Biochemical characterization and structural basis of reactivity and regioselectivity differences of Burkholderia thailandensis and Burkholderia glumae 1,6-didesmethyltoxoflavin N-methyltransferase

Abstract

Burkholderia glumae converts the guanine base of guanosine triphosphate into an azapteridine and methylates both the pyrimidine and triazine rings to make toxoflavin. Strains of Burkholderia thailandensis and Burkholderia pseudomallei have a gene cluster encoding seven putative biosynthetic enzymes that resembles the toxoflavin gene cluster. Four of the enzymes show sequence similarity to BgToxBCDE, which have been proposed to make 1,6-didesmethyltoxoflavin (1,6-DDMT). One of the remaining enzymes, BthII1283 in B. thailandensis E264, is a predicted S-adenosylmethionine (SAM)-dependent N-methyltransferase that shows low sequence identity to BgToxA, which sequentially methylates N6 and N1 of 1,6-DDMT to form toxoflavin. Here we show that, unlike BgToxA, BthII1283 catalyzes a single methyl transfer to N1 of 1,6-DDMT in vitro. In addition, we investigated the differences in reactivity and regioselectivity by determining crystal structures of BthII1283 with bound S-adenosylhomocysteine (SAH) or 1,6-DDMT and SAH. BthII1283 contains a Class I methyltransferase fold and three unique extensions used for 1,6-DDMT recognition. The active site structure suggests that 1,6-DDMT is bound in a reduced form. The plane of the azapteridine ring system is orthogonal to its orientation in BgToxA. In BthII1283 the modeled SAM methyl group is directed towards the p-orbital of N1 whereas in BgToxA it is first directed towards an sp2 orbital of N6 and then towards an sp2 orbital of N1 after planar rotation of the azapteridine ring system. Furthermore, in BthII1283, N1 is hydrogen bonded to a histidine residue whereas BgToxA does not supply an obvious basic residue for either N6 or N1 methylation.

INTRODUCTION

The rice pathogen Burkholderia glumae BGR1^1–4 produces toxoflavin^5–11 via expression of the gene cluster BgtoxABCDE.^{12, 13} We recently showed that BgToxA is an S-adenosylmethionine (SAM)-dependent N-methyltransferase that catalyzes the sequential methylation of N6 and N1 of 1,6-didesmethyltoxoflavin (1,6-DDMT, 1) to form toxoflavin (2, Figure 1B).¹⁴ Crystal structures of BgToxA complexes show a 140° rotation between 1,6-DDMT and toxoflavin that places N6 and N1 at locations consistent with sequential methylation.¹⁴ The results support earlier work which proposed that 1,6-DDMT is an intermediate produced by BgToxB-E starting from guanosine triphosphate.¹²

Figure 1. Toxoflavin biosynthesis and toxoflavin-like gene cluster in B. thailandensis E264.

(A) Diagrams of the toxoflavin biosynthetic gene cluster found in B. glumae BGRl (top) and related cluster identified in B. thailandensis E264 (bottom). The numbers in the B. thailandensis cluster refer to the last two digits of the locus tag (BTH_II12XX) assigned during genomic sequencing (e.g., 83=BTH_II1283). Homologues in the two clusters are depicted with identically colored arrows and assigned the gene name used for B. glumae BGRl. Colors denote predicted or experimentally verified gene functions: red, efflux or resistance; green, regulation; blue, biosynthesis; grey, transposition; orange and pink, unknown function. (B) Proposed biosynthetic pathway of toxoflavin in B. glumae BGRl and unknown molecule in B. thailandensis E264. The atom numbering scheme used in the manuscript can be seen illustrated on 6-DMT (4).

Burkholderia thailandensis,^{15, 16} a growing source of exciting natural products,¹⁷ and its close relative Burkholderia pseudomallei,¹⁸ the causative agent of melioidosis in humans,^{19, 20} are not known to produce toxoflavin but some strains contain a gene cluster that has similarity to the toxoflavin biosynthetic gene cluster. One such gene cluster containing 11 genes is found in B. thailandensis E264 (BTH_II1275–1285, Figure 1A, Table S1) and encodes seven putative biosynthetic proteins, BthII1279–1285. BthII1281, BthII1282, BthII1284, and BthII1285 show sequence similarity and conservation to BgToxE, BgToxD, BgToxC, and BgToxB, respectively, suggesting that B. thailandensis produces 1,6-DDMT as a biosynthetic intermediate en route to a cryptic secondary metabolite as proposed for toxoflavin biosynthesis.^{12, 21}

The remaining three putative biosynthetic gene products (BthII1283, BthII1280, and BthII1279) do not have homologs in the B. glumae cluster and suggest that toxoflavin is not the produced natural product. BthII1280 is a putative O-methyltransferase (PDB code 3MCZ), and BthII1279 is a putative member of the vicinal oxygen chelate (VOC) superfamily with a predicted metal chelating triad consisting of two histidines and one glutamate (Figure S1). BthII1283 is a putative SAM-dependent N-methyltransferase and is only 16 % identical to BgToxA, which raises the strong possibility that BthII1283 catalyzes a different transmethylation reaction.

In the present study, we initiated an investigation into the differences between the toxoflavin-like gene cluster in B. thailandensis E264 and the toxoflavin gene cluster in B. glumae BGR1 by examining the in vitro biochemical function of BthII1283 in the presence of 1,6-DDMT and 1-DMT, and the active site of BthII1283 after soaking BthII1283/S-adenosylhomocysteine (SAH) crystals with 1,6-DDMT. We find that in contrast to BgToxA, BthII1283 catalyzes methyl transfer only to N1 of 1,6-DDMT, and our crystal structures provide a structural basis for the differences in reactivity and regioselectivity.

MATERIALS AND METHODS

General.

All chemicals and solvents were purchased either from Sigma Aldrich, Alfa Aesar or Tokyo Chemical Industry Co., Ltd. (TCI). SAM was obtained as S-adenosyl-L-methionine disulfate tosylate (Sigma-Aldrich, St. Louis, MO) and used without purification. NdeI, HindIII, T4 DNA ligase, Phusion DNA polymerase, and Escherichia coli BL21(DE3)-T1R were purchased from New England Biolabs (Ipswich, MA) and used per the manufacturer’s recommendations. E. coli DH5α was purchased from Invitrogen (Carlsbad, CA). Lysogeny broth, Miller (LB) was purchased from EMD Millipore (Billerica, MA). All synthetic reactions were performed under an argon or nitrogen atmosphere employing oven-dried glassware when needed. Analytical thin-layer chromatography (TLC) was performed using silica plates (60 Å) with a fluorescent indicator (254 nm), which were visualized with a UV lamp. NMR spectra were recorded on a Bruker 700 MHz spectrometer with a 5 mm proton/carbon cryo-probe with z-axis gradient. 1H-NMR chemical shifts (δ) are expressed in ppm downfield from tetramethylsilane and are referenced to residual HDO in the D₂O NMR solvent (δ = 4.79 ppm), where s = singlet. ¹H-¹³C HMBC NMR spectra were recorded on a Bruker 700 MHz spectrometer with complete proton decoupling and lacking the low-pass filter. The lack of a low-pass filter allows direct correlations to be observed as large double resonances due to ¹J_CH. The UV absorptions were measured on an Eppendorf BioSpectrometer Kinetic and the cell used was Eppendorf mCuvette G1.0 (Eppendorf, New Haven, CT). All microbial procedures were performed using good microbial practices and aseptic technique. Kanamycin was used at 50 μg/mL for selection of plasmids in E. coli.

Creation of expression vector.

The DNA fragment corresponding to BTH_II1283 (protein accession number ABC33990.1) was amplified from B. thailandensis E264 genomic DNA by PCR using Phusion DNA polymerase and the oligonucleotides BTH1283Fwd (5′-TGGTGCCGCGCGGCAGCCATATGAGCGCCGCCGAACCGCA −3′) and BTH1283Rev (5′- TCGAGTGCGGCCGCAAGCTTTCATCGGCGTCCACTCGCGTC −3′) with an annealing temperature of 58 °C. The restriction sites of NdeI and HindIII are indicated by the underlined bases while the italicized bases indicate random bases added to increase restriction enzyme efficiency. The PCR amplicon was purified using the QIAquick PCR purification kit digestion with NdeI and HindIII. The restricted fragment was purified using the QIAquick PCR purification kit followed by ligation into pTHT, a modified pET-28 vector that supplies an N-terminal fusion tag MGSDKIHHHHHHSSGENLYFQGH, which had been previously digested with NdeI and HindIII, using T4 ligase. Chemically competent E. coli strain DH5α cells were transformed with a portion of the ligation mixture and the transformation mixture was plated on LB agar plates supplemented with kanamycin. Clones were grown overnight in LB media containing kanamycin and isolated using the QIAprep mini spin kit. The inserted sequence was verified by Sanger DNA sequencing resulting in plasmid pTBTH1283–2.

Overexpression, purification and crystallization of BthII1283.

E. coli BL21(DE3)-T1R cells containing pTBTH1283–2 were grown overnight at 37 °C on LB agar plates containing kanamycin. A single colony of each plate was inoculated into LB medium (5 mL) containing the above antibiotic and cultured in a shaker at 200 rpm at 37 °C for 8 h. The seed culture (1 mL) was transferred into LB medium (5 × 100 mL) in 5 × 500 mL Erlenmeyer flasks containing kanamycin and grown at 37 °C until the OD₆₀₀ reached 0.4. Then, the temperature was reduced to 18 °C and after 30 min adaptation, IPTG (0.1 mM final concentration) was added to each culture. After further growth for 16 h, the cells were harvested by centrifugation (2,200 x g, 20 min, 4 °C) and stored at −80 °C.

The frozen cells from 50 mL of culture were thawed on ice and then washed one time with cold sterile ddH₂O, followed by suspension in 5 mL lysis buffer [Tris-HCl (50 mM, pH 7.4), NaCl (300 mM), imidazole (10 mM)]. The suspension was sonicated using a Qsonica sonicator (9 mm probe, 30% amplitude, 15 sec, six times) on ice, then cell debris were removed by centrifugation at (19,745 x g, 4 °C for 20 min). The supernatant was applied to a Ni-NTA agarose resin (0.5 mL, BioRad, Hercules, CA) by batch binding at 4 °C for 30 min followed by pelleting the resin by centrifugation at 515 x g for 5 min. The supernatant was discarded and the resin was washed two times with 10 mL washing buffer [Tris-HCl (50 mM, pH 7.4), NaCl (300 mM), imidazole (20 mM)]. BthII1283 was eluted from the resin with elution buffer [Tris-HCl (50 mM, pH 7.4), NaCl (300 mM), imidazole (250 mM)]. The protein solution was desalted using an Econo-Pac 10DG Desalting column (BioRad) equilibrated with protein storage buffer [Tris-HCl (25 mM, pH 7.4), glycerol (10% w/v)]. The protein concentration was determined using an Eppendorf BioSpectrometer at (280 nm, 1 mm cell) with an extinction coefficient of 30,110 cm⁻¹ M^-1.

For crystallization studies, BthII1283 was isolated from cell pellets derived from 2 L of culture induced at an OD₆₀₀ of approximately 1.0 and the following modifications. The cells were suspended in lysis buffer [15 mM Tris, 500 mM NaCl, 0.1 mM SAH, and 2 mM dithiothreitol (DTT), pH 7.4, supplemented with one protease inhibitor tablet (Invitrogen) and 0.6 kU of benzonase] and sonicated on ice. The cell debris was pelleted by centrifugation at 60,000 x g for 40 min and the supernatant was subjected to nickel chelate chromatography using wash (20 mM Tris, 500 mM NaCl, 30 mM imidazole, 0.1 mM SAH, 2 mM DTT, pH 7.3) and elution (20 mM Tris, 500 mM NaCl, 250 mM imidazole, 0.1 mM SAH, 2 mM DTT, pH 7.3) buffers. The eluted protein was exchanged into wash buffer using an Econo-Pac 10DG desalting column (Bio-Rad) and was incubated at 4 °C for 24 h with tobacco etch virus protease (His₆-tagged) for cleavage of the hexahistidine tag. The proteolytic reaction mixture was subjected to subtractive nickel chelate chromatography, and the purified protein was exchanged into 5 mM HEPES, 25 mM NaCl, and 1 mM SAH, pH 7.0, concentrated to approximately 5 mg/mL, flash frozen in liquid nitrogen, and stored at −80 °C.

Crystallization.

BthII1283 and SAH were co-crystallized at 22 °C using the hanging drop vapor diffusion method. Drops were formed by combining protein (approximately 5 mg/mL) and reservoir solutions in a 1:1 ratio. Typical reservoir solutions consisted of 100 mM HEPES, pH 6–7, and 1–2 M (NH₄)₂SO₄; the (NH₄)₂SO₄ concentration was increased to 3.2 M prior to cryo-cooling the crystals with liquid nitrogen. Crystals used for experimental phasing were soaked in cryo-solutions supplemented with NaBr, with the concentration of NaBr increased gradually to 0.875 M. Crystals with bound 1,6-DDMT were obtained by soaking the crystals in cryo-solutions supplemented with 25 mM 1,6-DDMT, 6 mM SAH, and 5 mM DTT.

Data Collection and Processing.

Crystals of BthII1283 were irradiated with X-rays at 100 K at beam line NE-CAT 24-ID-C of the Advanced Photon Source (APS) and beam lines A1 and F1 of the Cornell High Energy Synchrotron Source (CHESS). X-ray diffraction experiments at NE-CAT 24-ID-C used X-rays having a wavelength of λ = 0.979 Å. Diffraction images were recorded with 1°/s oscillations on a PILATUS 6MF detector positioned 260 mm from the crystal. Experiments at A1 used X-rays having a wavelength of λ = 0.976 Å. Diffraction images were recorded for 1° oscillations and 7 s exposures on an ADSC Q210 CCD detector positioned 175 mm from the crystal. Finally, experiments at F1 used X-rays having a wavelength near the Br absorption edge (λ = 0.918 Å). Diffraction images were recorded for 1° oscillations and 5 s exposures on an ADSC Q270 CCD detector positioned 250 mm from the crystal. X-ray images were processed using XDS or HKL2000.^{22, 23} Table 1 gives a summary of the data collection statistics.

TABLE 1.

X-ray Data Collection Statistics.

	BthII1283 SAH Br	BthII1283 SAH	BthII1283 SAH 1,6-DDMT
Beamline	CHESS F1	APS 24-ID-C	CHESS A1
λ (Å)	0.9179	0.9793	0.9759
Space group	P61	P61	P61
a (Å)	111.7	112.3	112.7
c (Å)	78.3	78.2	78.4
Resolution of highest shell (Å)	1.80–1.83	1.39–1.44	1.77–1.83
Reflections	796,613	543,988	185,550
Unique reflections	51,464; 38,043a	112,155	54,274
Redundancy	15.5 (10.2)b	4.9 (3.7)	3.4 (3.0)
% Complete	99.9 (99.6)	99.5 (98.1)	98.1 (85.9)
<I>/<σI>c	21.4 (1.4)	26.7 (2.2)	17.3 (1.9)
<I/σI>	19.7	14.0 (2.2)	12.6 (2.6)
Rmerge (%)	11.9	4.7 (49.0)	5.8 (37.9)

^aNumber of Bijvoet pairs

^bValues in parentheses are associated with the highest resolution shell

^cAverage intensity divided by average error in intensity.

Structure Determination and Refinement.

Crystallographic phases were determined from combined molecular replacement and single-wavelength anomalous diffraction phasing using Phaser²⁴ based on Br sites located using SHELXD.^{25, 26} A truncated form of a methyltransferase (PDB ID 3CGG) was used as a search model. Density modification and auto-building were performed using PARROT²⁷ and BUCCANEER,²⁸ respectively. Iterative automated structure refinement with translation, libration, and screw vibrational motion of partitioned chains²⁹ was performed using PHENIX³⁰ and iterative manual model building was performed using COOT.³¹ Table 2 summarizes the structure refinement statistics.

TABLE 2.

Structure Refinement Statistics.

	BthII1283 SAH	BthII1283 SAH 1,6-DDMT
No. of reflections	112,141	54,199
No. of reflections in working set	106,620	51,534
Resolution (Å)	1.39	1.77
No. of protein atoms	3,636	3,594
No. of waters	573	462
No. of ligand atoms	152	176
RMSD bonds (Å)	0.005	0.006
RMSD angles (deg)	0.942	0.881
Rwork (%)	14.8	15.2
Rfree (%)	16.8	18.4
Ramachandran analysis
Most favored (%)	91.9	92.1
Additional allowed (%)	7.9	7.9
Generously allowed (%)	0.3	0.0

Size exclusion chromatography with multi-angle light scattering (SEC-MALS).

The oligomeric state of BthII1283/SAH was determined via SEC-MALS.^32–34 A frozen aliquot of BthII1283/SAH prepared for crystallization was thawed and injected into a WTC-050S5 SEC column connected to a DAWN HELEOS-II 18-angle light scattering detector and an Optilab T-rEX refractive index detector (Wyatt Technology Corporation). The column was operated at room temperature with a flow rate of 1 mL/min and was pre-equilibrated with a filtered solution of 25 mM Tris, 75 mM NaCl, 0.1 mM SAH, and 10 mM DTT, pH 7.5. A 5 mg/mL sample of monomeric bovine serum albumin (Sigma-Aldrich) was used as a molecular weight standard. Analysis of chromatograms was performed using ASTRA V software.

Synthesis of 1,6-Didesmethyltoxoflavin (1,6-DDMT) and 1-desmethyltoxoflavin (1-DMT).

The synthesis of 1-DMT was performed as previously reported.³⁵ The synthesis of 1,6-DDMT was performed as previously described with the exception that for the last step 0.3 eq. of triphosgene was used in place of phosgene.³⁶

In vitro enzymatic activity assay.

A reaction mixture containing 25 mM Tris-HCl buffer (pH 7.4), 0.1 mM 1,6-DDMT, 1 mM SAM, 1 mM DTT, and 15 μM BthII1283 in a final volume of 100 μL was mixed gently by pipetting and incubated at 28 °C. After 1 h the protein was removed by passing the reaction mixture through a VWR centrifugal filter (Modified PES 10K, 500 μL). All negative control reactions were performed in the same manner by replacing the omitted reagent with water. A portion of the reaction (50 μL) was then analyzed by HPLC as described below.

Enzymatic synthesis of toxoflavin.

A reaction mixture containing 25 mM Tris-HCl buffer (pH 7.4), 0.1 mM 1-DMT, 1 mM SAM, 1 mM DTT, and 15 μM BthII1283 in a final volume of 100 μL was mixed gently by pipetting and incubated at 28 °C. After 1 h the protein was removed by passing the reaction mixture through a VWR centrifugal filter (Modified PES 10K, 500 μL) and then was analyzed by HPLC as described below. HRMS for toxoflavin (ESI): m/z [M+H]⁺ calc. for C₇H₈N₅O₂ 194.0673, obs. 194.0669 (2 ppm error).

HPLC analysis of in vitro reaction mixtures.

All enzymatic reactions were analyzed on an Agilent 1100 HPLC instrument consisting of a vacuum degasser, quaternary pump, autosampler (cooled to 4 °C), column thermostat (maintained at 30 °C), and diode array detector using a Supelcosil LC-18-DB column (15 cm x 4.6 mm, 5 μm, Supelco, Bellefonte, PA) where line A was water, line C was 10 mM NH₄OAc, and line D was methanol at a flow rate of 1 mL/min. The column was pre-equilibrated with 90% A/10% C and held at this composition for 2 min. The composition was changed to 75% A/10% C/15% D over the next 13 min using a linear gradient. The composition was then adjusted to 10% A/10% C/80% D over 2 min and this composition was held for 5 min. The mobile phase was then changed to 90% A/10% C over 3 min and the composition was held for 6 min prior to the next injection.

Large scale enzymatic synthesis and partial purification of 6-desmethyltoxoflavin (6-DMT).

A reaction mixture of 50 mL containing potassium phosphate buffer (25 mM, pH 7.4), 0.1 mM 1,6-DDMT, 1 mM SAM, 1 mM DTT, and 15 μM BthII1283 was mixed and incubated at 28 °C for 1 h then the protein was removed by passing the reaction mixture through GE Healthcare centrifugal tubes (Vivaspin20, 5 kDa MWCO). The filtrate was concentrated under vacuum to 1.8 mL. The 1.8 mL filtrate was loaded to a PuriFlash 450 system (Interchim Inc., Los Angeles, CA). Separation was achieved using an Uptisphere Strategy RP 10 μm 150 × 21.2 mm Prep-LC column at flow rate 10 mL/min where line A was water, line C was 0.1% formic acid in water, and line D was methanol. The column was pre-equilibrated with 90% A/10% C and upon injection this composition was held for 2 min. The mobile phase was changed to 75% A/10% C/15% D over the next 13 min using a linear gradient and then altered to 10% A/10% C/80% D over 2 minutes. This composition was held for 5 min. Fractions containing the product were pooled and then concentrated to 500 μL using a miVac concentrator (SP Scientific, Warminster, PA) to give 0.9 mg of partially purified product. ¹H NMR (700 MHz, D₂O) δ 8.77 (s, 1H, H-3), 3.89 (s, 3H, N-CH₃). HRMS (ESI): m/z [M+H]⁺ calc. for C₆H₆N₅O₂ 180.0516, obs. 180.0512 (2 ppm error).

RESULTS AND DISCUSSION

Phylogenetic analysis and in vitro biological activity.

Phylogenetic analysis of BthII1283 showed that it closely clusters with other methyltransferases identified from the genomes of Burkholderia pseudomallei and is related to methyltransferases identified in Xenorhabdus and Pseudomonas strains, while BgToxA clustered in a different clade and is more closely related to methyltransferases identified in different Xenorhabdus strains as well as Streptomyces strains (Figure S2). BthII1283 and the most related proteins were found in putative clusters with genes encoding proteins with similarity to ToxB, ToxC, ToxD and ToxE (Figure S2). This suggests that BthII1283 utilizes 1,6-DDMT as a substrate for methylation. Due to the 16 % sequence identity to BgToxA, we hypothesize that BthII1283 has the potential to methylate 1,6-DDMT with a different regioselectivity.

Incubation of 1,6-DDMT (1) yields a single product that generates a protonated molecular ion of 180.0512, which is consistent with the molecular formula C₆H₆N₅O₂ (calc. 180.0516, 2 ppm error, Figure S3A,B), demonstrating that a single methyl group is transferred to 1,6-DDMT. Large-scale production, isolation and ¹H NMR analysis of the partially purified product in D₂O shows two singlets at 3.89 and 8.77 ppm (Figure S4). Further purification results in conversion of the product back to 1,6-DDMT, so characterization was performed on the partially purified product. The singlet at 3.89 ppm has an HSQC correlation to the carbon resonance at 43.1 ppm, while the singlet at 8.77 ppm has an HSQC correlation to an aromatic carbon resonance at 145.7 ppm. HMBC correlations from the aromatic signal show a correlation to a carbon resonance at 146.7 ppm while the methyl singlet shows a correlation to a carbon resonance at 152.2 ppm (Figure S5). Comparison of the observed chemical shifts to those published for toxoflavin suggests that the methyl group was located at N1 to yield 6-DMT (4, Figure 1B).³⁷

To provide further evidence that the methylation occurs at N1 we incubated 1-DMT with BthII1283. Reumycin (1-DMT) is converted to a single product with an identical retention time and UV-vis spectrum to a synthetic standard of toxoflavin and generated a protonated ion of 194.0669 (calc. 194.0673, 2 ppm error) demonstrating that the methylation occurred at N1 (Figure S3C,D).

Overall structure of BthII1283.

BthII1283 contains a Class I methyltransferase fold,^{38, 39} an N-terminal extension, and two internal extensions (Figure 2A, B). The N-terminal extension contains a β-hairpin and an 18-residue α-helix connected by a seven residue loop, which collectively form a ring-like structure. The first internal extension is located after the fifth β-strand (β7) of the Class I methyltransferase fold and contains a β-hairpin that forms hydrogen bonds with the β-hairpin in the N-terminal extension to make a four-stranded half β-barrel. The second internal extension consists of the top eight residues of the β-hairpin formed by the final two strands (β10 and β11) of the Class I methyltransferase fold, which protrude outwards. A search of the Protein Data Bank shows that the variable regions of BthII1283 are unique but resemble those of some other methyltransferases.⁴⁰ Of the ones characterized, human α-N-terminal methyltransferase 1 (HsNTMT1) (PDB 5E1B; ~19 % identical in primary sequence to BthII1283),⁴¹ which catalyzes the methylation of the α-amino group of Ser1 in the Ran guanine nucleotide-exchange factor, shows the highest degree of structural similarity (Figure 3A).^42–44 A notable difference is that the N-terminal extension of HsNTMT1 is shorter and lacks the 16-residue segment in BthII1283 that contains a β-hairpin, which results in an exposed substrate-binding cavity.

Figure 2. Crystal structure of BthII1283.

(A) Ribbon representation of BthII1283. The Class I methyltransferase fold, N-terminal extension, first internal extension, and second internal extension are colored light blue, salmon, light green, and light purple, respectively. SAH and 1,6-DDMTare shown as balls and sticks. (B) Structural topology diagram. Helices and β-strands are shown as cylinders and large arrows, respectively. (C) Crystallography asymmetric unit. Monomer A is on the left and monomer B is on the right.

Figure 3. Structural comparison of BthII1283 with other N-methyltransferases.

(A) Superimposition of HsTMT1 onto BthII1283. ( B) Superimposition of BgToxA onto BthII1283. The Class I methyltransferase fold is colored light gray. The N-terminal extension, first internal extension, and second internal extension of BthII1283 are colored salmon, light green, and light purple, respectively. The corresponding regions in HsNTMTl and BgToxA are colored dark red, dark green, and purple, respectively. SAH bound to BthII1283 is shown as balls and sticks.

The variable regions of BthII1283 show a slight resemblance to and occupy similar locations as the variable regions of BgToxA. However, the N-terminal extension of BgToxA is shorter and all-helical, the first internal extension of BgToxA is larger and contains a three-stranded β-sheet, and the second internal extension of BgToxA has a bigger loop that contains two short α-helicies (Figure 3B).¹⁴ Furthermore, the two internal extensions of BgToxA form a clamshell-like structure with the Class I methyltransferase fold that results in a substrate-binding cavity with a side entrance, and the N-terminal extension undergoes disorder-order transitions that open and close the cavity.¹⁴ In contrast, the N-terminal and internal extensions of BthII1283 form a cap over the Class I methyltransferase fold that encloses the substrate-binding cavity in both BthII1283/SAH and BthII1283/SAH/1,6-DDMT structures. However, the β-hairpin in the N-terminal extension lays directly over and interacts with 1,6-DDMT and its displacement would result in an exposed substrate-binding cavity, similar to the structure of HsNTMT1, which lacks the N-terminal β-hairpin. In addition, several residues in the N-terminal β-hairpin and the loop that follows it display high B-factors (Figure 4). These results suggest that the substrate-binding cavity of BthII1283 has a top entrance and that the N-terminal β-hairpin acts as a lid.

Figure 4. B-factors of residues of BthII1283 based on average for main chain atoms.

(A) Asymmetric unit of BthII1283/SAH crystals with monomer A on the left and monomer B on the right. (B) Asymmetric unit of BthII1283/SAH/1,6-DDMT crystals with monomer A on the left and monomer B on the right. The darkest shade of blue corresponds to the minimum residue B-factor (10.5 in panel A and 14.4 in panel B), white corresponds to the average residue B-factor (17.8 in panel A and 23.5 in panel B), and the darkest shade of red corresponds to a residue B-factor five or more standard deviations above the average (59.9 in panel A and 68.3 in panel B). SAH and 1.6-DDMT are shown as balls and sticks.

Crystal packing arrangement.

The BthII1283 crystals belonged to the space group P6₁ and the asymmetric unit contains two BthII1283 monomers (A and B) that display twofold non-crystallographic symmetry. The twofold axis runs between the α2 helices located at the protein-protein interface (Figure 2C). The remainder of the interface is formed from interactions between α3 and β10. Computer analysis of the interface suggests that this quaternary arrangement is disfavored when the ammonium sulfate ions from the structure are omitted.⁴⁵ Size exclusion chromatography with multi-angle light scattering (SEC-MALS) shows that BthII1283 elutes as a monomer in a solution containing 25 mM Tris, 75 mM NaCl, 0.1 mM SAH, and 10 mM DTT at pH 7.5 (Figure S6).

SAH binding site.

SAH binds to a conserved SAH/SAM-binding site in the Class I methyltransferase fold and to the α-helix of the N-terminal extension (Figure 5). The C-terminal end of the first β-strand of the fold (β₃) is followed by a conserved nucleotide-binding motif, GXGXG (G₇₄AG₇₆RG₇₈), and the C-terminal end of the second β-strand of the fold (β₄) is followed by a conserved acidic residue (Asp95) that forms hydrogen bonds with the ribose hydroxyls of SAH.⁴⁶ The adenine moiety of SAH is interposed between the side chain of Cys137 and the side chain of Leu96. N7 and the amino group of the adenine moiety form hydrogen bonds with the side chain amide of Gln141. N1 forms a hydrogen bond with the backbone NH of Val120, and N3 forms a hydrogen bond with the backbone amide NH of Leu96. Finally, the amino group of the homocysteine moiety of SAH forms hydrogen bonds to the backbone carbonyls of Gly74 and Asn135 and the carboxylate group of SAH makes a salt bridge with the guanidinium group of Arg79.

Figure 5. SAH-binding site of BthII1283.

Stereoview (left) and schematic (right) representation of interactions of SAH (balls and sticks) with residues of the Class I methyltransferase fold (light blue) and N-teminal extension (salmon). Dashed lines represent potential hydrogen bonds. Water molecules are shown as red spheres in the stereo diagram and ‘W’ in the schematic drawing.

Active site structure.

Electron density maps for BthII1283/SAH crystals soaked with 1,6-DDMT show 1,6-DDMT bound to BthII1283 in the presence but not in the absence of DTT (Figure S7). However, different binding modes occur in monomer A and monomer B (Figure 6). In both monomers, the azapteridine ring system is interposed between the side chain of Trp47 and the side chain of Phe168, with N1 being the closest nitrogen to Sδ of SAH (Sδ …N1, Sδ …N2, and Sδ …N8 distances are 3.8/3.7 Å, 4.0/4.2 Å, and 5.¼.5 Å, respectively). In addition, N6 forms a hydrogen bond with the carboxylate group of Asp10 (N6…Oδ2 distance is 2.6/2.9 Å), N4 forms a hydrogen bond to the hydroxyl group of Tyr219 (N4…OH distance is 3.0/2.8 Å), and O5 hydrogen bonds to a water molecule and is located near the guanidinium group of Arg212 (O5…NH2 distance is 3.4/3.5 Å).

Figure 6. 1,6-DDMT-binding site of BthII1283.

Stereoview ( left) and schematic (right) representations of interactions of 1,6-DDMT with residues of the Class I methyltransferase fold (light blue), N-teminal extension (salmon), first internal extension (light green), and second internal extension (light purple) in (A) monomer A and (B) monomer B of the asymmetric unit. SAH and 1,6-DDMT are shown as balls and sticks. Dashed lines represent potential hydrogen bonds. Water molecules are shown as red spheres in the stereo diagrams and ‘W’ in the schematic drawings. The side chain of Trp47 stacks against the azapteridine ring system of 1,6-DDMT but is omitted for clarity.

The main difference between the two binding modes occurs on the side of 1,6-DDMT containing N1, O7, and N8. In monomer A, N1 forms a hydrogen bond with Nε2 of His140 (N1… Nε2 distance is 2.8 Å), O7 forms a hydrogen bond with the hydroxyl group of Tyr185 (O7…OH distance is 2.6 Å), and O7 and N8 are within hydrogen bonding distance of the guanidinium group of Arg183 (O7…NH1 and N8…NH2 distances are 3.0 and 3.1 Å, respectively) (Figure 6A). In monomer B, N1 is within hydrogen bonding distance of the backbone carbonyl group of Gly136 (N1…O distance is 3.2 Å) and Nε2 of both His139 and His140 (N1…N_ε2 distances are 3.4 Å). N8 forms a hydrogen bond with Nε2 of His140 (N8… Nε2 distance is 2.8 Å). O7 is within hydrogen bonding distance of the hydroxyl group of Tyr185 (O7…OH distance is 3.4 Å) and the guanidinium group of Arg183 (O7…NH1 distance is 2.8 Å), which adopts two conformations (Figure 6B).

Substrate complex model.

A model of the enzyme-substrate complex based on the structure of BthII1283 monomer A shows that the methyl group of SAM is directed towards the p-orbital of N1 of 1,6-DDMT and that His140 is well-positioned near N1 to act as a base (Figure 7A). It is unclear if the binding mode of 1,6-DDMT in monomer B is biochemically significant or is different due to crystallization artifacts or the tautomeric state of ,6-DDMT; the side chain of Arg183 adopts two conformations (Figure 6B) and the N-terminal β-hairpin shows high B-factors (Figure 4). Nevertheless, the structure shows that proton transfer between His140 and N8 of 1,6-DDMT is structurally feasible.

Figure 7. Stereoview representations of enzyme-substrate complex models.

(A) Model of BthII1283 active site (monomer A) with bound SAM and 1,6-DDMT. (B) Model of BgToxA active site with bound SAM and 1,6-DDMT (top) or 1-DMT (bottom: model based on toxoflavin). In both panels. SAM is modeled based on SAH coordinates. Dashed lines represent potential hydrogen bonds.

The active site structure of both BthII1283 monomers is consistent with the results of the crystal soaking experiments, which suggest that 1,6-DDMT is bound in a reduced form (Figure 6). The hydroxyl group of Tyr219 is 3.0/2.8 Å from N4 of 1,6-DDMT and 2.6/2.5 Å from a carboxylate oxygen atom of Glu50, which suggests that N4 is likely to be protonated. In monomer A, N1 is 2.8 Å from Nε2 of His140 and Nε1 of His140 is 2.9 Å from the carbonyl oxygen atom of Leu184, which suggests that N1 is also likely to be protonated. On the other hand, the proximity of the guanidinium group of Arg183 to N8 suggests that N8 is likely to be deprotonated. In addition, O7 interacts closely with the guanidinium group of Arg183 and forms a hydrogen bond with the hydroxyl group of Tyr185, which raises the possibility of oxyanion stabilization. In monomer B, N8 is 2.8 Å from Nε2 of His140 and Nε1 of His140 is 2.8 Å from the carbonyl oxygen atom of Leu184, which suggests that N8 is likely to be protonated. In addition, N1 is 3.2 Å from the carbonyl oxygen atom of Gly136, which suggests that N1 is also likely to be protonated. The binding of the reduced form of 1,6-DDMT in the active site of BthII1283 is consistent with the fact that DTT was required for full in vitro activity (Figure S8). Here we propose that 1,6-DDMT (1) is first reduced to compound 3, which binds to enzyme with N8 deprotonated. The proton is abstracted from N1 to generate the dianion 7. This dianion then undergoes methylation at N1 followed by dissociation from the enzyme active site and oxidation, presumably through a reaction with molecular oxygen to generate 6-DMT (4). These results are illustrated in the schematic drawings of Figure 6 and in the proposed mechanism (Figure 8).

Figure 8. Proposed mechanism of methylation of 1,6-DDMT by BthII1283.

First 1,6-DDMT is reduced to 3, which binds to enzyme as the N8 anion. The proton is abstracted from Nl to generate dianion 7, which then undergoes méthylation at N1 followed by dissociation from the enzyme active site and oxidation, presumably through a reaction with molecular oxygen to generate 6-DMT(4).

Methylation specificity.

Both BgToxA and BthII1283 methylate 1,6-DDMT but result in methyl groups at distinct sites producing regioisomers. BgToxA methylates N6 to produce 1-DMT (reumycin)¹⁴ while BthII1283 methylates N1 to produce 6-DMT. Molecular models show that positioning N6 in various proper orientations for methylation results in unfavorable interactions and clashes with active site residues of BthII1283 (Figure S9) consistent with the specificity observed for the in vitro reaction.

BthII1283 and BgToxA both interpose 1,6-DDMT between a tryptophan side chain and a phenylalanine side chain but bind 1,6-DDMT in different orientations using a different set of residues.¹⁴ In BgToxA, the plane of the azapteridine is orthogonal to its orientation in BthII1283 (Figure 7B).¹⁴ In addition, unlike BthII1283, BgToxA does not supply an obvious basic residue for the catalysis of either N1 or N6 methylation. We proposed previously that 1,6-DDMT is bound to BgToxA with N6 deprotonated and that tyrosine and asparagine residues stabilize an O5 oxyanion. N1 is deprotonated in the likely solution form. Enzyme-substrate complex models for BgToxA show that the methyl group of SAM is directed towards an sp² lone pair on N6 and then towards an sp² lone pair on N1 after planar rotation of the azapteridine ring system (Figure 7B).¹⁴ The binding of deprotonated forms of 1,6-DDMT and 1-DMT enables BgToxA to use a similar active site geometry for both N6 and N1 methylations. BthII1283 only needs to catalyze a single methylation and binding of a protonated form of 1,6-DDMT enables BthII1283 to perform base-assisted methylation of N1.

Biological Significance.

Early studies on the toxoflavin biosynthetic gene cluster proposed that BgToxBCDE synthesize 1,6-DDMT and that BgToxA, the only methyltransferase encoded by the gene cluster, methylates 1,6-DDMT twice to form toxoflavin.¹² Our recent study on BgToxA confirmed the latter prediction, revealed the order of the methyl transfer reactions, and provided a structural basis of the dual methyltransferase activity.¹⁴ The high degree of sequence identity between BthII1285, BthII1284, and BthII1282 of B. thailandensis and BgToxBCD (as well as the moderate sequence identity between BthII1281 and BgToxE) led us to investigate whether 1,6-DDMT could also be a potential biosynthetic intermediate in B. thailandensis. Our biochemical results show that BthII1283 methylates N1 of 1,6-DDMT and our crystal structure of BthII1283 with bound 1,6-DDMT shows a high degree of structural complementarity between the enzyme and substrate. These results strengthen the plausibility that 1,6-DDMT is a bona fide intermediate and hence also suggest that 1,6-DDMT is a branch point to multiple natural products. The presence of genes encoding proteins with similarity to ToxBCDE in the genomic regions surrounding the methyltransferases identified by BLAST⁴⁷ analysis suggests that further investigations will reveal the presence of multiple natural products derived from GTP with 1,6-DDMT functioning as a key intermediate. However, to date we have been unable to isolate a natural product associated with this cluster from B. thailandensis E264 despite growth in multiple media under a variety of conditions and comparison of the wild-type strain to a strain containing a genetic disruption of BTH_II1282 (data not shown).

ABBREVIATIONS

1-DMT

1-desmethyltoxoflavin (reumycin)

1,6-DDMT

1,6-didesmethyltoxoflavin

6-DMT

6-desmethyltoxoflavin

D2O

deuterium oxide

DNA

deoxyribose nucleic acid

DTT

dithiothreitol

ESI

electrospray ionization

GTP

guanosine triphosphate

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HMBC

heteronuclear multiple bond correlation

HPLC

high performance liquid chromatography

HRMS

high resolution mass spectrometry

HsNTMT1

human α-N-terminal methyltransferase 1

HSQC

heteronuclear single quantum coherence

IPTG

isopropyl β-D-1-thiogalactopyranoside

kU

kilounit

LB

lysogeny broth

Ni-NTA

Ni-nitrilotriacetic acid

NMR

Nuclear magnetic resonance

OD₆₀₀

optical density measured at 600 nm

PCR

polymerase chain reaction

PDB

protein data bank

PEG

polyethylene glycol

PES

polyethersulfone

RMSD

root mean square deviation

RMSD

root-mean-squared deviation

rpm

revolutions per minute

SAH

S-adenosylhomocysteine

SAM

S-adenosylmethionine

SEC-MALS

Size exclusion chromatography with multi-angle light scattering

TLC

thin layer chromatography

Tris

tris(hydroxymethyl)aminomethane

UV

ultraviolet

x g

times gravity