Ferrochelatase π-Helix: Implications from Examining the Role of the Conserved π-Helix Glutamates in Porphyrin Metalation and Product Release

Abstract

Protoporphyrin ferrochelatase catalyzes the insertion of Fe²⁺ into protoporphyrin IX to form heme. To determine whether a conserved, active site π-helix contributes to the translocation of the metal ion substrate to the ferrochelatase-bound porphyrin substrate, the invariant π-helix glutamates were replaced with amino acids with non-negatively charged side chains, and the kinetic mechanisms of the generated variants were examined. Analysis of yeast wild-type ferrochelatase-, E314Q- and E318Q-catalyzed reactions, under multi- and single-turnover conditions, demonstrated that the mutations of the π-helix glutamates hindered both protoporphyrin metalation and release of the metalated porphyrin, by slowing each step by approximately 30–50%. Protoporphyrin metalation occurred with an apparent pK_a of 7.3 ± 0.1, which was assigned to binding of Fe²⁺ by deprotonated Glu-314 and Glu-314-assisted Fe²⁺ insertion into the porphyrin ring. We propose that unwinding of the π-helix concomitant with the adoption of a protein open conformation positions the deprotonated Glu-314 to bind Fe²⁺ from the surface of the enzyme. Transition to the closed conformation, with π-helix winding, brings Glu-314-bound Fe²⁺ to the active site for incorporation into protoporphyrin.

Graphical abstract

INTRODUCTION

Heme is an essential cofactor in most organisms (1,2), and even parasitic organisms that lack heme-synthesizing enzymes compensate for this deficiency by acquiring heme or its precursor, protoporphyrin IX (PPIX) (3). The terminal step of heme biosynthesis in metazoans, which is catalyzed by PPIX ferrochelatase (E.C. 4.99.1.1; hereon referred to as ferrochelatase), joins the porphyrin biosynthesis and iron transport pathways (4). Cellular toxicity of the two physiological substrates, Fe²⁺ and PPIX, and high cellular demand for the heme product call for the ferrochelatase-catalyzed step to be highly regulated. Not surprisingly, ferrochelatase gene mutations can result in enzyme variants with decreased activity, as manifested in erythropoietic protoporphyria (EPP) (5–7). Salient clinical features of EPP patients include PPIX accumulation in erythroid cells in the bone marrow and painful photosensitivity (6,8). Elimination of the PPIX excess can also have taxing effects on the liver, causing liver damage and terminal failure in rare cases (9,10).

Ferrochelatases from various organisms, including human, Mus musculus (mouse) and Saccharomyces cerevisiae (yeast), and Bacillus subtilis coproporphyrin ferrochelatase have been kinetically and structurally characterized (11–16). Despite the low amino acid sequence identity amongst the human and S. cerevisiae PPIX ferrochelatases and B. subtilis coproporphyrin ferrochelatase, the crystallographic structures of these proteins show that the overall tertiary structure is conserved (17–19). The human, mouse, and yeast enzymes are homodimers (20–22), whereas B. subtilis coproporphyrin ferrochelatase is a monomer (23). However, regardless of the oligomeric state and the porphyrin substrate (PPIX vs. coproporphyrin III), two similar domains in the monomeric unit are delimited by a cleft in which the porphyrin binds. Conserved residues in this cleft have important roles in catalysis (2,14,24–27). Amongst these conserved residues are two glutamates located on the same side of a π-helix (Fig. 1), which is a secondary structural element also conserved within the chelatase family of proteins (28–30). The crystal structures of human wild-type and mutated ferrochelatase and substrate- and product-bound ferrochelatase show that the π-helix adopts two conformations: unwound and wound (31). Further, the different positioning of a large number of π-helix residue side chains in these structures helped to track the interconversion between the two conformations (27,32) and led to the proposal that unwinding of the π-helix is necessary for product release (32).

FIGURE 1. Positioning of π-helix glutamates and other active site residues in the open and closed conformation of human ferrochelatase.

Panel A. Ferrochelatase in “open” conformation with product heme bound at the active site (PBD #2QD2). Panel B. Ferrochelatase (monomer) in “closed” conformation with substrate PPIX bound at the active site (PBD #2HRE). The π-helix and glutamate residues (Glu-343 and Glu-347) examined in this study are highlighted in yellow. Phe-337/His-341 are in magenta. His-263 and Met-76/Arg-164/Tyr-165 are in dark blue. The amino acid numbering (Glu-314 and Glu-318) in panels A and B is for S. cerevisiae ferrochelatase, and it corresponds to human ferrochelatase Glu-343 and Glu-347. In our proposed model, when the π-helix unwinds, catalytic Glu-314 projects out of the matrix side of S. cerevisiae ferrochelatase, where it picks up an iron atom from and Fe²⁺ chaperone, and the winding of the π-helix brings Glu-314-bound Fe²⁺ to the active site for incorporation into PPIX. Panel C. Amino acid sequence alignment of ferrochelatase from three organisms covering π-helix amino acids pertinent to this study. The π-helix glutamates are indicated in bold yellow in dark green-shaded boxes and the secondary metal ion-binding site His and Phe are in bold magenta. FC, ferrochelatase.

Both the crystal structures of S. cerevisiae ferrochelatase complexed with Co²⁺ or Cd²⁺ (17) and those of B. subtilis coproporphyrin ferrochelatase in complex with Zn²⁺ or Cd²⁺ (33) revealed two possibly interacting metal ion-binding sites 7 Å apart. The two sites, one at the surface of the protein and the other in the porphyrin-binding cleft, are connected through the π-helix. This structural arrangement is consistent with the initial hypothesis put forward by Al-Karadaghi and coworkers (33) to describe metal ion translocation from the surface to the active site of coproporphyrin ferrochelatase. Accordingly, the entering metal ion would exchange its solvent ligands with the coordinating, π-helix glutamates, and with the progressing ligand exchange along the π-helix acidic residues, the metal ion would reach the metal ion substrate-binding site and active site. There, the final ligand exchange, the substitution of the protein ligands for the pyrrole nitrogens, would occur and thus the metal ion would be inserted into the porphyrin macrocycle. The presence of an invariant His-Glu couple (S. cerevisiae ferrochelatase His-235 and Glu-314 and B. subtilis coproporphyrin ferrochelatase His-183 and Glu-264) in the inner or metal ion substrate-binding site corroborates with the findings from biochemical, spectroscopic and kinetic studies, which indicated that these residues are essential for metal ion binding and enzyme activity (15,24,25,34–36). In fact, no Fe²⁺ was coordinated in the active site of B. subtilis coproporphyrin ferrochelatase when the invariant His-183 was mutated to Ala (25). Nonetheless, the above interpretation for Fe²⁺ shuttling and coordination in the active site, prior to its incorporation in PPIX, does not have general acceptance (15,24,25,34–36). Based on analyses of the crystal structures of human ferrochelatase and variants of the enzyme complexed with various divalent metal ions, Dailey, Lanzilotta and co-workers (16,19,37) postulated that human ferrochelatase His-263 (S. cerevisiae ferrochelatase His-235 and B. subtilis coproporphyrin ferrochelatase His-183), along with Glu-343 (S. cerevisiae ferrochelatase Glu-314 and B. subtilis coproporphyrin ferrochelatase Glu-264) function in proton abstraction from the porphyrin macrocyle and not in Fe²⁺ coordination. More recently, these investigators further proposed that solvent/water-filled channels serve as a “conduit” for the transport of Fe²⁺ from the outer surface of ferrochelatase to the active site (38). According to their model, active site pocket Arg-164 and Tyr-165 (human ferrochelatase numbering), located opposite to His-263, coordinate the iron ion before porphyrin metalation. Arg-164 would also have a catalytic role by serving as a Lewis base (16); thus, Fe²⁺ would enter the active site from the opposite face of the porphyrin at the non-conserved residues Met-76/Arg-164/Tyr-165 rather than via the conserved His-263 and Glu-343.

In agreement with the ferrochelatase crystal structures showing a second bound metal ion (Mg²⁺ or Cd²⁺) (23,33), identification of a secondary conserved metal ion-binding site (14) also provided relevant information towards the understanding of the ferrochelatase chemical mechanism. A histidine (His-341 and His-287 in human and murine ferrochelatase, respectively), a perfectly conserved phenylalanine (Phe-337 and Phe-283 in human and murine ferrochelatase, respectively) and possibly other unidentified π-helix residues coordinate the metal ion in this secondary metal ion-binding site (14). Kinetic characterization of murine ferrochelatase variants with either the coordinating His-287 or Phe-337 mutated offered an explanation for the ferrochelatase substrate inhibition being only observed at high, and non-physiological, metal ion concentrations (14) [Scheme 1]. The second metal ion-binding site residues appear to enhance ferrochelatase activity at low micromolar metal ion concentrations, but inhibit activity at higher, non-physiological concentrations (14). In our mechanistic model, one of the conserved π-helix glutamates (i.e., murine ferrochelatatase Glu-289 or S. cerevisiae ferrochelatase Glu-314) binds the iron atom in the open conformation of the PPIX-bound enzyme (E_oPP; unwound π-helix) [Scheme 1 (1) and Fig. 1A]. During the transition to the closed conformation, the conserved His-287 (in murine ferrochelatase or His-312 in yeast ferrochelatase) coordinates Fe²⁺ transiently with the conserved Phe-283 (in murine ferrochelatase or Phe-308 in yeast ferrochelatase) enhancing the binding affinity. Upon catalysis, the heme-bound enzyme (EcPr) reverts to the open conformation (E_oHeme), and then heme is released from the enzyme [Scheme 1 (1) and Fig. 1B]. However, when the metal ion (e.g., Me²⁺ = Zn²⁺, Cu²⁺, Co²⁺ and even Fe²⁺) concentration is very high and non-physiologically significant, a second metal ion binds to the product-bound enzyme (E_oPr) prior to product release, yielding a conformational complex (E_xPrMe²⁺) where product release is inhibited.

SCHEME 1.

Hunter et al. proposed that, in vivo, the secondary metal ion-binding site is a processing site for the metal ion in route to the catalytic site, and the histidine-phenylalanine pair assists in catalyzing desolvation of Fe²⁺ immediately prior to reaching the catalytic site (14). In this report, we demonstrate the role of the glutamates in the conserved π-helix in both porphyrin metalation and release of the metalated porphyrin from ferrochelatase. These findings support the proposal that the glutamate residues coordinate the metal ion as it moves along the winding π-helix and accompany the unwinding of the π-helix to allow product release.

EXPERIMENTAL PROCEDURES

Reagents

MOPS, Tween 80, sodium chloride, cobalt chloride hexahydrate, zinc chloride, nickel chloride hexahydrate, and protoporphyrin IX were from Sigma. Ferrous chloride tetrahydrate was obtained from Fisher. Blue Sepharose was from GE Healthcare. Restriction enzymes, DNA polymerase, and dNTPs were obtained from New England Biolabs. PIPBS was obtained from GFS Pharmaceuticals until its production was discontinued; subsequently, it was synthesized in our laboratory.

Synthesis of PIPBS

The buffer was prepared from a mixture of piperazine hexahydrate, 1, 4-butane-sultone, and deionized water that was refluxed and chilled according to a previously describel protocol (39).

Construction of Expression Plasmids, Overexpression and Purification of Murine and S. cerevisiae π-Helix Ferrochelatase Variants

DNA fragments with various mutations of the π-helix glutamates were independently produced using megaprimer mutagenesis (40) and either murine ferrochelatase- or yeast ferrochelatase-harboring plasmids [i.e., pGF42 (41) and pCASS/HEM15 (42), respectively] as DNA templates. The purified PCR products containing the different mutations were digested with SgrAI and BstEII (pGF42 derived) or AgeI and BamHI (pCASS-HEM15 derived), which digest the ferrochelatase cDNA upstream and downstream of the mutated codon, respectively. The purified, digested and mutated PCR products were then sub-cloned into pGF42 or pCASS/HEM15, such that the wild-type ferrochelatase sequence was replaced with the mutated one. The sequence of the subcloned DNA and the expression plasmid construction were verified by DNA sequencing (Genewiz, Inc., New Brunswick, NJ and Sanger Sequencing, University of Florida, Gainesville). Recombinant wild-type murine or yeast ferrochelatase and respective variants were overexpressed, purified, and stored as previously described (41). Protein concentrations were determined spectroscopically using the previously calculated molar absorptivities of 48,500 M⁻¹•cm⁻¹ and 56,840 M⁻¹•cm⁻¹ at 280 nm, based on the amino acid sequences of the recombinant murine (14) and yeast (13) ferrochelatases, respectively.

Buffers

For steady-state kinetics experiments performed at pH 7.00, the ferrochelatase reaction buffer was 20 mM MOPS, 0.4 M NaCl, 0.2% (v/v) Tween 80, pH 7.00. As previously described (14), for the pH variation studies, the reaction buffer was 20 mM MOPS, 20 mM PIPBS, 0.4 M sodium chloride, and 0.2% (v/v) Tween 80, at pH 5.75–9.25. Ferrochelatase reaction buffer for pre-steady-state kinetics was 200 mM MOPS, 0.8 M NaCl, 40 mM PIPBS, 0.2% (v/v) Tween 80, pH 7.00.

Preparation of Protoporphyrin IX and Metal Ion Solutions

Stock solutions of 0.1 M FeCl₂•4H₂0, ZnCl₂, CoCl₂•6H₂0, and NiCl₂•6H₂0 were prepared in a 1.2 M HCl solution and serially diluted with deionized water on the day of the experiments to 10 and 1 mM solutions (13). Stock solutions of 2.0 mM PPIX were stored at 4°C in glass screw top tubes wrapped in aluminum foil for no longer than two weeks. The concentration of the PPIX stocks were determined spectrophotometrically in 2.7 M hydrochloric acid using a molar absorptivity of 262 mM⁻¹•cm⁻¹ at 408 nm as previously described (13). Immediately before use, aliquots of the stock solutions were diluted in the reaction buffer to give the PPIX concentrations used in the enzymatic assay.

Kinetic Traces, Determination of Initial Rates and Steady-State Kinetic Parameters

The progress of the PPIX metalation reaction, defined as formation of metalated PPIX, was followed using a Shimadzu UV2401PC spectrophotometer and by continuously monitoring the change in absorbance at a specific wavelength depending on the metal ion substrate. Formation of Ni²⁺-PPIX was monitored at 562 nm, heme (or Fe²⁺-PPIX) and Co²⁺-PPIX at 407 nm, and Zn²⁺-PPIX at 417 nm, as previously described (13). Reaction mixtures contained ferrochelatase reaction buffer (20 mM MOPS, 0.4 M NaCl, 0.2% (v/v) Tween-80, pH 7.00), 10 μM PPIX, 0.2–200 μM divalent metal ion, and 0.2 μM ferrochelatase. Reaction assays were performed in a final volume of 2.00 ml in a magnetically stirred, temperature-controlled (at 30°C) quartz cuvette and were initiated by addition of metal ion. The cuvette was cleaned between reactions by sequentially rinsing it with 0.6 M HCl and deionized water. The first 15 – 30 s of the reactions were analyzed using linear regression to obtain initial rates. For the calculation of the apparent steady-state kinetic parameters ( $V_{\text{max}}^{\text{app}}$, $k_{\text{cat}}^{\text{app}}$ and $K_{\mathrm{m}}^{\text{app}}$), the initial rates were determined by varying the concentration of the metal ion substrate, while keeping that of PPIX constant. Initial rate data points were fit to either the Michaelis-Menten equation (Equation 1) or a substrate-inhibited reaction equation (Equation 2) using SigmaPlot (Systat Software Inc.), as previously described (14).

$$\mathit{\text{rate}}=\frac{V_{\mathit{\text{max}}}^{\text{app}}\left[{\text{Me}}^{2+}\right]}{K_{\mathrm{m}}^{\text{app}}+[{\text{Me}}^{2+}]}$$ (1)

$$\mathit{\text{rate}}=\frac{V_{\text{max}}^{\text{app}}\left[{\text{Me}}^{2+}\right]}{K_{\mathrm{m}}^{\text{app}}+[{\text{Me}}^{2+}]+{[{\text{Me}}^{2+}]}^2/K_{\mathrm{i}}^{\text{app}}}$$ (2)

In Eq. 1 and Eq. 2, [Me²⁺] corresponds to the divalent metal ion concentration, and $K_{\mathrm{i}}^{\text{app}}$ represents the apparent inhibitory constant under the above conditions.

Pre-steady State Burst Kinetics and Data Acquisition and Analysis

Transient kinetic measurements were performed using a KinTek rapid mixing stopped-flow spectrophotometer (SF-2001) with a 0.1 cm light path and set in fluorescence mode. The ferrochelatase-catalyzed reaction was monitored by following the decrease in the fluorescence of the PPIX substrate. Upon excitation at 407 nm, the emitted light from the PPIX substrate was detected after passing through a 520 nm-long pass filter in the photomultiplier tube. Each time course for the reactions monitored at each substrate (or enzyme) concentration represents the average of a minimum of five data sets. Reactions were performed under burst conditions, in which substrate concentration is in excess over enzyme concentration and the first few turnovers of the enzyme are measured prior to the reaction reaching the steady state. One of the loading syringes contained enzyme and PPIX in 200 mM MOPS, 0.8 M NaCl, 0.2% (v/v) Tween-80, pH 7.0 while the other syringe contained Fe²⁺in Diaion™-treated deionized water. Equal volumes of the reactants in the two loading syringes were rapidly mixed in the reaction chamber, thus yielding the final concentrations of 15 μM PPIX, 100 μM Fe²⁺, and 1 μM enzyme. The reaction temperature was maintained at 30 °C.

The time course data were fit, using the nonlinear least-squares regression analysis program Sigmaplot (Systat Software Inc.), to equation 3. Equation 3 describes a pre-steady-state exponential decay phase, with amplitude A₁ and a rate k₁, followed by a linear steady-state phase with a steady-state turnover rate k_SS (approaching k_cat at saturating substrate). Y is the observed fluorescence at time t, and C is a constant.

$$Y=A_1{e}^{-k_{1}t}+k_{\mathit{\text{SS}}}t+C$$ (3)

Single-Turnover Experiments

The experimental setting was identical to that described for the burst kinetics related experiments with exception that the reactions were run under single turnover conditions, with the enzyme concentration in excess over the substrate concentration. The selected concentrations were based on the nanomolar K_D values for the binding of PPIX to ferrochelatase and variants. One of the syringes was loaded with a solution containing 40 μM enzyme (yeast wild-type ferrochelatase, E314A or E318A variant), 30 μM PPIX in 200 mM MOPS, 40 mM PIPBS, 0.8 M NaCl, 0.2% (v/v) Tween 80 pH 7.0, while the other syringe contained 0–600 μM Fe²⁺ in Diaion™-treated deionized water. The final reactant concentrations, i.e., the concentrations of the reactants in the assay reaction chamber, were 15 μM PPIX, 0–300 μM Fe²⁺, and 20 μM enzyme (yeast wild-type ferrochelatase, E314Q or E318Q variant). The time course data were fit, using the nonlinear least-squares regression analysis program Sigmaplot (Systat Software Inc.), to equation 4. This equation describes two successive first-order phases, where A₁ and A₂ are the amplitudes and k₁ and k₂ are the rates associated with phases 1 and 2, respectively. Y is the observed fluorescence at time t, and C is a constant.

$$Y=A_1{e}^{-k_{1}t}+A_2{e}^{-k_{2}t}+C$$ (4)

The pH dependences of the observed k₁ and k₂ rates was also determined. In these experiments, the final concentrations were 10 μM enzyme, 8 μM PPIX, and 100 μM Fe²⁺. The reactions were completed at 20°C, as it was determined that the rates of PPIX consumption became too rapid to monitor at high pH (7.5 and above) in the case of yeast wild-type ferrochelatase. Time course data from reactions performed covering the pH 6.0–9.0 range were fit as described above, and the observed k₁ rates were plotted against the pH of the reaction. The k₁ rates for reactions at the different pH values were fit to equation 5, with k being the observed rate rate, A the theoretical difference in the rates associated with the protonated and deprotonated species, pK_a the pH at which the ionizing group is 50% protonated, and B the rate of reaction when the ionizing group is 100% deprotonated. The nonlinear least-squares regression analysis program Sigmaplot (Systat Software Inc.) was used to fit the data.

$$k=\frac{A}{1+{10}^{(pH-\mathit{\text{pKa}})}}+B$$ (5)

RESULTS

Activity of π-helix Ferrochelatase Variants with the Concentration of the Fe²⁺ Substrate: Steady-State Kinetic Parameters

Mammalian (13) and S. cerevisiae (13,14) wild-type ferrochelatases and B. subtilis wild-type coproporphyrin ferrochelatase (12,43) are active with a variety of different divalent metal ions, such as Fe²⁺, Co²⁺, Zn²⁺, and Ni²⁺. However, aside from Fe²⁺, the purified enzyme is subject to substrate inhibition by these metal ions when the activity of the enzyme is assayed in the absence of metal ion complexing agents (13,44). To address the question of whether the ferrochelatase π-helix glutamates have a role in binding the metal ion substrate, two π-helix glutamates (Fig. 1) were substituted with amino acids with non-negatively charged side-chains. Mutation of either of the two glutamates in the yeast ferrochelatase π-helix 314-Glu-Thr-Leu-His-Glu-Ile-319 sequence (i.e., Glu-314 and Glu-318) with alanine yielded variants (E314A and E318A) with decreased enzymatic activities and increased $K_{\mathrm{m}}^{\text{Fe}}$ values in relation to those of yeast wild-type ferrochelatase (Fig. 2). Mutation of either of the two glutamates to the isosteric glutamine, which should nullify the charge without a significant side-chain size discrepancy, resulted in variants (E314Q and E318Q) with decreased V_max and increased K_m values (Table 1 and data not shown), much like the results with the alanine variants (Table 1). Similar to the wild-type enzyme, none of the π-helix variants were Fe²⁺-substrate inhibited, and, thus, the specific activity data were fit to the Michaelis-Menten equation (Eq. 1). The apparent V_max values for E314A and E318A were approximately 40% and 70% of that of yeast wild-type ferrochelatase, respectively, and the apparent K_m values were increased 3.5- to 4-fold over that of the wild-type enzyme (Table 1).

FIGURE 2. Specific activity of yeast ferrochelatase π-helix variants vs. Fe²⁺ concentration at pH 7.00 and 30 °C.

The data points correspond to: ○, yeast wild-type ferrochelatase; □, E314A yeast ferrochelatase variant; Δ, E318A yeast ferrochelatase variant. The enzyme concentration was 0.2 μM, while PPIX was held at 10 μM. PPIX consumption was monitored by following the change in absorbance at 407 nm. The lines represent the best fit to equation 1. Specific activity is in units of μmoles PPIX consumed per minute per μmole enzyme.

Table 1.

Steady-state kinetic parameters for murine and yeast ferrochelatase variants at pH 7.0

Ferrochelatase (wild-type or variant)	Fe2+			Ni2+			Zn2+
	$V_{\text{max}}^{\text{app}}$ (min−1)	$K_{\mathrm{m}}^{\text{app}}$ (μM−1)	$K_{\mathrm{i}}^{\text{app}}$ (μM−1)	$V_{\text{max}}^{\text{app}}$ (min−1)	$K_{\mathrm{m}}^{\text{app}}$ (μM−1)	$K_{\mathrm{i}}^{\text{app}}$ (μM−1)	$V_{max}^{\text{app}}$ (min−1)	$K_{\mathrm{m}}^{\text{app}}$ (μM−1)	$K_{\mathrm{i}}^{\text{app}}$ (μM−1)	$V_{\text{max}}^{\text{app}}$ (min−1)
murine WT FCa	24.6 ± 1.5	5.2 ± 1.2	N.O.b	3.3 ± 0.2	4.6 ± 0.9	100.3 ± 18.8	2.61 ±0.26	0.12 ± 0.09	14.8 ± 3.5	1.4 ± 0.0
E289A	34.2 ± 1.6	36.2 ± 3.6	N.O.	N.O.	N.O.	N.O.	0.27 ± 0.06	7.2 ± 3.0	53.9 ± 22.0	0.06 ± 0.01
E289Q	44.4 ± 1.9	25.8 ± 3.5	N.O.	N.O.	N.O.	N.O.	0.16 ± 0.01	66.6 ± 6.9	N.O.	N.O.
E293A	39.4 ± 2.6	25.0 ± 4.1	N.O.	0.5 ± 0.1	6.2 ± 4.8	N.O.	0.71 ± 0.02	0.19 ± 0.08	343.8 ± 68.0	0.7 ± 0.0
E293Q	70.9 ± 7.0	168.7 ± 29.3	N.O.	3.4 ± 0.1	48.2 ± 2.5	N.O.	3.2 ± 0.1	0.66 ± 0.13	328.9 ± 66.4	2.06 ± 0.03
Yeast WT FCa	77.0 ± 1.3	6.7 ± 0.4	N.O.	30.2 ± 1.2	10.2 ± 0.8	113.4 ± 9.9	23.0 ± 3.0	0.19 ± 0.06	1.8 ± 0.4	14.0 ±0.4
E314A	33.0 ± 2.2	24.1 ± 4.3	N.O.	N.O.	N.O.	N.O.	0.71 ± 0.02	4.7 ± 0.4	565.1 ± 92.7	0.45 ± 0.10
E318A	54.3 ± 1.8	25.7 ± 2.2	N.O.	1.1 ± 0.3	77.0 ± 49.3	N.O.	1.4 ± 0.1	0.08 ± 0.12	68.9 ± 16.1	1.0 ± 0.11
E318Q	35.6 ± 2.7	20.2 ± 4.3	N.O.	3.9 ± 0.3	67.6 ± 10.6	N.O.	4.09 ± 0.29	0.36 ±0.14	36.1 ± 7.0	2.49 ± 0.06

^aWT FC, wild-type ferrochelatase;

^bN.O., substrate inhibition Not Observed under the assay conditions described in the Methods section.

While the specific activity of yeast wild-type ferrochelatase, with Fe²⁺ as substrate, was approximately 3-fold greater than that of the murine counterpart (13) and (Table 1), the apparent K_m values of the corresponding murine ferrochelatase π-helix variants E289A and E293A (i.e., with mutation of either of the glutamates in the π-helix 289-Glu-Thr-Leu-Tyr-Glu-Leu-294) were about 5- to 7-fold greater than that of murine wild-type ferrochelatase. The resulting 8- and 5-fold decreases in the specificity constants $(k_{\text{cat}}/K_{\mathrm{m}}^{\text{Fe}})$ of yeast E314A and murine E289A variants relative to those of the yeast and murine wild-type enzymes, respectively, indicate that the substitution of the yeast ferrochelatase π-helix Glu-314 (or Glu-289 in the murine enzyme) significantly reduces the productive Fe²⁺ substrate binding. Albeit to a lesser degree, the 5- and 3-fold reduction of the specificity constants $(k_{\text{cat}}/K_{\mathrm{m}}^{\text{Fe}})$ of the yeast E318A and murine E293A variants in relation to those of the corresponding wild-type ferrochelatases imply that productive Fe²⁺ binding is also affected in these variants. The magnitude of the reduction of the $k_{\text{cat}}/K_{\mathrm{m}}^{\text{Fe}}$ values of the yeast E318Q and murine E289Q variants over those of the corresponding wild-type ferrochelatases was similar to those observed for the yeast E318A and murine E289A variants (Table 1).

Steady-state Wild-type Ferrochelatase and π-helix Variant Ferrochelatase Activity with Non-Fe²⁺ Divalent Metal Ions

The possibility of substrate inhibition of the various yeast and murine ferrochelatase π-helix variants by Ni²⁺, Zn²⁺ and Co²⁺ was examined by varying the divalent metal ion concentration while maintaining all of the other potential variables constant. The chelatase-catalyzed reactions of all of the variants (i.e., harboring mutations in yeast ferrochelatase Glu-314 or Glu-318 and murine ferrochelatase Glu-289 or Glu-293) were investigated with Ni²⁺, and then the activities of selected variants were further examined with Zn²⁺ and Co²⁺ substrates. Figure 3 illustrates the specific activities for the yeast wild-type ferrochelatase-, E314A- and E318A-catalyzed reactions as a function of Zn²⁺ concentration. Similar to the wild-type ferrochelatase reaction, the specific activity vs. [Zn²⁺] data for the E314A- and E318A-catalyzed reactions were fit to equation 2 (Eq. 2) for a substrate-inhibited reaction. The Zn²⁺ substrate is highly inhibitory, with inhibition constants of 1.8 ± 0.4 μM and 14.8 ± 3.5 μM for the yeast and murine wild-type ferrochelatases, respectively (Table 1). Alleviation of the Zn²⁺ substrate-inhibition was observed for the E314A- and E318A-catalyzed reactions (Fig. 3), as reflected by the inhibition constants of ~565 μM and ~36 μM for E314A and E318A, respectively (Table 1). The specific activity of the π-helix variants and yeast wild-type ferrochelatase were similar at high concentrations (≥ 100 μM) of Zn²⁺ (Fig. 3); however at concentrations lower than 20 μM Zn²⁺, the E314A and E318A variants were significantly less active than the yeast wild-type enzyme (Fig. 3).

FIGURE 3. Specific activity of yeast ferrochelatase π-helix variants vs. Zn²⁺ concentration at pH 7.00 and 30 °C.

The data points correspond to: ○, yeast wild-type ferrochelatase; □, E314A yeast ferrochelatase variant; Δ, E318A yeast ferrochelatase variant. The inset shows an expanded view of the data for the E314A and E318A variants. The enzyme concentration was 0.2 μM, while PPIX was held at 10 μM. ZnPP production was monitored by following the change in absorbance at 417 nm. The lines represent the best fit to the equation 1 (E314A) or equation 2 (yeast wild-type ferrochelatase and E318A). Specific activity is in units of μmoles ZnPP produced per minute per μmole enzyme.

The enhanced chelatase activity of yeast wild-type ferrochelatase over that of the murine enzyme was even more pronounced with Ni²⁺, Zn²⁺ and Co²⁺ than when Fe²⁺ was used as the metal ion substrate. As indicated in Table 1, the apparent turnover number of yeast wild-type ferrochelatase was approximately 10-fold greater than that of the wild-type murine enzyme with Ni²⁺, Zn²⁺, or Co²⁺ substrate, while a 3-fold enhancement was observed with Fe²⁺. Yeast ferrochelatase was also more strongly substrate-inhibited by Zn²⁺ and Co²⁺ than the murine enzyme. Ni²⁺ was the least inhibitory among the three alternative metal ion substrates tested, with wild-type yeast and murine ferrochelatases having similar values for their apparent inhibitory constants (Table 1).

Murine ferrochelatase E289D had no measurable activity with Ni²⁺, Zn²⁺ or Co²⁺ (data not shown), and thus was not further characterized with these metal ions. Yeast E318Q and all other murine ferrochelatase variants, namely, E289Q, E293A, E293Q (Table 1) and E293D (data not shown)) showed similar activity substrate concentration-dependence as those of the enzymes in Figure 3, although the activity values varied. Similarly, substitution of murine ferrochelatase Glu-297 with aspartate, alanine or glutamine yielded variants without drastically different dependence of specific Ni²⁺-chelatase activity on metal ion concentration relative to that of the wild-type enzyme. Specifically, when Ni²⁺ was used as the metal ion substrate, both the apparent V_max and K_m values of E297D were approximately one-half of those of murine wild-type ferrochelatase, whereas the apparent V_max and K_m values of E297Q were 1.9- and 1.4-fold of those of the murine wild-type enzyme (data not shown). The apparent inhibitory constants of E297D and E297Q were 0.9- and 1.4-fold of that of murine wild-type ferrochelatase, respectively (data not shown).

Pre-Steady State Burst Experiments of the Wild-type Ferrochelatase and π-Helix Variant Reactions

Stopped-flow analysis of the reaction of R115L human ferrochelatase•deuteroporphyrin IX complex with Fe²⁺, under multi-turnover conditions, indicated that a burst of metalloporphyrin formation (47) occurs prior to the steady-state, and therefore a step following product formation at the active site limited the overall reaction rate. To explore the events occurring at the active site and determine whether the π-helix glutamate residues affect either or both of the catalytic mechanism steps – porphyrin metalation and product release – stopped-flow experiments were performed to follow the reactions of yeast wild-type ferrochelatase, E314Q and E318Q variants. Yeast wild-type ferrochelatase and the yeast ferrochelatase variants were chosen for pre-steady state analysis due to their increased catalytic rates in comparison to the murine counterparts. In contrast to the heme product, PPIX fluoresces by emitting light at a wavelength longer than 520 nm when excited at 407 nm. Thus, stopped-flow studies can be used to follow consumption of PPIX in the ferrochelatase active site on a centisecond time scale. In these reactions, the enzyme (2.0 μM) was preincubated with 30.0 μM PPIX, and then 200 μM Fe²⁺ was added by rapid mixing. Since PPIX is stoichiometric with metalated PPIX (e.g., heme) in the ferrochelatase-catalyzed reaction, we assumed that metalated PPIX formation paralleled PPIX consumption. A pre-steady-state burst of product was observed in the reactions catalyzed by the yeast wild-type and variant ferrochelatases (Fig. 4). Thus, the time course data fit to a burst equation (Eq. 3). The presence of a product burst suggests that a step other than chemistry (presumably, product release) is at least partially rate-limiting (48). The time course for the yeast wild-type ferrochelatase reaction, described by a burst phase with a rate of 24.0 ± 3.0 s⁻¹ and a steady-state phase rate of 0.22 ± 0.01 s⁻¹ (Table 2), indicated that the initial observed rate constants and the steady-state rates were reduced by approximately 30–50% by the E314Q and E318Q mutations. These data suggest that substitution of either Glu-314 or Glu-318 with glutamine has modest but measurable deleterious effects on both the reaction chemistry and release of the product.

FIGURE 4. Pre-steady state kinetics of the reactions catalyzed by yeast ferrochelatase and π-helix variants.

Time courses of the reactions of yeast wild-type ferrochelatase (○), E314Q (□), and E318Q (Δ) under multi-turnover conditions at pH 7.00, 30 °C. The concentrations in the observation chamber following mixing were: 1 μM enzyme, 15 μM PPIX and 100 μM Fe²⁺. PPIX consumption was monitored by following the decrease in porphyrin fluorescence (λ_exc = 407 nm and λ_em > 520 nm). Lines represent the best fit to the equation 3.

Table 2. Yeast Wild-Type Ferrochelatase and π-Helix Variant Pre-Steady-State (Burst) Kinetics under Multi-turnover Conditions^a.

Porphyrin Metalation (Fe²⁺) and Product Release Rates for Yeast Wild-Type Ferrochelatase and π-Helix Variant Reactions

	Fe2+
	k1 (s−1)	kss (s−1)
Yeast WT FCb	24.0 ± 3.0	0.22 ± 0.01
E314Q	14.0 ± 2.0	0.11 ± 0.01
E318Q	16.0 ± 1.0	0.14 ± 0.01

^aThe data were fit to an exponential decay plus linear equation (Eq. 3), and the conditions used (e.g., pH 7.0 and 30 °C) were as summarized in the legend for Figure 4;

^bWT FC, wild-type ferrochelatase.

Pre-steady State Kinetics of the Wild-type Ferrochelatase and π-Helix Variant Reactions: Single-Turnover Conditions

Analysis of multi-turnover, particularly burst, kinetics can be complex due to the limitations resulting from, for example, substrate inhibition, which reduce the amplitude of the pre-steady state burst. Hence, we sought to examine more closely the events occurring at the yeast wild-type ferrochelatase and π-helix variant active sites. We performed stopped-flow experiments with enzyme concentration in excess over that of the substrate, i.e., under single-turnover conditions. The observation that porphyrin binding by ferrochelatase results in a decrease in porphyrin fluorescence allowed us to determine that the K_D value for the binding of PPIX to ferrochelatase is in the nanomolar range (data not shown). We set the final enzyme (yeast wild-type ferrochelatase, E314Q, or E318Q) and PPIX concentrations at 20 μM and 15 μM, respectively, such that under these conditions the amount of unbound PPIX was negligible, and we could safely assume that all PPIX was bound to yeast ferrochelatase. The final Fe²⁺ concentration, at 100 μM, was 6.7 (~7)-fold greater than that of the yeast ferrochelatase-PPIX complex, and thus single turnover conditions were observed. In addition, we could assume that the amount of Fe²⁺ consumed during the single-turnover reaction (1/7 ~14%) was also negligible and it could be assumed constant. The kinetic profiles for the reactions of the yeast wild-type ferrochelatase and π-helix variants are shown in Figure 5. The kinetic profiles for the reactions of yeast wild-type ferrochelatase and variants (E314Q and E318Q) were best described by a two-step kinetic mechanism, represented by equation 4 (Fig. 5). We ascribed the two steps of the kinetic mechanism of yeast wild-type ferrochelatase and E318Q variant to PPIX metalation (or heme formation) and metalated PPIX (heme) release from the enzyme. Comparison of the time courses for the reactions of the E314Q and E138Q variants with those of yeast wild-type ferrochelatase clearly indicate that the variants catalyze PPIX metalation at significantly lower rates (~11.0.s⁻¹ for E314Q and E138Q) than the wild-type enzyme (65.1 s⁻¹) (Figs. 5B and C vs. Fig. 5A; Table 3). The E314Q and E138Q mutations also caused a decrease on the product release rate (Table 3).

FIGURE 5. Single-turnover kinetics of the reactions catalyzed by yeast ferrochelatase and π-helix variants.

Time courses for the reactions of A, yeast wild-type ferrochelatase, B, E314Q and C, E318Q under single turnover conditions at pH 7.00 and 30 °C. The reactions were initiated by mixing enzyme (20 μM) with PPIX (15 μM) and Fe²⁺ (100 μM). [The concentration in parentheses are the final concentrations, i.e., those after mixing.] PPIX consumption was monitored by following the decrease in porphyrin fluorescence (λ_exc = 407 nm and λ_em > 520 nm). The curves were calculated by fitting the data to equation 4.

Table 3. Summary of Single Turnover Results from Pre-Steady-State Kinetic Experiments^a.

Porphyrin Metalation (Fe²⁺) and Product Release Rates for Yeast Wild-Type Ferrochelatase and π-Helix Variant Reactions

	Fe2+
	k1 (s−1)	k2 (s−1)
Yeast WT FCb	65.1 ± 0.8	8.1 ± 0.2
E314Q	11.0 ± 1.0	0.11 ± 0.01
E318Q	11.3 ± 0.5	1.37 ± 0.01

^aThe data were fit to a two-exponential equation (Eq. 4), and the conditions used (e.g., pH 7.0 and 30 °C) were as summarized in the legend for Figure 5;

^bWT FC, wild-type ferrochelatase.

The pH-dependence of the single turnover reaction kinetics, at 20 °C, for yeast wild-type ferrochelatase, E318Q and E314Q variants was investigated (Fig. 6). These studies indicated that pH changes affected substantially the kinetic data for the reactions of wild-type and E318Q enzymes (Figs. 6A and C), but not those of the E314Q variant (Fig. 6B). While the first kinetic step, which was also associated with the largest porphyrin fluorescence quenching signal, was clearly pH-dependent, the second step of PPIX decay was pH-independent over the range examined (Fig. 6A inset). The single-turnover pH-dependence of the first kinetic step of the yeast wild-type ferrochelatase reaction, plotted in the inset of Fig. 6A, could be accurately described by equation 5 for an acid-catalyzed reaction. A fit of the data to equation 5 resulted in an apparent kinetic pK_a of 7.3 ± 0.1. Similarly, the best fit of the single-turnover pH-dependence data for the first step of PPIX decay in the E318Q reaction (Fig. 6C inset) to equation 5 indicated that it occurred in an acid-catalyzed reaction with an apparent pK_a of 8.2 ± 0.1.

FIGURE 6. pH-dependence of single-turnover observed rates for porphyrin metalation as catalyzed by yeast ferrochelatase and π-helix variants at 20 °C.

Representative time course data with overlaid best fit curves for the reactions of A, yeast wild-type ferrochelatase, B, E314Q and C, E318Q at pH 7.00. In the insets A and C, the observed rates k₁ and k₂, obtained from fitting the PPIX consumption data in A and C, to equation 4, are plotted as function of pH: A, yeast wild-type ferrochelatase and C, E318Q. The observed rates k₁ for the first step of PPIX decay (○) were fit to equation 5, resulting in apparent pK_a values of A, 7.3 ± 0.1 for yeast wild-type ferrochelatase and C, 8.2 ± 0.1 for E318Q. The observed rates k₂ for the second step of PPIX decay (•) were pH-independent for A, yeast wild-type ferrochelatase and C, E318Q. B, time course data for the E314Q single-turnover with overlaid best fit curve derived from fitting the data to equation 4, and in the inset B, the nearly negligible observed rates (k₁and k₂), determined from data fitting in B, are shown to be pH-independent.

DISCUSSION

Fe²⁺ is the physiological substrate of ferrochelatase. How this oxidatively unstable substrate is delivered to ferrochelatase and translocated through the protein until it reaches the active site for incorporation into the PPIX substrate remain unresolved questions. Mitoferrin (49–52) and frataxin (2,53–56), a mitochondrial inner membrane Fe²⁺ transporter and a matrix Fe²⁺ chaperone, respectively, have been independently postulated to provide Fe²⁺ directly to ferrochelatase (49,53,56,57). Similarly, the route that the Fe²⁺ substrate takes from the surface of ferrochelatase to the active site of the enzyme is not unanimously accepted.

Taking into account the conformational changes undergone by ferrochelatase, the unwinding of the π-helix (31,32) and the presence of a conserved secondary metal ion-binding site in ferrochelatase (14,44), we (1,2,4,14) and others (17,28,33) previously proposed that the acidic residue-rich π-helix offers a suitable path for Fe²⁺ translocation to the active site. In this report, we demonstrate that 1) the π-helix glutamates are critical ligands of the Fe²⁺ substrate and 2) the first kinetic step of PPIX decay involves proton transfer, in line with a progressive exchange of Fe²⁺ ligands, including the π-helix glutamates.

Shuttling of the Fe²⁺ substrate along the π-helix is consistent with not only its conformational changes (31,58) and presence of a conserved secondary metal ion-binding site (13,14,44), but also with the π-helix unwinding during Fe²⁺ acquisition (14) and porphyrin-induced protein conformational change (59). Upon PPIX binding to ferrochelatase, the PPIX macrocycle is distorted, adopting a saddling deformation (26,60,61), which facilitates deprotonation of two pyrrole ring nitrogens and exposure of the pyrrole nitrogens to the incoming divalent metal ion (62,63). In addition to the porphyrin distortion, the protein undergoes a conformational transition, during which the cleft is widened, in a typical ‘induced-fit’ reaction mechanism (58,59). It is precisely the π-helix, that by unwinding and winding (Fig. 1), assumes the major conformational change within the protein, as revealed in the crystallographic structures of human ferrochelatase with either bound product (“open conformation”) or bound substrate (“closed conformation”) (31). We suggested that with the unwinding of the π-helix, Glu-314 (in yeast ferrochelatase or Glu-289 in murine ferrochelatase; Fig. 1C) projects into the mitochondrial matrix accepting the metal ion substrate from a Fe²⁺ donor (Fig. 1A). Then, with the winding of the π-helix, Fe²⁺ can be delivered directly to the active site (Fig. 1B). The results from this study indicate that substitution of yeast ferrochelatase Glu-314 or murine ferrochelatase Glu-289 with alanine or glutamine decreased the specificity constant of the mutated enzymes towards Fe²⁺ and hindered the porphyrin metalation step in a pH-dependent manner, thus supporting a role for this π-helix glutamate in insertion of Fe²⁺ into PPIX.

Within the beginning of the π-helix, at the base of the active site cleft, resides one of the ligands of a secondary metal ion-binding site, a well-conserved histidine (i.e., human ferrochelatase His-341; murine ferrochelatase His-287; S. cerevisiae ferrochelatase His-312; Fig. 1). Four amino acids towards the amino-terminus from this residue is a perfectly conserved phenylalanine. Cadmium-soaked crystals of B. subtilis coproporphyrin ferrochelatase revealed that the inhibitory cadmium ion bound to both this histidine and the active site, and catalytically relevant, histidine-glutamate pair, with a distance of 9 Å between the two metal atoms (33). The secondary metal ion-binding site, with the conserved His (i.e., human ferrochelatase His-341) functioning together with the perfectly conserved phenylalanine (i.e., human ferrochelatase Phe-337; Fig. 1C), and possibly other active site residues adjacent to the π-helix, prevents oxidation and promotes desolvation of Fe²⁺ in transit to the catalytic site. Further, this secondary metal ion-binding site mediates, at least in part, the metal ion substrate inhibition characteristic of ferrochelatase with non-physiologically, high substrate concentrations (2,13,14). Mutations E318A and E314A attenuated the strong Zn²⁺ substrate inhibition of yeast ferrochelatase by 40-fold and >2 orders of magnitude, respectively (Table 1). However, and in contrast to the His-287 and Phe-283 mutations of the secondary metal ion-binding site in murine ferrochelatase, substitution of the π-helix glutamates of the yeast and murine enzymes decreased the apparent V_max values, and not only increased the apparent K_m values (Table 1 and Fig. 3). Because Glu-293 in the π-helix of murine ferrochelatase is the closest conserved residue to the secondary metal ion-binding site capable of complexing a metal ion, Glu-293 has been proposed to be involved in metal ion-binding at this site (14). The findings presented here support this hypothesis. First, mutation of Glu-293 resulted in a ferrochelatase variant with an increased K_i and an increased K_m value for Zn²⁺. Second, the increase in both the K_m and V_max parameters when the murine variant uses Fe²⁺ as substrate indicates that while the enzyme has a lower affinity for the substrate, it also has a lower affinity for the product, and thus once substrate is bound, the product is more rapidly released. Together, these results corroborate the mechanistic model describing the metal ion substrate inhibition of ferrochelatase (14), according to which a second metal ion binds to the enzyme•product complex before the product is released (13,14) [Scheme 1 (2)]. Further, the results support the postulated function of the secondary metal ion-binding site as a processing site that enhances metal ion binding affinity at the primary catalytic site, as the reactions of E293 variants proceed at slower rates with lower concentrations of metal ion substrate.

A pre-steady state burst in PPIX consumption when yeast wild-type ferrochelatase was reacted with PPIX and Fe²⁺ indicates that a step following binding of substrates and formation of heme is rate-determining. Specifically, the 109-fold greater rate of chemical transformation (~24 s⁻¹) over the steady-state rate (Table 2) suggests that a step after the chemical transformation, ascribed to product release, limits the ferrochelatase reaction. Similar assignments were previously reported for the reaction of human ferrochelatase (R115L) with Fe²⁺ and deuteroporphyrin (47) and determined for the reaction of murine ferrochelatase with Fe²⁺ and PPIX (data not shown). While product dissociation remains limiting in the E314Q- and E318Q-catalyzed reactions, the introduced π-helix glutamate mutations slowed both the metal insertion and product release steps by between 30–50% (Table 2). Significantly, the rate associated with the second and slow step was pH-independent. Presumably, the pH-independent second step, which seemingly is not a chemical step, reflects the transition of the π-helix to the unwound conformation enabling dissociation of the metalated PPIX from the enzyme. The crystal structures further revealed that the π-helix of human ferrochelatase is unwound in product-bound structures (31), re-enforcing that this conformational change of ferrochelatase is associated with product release. However, unlike the second step, the first and faster step of PPIX consumption observed during the yeast wild-type ferrochelatase-and E318Q-catalyzed reactions, under single turnover conditions, was dependent on the pH value. This step, which was assigned to Fe²⁺ chelation by PPIX, occurred with an apparent kinetic pK_a of 7.3 ± 0.1 and 8.2 ± 0.1 for the yeast wild-type the yeast wild-type enzyme- and E318Q-catalyzed reactions. These data corroborate the conclusions drawn from the pK_a calculations and molecular dynamic simulations performed using human ferrochelatase (64): in the open conformation of ferrochelatase the equivalent, conserved Glu (i.e., yeast ferrochelatase Glu-314 and murine ferrochelatase Glu-289) is deprotonated, while in the closed conformation it remains protonated. In addition, based on the crystal structures of product- and substrate-bound ferrochelatase (31,32), we propose that in the open conformation (unwound π-helix), the side chain of this conserved Glu moves approximately 15Ǻ while rotating 180°, such that it ends up projecting from the surface of the enzyme and becoming exposed in the mitochondrial matrix. We have hypothesized that then the conserved Glu seizes Fe²⁺ from a mitochondrial Fe²⁺ chaperone or transporter and retains the Fe²⁺ atom bound during its translocation into the active site for the next round of catalysis (2). Essentially, the Fe²⁺ translocation from the surface of the enzyme into the active site would accompany the winding of the π-helix (2,14). However, the ferrochelatase region that we postulate to project from the enzyme has been proposed to be buried in the mitochondrial membrane by other investigators (38), who have attributed the open conformation of ferrochelatase solely to the “product release conformation”. This view on the role of the ferrochelatase π-helix in relation to the protein conformational states and enzyme function is distinct from ours, as we propose that both binding of Fe²⁺ substrate and product release are associated with the transition to the open conformation of the enzyme.

In summary, using steady- and pre-steady-state kinetics, we demonstrated here that the two ferrochelatase π-helix glutamate residues contribute to PPIX metalation and release of the metaled PPIX product. We propose that the binding of PPIX to ferrochelatase triggers unwinding of the π-helix resulting in the projection and exposure of the deprotonated yeast ferrochelatase Glu-314 carboxyl side-chain at the enzyme surface and the subsequent grasping of a Fe²⁺ atom from a mitochondrial Fe²⁺ donor (e.g., frataxin, mitoferrin). Transition from the open to the closed conformation is concomitant with the repositioning of Fe²⁺-bound Glu-314 into the active site for incorporation in the PPIX ring. Following the chemical step, the product-bound enzyme reverts to the open conformation to allow the release of the metalated PPIX product.

ABBREVIATIONS

MOPS

3-[N-morpholino]propanesulfonic acid

PIPBS

piperazine-N,N’-bis(4-butanesulfonic acid), PPIX, protoporphyrin IX