Structural analysis of a novel substrate-free form of the aminoglycoside 6′-N-acetyltransferase from Enterococcus faecium

The crystal structure of the aminoglycoside 6′-N-acetyltransferase from Enterococcus faecium is reported in a novel substrate-free form in order to understand the mechanism underlying its substrate recognition in detail. It is proposed that the enzyme sequentially undergoes conformational selection and induced fit for substrate binding.

Abstract

Aminoglycoside acetyltransferases (AACs) catalyze the transfer of an acetyl group between acetyl-CoA and an aminoglycoside, producing CoA and an acetylated aminoglycoside. AAC(6′)-Ii enzymes target the amino group linked to the 6′ C atom in an aminoglycoside. Several structures of the AAC(6′)-Ii from Enterococcus faecium [Ef-AAC(6′)-Ii] have been reported to date. However, the detailed mechanism of its enzymatic function remains elusive. In this study, the crystal structure of Ef-AAC(6′)-Ii was determined in a novel substrate-free form. Based on structural analysis, it is proposed that Ef-AAC(6′)-Ii sequentially undergoes conformational selection and induced fit for substrate binding. These results therefore provide a novel viewpoint on the mechanism of action of Ef-AAC(6′)-Ii.

1. Introduction  

Aminoglycosides such as kanamycin are the most popular antibiotics used for the treatment of a wide range of infections caused by Gram-positive and Gram-negative bacteria (Kotra et al., 2000 ▸; Magnet & Blanchard, 2005 ▸; Krause et al., 2016 ▸). However, the function of these compounds can be inhibited by chemical modification by various enzymes (Llano-Sotelo et al., 2002 ▸). These chemical modifications include acetylation, phosphorylation and adenylation, which weaken the affinity of aminoglycosides for the 30S ribosomal subunit (Kotra et al., 2000 ▸; Mingeot-Leclercq et al., 1999 ▸). Specifically, enzymes that modify aminoglycosides can be classified as aminoglycoside acetyltransferases [AAC(1), AAC(3), AAC(2′) and AAC(6′)] (Magnet & Blanchard, 2005 ▸; Mingeot-Leclercq et al., 1999 ▸; Wright, 1999 ▸; Azucena & Mobashery, 2001 ▸), aminoglycoside phosphotransferases [APH(4), APH(6), APH(9), APH(3′), APH(2′′), APH(3′′) and APH(7′′)] (Heinzel et al., 1988 ▸; Mckay & Wright, 1995 ▸; Culebras & Martínez, 1999 ▸; Wright & Thompson, 1999 ▸) and aminoglycoside nucleotidyltransferases [ANT(6), ANT(4′), ANT(2′′) and ANT(3′′)] (Kotra et al., 2000 ▸; Magnet & Blanchard, 2005 ▸; Hollingshead & Vapnek, 1985 ▸; Murphy, 1985 ▸; Ounissi et al., 1990 ▸), depending on the position of the targeted atom in the molecule.

AAC(6′) enzymes catalyze the transfer of an acetyl group from acetyl-CoA to an aminoglycoside, targeting the amino group linked to the 6′ C atom (Kotra et al., 2000 ▸). This enzymatic reaction produces CoA and an acetylated aminoglycoside. As such, AAC(6′) enzymes recognize two heterogeneous substrates for their function. Over the past two decades, much attention has been paid to AAC(6′)-Ii from Enterococcus faecium [Ef-AAC(6′)-Ii; the letters I and i symbolize the resistance profile that AACs confer on their host cells and an individual identifier, respectively] as an AAC(6′) enzyme from a Gram-positive bacterial pathogen (Wright & Ladak, 1997 ▸). Previous biochemical studies have shown that Ef-AAC(6′)-Ii has unique enzymatic properties, such as a broad range of substrate specificity and high catalytic efficiency (Draker et al., 2003 ▸; Draker & Wright, 2004 ▸). Therefore, Ef-AAC(6′)-Ii may be considered as an attractive target for the development of novel antibiotics.

Previous biochemical studies of Ef-AAC(6′)-Ii have been conducted focusing on enzyme kinetics and the roles of key residues in the active site (Draker et al., 2003 ▸; Draker & Wright, 2004 ▸). Draker and coworkers reported that acetyl-CoA binds first to Ef-AAC(6′)-Ii, sequentially followed by an aminoglycoside, and that it is the diffusion process of the aminoglycoside substrate, and not the transfer of the acetyl group, that corresponds to the rate-determining step in catalysis (Draker et al., 2003 ▸). Additionally, they elucidated the respective functional roles of Glu72, His74, Leu76 and Tyr147 in the active site of Ef-AAC(6′)-Ii by site-directed mutagenesis (Draker & Wright, 2004 ▸). Apart from these biochemical studies, structural studies of Ef-AAC(6′)-Ii have mainly been conducted by X-ray crystallography (Wybenga-Groot et al., 1999 ▸; Burk et al., 2003 ▸, 2005 ▸; Baettig et al., 2016 ▸). Four crystal structures of Ef-AAC(6′)-Ii have been deposited in the Protein Data Bank to date: a complex form with acetyl-CoA (PDB entry 1b87; Wybenga-Groot et al., 1999 ▸), two complex forms with CoA (PDB entries 1n71 and 2a4n; Burk et al., 2003 ▸, 2005 ▸) and a substrate-free form (PDB entry 5e96; Baettig et al., 2016 ▸).

These four structures have provided a wealth of information on the acetyl-CoA-binding site and on conformational changes in response to the substrate (Wybenga-Groot et al., 1999 ▸; Burk et al., 2003 ▸, 2005 ▸; Baettig et al., 2016 ▸). Comparative analysis of the previously reported conformers showed that the acetyl-CoA complex structure is nearly identical to the CoA complex structure and that Ef-AAC(6′)-Ii has an intrinsically flexible region (Burk et al., 2003 ▸, 2005 ▸). Baettig and coworkers pointed out that a partially unfolded region found in the substrate-free form corresponds to an intrinsically flexible region, which affects the dimerization and substrate binding of Ef-AAC(6′)-Ii (Baettig et al., 2016 ▸). However, the scarcity of structural information on the substrate-free form has limited the understanding of its structural dynamics. Therefore, another structure showing the substrate-free state is required to widen our understanding of the conformational distributions in the native state and the conformational changes in response to the substrate.

Here, we report the crystal structure of a novel substrate-free form of Ef-AAC(6′)-Ii. The structure was then compared with those of other conformers, such as the aforementioned substrate-free form and substrate-complexed form. Our results revealed that a specific region of Ef-AAC(6′)-Ii adopts nearly the same structural arrangement as in the complex form, which was not observed in the previous substrate-free form. This finding possibly implies that one of diverse conformers is selected as a suitable architecture for substrate binding. Based on comparative analysis of the diverse structures, the present study addresses how Ef-AAC(6′)-Ii recognizes acetyl-CoA as its substrate and undergoes structural changes accordingly.

2. Materials and methods  

2.1. Cloning, overexpression and purification  

To construct a recombinant vector for Ef-AAC(6′)-Ii, the gene was amplified using polymerase chain reaction (PCR). The PCR products were inserted into a pET-28a(+) vector at the NdeI and XhoI restriction sites. Escherichia coli strain BL21(DE3) competent cells were transformed with the resulting recombinant vector. The cells were cultured at 37°C in 2 l lysogeny broth containing 50 µg ml⁻¹ kanamycin. When the optical density at 600 nm reached 0.6, the temperature was adjusted to 20°C and 0.5 mM isopropyl β-d-1-thiogalacto­pyranoside was added to induce gene expression. The cells were then further cultured overnight. The cultured cells were harvested, resuspended in lysis buffer (20 mM Tris–HCl pH 8.0, 500 mM sodium chloride, 25 mM imidazole) and lysed by ultrasonication on ice. The cell lysate was centrifuged at 10 000g for 30 min at 4°C. The supernatant was mixed with nickel–nitrilotriacetic acid affinity resin for 2 h and the mixture was loaded onto a gravity-flow column. The column was subsequently washed with wash buffer (20 mM Tris–HCl pH 8.0, 500 mM sodium chloride, 60 mM imidazole). The protein was then eluted with elution buffer (20 mM Tris–HCl pH 8.0, 500 mM sodium chloride, 250 mM imidazole) and the eluted fractions were collected and concentrated using a centrifugal filter (Amicon; Millipore). For further purification, the protein solution was loaded onto a Superdex 200 Increase 10/300 column (GE Healthcare) pre-equilibrated with buffer (20 mM Tris–HCl pH 8.0, 150 mM sodium chloride). The resulting protein fractions were harvested and concentrated to 8 mg ml⁻¹. The purity of the protein was visually assessed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

2.2. Crystallization and X-ray diffraction data collection  

Crystallization conditions for Ef-AAC(6′)-Ii were explored in 24-well crystallization plates using the hanging-drop vapor-diffusion method. Commercial kits, such as Crystal Screen, Crystal Screen 2, Index, Natrix (Hampton Research) and Wizard Classic 1 and 2 (Rigaku Reagents), were used for crystallization screening. The protein sample (1 µl) was mixed with the crystallization solution (1 µl) and the droplets were equilibrated against 1 ml reservoir solution. Crystals were obtained using 50 mM HEPES–Na pH 7.0, 10 mM magnesium chloride, 1.6 M ammonium sulfate in approximately 60 days. A single crystal was selected and soaked in reservoir solution supplemented with 10%(v/v) glycerol for cryoprotection. X-ray diffraction data were collected at −178°C on beamline BL-5C at Pohang Accelerator Laboratory, Pohang, Republic of Korea. Data processing, including indexing, integration and scaling, was conducted using HKL-2000 (Otwinowski & Minor, 1997 ▸).

2.3. Structure determination and refinement  

The structure of Ef-AAC(6′)-Ii was determined using the molecular-replacement method with Phaser-MR (McCoy et al., 2007 ▸) in Phenix (Liebschner et al., 2019 ▸). The previously solved structure of Ef-AAC(6′)-Ii (PDB entry 5e96; Baettig et al., 2016 ▸) was used as a search model. After phase determination, models for most residues were built using AutoBuild (Terwilliger et al., 2008 ▸) in Phenix. Missing residues were manually fitted using Coot (Emsley et al., 2010 ▸). Model refinement was iteratively performed using phenix.refine (Afonine et al., 2012 ▸) in Phenix until the R _work and R _free values reached 22.6% and 26.5%, respectively. The stereochemistry of the final model was validated using MolProbity (Chen et al., 2010 ▸). The final atomic coordinates and structure factors for Ef-AAC(6′)-Ii have been deposited in the Protein Data Bank under accession code 7bxz. The structural figures in this manuscript were prepared using PyMOL (DeLano & Lam, 2005 ▸).

2.4. Multi-angle light-scattering analysis  

The absolute molecular weight of Ef-AAC(6′)-Ii in solution was measured by SEC-coupled multi-angle light scattering (SEC-MALS). The protein solution was loaded onto a Superdex 200 Increase 10/300 GL 24 ml column pre-equilibrated with SEC buffer (20 mM Tris–HCl pH 8.0, 500 mM NaCl). The flow rate of the buffer was controlled at 0.4 ml min⁻¹ at 20°C. A DAWN TREOS MALS detector was used connected to an ÄKTAexplorer system. The molecular weight of bovine serum albumin was measured as a reference. Data were processed and assessed using the ASTRA software.

3. Results and discussion  

3.1. Overall structure of Ef-AAC(6′)-Ii  

The structure of Ef-AAC(6′)-Ii was determined in a substrate-free form at 2.5 Å resolution. Diffraction data and refinement statistics for Ef-AAC(6′)-Ii are summarized in Table 1 ▸. The crystal belonged to space group C121, with three molecules (chains A–C) present in the asymmetric unit (Fig. 1 ▸ a). However, a binding prediction using the PDBePISA server (Krissinel & Henrick, 2007 ▸) showed that Ef-AAC(6′)-Ii presumably exists as a dimer and chains A and B most probably interact with each other. Several structures of Ef-AAC(6′)-Ii have previously been reported. We compared our dimeric structure (chains A and B) with other dimeric structures [PDB entries 1n71 (Burk et al., 2003 ▸), 2a4n (Burk et al., 2005 ▸) and 5e96 (Baettig et al., 2016 ▸)]. Our dimeric structure, including the interface, was consistent with these structures. In addition, we identified that chain C forms a dimer with a chain (chain C′) in the neighboring asymmetric unit. Specifically, we superimposed chain C onto chain C′ and chains AB onto chains CC′. The root-mean-square deviation (r.m.s.d.) value between chains C and C′ is 0.00 Å over 176 C_α atoms. This is a natural result as the the dimer CC′ is a crystallographic dimer. The r.m.s.d. between chains AB and chains CC′ is 0.88 Å over 352 C_α atoms. This result means that the dimers are also nearly identical to each other in terms of dimeric structure, even though chains A, B and C but not chain C′ are present in the asymmetric unit. Our SEC-MALS analysis also confirmed that Ef-AAC(6′)-Ii is present as a dimer in solution, in accordance with the crystallographic dimeric state. The SEC-MALS analysis showed a single peak at a retention volume of 38 ml in SEC, which corresponds to an absolute molecular weight of 48.0 kDa (Fig. 1 ▸ b). This value is almost in agreement with the theoretical molecular weight of dimeric Ef-AAC(6′)-Ii containing N- and C-terminal His₆ tags (48.1 kDa). Considering the multimeric state of Ef-AAC(6′)-Ii in solution, it is assumed that Ef-AAC(6′)-Ii plays a biological role in a dimeric form in the cellular environment. Thus, the structure of Ef-AAC(6′)-Ii as a functional unit constitutes a homodimer consisting of chains A and B (Fig. 1 ▸ c).

Table 1. Data-collection and refinement statistics for Ef-AAC(6′)-Ii.

Values in parentheses are for the outermost resolution shell.

Data collection
X-ray source	BL-5C
Wavelength (Å)	0.977872
Space group	C121
a, b, c (Å)	68.43, 90.32, 102.15
α, β, γ (°)	90, 90, 96.08
Total reflections	148028
Unique reflections	21196 (1041)
Multiplicity	7.0 (6.7)
Completeness (%)	98.9 (97.7)
Mean I/σ(I)	18.3 (2.3)
R merge † (%)	6.4 (47.2)
Resolution range (Å)	50.00–2.50 (2.54–2.50)
Refinement
Resolution range (Å)	29.90–2.50 (2.63–2.50)
No. of reflections in working set	21180
No. of reflections in test set	1037
R work (%)	22.6
R free (%)	26.5
Ramachandran plot
Favored (%)	98.66
Outliers (%)	0.00
Rotamer outliers (%)	0.00
R.m.s.d., bond lengths (Å)	0.004
R.m.s.d., angles (°)	0.827

^† R _merge = , where I _i(hkl) is the ith observed intensity of reflection hkl and 〈I(hkl)〉 is the average intensity obtained from multiple measurements.

Figure 1.

Overall structure of Ef-AAC(6′)-Ii. (a) Crystal structure of Ef-AAC(6′)-Ii in the asymmetric unit. The three molecules are represented as cartoons. (b) SEC-MALS profile of Ef-AAC(6′)-Ii. SEC-MALS data (red) are plotted as SEC elution volume (x axis) versus absolute molecular mass (y axis) distributions on the SEC chromatogram (black) at 280 nm. (c) The dimeric structure of Ef-AAC(6′)-Ii consisting of chains A and B. (d) The monomeric structure of Ef-AAC(6′)-Ii (chain B). (e) Surface representation of chains A and B viewed from two different perspectives.

A structural superimposition of the three chains showed that the r.m.s.d. values for 168–176 C_α atoms are 0.80–1.13 Å, signifying that the three chains are structurally nearly identical to each other. One monomer (chain B) comprises six helices (α1–α6), seven strands (β1–β7) and three loops (L1–L3) (Fig. 1 ▸ d). However, the region corresponding to the α2 helix in chain B is disordered in chains A and C, implying that the α2 region is possibly affected by the crystallographic environment. The β-strands form a twisted β-sheet in the center, which is incompletely surrounded by helices and loops. While the cluster consisting of α5–α6, β5–β7 and L1–L3 interacts with the neighboring chain, the other cluster forms a long cleft with the corresponding region of the neighboring chain (Fig. 1 ▸ e). Hence, these two clusters seem to play different roles both structurally and functionally.

3.2. Surface properties of Ef-AAC(6′)-Ii  

In our structure, 35 residues per subunit are involved in interactions with each other to form a dimer. These residues account for approximately 20% of the overall residues, thus forming relatively tight interactions between the two subunits. Moreover, these interactions contribute to creating a long cleft between the two subunits (Fig. 2 ▸ a) that corresponds to the binding site for aminoglycosides. This suggests that the formation of a dimer is indispensable for the enzymatic function of Ef-AAC(6′)-Ii by constructing the active site through such interactions.

Figure 2.

Surface properties of Ef-AAC(6′)-Ii. (a) Surface representation of chains A and B. The gray regions indicate the interacting residues between the two chains. (b) Surface electrostatic potential of Ef-AAC(6′)-Ii. The figures were generated using APBS Tools with a solvent-accessible surface calculation. The dashed circle and ellipse mark strongly electronegative regions. The scale bars indicate a range from −5kT e⁻¹ (red) to 5kT e⁻¹ (blue). (c) A cartoon representation of chain B colored by the degree of amino-acid sequence conservation.

A surface electrostatic potential analysis of Ef-AAC(6′)-Ii revealed its unique charge distribution. Remarkably, the cleft region was found to dominantly exhibit negative charge, whereas the remaining area did not show any particular charge distribution (Fig. 2 ▸ b). Such electrostatic properties imply that the cleft region can accept positively charged molecules as an active site. Indeed, aminoglycosides, including kanamycin, retain many amino groups which are prone to protonation in solution, thereby constituting positively charged molecules. To identify whether the residues related to this cleft region are evolutionarily conserved, we investigated the degree of sequence conservation across 150 enzymes homologous to Ef-AAC(6′)-Ii. We found that the residues contributing to this surface feature are evolutionarily well conserved across the homologues (Fig. 2 ▸ c). Therefore, it is assumed that the unique charge distribution of AAC(6′) enzymes, including Ef-AAC(6′)-Ii, may constitute a molecular strategy to facilitate substrate binding.

3.3. Structural comparison with other conformers  

Previous studies have reported diverse structures of Ef-AAC(6′)-Ii, including a substrate-free form (PDB entry 5e96; Baettig et al., 2016 ▸), an acetyl-CoA complex form (PDB entry 1b87; Wybenga-Groot et al., 1999 ▸) and two CoA complex forms (PDB entries 1n71 and 2a4n; Burk et al., 2003 ▸, 2005 ▸). We compared the structure of our novel substrate-free form with these structures in order to identify structural differences between our form and these other forms (Figs. 3 ▸ a and 3 ▸ b). Most remarkably, we found that the loop-rich region including α5 is ordered in our structure, in the acetyl-CoA complex form and in the CoA complex form (Fig. 3 ▸ a), but not in the other substrate-free form (PDB entry 5e96; Fig. 3 ▸ b). This region corresponds to a disordered region in the substrate-free form, probably owing to its extremely high flexibility. This finding indicates that the loop-rich region can exist in at least two conformational states, although acetyl-CoA does not bind to this region.

Figure 3.

Structural comparison of Ef-AAC(6′)-Ii. (a) Structural comparison between our substrate-free form (cyan) and other forms such as the acetyl-CoA-bound form (orange) and the CoA-bound form (pink). The gray and yellow sticks represent acetyl-CoA and CoA, respectively. (b) Structural comparison between our substrate-free form (cyan) and the other substrate-free form (yellow). The dashed ellipse indicates the ordered region shown in our structure. (c) B-factor distribution of Ef-AAC(6′)-Ii. Each chain is depicted in putty representation. The dashed circles indicate regions that are assumed to be highly flexible. The three structures are viewed from the same perspective.

Next, we investigated how the B factors in this region are distributed in the three molecules contained in the asymmetric unit, as B-factor values can provide structural information on local flexibility. Four distinct regions (FR1–FR4) exhibited relatively high B factors (Fig. 3 ▸ c), implying that they constitute intrinsically flexible regions. The loop-rich region corresponds to FR4, which seems to have been affected by the crystallo­graphic environment. This region in chain C showed higher values than those in chains A and B (Fig. 3 ▸ c). These results support our assumption that the FR4 region can exist in two conformational states in the substrate-free form.

Interestingly, we found that our structure differs from the other substrate-free form in terms of the spatial formation of the acetyl-CoA-binding site. In our structure, the ordered region seems to be spatially assigned as a hinge between the two monomers (Fig. 4 ▸ a). In the other substrate-free structure, however, this region is disordered, showing an incomplete architecture (Fig. 4 ▸ b). This finding suggests that our conformation is more suitable for binding acetyl-CoA than the other form. It is noteworthy that Lys84 and Glu141 in our form are closer to each other (Fig. 4 ▸ a) compared with those in the other form. In the acetyl-CoA complex form, these two residues form a salt bridge, thereby completing the formation of the acetyl-CoA-binding site (Fig. 4 ▸ c). Thus, our structure seems to represent an intermediate form between the substrate-free and substrate-bound forms.

Figure 4.

Structural differences between the two substrate-free forms and the acetyl-CoA-bound form. (a) The acetyl-CoA-binding site in our structure. The ordered FR4 region in chain A is colored pink and two residues at the edge of the disordered region in the other substrate-free form are colored red. The distance between Lys84 and Glu141 is 8.0 Å. (b) The other substrate-free structure. The color code for the two residues at the edge of the disordered region is the same as in (a). (c) The acetyl-CoA-binding site in the acetyl-CoA-bound structure. The pink and red color codes are the same as in (a). Acetyl-CoA is depicted in gray sticks. The distance between Lys84 and Glu141 is 3.9 Å.

3.4. A possible model for substrate recognition  

The structural properties of Ef-AAC(6′)-Ii have been investigated using various biophysical methods such as nuclear magnetic resonance (NMR; Freiburger et al., 2011 ▸), small-angle X-ray scattering (SAXS; Baettig et al., 2016 ▸) and circular dichroism (CD; Baettig et al., 2016 ▸), along with structural comparison. These studies showed that the two regions denominated FR1 and FR4 here are flexible and that substrate binding results in the ordering of these regions. Thus, these biophysical analyses also support our statement regarding the intrinsically flexible regions.

Considering that Ef-AAC(6′)-Ii exists as a dimer for its enzymatic activity, the issue of allosteric substrate-binding mechanisms needs to be addressed. Historically, three allo­steric mechanism models have been presented: the Monod–Wyman–Changeux (MWC) model (Monod et al., 1965 ▸), the Koshland–Némethy–Filmer (KNF) model (Koshland et al., 1966 ▸) and the Hilser and Thompson (HT) model (Hilser & Thompson, 2007 ▸). In the MWC model, substrate binding to a subunit induces conformational changes in all of the subunits in a dimeric enzyme and subsequently affects the binding affinity for another substrate. In the KNF model, on the other hand, a substrate-free subunit adopts a different conformation to a substrate-bound subunit. Cooperative activity between two subunits depends on the strength of subunit–subunit interactions. Lastly, in the HT model, the folding of intrinsically disordered regions is coupled to binding to their interaction partners. This model postulates that enzymes populate a conformational landscape in substrate-free subunits.

Allosteric mechanisms for Ef-AAC(6′)-Ii have been discussed in the literature (Freiburger et al., 2011 ▸, 2014 ▸). These studies presented a hybrid KNF–HT model and a hybrid MWC–KNF model based on NMR, CD and isothermal titration calorimetry analyses. In the hybrid KNF–HT model, cooperative acitity between the two subunits of Ef-AAC(6′)-Ii is modulated depending on the stability of the subunits. In addition, in the hybrid MWC–KNF model allosteric mechanisms vary depending on the binding sites of the substrates. Namely, the MWC model can be applied to a substrate binding to the dimer interface (for example paramomycin), whereas the effect of a substrate distant from the interface (for example acetyl-CoA) can be explained using the KNF model.

Previous structures of Ef-AAC(6′)-Ii have provided valuable information on conformational changes upon substrate and product binding (Wybenga-Groot et al., 1999 ▸; Burk et al., 2003 ▸, 2005 ▸; Baettig et al., 2016 ▸). Burk and coworkers showed that the acetyl-CoA-bound structure of Ef-AAC(6′)-Ii is similar to the CoA-bound structure (Burk et al., 2003 ▸). They also revealed that the aforementioned FR4 region is intrinsically flexible (Burk et al., 2005 ▸). Although our structure is of a different substrate-free form, this structure enables us to discuss a novel mechanism of substrate recognition by Ef-AAC(6′)-Ii.

We found that the substrate-free form of Ef-AAC(6′)-Ii can exist in at least two forms. Considering that the FR4 region is disordered in the unstable state, Ef-AAC(6′)-Ii is assumed to have a greater population in the substrate-free state. Most remarkably, the FR4 region in our structure resembles that in the acetyl-CoA-bound form, rather than that in the known substrate-free form, although the structural arrangements of residues adjacent to the acetyl-CoA-binding site are somewhat different. Given that Lys84 and Glu141 form a salt bridge, thereby confining acetyl-CoA, it would be reasonable to assume that such a salt bridge is formed sequentially after acetyl-CoA binds to the binding site. Therefore, we conclude that (i) Ef-AAC(6′)-Ii has diverse conformers, particularly in the FR4 region, (ii) an appropriate conformer close to the substrate-bound form is selected and (iii) conformational change occurs upon binding to the substrate in the induced-fit mode.

The conclusions drawn, however, are assumptions based on structural information, mainly using X-ray crystallography. Other analytical methods, such as molecular-dynamics simulations, may provide different information on the conformational changes of Ef-AAC(6′)-Ii during substrate binding and catalysis. Nonetheless, our results provide structural evidence to support the argument that substrate binding by Ef-AAC(6′)-Ii can be explained by an integrated mechanism of conformational change and induced fit.

Supplementary Material

PDB reference: aminoglycoside 6′-N-acetyltransferase, 7bxz

Funding Statement

This work was funded by Chung-Ang University grant . National Research Foundation of Korea grant 2017M3A9D8062960.