible handling of ureidoacrylate (B). (A) Rut pathway. RutG appears to be a pyrimidine nucleobase transporter. We infer that RutA catalyzes synthesis of ureidoacrylate peracid (see text). Although our work did not address the specific role of FMN, it is plausible that flavin hydroperoxide, a well-known intermediate in related reactions (40), would participate (37). We postulate that ureidoacrylate peracid is the primary substrate for RutB (see text). Activities of RutC, -D, and -E, which have not yet been studied biochemically, were inferred by a variety of other means. Whereas YdfG uses NADPH as a cofactor, RutE is predicted to be a flavoprotein (9, 27). Proposed names for Rut enzymes are in the inset. (B) Formation and use of ureidoacrylate. If some ureidoacrylate is formed by spontaneous reduction of ureidoacrylate peracid (37), RutB can hydrolyze it. We believe this auxiliary pathway, which was prominent in vitro (Fig. 5), plays a minor role in vivo (see text).