Abstract
IsdG is a member of the Staphylococcus aureus iron-regulated surface determinant system that degrades haem scavanged from a human host to a mixture of staphylobilin and biliverdin; the rate-limiting step of this enzyme is the rearrangment of a ferric–peroxohaem species to a hydroxylated ferryl=oxohaem speices. In this work, the role of conserved Asn and Trp residues in catalysing this reaction are interrogated by preparing and characterizing ferryl haem analogues for the hydroxylated ferryl=oxohaem speices in N7A and W67F IsdG. The ferryl haem form of N7A IsdG had an extremely short lifetime, and could only be detected by lowering the temperature to 5 °C. In contrast, the ferryl haem form of W67F IsdG could be observed at room temperature, but the lifetime of the ferryl haem form of W67F IsdG (110 s) was less than half that reported for wild-type (WT) enzyme (250 s). Magnetic circular dichroism (MCD) and electron paramagnetic resonance (EPR) characterization of the reaction mixture between ferric heme-bound N7A IsdG and meta-chloroperoxybenzoic acid (mCPBA) did not yield any evidence for a ferryl haem species, consistent with the extremely short lifetime of the ferryl haem form of N7A IsdG. Similarly, clear evidence for a ferryl haem form of IsdG was not obtained upon MCD characterization of the reaction mixture between the ferric haem form of W67F IsdG and mCPBA. However, EPR characterization of the W67F IsdG reaction provided clear evidence for the presence of a compound I or compound ES-like ferryl haem species. The lifetime measurements revealed that both Asn7 and Trp67 have important roles in stabilizing the ferryl haem form of IsdG, with Asn7 making a more significant contribution. Furthermore, the correlation between organic radical lifetime in the ferryl haem form of IsdG and the ratio of staphylobilin to biliverdin generated by the enzyme suggests that this organic radical may play a major role in determining the product selectivity of IsdG.
Introduction
The enzyme IsdG is a critical component of the iron-regulated surface determinant (isd) system used by Staphylococcus aureus to acquire iron from host-derived haem during human infection. The S. aureus isd system uses at least seven proteins to extract haem from human haemoglobin, import the host-derived haem into the bacterial cytoplasm, and degrade haem to yield non-haem iron plus various organic products.1–3 The expression of the isd pathway is regulated by the ferric uptake repressor (Fur),4–6 presumably because the pathway must establish a critical balance: haem acquisition and degradation by S. aureus must overcome nutritional immunity,7–12 the strategy by which a human host sequesters essential nutrients including iron,13, 14 but haem itself is toxic to bacteria at concentrations greater than 1 μM.15, 16 IsdG, a member of the isd pathway, has a dual role in controlling bacterial haem and iron levels both by inhibiting CpfC,17 the ferrochelatase utilized for S. aureus haem biosynthesis, and by degrading haem.18, 19 The primary products of IsdG-catalysed haem degradation are non-haem iron, staphylobilin,20–22 and formaldehyde.23, 24 In addition, IsdG also produces biliverdin and carbon monoxide as secondary products,25 which represent 10–15% of the heme degradation products for wild-type (WT) enzyme. As a consequence of its critical role in S. aureus iron homeostasis, IsdG has been identified as a novel antibiotic target for grampositive bacteria including S. aureus and Bacillus anthracis.
IsdG degrades haem to staphylobilin and formaldehyde using an enzymatic mechanism distinct from that employed by human heme oxygenase (hHO) to convert haem to biliverdin and carbon monoxide.26, 27 Overall, IsdG utilizes molecular oxygen and the reductase IruO to convert haem to non-haem iron,18, 28 staphylobilin,20 and formaldehyde via meso-hydroxyhaem and formyl-oxo-bilin intermediates.22–24 Notably, unlike like hHO, IsdG cannot use H2O2 as an alternate source of O2 and reducing equivalents since the reaction of IsdI and H2O2 results in non-specific haem degradation.29 The rate-limiting step for both the IsdG- and hHO-catalyzed haem oxygenation reactions is the rearrangement of a ferric–peroxohaem species to a hydroxylated ferryl=oxohaem.29–32 Hydroxylated ferryl=oxohame is a transient species that rapidly rearomatizes to form meso-hydroxyhaem, but ferryl haem analogues can be formed in IsdG and hHO following the addition of meta-chloroperoxybenzoic acid (mCPBA) to haem-bound enzyme. In hHO, ferryl haem has a short (1.7 s) lifetime and a compound I-like electronic structure with a porphyrin-based organic radical.33 In contrast, in IsdG, ferryl haem has a longer (250 s) lifetime and a compound ES-like electronic structure with a Trp67-based radical.34 Consequently, there is significant motivation to elucidate the origin(s) of the unique electronic structure and stability of the ferryl form of IsdG in order to better understand the kinetics and product selectivity of the enzyme.
We hypothesized that two conserved second-sphere residues, Asn7 and Trp67, may have important roles in stabilizing the ferryl haem form of IsdG and altering the course of the overall reaction. Asn7 has been shown to serve as a hydrogen bond donor to a distal ligand to haem in IsdG,30, 35 and a steric clash between the indole ring of Trp67 and the porphyrin ring of haem induces a b1u out-of-plane ruffling deformation of haem from idealized D4h symmetry (Fig. 1).36–38 Spectroscopic and computational characterization of WT and N7A IsdG has revealed long-distance electronic communication between Asn7 and the meso carbons of porphyrin; hydrogen bond donation from Asn7 weakens the Fe–distal bond and triggers partial electron transfer from porphyrin to iron.39 Similarly, studies of WT and W67F IsdG have demonstrated that porphyrin ruffling induced by Trp67 alters the relative energies of the iron 3d-based orbitals resulting in a change of electronic ground state.25, 40–42 The W67F substitution decreases the rate of IsdG-catalysed haem degradation, while the N7A substitution abolishes all enzymatic activity.43 In addition, the W67F substitution decreases staphylobilin production by IsdG and increases biliverdin production. As the ferryl haem form of IsdG has a hydrogen-bond accepting distal oxo ligand and a Trp67-based radical, it is reasonable to expect that Asn7 and Trp67 may perturb the electronic structure and properties of this species.
Fig. 1.

Overview of the chemical and biological structures relevant to this work. Predicted active-site structure for the ferryl haem form of IsdG based upon multi-scale modelling.34 The Asn7, Trp67, and His77 residues are all essential for IsdG-catalysed heme degradation.43 In this work, the ferryl haem forms of the second-sphere variants N7A and W67F IsdG were prepared and characterized (A). The known sequence of reaction intermediates in IsdG-catalysed haem degradation. Alternate isomers of formyl-oxo-bilin and staphylobilin are possible (B).20, 22 The rate-limiting step of IsdG-catalysed haem monooxygenation to yield meso-hydroxyhaem. The rate-limiting step produces a transient ferryl=oxohaem species (C).29, 30
Here we report the preparation and characterization of the ferryl haem forms of N7A and W67F IsdG. The ferryl haem forms of N7A and W67F IsdG were prepared by adding mCPBA to the ferric forms of these variants, and the stabilities of the resulting ferryl haem species were monitored using UV/Vis absorption (Abs) spectroscopy. Magnetic circular dichroism (MCD) spectroscopy was used to characterize the electronic structure of haem moiety for all components of the reaction mixtures. Finally, electron paramagnetic resonance (EPR) spectroscopy was employed to probe organic radical species generated by the reactions of N7A and W67F IsdG with mCPBA. The insights these data yield with regards to the origin of the relatively long-lived ferryl haem species of WT IsdG are discussed. In addition, the implications for the mechanism of IsdG-catalysed haem degradation are discussed. These data reveal important new functions for the conserved Asn7 and Trp67 residues in the enzymatic mechanism of IsdG.
Experimental
Unless otherwise noted, all materials used in this work were purchased from Fisher Scientific and used without further purification.
Protein expression and purification
N7A IsdG, W67F IsdG, and S219V tobacco etch virus (TEV) protease were expressed and purified as described previously. Briefly, pET15b (Ampr, Novagen) vectors encoding N7A and W67F IsdG,25, 39 and a pRK793(Ampr) plasmid encoding S219V TEV protease,44, 45 were over-expressed in Escherichia coli BL21-GOLD(DE3) cells (Agilent). N-terminal His6-tagged forms of N7A and W67F IsdG, along with S219V TEV protease, were purified using nickel affinity chromatography. Next, the N-terminal His6-tags of N7A and W67F IsdG were removed by incubation with S219V TEV protease. Finally, pure untagged forms of N7A and W67F IsdG were obtained from another round of nickel affinity chromatography. Sample purity was assessed using SDS-PAGE gel electrophoresis and apoprotein concentrations were determined via Bradford assay with bovine serum albumin (Pierce) as a standard. All spectroscopic data presented in this manuscript is derived from untagged forms of N7A and W67F IsdG.
UV/Vis Abs spectroscopy
The haem substrate was bound to N7A and W67F IsdG prior to spectroscopic characterization. As described previously for WT enzyme,30 haem-bound IsdG (IsdG–haem) was prepared by incubation of apoprotein with haem in 50 mM Tris pH 7.4, 150 mM NaCl. The mixture was exchanged into 125 mM potassium phosphate (KPi) pH 7.4 to remove unbound haem. In parallel, a 50 mM solution of mCPBA was prepared in ethanol as described previously.34 The IsdG–haem and mCPBA solutions were combined to yield 10 μM IsdG–haem, with 50 or 100 μM mCPBA, in 125 mM KPi pH 7.4. The reaction between IsdG–haem and mCPBA was characterized at various time points using UV/Vis Abs, MCD, and EPR spectroscopies, as described below.
The reactions of N7A and W67F IsdG–haem with mCPBA were monitored using UV/Vis Abs spectroscopy to identify ferryl haem species formed by the reaction and monitor their rate of decay. UV/Vis Abs spectra for room temperature reactions between IsdG–haem and mCPBA were acquired on a Cary 100 Bio UV-Vis Spectrophotometer for a wavelength range of 700–300 nm with a scan rate of 600 nm/min, a 1.0 nm data interval, and a 0.1 s integration time. In addition, UV/Vis Abs data in 625 to 375 nm range was acquired for the reaction between N7A IsdG–haem and mCPBA at 5 °C using a Cary 5000 instrument equipped with a Peltier temperature controller using the same instrumental parameters noted above. UV/Vis Abs spectra were acquired every minute for 90 minutes at room temperature, and every 30 s for 60 minutes at 5 °C. For W67F IsdG–haem, the decay of a shoulder observed in the UV/Vis Abs spectrum at 600 nm was fit to a first-order kinetic model in GraphPad Prism 9 to obtain the half-life of the ferryl haem form of this enzyme, as described previously for WT enzyme.34 For N7A IsdG–haem, the growth a band at 411 nm was fit to a first-order kinetic model to measure the half-life of compound X. With the sole exception of N7A IsdG compound X, all unique species formed upon addition of mCPBA to IsdG–haem were further characterized using MCD and EPR spectroscopy.
MCD spectroscopy
MCD spectroscopy was utilized to characterize the five unique species observed in the reactions of N7A and W67F IsdG–haem with mCPBA. In order to characterize the resting forms of N7A and W67F IsdG–haem, the IsdG–haem samples described above were combined with glycerol to yield 10 μM IsdG–haem samples in 50 mM KPi pH 7.4, 60% glycerol (v/v). These samples were loaded into homemade copper/quartz sample cells and flash-frozen in liquid nitrogen (AirGas) to yield transparent glasses. To characterize the final products of the reactions between IsdG–haem and mCPBA, the reactions described above were allowed to proceed for 90 minutes. At that time, the reaction mixtures were: combined with glycerol, loaded into MCD samples cells, and flash-frozen in liquid nitrogen. Finally, the short-lived ferryl haem form of W67F IsdG was trapped by repeating the procedure described above for the final product samples, except after one minute of reaction time. All five samples were stored under liquid nitrogen in a BioCane system to prevent ice formation prior to data collection.
MCD spectra were acquired on a home-built instrument. The sample temperatures were maintained by an Oxford SM4000–8T Spectromag with a Mercury iTC temperature controller. The magnetic field was provided by the same Oxford SM4000–8T Spectromag with a Mercury iPS power supply. The magnet was interfaced with a Jasco J-1500 spectropolarimeter using an optical rail and a series of four UV fused silica lenses (ThorLabs). Spectral data was acquired from 800–300 nm with a 1.0 nm bandwidth, 0.25 s integration time, 0.5 nm data pitch, and a 200 nm/min scanning speed. All MCD data presented in this work were processed to remove the contribution of the CD spectrum by subtracting the negative field data from the positive field data and dividing by two. These MCD spectra were compared to those reported previously for WT enzyme.34
EPR spectroscopy
EPR spectroscopy was also employed to characterize the five unique species observed in the reactions of N7A and W67F IsdG–haem with mCPBA. Higher concentration samples were desired for EPR spectroscopy, so 250 μM solutions of N7A and W67F IsdG–haem were prepared in 125 mM KPi pH 7.4. EPR samples for the resting forms of N7A and W67F IsdG–haem were prepared by loading these solutions into quartz EPR tubes (Wilmad) and flash-freezing the samples in liquid nitrogen. EPR samples of the final products were prepared by allowing the 250 μM solutions of N7A and W67F IsdG–haem to react with 2.5 mM mCPBA for 90 minutes prior to loading EPR tubes. Finally, an EPR sample of the ferryl haem form was prepared by flash-freezing the reaction mixture after one minute. As was the case with the MCD samples, the EPR samples were stored under liquid nitrogen in a BioCane prior to data collection.
EPR spectra were acquired on the X-band instrument in the University of Vermont magnetic resonance facility. X-band EPR spectra were acquired on a Bruker EMX Plus instrument with the sample temperature controlled by an Oxford continuous flow liquid helium cryostat. 10 K EPR data was collected with: 100 kHz modulation frequency, 10 G modulation amplitude, plus microwave powers ranging from 63 μW to 1 mW and time constants ranging from 5.12 to 20.48 ms. EasySpin 6.0.6 was used to simulate all five EPR spectra for N7A and W67F IsdG–haem and extract spin Hamiltonian parameters (Tables S1–S2). The complete parameter sets for all EPR spectral simulations are available in the Supporting Information.
Results
UV/Vis Abs spectroscopy
N7A IsdG.
UV/Vis Abs spectroscopy was employed to monitor the reaction of N7A IsdG–haem with mCPBA. In general, the UV/Vis Abs spectrum of the resting species was consistent with previous reports for N7A IsdG–haem (Fig. 2).30, 39, 43 The blue-shift of the Soret band and the additional low energy feature, compared to WT enzyme,34 are consistent with a high-spin species. Addition of 100 μM mCPBA to 10 μM N7A IsdG–haem yielded a new species (compound Y) with no intermediates observed in the UV/Vis Abs data (Fig. S1). Although the resting forms of WT and N7A IsdG–haem have different spin states, both fully react with mCPBA to yield ferryl haem species in under one minute, and any differences between the high- and low-spin IsdG–haem cannot be discerned from these UV/Vis Abs data. The UV/Vis Abs spectra of compound Y from N7A IsdG had characteristic Q-bands observed at 17,800 and 18,800 cm−1 (562 and 532 nm) with a Soret band at 24,400 cm−1 (410 nm). These spectroscopic features are remarkably similar to those reported previously for the compound Y form of WT IsdG, suggesting that both species have similar structures. Notably, unlike WT enzyme, the addition of mCPBA to N7A IsdG–haem converts the ferric resting state to compound Y without an observable compound X (ferryl haem) intermediate at room temperature.
Fig. 2.

UV/Vis Abs spectra for all unique species observed upon addition of 100 μM mCPBA to 10 μM IsdG–haem in 125 mM KPi pH 7.4 at room temperature. N7A IsdG converted the ferric resting state (red trace) to compound Y (green trace) without an observable intermediate (A). In contrast, W67F IsdG converted the ferric resting state (red trace) to compound Y (green trace) via compound X (gold trace) over the course of 90 minutes (B).
The reaction of N7A IsdG–haem was further investigated at lower temperatures to determine whether a ferryl haem intermediate, such as compound X, could be observed. Indeed, key differences were observed in the UV/Vis Abs spectra when 10 μM N7A IsdG–haem was reacted with 50 μM mCPBA at 5 °C (Fig. 3). Evidence for a partially-decayed ferryl haem (compound X) species is observed in the 30 s reaction time trace, which appears to fully convert to compound Y within 30 s. To quantify the rate of conversion from N7A IsdG–haem compound X to compound Y, the growth of the compound Y Soret band was fit to a first-order kinetic model (Fig. S2). This fit yielded a rate constant of 0.60 ± 0.02 min−1 and a half-life of 1.2 ± 0.4 min for the reaction at 5 °C. The half-life of N7A IsdG–haem compound X at 5 °C is nearly four-times shorter than that reported for WT enzyme at room temperature (4.0 ± 0.2 min).34 Thus, these data strongly suggest that an Asn7⋯oxo hydrogen bond stabilizes the ferryl=oxohaem (compound X) form of IsdG.
Fig. 3.

UV/Vis Abs spectra for the reaction of 10 μM N7A IsdG with 50 μM mCPBA in 125 mM KPi pH 7.4 at 5 °C. The initial ferric haem form of N7A IsdG (red trace) is rapidly oxidized to a ferryl haem species (gold trace) within 30 s. This species decays to compound Y (green trace) over the course of 60 min (intermediate traces depicted as dashed black lines).
W67F IsdG.
The reaction of W67F IsdG–haem with mCPBA was also monitored by UV/Vis Abs spectroscopy. Again, the UV/Vis Abs spectrum of W67F IsdG–haem is consistent with what has been reported previously for this species (Fig. 2).25 Similar to WT IsdG–haem, and dissimilar from N7A IsdG–haem, W67F WT IsdG–haem is a hydroxide-ligated low-spin ferric haem species.35 However, the UV/Vis Abs spectrum observed for the compound X intermediate of the W67F IsdG reaction is significantly different from that observed for WT IsdG compound X.34 Compared to the spectrum reported for WT IsdG compound X, the W67F IsdG compound X UV/Vis Abs spectrum has: a red-shifted Soret band, lacks the shoulder observed at 460 nm in WT enzyme, and has blue-shifted Q bands. In WT enzyme, the weak 650 nm band typically associated with compound I was not observed for compound X,46, 47 which can now be attributed to the compound ES-like electronic structure of this species. A weak 650 nm band is also not observed for W67F IsdG compound X, which would be consistent with a compound ES-like electronic structure assignment. Alternatively, it is possible that a weak band associated with a porphyrin cation radical may be shifted outside the spectra window considering the significant ruffling-induced red-shift of the Q-band induced by the IsdG active site. Finally, over the course of 90 minutes, compound X of W67F IsdG converted to a compound Y species whose spectral features are nearly indistinguishable from compound Y of WT enzyme (Fig. S3). Thus, the reactions of WT and W67F IsdG with excess mCPBA appear to yield similar products, albeit via different intermediates.
The kinetics for the reaction of W67F IsdG–haem with mCPBA were also assessed with UV/Vis Abs spectroscopy. For W67F IsdG, the best spectroscopic handle for quantifying the lifetime of compound X proved to be the unique UV/Vis Abs band observed at 600 nm. The decay of this band over time was monitored and fit to a first-order kinetic model yielding a rate constant of 0.39 ± 0.09 min−1 and a half-life of 1.8 ± 0.2 min (Fig. S4). Thus, both the N7A and W67F substitutions decrease the lifetime of IsdG compound X. Based upon the data reported here and elsewhere,34 the stability of a ferryl haem species decreases in the following order: WT IsdG > W67F IsdG > N7A IsdG. The spectroscopic and kinetic differences between the ferryl haem forms of these three variants motivate a detailed characterization of the species produced by the reaction of mCPBA with N7A and W67F IsdG–haem.
MCD spectroscopy
N7A IsdG.
The MCD spectrum of N7A IsdG–haem is consistent with a mixture of high- and low-spin haem. In contrast to the negatively-signed MCD band observed at 17,800 cm−1 (562 nm) in WT IsdG–haem,34 a derivative-shaped pseudo A-term was observed at this energy for N7A IsdG–haem (Fig. 4). This pseudo A-term is similar to that reported previously for the W67F variant of cyanide-inhibited IsdG (IsdG–haem–CN),25 which has been shown to have a 2Eg ground state. A second piece of evidence for the presence of a low-spin 2Eg species in N7A IsdG–haem is derived from the energy of the positively-signed MCD band observed at 20,600 cm−1 (485 nm). The energy of this band is closer to that observed for WT IsdG compound Y (20,800 cm−1), a mixture of species with a predominant contribution from a 2Eg species, versus WT IsdG–haem (20,300 cm−1), which is a relatively equal mixture of 2Eg and 2B2g species. Finally, the Soret band MCD intensity decreased two-fold from WT to N7A IsdG–haem. This observation is consistent with increased population of a high-spin S = 5/2 state.48 These spectral changes are all consistent with a perturbation of the axial water ligand pKa upon introduction of the N7A substitution, as suggested previously.39
Fig. 4.

The addition of mCPBA to N7A IsdG–haem converts high-spin ferric haem to low-spin ferric haem. The MCD spectrum of the resting ferric state (red trace) is consistent with a mixture of a low-spin 2Eg species and a high-spin (S = 5/2) species. Following the reaction with mCPBA, a low-spin 2Eg species dominates the MCD spectrum of N7A IsdG compound Y (green trace). The addition of mCPBA to WT and N7A IsdG–haem yields similar compound Y products.
Based upon the MCD data presented here, the reaction of N7A IsdG–haem with mCPBA yields a compound Y product with an increased low-spin ferric haem contribution compared to resting state enzyme. The same derivative-shaped pseudo A-term for the MCD-detected Q-band (18,200 cm−1, 550 nm), positively-signed MCD band at 20,700 cm−1, and derivate-shaped pseudo A-term Soret band (24,400 cm−1, 412 nm) were observed for N7A IsdG–haem and N7A IsdG compound Y (Fig. 4). As noted above, these spectral observations are consistent with the presence of a low-spin ferric haem species with a 2Eg electronic ground state based upon similarities to the MCD spectra of W67F IsdG–haem–CN and WT IsdG compound Y.25, 34 However, the intensity of all these spectral features is increased in N7A IsdG compound Y compared to N7A IsdG–haem. This observation indicates that N7A IsdG compound Y has a larger contribution from a 2Eg species compared to N7A IsdG–haem. Finally, it is important to note that the shapes of the WT and N7A compound Y MCD spectra are nearly identical. Thus, despite significant differences between the electronic structures of WT and N7A IsdG–haem, the addition of mCPBA yields similar products. These findings are consistent with our analysis of the UV/Vis Abs data (Fig. 2), but the MCD data provide more detailed insight into the natures of the species present in N7A IsdG–haem and compound Y.
W67F IsdG.
In contrast to WT and N7A IsdG–haem, the MCD spectrum of W67F IsdG–haem is consistent with predominance of a low-spin ferric haem with a 2Eg electronic ground state. The MCD spectrum of W67F IsdG–haem exhibited pseudo A-terms associated with the Q- (18,200 cm−1, 550 nm) and Soret (24,300 cm−1, 411 nm) bands plus a positively-signed MCD band at 21,000 cm−1 (Fig. 5). These spectral features were previously observed in WT IsdG compound Y, N7A IsdG–haem, and N7A IsdG compound Y where they were attributed to population of a 2Eg state (Fig. 4).34 Notably, the W67F IsdG–haem Soret band intensity was greater than any of the aforementioned species, or WT IsdG–haem, suggesting that W67F IsdG–haem is predominantly low-spin. Thus, there are subtle differences between the electronic structures of WT, N7A, and W67F IsdG–haem. WT IsdG–haem is a mixture of 2Eg and 2B2g species, N7A IsdG–haem is a mixture of high- and low-spin ferric haem, and W67F IsdG–haem is primarily a low-spin 2Eg species. Substitution of either of the conserved second-sphere amino acids in IsdG, Asn7 or Trp67,43 has a significant influence on the electronic structure of the resting state IsdG–haem form of the enzyme.
Fig. 5.

The addition of mCPBA to W67F IsdG–haem converts low-spin ferric haem to a different low-spin ferric haem with a different MCD spectrum. The MCD spectrum of W67F IsdG–haem is consistent with a low-spin haem possessing a 2Eg electronic ground state (red trace). Additional MCD spectra were acquired 1 min (gold trace) and 90 min (green trace) after the addition of mCPBA to W67F IsdG–haem. Both the MCD spectra acquired after 1 and 90 min of reaction time are consistent with a low-spin ferric haem possessing a 2Eg electronic configuration. There is a slight intensity gain from 1 min to 90 min indicating that a species with low MCD intensity is formed during the reaction.
The MCD spectrum of W67F IsdG compound Y is also consistent with the predominance of a low-spin ferric haem with a 2Eg ground state. Again, pseudo A-terms were observed in the MCD spectra for the Q- (18,100 cm−1, 552 nm) and Soret-(24,200 cm−1, 413 nm) bands, along with a positively-signed band at 20,900 cm−1 (Fig. 5). As discussed above, these observations can be attributed to a low-spin haem with a 2Eg electron configuration.25, 34 However, the 2.8-fold decrease in the Soret band MCD intensity from W67F IsdG–haem to W67F IsdG compound Y is notable. Based upon the short lifetime of W67F IsdG compound Y (Fig. 3), this is most likely due to haem chromophore destruction by the addition of mCPBA to W67F IsdG–haem. It is also noteworthy that the MCD spectra of WT, N7A, and W67F IsdG compound Y are very similar in both shape and intensity. These data, along with the UV/Vis Abs data presented above (Fig. 2), strongly suggest that the addition of mCPBA to WT, N7A, and W67F IsdG–haem yields similar products.
Finally, the MCD spectrum acquired 1 min after the addition of mCPBA to W67F IsdG–haem is also dominated by a low-spin ferric haem with a 2Eg electron configuration. Once again, diagnostic peaks for a 2Eg haem were observed: pseudo A-terms for the Q- (18,000 cm−1, 555 nm) and Soret (24,200 cm−1, 413 nm) bands and a positively-signed band at 20,900 cm−1 (Fig. 5). These spectral features are more similar to those observed for W67F IsdG compound Y than for W67F IsdG–haem, consistent with the observation that the initial reaction between W67F IsdG–haem and mCPBA is nearly complete after 1 min (Fig. 2). The temperature-dependence of the 700–500 nm region for this sample was carefully examined (Fig. S5), but this did not yield any evidence for a ferryl haem species. Most likely, a spectral signature for a ferryl haem was not observed in the MCD data due to the low intensity of compound I or the shorter lifetime of W67F IsdG compound X compared to WT enzyme (Fig. 3).34, 49 However, it is important to note that the MCD spectrum of the W67F IsdG–haem plus mCPBA reaction mixture does gain some intensity from the 1 min to the 90 min time point consistent with the reaction not yet being complete after 1 min. Nevertheless, MCD spectroscopy proved to be a poor tool for characterizing the electronic structure of W67F IsdG compound X and we turned to EPR spectroscopy instead.
EPR spectroscopy
N7A IsdG.
EPR data acquired for N7A IsdG–haem corroborates the MCD characterization of this species; N7A IsdG–haem is a mixture of high- and low-spin ferric haem. Similar to WT enzyme,34 simulation of the N7A IsdG–haem EPR spectrum revealed contributions from one S = 5/2 and two S = ½ species (Fig. 6, Table S1). Compared to WT enzyme, the contribution of high-spin ferric haem is greater in N7A IsdG–haem, which is consistent with the UV/Vis Abs and MCD data presented in this work (Figs. 2 and 4). The g values for the two low-spin components, g = [2.97, 2.24, 1.73] and g = [3.15, 2.29, 1.55], are consistent with ferric haem species possessing 2Eg electron configurations.50 Based upon visual inspection of the experimental and simulated EPR data, the statistically best fit does appear to underestimate the contributions of low-spin heme to the overall spectrum. Nevertheless, one explanation for the observation of two 2Eg species in N7A IsdG–haem is that they are derived from two distinct orientations of the hydroxide ligand, as previously reported for azide-inhibited IsdG (IsdG–haem–N3).39 Alternatively, these two species could correspond to two distinct conformations of the porphyrin ligand along the b1u ruffling coordinate,37 as observed for IsdG–haem–CN.30, 42 Nevertheless, a consistent description of the N7A IsdG–haem electronic structure arises from UV/Vis Abs, MCD, and EPR characterization of this species.
Fig. 6.

The lifetime of the ferryl haem form of N7A IsdG is extremely short. The EPR spectrum of N7A IsdG–haem (solid red trace) was fit to a mixture of one high-spin and two low-spin ferric haem species (dashed red trace). The EPR spectrum of N7A IsdG compound Y (solid green trace) was fit to a mixture of one high-spin and one low-spin ferric haem species (dashed green trace). No evidence for an organic radical was observed in these data.
An EPR sample of IsdG compound Y was prepared by allowing N7A IsdG–haem and mCPBA to react for 90 min before flash-freezing the sample in liquid nitrogen. The EPR spectra of IsdG compound Y could be fit to a mixture of high- and low-spin ferric haem (Fig. 6, Table S1), along with the usual contributions from aqueous iron and copper in the buffer. The weight of the high-spin contribution to the EPR simulation dropped approximately two-fold from N7A IsdG–haem to N7A IsdG compound Y, which is consistent with MCD observations (Fig. 4). Interestingly, the g values for the low-spin component, g = [2.93, 2.27, 1.55], are more consistent with the 2Eg species that persists in WT IsdG–haem following the addition of mCPBA than either of the two low-spin contributions to the EPR spectrum of N7A IsdG–haem.34 This explains the similarity of the WT and N7A IsdG compound Y MCD spectra. However, the novel 2B2g species produced by the reaction between WT IsdG–haem and mCPBA was not observed in the N7A variant. More significantly, no evidence for an organic radical was observed in the EPR spectrum of N7A IsdG compound Y. This is consistent with the extremely short lifetime of N7A IsdG.
W67F IsdG.
Based upon EPR characterization, the predominant species present in W67F IsdG–haem is a low-spin ferric haem species with a 2Eg electron configuration. Similar to N7A IsdG–haem (Fig. 6), the EPR spectrum of W67F IsdG–haem has contributions from: two low-spin ferric haem species, one high-spin ferric haem, aqueous iron, and aqueous copper (Fig. 7). The predominant species is a low-spin ferric haem with g = [3.11, 2.26, 0.98], and a smaller contribution from a low-spin ferric haem with g = [2.91, 2.07. 1.56] (Table S2). It is difficult to ascertain whether the double substitution of Ala7 and Trp67 for Asn7 and Phe67 primarily alters the distribution of porphyrin conformations, as reported for WT and W67F IsdG–haem–CN,25 or the distribution of distal ligand conformations, as reported for WT and N7A IsdG–haem–N3.30 The EPR spectrum of W67F IsdG–haem also has a contribution from a high-spin ferric haem. However, the contribution of high-spin haem to the EPR spectra of WT and W67F IsdG–haem is two orders of magnitude smaller than the contribution of high-spin haem to the EPR spectrum of N7A IsdG–haem,34 which is consistent with the UV/Vis Abs and MCD data for these species. In short, EPR spectroscopy confirms the MCD results and provides some additional details: W67F IsdG–haem is predominantly a low-spin ferric haem with a 2Eg ground state, and there are smaller contributions from a second low-spin haem species and a high-spin haem species.
Fig. 7.

A short-lived ferryl haem intermediate is observed 1 min after the addition of mCPBA to W67F IsdG–haem. The EPR spectrum of W67F IsdG–haem (solid red trace) was fit to a mixture of two low-spin and one high-spin ferric haem species (dashed red trace). The EPR spectrum of W67F IsdG compound Y (solid green trace) was also fit to a mixture of two low-spin and one high-spin ferric haem species (dashed green trace), but the low-spin ferric haem signals observed in W67F IsdG–haem and W67F IsdG compound Y were distinct. Finally, the EPR signal of the W67F IsdG–haem plus mCPBA reaction mixture flash-frozen after 1 min (solid gold trace) was fit to a mixture of W67F IsdG compound Y and a novel radical signal attributed to a ferryl haem species (dashed gold trace).
As was the case for N7A IsdG, an EPR sample of W67F compound Y was prepared by reacting W67F IsdG–haem with mCPBA for 90 minutes before quenching the reaction. Similar to observations for W67F IsdG–haem, W67F IsdG compound Y is a mixture of two low-spin ferric haem species and one high-spin ferric haem (Fig. 7). The largest contribution was a 2Eg ferric haem with g = [2.93, 2.27, 1.55] (Table S2). These g values are similar to those reported for the primary contributor to the EPR spectra of WT and N7A IsdG compound Y (Table S1),34 strongly suggesting that all three variants yield similar products following oxidation with mCPBA. A second, smaller, contribution to the N7A IsdG compound Y EPR spectrum is a low-spin ferric haem with a 2B2g ground state.50 A similar species was observed as a minor contributor to the WT IsdG compound Y EPR spectrum. Thus, EPR data reveal that WT, N7A, and W67F IsdG all produce a similar low-spin ferric haem product with a 2Eg electronic ground state. In addition, WT and W67F IsdG produce a second minor product with a rare 2B2g electronic ground state.
Unlike N7A IsdG, a distinct ferryl haem intermediate was observed 1 min after the addition of mCPBA to W67F IsdG–haem. All five species observed in the EPR spectrum of W67F IsdG compound Y are also present in the 1 min reaction time sample: a low-spin ferric haem with a 2Eg ground state, a low-spin ferric haem with a 2B2g ground state, a high-spin ferric haem, aqueous iron, and aqueous copper (Figure 7, Table S2). Importantly, the ratios of all these species are similar in the one min reaction time and W67F IsdG compound Y samples, strongly suggesting that an additional species is not masked by the W67F IsdG compound Y signal. It is also interesting to note that neither of the low-spin 2Eg species present in the W67F IsdG–haem EPR sample are present after 1 min of reaction with mCPBA. This indicates that the initial reaction between W67F IsdG–haem and mCPBA is quite rapid. Most significantly, the 1 min reaction time sample contains an organic radical signal that was not present in either W67F IsdG–haem or W67F IsdG compound Y. Most likely, this species we will label compound X is a short-lived ferryl haem species with a compound I- or compound ES-like electronic structure. Previous studies of ferryl=oxohaem species have demonstrated key differences between the bandshapes of porphyrin-, tryptophan-, and tyrosine-based radicals.51–53 Unfortunately, the organic radical bandshape in W67F IsdG compound X is masked by the significant contribution of residual W67F IsdG–haem at the 1 min reaction time point. Thus, we turned to complete spectral simulation of the mixture of W67F IsdG–haem, compound X, and compound Y species to extract key data regarding the electronic structure of W67F IsdG compound X.
Analysis of the radical signal observed for W67F IsdG compound X yielded additional insight into the electronic structure of the organic radical in this species. The zero-field splitting D value of WT IsdG compound X was estimated to be +40 ± 20 cm−1 based upon analysis of the MCD data,34 but that was not possible for W67F IsdG compound X due to low MCD signal intensity (Fig. 5). When the D value was held fixed at +40 cm−1, the best fit value of −3600 MHz did a better job of reproducing the experimental line shape compared to the value of J = −36 MHz used to fit WT IsdG compound X (Fig. S6). However, previous work has shown that the zero-field splitting D value of horseradish peroxidase (HRP) compound I is approximately half that reported for HRP compound II.49, 54 Thus, the EPR spectrum of W67F IsdG compound X was also fit with a D value of 20 cm−1, yielding a notable improvement in the fit of the derivative-shaped signal at 3350 G. Minor improvements to the overall quality of fit were achieved by using a J value previously reported by cytochrome c peroxidase,52 either directly (J = −2300 MHz) or scaled to account for the difference in D value (J = −1800 MHz). Ultimately, the overall best fit J value for W67F IsdG compound X was −7500 MHz. Importantly, regardless of whether the D value for W67F IsdG compound X is closer to +40 or +20 cm−1, the magnitude of the anti-ferromagnetic exchange coupling between the S = 1 Fe(IV) center and the organic radical is significantly larger in W67F IsdG compound X compared to WT enzyme. However, the shape of the organic radical is not consistent with that reported for compound I of cytochrome P450,55 and accurate simulation of the signal at 3250 G is complicated by possible contributions from ferric heme. Thus, W67F IsdG compound X is most likely a compound I-like species with a perturbed electronic structure due to ruffling of the porphyrin ligand, but we cannot rule out a compound ES-like electronic structure.
Discussion
Asn7 and Trp67 work in concert to stabilize the ferryl form of IsdG
The data presented in this work demonstrates that the second-sphere residue Asn7 plays a particularly important role in stabilizing the ferryl haem form of IsdG. UV/Vis Abs, MCD, and EPR characterization revealed that the addition of mCPBA to N7A IsdG cleanly converts a high-spin ferric haem to a low-spin ferric haem (Figs. 2, 4, and 6). Based upon the data acquired at 5 °C, this reaction proceeds through a short-lived ferryl haem intermediate. However, whereas the ferryl haem form of WT enzyme has a lifetime of 250 s,34 the lifetime of the ferryl haem form of N7A IsdG was too short to be detected at room temperature. Most likely, this is due to the loss of an Asn7⋯oxo hydrogen bond as hydrogen bond donation from Asn7 has previously been shown to alter the electronic structure of IsdG–haem–N3 and perturb the pKa of water/hydroxide-bound IsdG.30, 39 The loss of the hydrogen bond would raise the energy of the ferryl haem form of N7A IsdG relative to WT IsdG and increase its rate of spontaneous decay (Fig. 8). Thus, the reason that Asn7 is critical for the stability of the ferryl haem form of IsdG is most likely its contribution of a stabilizing hydrogen bond.
Fig. 8.

Current understanding for the potential energy surface of the rate-limiting step in IsdG-catalysed haem hydroxylation. Based upon previously published spectroscopic and computational data, the N7A substitution is expected to destabilize the ferric–peroxohaem “reactant” species while the W67F substitution stabilize the haem moiety.25, 39 Based upon the data presented in this manuscript, and a previous study of WT enzyme,34 the W67F and N7A substitutions are both expected to destabilize the hydroxylated ferryl=oxohaem “product” species with the N7A substitution having a more dramatic effect. Application of the Hammond postulate suggests that progressive “product” destabilization in the cases of W67F and N7A IsdG results in decreased haem hydroxylation rates,56 which correlates with experimental observations.43
The second-sphere Trp residue has a less significant influence on the lifetime of the ferryl haem form of IsdG, but a more dramatic impact on the electronic structure of that species. The lifetime of the ferryl haem form of IsdG is reduced over two-fold from 250 s to 110 s (Fig. 3),34 which is still significantly longer than the lifetime of the ferryl haem form of hHO (1.7 s).33 Most likely, this is because the W67F variant of IsdG retains an Asn7⋯oxo hydrogen bond.30, 39 But, more notably, the electronic structure of the ferryl haem form of W67F IsdG is perturbed relative to WT enzyme.57 Based upon the decreased lifetime of the ferryl haem form of W67F IsdG compared to WT enzyme, either the increased ruffling of the porphyrin ligand or electron transfer from Trp67 to porphyrin to quench a porphyrin-based cation radical stabilizes the ferryl haem form of WT IsdG (Fig. 8). Consequently, Trp67 changes the nature and stability of the ferryl haem form of IsdG.
Implications for IsdG-catalysed haem oxygenation
The data presented in this work has important implications for the kinetics of IsdG-catalysed haem degradation. The rate-limiting step of IsdG-catalysed haem oxygenation is understood to be the conversion of a ferric–peroxohaem species to a hydroxylated ferryl=oxohaem species,18, 29, 30, 35 which rapidly rearomatizes to form meso-hydroxyhaem,22, 24 via a concerted mechanism following the reduction of ferrous–oxyhaem by IruO.28 The N7A substitution has been shown to abolish enzymatic turnover,43 and the W67F substitution decreases the rate of haem oxygenation over two-fold.25 This work has shown that the lifetime of the ferryl=oxohaem analogue for the hydroxylated ferryl=oxohaem of the rate-limiting step in the N7A variant is shorter than can be observed at room temperature (Fig. 2). For W67F IsdG, the lifetime of the ferryl=oxohaem species is less than half that of WT enzyme (Fig. 3).34 Thus, there is a clear inverse correlation between ferryl haem lifetime and enzymatic rate for WT, W67F, and N7A IsdG. These data indicate that stabilization of the hydroxylated ferryl=oxohaem yielded by the rate-limiting step is a major determinant for the rate of IsdG-catalysed haem degradation.
The electronic structures of the ferryl haem forms of WT and W67F IsdG may also yield insight into the product selectivity of the enzyme. Somewhat surprisingly, considering the differences from hHO, the major organic products of WT IsdG have been identified to be staphylobilin and formaldehyde (Fig. 9).20, 23 In contrast, the major product of W67F IsdG has been reported to be biliverdin,25 similar to what is observed in the hHO-catalyzed and enzyme-free reactions.26, 58 On several occasions, it has been hypothesized that perturbation of the electronic structure of the meso-hydroxyhaem intermediate common to all three reactions is responsible for the divergent reactivities.22–25, 35 Indeed, WT IsdG has access to a redox-active Trp residue that can generate a one-electron reduced form of meso-hydroxyhaem.34 In W67F IsdG, an organic radical may be delocalized throughout the porphyrin moiety as observed previously for the hHO-bound and enzyme-free forms of meso-hydroxyhaem.59, 60 Consequently, the ability of IsdG to access a Trp radical may be the origin of the unique reactivity of meso-hydroxyhaem upon binding to IsdG.
Fig. 9.

The accessibility of a Trp-radical by the meso-hydroxyhaem intermediate may be the origin of the unique haem degradation products yielded by IsdG. WT IsdG can form a Trp-based radical within the enzyme active site,34 and the primary organic products are staphylobilin and formaldehyde.20, 23 In contrast, hHO and enzyme-free haem cannot host neighbouring Trp radicals.33 These systems yield biliverdin and carbon monoxide as major haem degradation products.25, 26 Note, the relevant isomer(s) of meso-hydroxyhaem, staphylobilin, and biliverdin vary with enzyme variant.
Conclusions
In summary, this article reports a detailed kinetic and spectroscopic characterization of the ferryl haem forms of N7A and W67F IsdG. The ferryl haem form of IsdG is an experimentally-accessible analogue for the transient hydroxylated ferryl haem species that is generated by the rate-limiting step of the enzyme.18, 29, 30, 35 The reaction of N7A and W67F IsdG–haem with mCPBA was monitored by UV/Vis Abs spectroscopy (Fig. 2), which revealed that the ferryl haem form of both variants has a reduced lifetime compared to WT enzyme (Fig. 3).34 These data reveal an inverse correlation between haem oxygenation rate and ferryl haem lifetime for three variants of IsdG. MCD and EPR characterization of the ferryl haem form of W67F IsdG revealed that the species has a distinct electronic structure compared to the ferryl haem form of WT enzyme. In addition, the distinct major products for W67F IsdG,25 compared to WT IsdG,20, 23 suggest that the redox-active Trp67 residue positioned in the active site of WT IsdG may be a major determinant of product selectivity (PDB ID 2ZDO).36 The novel mechanistic features of S. aureus IsdG relative to hHO revealed by this study are interesting from the view of fundamental inorganic complex reactivity, and intriguing from the perspective of antibiotic drug design.
Supplementary Material
† Supplementary Information available: Additional UV/Vis Abs and MCD data. EPR simulation parameters.
Acknowledgements
M.D.L. thanks the National Institutes of Health (R35-GM139516) and the National Science Foundation (CHE-1919417) for financial support. The authors thank Eric Skaar (Vanderbilt University) and Celia Goulding (University of California-Irvine) for the pET15b vector encoding IsdG. We also thank David Waugh (National Cancer Institute) for the pRK793 plasmid encoding S219V TEV protease. Finally, we thank Monika Ivancic (University of Vermont) for assistance with EPR data acquisition.
Footnotes
Conflicts of interest
There are no conflicts to declare.
Data availability
The data supporting this article have been included as part of the Supplementary Information.
References
- 1.Mazmanian SK, Skaar EP, Gaspar AH, Humayun M, Gornicki P, Jelenska J, Joachmiak A, Missiakas DM and Schneewind O, Science, 2003, 299, 906–909. [DOI] [PubMed] [Google Scholar]
- 2.Skaar EP, Humayun M, Bae T, DeBord KL and Schneewind O, Science, 2004, 305, 1626–1629. [DOI] [PubMed] [Google Scholar]
- 3.Grigg JC, Ukpabi G, Gaudin CFM and Murphy MEP, J. Inorg. Biochem, 2010, 104, 341–348. [DOI] [PubMed] [Google Scholar]
- 4.Skaar EP and Schneewind O, Microbes Infect., 2004, 6, 390–397. [DOI] [PubMed] [Google Scholar]
- 5.Reniere ML and Skaar EP, Mol. Microbiol, 2008, 69, 1304–1315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lojek LJ, Farrand AJ, Weiss A and Skaar EP, Int. J. Med. Microbiol, 2018, 308, 582–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hammer ND and Skaar EP, Annu. Rev. Microbiol, 2011, 65, 129–147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Cassat JE and Skaar EP, Semin. Immunopathol, 2012, 34, 215–235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hammer ND and Skaar EP, Curr. Opin. Microbiol, 2012, 15, 10–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hood MI and Skaar EP, Nat. Rev. Microbiol, 2012, 10, 525–537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cassat JE and Skaar EP, Cell Host Microbe, 2013, 13, 509–519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Choby JE and Skaar EP, J. Mol. Biol, 2016, 428, 3408–3428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ratledge C and Dover LG, Annu. Rev. Microbiol, 2000, 54, 881–941. [DOI] [PubMed] [Google Scholar]
- 14.Palmer LD and Skaar EP, Ann. Rev. Genet, 2016, 50, 67–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Skaar EP, Gaspar AH and Schneewind O, J. Bacteriol, 2006, 188, 1071–1080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Reniere ML, Torres VJ and Skaar EP, Biometals, 2007, 20, 333–345. [DOI] [PubMed] [Google Scholar]
- 17.Videira MAM, Lobo SAL, Silva LSO, Palmer DJ, Warren MJ, Prieto M, Coutinho A, Sousa FL, Fernandes F and Saravia LM, Mol. Microbiol, 2018, 109, 385–400. [DOI] [PubMed] [Google Scholar]
- 18.Skaar EP, Gaspar AH and Schneewind O, J. Biol. Chem, 2004, 279, 436–443. [DOI] [PubMed] [Google Scholar]
- 19.Conger MA, Pokhrel D and Liptak MD, Metallomics, 2017, 9, 556–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Reniere ML, Ukpabi G, Harry SR, Stec DF, Krull R, Wright DW, Bachmann BO, Murphy ME and Skaar EP, Mol. Microbiol, 2010, 75, 1529–1538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Haley KP, Janson EM, Heilbronner S, Foster TJ and Skaar EP, J. Bacteriol, 2011, 193, 4749–4757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Streit BR, Kant R, Tokmina-Lukaszewska M, Celis AI, Machovina MM, Skaar EP, Bothner B and DuBois JL, J. Biol. Chem, 2016, 291, 862–871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Matsui T, Nambu S, Ono Y, Goulding CW, Tsumoto K and Ikeda-Saito M, Biochemistry, 2013, 52, 3025–3027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Matsui T, J. Biol. Chem, 2023, 299, 104648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Schuelke-Sanchez AE, Cornetta AR, Kocian TAJ, Conger MA and Liptak MD, J. Inorg. Biochem, 2022, 230, 111775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wilks A, Black SM, Miller WL and Ortiz de Montellano PR, Biochemistry, 1995, 34, 4421–4427. [DOI] [PubMed] [Google Scholar]
- 27.Wilks A and Heinzl G, Arch. Biochem. Biophys, 2014, 544, 87–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Loutet SA, Kobylarz MJ, Chau CHT and Murphy MEP, J. Biol. Chem, 2013, 288, 25749–25759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Takayama SJ, Loutet SA, Mauk AG and Murphy ME, Biochemistry, 2015, 54, 2613–2621. [DOI] [PubMed] [Google Scholar]
- 30.Lockhart CL, Conger MA, Pittman DS and Liptak MD, J. Biol. Inorg. Chem, 2015, 20, 757–770. [DOI] [PubMed] [Google Scholar]
- 31.Davydov R, Kofman V, Fujii H, Yoshida T, Ikeda-Saito M and Hoffman BM, J. Am. Chem. Soc, 2002, 124, 1798–1808. [DOI] [PubMed] [Google Scholar]
- 32.Chen H, Moreau Y, Derat E and Shaik S, J. Am. Chem. Soc, 2008, 130, 1953–1965. [DOI] [PubMed] [Google Scholar]
- 33.Matsui T, Kim SH, Jin H, Hoffman BM and Ikeda-Saito M, J. Am. Chem. Soc, 2006, 128, 1090–1091. [DOI] [PubMed] [Google Scholar]
- 34.Grover A, Conger MA and Liptak MD, J. Am. Chem. Soc, 2025, 147, 10598–10611. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Takahashi S, Nambu S, Matsui T, Fujii H, Ishikawa H, Mizutani Y, Tsumoto K and Ikeda-Saito M, Biochemistry, 2020, 59, 3918–3928. [DOI] [PubMed] [Google Scholar]
- 36.Lee WC, Reniere ML, Skaar EP and Murphy ME, J. Biol. Chem, 2008, 283, 30957–30963. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jentzen W, Song X-Z and Shelnutt JA, J. Phys. Chem. B, 1997, 101, 1684–1699. [Google Scholar]
- 38.Graves AB, Graves MT and Liptak MD, J. Phys. Chem. B, 2016, 120, 3844–3853. [DOI] [PubMed] [Google Scholar]
- 39.Conger MA, Cornetta AR and Liptak MD, Inorg. Chem, 2019, 58, 15455–15465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Takayama SJ, Ukpabi G, Murphy ME and Mauk AG, Proc. Natl. Acad. Sci, 2011, 108, 13071–13076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ukpabi G, Takayama SJ, Mauk AG and Murphy MEP, J. Biol. Chem, 2012, 287, 34179–34188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Graves AB, Horak EH and Liptak MD, Dalton Trans., 2016, 45, 10058–10067. [DOI] [PubMed] [Google Scholar]
- 43.Wu R, Skaar EP, Zhang R, Grazyna J, Piotr G, Schneewind O and Joachmiak A, J. Biol. Chem, 2005, 280, 2840–2846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Kapust RB, Tözsér J, Fox JD, Anderson DE, Cherry S, Copeland TD and Waugh DS, Protein Eng., 2001, 14, 993–1000. [DOI] [PubMed] [Google Scholar]
- 45.Nallamsetty S, Kapust RB, Tözsér J, Cherry S, Tropea JE, Copeland TD and Waugh DS, Protein Expr. Purif, 2004, 37, 108–115. [Google Scholar]
- 46.Hewson WD and Hager LP, J. Biol. Chem, 1979, 254, 3182–3186. [PubMed] [Google Scholar]
- 47.Lad L, Mewies M and Raven EL, Biochemistry, 2002, 41, 13774–13781. [DOI] [PubMed] [Google Scholar]
- 48.Paulat F and Lehnert N, Inorg. Chem, 2008, 47, 4963–4976. [DOI] [PubMed] [Google Scholar]
- 49.Browett WR, Gasyna Z and Stillman MJ, J. Am. Chem. Soc, 1988, 110, 3633–3640. [Google Scholar]
- 50.Walker FA, Coord. Chem. Rev, 1999, 185–186, 471–534. [Google Scholar]
- 51.Ivancich A, Jouve HM, Sartor B and Gaillard J, Biochemistry, 1997, 36, 9356–9364. [DOI] [PubMed] [Google Scholar]
- 52.Ivancich A, Dorlet P, Goodin DB and Un S, J. Am. Chem. Soc, 2001, 123, 5050–5058. [DOI] [PubMed] [Google Scholar]
- 53.Ivancich A, Mazza G and Desbois A, Biochemistry, 2001, 40, 6860–6866. [DOI] [PubMed] [Google Scholar]
- 54.Schulz CE, Devaney PW, Winkler H, Debrunner PG, Doan N, Chiang R, Rutter R and Hager LP, FEBS Lett., 1979, 103, 102–105. [DOI] [PubMed] [Google Scholar]
- 55.Rittle J and Green MT, Science, 2010, 330, 933–937. [DOI] [PubMed] [Google Scholar]
- 56.Hammond GS, J. Am. Chem. Soc, 1955, 77, 334–338. [Google Scholar]
- 57.Moody PCE and Raven EL, Acc. Chem. Res, 2018, 51, 427–435. [DOI] [PubMed] [Google Scholar]
- 58.Sano S, Sano T, Morishima I, Shiro Y and Maeda Y, Proc. Natl. Acad. Sci, 1986, 83, 531–535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Morishima I, Fujii H and Shiro Y, Inorg. Chem, 1995, 34, 1528–1535. [Google Scholar]
- 60.Liu Y, Moënne-Loccoz P, Loehr TM and Ortiz de Montellano PR, J. Biol. Chem, 1997, 272, 6909–6917. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting this article have been included as part of the Supplementary Information.
