Structural Basis for 3-Amino-3-carboxypropyl Transfer in Nocardicin Biosynthesis

Abstract

S-Adenosyl-l-methionine (SAM) is well-known as a methyl donor for methyltransferases but also functions as a 3-amino-3-carboxypropyl (3-ACP) donor for 3-ACP transferases. NAT is a 3-ACP transferase which accepts β-lactam antibiotic nocardicin G (1) and SAM to produce isonocardicin C. Due to the lack of structural information about this enzyme, its reaction mechanism has not been fully identified. In this study, we report two X-ray crystal structures of NAT, including its apo and complex structure with 1 and SAH. Examination of them identified the structural basis for substrate recognition. Comprehensive approach integrating site-directed mutagenesis, thermal shift assay, MD simulation, and QM/MM calculation revealed that the Cα-amino group of SAM functions as a Brønsted base to enhance the nucleophilicity of the C6′-OH of 1, with the assistance of E143, thereby facilitating S_N2 attack on the Cγ of SAM. This study presents structural and computational analysis leading to more precise understanding of 3-ACP transfer.

Graphical Abstract

INTRODUCTION

S-Adenosyl-l-methionine (SAM) is a sulfonium primary metabolite found in all living organisms.¹ SAM is best known for a role it serves as a methyl donor in many biological reactions. Three million protein sequences have been annotated as SAM-dependent methyltransferases (MTs) in the InterPro database.² Most of them catalyze a methyl-transfer reaction by facilitating a nucleophilic attack at the methyl group of SAM, releasing S-adenosylhomocysteine (SAH) as a leaving group (Figure 1a). On the other hand, a part of them catalyze transfer of a 3-amino-3-carboxypropyl (3-ACP) group to an acceptor substrate by facilitating a nucleophilic attack at the Cγ methylene of SAM, releasing 5′-methylthioadenosine (MTA) as a leaving group (Figure 1a).³ The regioselectivity of the enzyme in regulating the relative positions of the alkyl donor and acceptor is of interest in determining the regioselectivity of the methyl group transfer and the 3-ACP transfer. The skeletal modification by 3-ACP transfer plays an important role in the bioactivity of the natural products or RNA. For example, YfiP transfers 3-ACP to position 47 in the variable loop of Escherichia coli tRNAs, improving their thermostability and maintaining canonical function.^4–6 In plants, nicotianamine synthase (NAS) cyclizes one molecule of SAM to azetidine and condenses two 3-ACP group to the azetidine to produce a siderophore molecule.^7–9 This molecule plays an essential role in the maintenance of metal homeostasis. In the biosynthesis of the antibiotic microcin C, MccD catalyzes 3-ACP transfer on the phosphate of the phosphoramide linkage of the intermediate McC¹¹²⁰ which is followed by decarboxylation by the PLP-dependent enzyme MccE¹⁰ to produce the aminopropyl-linked phosphate moiety of microcin C. The activity of microcin C is 10-fold higher than that of McC¹¹²⁰,¹¹ suggesting that 3-ACP transfer is important from the perspective of enzymatic modifications of antibiotics.

Figure 1.

Reactions of SAM-dependent transfer. (a) Methyltransfer catalyzed by SAM-dependent methyltransferase and 3-ACP transfer by 3-ACP transferase. (b) 3-ACP group transfer reaction catalyzed by NAT.^12–15 NocJ is an isomerase that acts on the amino group of 3-ACP, and NocL is an oxidase that acts on the amino group of p-hydroxyphenylglycine.

Nocardicin A is one of the most bioactive monocyclic β-lactam antibiotics isolated from Actinomycete Nocardia uniformis subsp. tsuyamanesis.¹⁶ In the biosynthesis of nocardicin A, SAM:nocardicin 3-ACP transferase (NAT) catalyzes the transfer of 3-ACP to the C6′-OH of nocardicin G (1) or nocardicin E (2), to produce isonocardicin C (3) or isonocardicin A, respectively (Figure 1b). The NH₂ group of the 3-ACP moiety of 3 is epimerized by NocJ, and the 2′-NH₂ group of NocJ product is oxidized by NocL to nocardicin A.^17–20 While 1 is inactive as an antibiotic, nocardicin A has antimicrobial activity, suggesting that the 3-ACP moiety is essential for its antibiotic activity.^16,18 By Townsend’s group, NAT was purified to homogeneity by SAH and nocardicin A affinity columns, and the transfer reaction was characterized biochemically. A kinetic analysis of NAT was carried out, and it was shown to catalyze the reaction in an ordered sequential Bi–Bi manner in which the 3-ACP acceptor firstly binds to the enzyme, SAM secondly binds to the enzyme–substrate complex, and the product is finally released from the enzyme.¹⁸ While a wealth of information is available on the biochemical properties of NAT, the lack of structural information has limited our understanding of its substrate recognition and reaction mechanism and how these differ from conventional MTs.

Here, we report two X-ray crystal structures of NAT, apo, and the structure complexed with SAH and 1. This is the first structural study of 3-ACP transferase complexed with both SAH and 3-ACP acceptor among the reported X-ray structures, including MtNAS,⁸ TYW2,²¹ CntL,²² MccD,¹⁰ and Tsr3.²³ In addition, MD simulation and QM/MM calculations provided insight into the detailed reaction mechanism of NAT. The Cα-NH₃⁺ of SAM is initially deprotonated by the side chain of E143, generating an NH₂ group that attracts the proton of the C6′-OH of 1. This interaction facilitates the nucleophilic attack of the C6′-oxygen atom of 1 on the Cγ of the SAM. The chemically synthesized deamino-SAM did not react with 1, supporting this hypothetical reaction mechanism. Taken together, these findings reveal the molecular basis and dynamics of the reaction of 3-ACP transferase in nocardicin biosynthesis, in which the amino group of one substrate acts as a base for the hydroxy group of the other substrate.

RESULTS

Comparative Structural Analysis of NAT (apo) and Nocardicin G (1)-SAH Ternary Complex.

NAT has little sequence identity to either reported SAM-dependent O-methyltransferases or other 3-ACP transferases^{3,8,10,21–24} (14.1% identity with MfnG, 7.7% identity with MtNAS, 7.4% identity with TYW2, and 5.8% identity with CntL, Scheme S1, Figure S1). Thus, we utilized the AlphaFold2 model as a template for molecular replacement. Finally, we solved apo (apo structure) and complex structure with 1 and SAH (1+SAH structure) at 1.8 and 2.1 Å, resolution, respectively (Figures 2, S2, and S3 and Table S3).

Figure 2.

Overall structural comparison of apo and complex structure of NAT. Cyan: complex-Cterm. Brown: Apo-Cterm. Light blue: complex-N-term. Orange: apo-N-term. Green: SAH. Blue: 1.

From the structures, we found that NAT forms a dimer, and each monomer has two regions: an N-terminal region consisting of eight β-sheets (β1–β8) and one α-helix for 1 binding (residues 1–125), and a Rossmann fold C-terminal region² consisting of nine β-sheets (β9–β15) for SAM binding (residues 126–301) (Figures 2 and S3). The C-terminal region is well-conserved among known 3-ACP transferases, TYW2,²¹ MtNAS,⁸ and CntL²² (Figures S4 and S5), but the N-terminal region shows little identity among them. This low identity reflects the different structures of the 3-ACP acceptors. A homology search in the DALI²⁵ server indicated that NAT is most similar to the tyrosine O-MT in marformycin biosynthesis (RMSD = 4.8 Å, PDB code 7UX8)²⁶ (Figure S6).

A comparison between the apo and 1+SAH structures revealed positional changes in the C-terminal region, as reported in other 3-ACP transferases, with helix 2 (residues 133–156) in the C-terminal region moving ~3.5 Å toward the N-terminal region (Figure 2). Furthermore, loop 7 between the N-terminal and C-terminal regions (residues 122–133) moves upon the binding of 1 and SAH. The movement of this loop region helps 1 and SAM bind to the enzyme. In addition, loop 4 in the Nterminal region, which was not observed in the apo structure, becomes visible in the 1+SAH structure, indicating the movement of the loop4 region when the enzyme binds to 1.

A comparison of the substrate binding sites among 3-ACP transferases, including NAT, MtNAS,⁸ TYW2,²¹ and CntL,²² revealed that the binding conformations of SAM are common in the Rossmann fold MT C-terminal region, although the binding residues are not conserved (Figures S5 and S7). The conformations of the C-terminal regions between NAT and TYW2 share structure similarity with an RMSD value of 3.0 Å. However, the conformations of the N-terminal regions are very different, with an RMSD value of 4.7 Å, and the residues for the 3-ACP acceptor binding are not conserved (Figures S5 and S7).

Substrate Recognition and Active Site Architecture in NAT.

We next investigated the substrate binding site in the NAT structure in detail. The C5-CO₂⁻ group of 1 is bound to the N-terminal region of NAT via salt bridges and hydrogen bonds with K23, Y25, and K60. In addition, the C1-OH group of 1 interacts with K41 via a water molecule (w1) (Figure 3a). The C2′-NH₃⁺ group of 1 is bound to the C-terminal region of NAT by hydrogen bonding with one molecule of water (w2), and the main chain of F233 and the C6′-OH group of 1 are bound with a water molecule (w3) hydrogen bonded to T263 and N231 (Figure 3a).

Figure 3.

Substrate binding residues of NAT. (a) Active site structure of NAT in complex with 1 and SAH. Dashed red lines represent hydrogen bonds and salt bridges. Water molecules are shown as red nb_spheres. (b) In vitro assays using the NAT variants (group 1, amino acid (aa) residues around the C-terminal of 1; group 2, residues around the N-terminal of 1; group 3, residues around 3-ACP of SAH; group 4, residues around adenosine of SAH). The peaks of the product were detected by LC-MS, and the intensities of the peak were compared.

Next, we looked into the SAH binding site in the C-terminal region. SAH extends deeply inside the pocket, which consists of α4, α6, and α7, and the loops between α2 and α4, β9 and α5, β10 and α6, β11 and β12, and β12 and α6 (Figure S3). The Cα-NH₃⁺ of SAH interacts with E143 and N231, while the Cα-CO₂⁻ of SAH forms a salt bridge with R228 that is supported by a hydrogen bond network with D140, E143, and one water molecule (w4). The C2′- and C3′-OH groups of the SAH ribose form hydrogen bonds with D191, and the C6-NH₂ group of the adenine moiety is hydrogen-bonded to D211 (Figure 3a). The possible π–cation interaction between the sulfonium cation and W129 is expected to support the conformation of SAH (Figure 3a).

Most importantly, the distance between C6′-OH of 1 and Cγ methylene of SAH is 3.2 Å. These data indicate that the nucleophilic attack on the Cγ methylene of SAM is plausible (Figure 3a). This is the first study where the distance between the acceptor and donor is shown in a 3-ACP transferase structure, because there has been no reported complex structure with both 3-ACP acceptor and SAH/SAM. Furthermore, in the docking model of SAM, the distance between the C6′-OH of 1 and the Cγ methylene of SAM (3.6 Å) is shorter than the distance from the C6′-OH of 1 to the methyl group of SAM (6.0 Å) (Figure S8), suggesting that 3-ACP transfer is favored over methyl group transfer.

Functional Roles of Key Amino Acids Identified via Site-Directed Mutagenesis.

To identify the roles of residues in the substrate-binding site, we performed a site-directed mutagenesis experiment (Figure 3a, Figures S9 and S11, and Tables S1 and S2). First, the residues around C1-OH and C5-CO₂H of 1, K23, Y25, K41, and K60 were mutated (Figure 3b, group 1). The assays were performed by using 2 as a substrate instead of 1 due to the limited availability of 1. The K60A variant dramatically decreased the activity by 88%; K41A decreased the activity by 75%, while the K23A and Y25F variants maintained activities comparable to that of WT. The data indicated that K41 and K60 are important for the binding of 1, while K23 and Y25 are not. Next, we also mutated the residues around C6′-OH, T263 and N231, to alanine (Figure 3b, group 2). T263A and N231A reduced the activity by >90% and 70%, respectively. These data indicated the importance of water 3 for 1 binding and orienting C6′-OH of 1 to SAM. Next, D140, E143, and R228, forming a hydrogen bonding network with the carboxy group of SAH and water 4, were mutated (Figure 3b, group 3). R228A, E143A, and E143Q almost abolished enzyme activity, while D140A retained 38% of the activity. These data indicate that the hydrogen bonds between the carboxyl group and R228 and that between the amino group and E143 are essential. E143 was expected to function as a base for the deprotonation of the Cα-NH₃⁺ of SAM, because it interacts with the amino group. D191, D211, and W129 around adenosine and the sulfonium cation were also mutated (Figure 3b, group 4). D191A completely abolished enzyme activity, and the activities of W129A were reduced by 62%. On the other hand, D211A did not affect the transferase activity. These data indicate that D191 is required for adenosine binding, W129 is important for orientation of the SAM to 1, but D211 is not essential for adenosine binding (Figure 3b, group 4).

For the deprotonation of the C6′-OH of 1 to initiate the 3-ACP transfer reaction, we initially hypothesized that the hydroxyl group of T263 is deprotonated by H262, and the generated alkoxide anion abstracts the proton from C6′-OH as a base. The reactions of T263 variants indicated that the enzyme activity of T263A was severely reduced by >90%, whereas that of H262A was similar to that of the wild type (Figures S9 and S11), unexpectedly. These data indicated that H262 is not involved in the deprotonation of T263 or C6′-OH of 1. Since the distance between the carboxyl group of E143 and the Cα-NH₂ of SAH is 2.8 Å (Figure 3a), we next hypothesized that E143 abstracts the proton from the Cα-NH₃⁺ of SAM as a base to initiate the reaction. The reaction profiles of the E143 variants showed that E143Q and E143A were severely reduced by 93 and 100%, respectively, suggesting that E143 should work as a base as expected (Figure 3b, group 3).

Substrate Binding Affinity of Variants Determined by Thermal Shift Assays.

To investigate the substrate binding capacity of each mutant, thermal shift assays were conducted for WT, and the mutants K60A, E143Q, E143A, D191A, R228A, N231A, and T263A, where the activity was significantly abolished. When 2 was added to the WT enzyme, a clear melting temperature peak appeared, indicating that 2 can bind to the enzyme (Figure S12). The further addition of SAM increased the melting temperature by 7°C (Figure S12). These data are consistent with the kinetics data of NAT which shows an ordered Bi–Bi pattern in which 1 first binds to the enzyme and SAM second binds to E-1 complex. The same effects were also observed in E143Q, E143A, N231A, R228A, and T263A, suggesting that these variants can bind to 2 and SAM. In particular, E143Q showed the same melting point as WT, implying that E143Q can bind to the substrate with the same affinity as WT. The data show that E143 is not responsible for substrate binding and reinforce the hypothesis that E143 acts as a base. In contrast, K60A did not exhibit a clear melting temperature when 2 was added, and D191A did not exhibit a clear peak when SAM was added. These data indicate that K60 and D191 are essential for binding of 2 and SAM binding, respectively. As this example shows, in addition to assays, biophysical analysis is useful for obtaining information when distinguishing between substrate binding and catalytic roles, identifying amino acid residues involved in substrate binding when two substrates are accepted, and demonstrating the mechanism of Bi–Bi reactions.

Simulation of NAT Reactions Using MD and QM/MM Calculations.

To explore the detailed mechanism underlying the S_N2 reaction catalyzed by NAT, we employed a computational approach that combined MD simulations and QM/MM calculations. We modeled SAM into SAH in the crystal structure of NAT in complex with 1 and SAH, and we conducted MD simulations on the model (Figure S13). During three independent 500 ns MD simulations, we analyzed the formation of a near attack conformation (NAC), which was defined based on the distance and angle among the nucleophile, electrophile, and leaving group (Figure S14A). The initial protonation states upon binding of SAM and 1 to the enzyme were considered to be C6′-OH for 1, Cα-NH₃⁺ for SAM, and COO⁻ for the side chain of E143, resulting in a low NAC population of up to 37.8% (Figure S14C, condition 1). Even after proton transfer from the Cα-NH₃⁺ of SAM to the E143 side chain, the occurrence of NAC did not increase (Figure S14C, conditions 2 and 3). Therefore, we examined the state wherein the proton on E143 is further transferred to D140 (Figure S14C, conditions 4 and 5). As expected, this protonation state led to a higher NAC population, reaching up to 97.5%, which suggested that these states were plausible for initiating the S_N2 reaction. Using k-means clustering, we selected three representative snapshots of the NAC in the optimal protonation states (Figure S15). We then performed QM/MM calculation using the ONIOM method in triplicate to estimate the energy profile of the S_N2 reaction (Figure S16, at the M06–2X/6–311++G-(2df,p):AMBER//M06–2X/6–31G(d,p):AMBER level).^27–30 The activation free energies (ΔG^‡) were 15.9, 18.1, and 18.8 kcal mol⁻¹ in the three trials with negative reaction free energies (ΔG < 0), supporting that the reaction proceeds spontaneously at room temperature. During C–O bond formation between the C6′-oxygen of 1 and the Cγ of SAM, a proton was transferred from the C6′-oxygen to the Cα-nitrogen of SAM, forming a six-membered ring transition state (Figure S17A). Further noncovalent interaction analysis revealed that the carboxylate of E143 formed a hydrogen bond with the Cα-amino group of SAM, either directly or mediated by a water molecule (Figure S18). In addition, the activation free energy increased by 5.0 to 8.0 kcal mol⁻¹ when E143 was excluded from the QM regions but was treated at the MM level (Figure S16C). In the transition state lacking E143, a hydrogen atom was positioned close to the C6′-oxygen of 1, rather than the Cα-nitrogen of SAM (Figure S17C, d_O–H and d_N–H). These results suggested that E143 facilitated the S_N2 reaction by increasing the Brønsted basicity of Cα-NH₂ of SAM, thereby promoting the deprotonation of C6′-OH of 1 and stabilizing the resulting Cα-NH₃⁺ species of SAM (Figure S17D).

Subsequently, we investigated possible pathways of the initial proton transfer from the Cα-NH₃⁺ of SAM to D140, which established the plausible protonation states required to initiate the S_N2 reaction (Figure S19). Each protonation state was modeled from the reactant of the S_N2 reaction and optimized by a QM/MM calculation. We identified two thermodynamically favorable proton transfer pathways: one mediated by E143 and the other involving the carboxy group of SAM (Figure S19, routes 1 and 2, respectively). Given the surface location of D140, the proton transferred to D140 may ultimately be released into the bulk solvent, thereby completing the catalytic cycle. Judging from the experimental result that the D140A mutant exhibited decreased catalytic activity of 38% of the wild type, while D140 is primarily responsible for accepting the proton from E143, an alternative pathway that does not rely on D140 is also likely to exist. We also conducted MD simulations of the N231A and T263A mutants, which resulted in a reduced population of NAC compared to that of the wild type (Figure S20). Combining the results of the mutation experiments and thermal shift assays, N231 and T263 appear to maintain the preferred conformation of 1 to initiate the S_N2 reaction, making them critical for catalytic activity.

DISCUSSION

By integrating structural, biophysical, and computational analyses, we propose the reaction mechanism of NAT in Figure 4. Upon binding of SAM and 1 to NAT, the reaction proceeds through two consecutive steps, as discussed below. First, protons are relayed between the SAM and enzyme to form a reactive complex (Figure 4a). In route 1, a proton was first transferred from the Cα-NH₃⁺ of SAM to the carboxylate of E143, either directly or via a water molecule (Figure S19, states 1 to 2). In route 2, the proton was directly accepted by the carboxylate of SAM (Figure S19, states 1 to 4). The second proton relay occurred from the CO₂H group of either E143 (route 1; Figure S19, states 2 to 3) or SAM (route 2; Figure S19, states 4 to 5) to the carboxylate of D140, possibly mediated by one or two water molecules. A hydrogen-bonding network was formed between the carboxy groups of SAM, D140, and E143 through two or three water molecules, which appeared to facilitate the second proton transfer (Figure 4a). In the reactant, SAM and 1 form the six-membered ring transition state via Cα-NH₂ of SAM, whose Brønsted basicity is increased by E143 and C6′-OH of 1. The phenoxide anion generated by deprotonation of the C6′-OH attacks Cγ of SAM, resulting in C–S bond cleavage to produce 3 and MTA (Figure 4b). In the crystal structure, the distances between Cα-NH₂ of SAH and E143, the distance between the NH₂ and C6′-OH of 1, and the distance between the C6′-OH and Cγ of SAH are 2.8, 3.0, and 3.2 Å, respectively (Figure 3a). In addition, three water molecules including w4 are positioned to potentially mediate the initial proton relay between SAM and D140. These results support the proposed reaction mechanism of NAT.

Figure 4.

Proposed reaction mechanism of NAT. The proton relay between SAM and enzyme to generate the reactant (a) and S_N2 attack in the reactant (b) is depicted.

In MtNAS, deprotonation of the substrate is driven by a tyrosine residue Y107 that is activated by a nearby glutamic acid residue E81.⁸ In Tsr3, D70 is utilized as a base for the deprotonation of substrate.²³ In NAT, E143 or SAM-CO₂⁻ abstracts a proton from the NH₃⁺ group of SAM, and the NH₂ group abstracts a proton from the C6′-OH as a base. This is the first case in which the amino group of SAM acts as a base for the S_N2 attack in the 3-ACP transfer. To test whether the amino group of SAM works as a base in the reaction, we synthesized deamino-SAM (da-SAM) (Schemes S2 and S3, Figure S21).³¹ As expected, the reaction with deamino-SAM and nocardicin E did not yield any product (Figure S23). The thermal shift assay showed that deamino-SAM can bind to the enzyme to a similar extent as SAM (Figure S12), indicating that the loss of the amino group does not affect the binding affinity to the enzyme. These data indicated that the amino group of SAM is necessary for the reaction to occur and potentially acts as a base as we hypothesized. We also synthesized decarboxyl-SAM (dc-SAM), but it did not react with nocardicin E (Scheme S4 and Figures S22 and S23). The thermal shift assay showed that decarboxyl-SAM cannot bind to NAT (Figure S12), indicating the importance of hydrogen bonding between the carboxyl group of SAM and R228 for substrate binding.

To investigate the importance of amino group for another 3-ACP transferase, we assayed CntL²² with da-SAM and 3-ACP acceptor d-histidine (d-his), but the reaction did not yield any product (Figure S24). The thermal shift assay showed that da-SAM can bind to the enzyme to the same extent as SAM (Figure S12). However, while the melting point increased when CntL was combined with SAM and d-his compared to when it was combined with SAM alone, no increase in melting point was observed when CntL was combined with da-SAM and d-his compared to that when it was combined with da-SAM alone (Figure S12). This result suggests that the CntL-da-SAM complex cannot bind to d-His and the amino group of SAM is important for binding to d-His in CntL. Our strategy based on the biochemical analysis and thermal shift assay provides useful information to understand the reaction of CntL.

We also investigated the product profiles of NAT variant reactions expecting that methyl transfer might occur instead of 3-ACP transfer owing to the mutations in the substrate binding region, which should affect their reactivity. However, none of the constructed variants catalyzed methyltransfer to produce a methylated product, indicating that the orientation of SAM in relation to 1 is tightly regulated or the two substrates themselves define the orientation of the two substrates in NAT (Figures S10 and S11). To examine the substrate specificity of NAT and its variants, we utilized 1 analogs as 3-ACP acceptors, including amoxicillin and cefadroxil (Figures S25–S28). NAT accepted both of them, to produce the 3-ACP-β-lactam products, but the yields of these products were very low (<1%) as compared to the natural substrate 1. These data further confirmed that the substrate specificity of the NAT is strict. We also tested several variants, K23A, Y25F, and K60A from group 1, and T263A from group 2 (Figures S27 and S28). In the assay with cefadroxil, the Y25F variant showed increased activity, and K23A exhibited activity comparable to the wild type, while T263A showed a complete loss of activity. In contrast, none of the mutants retained activity toward amoxicillin. As expected, the mutation onto groups 1 and 2 affected the substrate binding of the 3-ACP acceptor and has the potential to produce useful biocatalysts.

CONCLUSIONS

Taken together, our X-ray structural and computational studies of the 3-ACP transferase NAT provided detailed information to understand the catalysis of NAT. This is the first enzyme structure complexed with both SAH and its 3-ACP acceptor and displays the structural basis for the orientation of the acceptor and donor SAM, which distinguishes a 3-ACP transferase from a methyltransferase. The QM/MM calculation suggested that 1 is deprotonated by the activated Cα-NH₂ of SAM, which is distinct from the reported 3-ACP synthase.⁸ The thermal shift assay demonstrated that E143 functions as a base rather than a substrate-binding amino acid residue. Furthermore, although the synthesized diamino-SAM was shown to bind to the enzyme by a thermal shift assay, the reaction did not proceed in the assay using this substrate, supporting deprotonation by Cα-NH₂ of SAM. Our analysis is the first computational chemical analysis of the S_N2 reaction mechanism of 3-ACP transferase, providing important insight into understanding its catalysis. The knowledge of enzyme reaction dynamics obtained from such computational analysis complements the information obtained from X-ray crystal structural analysis, deepening our understanding of enzyme reactions. The combination of structural, biochemical, biophysical, and computational analyses provides expanded insights into the catalysis of 3-ACP transferases. Future engineering of NAT based on the findings obtained in this study will enable the production of 3-ACP-modified analogues or methylated analogues, thereby expanding the potential for drug development using SAM-mediated enzyme modification methods. The knowledge and strategies based on the biochemical, biophysical, computational, and structural biological analysis in this study are expected to contribute to a broader understanding of enzymes that catalyze similar S_N2-type reactions involving two substrates.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c08367.

Experimental details for X-ray crystallography, enzymatic assays, LC-MS chromatograms, thermal shift assays, MD simulations, QM/MM calculations, and NMR spectra of the synthetic compound (PDF)

ABBREVIATIONS

3-ACP

3-amino-3-carboxypropyl

SAM

S-adenosyl-l-methionine

MT

methyltransferase

SAH

S-adenosylhomocysteine

MTA

5′-methylthioadenosine

MD

molecular dynamics

QM/MM

quantum mechanics/molecular mechanics