Substrate activation for O₂ reactions by oxidized metal centers in biology

Abstract

The uncatalyzed reactions of O₂ (S = 1) with organic substrates (S = 0) are thermodynamically favorable but kinetically slow because they are spin-forbidden and the one-electron reduction potential of O₂ is unfavorable. In nature, many of these important O₂ reactions are catalyzed by metalloenzymes. In the case of mononuclear non-heme iron enzymes, either Fe^II or Fe^III can play the catalytic role in these spin-forbidden reactions. Whereas the ferrous enzymes activate O₂ directly for reaction, the ferric enzymes activate the substrate for O₂ attack. The enzyme–substrate complex of the ferric intradiol dioxygenases exhibits a low-energy catecholate to Fe^III charge transfer transition that provides a mechanism by which both the Fe center and the catecholic substrate are activated for the reaction with O₂. In this Perspective, we evaluate how the coupling between this experimentally observed charge transfer and the change in geometry and ligand field of the oxidized metal center along the reaction coordinate can overcome the spin-forbidden nature of the O₂ reaction.

Mononuclear non-heme iron enzymes perform a wide range of important biological functions involving O₂ in parallel to those of heme enzymes (1–5). Although the uncatalyzed reactions of O₂ (S = 1) with organic substrates (S = 0) are thermodynamically favorable, the reactions are kinetically slow because they are spin-forbidden and the one-electron reduction potential of O₂ is unfavorable. Most of these enzymes catalyze the O₂ reaction by using a high-spin Fe^II to activate O₂ through a redox process that also involves the substrate or an additional cofactor providing the required number of electrons. These ferrous O₂-activating enzymes include extradiol dioxygenases, pterin-dependent hydroxylases, α-ketoglutarate-dependent enzymes, and Rieske dioxygenases. However, some mononuclear non-heme iron enzymes possess an oxidized iron active site and use the Fe^III to activate substrate for the O₂ reaction. These ferric substrate-activating enzymes presently include the lipoxygenases (LO) and intradiol dioxygenases

LO employ a relatively straightforward substrate activation mechanism. They catalyze the regio- and stereospecific hydroperoxidation of 1,4-Z,Z-pentadiene-containing polyunsaturated carboxylic acids (6–10). The resting site of LO consists of a ferrous center coordinated to three histidine ligands, a monodentate C-terminal carboxylate of isoleucine, a carbonyl O of asparagine, and a water (11–13). Resting Fe^IILO is oxidized by the hydroperoxide product to Fe^IIILO, and the water ligand is deprotonated to hydroxide. This Fe^III(OH) species catalyzes the first step of the reaction, namely, hydrogen-atom abstraction from the substrate to form a substrate radical, which is accompanied by the reduction of the Fe^III(OH) to Fe^II(H₂O). The substrate bound near this Fe^II site is activated to a radical species (S = 1/2), and its reaction with O₂ (S = 1) is spin-allowed. This step forms a peroxy radical, and then the hydroperoxide product as Fe^II(H₂O) is oxidized to Fe^III(OH). This mechanism is summarized in Fig. 1 (14, 15). The driving force for the hydrogen-atom abstraction is related to the strength of the OH bond produced in the abstraction intermediate, which is directly related to the reduction potential of the Fe^III/Fe^II couple (E°) and the pK_a of the abstracted proton in the ferrous species. A high reduction potential and a high pK_a of the coordinated water in the reduced species increases the hydrogen-atom abstraction driving force and, hence, the catalytic rate (15–17).

Fig. 1.

Mechanism for LO.

The substrate activation mechanism for intradiol dioxygenases is more complex. These enzymes catalyze the cleavage of molecular oxygen accompanied by insertion of both oxygen atoms between the vicinal hydroxyl groups of the catecholic substrate, resulting in ring cleavage to yield muconic acid derivatives (3–5, 18, 19). The resting state of intradiol dioxygenases contains a high-spin ferric center in a distorted trigonal bipyramidal geometry, with Tyr and His as the axial ligands and Tyr, His, and a hydroxide ligand defining the equatorial plane (20–22). Upon anaerobic substrate binding, the active site shifts to a square pyramidal geometry in which the axial Tyr and equatorial OH⁻ are displaced by the substrate, which binds bidentate in its doubly deprotonated form (23, 24). The open coordination position is trans to the equatorial His, and the substrate binds asymmetric to the Fe^III center with the longer bond trans to the equatorial Tyr. EPR and Mössbauer studies have shown that that the iron center remains oxidized and high-spin upon substrate binding (25, 26). Thus direct interaction between the triplet oxygen and the singlet catecholate substrate is still spin-forbidden. Based on different electronic descriptions of the enzyme–substrate (ES) complex, various mechanisms have been proposed for the substrate activation step for the initial O₂ attack in the enzymatic reaction (bracket in Fig. 2): (i) Fe^II-semiquinone character with O₂ attacking the iron site; (ii) Fe^II-semiquinone character with O₂ attacking the substrate through radical coupling, and (iii) Fe^III-catecholate with strong ketonized-character promoted by lengthening the Fe–O^catechol bond due to the trans effect of the equatorial Tyr, with O₂ directly attacking the substrate through an electrophilic activation (27–33). The net result of the O₂ attack is generally thought to be the formation of a peroxy-adduct bridging between the iron site and one of the hydroxylated carbons of the now quinone that requires a spin-forbidden two-electron transfer of the electron pair of the substrate to the triplet O₂ molecule (23). It is important to determine the mechanism of substrate activation for O₂ reaction in the intradiol dioxygenases both for this class of enzymes and for others where similar mechanisms have been invoked as, for example, in cofactor biogenesis in copper-containing amine oxidases (34, 35).

Fig. 2.

Intradiol dioxygenases mechanism.

Because the differences among the substrate activation proposals of intradiol dioxygenases (Fig. 2) lie in the electronic structure description of the ES complex, it was essential to determine the nature of this ES complex experimentally, particularly its degree of radical character. In our previous spectroscopic studies on protocatechuate 3,4-dioxygenase (3,4-PCD) with protocatechuate (PCA) bound, we found that the ES complex is best described as a highly covalent Fe^III-catecholate complex with the covalency distributed unevenly among the four valence donor orbitals of the catecholate (36). Because the catecholate has not been fully oxidized to a semiquinone and the oxidation state of the iron remains at the Fe^III level (S = 5/2), the reaction with triplet oxygen is still formally spin-forbidden. The combined absorption and magnetic circular dichroism spectra of the ES complex shown in Fig. 3 reflect a very low energy β electron charge transfer transition (the high-spin Fe^III has five d electrons with α spin) from the PCA substrate to the Fe^III center involving the π_op-sym (out-of-plane, symmetric) highest occupied molecular orbital, which is the frontier orbital of the catecholate. This low-energy catecholate to Fe^III charge transfer (arrow in Fig. 3) provides a mechanism in which both the Fe center and the catecholic substrate are activated for the reaction with O₂ (36).

Fig. 3.

Absorption (ABS) and magnetic circular dichroism (MCD) spectra of the ES complex of 3,4-PCD.

In this Perspective, we will consider how this β charge transfer between PCA and the Fe^III center and the ligand field of the oxidized metal site in intradiol dioxygenases can overcome the spin-forbidden nature of the two-electron transfer of an electron pair from the highest occupied molecular orbit of the substrate to the triplet O₂ molecule. A reaction coordinate leading to the final intradiol-cleaved product also will be considered.

Nature of the ESO₂ Complex

A two-electron transfer between PCA and O₂ accompanies the formation of the peroxy intermediate bridging between the Fe center and one of the hydroxyl carbons on the catecholic substrate. Direct interaction between the triplet O₂ (S = 1) and the singlet catecholic substrate (S = 0) is formally spin-forbidden. However, EPR studies showed that the ES complex contains a high-spin Fe^III center with S_tot = 5/2 (25, 26). Hence, the accessible spin states for the O₂ (S = 1) reaction with the ES complex are 3/2, 5/2, and 7/2.

To obtain S_tot = 7/2, the O₂ molecule has to be in its ³Σ_g⁻ ground state and ferromagnetically coupled to the high-spin Fe^III center for the initial O₂ attack (Fig. 4A). This would allow the transfer of a β electron from the occupied PCA π_op-sym orbital to the O₂ π* orbital either directly or through the Fe center. However, the remaining half-occupied PCA π_op-sym orbital has an α spin. Transfer of this α electron to the remaining half-occupied O₂ π* orbital to complete the two-electron redox process is therefore restricted. Geometry optimization of ESO₂ with S_tot = 7/2 shows very weak O₂ interactions with both the Fe center and the substrate (Fe–O₂ = 2.42 Å, Fe–C3 = 2.78 Å, Fe–C4 = 2.80 Å). These limited bonding interactions with the ES complex suggests that the S_tot = 7/2 surface is not a favorable path for the initial O₂ reaction.

Fig. 4.

Schematics of electron transfer in the initial O₂ reaction of intradiol ctron transfer in the initial O₂ reaction of intradiol dioxygenases at various spin states: S_tot = 7/2 (A), S_tot = 5/2 (B), S_tot = 3/2 (C).

For S_tot = 5/2, the O₂ molecule has to be partially in an excited singlet state (mixed ¹Δ_g and ¹Σ_g⁺) at the start of the O₂ reaction (Fig. 4B). Because the ¹Δ_g and ¹Σ_g⁺ states are ≈1 eV and 1.6 eV higher in energy than the ³Σ_g⁻ ground state of O₂, the initial O₂ reaction with S_tot = 5/2 is energetically highly unfavorable.

To obtain S_tot = 3/2, the O₂ molecule is in its ³Σ_g⁻ ground state and antiferromagnetically coupled to the high-spin Fe^III center for the initial O₂ attack (Fig. 4C, d⁵ Fe has five α spins and triplet O₂ has two β spins). Spectroscopic data on the ES complex of 3,4-PCD (Fig. 3) (36) showed that the doubly occupied π_op-sym orbital of the catecholic substrate PCA is the highest occupied molecular orbital. The high-lying π_op-sym orbital of the substrate can readily donate into the O₂ π* orbital for an electrophilic attack by O₂. However, because O₂ is in its ³Σ_g⁻ state, only the α electron from the doubly occupied π_op-sym orbital can be donated directly to the O₂ π* orbital in the formation of a C^PCA–O bond. The second electron required to reduce O₂ to O₂²⁻ has to be donated either by the substrate π_op-sym orbital, which would require a forbidden spin-flip of the β electron or by an occupied d orbital (d_xz) of the Fe^III center, which has α spin and can interact with the remaining half-occupied π* orbital of O₂ due to their antiferromagnetic coupling. The strong covalent interaction between the PCA π_op-sym and d_xz orbitals in the β manifold, reflected by the low-energy ligand-to-metal charge transfer transition observed in the spectroscopic data (Fig. 3), results in significant transfer of the β electron from the occupied PCA π_op-sym orbital into the unoccupied β d_xz orbital. This transfer would compensate the Fe for the donation of an α electron to O₂. However, because the d_xz orbital is the lowest-energy orbital in the Fe d manifold, the transfer of a β electron from the PCA to Fe and an α electron from Fe to O₂ results in an excited intermediate spin state (S = 3/2) on the Fe center (Fig. 4C).

In the following sections, we consider the O₂ reaction coordinate leading to the bridging peroxo complex along the S_tot = 3/2 surface and explore the geometric changes involved in stabilizing the intermediate spin on the Fe^III center. We also will discuss this reaction coordinate in the context of the mechanisms for overcoming the spin-forbidden nature of the O₂ reaction in the intradiol dioxygenases.

Reaction Coordinate to Form the Peroxy ESO₂ Complex

O₂ can interact with the Fe center and the hydroxylated carbon of catecholic substrate either sequentially or in a concerted manner in the formation of the peroxy ESO₂ complex. Previous experimental and theoretical studies showed that the ES complex of 3,4-PCD has little Fe^II character (20, 25, 26, 36); thus, direct interaction of O₂ at the iron center as the first step of the reaction is unlikely. To determine whether the substrate is the initial site of O₂ attack, we performed density functional theory calculations using an experimentally calibrated methodology developed in ref. 36 on hypothetical Ga^III catecholate and semiquinone complexes. These complexes have a similar ligand environment to the enzyme active site. Because Ga^III is diamagnetic with the d¹⁰ manifold strongly stabilized relative to the valence orbitals of the catecholic substrate, the metal d orbitals cannot participate in an interaction with O₂, which allows evaluation of whether triplet oxygen can directly attack the substrate in its two limiting electronic descriptions (catecholate and semiquinone). Geometry optimizations of these Ga^III complexes with O₂ showed that triplet O₂ does not interact with Ga^III-catecholate. While a bridging superoxy complex is formed with Ga^III-semiquinone, this adduct is 13 kcal/mol higher in energy than the nonbonding species, indicating that the direct attack of O₂ on PCA in the ES complex of 3,4-PCD also is unlikely. This result is consistent with previous theoretical studies (32, 37) and suggests that O₂ does not attack the iron center or the catecholate of the ES complex in a sequential manner.

Binding O₂ to the ES complex of 3,4-PCD results in a S_tot = 3/2 complex with Fe–O^O2 = 2.11 Å, O–C3^PCA = 2.61 Å and O–C4^PCA = 2.63 Å (Fig. 5A). Formation of this O₂ complex is exothermic by 3.5 kcal/mol, and molecular orbital analysis shows that O₂ interacts with both Fe and PCA in this complex (Fig. 6). This orbital interaction suggests that O₂ attacks the iron and the catecholate in a concerted fashion. To form the bridging peroxo intermediate, the distal oxygen has to attack either of the hydroxylated carbons (C3 or C4). Reaction coordinates along the Fe–O^O2 bond and either of the O–C3^PCA and O–C4^PCA bonds for S_tot = 3/2 were calculated, and the 2D energy profile along the O–C4^PCA reaction coordinate with a double-ξ basis set (6–31G*/3–21G*) in vacuum is shown in Fig. 7. A transition state is located at Fe–O^O2 = 1.83 Å and O–C4^PCA = 1.92 Å, with an imaginary frequency of −300 cm⁻¹ (Fig. 5B). The vibrational motion of this mode corresponds to the distal O atom moving toward C4^PCA and the O3^PCA moving away from the Fe center. After the transition state, the peroxo-bridged complex is formed. The peroxy bridged quinone substrate now binds monodentate through O4^PCA to form a five-coordinate (5C) Fe center (Fig. 5C). More accurate energies of all of the optimized structures were calculated with a triple-ξ basis set (TZVP), including solvation effects (polarized continuum model with a dielectric constant ε = 4.0) and zero-point and thermal corrections and are reported in this Perspective unless otherwise specified. The transition state for the formation of the peroxo intermediate bridging to C4^PCA (Fig. 5B) is 12.2 kcal/mol higher in energy than the O₂ complex (Fig. 5A). The overall reaction is slightly exothermic by 1.9 kcal/mol. The barrier of the peroxo formation can be attributed to the geometry change along the reaction coordinate. As the reaction proceeds from the six-coordinate (6C) O₂ complex to the transition state, the Fe–O3^PCA bond lengthens from 2.05 Å to 2.61 Å while the Fe–O^O2 bond shortens from 2.11 Å to 1.83 Å. This leads to a 5C site at the Fe^III center. However, because the bond between the distal oxygen and the substrate has not yet formed (O–C4^PCA = 1.92 Å), this 5C structure is destabilized relative to the 6C O₂ complex. As the O–C4 distance shortens, the bond between the distal oxygen and the substrate forms and the energy of the final peroxo product is stabilized relative to the transition state. Intrinsic reaction coordinate calculations toward both reactant and product directions at the transition state result in structures that look similar to the 6C O₂ and the 5C peroxy complexes, respectively. This result confirms the presence of an accessible reaction pathway for the geometric change observed in the enzyme active site.

Fig. 5.

Geometry optimized structures of the O₂ complex (A), transition state (B), and peroxo intermediate (C) obtained in the reaction coordinate calculation along Fe–O^O2 and O–C4^PCA. Their corresponding axes are included with the structures.

Fig. 6.

Molecular orbital demonstrating the concerted interactions of Fe and PCA with O₂ in the S_tot = 3/2 O₂ complex.

Fig. 7.

O₂ reaction coordinate for 3,4-PCD. (A) Two-dimensional potential energy surface along R(O–C4^PCA) and R(Fe–O^O2) reaction coordinates for S_tot = 3/2 and S_tot = 5/2 (red). (B) The two crossing points (1 and 2) between S_tot = 3/2 and S_tot = 5/2. The energies are obtained with BP86 + 10% Hartree–Fock exchange and 6–31G*/3–21G* basis set in vacuum.

The energy profile along the O–C3^PCA reaction coordinate with S_tot = 3/2 (data not shown) also leads to the formation a 5C bridging peroxo intermediate, but the monodentate substrate binds through O3^PCA instead of O4^PCA. The formation of this peroxo intermediate is exothermic by 0.5 kcal/mol. Hence, in terms of the energetics for the formation of the peroxo intermediate, there is no preference in C3 versus C4 attack by the distal oxygen, but the asymmetry of the Fe–O^PCA bonds (Fe–O3^PCA = 2.47 Å, Fe–O4^PCA = 1.97 Å) observed in the crystal structure of the ES complex would facilitate the dissociation of the Fe–O3^PCA bond in the attack of O₂ at C4^PCA, favoring this reaction. In this and the next section, we focus on the coordinate for O₂ attack at C₄^PCA to understand the mechanism for overcoming the spin-forbiddeness of the O₂ reaction.

As discussed in the previous section, the initial O₂ attack with S_tot = 5/2 is unfavorable because the O₂ molecule has an excited singlet state character. However, the total energy of the S_tot = 5/2 system decreases dramatically as the interaction between the singlet O₂ molecule and the ES complex is increased. Upon formation of the O₂ complex, the total energy of the S_tot = 5/2 system decreases from being 38.3 kcal/mol higher than the S_tot = 3/2 complex to only 3.6 kcal/mol. The energy profile along the O–C4^PCA reaction coordinate of the S_tot = 5/2 O₂ complex (Fig. 7, red) shows that the energy barrier for the formation of the 5C peroxo intermediate is only 3.7 kcal/mol. This is significantly lower than the 12.2 kcal/mol observed in the S_tot = 3/2 reaction but the total energy of the final 5C peroxo species in the S_tot = 5/2 reaction is 4.1 kcal/mol less stable than the S_tot = 3/2 product. Thus, the S_tot = 5/2 surface intersects with the S_tot = 3/2 surface at two different positions (Fig. 7B); the first is near the beginning of the reaction coordinate (Fe–O^O2 = 1.95 Å, O–C4^PCA = 2.40 Å) and the second is after the S_tot = 3/2 transition state (Fe–O^O2 = 1.90 Å, O–C4^PCA = 1.80 Å). Although S_tot = 3/2 is the more favorable reaction path for the initial O₂ attack, intersystem crossing from S_tot = 3/2 to S_tot = 5/2 at the first intersection point of the two potential surfaces could, in principle, favor the overall energetics of the peroxy formation. However, differences in electronic as well as geometric structure between the two isoenergetic structures at the crossing point govern the feasibility of the intersystem crossing between these two spin surfaces. In the section below, we evaluate the viability of this intersystem crossing and present an alternative description of the mechanism for overcoming the spin-forbidden nature of the O₂ reaction.

Mechanism for Overcoming the Spin-Forbidden Reaction

For facile spin-crossover from the S_tot = 3/2 to S_tot = 5/2 surfaces, the two species are required to have similar geometries and energies. Moreover, the electronic structures must be able to spin-orbit couple. Spin-orbit coupling (SOC) is effectively a localized, single-center, one-electron operator and can be written as

The L₊S₋ + L₋S₊ operator in Eq. 1 performs a spin-flip and this process is accompanied by a change in the orbital due to the associated L₊/L₋ raising/lowering operator. Hence, the two orbitals of opposite spins involved in SOC have to be on the same center with different spatial components. SOC can also proceed through the L_zS_z operator between microstates having the same M_s for the two different S_tot states. In our case, this means coupling between the M_s = 3/2 microstates of the S_tot = 3/2 and 5/2 states. This mechanism of SOC is feasible only if the two microstates differ solely in the occupation of two orbitals with the same spin and these two orbitals can couple through L_z.

The two isoenergetic structures for S_tot = 5/2 and 3/2 at the second crossing point along the O–C4^PCA reaction coordinate have very different geometries; hence, intersystem crossing is not feasible. However, as shown in Fig. 8, the two isoenergetic structures for S_tot = 5/2 and 3/2 at the first crossing point have similar geometries. It only takes 3.1 kcal/mol to excite the S_tot = 3/2 electronic configuration in its optimized structure into a S_tot = 5/2 state. Hence, it is energetically feasible for intersystem crossing. Orbital analysis on the S_tot = 5/2 structure shows significant overlap between the doubly occupied PCA π_op-sym orbital and the O₂ molecule (m_s = 0) in which the two π* orbitals are occupied by two electrons of opposite spin (β in π_ip* and α in π_op*, with the in-plane (ip) and out-of-plane (op) orbitals defined by the Fe–O–O plane). For the S_tot = 3/2 structure, a covalent interaction between the PCA π_op-sym and O₂ π_ip* in the α manifold is observed while the O₂ π_op* orbital remains singly occupied by a β electron. For both spin states, the iron center remains high-spin ferric with strong bonding interactions with PCA. Hence, the major difference in the electronic structure between the two spin states at the crossing point lies in the spin of the electron residing in the singly occupied π_op*, with α for S_tot = 5/2 and β for S_tot = 3/2 (Fig. 8). Although both of these spin orbitals are localized on O₂, they have the same spatial component in their wave functions (π_op*). Hence, the S_tot = 3/2 surface cannot effectively intersystem cross to the S_tot = 5/2 surface through the SOC mechanism at this point of the potential energy surface as the orbital angular momentum operators associated with SOC in Eq. 1 require a change in orbital occupation. Thus, the O₂ reaction with the ES complex should remain on the S_tot = 3/2 surface from the initial attack at the Fe center through the formation of the bridging peroxo complex.

Fig. 8.

Optimized geometry and electronic structures of the ESO₂ complex at the crossing point (Fe–O = 1.95 Å, O–C4 = 2.4 Å) of the S_tot = 5/2 and S_tot = 3/2 O-C4^PCA reaction coordinates: S_tot = 5/2 (A) and S_tot = 3/2 (B).

To understand how the spin forbiddeness of the O₂ reaction is overcome on the S_tot = 3/2 surface, electronic structures at three points along the O–C4 reaction coordinate are considered: (i) the O₂ complex, (ii) the transition state, and (iii) the final bridged peroxo intermediate. The amount of charge transfer among the five Fe d orbitals, the PCA π_op-sym orbital and the two O₂ π* orbitals were quantified by calculating the total percentage of orbital occupancy and assigning the major bonding partner(s) for each of the orbitals in both α and β spins. For example, in the ES complex, the β PCA π_op-sym orbital is 61% occupied (summed over all occupied molecular orbitals), and the remaining 39% is contributed in the three unoccupied Fe β dπ orbitals, with 18% on d_xz. Thus the Fe β d_xz orbital is the major bonding partner of the β PCA π_op-sym, and a total of 0.39 e⁻ has been transferred from the β PCA π_op-sym to the Fe β dπ manifold. The electron transfers among orbitals for the three different points along the reaction coordinates are summarized in Fig. 9.

Fig. 9.

Schematics of electron transfer at three different points along the O–C4^PCA reaction coordinates: O₂ complex (A), transition state (B), and peroxo intermediate (C).

O₂ Complex [Figs. 5A (Geometry) and 9A (Electronic Structure)].

The Fe–O₂ bond is defined as the z axis with the two Fe–O^PCA bonds defining the x and y axes. The energy ordering of the Fe d orbitals is as follows: d_xz < d_yz < d_xy < d_z² < d_x² − y². Because the dihedral angle between the Fe–O–O and O–O–C4 planes is 14.5°, our frontier molecular orbital analysis in ref. 36 predicted that O₂ π_ip* will directly overlap with the PCA π_op-sym orbital. Consistent with this model, 0.42 e⁻ is donated directly from the α PCA π_op-sym to the α O₂ π_ip*, whereas 0.37 e⁻ is donated through covalent bonding between β PCA π_op-sym and Fe β d_xz, as observed in our spectroscopic studies of the ES complex (Fig. 3) (36). The small charge transfer from the occupied β O₂ π* orbitals to the unoccupied Fe β d orbitals reflects the covalent interactions between O₂ and Fe. The net result is that PCA starts to gain more semiquinone character as a total of 0.79 e⁻ is lost from the PCA π_op-sym and the electronic description of the Fe–O₂ unit is between the two limiting cases, Fe^II–O₂⁻ and Fe^III–O₂.

Transition State [Figs. 5B (Geometry) and 9B (Electronic Structure)].

Due to the lengthening of the Fe–O3^PCA bond and the contraction of the Fe–O^peroxide bond, the active site now adopts an intermediate structure between square pyramidal and trigonal bipyramidal. The energy ordering of the Fe d orbitals in the same molecular coordinate system as defined in the O₂ complex is as follows: d_xz < d_yz < d_xy < d_x² −_y² < d_z². The dissociation of the Fe–O3^PCA bond allows the remaining three equatorial ligands (Tyr-408, His-462, and O4^PCA) to reorganize in the plane and leads to the stabilization of the d_xz orbital. On the other hand, the destabilization of the d_z² orbital is attributed to the strong Fe–O^peroxide bond. Due to the large geometric rearrangement to produce the 5C Fe center, the Fe–O–O–C4^PCA dihedral angle increases in magnitude from 14.5° to −35.0°, which allows direct overlap between the α PCA π_op-sym and O₂ π_op* orbitals. The charge transfer pattern in the O₂ complex also is observed in the transition state: 0.60 e⁻ from the α PCA π_op-sym to the α O_{2 op}*, 0.74 e⁻ from β PCA π_op-sym to Fe β d_xz and 0.41 e⁻ from the β O₂ π* to the Fe β d orbitals. There also is charge transfer of 0.36 e⁻ from the Fe α d_z² orbital to the α O₂ π_ip* orbital at the transition state. These results demonstrate that the two α electrons involved in reducing O₂ to O₂²⁻ come from two different sources, one from the catecholic substrate and the other from the Fe center. Due to the ligand rearrangement, the α electron transfer from the Fe proceed through the good overlap between the highest energy d orbital (d_z²) and the O₂ π_ip* orbital. However, this is compensated by the β donation from the PCA to the Fe dπ orbitals and results in an S = 3/2 intermediate spin electronic configuration at the Fe^III center. The S = 3/2 is now in its ground state as the distorted trigonal bipyramidal structure around the Fe center stabilizes this electronic configuration. Considering the total e⁻ gain and loss in the O₂ π* and PCA π_op-sym orbitals, respectively, the O₂ has been reduced to a superoxide, while the PCA is at the semiquinone level at the transition state.

Final Peroxo Intermediate [Figs. 5C (Geometry) and 9C (Electronic Structure)].

The active site becomes more trigonal bipyramidal with the axial direction defined by the short Fe–O^peroxide bond. The charge transfer pattern of this peroxo intermediate is the same as in the transition state, except that the two-electron redox process between the PCA and the O₂ molecule is now complete. The same amount of α and β spins of the PCA π_op-sym orbital are transferred to the α O₂ π_ip* and Fe β d_xz orbitals, respectively. The α electron transfer from the Fe d_z² orbital to the O₂ π_ip* has increased to 0.50 e⁻. These results show that three different electron transfer processes (one β and two α) are involved with no spin flip in the formally spin-forbidden, two-electron redox process between the singlet catecholate and the triplet O₂.

While the α electron of the singlet substrate can have direct overlap with the triplet O₂ molecule, its β counterpart cannot be transferred directly to the O₂ molecule without a spin-flip process. The Fe^III center in the intradiol dioxygenases thus plays a critical role in overcoming the spin-forbidden nature of the O₂ reaction. The strong covalent interaction between the substrate and the iron center, as reflected in the low-energy spectroscopic feature of the ES complex (Fig. 3), allows the transfer of this β electron from the substrate to the metal. With the metal acting as a buffer, an electron of the appropriate spin (α) can be transferred to the O₂ molecule to complete the two-electron redox process. However, this would leave the iron center in an excited intermediate spin state (S = 3/2). The ligand field geometry change at the Fe center along the reaction coordinate is key in stabilizing this intermediate spin configuration. As the catecholic substrate is oxidized, the iron loses the Fe–O3^PCA bond in the equatorial plane and the bonding interaction between the Fe and O₂ is strengthened in the axial direction, producing a trigonal bipyramidal ligand field with a strong axial bond, which destabilizes the d_z² orbital and stabilizes the intermediate S_tot = 3/2 spin state on the ferric center.

The Reaction Coordinate

As developed in the previous section, the O₂ reaction with the ES complex remains on the S_tot = 3/2 surface from the initial attack at the Fe center all of the way through the formation of the bridging peroxo complex. However, experimental data show that the ferric center is high-spin, S = 5/2, at the beginning and end of the catalytic cycle; hence, the 5C S_tot = 3/2 peroxo intermediate must return to the S_tot = 5/2 reaction surface. Inspection of the 6C O₂ complex and the 5C peroxo species shows that, although the three endogenous ligands show very little change in their positions, the O^Tyr-408–Fe–O4^PCA angle has increased from 100.7° to 135.7° as the peroxo is formed and the Fe–O3^PCA bond is lost. This geometric change is due to the swinging of the Fe–O4^PCA bond away from the trans position of His-462. Such geometric arrangement is very similar to those observed in the crystal structures of enzyme–inhibitor complexes (38). In these inhibitor complexes, Tyr-447 becomes the ligand trans to His-462. Opening the O^Tyr-408–Fe–O4^PCA angle in the peroxy species would allow coordination of a sixth ligand, either Tyr-447 or a solvent water molecule. Kinetics studies of the O₂ reaction of both the wild-type enzyme and the Y447H mutant show similar spectroscopic features for the two oxygen intermediates, ESO₂ and ESO₂*, with the latter being the enzyme–product complex (39). The absence of additional spectral features in the oxygen intermediates of the wild-type enzyme relative to the Y447H mutant strongly suggests the sixth ligand that binds to the 5C peroxo species is a water molecule. In the previous section, we emphasized how the intermediate spin state on the Fe^III center is stabilized by a change in coordination number from 6 to 5 and the associated ligand field as the peroxo complex is formed. Thus, we expect the change in ligand field as the sixth ligand binds to the 5C peroxo complex would allow the S_tot = 3/2 peroxo complex to return to S_tot = 5/2 and complete the catalytic reaction.

Attempts to bind a water to the open site of the S_tot = 3/2 peroxo species lead to a very long Fe–O^H2O distance of 3.61 Å, where the water is held weakly to Fe^III active site through hydrogen bonding with the catecholic substrate. However, if the water molecule is added to the active site such that it is allowed to hydrogen-bond to the peroxo bridge, an interesting geometric change is observed. The O–O bond vector changes from pointing toward C5^PCA, as in the 5C peroxo complex, to pointing toward C3^PCA (Fig. 10). Such a rearrangement for the S_tot = 5/2 species has been proposed by Siegbahn et al. (37) to be important in producing intradiol-cleaved product. Upon formation of this rearranged 5C peroxo species, proton donation from H₂O to the peroxo unit results in OH⁻ binding to the open coordination site. Geometry optimization of the 6C bridging hydroperoxy complex at both S_tot = 3/2 and S_tot = 5/2 shows that the latter species is now more stable by 2.8 kcal/mol. The geometric and electronic structures of these two bridging hydroperoxy adducts are shown in Fig. 11. Both of the structures are octahedral at the Fe^III center, with the Fe–O^hydroperoxo bond lengthened to ≈2.3 Å. The 5C strong axial trigonal bipyramidal stabilization for the intermediate spin we observe for the peroxo species is no longer present. Comparing the electronic structures of the 6C hydroperoxo species in the two different spin states and using the coordinate system in Fig. 11, we can see that SOC between d_xz′ and d_z²′ in the S_tot = 3/2 adduct through L_y would allow for intersystem crossing and the Fe^III center can now return to the S_tot = 5/2 state. We have considered the energetics for O–O bond cleavage following the formation of this S_tot = 5/2 hydroperoxy adduct that lead to the formation of an anhydride intermediate. The results are similar to those reported by Siegbahn et al. (37) and, thus, will not be further discussed in this Perspective.

Fig. 10.

Geometries of peroxo intermediates: optimized peroxo intermediate (A), rearranged peroxo intermediate optimized with hydrogen bonding from water (B). The black arrow indicates the O–O bond vector.

Fig. 11.

Optimized geometry and electronic structures of the 6C hydroperoxo complex for both S_tot = 5/2 (A) and S_tot = 3/2 (B).

In summary, coordination of the sixth ligand to the 5C S_tot = 3/2 peroxo adduct results in a ligand field change that allows the Fe^III center to return to the S_tot = 5/2 reaction surface and complete the catalytic cycle.

Concluding Comments

We have developed a mechanism in which the high-spin ferric active site of the intradiol dioxygenases can use its coordination environment to tune the ligand field around the metal center to allow a change in spin state from S = 5/2 to S = 3/2 along the reaction coordinate, which involves the transfer of three electrons without a change in their spins. The intermediate S = 3/2 spin state is accessible to the Fe^III center upon O₂ binding and dissociation of an Fe–O^catecholate bond. This provides a mechanism for overcoming the spin-forbidden nature of the reaction between the triplet O₂ and singlet organic substrates. This 5C S_tot = 3/2 peroxo intermediate must return to the S_tot = 5/2 surface to complete the reaction and this can be accomplished by coordination of H₂O to the open position in the Fe^III center. H₂O coordination also provides a proton for the O–O bond cleavage and the insertion of an O atom for the intradiol dioxygenation.

Abbreviations

LO

lipoxygenase

ES

enzyme–substrate

3,4-PCD

protocatechuate 3,4-dioxygenase

PCA

protocatechuate

5C

5-coordinate

6C

6-coordinate

SOC

spin-orbit coupling.