Catalytic diversity and homotropic allostery of two Cytochrome P450 monooxygenase like proteins from Trichoderma brevicompactum

  • Razak Hussain
  • Indu Kumari
  • Shikha Sharma
  • Mushtaq Ahmed
  • Tabreiz Ahmad Khan
  • Yusuf Akhter
Original Paper

Abstract

Trichothecenes are the secondary metabolites produced by Trichoderma spp. Some of these molecules have been reported for their ability to stimulate plant growth by suppressing plant diseases and hence enabling Trichoderma spp. to be efficiently used as biocontrol agents in modern agriculture. Many of the proteins involved in the trichothecenes biosynthetic pathway in Trichoderma spp. are encoded by the genes present in the tri cluster. Tri4 protein catalyzes three consecutive oxygenation reaction steps during biosynthesis of isotrichodiol in the trichothecenes biosynthetic pathway, while tri11 protein catalyzes the C4 hydroxylation of 12, 13-epoxytrichothec-9-ene to produce trichodermol. In the present study, we have homology modelled the three-dimensional structures of tri4 and tri11 proteins. Furthermore, molecular dynamics simulations were carried out to elucidate the mechanism of their action. Both tri4 and tri11 encode for cytochrome P450 monooxygenase like proteins. These data also revealed effector-induced allosteric changes on substrate binding at an alternative binding site and showed potential homotropic negative cooperativity. These analyses also showed that their catalytic mechanism relies on protein–ligand and protein–heme interactions controlled by hydrophobic and hydrogen-bonding interactions which orient the complex in optimal conformation within the active sites.

Keywords

Trichoderma brevicompactum Cytochrome P450 monooxygenase Trichodiene 12, 13-epoxytrichothec-9-ene Homotropic cooperativity 

Abbreviations

MD

Molecular dynamics

TDN

Trichodiene

EPT

12, 13-epoxytrichothec-9-ene

SMs

Secondary metabolites

SRSs

Substrate recognition sites

P450s

Cytochrome P450 monooxygenases

Introduction

The use of biological control agents for agriculture is a sustainable alternative to the synthetic pesticides that are costly and have negative environmental impacts. The alternative use of biocontrol agents such as Pseudomonas and Trichoderma as biofertilizers and also for plant disease control is a recent advancement in modern agriculture towards sustainable development [1]. Many of the soil-borne bacteria and fungi colonizing the plant roots may be proven beneficial to the crop plants [2]. Apart from the mycorrhizal fungi, other plant growth promoting genera of soil fungi such as Trichoderma spp. can be used for crop plant growth stimulation and disease suppression [2]. These biocontrol activities of Trichoderma are attributed to its ability to produce secondary metabolites (SMs) [1, 2, 3]. Filamentous soil fungi like Trichoderma spp. are known to produce agriculturally important SMs [4, 5]. Trichothecenes are sesquiterpenoid epoxide SMs produced by Trichoderma spp. [6, 7]. Trichothecenes biosynthetic pathways are well studied in Trichothecium, Fusarium, Myrothecium, and Trichoderma [8, 9] (Fig. 1). In Trichoderma spp., the trichothecenes are synthesized from the derivative compound farnesyl pyrophosphate through a complex pathway involving various oxygenation, cyclization, and esterification reactions [10]. Several Trichoderma spp. have been reported to produce trichothecenes namely, trichodermin, and harzianum A, including, T. brevicompactum, T. arundinaceum, T. turrialbense, T. protrudens, T. viride, T. longibrachiatum, T. erinaceum, and T. citrinoviride [11]. The trichothecene molecule, trichodermin, has been reported to possess antibiotic and antitumor activity [12], while T. brevicompactum was reported to induce the leaf necrotic lesions produced by B. cinerea in tomato seed [11]. However, trichodermin showed a relatively poor inhibitory effect against Colletotrichum lindemuthianum [13]. Trichodermin may also act as microbial- or pathogen-associated molecular pattern to stimulate the expression of pathogen-related proteins [14]. Furthermore, trichothecene compounds are reported to induce the expression of pathogen-related proteins like PR1b1 and PR-P2 genes (defence-related genes) of the salicylic acid pathway. This was finally observed to show a decreased growth of both Botrytis cinerea and Rhizoctonia solani [8, 15, 16]. Trichodermin is reported for protein synthesis inhibition in eukaryotes, in which it is reported to disrupt the chain elongation and termination steps on the larger subunit of the 80S ribosome [9, 17]. Tri cluster genes involved in trichothecene biosynthetic pathway in Trichoderma spp. are assigned with different functions [8]. However, there are no detailed studies carried out so far; on all the tri cluster, genes encoded proteins involved in the trichothecene biosynthetic pathway in Trichoderma. Tri4 and tri11 have been annotated as cytochrome P450 monooxygenase like protein encoding genes in Trichoderma [10]. Cytochrome P450s are heme-b type thiolate proteins [18] present in prokaryotes and eukaryotes, and constitute a superfamily of metalloproteins catalyzing the monooxygenation reactions by introducing an oxygen atom in the substrates [19]. The tri4 and tri11 catalyze monooxygenation reactions in trichothecenes biosynthetic pathway of Trichoderma spp. [10] After the production of trichodiene (TDN) in trichothecene biosynthetic pathway, it is further oxygenated by tri4-encoded cytochrome P450 monooxygenase like protein in three consecutive oxygenation steps at C-2, C-13, and C-11 positions to produce isotrichodiol, whereas tri11 encoded another cytochrome P450 monooxygenase like protein catalyzes a single-step hydroxylation of 12,13-epoxytrichothec-9-ene (EPT) at position C-4 to produce trichodermol in Trichoderma [9, 10] (Fig. 1). The events operating at the molecular level that lead to the synthesis of different products catalyzed by these similar proteins are not elucidated in details as yet. Therefore, in the present work, we investigated the possible reasons of product diversity shown by tri4 and tri11. The structures of two cytochrome P450-like heme proteins, tri4 and tri11, were homology modelled. Molecular dynamics (MD) simulations were then performed on the free proteins and complexes with their substrates, TDN and EPT, to elucidate the structural changes that may lead to catalysis by these proteins. The heme-b type P450 like proteins tri4 and tri11 of Trichoderma contain all the conserved structural features described for heme thiolate proteins. The results indicated that monooxygenation reactions catalyzed by these proteins are mediated by the heme cofactor. The molecular interactions of active site amino acid residues through hydrogen-bonding and hydrophobic contacts were observed with the cofactor heme and the ligands, TDN and EPT. It is suggested that the catalysis is mediated by repositioning of substrates in the active site and electron transfer. We have also observed the disappearance of active site water molecules and helix-I structural changes to provide a more hydrophobic environment in these proteins. The inter-molecular interactions between active site amino acid residues with the heme showed conformational changes in heme propionates, which is important in maintaining the flexibility of binding cavity and also plays an important role in ligand binding, electron transfer process, reduction of heme, maintaining overall catalytic activity, and release of water after the catalysis. Out of six potential substrate recognition sites (SRSs) reported earlier [20]. SRS-5 lies most closer to the heme [21]. Therefore, it was targeted for allosteric modulation studies. The substrates of tri4 and tri11 proteins at this high binding affinity substrate recognition site, besides the productive active site in the ternary complex (protein-heme-2TDN/2EPT), showed allosteric effects of homotropic negative cooperativity.
Fig. 1

Trichothecene biosynthetic pathway in Trichoderma spp.: the figure is showing various tri cluster genes encoded enzymes involved in cyclization, hydroxylation, and esterification reactions to produce trichodermin and harzianum A. Tri4 and tri11 are the cytochrome P450 monooxygenase like proteins, that catalyze various monooxygenation reactions during the pathway, tri4 catalyzes three consecutive oxygenation reactions at C11, C13, and C2 of TDN and tri11 catalyzes a single-step oxygenation at C4 of EPT [10, 17, 46]

Methods

Sequence analysis and data sets

The sequences for tri4 and tri11 proteins from Trichoderma brevicompactum were downloaded in FASTA format from the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov.in). Secondary structure analysis was carried out using PSIPRED [22]. The PSIPRED predicts the presence of α- helices, β-sheets and loops within a protein sequence. Multiple sequence alignment (MSA) was performed using MultAlin [23] between tri4, tri11, and cytochrome P450 s from other diverse organisms to find out conserved residues, motifs, and signature sequences. MultAlin uses an algorithm for the multiple alignments of sequences that is both accurate and precise. 3-D structures of ligands, TDN, and EPT were obtained from PubChem [24].

Homology modelling and structural analysis

Tertiary structure models for tri4 and tri11 proteins were generated by homology modelling using Phyre2 [25]. The template used was a fungal cytochrome P450, lanosterol 14α-demethylase from yeast Saccharomyces cerevisiae (PDB ID; 4LXJ). Phyre2 predicted protein structure based on template 4LXJ with 100% confidence and 94% coverage for the query sequence of the tri4 protein and with 100% confidence and 98% coverage for the query sequence of the tri11 protein. The tri4 and tri11 protein sequences showed 37% and 44% overall homology with the template structure 4LXJ [25, 26, 27]. The modelled protein structures for both tri4 and tri11 proteins were energy minimized using Gromacs 4.6 package [9, 18, 19]; using the Gromos96 53a6 force field with a steepest descent optimization algorithm, protein immersed in water was subjected to energy minimization as reported earlier [17, 28, 29]. The system was immersed in a cubic box of simple point charge water molecules. Some of the Na+ counter-ions were added by replacing water molecules to keep charge neutrality of the simulated systems. The ligands were also energy minimized to obtain a stable structure. The 3-D models of proteins were validated by Ramachandran plot using RAMPAGE (http://mordred.bioc.cam.ac.uk/rapper/rampage.php) to assess the stereochemical quality of models.

Docking of TDN and EPT into the protein

TDN and EPT (ligands) were docked into the active sites of tri4 and tri11 proteins, respectively, to yield a binary complex using AutoDock Vina-2.0 software which is Monte Carlo-based. The ligands were docked at the conserved cysteine residue of CXG motif, reported for heme and substrate binding [30]. The pdb files of proteins and ligands were converted to pdbqt files using AutoDock utility of MGLTools-1.5.6 [9, 21, 31] and the polar hydrogen atoms were added to all the receptor protein molecules. In AutoDock Vina, the coordinates of the predicted binding site residues were used to assign binding positions for the ligands. These pdbqt files of proteins and ligands were added to the conf.txt file which was used as input for AutoDock Vina. A grid was generated around the binding site of the dimensions 30, 30, and 30 in x-, y-, and z-directions, respectively. AutoDock Vina uses semi-empirical free energy force field to evaluate the ligand-binding conformation. To analyse allosteric behavior, another molecule of both the ligands was also docked at SRS-5 of tri4 and tri11 to yield a ternary complex.

Molecular dynamics simulations

We have carried out six molecular dynamics simulations of 30 ns each (Table 1) for native proteins and binary and tertiary protein ligand complexes, three each for tri4 and tri11, respectively, with GROMACS 4.6.5 (GROningen MAchine for Chemical Simulations) package using the GROMOS96 53a6 (native protein) and GROMOS96 43a1 (protein–ligand complex) force fields. To study the tri4 and tri11 catalyzed oxygenation reactions, these six sets of simulations consisted of protein–heme complexes (tri4-heme and tri11-heme), the binary complexes with one ligand molecule of each (TDN and EPT) in the active sites (tri4-heme-TDN and tri11-heme-EPT), and the ternary complexes containing two ligand molecules one bound to the active site and second ligand molecule bound to the allosteric site SRS-5 (tri4-heme-2TDN and tri11-heme-2EPT) (Table 1). The topology parameters for native proteins were created by GROMACS program and the coordinates for ligands (TDN and EPT) were generated using PRODRG [32] server in gromos format using default settings except that the energy minimization option was kept off as the ligands were already energy minimized before docking [29]. The heme moiety was also included in the simulation and the topology parameters were generated using GROMACS-GROMOS96 43a1 force field [28], which considers an overall charge on heme as neutral, while Fe in its 2+ oxidation state as also followed in earlier studies [33, 34], while original calculations for Fe in its 2+ oxidation state were reported previously [35]. The native proteins and protein–ligand complexes were solvated in a cubic box maintaining a distance of 0.7 nm between the box edges and the proteins and protein–ligand complexes, respectively. The system was neutralized by adding counter Na+ ions in proteins and protein–ligand complexes. To avoid the high-energy interactions and steric clashes the system was subjected to a steepest descent energy minimization. To equilibrate the system, the ligand was subjected to position-restrained dynamics simulation (NVT and NPT) at a temperature of 300 k for 100 ps. Finally, the system was subjected to MD run at 300 k temperature and 1 bar pressure for 30 ns.
Table 1

Summary of tri4 and tri11 proteins simulation systems

Name

Description

Distancea

tri4 protein (protein–heme)

Substrate free

NA

tri4 protein–ligand binary complex

TDN in the active site

7.0 Å

tri4 protein–ligand ternary complex

TDN in the SRS-5

9.0 Å

tri11 protein (protein–heme)

Substrate free

NA

tri11 protein ligand complex

EPT in the active site

7.0 Å

tri11 protein–ligand ternary complex

EPT in the SRS-5

18.0 Å

a Distance between center of mass of heme group and ligands (TDN and EPT) at the start of the simulations

Trajectory analysis

For analysis, the atom coordinates were recorded at every 10 ps during the simulation. Various parameters like root-mean-square deviations (RMSDs), root-mean-square fluctuations (RMSFs), hydrogen bonds, the distance between the center of masses of heme and ligands (TDN and EPT), and distances between interacting atoms of ligands and heme–iron were considered for analysis. Different time frames containing the reaction intermediates from simulation trajectories were analyzed using LigPlot+ and PyMOL. The hydrogen bonds mediated interactions and the hydrophobic contacts were analyzed using LigPlot+ and the distances were measured in PyMOL.

Results

Sequence analysis of tri4 and tri11 proteins in T. brevicompactum showed conserved catalytic motifs and alternative SRSs

The MSA showed that both cytochrome P450 monooxygenase like proteins possess highest sequence conservation in the catalytic region of the protein around the heme which indicated that these enzymes may follow a common catalytic mechanism. MSA analysis showed that this core region of proteins comprised of a heme-binding loop with conserved motifs FSQGSRQCIG (amino acids positions 445–454) in tri4 and FNVGPRNCVG (amino acid positions 435–444) in tri11 which is one of the characteristics of P450 monooxygenases; both of these sequences possess absolutely conserved cysteine residue in the CXG motif. The conserved motifs EGLR (amino acids positions 369–372) in tri4 and EAFR (amino acids positions 352–355) in tri11, and PER, FKPERW (amino acids 423–428) in tri4, and FAPERW (amino acids 407–412) in tri11protein, the EXXR, and PER motifs form the P-R-R triad that is important for locking the heme pocket into position crucial for stabilizing the protein–ligand complex, and the central part of the protein containing the P450 motif AGTET (amino acid positions 311–315) in tri4 and AGSET (amino acid positions 294–298) in tri11, this motif is important for oxygen binding and activation as reported earlier [5, 10] (Fig. 2a, b; Fig. S1). In spite of the greater conservation in amino acid residues present in the catalytic site and conserved motifs for P450 monooxygenases between tri4 and tri11, it was observed that these proteins possess only 19% amino acid sequence similarity with each other. This showed that proteins have lower similarity in the complete sequence; however, the active site and other functionally important regions are conserved and they also possess the same catalytic mechanism and similar ligand-binding sites. The structures of both P450 like proteins were verified using RAMPAGE which predicted 96.7 and 96.6% residues in the allowed regions for tri4 and tri11 proteins, respectively, and the same structures were used for further analysis. The modelled proteins consist both of α-helices and β-sheets. The MSA was carried out to map six SRSs regions present in both tri4 and tri11 proteins for substrate binding (Fig. 4b, c). These SRSs were determined previously for other cytochrome P450 monooxygenases [1, 2].
Fig. 2

Conserved P450 motifs of tri4 and tri11 proteins shown in the structure of protein: a The figure is showing the positions of different conserved P450 motifs in tri4 protein tertiary structure AGXXT (blue) from A311 to T315, EXXR (yellow) from E369 to R372, PER (orange) from F423 to W428, and CXG (magenta) from F445 to G454 motif; heme is shown as sticks (red). b Positions of conserved P450 motifs in tri11 protein tertiary structure AGXXT from A294 to T298, EXXR from E352 to R355; PER from F407 to W412; and CXG from F435 to G444; the colour codes are same as in (a). In figures c and d, we are showing cartoon representation of the tertiary structure of tri4 and tri11 protein showing docking position of trichodiene (TDN) and 12, 13-epoxytrichothec-9-ene (EPT)

Interactions between proteins and ligands in the active site showed the amino acid residues involved in the catalysis

The ligands TDN and EPT were docked into an already reported active site that is present deep inside the proteins around the heme-binding region (Fig. 2c, d). The ligand TDN showed hydrophobic interactions with amino acid residues K110, T114, F118, A123, A125, A126, H133, V307, and A311 of tri4. The ligand EPT showed hydrophobic interactions with the active site residues R103, S107, I114, F115, T290, T293, A294, H480, and F481 of tri11 protein after docking. There were no hydrogen-bonding interactions between the proteins and the ligands observed after molecular docking.

Spatial arrangement of amino acid residues of heme-binding P450s like proteins to accommodate different reactants (TDN and EPT)

The diversity in the binding pocket of tri4 and tri11 proteins was analyzed by docking a single (TDN/EPT) and two (2TDN/2EPT) ligand molecules in binary and ternary complexes, respectively, as explained above in methods section. Preliminary analyses of the stability of the systems were examined by means of RMSDs with respect to the initial structure of the enzymes. The RMSDs of protein and protein–ligand complexes indicated that the protein–ligand complexes attain equilibrium around 8 ns and 4 ns of simulation for tri4 and tri11 ligand complexes, respectively. Therefore, we have used the trajectories after equilibrated conformations of protein–ligand complexes for the analysis (Fig. S2a, b). It was observed that the protein–ligand complexes possess more stability than the native proteins. The RMSFs analysis showed that the protein–ligand complexes have less fluctuation in comparison to the native proteins. However, the hydrogen-bonding interactions were observed between amino acid residues and heme (Table 2). Furthermore, the residues (A218, H480, R111, R239, and S107) which showed interactions with ligand EPT possessed lower RMSFs in comparison to the native version of tri proteins (Fig. S2c, d). It showed that the binding of a ligand to the protein has led to the stability of protein–ligand complexes which may aid in the catalysis. The ligands, TDN and EPT, showed significant interactions with respective proteins. In case of tri11 protein, the EPT showed hydrogen-bonding interactions with residues S107, R111, A219, and H480 throughout the simulation besides the hydrophobic contacts that were also observed during the simulations (Fig. S3c, Table 2). However, in case of tri4, no hydrogen-bonding interactions were observed and only hydrophobic contacts were present. The hydrogen-bonding and hydrophobic interactions of active site amino acid residues reoriented the ligands in an appropriate position for monooxygenation reaction during simulation.
Table 2

Summary of inter-molecular interactions of ligands (TDN and EPT) and heme with tri4 and tri11 encoded cytochrome P450 monooxygenase like proteins during simulations

Proteins

Hydrogen-bond interactions between catalytic residues and heme propionates

Hydrogen-bond interactions between catalytic residues and TDN/EPT

Hydrophobic contacts between protein and TDN/EPT

tri4

K110, T127, V128, D129, H130, D131, H133, R137, R380, S381, R450, Q451

 

T114, F118, A123, A125, A126, R221, F144, V307, P308, A311, G312, T315, P377, R491

tri11

Y94, R103, D121, H122, R126, M364, N436, N441, V443, G444

R111, S107, A219, H480

T108, R111, I114, F115, F205, A218, A219, M222, T293, A294, S296, T298, V360, M364, H480, F481

The conformational flexibility of heme propionates is controlled by amino acid residues present in the catalytic site

Conserved cysteine residues were observed in the CXG motif of both the proteins that serve as a ligand to interact with heme–iron; in addition, in the case of tri4, sulfur atom of conserved cysteine residue was observed to be covalently linked with heme iron. During simulation, heme propionates showed hydrogen-bonding network with different amino acid residues, viz., K110, T127, V128, D129, H130, D131, H133, R137, R380, S381, R450, and Q451 of tri4 protein and Y94, R103, D121, H122, R126, M364, N436, N441, V443, and G444 of tri11 protein. Hydrophobic interactions of heme were also observed with different amino acid residues (Table 2). The hydrogen-bonding amino acid residues induced various conformations of heme propionates as observed during simulations. This conformational flexibility of heme propionates is mainly driven by interacting amino acid residues of catalytic site and is an important factor for P450 catalytic activity as reported earlier [36]. Heme propionate interacted with arginine and asparagine repeatedly. The positive charge and amide group of asparagines could be the reason of these interactions. Since these residues have longer side chains, it may provide conformational flexibility for the interactions with heme propionates and help in accommodation of the substrates.

Changes in the active site of proteins caused by TDN and EPT binding

The ligand TDN was docked into the active site (Fig. 2c) of tri4 protein forming a binary complex and it remained in the active site with its C11, C13, and C2 at a distance of 9.1, 6.8, and 4.7 Å from the nearest side chain atoms (Fig. 3b.1, c.1, a.1), respectively, pointing away from the heme–iron. However, the ligand TDN oriented in different positions with respect to heme–iron during simulations and the hydrophobic interactions of active side residues reoriented the ligand with its C11, C13, and C2 at a distance of 3.7, 3.0, and 4.4 Å (Fig. 3b.2, c.2, a.2), respectively, pointing towards the heme–iron, the potential distance for accommodating an oxygen molecule between potential carbon atoms of TDN (C11, C13, and C2) and heme–iron as described earlier [33, 37]. Similarly, the ligand EPT for tri11 protein was docked into the active site in a binary complex (Fig. 2d); the EPT was located in the active site closer to the heme group with its C4 atom at a distance of 7.9 Å pointing away from the heme–iron (Fig. 4a.1); however, the ligand (EPT) oriented in different positions with respect to heme–iron during simulation and the hydrogen-bonding and hydrophobic interactions of active site residues repositioned the ligand with its C4 atom at a distance of 5.6 Å from the heme iron (Fig. 4a.1) which is within a potential distance for oxygenation at C4 atom of EPT for catalysis as reported earlier [33, 37].
Fig. 3

Decrease in distance between TDN and heme–iron within the active site and repositioning of the ligand to favor the possible oxygenation and further possible mechanism of negative allostery: In the figures a.1, b.1 and c.1, we are showing the distances between C2, C11 and C13 of TDN and heme–iron at the initial time of simulation and figures a.2, b.2 and c.2 at time frame 8800, 20020, and 277710 ps. TDN and heme are shown as sticks, coloured in cyan and orange, respectively, and iron is shown as sphere and coloured in magenta. The figure indicates that the TDN attained the required orientation, so that the potential atoms of TDN lie at an appropriate distance from heme–iron for monooxygenation reaction. The figures a.3, b.3, and c.3 are showing the distances between TDN and heme–iron in a ternary complex. The figure indicates that after the allosteric modulator (TDN) was bound to the SRS-5 site, the potential atoms were observed to be repositioned within a non-optimal distance from heme–iron for monooxygenation reaction

Fig. 4

Spatial orientation of catalytic site residues to accommodate the substrates: In the figures a.1 and a.2, we are showing the distance between C4 of EPT and heme–iron at the initial time of simulation and at time frame 4070 ps. The figure indicates that the EPT attained required conformation, so that the C4 atom (potential monooxygenation site) of EPT lies within an appropriate distance from heme–iron for monooxygenation reaction. The figure a.3 is showing the distance between EPT and heme–iron in the ternary complex (with negative allosteric modulation). The figure indicates that potential oxygenation atom of EPT lies at a non-optimal distance from the heme–iron for monooxygenation reaction. EPT and heme are shown as sticks, coloured in cyan and orange, respectively. Iron is shown as sphere coloured in magenta. The figures b and c are showing conserved substrate recognition sites (coloured in black) in tri4 and tri11 proteins, respectively [24]. The allosteric site SRS-5 is labelled

Helix-I present near heme exhibited catalytically important structural changes

The helix-I running over the distal surface of heme possesses conserved motif AGXXT, the helix-I exhibited structurally transitional changes during simulation, in AGXXT motif, a transition from helix to loop formation was observed in AGXXT motif of helix-I in both tri4 and tri11 proteins (Fig. 5a, b). Moreover, we have also observed the presence of two solvent molecules in the active site located in the vicinity of alanine residue of AGXXT motif in tri4 and tri11 as reported previously for similar proteins [38]. These active site water molecules lie within hydrogen-bonding distance from the conserved alanine residue of AGXXT motif that may mediate interactions with these solvent molecules (Fig. 5c, d). The center of mass distances was calculated between ligands (TDN and EPT) and heme–iron. In tri4 protein, the ligand TDN initially moved away from the heme–iron; however, after 2 ns of simulation time, the ligand moved closer to the heme–iron within a center of mass distance of less than 0.5 nm and remained within the same range throughout the 30 ns simulation. Similarly, in tri11 protein, the center of mass distance between ligand EPT and heme–iron was measured and the ligand was observed to be initially located at a distance of 0.7 nm, and during simulation, the ligand initially moved away from the heme–iron, and after 4 ns simulation time, the ligand moved closer to the heme–iron around 0.5 nm and remained at the same distance throughout the 30 ns simulation.
Fig. 5

Helix-I transitional changes at the neighbouring site of the binding pocket of protein aids in catalysis: In the figures a and b, we are showing the helix-I AGXXT motif unfolding into a loop in tri4 and tri11 proteins, respectively. c, d the figure is showing the active site water molecules and alanine-mediated interactions of AGXXT motif for tri4 and tri11 proteins, respectively. The TDN, EPT (blue), and heme (cyan) are shown as sticks and iron (magenta) is shown as a sphere. The figure is showing that the helix-I unfolding-to-loop formation is associated with the release of active site water molecules

Effect of allostery on P450 catalytic mechanism

Ligands were docked at the two binding sites reported earlier in the enzymes [21]. Docking was carried out one at the catalytic site and another at the chosen allosteric site SRS-5. From the docking scores, it was demonstrated that at the SRS-5 site, the ligands showed more tight binding in comparison to the active site. The high binding affinity site SRS-5 is specific for substrate orientation in the active site; as reported earlier [21], the TDN and EPT were docked at this potential site of tri4 and tri11 proteins to yield a ternary complex to analyse potential allostery induced changes on the substrate catalysis.

The MD trajectory analyses of ternary complexes with two ligand molecules showed that high-affinity ligand-binding site SRS-5 functions as an allosteric site that may affect the catalysis at the active site ligand molecule. The EPT molecule bound at the active site showed a lower number of hydrogen bonds (with only amino acid residues H480 and R239) in the ternary complex in the presence of allosteric EPT molecule which has interacted with T362 at SRS-5. When the second molecule of TDN/EPT was docked at allosteric site SRS-5 of tri4 and tri11 in the ternary complex, the trajectory analyses showed that the relative distance between potential oxygenation atoms of TDN molecule bound at the active site and heme–iron was increased significantly, viz., C11, C13, and C2 at a distance of 5.8, 7.2, and 7.7 Å, respectively (Fig. 3b.3, c.3, a.3). Similarly, the trajectories analyses of tri11 ternary complex showed that the distance between C4 of active site EPT and heme–iron was increased to 6.6 Å from 5.6 Å (Fig. 4a.3). Two of the active site water molecules that were observed to show interactions with heme during MD trajectories analyses of binary complexes, disappeared in the ternary complexes, when the second ligand molecule was bound to allosteric site SRS-5 of both tri4 and tri11 proteins (Fig. 5c, d). The center of mass distances of ternary complexes indicated fewer fluctuations in active site ligand molecules for appropriate orientations for productive reaction. All these transitions of ternary complexes from binary complexes together pointed out at negative homotropic cooperativity which might be due to allosteric modulation of active site of protein by the incoming second ligand molecule located at a high binding affinity site SRS-5 (− 6.9 and − 7.6 kcal/mol) for tri4 and tri11 ternary complexes, respectively as reported earlier [39].

Discussion

In this work, docking and MD simulations were performed to explore structural features that lead to the catalytic events occurring at the molecular levels in two cytochrome P450 monooxygenase like proteins (tri4 and tri11) from T. brevicompactum. These proteins showed common structural features and conserved sequence motifs that are important for the reaction. MSA of tri4 and tri11 proteins with other cytochrome P450s showed conserved amino acid residues. These conserved regions were observed to contain the catalytic site of the enzyme and other structurally important functional sequences. Our analysis showed that highly conserved motif FXXGXRXCXG is involved in heme-binding which contains the axial cysteine ligand of heme. This region contains the highly conserved cysteine residue which serves as fifth ligand residue in addition to the four coordination atoms from the porphyrin rings that coordinate with the heme iron and is located at position 452 in tri4 protein and at 442 in tri11 protein. The crystal structure of many P450s showed that this cysteine is absolutely conserved and is essential for P450 monooxygenase activity [40]. The region around the heme possesses the highest conservation of amino acid residues as reported for other P450s, and most of these conserved residues contribute to the active site cavity. The motifs EXXR and PER are present in both tri4 and tri11 proteins, as reported earlier for E-R-R triads, the role of this triad has been reported in fungal P450s for locking the heme pocket in the active site cavity and stabilization of the active site core structure [5]. Another motif AGXXT of cytochrome P450 monooxygenases is present in both tri4 and tri11 proteins which have been reported to be important for oxygen binding and activation. The C-terminal threonine residue of this motif lies in the active site cavity and may have a role in substrate binding during catalysis.

The cytochrome P450 monooxygenase catalytic cycle begins by binding of substrate in the active site in a favourable orientation with heme–iron, and the catalytic efficiency of the enzyme depends on substrate binding orientation and the distance of the potential atom of the substrate for oxygenation from the heme–iron. In the present study, the docking and MD simulation analysis on tri4 protein and protein–ligand complexes showed that after adopting the stability of system, the hydrophobic interactions of active site amino acid residues F118, A123, A125, A126, V307, F308, A311, P377, G312, and I491 with TDN in a binary complex positioned the substrate in a favourable orientation within the active site cavity with its potential atoms for hydroxylation and epoxidation reaction within a distance of 6 Å from the heme–iron (Fig. 3a.1, b.1, c.1), that is within an optimal distance for the reaction as reported earlier [23, 24]. The hydrophobic interactions with the ligand TDN are contributed by non-polar side chains of F118, A123, A125, A126, V307, F308, A311, P377, G312, and I491. These hydrophobic interactions may stabilize the substrate TDN in the binding site of tri4 encoded Cytochrome P450 monooxygenase like protein. The hydrophobic interactions of TDN with the polar side chain amino acid residues R221, R491, T315, and heme group could help in the reaction catalysis (Table 2). Arginine is a positively charged amino acid with guanidino group in the side chain, whereas threonine has a hydroxyl group and these amino acid residues may act as electrophiles as indicated earlier [41] and participate in proton transfer process that may lead to the reaction catalysis. We have not observed any hydrogen-bonding interactions between the substrate TDN and tri4 protein. Similarly, in tri11 protein, the docked protein–ligand complex showed that hydrogen-bonding interactions with various active site residues S107, R111, R239, A219, and H480, and EPT repositioned the ligand in a favourable orientation with its potential atoms for hydroxylation within a distance of 6 Å (Fig. 4a.2) from the heme–iron, which is chemically feasible distance for the reaction between C4 of EPT and heme–iron [33, 37]. S107 hydroxyl group interacted with the oxygen atom of EPT through hydrogen bonding; similarly, hydrogen bonding was also observed between guanidino group of R111 and the oxygen atom of EPT, between the amino group of A219 and the oxygen atom of EPT and between imidazole group of H480 and the oxygen atom of EPT. The side chains of these amino acid residues were observed to interact with their electrophilic atoms (N and O) with the ligand EPT. The electrophiles are involved in proton transfer process that leads to the oxygenation of the substrate. The changing hydrogen-bonding network could be correlated with the repositioning of ligand for better binding and the proton transfer process that ultimately lead to the catalysis [36]. It was observed that the hydrophobic contacts by active site residues were established to block the entry of solvent which provided a hydrophobic environment for the reaction catalysis. Ionic interactions were also observed in the active site region of tri4 protein with heme propionate between positively and negatively charged amino acid residues, carboxylate group of D129 and the amino group of H133 and between D131 and H133, respectively, and ionic interaction between carboxylate group of D121 and amino group H122 of tri11 protein. These ionic interactions may provide stability to the cofactor heme and lock up the substrates TDN and EPT in the active site cavity. However, the ionic interactions between D121 and H122 of tri11 were lost in the ternary complex. Like substrate interacting residues, many active site residues were also observed to be forming hydrogen-bonding and hydrophobic interactions with the heme moiety in the binary complexes of both tri4 and tri11 (Table 2). More number of hydrogen bonds was observed between positively charged amino acid residues and heme propionates during the simulations; this could be correlated with the efficient process of proton transfer required for the reaction. This hydrogen-bonding network of active site residues with the heme propionates may play an important role in maintaining the catalytic activity of P450s as reported earlier for other cytochrome P450s [36, 42, 43]. The RMSDs of Cα atoms indicated that all the heme and substrate interacting residues were found to be stable during the simulation in both tri4 and tri11 proteins (Fig. S2).

The catalytic mechanism of cytochrome P450 like proteins from T. brevicompactum involves various steps for substrate entry and product release as presented schematically in Fig. 6. The helix-I-to-loop transition of AGXXT motif suggested that this loop formation by unfolding mediated by glycine residue of this motif and bending in this helix is important for the release of axial water molecule present in the active site after substrate entry by a similar mechanism as reported earlier [44]. The alanine residue of this motif mediates interaction with its carbonyl oxygen atom with water molecule present in the active site (Fig. 5c, d), and this interaction is the prerequisite step for P450 catalysis. The terminal threonine residue of this motif mediates interactions with an iron bound oxygen atom and these interactions are important for the protonation of oxygen. The water molecules observed in the active site at the initial stage of simulation disappeared with the movement of ligands in the active site and the helix-I-to-loop formation, also suggesting that hydrophobic environment may be required for the catalytic reaction of P450s like proteins. Similar results were shown for other P450 like proteins in the earlier studies [38].
Fig. 6

Proposed catalytic mechanism of Cytochrome P450 like proteins from T. brevicompactum: The figure is showing various steps of the substrate (RH) entry and product release in the monooxygenation reaction. Alanine-mediated interactions were observed to be necessary for the release of active site water molecule and threonine-mediated interactions with oxygen seem to be required for its protonation, and formation of hydroxylate product (ROH) through a sequence of reactions

The center of mass distances between ligands (TDN and EPT) and the heme–iron indicated that the ligands were located at a distance of 0.7 nm (Fig. S4a, b) in the active site cavity. This orientation of ligands may help them to interact with the heme–iron. However, in the case of tri4 protein, the ligand TDN again moved away from the heme–iron around 20 ns time of simulation. This fluctuation may be attributed to the little degree of motion in the active site and rearrangement of active site residues [33]. The center of mass distances between ligands (TDN and EPT) and heme indicated that the movement of ligands and heme closer to each other within an ideal distance during simulation is associated with the stability of protein–ligand complexes and prepared the active site for productive catalysis.

The effect of allostery on P450 catalysis was also analyzed by docking a second ligand molecule in ternary complexes at allosteric site SRS-5 of tri4 and tri11 (Fig. 4b, c).This potential substrate binding site located around the EXXR motif has been reported for conserved residues that affect substrates orientation in the active site. The binding of second ligand molecule at this site may affect the proper positioning of ligand molecule in the active site as reported earlier, and similar results were obtained in the present study [21, 45]. The MD trajectories analyses of ternary complexes showed that the distance between the active site TDN/EPT and heme–iron has been increased significantly which may reduce the efficiency of the oxygenation reaction (Figs. 3a.3, b.3, c.3 and 4 a.3).The presence of second ligand molecule at SRS-5 and less hydrogen-bonding and hydrophobic interactions have limited the active site ligand molecule at an appropriate distance from heme–iron and the strong hydrophobic and hydrogen-bonding interactions were partially lost as compared to binary complexes [39]. The active site EPT showed limited hydrogen-bonding interactions with only H480 and R239. The RMSDs and RMSFs analyses of ternary complexes showed more stability compared to the binary complexes. However, the RMSFs of tri4 ternary complex showed higher fluctuations than binary complex. This might be due to the nearest location of second TDN molecule to the active site (at a distance of 9 Å from heme–iron); this close proximity altered the active site more leading to an overall change in protein conformation, suggesting its allosteric behavior (Fig. S2c).The center of mass distances between the active site ligand molecules and heme–iron in ternary complexes during simulation indicated that ligands exhibited non-optimal orientations for a productive-binding site with their potential atoms moving away from the heme–iron (Fig. S4a, b).These results suggested that presence of second ligand molecules at allosteric site SRS-5 of tri4 and tri11 in ternary complexes leads to a conformational change at the active site of proteins that indicated the negative type of homotropic cooperativity.

Conclusions

Cytochrome P450 monooxygenases are important bio-catalysts involved in the substrate oxygenation and hydroxylation reaction steps during the biosynthesis of several fungal SMs. Our results on cytochrome P450 monooxygenase like proteins (tri4 and tri11) from T. brevicompactum which are involved in the trichothecenes biosynthetic pathway (Fig. 1) demonstrated that these two P450 like proteins may catalyze monooxygenation reactions mediated by heme cofactor and many of the inter-molecular interactions may lead to the catalysis. The MD analysis indicated that in tri4 and tri11 P450 like proteins, catalytic mechanism relies on the fact that direct protein–ligand and protein–heme interactions are promoted by hydrophobic and electrostatic forces with the efficiently controlled orientation of substrate within the active site and less structural and conformational changes in other parts of the proteins were observed. Our results also showed that homotropic negative cooperativity in ternary complexes of tri4 and tri11 proteins due to allosteric modulation of the active site in the presence of a second ligand molecule at alternative binding site, SRS-5 of both proteins may provide a way of feedback type of enzyme activity regulation. This work may be helpful to design in vitro experiments to elucidate further the catalytic mechanism of these two tri cluster genes encoded P450 like proteins involved in the production of agriculturally important SMs. This information may be helpful to the scientific workers in the field to better understand the catalytic mechanism of P450s which are involved at two crucial reaction steps of the biosynthetic pathway of trichothecene compounds. Therefore, this in silico study about their mechanisms may help to devise strategies to enhance/suppress the overall production of these molecules in biotechnological settings.

Notes

Acknowledgements

We acknowledge Central University of Himachal Pradesh and Bioinformatics Resources and Applications Facility, Centre for Development in Advanced Computing, Pune for providing the computational infrastructure. RH acknowledges National Fellowship for Higher Education from University Grants Commission, Govt. of India (UGC). SS receives research stipend from UGC. Research in YA lab is supported by extramural research funds from UGC, Science and Engineering Research Board (DST, Govt. of India), and Indian Council of Medical Research. We thank Dr. P. Aparoy for his generous help during the revision. Prof. Claudio Luchinat (editor-in-chief) and two anonymous referees are also sincerely acknowledged, whose insightful comments and advice during the editorial review helped us to improve our work enormously.

Supplementary material

775_2017_1496_MOESM1_ESM.pdf (1.8 mb)
Supplementary material 1 (PDF 1874 kb)

References

  1. 1.
    Reino JL, Guerrer ORF, Hrnndez-Galn RIG, Collado IG (2008) Phytochem Rev 7:89–123CrossRefGoogle Scholar
  2. 2.
    Hermosa R, Viterbo A, Chet I, Monte E (2012) Microbiology 158:17–25CrossRefPubMedGoogle Scholar
  3. 3.
    Vinale F, Sivasithamparam K, Ghisalberti EL, Marra R, Barbetti MJ, Li H, Woo SL, Lorito M (2008) Physiol Mol Plant Pathol 72:80–86CrossRefGoogle Scholar
  4. 4.
    Vinale F, Sivasithamparam K, Ghisalberti EL, Marra R, Woo SL, Lorito M (2008) Soil Biol Biochem 40:1–10CrossRefGoogle Scholar
  5. 5.
    Chen W, Lee MKC, Jefcoate CS-C, Kim SC, Chen FYuJH (2014) Genome Biol Evol 6:1620–1634CrossRefPubMedPubMedCentralGoogle Scholar
  6. 6.
    Bowen GD, Rovira AD (1999) Adv Agron 66:1–102CrossRefGoogle Scholar
  7. 7.
    Jain A, Singh A, Singh S, Singh HB (2013) J Plant Growth Regul 32:388–398CrossRefGoogle Scholar
  8. 8.
    Malmierca MG, Cardoza RE, Alexander NJ, Mccormick SP, Hermosa RE, Monte E, Gutiérrez S (2012) Appl. Environ Microbiol 78:4856–4868CrossRefGoogle Scholar
  9. 9.
    Cundliffe E, Cannon M, Davies J (1974) Proc Natl Acad Sci USA 71:30–34CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Cardoza RE, Malmierca MG, Hermosa MR, Alexander NJ, Mccormick SP, Proctor RH, Tijerino AM, Rumbero A, Monte E, Gutie S (2011) Appl. Environ Microbiol 77:4867–4877CrossRefGoogle Scholar
  11. 11.
    Degenkolb T, Dieckmann R, Nielsen KF, Gräfenhan T, Theis C, Zafari D, Chaverri P, Ismaiel A, Brückner H, Döhren HV (2008) Mycol Prog 7:177–219CrossRefGoogle Scholar
  12. 12.
    Tijerino A, Cardoza RE, Moraga J, Malmierca Vicente F, Aleu J, Collado IG, Gutiérrez S, Monte S, Hermosa R (2011) Fungal Genet. Biol 48:285–296Google Scholar
  13. 13.
    Shentu XP, Yuan XF, Liu WP, Xu JF, Yu XP (2015) Am J Biochem Biotechnol 11:169CrossRefGoogle Scholar
  14. 14.
    Hermosa R, Rubio MB, Cardoza RE, Nicolas C, Monte E, Gutiérrez S (2013) Int Microbiol 16:69–80PubMedGoogle Scholar
  15. 15.
    Malmierca MG, Izquierdo I, Bueno McCormick SP, Cardoza RE, Alexander NJ, Barua J, Lindo L, Casquero PA, Collado IG, Monte E (2016) Environ. Microbiol 18:3991–4004Google Scholar
  16. 16.
    Pusztahelyi T, Holb IJ, Pocsi I (2015) Front Plant Sci 6:573CrossRefPubMedPubMedCentralGoogle Scholar
  17. 17.
    Kumari I, Ahmed M, Akhter Y (2016) Int J Biochem Cell Biol 78:370–376CrossRefPubMedGoogle Scholar
  18. 18.
    Hlavica P (2017) J Inorg Biochem 167:100–115CrossRefPubMedGoogle Scholar
  19. 19.
    Cresnar B, Petric S (2011) Biochim Biophys Acta (BBA)-Proteins Proteom 1814:29–35CrossRefGoogle Scholar
  20. 20.
    Gotoh O (1992) J Biol Chem 267:83–90PubMedGoogle Scholar
  21. 21.
    Seifert A, Pleiss J (2009) Proteins Struct Funct Bioinform 74:1028–1035CrossRefGoogle Scholar
  22. 22.
    McGuffin LJ, Bryson K, Jones DT (2000) Bioinformatics 16:404–405CrossRefPubMedGoogle Scholar
  23. 23.
    Corpet F (1988) Nucleic Acids Res 16:10881–10890CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    Kim S, Thiessen PA, Bolton EE, Bryant SH (2015) Nucleic Acids Res 43(W1):W605–W611CrossRefPubMedPubMedCentralGoogle Scholar
  25. 25.
    Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJE (2015) Nat Protoc 10:845–858CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    Eswar N, Webb B, Marti-Renom MA, Madhusudhan MS, Eramian D, Shen MY, Pieper U, Sali A (2006) Curr Protoc Bioinformatics. doi:10.1002/0471250953.bi0506s15 PubMedPubMedCentralGoogle Scholar
  27. 27.
    Krieger E, Nabuurs SB, Vriend G (2003) Methods Biochem Anal 44:509–524PubMedGoogle Scholar
  28. 28.
    Berendsen HJC, Spoel Van Der D, Drunen Van R (1995) Comput Phys Commun 91:43–56CrossRefGoogle Scholar
  29. 29.
    Sandhu P, Akhter Y (2016) Arch Biochem Biophys 592:38–49CrossRefPubMedGoogle Scholar
  30. 30.
    Guengerich FP (2001) Chem Res Toxicol 14:611–650CrossRefPubMedGoogle Scholar
  31. 31.
    Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Goodsell DS, Olson AJ (2009) J Comput Chem 30:2785–2791CrossRefPubMedPubMedCentralGoogle Scholar
  32. 32.
    SchuÈttelkopf AW, Aalten Van DMF (2004) Acta Crystallogr Sect D Biol Crystallogr 60:1355–1363CrossRefGoogle Scholar
  33. 33.
    Vohra S, Musgaard M, Bell SG, Wong L, Zhou W, Biggin PC (2013) Protein Sci 22:1218–1229CrossRefPubMedPubMedCentralGoogle Scholar
  34. 34.
    Oostenbrink C, Villa A, Mark AE, Van Gunsteren WF (2004) J Comput Chem 25:1656–1676CrossRefPubMedGoogle Scholar
  35. 35.
    Micaelo NM, Macedo AL, Goodfellow BJ, Felix V (2010) J Mol Graph Model 29:396–405CrossRefPubMedGoogle Scholar
  36. 36.
    Cojocaru XYUV, Mustafa G, Salo Ahen OMH, Lepesheva GI, Wade RC (2015) J Mol Recognit 28:59–73CrossRefPubMedPubMedCentralGoogle Scholar
  37. 37.
    Graaf CDE, Oostenbrink C, Keizers PHJ, Tvander WIJST, Jongejan A, Vermeulen NPE (2006) J Med Chem 49:2417–2430CrossRefPubMedGoogle Scholar
  38. 38.
    Podust LM, Ouellet H, Kries von JP, Montellano de PRO (2009) J Biol Chem 284:25211–25219CrossRefPubMedPubMedCentralGoogle Scholar
  39. 39.
    Zhang H, Gay SC, Shah M, Foroozesh M, Osawa J, Liu Y, Zhang Q, Stout CD, Halpert JR, Hollenberg PF (2013) Biochemistry 52:355CrossRefPubMedPubMedCentralGoogle Scholar
  40. 40.
    Guengerich FP, Munro AW (2013) J Biol Chem 288:17065–17073CrossRefPubMedPubMedCentralGoogle Scholar
  41. 41.
    Bischoff R, Schlüter H (2012) J Proteom 75:2275–2296CrossRefGoogle Scholar
  42. 42.
    Hasemann CA, Kurumbail RG, Boddupalli SS, Peterson JA, Deisenhofer J (1995) Structure 3:41–62CrossRefPubMedGoogle Scholar
  43. 43.
    Guallar V, Baik MH, Lippard SJ, Friesner RA (2003) Proc Natl Acad Sci 100:6998–7002CrossRefPubMedPubMedCentralGoogle Scholar
  44. 44.
    Podust LM, Sherman DH (2012) Nat Prod Rep 29:1251–1266CrossRefPubMedPubMedCentralGoogle Scholar
  45. 45.
    Zanger UM, Schwab M (2013) Pharmacol Ther 138:103–141CrossRefPubMedGoogle Scholar
  46. 46.
    Kimura M, Tokai T, Takahashi N, Ohsato S (2007) Biosci Biotechnol Biochem 71:2105–2123CrossRefPubMedGoogle Scholar

Copyright information

© SBIC 2017

Authors and Affiliations

  1. 1.Department of BotanyAligarh Muslim UniversityAligarhIndia
  2. 2.Department of Environmental Science, School of Earth and Environmental SciencesCentral University of Himachal PradeshKangraIndia
  3. 3.Centre for Computational Biology and Bioinformatics, School of Life SciencesCentral University of Himachal PradeshKangraIndia

Personalised recommendations