Evidence That the Heme Regulatory Motifs in Heme Oxygenase-2 Serve as a Thiol/Disulfide Redox Switch Regulating Heme Binding^*s

Cloning Overexpression, and Purification of Human HO-2

The human full-length HO-2 cDNA in a pGEX-4T-2 vector (Amersham Biosciences), which was graciously contributed by Dr. Mahin D. Maines (University of Rochester, School of Medicine), was transformed into Escherichia coli strain BL21. In this construct, the human HO-2 fragment is fused to glutathione S-transferase (GST) and thus can be overexpressed and purified using a glutathione affinity column.

BL21 cells carrying the pGEX-4T-2/HO-2 plasmid were cultured in 1 liter of LB medium containing 100 μg/ml ampicillin at 37 °C. When the absorbance at 600 nm reached 0.8–1.0, 60 mg of 5-aminolevulinic acid hydrochloride (Sigma) was added along with 300 μl of 1 m isopropyl β-d-thiogalactopyranoside (Gold BioTechnology, St. Louis, MO) to induce the expression of HO-2. The cells were incubated for another 15 h at 25 °C and harvested by centrifugation at 7000 rpm for 10 min at 4 °C in a Beckman J2-HS centrifuge.

The cell pellet was resuspended in 5 volumes of 1× phosphate-buffered saline (PBS; 140 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, and 1.8 mm KH₂PO₄, pH 7.4), containing Triton X-100 (0.04% (v/v); Sigma), protease inhibitor (1 tablet/50 ml of extraction solution; Roche Applied Science), lysozyme (0.5 mg/ml; Sigma), and DNase I (5 units/ml; Sigma) at 4 °C. Cells were then lysed by sonication, and the suspension was centrifuged at 17,000 rpm for 1 h at 4 °C in the Beckman J2-HS centrifuge. The supernatant was loaded onto a 10-ml glutathione-Sepharose 4B column (Amersham Biosciences), which was washed with 1× PBS containing 0.04% (v/v) Triton X-100. The HO-2/GST fusion protein was eluted with buffer containing 50 mm Tris-HCl, 0.04% (v/v) Triton X-100, and 10 mm reduced glutathione, pH 7.4; and the combined fractions containing the HO-2/GST fusion protein were extensively dialyzed with 1 × PBS containing 0.04% (v/v) Triton X-100 to remove glutathione. The GST domain then was cleaved by limited proteolysis using thrombin, and HO-2 was separated from GST by chromatography on a glutathione-Sepharose 4B column. All of these steps were carried out at 4 °C.

Site-directed Mutagenesis of HO-2

As described below, the truncated HO-2 variants are more stable and soluble than the full-length protein. A number of variants were generated to determine the functional roles of the HRMs in HO-2. Two kinds of truncated HO-2 were generated: one lacking the C-terminal membrane-binding region (HO-2Δ289-316) the other lacking the two C-terminal HRMs and the membrane-binding region (HO-2Δ265–316). The C265A, C282A, C127A, C265A/C282A, and C127A/C282A variants were generated from HO-2Δ289–316, and C127A/Δ265–316 was from HO-2Δ265–316. All mutations were generated using the QuikChange site-directed mutagenesis protocol (Stratagene, La Jolla, CA). The pGEX-4T-2/HO-2 plasmid was the template for PCRs using primers from Integrated DNA Technologies (25). All of these variants were purified by the same procedure described above for the wild-type enzyme.

HO-2 Assay

The HO-2 enzymatic assay was slightly modified from that described previously (26). The 1-ml reaction mixture contained 0.1 m potassium phosphate buffer, pH 7.4, 20 μm freshly prepared hemin (from a 2 mm stock containing 0.1 m NaOH and 5% Me₂SO), 0.1 mg/ml bovine serum albumin, 30 μg of NADPH-cytochrome P450 reductase, 60 μg of purified human biliverdin reductase, and 30–100 μg of HO-2. The mixture was incubated at 37 °C, and the reaction was started by the addition of 10 μl of 100 mm NADPH and monitored by following the increase in absorbance at 468 nm using a Beckman DU7400 spectrophotometer. The specific activities of the various wild-type and variant HO-2 proteins were calculated using the difference extinction coefficient between heme and bilirubin at 468 nm (43.5 mm⁻¹ cm⁻¹). P450 reductase was donated by Dr. Bettie Sue Masters (University of Texas Health Sciences Center, Arlington, TX). Biliverdin reductase was purified from E. coli cells containing pGEX-4T-2/BVR. The BVR gene was purchased from American Type Culture Collection (Manassas, VA) and then subcloned into the pGEX-4T-2 plasmid.

Quantification of Heme Binding by the Pyridine Hemochrome Assay

HO-2·heme complexes were prepared by incubating purified HO-2 variants with a 5-fold molar excess of Fe³⁺ heme from a stock that was freshly prepared in 5% Me₂SO to prevent aggregation. After incubation at 4 °C for 10 min, excess free heme was removed by chromatography on a Sephadex G-15 gel filtration column (Sigma). The amount of heme bound to HO-2 was calculated using a difference extinction coefficient at 556 nm of 28.32 mm⁻¹ cm⁻¹ (27).

Determination of Free Thiol Groups by the 5,5′ -Dithiobis(ni-trobenzoic Acid) (DTNB) Assay

Free thiol quantification of HO-2 was conducted by the DTNB assay basically as described (28, 29). After incubation of the 1-ml reaction mixture containing HO-2,100 mm Tris-HCl, pH 8.0, and 100 μm DTNB at room temperature for 15 min, the absorbance at 412 nm was recorded, and the free thiol concentrations were calculated by reference to a standard curve generated using dithiothreitol (DTT). The DTNB titration was performed in the presence and absence of 8 m urea. When the thiol groups in DTT-reduced HO-2 were measured, the protein was extensively dialyzed in an anaerobic chamber (Vacuum Atmospheres Co., Hawthorne, CA) before the assay.

Alkylation of Cysteine Residues and Mass Spectrometric Analysis

One- and two-step alkylation reactions were performed as described previously (30). Samples (2 mg/ml) of reduced (with 10 mm DTT) and oxidized (lacking DTT) HO-2 were dissolved in 0.5 ml of 8 m urea, 50 mm Tris-HCl, pH 8.2, and 1 mm Na-EDTA. The first alkylation step was performed by the addition of 50 mm iodoacetamide (final concentration) and incubation for 30 min at room temperature. The excess iodoacetamide was removed by precipitating the protein by adding a solution containing cold acetone and 1 n HCl/H₂O (98:2:10), followed by centrifugation at 5000 rpm for 10 min at 4 °C. After washing the pellet three times with the same solution, the protein was reduced by reacting with 10 mm DTT for 30 min at 37 °C, and the second alkylation step was performed by the addition of 2 μl of 9.5 m 4-vinylpyridine and incubation for 20 min at room temperature. The 4-vinylpyridine addition was repeated three times. The unbound 4-vinylpyridine was removed using the acidic acetone solution as described above. After the one- and two-step alkylation reactions, samples were digested with trypsin overnight at room temperature, and the tryptic peptides were isolated and analyzed by nano-liquid chromatography/tandem mass spectrometry using a QSTAR XL mass spectrometer (Applied Biosystems). Peptide isolation was accomplished with a C18 PepMap100 column (75 μm × 15 cm, 3 μm, 100 Å; LC Packings, Sunnyvale, CA) using a linear gradient from 0.3% formic acid in H₂O to 0.3% formic acid in acetonitrile at a flow rate of 170 nl/min. The mass spectroscopic results were analyzed using the MASCOT search engine.

Spectroscopic and Analytical Methods

Protein concentration was calculated based on the rose bengal method (30) using a standard curve generated using known amounts of HO-2, the concentration of which was determined by dry weight.

The heme binding affinity of each of the HO-2 variants was determined by adding heme to the reference and sample (containing HO-2) cuvettes and measuring the difference spectrum from 350 to 750 nm in an Olis updated Cary 14 double-beam spectrophotometer. To determine the binding affinity of Fe³⁺ heme, freshly prepared hemin (from a stock solution of 500 μm hemin in 5% Me₂SO) was added in 1–2-μl aliquots to the sample cuvette containing 8 μm HO-2 in 1 ml of 100 mm potassium phosphate buffer, pH 7.4, and to the reference cuvette containing the same buffer. The stock hemin (Fe³⁺ heme) solution was prepared by dissolving hemin in 0.1 m NaOH and 5% Me₂SO and filtration with a 0.2-μm syringe filter (Amicon, Beverly, MA). The heme concentration was calculated using an extinction coefficient at 385 nm of 58.4 mm⁻¹ cm⁻¹ (31).

To determine the binding affinity of reduced HO-2 for hemin, the as-isolated protein was incubated with a 50-fold molar excess of DTT in the anaerobic chamber for 30 min. DTT was then removed by dialyzing the protein in anaerobic buffer (50 mm Tris-HCl and 50 mm KCl, pH 7.4). Titration was performed as just described for the oxidized protein, except that it was performed under anaerobic conditions in serum-stoppered cuvettes.

To determine the binding affinity of HO-2 for ferrous heme, freshly prepared sodium dithionite (2 mm final concentration) was added to a stock solution of 200 μm Fe²⁺ heme in 20% Me₂SO in the anaerobic chamber. Titrations with the reduced protein and Fe²⁺ heme were performed as described above in serum-stoppered cuvettes to maintain anaerobic conditions.

The heme titration data were plotted and fit to a one-binding site model to determine the dissociation constant (Equations 1 and 2),

where ΔA is the absorbance difference between sample and reference cuvettes; Δ∈ is the difference extinction coefficient between bound and free heme; and EL is the concentration of the HO-2·heme complex, which is calculated using the quadratic Equation 2, in which E_O is the total HO-2 concentration, L_O is the total heme concentration, and K_d is the dissociation constant.

EPR measurements were performed on a Bruker ESP 300E spectrometer recently upgraded to an EMX, operating at ∼ 9.39 GHz and equipped with an Oxford ITC4 temperature controller, a Hewlett-Packard Model 5340 automatic frequency counter to monitor microwave frequency, and a Bruker gaussmeter to determine the magnetic flux density. All of the EPR samples were prepared in 50 mm phosphate buffer at pH 7.4.

Protein Stability Measured by Differential Scanning Calorimetry (DSC)

Heat-induced unfolding was performed in a DSC microcalorimeter (MicroCal, LLC, Northhampton, MA). Purified protein was dialyzed against 1 × PBS, pH 7.4, to remove the Triton X-100, and a 0.2 mg/ml protein solution was degassed for 5 min at room temperature before loading into the cells. The scan was performed at a rate of 1 °C/min from 20 to 80 °C. Thermogram analysis was performed using Origin Version 7.0 software supplied with the instrument.

Effect of HRMs on HO-2 Stability

Similar amounts of purified HO-2 were recovered from cells expressing wild-type protein or any of the variants. Each of the purified proteins showed predominantly a single band when analyzed by SDS-PAGE, with a purity between 91 and 96% based on densitometric analysis (Fig. 2A). In addition, DSC experiments were carried out to evaluate the influence of HRMs on HO-2 stability. The DSC profiles for the variants overlay that for the wild-type protein and fit a non-two-state model (Fig. 3) with similar melting temperatures in the range of 52.4-53.7 °C (Table 1). These results indicate that neither the HRMs nor the sequence between the HRMs at the C terminus affects protein stability.

Effect of HRMs on HO-2 Activity

The steady-state enzymatic activities of the various forms of HO-2 were determined by following biliverdin formation using the assay described under “Experimental Procedures.” In all experiments, 5% Me₂SO was present in the heme stock solution (0.05% in the final assay mixture) to prevent heme aggregation. We established that the addition of Me₂SO in the assay mixture at concentrations up to 0.2% had no effect on HO-2 activity. As shown in Fig. 2B, both full-length HO-2 and the HO-2Δ289–316 truncation variant, which lacks the membrane-spanning region, exhibited similar activities of ∼5.2 nmol/min/mg. Therefore, because the C-terminal membrane-binding region does not affect enzymatic activity, Cys substitutions were generated in the more soluble and stable HO-2Δ289–316 truncation variant to explore the function(s) of the HRMs in HO-2 (Fig. 2B). To simplify their descriptions, we have omitted the Δ289-316 designation from the variants; thus, substitution of Cys²⁶⁵ with Ala in HO-2Δ289–316 generated a variant that we labeled simply C265A. All of the Cys variants, including C265A/C282A, C127A, C265A, C282A, and C127A/C282A, exhibited similar heme degradation activities in the range of 5.0–5.5 nmol/min/mg. The HO-2Δ265–316 truncation variant, lacking the entire HRM1-HRM2 region, and C127A/Δ265–316, lacking all three Cys residues, had similar enzymatic activities of 5.5 and 5.3 nmol/min/mg, respectively. These results strongly indicate that HRM1, HRM2, and HRM3 do not affect steady-state HO-2 activity.

Evaluation of Heme Binding to the HRMs in HO-2

A previous study indicated that rat HO-2 binds three hemes: one heme at the active site (coordinated by His⁴⁵) plus one each at the two C-terminal HRMs (24). However, the UV-visible spectra of the Fe(III), Fe(II), and Fe(II)-CO states of all the HRM variants are identical to those of the wild-type enzyme (Fig. 4, A–C). The spectra show a Soret peak at 406 nm and four long wavelength range peaks at 503, 532, 571, and 630 nm (Fig. 4A). The Fe(II) states exhibit a Soret peak at 430 nm and a long wavelength peak at 557 nm with a shoulder at 527 nm (Fig. 4B), whereas the Fe(II)-CO adducts show a Soret peak at 422 nm and two long wavelength peaks at 541 and 572 nm (Fig. 4C). These spectra, including the extinction coefficients, are very similar to those of the Fe(III), Fe(II), and Fe(II)-CO states of human HO-1 (32) and of myoglobin (33), which bind only one heme/monomer and do not contain any HRMs, strongly indicating that HO-2 binds only one heme/monomer.

To further assess the possibility of heme binding to the HRMs, a 5-fold excess of heme was incubated with the wild-type enzyme and each of the variants; unbound heme was removed by gel filtration chromatography; and the heme content was measured by the pyridine hemochrome assay (Fig. 5). Oxidized HO-2Δ289–316 and C282A at 1.58 and 1.51 μm, respectively, were found to bind ∼1.5 μm ferric heme. The other HO-2 variants also exhibited 1:1 binding of ferric heme (supplemental Table 1). In summary, the titration and pyridine hemochrome results strongly indicate that HO-2, like HO-1, can only bind one heme/monomer and that the HRMs do not bind additional heme.

As determined by titrations with ferric heme, DTT-reduced HO-2Δ289–316, like the oxidized protein, was saturated at a stoichiometry of 1 mol of heme/mol of monomeric protein (Fig. 6). Similarly, all other HO-2 variants exhibited saturation at a 1:1 ratio in the oxidized and DTT-reduced states (supplemental Fig. 4). Therefore, these results strongly indicate that both oxidized and reduced states of HO-2 bind only one heme.

HRM1 and HRM2 Undergo Thiol/Disulfide Redox Interconversion

The three Cys residues at positions 256, 282, and 127 in the HO-2 sequence are part of HRM1, HRM2, and HRM3, respectively. To determine the redox state of these residues, we determined the number of free thiol groups in HO-2 and variants in their as-isolated (oxidized) and DTT-reduced states using the DTNB assay (Table 2). In all cases, DTT was removed before the DTNB assay. This assay was performed with native protein and with urea-denatured protein. In nearly all experiments, the full recovery of thiols required urea treatment, indicating that none of the Cys residues are completely accessible in the native state. For as-isolated HO-2Δ265–316, a single thiol group was detected; however, 3 thiols were measured for the DTT-reduced protein. The C127A variant was found to contain no free thiol groups in the oxidized state, but two in the DTT-reduced state. Correspondingly, HO-2Δ265–316 and the C265A/C282A variant, with only one Cys at position 127, contained one thiol group in both the oxidized and DTT-reduced states. Thus, the combined results of the DTNB titrations strongly indicate that as-isolated HO-2 contains an intramolecular disulfide bond between Cys²⁶⁵ and Cys²⁸² to link HRM1 and HRM2.

Consistent with this conclusion, when the C265A variant was reduced with DTT, 1.93 free thiols were measured by the DTNB assay. However, the results with the oxidized C265A variant were more ambiguous, exhibiting 1.39 thiols, instead of 2.0 (Table 2), suggesting that, when HRM2 is absent, Cys²⁸² has some ability to form a disulfide bond with Cys¹²⁷. We excluded the possibility of intersubunit cross-links by performing nonreducing SDS-PAGE (supplemental Fig. 1). Even when these variants (or the wild-type protein) were treated with diamide, <5% of the protein was observed at the position of the dimer, and there was no evidence for the formation of higher oligomeric states.

The intramolecular disulfide bond formed between the C-terminal HRMs was confirmed by mass spectrometric analysis using one- and two-step alkylation protocols (supplemental Table 3). HO-2 was reacted with iodoacetamide in the first step to label the free cysteine thiols; the protein was then reduced with DTT and reacted with 4-vinylpyridine to label the cysteines that were initially linked by disulfide bonds. Samples were removed after both alkylation steps for nano-liquid chromatographic/tandem mass spectrometric analysis. After the one-step alkylation reaction, coverage of the HO-2 sequence was 73% for the oxidized protein and 81% for the reduced protein. After the two-step procedure, sequence coverage was 76%. In the first alkylation reaction, only Cys¹²⁷ of oxidized HO-2 was S-carboxamidomethyl-labeled by iodoacetamide, and all three cysteines (Cys¹²⁷, Cys²⁶⁵, and Cys²⁸²) of reduced HO-2 contained the S-carboxamidomethyl label. Correspondingly, with the two-step alkylation protocol, Cys¹²⁷ was predominantly labeled by iodoacetamide, and Cys²⁶⁵ and Cys²⁸² were S-4-vinylpyridylethyl-labeled by 4-vinylpyridine. Specifically, after the two-step procedure, 86% of the peptides containing Cys¹²⁷ were labeled by iodoacetamide, and 94% of the peptides containing Cys²⁶⁵ or Cys²⁸² were labeled by 4-vinylpyridine.

Surprisingly, for the C282A and C127A/C282A variants in the oxidized state, even in the presence of urea, the DTNB assay showed only 0.91 and0.14thiols (instead of 2.0 and 1.0), respectively (Table 2). However, after DTT treatment, 1.88 and 0.93 thiols were observed. As for the other variants, nonreducing SDS-PAGE experiments showed that <5% of these two variants were in a dimeric state. Thus, Cys²⁶⁵ in these oxidized HO-2 variants does not appear to be in the form of a thiolate or a disulfide because Cys²⁸² is absent in both variants and because Cys²⁶⁵ is the only Cys residue in the one variant. However, the thiol state was recovered after treatment with DTT. These properties are consistent with those observed for cysteine sulfenate.

Thus, we performed experiments with 7-chloro-4-nitro-benz-2-oxa-1,3-diazole (NBD-Cl) to assess the possibility that Cys²⁶⁵ is present in the sulfenic acid form. Upon reaction with NBD-Cl, the oxidized C127A/C282A variant exhibited a peak at 352 nm (supplemental Fig. 2), which is characteristic of the Cys-S(O)-NBD derivative (34). Upon reduction with DTT, the 352 nm peak was markedly reduced, and a peak at 418 nm appeared, which is characteristic of the Cys-S-NBD adduct. The 418 nm peak is absent in the spectrum of the oxidized C127A/C282A variant, indicating that Cys²⁶⁵ is mostly oxidized. In addition, the 352 nm peak appears to be absent in the C127A/Δ265–316 variant (lacking all three Cys residues). We also have not seen evidence for a sulfenic acid derivative in variants that contain Cys²⁸². These results indicate that, in the oxidized variant, Cys²⁶⁵ is present in the sulfenate form and can be reduced by DTT to the thiol. These properties are characteristic of cysteine sulfenates because the sulfinate and sulfonate are not reducible by DTT (34,35). Our results are consistent with Scheme 1, showing that, under oxidizing conditions, Cys²⁶⁵ can form a sulfenate that reacts with Cys²⁸² to form the disulfide that is observed in the oxidized wild-type protein.

Effect of Redox State of the HRMs on Heme Binding

The effect of the redox state of the HRMs on heme binding was determined. Note that “oxidized” and “reduced” will refer to the state of the disulfide bonds of the HRMs, and the oxidation state of the iron in the heme will be explicitly identified as (II) or (III). The UV-visible spectra of the oxidized Fe(III) (Fig. 4A) Fe(II) (Fig. 4B), and Fe(II)-CO (Fig. 4C) states of all of the HRM variants are identical to those of oxidized wild-type HO-2 and overlay those of the corresponding DTT-reduced proteins (supplemental Fig. 3). The extinction coefficients of HO-2 are presented in supplemental Table 2. Furthermore, the absorption spectra of the Fe(III), Fe(II), and Fe(II)-CO states of DTT-reduced HO-2Δ289–316 (Fig. 4D) are indistinguishable from those of the oxidized protein (Fig. 4, A–C). Thus, the UV-visible spectra of HO-2 in the Fe(III), Fe(II), and Fe(II)-CO states are independent of the redox state of the HRMs.

As described above, HO-2 binds only 1 mol of heme/mol of monomer in both the oxidized and reduced states. However, the titration curves of the oxidized and reduced states of HO-2 are noticeably different. The oxidized proteins showed a sharp break at the equivalence point, whereas the reduced enzymes exhibited a noticeable bend, indicating that oxidized HO-2 binds Fe³⁺ heme much more tightly than does the DTT-reduced protein. Thus, the dissociation constants (K_d) were measured for the oxidized and DTT-reduced states of HO-2Δ289–316 and each of the HRM variants.

Based on Fe³⁺ heme titrations for each of the HO-2 variants in the oxidized and reduced states (Fig. 2 and supplemental Fig. 3), K_d values were calculated based on a one-site binding model (Table 3). The K_d values of all the oxidized HO-2 proteins were very similar, varying from 0.028 to 0.050 μm. Because the K_d values were lower than the protein concentrations used in the titration, the data were analyzed by a quadratic equation routinely used in such cases in which the free and bound hemes are taken into account. To ensure that the K_d values are valid, we repeated our experiments at a 10-fold lower HO-2 concentration (1.0 μm) and obtained the same K_d values (data not shown). The hemin binding affinity of DTT-reduced HO-2Δ289–316 (K_d = 0.348 μm) was 10-fold lower than that of the oxidized protein. The oxidized C127A variant exhibited the same 10-fold change in heme affinity, with K_d values of 0.031 and 0.323 μm in the oxidized and reduced states, respectively. However, the oxidized and reduced states of HO-2Δ265–316, C127A/Δ265–316, C265A, C282A C265A/C282A, and C127A/C282A, all of which lack the intramolecular disulfide bond between HRM1 and HRM2, showed nearly equal binding affinity for ferric heme. These results strongly indicate that HRM1 and HRM2 may act as a redox switch that controls ferric heme binding to HO-2. Whether or not this switch might have physiological relevance is addressed under “Discussion.”

Because the redox potential of the Fe³⁺/Fe²⁺ heme couple is near the ambient intracellular potential (see “Discussion”) and because ferrous heme is the form that binds O₂ in the HO reaction, we determined the binding affinity of the HO-2 variants for ferrous heme (Table 3). We were able to perform these titrations only with the reduced HO-2 proteins because any reductants present in the heme solution also reduce the disulfide bonds in HO-2. Surprisingly, unlike the titrations of reduced HO-2 with Fe³⁺ heme, all of the reduced HRM variants exhibited similar K_d values for ferrous heme within a range of 0.16 -0.58 μm. Thus, the redox state of the HRMs appears to affect the affinity of HO-2 for only ferric (but not ferrous) heme. The possible implications of these surprising results are described below.

Redox Changes in the HRMs Affect the Spin State of Ferric Heme

EPR measurements were performed to characterize the axial ligand for the ferric heme in oxidized (Fig. 7A) and reduced (Fig. 7B) HO-2 and the HRM variants. All of the EPR measurements were performed at 10 K because the low-spin heme relaxes very fast. The spectra of oxidized HO-2Δ289–316 indicate a predominantly six-coordinate high-spin state with g-values for the low-field feature of the high-spin component at 5.80 and for the low-spin component at 2.87,2.26, and 1.99 (Fig. 7A). These values are indicative of ligation by His and water (36), an in HO-1 (37). The EPR spectra of other oxidized variants are identical to those of oxidized HO-2Δ289–316.

Compared with the EPR spectra of oxidized HO-2, those of the DTT-reduced enzyme are more complicated (Fig. 7B). Although C265A/C282A, C265A, HO-2Δ265–316, and C127A/Δ265–316 are predominantly in the high-spin state, the EPR spectra of reduced HO-2Δ289–316 and the C127A/C282A variant exhibit a mixture of high-spin heme (g-value at 5.78) and low-spin heme (g-values at 2.41, 2.26, and 1.91). The DTT-reduced forms of the C127A and C282A variants also exhibit a large proportion of low-spin Fe³⁺ heme, with g-values at 2.41, 2.26, and 1.91. These values are indicative of thiolate/water ligation (36). Thus, when the HRMs are in the disulfide state, the low- and high-spin ferric hemes appear to have His and water (or OH) ligands (g-values at 2.87, 2.26, and 1.99), as concluded previously on the basis of resonance Raman spectroscopy (26, 38). It is only upon reduction of the HRM that the EPR spectra of the low-spin heme switch to the values characteristic of thiol ligation (g-values at 2.41, 2.26, and 1.91).