A Biomimetic System for Studying Salicylate Dioxygenase

Abstract

We report the characterization of [Fe(T1Et4iPrIP)(sal)] (2) (T1Et4iPrIP = tris(1-ethyl-4-isopropyl-imidazolyl)phosphine; sal²⁻ = salicylate dianion), which serves as a model for substrate-bound salicylate dioxygenase (SDO). Complex 2 crystallizes in the monoclinic space group P2₁/n with a = 10.7853(12) Å, b = 16.5060(19) Å, c = 21.217(2) Å, β = 94.489(2)°, and V = 3765.5(7) ų. The structure consists of Fe^II bonded in distorted square pyramidal geometry (τ = 0.32) with two salicylate oxygens and two T1Et4iPrIP nitrogens serving as the base and the apical position occupied by the other ligand nitrogen. [Fe(T1Et4iPrIP)(OTf)₂] (1), the precursor for 2, catalyzes the cleavage of 1,4-dihydroxy-2-naphthoate in the presence of O₂. Complex 1 is also capable of cleaving the salicylate aromatic ring in the presence of H₂O₂. The progression of this reaction toward product formation involves an Fe^III–phenoxide species.

Introduction

The deposition of aromatic waste products into the environment by either intentional or accidental means is a threat to human health and may be responsible for a number of acute or chronic diseases. Sources of these hydrocarbons include vehicle emissions and industrial exhaust while other sources include wild fires, burning of plastics and pesticides, incomplete burning of coal or other fuels, and oil spills (1,2). Bacteria have evolved to utilize these waste products as a carbon source. Activation of an aromatic substrate in aerobic microorganisms is usually affected by hydroxylation of the ring and subsequent dearomatization. The typical ring-fission substrates are catechol, protocatechuate, homoprotocatechuate, hydroquinone, gentisate, or homogentisate. These substrates are converted to either mono- or dicarboxylic acids, depending on the mode of cleavage (3). Hydroquinones, homogentisate, and gentisate undergo fission of the aromatic ring adjacent to the hydroxyl group by dioxygenases. For hydroquinones-bearing carboxyl groups (e.g., homogentisate and gentisate), cleavage usually occurs at the 1,2-position between the carboxyl and hydroxyl group (4,5). Intradiol Fe^III dioxygenases are usually present in these ring-fission dioxygenases while Fe^II is present for para-substituted diols-cleaving enzymes and extradiol-cleaving enzymes (3, 6–8). A third type of dioxygenase that cleaves aromatic compounds (9–11) has been assigned to a new class (Class III) (12). These enzymes catalyze the dioxygenation of gentisate, homogentisate, 3-hydroxy-anthranilate, 1-hydroxy-2-naphthoate, 2-aminophenol, salicylate, 5-aminosalicylate, 5-chlorosalicylate, and 5-nitrosalicylate (13). Gentisate, salicylate, 5-aminosalicylate, and 5-chlorosalicylate are converted via the Class III dioxygenases while the rest are oxidized by Class I and II enzymes. The Class III ring-cleaving dioxygenases are part of the cupin superfamily (14,15). Enzymes containing a 3His metal binding site are gentisate 1,2-dioxygenase (GDO), salicylate 1,2-dioxygenase (SDO), 3-hydroxyanthranilate 3,4-dioxygenases, 1-hydroxy-2-naphthoate dioxygenase (HNDO), and Type II hydroquinone 1,2-dioxygenase.

Gentisate 1,2-dioxygenase converts gentisate to maleylpyruvate (Scheme 1) (16,17). GDO contains Fe^II and has a high specificity for gentisate. Alternately, SDO (Scheme 1) catalyzes the dioxygenation of many monohydroxylated aromatic substrates, which include 3-, 4-, and 5-chlorosalicylate, and 1-hydroxy-2-naphthoate (9,18). Other substrates include: 5-fluorosalicylate; 3,5-dichlorosalicylate; 3-, 4-, 5-bromosalicylate; 3-, 4-, and 5-methylsalicylate (9); 1-hydroxy-2-naphthoate; 3- and 5-aminosalicylate; and 3- and 4-hydroxysalicylate (9,18). Figure 1 shows the structures of GDO (PDB 3BU7) (19) and SDO with bound gentisate (PDB 3NL1), bound salicylate (PDB 3NJZ), and bound hydroxy naphthoate (PDB 3NKT) (12). No substrate-bound adduct has been structurally characterized for GDO; however, hydrogen bonding interactions between the active site residues (His162, Tyr190, Gln108, and Asp175) and two iron-bound water ligands are clearly seen.

Scheme 1.

Class III dioxygenase enzymes: GDO, SDO, and HNDO.

Figure 1.

X-ray structure for the active site of GDO (PDB 3BU7) and SDO with bound gentisate (PDB 3NL1), salicylate (PDB 3NJZ), and naphthoate (PDB 3NKT).

Substrate-bound SDO active site structures also exhibit extensive hydrogen bonding interactions between active site residues and substrates. In all cases, His162 is within bonding distance to the carbonyl oxygen of the substrate and Arg83 also provides an H-donor to the bound substrate oxygen of the carboxylate group. Arg127 also interacts in the salicylate and gentisate structures. In the case of SDO, it has been proposed that His162 acts as proton donor to the evolving peroxidate intermediate (20). It is quite probable that successful synthetic models for these active sites will require outer-sphere H-bonding interactions. Interestingly, Eppinger et al. reported that a GDO mutant (Ala112Gly) allowed GDO to function as SDO in its substrate preference (21). This suggests that in GDO, salicylate is unable to bind productively (with both the deprotonated carboxyl and hydroxyl groups) due to the Ala112 group (located distally to the active site); therefore, no turnover is observed.

The catalytic mechanism for GDO is believed to first involve coordination of the substrate followed by dioxygen binding. The O₂ attacks at carbon-1 of the bound gentisate to form an alkylperoxo intermediate (Scheme 2). Heterolytic cleavage of the O–O bond and insertion of one oxygen atom into the ring C1–C2 bond (Criegee rearrangement) would be enhanced by conversion of the hydroxyl group to a ketone at C5 and would result in formation of a cyclic lactone. Hydrolysis by coordinated hydroxide could occur, resulting in formation of the final product (19,22,23).

Scheme 2.

Proposed mechanism for dioxygenation of gentisate by GDO (23).

For SDO, gentisate and 5-aminosalicylate are converted to products at a much faster rate (>100-fold for gentisate) to products compared to salicylate. The rate constant for wild-type SDO (0.9 s⁻¹) (24) corresponds to ~68 kJ/mol. The key factor responsible for the enhanced catalytic rate could be that gentisate possesses a hydroxyl group at the 5-position. The presence of an amino group for 5-aminosalicylate could also be key since deprotonation could occur here to stabilize the forming intermediate. The hydrophobic active site pocket could be important in transferring electron density to the ring. For gentisate-bound SDO (Figure 1, 3NL1), Trp104 is typically a tyrosine in many GDOs. The presence of extensive H-bonding interactions in SDO between the substrate and binding pocket could be an important factor in activation of the substrate; however, this point must be investigated with model chemistry. The mechanism believed operative in GDO may not account for the reactivity toward salicylate since no 5-OH group is present. The mechanism for salicylate 1,2-dioxygenase has been investigated by a quantum mechanical and molecular mechanical (QM/MM) study (25). The study reveals that an active site His (His162) plays a role as an acid/base catalyst supplying a proton to the substrate during catalysis. Roy et al. also conducted QM/MM calculations on the mechanism of SDO and found that it is possible for a GDO-like mechanism to occur (Scheme 2) without the need for the 5-OH group (26).

To study and provide support for mechanistic proposals we have utilized tris(1-ethyl-4-isopropyl-imidazolyl)phosphine (T1Et4iPrIP, Scheme 3) as our 3His metal binding site and have synthesized [Fe(T1Et4iPrIP)(OTf)₂] (1) to serve as a model for the iron site in SDO.

Scheme 3.

Structures of T1Et4iPrIP and [Fe(T1Et4iPrIP)(OTf)₂] (1).

Results and Discussion

Complex 1, first reported by our group (27), was reported to be a good model for iron–3His sites from a structural and electronic standpoint. Complex 1 was shown to reversibly bind NO (similar to other iron–3His enzyme active sites such as GDO and other mononuclear iron sites) (23, 28). We studied the LFe + NO ↔ LFeNO equilibrium process spectroscopically by titrating a solution of NO (7 mM) in THF at 25 °C with 1. The study yielded the equilibrium constant (K_eq = 470 M^–1) (29) and is in line with similar studies (30).

We sought to synthesize a model for the enzyme-substrate adduct to investigate and understand the mechanism for SDO. To do this we reacted 1 with one equiv bis(tetraethylammonium) salicylate in dichloromethane. After diffusion in an ether chamber and crystallization at −20 °C, light-green crystals of [Fe(T1Et4iPrIP)(sal)] (2) deposited and were isolated in moderate yield. The crystallographic parameters (Table 1) and bond distances and angles (Table 2) for 2 are given. The structure is shown in Figure 2. The structure reveals Fe^II in distorted square pyramidal geometry (τ = 0.32) (31). The base comprises oxygens from phenoxy and carboxy groups of the salicylate and two nitrogens from the imidazole groups of the ligand. The apical position is occupied by a third nitrogen from T1Et4iPrIP. The average Fe–N bond distance (2.190 Å) is slightly longer compared to 1 (2.140 Å) (27) while the Fe–O distances (1.9519(12) and 2.0456(11) Å) are slightly shorter compared to another iron salicylate complex (1.958(1) and 2.060(1) Å) (32).

Table 1.

Crystallographic Data for [Fe(T1Et4iPrIP)(sal)]·(Et₂O) (2·Et₂O)

Complex	2 · Et2O
chemical formula	C35H53FeN6O4P
formula weight (g/mol)	708.65
temperature /K	100.0(5)
λ/Å	0.71073
Crystal system	Monoclinic
space group	P21/n
a /Å	10.7853(12)
b /Å	16.5060(19)
c /Å	21.217(2)
α /°	90
β /°	94.489(2)
γ /°	90
V /Å3	3765.5(7)
Z	4
μ /mm−1	0.487
Dcalcd /Mg/m3	1.250
R1 ((I > 2σ(I))	0.0485
wR2 ((I > 2σ(I))	0.1017

R1 = Σ||F_o|-|F_c||/Σ|F_o|, wR2 = [Σ[w(F_o²-F_c²)²]/Σ[w(F_o²)²]]^1/2

w = q/[σ²(F_o²)+(a·P)²+b·P+d+e·sin(θ)]

Table 2.

Selected Bond Lengths (Å) and Angles (°) for [Fe(T1Et4iPrIP)(sal)]·(Et₂O) (2·Et₂O)

Bond Lengths (Å)
Fe(1)–O(2)	1.9519(12)	P(1)–C(1)	1.8157(16)
Fe(1)–O(1)	2.0456(11)	P(1)–C(2)	1.8219(15)
Fe(1)–N(2)	2.1598(13)	O(1)–C(25)	1.2805(18)
Fe(1)–N(1)	2.1786(13)	O(2)–C(31)	1.3147(19)
Fe(1)–N(3)	2.2320(12)	O(3)–C(25)	1.2457(18)
P(1)–C(3)	1.8142(15)
Bond Angles (°)
O(2)–Fe(1)–O(1)	87.07(5)	N(2)–Fe(1)–N(3)	84.16(5)
O(2)–Fe(1)–N(2)	151.58(5)	N(1)–Fe(1)–N(3)	89.82(5)
O(1)–Fe(1)–N(2)	93.40(5)	C(3)–P(1)–C(1)	97.93(7)
O(2)–Fe(1)–N(1)	117.36(5)	C(3)–P(1)–C(2)	99.21(7)
O(1)–Fe(1)–N(1)	99.34(5)	C(1)–P(1)–C(2)	97.49(7)
N(2)–Fe(1)–N(1)	90.64(5)	C(25)–O(1)–Fe(1)	130.69(10)
O(2)–Fe(1)–N(3)	90.75(5)	C(31)–O(2)–Fe(1)	128.11(10)
O(1)–Fe(1)–N(3)	170.56(5)

Figure 2.

Oak Ridge thermal ellipsoid plot of 2 (50% probability). H-atoms have been removed for clarity.

Reactivity Studies

A model for SDO has previously been reported (32). [Fe^II(Tp^Ph2)]⁺ (Tp^Ph2 = hydrotris(3,5-diphenylpyrazole-1-yl)borate) has been shown to catalyze the aromatic ring scission of 1,4-dihydroxy-2-naphthoate using O₂ as the oxidant (Scheme 4) (33). Incorporation of both oxygen atoms into the substrate was demonstrated. It was noted that the presence of 4-hydroxy group was critical for the reaction and that substitution by −NH₂ or −OCH₃ did not lead to ring scission.

Scheme 4.

Dioxygenation of 1,4-dihydroxy-2-naphthoate (33).

The model complex was not capable of degrading salicylate or 1-hydroxy-2-naphthoate as is the case with SDO and HNDO. In the reaction, the absorption band at 340 nm disappears and two new charge-transfer bands appear at 720 nm and 920 nm, following a pseudo-first order rate. The proposed mechanism (Scheme 5) (33) involves initial outer-sphere oxidation of Fe^II to Fe^III by oxygen followed by an internal redox reaction where iron is reduced back to Fe^II creating a radical on the ligand. Oxygen then reacts with the iron center affording an Fe^III alkyl peroxide moiety. The 4-hydroxy group then directs heterolytic cleavage of the O–O bond via a Criegee type rearrangement to generate an anhydride intermediate. The resulting coordinated hydroxide participates in hydrolysis of the anhydride yielding the ring-opened product. [Fe^II(Tp^Ph2)]⁺ was also shown to be unreactive toward 4-methylsalicylate and 5-aminosalicylate (33). Using complex 1, we were also able to catalyze the oxidative scission of 1,4-dihydroxy-2-naphthoate using O₂ as the oxidant in CH₂Cl₂. Next we focused our attention on examining whether 1 is capable of catalyzing the ring scission of salicylate in the presence of O₂. NMR studies and gas chromatography–mass spectrometry (GCMS) analysis indicated that although the solution quickly changes in color from yellow to brownish-red upon exposure to oxygen, the salicylate was unaffected in these reactions. We then pursued using H₂O₂ as the oxidant. The addition of 20 equiv H₂O₂ to 1 and bis(tetraethylammonium) salicylate in methanol or CH₂Cl₂ at 25 °C resulted in the immediate formation of a deep reddish-purple solution (Figure 3). After 10 min the reaction became reddish-brown. Analysis of the reaction mixture by acidification and extraction with CH₂Cl₂ indicated that the aromatic ring had been cleaved (Figure 4).

Scheme 5.

Proposed mechanism for the dioxygenation of 1,4-dihydroxy-2-naphthoate (33).

Figure 3.

Initial UV–vis spectrum of 1 and bis(tetraethylammonium) salicylate at a ratio of 1:1 in methanol (−), spectrum in methanol after addition of 20 equiv H₂O₂ (⋯), and spectrum after 10 min (- - -).

Figure 4.

¹H NMR spectra (25 °C) of extracted reaction mixtures in CDCl₃. Fe(OTf)₂·2CH₃CN (control, top) and 1 reacted with 1 equiv bis(tetraethylammonium) salicylate and 20 equiv H₂O₂ in methanol. After 14 h, reactions were acidified with 1 N HCl and extracted with CH₂Cl₂. The CH₂Cl₂ was removed in vacuo and the residues were redissolved in CDCl₃ for analysis.

A control experiment was run by replacing 1 with Fe(OTf)₂·2MeCN. This reaction showed no cleavage of the salicylate aromatic ring (Figure 4). The intermediate spectrum in Figure 3 is anticipated to be related to an Fe^III–phenoxide band arising from oxidation of the Fe by H₂O₂ (34,35). The disappearance of this band is expected coincide with the cleavage of the aromatic ring.

Summary and Conclusions

In this work we reported the synthesis and characterization of 2, a structural model for substrate-bound SDO. The precursor complex 1, was found to be active in the cleavage of 1,4-dihydroxy-2-naphthoate in the presence of O₂ and to affect the ring scission of salicylate in the presence of H₂O₂. An intermediate species was observed during the oxidation of salicylate in the presence of H₂O₂, and is believed to be an Fe^III–phenoxide species with an intermediate in the ring scission reaction. Low temperature studies are planned to determine if any intermediates are present prior to the intermediate observed at 25 °C. Density functional theory studies will also help generate a proposed mechanism for the reaction.

Experimental

Fe(OTf)₂·2CH₃CN was prepared according to literature procedure (36). 1 was also prepared according to a published method (37). Anhydrous THF, pentane, dichloromethane, and ether were obtained using a solvent purification system (Innovative Technologies, Inc., Amesbury, MA). Methanol (CH₃OH) was dried and distilled according to published procedures (38). Elemental analysis was performed on pulverized crystalline samples that were placed under vacuum and sealed in a glass ampule prior to submission (Atlantic Microlabs, Inc., Norcross, GA). H₂O₂ (80%, v/v) was prepared by adding anhydrous MgSO₄ to 30% H₂O₂ (Alfa Aesar, Haverhill, MA) and storing this mixture at 5 °C. The purity was determined by density measurements. Unless otherwise noted, all the reaction chemicals and tetramethylsilane (TMS) were AR grade (Sigma-Aldrich, St. Louis, MO or Alfa Aesar) and used as received.

Synthesis

Bis(tetraethylammonium) Salicylate

To a stirred solution of salicylic acid (2.76 g, 0.02 mol) in CH₃OH (10 mL) was added aqueous Et₄N(OH) (wt.% = 20%, 29.45 g, 0.04 mol). The solution was stirred for 30 min, evaporated under high vacuum overnight to remove the solvent and affording a pale-white solid as the product. Yield: 7.54 g (95%). ¹H NMR (400 MHz, CDCl₃, δ (ppm)) = 7.92 (d, J = 2.0 Hz, 1H), 7.18–7.22 (m, 1H), 6.78 (d, J = 4.0 Hz, 1H), 6.69–6.73 (m, 1H), 3.29–3.34 (m, 16H), 1.27–1.30 (m, 24H).

[Fe(T1Et4iPrIP)(sal)] (2)

To a stirred solution of [Fe(T1Et4iPrIP)(OTf)₂] (50 mg, 0.063 mmol) in CH₃CN (2 mL) was added a prepared solution of bis(tetraethylammonium) salicylate (25 mg, 0.063 mmol) with 2.5 equiv trimethylamine (22 μL, 0.158 mmol) in CH₃CN (2 mL) under dry nitrogen. The solution was stirred for 30 min then filtered. Ether was added to the filtrate until it became slightly cloudy. The resulting mixture was filtered and placed at −30 °C. After a few days, light-green crystals were deposited as the product. Yield: 25 mg (75%). IR (KBr pellet, cm^–1): ṽ = 3124 (m), 3072(m), 2960 (s), 2932 (sh), 2869 (m), 2575(w), 1601 (vs), 1571(vs), 1465(vs), 1447 (vs), 1424 (s), 1383 (s), 1352 (vs), 1323 (sh), 1259 (s), 1201(m), 1195 (m), 1147 (m), 1132 (m), 1107 (w), 1084(w), 1031(w), 1101(w), 965(w), 923(w), 879(m), 818(m), 757(m), 727(w), 707(w), 655(w), 610 (w), 569 (w), 541 (s), 490 (w), 446 (w), 407 (w). UV–vis [λ_max, nm (ε, M^–1 cm^–1) in CH₃OH] 409 (598), 262 (130500). Anal. Calcd for C₃₁H₄₃N₆O₃PFe: C, 58.68; H, 6.83; N, 13.24. Found: C, 58.03; H, 6.75; N, 12.86. Magnetic measurements, μ_eff (CDCl₃ solution, 298 K): 5.20 μ_B.

Physical Measurements

A Cary 50 ultraviolet–visible (UV–vis) spectrophotometer was used to collect optical spectra. FT-IR spectra were acquired on a Varian 3100 Excalibur Series and a Bruker ATR Alpha P spectrometer. NMR spectra were monitored at 25 °C on a Brüker Avance II 400 MHz instrument and sample peaks were referenced to TMS (CDCl₃). GCMS experiments were performed on an Hewlett-Packard 6890/5973 GCMS. Solution magnetic susceptibility measurements on 1 at 298 K were obtained using the Evans NMR method (39) with CDCl₃ containing 5% CH₃CN as the reference. Mass susceptibility, χ_g, was calculated from the following equation:

$${\chi }_g=\frac{-3\Delta f}{4\pi fm}+{\chi }_{\text{o}}\left[1+\frac{\left(d_{\text{o}}-d_s\right)}{m}\right]$$ (1)

Where Δf is the frequency shift in Hz of the reference compound, f is the fixed probe frequency of the spectrometer, χ_o is the mass susceptibility in cm³ g⁻¹ of the solvent, m is the mass in g of the complex in 1 mL of solution, and d_o and d_s are the densities of the solvent and solution, respectively.

X-ray Crystallography

Green blocks of 2 were obtained by allowing a saturated CH₃CN solution of 2 with ether to cool at −30 °C. A crystal of 2 was placed onto the tips of a thin glass optical fiber and mounted on a Bruker SMART APEX II CCD platform diffractometer for data collection at 100.0(5) K (40). The full data collection was carried out using MoKα radiation (graphite monochromator). The intensity data were corrected for absorption (41). Final cell constants were calculated from the xyz centroids after integration (42). The structures were solved using SIR2011 (43) and refined using SHELXL-2014/7 (44). A direct-methods solution was calculated and provided most nonhydrogen atoms from the E-map. Full-matrix least squares/difference Fourier cycles were performed which located the remaining nonhydrogen atoms. All nonhydrogen atoms were refined with anisotropic displacement parameters. All other hydrogen atoms were placed in ideal positions and refined as riding atoms with relative isotropic displacement parameters. The asymmetric unit for each compound contains one iron complex and one cocrystallized diethyl ether molecule, both in general positions. Refer to Table 1 for additional crystal and refinement information. Selected bond lengths and angles are listed in Table 2.