Molecular Dynamics Simulations Elucidate Conformational Dynamics

Subscriber access provided by EMORY UNIV

Article

Molecular Dynamics Simulations Elucidate Conformational Dynamics Responsible for the Cyclization Reaction in TEAS Fan Zhang, Nanhao Chen, and Ruibo Wu J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.6b00091 • Publication Date (Web): 15 Apr 2016 Downloaded from http://pubs.acs.org on April 18, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Information and Modeling is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

Molecular Dynamics Simulations Elucidate Conformational Dynamics Responsible for the Cyclization Reaction in TEAS Fan Zhang, Nanhao Chen, Ruibo Wu*

School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou 510006, P.R. China

*To whom correspondence should be addressed. E-mail: [email protected]

TOC

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract: The Mg-dependent 5-epi-aristolochene synthase from Nicotiana tabacum (called as TEAS) could catalyze the linear Farnesyl Pyrophosphate (FPP) substrate to form the bicyclic hydrocarbon 5-epi-aristolochene. The cyclization reaction mechanism of TEAS was proposed based on static crystal structures and quantum chemistry calculations in couples of previous studies, but substrate FPP binding kinetics and protein conformational dynamics responsible for the enzymatic catalysis are still unclear. Herein, by elaborative and extensive molecular dynamics simulations, the loop conformation change and several crucial residues promoting the cyclization reaction in TEAS are elucidated. It is found that the unusual non-catalytic NH2-terminal domain is essential to stabilize the Helix-K and adjoining J-K loop of catalytic COOH-terminal domain. It is also illuminated that the induce-fit J-K/A-C loop dynamics is triggered by Y527 and optimum substrate binding mode in a “U-shape” conformation. And the U-shaped ligand binding pose is maintained well with the cooperative interaction of the three Mg2+ containing coordination shell and conserved residue W273. Furthermore, the conserved Arg residue pair R264/R266 and aromatic residue pair Y527/W273, whose spatial orientations are also crucial to promote the closure of active site to be a hydrophobic pocket, as well as to form π-stacking interactions with ligand, would facilitate the carbocation migration and electrophilic attack involved catalytic reaction. Our investigation prove the greater roles of protein local conformational dynamics than hints from the static crystal structures observations, and thus these founding would act as a guide to new protein engineering strategy on diversifying the sesquiterpene products for drug discovery.

Keywords: farnesyl pyrophosphate cyclase (FPPC); molecular dynamics; conformational dynamics; binding modes; domain-domain interaction


Page 2 of 17

Page 3 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Germacrenes, including five isomers (germacrene A/B/C/D/E, germacrene A and D are the two most prominent molecules), are typically produced in many of plant species for their antimicrobial and insecticidal properties.1 The germacrene A (see Scheme 1) is yield as an intermediate in the cyclization reaction catalyzed by 5-epi-aristolochene synthase (a sesquiterpene synthases, also called as Farnesyl Pyrophosphate cyclase, FPPC).2 FPPC belongs to terpene synthases superfamily (also referred to as terpene cyclases) which is famous for its functional role in catalyzing linear sesquiterpene substrate Farnesyl Pyrophosphate (FPP, see Scheme 1) to generate sesquiterpene products with more than 300 different carbon skeletons, including the monocyclic, bicyclic, and tricyclic skeletons. 3-5 Then the cyclized products can undergo many different modifications in plants, further increasing the diversity of natural products with a number of unusual configurations/conformations which are challenging for chemical synthesis.6 Meanwhile, nature products especially those with medium-sized rings are of great interest as novel scaffolds in drug discovery.7-9 Therefore, it is of great importance to understanding the structural basis and catalytic mechanism of FPPC. As one kind of FPPC, the 5-epi-aristolochene synthase from Nicotiana tabacum (call as TEAS) could catalyze the cyclization of linear FPP to form the bicyclic hydrocarbon 5-epi-aristolochene (see Scheme 1), which comprises the first committed step in the biosynthesis of the antifungal phytoalexin capsidiol.10, 11 Many X-ray crystal structures of TEAS have been reported to date.12-14 As seen from the TEAS-substrate complex (Figure 1a), there are three magnesium ions (Mg2+) coordination shell at the entrance of the active site, surrounded by five negatively charged amino residues and diphosphate group of FPP in the catalytic COOH catalytic domain. It is reported that the initial enzymatic reaction step (namely the release of diphosphate group and 1,10-closure cyclization, see Scheme 1) is triggered by the metal-ligand coordination interactions and hydrogen bond interactions around the diphosphate group of FPP.3 It should be pointed out that all the three Mg2+ make direct coordination interactions with diphosphate group and the coordination shell in most FPPC,15-17 but only two Mg2+ (Mg2+A and Mg2+C in Figure 1b) are coordinated with FPP directly in TEAS. The nonpolar hydrocarbon group of FPP are located in the aromatic pair area which is surrounded by Tyr527 of J-K loop (521-534) & Trp273 of helix C at the bottom of the active site (see Scheme 2 and Figure 1c). It is also reported that Tyr520 (see Scheme 2) might be a general acid to donate a proton on the intermediate germacrene A to form the bicyclic eudesmane cation by 2,7-closure cyclization.12 Finally, the 5-epi-aristolochene is yield followed by an intramolecular hydride shift and methyl migration (see Scheme 1).

Scheme 1. Proposed 5-epi-aristolochene biosynthetic mechanism catalyzed by Nicotiana tabacum 5-epi-aristolochene synthase (TEAS).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 17

Scheme 2. The bottom of the pocket of TEAS. Although the sequence similarity is not high, the catalytic domains are rich in α-helix structures and share high architecture-conservatism in most sesquiterpene synthases. Most bacterial and fungal sesquiterpene synthases such as pentalenene synthase from Streptomyces exfoliatus and aristolochene synthase from Penicillium roqueforti,18, 19 only have this α-helix rich catalytic domain.20 Whereas in TEAS, besides the α-helix rich catalytic domain (see details in Figure S1), it also contains an additional α-helical rich NH2-terminal domain.12 See from Figure 1a, this domain does not directly neighbor to the J-K loop as well as the Mg coordination shell. However, considering that the closure of active pocket would take place upon FPP binding into the FPPC21, it is necessary to understand the domain-domain interaction as well as its potential remote regulation on the ordering of J-K loop. Moreover, it is found that a significant conformation change would occur and lead to active site closure with root-mean-square deviation (rmsd) of ~ 1.5 Å in fungal terpenoid cyclases,22, 23 whereas this conformation change only results in partial ordering of loop residues to cover the active site entrance upon ligand binding to the active site of plants sesquiterpene synthases,17 especially 5-epi-aristolochene synthase, with an rmsd of 0.43 Å for 308 Cα atoms in catalytic domain.12 This distinct phenomenon on the active site closure has not been clearly understood since the structure basis of the conformation change upon substrate binding into these cyclases are still not clear. In addition, the functional role of the conserved Arg264/266 pairs in A-C loop (see Scheme 2) is also unknown. b

a

E452 T448 Mg2+B

J-K loop

D305

Mg2+A

Residue 1-36

Mg2+C

D444

D301

A-C loop FPP

c FPP

Y520

Y527 NH2-terminal domain

W273

Figure 1. (a) Structure of TEAS-substrate complex. NH2-terminal domain is rendered in blue. J-K loop and A-C loop are rendered in magenta and green. Other key residues are rendered in cyan. The Mg coordination shell and aromatic area has been highlighted at (b) and (c).


Page 5 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Although reaction mechanisms of FPPCs is extensively studied by quantum chemical calculations with the gas phase model in the past ten years,24-35 how enzyme conformation dynamics regulate FPPCs’ catalytic function is rarely illuminated. As we known, molecular dynamics (MD) simulations is an efficient and powerful tool to reveal the detailed structure basis at the atomic level. As one of the first and unique two domain containing crystal structures of sesquiterpene synthases, it is important to understand the active pocket closure process induced by substrate-loop and domain-domain interaction in TEAS. Thus we perform extensive MD simulations on various TEAS models to illuminate the structure function correlativity for substrate-loop and domain-domain interaction. Especially, the powerful Definition of Secondary Structure of Proteins (DSSP) and Principal Component Analysis (PCA) methods are introduced to figure out the relationship between NH2-terminal domain and J-K loop, as well as the induce-fit mechanism between substrate binding modes and J-K/A-C loop conformational exchange dynamics.

Methods System setup Although many X-ray crystal structures of TEAS have been reported to date, as summarized in Table S1, only a few of them reveal the J-K loop residues with side chains (PDB entry 5EAT and PDB 4RNQ).12, 14 Some structures contain substrate analogs or protein mutations (PDB entry 3M00, 3M01 and 3M02).13 The X-ray crystal structures of TEAS were all collected and downloaded from RCSB PDB. Because 3M02 contains all residues except side chains of J-K loop residues and reveals the binding mode between all three Mg2+ ions and the substrate analogue, we choose PDB entry 3M02 and the J-K loop residues 521-534 of the recently released PDB entry 4RNQ14 to rebuild the protein model of TEAS. In addition, it is reported that Mg2+ ions would not bind into the active site until substrate FPP enter into the active site,16, 21 thus none Mg2+ ion were included in the apo state simulations herein. Since the ligand in 3M02 is a cis-trans substrate analogue, we employed the substrate analogue (2-trans,6-trans)-2-fluorofarnesyl diphosphate of PDB entry 3M01 to regenerate the original substrate FPP by replacing the fluorine substituent group with hydrogen and superposing it into the rebuild TEAS scaffold. Based on the protein and substrate structure rebuilt above, different simulation models were constructed. NH2-terminal domain (see Figure S2) is cut off to make single domain (SD) systems TEAS-apo-SD and TEAS+FPP+3Mg2+-SD. So far, these reconstructed modes are the most complete enzyme-substrate/apo-enzyme complex for TEAS due to the limitation of current available XRD structures, and all simulation model sets are listed in Table 1. Table 1. Simulation model sets. Serial number 1 2 3 4 5 6 7

System TEAS-apo TEAS-apo-SD TEAS+FPP+3Mg2+ TEAS+FPP+3Mg2+-SD TEAS+FPP+ Mg2+A&C TEAS+FPP+ Mg2+A&B TEAS+FPP+ Mg2+B&C


Simulation time (ns) 200 >100 >100 >100 100


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8 9 10 11

TEAS-Y527A TEAS-Y527F TEAS-W273A TEAS-W273F

Page 6 of 17

100

Classical MD simulations and trajectory analysis. For these models, the AMBER03 force field36 was employed for the protein and the TIP3P model was used for water molecules37. The force field parameter of ligand FPP was generated from AMBER GAFF force field38, and the partial atomic charge of substrates was calculated by the restrained electrostatic potential (RESP)39 charge from HF/6-31G* calculation with the Gaussian 09 package.40 The reliability of parameter fitting procedure for the ligand was proved in our previous studies.41 The initial coordinates and topology files were generated by the tleap program in AMBER12. The MD simulations were carried out using the AMBER12 molecular simulation package.42 First, 10000 steps of minimization (including 4000 steps of steepest descent and 6000 steps of conjugate gradient) were carried out to relax the solvent, while all protein and substrates atoms were constrained by a potential of 1000 kcal/mol·Å2. Second, another 10000 steps of minimization stage was conducted with the protein backbone and heavy atoms of the substrates constrained (20 kcal/mol·Å2). Each system was then submitted to the third minimization of 10000 steps without any constraint (both second and third minimization include 5000 steps of steepest descent and 5000 steps of conjugate gradient). After minimization, each system was gradually heated from 0 to 300K over a period of ~100 ps, followed by another 100 ps of NPT MD simulations at 300 K. In the NPT ensemble, the Berendsen Thermostat method was used to control the system temperature.43 Afterward, NVT production MD simulation with a target temperature of 300 K were performed for each system to produce trajectories. Simulation time of different systems were listed in Table 1. During the MD simulations, the SHAKE algorithm44 was applied to constrain all hydrogen-containing bonds and the periodic boundary was employed. A time step of 1.0 fs was used for all systems. In order to characterize the structure transformations of J-K loop and adjoin Helix-K residues, Definition of Secondary Structure of Proteins (DSSP) and Principal Component Analysis (PCA) were performed by ptraj, a simulation analysis tools implemented in AMBER12. DSSP recognizes the structure type of selected residues and outputs a series of values representing the probability of different secondary structure type. In this work, DSSP was performed on J-K loop and adjoining Helix-K residues (521-540) of the first four systems in Table 1. Principal Component Analysis (PCA) is a good choice to further trace the conformation change process. Generally, PCA uses two or more principal components for clustering while distance or volume measurement only based on a single variable. PCA mathematically converts a number of possibly correlated multi variables into sets of linearly uncorrelated variables through orthogonal transformation, thereby reducing the dimensionality, or the number of variables, of the dataset.45 Herein PCA was performed on both J-K loop and Helix-K residues and the whole protein of the first four systems in Table 1 to detect both local and overall structure changes. The obtained data are converted to scatterplot according to simulation time.

Results and discussion 1. Function of NH2-terminal domain in TEAS .


Page 7 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In normal apo state TEAS enzyme model (namely system 1), Helix-K is stable in the 200 ns MD simulations, and the R266-E531 salt bridge is maintained well in the simulations. Strikingly, it represents quit different conformation dynamics of Helix-K and J-K loop due to the absence of NH2-terminal domain in system 2. As shown in Figure 2, the distance of Cα atoms between crucial residues Y527 and R266 is chosen to represent the connection/disconnection states between J-K loop and A-C loop in COOH-terminal domain. The two loops are closely connected (with the relative distance around 8 Å) through R266-E531 salt bridge at the beginning of the simulations. But around 100 ns, R266-E531 salt bridge is totally broken, followed by several residues flipping and rotation. As a result, Helix-K begins to untwist and comes to the maximum despiralization (see details infra). Meanwhile, J-K loop departs from A-C loop and the relative distance between two loops reaches the maximum value of 13.5 Å. When simulation is further extended, despiralization/spiralization of Helix-K are continued and thus relative distance between two loops is fluctuating in a large scale, that is, the disruption of R266-E531 salt bridge is unrecoverable. In sum, the domain-domain interaction is essential to stabilize the loop-loop interaction between J-K/A-C loops, otherwise, the two loops would be totally disconnected.

Figure 2. Representative structures of TEAS-apo-SD system during the simulation. Helix-K and J-K loop residues were rendered in green. For concision, only alpha-helix column was showed in DSSP analysis herein(See complete data in Table S2 and Table S3). In addition, the DSSP analysis was performed to further confirm the stability of Helix-K. As



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

shown in right table in Figure 2, for system 1 (normal apo form of TEAS, namely TEAS-apo model), the structural conservation of alpha-helix for each residue is very close to 100%. Whereas for system 2, in which domain-domain interaction is nonexistent, the distinct despiralization of Helix-K is observed, especially for 535~537 residues. Furthermore, 2-dimension PCA is employed to probe the secondary structures evolution of protein. The scattered distribution according to the simulation time in the PCA for J-K loop and Helix-K (see Figure 3a), is consistent with the conformation change process of J-K loop and Helix-K as discussed above in Figure 2. The two distinguishable areas in the PCA for whole protein (see green/red dash line boxes in Figure 3b), represent the connected/disconnected states of J-K loop and A-C loop as shown in Figure 2. Apparently, it confirms that connection/disconnection of two loops is taking place around 100 ns refer to the simulation time color bar, which is also in agreement with above structural basis analysis as shown in Figure 2.

Figure 3. PCA analysis of (a) J-K loop & Helix-K and (b) whole protein in TEAS-apo-SD system. The color gradual changes from red to blue corresponding to simulation time scale, while the horizontal axis and the vertical axis describe the largest two variances in the dataset. The representative connected and disconnected state of J-K and A-C loop, are referred to Figure 2. In summary, if NH2-terminal domain is absent, Helix-K would be unstable and untwist to bring the adjoining J-K loop more flexibility, and it would gradually be set apart from A-C loop. Indeed, even if the substrate FPP bound in the active site, J-K loop is also disconnected with A-C loop if NH2-terminal domain is not present (system 4, see Figure S3). In other words, J-K loop would be disordered if domain-domain interaction is absent in TEAS. Whereas for system 1&3 with the


Page 8 of 17

Page 9 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


presence of NH2-terminal domain, J-K loop is ordered to participate in the subsequent active site closure (vide infra). Therefore, NH2-/COOH-terminal domain-domain interaction has an essential effect on stabilizing Helix-K and J-K loop, which are responsible for the following substrate binding, active site closure and cyclization reaction in TEAS. 2. Induce-fit between the optimum substrate binding mode and J-K/A-C loops Many induce-fit effects between the substrate binding and active sites have been reported in FPPCs.46 Regarding to TEAS, it is proposed that the presence of substrate analogue FHP (farnesyl hydroxyphosphonate) would induce the J-K loop to form a lid that clamp down over the active site entrance based on the static XRD structure12. Herein, induce-fit phenomenon between J-K loop and substrate FPP is observed directly at atomic level by our full-atom MD simulations, as shown in Figure 4ab, a dispersive distribution along PC1 can be found in the whole protein PCA of apo state (system 1) whereas both convergent distribution along PC1 and PC2 axis for substrate bound state (system 3). In comparison between Figure 4c and 4d, it is further confirmed that J-K loop is flexible in apo state while it is stable and ordered after substrate FPP binding into the active site. In addition, the root-mean-square fluctuation curve of carbon atoms for systems 1&2&3 is plotted in Figure S4. It also verified that J-K loop is very flexible in apo form and it will become ordered after substrate FPP bound. Therefore, we concluded that substrate binding has distinct induce-fit effect on the conformational change of J-K loop and then J-K loop would participate in the active site closure process.

Figure 4. Whole protein (a,b) and J-K loop (c,d) 2-dimention PCA analysis for TEAS-apo and TEAS+FPP+3Mg2+ model. Regarding to closure state of the active site, as shown in Figure 5, the more flexible J-K loop serve as the “door-panel” and less flexible A-C loop like as “door-frame”, and the loop-loop



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

interactions between them is like the “door-latch”. Particularly, R266 is the “key” to open/close the “door” which would lead to a “room” (binding pocket) for accommodating the substrate FPP. As shown in Figure S5, R266 firstly interacts with the main chain oxygen of Y527 and T528, it could be thought as a “half-close” state, same as some crystal structures observation12, 14, 47. Whereas R266 would turn outwards and make a salt bridge interaction with E531 to restraint the conformational dynamics of J-K loop and thus keep the active site in a “full-close” state. Unfortunately, the “full-close” XRD structure is not available to verify this computationally detected “full-close” state. Nevertheless, as this kind of R-E salt bridge is so stable and conserved in various TEAS dimmer models, it is believed that formation of R266-E531 salt bridge is the essential step to close the active site. Besides, R264 is another conserved polar residue among many different terpene cyclases. Herein in TEAS, positively charged R264 directly interacts with negatively charged phosphate group of FPP via electrostatic interaction (also namely salt bridge), combined with the Mg2+ coordination interaction as well as π-stacking of W273 (will be discussed in infra), leading to the “U-shape” conformation of FPP. In sum, this accurate substrate binding mode firstly stimulates the conformation transformation of J-K loop to be ordered, and the conserved Arg264/266 pair are also important for the induced-fit stimulated active site closure process, Arg264 takes in charge of accurate substrate binding and Arg266 directly involves in the subsequent active site full close. Considering the conservation of the Arg pair in most FPPCs, the functional roles of Arg264/266 might be also existed in other FPPCs.

Figure 5. Representative full-close active site in TEAS+FPP+3Mg2+ model. J-K loop and A-C loop are rendered in green and cyan respectively. Moreover, on the basis of the proposed reaction mechanism as shown in Scheme 1, the above mentioned U-shaped conformation of FPP is the most optimum binding mode in TEAS, as it is a reasonable “Near-Attack-Conformation” for the initial enzymatic catalysis which is beginning by the release of diphosphate group and 1,10-closure cyclization. As only when substrate FPP binding into the active site in the optimum binding pose with a U-shaped conformation, the above mentioned induce-fit conformation change between J-K/A-C loops would happen. To answer the query how does enzyme keep FPP binding in “U-shape” conformation, different Mg2+ binding systems are considered (system 5-7 in Table 1). See from Figure 6ab, the binding pose of FPP and the orientation of W273 and Y527 are similar either in Mg2+B-absent or Mg2+B-present models. Strikingly, if Mg2+A ion is absent, the FPP would adopt an incorrect binding mode with a linear conformation of farnesyl chain, and lead to a unconventional spatial orientation of W273 (Figure 5ac). Regarding to nonexistence of Mg2+C ion, the FPP will further enter into the bottom of the


Page 10 of 17

Page 11 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


active site, as a result, Y527 will not facing towards FPP farnesyl chain (Figure 5ad). Therefore, the two ions (Mg2+A and C in our models) serve as important anchors to restraint FPP in a specific “U-shape” conformation by strong coordination interaction, whereas the Mg2+B has indistinctive effect, which is similar to the three-Mg containing geranyl pyrophosphate synthase in our previous quantum mechanics/molecular mechanics (QM/MM) MD simulations.48 Nevertheless, all the three Mg2+ ions would serve as Lewis acid to polarize the pyrophosphate group in the subsequent cyclization reaction.

Figure 6. FPP binding modes comparison among various TEAS. (a) Normal wild type TEAS. (b) TEAS without Mg2+ B. (c) TEAS without Mg2+ A. (d) TEAS without Mg2+ C. 3. Induce-fit between substrate and W273/Y527 aromatic pair. The induce-fit conformation change is also occurring at the bottom aromatic area of active pocket, as shown in Figure 7, in contrast to the apo state, when substrate binding into the pocket, Y527 will approach to W273 (from 8.5 to 5.1 Å), and the “T-shape” (the aromatic side chain dihedral is close to 90゜) π-π stacking between W273 and Y527 in apo state, while it would be “shift-parallel” π-π stacking with the smaller aromatic-ring dihedral, leading to an extended aromatic box at the bottom of the active site pocket. In details, W273 makes π-π stacking with C6-C7 double bond. Meanwhile, Y527 makes critical cation-π stacking with carbocation after the protonation steps of reaction intermediate germacrene A (See Scheme 1). Both the above mentioned π-stacking interaction are favorable to the involved electrophilic reaction. Thus, it is suspected that FPP binding has strong induction effect on the aromatic pair W273 and Y527 to provide a rational reactive binding site.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7. The Y527/W273 aromatic pairs in TEAS-apo (left) and TEAS+FPP+3Mg2+ (right). The center of mass distance between two aromatic rings and its dihedral angle are measured.

Figure 8. Superposition of wild type TEAS and mutations on (a) W273A, (b) W273F, (c) Y527A and (d) Y527F (cyan for wild type and yellow for mutant model). In order to understand the roles of aromatic W273/Y527 residues, various mutants on wild type TEAS-substrate complex were performed herein. See from the structure superposition between wild type and mutants (Figure 8), the interaction between J-K loop and A-C loop is lost and the rational FPP binding conformation is damaged by W273A mutant. While in W273F model, the loop-loop interaction is maintained but conformation dynamics of J-K loop would be different


Page 12 of 17

Page 13 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


from that in wild type, with Y527 in an inappropriate orientation towards to FPP. As for Y527 mutants, the U-shaped FPP conformation is maintained well but J-K loop becomes more flexible, as a result, J-K loop is disconnected with A-C loop to form a “open” active site. In contrast, the bottom aromatic area is surround by W273/Y527 and the whole active pocket is closed by J-K/A-C loops, then a solvent-inaccessible pocket is constructed in the wild type TEAS, with the nonpolar hydrocarbon group of FPP is relatively buried in the hydrophobic active site pocket. As shown in Figure 9, in normal wild type TEAS, solvent number maintains at a low level. Actually only one water molecule is found around the substrate (Figure S6), and this water molecule is in the initial experimental crystal structure. Whereas for the single domain TEAS, as it would result in an open, solvent-exposed pocket, many water molecules would come into the pocket finally, therefore it further confirms the essential role of domain-domain interaction to decrease the J-K loop flexibility and construct a hydrophobic pocket. When Y527 is mutated, FPP binding would not further induce J-K loop conformation change to a close state, thus corresponding to an early and rapid active site solvent molecule growth curve as shown in Figure 9. As we known, the low dielectric, that is hydrophobic environment, is necessary for this kind of cyclization reaction in which involves proton/carbocation immigration in our previous study.49 The above discussed closed protein conformation of wild type would shield reactive carbocation intermediates from solvent Otherwise, it will be unfavorable for intermediate stabilization and may results in premature quenching of the reaction. Meanwhile, based on the previous proposed reaction mechanism (Scheme 1), it is believed that W273 and Y527 are very crucial in the active site. The above mentioned π-stacking by W273 is preferable to locate the electrophilic center at C7 in methyl migration12. And W273 could also help anchor FPP in a U-shaped conformation with C6-C7 double bond approaching to C1/C2/C3, which would promote the initial 1,10-closure and following protonation transfer reaction. Regarding to Y527, as an inductor, it would trigger the “door-panel” J-K loop gradually serve as a lid to cover the active site and the R266-E531 salt bridge and lock the active site in a close state to bring a hydrophobic pocket. Meanwhile, the subsequent 2,7-closure cyclization reaction would be promoted by Y527, since the methyl group on C7 is perpendicular to aromatic ring plane of Y527 to provide maximum reactive cation-π interaction (as shown in Figure 5 and Figure 7(right)).

Figure 9. Analysis of water number around TEAS active site. See details in Figure S6.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Conclusion By extensive molecular dynamics simulations, it is elucidated that the unusual NH2-terminal domain is essential to stabilizing Helix-K and ordering J-K loop in TEAS, to protect the loop-loop interaction especial R266-E531 salt bridge between J-K and A-C loop. One hand, the two of three Mg2+ ions anchored the head (phosphate group) of substrate FPP by strong coordination shell, the other hand, W273 contact to the tail of FPP by the stacking interaction, leading the FPP to be a specific “U-shape” conformation (namely “Near-Attack-Conformation”) that facilitate the first step of cyclization reaction. This optimum substrate binding mode is also suspected to play important roles in induce-fit J-K loop conformation dynamics through the inductor, Y527. Finally, this crucial domain-domain interaction, combined with the presence of an “aromatic box” surround by W273/Y527, then a solvent-inaccessible pocket is constructed and kept well, and it is favorable for the subsequent proton/carbocation-migration implicated cyclization reaction. These insights on understanding the TEAS structural features and conformation dynamics as well as its functional roles would provide us useful clues on further studying the detailed catalytic cyclization reaction mechanism. Besides, since it is desirable to increase the natural products complexity for modern drug discovery, these founding also shed lights on designing new protein engineering strategy to modify the product diversity of plant sesquiterpene synthases by rational mutagenesis.

ASSOCIATED CONTENT Supporting Information Figure S1-S6 and Table S1-S3. TEAS active pocket; NH2-terminal domain; representative structures of TEAS-FPP+3Mg2+-SD model; RMSF curve of COOH-terminal domain; formation of E-R salt bridge; water shell analysis around active site; Summary of TEAS crystal structures; DSSP analysis of TEAS-apo and TEAS-apo-SD. This information is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Authors: *E-mail: [email protected]

Notes The authors declare no competing financial interest.

Acknowledgments This work was supported by the National Science Foundation of China (21272289) and Pearl River S&T Nova Program of Guangzhou (2014J2200062). We also thank the National Supercomputing Centers in Shenzhen and Guangzhou for providing the computational resources.


Page 14 of 17

Page 15 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References (1) Adio, A. M., Germacrenes A–E and Related Compounds: Thermal, Photochemical and Acid Induced Transannular Cyclizations. Tetrahedron 2009, 68, 1533-1552. (2) Deguerry, F.; Pastore, L.; Wu, S.; Clark, A.; Chappell, J.; Schalk, M., The Diverse Sesquiterpene Profile of Patchouli, Pogostemon Cablin, Is Correlated with a Limited Number of Sesquiterpene Synthases. Arch. Biochem. Biophys. 2006, 454, 123-136. (3) Christianson, D. W., Structural Biology and Chemistry of the Terpenoid Cyclases. Chem. Rev. 2006, 106, 3412-3442. (4) Davis, E. M., and Croteau, R., Cyclization Enzymes in the Biosynthesis of Monoterpenes, Sesquiterpenes, and Diterpenes. Top. Curr. Chem. 2000, 209, 53−95. (5) Miller, D. J.; Allemann, R. K., Sesquiterpene Synthases: Passive Catalysts or Active Players? Nat. Prod. Rep. 2012, 29, 60-71. (6) Marsault, E.; Peterson, M. L., Macrocycles Are Great Cycles: Applications, Opportunities, and Challenges of Synthetic Macrocycles in Drug Discovery. J. Med. Chem. 2011, 54, 1961-2004. (7) Parenty, A.; Moreau, X.; Campagne, J. M., Macrolactonizations in the Total Synthesis of Natural Products. Chem. Rev. 2006, 106, 911-939. (8) Driggers, E. M.; Hale, S. P.; Lee, J.; Terrett, N. K., The Exploration of Macrocycles for Drug Discovery--an Underexploited Structural Class. Nat. Rev. Drug Discovery 2008, 7, 608-624. (9) Mallinson, J.; Collins, I., Macrocycles in New Drug Discovery. Future Med. Chem. 2012, 4, 1409-1438. (10) Vogeli, U.; Freeman, J. W.; Chappell, J., Purification and Characterization of an Inducible Sesquiterpene Cyclase from Elicitor-Treated Tobacco Cell Suspension Cultures. Plant Physiol. 1990, 93, 182-187. (11) Whitehead, I. M.; Ewing, D. F.; Threlfall, D. R.; Cane, D. E.; Prabhakaran, P. C., Synthesis of (+)-5-Epi-Aristolochene and (+)-1-Deoxycapsidiol from Capsidiol. Phytochemistry 1990, 29, 479-482. (12) Starks, C. M., Structural Basis for Cyclic Terpene Biosynthesis by Tobacco 5-Epi-Aristolochene Synthase. Science 1997, 277, 1815-1820. (13) Noel, J. P.; Dellas, N.; Faraldos, J. A.; Zhao, M.; Hess, B. A., Jr.; Smentek, L.; Coates, R. M.; O'Maille, P. E., Structural Elucidation of Cisoid and Transoid Cyclization Pathways of a Sesquiterpene Synthase Using 2-Fluorofarnesyl Diphosphates. ACS Chem. Biol. 2010, 5, 377-392. (14) Rising, K. A.; Crenshaw, C. M.; Koo, H. J.; Subramanian, T.; Chehade, K. A.; Starks, C.; Allen, K. D.; Andres, D. A.; Spielmann, H. P.; Noel, J. P.; Chappell, J., Formation of a Novel Macrocyclic Alkaloid from the Unnatural Farnesyl Diphosphate Analogue Anilinogeranyl Diphosphate by 5-Epi-Aristolochene Synthase. ACS Chem. Biol. 2015, 10, 1729-1736. (15) Vedula, L. S.; Zhao, Y.; Coates, R. M.; Koyama, T.; Cane, D. E.; Christianson, D. W., Exploring Biosynthetic Diversity with Trichodiene Synthase. Arch. Biochem. Biophys. 2007, 466, 260-266. (16) Shishova, E. Y.; Yu, F.; Miller, D. J.; Faraldos, J. A.; Zhao, Y.; Coates, R. M.; Allemann, R. K.; Cane, D. E.; Christianson, D. W., X-Ray Crystallographic Studies of Substrate Binding to Aristolochene Synthase Suggest a Metal Ion Binding Sequence for Catalysis. J. Biol. Chem. 2008, 283, 15431-15439. (17) Gennadios, H. A.; Gonzalez, V.; Di Costanzo, L.; Li, A.; Yu, F.; Miller, D. J.; Allemann, R. K.; Christianson, D. W., Crystal Structure of (+)-Delta-Cadinene Synthase from Gossypium Arboreum and Evolutionary Divergence of Metal Binding Motifs for Catalysis. Biochemistry 2009, 48, 6175-6183. (18) Lesburg, C. A., Crystal Structure of Pentalenene Synthase: Mechanistic Insights on Terpenoid Cyclization Reactions in Biology. Science 1997, 277, 1820-1824.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(19) Caruthers, J. M.; Kang, I.; Rynkiewicz, M. J.; Cane, D. E.; Christianson, D. W., Crystal Structure Determination of Aristolochene Synthase from the Blue Cheese Mold, Penicillium Roqueforti. J. Biol. Chem. 2000, 275, 25533-25539. (20) Wendt, K. U.; Schulz, G. E., Isoprenoid Biosynthesis: Manifold Chemistry Catalyzed by Similar Enzymes. Structure 1998, 6, 127-133. (21) van der Kamp, M. W.; Sirirak, J.; Zurek, J.; Allemann, R. K.; Mulholland, A. J., Conformational Change and Ligand Binding in the Aristolochene Synthase Catalytic Cycle. Biochemistry 2013, 52, 8094-8105. (22) Rynkiewicz, M. J.; Cane, D. E.; Christianson, D. W., Structure of Trichodiene Synthase from Fusarium Sporotrichioides Provides Mechanistic Inferences on the Terpene Cyclization Cascade. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 13543-13548. (23) Shishova, E. Y.; Di Costanzo, L.; Cane, D. E.; Christianson, D. W., X-Ray Crystal Structure of Aristolochene Synthase from Aspergillus Terreus and Evolution of Templates for the Cyclization of Farnesyl Diphosphate. Biochemistry 2007, 46, 1941-1951. (24) Hong, Y. J.; Tantillo, D. J., Which Is More Likely in Trichodiene Biosynthesis: Hydride or Proton Transfer? Org. Lett. 2006, 8, 4601-4604. (25) Gutta, P.; Tantillo, D. J., Theoretical Studies on Farnesyl Cation Cyclization: Pathways to Pentalenene. J. Am. Chem. Soc. 2006, 128, 6172-6179. (26) Lodewyk, M. W.; Gutta, P.; Tantillo, D. J., Computational Studies on Biosynthetic Carbocation Rearrangements Leading to Sativene, Cyclosativene, Alpha-Ylangene, and Beta-Ylangene. J. Org. Chem. 2008, 73, 6570-6579. (27) Hong, Y. J.; Tantillo, D. J., Modes of Inactivation of Trichodiene Synthase by a Cyclopropane-Containing Farnesyldiphosphate Analog. Org. Biomol. Chem. 2009, 7, 4101-4109. (28) Zu, L.; Xu, M.; Lodewyk, M. W.; Cane, D. E.; Peters, R. J.; Tantillo, D. J., Effect of Isotopically Sensitive Branching on Product Distribution for Pentalenene Synthase: Support for a Mechanism Predicted by Quantum Chemistry. J. Am. Ceram. Soc. 2012, 134, 11369-11371. (29) Hong, Y. J.; Irmisch, S.; Wang, S.; Garms, S.; Gershenzon, J.; Zu, L.; Kollner, T.; Tantillo, D. J., Theoretical and Experimental Analysis of the Reaction Mechanism of Mrtps2, a Triquinane-Forming Sesquiterpene Synthase from Chamomile. Chem. Eur. J. 2013, 19, 13590-13600. (30) Isegawa, M.; Maeda, S.; Tantillo, D. J.; Morokuma, K., Predicting Pathways for Terpene Formation from First Principles – Routes to Known and New Sesquiterpenes. Chem. Sci. 2014, 5, 1555. (31) Hong, Y. J.; Giner, J.-L.; Tantillo, D. J., Bicyclobutonium Ions in Biosynthesis – Interconversion of Cyclopropyl-Containing Sterols from Orchids. J. Am. Chem. Soc. 2015, 137, 2085-2088. (32) Hamlin, T. A.; Hamann, C. S.; Tantillo, D. J., Delocalization of Charge and Electron Density in the Humulyl Cation-Implications for Terpene Biosynthesis. J. Org. Chem. 2015, 80, 4046-4053. (33) Wedler, H. B.; Pemberton, R. P.; Tantillo, D. J., Carbocations and the Complex Flavor and Bouquet of Wine: Mechanistic Aspects of Terpene Biosynthesis in Wine Grapes. Molecules 2015, 20, 10781-10792. (34) Pemberton, R. P.; Ho, K. C.; Tantillo, D. J., Modulation of Inherent Dynamical Tendencies of the Bisabolyl Cation Via Preorganization in Epi-Isozizaene Synthase. Chem. Sci. 2015, 6, 2347-2353. (35) Hare, S. R.; Tantillo, D. J., Dynamic Behavior of Rearranging Carbocations – Implications for Terpene Biosynthesis. Beilstein J. Org. Chem. 2016, 12, 377-390. (36) Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M. C.; Xiong, G.; Zhang, W.; Yang, R.; Cieplak, P.; Luo, R.; Lee, T.; Caldwell, J.; Wang, J.; Kollman, P., A Point-Charge Force Field for Molecular Mechanics Simulations


Page 16 of 17

Page 17 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


of Proteins Based on Condensed-Phase Quantum Mechanical Calculations. J. Comput. Chem. 2003, 24, 1999-2012. (37) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L., Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926-935. (38) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A., Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004, 25, 1157-1174. (39) Bayly, C. I.; Cieplak, P.; Cornell, W.; Kollman, P. A., A Well-Behaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: The Resp Model. J. Phys. Chem. 1993, 97, 10269-10280. (40) Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.;; Robb, M. A. C., J. R.; Scalmani, G.; Barone, V.; Mennucci,; B.; Petersson, G. A. N., H.; Caricato, M.; Li, X.; Hratchian, H.; P.; Izmaylov, A. F. B., J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;; Ehara, M. T., K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,; T.; Honda, Y. K., O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;; Peralta, J. E. O., F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,; K. N.; Staroverov, V. N. K., R.; Normand, J.; Raghavachari, K.;; Rendell, A. B., J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,; N.; Millam, J. M. K., M.; Knox, J. E.; Cross, J. B.; Bakken, V.;; Adamo, C. J., J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;; Austin, A. J. C., R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;; Morokuma, K. Z., V. G.; Voth, G. A.; Salvador, P.;; Dannenberg, J. J. D., S.; Daniels, A. D.; Farkas, O.;; Foresman, J. B. O., J. V.; Cioslowski, J.; Fox, D. J., Gaussian 09, Gaussian Inc.: Wallingford, CT, 2009. (41) Liu, Z.; Zhou, J.; Wu, R.; Xu, J., Mechanism of Assembling Isoprenoid Building Blocks 1. Elucidation of the Structural Motifs for Substrate Binding in Geranyl Pyrophosphate Synthase. J. Chem. Theory Comput. 2014, 10, 5057-5067. (42) Case, D. A. D., T. A.; Cheatham, T. E., III; Simmerling, C.; L.; Wang, J. D., R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K.; M.; Roberts, B. H., S.; Roitberg, A.; Seabra, G.; Swails, J.; Goetz, A.; W.; Kolossvary, I. W., K. F.; Paesani, F.; Vanicek, J.; Wolf, R. M.;; Liu, J. W., X.; Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye,; X.; Wang, J. H., M. J.; Cui, G.; Roe, D. R.; Mathews, D. H.; Seetin,; M. G.; Salomon-Ferrer, R. S., C.; Babin, V.; Luchko, T.; Gusarov,; S.; Kovalenko, A. K., P. A., AMBER12, University of California: San Francisco, CA, 2012. (43) Berendsen, H. J. C. P., J. P. M.; Van Gunsteren, W. F.;; DiNola, A. H., J. R., Molecular Dynamics with Coupling to an External Bath. J. Chem. Phys 1984, 81, 3684-3690. (44) Ryckaert, J. P. C. G. B., H. J. C., Numerical Integration of the Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of N-Alkanes. J. Comput. Phys. 1977, 23, 327-341. (45) Liu, Y. L.; Lindert, S.; Zhu, W.; Wang, K.; McCammon, J. A.; Oldfield, E., Taxodione and Arenarone Inhibit Farnesyl Diphosphate Synthase by Binding to the Isopentenyl Diphosphate Site. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, E2530-2539. (46) Baer, P.; Rabe, P.; Fischer, K.; Citron, C. A.; Klapschinski, T. A.; Groll, M.; Dickschat, J. S., Induced-Fit Mechanism in Class I Terpene Cyclases. Angew. Chem., Int. Ed. 2014, 53, 7652-7656. (47) Kersten, R. D.; Diedrich, J. K.; Yates, J. R., 3rd; Noel, J. P., Mechanism-Based Post-Translational Modification and Inactivation in Terpene Synthases. ACS Chem. Biol. 2015, 10, 2501-2511. (48) Zhou, J.; Wang, X.; Kuang, M.; Wang, L.; Luo, H.-B.; Mo, Y.; Wu, R., Protonation-Triggered Carbon-Chain Elongation in Geranyl Pyrophosphate Synthase (Gpps). ACS Catal. 2015, 5, 4466-4478. (49) Chen, N.; Wang, S.; Smentek, L.; Hess, B. A., Jr.; Wu, R., Biosynthetic Mechanism of Lanosterol: Cyclization. Angew. Chem., Int. Ed. 2015, 54, 8693-8696.


Molecular Dynamics Simulations Elucidate Conformational Dynamics

Recommend Documents