Fisetin

Characterization of Regioselective Flavonoid O-methyltransferase from the Streptomyces sp. KCTC 0041BP

Sumangala Darsandhari, Dipesh Dhakal, Biplav Shrestha, Prakash Parajuli, Joo-Hyun Seo, Tae-Su Kim, Jae Kyung Sohng

ABSTRACT

A flavonoid comprises polyphenol compounds with pronounced antiviral, antioxidant, anticarcinogenic, and anti-inflammatory effects. The flavonoid modification by methylation provides a greater stability and improved pharmacokinetic properties. The methyltransferase from plants or microorganisms is responsible for such substrate modifications in a regiospecific or a promiscuous manner. GerMIII, originally characterized as a putative methyltransferase in a dihydrochalcomycin biosynthetic gene cluster of the Streptomyces sp. KCTC 0041BP, was tested for the methylation of the substrates of diverse chemical structures. Among the various tested substrates, flavonoids emerged as the favored substrates for methylation. Further, among the flavonoids, quercetin is the most favorable substrate, followed by luteolin, myricetin, quercetin glucoside, and fisetin, while only a single product was formed in each case. The products were confirmed by HPLC and mass-spectrometry analyses. A detailed NMR spectrometric analysis of the methylated quercetin and luteolin derivatives confirmed the regiospecific methylation at the 4′-OH position. Modeling and molecular docking provided further insight regarding the most favorable mechanism and substrate architecture for the enzymatic catalysis. Accordingly, a double bond between the C2 and the C3 and a single-ring-appended conjugate-hydroxyl group are crucial for the favorable enzymatic conversions of the GerMIII catalysis. Thus, in this study, the enzymatic properties of GerMIII and a mechanistic overview of the regiospecific modification that was implemented for the acceptance of quercetin as the most favorable substrate are presented.

Key words: O-methyltransferase, flavonoid, region-selective methylation, enzyme kinetics

1. Introduction

The natural products (NPs) that are derived from microbial or plant sources are a starting point for drug discovery, and they generally exhibit a wide range of pharmacophores and a high degree of stereochemistry [1]. These NPs comprise basic chemical skeletons, or modifications of their forms that have been produced by hydroxylation, glycosylation, and methylation [2]. Generally, methylation reactions are catalyzed by the O-methyltransferase (OMT) that catalyzes the transfer of the methyl group of S-(5′-Adenosyl)-L-methionine (SAM) to the hydroxyl groups of such NPs. Numerous OMTs that are either regioselective/stereoselective or substrate-promiscuous have been characterized from plants or microbial sources. Specifically, based on their size, amino-acid sequence, and cation dependency, plant OMTs are categorized into the following two major groups: Class I and Class II. The molecular-mass values of the Class-I OMTs (caffeoyl- coenzyme A OMTs [CCoAOMT]) from 26–29 KDa are lower, and this class methylate the lignin precursors viz caffeoyl-coenzyme A (CCoA) and 5-hydroxyferuloyl CoA in the presence of the magnesium (Mg) ion Mg2+. The molecular weights of the Class-II OMT (caffeic-acid OMT [COMT]) from 38-45 KDa, which can catalyze the methylation reaction of flavonoids as well as the 3-hydroxyl- and 5-hydroxyl-containing phenylpropanoid-derived lignin precursors without the presence of a metal cation, are higher [3]. The roles of these plant OMTs regarding the physiological and biochemical properties of plants are diverse. Similarly, the OMTs from microorganisms have been frequently characterized in the biosynthetic pathways of secondary metabolites, or they have been utilized for the biotechnological modifications of a number of compounds like flavonoids, alkaloids, and antibiotics. Consequently, an elevated interest in the understanding of the biological importance of microbial OMTs and the study of their enzymatic properties has emerged.

Flavonoids are polyphenol compounds with important medicinal values including antioxidant, anticarcinogenic, antimicrobial, anti-inflammatory, and antiviral effects [4]. Generally, it is thought that O-methylation plays an important role in the deactivation of the reactive hydroxyl groups of flavonoids and the alteration of their solubility and intracellular compartmentation. Although the structural diversity among flavonoids is naturally high, the production of novel or biologically potent flavonoid derivatives is a research-focus area. Numerous researches on both the synthetic approach [5, 6] and the metabolic engineering using OMTs derived from plant or microbial sources [7, 8] have been completed. The chemical synthesis for the O-methylation of a specific OH group is challenging, however, because of the requisite harsh catalytic conditions, a long reaction time, the protection of unwanted groups, and the monetary expense of the reagents. Alternatively, microorganisms are capable of the O-methylation of compounds for the production of novel bioactive compounds under much milder reaction conditions in a one-step reaction. These approaches for the generation of O-methylated compounds either by in vitro reactions or the whole-cell biotransformation have been successful for the production of valuable medical compounds, especially regarding the methylated derivatives of hydroxylated molecules. Due to the production capability of the versatile secondary metabolites, different Actinomyces spp. are being explored for the identification and characterization of the OMT genes that are involved in the methylation of diverse substrates. Due to the usage potential in the cosmetic, pharmaceutical, and food industries, flavonoids are commonly selected as substrates. The Streptomyces coelicolor OMT can modify flavonoids with a broad specificity [8]. Similarly, the identification of the methylation of various flavonoids at the 7-OH and the 4′-OH occurred in the in vitro and in silico studies of DnrK, an OMT of the Streptomyces peucetius doxorubicin- biosynthesis pathway [9]. The different OMTs that have been derived from the Streptomyces sp. are capable of the methylated modification of flavonoids, as listed in Table S1. Regioselectivity/sterioselectivity is the most desirable feature of any OMT, as this allows the targeted substrate to be modified for the attainment of the desired product. These types of enzymes are suitable for substrate modifications at the industrial scale, which requires rigorous screening and selection processes. SaOMT-2, which is from Streptomyces avermilitis, is a regioselective OMT that can convert different flavonoids into their corresponding 7-O- methylated form. SaOMT-2 was used for the biotechnological production of the antifungal compound sakuranetin [7].

Generally, flavonoids are subjected to autooxidation or are oxidatively degraded, while the methylation at the 4′-hydroxyl group of the B-ring of flavonoids confers a structural constraint for auto-oxidation; therefore, the importance of the regiospecific 4′-OMT is significant [10]. SOMT-2 (glycine max) has been reported as a 4′OMT [2] for which naringenin is the best substrate. Recently, the 4′OMT (Pa 4′OMT) of liverworts (plant source) was characterized, and it was used to effectively catalyze apigenin to acacetin [11]. A few of the OMTs from Streptomyces have been examined for the methylation of flavonoids, and to date, none of the Streptomyces methyltransferases (MTs) have been characterized as the regiospecific 4′OMT. GerMIII (GenBank accession no. AY118081) is present on the biosynthetic gene cluster of dihydrochalcomycin, a macrolide antibiotic that is produced by the Streptomyces sp. KCTC 0041BP. It has not been functionally characterized, but it has been proposed as a putative sugar OMT [12]. In this study, GerMIII was heterologously expressed in E. coli and was screened against the substrates of diverse chemical structures. The enzyme can accept flavonoids as a suitable substrate, so the reactions were performed with various flavonoids, thereby leading to the selection of quercetin as the most favorable substrate. Based on the studies on the enzymatic properties and the product characterization, GerMIII was established as a novel regiospecific flavonoid 4′OMT. GerMIII showed the best activity for the conversion of quercetin to tamarixetin, with the conversion of 85 % of 2 mM quercetin that was fed in vitro; this is the best conversion rate of any OMT for the 4′-OH methylation. Further, in silico modeling and docking studies were performed to obtain a deep insight regarding the interesting features of the GerMIII- modifying flavonoid in a regiospecific manner.

2. Materials and Methods

2.1 Chemicals and reagents

High performance liquid chromatography (HPLC)-grade acetonitrile, HPLC-grade trifluoroacetic acid (TFA) and water were purchased from Mallinckrodt Baker (U.S.A.). Isopropryl β-D-1- thiogalactopyranoside (IPTG) was provided by GeneChem Inc. (South Korea), while SAM chloride dihydrochloride, the flavonoids and the anthraquinones were purchased from Sigma- Aldrich (U.S.A.). The other chemicals were of a high-grade quality and were purchased from commercially available sources.

2.2 Bioinformatic analysis and cloning of GerMIII

The in silico analyses and comparisons of the nucleotide and protein sequences were performed using the Basic Local Assignment Search Tool (BLAST), developed by the National Library of Medicine (U.S.A.), and a FASTA (University of Virginia, U.S.A.) program, and the multiple sequence alignments were performed using the ClustalX 2.0 software (Conway Institute UCD Dublin, Republic of Ireland) [13]. A phylogenetic tree was generated using the neighbor-joining method for which the MEGA 7.0 software (MEGA, U.S.A.) was employed [14]. After a properanalysis, the gene was cloned for the heterologous expression in E. coli. All of the strains, plasmids, and polymer chain reaction (PCR) primers that were used in this study are presented in Table 1. The DNA preparation, digestion, ligation, and any other DNA manipulation, for E. coli were performed using the standard techniques. The pGEM®-T Easy Vector system (Promega Corporation, U.S.A.) was used to clone the PCR products. The DNA manipulation was carried out using E. coli XL1-Blue (Agilent Technologies, U.S.A.). E. coli BL21 (DE3), also sourced from Agilent Technologies (U.S.A.), was used as the host for the overexpression of the proteins. The genomic DNA from the Streptomyces sp. KCTC 0041BP was isolated using the DNeasy tissue kit (Quiagen, U.S.A.), and it was then used as a template for the PCR amplification of the GerMIII gene, for which the Pyrococcus furiosus (Pfu)-DNA polymerase was used under the following conditions: 30 cycles of a 1-min denaturation at 92 °C, a 1-min annealing at 58 °C, and a 1-min extension at 72 °C, while the initial denaturation was performed at 95 °C, and the final extension was performed for 10 min/step at 72 °C. The PCR restriction enzymes of the forward primer (GAATTCATGGTCAAGCACGCCCCGAAC) and the reverse primer (AAGCTTTCATTCGTGTCCCCCCGAAGA) are EcoRI and HindIII, respectively. The PCR products were subcloned into the pGEM-T Easy Vector system. The resultant plasmid was sequenced and the vector pET-28a(+) was used for the cloning and the subcloning, as listed in Table 1.

2.3 Expression and purification of the recombinant proteins

The recombinant plasmid pET28-GerMIII was formed by cloning the GerMIII gene into the pET28 vector. The gene was confirmed by the restriction of the enzymatic digestion, resulting in the transformation into E. coli BL21 (DE3). The host was named E. coli S, and for the protein expression, the E. coli S was cultured and grown at 180 rpm in Luria Bertani (LB) media that had been supplemented with 35 μg/mL of kanamycin at 37 °C until the OD600 reached 0.8; then, it was induced with IPTG with a final concentration of 0.4 mM and grown at 20 °C for 18 hr. The cells were collected using centrifugation (20 min, 3000 rpm), washed once under the cold condition at 4 °C with a 50-mM Tris-Cl buffer (pH 7.8) containing 100-mM sodium chloride (NaCl), and they were then suspended in the same buffer. Finally, the cells were disrupted by sonication. The cell debris was removed by a 20-min centrifugation at 10,000 g. The target protein was purified using a kit with the TALON Cobalt resin (Takara Bio Company, Japan) for the polyhistadine (His)-tag purification. The protein concentration in the samples was determined using the Bradford method [15], and the purified protein was preserved in -70 oC for further use. For the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, 12-% (w/v) polyacrylamide was used in the separating gel, and the protein bands were visualized using the Coomassie Brilliant Blue R-250 dye (Tokyo Chemical Industry, Japan).

2.4 Characterization of the methylation activity of GerMIII

To determine the activity of the GerMIII MT, in vitro enzymatic reactions were conducted in 50 ul of a reaction mixture. The reaction mixture consists of a 100-mM Tris buffer (pH 7.5), a 2 mM SAM, a 2 mM substrate, a 20 mM magnesium sulfate (MgSO4), and 1 mg/mL protein. The following substrates were among those that were used to determine the methylation activity: flavone (luteolin, apigenin); flavonol (quercetin, myricitin, morin, fisetin); flavanone (naringenin); isoflavone (daidzein); O-methylated isoflavone (formononetin); dihydrochalcone (phloretin); stilbenoid (resveratrol), querecetin glucoside; rutin; anthraquinones (alizarin, emodin, anthraflavic acid, purpurin, quinizarin, 2-amino-3-hydroxy anthraquinone); polyketides (erythromycin, nargenicin); naphthalene; 1,2-dihydro naphthalene; 1,6-dihydro naphthalene; 2 napthol; 1,2-dihydrobenzene; methyl 3,4,5-trihydroxy benzoate; resorcinol; and quinolone. The chemical structures of some of the flavonoids that were successfully methylated and used in this study are shown in Table 2. The reaction mixture was incubated for 2 hr at 37 oC, and it was finally quenched by the addition of a double volume of chilled methanol. The precipitated proteins were then removed by a 12000 rpm centrifugation at room temperature that lasted 30 min, after which time, the supernatant was collected for the product analysis.

2.5 Study of the effects of the temperature, cofactor, and pH on the enzymatic activity

The temperature profile for GerMIII was measured for 2 hr in the range between 4 and 75 oC using a Tris buffer (pH 7.5), 2 mM SAM, 20 mM magnesium chloride (MgCl2), and 2 mM quercetin, and the formed product was quantified using HPLC. A linear plot was then constructed between the product amount and the reaction temperature. The obtained graph shows the dependence of the enzymatic activity on temperature. Similarly, the effect of the pH on the activity of the enzyme was investigated using a Tris buffer in the pH region of 4.5–10.0 at 37 °C. To determine the metal-ion dependence of GerMIII, an activity test was carried out for the methylation of the 2-mM quercetin at 37 °C at a pH of 7.5 in the presence of different metal ions such as Mg2+, Mn2+, Ca2+, Cu2+, Pb2+, Fe2+, Zn2+, Co2+ and EDTA was used as negative control.

2.6 Study of the enzymatic kinetics of GerMIII

The kinetic analyses were performed using various SAM concentrations from 1-7 mM. The quercetin concentration is from 1-4 mM, and 20 mM MgCl2 and 100 mM Tris-hydrogen chloride (HCl) of a pH of 7.5 were used for each reaction for 2 hr at 37 oC. All of the reactions were terminated by the addition of a double volume of chilled methanol, followed by a centrifugation, and the supernatant was analyzed using HPLC. This experiment was repeated three times and the data were reported as the mean ± standard deviations. The values of the Vmax and the Michaelis– Menten constant Km were determined using the Lineweaver–Burk plot. In addition, the kcat of kcat = Vmax/(Et), where (Et) represents the initial enzymatic concentration, was also obtained [16].

3. Analytical Methods

Reverse-phase HPLC was performed at an ultraviolet (UV) absorbance of 312 nm using the Mightysil RP-18 GP C18 column (Kanto, U.S.A.) that is composed of a column size and length of 250 and 4.6 mm, respectively, and a mean particle size of 5 μm. The mobile phase consists of the solvent A, comprising water with 0.05-% trifluoroacetic acid (TFA), and the solvent B, comprising acetonitrile (C2H3N). The HPLC program with a flow rate of 1 ml/min is as follows: 10-% C2H3N at 0 min, 30-% C2H3N at 5 min, 50-% C2H3N at 10 min, 90-% C2H3N at 15 min, 70-% C2H3N at 18 min, and 10-% C2H3N at 25 min. The high-resolution quadruple time-of-flight electrospray ionization/mass spectrometry (HR-QTOF ESI/MS) analysis was performed in the positive-ion mode using the ACQUITY ultra-performance liquid chromatography (UPLC) water column (Waters, U.S.A.) coupled with the SYNAPT G2-S mass spectrometer (Waters, U.S.A.).

3.1 Purification and characterization of the methoxy derivatives

The purification of the enzymatically methylated compounds was performed using a preparative- HPLC (prep-HLPC) instrument equipped with the YMC-Pack ODS-AQ C18 column, with a 150 × 20 mm I.D. and a mean particle size of 10 μm (YMC America, Inc., U.S.A.), and a connected UV detector (312 nm); here, a 40-min binary program with the implementation of 10-% (0-5 min), 40-% (5-10 min), 40-% (10-15 min), 90-% (15-25 min), 90-% (25-30 min), and 10-% (30- 35 min) C2H3N at a flow rate of 10 mL/min was used. The purified products were dried, lyophilized, and dissolved in dimethyl sulfoxide (DMSO)-d6 solvent. Several nuclear magnetic resonance (NMR) methods including proton (1H) NMR, carbon-13 (13C) NMR, heteronuclear multiple quantum coherence (HMQC), and heteronuclear multiple bonded connectivity (HMBC) were carried out as needed.

3.2 Homology-modeling studies

With the utilization of the X-ray crystal structures of the Micromonospora griseorubida OMT (3SSM) as the templates, a three-dimensional (3D) multitemplate homology model was generated in the Discovery Studio 3.5 (DS 3.5) application (BIOVIA, U.S.A.) using the Build homology models of the MODELLER module (UCSF, U.S.A.) [17]. The 65 % sequence identity of GerMIII, with the M. griseorubida protein sequence (PDB code: 3SSM; resolution: 1.90 Å), is the highest [18]. Significant 3D similarities were exhibited between the generated model and its templates. The 3D homology modeling and the sequence analysis of GerMIII were performed as previously described [19], unless otherwise specified. The model was validated using the score from the Profiles 3D function and the ProStat (Poly Software International, U.S.A.) inspection of the ø and Ψ angles.

3.3 Cofactor and substrate docking

The validated 3D model of GerMIII was used for the docking and post-docking analyses. Hydrogen (H) atoms were first added to the 3D model, and then these added H atoms were minimized for a stable energy conformation and the relaxation of the conformation from the close contacts under the Chemistry at HARvard Macromolecular Mechanics (CHARMM) force field (CHARMM, U.S.A.) [20]. Iterative docking rounds were used to identify the optimum substrate-docked conformation. First, the SAM and substrate (quercetin) molecules were energy- minimized under the CHARMM force field, and then they were docked into the cofactor and substrate-binding pockets of GerMIII using the Dock Ligands (LigandsFit) approach (Dassault Systemes, U.S.A.), a molecular dynamics (MD) simulation-annealing-based algorithm module from DS 3.5 [21].
The SAM and magnesium (Mg) ions were then docked into the coenzyme-binding domain of the GerMIII model, and this was followed by the formation of the docking receptor molecule with its natural quercetin substrate by the energy-minimized enzyme complex, SAM, and Mg ions. Before the docking, however, the structures of the protein and the substrate, as well as their complexes, were subjected to an energy minimization using the CHARMM force field that was implemented in DS 3.5. Then, a full-potential final minimization was used to refine the substrate (ligand) orientation. The substrate orientation with the lowest interaction energy was chosen and used as the starting conformation for the next docking round until a similar docked-substrate orientation was observed between two consecutive docking-experiment rounds. The docked conformation of the substrate with the lowest energy for the postdocking analysis was retrieved from the Dock Ligands (LigandFit) module.

4. Results and Discussion

4.1 Bioinformatics studies of GerMIII

The open reading frame of GerMIII revealed its extent of 1,212 bp and a molecular mass of ~44 KDa; moreover, the BLAST analysis of GerMIII showed several homologous sequences in its sequence alignment with other OMTs. The primary GerMIII sequence and its sequence alignment with several OMTs are shown in Fig. 1 (A). The identity of the GerMIII MT with the OMT from Streptomyces bikiniensis is 95 %, its identity with the OMT from Streptomyces fradiae is 72 %, its identity with the OMT from Micromonospora griseorubida is 69 %, and its identity with the OMT from Saccharopolyspora pogona is 47 %. Its binding sites were predicted based on the sequence alignment and the comparison with the OMT from Micromonospora griseorubida, whose crystal structure is already available [18]. The GerMIII sequence alignments with the homologous sequences showed that the active sites tyrosine (Tyr) and histodine (His), the Mg2+, and the SAM-binding sites were conserved. The phylogenetic analysis that is shown in Fig. 1 (B) shows that GerMIII is closest to the OMTs from the Streptomyces sp. CB03578 (WP_0737965977.1), and also to those from the S. sp. Mg1 (WP_008741999.1).

4.2 Cloning, protein expression, and purification

The biosynthetic gene cluster of the Streptomyces sp. KCTC 0041BP has already been identified, but many biosynthetic genes are still not characterized [22]. GerMIII was previously characterized as a putative sugar-OMT and was heterologously expressed in E. coli after its cloning into the pET28 vector. The SDS-PAGE analysis of the purified enzyme showed a high purity (Fig. S1), so the purified fraction of the expressed GerMIII was proceeded for further studies on the enzyme characterization and the production of the methoxy derivatives of different substrates.

4.3 Determination of the GerMIII-substrate specificity

The purified protein of GerMIII was used for the in vitro enzymatic reactions with several substrates. The methoxide derivative of some of the substrates could be obtained, as shown in Table 2 and Fig. S5. Among all of the tested flavonoids, quercetin is the most preferred substrate with a conversion rate of 85 %, followed by luteolin (67 %), myricetin (20 %), quercetin glucoside (16 %), and fisetin (8 %), as shown in Table 3. Of all of the anthraquinones that were tested, only alizarin was methylated by the GerMIII at a detectable amount, whereas the methylation of all of the other remaining compounds was not evident. The common feature among the GerMIII substrates is the existence of two neighboring hydroxyl groups in the presence of a double bond between the C2 and the C3, as found in the quercetin, luteolin, myricetin, fisetin, and quercetin glucoside. GerMIII did not methylate many substrates including the close flavonoid relative catechin, which lacks the double bond between the C2 and the C3 (Table 2.).

4.4 Effects of the incubation temperature, pH, and cofactor on the enzymatic activity

The effects of the incubation temperature, pH, and cofactor dependence on the GerMIII activity for the quercetin conversion were observed. As shown in Fig. S2, GerMIII exhibited the highest methylation activity at 37 oC. A decrease in the methylation activity of approximately 64 % at 45 oC was observed, as can be seen in Fig. S2 (A). In the pH-dependence experiment, the GerMIII displayed a pH optimum of 7.5, as shown in Fig. S2 (B). Among the nine different divalent cations that were tested at 20 mM, Ni2+, Zn2+, Co2+, Cu2+, Pb2+, Mn2+, and Fe2+ rendered the GerMIII virtually inactive. The GerMIII activity was even observed in the absence of a cofactor, but its activity is less than half of that of the Mg2+-containing GerMIII, as is evident in Fig. 2 (C). It was suspected that the reaction in control may be mediated by divalent cations co-purified along with the protein because the medium used for cell growth is rich in divalent salts. This was proved by using EDTA for GerMII activity. There was a significant loss in the GerMIII activity (<2%) when EDTA (20mM) was used in the reaction. This clearly corroborated on the divalent dependency of GerMIII for its enzymatic reaction. Based on its dependency on the metallic- cofactor requirement for its reaction, it resembles the Class-II plant OMTs. 4.5 Kinetic-parameter studies The in vitro methylation for which different quercetin concentrations were used for the substrate was performed using the purified enzyme at 37 oC with a pH of 7.5; this was followed by the use of different SAM concentrations. The data concerning the apparent Km, Vmax, and kcat are shown in Table 4. The Km and Vmax of quercetin are 162.81 (μM) and 160.56 (µM min-1mg-1), respectively, and the Km and Vmax of SAM are 1.83 μM and 64.8 (µM min-1mg-1), respectively. The Kcat values of quercetin and SAM are 0.027 and 0.011 s-1, respectively. The Vmax/Km values are approximately 1, thereby indicating that the flavonoid is not preferable as an endogenous substrate, unlike the plant OMTs where the values are far higher than 1 [23]. An approximate 85 % conversion of the flavonoids in a regiospecific manner, however, is a very beneficial feature for the target modification in terms of flavonoids. 4.6 Analysis and characterization of the 4′-O-methoxy products The HPLC chromatogram of the quercetin reaction shows the retention time of the standard quercetin at 11.9 min and a new peak at the retention time of 13.4 min under a 320-nm UV absorbance (Fig. 2A). Likewise, the HPLC chromatogram of the luteolin reaction shows a single new peak at the retention time of 10.7 min, while the luteolin was also observed at the retention time of 9.5 min (Fig. 2B). The new peaks were analyzed using electrospray ionization-mass spectrometry (ESI-MS)/mass spectrometry (MS). The positive-ion ESI-MS/MS-spectra analysis of the peaks showed total-mass values of (M + H) + m/z = 317.0658 and (M + H) + m/z = 301.0703, which resemble the total molecular weights of the methylated quercetin and luteolin, respectively. Similarly, the methylated products of a number of other flavonoids were also obtained (Fig. S5). To reveal the methylated position of the GerMIII-formed product, and for a structural elucidation of the product, the compounds were purified from the crude extract of the in vitro reaction using prep-HPLC. The new peaks that were observed in the quercetin and luteolin reactions were collected using prep-HPLC. The purified products were then freeze-dried and dissolved in 400 μL of the DMSO-d6 for an NMR analysis (Figs. S3 and S4). The structure of the 4′-methylated quercetin was confirmed using various NMR methods including the one-dimensional (1D) NMR (1H-NMR and 13C-NMR) and two-dimensional (2D) NMR (HMBC and HSQC) analyses, as shown in Fig. S3. The single -OCH3 spectra are visible in both the proton- and carbon-NMR results at 3.85 ppm (3H, s) and 56 ppm, respectively, while the cross-peak of this O-methyl group was found with the quercetin 4′-hydroxy position at 149 ppm, as indicated by the HMBC and HSQC correlations; this confirmed that the product is 4′-O-methyl-quercetin. Similarly, by comparing the 1H-NMR, 13C-NMR, and HMBC data to the standard diosmetin data, it was confirmed that the methylated position of luteolin is 4′-O-methoxy-luteolin (Fig. S4). The 4′-O- methoxy-quercetin and the 4′-O-methoxy-luteolin are known as tamarixetin and diosmetin, respectively. Based on these observations, it was verified that GerMIII is a regioselective O-MT. 4.7 Mechanistic studies by modeling and docking The regiospecific preference of the flavonoid, quercetin, is particularly interesting, and this prompted the attainment of a detailed insight regarding the protein structure and the molecular mechanism for the favoring of such regiospecific reactions. The GerMIII model was generated in the DS 3.5 application. The model was validated according to the consistency of the model using the validation tool PROCHECK (EMBL-EBI, U.K.) [24]. The calculated Ramachandran plots suggested that 91.7 %, 5.8 %, and 2.5 % of the residues in the derived model are in the favored, allowed, and disallowed regions, respectively. The built model was also evaluated according to its superimposition onto the template crystal structures, while the active-site residues of GerMIII were effectively superimposed upon the coenzyme-binding motif. The molecular-docking experiment was carried out using the generated GerMIII model, with the utilization of the quercetin, SAM, and Mg2+ as ligands. The Mg2+ comprises an octahedral coordination environment that is suitable for the two aspartic-acid residues Asp279 and Asp308 and two water molecules (not shown). The two free coordination sites are occupied by the two quercetin hydroxyl groups, 3′-OH and 4′-OH. In the case of Mg2+-dependent OMTs, the Mg ions bind the two hydroxyl groups of the orthophenol substrate and help in the placement of the substrate into the exact position of SAM, commonly referred to as AdoMet (C15H22N6O5S+) [25]. Thus, the Mg2+ is responsible for “anchoring” the substrate (3'- and 4'-hydroxyl group positions) molecules in catalytic positions. The two quercetin hydroxyls form additional H bonds with the side chains of the residues His282 and Asp308. In the case of luteolin and myricetin, they are maintained in the same ring, as well as the conservation of the conjugate hydroxyl groups (3'- and 4'-hydroxyl group positions), which is consistent with the preferred methylation of such substrates. In the case of the representative flavonoids that retain two hydroxyl groups (such as fisetin and morin) in the same ring, but do not retain the 3'- and 4'-position pairings, the methylation is detectable. Regarding the other flavonoids, for which such hydroxyl pairs are not retained in the same ring, they are not preferable for methylation. Interestingly, in the case of quercetin glucoside, a methylated derivative was obtained as well, thereby signifying that the attachment of the other group at the next hydroxyl position, and besides the retention of the 3'- and 4'-hydroxyl pairs in the same ring, does not disturb the anchoring of the substrates in catalytic positions. This feature provides new avenues for the structural modification of the diverse derivatives of the selected flavonoids. Generally, the SAM-dependent OMTs employ acid-base chemistry to deprotonate the hydroxyl methyl that is the acceptor prior to the attack of the SAM methyl group. A distance of 2.7 Å was observed between the nitrogen atoms of the imidazole ring of the catalytic His282 and the quercetin 4′-hydroxyl group, while the latter is sufficiently close (2.7 Å) to the SAM methyl group (Figs. 3 and 4). His282 acts as a general base, removing the proton from the 4′-OH and leaving a hydroxylated anion, which is stabilized by the interactions with the Mg cation. Next, the 4′-OH extracts the methyl group from SAM to complete the reaction. Therefore, the most favorable configuration between histidine (His278) and quercetin is the one in which the quercetin 4′-OH group is positioned close to the donor of the SAM methyl group. Apart from the substrate-binding motif and the active tetrad of tyrosine-asparagine-histidine-aspartic acid (Tyr212–Asn279–His282–Asp308), was conserved in the GerMIII 3D model. The molecular docking of quercetin into the GerMIII modeled structure showed that the distance between the nitrogen atoms of the imidazole ring of the catalytic histidine (His282), which transfers its hydride to the quercetin during the course of the reaction, and the quercetin 4'-OH group is 2.7 Å, which is suitable for a hydride transfer. Further, the distance between the quercetin 4'-OH group and the SAM methyl group is also 2.7 Å. Based on these observations, a mechanistic approach for the regioselective modifications of flavonoids like quercetin was predicted (Fig. 4). 5. Conclusions It is established that methylation can be crucial enzymatic modification of diverse NPs like flavonoids for increasing their stability, promoting the pharmaceutical values and improving their biological/chemical properties. The regioselective MT has been particularly interesting for the diversification of the desired substrate. GerMIII was previously reported as a putative MT in terms of the biosynthetic gene cluster of dihydrochalcomycin, a macrolide compound produced by the Streptomyces sp. KCTC 0041BP. From this study it has been established as the flavonoid MT that is capable of adding methyl group to 4′-OH in a regioselective manner. Generally, flavonoids with methylation at 4′-OH, tend to be biologically potent and chemically stable. Hence, this MT can be utilized for the production of these diverse bioactive molecules and their derivatives. The detailed study that utilized modeling and docking provided insight regarding the preserved environmental condition that is required in the active site for methylation reactions. It became evident that the molecules of the single-ring conjugate-hydroxyl group can be successfully methylated at the region-selective 4′-OH position, which is key substrate-architecture information that is required for such modifications. The knowledge about the enzymatic mechanism and the substrate architecture can be utilized to devise strategies for the increasing of the enzymatic activities for which site-directed mutations or enzymatic evolutions can be employed. In addition, diverse chemical and biological approaches can be utilized using this enzyme for the creation of libraries of new analogs with enhanced pharmaceutical properties. Acknowledgments This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (NRF-2017R1A2A2A05000939). Reference 1. D. Dhakal, J.K. Sohng, Coalition of biology and chemistry for ameliorating antimicrobial drug discovery, Front Microbiol. 8 (2017) 734. 2. D.H Kim, B.G. Kim, Y. Lee, J.Y. Ryu, Y. Lim, H.G. Hur, J.H. Ahn, Regiospecific methylation of naringenin to ponciretin by soybean O-methyltransferase expressed in Escherichia coli, J Biotechnol. 119 (2005) 155-162. 3. R. Edwards, A.D Richard, Purification and characterization of S-adenosyl-L-methionine: caffeic acid 3-O-methyltransferase from suspension cultures of alfalfa (Medicago sativa L.), Arch Biochem Biophys. 287 (1991) 372-379. 4. L.H. Cazarolli, L. Zanatta, E.H. Alberton, B. Figueiredo, M.S. Reis, P. Folador, R.G. Damazio, M.G. Pizzolatti, B. Silva, F.R. Mena, Flavonoids: prospective drug candidates, Mini Rev Med Chem. 8 (2008) 1429-1440 5. K. Imai, I. Nakanishi, K. Ohkubo, Y. Ohba, T. Arai, M. Mizuno, S. Fukuzumi, K.I. Matsumoto, K. Fukuhara, Synthesis of methylated quercetin analogues for enhancement of radical-scavenging activity, RSC Advances. 7 (2017) 17968-17979. 6. N.G. Li, Z.H. Shi, Y.P. Tang, J.P. Yang, T.L. Lu, F. Zhang, Y.W. Huang, Z.J. Wang, J.A. Duan, Synthetic studies on the construction of 7-O-methylquercetin through regioselective protection and alkylation of quercetin, Chin Chem Lett. 22 (2011) 5-8. 7. B.G. Kim, B.R. Jung, Y. Lee, H.G. Hur, Y. Lim, J.H. Ahn, Regiospecific flavonoid 7-O- methylation with Streptomyces avermitilis O-methyltransferase expressed in Escherichia coli, J Agric Food Chem. 54 (2006) 823-828. 8. Y. Yoon, Y.S. Yi, Y. Lee, S. Kim, B.G. Kim, J.H. Ahn, Y. Lim, Characterization of O- methyltransferase ScOMT1 cloned from Streptomyces coelicolor, Biochim. Biophys. Acta. 1730 (2005) 85-95. 9. N.Y. Kim, J.H. Kim, Y.H. Lee, E.J. Lee, J.Y. Kim, Y. Lim, Y. Chong, J. Ahn, O- Methylation of flavonoids using DnrK based on molecular docking. Journal of microbiology and biotechnology, 17(12) (2007) 1991. 10. J.P. Spencer, G.G. Kuhnle, R.J. Williams, R.E. Catherine, Intracellular metabolism and bioactivity of quercetin and its in vivo metabolites, Biochem J. 372 (2003) 173-181. 11. H. Liu, R.X. Xu, S. Gao, A.X. Cheng, The Functional Characterization of a Site-Specific Apigenin 4′-O-methyltransferase synthesized by the Liverwort Species Plagiochasma appendiculatum, Molecules. 22 (2017) 759. 12. T.T.T. Thuy, K. Liou, T.J. Oh, D.H. Kim, D.H. Nam, J.C. Yoo, J.K. Sohng, Biosynthesis of dTDP-6-deoxy-β-d-allose, biochemical characterization of dTDP-4-keto-6- deoxyglucose reductase (GerKI) from Streptomyces sp. KCTC 0041BP, Glycobiology. 17 (2006) 119-126. 13. J.D. Thompson, T.J. Gibson, F. Plewniak, F. Jeanmougin, D.G. Higgins, The CLUSTAL_ X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25 (1997) 4876-4882. 14. S. Kumar, G. Stecher, K. Tamura, MEGA7: Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets, Mol Biol Evol. 33 (2016) 1870-1874. 15. N.J. Kruger, The Bradford method for protein quantitation. Basic protein and peptide protocols. Methods, Mol Biol. 32 (1994) 9-15. 16. R. Copeland, Enzyme: a practical introduction to structure, mechanism and data analysis, second ed., Wiley, UK, 2000. 17. A. Šali, L. Potterton, F. Yuan, H. van Vlijmen, M. Karplus, Evaluation of comparative protein modeling by MODELLER. Proteins, 23 (1995) 318-326. 18. D.L. Akey, S. Li, J.R. Konwerski, L.A. Confer, S.M. Bernard, Y. Anzai, F. Kato, D.H. Sherman, J.L. Smith, A new structural form in the SAM/metal-dependent o‑methyltransferase family: MycE from the mycinamicin biosynthetic pathway, J Mol Biol. 413 (2011) 438-450. 19. M.K. Tiwari, H.J. Moon, M. Jeya, J.K. Lee, Cloning and characterization of a thermostable xylitol dehydrogenase from Rhizobium etli CFN42, Appl Microbiol Biotechnol. 87 (2010) 571-581. 20. B.R. Brooks, R.E. Bruccoleri, B.D. Olafson, D.J. States, S.A. Swaminathan, M. Karplus, CHARMM: a program for macromolecular energy, minimization, and dynamics calculations, J Comput Chem. 4 (1983) 187-217. 21. G. Wu, D.H. Robertson, C.L. Brooks III, M. Vieth, Detailed analysis of grid-based molecular docking: A case study of CDOCKER-A CHARMm-based MD docking algorithm, J Comput Chem. 24 (2003) 1549-1562 22. B.B. Pageni, D. Simkhada, T.J. Oh, J.K. Sohng, Biosynthesis of dihydrochalcomycin: characterization of a deoxyallosyltransferase (gerGTI), Mol Cells. 29 (2010) 153-158. 23. Y. Yoon, Y. Park, Y. Lee, Y.S. Yi, G. Jo, J.C. Park, J.H. Ahn, Y. Lim, Characterization of an O-methyltransferase from Streptomyces avermitilis MA-4680, J Microbiol Biotechnol. 20 (2010) 1359-1366. 24. R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, Fisetin ,PROCHECK: a program to check the stereochemical quality of protein structures, J Appl Crystallogr. 26 (1993) 283-291.
25. L. Hoffmann, S. Maury, M. Bergdoll, L. Thion, M. Erard, M. Legrand, Identification of the enzymatic active site of tobacco caffeoyl-coenzyme A O-methyltransferase by site- directed mutagenesis, J Biol Chem. 276 (2001) 36831-36838.