Structural and Functional Analysis of SAM-Dependent N-Methyltransferases Involved in Ovoselenol and Ovothiol Biosynthesis

SUMMARY

Thio/selenoimidazole Nπ-methyltransferases are an emerging family of enzymes catalyzing the final step in production of the S/Se-containing histidine-derived antioxidants ovothiol and ovoselenol. These enzymes, prevalent in prokaryotes, show minimal sequence similarity to other methyltransferases, and the structural determinants of their reactivities remain poorly understood. Herein, we report ligand-bound crystal structures of OvsM from the ovoselenol pathway as well as a member of a previously unknown clade of standalone ovothiol-biosynthetic Nπ-methyltransferases, which we have designated OvoM. Unlike previously reported ovothiol methyltransferases, which are fused as a C-terminal domain to the sulfoxide synthase OvoA, OvoMs function independently. Comparative structural analyses reveal conserved, ligand-induced conformational changes, suggesting similar behavior in dual-domain OvoA enzymes. Mutagenesis supports a model where OvoA domain rearrangement facilitates substrate recognition via a critical Tyr residue in the domain linker. Biochemical studies identify an essential active site Asp, likely serving as a catalytic base in the S_N2-like nucleophilic substitution reaction.

Graphical Abstract

eTOC blurb

Ireland et al. report X-ray crystal structures of the SAM-dependent N-methyltransferases responsible for the final transformation in ovothiol and ovoselenol biosynthesis, S- and Se-containing natural product antioxidants. The results elucidate ligand recognition and reactivity in these enzymes while shedding light on dual-domain, bifunctional OvoA enzymes involved in ovothiol biosynthesis.

INTRODUCTION

Thio/selenoimidazoles (TSIs) are S- and Se-containing biomolecules derived from histidine and synthesized by bacteria, fungi, and archaea (Fig. 1A).¹ Their antioxidant qualities play a crucial role in the oxidative stress responses of their producers and hold promise for the prevention and treatment of a spectrum of inflammatory conditions in humans.^2–6 TSIs characterized to date include ergothioneine and ovothiol, along with their Se isologs selenoneine and the recently identified molecule ovoselenol. Following its discovery, the biosynthetic pathway of ovoselenol was elucidated (Fig. 1B).⁷ This pathway begins with the formation of N-acetyl-1-seleno-β-D-glucosamine (SeGlcNAc), catalyzed by homologs of the selenophosphate synthetase SenC and the selenosugar synthase SenB from the sen biosynthetic gene cluster.⁸ Oxidative coupling of SeGlcNAc with histidine is then catalyzed by a member of the non-heme iron-dependent sulfoxide/selenoxide synthase (NHISS) family. This enzyme, OvsA, generates a selenoxide intermediate via the installation of a C–Se bond at the imidazole 5′-carbon of histidine. The selenoxide intermediate rapidly undergoes syn elimination to release the sugar moiety followed by reduction to afford 5-selenohistidine (5-SeHis). A final S-adenosylmethionine (SAM)-dependent methylation at the imidazole π-nitrogen of 5-SeHis, mediated by OvsM, completes the pathway to yield the mature natural product. While the biosynthesis of ovothiol proceeds similarly, the required NHISS and methyltransferase functionalities were reported to be integrated within the bifunctional enzyme OvoA.^9,10 Unlike OvsM, however, the methyltransferase domain of OvoA does not directly accept the product of its NHISS domain; instead, an additional lyase (OvoB) is needed to cleave the cysteinyl C–S bond prior to methylation (Fig. 1B).

Figure 1. Thio/selenoimidazoles and associated biosynthetic pathways.

(A) Structures of the four currently known TSIs. (B) Biosynthetic pathways for ovothiol and ovoselenol.

Our recent crystal structures of SenB and OvsA offer insights into ligand recognition and associated conformational changes important for selenosugar biosynthesis, as well as elucidate the molecular basis for 5’-specific C–Se bond formation en route to ovoselenol.^7,11 In this study, we further illuminate ovoselenol biosynthesis by providing structural insights into the final enzymatic transformation. While Wang et al. clarified the mode of ligand binding within the NHISS domain of a bifunctional OvoA from Hydrogenimonas thermophila (OvoA_Th2), the structure of the methyltransferase domain was reported only in the apo state.¹² Herein, we present the first visualization of ligand binding by Nπ-methyltransferases involved in TSI biosynthesis. Structures of OvsM are reported, together with bioinformatic identification and structural characterization of a standalone Nπ-methyltransferase involved in ovothiol biosynthesis, which we have termed OvoM. Together with the published model of OvoA_Th2, these results enable a comparative structural analysis of the three methyltransferase classes involved in ovothiol and ovoselenol biosynthesis. Mutagenesis studies provide further insights into ligand recognition by the methyltransferase domain of bifunctional OvoA_Th2, as well as the reaction mechanism of these enzymes.

RESULTS AND DISCUSSION

Bioinformatic analysis reveals monofunctional Nπ-methyltransferases in ovothiol biosynthesis

We initiated our investigation of SAM-dependent methyltransferases involved in TSI biosynthesis with a comprehensive bioinformatic study. The vast majority of ovsM homologs were found to be either fused to ovoA or as discrete genes co-encoded with ovsA. Unexpectedly, we also discovered a third category within a small family of Campylobacterota that harbor a split version of the ovoA gene, with NHISS and methyltransferase domains separated by a large chromosomal distance (Fig. 2A). To the best of our knowledge, standalone 5-thiohistidine N-methyltransferases have not been previously reported; we therefore designate these as OvoM. It is tempting to speculate that the monofunctional enzymes OvoM and OvsM may share a more recent common ancestor, possibly arising from a single ovoA-gene-splitting event. However, all three classes of TSI-biosynthetic Nπ-methyltransferases (OvoA C-terminus, OvoM, and OvsM) are highly similar, with an average pairwise sequence identity of around 50% (Fig. 2B). Furthermore, they appear to be biosynthetically interchangeable as all three efficiently methylate both 5-thioHis and 5-SeHis. Specifically, we recombinantly expressed and purified a member from each category and monitored methyltransferase activity with 5-thioHis or 5-SeHis. Similar conversions were observed with all three enzymes, with slightly higher yields with 5-SeHis, likely resulting from enhanced electron donation into the imidazole ring system rather than an enzymatic preference (Fig. S1).

Figure 2. Phylogenetic analysis of SAM-dependent N-methyltransferases involved in the biosynthesis of ovoselenol and ovothiol.

(A) Organization and context of genes encoding three types of TSI-biosynthetic Nπ-methyltransferases (MT): OvoA C-terminus, OvoM, and OvsM. (B) Average pairwise % sequence identity between the three different MT types. (C) Phylogenetic tree, including four distant homologs involved in N-methylation of related compounds. Specific proteins analyzed in this study are labeled with circles. (D) Products of the MTs displayed in panel C, each of which features a methylated N-heterocycle (shaded green).

To gain additional insights, we constructed a phylogenetic tree of the family, which suggests that OvoM and OvsM arose from separate evolutionary events (Fig. 2C). Remarkably, this gene-splitting event seems to have occurred at least three times in their evolution, as evidenced by two distinct clades of OvsM and a single clade of OvoM, all three of which diverged from different regions of the OvoA C-terminus phylogeny. The four previously characterized methyltransferases with the highest sequence similarity to OvsM are also included in the phylogenetic tree. Although these enzymes (BorM1, HrsC, MitM, and CyaF) likewise methylate N-heterocycles, their similarity to OvsM is surprisingly marginal (Fig. 2C, D).^13–16 It remains to be seen whether other methyltransferases with intermediate homology catalyze imidazole methylation in other natural products.

Overall architectures of OvsM and OvoM

Given their limited homology to characterized systems, structural insights are crucial to understand the reactivity of the TSI-biosynthetic Nπ-methyltransferases. We, therefore, crystallized representative homologs from Marinimicrobium koreense (OvsM_Mko) and Sulfuricurvum sp. isolate STB_99 (OvoM_Scu). The resultant structures, solved to 1.48 and 2.27 Å resolution, respectively (PDB IDs 9BXH and 9BXM; Table 1, 2), showcase remarkably similar overall architecture (Fig. 3A, B; S2A), with a root-mean-square deviation (RMSD) of 0.69 Å for the peptide backbone (over 825 atoms). Their topology is characteristic of the conserved Rossmann fold frequently utilized by SAM-dependent methyltransferases. In each case, a seven-stranded β-sheet sits at the center of the molecule, surrounded by six major α-helices flanked by two additional two-stranded β-sheets. Typical of class I Rossmann-fold methyltransferases,¹⁷ the β-strand order for the central β-sheet is 3–2–1–4–5–7–6, with the C-terminal β-strand oriented antiparallel to the others. The active site is located at the center of the molecule, shaped primarily by loops, α1, and the C-terminal region of the central β-sheet. Both models closely align with the methyltransferase domain of the bifunctional enzyme OvoA_Th2 (Fig. S2B), displaying only minor differences in secondary structure (Fig. S3). RMSDs of the peptide backbone range from 0.53 Å for OvoM_Scu (over 802 atoms) to 0.65 Å for OvsM_Mko (over 847 atoms). Their structures also closely align with the carnosine N-methyltransferase CARNMT1 (PDB ID 5YF0), yielding backbone RMSDs of 1.89 Å (over 593 atoms) and 2.05 Å (over 585 atoms) for OvoM_Scu and OvsM_Mko, respectively (Fig. S2B). CARNMT1 catalyzes a similar reaction, methylating the imidazole π-nitrogen of its histidine-derived substrate, carnosine.¹⁸ These findings are striking given that CARNMT1 shares virtually no detectable sequence homology with TSI-biosynthetic methyltransferases (Table S1, S2).

Table 1.

Crystallographic data processing and refinement statistics for OvsM_Mko structures.

PDB ID (Ligand)	9BXH (OvsMMko)	9BXJ (OvsMMko + SAM)	9BXK (OvsMMko + 5-SeHis + SAH)
Data Collection a
Space group	C 2 2 21	C 2 2 21	C 2 2 21
Unit cell (Å)	a = 83.8, b = 90.1, c = 80.3	a = 84.0, b = 90.6, c = 79.5	a = 83.8, b = 90.1, c = 80.3
	α = β = γ = 90	α = β = γ = 90	α = β = γ = 90
Wavelength (Å)	0.9686	0.9686	0.9686
Resolution range (Å)	29.00 – 1.48 (1.53 – 1.48)	28.87 – 1.56 (1.62 – 1.56)	29.00 – 1.66 (1.72 – 1.66)
Total observations	100582 (9091)	83063 (8223)	68585 (6513)
Total unique observations	50394 (4614)	43317 (4241)	35794 (3373)
CC1/2	0.999 (0.560)	0.999 (0.602)	0.997 (0.521)
I/σ1	11.38 (1.10)	12.40 (1.25)	9.95 (1.20)
Completeness (%)	98.86 (90.99)	99.64 (98.46)	98.75 (95.33)
Rmerge	0.0293 (0.6790)	0.0398 (0.6858)	0.0533 (0.6965)
Rpim	0.0293 (0.6790)	0.0398 (0.6858)	0.0533 (0.6965)
Redundancy	2.0 (2.0)	1.9 (1.9)	1.9 (1.9)
Refinement Statistics a
Resolution range (Å)	29.00 – 1.48	28.87 – 1.56	29.00 – 1.66
Reflections (total)	50355 (4594)	43282 (4233)	35765 (3366)
Reflections (test)	1912 (169)	1639 (151)	1362 (130)
Total atoms refined	2153	2172	2236
Solvent	286	221	282
Rwork/Rfree	0.1765/0.1965	0.1927/0.2137	0.1776/0.2051
RMSDs
Bond lengths (Å)/angles (°)	0.005/0.78	0.007/1.00	0.011/1.21
Ramachandran plot
Favored/allowed (%)	99.16/0.84	99.17/0.83	99.18/0.82
Mean B values (Å2)
Protein	21.54	23.06	23.24
Ligands	30.67	27.62	22.19
Solvent	35.97	34.54	35.60

^aValues in parentheses refer to the high-resolution shell.

Table 2.

Crystallographic data processing and refinement statistics for OvoM_Scu structures.

PDB ID (Ligand)	9BXM (OvoMScu)	9BXL (OvoMScu + SAM)	9BXN (OvoMScu + 5-thioHis + SAH)
Data Collection a
Space group	P 21 21 21	P 21 21 21	P 21 21 21
Unit cell (Å)	a = 70.2, b = 92.2, c = 96.0	a = 70.2, b = 92.2, c = 96.0	a = 69.8, b = 92.0, c = 96.2
	α = β = γ = 90	α = β = γ = 90	α = β = γ = 90
Wavelength (Å)	1.0	0.9795	0.9530
Resolution range (Å)	39.62 – 2.27 (2.35 – 2.27)	48.27 – 2.09 (2.17 – 2.09)	48.10 – 1.70 (1.76 – 1.70)
Total observations	52934 (3988)	74766 (7063)	131672 (13153)
Total unique observations	28478 (2306)	37433 (3567)	68782 (6777)
CC1/2	0.999 (0.764)	0.997 (0.644)	0.999 (0.597)
I/σ1	16.63 (1.93)	9.49 (1.11)	13.00 (1.11)
Completeness (%)	96.68 (79.83)	99.56 (96.36)	99.97 (99.96)
Rmerge	0.0325 (0.3115)	0.0543 (0.5509)	0.0276 (0.6035)
Rpim	0.0325 (0.3115)	0.0543 (0.5509)	0.0276 (0.6035)
Redundancy	1.9 (1.7)	2.0 (2.0)	1.9 (1.9)
Refinement Statistics a
Resolution range (Å)	39.62 – 2.27	48.27 – 2.09	48.10 – 1.70
Reflections (total)	28475 (2307)	37403 (3552)	68773 (6776)
Reflections (test)	1422 (116)	1863 (166)	3424 (329)
Total atoms refined	3998	4229	4381
Solvent	228	290	400
Rwork/Rfree	0.1865/0.2354	0.1913/0.2222	0.1685/0.1930
RMSDs
Bond lengths (Å)/angles (°)	0.004/0.68	0.008/1.05	0.010/1.16
Ramachandran plot
Favored/allowed (%)	98.34/1.66	98.76/1.24	99.59/0.41
Mean B values (Å2)
Protein	36.90	32.74	27.59
Ligands	46.01	30.94	33.63
Solvent	43.99	38.22	41.38

^aValues in parentheses refer to the high-resolution shell.

Figure 3. Crystal structures of OvsM_Mko and OvoM_Scu in various ligand-bound states.

Overall structures of (A) apo-OvsM_Mko (PDB ID 9BXH), (B) apo-OvoM_Scu (PDB ID 9BXM), (C) OvsM_Mko·SAM complex (PDB ID 9BXJ), (D) OvoM_Scu·SAM complex (PDB ID 9BXL), (E) OvsM_Mko·SAH·5-SeHis complex (PDB ID 9BXK) and (F) OvoM_Scu·SAH·5-thioHis complex (PDB ID 9BXN). Insets for (E) and (F) show H-bonding interactions for both ligands. Secondary structure elements are labeled, and the ligands are shown in gray sticks. The DxGxAxG sequence (motif I) connecting β1 and α3 is highlighted in yellow, motif II (located at the end of β2) highlighted in red, and motif III (in the loop connecting β5 and β6) highlighted in purple.

Ligand recognition by OvsM and OvoM

To elucidate the active site features that facilitate OvsM- and OvoM-catalyzed methylation at the imidazole π-nitrogen, we subsequently crystallized the enzymes with either the methyl donor SAM or their respective histidine-derived substrate and the co-product S-adenosylhomocysteine (SAH). Structures of the OvsM_Mko•SAM, OvsM_Mko•SAH•5-SeHis, OvoM_Scu•SAM, and OvoM_Scu•SAH•5-thioHis complexes were determined to 1.56, 1.66, 2.09, and 1.70 Å resolution, respectively (PDB IDs 9BXJ, 9BXK, 9BXL, and 9BXN; Fig. 3C–F; S4). Despite the evolutionary distance between the three TSI-biosynthetic methyltransferase subfamilies (Fig. 2C), they feature strong sequence conservation of active site residues (Fig. S5). Inspection of the binary complexes reveals that the enzymes exhibit conserved SAM-binding motifs typical of nucleoside-binding Rossmann-fold enzymes (Fig. 3E, F).^17,19,20 Motif I is characterized by a Gly-rich loop (DxGxAxG) in which the Asp and first Gly residue (D51/D55 and G53/G57 in OvsM_Mko/OvoM_Scu) interact with the α-amino moiety of SAM, with the former mediated by a water molecule. The backbone amide of the Ala residue (A55/A59), by contrast, interacts with a ribose hydroxyl group. Motifs II and III consist of acidic residues that interact with the nucleoside region of SAM. The first of these (D74/D78) forms hydrogen (H)-bonds with the ribose hydroxyl groups, while the second (D128/D132) interacts with the adenine ring.

Beyond the conserved SAM-binding motifs I-III, a plethora of variable residues located in active site loops and α1 further stabilize SAM through both direct and water-mediated H-bonds (Fig. 3E, F; S6A, B). In OvsM_Mko/OvoM_Scu, residues N4/7 (mediated by a water molecule), L75/Y79, A129/133, and R150/154 (water-mediated) interact with the nucleoside portion. Residues Y15/18, L52/56, Q22 in OvoM_Scu, and S59 in OvsM_Mko (where OvoM_Scu features a non-interacting Ala residue) form water-mediated H-bonds with SAM’s amino acid moiety. G145/T149, N146/150, and H19 in OvsM_Mko (analogous to Q22 in OvoM_Scu) also interact with the amino acid moiety through direct H-bonds.

The ternary complexes depict the histidine-derived substrates adjacent to SAM, appropriately positioned for methyl group transfer to the π-nitrogen. As with SAM binding, OvsM_Mko/OvoM_Scu-mediated recognition of 5-SeHis/5-thioHis primarily involves interactions with active site loops and α1 (Fig. 3E, F; S6C, D), in addition to interactions with a Gln residue (Q235/Q239) from the C-terminal β-strand (β11). The 5-SeHis/5-thioHis binding contacts, however, are significantly more conserved compared to SAM binding (Fig. S6). These include Y6/9, Q14/17 (mediated by a water), and Y15/18, which H-bond to the amino acid moiety; as well as D149/153, which H-bonds to the imidazole ring; and T178/182, which H-bonds to the S/Se atom. Additionally, R150/154 forms multiple H-bonds with the substrate, interacting with the S/Se atom, imidazole ring, and carboxylate moiety. The distinguishing feature of 5-SeHis/5-thioHis binding between the two enzymes lies in the additional interaction between the imidazole side chain of the substrate and N146 in OvsM_Mko. The analogous Asn from OvoM_Scu (N150), while positioned similarly in the active site, is prohibitively distant at approximately 4.3 Å from the closest imidazole nitrogen. Notably, there are no apparent differences that would cause a preference for S vs. Se, consistent with the observation that these enzymes are biosynthetically interchangeable.

Ligand-induced conformational changes in OvsM and OvoM

Superposition of substrate-free OvsM_Mko and OvoM_Scu with their binary and ternary complexes reveals several structural changes that accompany substrate binding (Fig. 4A, B). These changes primarily occur at the side chain level, while the backbone shows minimal deviation. RMSDs between the apo and binary/ternary complexes are 0.14 Å (over 829 atoms) and 0.20 Å (863 atoms) for OvsM_Mko, and 0.22 Å (891 atoms) versus 0.29 Å (887 atoms) for OvoM_Scu. In the apo state, however, electron density for the N-termini was absent, suggesting they sample multiple conformations and serve as a localized exception to the overall backbone stability. Upon binding SAM, these regions undergo stabilization, uniformly orienting toward the active site. This reorganization serves multiple purposes: shaping the solvent-accessible substrate-binding pockets (Fig. 4C, D), and enabling interactions between N4/7 (OvsM_Mko/OvoM_Scu) and SAM as well as Y6/9 and 5-SeHis/5-thioHis. In order to accommodate N-terminal motion, the side chain of R150/154 shifts away from Y6/9. This reorientation places the side chain parallel with the substrate imidazole upon 5-SeHis/5-thioHis binding, thereby shaping the binding cleft (Fig. 4E, F) and facilitating H-bonding with the substrate. Thus, SAM-binding appears to modulate recognition of the histidine-derived substrate.

Figure 4. Substrate-induced conformational changes in OvsM_Mko and OvoM_Scu.

(A) SAM binding (purple) and 5-SeHis/SAH binding (green) stabilize the N-terminus of OvsM_Mko for ligand recognition and trigger reorientation of active site residues. (B) SAM binding (coral) and 5-thioHis/SAH binding (blue) initiate similar stabilization of the N-terminus and side chain movements in OvoM_Scu. The dashed lines in panels A and B represent the unstructured region of the N-terminus in the apo structures. Surface representations of (C) OvsM_Mko and (D) OvoM_Scu, highlighting the contribution of the N-termini to the structure of the substrate-binding site. Shaping of the 5-SeHis/5-thioHis-binding cleft upon SAM-triggered reorientation of (E) R150 in OvsM_Mko and (F) R154 in OvoM_Scu, and 5-SeHis/5-thioHis-triggered reorientation of (G) Q235 in OvsM_Mko and (H) Q239 in OvoM_Scu. Note that panels E – H display SAM and 5-SeHis/5-thioHis irrespective of the ligand state, to provide context on how the side chain movements influence the shape of the binding cleft.

This sequential binding model is further confirmed by thermal denaturation analysis using differential scanning calorimetry (DSC), whereby we probed the protein stabilization effects induced by substrate binding. As expected, we observe two progressive increases in the melting temperature of OvsM_Mko upon addition of SAH followed by 5-SeHis, but no change upon the addition of 5-SeHis alone, indicating a clear order of binding events (Fig. S7). Examination of the OvsM_Mko/OvoM_Scu ternary complexes reveals that following SAH binding, subsequent interaction with 5-SeHis/5-thioHis is accompanied by a rotameric change in Q235/239 (Fig. 4A, B, G, H). Y15/18 is also subtly reoriented to optimize the distances for H-bonding with both substrates. In addition to these conserved movements, the OvoM_Scu active site undergoes two additional side chain rotations of T149 and T182 upon SAM and 5-thioHis binding, respectively (Fig. 4B).

Methyltransferase activity in bifunctional, dual-domain OvoA requires domain rearrangement

Comparison with the methyltransferase domain of the bifunctional enzyme OvoA_Th2 highlights a strikingly similar active site composition (Fig. 5A; S8). This observation suggests that the mode of substrate recognition is likely conserved despite the enzymes’ evolutionary separation and modest sequence identities (44-49%; Table S1, S2). A key difference, however, lies in the placement of the N-terminal Tyr residue (Y6/9) responsible for H-bonding with the carboxylate of 5-SeHis/5-thioHis in OvsM_Mko and OvoM_Scu. In OvoA_Th2, the corresponding Tyr residue (Y467) is located distant from the active site in the linker region connecting its two domains. Given that SAM binding to OvsM_Mko and OvoM_Scu triggers stabilization of their N-termini to prime the active site for the histidine-derived substrate, we suspect that OvoA_Th2 may experience similar SAM-induced conformational rearrangements. Analogous Y467-mediated binding of 5-thioHis, for example, would necessitate rearrangement of the two domains relative to the published crystal structure (Fig. 5B; PDB ID 8KHQ).¹²

Figure 5. Structural comparison of the methyltransferase domain of OvoA_Th2 with OvsM_Mko and OvoM_Scu.

(A) Superposition of the methyltransferase domain of bifunctional OvoA_Th2 (PDB ID 8KHQ, maroon) with the ternary complexes of OvsM_Mko (green) and OvoM_Scu (chain A in blue; chain B, exhibiting a catalytically incompetent conformation, in gray). For clarity, chain B is omitted from the top image. (B) Hypothetical model for domain rearrangement in OvoA_Th2 to facilitate Y467-mediated 5-thioHis recognition. Lower right corner: percent conversion of wild-type and mutant OvoA_Th2 and OvsM_Mko (OvoA_Th2 activity tested using 5-thioHis, OvsM_Mko using 5-SeHis). Error bars represent standard deviation of three replicates.

Curiously, the N-terminus of the chain B OvoM_Scu model follows the divergent orientation of the OvoA_Th2 linker, pointing away from the active site (Fig. 5A; S9). The absence of electron density for 5-thioHis in this monomer supports our analysis above and highlights the essential role of Y9 in substrate recognition. To investigate whether substrate binding by OvoA_Th2 relies on equivalent interactions, we generated an OvoA_Th2-Y467F variant for comparison. The mutant exhibited a reduction in catalytic activity of over 50% (Fig. 5B), supporting this hypothesis and indicating the necessity of domain rearrangement for methyltransferase activity in OvoA_Th2. Comparable results were obtained for OvsM_Mko-Y6F (Fig. 5B). The orientations of Y476, R611, and Q695 in OvoA_Th2 (Fig. 5A) further substantiate this notion, as these residues closely resemble their counterparts from the apo structures of OvsM_Mko and OvoM_Scu (Fig. 4A, B). Similar side chain repositioning is therefore necessary to achieve optimal H-bond distances for 5-thioHis binding by the OvoA_Th2 methyltransferase domain.

Histidine Nπ-methyltransferases utilize a catalytic Asp residue for imidazole deprotonation

In addition to providing unique insights into the dynamics associated with substrate binding, our models of the OvsM_Mko/OvoM_Scu reactant complexes allowed us to interrogate the mechanism of these intriguing enzymes. SAM-dependent methyltransferases commonly employ S_N2-like nucleophilic substitution chemistry.^21,22 Evaluation of our ligand-bound crystal structures supports such a mechanism in OvsM/OvoM, as the imidazole π-nitrogen nucleophilic center is positioned at a distance and orientation optimal for interaction with the electron-deficient methyl group of SAM. Leveraging these structural insights, we sought to determine whether the enzymes employ base-mediated catalysis, and, if so, to identify the active site base involved. Analysis of the residues proximal to the substrates 5-SeHis and 5-thioHis point to D149/153 and R150/154 as potential candidates.

To assess the role of D149 as the catalytic base in OvsM_Mko, we mutated it to Asn, attempting to preserve its capacity to H-bond with 5-SeHis while eliminating the negative charge. Subsequent activity assays demonstrated a near-complete loss of enzymatic activity for the D149N variant (Fig. 6A). Unexpectedly, DSC measurements indicate that this mutation also abolished binding of 5-SeHis (Fig. 6B). We then investigated the catalytic involvement of R150 by generating two additional OvsM_Mko variants: R150A and R150Q. Both substitutions resulted in total loss of activity, and the R150Q variant was unable to bind 5-SeHis (Fig. 6A, B). Although we could not decouple the binding and catalytic functions of these residues, based on the positioning of D149/153 in the OvsM_Mko/OvoM_Scu active site relative to the histidine-derived substrate (Fig. 3), we hypothesize that this residue deprotonates the imidazole τ-nitrogen to initiate nucleophilic attack on the methyl group of SAM by the imidazole π-nitrogen (Fig. 6C). By contrast, we suspect that R150/154 fulfills a crucial role in binding 5-SeHis/5-thioHis, rather than in catalysis, by forming multiple H-bonds and contributing to shaping the substrate-binding cleft (Fig. 4A, E).

Figure 6. Proposed mechanism of OvsM/OvoM-catalyzed imidazole π-nitrogen methylation.

(A) Catalytic activity of wild-type and mutant OvsM_Mko, measured via percent conversion of 5-SeHis. Error bars represent standard deviation of three replicates. (B) Thermal denaturation analysis of wild-type and mutant OvsM_Mko using DSC. The protein melting temperature is unaffected by the addition of 5-SeHis in both the D149N and R150Q mutants, indicating that 5-SeHis binding is abolished in these variants. (C) S_N2 methyl transfer from SAM to 5-SeHis/5-thioHis. In this model, D149/153 in OvsM_Mko/OvoM_Scu serves as the catalytic base to deprotonate the imidazole τ-nitrogen, initiating nucleophilic attack by the imidazole π-nitrogen on the methyl group of SAM and cleavage of the C–S bond to yield ovoselenol/ovothiol and SAH. Dashed lines represent H-bonds between the substrate and R150/154.

Beyond TSI-biosynthetic methyltransferases and CARNMT1, the histidine Nπ-methyltransferase METTL9 represents another structurally characterized enzyme of this type.^23,24 CARNMT1 and METTL9 demonstrate nearly undetectable sequence homology with OvsM/OvoM/OvoA, yet they both feature a similarly positioned active site Asp residue implicated in imidazole deprotonation (Fig. S10).^18,23 By contrast, METTL9 hosts an Arg residue (R214) analogous to that providing critical interactions with 5-SeHis/5-thioHis in OvsM_Mko/OvoM_Scu (R150/154), while CARNMT1 employs a His residue (H347) for substrate binding. These observations suggest that while conservation of the catalytic Asp residue is essential for imidazole Nπ-methylation, the substrate-binding residues exhibit flexibility, allowing for variation in their identity.

CONCLUSIONS

In this study, we elucidate ligand recognition and reactivity in the Nπ-methyltransferases that catalyze the final step in the biosynthesis of the antioxidants ovothiol and ovoselenol. Through bioinformatic analysis, we uncovered a new subfamily of monofunctional Nπ-methyltransferases involved in biosynthesizing ovothiol, which we have termed OvoM. Subsequent phylogenetic analysis of the three TSI-biosynthetic Nπ-methyltransferase classes suggests that OvoM and OvsM arose from separate evolutionary events involving splitting of the bifunctional ovoA gene. Our structural characterization of OvoM_Scu and OvsM_Mko reveal that they belong to the Rossmann-fold class of SAM-dependent methyltransferases and exhibit close structural homology to the methyltransferase domain of the bifunctional, ovothiol-biosynthetic enzyme OvoA_Th2. The 5-SeHis/5-thioHis binding contacts are particularly well conserved among these enzymes, which is consistent with our results indicating that they are functionally redundant.

Analysis of ligand-binding contacts reveals an orchestrated series of substrate-induced conformational movements, the most dramatic of which involves reorientation of the N-terminus upon SAM binding to prepare the active site for the histidine-derived substrate. Further examination of the substrate-binding residues offers mechanistic insights into the S_N2-like nucleophilic substitution reaction employed by these enzymes. Analysis of the active site geometry and activity assays on OvsM_Mko variants support a mechanism in which D149/153 (in OvsM_Mko/OvoM_Scu) serves as the catalytic base for imidazole deprotonation. However, this residue appears to be vital for substrate binding, and these two functionalities could not be fully uncoupled.

The structures reported herein also suggest that methyltransferase activity by OvoA_Th2 necessitates significant conformational rearrangement of the NHISS and methyltransferase domains relative to the published crystal structure. In our proposed model, binding of SAM to the methyltransferase domain triggers repositioning of the two domains. The associated reorientation of a key Tyr residue (Y467) located in the linker connecting the two domains facilitates 5-thioHis recognition. Mutagenesis studies support this hypothesis and validate the residue’s essential role in methyltransferase activity, while motivating further investigation into the unique functional dynamics of dual-domain OvoA enzymes. Collectively, our findings unveil the molecular basis for Nπ-specific methylation in the ovoselenol and ovothiol biosynthetic pathways, representing a significant advancement in understanding how nature produces these important antioxidants. Furthermore, our identification of a new subfamily of standalone ovothiol-biosynthetic Nπ-methyltransferases, six years after the complete in vitro reconstitution of the ovothiol pathway,¹⁰ highlights the underlying complexity of TSI biosynthesis and suggests that additional aspects may still await discovery.

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Katherine M. Davis (katherine.davis@emory.edu).

Materials availability

Protein expression plasmids are available upon request. All unique/stable reagents generated in this study are available from the lead contact with a completed Materials Transfer Agreement.

Data and code availability

The atomic coordinates and structure factors for the OvsM_Mko crystal structures have been deposited in the Protein Data Bank under accession numbers 9BXH (apo), 9BXJ (in complex with SAM), and 9BXK (in complex with 5-SeHis and SAH). Atomic coordinates and structure factors for the OvoM_Scu crystal structures have been deposited in the Protein Data Bank under accession codes 9BXM (apo), 9BXL (in complex with SAM), and 9BXN (in complex with 5-thioHis and SAH). All atomic coordinates and structure factors are publicly available as of the date of publication. Accession numbers are also listed in the key resources table. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Bacterial and virus strains
Escherichia coli DH5α	New England Biolabs	Cat# C2987I
Escherichia coli BL21 (DE3)	New England Biolabs	Cat# C2527I
Chemicals, peptides, and recombinant proteins
NdeI restriction endonuclease	New England Biolabs	Cat# R0111S
XhoI restriction endonuclease	New England Biolabs	Cat# R0146S
HiFi DNA Assembly Master Mix	New England Biolabs	Cat# E2621L
His-tagged-OvsMMko	This study	N/A
His-tagged-OvoMScu	This study	N/A
His-tagged-OvoATh2	This study	N/A
His-tagged-OvsMMko variants	This study	N/A
His-tagged-OvoATh2 variants	This study	N/A
5-thiohistidine	This study	N/A
5-selenohistidine	Kayrouz et al.7	N/A
S-adenosylmethionine	Sigma-Aldrich	Cat# A7007
S-adenosylhomocysteine	Sigma-Aldrich	Cat# A9384
Monobromobimane	MilliporeSigma	Cat# 596105
Tacsimate pH 7	Hampton Research	Cat# HR2-755
PEG monomethyl ether 5000	Sigma-Aldrich	Cat# 81323
PEG 6000	Fisher Scientific	Cat# AC192280010
PEG 400	Sigma-Aldrich	Cat# PX1286B-2
Deposited data
OvsMMko crystal structure	This study	PDB: 9BXH
OvsMMko■SAM crystal structure	This study	PDB: 9BXJ
OvsMMko■5-SeHis■SAH crystal structure	This study	PDB: 9BXK
OvoMScu crystal structure	This study	PDB: 9BXM
OvoMScu■SAM crystal structure	This study	PDB: 9BXL
OvoMScu■5-thioHis■SAH crystal structure	This study	PDB: 9BXN
OvsA crystal structure	Kayrouz et al.7	PDB: 8U42
OvoATh2•Co•histidine•cysteine crystal structure	Wang et al.12	PDB: 8KHQ
CARNMT1■;SAM crystal structure	Cao et al.18	PDB: 5YF0
Recombinant DNA
pET28b vector	MilliporeSigma	Cat# 69865
pET28b-His6x-OvsMMko	This study	N/A
pET28b-His6x-OvoMScu	This study	N/A
pET28b-His6x-OvoATh2	This study	N/A
pET28b-His6x-OvsMMko variants	This study	N/A
pET28b-His6x-OvoATh2 variants	This study	N/A
Software and algorithms
CD-Hit	Fu et al.25	https://sites.google.com/view/cd-hit
MAFFT	Katoh et al.26	https://mafft.cbrc.jp/alignment/software/
FastTree	Price et al.27	http://www.microbesonline.org/fasttree/
FigTree	Rambaut Lab, University of Edinburgh	http://tree.bio.ed.ac.uk/software/figtree
Logomaker	Tareen and Kinney28	http://github.com/jbkinney/logomaker
MassHunter Software Suite	Agilent	https://www.agilent.com/en/product/software-informatics/mass-spectrometry-software
MestReNova	Mestrelab Research	https://mestrelab.com/download/mnova/
XDS	Kabsch30	https://xds.mr.mpg.de
AIMLESS	Evans and Murshudov31	https://www.ccp4.ac.uk/html/aimless.html
AlphaFold2	Jumper et al.32	https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb
Coot	Emsley et al.33	https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
Phenix	Adams et al.34	https://phenix-online.org/
Molprobity	Williams et al.35	http://molprobity.biochem.duke.edu/
PyMOL	Schrodinger, LLC36	http://www.pymol.org
LigPlot+	Laskowski and Swindells37	https://www.ebi.ac.uk/thornton-srv/software/LigPlus/
Other
HisPur Ni-NTA resin	ThermoFisher	Cat# 88222
Sephadex G-25 medium resin	Cytiva	Cat# 17003301
HiPrep 16/60 Sephacryl S-200 HR column	Cytiva	Cat# 17116601
Cary 60 UV-Visible Spectrophotometer	Agilent	Cat# G6860A
MicroCal PEAQ-DSC	Malvern Panalytical	N/A

STAR Methods

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Bacterial strains

Escherichia coli (E. coli) BL21 (DE3) cells were used for producing the recombinant proteins in this study. The cells were cultured in standard media (LB broth) at 37 °C. The method details section gives more comprehensive details about protein expression and purification.

METHOD DETAILS

Materials

All materials were purchased from Millipore-Sigma or Fisher Scientific unless otherwise specified. Cloning reagents were purchased from New England BioLabs (NEB). Codon-optimized gene fragments were purchased from Twist Biosciences. Synthesis of 5-selenohistidine (5-SeHis) was carried out as previously described.⁷

General Experimental Procedures

High-performance liquid chromatography-coupled mass spectrometry (HPLC-MS) was performed on an Agilent instrument equipped with a 1260 Infinity Series HPLC, an automated liquid sampler, a JetStream ESI source, and the 6540 Series Q-tof mass spectrometer. HPLC-MS data were acquired and analyzed with Agilent MassHunter software. HPLC purifications were carried on an Agilent 1260 Infinity Series HPLC system equipped with a photodiode array detector and an automated fraction collector. Solvents for all LC-MS and HPLC experiments were water + 0.1% formic acid (Solvent A) and MeCN + 0.1% formic acid (Solvent B). HPLC data were acquired and analyzed with Agilent OpenLab software. Nuclear magnetic resonance (NMR) spectra were collected at the Princeton University Department of Chemistry’s NMR facility on a Bruker Avance III 500 MHz NMR spectrometer equipped with a DCH double resonance cryoprobe. NMR data were acquired using Bruker Topspin software and processed/analyzed using MestReNova software.

Bioinformatics

TSI-biosynthetic methyltransferases (OvsM, OvoM, and OvoA C-terminus) were retrieved from the NCBI database by performing BLASTp searches with OvsM_Mko as a seed sequence and an E-value cutoff of 1e-50. Sequences with an N-terminal NHISS domain were classified as OvoA (4,434 total). For all further sequence analyses, the NHISS domains were manually excised, leaving only the C-terminal methyltransferase domains. Standalone methyltransferases were classified as OvsM if the strain also contains an ovsA gene and no gene encoding a standalone OvoA NHISS domain (235 total). Sequences were classified as OvoM if the strain harboring it lacks an ovsA gene and instead encodes a standalone OvoA NHISS domain (22 total). All-by-all BLASTp searches were performed to generate pairwise percent identity statistics among the three groups.

For construction of phylogenetic trees, first the CD-hit tool²⁵ was used to cluster all OvoA C-terminus sequences at 55% identity and all OvsM and OvoM sequences at 75% identity to improve visual readability. Clustering resulted in 42, 24, and 4 diverse OvoA C-terminus, OvsM, and OvoM sequences, respectively. Four additional N-heterocycle methyltransferases (BorM1, CyaF, HrsC, and MitM) were added to the list of sequences to serve as an outgroup and illustrate their distant homology to TSI-biosynthetic methyltransferases. Trees were then generated by first performing multiple sequence alignments using MAFFT²⁶ with the L-INS-I algorithm, which were then used as input to FastTree²⁷ for maximum-likelihood phylogeny approximation with parameters -gamma, -mlacc 2, and -slownii. Trees were visualized using FigTree (http://tree.bio.ed.ac.uk/software/figtree). The bifurcation between the TSI methyltransferases and the outgroup was selected to be the tree root. Logo alignment diagrams were generated using the Logomaker python script.²⁸

OvsM_Mko Protein Sequence

MGSSHHHHHHSSGLVPRGSHMTDNPYETDELLGQYLDFHYGPGHFDVPNYPKACIEQALAHHTGTTGRALDLGCAVGRSSFELARRFDEVIGIDLSRRFIDSATRLAEQGQLQYQVTLEGELIERRTADLAALELSNTAGRTRFQVGDACALDDTLGRFDLIFAGNLIDRLPDPAAFLAQLPALVRPGGLLMITSPYTLLPEFTPRERWIGGFERNGQPVRMLDGLRHHLEPDFVLLEPTRDIPFVIRETTRKYQHTVAEASLWRRA

OvoM_Scu Protein Sequence

MGSSHHHHHHSSGLVPRGSHMIFKEKNIYETDESVSQYCDFQYGEDCFGVLNFALACATKAIGYTKETPKNRALDLGCATGRASFELARSYKHVDGVDYSQAFVDAATELQKNGCIGYSQNGEGELKNYKVIDREGYAFRDSFTKVEFFQGDACNLTPQFKEYDLIMATNLIDRLYEPRLFLENIHKRINEKGYLILTSPYTWREEYTAKKFWIGGYVDENGKEVSTLEGLKEILEIHFELVATEDIPFVIRETSRKFQHTISQMSVWKVIST

OvoA_Th2 Protein Sequence

MGSSHHHHHHSSGLVPRGSHMGDRGPEFNKILLRPLLLKQKNPENLRQLIKKSFHRTFDTFESLFSMLRNDEAFYNRPEPLRHPHIFYFGHTAVFFINKLILSKIIDTRINAKMESIFAIGVDEMSWDDLNDDHYEWPSVEETRLYRNRVREVVDNLINTLPLELPITWDSPWWIILMGIEHERIHIETSSVLIRQTDISLVLPQPEWSKCNVSGKAPENELLFVPGGEIEIGKYKSDDYYGWDNEYGKHKTVIPDFKASKYLVSNGEFMEFVKDGGYENDLWWEEEGLAWRNFKKAKHPIFWIPFKNEYRYRTLTEIVDMPLDWPVDVNYHEAKAFCNWLSAKKGKPIRLPVEDEWYRLKEYCNVPDVSKWDEKAPANINLEHYASACPVTQFSFGNFYDVIGNVWQWTETPIYPFNGFKIHPIYDDFSTPTFDNRHNLIKGGSFISTGNEILASSRYAFRRHFFQHAGFRYVESSYKEKINSSGYESDTQVSQYCEFGWGDRYFGIENYPKRCAKICIEVTEGKPRKKALDVGCAIGRSTLELATSFESVTGLDFSARFIEMAERMRKDGSIRYTITTEGELVEYKEATLPKRLAKVVDRVEFWQADACNLKPIFTGYDLVFAGNLIDRLYDPAKFLNDIGKRINSGGMLILTSPYTWLEEFTPKQKWLGGFKQDGEPVKSIDGLKSHLKDSFKLIETRDIEFVIRETARKFQHSVAQMSIWEKILEVDLQGDHGLSAWSHPQFEK

Synthesis of 5-thiohistidine

The chemical synthesis of 5-thiohistidine (5-thioHis) was carried out according to previous methods,²⁹ with slight modifications. First a 25 mL aqueous solution of histidine (3.17 mmol, 1 equiv.) and HCl (3.17 mmol, 1 equiv.) was cooled to 5 °C under stirring, after which N-bromosuccinimide (4.12 mmol, 1.3 equiv.) was added. After 2.5 min, thioacetic acid was added (12.68 mmol, 4 equiv.) and the reaction was left to stir for 1 hr at 5 °C. Next, 3-mercaptopropionic acid (19.02 mmol, 6 equiv.) was added and the reaction was allowed to stir under reflux at 100 °C for 16 hours. The reaction was then cooled to room temperature and the precipitate was removed by vacuum filtration. The filtrate was fractionated on a Luna Omega Polar C18 column (Phenomenex, 150 x 21.2 mm, 5 μm) with a flow rate of 15 mL/min and 100% Solvent A. Fractions containing 5-thioHis were further purified by fractionation on an XBridge BEH Amide OBD column (Waters, 10 x 250 mm, 5 μm) with a flow rate of 4 mL/min and 60% Solvent B. Pure 5-thioHis, in disulfide form, was lyophilized and stored at −20 °C until further use.

¹H NMR (500 MHz, D₂O): δ3.04 (br, 2H), 3.93 (br, 1H), 8.11 (s, 1H).

Protein Expression and Purification

Genes encoding OvsM_Mko, OvsM_Mko variants, OvoM_Scu, OvoA_Th2, and OvoA_Th2 variants were obtained as synthetic DNA fragments, codon-optimized for expression in Escherichia coli and with overhangs to allow for assembly into pET28b(+). Gene fragments were assembled into pET28b(+) using HiFi DNA Assembly Master Mix (NEB) following the manufacturer’s instructions. Assembled plasmids were transformed into chemically competent E. coli BL21(DE3) for protein expression, which was carried out in 2.8 L baffled Fernbach flasks containing 1.5 L LB broth supplemented with 50 mg/L kanamycin. Production cultures were incubated at 37 °C and 200 RPM until an OD₆₀₀ of 0.5-0.6 was reached, after which protein expression was induced with 0.2 mM isopropyl β-d-1-thiogalactopyranoside for an additional 12 to 24 h at 18 °C. Cells were pelleted by centrifugation and stored at −80 °C until purification.

All purification steps were carried out in a cold room at 4 °C. Cells were resuspended in lysis buffer (5 mL/g cell paste), which consisted of 25 mM Tris-HCl (pH 8), 500 mM NaCl, 10 mM imidazole, 10% glycerol, supplemented with 1 mM phenylmethylsulfonyl fluoride. Once homogenous, 0.1 mg/mL deoxyribonuclease I (Alfa Aesar) was added, and the cells were lysed by the addition of 5 mg/mL lysozyme followed by sonication using 30% power (^~150 W) in 15 s on/15 s off cycles for a total of 4 min. This process was repeated three times. The lysate was then clarified by centrifugation (17,000 g, 15 min, 4 °C) and loaded onto a 7 mL Ni-NTA column pre-equilibrated in lysis buffer. The column was washed with lysis buffer and His-tagged proteins were e luted with elution buffer consisting of 25 mM Tris-HCl (pH 8), 500 mM NaCl, 300 mM imidazole, 10% glycerol. Eluted proteins were then buffer-exchanged using a 75 mL column of Sephadex G-25 (Cytiva) into storage buffer consisting of 50 mM Tris-HCl (pH 8), 150 mM NaCl. For crystallography, OvsM_Mko and OvoM_Scu were further purified by fractionation on a Sephacryl S-200 HR HiPrep 16/60 column (Cytiva) with a running buffer consisting of 50 mM Tris-HCl (pH 8), 150 mM NaCl. Purified proteins were stored at −80 °C. Protein concentrations were determined spectrophotometrically on a Cary 60 UV-visible spectrophotometer (Agilent) using calculated molar extinction coefficients at 280 nm. Enzymes were stored in 50 mM Tris (pH 8), 150 mM NaCl for crystallization.

Enzyme Activity Assays

Assays were carried out in triplicate in 150 μL reactions containing 50 mM sodium phosphate buffer (pH 8), 4 mM dithiothreitol (DTT), 2 mM SAM, and 1 mM histidine-derived substrate (5-thioHis or 5-SeHis, monomer basis). Reactions were initiated by addition of the methyltransferase (OvsM_Mko, OvoM_Scu, OvoA_Th2, or variant thereof) to a final concentration of 10 μM and allowed to proceed for 1 hr at room temperature, after which the enzyme was immediately removed by centrifugal spin filter (3 kDa MWCO for OvsM/OvoM or 10 kDa MWCO for OvoA). Reactions were then mixed with an equal volume of 10 mM mBBr in MeCN for 30 minutes at room temperature to facilitate complete derivatization. The samples were then filtered and analyzed by HPLC-MS using a Synergi Fusion-RP C18 column (Phenomenex, 250 x 4.6 mm, 4 μm) with a flow rate of 1.5 mL/min and 8% Solvent B for 8 min. Conversion was determined by integrating extracted ion chromatogram peak areas of mBBr-derivatized substrates and products.

Thermal Denaturation Assays Using Differential Scanning Calorimetry (DSC)

Assays were carried out using 250 mL samples containing 3 mg/mL of OvsM_Mko and its D149N and R150Q variants (100 mM) in buffer consisting of 50 mM Tris-HCl (pH 8), 150 mM NaCl. SAH and 5-SeHis were added to final concentrations of 800 μM (8 equiv.) and 1.6 mM (16 equiv., monomer basis), respectively, when present. DTT was added to each sample to create a final concentration of 4 mM. Samples were incubated at room temperature for 40 minutes to facilitate binding. Measurements were carried out using a Malvern Panalytical semi-automated MicroCal PEAQ-DSC from a temperature range of 25 °C to 100 °C with a scanning rate of 180 °C/hr using a 250 mL reference cell of 50 mM Tris-HCl (pH 8), 150 mM NaCl, 4 mM DTT buffer. Data were analyzed through Malvern Panalytical MicroCal PEAQ-DSC Analysis Software ver. 15.1 using a sample of DTT-containing buffer as a blank and the Line baseline parameters.

Crystallization of OvsM_Mko

Crystals of OvsM_Mko were grown using the sitting drop vapor diffusion method at room temperature. In the absence of substrate, OvsM_Mko (12 mg/mL) was mixed 1:1 with a precipitant solution of 0.1 M HEPES (pH 7), 5% (v/v) tacsimate (pH 7), 10% (w/v) polyethylene glycol (PEG) monomethyl ether 5,000. To obtain crystals of the enzyme complexed with SAM, 5 mM SAM was added to 14.8 mg/mL OvsM_Mko before mixing 1:1 with a precipitant solution of 0.1 M BICINE (pH 9), 10% (w/v) PEG 6,000. To obtain crystals of the enzyme complexed with 5-SeHis and SAH, substrate-free crystals were first grown by mixing OvsM_Mko (10 mg/mL) 1:1 with a precipitant solution of 0.1 M Tris-HCl (pH 8.5), 0.2 M tri-sodium citrate, 30% (v/v) PEG 400. The resulting crystals were subsequently soaked with 30 mM SAH for 4 h, then soaked with 30 mM 5-SeHis and 60 mM DTT for 6.5 h before looping. Thin plates appeared within 2 days for all conditions and were fully formed within 1 week. For harvesting, crystals were looped and briefly transferred into cryoprotectant before flash freezing in liquid N₂. Cryoprotectants were comprised of the precipitant and 25-30% (v/v) ethylene glycol, along with any presumed ligands (18 mM SAM for the binary complex; 12 mM 5-SeHis, 12 mM SAH, and 40 mM DTT for the ternary complex).

Crystallization of OvoM_Scu

Crystals of OvoM_Scu were grown using the sitting drop vapor diffusion method at room temperature, using 15 mg/mL OvoM_Scu for all crystallization trials. In the absence of substrate, OvoM_Scu was mixed 1:1 with a precipitant solution of 0.1 M citrate (pH 5), 19-24% (v/v) PEG 6000. To obtain crystals of the enzyme complexed with SAM, 5 mM SAM was added to the protein, before mixing 1:1 with a precipitant solution of 0.1 M citrate (pH 5), 20% (v/v) PEG 6000. To obtain crystals of the enzyme complexed with 5-thioHis and SAH, substrate-free crystals were soaked with 30 mM SAH for 3 h, then soaked with 30 mM 5-thioHis and 60 mM DTT for 6.5 h before looping. Small, fully formed plates appeared after 1-7 weeks for all conditions. For harvesting, crystals were looped and briefly transferred into cryoprotectant before flash freezing in liquid N₂. Cryoprotectants were comprised of the precipitant and 25-30% (v/v) ethylene glycol, along with any presumed ligands (20 mM SAM for the binary complex; 12 mM 5-thioHis, 12 mM SAH, and 40 mM DTT for the ternary complex).

X-ray Data Collection and Processing

Crystals were maintained at 100 K to minimize X-ray-induced damage during data collection. All diffraction data for the OvsM_Mko structures were collected sequentially (Δϕ = 0.25°) with an incident wavelength of 0.9686 Å at beamline ID7B2 of the Cornell High Energy Synchrotron Source using an Eiger2 16M (Dectris) detector. Diffraction data for the apo-OvoM_Scu structure were collected sequentially (Δϕ = 0.25°) with an incident wavelength of 1.0 Å at beamline 8.2.1 of the Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory using an Eiger2 9M (Dectris) detector. Diffraction data for the OvoM_Scu binary complex were collected sequentially (Δϕ = 0.25°) with an incident wavelength of 0.9795 Å at beamline 8.2.2 of the ALS using a Pilatus3 2M detector. Diffraction data for the OvoM_Scu ternary complex were collected sequentially (Δφ = 0.2°) with an incident wavelength of 0.9530 Å at the CMCF beamline 08ID at the Canadian Light Source using an Eiger X 9M detector.

All data were indexed, integrated, and scaled with XDS before merging with AIMLESS.^30,31 Structures of substrate-free OvsM_Mko and SAM-bound OvoM_Scu were first solved via molecular replacement with PHASER, using the AlphaFold2-predicted structure as the search model after removing all residues with per-residue confidence scores (pLDDT) below 90%.³² All subsequent structures were solved via isomorphous replacement after transferring and truncating the R_free flags from the apo-OvsM_Mko and SAM-bound OvoM_Scu structures. Model building was conducted in Coot,³³ the structures refined in Phenix,³⁴ and model quality assessed using Molprobity.³⁵ Data processing and refinement statistics can be found in Table 1 for the OvsM_Mko structures and Table 2 for the OvoM_Scu structures. All OvsM_Mko crystals were in the orthorhombic C 2 2 2₁ space group and contain one molecule in the asymmetric unit, while the OvoM_Scu crystals were in the orthorhombic P 2₁ 2₁ 2₁ space group and contain two molecules in the asymmetric unit. Figures depicting the structures were generated with PyMOL³⁶ and visualize chain A from all structures, unless otherwise stated. 2-D interaction diagrams were generated with LigPlot+.³⁷

QUANTIFICATION AND STATISTICAL ANALYSIS

Data processing and refinement statistics for structural data reported in this article are summarized in Table 1 and Table 2.

Highlights.

• Bioinformatic analysis uncovers monofunctional ovothiol Nπ-methyltransferases

• OvsM and OvoM feature conserved ligand-binding motifs and conformational dynamics

• Activity in dual-domain bifunctional OvoA enzymes necessitates domain rearrangement

• Methyl transfer appears to rely on an active site Asp base and an S_N2-like reaction