What Drives Radical Halogenation versus Hydroxylation in Mononuclear Nonheme Iron Complexes? A Combined Experimental and Computational Study

Abstract

Nonheme iron halogenases are unique enzymes in nature that selectively activate an aliphatic C-H bond of a substrate to convert it into C-X (X = Cl/Br, but not F/I). It is proposed that they generate an Fe^III(OH)(X) intermediate in their catalytic cycle. The analogous Fe^III(OH) intermediate in nonheme iron hydroxylases transfers OH^• to give alcohol product, whereas the halogenases transfer X^• to the carbon radical substrate. There remains significant debate regarding what factors control their remarkable selectivity of the halogenases. The reactivity of the complexes Fe^III(BNPA^Ph2O)(OH)(X) (X = Cl, Br) with a secondary carbon radical (R^•) is described. It is found that X^• transfer occurs with a secondary carbon radical, as opposed to OH^• transfer with tertiary radicals. Comprehensive computational studies involving density functional theory were carried out to examine the possible origins of this selectivity. The calculations reproduce the experimental findings, which indicate that halogen transfer is not observed for the tertiary radicals because of a nonproductive equilibrium that results from the endergonic nature of these reactions, despite a potentially lower reaction barrier for the halogenation pathway. In contrast, halogen transfer is favored for secondary carbon radicals, for which the halogenated product complex is thermodynamically more stable than the reactant complex. These results are rationalized by considering the relative strengths of the C-X bonds that are formed for tertiary versus secondary carbon centers. The computational analysis also shows that the reaction barrier for halogen transfer is significantly dependent on secondary coordination sphere effects, including steric and H-bonding interactions.

Scheme 1. General Reaction Catalyzed by Nonheme Iron Halogenases, with αKG = α-Ketoglutarate and R–H as the Substrate

Scheme 2. Iron Complexes, Carbon Radicals, and Reactions Studied in This Work

Figure 1

Top: ¹H NMR spectra obtained in CDCl₃ of isolated C₆H₅CH(Cl)CH₃ from the reaction of 1_Cl with C₆H₅CH(CH₃)N=N(CH₃)CHC₆H₅ (a), from the reaction of 2_Cl with C₆H₅CH(CH₃)N=N(CH₃)CHC₆H₅ (b), and authentic C₆H₅CH(Cl)CH₃ (c). Residual solvent signals are marked with a red asterisk (*). Bottom: Reactivity of 1_Cl with tertiary and secondary carbon radicals.

Scheme 3. Reactions of 1_OTf with Tertiary and Secondary Carbon Radicals

Figure 2

UB3LYP/BS1-optimized geometries of the isolated reactants ⁶1_Cl/⁶1_Br as obtained in Gaussian with bond lengths given in Å. Relevant molecular valence orbitals are shown on the right.

Figure 3

Potential energy landscape for halogen versus hydroxyl transfer in complexes 1_Cl/1_Br/1_F as obtained at UB3LYP/BS2//UB3LYP/BS1 level of theory with the solvent and zero-point corrections included. Energies relative to the reactant complexes are in kcal mol^–1, while transition-state structures give bond lengths in Å, angles in degrees, and the imaginary frequency in cm^–1.

Scheme 4. Halogenated and Hydroxylated Product Complexes Showing Feasible Reverse Reactions for Br Transfer and Not for OH Transfer

Figure 4

Optimized geometries of TS_Cl2,2Cl and TS_Cl3,2Cl as obtained in Gaussian. Energies (ΔE + ZPE) are relative to the reactant complexes in kcal mol^–1, while transition-state structures give bond lengths in Å, angles in degrees, and the imaginary frequency in cm^–1.

Scheme 5. Reactions of 2_Cl with Tertiary and Secondary Carbon Radicals

Figure 5

Potential energy landscape for halogen versus hydroxyl transfer after isomerization of complex 1_Cl into 2_OH as obtained at UB3LYP/BS2//UB3LYP/BS1 level of theory with the solvent and zero-point corrections included. Energies are in kcal mol^–1, while structures give bond lengths in Å, angles in degrees, and the imaginary frequency in cm^–1. The constraint geometry scan for isomerization from 1_Cl to 2_OH is shown in the inset.

Figure 6

OH and Cl rebound barriers from ⁵Re_1Cl and truncated models with steric or hydrogen-bonding interactions (or both) removed. Energies obtained at UB3LYP/BS2//UB3LYP/BS1 with zero-point and solvent corrections included. Data in kcal mol^–1 with respect to the reactant complexes. Also shown are overlays of the Cl transfer and OH transfer transition states with the following color coding: full system (khaki), without sterics (pink), without H-bond (cyan), and without sterics/H-bond (light green).

Figure 7

Potential energy landscape for halogen versus hydroxyl transfer from complexes Re_1Cl/Re_1Br to the secondary radicals (C₆H₅)₂CH^• (PP) and (C₆H₅)(CH₃)CH^• (PM) as calculated at UB3LYP/BS2//UB3LYP/BS1 level of theory with the solvent and zero-point corrections included. Energies are in kcal mol^–1. Values in parentheses are for the reaction from Re_1Br.

Figure 8

Valence bond diagram for the prediction of OH and Cl-transfer barriers in ⁵Re_1Cl. Dots are electrons and a bar between two dots is a chemical bond occupied with two electrons.