Figures
Abstract
Flavonoids are a class of plant secondary metabolites that act as storage molecules, chemical messengers, as well as participate in homeostasis and defense processes. They possess pharmaceutical properties important for cancer treatment such as antioxidant and anti-tumor activities. The drug-related properties of flavonoids can be improved by glycosylation. The enzymes glycosyltransferases (GTs) glycosylate acceptor molecules in a regiospecific manner with the help of nucleotide sugar donor molecules. Several plant GTs have been characterized and their amino acid sequences determined. However, three-dimensional structures of only a few are reported. Here, phylogenetic analysis using amino acid sequences have identified a group of GTs with the same regiospecific activity. The structures of these closely related GTs were modeled using homologous GT structures. Their substrate binding sites were elaborated by docking flavonoid acceptor and UDP-sugar donor molecules in the modeled structures. Eight regions near the acceptor binding site in the N- and C- terminal domain of GTs have been identified that bind and specifically glycosylate the 3-OH group of acceptor flavonoids. Similarly, a conserved motif in the C-terminal domain is known to bind a sugar donor substrate. In certain GTs, the substitution of a specific glutamine by histidine in this domain changes the preference of sugar from glucose to galactose as a result of changed pattern of interactions. The molecular modeling, docking, and molecular dynamics simulation studies have revealed the chemical and topological features of the binding site and thus provided insights into the basis of acceptor and donor recognition by GTs.
Citation: Sharma R, Panigrahi P, Suresh C (2014) In-Silico Analysis of Binding Site Features and Substrate Selectivity in Plant Flavonoid-3-O Glycosyltransferases (F3GT) through Molecular Modeling, Docking and Dynamics Simulation Studies. PLoS ONE 9(3): e92636. https://doi.org/10.1371/journal.pone.0092636
Editor: Annalisa Pastore, National Institute for Medical Research, Medical Research Council, London, United Kingdom
Received: November 25, 2013; Accepted: February 23, 2014; Published: March 25, 2014
Copyright: © 2014 Sharma et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: 1) Council of Scientific and Industrial Research (CSIR), New Delhi. 2) CSIR-National Chemical Laboratory Centre of Excellence in Scientific Computing (CoESC). The funding agencies have no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Glycosyltransferases (GTs; EC 2.4.x.y) belong to the transferase family of enzymes which catalyze the transfer of sugar moiety from an activated nucleotide sugar donor molecule to a saccharide or nonsaccharide acceptor substrate to synthesize oligosaccharides, polysaccharides or glycoconjugates [1] (Figure 1). Sugar donors are activated nucleotide molecules possessing a variant sugar moiety, and an invariant UDP group. The acceptor molecules are generally flavonoids consisting of flavanones, flavanonol, flavans, flavones, flavonols, and anthocyanidins. Flavonoids, a class of plant secondary metabolites, play a prominent role as anti-cancer and antioxidant agents [2]. UDP-glycosyltransferases (UGTs) are GTs which transfer the glycosyl group (glucose, galactose, xylose, rhamnose, etc) from a nucleoside diphosphate sugar donor (UDP-sugar) to a diverse array of acceptor substrates [3]. In addition to the specific sugars they transfer, some UGTs are also highly specific to acceptor molecules and they catalyze glycosylation of only one or two types of molecules [4], whereas some others catalyze glycosylation of a broad range of acceptors [5]. Flavonoids possess a polyphenolic structure bearing multiple -OH groups (7-OH, 5-OH, 3-OH, 4′-OH etc) on the phenyl rings (Figure 1). Some UGTs specifically glycosylate only one of these -OH groups, whereas others glycosylate multiple -OH groups.
Reaction catalyzed by UGT is shown with Kaempferol and UDP-glucose as substrates that react to form Kaempferol-3-O-glucoside along with the release of UDP moiety. The numbering of the OH groups- 7-OH, 5-OH, 3-OH and 4′-OH in flavonoids is shown.
On the basis of amino acid sequence similarity, GTs have been classified into 94 CAZy families [6]. The amino acid sequence is poorly conserved across GTs [7], [8]. However, despite lack of sequence conservation, available three-dimensional crystal structures of plant UGTs show highly conserved secondary and tertiary structures [8], [9], [10]. Only two types of structural folds of GTs have been identified, namely GT-A and GT-B. Both have a Rossmann-like fold architecture, a characteristic feature of nucleotide binding proteins. The Rossmann fold has an arrangement of alternate β/α/β secondary structure elements which form a sandwich structure with nucleotide binding site between the layers [5] (Figure 2). The GT-A fold has a single Rossmann domain that requires a divalent cation such as Mg2+ or Mn2+ for its catalytic activity [11]. The GT-B fold has two Rossmann domains connected by a linker region, and the substrate binding pocket is in a narrow cleft between the domains. The C-terminal domain (CTD) of GT-B proteins has a long stretch of 44 amino acid residues called Plant Secondary Product Glycosyltransferase [12] or Putative Plant Secondary Glycosyltransferase (PSPG) motif [3] (Figure S1). This motif plays a crucial role in sugar recognition by facilitating the binding of a sugar donor molecule in the active site. For example, in the Vitis vinifera UGT crystal structure (PDB-2C1Z), out of the ten residues interacting with UDP-sugar, seven interact with the invariant part of the donor. The remaining three interact with the variant part of the donor. The conserved Trp residue is the 22nd amino acid of the PSPG motif, while the last two residues, 43rd and 44th amino acids of PSPG, are Asp/Glu/Ser and Gln/Glu/His/Asn. These last two residues are aspartic acid (D) or glutamic acid (E) in the first place and glutamine (Q) at the second place in four available plant GT structures. The residues in the above three sites W, D/E, and Q are positioned to form hydrogen bonds with the sugar part of the donor [13]. The remaining amino acids of PSPG box, which are not involved in the protein-ligand interactions, help in maintaining the 3-D structure of the protein.
Topology diagram of UGT from F. ananassa shows CTD and NTD having an architecture of Rossmann fold. An arrangement of alternate β/α/β secondary structure elements can be seen in the diagram along with the key catalytic and binding site residues of acceptor and sugar donor shown by square mark and labeled.
Contrary to the existence of a conserved CTD binding site, there are pronounced variations in the loops and helix regions of the N-terminal domain (NTD) near the acceptor binding site, presumably to accommodate a wide range of acceptor molecules. Even the reaction mechanism of these enzymes is conserved. The transfer of sugar to an acceptor substrate occurs by an oxycarbenium ion formation. It catalyzes the transfer of glucose moiety from a sugar donor to an acceptor molecule with a direct displacement SN2 mechanism. The conserved histidine residue, close to the acceptor and sugar donor, acts as a catalytic base that facilitates the deprotonation of the acceptor molecule (Figure S1, Region N1). The deprotonated acceptor, which turns into a nucleophilic oxyanion, attacks the C1 carbon of the sugar and forms the respective β-glucosides. A neighboring aspartic acid residue interacts with the conserved histidine by forming a hydrogen bond and helps to balance the charge on the histidine subsequent to deprotonation of the acceptor substrate [3] (Figure S1, Region N4).
A phylogenetic, sequence and structural analysis of GT-B family of enzymes from various plant sources has been carried out to identify motifs and crucial amino acids which play an important role in the binding and recognition of specific acceptors, also to identify amino acid residues involved in maintaining the 3-D structure.
Materials and Methods
Sequence Comparison and Phylogenetic Analysis
Protein sequences of plant GTs of sequence length 430–480 amino acids were retrieved from UniProtKB/Swiss-Prot database (http://www.uniprot.org/) using query search. A dataset of 101 protein sequences of plant GTs was obtained. Protein sequences were initially aligned using MUltiple Sequence Comparison by Log- Expectation (MUSCLE) [14] employing the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) based clustering method with gap open and gap extension penalties set to −2.9 and 0. This alignment was used to build the phylogenetic tree using the Neighbor Joining (NJ) algorithm of Molecular Evolutionary Genetics Analysis 5.1 (MEGA) [15] with the Dayhoff substitution matrix (PAM250) and bootstrap value set to 1000. A well defined cluster for flavonoid-3-O-GT (F3GT) was observed in the dendrogram comprising thirty protein sequences. Multiple sequence alignment (MSA) of these thirty F3GT protein sequences was done to analyze the extent of conservation among the sequences using the ClustalX program [16]. A percent identity matrix of F3GTs was prepared with the help of ClustalX program.
Molecular Modeling of F3GT Proteins
A search using Basic Local Alignment Search Tool (BLAST) algorithm [17] was carried out against the Protein Data Bank (PDB) [18] to select the high resolution crystal structures of homologous proteins. The sequence identity cut off was set to ≥30% (E-value cut off = 1). Homology modeling of F3GT protein sequences was carried out using the Composite/Chimeric model type of Prime 3.1 [19], by taking structures homologous to the target proteins as templates, in order to study their structural features, binding mode and affinity with the substrates.
Model Validation and Refinement
The initial models were evaluated for the stereochemical quality of the protein backbone and side chains using PROCHECK [20]. Environment of the atoms in the protein model was checked by ERRAT server [21]. Verify3D plot [22] showed the compatibility of the 3-D structure with respect to the protein sequence. Errors in the model structures were also checked with ProSA server [23]. After model validation, initial models were refined using impref minimization of protein preparation wizard [19] and Impact 5.8 [19] minimization. These energy minimized final models were further used for the binding studies with their substrates.
Ligand Design, Protein and Ligand Preparation
The 2-D sketcher utility of Maestro 9.3 was used to design the structures of ligand molecules in two-dimensions, which were then converted into three-dimensional structures in mol2 format. Before proceeding with the docking studies, water and other hetero atom groups from the protein structures were removed using protein preparation utility of Maestro. Hydrogens were added subsequently to carry out restrained minimization of the models. The minimization was done using impref utility of Maestro in which the heavy atoms were restrained such that the strains generated upon protonation could be relieved. The root mean square deviation (RMSD) of the atomic displacement for terminating the minimization was set as 0.3 Å. Similarly, ligands were refined with the help of LigPrep 2.5 [24] to define their charged state and enumerate their stereo isomers. The processed receptors and ligands were further used for the docking studies using Glide 5.8 [25].
Modeling Enzyme-ligand Interactions
Reported wet lab experiments have shown that some of the F3GT enzymes have high binding affinity and specificity for 3-OH group of flavonoids, more specifically flavonols and anthocyanidins, which include kaempferol (kmp), myricetin, quercetin, delphinidin and many more. Therefore, kmp was docked in the acceptor binding site cavity of the prepared receptor molecules. A grid was made either by taking the reference ligand or by selecting the active site residues involved in the binding of the substrate. Flexible ligand docking was carried out using the standard precision option. The sugar donor was docked in the active site by creating a grid around the bound reference ligand i.e. U2F (Uridine-5′-Diphosphate-2-deoxy-2-fluoro-alpha-D-glucose). Final products of the reaction were also docked in the V. vinifera GT crystal structure. A total of 20 poses generated were scored on the basis of their glide score and E-model values. The hydrogen bond interactions between the protein and ligands were visualized using PyMOL [26].
Molecular Dynamics Simulations to Assess the Stability of Docked Complexes
The docked complexes were subjected to molecular dynamics simulations using the GROningen Machine for Chemical Simulations V4.5.4 (GROMACS) [27], [28]. GROMOS96 43a1 force field was applied on 31 docked complexes placed in the centre of the dodecahedron box solvated in water. Topology files and other force field parameter files for the ligands were created using PRODRG2 server [29]. Protein was immersed in dodecahedron water box of SPC216 water model with distance between the solute and the box was set to 10 Å. The initial simulation cell dimensions set up were approximately 90×90×90 Å for all the simulations (Table S1). The system was initially energy minimized by steepest descent minimization for 50,000 steps until a tolerance of 10 kJ/mol to avoid the high energy interactions and steric clashes. Total negative charges on the docked structures were balanced by suitable number of Na+ ions to make the whole system neutral using genion program of GROMACS. After adding ions, the system was again energy minimized by steepest descent minimization retaining the same parameters. The V-rescale, a modified Berendsen thermostat, temperature coupling [30] and Parrinello-Rahman pressure coupling [31] methods were used to keep the system stable at 300 K temperature and pressure of 1 bar. The Particle Mesh Ewald (PME) method [32] was selected to compute long range electrostatic interactions. A cut off distance of 9 Å and 14 Å was set for Coulombic and van der Waals interactions. Rotational constraint was applied to bonds by LINCS algorithm [33].
The analysis was performed on the ensemble of system configurations extracted at 2 ps time intervals from the simulation and was equilibrated for 2 ns time period, to guarantee a completely equilibrated system. The analysis was carried out relative to the last 3 ns time period on the C-alpha atoms of the structures. No positional constraints were applied on the system. Periodic boundary conditions were applied in all three directions. The trajectories were visualized using Visual Molecular Dynamics program (VMD) [34].
Results and Discussion
Amino Acid-based Phylogenetic Analysis of GTs
After creating a dataset of 101 protein sequences of plant GTs, a dendrogram was built using neighbor joining method of MEGA in order to study their evolutionary relation (Table S2). Bootstrapping (value: 1000) was performed to examine how often a particular cluster in a tree appeared on re-sampling amino acids.
The GT sequences clustered into few distinct groups in the output of phylogenetic analysis. Out of them, the maximum number of GTs clustered in the F3GT specific clade with a significantly high bootstrap value (Figure 3). The members of the remaining clusters showed mixed specificity/activity for different -OH groups of acceptors. Here, the focus of study is the F3GT specific cluster, as the members of this group are closely related to each other showing high sequence identity and present in significant numbers. The percentage identity matrix revealed a significant level of conservation with approximate 40% sequence identity among them (Table S3).
Dendrogram of 101 UGT protein sequences shows distinct clusters of F3GT members marked with an arrow. The underlined taxon names are annotated to be flavonoid-3-OH glycosyltransferases and they form a separate clade.
Molecular Models of F3GT Proteins
The molecular structures of F3GT proteins were modeled in order to study, from a structural point of view, the role of conserved regions in the close vicinity of acceptor and their interactions that favor formation of protein-ligand complex. BLAST search against PDB identified homologous structures to be used as templates to build the homology models (Table S4). Crystal structures of GTs from V. vinifera (PDB-2C1Z), Medicago truncatula (PDB-3HBF) and Clitoria ternatea (PDB-3WC4) of resolution 1.90 Å, 2.10 Å and 1.85 Å, respectively, shared an identity of ≥40% (except Iris_2 and Hordeum) and query coverage of ∼>90% with the target proteins, and were thus used for modeling studies. Secondary structure prediction by Psipred showed the similarity between the templates and the target with respect to the arrangement of secondary structure elements. Very few gaps were present in the alignment of the target and template sequences (Figure S2). The bound ligands of 2C1Z i.e., kmp and U2F were incorporated in the target models during the modeling procedure. Few N-terminal residues (4–5 amino acids) couldn’t be modeled for some of the target sequences because of the absence of corresponding regions in the templates (Missing residues in 2C1Z: 1–5, 3HBF: 1–11).
Model Validation and Refinement
The initial models were evaluated for their stereochemical parameters. More than 99% of all residues were in the allowed regions of Ramachandran plot. The goodness factor of the models, a log odd score calculated based on the stereochemical parameters, were in the range of 0 to −0.4. Verify3D plot for the models showed that more than 95% of the residues had a positive 3D-1D averaged score, which revealed that the models were folded correctly. RMSD of the C-alpha atoms of the models and the templates were close to 1 Å. The z-score value calculated by ProSA server is within the z-scores of experimentally determined PDB structures. The ERRAT results also showed that the overall quality score of the models was good (Table 1).
The model refinement phase involved preprocessing the initial models by adding hydrogens, assigning bond order, and filling missing loops and side chains. Next, the models were subjected to restrained minimization by applying the constraint to converge the non-hydrogen atoms to an RMSD of 0.3 Å using OPLS 2005 [35] force field. After that, the models were further subjected to 500 steps of steepest descent energy minimization followed by 1000 steps of conjugate gradient energy minimization using the same force field. These energy minimized models were further used for docking and molecular dynamics studies.
Sequence Conservation and Structural Integrity of F3GT Enzymes
MSA of the F3GT proteins was carried out using ClustalX program (Figure S1). Sequence alignment and construction of percent identity matrix showed higher sequence conservation of both CTD and NTD (Figure S1 and Table S3). The residue numbers used in this analysis are with respect to the sequence of transferase from V. vinifera source (PDB-2C1Z). An analysis of the conserved residues using the V. vinifera GT structure revealed that these residues interacted through hydrophobic, π-π stacking and hydrogen bonding interactions for maintaining the required framework of interacting residues in the binding site.
One key conserved residue involved in catalysis is His20, which steers the deprotonation of acceptor substrates. An important interaction near the binding site is the side chain-side chain hydrogen bonding interaction between His9 and Ser41, which in turn helps Phe42 to make a π-π interaction with Phe53. Similarly, Phe40 is surrounded by hydrophobic amino acids of the adjacent α-helix and β-sheets that provide stability to the folded structure. Pro16 present near the acceptor binding site provides rigidity to the binding site loop and helps in orienting the neighboring residues Phe15 and Phe17 to provide a hydrophobic environment to the acceptor. Similarly, the conserved Gly72 residue provides flexibility to a binding site loop. Another important interaction involves residue Pro74 that orients the side chain of Glu75, making it accessible to the solvent. Phe90 is surrounded by a series of hydrophobic and aromatic amino acids such as Phe15, Pro16, Phe17, Ile69, Phe98, etc. Side chain of Asn49 interacts with Ser46 and Asp68 of neighboring β-sheet. The interaction of Asp119 with the key catalytic residue His20 balances its acquired charge after deprotonation of the acceptor. Phe121 provides hydrophobic environment to one of the phenyl rings of the acceptor molecule. Conserved Trp123 and Phe124 are involved in π-π interaction while Pro95 forms a stacking interaction with Phe124. The indole ring of Trp140 interacts with Gly143 and Ser146, its aromatic ring providing a hydrophobic environment to the acceptor substrate. His150 being a polar amino acid provides a hydrophilic environment to the 4′-OH group of the flavonoid acceptor. Arg157 connects the other loop by making hydrogen bonds with residues Gly190 and Ile191 and also interacting with the solvent molecule. Leu204 also contributes to the hydrophobicity of the acceptor binding pocket. Conserved Asn220 and Ser221 stabilize a gamma turn that folds the polypeptide chain, bringing together and allowing interactions between regular secondary structure elements (Figure 4a).
Stereo image representing cartoon drawing of UGT from V. vinifera with conserved amino acids of the NTD shown in stick form. b Stereo image representing cartoon drawing of UGT from V. vinifera with conserved amino acids of the CTD shown in stick form.
The CTD shows a higher level of sequence conservation and binds the nucleotide sugar donor substrate. Side chain of Tyr275 interacts with the conserved Gln335 residue and helps in maintaining the structure by connecting the β-sheet with the α-helix. In addition, conserved Phe278 is involved in hydrophobic interactions with Pro284, Phe371 and Phe397. The residues Glu288 and Arg369 are involved in ionic interactions that stabilize the structure by connecting the α-helix to a coil. Pro334 present at the beginning of the α-helix helps the neighboring Gln335 to interact with Tyr275 of a proximate β-sheet. Trp332 stacks with the uridine ring of the UDP-sugar substrate. Residues Thr19, Ala333, His350, Trp353, Asn354, Ser355, Glu358, Asp374 and Gln375 also interact with the sugar donor. Asn378 interacts with Asp374 and Asn145 thus helping in maintaining the 3-D structure by connecting two secondary structural elements (Figure 4b). Conserved amino acid residues with less bulky side chains provide required flexibility to the loops surrounding the binding site. These features are conserved or semi-conserved in most of the proteins of the F3GT cluster.
Interaction Studies using Modeled Complexes
Interactions between enzyme and substrates.
Prior to the docking, the models and the ligands were first pre-processed to check for any problems in the structures related to missing hydrogens, side chains, improper bond orders, overlapping atoms etc. The refined sugar donor substrate was docked in the binding site cavity of the processed target by using standard precision mode of Glide. The bound ligand of template was used to mark the binding site by creating a grid around it. A total of 20 docked complexes were generated out of which the one with the highest glide score and forming all possible interactions with the conserved residues of PSPG motif was selected as the best pose (Table 2).
After the preparation of acceptor through LigPrep, kmp was docked in the acceptor binding pocket of the protein-UDP sugar complex by placing the grid around the reference ligand i.e. kmp. Standard precision mode of glide was used for docking studies. Similarly, 20 docked poses were generated and the one with a better glide score value and that interacted with the catalytic histidine was chosen as the best pose (Table 2). Docking studies have shown that the environment surrounding the acceptor is conserved or similar for all the docked structures. Some changes have been seen near the 5-OH and 4′-OH groups of kmp. In some structures 5-OH group is stabilized by the hydrophilicity provided by the side chain of Ser18 (w.r.t. PDB-2C1Z) while in some structures non-polar amino acids like glycine and alanine are present in that position. In the crystal structure 3HBF, glycine occupies this position but neighboring water molecule stabilizes the ligand molecule by forming a hydrogen bond with its 5-OH group. Amino acid Gln188 interacts with the O4′ group of kmp. In the F3GT proteins this glutamine is replaced by proline (close to N6; Figure S1), so this interaction is lost in all the docked complexes. This is eventually compensated by another interaction formed by His150 (Figure S1, Region N5) (Figure 5). The glide score value was higher for the pose in which 3-OH group of kmp interacted with the catalytic histidine residue while the other poses in which 7-OH, 4′OH and 5-OH groups of kmp interacted with the histidine had a low glide score. Docking studies aided in the identification of eight conserved regions present in the loops and helix near the acceptor binding regions (N1–N6 regions in the NTD and C1–C2 regions in the CTD) which might decide their substrate/glycosylation preference (Figure 6, Figure S1). This observation suggests that the environment topology observed here favors the binding of flavonoids to facilitate the glycosylation of 3-OH group of flavonoids.
Stereo view of the docked complex of UGT from F. ananassa with kaempferol shown in stick form. The arrow shows the 3-OH group of kaempferol which take part in glycosylation event. UDP-glucose is also shown in stick form.
Ribbon view of UGT from V. vinifera with docked acceptor and sugar donor in stick form is shown. Six conserved regions from N1 to N6 at the NTD and two regions C1 and C2 at the CTD marked with an arrow play a crucial role in holding the acceptor in the binding pocket.
Four protein sequences of Fragaria ananassa (Swiss-Prot ID-Q2V6K0), Manihot esculenta, Hevea brasiliensis and M. truncatula (Swiss-Prot ID-Q5IFH7) were annotated as flavonoid-3-O-glycosyltransferases in the protein sequence database but they were present outside the F3GT cluster in the dendrogram (Figure 3). Modeling and docking studies of these proteins showed that their activity might be less towards the flavonoid-3-OH group as compared to the other annotated members of F3GT. This is an additional evidence to show that the environment provided by the conserved acceptor binding site residues of F3GTs is more favorable for orienting the acceptor molecule for glycosylation at 3-OH group of the flavonoid molecule than sequences outside the F3GT cluster.
Docking studies with reaction products: Kaempferol-3-O-glucoside and UDP.
The crystal structure from V. vinifera (PDB-ID: 2C1Z) was taken as the target molecule to dock the glycosylated flavonoid, kaempferol-3-O-glucoside (KMG) and UDP in their respective binding pockets (Figure 7). Grid was created around the reference ligands, kmp (for KMG) and U2F (for UDP). The best pose for UDP and KMG had glide scores of −9.43 and −6.20 (Figure 8). The sugar moiety of KMG interacts with the sugar specificity determining amino acids of the PSPG motif.
Ribbon view of UGT from F. ananassa with docked Kaempferol-3-O-glucoside and UDP shown in stick form.
Stereo view of docked complex of UGT from F. ananassa with Kaempferol-3-O-glucoside shown in stick form.
Molecular Dynamics Simulations of the Docked Complexes
Molecular dynamics simulations were performed on 31 docked complexes in order to check their stability. Their trajectories were visually inspected for the binding of ligands and it was observed that the substrates were firmly bound in their respective binding pockets (Video S1). Here we have explained the interactions and the stability patterns of the ligands for one of the F3GT members i.e. F. ananassa GT (Swiss-Prot ID-Q5UL10). The RMSD graph was drawn by assigning time in ps on the X-axis and RMSD values of the Cα atoms in nm on the Y-axis (Figure 9). From the graph, we can conclude that the RMSD of Cα atoms of the generated structures over the complete trajectory is stable. To check the fluctuations of kmp in binding site, the distance between kmp and three key residues H20, Q84 and H152 which form hydrogen bonds with 3-OH, 5-OH and 4′OH group of kmp was calculated using g_mindist program of GROMACS. The graph showed that kmp fluctuates in the binding pocket within a distance of only 3 Å to 5 Å with respect to three key residue considered in the analysis for the last 2 ns of the trajectory (Figure 10). The life of hydrogen bond between the catalytic H20 and D120 remained for 70% of the total simulation time after equilibration. The root mean square fluctuation (RMSF) plot showed negligible fluctuations of the catalytic and binding site residues of the acceptor and donor (Figure 11). A highly fluctuating loop region, not part of the binding site region, was observed in the plot. Molecular dynamics simulation showed that the final products also form a stable complex (Video S2).
RMSD plot of Cα atoms of generated structures of F. ananassa UGT with time shown in ps along X-axis and RMSD values in nm along Y-axis.
Minimum distance plot between kaempferol and acceptor binding residues of F. ananassa UGT.
RMSF plot F. ananassa UGT shows negligible fluctuation of catalytic and binding site residues of acceptor and sugar donor substrate.
Experimental Evidence for F3GTs
Experimental evidence for the flavonoid 3-OH activity is known for UGTs enzymes of Dianthus_1, Diospyros, Petunia_2, Gentiana, Iris_2, Vigna, Actinidia, and Perilla_3 included in the analysis [36]–[43]. Out of them, UGTs of Dianthus_1, Diospyros and Petunia_2 were considered for analyzing their specificity against various flavonoid substrates. Few positive controls (3-OH group present: Kmp, Myricetin, Quercetin, and Fisetin) and negative control (3-OH group absent/present in different orientation: Apigenin, Naringenin, Dihyrdroquercetin (DHQ), Catechin, and Epicatechin) were selected to dock in the acceptor binding pocket of the protein. It has been observed that wherever the favorable binding pose was attained the flavonoid had its 3-OH present in close proximity of the catalytic histidine and sugar donor which is known to facilitate the nucleation during glycosylation reaction. However, the neighboring regions of the protein in the vicinity of catalytic base were not conducive for binding any other hydroxyl group of flavonoid (Figure S3, S4 and S5).
In order to further explore the importance of these crucial binding regions, flavonoid 7-O glucosyltransferase (F7GT) of Scutellaria baicalensis (Swiss-Prot ID: Q9SXF2), for which the flavonoid specificity was known, was selected for the study [44]. The three-dimensional structure of F7GT was modeled (Template: 2VCE; Resolution 1.9 Å, Identity: 30%) and three flavonoids namely baicalein, scutellarein, and kmp were chosen for the docking studies. The hydroxyl group is present on the C3 carbon atom of kmp but absent in other two. In spite of this difference the docked complexes showed that all the three ligands attained the favorable pose having only 7-OH closer to the catalytic histidine and sugar donor (Figure S6). Considering the flavonoid kmp, irrespective of it’s having both 7-OH and 3-OH, as per our classification F3GT favorably binds for 3-OH glycosylation and F7GT binds favoring 7-OH glycosylation. These results show how the acceptor binding site of UGTs decides the favorable binding mode of flavonoids and the nature of the glycosylation product formed.
Conclusions
Homology modeling and docking studies of the F3GTs showed that the environment in the close vicinity of the acceptor is conserved in all the members of this group and that these amino acids favor the binding of a flavonoid such that the glycosylation of the 3-OH group is preferred over other hydroxyl groups. Eight regions in the acceptor binding site were identified as important for acceptor specificity. Alterations in these regions may affect the type of molecule to be glycosylated and change the acceptor specificity from one molecule to another or from one side group for glycosylation to another side group. These observations have been further confirmed using molecular dynamics simulations of the docked complexes. The binding site information was successfully exploited to predict UGTs with experimentally known 3-OH specificity as positive control and 7-OH specific UGTs along with different ligands as negative control. This knowledge can be implemented to design engineered GTs which can glycosylate flavonoids in a regiospecific manner. In this way, the biochemical and pharmaceutical properties of flavonoids can be altered to prepare more effective drug molecules in a desired chemical form.
Supporting Information
Figure S1.
Multiple sequence alignment of 30 F3GT protein sequences.
https://doi.org/10.1371/journal.pone.0092636.s001
(PDF)
Figure S2.
Multiple sequence alignment between query (30F3GTs) and template (PDB ID: 2C1Z, 3HBF and 3WC4).
https://doi.org/10.1371/journal.pone.0092636.s002
(PDF)
Figure S3.
Image showing docked complexes of three positive and two negative control ligands in the acceptor binding pocket of Dianthus caryophyllus (Dianthus_1).
https://doi.org/10.1371/journal.pone.0092636.s003
(TIF)
Figure S4.
Image showing docked complexes of three positive and three negative control ligands in the acceptor binding pocket of Diospyros kaki.
https://doi.org/10.1371/journal.pone.0092636.s004
(TIF)
Figure S5.
Image showing docked complexes of three positive and three negative control ligands in the acceptor binding pocket of Petunia hybrida (Petunia_2).
https://doi.org/10.1371/journal.pone.0092636.s005
(TIF)
Figure S6.
Image showing docked complexes of three positive control ligands in the acceptor binding pocket of Scutellaria baicalensis.
https://doi.org/10.1371/journal.pone.0092636.s006
(TIF)
Table S1.
Details of box size and number of water molecules added during simulation process.
https://doi.org/10.1371/journal.pone.0092636.s007
(XLS)
Table S2.
Percent identity matrix of 30 F3GT protein sequences.
https://doi.org/10.1371/journal.pone.0092636.s008
(XLS)
Table S3.
Dataset of 101 GT protein sequences taken for the phylogenetic study.
https://doi.org/10.1371/journal.pone.0092636.s009
(XLS)
Video S1.
Multimedia file showing UGT docked with kaempferol and UDP-glucose.
https://doi.org/10.1371/journal.pone.0092636.s011
(MPG)
Video S2.
Multimedia file showing UGT docked with kaempferol-3-O-glucoside and UDP.
https://doi.org/10.1371/journal.pone.0092636.s012
(MPG)
Author Contributions
Conceived and designed the experiments: CGS. Performed the experiments: RS. Analyzed the data: RS PP. Contributed reagents/materials/analysis tools: RS. Wrote the paper: RS.
References
- 1. Taniguchi N, Honke K, Fukuda M (2003) Handbook of Glycosyltransferases and Related Genes. Springer 68: 707–708.
- 2. Kanadaswami C, Lee LT, Lee PPH, Hwang JJ, Ke FC, et al. (2005) The Antitumor Activities of Flavonoids. in vivo 19: 895–910.
- 3. Wang X (2009) Structure, mechanism and engineering of plant natural product glycosyltransferases. FEBS Letters 583: 3303–3309.
- 4. Shao H, He X, Achnine L, Blount JW, Dixon RA, et al. (2005) Crystal structures of a multifunctional triterpene/flavonoid glycosltransferase from Medicago truncatula. Plant Cell 17: 3141–3154.
- 5. Osmani SA, Bak S, Moller BL (2009) Substrate specificity of plant UDP-dependent glycosyltransferases predicted from crystal structures and homology modeling. Phytochem 70: 325–347.
- 6. Coutinho PM, Deleury E, Davies GJ, Henrissat B (2003) An evolving hierarchical family classification for glycosyltransferases. J Mol Biol 328: 307–317.
- 7. Campbell JA, Davies GJ, Bulone V, Henrissat B (1997) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem J 326: 929–939.
- 8. Hu Y, Walker S (2002) Remarkable Structural Similarities between Diverse Glycosyltransferases. Chemistry and Biology 9: 1287–1296.
- 9. Unligil UM, Rini JM (2000) Glycosyltransferase structure and mechanism. Current Opinion in Structural Biology 10: 510–517.
- 10. Zhang Z, Kochhar S, Grigorov M (2003) Exploring the sequence-structure protein landscape in the glycosyltransferase family. Protein Sci 12: 2291–2302.
- 11. Lairson LL, Henrissat B, Gavies GJ, Withers SG (2008) Glycosyltransferases: Structure, Function, and Mechanisms. Annu Rev Biochem 77: 521–55.
- 12. Masada S, Terasaka K, Mizukami H (2007) A single amino acid in the PSPG-box plays an important role in the catalytic function of CaUGT2 (Curcumin glucosyltransferase), a Group D Family 1 glucosyltransferase from Catharanthus roseus. FEBS Lett 581: 2605–10.
- 13. Li L, Modolo LV (2007) Crystal structure of Medicago truncatula UGT85H2– Insight into the Structural Basis of a Multifunctunal (Iso)flavonoid Glycosyltransferase. J Mol Biol 370: 951–963.
- 14. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32: 1792–97.
- 15. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28: 2713–9.
- 16. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25: 4876–4882.
- 17. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215: 403–410.
- 18. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, et al. (2000) The protein data bank. Nucleic Acids Res 28: 235–242.
- 19.
Schrödinger Suite 2011 Protein Preparation Wizard; Epik version 2.3, Schrödinger, LLC, New York, NY, 2012; Impact version 5.8, Schrödinger, LLC, New York, NY, 2012; Prime version 3.1, Schrödinger, LLC, New York, NY, 2012.
- 20. Laskowski RA, MacArthur MW, Moss DS, Thornton JM (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Cryst 26: 283–291.
- 21. Colovos C, Yeates TO (1993) Verification of protein structures: Patterns of nonbonded atomic interactions. Protein Sci 2: 1511–1519.
- 22. Luthy R, Bowie JU, Eisenberg D (1992) Assessment of protein models with three-dimensional profiles. Nature 356: 83–85.
- 23. Wiederstein M, Sippl MJ (2007) ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucleic Acids Res 35: W407–W410.
- 24.
Suite 2012: LigPrep, version 2.5, Schrödinger, LLC, New York, NY, 2012.
- 25. Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, et al. (2004) “Glide: A New Approach for Rapid, Accurate Docking and Scoring. 1. Method and Assessment of Docking Accuracy”. J Med Chem 47: 1739–1749.
- 26.
DeLano WL (2002) The PyMOL Molecular Graphics System. http://www.pymol.org.
- 27. Spoel D Van der, Lindahl E, Hess B, Groenhof G, Mark AE, et al. (2005) GROMACS: Fast, Flexible and Free. J Comp Chem 26: 1701–1719.
- 28. Berendsen HJC, Spoel D Van der, Drunen R Van (1995) GROMACS: A message-passing parallel molecular dynamics implementation. Comp Phys Comm 91: 43–56.
- 29. Schüttelkopf AW, Aalten van DMF (2004) PRODRG: a tool for high-throughput crystallography of protein-ligand complexes. Acta Cryst D60: 1355–1363.
- 30. Berendsen HJC, Postma JPM, Gunsteren van WF, DiNola A, Haak JR (1984) Molecular-dynamics with coupling to an external bath. J Chem Phys 81: 3684–3690.
- 31. Martonàk R, Laio A, Parrinello M (2003) Predicting crystal structures: the Parrinello-Rahman method revisited. Physiol Rev Lett 90: 075503.
- 32. Darden T, York D, Pedersen LG (1993) Particle mesh Ewald: An Nlog(N) method for Ewald sums in large systems. J Chem Phys 98: 10089–10092.
- 33. Hess B, Bekker H, Berendsen HJC, Fraaije JGEM (1997) LINCS: A linear constraint solver for molecular simulations. J Comp Chem 18: 1463–1472.
- 34. Humphrey W, Dalke A, Schulten K (1996) “VMD – Visual Molecular Dynamics”. J Mol Graph 14: 33–38.
- 35. Jorgensen WL, Tirado-Rives J (1988) The OPLS [optimized potentials for liquid simulations] potential functions for proteins, energy minimizations for crystals of cyclic peptides and crambin. J Am Chem Soc 110: 1657–1666.
- 36. Ikegami A, Akagi T, Potter D, Yamada M, Sato A, et al. (2009) Molecular identification of 1-Cys peroxiredoxin and anthocyanidin/flavonol 3-O-galactosyltransferase from proanthocyanidin-rich young fruits of persimmon (Diospyros kaki Thunb.). Planta 230: 841–855.
- 37. Miller KD, Guyon V, Evans JNS, Shuttleworth WA, Taylor LP (1999) Purification, Cloning, and Heterologous Expression of a Catalytically Efficient Flavonol 3-O-Galactosyltransferase Expressed in the Male Gametophyte of Petunia hybrida. J Biol Chem 274: 34011–34019.
- 38. Yoshihara N, Imayama T, Mizutani-Fukuchi M, Okuhara H, Tanaka Y, et al. (2005) cDNA cloning and characterization of UDP-glucose: Anthocyanidin 3-O-glucosyltransferase in Iris hollandica. Plant Sci 169: 496–501.
- 39. Gong Z, Yamazaki M, Sugiyama M, Tanaka Y, Saito K (1997) Cloning and molecular analysis of structural genes involved in anthocyanin biosynthesis and expressed in a forma-specific manner in Perilla frutescens. Plant Mol Biol 35: 915–927.
- 40. Ogata J, Itoh Y, Ishida Y, Yoshida H, Ozeki Y (2004) Cloning and heterologous expression of cDNAs encoding flavonoid glucosyltransferase from Dianthus caryophyllus. Plant Biotechnol 21: 367–375.
- 41. Mato M, Ozeki Y, Itoh Y, Higeta D, Toshitama K, et al. (1998) Isolation and Characterization of a cDNA of UDP-Galactose: Flavonoid 3-O-Galactosyltransferase (UF3GaT) Expressed in Vigna mungo Seedlings. Plant Cell Physiol 39: 1145–1155.
- 42. Tanaka Y, Yonekura K, Mizutani-Fukuchi M, Fukui Y, Fujiwara H, et al. (1996) Molecular and Biochemical Characterization of Three Anthocyanin Synthetic Enzymes from Gentiana triflora. Plant Cell Physiol 37: 711–716.
- 43. Montefiori M, Espley RV, Stevenson D, Cooney J, Datson PM, et al. (2011) Identification and characterisation of F3GT1 and F3GGT1, two glycosyltransferase responsible for anthocyanin biosynthesis in red-fleshed kiwifruit (Actinidia chinensis). Plant J 65: 106–18.
- 44. Hirotani M, Kuroda R, Suzuki H, Yoshikawa T (2000) Cloning and expression of UDP-glucose:flavonoid 7-O-glucosyltransferase from hairy root cultures of Scutellaria baicalensis. Planta 210: 1006–13.