Identification of 7-hydroxyindole as an alternative substrate of MauG by in silico and in vitro analysis

Abstract

MauG catalyzes the six-electron oxidation of pre-tryptophan tryptophylquinone (preTTQ) cofactor in methylamine dehydrogenase (MADH) to form mature tryptophan tryptophylquinone (TTQ) via long-range electron transfer. To identify alternative substrates for MauG, docking models for 10 tryptophan-like compounds were constructed using Autodock Vina. These demonstrated spontaneous binding to the preTTQ binding site of MauG, with hydroxyindoles most frequently sharing the natural substrate binding site of MauG. To confirm the result of in silico analysis, 7-hydroxyindole was reacted with bis-Fe^IV of MauG. The spectroscopic change, representing the reactivity of MauG, revealed the highly increased reaction rate (k₃) toward 7-hydroxyindole, suggesting that bis-Fe^IV MauG extracted an electron from the 7-hydroxyindole and then oxidized to di-ferric MauG.

Introduction

Heme, incorporated within heme proteins, has diverse functions such as storage, transfer, and activation of oxygen; electron transfer; and novel catalytic center [1]. The heme in oxygenases catalyzes the incorporation of oxygen atoms into organic molecules. Oxygenases are classified into two types, monooxygenases and dioxygenases, according to the number of oxygen atoms incorporated into a substrate [2]. MauG is a c-type diheme oxygenase, catalyzing the posttranslational modifications of methylamine dehydrogenase precursor (preMADH) (Fig. 1A). During the posttranslational modification, MauG matures the protein-derived cofactor, tryptophan tryptophylquinone (TTQ), of methylamine dehydrogenase (MADH) by six-electron oxidation (Figs. 1B, 2) [3, 4]. The maturation of TTQ involves three steps: crosslinking between the βTrp108 and βTrp57 of MADH, insertion of an oxygen atom into monohydrated βTrp57, and oxidation of the quinol into a quinone [5]. The reaction requires bis-Fe^IV formation of both hemes within MauG and is accomplished by long-range electron transfer via a hole hopping mechanism, without any direct contact between the two hemes in MauG and preTTQ in preMADH [6].

Fig. 1

The structure of the MauG/preMADH complex. A indicates the overall structure of the complex consisting of MauG (pink), preMADH α-(blue), and β-(green) subunits. The arrows indicate the position of the hemes within MauG, colored by atom (hot pink). The gray box indicates the location of preTTQ in MauG, colored by atom (yellow). B is the magnified structure of preTTQ. βTrp108 and βTrp57 are represented in yellow, colored by atom. The figure is produced by PyMOL 2.0

Fig. 2

The overall scheme representing the maturation of preTTQ to TTQ catalyzed by MauG. βTrp108 and monohydrated βTrp57 of preTTQ convert to TTQ via six-electron oxidation

Oxygenases are industrially important for enzymatic organic synthesis because of their selectivity, thus avoiding potential overoxidation [7, 8]. As previously mentioned, MauG functions as an oxygenase, catalyzing the crosslinking and oxidation of preTTQ [9, 4]. However, the substrate range of MauG has been limited to the natural substrate, preTTQ, and potential alternative substrates have been poorly investigated. If novel substrates of MauG including tryptophan are identified, MauG could be utilized as a novel oxygenase in diverse industries, such as the pharmaceutical and food industries.

The traditional screening method to identify substrate-like molecules for an enzyme is a lengthy procedure, requiring numerous trial and error. However, the development of in silico analysis shortens the time for screening the chemical candidates by predicting the interactions between protein and compounds in virtual force fields. Here, the alternative substrates for MauG were rapidly screened from tryptophan-like molecules by confirming potential binding using docking simulation. Docking is a molecular modeling method that predicts the possible interaction between proteins and small molecules (ligands) [10]. It is important to analyze the conformation and orientation of the molecules and predict the currently unknown binding site within the protein. For drug discovery, docking can be analyzed by a range of computer software such as Autodock Vina, DockThor, and Molegro Virtual Docker [11], [12], [13]. Among these, the Autodock Vina is an open-source program, based on a simple scoring function that highly improves the average accuracy of binding predictions [14, 15].

In this study, we used Autodock Vina to screen tryptophan-like molecules, which bound to the TTQ binding site of MauG, and visualized the binding sites using PyMOL. Among the molecules, those with a higher binding frequency to the TTQ binding site were selected for in vitro analysis. The association rate and binding affinity to MauG of the selected molecules, including 7-hydroxyindole, was determined to find alternative subatrate of MauG.

Results

Substrate screening by docking simulation

Most of the c-type diheme proteins mediate the electron transport between proteins. However, MauG can perform several catalytic reactions in addition to electron transport. MauG matures preTTQ, the protein-derived cofactor of MADH, to the activated TTQ form. During maturation, the dihemes of MauG oxidize to high-valence iron species by reacting with the oxygen species such as O₂ and H₂O₂ [16]. Donating the electron induces the crosslinking between βTrp108 and mono-hydroxylated βTrp57 in pre-TTQ. The diferric heme then repeats the reduction by accepting the electron from TTQ. The maturation of TTQ is completed after the addition of oxygen and conversion of quinol to quinone. In this study, 10 indole compounds, similar to mono-hydroxylated βTrp57, were selected as alternative substrates of MauG (Table 1). The docking simulation was performed using two states of ligand molecules, rotation and non-rotation. The rotation state involves the rotation of all rotatable bonds of the ligand, indicating the ligands in biological conditions. The non-rotation state designates that the bonds are non-rotatable, indicating the ligand is fixed in a protein crystal form. The results showed that the binding energy of the compounds exhibits a negative value, indicating spontaneous binding to MauG. Additionally, the energy values of all ligands were similar, regardless of the rotational states of the ligands (Table 2). Furthermore, all ligand models possessed a similar location on MauG regardless of the torsion states. This suggests that the rotational state of the ligand model might not significantly affect the docking results because most ligands lack the rotatable bond.

Table 1 Docking ligand structure and PubChem CID number

Table 2 Docking energy between MauG and ligand in rotated and non-rotated states

As previously mentioned, MauG functions as an oxygenase and matures the preTTQ of preMADH to TTQ by transferring electrons to Trp108 and Trp57 of preMADH. During the maturation of TTQ, the MADH attaches to the MauG. Therefore, if the indole compounds frequently bind to the MADH binding site of MauG, the compound can be similarly oxidized by MauG to Trp54 and Trp108 of preTTQ. The binding models of candidate compounds for the MADH binding site of MauG were included among the 20 positions of each docking model (Fig. 3). As a result, all docking results for each compound represent at least three binding models with the preTTQ binding site of MauG (Fig. 3; Table 3). Importantly, more than half of the docking models bound to the catalytic area of the MauG in 5-hydroxyindole, 6-hydroxyindole, and 7-hydroxyindole (Fig. 3F–H). The frequency of docking hydroxyindole compounds at the active site of MauG was two or three times greater than that of the other compounds (Table 3). This indicated that hydroxyindole represents a similar structure to mono hydroxylated Trp57, therefore functioning as an alternative substrate.

Fig. 3

The binding models of the rotated tryptophan-like compounds on MauG. Docking models were visualized by PyMOL. The colored ligands indicate the tryptophan-like chemicals; A: 4-benzyloxyindole, B: 5-benzyloxyindole, C: 6-benzyloxyindole, D: 6-methoxyindole, E: methyl-L-tryptophanate, F: 5-hydroxyindole, G: 6-hydroxyindole, H: 7-hydroxyindole, I: D-tryptophan, J: L-tryptophan. The square box indicates the preTTQ binding site of MauG, and the red box indicates the most frequently bound ligands on the preTTQ binding site among the candidates

Table 3 Frequency of ligands on the active site of MauG based on Fig. 2

The reactivity of 7-hydroxyindole toward MauG

To identify the reactivity of 7-hydroxyindole toward MauG, the reaction was measured by UV/Vis-spectrophotometer at 20 °C, after titration of 7-hydroxyindole. When hydrogen peroxide was titrated with MauG, the absorbance at 405 nm was decreased due to the oxidation of the dihemes in MauG to form bis-Fe ^IV. The addition of 7-hydroxyindole reversed this decrease in absorbance (Fig. 4A). This indicates that the oxidized diheme of MauG accepts the electrons from 7-hydroxyindole, and the tetravalent diheme was reduced to a diferric state by 7-hydroxyindole. It also demonstrates that 7-hydroxyindole can act in the role of preTTQ, the original substrate of MauG. The k_obs value was calculated by single-phase exponential analysis of the ligand binding to MauG. This indicated that the observed rate of the reaction with 7-hydroxyindole represents saturation behavior depending on the substrate concentration (Fig. 4B). The k_obs value was fitted to Eq. (1) to obtain k₃ and k₄ values, and each value was compared to that for the MauG-dependent TTQ synthesis from published data (Table 4). The reaction rate (k₃) for 7-hydroxyindole were 1.64 ± 0.17 s⁻¹. This was highly increased, even after incubation at 20 °C, compared to the value of 0.8 s⁻¹ for the association rate of the preMADH original substrate initiated by the addition of H₂O₂ [17]. However, the binding affinity toward 7-hydroxyindole was highly decreased compared to the preMADH value which did not exceed 1.5 μM. The dissociation rate (k₄) represents -0.32 ± 0.07 s⁻¹ which indicates the negligible reverse reaction.

Fig. 4

The kinetic analysis of the reaction between 7-hydroxyindole and MauG, measured by UV/Vis-spectrophotometer. A represents the spectroscopic change of MauG after sequential titration of H₂O₂ and 7-hydroxyindole. The gray line indicates the original absorbance of MauG before titration of H₂O₂, and the dashed line is the absorbance after titration of H₂O₂. The black line represents the absorbance of MauG after the reaction with 7-hydroxyindole. The black arrow represents the decrease of the absorbance at 403 nm, and the gray arrow indicates the recovery of the absorbance at 403 nm. B is the plot of the concentration-dependent reaction of MauG with 7-hydroxyindole

Table 4 Single-turnover kinetic parameters for the reaction of MauG toward the novel and original substrate

Discussion

The use of oxygenases is industrially important because of their properties that prevent potential overoxidation [7]. MauG is an atypical, c-type diheme oxygenase that matures preTTQ to TTQ by six-electron oxidation using three oxygen atoms, but its substrate range has been limited to preTTQ [4]. It is vital to investigate alternative substrates of MauG, to determine the potential of MauG as an oxygenase in a broad range of industrial settings.

To investigate alternative substrates for MauG, we first used in silico docking analysis to screen and identify candidate compounds. The analysis was a structure-based method that predicts the potential noncovalent binding between protein and ligands by calculating the binding energy using scoring functions [18]. It is significant for drug development because it can minimize the time taken to screen chemical candidates using the vast amount of information available from online databases [10]. Here, we used Autodock Vina to screen for alternative substrates of MauG. Autodock Vina is a docking program with improved accuracy and speed compared to previously released software [15]. The compounds containing an indole ring were selected as candidates, because tryptophan, the original MauG substrate, contains an indole ring in its functional group. The ligand files were set to both rotation and non-rotation mode, considering the condition in which ligands are placed. The binding model was constructed after the docking site was set to cover the whole MauG protein. Both non-rotated and rotated forms of all the candidates exhibited similar docking energy values, with the highest and lowest energy with the range of − 7.0 to − 5.0 kcal/mol and − 5.1 to − 3.9 kcal/mol, respectively. The negative highest and lowest energy values indicated that the ligands can spontaneously bind on the surface of MauG including the TTQ binding site. From the binding energy data alone, it is hard to predict which of the candidate compounds can act as an alternative substrate because it represents the ligand binding energy toward the whole MauG protein.

In previous studies, the reaction in the MauG and MADH complex was demonstrated in crystallo, revealing the TTQ binding site and the distance for electron transfer [19]. Additionally, site-directed mutagenesis of MauG has identified the significant residues involved in TTQ biosynthesis and stabilization of bis-Fe^IV redox state [20,21,22]. To identify potential compounds that can act as alternative substrates, the docking models were screened for interactions with the TTQ binding site of MauG. The ligand binding frequencies on the TTQ binding site were calculated by confirming the ligand binding patterns using PyMOL. The docking models for all candidate compounds contained ligand molecules bound to the TTQ binding site. However, the compounds containing a mono-hydroxylated indole ring demonstrated the highest binding frequency to the TTQ binding site. During TTQ biosynthesis, MauG utilizes three moles of H₂O₂ to catalyze the total six-electron oxidation reaction, equating to three cycles of two-electron oxidation [5]. During the oxidation, MauG introduces a hydroxyl group into mono-hydroxylated βTrp57. This indicates that the original substrate contains a hydroxyl group on its indole ring. Because the mono-hydroxylated βTrp57 possesses the hydroxyl group on the C7 position, the hydroxyindole candidate compounds, which have a similar formation of the indole ring, can bind to the TTQ binding site of MauG with a higher frequency than other tryptophan-like compounds [23]. Among the hydroxyindole compounds, 7-hydroxyindole was selected because it shared the same indole formation as mono-hydroxylated βTrp57.

During TTQ biosynthesis, the heme of MauG undergoes a spectra change depending on its redox state [24]. With the addition of H₂O₂, the diferric state of heme in MauG converts to the bis-Fe^IV state, and the absorbance in the Soret region is decreased. When MauG reacts with preTTQ, the bis-Fe^IV state of MauG converts to the diferric state by incorporating the second oxygen from the solvent into C6 of βTrp57 [19]. This reaction leads to the restoration of the Soret peak of MauG and crosslinking of βTrp57 and βTrp108 in MADH. In our study, the change of iron redox state in heme of MauG was investigated to estimate the function of 7-hydroxyindole as a substrate toward MauG. The Soret peak was decreased when H₂O₂ was titrated to MauG, and the absorbance in the Soret region was recovered during the reaction with 7-hydroxyindole. This indicated that the diferric state of MauG was converted to the bis-Fe^IV state by H₂O₂, and the bis-Fe^IV state could be reduced to the diferric state by the addition of 7-hydroxyindole. The association rate was highly increased whereas the binding affinity was highly decreased, compared to that of the reaction in preTTQ. The binding affinity of 7-hydroxyindole toward the bis-Fe^IV state of MauG is stronger than that of 6-hydroxyindole and 5-hydroxyindole [13]. Furthermore, the rate increased with an increase in the ligand concentration. It should be noted that the bis-Fe^IV state of MauG can automatically reduce to the diferric state in the absence of preTTQ, but the conversion of bis-Fe^IV to diferric has a much slower transition rate than that in the presence of 7-hydroxyindole [25]. Our results suggest that 7-hydroxyindole can function as an alternative substrate because it can influence the valence state of iron in the heme of MauG. Additionally, the results represent the correlation between docking simulation and in vitro analysis. This demonstrates that molecular docking is a valuable tool to predict the unknown interaction between two different molecules.

In summary, the alternative substrate of MauG was investigated in the hope of utilizing the enzyme in various industrial applications. Among the tryptophan-like compounds, the candidates that could potentially function as alternative substrates were screened by docking using Autodock Vina. Additionally, the reactivity of MauG toward the candidate compound, 7-hydroxyindole, was demonstrated by measuring the spectra change of MauG using a UV/Vis spectrophotometer. In this study, we demonstrated the correlation between docking simulation and in vitro analysis. The association and dissociation rates for 7-hydroxyindole were demonstrated using the initial spectra conversion rate of MauG and our results suggest that hydroxyindole compounds can act as alternative substrates for MauG. Therefore, this study shows that MauG can be utilized for the oxidation of indole rings in various applications. In this study, it was not possible to distinguish whether the structure of the candidate compound was modified by MauG. Therefore, structural analysis using NMR is required further to identify the incorporation of the oxygen into the selected molecules by MauG.

Methods

In silico docking of tryptophan-like compounds to MauG

The interaction between MauG and tryptophan-like compounds was analyzed using Autodock Vina and PyMOL [26]. For the protein model, the MauG structure was obtained from the file ‘Crystal structure of the MauG/pre-Methylamine Dehydrogenase Complex (PDB code: 3L4M)’ in the Protein Data Bank (PDB). The structure in the file was a dimer form; therefore, it was modified to a monomer using UCSF Chimera [27]. The PDB file was converted to the PDBQT file format using Autodock tools program version 1.5.6. The docking grid was then determined to select the docking area in the protein (Table 5). Although the accurate location of preTTQ/TTQ binding site of MauG was previously demonstrated, the grid was selected to cover the whole MauG protein to investigate the binding pattern of the candidate compounds to MauG. The tryptophan-like compounds were selected and their binding, as alternative substrates, to MauG was investigated. The ligand files were obtained from PubChem (National Institutes of Health, US) as 3D files (Table 5), and each compound 3D file was converted to Mol2 file format using the Open Babel Package, version 2.3.1, http://openbabel.org [28]. To match the compound file format to the protein file, the Mol2 file was converted to the PDBQT file, using Autodock tools -1.5.6 program. The docking of each complex was performed by Autodock Vina, using Monte Carlo simulated annealing (SA) and Lamarckian Genetic Algorithm (LGA) to present the predicted values. Twenty different docking models were exported in the PDB file format in order of lowest energy value (ΔG°). The simulated models were visualized using the UCSF Chimera 1.11.2 program to confirm the overall binding state. Additionally, the electrostatic force and binding residue between protein and ligand were confirmed using the Ligplot + program [29], which can visualize the interactions such as hydrogen bond and hydrophobic interaction.

Table 5 The center and size of the 3D grid of MauG

Measurement of the reactivity of MauG to alternative substrates

Absolute ethanol was used to dissolve 7-hydroxyindole at a 10 mM concentration. The ligand solution was further diluted to 1 mM using 0.01 M potassium phosphate buffer at pH 7.5. MauG and 7-hydroxyindole were reacted in 0.01 M potassium phosphate buffer (pH 7.5) at 20 °C. The absorbance of 4.5 μM MauG was measured between 250 and 700 nm using UV/Vis-spectrophotometry. Next, 10 μM of H₂O₂ was added to the MauG, and the absorbance was measured with the same range of wavelength. When the peak at 406 nm had visibly decreased, varying concentration of 7-hydroxyindole was added to the solution and the absorbance at 406 nm was measured for 15 min at 1-s intervals. The observed rate was best fit to a single exponential. The limiting first-order rate constant for the reaction of MauG with 7-hydroxindole (k₃) and single turnover kinetic parameters were determined from the concentration dependence of the observed rate using Eq. (1).

$$k_{obs} = \frac{k_3 [S]}{{K_d + [S]}} + k_4$$

(1)

k_obs indicates an observed rate constant at which the association data was obtained to single association equation. K_d represents a substrate concentration at which the k_obs was half of k₃. k₄ is a dissociation rate constant and [S] indicates the concentration of 7-hydroxyindole.