Free energy calculations elucidate substrate binding, gating mechanism, and tolerance‐promoting mutations in herbicide target 4‐hydroxyphenylpyruvate dioxygenase

Christina E M Schindler; Eva Hollenbach; Thomas Mietzner; Klaus‐Jürgen Schleifer; Martin Zacharias

doi:10.1002/pro.3612

. 2019 Apr 19;28(6):1048–1058. doi: 10.1002/pro.3612

Free energy calculations elucidate substrate binding, gating mechanism, and tolerance‐promoting mutations in herbicide target 4‐hydroxyphenylpyruvate dioxygenase

Christina E M Schindler ^1,², Eva Hollenbach ³, Thomas Mietzner ⁴, Klaus‐Jürgen Schleifer ⁴, Martin Zacharias ^1,^2,^✉

PMCID: PMC6511742 PMID: 30945368

Abstract

4‐Hydroxyphenylpyruvate dioxygenase (HPPD) catalyzes the second reaction in the tyrosine catabolism and is linked to the production of cofactors plastoquinone and tocopherol in plants. This important biological role has put HPPD in the focus of current herbicide design efforts including the development of herbicide‐tolerant mutants. However, the molecular mechanisms of substrate binding and herbicide tolerance have yet to be elucidated. In this work, we performed molecular dynamics simulations and free energy calculations to characterize active site gating by the C‐terminal helix H11 in HPPD. We compared gating equilibria in Arabidopsis thaliana (At) and Zea mays (Zm) wild‐type proteins retrieving the experimentally observed preferred orientations from the simulations. We investigated the influence of substrate and product binding on the open–closed transition and discovered a ligand‐mediated conformational switch in H11 that mediates rapid substrate access followed by active site closing and efficient product release through H11 opening. We further studied H11 gating in At mutant HPPD, and found large differences with correlation to experimentally measured herbicide tolerance. The computational findings were then used to design a new At mutant HPPD protein that showed increased tolerance to six commercially available HPPD inhibitors in biochemical in vitro experiments. Our results underline the importance of protein flexibility and conformational transitions in substrate recognition and enzyme inhibition by herbicides.

Keywords: dioxygenase, Arabidopsis thaliana, gating, enzyme mutation, conformational change, molecular dynamics, free energy calculations, enhanced sampling, rational enzyme design, herbicide tolerance

Introduction

4‐hydroxyphenylpyruvate dioxygenase (HPPD) is a member of the non‐haem Fe(II)/2‐oxoacid‐dependent oxygenase superfamily1 and a key enzyme in tyrosine catabolism. It catalyzes the conversion of hydroxyphenylpyruvate (HPP) to homogentisate (HGA).2 In plants, HGA is an intermediate in the biosynthesis of cofactor plastoquinone and tocopherols. Tocopherols are antioxidants required for plant growth and stress tolerance and plastoquinone is essential for photosynthetic electron transfer and the biosynthesis of carotenoids.3 The inhibition of carotenoid biosynthesis results in bleaching, necrosis, and subsequent death of the plants.4, 5, 6 Hence, for many years, HPPD has been a promising target for the agrochemical industry and a variety of HPPD inhibitors (herbicides) have been developed to date.7, 8, 9, 10 In addition, developing crops tolerant to HPPD inhibitors has also become an active field of agrochemical research. Herbicide tolerance is typically assessed by comparing the inhibition of wild type and mutant HPPD proteins by different inhibitors. A range of different mutations has been proposed for HPPD proteins in mousear cress, rice, and soybean.11, 12, 13, 14, 15 Recently, Siehl et al. presented an approach to develop transgenic soybean plants tolerant to a variety of HPPD inhibitors by directed evolution of maize HPPD.15 Interestingly, the resulting variant with the highest fitness showed 26 different amino acid substitutions distributed throughout the entire sequence. These mutations did not involve any residues reported to be related to catalytic function.16 The authors concluded that the improved herbicide tolerance occurred “through the accumulated effect of many small changes outside the active site that influence the catalytic properties of the enzyme in a manner that is not obvious upon inspection of the structure.”15 Hence, a detailed understanding of the molecular phenomena promoting herbicide tolerance and especially the effect of mutations outside of the active site is lacking to date.

A number of HPPD crystal structures from different organisms have been resolved and are available in the Protein Data Bank (PDB, www.rcsb.org).17, 18, 19, 20 The eukaryotic HPPD enzyme is active as a homodimer with a subunit molecular mass of about 45 kDa. The topology and the overall fold of all crystallized structures are very similar (Fig. 1).17, 18, 19, 20 The monomer is folded into two structural domains that are arranged as an N‐terminal and C‐terminal open β‐barrel of eight β‐strands each. The active site is located inside the β‐barrel of the C‐terminal domain and contains a 2‐His‐1‐carboxylate motif that non‐covalently binds a Fe²⁺ ion.18 Previously, Raspail et al. were able to elucidate the positioning of the substrate and the product in the active site by identifying catalytic residues by quantum mechanical calculations and mutational experiments.16

The enzyme 4‐hydroxyphenylpyruvate dioxygenase (HPPD). (a) HPPD catalyzes the second reaction in the tyrosine catabolism converting hydroxyphenylpyruvate to homogentisate. In eukaryotes, the protein is active as a homodimer. The active site Fe²⁺ ion is shown in orange, C‐terminal gating helix H11 in yellow (closed state) and red (open). (b) Superposition of structures of HPPD from different organisms.17, 18, 19, 20 Helix H11 can adopt two distinct conformations keeping the active site shielded from the solvent (closed, yellow) or accessible (open, red). Inset: Locations of investigated mutations (green). All the mutations are located outside of the active site (magenta, shown in part). (c) Sequence alignment of *Zea mays* (Zm) and *Arabidopsis thaliana* (At) HPPD. The proteins have more than 60% sequence identity. Identical residues are shaded in blue. Mutations investigated in this work are highlighted by a red frame.

Although the crystal structures of all characterized HPPD proteins are very similar, there is a striking difference in the conformation of the highly conserved C‐terminal α‐helix H11. H11 can be found in two different orientations, either in a closed conformation (i.e., shielding the active site from the solvent) or an open conformation that allows access to the active site (Fig. 1). Fritze et al. hypothesized that these different H11 orientations can function as a gate regulating binding to the active site.18 However, the dynamic properties of this gating mechanism and its role in catalysis have yet to be elucidated. A range of HPPD inhibitors have also been co‐crystallized with the enzyme.19, 20 Most inhibitors are bulky molecules compared to the substrate HPP and for sterical reasons bind to the active site with the C‐terminal α‐helix H11 in its open conformation only.10 This observation suggests that inhibitor binding and thus herbicide tolerance could also be related to the accessibility of the active site and the conformational equilibrium of the C‐terminal helix H11.

In this work, we investigated the gating mechanism of At and Zm HPPD via a conformational change in its C‐terminal α‐helix H11 using molecular dynamics (MD) simulations and free energy calculations. We studied both the influence of ligand binding and of amino acid substitutions that correspond to sequence differences between At and Zm HPPD on the H11 transition. These findings helped to understand some of the observed difference in herbicide tolerance of At and Zm HPPD and to rationalize the tolerance of one At mutant HPPD protein caused by mutations outside of the active site of the enzyme. Furthermore, our analysis of the interactions formed by H11 in Zm wild‐type HPPD resulted in the creation of a new At HPPD variant that displayed increased tolerance in biochemical in vitro experiments.

Results

Identifying HPPD mutant proteins that display tolerance toward known HPPD inhibitors is an active field of agrochemical research. However, it remains unclear how sequence variations between HPPD proteins from different organisms that are located outside of the active site of the enzyme can contribute to herbicide tolerance. The tolerance is indicated as tolerance index (TI) that considers the sensitivity for a given herbicide but also the respective activity of the enzyme or enzyme variant (see Methods section for the definition of the TI). Initially, we noted that the Zea mays (Zm) HPPD protein that displays higher natural herbicide tolerance compared to Arabidopsis thaliana (At) HPPD (Table 3), had been crystallized with the C‐terminal helix H11 in a closed conformation,18 whereas At HPPD protein is found in the open H11‐conformation in all available crystal structures.18, 19 We hypothesized that sequence differences between the two enzyme variants outside of the active site may affect protein conformational dynamics, specifically the proposed gating of H11 that blocks access to the active site18 and through this mechanism promote herbicide tolerance. Note that H11 gating is clearly not the only driver of herbicide tolerance in HPPD. In fact, mutations in the active site and closeby have a strong effect on herbicide tolerance due to their direct effect on inhibitor binding. For herbicide tolerance in Zm HPPD, both effects likely play a role. To better understand the conformational aspect, we set out to investigate H11 gating equilibrium in At wild‐type HPPD, Zm wild‐type HPPD, and At mutant HPPD protein 1, and in the presence of substrate and product using MD simulations and free energy calculations (Table 1).

Table 3.

Enzyme Kinetics and Herbicide Tolerance of At Wild‐Type HPPD, Zm Wild‐Type HPPD, and At Mutant HPPD Proteins

Protein	K _m (mol)	k_cat (min^–1)	Mesotrione IC50	Topramezone IC50	Sulcotrione IC50	Isoxaflutole IC50	Bicyclopyrone IC50	Tembotrione IC50
At WT	7.70 × 10⁻⁵	10.96	1.8 × 10⁻⁸	1.3 × 10⁻⁸	1.7 × 10⁻⁸	1.3 × 10⁻⁸	5.1 × 10⁻⁸	2.2 × 10⁻⁸
Zm WT	3.22 × 10⁻⁴	30.51	6.2 × 10⁻⁷	7.2 × 10⁻⁸	9.4 × 10⁻⁷	4.2 × 10⁻⁷	2.3 × 10⁻⁷	8.7 × 10⁻⁸
1	1.34 × 10⁻⁴	10.62	2.7 × 10⁻⁸	1.6 × 10⁻⁸	4.1 × 10⁻⁸	2.5 × 10⁻⁸	1.5 × 10⁻⁷	4.0 × 10⁻⁸
2	9.77 × 10⁻⁵	6.96	6.8 × 10⁻⁸	6.1 × 10⁻⁸	1.2 × 10⁻⁷	7.5 × 10⁻⁸	4.3 × 10⁻⁷	1.0 × 10⁻⁷

Open in a new tab

Tolerance was tested with six different inhibitors.

Table 1.

Systems with Different HPPD Proteins Investigated in this Work

Name	Organism	Ligands/mutations
At WT	At	–
Zm WT	Zm	–
Substrate	At	HPP
Product	At	HGA
1	At	M335H P336A E363Q
2	At	M335H P336A P339T T341D E363Q E426Q

Open in a new tab

Unrestrained MD simulations were run for each system for up to 200 ns. The free energy profiles were evaluated using the angle between the center‐of‐mass of the helix H11 and subsets of atoms in the H11‐preceding loop and the active site as a reaction coordinate (Fig. S1). The free energy calculations were based on Hamiltonian Replica Exchange Umbrella Sampling (H‐REUS) simulations, in which the sampling was distributed along the chosen reaction coordinate. Unfavorable trapping of simulations in local free‐energy minima was avoided by allowing configurations of neighboring windows to exchange according to a Metropolis criterion (see Methods for details). We simulated each window for 100 ns and splitting the data showed reasonable convergence in the free energy profiles (Fig. S2).

We first performed unrestrained MD simulations for At wild‐type HPPD and Zm wild‐type HPPD starting from the H11 conformations observed in the crystal structures (open and closed, respectively, PDB ID: 1SP9, 1SP8).18 Throughout the simulations, the helix H11 remained stable in its initial orientation and we observed no transition between open and closed state (Fig. S3). We also ran simulations on modeled structures of At HPPD and Zm HPPD with H11 in the closed and open state, respectively (see Methods for details). However, these simulations still showed no complete transitions, although there was a slight shift from the open conformation toward a half‐closed state for Zm HPPD (Fig. S3). The fact that we were not able to observe a transition in the unrestrained MD simulations was not unexpected. Large‐scale domain rearrangements are not only influenced by the associated change in free energy but also by the diffusivity along a given collective transition coordinate. The length of the simulation was apparently too short for protein domain rearrangements that may occur on the microsecond time‐scale or even longer time scales. We then used enhanced sampling H‐REUS simulations to calculate the free energy profiles of the H11 gating motion. The results are shown in Figure 2. For At HPPD, we found that the open H11 state is favored, whereas in Zm HPPD, the closed H11 state has a lower free energy. These observations match the orientations found in the respective crystal structures.18

Free energy profile for At wild‐type HPPD and Zm wild‐type HPPD. The observed preferred H11 state corresponds to the conformations in the respective crystal structures. The free energy profiles were calculated using enhanced sampling H‐REUS simulations.

In order to understand what drives these different preferences, we analyzed hydrogen‐bonding patterns between H11 and other parts of the protein for four unrestrained MD simulations (At HPPD with H11 in open state, At HPPD with H11 in closed state, Zm HPPD with H11 in “half‐closed” state, and Zm with H11 in closed state). For both At HPPD and Zm HPPD, H11 formed more hydrogen bonds with the rest of the protein in the closed state than in the open state. Interestingly, the total average number of hydrogen bonds in the closed H11 state was markedly higher in Zm HPPD than in the At HPPD protein (on average n_hb = 7.90 in At HPPD, n_hb = 9.23 in Zm HPPD). We then compared high occupancy H11 hydrogen bonds in the closed H11 state (>5% of the simulation time) in more detail concentrating on contacted residues that differ between At HPPD and Zm HPPD sequences. We identified a set of stable contacts that were present in the closed H11 form of Zm HPPD but not in the closed state of H11 in At HPPD, since the sites correspond to sequence differences in both enzymes (Table 2). The analysis suggests that the open H11 state in At HPPD may be favored due to higher configurational entropy and solvation of polar residues, whereas in Zm HPPD, H11 is stabilized in the closed state by forming additional contacts/hydrogen bonds.

Table 2.

Contacts of H11 with Residues Outside the Active Site in Zm HPPD with H11 in Closed State

Residue in At HPPD	H11 contacts in Zm HPPD with H11 in closed state	Corresponding amino acid in Zm HPPD
P336	S430, Y434	A
S337	Y434	P
P339	K429, E435	T
S340	E435	P
T341	Q426, K429, Y434	D

Open in a new tab

To better understand the effect of H11 gating mechanism on HPPD enzymatic activity, we simulated At HPPD in the presence of its substrate HPP and its product HGA (Fig. 3) and compared these simulations to those of the apo form. Interestingly, in unrestrained MD simulations, the substrate bound in a position as proposed by Raspail et al.16 induced a transition of H11 from the open to the closed state (Fig. 4). A video of the trajectory showing the conformational transition is provided as Supporting Information. We calculated the free energy profiles for each complex (Fig. 5) and found that substrate binding indeed shifted the preferred H11 orientation to the closed state, whereas product‐bound HPPD still favored the open state.

Modeling of substrate and product binding in HPPD. The substrate hydroxyphenylpyruvate is shown in cyan, the product homogentisate in yellow. The position of the product was modeled according to the position of a product analog in the hydroxymandelate synthase (PDB 2R5V, drawn in magenta).21 Important residues for substrate binding (Q293 and Q307) and product formation (S267 and N282) identified by Raspail et al.16 are drawn in white. The C‐terminal helix H11 and the iron ion are shown as a reference in red and orange. The figure was created with PyMOL.22

The orientation θ of helix H11 during the unrestrained molecular dynamics simulation of At HPPD with its substrate hydroxyphenylpyruvate bound to the active site. The helix H11 transitioned from the open to the closed state.

Free energy profiles for H11 gating of At HPPD in the presence of different ligands. Substrate binding shifts the conformational equilibrium toward the closed state, whereas product‐bound At HPPD still favors the open state. The free energy profiles were calculated using enhanced sampling H‐REUS simulations.

We then turned to investigate H11 gating equilibrium in At mutant HPPD protein 1 (Table 1). In this triple mutant HPPD protein, residues 336 and 363 are substituted by the corresponding residues of Zm HPPD and residue 335 has been mutated from methionine to histidine. This HPPD variant was initially discovered at BASF displaying herbicide tolerance in in vitro biochemical experiments toward in‐house HPPD inhibitors (data not shown) and to two commercially available HPPD inhibitors (Table 3, more than twofold increase in IC50 for sulcotrione and tembotrione compared to At wild‐type HPPD). 1 has been subject of a patent application.23 It is important to note that all substitutions in 1 are located outside of the active site (Fig. 1) and that the mutated residues were not implicated in catalytic functions.16 The M335H substitution most likely directly disrupts hydrophobic contacts with HPPD inhibitors (for binding modes as observed by Yang et al.19 and Brownlee et al.20). However, for the other two point mutations, it is not obvious how they could promote herbicide tolerance.

When calculating the free energy profile for 1 (Fig. 6), we found that increased herbicide tolerance in 1 compared to At wild‐type HPPD (Table 3) matched with a clear shift in the free energy profile toward favoring the closed H11 state in comparison to At wild‐type HPPD (Fig. 2). In principle, a change in the H11 gating free energy landscape can occur either by destabilizing the open state or by stabilizing the closed state by additional favorable interactions. Since the two P336A and E363Q substitutions in 1 of At HPPD correspond to the residues found in Zm HPPD, we hypothesized that these mutations may modify and bias H11 gating equilibrium toward the closed state similar to Zm HPPD. We then turned to investigate the molecular effects of these mutations in more detail.

Free energy profiles of H11 gating for At mutant HPPD. 1 and 2 clearly show a shift of the conformational H11 equilibrium toward the closed state. The free energy profiles were calculated using enhanced sampling H‐REUS simulations.

When comparing the unrestrained MD simulations of At HPPD and Zm HPPD, we identified a set of stable H11 contacts that were present in the closed form of Zm HPPD (Table 2). These sites also include P336 (corresponding to an alanine in Zm HPPD and one of the substitutions used in 1). During the MD simulation of Zm HPPD, A336 frequently formed main chain hydrogen bonds with H11, whereas no such contacts were found for the corresponding proline residue in the simulation of At HPPD with H11 in the closed state. We further analyzed the hydrogen bonds formed during the umbrella sampling simulations and observed that for Zm HPPD with H11 in the closed state, it was also possible for the neighboring residue P337 to form stable contacts with H11 (hydrogen bonds were formed either with A336 or with P337). In contrast, in At HPPD, for the corresponding residue S337, we did not detect any stable contacts with H11. However, in At mutant HPPD 1, hydrogen bond analysis showed that the mutated residue P336A formed significantly more stable contacts with H11 than in At wild‐type HPPD. The additional substitution E363Q in 1 is far from the active site and cannot form any direct contacts with H11. From the crystal structures, it is not clear how such a distant mutation could affect H11 gating. We therefore analyzed the contacts formed by E363Q in 1 and Zm wild‐type HPPD and compared it to the behavior in At wild‐type HPPD. Interestingly, in Zm HPPD, residue 363 formed hydrogen bonds to the loop in which P336A and other residues that form hydrogen bonds with H11 are located [Fig. 8(a)]. In contrast, in the At HPPD simulation, the corresponding glutamate formed salt bridges with a neighboring lysine residue and did not contact the P336 loop at all [Fig. 8(b)]. The stabilization of the loop also led to a slight widening of the active site entrance by 1 Å (data not shown) which probably makes it easier to accommodate H11 in the closed state (this slight widening can also be seen in the crystal structures of At HPPD and Zm HPPD18). Hence, by stabilizing the loop that forms hydrogen bonds with H11, the E363Q mutation might indirectly stabilize the closed H11 state.

Indirect effect of E363Q mutation on hydrogen bonding to H11. Polar contacts are drawn in yellow, H11 is shown in red. (a) When mutated to glutamine, residue 363 stabilizes a loop, which forms hydrogen bonds with H11 in the closed state in Zm HPPD but not in At HPPD. (b) In At HPPD, the corresponding glutamate forms a salt bridge with an adjacent lysine residue. The figure was created with PyMOL.22

Based on stabilizing H11 contacts identified in the previous analysis (Table 2), we designed a new At mutant HPPD protein (2, Table 1). This mutant HPPD protein 2 contains three additional mutated residues compared to 1 (P339T, T341D, and E426Q). Similar to the enzyme variant 1 the substitutions correspond to residues found in Zm at the same locations. In the Zm HPPD simulations, these residues were engaged in contacts with H11. For T341D, a contact was also detected in the crystal structure of Zm HPPD.18 Since we introduced additional residues that could promote contacts between H11 and the rest of the protein in the closed H11 state, we expected to see an even stronger shift in H11 equilibrium toward the closed form for 2 compared to 1. The free energy profile of 2 is shown in Figure 6. Indeed as expected, 2 had a clear preference for the closed H11 state in the simulations (Fig. 6). This result confirmed that we were able to correctly identify important contacts that stabilized the closed form of helix H11. 2 was then tested in biochemical in vitro experiments and also displayed increased tolerance compared to the At wild‐type HPPD and 1 (Table 3, all six inhibitors show more than twofold increase in IC50 compared to At wild‐type HPPD). The TI was especially favorable, since 2 also retained high catalytic activity (Fig. 7 and Table 3). We hence successfully managed to modify the gating equilibrium of At HPPD by introducing selected mutations and based on these modifications obtained an enzyme with increased herbicide tolerance. This successful prediction further validates our previous analysis on the role of the open‐to‐close equilibrium in herbicide tolerance and gives detailed insight into the inhibition mechanisms in HPPD.

IC50 and tolerance index for At HPPD, Zm HPPD, and At mutant HPPD proteins tested on six different commercially available inhibitors. The tolerance index was calculated as $\frac{k_{cat} IC 50}{K_{m} + S}$ for a substrate concentration of S = 5 × 10⁻⁵.

Discussion

HPPD is an important herbicide target, however, an understanding of how protein conformation affects substrate binding and inhibition is lacking to date. Here, we studied active site gating of the C‐terminal helix H11 by MD simulations and free energy calculations. We were able to reproduce experimentally observed preferences for open and closed H11 states in At HPPD and Zm HPPD crystal structures. In this work, we also analyzed how the presence of substrate and product affects the H11 open–closed equilibrium and presented evidence for a ligand‐mediated conformational switch. Our results suggest a model for HPPD gating throughout the catalytic cycle (Fig. 5). As an initial step in catalysis, the substrate binds to the active site with helix H11 in its open conformation. Substrate binding subsequently triggers a fast conformational transition into the closed state, where the active site is shielded from the solvent and catalysis can take place. Conversion of HPP to HGA with H11 in its closed state then promotes reopening of the active site for (fast) product release and subsequent substrate binding. In this way, the sensitivity of H11 gating to ligand binding could help to advance the enzyme through its catalytic cycle and increase the efficacy of the enzyme.

Furthermore, our simulations revealed changes in conformational equilibrium that could be linked to experimental observations on the effect of mutations on herbicide tolerance. Again, note that H11 gating equilibrium is certainly not the only important aspect in promoting herbicide tolerance. Despite this limitation, our work allowed us to explain in atomistic detail how residue substitutions distant from the active site can affect the H11 open–closed equilibrium and modulate binding of inhibitors, which preferentially associate to the open state. We further identified important contacts that stabilize the closed H11 state in Zm HPPD and successfully transferred these to design an At mutant HPPD protein with increased herbicide tolerance. This success demonstrates that the accuracy of the molecular mechanics simulations is in principle sufficient to correctly represent the properties of the HPPD systems. To our knowledge, this is the first time that a herbicide‐tolerant enzyme was designed based on rationally modifying its conformational equilibrium.

Our study further illustrates the importance of conformational transitions in enzyme inhibition. We were able to trace the molecular effects of each mutation due to the atomic resolution of the simulations. Our results uncovered long‐range interactions/an allosteric effect on gating by mutating residue 363; an effect that could not have been predicted based on the static crystal structures alone. The time‐scales covered in conventional MD simulations are still short compared to many interesting biological processes such as domain rearrangements and this limits the interpretation of observed events. However, with enhanced sampling methods, we were able to characterize the gating free energy in HPPD making it possible to compare different systems. With modern graphical processing unit clusters, free energy calculations can be performed in a few days. Hence, routine applications for designing herbicide‐tolerant enzymes based on different conformations are coming within reach.

Materials and Methods

Residue numberings are always given according to the At HPPD sequence from crystal structure PDB 1SP9.

Force field

The Amber99SB‐ILDN force field24 was used for the protein description and the TIP3P model25 for the water molecules. The Fe²⁺ ion was described as a single, doubly charged van der Waals particle following a strategy used in earlier work.26 The GAFF force field27 was used to describe the HPP and the HGA molecule. Force field parameters for HPP and HGA were determined with ANTECHAMBER28 via the acpype interface.29

Starting structures

Simulations were carried out using as initial structures the crystal structures of At HPPD (PDB 1SP9) and Zm HPPD (PDB 1SP8).18 We only simulated the monomer (chain A from PDB entry 1SP8 and chain B from PDB entry 1SP9). Missing residues were built with MODELLER30, 31 via its UCSF Chimera interface32 and mutations of individual residues were carried out with PyMOL.22 A model of an open H11 state for the Zm HPPD protein was built by superimposing the At HPPD coordinates onto the Zm HPPD protein, extracting the coordinates of the C‐terminal helix H11 from At HPPD, and introducing the necessary mutations in PyMOL.22 A model of the closed H11 state in At HPPD was extracted as a snapshot from the trajectory of substrate‐bound At HPPD. Structural data for the ligands were obtained from the HIC‐Up database Release 12.1.33 Substrate and product binding in the active site were modeled according to the study by Raspail et al.16 (Fig. 3). The proteins and ligands were solvated in dodecahedral boxes using a minimum solute to wall distance of 1.0 nm. Required Na⁺ and Cl⁻ ions were added to maintain electroneutrality. The resulting systems comprised up to 75,000 atoms. The structures were energy‐minimized and equilibrated in the constant pressure and temperature (NPT) ensemble for 5 ns keeping protein (and, if present, ligands) restrained using GROMACS 4.6.34, 35, 36 In total, we simulated six different systems (Table 1 and Fig. 1).

Conventional molecular dynamics

All simulations are performed in the NPT ensemble using GROMACS 4.6 (www.gromacs.org).34, 35, 36 Simulations were run for 100 ns (protein with HPP for 200 ns). The system was maintained at constant temperature (300 K) by the V‐rescale method with a coupling constant τ _t = 0.1 ps37, 38 and constant pressure (1 atm) using the Parrinello–Rahman isotropic pressure coupling with τ _p = 2.0.39 During the simulation, H‐atoms were constrained using the LINCS algorithm40 enabling a time step of 2 fs. Nonbonded interactions were calculated within a cutoff of 1.4 nm. Long‐range electrostatics were treated by the particle‐mesh Ewald electrostatics41 with the real space cutoff fixed at 1.4 nm and the highest magnitude of wave vectors were controlled by Fourier spacing with a parameter held at 0.16 nm.

Free energy calculations

H‐REUS was used to calculate the free energy profile for the gating movement of the C‐terminal helix H11. All simulations were carried out using the GROMACS 4.6 simulation package34, 35, 36 with the PLUMED2 plugin.42, 43 The umbrella sampling was performed using simulation windows in which the H11 helix was restrained to different orientations. We chose the H11 angle θ as a reaction coordinate for the free energy calculations. θ was defined between the centers of mass of three groups of residues: one group containing residues of H11 helix (423–442), one group containing residues in the H11 preceding loop (410–420), and one group containing residues in the active site (305–308, 389–392) (Fig. S1). It was ensured that the selected residues were present in both sequences. We defined the open state of the H11 helix by θ > 0.9 rad and the closed state by θ < 0.7 rad. The crystal structure of At HPPD (PDB 1SP9 18) and the crystal structure of Zm HPPD (PDB 1SP818) have θ = 1.03 rad and θ = 0.67 rad, respectively.

The initial configuration for each window was extracted from a 10 ns steered MD simulation. We used 15 simulation windows for each system spaced equally 0.04 rad apart in the range of 0.49–1.05 rad. In each of these simulation windows, we employed a harmonic umbrella potential with a force constant of 10,000 kJ mol^–1 rad^–2. Exchanges between neighboring windows were attempted every 2 ps according to a Metropolis criterion. The Monte Carlo exchange scheme avoids trapping of simulations in local minima of the free energy landscape and allows for more efficient sampling of degrees of freedom orthogonal to the reaction coordinate. Simulations were run for 100 ns per window discarding the first 50 ns as equilibration. To calculate the free energy profile from the umbrella windows, we employed the weighted histogram analysis method using the wham program by Grossfield44 with a tolerance of 0.000001. Error estimates for free energy profile were obtained according to the method proposed by Zhu and Hummer.45 Errors for each umbrella window were estimated using block averaging with the GROMACS g_analyze tool.

Analysis

Analysis was conducted using the tools provided by the GROMACS package and VMD.46 For hydrogen bond analysis, we used cutoffs of 3.5 Å and 30°. Images were generated using PyMOL22 and TeXshade.47 Plots were generated using the LaTeX pgfplots package.

Cloning and protein expression

All HPPD encoding genes and mutants thereof were synthesized by an external service provider and subcloned into a modified pET24D (Novagen, www.novagen.com) or pEXP5‐NT/TOPO® expression vector resulting in N‐terminally His‐tagged expression constructs. Recombinant protein expression of HPPD enzymes was done in Escherichia coli. To this end, chemically competent BL21 (DE3) cells (Invitrogen, Carlsbad, CA, www.thermofisher.com) were transformed with expression vectors according to the manufacturer's instructions. Transformed cells were grown in auto‐induction medium (ZYM 5052 supplemented with 100 μg/mL kanamycin or ampicillin) for 6 h at 37°C followed by 24 h at 25°C. Cells were harvested by centrifugation (8000 g) and the pellet was resuspended in Lysis‐Equilibration‐Wash Buffer (Machery‐Nagel, www.mm-net.com) supplemented with 1 mg/mL lysozyme, 5 μg/mL DNase I, and complete EDTA free protease inhibitor mix (Roche‐Diagnostics, lifescience.roche.com). For complete breakage of bacterial cells, the lysate was homogenized via sonication (6 cycles of 30 sec, on ice, 90% amplitude). The homogenate was cleared by centrifugation (40,000 g). His6‐tagged HPPD proteins were purified by affinity chromatography on a Protino Ni‐IDA 1000 Packed Column (Macherey‐Nagel) according to the manufacturer's instructions. Purified HPPD proteins were either dialyzed against 100 mM potassium phosphate buffer pH 7.0 overnight or via PD‐10 desalting columns (GE Healthcare, www.gelifesciences.com). Purified recombinant proteins are supplemented with 10% glycerin and stored at −80°C. Protein content was determined according to Bradford48 using the Bio‐Rad protein assay (Bio‐Rad Laboratories, Hercules, CA, www.bio-rad.com) and the purity of the enzyme preparation was estimated by gel electrophoresis.

Assay for HPPD activity

The activity assay for HPPD was based on the analysis of HGA by reversed phase HPLC. The assay mixture contained 150 mM potassium phosphate buffer pH 7.0, 50 mM l‐ascorbic acid, 100 μM‐catalase (Sigma‐Aldrich, www.sigmaaldrich.com), 1 μM FeSO₄, and 0.2 units of purified HPPD enzyme. 1 unit is defined as the amount of enzyme that is required to produce 1 nmol of HGA per minute at 20°C. IC50 values were determined by dilution series of HPPD inhibitors in comparison to control treatments. Inhibitors were dissolved in dimethylsulfoxide and used with 5 × 10⁻⁶ M at the highest concentration. Results were normalized by setting the uninhibited enzyme activity to 100% and IC50 values were calculated using nonlinear regression. After a 30 min pre‐incubation of enzyme assay mix and inhibitor, the reaction was started by adding 4‐hydroxyphenylpyruvate to a final concentration of 0.05 mM. The reaction was allowed to proceed for 45 min at room temperature followed by the addition of phosphoric acid to a final concentration of 400 mM. Afterward, the 96 well plates were filtered using a 0.2 μM pore size PVDF filtration device. Five microliters of the cleared sample was analyzed on an UPLC HSS T3 column (particle size 1.8 μm, dimension 2.1 × 50 mm; Waters) by isocratic elution using 90% 20 mM NaH₂PO₄ (pH 2.2) and 10% methanol (v/v). HGA was detected electrochemically at 750 mV (mode: DC; polarity: positive) and quantified by integrating peak areas (Empower software, Waters Corporation, www.waters.com). Apparent Michaelis constants (K _m) and maximal reaction velocities (V _max) were calculated by nonlinear regression with the software GraphPad Prism 5 (GraphPad Software, La Jolla, CA, www.graphpad.com) and apparent k_cat values were calculated from V_max assuming 100% purity of the enzyme preparation. Weighted means (by standard error) of K _m and IC50 values were calculated from at least three independent experiments. The Cheng–Prusoff equation for competitive inhibition49 was used to calculate dissociation constants (K _i). Easy comparison and ranking of HPPD enzymes was enabled by using the tolerance index TI that not only depends on lack of sensitivity toward HPPD inhibitors but also on the respective activity of the proteins. TI is given by

TI = \frac{k_{cat} K_{i}}{K_{m}} = \frac{k_{cat} IC 50}{K_{m} + S}

with S being the substrate concentration.

Supporting information

Figure S1 Definition of the reaction coordinate H11 angle θ. The centers of mass of the three residue groups are shown as spheres. The helix H11 is drawn in red, the iron ion in orange. For clarity, parts of the protein are not shown. We defined the open H11 conformation as θ > 0:9 rad and the closed as θ < 0:7 rad. The conformational transition corresponded to roughly a 30° rotation in this reaction coordinate.

Figure S2 Cumulative free energy pro_le for At HPPD illustrating the convergence of the simulations.

Figure S3 Orientation of helix H11 θ during the unrestrained molecular dynamics simulations. The simulations for At HPPD and Zm HPPD were started from the orientation found in the respective crystal structures (PDB ID: 1SP9, PDB ID: 1SP8). The closed form of At HPPD and the open form of Zm HPPD were modeled (see Materials and Methods for details). The open form of Zm HPPD quickly adopted a half‐closed conformation.

Click here for additional data file.^{(685.6KB, pdf)}

Acknowledgments

M.Z. and C.S. thank the Center for Integrated Protein Science Munich funded by the Deutsche Forschungsgemeinschaft for financial support. Computer resources for this project have been provided by the Leibniz Supercomputing Centre under Grant pr27za.

Significance: In this work, we used computational methods to study substrate binding, gating, and mutations leading to tolerance in an important herbicide target. We were able to link our computational results back to experiments by designing a new mutant protein. To our knowledge, we report the first rational design of a herbicide‐tolerant enzyme by modifying its conformational equilibrium. Our study further offers an example of applying free energy calculations and molecular dynamics simulations in agrochemical research.

References

1. Hausinger RP (2004) Fe(II)/α‐ketoglutarate‐dependent hydroxylases and related enzymes. Crit Rev Biochem Mol Biol 39:21–68. [DOI] [PubMed] [Google Scholar]
2. Moran GR (2005) 4‐Hydroxyphenylpyruvate dioxygenase. Arch Biochem Biophys 433:117–128. [DOI] [PubMed] [Google Scholar]
3. Goodwin TW, Mercer EI. Introduction to plant biochemistry. 2nd ed. Sydney, Australia: Pergamon Press, 1983. [Google Scholar]
4. Prisbylla M, Onisko B, Shribbs J (1993) The novel mechanism of action of the herbicidal triketones. Brighton Crop Protection Conference Weeds, Vol. 2. Brit Crop Protection Council, pp. 731–731.
5. Schulz A, Ort O, Beyer P, Kleinig H (1993) SC‐0051, a 2‐benzoyl‐cyclohexane‐1, 3‐dione bleaching herbicide, is a potent inhibitor of the enzyme p‐hydroxyphenylpyruvate dioxygenase. FEBS Lett 318:162–166. [DOI] [PubMed] [Google Scholar]
6. Hawkes T, Kramer W, Schirmer U. Modern crop protection compounds. Weinheim: Wiley‐VCH Verlag GmbH & Co KGaA, 2007. [Google Scholar]
7. Lock E, Ellis M, Gaskin P, Robinson M, Auton T, Provan W, Smith L, Prisbylla M, Mutter L, Lee D (1998) From toxicological problem to therapeutic use: the discovery of the mode of action of 2‐(2‐nitro‐4‐trifluoromethylbenzoyl)‐1, 3‐cyclohexanedione (NTBC), its toxicology and development as a drug. J Inherited Metab Dis 21:498–506. [DOI] [PubMed] [Google Scholar]
8. Grossmann K, Ehrhardt T (2007) On the mechanism of action and selectivity of the corn herbicide topramezone: a new inhibitor of 4‐hydroxyphenylpyruvate dioxygenase. Pest Manag Sci 63:429–439. [DOI] [PubMed] [Google Scholar]
9. Beaudegnies R, Edmunds AJ, Fraser TE, Hall RG, Hawkes TR, Mitchell G, Schaetzer J, Wendeborn S, Wibley J (2009) Herbicidal 4‐hydroxyphenylpyruvate dioxygenase inhibitors–a review of the triketone chemistry story from a syngenta perspective. Bioorg Med Chem 17:4134–4152. [DOI] [PubMed] [Google Scholar]
10. Ahrens H, Lange G, Müller T, Rosinger C, Willms L, van A (2013) 4‐Hydroxyphenylpyruvate dioxygenase inhibitors in combination with safeners: solutions for modern and sustainable agriculture. Angew Chem Int Ed 52:9388–9398. [DOI] [PubMed] [Google Scholar]
11. Busch M, Fischer K, Laber B, Sailland A (2011) New mutated hydroxyphenylpyruvate dioxygenase, DNA sequence and isolation of plants which are tolerant to Hppd inhibitor herbicides, European patent EP2268815A1.
12. Hinga M, Moon MS, Channarayappa VR, Rasmussen RD, Cuevas F (2014) Rice resistant/tolerant to HPPD inhibiting herbicides, US patent 2014/0059721 A1.
13. Coulombier F, Eckert H, Favre Y, Pelissier B (2012) Soybean transformation using hppd inhibitors as selection agents, US patent WO2011095460A1
14. Busch M, Fischer K, Laber B, Sailland A (2014) Mutated hydroxyphenylpyruvate dioxygenase, DNA sequence and isolation of plants which are tolerant to HPPD inhibitor herbicides, US patent WO2009144079.
15. Siehl DL, Tao Y, Albert H, Dong Y, Heckert M, Madrigal A, Lincoln‐Cabatu B, Lu J, Fenwick T, Bermudez E, Sandoval M, Horn C, Green JM, Hale T, Pagano P, Clark J, Udranszky IA, Rizzo N, Bourett T, Howard RJ, Johnson DH, Vogt M, Akinsola G, Castle LA (2014) Broad 4‐hydroxyphenylpyruvate dioxygenase inhibitor herbicide tolerance in soybean with an optimized enzyme and expression cassette. Plant Physiol 166:1162–1176. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Raspail C, Graindorge M, Moreau Y, Crouzy S, Lefèbvre B, Robin AY, Dumas R, Matringe M (2011) 4‐hydroxyphenylpyruvate dioxygenase catalysis: identification of catalytic residues and production of a hydroxylated intermediate shared with a structurally unrelated enzyme. J Biol Chem 286:26061–26070. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Serre L, Sailland A, Sy D, Boudec P, Rolland A, Pebay‐Peyroula E, Cohen‐Addad C (1999) Crystal structure of Pseudomonas fluorescens 4‐hydroxyphenylpyruvate dioxygenase: an enzyme involved in the tyrosine degradation pathway. Structure 7:977–988. [DOI] [PubMed] [Google Scholar]
18. Fritze IM, Linden L, Freigang J, Auerbach G, Huber R, Steinbacher S (2004) The crystal structures of Zea mays and Arabidopsis 4‐hydroxyphenylpyruvate dioxygenase. Plant Physiol 134:1388–1400. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Yang C, Pflugrath JW, Camper DL, Foster ML, Pernich DJ, Walsh TA (2004) Structural basis for herbicidal inhibitor selectivity revealed by comparison of crystal structures of plant and mammalian 4‐hydroxyphenylpyruvate dioxygenases. Biochemistry 43:10414–10423. [DOI] [PubMed] [Google Scholar]
20. Brownlee JM, Johnson‐Winters K, Harrison DH, Moran GR (2004) Structure of the ferrous form of 4‐hydroxyphenylpyruvate dioxygenase from streptomyces avermitilis in complex with the therapeutic herbicide ntbc. Biochemistry 43:6370–6377. [DOI] [PubMed] [Google Scholar]
21. Brownlee J, He P, Moran GR, Harrison DH (2008) Two roads diverged: the structure of hydroxymandelate synthase from amycolatopsis orientalis in complex with 4‐hydroxymandelate. Biochemistry 47:2002–2013. [DOI] [PubMed] [Google Scholar]
22. Schrödinger, LLC (2014) The PyMOL molecular graphics system, Version 1.7r1.
23. Hutzler J, Tresch S, Mietzner T, Witschel M, Lerchl J, Aponte R, Parra Rapado L, Paulik JM (2013) Plants having increased tolerance to herbicides, US patent WO2015022640A3.
24. Lindorff‐Larsen K, Piana S, Palmo K, Maragakis P, Klepeis JL, Dror RO, Shaw DE (2010) Improved side‐chain torsion potentials for the amber ff99sb protein force field. Proteins 78:1950–1958. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML (1983) Comparison of simple potential functions for simulating liquid water. J Chem Phys 79:926–935. [Google Scholar]
26. De Beer SB, Glättli A, Hutzler J, Vermeulen NP, Oostenbrink C (2011) Molecular dynamics simulations and free energy calculations on the enzyme 4‐hydroxyphenylpyruvate dioxygenase. J Comput Chem 32:2160–2169. [DOI] [PubMed] [Google Scholar]
27. Wang J, Wolf RM, Caldwell JW, Kollman PA, Case DA (2004) Development and testing of a general amber force field. J Comput Chem 25:1157–1174. [DOI] [PubMed] [Google Scholar]
28. Wang J, Wang W, Kollman PA, Case DA (2006) Automatic atom type and bond type perception in molecular mechanical calculations. J Mol Graph Model 25:247–260. [DOI] [PubMed] [Google Scholar]
29. Sousa da Silva A, Vranken W (2012) ACPYPE ‐ AnteChamber PYthon Parser interface. BMC Res Notes 5:367. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Šali A, Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234:779–815. [DOI] [PubMed] [Google Scholar]
31. Webb B, Sali A (2014) Comparative protein structure modeling using modeller. Curr Prot Bioinf 54:5–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) UCSF chimera–a visualization system for exploratory research and analysis. J Comput Chem 25:1605–1612. [DOI] [PubMed] [Google Scholar]
33. Kleywegt GJ (2007) Crystallographic refinement of ligand complexes. Acta Crystallogr Sect D Biol Crystallogr 63:94–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Berendsen HJ, van der Spoel D, van Drunen R (1995) GROMACS: a message‐passing parallel molecular dynamics implementation. Comput Phys Commun 91:43–56. [Google Scholar]
35. Hess B, Kutzner C, van Der Spoel D, Lindahl E (2008) GROMACS 4: algorithms for highly efficient, load‐balanced, and scalable molecular simulation. J Chem Theory Comput 4:435–447. [DOI] [PubMed] [Google Scholar]
36. Pronk S, Páll S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, Shirts MR, Smith JC, Kasson PM, van der Spoel D, et al. (2013) GROMACS 4.5: a high‐throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29(7):845–854. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Berendsen HJ, Postma JPM, van Gunsteren WF, DiNola A, Haak J (1984) Molecular dynamics with coupling to an external bath. J Chem Phys 81:3684–3690. [Google Scholar]
38. Bussi G, Donadio D, Parrinello M (2007) Canonical sampling through velocity rescaling. J Chem Phys 126:014101. [DOI] [PubMed] [Google Scholar]
39. Parrinello M, Rahman A (1981) Polymorphic transitions in single crystals: a new molecular dynamics method. J Appl Phys 52:7182–7190. [Google Scholar]
40. Hess B (2008) P‐LINCS: a parallel linear constraint solver for molecular simulation. J Chem Theory Comput 4:116–122. [DOI] [PubMed] [Google Scholar]
41. Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen LG (1995) A smooth particle mesh Ewald method. J Chem Phys 103:8577–8593. [Google Scholar]
42. Bonomi M, Branduardi D, Bussi G, Camilloni C, Provasi D, Raiteri P, Donadio D, Marinelli F, Pietrucci F, Broglia RA, Parrinello M (2009) Plumed: a portable plugin for free‐energy calculations with molecular dynamics. Comput Phys Commun 180:1961–1972. [Google Scholar]
43. Tribello GA, Bonomi M, Branduardi D, Camilloni C, Bussi G (2014) {PLUMED} 2: new feathers for an old bird. Comput Phys Commun 185:604–613. [Google Scholar]
44. Lin D, Grossfield A (2014) Thermodynamics of antimicrobial lipopeptide binding to membranes: origins of affinity and selectivity. Biophysical Journal 107:1862–1872. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Zhu F, Hummer G (2012) Convergence and error estimation in free energy calculations using the weighted histogram analysis method. J Comput Chem 33:453–465. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Humphrey W, Dalke A, Schulten K (1996) VMD – visual molecular dynamics. J Mol Graph 14:33–38. [DOI] [PubMed] [Google Scholar]
47. Beitz E (2000) TeXshade: shading and labeling of multiple sequence alignments using LaTeX2e. Bioinformatics 16:135–139. [DOI] [PubMed] [Google Scholar]
48. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein‐dye binding. Anal Biochem 72:248–254. [DOI] [PubMed] [Google Scholar]
49. Cheng YC, Prusoff WH (1973) Relationship between the inhibition constant (k _i) and the concentration of inhibitor which causes 50 per cent inhibition (ic ₅₀) of an enzymatic reaction. Biochem Pharmacol 22:3099–3108. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S2 Cumulative free energy pro_le for At HPPD illustrating the convergence of the simulations.

Click here for additional data file.^{(685.6KB, pdf)}

[pro3612-bib-0001] 1. Hausinger RP (2004) Fe(II)/α‐ketoglutarate‐dependent hydroxylases and related enzymes. Crit Rev Biochem Mol Biol 39:21–68. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0002] 2. Moran GR (2005) 4‐Hydroxyphenylpyruvate dioxygenase. Arch Biochem Biophys 433:117–128. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0003] 3. Goodwin TW, Mercer EI. Introduction to plant biochemistry. 2nd ed. Sydney, Australia: Pergamon Press, 1983. [Google Scholar]

[pro3612-bib-0004] 4. Prisbylla M, Onisko B, Shribbs J (1993) The novel mechanism of action of the herbicidal triketones. Brighton Crop Protection Conference Weeds, Vol. 2. Brit Crop Protection Council, pp. 731–731.

[pro3612-bib-0005] 5. Schulz A, Ort O, Beyer P, Kleinig H (1993) SC‐0051, a 2‐benzoyl‐cyclohexane‐1, 3‐dione bleaching herbicide, is a potent inhibitor of the enzyme p‐hydroxyphenylpyruvate dioxygenase. FEBS Lett 318:162–166. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0006] 6. Hawkes T, Kramer W, Schirmer U. Modern crop protection compounds. Weinheim: Wiley‐VCH Verlag GmbH & Co KGaA, 2007. [Google Scholar]

[pro3612-bib-0007] 7. Lock E, Ellis M, Gaskin P, Robinson M, Auton T, Provan W, Smith L, Prisbylla M, Mutter L, Lee D (1998) From toxicological problem to therapeutic use: the discovery of the mode of action of 2‐(2‐nitro‐4‐trifluoromethylbenzoyl)‐1, 3‐cyclohexanedione (NTBC), its toxicology and development as a drug. J Inherited Metab Dis 21:498–506. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0008] 8. Grossmann K, Ehrhardt T (2007) On the mechanism of action and selectivity of the corn herbicide topramezone: a new inhibitor of 4‐hydroxyphenylpyruvate dioxygenase. Pest Manag Sci 63:429–439. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0009] 9. Beaudegnies R, Edmunds AJ, Fraser TE, Hall RG, Hawkes TR, Mitchell G, Schaetzer J, Wendeborn S, Wibley J (2009) Herbicidal 4‐hydroxyphenylpyruvate dioxygenase inhibitors–a review of the triketone chemistry story from a syngenta perspective. Bioorg Med Chem 17:4134–4152. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0010] 10. Ahrens H, Lange G, Müller T, Rosinger C, Willms L, van A (2013) 4‐Hydroxyphenylpyruvate dioxygenase inhibitors in combination with safeners: solutions for modern and sustainable agriculture. Angew Chem Int Ed 52:9388–9398. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0011] 11. Busch M, Fischer K, Laber B, Sailland A (2011) New mutated hydroxyphenylpyruvate dioxygenase, DNA sequence and isolation of plants which are tolerant to Hppd inhibitor herbicides, European patent EP2268815A1.

[pro3612-bib-0012] 12. Hinga M, Moon MS, Channarayappa VR, Rasmussen RD, Cuevas F (2014) Rice resistant/tolerant to HPPD inhibiting herbicides, US patent 2014/0059721 A1.

[pro3612-bib-0013] 13. Coulombier F, Eckert H, Favre Y, Pelissier B (2012) Soybean transformation using hppd inhibitors as selection agents, US patent WO2011095460A1

[pro3612-bib-0014] 14. Busch M, Fischer K, Laber B, Sailland A (2014) Mutated hydroxyphenylpyruvate dioxygenase, DNA sequence and isolation of plants which are tolerant to HPPD inhibitor herbicides, US patent WO2009144079.

[pro3612-bib-0015] 15. Siehl DL, Tao Y, Albert H, Dong Y, Heckert M, Madrigal A, Lincoln‐Cabatu B, Lu J, Fenwick T, Bermudez E, Sandoval M, Horn C, Green JM, Hale T, Pagano P, Clark J, Udranszky IA, Rizzo N, Bourett T, Howard RJ, Johnson DH, Vogt M, Akinsola G, Castle LA (2014) Broad 4‐hydroxyphenylpyruvate dioxygenase inhibitor herbicide tolerance in soybean with an optimized enzyme and expression cassette. Plant Physiol 166:1162–1176. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0016] 16. Raspail C, Graindorge M, Moreau Y, Crouzy S, Lefèbvre B, Robin AY, Dumas R, Matringe M (2011) 4‐hydroxyphenylpyruvate dioxygenase catalysis: identification of catalytic residues and production of a hydroxylated intermediate shared with a structurally unrelated enzyme. J Biol Chem 286:26061–26070. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0017] 17. Serre L, Sailland A, Sy D, Boudec P, Rolland A, Pebay‐Peyroula E, Cohen‐Addad C (1999) Crystal structure of Pseudomonas fluorescens 4‐hydroxyphenylpyruvate dioxygenase: an enzyme involved in the tyrosine degradation pathway. Structure 7:977–988. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0018] 18. Fritze IM, Linden L, Freigang J, Auerbach G, Huber R, Steinbacher S (2004) The crystal structures of Zea mays and Arabidopsis 4‐hydroxyphenylpyruvate dioxygenase. Plant Physiol 134:1388–1400. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0019] 19. Yang C, Pflugrath JW, Camper DL, Foster ML, Pernich DJ, Walsh TA (2004) Structural basis for herbicidal inhibitor selectivity revealed by comparison of crystal structures of plant and mammalian 4‐hydroxyphenylpyruvate dioxygenases. Biochemistry 43:10414–10423. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0020] 20. Brownlee JM, Johnson‐Winters K, Harrison DH, Moran GR (2004) Structure of the ferrous form of 4‐hydroxyphenylpyruvate dioxygenase from streptomyces avermitilis in complex with the therapeutic herbicide ntbc. Biochemistry 43:6370–6377. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0021] 21. Brownlee J, He P, Moran GR, Harrison DH (2008) Two roads diverged: the structure of hydroxymandelate synthase from amycolatopsis orientalis in complex with 4‐hydroxymandelate. Biochemistry 47:2002–2013. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0022] 22. Schrödinger, LLC (2014) The PyMOL molecular graphics system, Version 1.7r1.

[pro3612-bib-0023] 23. Hutzler J, Tresch S, Mietzner T, Witschel M, Lerchl J, Aponte R, Parra Rapado L, Paulik JM (2013) Plants having increased tolerance to herbicides, US patent WO2015022640A3.

[pro3612-bib-0024] 24. Lindorff‐Larsen K, Piana S, Palmo K, Maragakis P, Klepeis JL, Dror RO, Shaw DE (2010) Improved side‐chain torsion potentials for the amber ff99sb protein force field. Proteins 78:1950–1958. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0025] 25. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML (1983) Comparison of simple potential functions for simulating liquid water. J Chem Phys 79:926–935. [Google Scholar]

[pro3612-bib-0026] 26. De Beer SB, Glättli A, Hutzler J, Vermeulen NP, Oostenbrink C (2011) Molecular dynamics simulations and free energy calculations on the enzyme 4‐hydroxyphenylpyruvate dioxygenase. J Comput Chem 32:2160–2169. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0027] 27. Wang J, Wolf RM, Caldwell JW, Kollman PA, Case DA (2004) Development and testing of a general amber force field. J Comput Chem 25:1157–1174. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0028] 28. Wang J, Wang W, Kollman PA, Case DA (2006) Automatic atom type and bond type perception in molecular mechanical calculations. J Mol Graph Model 25:247–260. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0029] 29. Sousa da Silva A, Vranken W (2012) ACPYPE ‐ AnteChamber PYthon Parser interface. BMC Res Notes 5:367. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0030] 30. Šali A, Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234:779–815. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0031] 31. Webb B, Sali A (2014) Comparative protein structure modeling using modeller. Curr Prot Bioinf 54:5–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0032] 32. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) UCSF chimera–a visualization system for exploratory research and analysis. J Comput Chem 25:1605–1612. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0033] 33. Kleywegt GJ (2007) Crystallographic refinement of ligand complexes. Acta Crystallogr Sect D Biol Crystallogr 63:94–100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0034] 34. Berendsen HJ, van der Spoel D, van Drunen R (1995) GROMACS: a message‐passing parallel molecular dynamics implementation. Comput Phys Commun 91:43–56. [Google Scholar]

[pro3612-bib-0035] 35. Hess B, Kutzner C, van Der Spoel D, Lindahl E (2008) GROMACS 4: algorithms for highly efficient, load‐balanced, and scalable molecular simulation. J Chem Theory Comput 4:435–447. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0036] 36. Pronk S, Páll S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, Shirts MR, Smith JC, Kasson PM, van der Spoel D, et al. (2013) GROMACS 4.5: a high‐throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29(7):845–854. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0037] 37. Berendsen HJ, Postma JPM, van Gunsteren WF, DiNola A, Haak J (1984) Molecular dynamics with coupling to an external bath. J Chem Phys 81:3684–3690. [Google Scholar]

[pro3612-bib-0038] 38. Bussi G, Donadio D, Parrinello M (2007) Canonical sampling through velocity rescaling. J Chem Phys 126:014101. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0039] 39. Parrinello M, Rahman A (1981) Polymorphic transitions in single crystals: a new molecular dynamics method. J Appl Phys 52:7182–7190. [Google Scholar]

[pro3612-bib-0040] 40. Hess B (2008) P‐LINCS: a parallel linear constraint solver for molecular simulation. J Chem Theory Comput 4:116–122. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0041] 41. Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen LG (1995) A smooth particle mesh Ewald method. J Chem Phys 103:8577–8593. [Google Scholar]

[pro3612-bib-0042] 42. Bonomi M, Branduardi D, Bussi G, Camilloni C, Provasi D, Raiteri P, Donadio D, Marinelli F, Pietrucci F, Broglia RA, Parrinello M (2009) Plumed: a portable plugin for free‐energy calculations with molecular dynamics. Comput Phys Commun 180:1961–1972. [Google Scholar]

[pro3612-bib-0043] 43. Tribello GA, Bonomi M, Branduardi D, Camilloni C, Bussi G (2014) {PLUMED} 2: new feathers for an old bird. Comput Phys Commun 185:604–613. [Google Scholar]

[pro3612-bib-0044] 44. Lin D, Grossfield A (2014) Thermodynamics of antimicrobial lipopeptide binding to membranes: origins of affinity and selectivity. Biophysical Journal 107:1862–1872. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0045] 45. Zhu F, Hummer G (2012) Convergence and error estimation in free energy calculations using the weighted histogram analysis method. J Comput Chem 33:453–465. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pro3612-bib-0046] 46. Humphrey W, Dalke A, Schulten K (1996) VMD – visual molecular dynamics. J Mol Graph 14:33–38. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0047] 47. Beitz E (2000) TeXshade: shading and labeling of multiple sequence alignments using LaTeX2e. Bioinformatics 16:135–139. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0048] 48. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein‐dye binding. Anal Biochem 72:248–254. [DOI] [PubMed] [Google Scholar]

[pro3612-bib-0049] 49. Cheng YC, Prusoff WH (1973) Relationship between the inhibition constant (k _i) and the concentration of inhibitor which causes 50 per cent inhibition (ic ₅₀) of an enzymatic reaction. Biochem Pharmacol 22:3099–3108. [DOI] [PubMed] [Google Scholar]

PERMALINK

Free energy calculations elucidate substrate binding, gating mechanism, and tolerance‐promoting mutations in herbicide target 4‐hydroxyphenylpyruvate dioxygenase

Christina E M Schindler

Eva Hollenbach

Thomas Mietzner

Klaus‐Jürgen Schleifer

Martin Zacharias

Abstract

Introduction

Figure 1.