**1. Introduction**

Methyltransferases represent a ubiquitous class of enzymes that methylate a vast array of small molecules and macromolecules and participate in numerous biological processes, including metabolism, signal transduction, and gene expression [1–3]. The majority of these enzymes utilize the methyl donor S-adenosylmethionine (AdoMet) whose methyl group is rendered highly reactive through its bonding to a sulfonium cation in the substrate. AdoMet-dependent methyltransferases catalyze an SN2 reaction wherein a nucleophilic atom, such as oxygen, nitrogen, or sulfur, attacks the electrophilic methyl carbon atom of AdoMet, with the sulfur atom displaced as the leaving group [4]. The reaction mechanism of these enzymes has been a subject of intense study for over 40 years [5] and has led to the proposal of several different models for catalysis. These models include (1) compression or compaction of nucleophile, electrophile, and leaving groups along the reaction coordinate [6–9], (2) formation of near attack conformers (NACs) that align the nucleophile and methyl group in a productive geometry for the SN2 reaction [10–13], (3) electrostatic pre-organization within

the active site that promotes methyl transfer [14,15], and (4) cratic effects involving the free energy of association of the substrates in a catalytically favorable alignment within the active site [16,17]. Despite these models, the methyltransferase mechanism remains a topic of active debate.

Recent structure–function studies of methyltransferases have explored the interactions between their active sites and the AdoMet sulfonium cation. A survey of high-resolution crystal structures of methyltransferases in the Protein Databank (PDB) identified unconventional carbon–oxygen (CH···O) hydrogen bonds between the AdoMet methyl group and oxygen atoms within the active sites of different classes of these enzymes [18]. Quantum mechanical (QM) calculations demonstrated that the AdoMet methyl group forms relatively strong CH···O hydrogen bonds due to its polarization by the neighboring sulfonium cation [18–20]. Correlatively, structural and biochemical characterization of the protein lysine N-methyltransferase (KMT) SET7/9 and the reactivation domain of methionine synthase demonstrated that these hydrogen bonds promote high affinity binding to AdoMet compared to the methyl transfer product S-adenosylhomocysteine (AdoHcy), thus mitigating product inhibition [18,21]. Moreover, CH···O and CH···N interactions with the AdoMet methyl group have been proposed to contribute to transition state stabilization in several methyltransferases, including SET7/9, SET8, NSD2, and glycine N-methyltransferase [18,22–24].

In addition to unconventional hydrogen bonding, chalcogen bonding between the AdoMet sulfur cation and the active sites of methyltransferases has also been observed [25]. A chalcogen bond is defined as a non-bonded interaction wherein a Lewis base donates a lone pair of electrons into the sigma antibonding (σ\*) orbital of an atom from the Group VI elements (oxygen group) of the periodic table [26]. Structural and functional characterization of an S···O chalcogen bond between AdoMet and an asparagine residue in the active site of SET7/9 demonstrated that this interaction enhances the binding affinity for the substrate relative to AdoHcy and modestly augments the rate of methyl transfer [25]. Together, these results illustrate that carbon hydrogen bonding and sulfur chalcogen bonding between the AdoMet sulfonium cation and residues in the methyltransferase active site can enhance the enzyme's binding affinity for the substrate and promote the methyl transfer reaction.

Beyond hydrogen bonding and chalcogen bonding, there is a third unconventional interaction that can occur with sulfonium cations involving a σ\* orbital of a carbon atom [27]. This interaction is termed a tetrel bond and occurs when an atom from the Group IV elements (carbon group) of the periodic table accepts a lone pair of electrons from an electronegative atom [28,29]. In the case of AdoMet, this interaction can occur with the σ\* orbital of the methyl carbon atom that corresponds to the sulfur–carbon (S–CH3) bond.

Although aliphatic carbon atoms typically form weak tetrel bonds compared to other Group IV elements, QM calculations have demonstrated that a methyl carbon atom bonded to a sulfonium ion can form relatively strong tetrel interactions due to polarization by the adjacent cation [27]. Notably, the geometry of the tetrel bond, in which the interaction angle between the Lewis base (X) and S–CH3 bond is approximately linear, precludes strong methyl CH···X hydrogen bonding due to acute hydrogen bond angles [27,30]. Experimental evidence for carbon tetrel interactions first emerged from a survey of the Cambridge Structural Database, which identified over 700 small molecule crystal structures displaying C···O tetrel bonds, including multiple interactions involving methyl groups [31]. In addition, recent studies by Frontera and colleagues have revealed crystallographic evidence of methyl and trifluoromethyl C···O tetrel bonding between proteins and various ligands [30,32]. Pertinent to methyltransferases, a computational analysis by Grabowski directly implicated tetrel bonding between an electrophilic tetrel atom and a nucleophile as a preliminary state that precedes the transition state in SN2 reactions, including methyl transfer [33]. Collectively, these findings prompted us to examine structures of AdoMet-dependent methyltransferases to ascertain whether methyl tetrel bonding occurs in these enzymes. The results of our structural survey coupled with corroborative QM calculations demonstrate the existence of the tetrel bonding in methyltransferases, furnishing insights into the potential roles of these interactions in ligand binding and SN2 catalysis.

#### **2. Material and Methods**

#### *2.1. PDB Survey*

Crystal structures of methyltransferase/AdoMet complexes with a resolution of ≤2.50 Å were downloaded from the PDB and visually examined for the presence of carbon tetrel bonding to the AdoMet methyl group. Tetrel bonds between the AdoMet methyl group and an electronegative atom (X) of a small molecule inhibitor, solvent molecule, or ion were defined as exhibiting: (1) an θ(S–C···X) interaction angle between 160◦ and 180◦ (where S and C are the sulfur and methyl carbon atoms of AdoMet, respectively) and (2) a C···X interaction distance less than or equal to sum of the van der Waals radii of the carbon and electronegative atoms, specifically R(C···O) ≤3.25 Å and R(C···Cl) ≤3.5 Å (carbon, oxygen, and chlorine van der Waals radii were defined as 1.75 Å, 1.5 Å, and 1.75 Å, respectively) [34]. These geometric parameters are consistent with the formal definition of halogen bonding, a related category of interactions that are considered an archetype for σ-hole bonding [35]. For crystal structures displaying potential carbon tetrel bonds, the electron density corresponding to AdoMet and the electron donor were visually inspected using the program Coot to confirm the integrity of the model [36,37]. Structures that displayed ambiguous electron density for the ligands were omitted from the survey. For the structure of the DhpI phosphonate *O*-methyltransferase (accession code 3OU6.pdb), the AdoMet molecules were remodeled in the electron density maps using the real space refinement and geometry tools in Coot. The remodeled AdoMet coordinates were then used to measure the tetrel bond geometries (Table 1). Finally, in cases where two or more structures of a given methyltransferase possess the same tetrel bond donor, such as interactions involving water molecules and the COMT/AdoMet/DNC/Mg2+ complexes, only the highest resolution structure of the wild type enzyme is reported in Table 1.


**Table 1.** Crystallographic survey of methyl tetrel bonding in AdoMet-dependent methyltransferases.


**Table 1.** *Cont.*

Note: A: A, B, and C denote the protein chains in the asymmetric unit of the crystal structure; B: (3,4- dihydroxy-2-nitrophenyl)(phenyl)methanone; C: (3,4-dihydroxy-5-nitrophenyl)(4- methylphenyl)methanone; D: 1-(biphenyl-3-yl)-3-hydroxypyridin-4(1H)-one; E: [1-(biphenyl-3-yl)-5-hydroxy-4-oxo-1,4-dihydropyridin-3-yl]boronic acid; F:

N-cyclohexyl-N~3~-[2-(3,4-dichlorophenyl)ethyl]-N-(2-{[2-(5-hydroxy-3-oxo-3,4-dihydro-2H-1,4-benzoxazin-8 yl)ethyl]amino}ethyl)-beta-alaninamide; G: N-[1-(N'-cyano-N-[3-(difluoromethoxy)phenyl]carbamimidoyl)-3- (3,4-dichlorophenyl)-4,5-dihydro-1H-pyrazol-4-yl]-N-ethyl-2-hydroxyacetamide; H: 2-oxidanylidene-N-piperidin-4-yl-1,3-dihydroindole-5-carboxamide; <sup>I</sup> : 6-chloranyl-2-oxidanylidene-N-[(1S,5R)-8- [4-[(phenylmethyl)amino]piperidin-1-yl]sulfonyl-8-azabicyclo[3.2.1]octan-3-yl]-1,3-dihydroindole-5-carboxamide.
