The Lreu_1276 protein from Limosilactobacillus reuteri represents a third family of dihydroneopterin triphosphate pyrophosphohydrolases in bacteria

ABSTRACT

Tetrahydrofolate is a cofactor involved in C₁ metabolism including biosynthesis pathways for adenine and serine. In the classical tetrahydrofolate biosynthesis pathway, the steps removing three phosphate groups from the precursor 7,8-dihydroneopterin triphosphate (DHNTP) remain unclear in many bacteria. DHNTP pyrophosphohydrolase hydrolyzes pyrophosphate from DHNTP and produces 7,8-dihydroneopterin monophosphate. Although two structurally distinct DHNTP pyrophosphohydrolases have been identified in the intestinal bacteria Lactococcus lactis and Escherichia coli, the distribution of their homologs is limited. Here, we aimed to identify a third DHNTP pyrophosphohydrolase gene in the intestinal lactic acid bacterium Limosilactobacillus reuteri. In a gene operon including genes involved in dihydrofolate biosynthesis, we focused on the lreu_1276 gene, annotated as Ham1 family protein or XTP/dITP diphosphohydrolase, as a candidate encoding DHNTP pyrophosphohydrolase. The Lreu_1276 recombinant protein was prepared using E. coli and purified. Biochemical analyses of the reaction product revealed that the Lreu_1276 protein displays significant pyrophosphohydrolase activity toward DHNTP. The optimal reaction temperature and pH were 35°C and around 7, respectively. Substrate specificity was relatively strict among 17 tested compounds. Although previously characterized DHNTP pyrophosphohydrolases prefer Mg²⁺, the Lreu_1276 protein exhibited maximum activity in the presence of Mn²⁺, with a specific activity of 28.2 ± 2.0 µmol min⁻¹ mg⁻¹ in the presence of 1 mM Mn²⁺. The three DHNTP pyrophosphohydrolases do not share structural similarity to one another, and the distribution of their homologs does not overlap, implying that the Lreu_1276 protein represents a third structurally novel DHNTP pyrophosphohydrolase in bacteria.

IMPORTANCE

The identification of a structurally novel DHNTP pyrophosphohydrolase in L. reuteri provides valuable information in understanding tetrahydrofolate biosynthesis in bacteria that possess lreu_1276 homologs. Interestingly, however, even with the identification of a third family of DHNTP pyrophosphohydrolases, there are still a number of bacteria that do not harbor homologs for any of the three genes while possessing other genes involved in the biosynthesis of the pterin ring structure. This suggests the presence of an unrecognized DHNTP pyrophosphohydrolase gene in bacteria. As humans do not harbor DHNTP pyrophosphohydrolase, the high structural diversity of enzymes responsible for a reaction in tetrahydrofolate biosynthesis may provide an advantage in designing inhibitors targeting a specific group of bacteria in the intestinal microbiota.

INTRODUCTION

It is well established that intestinal bacteria are closely related to human health (1–3). Beneficial intestinal bacteria including Limosilactobacillus reuteri (Lactobacillus reuteri) (4–6) synthesize vitamins and supply them to the host (1, 2, 7), suppress the growth of harmful bacteria (1, 2, 8), and modulate the human immune system (1, 2). On the other hand, some intestinal bacteria that produce toxic substances such as Clostridium perfringens (9–11) are also present in the intestines. Various bacteria together with archaea, eukaryotes, and viruses from the intestinal microbiota and its composition can differ according to the individual and the period in life (12, 13). Some of the beneficial effects are known to be induced by metabolites produced by intestinal bacteria from oligosaccharides undegradable by humans and nutrients slipping through degradation by the human digestive system (14). For example, short-chain fatty acids including acetate and compounds such as lactate provoke various beneficial effects. Acetate produced by Bifidobacterium longum subsp. longum JCM 1217^T can protect germ-free mice from lethal infection of Escherichia coli O157:H7 (15). It is also known that humans acquire some vitamins, including thiamine, riboflavin, biotin, and folate, from intestinal bacteria as well as from food. Therefore, in addition to the ecology of gut bacterial flora, understanding the metabolism of individual intestinal bacteria should provide insight into their contribution to the host as well as to the sustainment of the bacterial community.

Folate is a water-soluble vitamin classified into the vitamin B group. Tetrahydrofolate, the reduced form of folate, is the functional cofactor for enzymes. The cofactor is involved in the biosynthesis of adenine nucleobase and amino acids such as serine and methionine. The C₂ and C₈ carbons in adenine are generally derived from the carbon of N⁵N¹⁰-methenyltetrahydrofolate and N¹⁰-formyltetrahydrofolate, respectively. Plants, yeasts, some bacteria, and some protists are considered to be able to synthesize tetrahydrofolate. On the other hand, although humans and other higher animals utilize tetrahydrofolate, they are unable to synthesize the cofactor. Figure 1 shows the tetrahydrofolate biosynthesis pathway in L. reuteri, which is predicted based on the presence of gene homologs, whereas the classical tetrahydrofolate biosynthesis pathway based on knowledge summarized in previous reviews (16, 17) and alternative pterin biosynthesis pathways are shown in Fig. S1 (red and other colored arrows, respectively). In the first reaction of the classical pathway, GTP is converted to 7,8-dihydroneopterin triphosphate (DHNTP) by GTP cyclohydrolase I (FolE). The triphosphate of DHNTP is then sequentially removed by DHNTP pyrophosphohydrolase (FolQ) and an unknown phosphatase or by a non-specific alkaline phosphatase to generate 7,8-dihydroneopterin (DHN). The two enzymes dihydroneopterin aldolase (FolB) and 2-amino-4-hydroxy-6-hydroxymethyldihydropteridine diphosphokinase (FolK) convert DHN to 6-hydroxymethyl-7,8-dihydropterin diphosphate (DHPPP) via 6-hydroxymethyl-7,8-dihydropterin (6-HMD). To synthesize dihydrofolate, DHPPP is condensed with p-aminobenzoate (pABA) and ligated with glutamate by dihydropteroate synthase (FolP) and dihydrofolate synthase (FolC). Dihydrofolate is oxidized or reduced to folate or tetrahydrofolate, respectively, by dihydrofolate reductase (FolA). As mammals including Homo sapiens lack at least FolK and FolP, and proteins synthesizing pABA from chorismate, they are considered unable to synthesize the cofactor de novo. Alternative pathways to synthesize pterin rings have been reported (Fig. S1). In some archaea, DHNMP is synthesized from GTP via 7,8-dihydroneopterin 2’,3’-cyclic phosphate (18, 19), and conversion of DHN to DHPPP is catalyzed by archaea-type enzymes (20). In the bacterial genus Chlamydia, DHNMP is synthesized from GTP through 2,5-diamino-6-(5-phospho-d-ribosylamino)pyrimidin-4(3H)-one (DARP), a precursor of riboflavin, by GTP cyclohydrolase II (GCH-II) and N-(5'-phosphoribosyl)anthranilate isomerase (TrpF_CtL2) (21). It has also been reported that 6-pyruvoyltetrahydropterin synthase (PTPS-III) paralogs, which can be found on genomes of some protists, archaea, and bacteria, catalyze the one-step synthesis of 6-HMD from DHNTP (22, 23).

Fig 1.

The predicted tetrahydrofolate biosynthesis pathway in L. reuteri. Gene locus tags are shown in parenthesis. Red broken arrows indicate reactions catalyzed by enzymes that have not been identified in L. reuteri. Abbreviations: DHNTP, 7,8-dihydroneopterin triphosphate; DHN, 7,8-dihydroneopterin; 6-HMD, 6-hydroxymethyl-7,8-dihydropterin; DHPPP, 6-hydroxymethyl-7,8-dihydropterin diphosphate; DHP, 7,8-dihydropteorate.

Among the reactions in classical folate biosynthesis, the enzymes involved in the removal of the three phosphate groups of DHNTP to form DHN are unclear in many bacteria (Fig. 1; Fig. S1). In the intestinal bacteria Lactococcus lactis (24) and Escherichia coli (25, 26), two structurally distinct DHNTP pyrophosphohydrolases have been identified. To emphasize the differences in structure between the two proteins and the protein identified in this study, we designate the former as Ll-FolQ1 (encoded by ylgG) and the latter as Ec-FolQ2 (encoded by nudB). Although these two proteins harbor the Nudix motif, a characteristic domain in proteins cleaving phosphate ester bonds between nucleoside diphosphate and other moieties (27), they share only 13% identity with each other. The distribution of homologs encoding FolQ1 or FolQ2 proteins is confined to a limited number of bacteria, and the DHNTP pyrophosphohydrolase gene is still unidentified in many bacteria including L. reuteri.

Using genome information, here, we identified a candidate gene encoding the missing DHNTP pyrophosphohydrolase in L. reuteri, a representative of well-characterized intestinal bacteria. The function of the gene was unidentified, and the gene was present within a gene cluster consisting of multiple genes related to folate biosynthesis. Here, we carried out biochemical analyses on the gene product and showed that the protein displays DHNTP pyrophosphohydrolase activity and represents a third family of DHNTP pyrophosphohydrolases.

RESULTS

Candidate gene encoding DHNTP pyrophosphohydrolase in L. reuteri

Although two DHNTP pyrophosphohydrolases, Ll-FolQ1 and Ec-FolQ2, have been identified in intestinal bacteria L. lactis and E. coli, respectively, the distribution of their homologs is limited in bacteria (Fig. 2; Table S1). A number of bacteria including Limosilactobacillus reuteri, Corynebacterium glutamicum, Streptococcus pyogenes, and other diverse species shown in Fig. 2 do not possess genes encoding a protein displaying significant similarity with the two FolQ enzymes. In this study, we aimed to identify a gene encoding DHNTP pyrophosphohydrolase in the intestinal lactic acid bacterium L. reuteri. In searching for a candidate gene, we first focused on the genes in an operon including other dihydrofolate biosynthesis genes, folP, folC, folE, folK, and folB, (lreu_1275, lreu_1277, lreu_1278, lreu_1279, and lreu_1280, respectively) on the L. reuteri DSM 20016 genome (Fig. 3). The lreu_1276 gene is located in the dihydrofolate biosynthetic operon, but its function is unclear. The Lreu_1276 protein is annotated as Ham1 family protein in GenBank and as XTP/dITP diphosphohydrolase in KEGG database. The possibility that the Lreu_1276 protein displays DHNTP pyrophosphohydrolase activity has been pointed out (17) but has not been experimentally established. The amino acid sequence of the lreu_1276 gene product does not display significant identity with those of Ll-FolQ1 and Ec-FolQ2 proteins (15% and 10%) and does not seem to possess the Nudix motif, which is found in the two previously identified FolQ proteins (Fig. S2).

Fig 2.

Distribution of genes involved in dihydrofolate biosynthesis. Black and white circles indicate the presence and absence of gene homolog, respectively. In folEBKP column, gray circles with EKP and BKP show the presence of folEKP and folBKP gene homologs, respectively. G-T indicates a set of GCH-II and TrpF_CtL2. Proteins displaying e-values less than 1 × 10e-5, 6 × 10e-25, and 2 × 10e-17 toward PTPS-III (CLC_0882), GCH-II (CT_731), and TrpF_CtL2 (Ctl0581) were recognized as their homologs, respectively. The PTPS-III protein and GCH-II/TrpF_CtL2 proteins can form pathways that bypass those composed of FolQB and FolEQ proteins, respectively. Red letters indicate organisms harboring homologs of the folQ3 gene identified in this study, whereas blue letters indicate organisms without any already-identified folQ genes, although they possess other genes predicted to be involved in dihydrofolate biosynthesis (folE, folB, folK, and folP) and do not possess possible bypass pathways. †Organisms with GTP cyclohydrolase IB homolog instead of FolE homolog. *Organisms harboring a set of FolE, FolB, FolK, and FolP homologs, although similarity between some of the proteins and corresponding query proteins is lower than the threshold in Table S1.

Fig 3.

Schematic diagram of the predicted dihydrofolate biosynthesis operon in L. reuteri. Arrowed boxes indicate genes in the predicted dihydrofolate biosynthesis gene cluster on the L. reuteri genome. lreu_1275 and lreu_1277-lreu_1280 genes are predicted to encode proteins involved in dihydrofolate biosynthesis shown in Fig. 1. The lreu_1276 gene, forming an operon with these genes, is annotated as Ham1 family protein in GenBank and as XTP/dITP diphosphohydrolase in KEGG database and its function is unclear.

Preparation of the purified recombinant Lreu_1276 and Ec-FolE proteins

In order to examine the enzyme activity of the Lreu_1276 recombinant protein, the lreu_1276 gene (Fig. S3) was expressed in E. coli. In addition, a gene encoding the GTP cyclohydrolase I protein from E. coli (Ec-FolE) was also expressed in E. coli for its use in enzymatic synthesis of DHNTP, the substrate for DHNTP pyrophosphohydrolase, which is not commercially available. The recombinant proteins were purified by nickel affinity chromatography and gel filtration chromatography. Both proteins were purified to apparent homogeneity (Fig. 4).

Fig 4.

Purified Ec-FolE and Lreu_1276 recombinant proteins. One microgram of Ec-FolE (A) and 4 μg of Lreu_1276 (B) recombinant proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Gels were stained with Coomassie Brilliant Blue. M indicates molecular mass marker.

Enzymatic synthesis of the substrate for DHNTP pyrophosphohydrolase

DHNTP, the substrate for DHNTP pyrophosphohydrolase, was prepared from GTP using the Ec-FolE recombinant protein. The reaction product was analyzed by high-performance liquid chromatography(HPLC), and compounds displaying absorbance at 330 nm, characteristic of the pterin ring structure, were detected. As a result, a new peak was observed only when the enzyme was added into the reaction mixture, implying that DHNTP was synthesized by the enzymatic reaction (Fig. 5A). In addition, liquid chromatography-mass spectrometry (LC-MS) analysis of the reaction product revealed that a compound with the exact mass corresponding to that of deprotonated DHNTP ([M-H]^- = 493.989 ± 5 ppm m/z) was detected only in the presence of the Ec-FolE enzyme (Fig. 5B and C), indicating that the reaction product is DHNTP.

Fig 5.

The HPLC and LC-MS analyses of Ec-FolE reaction products. (A) HPLC analysis was carried out toward reaction products without and with Ec-FolE enzyme indicated with blue and red lines, respectively. LC-MS analysis was performed toward reaction products without (B) and with the enzyme (C).

DHNTP pyrophosphohydrolase activity of the Lreu_1276 recombinant protein

The pyrophosphohydrolase activity of the Lreu_1276 recombinant protein was examined using the enzymatically synthesized DHNTP. The reaction product was analyzed by HPLC. When the Lreu_1276 recombinant protein was added into the reaction mixture, the peak area of DHNTP decreased, and we could observe a new peak whose retention time was almost identical to that of the DHNMP standard (Fig. 6A). A shoulder peak observed in the DHNTP peak (blue line in Fig. 6A) also decreased and was most likely converted to DHNMP in the presence of the enzyme (red line in Fig. 6A), suggesting that the shoulder peak is not a degradation product of DHNTP but DHNTP itself. It is presumed that DHNTP forms two charged states, which are separated by liquid chromatography.

Fig 6.

The HPLC and LC-MS analyses of Lreu_1276 reaction products. (A) HPLC analysis was carried out toward Lreu_1276 reaction product. Black, blue, and red lines indicate the DHNMP standard compound, reaction product without the enzyme, and reaction product with the enzyme, respectively. LC-MS analysis was performed toward reaction products without (B) and with the enzyme (C).

LC-MS analysis of the reaction product was also carried out. As a result, in the mass chromatogram of the exact mass of deprotonated DHNMP ([M-H]^- = 334.056 ± 5 ppm m/z), a peak corresponding to DHNMP was clearly observed in the presence of the enzyme in the reaction mixture (Fig. 6B and C). Although it is unclear whether the released product is phosphate or pyrophosphate, the results indicated that this enzyme catalyzes the synthesis of DHNMP from DHNTP.

Metal ion dependency of the Lreu_1276 recombinant protein

Since Mg²⁺ cations were utilized as metal ions for two previously characterized DHNTP pyrophosphohydrolases (24, 26), Mg²⁺ was added to the reaction mixture in the experiments described above. Nevertheless, metal cation specificity of the Lreu_1276 protein was examined (Fig. 7A). Activity in the presence of MnCl₂, MgCl₂, ZnCl₂, CaCl₂, CoCl₂, NiCl₂, CuCl₂ (1 mM each), or no divalent metal ion was tested. The Lreu_1276 protein exhibited detectable pyrophosphohydrolase activity in the presence of Mn²⁺, Mg²⁺, and Zn²⁺ ions (9.8 ± 2.5, 0.47 ± 0.06, and 0.13 ± 0.03 µmol min⁻¹ mg⁻¹, respectively) (Fig. 7A). Interestingly, the enzyme exhibited the highest activity in the presence of Mn²⁺, suggesting that the Lreu_1276 protein is a Mn²⁺-dependent enzyme, whereas FolQs identified in previous studies have been reported to prefer Mg²⁺ to catalyze the phosphohydrolysis reaction of DHNTP.

Fig 7.

Metal ion dependency of DHNTP pyrophosphohydrolase reaction catalyzed by the Lreu_1276 recombinant protein. (A) DHNTP pyrophosphohydrolase activity was measured in the presence of various metal ions (1 mM each). For examination of Mn²⁺, Mg²⁺, and Zn²⁺ ions, 100 mM DHNTP was used as a substrate, whereas for other ions, 150 mM DHNTP was used as a substrate. (B) DHNTP pyrophosphohydrolase activity was measured in the presence of various concentrations of Mn²⁺ ions. Error bars indicate the standard deviations of three independent experiments.

In order to examine Mn²⁺ ion dependency, specific activity in the presence of various concentrations of Mn²⁺ ion was measured (Fig. 7B; Fig. S4). The pyrophosphohydrolase activity toward DHNTP was enhanced with increasing Mn²⁺ concentration, at least up to 8 mM.

Reaction temperature and pH dependencies of the Lreu_1276 recombinant protein

The effects of reaction temperature on activity of the Lreu_1276 protein were examined between 10°C and 45°C. The enzyme showed the highest activity at 35°C (Fig. 8A). The effects of pH of the reaction mixture on activity were also investigated between pH 5.5 and 8.0. The Lreu_1276 protein showed higher activity around pH 6.5–7.0 (Fig. 8B). When the reaction was carried out in 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.0) and 37°C with 1 mM Mn²⁺, the specific activity was 28.2 ± 2.0 µmol min⁻¹ mg⁻¹. Based on the specific activity, frequency to convert substrate to product was 10.6 s⁻¹, which is comparable to the k_cat value of Ec-FolQ2 (11.6 s⁻¹) and higher than that of Ll-FolQ1 (0.21 s⁻¹) (24, 25).

Fig 8.

Enzymatic properties of Lreu_1276 recombinant protein. (A) Effects of reaction temperature on pyrophosphohydrolase activity of Lreu_1276 protein were examined. (B) Effects of pH on pyrophosphohydrolase activity were investigated. Error bars represent the standard deviations of three independent experiments.

Reaction stoichiometry and reaction mechanism

We examined the products and stoichiometry of the reaction. To avoid the titration and removal of pyrophosphate and phosphate by manganese, reactions were performed with minimized concentrations of Mn²⁺ ions (0.1 mM). After an enzyme reaction using 107 µM DHNTP as substrate, 106 ± 1 µM DHNMP and 88 ± 6 µM pyrophosphate were detected. We did not observe the production of free phosphate. This result suggested that the enzyme is not a DHNTP phosphatase subsequently releasing two phosphate groups but a DHNTP pyrophosphohydrolase, releasing pyrophosphate.

Substrate specificity of the Lreu_1276 recombinant protein

Substrate specificity of the Lreu_1276 protein was examined using various compounds as a substrate at a concentration of 100 µM (Table 1). The Lreu_1276 protein possesses the Ham1 domain, and enzymes with this domain have been reported to remove pyrophosphate from non-standard nucleotides such as XTP and ITP (28). On the other hand, in a previous study (24), the At1g68760 protein from Arabidopsis thaliana, which is a homolog of Ll-FolQ1, was found to be active toward some Nudix-type substrates. Therefore, in addition to DHNTP, ten (deoxy)nucleoside 5’-triphosphates including GTP and six Nudix-type substrates such as NAD⁺, NADH, diadenosine diphosphate, diadenosine tetraphosphate, ADP-ribose, and UDP-glucose were tested as substrates. As all pyrophosphohydrolase or phosphohydrolase reactions provide nucleoside 5’-monophosphate (NMP) as a product, activity was measured by HPLC analysis detecting NMPs, which display absorbance at 254 nm. When GTP, ITP, XTP, and dTTP as well as DHNTP were tested, reaction products could be detected (Fig. S5). However, no detectable activity was found when other substrates were examined. The concentrations of reaction products, GMP, IMP, XMP, and dTMP, were determined based on the calibration curves constructed with each standard compound. In particular, when DHNTP and XTP were used as substrates, higher concentrations of reaction products, DHNMP and XMP, were detected. Therefore, specific activities of the enzyme were determined toward these two substrates. The specific activity of Lreu_1276 recombinant protein toward DHNTP was 28.2 ± 2.0 µmol min⁻¹ mg⁻¹ as described above, whereas that toward XTP was 0.7 ± 0.2 µmol min⁻¹ mg⁻¹. These results implied that the substrate specificity of Lreu_1276 protein is relatively strict, suggesting that the Lreu_1276 protein represents a structurally novel third family of bacterial DHNTP pyrophosphohydrolases and was thus designated Lr-FolQ3.

TABLE 1.

Substrate specificity of the Lreu_1276 protein

Substrate 100 µM	Concentration of the product [μM]
DHNTP	94.1
GTP	2.5
ITP	3.4
XTP	13.0
ATP	Not detected
CTP	Not detected
UTP	Not detected
dTTP	3.1
dATP	Not detected
dCTP	Not detected
dGTP	Not detected
NAD+	Not detected
NADH	Not detected
Diadenosine triphosphate	Not detected
Diadenosine tetraphosphate	Not detected
ADP-ribosea	Not detected
UDP-glucose	Not detected

^a ADP-ribose was synthesized by heat treatment of 100 µM NAD⁺, and its concentration was not determined.

Three-dimensional structures of DHNTP pyrophosphohydrolases

The three-dimensional structures of the three DHNTP pyrophosphohydrolases were compared. Structure models of Ll-FolQ1 and Lr-FolQ3 predicted with AlphaFold2 (29) and deposited in AlphaFold Protein Structure Database (AF-P0CI35-F1 and AF-A5VL09-F1, respectively) were utilized, along with the crystal structure of the monomeric form of Ec-FolQ2 (PDB-ID: 2O5W) reported in a previous study (25). The predicted structure of Lr-FolQ3 possesses β-sheets at the center surrounded by multiple α-helices (Fig. 9A), whereas those of Ll-FolQ1 (Fig. 9B) and Ec-FolQ2 (Fig. 9C) also harbor similar components. The three-dimensional structures of Ll-FolQ1 and Ec-FolQ2 overlapped relatively well (root mean square deviation [RMSD] = 4.7 Å, 639 to 639 atoms) (Fig. 9D), whereas the Lr-FolQ3 protein exhibited no structural similarity with Ll-FolQ1 (RMSD = 7.6 Å, 238 to 238 atoms) (Fig. 9E). When Lr-FolQ3 and Ec-FolQ2 were compared, although partial structures overlapped (RMSD = 4.6 Å, 116 to 116 atoms), the overall structures clearly differed (Fig. 9F). These structural analyses also suggest the structural novelty of Lr-FolQ3 among DHNTP pyrophosphohydrolases.

Fig 9.

Structural analysis of Ll-FolQ1, Ec-FolQ2, and Lr-FolQ3 proteins. The structures of DHNTP pyrophosphohydrolases from L. reuteri (Lr-FolQ3) (A) and L. lactis (Ll-FolQ1) (B) predicted by AlphaFold2 and deposited in AlphaFold Protein Structure Database (AF-A5VL09-F1-model_v4 and AF-P0CI35-F1-model_v4, respectively) were utilized. Structure of the enzyme from E. coli (monomeric form Ec-FolQ2), which has been published in PDB database (2O5W), was also utilized (C). Structures of Ll-FolQ1 (magenta) and Ec-FolQ2 (green) (D), Lr-FolQ3 (cyan) and Ll-FolQ1 (magenta) (E), and Lr-FolQ3 (cyan) and Ec-FolQ2 (green) (F) were superimposed. Superimposition was performed utilizing PyMOL software with a cutoff of 2.0.

Distribution of the third bacterial DHNTP pyrophosphohydrolase

Distribution of gene homologs encoding Ll-FolQ1, Ec-FolQ2, and the third DHNTP pyrophosphohydrolase identified in this study (Lr-FolQ3) was examined in 7,983 bacterial species whose genomic information is deposited in KEGG database as of September 26, 2023 (Table S1). Classification and numbers of organisms possessing the three DHNTP pyrophosphohydrolase homologs are listed in Table S2. In addition, the distribution of the three DHNTP pyrophosphohydrolases and other pterin biosynthesis genes, folE, folB, folK, and folP in 55 selected species, indicated by the yellow column in Table S1, are shown in Fig. 2. The number and distribution of Ll-folQ1 gene homologs (three species in Fig. 2 and 41 species in Table S1) are relatively restricted to 36 species in Lactobacillales Order of Bacilli Class and five species in Fusobacteriales Order of Fusobacteriia Class (Table S2). In terms of Ec-folQ2 homologs, they are widely distributed in three classes with members in 14 Orders (six species in Fig. 2 and 1148 species in Table S1); 8 Orders in the Gammaproteobacteria Class (Enterobacterales, Pasterellales, Alteromonadales, Chromatiales, Aeromonadales, Acidiferrobacterales, Orbales, and Thiotrichales), 5 Orders in the Betaproteobacteria Class and one Order in the Acidithiobacillia Class (Acidithiobacillales) (Table S2). On the other hand, although the number of organisms with homologs of the Lr-FolQ3 gene is relatively small (13 species in Fig. 2 and 157 species in Table S1) compared with those of Ec-FolQ2, homologs were observed in a wide range of bacteria, 22 Classes (including Bacilli, Bacteroidia, Flavobacteriia, Deinococci, Sphinobacteriia, Chitinophagia, Cytophagia, and Terriglobia) (Table S2). Although the Order of some species was not designated, the 22 Classes included 23 Orders such as Lactobacillales, Bacteroidales, Flavobacteriales, Sphigobacteriales, and Deinococcales. The presence of the three DHNTP pyrophosphohydrolase homologs does not overlap, adding further support that Lr-FolQ3 and its homologs function as DHNTP pyrophosphohydrolase.

DISCUSSION

In this study, we identified a third family of DHNTP pyrophosphohydrolases encoded by Lr-folQ3 (lreu_1276) in the intestinal lactic acid bacterium L. reuteri by focusing on a gene in a dihydrofolate biosynthesis gene cluster, which also includes folB, folE, folK, folP, and folC genes (Fig. 3), similar to the case of Ll-folQ1, which resides in a dihydrofolate biosynthesis operon together with folKE, folP, and folC genes on the genome of L. lactis subsp. cremoris MG1363 (24). On the other hand, Ec-folQ2 homologs in E. coli K-12 MG1655 and E. coli O157:H7 Sakai do not constitute operons with other tetrahydrofolate biosynthesis genes. It is noted that primary sequences of the DHNTP pyrophosphohydrolases identified in previous studies (Ll-FolQ1 and Ec-FolQ2) and this study do not exhibit significant similarity with each other (13% identity between Ll-FolQ1 and Ec-FolQ2, 15% identity between Ll-FolQ1 and Lr-FolQ3, and 10% identity between Ec-FolQ2 and Lr-FolQ3). This implies that enzymes catalyzing the step releasing pyrophosphate from DHNTP are diverse in structure.

The distribution of the three DHNTP pyrophosphohydrolase proteins in nature does not seem to correlate with ecological niches. We do observe that the majority of Ec-folQ2 homologs are confined to the gamma- and beta-proteobacteria, and this may reflect the emergence of the protein in a common ancestor of these Classes (30, 31), combined with gene loss in some members. In terms of Ll-folQ1 homologs, the majority are confined to Lactobacillales, and the protein may have recently emerged in this Order. The distribution of Lr-folQ3 homologs is difficult to interpret in terms of evolution, as they are found in a particularly wide range of bacteria (Fig. 2). In many cases, only single/several species of a given genus harbor the Lr-folQ3 homolog (Tables S1 and S2), which may reflect the occurrence of widespread horizontal gene transfer.

As described above, Ll-FolQ1, Ec-FolQ2, and Lr-FolQ3 homologs were found in 41, 1,148, and 157 species, respectively, and the distribution of their homologs do not overlap in any of the 1,346 species, seeming to complement each other. Therefore, DHNTP pyrophosphohydrolases in the 1,346 bacterial species have most likely been identified among the 7,983 bacterial species shown in Table S1. However, this is probably an underestimation as we can consider the possibility that there are additional DHNTP pyrophosphohydrolases displaying e-values higher than the thresholds we set in Table S1 (2 × 10e-34, 5 × 10e-18, and 3 × 10e-17 for Ll-FolQ1, Ec-FolQ2, and Lr-FolQ3 proteins, respectively). Identifying FolQ1 homologs was straightforward, as we observed a clear drop in the degree of similarity among proteins obtained in the homology search with the protein basic local alignment search tool (BLASTp). Forty-one proteins displayed e-values below 2 × 10e-34, whereas the next most similar proteins only had e-values of 2 × 10e-18 or higher, revealing a large gap in similarity. On the other hand, no such gaps in e-values were observed in the search for FolQ2 and FolQ3 homologs, and there is a possibility that proteins displaying e-values higher than our thresholds may also be DHNTP pyrophosphohydrolases. Even considering this underestimation, there still seems to be a large number of bacteria that do not harbor any of the three FolQ proteins. Therefore, to evaluate the presence or absence of a new family of DHNTP pyrophosphohydrolase, gene distribution of folEBKP homologs was re-examined in the organisms shown in Fig. 2. As a result, we could identify that additional organisms, indicated with asterisks in Fig. 2, harbor complete sets of folEBKP genes by lowering our e-value threshold, whereas in organisms, indicated with daggers in Fig. 2, FolE protein is replaced with an alternative GTP cyclohydrolase IB enzyme (32). In addition, we could find that some organisms harbor a gene most likely encoding PTPS-III and a set of genes encoding GTP cyclohydrolase II and TrpF_CtL2 (“PTPS-III” and “G-T” columns in Fig. 2, respectively). These proteins can form metabolic pathways that bypass the reaction catalyzed by DHNTP pyrophosphohydrolase (Fig. S1). Although some organisms including Aquifex aeolicus, Hydrogenobacter thermophilus, Bacillus subtilis, Staphylococcus aureus, and Synechococcus elongatus lack the gene encoding DHNTP pyrophosphohydrolase, it is likely that their PTPS-III proteins can complement the enzyme (Fig. S1). Characterization of PTPS-III homologs in these organisms will clarify this hypothesis.

On the other hand, organisms, with FolEBKP proteins but with neither the three DHNTP pyrophosphohydrolases nor bypass pathway (indicated by blue letters in Fig. 2) can be predicted to harbor unrecognized family of DHNTP pyrophosphohydrolase and/or new bypass pathway(s). These include Bifidobacterium longum, Streptomyces coelicolor, Corynebacterium glutamicum, Cutibacterium acnes, Streptococcus pyogenes, Clostridium perfringens, Thermus thermophilus, Shewanella oneidensis, and Legionella pneumophila. Further studies will be necessary to identify the enzyme catalyzing the step(s) and/or the bypass pathway(s).

Our results suggested that Lr-FolQ3 is Mn²⁺-dependent. In previous studies of Ll-FolQ1 and Ec-FolQ2, it has been reported that DHNTP is spontaneously degraded in the presence of Mn²⁺ ions and the effects of the metal ion on the enzyme activities were not tested (24, 25). Although we also confirmed DHNTP degradation in the presence of Mn²⁺, no DHNMP was detected after incubation in the reaction mixture including 1 mM Mn²⁺ ions for 1 min, 2 min, 3 min, 1 h, or 5 h. In contrast, under the conditions with the Lr-FolQ3 enzyme, DHNMP could be detected, indicating that the enzyme contributes to significantly accelerate the generation of DHNMP from DHNTP. The Lr-FolQ3 protein has been annotated as Ham1 family protein in GenBank and harbors the Ham1p-like domain. A previous study reported that an Mj0226 protein from Methanocaldococcus jannaschii harbors the Ham1p-like domain and could utilize Mn²⁺ as well as Mg²⁺ to remove pyrophosphate from non-standard nucleotides such as XTP and ITP (28). Therefore, the domain may be related to Mn²⁺-dependent phosphohydrolase activity. Although, to our knowledge, the precise intracellular Mn²⁺ concentration in L. reuteri cells is unknown, it was reported to be several hundreds of μM in E. coli cells (33). On the L. reuteri genome, we could find genes predicted to be involved in manganese transport. Lreu_0216/Lreu_0217/Lreu_0218 are annotated as zinc/manganese transport system substrate-binding protein/ATP-binding protein/permease protein, respectively, in KEGG database while both Lreu_1555 and Lreu_1861 genes are annotated as manganese transport protein in KEGG database and Mn²⁺/Fe²⁺ transporter, NRAMP family, in GenBank. Although their functions seem not to be examined experimentally, there is the possibility that their gene products facilitate uptake of exogenous manganese ions. These imply that L. reuteri cells may harbor sufficient concentrations of Mn²⁺ to be utilized by Lr-FolQ3 in vivo.

As tetrahydrofolate is involved in biosynthesis of key molecules such as thymidine, purines, methionine, serine, and glycine, enzymes for folate biosynthesis are considered druggable targets to inhibit bacterial proliferation (16, 34). A prodrug, para-aminosalicylic acid, whose metabolite inhibits dihydrofolate reductase (DHFR; FolA), is known to suppress growth of Mycobacterium tuberculosis in middlebrook 7H9 or 7H10 synthetic media (35, 36), although this compound also turned out to inhibit FolP and flavin-dependent thymidine synthase (37). A number of other inhibitors for FolA have been examined as anti-tuberculosis drugs (37–40). Whether other enzymes of the folate biosynthesis pathway can be efficiently targeted with inhibitors will require further analysis. Although E. coli mutant strains, in which folate biosynthesis is shut down by gene disruption of folE, folB, folK, or folP gene, display auxotrophy toward compounds such as thymidine, adenine, and methionine (21, 41–43), there have been reports that disruption of DHNTP pyrophosphohydrolase genes do not necessarily display clear phenotypes in E. coli and L. lactis (24, 25). Confirming the absence of functionally redundant proteins encoding compensatory functions will be a prerequisite for utilizing FolQ and other folate biosynthesis proteins as targets of inhibition. The identification of Lr-FolQ3 and the distribution of the three DHNTP pyrophosphohydrolases suggest that the enzymes catalyzing this reaction are diverse in bacteria. This unusual trait may provide us the opportunity to develop a compound inhibiting the enzyme and suppressing the growth of specific bacteria to regulate the composition of intestinal microbiota. If a certain group of harmful bacteria harbored a common type of DHNTP pyrophosphohydrolase and inhibition of the enzyme led to growth suppression, this would provide a straightforward means to suppress the growth of these bacteria in the intestine.

MATERIALS AND METHODS

Chemicals, strains, and media

Unless mentioned otherwise, chemical reagents were purchased from Fujifilm Wako Pure Chemicals (Osaka, Japan), Nacalai Tesque (Kyoto, Japan), or Sigma-Aldrich (St. Louis, MO). Primers used in this study are listed in Table S3. E. coli DH5α (TaKaRa, Ohtsu, Japan) used for plasmid construction and E. coli BL21-CodonPlus(DE3)-RIL (Agilent Technologies, Santa Clara, CA) used for gene expression were cultured at 37°C in lysogeny broth medium supplemented with ampicillin (100 mg liter⁻¹) (LB medium). Lysogeny broth medium was composed of 10 g liter⁻¹ sodium chloride, 10 g liter⁻¹ tryptone, and 5 g liter⁻¹ yeast extract. When necessary, 40 mg liter⁻¹ 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) was supplemented. To prepare solid medium, 15 g liter⁻¹ agar was added to the medium. All media were sterilized by autoclave at 121°C for 20 min before use. After cooling, ampicillin and X-gal were added.

Construction of plasmids for gene expression of Ec-folE and lreu_1276

The genes encoding Ec-FolE and Lreu_1276 proteins were designed to change rare codons for E. coli to codons used frequently in E. coli in order to enhance gene expression efficiency (Fig. S3). NdeI and BamHI restriction sites were added at 5’- and 3’-termini, respectively. In addition, gene sequence encoding His tag was inserted to prepare N-terminal His-tag fusion proteins. The designed sequences shown in Fig. S3 were chemically synthesized and inserted into pUCIDT-AMP plasmid by Integrated DNA Technologies (Coralville, IA). The synthesized genes were excised from the plasmids with NdeI and BamHI and inserted into the same restriction site in pET21a(+) vector (Merck, Darmstadt, Germany) individually. The resulting plasmids including Ec-folE and lreu_1276 genes were designated pET-His-Ec-FolE and pET-His-Lreu_1276, respectively. By DNA sequencing using primers listed in Table S3 (T7P modified/T7T modified/Ec3110seq-F1/-F2/-F3/-F4/-R1/-R2/-R3/-R4 for pET-His-Ec-FolE and T7P modified/T7T modified/Lreu1276seq-F1/-F2/-F3/-R1/-R2/-R3 for pET-His-Lreu_1276), we confirmed the absence of no unintended mutation in coding regions.

Overexpression of Ec-folE and lreu_1276 genes in E. coli

Expression plasmids were introduced into E. coli strain BL21 CodonPlus (DE3)-RIL, and then, the transformed cells were inoculated onto LB plates. Colonies grown on the plates were picked up and subcultured into 5 mL of liquid LB medium and cultivated overnight. Two mL of the cultures were inoculated into 200 mL of liquid LB medium and incubated at 37°C. After their optical density at 660 nm reached 0.4–0.8, gene expression was induced by adding 200 µL of 100 mM IPTG (final concentration 0.1 mM). For Ec-FolE gene expression, cells were incubated at 37°C for 4 h, whereas for Lreu_1276 gene expression, cells were incubated at 16°C for 20 h. Then cells were harvested by centrifugation (4°C, 5,000 × g, 15 min), washed with 50 mM Tris-HCl (pH 7.5) including 150 mM NaCl, and harvested again under the same condition described above.

Purification of recombinant proteins

Cells overexpressing genes encoding Ec-FolE and Lreu_1276, collected from 125 mL of medium, were resuspended with 10 mL of buffer A containing 20 mM Tris-HCl (pH 7.4), 500 mM NaCl, and 20 mM imidazole and disrupted by sonication. The crude extract was centrifuged (4°C, 15, 000 × g, 15 min). The target protein in the supernatant was purified with a nickel affinity column, His GraviTrap (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom), which was equilibrated with buffer A. After the supernatant was applied to the column, the column was washed with 10 mL of buffer A. To elute the protein, 3 mL of buffer B containing 20 mM Tris-HCl (pH 7.4), 500 mM NaCl, and 500 mM imidazole was added to the column. The protein solution was further purified with a gel filtration column, Superdex 200 Increase (GE Healthcare). A Superdex 200 Increase column was equilibrated with 50 mM Tris-HCl (pH 7.5) including 150 mM NaCl. The same buffer was used as a mobile phase. Protein elution was monitored by ultraviolet (UV) absorbance (280 nm). In order to determine which fractions include the target protein, fractions were analyzed by SDS-PAGE. Protein concentration was determined with the Protein Assay System (Bio-Rad, Hercules, CA), using bovine serum albumin (Thermo Fisher Scientific, Waltham, MA) as a standard.

Preparation of DHNTP

The substrate for DHNTP pyrophosphohydrolase, DHNTP, was prepared from GTP through the enzymatic reaction catalyzed by the Ec-FolE recombinant protein as described previously (44) with some modifications. The reaction mixture (100 µL) included 50 mM Bicine-NaOH (pH 9.0), 0.5 mM GTP, 100 mM KCl, and 15 µg of the purified Ec-FolE recombinant enzyme. After incubation at 37°C for 3 h, the reaction was stopped by rapid cooling on ice for 5 min. The enzyme was then removed by ultrafiltration with Amicon Ultra-0.5 Centrifugal Filter Unit (MWCO: 10 kDa) (EMD Millipore, Billerica, MA). The reaction product was analyzed by HPLC using a COSMOSIL 5C18-PAQ column (4.6 I. D. × 250 mm). The column was equilibrated with 10 mM sodium phosphate buffer (pH 6.0). Compounds in the reaction mixture were separated with the column and monitored by UV absorbance (330 nm) with an SPD-20A detector (Shimadzu, Kyoto, Japan). The concentration of DHNTP was calculated utilizing the molar extinction coefficient of DHNTP, which is 6,300 M⁻¹ cm⁻¹ at 330 nm (45). Aliquots containing DHNTP were frozen in liquid N₂ and stored at −80°C. After melting, aliquots were used as a substrate.

LC-MS analysis

LC-MS analysis to examine the Ec-FolE reaction product was performed as follows. The reaction mixture included 50 mM Bicine-NaOH (pH 9.0), 0.5 mM GTP, and 60 µg of the purified Ec-FolE recombinant enzyme which was desalted by ultrafiltration using 50 mM Tris-HCl (pH 7.5) and Amicon Ultra-0.5 Centrifugal Filter Unit (MWCO: 10 kDa). The desalting process of Ec-FolE is started by applying 200 µL each of Ec-FolE solution and Tris-HCl (pH 7.5) to the ultrafiltration unit. By centrifugation (7,000 × g for 6 min), the volume of the solution was reduced to approximately 200 µL. Then, further 200 µL of Tris-HCl (pH 7.5) was added to the unit again, and centrifugation was performed. These procedures were repeated totally eight times. As a result, the concentration of NaCl in the enzyme solution of Ec-FolE was reduced to approximately 590 µM from 150 mM. After incubation of enzyme reaction mixture at 37°C for 3 h, the reaction was stopped by rapid cooling on ice for 5 min. Then, the enzyme was removed by ultrafiltration with Amicon Ultra-0.5 Centrifugal Filter Unit (MWCO:10 kDa). The filtrate, reaction product, was separated by LC using COSMOSIL 5C18-PAQ (4.6 I. D. × 250 mm) as a column. The column was equilibrated with 10 mM ammonium bicarbonate-formic acid (pH 7.0). The same buffer was used as a mobile phase. Mass of separated compounds was analyzed using Fourier-transform orbitrap mass spectrometer, Exactive Plus (Thermo Fisher Scientific), with an electrospray ionization (ESI) source in negative ion mode. The parameters were set as follows: Aux gas heater temperature at 350°C, a capillary temperature at 350°C, a spray voltage at 2.0 kV, and a mass range between m/z 105–1500.

DHNTP pyrophosphohydrolase activity assay of the Lreu_1276 recombinant protein

DHNTP pyrophosphohydrolase activity of the Lreu_1276 protein was measured with HPLC as described in previous report (24) with some modifications. The reaction mixture (100 µL) included 20 mM Tris-HCl (pH 7.4), 50 µL of Ec-FolE reaction product (DHNTP final conc.: 150 µM), 1 mM MgCl₂, and 4 µg of the purified Lreu_1276 recombinant protein. After pre-incubation at 37°C for 3 min without the substrate, the reaction was started by adding the substrate. After incubation for 15 min, the reaction was stopped by rapid cooling on ice for 20 min. Then, the enzyme was removed by ultrafiltration as described above. The reaction products were analyzed by HPLC as described in the analysis of Ec-FolE reaction product. The compounds in the reaction mixture including DHNTP and DHNMP were separated with the column and monitored by UV absorbance (330 nm) with an SPD-20A detector. DHNMP, purchased from Schircks Laboratories (Bauma, Switzerland), was utilized as a standard compound and to construct a standard curve.

LC-MS analysis to examine the reaction product of the Lreu_1276 protein was also carried out as described for LC-MS analysis of Ec-FolE reaction product. The reaction mixture (100 µL) included 100 mM HEPES-NaOH (pH 7.0), 50 µL Ec-FolE reaction product, 0.1 mM MnCl₂, and 10 µg of the purified Lreu_1276 recombinant protein. Ec-FolE reaction mixture (100 µL) included 50 mM Bicine-NaOH (pH 9.0), 0.5 mM GTP, and 120 µg of the purified Ec-FolE recombinant enzyme, which was desalted by ultrafiltration using Amicon Ultra-0.5 Centrifugal Filter Unit (MWCO: 10 kDa) and 50 mM Bicine-NaOH (pH 9.0). DHNTP pyrophosphohydrolase reaction catalyzed by Lreu_1276 recombinant protein, which was desalted with the method described above, was carried out as described above. After incubation at 37°C for 10 min, the reaction was stopped by rapid cooling on ice for 5 min. Then, the enzyme was removed by ultrafiltration. The reaction products were analyzed with LC-MS.

Metal ion dependency of the DHNTP pyrophosphohydrolase reaction catalyzed by the Lreu_1276 recombinant protein

Metal ion specificity of the Lreu_1276 protein was examined as follows. The reaction mixture (100 µL) included 100 mM Tris-HCl (pH 7.4), 50 µL of Ec-FolE reaction mixture (DHNTP final conc.: 100–150 µM), 1 mM each MnCl₂, MgCl₂, ZnCl₂, CoCl₂, CaCl₂, NiCl₂, CuCl₂, or no divalent metal ion, and 0.05–4 µg of the purified Lreu_1276 recombinant protein. After pre-incubation at 37°C for 3 min without the substrate, the reaction was started by adding the substrate and carried out for 1–3 min. After incubation, the reaction was stopped by rapid cooling on ice for 5 min and removing enzyme from reaction mixture by ultrafiltration. The reaction products were analyzed by HPLC as described above.

Specific activities in the presence of various Mn²⁺ concentrations were determined as described above. The reaction mixture (100 µL) included 100 mM Tris-HCl (pH 7.4), 50 µL of Ec-FolE reaction mixture (DHNTP final conc.: 100 µM), 0.5, 1.0, 1.5, 2.0, 4.0, or 8.0 mM MnCl₂, and 0.01–0.1 µg of the purified Lreu_1276 recombinant protein. Reactions were performed for 1–3 min. Following procedures are the same as those described above. As keeping reaction mixtures including more than 1 mM Mn²⁺ led to the spontaneous degradation of DHNTP and DHNMP, quantification of the reaction product by HPLC was started immediately after the reaction.

Reaction temperature and pH dependencies of the DHNTP pyrophosphohydrolase reaction catalyzed by the Lreu_1276 recombinant protein

In the examination of reaction temperature effects on activity, the reaction mixture (100 µL) included 100 mM Tris-HCl (pH 7.4), 50 µL of Ec-FolE reaction mixture (DHNTP final conc.: 100 µM), 1 mM MnCl₂, and 0.05 µg of the purified Lreu_1276 recombinant protein. Reactions were performed at various temperature conditions and for 1–15 min. For the examination of pH effects on activity, the reaction mixture (100 µL) included 100 mM each MES-NaOH (pH 5.5, 6.0, 6.5, and 7.0) and HEPES-NaOH (pH 7.0, 7.5, and 8.0), 50 µL Ec-FolE reaction mixture (DHNTP final conc.: 100 µM), 1 mM MnCl₂ and 0.05 µg of the purified Lreu_1276 recombinant protein. Reactions were carried out for 1–3 min. The following procedures to quantify DHNMP were the same as those described above.

Substrate specificity of the Lreu_1276 protein

In an examination of substrate specificity, the reaction mixture (100 µL) included 100 mM HEPES-NaOH (pH 7.0), each 100 µM substrate (except ADP-ribose), 1 mM MnCl₂, and 0.05 µg of the purified Lreu_1276 recombinant protein. As a substrate, in addition to 50 µL Ec-FolE reaction mixture (DHNTP final conc.: 100 µM), GTP, ITP, XTP, ATP, CTP, UTP, dTTP, dATP, dCTP, dGTP, NAD⁺, NADH, diadenosine triphosphate, diadenosine tetraphosphate, ADP-ribose, and UDP-glucose were tested. ITP was purchased from Tokyo Chemical Industry (Tokyo, Japan), whereas adenosine 5’-triphosphate and adenosine 5’-tetraphosphate were purchased from Jena Bioscience (Jena, Germany). ADP-ribose was prepared by heat treatment of 100 µM NAD⁺ for 30 min at 85°C (46). The phosphohydrolase reaction was performed for 20 min. DHNTP and DHNMP were monitored by UV absorbance (330 nm) and other substrates and products were monitored by UV absorbance (254 nm). For the determination of phosphohydrolase activity toward XTP, reactions were carried out for 10–30 min, and the reaction mixture without protein was set as a blank control.

Quantification of pyrophosphate and phosphate concentrations

To examine stoichiometry between substrates and reaction products, DHNTP pyrophosphohydrolase reaction was carried out as follows. Reaction mixture (100 µL) included 100 mM Tris-HCl (pH 7.4), 50 µL of Ec-FolE reaction mixture (DHNTP final conc.: 100 µM), 0.1 mM MnCl₂, 10 µg of the purified Lreu_1276 recombinant protein. After pre-incubation at 37°C for 3 min, the reaction was started by adding DHNTP and carried out for 10 min. The following procedures and DHNMP quantification were performed as described above.

Free pyrophosphate was quantified utilizing coupling enzymes UDP-glucose pyrophosphorylase (UGPase), phosphoglucomutase (PGM), and glucose-6-phosphate dehydrogenase (G6PDH) producing NADPH from NADP⁺. Coupling reaction mixture (100 µL) was composed of 100 mM Tris-HCl (pH 7.4), 50 µL of Lreu_1276 reaction mixture, 0.2 mM UDP-glucose, 0.4 mM NADP⁺, 0.05 mM glucose 1,6-bisphosphate, 3 mM MgCl₂, 40 mU UGPase, 60 mU PGM, and 0.6 mU G6PDH. After measuring A₃₄₀ of the reaction mixture without UGPase, reaction was started by adding 4 µL of UGPase. After incubation at room temperature for 40 min, A₃₄₀ was measured again. Based on the difference of A₃₄₀ before and after reaction, concentration of produced NADPH, equal to pyrophosphate, was quantified utilizing the molar extinction coefficient of NADPH (6,220 M⁻¹ cm⁻¹).

Free phosphate was quantified using BIOMOL Green Reagent (Enzo Life Sciences, Farmingdale, NY). After removing protein by ultrafiltration as described above, 50 µL of reaction product was mixed with 100 µL of BIOMOL Green Reagent and incubated at room temperature for 22–30 min. A₆₂₀ was then measured by a microplate reader, Multiskan SkyHigh Microplate Spectrophotometer (Thermo Fisher Scientific). Calibration curve was constructed utilizing solutions of known concentrations of phosphate. The reaction mixture without protein was set as a blank control.

Structural analysis

The three-dimensional structure models of Lr-FolQ3 and Ll-FolQ1 predicted by the AlphaFold2 (29) and deposited in AlphaFold Protein Structure Database (https://www.alphafold.ebi.ac.uk/) (AF-A5VL09-F1-model_v4 and AF-P0CI35-F1-model_v4, respectively) were utilized in this study. The superimposition diagrams of the three-dimensional structures were generated by PyMOL 2.5.0 with a cutoff value of 2.0.

Construction of phylogenetic tree

Sequences of 16S ribosomal RNA genes from 55 bacterial species were acquired from the Kyoto Encyclopedia of Genes and Genomes (KEGG) GENES database. When there were multiple 16S ribosomal RNAs in one organism, sequence displaying the highest similarity with rrsA gene from E. coli K-12 MG1655 (b3851) was selected. Phylogenetic analysis was carried out using the Multiple Sequence Alignment by CLUSTALW (https://www.genome.jp/tools-bin/clustalw), and the phylogenetic tree was constructed utilizing the TreeViewX Version 0.5.0.

DATA AVAILABILITY

The NCBI protein ID of Lreu_1276 protein is ABQ83533. Other relevant data are described in the paper.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.00814-24.

Supplemental figures. aem.00814-24-s0001.pdf.

Fig. S1 to S5.

Table S1. aem.00814-24-s0002.xlsx.

Distribution of genes involved in dihydrofolate biosynthesis.

Table S2. aem.00814-24-s0003.xlsx.

The number of organisms harboring three DHNTP pyrophosphohydrolase homologs.

Table S3. aem.00814-24-s0004.xlsx.

Primers used in this study.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.