Biosynthesis of Spinosyn A: A [4 + 2] or [6 + 4] Cycloaddition? - ACS

Feb 1, 2018 - 2016 , 138 , 3631−3634) shows that the exact mechanism of SpnF becomes even more interesting as well as now being controversial...
0 downloads 0 Views 1MB Size
Subscriber access provided by READING UNIV

Article

Biosynthesis of Spinosyn A: a [4+2] or [6+4] Cycloaddition? Nan-Hao Chen, Fan Zhang, Ruibo Wu, and bernard Andes hess Jr ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b03908 • Publication Date (Web): 01 Feb 2018 Downloaded from http://pubs.acs.org on February 2, 2018

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.

ACS Catalysis 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 7 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

ACS Catalysis

Biosynthesis of Spinosyn A: a [4+2] or [6+4] Cycloaddition? Nanhao Chen†,‡, Fan Zhang†, Ruibo Wu*,†, B. Andes Hess, Jr.*,§ †School

of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou 510006, P.R. China of Chemistry, University of California, Davis, California 95616, United States §Department of Chemistry, Vanderbilt University, Nashville, TN37235, United States ‡Department

ABSTRACT: SpnF, one of the Diels-Alderases, produces spinosyn A, which has been demonstrated that its sole function is to catalyze the [4+2] cycloaddition (Fage C. D. et al. Nat. Chem. Bio. 2015, 11, 256-258). Furthermore, the potential existence of a [6+4] cycloaddition bifurcation from previous theoretical calculations on the non-enzyme model (Patel A. et al. J. Am. Chem. Soc. 2016, 138, 3631-3634), shows that the exact mechanism of SpnF becomes even more interesting as well as now being controversial. Herein QM(DFT)/MM MD simulations on the full-enzyme model reveal three significant residues collaborating with other residues that control the direction of the cycloaddition, namely Tyr23, Thr196 and Trp256. These residues force the substrate into a reactive conformation that causes the cycloaddition reaction to proceed through a [4+2] pathway instead of the [6+4] one. The mechanistic insights deciphered here is fundamentally important for the rational design of Diels-Alderases and biomimetic syntheses.

The Diels-Alder (DA) [4+2] cycloaddition1 is a widely used and powerful strategy in synthetic organic chemistry.2 It forms the cyclohexene system through a reaction between a conjugated diene and an alkene. This reaction has also been proposed as a key transformation in the biosynthesis of a number of cyclohexene-containing metabolites. However, it has rarely been adopted by enzymes in the synthesis of natural products.3 There are only a few enzymes that have been shown to promote the DA cycloadditions4,5, including Lovastatin Synthase (LovB),6 riboflavin synthase,7 macrophomate synthase8-10 and Spinosyn A synthase F (SpnF)11-13. Among these enzymes, the SpnF is the only one whose sole function has been shown to be the catalysis of the [4+2] cycloaddition.4 This system has recently received much attention in order to discover the fundamental mechanism of this enzyme. As shown in Figure 1(a), a previous theoretical study of this reaction by Hess and Smentek revealed that the cycloaddition reaction in SpnF was concerted but highly asynchronous with a 21.0 kcal/mol barrier.14 In 2016, owing to the existence of a triene and diene in the ligand, Houk proposed a different reaction pathway that is likely to first yield a [6+4] intermediate (highlighted in Figure 1(a)), which subsequently generates the [4+2] cycloaddition product. Based on the gas phase and dynamics calculation, a “bis-pericyclic” transition structure was located that connects the [4+2] substrate and the [6+4] intermediate and that features two sets of stabilizing cyclic aro-

matic orbital interactions.15 However, the barrier is much higher than the experimental results (27.6 kcal/mol vs. 18.6 kcal/mol). In addition, Quantum Mechanical/Molecular Mechanical Molecular Dynamics (QM/MM MD) simulations (without considering the whole enzyme in the modeling) were employed to demonstrate the probability of this [6+4] reaction.15 In the early of 2017, Michael Medvedev and his coworkers applied exhaustive QM MD simulations in the implicit water model and proposed that although the DA TSs are more easily accessible, the bis-pericyclic TSs are thermally dominant overall.16 Even if this bis-pericyclic pathway is a potentially reasonable reaction pathway, it is still unclear how the SpnF enzyme realizes the dominant outcome of the ultimate [4+2] product. Does the reaction go through the [6+4] pathway or by the [4+2] pathway directly? How might the

Figure 1. (a) The function and catalytic mechanism hypothesis of the SpnF, (b) the model definition based on different QM sites, (c) the substrate structure and definition of the reaction coordinate. The black dashed lines in (b) and (c) indicate the hydrogen bonds between atoms, while blue dashed lines show the reaction coordinates.

ACS Paragon Plus Environment

ACS Catalysis 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 2. (a) The free energy profiles of Model C with M06-2X and B3LYP methods with the 6-31G* basis set, and (b and c) their [4+2] pathway TS structures. The dashed arrows in (a) point to the values of the reaction coordinate in different models and the numbers beside the free energy curves are the energy values of the key states in the different models. For the structures, the hydrogen bonds are shown in black dashed line and the key distances are in blue dashed lines. The numbers are the average values of the key distances and the ones in the parentheses are the standard error values. enzyme direct the reaction through the [6+4] intermediate, which is also a stable state? Recently, the semi-empirical QM/MM MD simulations by Yiying Zhang and Walter Thiel suggested that SpnF does behave as a Diels-Alderase and the residue, Thr196, played a key role in promoting the DielsAlder reaction while impeding the [6+4] cycloaddition. Even though the whole enzyme has been considered in the semiempirical simulations, the roles of the other residues and the product selectivities have not been well discussed.17 These remarkable chemical selectivity issues will be addressed in this Communication by using two-dimensional umbrella sampling, with more than 20 ns QM/MM MD simulations of the whole enzyme environment. Fortunately, the X-ray crystal structure of SpnF (PDB ID: 4PNE) was solved in early 2015, which makes it possible to apply bio-simulations to study the role of the protein.13 According to the crystal structure and the docking results (in Figure 1(c)), several residues likely play a crucial role in the enzyme catalysis. Two aromatic residues, Tyr23 and Trp256, are located around the diene and triene while three polar residues form hydrogen bonds directly with the ligand, especially Thr196, which is close to the catalytic area. As in Houk’s paper15, the intramolecular hydrogen bond was found in the SpnF substrate. However, this intramolecular interaction disappeared when considering the existence of the enzyme, which is consistent with Fage’s13 and Zhang’s17 results. In this study, QM/MM MD simulations18 have been used to investigate the detailed function of SpnF. In order to precisely reveal the roles of some key residues, three models were prepared based on different QM regions (in Figure 1(b)). In Model A only the ligand and two residues, Glu152 and Trp256, were included in the QM area so that we could focus on the function of the aromatic residue Trp256 and the negatively charged residue Glu152. Another aromatic residue, Tyr23, was added to the QM area for Model B. Also, Thr196, which forms a hydrogen bond with the ligand, was described by the QM method in Model C. The enzyme-substrate complex was built by the Schrödinger suite.19 All the QM/MM MD simulations were run with QChem20 and Amber21. Houk has reported that dispersion is significant in obtaining the [6+4] product.15 Herein, two methods have been employed, namely B3LYP22,23 and M062X24,25. M06-2X has some implicit dispersion corrections while B3LYP does not contain any dispersion correction.26,27 Additional computational details are given in supporting information.

As illustrated in Figure 2, the free energy profiles along RC[4+2] indicate a concerted mechanism of this enzyme, which is consistent with the previous calculation.14 Thermodynamically, different models share a similar reaction barrier with about 20 kcal/mol by either M06-2X or B3LYP methods (in Figure 2(a) and Figure S1), which is in good agreement with the experimental result (18.6 kcal/mol).12 However, the relative stability of the product is quite different for these two methods. Only about 10 kcal/mol energy is released with B3LYP, while more than 20 kcal/mol by M06-2X. This large discrepancy (10.0 vs. 20.4 kcal/mol) in the energy released is not surprising since Matsuda has pointed out that the B3LYP functional greatly underestimates the energy release when a π-bond is converted to a σ-bond.28 Thus, M06-2X is more suitable for modeling this system than B3LYP. According to the [4+2] transition structures (TS) in Figure 2(b and c), the distance changes of C7-C11 and C4-C12 (in Figure S2) and the electrostatic potential (ESP) charge in Table S1, this reaction is an asynchronous reaction with a single TS, although the two TS structures have different distances of C2-C14 and C4-C12. Some conformational changes in the ligands occur along the reaction pathway (see details in Figure S3 and S4). Based on these two QM/MM TS structures, the gas phase TS optimizations and IRC (Intrinsic Reaction Coordinates) calculations have been used to calculate the reaction pathway with M06-2X(D3)/6-31G*. Surprisingly, the two similar TS structures result in two different reaction pathways. In Figure S5, the structure based on M06-2X/MM model will go through the [6+4] pathway and generate the [6+4] product (the distance of C2-C14 decreases during the calculation), whereas the TS structure based on B3LYP/MM model will go directly through the [4+2] pathway to obtain the usual TS DA product. In addition, the similarity between Houk’s TS7 and our TS in the M06-2X/MM models (Figure S6) is evidence that the TS in the B3LYP/MM model prefers to go through the [4+2] pathway while the TS in M06-2X/MM tends to follow the [6+4] pathway. Similarly, in QM/MM MD simulations, the [6+4] pathway is dominant in Model A with the M06-2X method (Figure 3(a)). After unrestrained sampling around the [4+2] TS, about 74.2% sampling structures are the [6+4] product which matches the results of the gas phase IRC. However, when Tyr23 is included in the QM area (Model B, Figure 3(b)), the probability of the [6+4] pathway decreases to 53.7%. Even more dramatic is the decrease of this [6+4] pathway probability to 5.6% when Thr196 in included in the QM area (Model C,

ACS Paragon Plus Environment

Page 2 of 7

Page 3 of 7 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

ACS Catalysis

FIGURE 3. The free energy maps and the scatters of the sampling around the [4+2] TS along the distances of C7-C11, C4-C12, and C2-C14 in M06-2X models (Model A (a), Model B (b), and Model C (c)) and (d) the reaction diagram of Model C. Here, the energy of the [4+2] TS is normalized to 20.1 kcal/mol, which is the same as the barrier in Figure 2(a). The increment of the two adjacent contours in all the free energy maps is 0.5 kcal/mol. The key states are highlighted, and the reaction paths are drawn in white dashed lines. In Model A, the gradient around the TS is about 2 kcal/(mol•Å) (0.14 kcal/mol / 0.07 Å) while it is about 5.9 kcal/(mol•Å) (~1 kcal/mol / 0.17 Å) in Model B. It increases to about 12.5 kcal/(mol•Å) (~2.5 kcal/mol / 0.2 Å) in Model C. Also, one transition region exists that connects the [6+4] product and the [4+2] pathway. For the scatter figures, the [6+4] product is defined based on the distance of C2-C14 (