The ybcF Gene of Escherichia coli Encodes a Local Orphan Enzyme, Catabolic Carbamate Kinase

Escherichia coli can use allantoin as its sole nitrogen source under anaerobic conditions. The ureidoglycolate produced by double release of ammonia from allantoin can flow into either the glyoxylate shunt or further catabolic transcarbamoylation. Although the former pathway is well studied, the genes of the latter (catabolic) pathway are not known. In the catabolic pathway, ureidoglycolate is finally converted to carbamoyl phosphate (CP) and oxamate, and then CP is dephosphorylated to carbamate by a catabolic carbamate kinase (CK), whereby ATP is formed. We identified the ybcF gene in a gene cluster containing fdrA-ylbE-ylbF-ybcF that is located downstream of the allDCE-operon. Reverse transcription PCR of total mRNA confirmed that the genes fdrA, ylbE, ylbF, and ybcF are co-transcribed. Deletion of ybcF caused only a slight increase in metabolic flow into the glyoxylate pathway, probably because CP was used to de novo synthesize pyrimidine and arginine. The activity of the catabolic CK was analyzed using purified YbcF protein. The Vmax is 1.82 U/mg YbcF for CP and 1.94 U/mg YbcF for ADP, and the KM value is 0.47 mM for CP and 0.43 mM for ADP. With these results, it was experimentally revealed that the ybcF gene of E. coli encodes catabolic CK, which completes anaerobic allantoin degradation through substrate-level phosphorylation. Therefore, we suggest renaming the ybcF gene as allK.


Introduction
Carbamoyl phosphate (CP) is an essential metabolite involved in the biosynthesis of pyrimidine nucleotide and arginine/urea and also in the synthesis of antibiotics [1][2][3]. CP is composed of only three functional groups, ammonia, carbonate, and phosphate, and it is metabolically reactive with high energy. Walsh et al. called CP one of the eight key metabolites at the core of cell metabolism [4].
CP is produced in various pathways using several unique enzymes. The first pathway is irreversible formation of CP by carbamoyl phosphate synthetase (CPS) from ammonia (or glutamine), bicarbonate, and ATP, observed in most organisms. In the urea cycle of mammalian mitochondria, CP is formed from inorganic ammonium and bicarbonate by CPS catalysis, which consumes 2 ATP (reaction 1). This CPS using inorganic ammonia as the nitrogen donor is called CPS I: The biosynthesis of arginine and pyrimidine uses CP produced by CPS to catalyze the transfer of ammonia from glutamine to bicarbonate and also requires 2 ATP (reaction 2). This type of CPS using glutamine as the nitrogen donor is called CPS II: Carbamate + ATP ↔ CP + ADP (CK-like CPS, EC 2.7.2.2) ΔG 0 ′ = +8.9 kJ/mol (3) NH 4 + + HCO 3 -↔ Carbamate + H 2 O (Chemical) ΔG 0 ′ = -1.6 kJ/mol (4) The third case of CP generation uses a degradation (catabolic) pathway rather than a biosynthetic (anabolic) pathway. Shi et al. described this pathway as a collection of catabolic transcarbamylase reactions [3]. One wellstudied example is CP production by catabolic ornithine transcarbamylase (OTCase) in the arginine deiminase (ADI) pathway, which is found in Pseudomonas aeruginosa, Enterococcus faecalis, and in several bacteria in the order Lactobacillales [9][10][11][12]. In Lactobacillus brevis, arginine is anaerobically degraded by ADI into ammonia and citrulline that is further converted into ornithine and CP by OTCase [12]. These are coupled with the reaction of catabolic CK, which transfers a phosphate group to generate ATP and leave carbamate (CP + ADP → Carbamate + ATP: catabolic CK, E.C. 2.7.2.2). In this pathway, CP is formed by OTCase and used for substrate-level phosphorylation, a thermodynamically favorable pathway. Moreover, in the genome of L. brevis, the gene encoding CK is clustered with the genes for ADI, OTCase, and the related enzymes [12], implying that the main function of CK is related to catabolism in this strain. A similar catabolic transcarbamylase reaction produces CP by anaerobic agmatine degradation via agmatine deiminase [13,14].
E. coli is another organism that uses catabolic transcarbamylase to produce CP via the oxamic transcarbamylase (OXTCase) reaction in anaerobic allantoin degradation. It has long been known that this pathway exist in E. coli, but the genes for OXTCase and catabolic CK have not yet been experimentally identified. The former is a global orphan, and the latter is a local orphan. Though there have been attempts to identify the protein YgeW as OXTCase, they were not successful [15]. Recently, we presented a strong candidate gene, fdrA, for OXTCase with evidence that ΔfdrA failed to convert oxalurate to CP and oxamate [16]. Near fdrA lies the ybcF gene that Smith et al. also suggested as a candidate for a CK [17]. If that is correct, the gene for catabolic CK of E. coli is clustered with allantoin pathway, unlike the CK of L. brevis, in which the gene is located with the ADI pathway together. In this work, we demonstrate that the ybcF gene of E. coli belongs to an fdrA operon and encodes the missing catabolic CK. This study is the first to experimentally validate that YbcF of E. coli has catabolic carbamate kinase activity.
To produce a concentrated cell suspension, E. coli cells grown in the above culture conditions were harvested by centrifugation for 5 min at 5,000×g. The cells were immediately resuspended in nitrogen-deficient 3-(Nmorpholino) propanesulfonic acid (MOPS) buffer (no ammonium chloride added) [19] to OD 600 = 20 and anaerobically changed with a 95% N 2 + 5% H 2 gas mixture. Allantoin degradation was initiated by mixing the cells with the same volume of the MOPS containing glycerol (20 mM), DMSO (20 mM), and allantoin (20 mM), and the cell suspension was incubated at 37°C and analyzed by high performance liquid chromatography (HPLC).

Inactivation of ybcF
The ybcF gene in the genome of E. coli K-12 MG1655 (Fλ -ilvGrfb-50 rph-1) was deleted using the method of Datsenko and Wanner [20]. For insertional inactivation, the PCR product of the catR chloramphenicol resistance gene from pKD3 was used and flanked by FRT sequences. Primers fdrA_H1P1 [5′-CGC TGC TGG GGT TCT GGC TTG GCC AAC AAT TAT TAC AAG GAA AAC CAT GAG TGT AGG CTG GAG CTG CTT C -3′] and fdrA_H2P2 [5′-GAT AAG ACG CGT CAA GCG TCG CAT CAG GCA CAA ATG TCT AAT GCC TAC GAC ATA TGA ATA TCC TCC TTA G -3′] contain parts of the regions adjacent to FRT and the 5′ and 3′ regions of the amplified ybcF, respectively. The strain in which ybcF was replaced by catR was designated as LMB111.

Analysis of Fermentation Products by HPLC
Culture supernatants were analyzed using an Hitachi LaChrom Elite HPLC System (Hitachi High-Tech Corp., Japan), equipped with a pump (L-2130), column oven (L-2350), autosampler (L-2200), and an Aminex HPX-87H ion-exclusion column (300 × 7.8 mm; Bio-Rad, USA). The mobile phase was 4 mM H 2 SO 4 supplied at a constant flow rate of 0.55 ml/min. The sample was injected with 10 μl and run for 25 min. The column temperature was adjusted to 18°C for allantoin and oxamate and 27°C for oxalate and other organic acids. The quantitative determination was carried out using an L-2490 refractive index detector.

Expression and Purification of YbcF
The ybcF overexpression plasmid pCA24N::ybcF (JW0510-AM, NBRP, NIG, Japan), which includes six histidines at the N-terminus, was transformed to MG1655 and aerobically grown in LB broth (Difco, USA) to OD 600 of about 0.3 at 37°C. Induction was performed with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for 2 h at 30°C. For purification of the 6X His-tagged proteins, cell pellets were suspended in lysis buffer (20 mM Tris-Cl pH 8.3, 0.5% triton X-100, 1 mg/ml lysozyme), and cell disruption was achieved by six repetitions of vortexing with acid-washed glass beads (G1145, Sigma-Aldrich) and icing for 5 min. Cell debris was removed by centrifugation (13,000 ×g, 4°C, 20 min), and the supernatant cell lysate was collected. The purification of YbcF was performed via affinity chromatography with Ni-NTA Spin Columns (31014, Qiagen) according to the manufacturer's directions. Protein concentrations were determined with the Bradford method using Protein Assay Dye (5000006, Bio-Rad). The standard curve was generated using bovine serum albumin (BSA).

Western Blotting
Elution fractions were analyzed with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting. Samples were subjected to SDS-PAGE (Bio-Rad) using a 12% running gel. Electrophoretic transfer was performed on an Amersham Hybond P western blotting membrane (GE Health, USA) with a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). After blotting (15 V, 1 h 20 min), the membrane was blocked for 1 h at room temperature. Then, the membrane was incubated overnight at 4°C with 6X His-tag monoclonal primary antibody (MA1-21315, Invitrogen, USA), which was diluted 1:2000. Subsequently, horseradish peroxidasecoupled goat anti-mouse secondary antibody (G21040, Invitrogen) was diluted 1:100000 and incubated with the membrane at room temperature for 1 h. Dyne ECL Pico Plus (GBE-P200, Dynebio, Republic of Korea) was used to detect peroxidase activity. Development was performed using an ImageQuant800 (USA).
To determine the concentration-dependent kinetics, CK activity was measured using 100 nM YbcF by varying the concentration of CP from 0 to 3 mM while holding ADP fixed at 2 mM, and vice versa. After 60 min at 28°C, the ATP concentration in the reaction mixture was measured with an ATP Determination Kit (A22066, Molecular Probes Inc., USA). The ATP standard (10 μl) or carbamate kinase assay sample (10 μl) was added to the Standard Reaction Solution (90 μl, 0.5 mM D-luciferin, 1.25 μg/ml firefly luciferase, 25 mM tricine buffer, pH 7.8, 5 mM MgSO 4 , 100 μM ethylenediaminetetraacetic acid, 1 mM dithiothreitol). Then they were incubated at 28°C for 15 min. The amount of ATP was analyzed with a Spectramax i3 bioluminescence reader (Molecular Devices, USA) at 560 nm.
The Michaelis-Menten constant (K M ) and maximum velocity (V max ) of ADP and CP were calculated using the Michaelis-Menten equation and Lineweaver-Burk equation, respectively.

Statistical Analysis
Statistical analyses were performed using PASW Statistics 18 (SPSS, Inc.). Data were analyzed using unpaired two-tailed Student's t-tests and one-way analysis of variance (ANOVA). Post hoc analyses were performed using Duncan's multiple comparison test. Statistical significance was defined as p < 0.05.

The ybcF gene Is Part of the fdrA Operon
The possible promoter regions for gene clusters in the E. coli genome around ybcF were investigated by calculating the G+C content percentage in GC Content Calculator (https://www.biologicscorp.com/tools/ GCContent/). Low G+C content provides a relative low stacking energy between strands of double-stranded DNA, which enables it to open at the initiation of transcription. We found a low G+C content region ca. 400 bps in length in front of the fdrA gene, which is only the promoter region we expected. The region in front of ybcF shows slightly low G+C content, but it is very short (Fig. 1A). In addition, two bioinformatics tools using iPro54-PseKNC and iPro70-FMWin predicted binding motifs for δ54 and δ70 upstream 334 bps (score: 0.9121) and upstream 271 bps (score: 0.9978) from the fdrA gene, respectively, which were consistent with the G+C content prediction [21,22].
We investigated and found that ybcF was co-transcribed with fdrA, ylbE, and ylbF, i.e., they were transcribed to one polycistronic mRNA. Total mRNA was isolated from a wild-type strain of E. coli grown anaerobically on allantoin as the sole nitrogen source. The total mRNA was reverse-transcribed into cDNA that was then amplified with every conceivable combination of primer pairs that bind to the fdrA, ylbE, ylbF, and ybcF genes (Fig. 1B). The results showed that each of the four genes, fdrA, ylbE, ylbF, and ybcF, was well and comparably expressed under allantoin degradation conditions (lanes 1 to 4). Lanes 5 to 10 show that the total mRNA contained all possible combinations of these neighboring gene transcripts, including two-gene transcripts (fdrA-ylbE, ylbE-ylbF, ylbF- ybcF), three-gene transcripts (fdrA-ylbE-ylbF, ylbE-ylbF-ybcF), and the four-gene transcript (fdrA-ylbE-ylbF-ybcF), indicating that the ybcF gene belongs to the fdrA operon.

Deletion of the ybcF Gene Does Not Affect Allantoin Degradation
Because ybcF is part of the fdrA operon, its expression will be controlled by a common promoter located upstream of fdrA. In a recent study, we analyzed the expression of the fdrA operon using fdrA-lacZ reporter gene fusion and found that its expression was dramatically induced by the presence of allantoin under anaerobic conditions [16]. Therefore, ybcF (as well as fdrA) was expected to play a role in the anaerobic degradation of allantoin ( Fig. 2A). Moreover, Smith et al. used bioinformatic tools to consider the genome and metabolic context with sequence orthology and predicted that ybcF would encode a carbamate kinase [17]. Therefore, we also assumed that YbcF could be a catabolic carbamate kinase associated with anaerobic allantoin degradation. The proposed reaction would produce ATP from carbamoyl phosphate (CP): CP + ADP → carbamate + ATP. To verify that reaction, we cultivated the wild-type and ΔybcF (LMB111) strains of E. coli anaerobically on allantoin with glycerol plus DMSO in nitrogen-deficient M9 medium for 48 h. However, no significant change in allantoin degradation due to ybcF deletion was observed (Table S1). Next, allantoin degradation by the ΔybcF strain was observed using a concentrated cell suspension, and no noticeable changes were observed, similar to the results in growth culture (Table 1, Table S1). Allantoin is degraded into oxalurate and then divided into equivalent amounts of oxamate and CP ( Fig. 2A). CP is a very unstable substance, whereas oxamate is a dead-end metabolite [23]. Therefore, we quantified oxamate and estimated that the same amount of CP was generated. It could be that the allantoin pathway was not interrupted in the ΔybcF strain because of the rapid consumption of CP for arginine and pyrimidine biosynthesis under the tested nitrogen-deficient growth conditions. Therefore, we conducted the following direct carbamate kinase assay to identify the enzyme YbcF.

E. coli YbcF Shows Activity as a Catabolic Carbamate Kinase (CK)
During anaerobic allantoin degradation, oxalurate is degraded into CP and oxamate by OXTCase (Fig. 2A). The phosphate group of CP is transferred to ADP by CK to produce ATP, leaving carbamate, which is then spontaneously mineralized to HCO 3 and NH 4 + (Fig. 2A). The CK activity in this study was determined by quantifying the ATP produced, which we measured by coupling it to a recombinant firefly luciferase and its substrate D -luciferin (Fig. 2B) [24].
First, the CK activity of YbcF was evaluated using crude cell lysate (Fig. 3). The cell lysate containing overproduced YbcF (Fig. 3A) clearly showed higher CK activity (38 μmole/min/g total protein) than the wild-type and ΔybcF strains (about 16 μmole/min/g total protein) (Fig. 3B). The lack of difference in CK activity between wild-type and ΔybcF strains could be due to the high background caused by the use of crude cell lysate, as reflected by the following two factors. This enzyme assay was unable to distinguish CK activity from acetate kinase and carbamoyl phosphate synthetase activity [25], and ATP was already present in the cell lysates. Therefore, their sum was measured as CK activity, and that value represented quite a high background. Thus, we found that YbcF had CK activity by using crude cell lysate from an YbcF-overexpressing strain.  The bacterial strains were anaerobically grown on glycerol (50 mM), DMSO (50 mM), and allantoin (20 mM) in nitrogendeficient M9 medium and resuspended in nitrogen/carbon-deficient MOPS buffer to an OD 600 of 20. The reaction was started by adding allantoin (20 mM), glycerol (20 mM), and DMSO (20 mM). WT, wild-type; ΔybcF, ybcF deletion mutant; gCDW, g cell dry weight The OD 600 of the cell suspension maintained; initial pH value of 7.0 increased to 7.4 at the end of incubation of both WT and ΔybcF. a Unpaired two-tailed student's t-tests were performed to analyze WT and ΔybcF. b Values represent avg ± SD for 3 replicates.

Carbamate Kinase Kinetics of YbcF
The optimal temperature and pH value for the CK enzyme assay were determined with purified YbcF. The overproduction and purification of YbcF were confirmed by SDS-PAGE and Western blotting (Fig. 4), and 100 nM YbcF was used for each assay. The temperature was tested at 28°C and 37°C, the former being the recommended temperature for ATP quantification (A22066), and the latter the optimum temperature for the  carbamate kinase reaction (E.C.2.7.2.2). Because CK activity was twice as high at 28°C (1.29 μmole/min/mg YbcF) than at 37°C (0.61 μmol/min/mg YbcF), the temperature selected was 28°C (Fig. 5A). The chosen pH value was pH 8 (1.37 μmole/min/mg YbcF). Although there was no significant difference at pH 6, 7, and 8 (Fig. 5B), the luminescence-based CK assay previously was performed at pH 8 [24].
The CK kinetics of YbcF were investigated at 28°C and pH 8 for 60 min, and we calculated the amount of ATP produced as U/mg YbcF protein. CK has two substrates in the forward direction for allantoin degradation: CP and ADP. Thus, CK activity was measured depending on the concentration of those two substrates. Keeping ADP fixed at 2 mM, activity was measured as we changed the concentration of CP from 0-3 mM, and vice versa (Fig. 6). When CP was varied, V max_CP was 1.82 μmol/min/mg YbcF, and K M_CP was 0.47 mM. The values were similar when ADP was varied: V max_ADP was 1.94 μmol/min/mg YbcF, and K M_ADP was 0.43 mM (Fig. 6).

Discussion
We found in this study that catabolic CK is encoded by the ybcF gene belonging to an operon (b0518-0521) downstream of the allantoin cluster in the E. coli genome. That operon begins with the fdrA gene encoding a potential OXTCase, a catabolic transcarbamoylase [16]. This is somehow similar to that in L. brevis, where the arcC encoding the catabolic CKs is located together with the arcB for the catabolic transcarbamoylase OTCase in the ADI operon [12]. However, the catabolic CK of E. coli (ybcF) should be distinguished from arcC of L. brevis in that it is on the allantoin degradation pathway, not the arginine deiminase pathway.
The K M value of YbcF (0.47 mM) for CP is relatively low compared with those of catabolic CKs determined in other bacteria: S. pyogenes (0.65 mM), E. faecalis (1.40 mM), P. aeruginosa (5 mM), and L. buchneri (1.1 mM, 1.53 mM) ( Table 2) [26][27][28][29]. On the other hand, in CKs of two parasitic/pathogenic protozoa, the K M values for CP are very low, G. intestinalis (0.085 or 0.34 mM) and T. vaginalis (0.13 mM) ( Table 2), which is thought to indicate that they evolved to easily use the energy-rich CP (ΔG 0 ′ = -39 kJ/mol) produced by the host as an energy source [30][31][32]. In addition, the K M value of YbcF in E. coli for ADP (0.43 mM) is similar to that for CP (0.47 mM)  ( Table 2), showing that YbcF has similar affinity for the two substrates, so it seems easy to perform catabolic CK. Overall, the K M values indicate that YbcF, as a catabolic CK, has a higher affinity for the substrates CP and ADP than those of other nonpathogenic bacteria.
E. coli can use allantoin, the main product of purine degradation, as a sole nitrogen source, but only under anaerobic conditions [33]. The conversion steps from allantoin to oxalurate are ring opening, deamination, and oxidation, and they are catalyzed by enzymes encoded by allB (b0512), allC (b0516), allE (b0515), and allD (b0517) (Fig. 7, upper vertical route). Subsequently, oxalurate is converted to oxamate and CP by potential OXTCase (probably fdrA, b0518), and CP is dephosphorylated by CK (b0521) to carbamate, which spontaneously degrades (Fig. 7, lower vertical route). Alternatively, ureidoglycolate is divided into urea and glyoxylate by ureidoglycolate lyase (allA, b0505), and the latter is then oxidized to oxalate [16,33]. We predicted that the ybcF deletion strain (ΔybcF) would not decompose CP, and the path would be bypassed from ureidoglycolate to glyoxylate, resulting in an increase in oxalate and decrease in oxamate (and CP). Although changes that matched those predictions did occur, the decrease in oxamate (and CP) was very slight ( Table 1, Table S1), indicating that CP continues to be used in ΔybcF mutant. The CP used in biosynthesis is produced by CPS in E. coli, with the amino group coming from glutamine, and the reaction requires 2 ATP (Fig. 7). Furthermore, the culture medium used in this study contained no nitrogen source other than allantoin. Therefore, if CP is available in the cell, it would be beneficial to use it first. Compared with CK YbcF (K M 0.47 mM), both of the OTCase isoenzymes, ArgF (K M 0.36 mM) and ArgI (K M 0.05 mM), exhibit higher affinity for CP in arginine biosynthesis (Fig. 7, Table 3). Also, ATCase (K M 0.20 mM) shows a higher affinity for CP in pyrimidine biosynthesis (Fig. 7, Table 3). Therefore, the CP in the cell is preferentially used for de novo synthesis of arginine or pyrimidine, and the remaining CP is used for ATP generation via the catabolic carbamate kinase because energy sources are rare in anaerobic environments.