Biochemical and structural characterization of meningococcal methylenetetrahydrofolate reductase

Abstract

Methylenetetrahydrofolate reductase (MTHFR) is a key metabolic enzyme in colonization and virulence of Neisseria meningitidis, a causative agent of meningococcal diseases. Here, the biochemical and structural properties of MTHFR from a virulent strain of N. meningitidis serogroup B (NmMTHFR) were characterized. Unlike other orthologs, NmMTHFR functions as a unique homohexamer, composed of three homo‐dimerization partners, as shown in our 2.7 Å resolution crystal structure. Six active sites were formed solely within monomers and located away from the oligomerization interfaces. Flavin adenine dinucleotide cofactor formed hydrogen bonds with conserved sidechains, positioning its isoalloxazine ring adjacent to the overlapping binding sites of nicotinamide adenine dinucleotide (NADH) coenzyme and CH₂‐H₄folate substrate. NmMTHFR utilized NADH (K _m = 44 μM) as an electron donor in the NAD(P)H‐CH₂‐H₄folate oxidoreductase assay, but not nicotinamide adenine dinucleotide phosphate (NADPH) which is the donor required in human MTHFR. In silico analysis and mutagenesis studies highlighted the significant difference in orientation of helix α7A (Phe215–Thr225) with that in the human enzyme. The extended sidechain of Met221 on helix α7A plays a role in stabilizing the folded structure of NADH in the hydrophobic box. This supports the NADH specificity by restricting the phosphate group of NADPH that causes steric clashes with Glu26. The movement of Met221 sidechain allows the CH₂‐H₄folate substrate to bind. The unique topology of its NADH and CH₂‐H₄folate binding pockets makes NmMTHFR a promising drug target for the development of new antimicrobial agents that may possess reduced off‐target side effects.

Short abstract

PDB Code(s): 7RML;

1. INTRODUCTION

Invasive meningococcal disease (IMD) is an active stage of the diseases caused by Neisseria meningitidis, which results in high rates of mortality due to septicemia and morbidity from arthritis and meningitis—persistent neurological defects—particularly among infants and young children (Olbrich et al., 2018). The World Health Organization has given significant attention to the prevalence of IMD by introducing the “Defeating meningitis by 2030” campaign in 2015 (World Health Organization, 2018). Although the treatment of IMD patients by the parenteral administration of ß‐lactams and quinolones is effective in general, incidences of decreasing drug susceptibility and the failure of low‐dose treatment regimens have been continuously reported (Antignac et al., 2003; Jorgensen et al., ²⁰⁰⁵; Rainbow et al., ²⁰⁰⁵; Zerouali et al., ²⁰⁰²). The discovery of ß‐lactamase in the causative pathogen and an increase in infections with either penicillin‐, ciprofloxacin‐, or rifampin‐resistant strains of N. meningitidis, have raised concerns of antibiotics resistance worldwide (Jorgensen et al., 2005; Oppenheim, ¹⁹⁹⁷; Rainbow et al., ²⁰⁰⁵; Wu et al., ²⁰⁰⁹; Zerouali et al., ²⁰⁰²). A new drug class is likely to be imminently required. Identification of new drug targets is therefore a strategy that should be explored to combat the existing and emerging drug resistance.

5,10‐Methylenetetrahydrofolate reductase (MTHFR; EC1.5.1.20) is one of the metabolic signatures of nasopharyngeal colonization and invasiveness of N. meningitidis (Grifantini et al., 2002; Schoen et al., ²⁰¹⁴). The expression of the MTHFR‐encoding gene, metF, during colonization in nasopharyngeal cells is believed to be an essential initial step for bacteremia in an infant rat model (Sun et al., 2000). Transcriptomic analysis also showed increased expression of metF upon incubation of bacterial cells with human blood (Echenique‐Rivera et al., 2011; Hedman et al., ²⁰¹²). These expression profiles suggested that N. meningitidis possesses a dependence on MTHFR activity to supply a methyl group from 5‐methyltetrahydrofolate in the final step of de novo methionine biosynthesis. Genes encoding other enzymes in the methionine pathway were also identified to have significant roles in pathogenesis (Schoen et al., 2014). The final product, methionine, is not only an important building block for protein synthesis, but also the precursor of S‐adenosylmethionine (SAM or AdoMet), which serves as the universal methyl donor in cells. NmMTHFR therefore has the potential to be an interesting drug target for the development of a new antimicrobial drug class for the treatment of drug resistant N. meningitidis infections.

While the biochemical and structural properties of NmMTHFR have until now remained undetermined, studies of MTHFR from various other organisms have been performed. Bacterial MTHFR is comprised of a catalytic domain (~33 kDa) in a homo‐dimer, ‐tetramer or ‐octamer and utilizes nicotinamide adenine dinucleotide (NADH) as the electron donor in the reaction (Guenther et al., 1999; Igari et al., ²⁰¹¹; Lee et al., ²⁰⁰⁹; Pejchal et al., ²⁰⁰⁵; Pejchal et al., ²⁰⁰⁶). In contrast, mammalian, yeast, and plant MTHFRs form a homodimer of 66–70 kDa monomers, each containing an N‐terminal catalytic domain linked to a C‐terminal regulatory domain (Chan & Appling, 2003; Daubner & Matthews, ¹⁹⁸²; Froese et al., ²⁰¹⁸; Roje et al., ¹⁹⁹⁹). The mammalian MTHFR is allosterically inhibited by SAM, and nicotinamide adenine dinucleotide phosphate (NADPH) is the preferred electron donor. Despite these differences, all MTHFR enzymes catalyze the conversion of 5,10‐methylenetetrahydrofolate (CH₂‐H₄folate) to 5‐methyltetrahydrofolate (CH₃‐H₄folate) in the folate cycle. An enzyme‐bound Flavin adenine dinucleotide (FAD) cofactor functions as an intermediate electron acceptor and donor that transfers electrons from NAD(P)H to CH₂‐H₄folate via a ping‐pong Bi‐Bi mechanism (Trimmer et al., 2001), as depicted in Scheme S1. In the first half‐reaction, NAD(P)H binds against the si face of FAD, in the adjacent pocket, resulting in its 4S‐hydrogen transferred as a hydride ion to the N5 of FAD and the concomitant release of NAD(P)⁺. When CH₂‐H₄folate subsequently binds to the former NAD(P)H pocket, the reduced FAD then donates electrons to the C11 of CH₂‐H₄folate, leading to the production of CH₃‐H₄folate and the regeneration of oxidized FAD in the second half‐reaction (Igari et al., 2011; Pejchal et al., ²⁰⁰⁵; Sheppard et al., ¹⁹⁹⁹; Trimmer et al., ²⁰⁰¹).

This study represents the first biochemical and structural characterization of NmMTHFR. Unusually, size exclusion chromatography and mass spectrometry revealed that the native form of NmMTHFR is homohexamer. The precise arrangement of this hexameric complex was uncovered by determination of the x‐ray crystal structure of NmMTHFR in complex with the FAD cofactor. The preference for NADH was also confirmed using kinetic and mutagenesis studies. Together with in silico analysis, the studies highlighted an important role for Met221 on helix α7A (Phe215–Thr225) in the binding of both conformationally folded NADH and the CH₂‐H₄folate substrate, which makes the NADH and CH₂‐H₄folate binding pockets of NmMTHFR different from those in human MTHFR (hMTHFR). This therefore offers a distinct and potentially druggable pocket on NmMTHFR that can be used for future development of anti‐meningococcal agents.

2. RESULTS AND DISCUSSION

2.1. NmMTHFR co‐purifies with FAD

The full‐length metF gene from the virulent strain of N. meningitidis serogroup B (strain MC58) was amplified by PCR using gene‐specific primers designed to add a C‐terminal 6xHis‐tag for purification purposes. The recombinant NmMTHFR protein (298 residues) was overexpressed as a soluble protein in the Escherichia coli BL21(DE3) strain and was purified as a yellow homogeneous solution in a single step using nickel affinity chromatography (Figure S1; Lane 1). A protein band at ~30 kDa, consistent with the calculated monomeric mass of NmMTHFR (33 kDa), was observed on a 10% SDS‐PAGE gel. Peptides of tryptic digested purified wildtype protein identified using Orbitrap mass analyzer were matched to 5,10‐methylenetetrahydrofolate reductase from N. meningitidis serogroup B (strain MC58), UniProtKB Accession No. Q9JZQ3, with 90% sequence coverage (Table S1, Figure S2A). Both Met221Ala and Met221Trp mutants were expressed and purified in the same manner as the wildtype protein (Figure S1; Lane 2–3). The alanine and tryptophan substitutions were confirmed in the tryptic digested peptides of the mutants, with the respective M/A/W‐AQVTNVK peptides observed in each case (Table S1, Figures S2B and S2C). An FAD absorbance peak at about 447 nm was identified in the UV–Vis spectrum for all purified proteins (Figure S3), as observed previously for other flavoproteins with embedded oxidized FAD, including the FAD‐bound E. coli MTHFR (EcMTHFR) (Macheroux, 1999; Sheppard et al., ¹⁹⁹⁹; Trimmer et al., ²⁰⁰¹). This demonstrated that NmMTHFR co‐purifies with FAD following cellular expression, consistent with the yellow color of the protein solution. The similar spectral properties of the wildtype NmMTHFR and mutants and no remarkable difference in the magnitudes of absorbance peak at 447 nm, for equivalent concentrations of purified proteins (1 mg/mL), suggested that the substitutions did not perturb the protein structure significantly.

2.2. NmMTHFR prefers NADH over NADPH

The activity of MTHFR generally requires NAD(P)H as an electron donor and CH₂‐H₄folate substrate as an electron accepter in the two‐half reactions. According to the kinetic traces at 375 nm obtained from NmMTHFR assay based on NAD(P)H‐CH₂‐H₄folate oxidoreductase assay, NmMTHFR had a clear preference for NADH over NADPH as shown K _m for NADH = 44 ± 9 μM, but no detectable activity in the reaction containing NADPH at 100 μM (Figure S4, Table 1). This preference is the same as EcMTHFR (Trimmer et al., 2001) but with about 26‐fold difference in catalytic efficiency, represented by the k _cat/K _m values (0.02 for NmMTHFR [Table 1, Figure S5A] and 0.52 μM⁻¹ s⁻¹ for EcMTHFR [Trimmer et al., ²⁰⁰¹]). This however contrasted to that of hMTHFR, which preferably utilizes NADPH (K _m for NADPH = 35.5 ± 2.4 μM, K _m for NADH = 3760 ± 410 μM) (Froese et al., 2018).

TABLE 1.

Kinetic parameters for NADH.

Kinetic parameters	NmMTHFR
	Wildtype c	Met221Trp c	Met221Ala d
k cat (s−1)
CH2‐H4folate a	0.75 ± 0.08	0.55 ± 0.08	0.39 ± 0.06
NADH b	0.72 ± 0.05	0.55 ± 0.06	0.40 ± 0.03
K m (μM)
CH2‐H4folate a	83 ± 25	58 ± 26	184 ± 55
NADH b	44 ± 9	50 ± 16	99 ± 17
k cat/K m (×103 M−1 s−1)
CH2‐H4folate a	9.0 ± 2.9	9.5 ± 4.5	2.1 ± 0.7
NADH b	16.4 ± 3.5	11.0 ± 3.7	4.0 ± 0.8

Note: The kinetic parameters were determined in NmMTHFR assay based on NAD(P)H‐CH₂‐H₄folate oxidoreductase assay. The reactions were performed in phosphate buffer pH 7.2 and 0.3 mM EDTA in anaerobic condition at room temperature. The value was calculated from three independent experiments and reported as mean ± SD.

^a The kinetic parameters were determined at varying concentrations of CH₂‐H₄folate (10–250 μM) and a fixed concentration of 200 μM NADH.

^b The kinetic parameters were determined at varying concentrations of NADH (10–200 μM) and a fixed concentration of 200 μM CH₂‐H₄folate.

^c The total enzyme concentration used in the reaction was 1 μM.

^d The total enzyme concentration used in the reaction was 2 μM.

2.3. Analytical‐SEC and native MS indicate NmMTHFR exists as a hexamer

A combination of analytical‐SEC and native mass spectrometry was used to determine the oligomeric state of NmMTHFR. Analytical‐SEC of purified wildtype NmMTHFR showed a single, broad peak eluting at 12.2 mL (~177 kDa) across a range of concentrations (from 1 mg/mL down to 0.1 mg/mL), consistent with either a pentameric or hexameric oligomeric state (calculated molecular mass of a pentamer and a hexamer is 165 and 198 kDa, respectively) (Figure 1a). Native MS, an extremely powerful analytical tool for the accurate determination of protein oligomeric states (Benesch & Ruotolo, 2011; Jovcevski et al., ²⁰¹⁵; Wang et al., ²⁰²⁰), subsequently showed that NmMTHFR exclusively adopts a hexameric quaternary structure (~203 kDa, which is closer to the calculated 198 kDa) (Figure 1b). This result contrasts with other MTHFR orthologs, reported as a dimer (from Homo sapiens [Froese et al., ²⁰¹⁸], Sus domesticus [Daubner & Matthews, ¹⁹⁸²], Saccharomyces cerevisiae [Chan & Appling, ²⁰⁰³], Arabidopsis thaliana [Roje et al., ¹⁹⁹⁹], and Thermus thermophilus [Igari et al., ²⁰¹¹]), a tetramer (from E. coli) (Pejchal et al., 2005), or an octamer (from Peptosteptococcus productus) (Wohlfarth et al., 1990).

FIGURE 1.

Analytical‐SEC and native MS reveal NmMTHFR adopts a hexameric quaternary state. (a) Analytical‐SEC of NmMTHFR in 50 mM phosphate buffer (pH 7.4) at 0.1 mg/mL (purple), 0.5 mg/mL (green) and 1 mg/mL (black), demonstrating the oligomeric state of NmMTHFR is preserved nearing physiological concentrations. Elution volumes of SEC mass standards are indicated above. (b) Native MS shows that NmMTHFR (0.5 mg/mL) is hexameric (green). Dominant charge states that correspond to the hexamer are noted above.

2.4. NmMTHFR tertiary structure resembles that of known orthologs, while quaternary structure differs

The structure of NmMTHFR was determined by x‐ray crystallography at 2.7 Å resolution. Crystallographic details are provided in the Materials and Methods, and statistics in Table S2. Three NmMTHFR monomers (chain A, B, and C) crystallized in the asymmetric unit with one FAD bound per monomer (Figure S6). As in other orthologs, the FAD binding pockets of NmMTHFR is located across the conserved (β/α)₈ TIM barrel (Figures 2a and S7). Of the monomers, chain A had the best electron density map with the most complete polypeptide, missing only the initiating methionine and parts of two loop regions: residues 61–65 in loop 4 and 121–127 in loop 8 (Figure 2a). To be noted, the protein at 30 mg/mL was used in crystallization, which is much higher than concentrations used in biophysical assays (0.1, 0.5, and 1.0 mg/mL in analytical SEC or 0.5 mg/mL in native MS) and the crystals could not be obtained at those nearing physiological concentrations. Therefore, the final model of NmMTHFR was built by packing of two crystal asymmetric units as a homohexamer according to its verified oligomeric state. The hexamer composed of three homo‐dimerization partners (Figure 2b,c). Since there is no evidence that the dimer of NmMTHFR exists on its own and the hexamer is the physiological state of NmMTHFR, we speculate that the hexamer likely forms for stability rather than to allow catalysis to occur. Similarly, in EcMTHFR, both the dimeric and tetrameric forms process catalytic activity in the presence of FAD (Pejchal et al., 2006; Misra & Bhakuni, ²⁰⁰³). As well as different multimeric arrangements, the symmetry of dimeric partners also varies between orthologs. NmMTHFR shared a similar dimeric symmetry with EcMTHFR (PDBID: 1ZP3), having overall RMSD of chains A at 0.934 Å and the partner chains at 1.858 Å, with the catalytic sites located away from the oligomerization interfaces (Figure S8A). The oligomerization interface was most dissimilar to that of TtMTHFR (PDBID: 3APY) (Igari et al., 2011), leading to overall RMSD of chains A at 1.407 Å and the partner chains at 3.238 Å. In contrast, hMTHFR (PDBID: 6FCX) monomers interacted at the C‐terminal regulatory domains instead of the catalytic domains (Figure S8B) (Froese et al., 2018). By focusing on the catalytic domains of MTHFRs, the structural arrangement of NmMTHFR was the most similar to that of EcMTHFR (PDBID: 1ZP3) than TtMTHFR (PDBID: 3APY) and hMTHFR (PDBID: 6FCX) as shown the RMSD at 0.761, 1.327, and 1.516 Å, respectively (Figure S9).

FIGURE 2.

Structural overview of NmMTHFR. (a) Ribbon drawing of NmMTHFR subunit containing α helices (green), β sheets (purple), and loops (gray). Dotted gray lines indicate disordered regions that are not modeled in the structure. The structure is composed of a conserved (β/α)₈ TIM barrel with the additional N‐terminal αA helix and three fragments of α7 (α7A, B, and C) helices. Bound FAD is shown in yellow stick representation. (b) Dimeric symmetry of NmMTHFR. The dimerization is formed by interactions between helices α7C (purple), helices α7B (orange), and helices α8 and loop 19 (blue). Helices αA (red) and α6 (pink) are involved in hexameric formation. (c) Orthogonal view of hexameric NmMTHFR composed of three homo‐dimerization partners (green, blue, and gold).

2.5. FAD interactions with NmMTHFR resemble those found in other orthologs

FAD was stabilized through hydrogen bonding with both residue backbone atoms (Tyr58, Arg116, Gly117, Asp118, and Ala130) and conserved sidechains (His86, Tyr150, His154, Asp163, and Lys170) of NmMTHFR. The isoalloxazine ring was embedded in the middle of the (β/α)₈ TIM barrel by hydrogen bonding with Tyr58 and His86, with the si face of the ring turned toward the adjacent NADH/CH₂‐H₄folate binding pocket (Figure 3). Most conserved sidechain interactions remained the same as those observed in hMTHFR and EcMTHFR (Figure S9) (Pejchal et al., 2005; Froese et al., ²⁰¹⁸). EcMTHFR made an additional hydrogen bond between the backbone nitrogen of Asn168 and the phosphate of the adenine moiety. A corresponding asparagine residue is also located in NmMTHFR (Asn166); however, this did not form the equivalent interaction. Human MTHFR does not contain this conserved asparagine. Here it is replaced by His213, whose sidechain orients away from the phosphate and π‐π stacked instead with the adenine ring of FAD. Hydrophobic interactions between the FAD adenine ring and a conserved tyrosine sidechain (Tyr129, Tyr131, Tyr174 in NmMTHFR, EcMTHFR, and hMTHFR, respectively) also contribute to FAD binding. Due to having high similarity with the FAD pocket of hMTHFR, this binding pocket in NmMTHFR might not be an ideal target for drug development.

FIGURE 3.

Structural examination of FAD binding in NmMTHFR. (a) Structural alignment of FAD bound‐NmMTHFR (PDBID: 7RML [green] vs. FAD [yellow]), ‐EcMTHFR (PDBID: 1ZP3) (pink), and ‐hMTHFR (PDBID: 6FCX) (gray). Residues contributing to hydrogen bonding with FAD are indicated in black. For clarity, only the superimposition of helices α5 (equivalent to α6 in hMTHFR) is shown to highlight the variant residue, Asn166 (NmMTHFR), Asn168 (EcMTHFR) and His213 (hMTHFR). (b) Structure‐based alignment of amino acid sequences corresponding the helices α5 in NmMTHFR, Ec‐MTHFR and the helix α6 in hMTHFR (AQADLINLKRKID, AQADLLNLKRKVD, and FEADLKHLKEKVSA, respectively). The alignment was performed using Match → Align function in Chimera software. The RMSD of 7RML with 1ZP3 and 6FCX was 0.768 and 1.522, respectively.

2.6. Met221 plays a role in NADH utilization

In silico analysis and mutagenesis studies were performed to search for structural features that are essential for NADH specificity in NmMTHFR. Superimposing the FAD removed structure of NmMTHFR onto the NADH bound EcMTHFR (PDB: 1ZRQ) (Pejchal et al., 2005) revealed the likely binding site of NADH in the hydrophobic box formed by conserved residues Glu26, Phe28, Thr57, Gln181, Tyr273, and Leu275, and non‐conserved residues Met221 and Val224 (Figures 4a, S9 and S10 [for amino acid conservation]). The corresponding residues in EcMTHFR (Glu28, Thr59, Gln183, and Phe223) were previously suggested to play important roles in NADH binding and orientation (Pejchal et al., 2005). Overlaying NADH into the equivalent NmMTHFR hydrophobic box showed that it can be accommodated by being wedged between the isoalloxazine ring of FAD and the sidechain of Met221 to make a 4‐layer sandwich (Figure 4b,c [for structural alignment of amino acids]). A phenylalanine residue (Phe223) plays the equivalent role in EcMTHFR, where the aromatic sidechain of Phe223 supports the folded conformation of NADH by making π‐π interactions with the adenine ring (Lee et al., 2009). Substitution of Phe223 with leucine resulted in a 12‐fold increase in K _m value, suggesting hydrophobic interaction with leucine was not sufficient to replace the π‐π interactions. Meanwhile, the substitution with alanine less likely affect as shown a 7‐fold increase in K _m (Lee et al., 2009). In contrast, the replacement of the methionine residue in NmMTHFR with an aromatic residue, tryptophan, resulted in a similar K _m value (K _m = 50 ± 16 μM). The elimination of the sidechain by alanine substitution caused two‐fold increase in K _m (K _m = 99 ± 17 μM) (Table 1, Figure S5). These suggest that the sidechain at position 221 in NmMTHFR plays a role in NADH utilization and might influence the substrate binding via methionine‐aromatic interaction (Valley et al., 2012) (or can be replaced with π‐π stacking). However, Met221 seems to be less important in NmMTHFR compared to the equivalent position, Phe223, in EcMTHFR. Val224 at the front of the box, and Tyr273 at the back, provided additional hydrophobic supports to stabilize the adenine ring. On the other side of the sandwich layer, the conserved glutamine (Gln181 in NmMTHFR or Gln183 in EcMTHFR) plays a major role in orientating NADH correctly by forming hydrogen bonds from its sidechain to the O7 and N7 of the nicotinamide ring (Figure 4b). Thr59 and Glu28 in EcMTHFR were suggested to participate in the direct or indirect hydrogen bonding to the O3' of adenine ribose and the N3 of the adenine ring, respectively (Pejchal et al., 2005), although there were no predicted equivalent interactions at Thr57 and Glu26 in NmMTHFR (Pejchal et al., 2005). With these hydrogen bonds, the nicotinamide ring was aligned and made π‐π interactions to the isoalloxazine ring of FAD, and its 4S‐hydrogen was presented for the hydride transfer. Based on in silico analysis, Glu26 in NmMTHFR potentially plays another role in the NADH specificity. The O2' of the adenine moiety of the folded NADH was comfortably surrounded by residues Thr57, Phe28, Glu26, Leu275, and Tyr273 in the hydrophobic box. In contrast, the addition of a phosphate group, to become NADPH, could possibly cause electrostatic clashes with the glutamic acid (Figure 4d). Together with the presence of Met221, this provides an explanation for the specificity for NADH in the NmMTHFR reaction. Meanwhile, in the hMTHFR reaction where NADPH is preferred (Froese et al., 2018), the equivalent helix (His263–Ser272) is positioned at a distinct location from that of helix α7A (Phe215–Thr225) in NmMTHFR, resulting in severe steric clashes between the overlayed NADH and Leu271. This suggests that the pocket topology of hMTHFR is different from that of NmMTHFR, and NADPH would bind to the pocket differently to the sandwiched NADH structure in EcMTHFR and NmMTHFR. There is no obvious feature to explain the selectivity for NADPH in hMTHFR as previously discussed (Froese et al., 2018).

FIGURE 4.

In silico investigation of the NADH binding pocket in NmMTHFR. (a) A hydrophobic box on NmMTHFR. Residues (Glu26, Phe28, Thr57, Gln181, Met221, Val224, Tyr273, and Leu275) forming the hydrophobic box or the NADH binding pocket are labeled. The black mesh indicates the predicted binding site of NADH, which is taken from an overlay of EcMTHFR (PDBID: 1ZRQ) with NmMTHFR. The co‐crystallized FAD in NmMTHFR is shown in yellow sticks. (b) Effect of equivalent helix α7A on folded NAD(P)H binding. For clarity, only the superimposition of helices α7A of NmMTHFR (PDBID: 7RML, green) and EcMTHFR (PDBID: 1ZRQ, pink) with the equivalent helix 8 of hMTHFR (PDBID: 6FCX, gray) is shown to indicate the distinct local difference. The overlayed NADH of EcMTHFR is shown in blue sticks. The predicted hydrogen bond between NADH and Gln181 are indicated with black dashed lines. The superimposition of helices α7A from NmMTHFR and EcMTHFR reveals variant residues Met221 and Phe223, are expected to contribute the stacking of the folded NADH. (c) Structure‐based alignment of amino acid sequences corresponding the helices α7A in NmMTHFR, EcMTHFR and the helix α8 in hMTHFR (FKQLGKMAQVT, FKQAKKFADMT and YHSLRQLVKLS, respectively). The alignment was performed using Match → Align function in Chimera software. The RMSD of 7RML with 1ZRQ and 6FCX was 0.806 and 1.500, respectively. (d) The predicted electrostatic clashes between a phosphate group of NADPH (an in silico modified NADH) and the hydrophobic box. Residues contributing to the restriction of the phosphate binding in the hydrophobic box are labeled. The overlayed NADH was modified using Chimera software (UCSF Chimera) by replacing the hydrogen at O2' of adenine ribose with the phosphate group (shown within black dashed circle).

2.7. Met221 and Gln217 play important roles in CH₂‐H₄folate substrate binding

Currently, there is no available structure of CH₂‐H₄folate bound MTHFR. To evaluate the similarities and differences between the pocket topologies of NmMTHFR and orthologs, the EcMTHFR structure in complex with CH₃‐H₄folate product (PDBID: 1ZP4) (Pejchal et al., 2005) was the best option used for studying the binding mode of CH₂‐H₄folate substrate to NmMTHFR. The binding sites of the NADH coenzyme and CH₂‐H₄folate substrate overlap and are located adjacent to the isoalloxazine ring of FAD to facilitate electron transfer via the ping‐pong Bi‐Bi mechanism (Figure 5a). Unlike folded NADH, the folate molecule is in an extended conformation, aligned on the surface exposed pocket formed by Glu26, Asp118, Gln181, Gln217, Met221, Tyr273, Leu275, and Arg277 (Figure S10 [for amino acid conservation]). Upon the folate binding, residues Met221 and Gln217 were predicted to move in an analogous manner to the movement of Phe223 and Gln219 in EcMTHFR. The Met221 sidechain shifts approximately 6 Å, resulting in a parallel alignment to the p‐aminobenzoate (PABA) ring of folate (Figure 5b,c [for structural alignment of amino acids]). Together with Tyr273, Met221 would likely provide hydrophobic interactions for the PABA ring stacking. To accommodate the monoglutamate of folate, the conserved sidechain of Gln217 would move similarly to that of Gln219 in EcMTHFR, to form a hydrogen bond between the sidechain nitrogen and the O2 of the monoglutamate. Meanwhile, Arg277 is predicted to form hydrogen bond with the OE1 of the monoglutamate (Figure 5b). An equivalent interaction with conserved Arg279 was not found for EcMTHFR (Pejchal et al., 2005). The tetrahydropterin ring is predicted to embed in NmMTHFR by hydrogen bonding through conserved Gln181 and Asp118 residues. Although interactions with Asp118 were not observed, according to the criteria used by Chimera software, the orientation of this sidechain was highly similar to that of Asp120 in folate‐bound EcMTHFR. This suggested potential interactions between the carboxyl group of Asp118 and the N2 and O4 of the tetrahydropterin ring (Figure 5b). As a result, the C11 of the tetrahydropterin ring is located close enough to accept electrons transferred from the N5 of FADH₂. To be noted, no interaction occurred at the methyl group of the tetrahydropterin ring of CH₃‐H₄folate; however, this might be similar or different in CH₂‐H₄folate where the methyl is replaced by the methylene. In comparison to hMTHFR (Froese et al., 2018), the main difference was in the binding pose of the O of the PABA and the monoglutamate tail of the overlayed folate, which severely clashed with the sidechain of Gln267 on the equivalent helix (Figure 5d). The C_β of Gln267 sidechain is positioned at a close distance to the O of the PABA and possibly limits the movement of the sidechain to allow the folate to bind. This suggested either the helix in hMTHFR would be orientated differently in the folate bound structure or the CH₂‐H₄folate would bind with slightly different structural arrangement by shifting away the Gln267 sidechain.

FIGURE 5.

In silico investigation of CH₂‐H₄folate binding pocket in NmMTHFR. (a) The predicted binding site of CH₂‐H₄folate substrate (mesh highlighted in both purple and black), which is taken from an overlay of CH₃‐H₄folate product in EcMTHFR (PDBID: 1ZP4) with NmMTHFR. The purple mesh indicates the overlapped binding site of NADH. Residues of NmMTHFR that contribute hydrophobic interactions or hydrogen bonds to the folate are labeled. The co‐crystallized FAD in NmMTHFR is shown in yellow sticks. (b) Binding mode of CH₂‐H₄folate substrate (or CH₃‐H₄folate product). The overlayed CH₃‐H₄folate is shown in purple sticks. Residues involved in hydrogen bonding are indicated in black (NmMTHFR) and pink (EcMTHFR). The predicted hydrogen bonds at Gln181 and Arg277 of NmMTHFR are indicated with black dashed lines while the bonds between EcMTHFR and the substrate are shown as black solid lines. The red dashed arrows represent the predicted sidechain movement of residues; Met221 to Phe223 (5.946 Å) and Gln217 to Gln219 (4.990 Å). (c) Structure‐based alignment of amino acid sequences corresponding the helices α7A in NmMTHFR, EcMTHFR and the helix α8 in hMTHFR (FKQLGKMAQVT, FKQAKKFADMT and YHSLRQLVKLS, respectively). The alignment was performed using Match → Align function in Chimera software. The RMSD of 7RML with 1ZP4 and 6FCX was 0.814 and 1.502, respectively. (d) The predicted electrostatic clashes between Gln267 of hMTHFR and CH₂‐H₄folate substrate (or CH₃‐H₄folate product). The overlayed CH₃‐H₄folate is shown in purple sticks. The helices α7A in NmMTHFR (cyan) and EcMTHFR (pink) were aligned with the helix α8 in hMTHFR (gray).

3. CONCLUSION

Drug targeting and gene knockdown of MTHFR have been proposed as treatments for human cancer (Stankova et al., 2008), but the idea of targeting MTHFR for the development of new antimicrobial agents has not previously been proposed, despite other enzymes from the folate cycle such as dihydrofolate reductase, dihydropteroate synthase, and flavin‐dependent thymidylate synthase receiving much attention as drug targets for both cancer and bacterial infections (Fernandez‐Villa et al., 2019; Mishanina et al., ²⁰¹⁶; Visentin et al., ²⁰¹²). Several drugs targeting these enzymes have been used widely for decades as antitumor and antimicrobial agents which have either a single‐ or multi‐targeted mechanism of action (Hajian et al., 2019; Zheng et al., ²⁰¹³). However, resistance to these drugs (e.g., trimethoprim, methotrexate, pyrimethamine, and trimetrexate) has been shown to arise in a variety of bacterial pathogens including Mycobacterium tuberculosis (Nixon et al., 2014). Therefore, the idea of exploring MTHFR; for example, NmMTHFR, in the folate cycle as a novel antimicrobial drug target due to its importance in meningococcal pathogenesis, would be a promising strategy to overcome the existing drug resistance.

The differences in both structural features and functional roles between NmMTHFR, EcMTHFR, and hMTHFR identified in this study suggest the possibility to develop antimicrobial agents with selectivity for NmMTHFR, in which the dual substrates (NADH and CH₂‐H₄folate) binding pocket is proposed as a druggable site. NmMTHFR possesses unique topology regarding the orientation of helix α7A Phe215–Thr225. Together with Glu26, Met221 on this helix plays a role in accepting the folded NADH, but excluding NADPH, suggesting the opportunity for drug selectivity. A new antimicrobial drug class could be developed using compounds that can specifically react with the sulfur in the methionine sidechain, such as compounds containing an oxaziridine group that can form covalent sulfimide (S=N) linkage with the sulfur (Lin et al., 2017). However, one potential drawback is that the Met221Trp mutant in this study displayed similar activity, suggesting the possibility of the antibiotic resistance for this class of drug through such a mutation. Targeting multiple sites of the enzyme may offer an alternative strategy for the development of a new drug. In addition to Met221, the amide group of conserved Gln217 could be targeted for making specific interactions with a small molecule compound. The conserved glutamine (Gln267) in the equivalent helix from hMTHFR locates in a position that would cause steric clashes with compounds designed to interact with Gln217 in NmMTHFR. Such strategies may allow the development of antimicrobial agents that are specific for NmMTHFR and unable to bind hMTHFR, and as a result less likely to cause off‐target side effects.

4. MATERIALS AND METHODS

4.1. Construction of NmMTHFR wildtype and mutant plasmids

The metF gene encoding NmMTHFR was amplified by PCR using forward primer (5′‐GGTACGCACATTGAATTACGCAAAAGAAATCAATGCG‐3′), reverse primer (5′‐GGAATTCTTAGTGGTGATGGTGATGATGAGGGCGCACGCCTAAAATATGG‐3′), and the genomic DNA obtained from N. meningitidis MC58 serogroup B (ATCC® BAA‐335D‐5™) (American Type Culture Collection) as a template. The forward primer contains an NdeI recognition site while the reverse primer has an EcoRI recognition site and codons for a C‐terminal hexa‐histidine tag. The amplified fragment was ligated into the pGEM‐T Easy vector (Promega) and then subcloned into NdeI and EcoRI treated pET17b expression vector (Novagen). Site‐directed mutagenesis was performed to generate Met221Ala and Met221Trp mutants using the recombinant pGEM‐T Easy plasmid harboring the 6xHis tagged wildtype metF as a template. The Met221Trp mutation was generated using complementary oligonucleotides containing the BanII restriction site (5′‐GCAGCTCGGTAAATGGGCTCAAGTAACC‐3′), while the Met221Ala mutation was created using complementary oligonucleotides containing the HaeIII restriction site (5′‐GCAGCTCGGTAAAGCGGCCCAAGTAACCAAC‐3′). The mutated metF gene was subcloned into the NdeI and EcoRI treated pET17b expression vector (Novagen). All constructs were confirmed by restriction enzyme digestion and DNA sequencing (Humanizing Genomics Macrogen).

4.2. Protein expression and purification

NmMTHFR recombinant proteins were expressed in E. coli BL21(DE3) system (Invitrogen). A single colony was grown overnight in LB medium containing 100 μg/mL ampicillin at 37°C. The culture was diluted 1:100 in LB broth containing 100 μg/mL ampicillin and incubated at 37°C with shaking at 200 rpm until reaching an OD₆₀₀ of 0.5–0.6. Protein expression was induced by adding 0.1 mM IPTG into the culture at 18°C for 18–20 h. Cells were harvested and resuspended in 50 mM phosphate buffer pH 8.0, 300 mM KCl, 5 mM imidazole, 10 μM PMSF. The cell suspension was added with 100 μg/μL lysozyme and incubated overnight at 4°C prior to sonication. Cell debris was removed by centrifugation, followed by filtration using a 0.25 μm membrane filter (GE Healthcare). The 6xHis‐tagged NmMTHFR protein in the supernatant was purified by affinity chromatography using HisTrap™ FF column (Cytiva). The protein was eluted in 50 mM phosphate buffer pH 8.0, 300 mM KCl, 250 mM imidazole and further exchanged into 10 mM Tris–HCl pH 8.0 using cellulose membrane for dialysis. For the study of protein structure, an additional size exclusion chromatography step using a Superdex™ 200 10/300 GL (GE Healthcare) was included to purify the protein together with buffer exchange into the storage buffer (25 mM Tris–HCl buffer pH 8.0, 30 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol (DTT), 5% (v/v) glycerol). The purified proteins were concentrated using centrifugal filter units (Merck Millipore), followed by centrifugation at 14,000 rpm for 15 min to discard protein aggregates. Protein concentration was detected using either a Bradford assay (Bradford, 1976) with bovine serum albumin (Sigma Aldrich®) as a standard or NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) based on absorbance measured at 280 nm. The purity of purified proteins was examined using 10% SDS‐PAGE.

4.3. Analytical size‐exclusion chromatography

The average molecular weight of purified NmMTHFR was determined by analytical SEC. Samples (0.1, 0.5 and 1 mg/mL) were loaded onto a Superdex 200 10/300 GL analytical SEC (GE Healthcare) which was equilibrated in 50 mM phosphate buffer (pH 7.4) at a flow rate of 0.4 mL/min at room temperature. The size‐exclusion column was calibrated using standards (Sigma) containing bovine thyroglobulin (670 kDa), equine ferritin (440 kDa), rabbit aldolase (158 kDa), BSA (66 kDa), horse myoglobin (17 kDa) and bovine ubiquitin (9 kDa).

4.4. Native mass spectrometry

The oligomeric state of NmMTHFR was examined using a Synapt G1 HDMS (Waters) compatible with a nanoelectrospray ionization source. The purified NmMTHFR (1 mg/mL) was buffer exchanged into 200 mM ammonium acetate (pH 6.8) using an Amicon Ultra‐Centrifugal Filter (10,000 MWCO) at 4°C. Protein was diluted to 0.5 mg/mL and loaded (3 μL) into platinum‐coated borosilicate glass capillaries prepared in‐house. Key instrument parameters were as follows: capillary voltage (kV): 1.65; sampling cone (V): 30; extraction cone (V): 1.5; trap/transfer collision energy (V/V): 15/10; trap gas (l/h): 5.5; backing gas (mbar): ~4.5.

4.5. Peptide mass spectrometry

Peptides from protein of interest were prepared using a tryptic digestion of in‐gel protein method (details are provided in supporting information). Approximately 100 ng of digested peptides were submitted to LC–MS/MS analysis using Easy‐nLC 1200 (Thermo Fisher Scientific) coupled with an Orbitrap Fusion Tribrid system (Thermo Fisher Scientific) using the following parameters: For full MS scan, scan range was set from m/z 350–1800 followed by a resolution of 120,000 and a mass tolerance of 10 ppm. Ions with charge states of 2+, 3+, and 4+ were sequentially fragmented by collision induced dissociation with collision energy of 30% to obtain the MS/MS scan data. The raw spectral data were analyzed using Mascot search engine (ver. 2.5.1) (Matrix Science) (Perkins et al., 1999). The acquired MS/MS spectrum were searched against the Swiss‐Prot database (Last update: 11 September 2020; all entries; 563,082 sequences). The parameters were described as follows: enzyme–trypsin; missed cleavage–1; fixed modification–carbamidomethylation of cysteine; variable modification–oxidation of methionine; MS precursor ion mass tolerance–10 ppm; MS/MS fragmentation tolerance–0.02 Da; and peptide charge: 2+, 3+ and 4+. The targeted decoy was implemented to estimate false discovery rate to less than 1%.

4.6. FAD spectrophotometry

The purified protein (1 mg/mL) dissolved in 10 mM Tris–HCl pH 8.0 were subjected to Agilent Cary 60 UV/vis spectrophotometer (Agilent Technology) to monitor changes in absorbance spectrum between the wavelength of 300 and 600 nm. The enzyme‐bound FAD was expected at 450 nm spectrum (Macheroux, 1999).

4.7. NmMTHFR assay

The apparent kinetics of NmMTHFR wildtype and mutants were carried out by NAD(P)H‐CH₂‐H₄folate oxidoreductase assay as described previously (Zuo et al., 2018), with slight modifications. The reaction kinetics of NmMTHFR enzymes (wildtype, 1 μM; Met221Trp, 1 μM; and Met221Ala, 2 μM) were performed in 50 mM phosphate buffer, pH 7.2 and 0.3 mM EDTA under anaerobic condition in anaerobic glovebox (<5 ppm O₂) at room temperature (Belle Technology) at either varying concentrations of CH₂‐H₄folate (10–250 μM) and a fixed concentration of NAD(P)H (200 μM), or varying concentrations of NAD(P)H (10–200 μM) and a fixed concentration of CH₂‐H₄folate (200 μM). The control reaction was performed accordingly, excepting that enzyme was omitted. The high quality (6R)‐5,10‐CH₂‐H₄folate or CH₂‐H₄folate was purchased from Merck Eprova AG (Schafhausen, Switzerland). The solution of CH₂‐H₄folate was prepared in an oxygen free 50 mM potassium phosphate buffer, pH 7.2 containing 0.3 mM EDTA in anaerobic glovebox (<5 ppm O₂) (Belle Technology) and the concentration was calculated based on the molar extinction coefficient of 32,000 M⁻¹ cm⁻¹ at 297 nm (Zuo et al., 2018). To avoid interference from absorbance of folate oxidation and signal saturation of NAD(P)H used, the reaction progress was monitored via absorbance decrease of NAD(P)H at 375 nm instead of 340 nm by a charged coupled device (CCD) spectrometer equipped with a UV–VIS–NIR light source (Model DH‐2000‐BAL, Ocean Optics, USA) and DET2B‐200‐1100 detector (Model MG‐7000, TgK Scientific). Slopes of the reaction traces were analyzed by Kinetic Studio software (TgK Scientific). The initial velocities (ν) were calculated using a molar extinction coefficient of 1920 M⁻¹ cm⁻¹ at 375 nm (Maenpuen et al., 2015). The apparent kinetic parameters were determined from the Michaelis–Menten plots of ν/e versus substrate concentration, which was fitted with Michaelis–Menten equation, where e represent total concentration of enzyme, k _cat denotes the catalytic rate constant or turnover number, K _m is the Michaelis–Menten constant, and S is substrate concentration.

$$\frac{v}{e}=\frac{k_{\mathrm{cat}}\bullet S}{K_{\mathrm{m}}+S}$$

4.8. Protein crystallization, data collection, and structure determination

NmMTHFR was crystallized using the hanging‐drop vapor diffusion technique in 24 well plates at 16°C. Wells contained 0.5 mL of 10% PEG 3350 reservoir solution with a drop of 1 μL reservoir solution and 1 μL purified NmMTHFR (30 mg/mL). Yellow rod‐shaped crystals were cryoprotected with Paratone‐N and flash cooled in liquid nitrogen. Diffraction images were collected on the MX2 beamline of the Australian Synchrotron at 100 K. Indexing and integration were performed using XDS, with scaling and merging completed using Aimless (CCP4) (Winn et al., 2011). The phase problem was solved by molecular replacement in Phaser using the crystal structure of Haemophilus influenzae MTHFR (PDB: 5UME) with ligands and solvent removed as the search model (McCoy et al., 2007). Model building was completed in Coot followed by refinement in Phenix (Emsley & Cowtan, 2004; Adams et al., ²⁰¹⁰). The final model has greater than 97% of peptide bonds within the favored region of the Ramachandran plot. Only a single Ramachandran outlier, Val153, is present in each chain. The secondary structure of residues 205–209 was not presented clearly; however, it was assigned as β7 strand based on the structural alignment with EcMTHFR structures (Pejchal et al., 2005). Electron density of the modeled FAD cofactors is presented in Figure S5.

AUTHOR CONTRIBUTIONS

Wanita Pantong: Conceptualization (supporting); formal analysis (supporting); funding acquisition (supporting); methodology (equal); writing – original draft (supporting); writing – review and editing (supporting). Jordan L. Pederick: Formal analysis (supporting); methodology (equal); writing – review and editing (supporting). Somchart Maenpuen: Formal analysis (supporting); funding acquisition (supporting); methodology (equal); writing – review and editing (supporting). Ruchanok Tinikul: Formal analysis (supporting); funding acquisition (supporting); methodology (equal); writing – review and editing (supporting). Jaime J. Jayapalan: Formal analysis (supporting); funding acquisition (supporting); methodology (equal); resources (supporting); writing – review and editing (supporting). Blagojce Jovcevski: Formal analysis (supporting); methodology (equal); resources (supporting); writing – review and editing (supporting). Kate L. Wegener: Methodology (supporting); resources (supporting); writing – review and editing (lead). John B. Bruning: Funding acquisition (supporting); resources (supporting); writing – review and editing (supporting). Wanisa Salaemae: Conceptualization (lead); formal analysis (lead); funding acquisition (lead); methodology (equal); resources (lead); supervision (lead); validation (lead); writing – original draft (lead); writing – review and editing (lead).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Table S1. Protein identification by peptide mass spectrometry.

Table S2. Crystallographic data collection and refinement statistics.

Scheme S1. MTHFR catalytic reactions.

Figure S1. SDS‐PAGE analysis Methodology of the tryptic digestion of in‐gel protein mass spectrometry.

Figure S2. Annotated MS/MS spectrum for tryptic digested peptide obtaining the residue at position 221.

Figure S3. UV visible absorbance spectra for purified MTHFRs.

Figure S4. Comparison of NADH versus NADPH oxidation kinetics from CH₂‐H₄folate oxidoreductase assay monitored at 375 nm of NmMTHFR reactions.

Figure S5. Apparent steady‐state kinetics of NmMTHFR wildtype and mutants based on NADH‐CH₂‐H₄folate oxidoreductase assay.

Figure S6. Walleye stereo view of simulated annealing composite omit electron density map (2F_o – F_c; 1.0σ) for the modeled FAD cofactors.

Figure S7. The (α/β)₈ of NmMTHFR.

Figure S8. Dimeric symmetry of MTHFR orthologs.

Figure S9. Structure‐based alignment of the catalytic domains of MTHFR orthologs.

Figure S10. Amino acid sequence alignment of the catalytic domains of MTHFR orthologs.

Tables S1 and S2, Scheme S1, Figures S1–S10, and methodology of the tryptic digestion of in‐gel protein MS spectrometry are obtained in supporting data. Atomic coordinates and structure factors of NmMTHFR (PDBID: 7RML) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ.

Pantong W, Pederick JL, Maenpuen S, Tinikul R, Jayapalan JJ, Jovcevski B, et al. Biochemical and structural characterization of meningococcal methylenetetrahydrofolate reductase. Protein Science. 2023;32(6):e4654. 10.1002/pro.4654

DATA AVAILABILITY STATEMENT

The data that supports the findings of this study are available in the supplementary material of this article.