Molecular structure of a 5,10‐methylenetetrahydrofolate dehydrogenase from the silkworm Bombyx mori

The enzyme 5,10‐methylenetetrahydrofolate dehydrogenase (MTHFD) is essential for the production of certain amino acids (glycine, serine, and methionine) and nucleic acids (thymidylate and purine). Here, we identified a cDNA encoding this enzyme from the silkworm Bombyx mori. The recombinant B. mori MTHFD (bmMTHFD) expressed in Escherichia coli recognized 5,10‐methylenetetrahydrofolate and 5,10‐methenyltetrahydrofolate as substrate in the presence of NADP + as well as NAD +. The bmMTHFD structure was determined at a resolution of 1.75 Å by X‐ray crystallography. Site‐directed mutagenesis indicated that the amino acid residue Tyr49 contributed to its catalytic activity. Our findings provide insight into the mechanism underlying the activity of MTHFD from B. mori and potentially other insects and may therefore facilitate the development of inhibitors specific to MTHFD as insecticides.

The enzyme 5,10-methylenetetrahydrofolate dehydrogenase (MTHFD) is essential for the production of certain amino acids (glycine, serine, and methionine) and nucleic acids (thymidylate and purine). Here, we identified a cDNA encoding this enzyme from the silkworm Bombyx mori. The recombinant B. mori MTHFD (bmMTHFD) expressed in Escherichia coli recognized 5,10-methylenetetrahydrofolate and 5,10-methenyltetrahydrofolate as substrate in the presence of NADP + as well as NAD + . The bmMTHFD structure was determined at a resolution of 1.75 A by X-ray crystallography. Site-directed mutagenesis indicated that the amino acid residue Tyr49 contributed to its catalytic activity. Our findings provide insight into the mechanism underlying the activity of MTHFD from B. mori and potentially other insects and may therefore facilitate the development of inhibitors specific to MTHFD as insecticides.
One-carbon metabolism is involved in the synthesis of amino acids like alanine, glycine, and serine, as well as purine and pyrimidine bases [1,2]. 5,10-methylenetetrahydrofolate (MTHF) dehydrogenase (MTHFD; EC 1. 5.1.5) is an enzyme involved in one-carbon metabolism in eukaryotes [2]. To date, cytosolic and mitochondrial versions of this protein have been reported. For example, MTHFD1 is a cytoplasmic protein [3], whereas MTHFD1L, MTHFD2, and MTHFD2L are mitochondrial proteins [4][5][6]. Human MTHFD1 is a trifunctional enzyme with dehydrogenase (D), cyclohydrolase (C), and synthetase (S) activities that catalyze the oxidation of MTHF to 5,10-methenyl-THF, which is then hydrolyzed to 10-formyl-THF, and finally converted to THF and formate [3]. The 3D structure of the D/C domain of MTHFD1, referred to as DC301, has been reported [3]. MTHFD2 and MTHFD2L are bifunctional enzymes [7,8], whereas MTHFD1L is a monofunctional enzyme [9]. MTHFD often requires NADP + or NAD + as the cofactor for their activity. MTHFD1 requires NADP + [3], MTHFD2 and MTHFD2L use either NADP + or NAD + [7,8], whereas MTHFD1L is monofunctional with only S activity and does not use either cofactors [9]. Likewise, the prokaryotic MTHFD of Escherichia coli is a bifunctional enzyme that uses NADP + [10], and the monofunctional enzyme of Peptostreptococcus productus requires NADP + as the cofactor [11]. Although one-carbon metabolism has been studied in vertebrates, there are no reports from invertebrates, including silkworm and other insects.
To characterize one-carbon metabolism in insects, we isolated mRNA encoding an MTHFD of the silkworm Bombyx mori MTHF dehydrogenase (bmMTHFD), which is an important lepidopteran insect model. The structure-function relationships of insect MTHFDs have not been studied in detail. Since many agricultural pests are lepidopteran insects, it is useful to investigate the amino acid residues present in the active site of bmMTHFD. Further, because MTHFD is involved in the synthesis of important biomolecules such as amino acids and purine and pyrimidine bases, the inhibitors could be effective insecticides against agricultural pests. Here, we determined the three-dimensional structure of bmMTHFD to identify the amino acid residues important for bmMTHFD activity and conducted mutation analysis of bmMTHFD to determine the role of the amino acids lining the substrate-binding site. Examination of bmMTHFD catalytic activity indicated that it participates in the D and C activities. The active site in bmMTHFD was then determined to better understand the structural basis for this conversion. As described, mammalian MTHFDs are key enzymes involved in the synthesis of amino acids and purine and pyrimidine bases, which are crucial biomaterials for survival. Analysis of inhibition of insect MTHFDs would aid in the design of pesticides and insecticides. The crystal structure of bmMTHFD and the identification of the amino acid residues involved in catalytic function in the current study may provide insights into the mechanism underlying MTHFD activity and could facilitate the development of inhibitors specific to MTHFD as insecticides. To the best of our knowledge, this study is the first to report on MTHFD in insects.

Insects
Bombyx mori larvae (p50T strain) were reared at the Kyushu University Graduate School (Fukuoka, Japan) and fed mulberry leaves. Day-3 fifth-instar larvae were dissected on ice, and fat body was stored at À80°C until use.
RNA extraction, cloning, and sequencing of cDNA encoding bmMTHFD Total RNA was isolated from the fat body using RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) and was analyzed by reverse transcription-PCR. First-strand cDNA was obtained using SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and an oligo-dT primer. The resulting cDNA was used as a PCR template with the following oligonucleotide primers: 5 0 -CAACAGCCATATG GCGCGTATCCTCGATGG-3 0 (sense) and 5 0 -CCGGATC CTTAATTGGATTTGTTTGCTTGA-3 0 (antisense). The primer designs were based on a partial sequence obtained from the SilkBase database (http://silkbase.ab.a.u-tokyo.ac. jp/cgi-bin/index.cgi). The underlined and double-underlined regions indicate NdeI and BamHI restriction enzyme sites, respectively, which were used for insertion of the PCR product into an expression vector. The PCR program was as follows: 94°C for 2 min, 35 cycles of 94°C for 1 min, 59°C for 1 min, and 72°C for 2 min, and 72°C for 10 min. The bmMTHFD cDNA (bmmthfd) was ligated into the pGEM-T Easy Vector (Promega, Madison, WI, USA) and transformed into E. coli DH5a.
To obtain the complete sequence of bmmthfd and to deduce its amino acid sequence, the GENETYX-MAC software (ver. 14.0.12; GENETYX Corporation, Tokyo, Japan) was used. Homology alignment was performed using CLUSTALW (ver. 1.83; DNA Data Bank of Japan, Shizuoka, Japan), with 10 and 0.2 as the gap creation penalty and gap extension, respectively. A phylogenetic tree was generated using neighbor-joining plot software (http://d oua.prabi.fr/software/njplot).

Overexpression and purification of recombinant protein
The bmmthfd clone was digested with NdeI and BamHI, subcloned into the expression vector pET-15b (Merck Millipore, Darmstadt, Germany), and transformed into competent E. coli Rosetta (DE3) pLysS cells (Merck Millipore). The cells were then grown at 37°C in Luria-Bertani media containing 100 lgÁmL À1 ampicillin. After the cell density reached an OD 600 of 0.7, isopropyl-1-thio-b-D-galactoside was added to a final concentration of 1 mM to induce recombinant protein production. The culture was further incubated for 3 h, and the cells were harvested by centrifugation at 10 000 g for 15 min. Bacteria were resuspended in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , and 1.76 mM KH 2 PO 4 , pH 7.4) containing 4 mgÁmL À1 lysozyme and were subsequently disrupted by sonication. Sonication was performed three times for 1 min each on ice using a 3-mm tapered microtip probe. The supernatant containing the recombinant protein was clarified by centrifugation at 10 000 g for 15 min and subjected to Ni 2+affinity chromatography equilibrated with the same buffer. After washing with the same buffer, the samples were eluted with a linear gradient of 0-0.5 M imidazole. The enzyme-containing fractions, assayed as described below, were pooled, concentrated using a centrifugal filter (10 000 molecular weight cutoff; Millipore Corp., Billerica, MA, USA), and applied to a Superdex 200 column (GE Healthcare Bio-Sciences, Buckinghamshire, UK) equilibrated with 20 mM Tris/HCl buffer (pH 8.0), with the addition of 0.2 M NaCl. The purity of the pooled material was analyzed by sodium dodecyl sulfate (SDS)/PAGE performed using a 15% polyacrylamide slab gel containing 0.1% SDS, according to the method reported by Laemmli [12]. Protein bands were visualized by Coomassie Brilliant Blue R-250 (Sigma-Aldrich, St. Louis, MO, USA) staining.

Protein crystallization and X-ray diffraction data analysis
Recombinant bmMTHFD was purified as described in 'Overexpression and purification of recombinant protein' and then concentrated using a centrifugal filter (Millipore) to 10 mgÁmL À1 in 20 mM Tris/HCl buffer, pH 8.0, containing 0.2 M NaCl. Crystallization was performed using the sittingdrop vapor diffusion method at 20°C with Crystal Screen Kits (Hampton Research, Aliso Viejo, CA, USA). Each drop was formed by mixing an equal (1 lL) or twofold greater volume (1 : 2 lL) of protein and reservoir solutions, respectively. Crystals suitable for X-ray analysis were grown for 4 weeks in 0.2 M ammonium sulfate, 0.1 M sodium cacodylate trihydrate, and 30% PEG 8000 (w/v). The crystal was transferred to reservoir solution with 25% (v/v) ethylene glycol using a cryoloop and flash frozen with liquid nitrogen before data collection. X-ray diffraction data collection was performed using synchrotron radiation on SPring-8 beamline BL44XU with a wavelength of 0. 9 A, in a nitrogen vapor stream at 100 K [13]. The data set was integrated and scaled using the DENZO and SCALEPACK programs, as implemented in the HKL2000 program package [14].

Determination of structure
The initial structure was determined using a molecular replacement method with the program PHENIX PHASER-MR [15], using the polyalanine model of E. coli MTHFD (PDB ID: 1B0A) as a search model. The ARP/WARP program [16] was used to build missing side-chain atoms and to add water molecules. After manual adjustment to the electron density map with the program COOT [17], refinement was performed using the program PHENIX.REFINE [18]. Figures were prepared using the PyMOL software (http://py mol.sourceforge.net). The atomic coordinates and structure factor of bmMTHFD have been deposited in the Protein Data Bank (PDB ID: 5ZF1). The 3D structure alignment was performed using the MATRAS server [19].

Measurements of enzyme activity
Dehydrogenase activity of bmMTHFD was spectrophotometrically measured using MTHF (Toronto Research Chemicals, North York, ON, Canada) as substrate and NADP + or NAD + as cofactor. MTHF, NADP + , and NAD + were stored at À80°C until use. The reaction mixture (100 lL) included 20 mM sodium phosphate buffer (pH 6.5), 30 mM 2-mercaptoethanol, and various concentrations of substrate, NADP + or NAD + . The assay was initiated by addition of the enzyme and then incubated at 28°C for 10 min. Production of methenyltetrahydrofolate was measured at 350 nm by an endpoint assay and a background rate determined in enzyme-free assays. The substrate concentrations were calculated using the extinction coefficient of 24.9 mM À1Á cm À1 [21]. Kinetic parameters were measured with a nonlinear least squares data fit to the Michaelis-Menten equation using KALEIDA-GRAPH (Synergy Software; HULINKS Inc., Tokyo, Japan).
Cyclohydrolase activity of bmMTHFD was assayed with 5,10-methenyl-THF (Santa Cruz Biotechnology, Inc., Dallas, TX, USA), that was stored at À80°C until use. Reaction mixture (100 lL) contained 0.1 M potassium maleate (pH 7.4), 0.02 M 2-mercaptoethanol, and 0.1 mM 5,10methenyl-THF. The reaction was started by addition of the enzyme and then incubated at 30°C for 10 min. The assay was performed by measuring the decrease in absorption at 355 nm [22], and the values were compared to those from an enzyme-free assay.

Structural description of bmMTHFD
The crystal structure of bmMTHFD was solved by the molecular replacement method, and the structure was refined against 1.75 A resolution data. The data collection and refinement statistics are summarized in Table 1. Two molecules were included in the crystallographic asymmetric unit and showed similar structures with root-mean-square deviation (rmsd) between the corresponding atoms of 1.05 A. The residues 44-46 and 246-253 of the A chain and 244-252 of the B chain were not modeled because of poor electron density. Figure S1 shows that bmMTHFD eluted at 205 mL, corresponding to a molecular size of 63 kDa, and appears to be a homodimeric protein based on the elution profile. Dali search for protein structure comparison between the determined structure and structures already deposited in the Protein Data Bank [23] revealed that DC301 (PDB ID: 1A4I) with rmsd of 1.40 A and hMTHFD2 (PDB ID: 5TC4) with rmsd of 1. 10 A were most similar. The sequence alignment of bmMTHFD to that of DC301 showed that the a-helices and b-strands of bmMTHFD were conserved across the structures (Fig. 1). bmMTHFD is composed of two domains containing N-terminal and C-terminal domains with an a/b fold. In DC301, hydrogen bonds are crucial for stabilizing interactions of the DC301 monomers [3]. Thr191, Val192, Thr193, and Thr194 of one subunit of DC301 interact with Thr199, Thr194, Thr193, and Val192 of the other subunit via hydrogen bonds [3]. The sequence alignment of DC301 to bmMTHFD (Fig. 1) indicated that Thr193, Val194, Thr195, and Ile196 of the B. mori residues interact with Thr201, Ile196, Thr195, and Val194 of the other subunit in a similar manner to DC301. These amino acid residues form b-strand in the C-terminal region and provide bridging hydrogen bonds in bmMTHFD that contribute to stabilizing the dimer of bmMTHFD. There are hydrogen bonds observed between the side chain of Thr193 in one subunit and side chain of Thr201 in the other subunit, main chain of Val194 of one subunit and main chain of Ile196 in the other subunit, side chain of Thr195 in one subunit and the counterpart in the other subunit, and main chain of Ile196 in one subunit and main chain of Val194 in the other subunit (Fig. 1).

Amino acid residues involved in substrate and cofactor binding
Bombyx mori MTHFD produced in bacteria was purified to homogeneity, yielding a single band in SDS/ PAGE with a molecular size of~32 000 Da. This was close to the estimated size based on the deduced amino acid sequence.

Discussion
This study aimed to detect the presence of one-carbon metabolism in the silkworm B. mori and to analyze whether MTHFD is relevant for insecticide design. The amino acid sequence of bmMTHFD revealed high homologies to those of mitochondrial MTHFD2 (49%) and MTHFD2L (51%). However, bmMTHFD did not possess a predicted N-terminal mitochondrial targeting sequence (Fig. 1). According to substrate specificity, MTHFD proteins are divided into mono-, bi-, and trifunctional proteins. For instance, MTHFD1 is a trifunctional enzyme [25], MTHFD2 is a bifunctional enzyme [26], and MTHFD1L is a monofunctional enzyme [9]. Our experiments revealed that bmMTHFD is bifunctional, exhibiting D and C activities ( Table 2). In DC301, Tyr52-XXX-Lys56 and/or Ser49, Gln100, and Pro102 were proposed as C active site. Since we found the corresponding residues in bmMTHFD to be Tyr49-XXX-Lys53 and/or Ser46, Gln97, and Pro99 [3], we predicted that bmMTHFD possesses the C domain (Fig. 1). The amino acid residues required for S activity are located in the S domain of the 70 kDa MTHFD1. However, bmMTHFD lacks the S domain for S activity (Fig. 1).
Elucidation of the bmMTHFD tertiary structure identified a globular shape that was similar to those of other known MTHFDs. In addition, the conserved amino acid residues are critical for stabilizing the interaction of bmMTHFD monomers. Our study indicates that Thr193, Val194, Thr195, Ile196, and Thr201 are involved in hydrogen bond interactions, suggesting that these amino acid residues are involved in maintaining the homodimer of bmMTHFD.
Our findings show that bmMTHFD contains the catalytic site where the substrate binds in the N-terminal domain and displays cofactor binding in the Cterminal domain, like other MTHFDs. bmMTHFD is composed of 10 a-helices and nine b-strands, and their location is highly conserved in DC301 (Fig. 1).
The K m value of bmMTHFD was also found to be similar to those for Pseudomonas aeruginosa (26 lM for MTHF and 176 lM for NADP + ) [27], and Trypanosoma brucei (35 lM for MTHF and 70 lM for NADP + ) [28]. bmMTHFD uses either NADP + or NAD + for catalysis with similar k cat /K m values ( Table 2). Similar results were obtained with mammalian MTHFDs, where hMTHFD2 and mMTHFD2L were reported to utilize both NADP + and NAD + [7,8]. Superimposition between bmMTHFD and DC301 requiring NADP + indicated that a single cofactor bound to the site contained Arg166, Ser167, Arg199, and Ile242 (Fig. 4A). Superimposition between bmMTHFD and hMTHFD2 requiring NAD + showed that Arg166, Ser167, Asn169, Arg199, Ile242, and Thr281 were present in bmMTHFD in a dinucleotide-binding motif (Fig. 4B). In bmMTHFD, the motif (165GRSRQVG171) does not conform to the classical GXGXXG fingerprint pattern. The sequence comparisons did not suggest the amino acid residues that may be involved in NAD + utilization. Cocrystallization with NADP + or NAD + would allow improved prediction of specificity.
Tyr52, Lys56, Gln100, and Asp125 in hMTHFD1 have been identified as critical for its activity [29]. Superimposition between DC301 and bmMTHFD revealed that the corresponding residues in bmMTHFD were Tyr49, Lys53, Gln97, and Asp120, respectively. Tyr52 (bmMTHFD Tyr49) was implicated in substrate binding, and Lys56 (bmMTHFD Lys53) and Gln100 (bmMTHFD Gln97) were proposed to work together in establishing an environment for hydride transfer [3,29,30]. Asp125 (bmMTHFD Asp120) serves in substrate binding. Mutations at Tyr49 eliminated affinities to MTHF and 5,10-methenyl-THF. Mutagenesis of Asp120 influences the D activity. The results are consistent with a key role for Tyr52, and Asp125 in hMTHFD1 to correctly bind the folate substrate. Since MTHFDs have an aspartate residue at the 125 position, the carboxylate is invariant. The side chain would then be in position to interact with the edge of the pterin ring at N3 and at the exocyclic amine of C2 of pterin [31]. In bmMTHFD, the electron density from Pro44 to Ser46 of A chain was disordered. In DC301, there is a hydrogen-bonding network near these amino acid residues [3]. To examine whether Ser46 contributes to bmMTHFD activity, we mutated the amino acid residue to Ala. Subsequent mutagenesis results showed that their kinetic parameters were not changed significantly ( Table 2). In E. coli MTHFD, mutation of Asp121 (bmMTHFD Asp120) decreases the k cat /K m values toward MTHF to 0.17%, while the kinetic parameters of mutating Tyr50 (bmMTHFD Tyr49) could not be determined, due to very low activities [32]. The results indicate that the catalytic mechanism could be similar to that in E. coli MTHFD.
In conclusion, this study identified and biochemically characterized bmMTHFD, a MTHFD in the silkworm B. mori. To the best of our knowledge, this study is the first to report an MTHFD in insects. Furthermore, we identified bmMTHFD amino acid residues that are likely to play roles in catalysis. We are currently attempting the cocrystallization of bmMTHFD with a suitable substrate analogue conjugate to aid in the determination of amino acid residues involved in bmMTHFD catalysis. Our findings provide insight into the mechanism underlying bmMTHFD activity and potentially that of other insects and may therefore facilitate the development of more effective and safe insecticides.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.  Fig. S2. Enzymatic properties of bmMTHFD. The enzymatic properties of bmMTHFD were analyzed using 5, 10-methylenetetrahydrofolate as substrate and NADP + as cofactor. The MTHFD activity was assayed under standard conditions, as described in Experimental Procedures, unless otherwise indicated. The maximum value obtained was set to 100%. Data represent the mean with AE SD from three independent experiments. (A) pH stability was assessed by preincubation of the enzyme solution at various pH values at 4°C for 24 h before the residual activity was assayed.
(B) Thermostability was determined by preincubation of the enzyme solution at various temperatures at pH 6 for 30 min before the residual activity was assayed.
(C) Optimum pH levels for the activities were assayed at 28°C using citrate-phosphate-borate buffer at various pH value.