The Role of Conserved Elements in an Active Site α-Helix of Coproheme Decarboxylase

Abstract

The final step in the coproporphyrin-dependent (CPD) branch of the heme biosynthesis pathway involves the oxidative decarboxylation of coproheme to form heme b. This reaction, catalyzed by coproheme decarboxylase (ChdC), requires two equivalents of hydrogen peroxide to complete the synthesis of one b-type heme molecule. The CPD pathway is limited to Gram-positive bacteria and some archaea, and the precise mechanism of ChdC differs between Firmicutes and Actinobacteria. These variations highlight the importance of studying ChdCs from diverse organisms. The reaction proceeds through two sequential oxidative decarboxylations via the intermediate monovinyl monopropionate deuteroheme (MMD). Previous studies suggest that MMD does not exit the active site but instead undergoes a 90-degree rotation before another equivalent of hydrogen peroxide binds and initiates the second oxidative decarboxylation. This mechanism requires a high degree of specificity to distinguish between substrate, intermediate, and final product. To further understand this selectivity, we present biochemical and structural analyses of wild-type and variant forms of ChdC from Streptomyces coelicolor (ScChdC). We hypothesize that a conserved active site element within an alpha helix contributes to porphyrin specificity/selectivity and conformation and investigate how this influences an active site loop. Our data provides new insight into the role of this loop in substrate recognition, rotation, and catalysis. The substrate selectivity model for ChdC developed in this study will inform future mechanistic investigations and provide insights into key functional interfaces, highlighting potential targets for drug development.

Graphical Abstract

Conformational changes in the active site loop of Actinobacteria Coproheme Decarboxylase

Introduction

Heme is an essential cofactor conserved across all domains of life, and its synthesis is vital for metabolism. The knockout of heme biosynthesis is lethal for most organisms (1), highlighting its critical role. There are three known pathways for heme synthesis (2), with the coproporphyrin-dependent (CPD) pathway identified in several Gram-positive pathogens. This pathway represents an evolutionary transition from the siroheme-dependent (SHD) pathway in archaea to the canonical heme biosynthesis pathway (protoporphyrin dependent pathway- PPD) found in eukaryotes and Gram-negative bacteria (3). Given the exclusive occurrence of CPD pathway enzymes in Gram-positive bacteria (4), many of which are pathogenic, a detailed atomic understanding of their enzymatic mechanisms and regulation is desirable.

The final step of the CPD pathway involves the oxidative decarboxylation of coproheme to form heme b, catalyzed by coproheme decarboxylases (ChdCs) (2). Biochemical and structural studies have shown that ChdC enzymes function as homopentamers and share structural similarity with chlorite dismutase and dye-decolorizing peroxidases (5, 6). The reaction is hydrogen peroxide-dependent, and, like many peroxidases, the iron atom of the substrate is coordinated by a single conserved histidine residue (7). A unique feature of the CPD pathway is the hypothesis that the hydrogen peroxide required for ChdC activity is produced by coproporphyrinogen oxidase (CgoX), an upstream enzyme that oxidizes coproporphyrinogen III to coproporphyrin III while generating hydrogen peroxide as a byproduct (2),(8). Coproporphyrin ferrochelatase (CpfC) subsequently inserts iron into coproporphyrin III to form coproheme, the substrate for ChdC (9). It is plausible that both hydrogen peroxide (from CgoX) and coproheme (from CpfC) are immediately consumed by ChdC, suggesting that in vivo enzymatic activity may not conform to Michaelis-Menten kinetics. Indeed, excess H₂O₂ has been shown to cause irreversible damage to ChdC and bleaching of the heme product (10–12). Furthermore, heme is hydrophobic and poorly soluble in water, supporting a mechanism where product release could be the rate limiting step. Regardless, both mechanistic attributes make performing steady-state kinetic analysis impossible.

Mechanistically, ChdC facilitates oxygen-oxygen bond cleavage, water loss, and the formation of a porphyrin cation radical (compound 1) (13) (Figure 1). As in other peroxidases, proton transfer is essential to catalysis, though the specific amino acids involved vary (14). In Actinobacteria, such as the ChdC studied here, a conserved histidine acts as a distal base in hydrogen peroxide heterolysis (15). This histidine deprotonates hydrogen peroxide and donates a proton to the departing water molecule, forming an oxyferryl porphyrin cation radical (compound I). The radical then abstracts a hydrogen atom from a strictly conserved tyrosine residue near the propionate at position 2 (p2) of coproheme (Figure 1), generating a tyrosyl radical (16). This tyrosyl radical extracts a hydrogen atom from the p2 propionate, forming a secondary carbon radical. The radical re-enters the porphyrin ring, reducing Fe (IV)=O to Fe (III)-OH and leaving behind an unstable carbocation. The carboxylate carbon-alpha carbon bond then cleaves, releasing carbon dioxide and forming a vinyl group, yielding the intermediate-monovinyl, monopropionate, deuteroheme (MMD) (17).

Figure 1.

Proposed reaction mechanism for ChdC enzymes, with amino acid numbering for Streptomyces coelicolor (ScChdC).

For the next oxidative decarboxylation to occur, the MMD intermediate must be repositioned and previous studies suggest that the MMD remains in the active site, rotating clockwise by 90°(12). Evidence for this rotation includes observations that active site variants can lead to MMD accumulation (10, 12, 18). Moreover, in several cases, crystal structures confirm a 90° rotation (6, 19) positioning propionate four (P₄, Figure 1) adjacent to the catalytic tyrosine. Additionally, computational studies of a firmicute ChdC have shown that the energy barrier to rotation is much lower than a release/rebinding mechanism, and that specific residues form a “gate” to facilitate the clockwise rotation of MMD, preventing a counterclockwise rotation (20). Following rotation, hydrogen peroxide binding to MMD initiates a second catalytic cycle, decarboxylating p4 to yield the final product, heme b (13). Given that two different substrates undergo the same oxidative decarboxylation reaction, a central tenant of the mechanism is that the enzyme must catalyze turnover of coproheme and MMD but not heme b. Supporting this tenant is the fact that a recent Cryo-EM structure of ChdC with heme b bound (PDB ID 8QUO) demonstrates no further rotation of the substrate following the second oxidative decarboxylation (21). This implies a selective active site that modulates binding affinity to either promote substrate binding, product release or, in the case of MMD, reposition the intermediate for another oxidative decarboxylation. Once coproheme is bound, compound I formation and generation of the catalytic tyrosyl radical occurs quickly (22), suggesting that MMD rotation could be rate-limiting, though experimental evidence is lacking. Therefore, the multiple-step mechanism of ChdC necessitates that the active site can distinguish between the substrate, intermediate, and product, while initiating compound I formation for two of the metalloporphyrins, but not heme b.

We hypothesize that the structural selectivity described above arises from conserved elements in the active site. Specifically, conserved residues located in an active site loop and α-helix (Figure 2). The loop provides the acid/base chemistry required for compound I formation (H128 in ScChdC) (17) and has been proposed to act as a gate for substrate transfer from CpfC (23), though a substrate-bound, open-loop conformation has not been observed. The structurally conserved active site helix, located beneath the porphyrin ring, has been implicated in porphyrin selectivity in other enzymes (24, 25). This helix also contains key residues, including the substrate’s proximal ligand (H168) and a strictly conserved glycine (G169) (Figure 2). While a glycine residue in an α-helix is not unprecedented, it will introduce flexibility (26) and create a helix notch that accommodates the edge of the porphyrin ring, locking reactants in place, facilitating rotation of the intermediate, and maintaining the proper orientation of the porphyrin. The conservation of these residues across ChdCs (Figure 2) underscores their functional significance and a potential role in modulating porphyrin conformation and interactions with the active site loop, which has been hypothesized to confer selectivity (21).

Figure 2. Sequence alignments for ChdC enzymes with available structures, highlighting strictly conserved residues in the active site alpha helix and loop.

Identical residues are colored red and similar residues, across all sequences, are colored blue. Conserved histidines, which serve as a proximal ligand to substrate and the subsequent glycine residue are indicated in red in the top panel, and the conserved histidine which serves as a distal base to deprotonate hydrogen peroxide in the loop is indicated with a *.

To investigate this hypothesis and address whether or not the positioning/conformation of the porphyrin (for substrate and intermediate) influences loop structure, we have characterized ChdC from Streptomyces coelicolor (ScChdC), a Gram-positive Actinobacteria typically used to produce clinical antibiotics and secondary metabolites (27). In this work we directly measured compound I formation in ScChdC and provide additional biochemical and biophysical data for an Actinobacterial ChdC. Using site-directed mutagenesis, we crowded the heme binding site by introducing aliphatic side chains into the notch formed by G169. This was done with either a helix-stabilizing amino acid (leucine) or a helix destabilizing amino acid (valine). The iron ligand was also removed through a H168C variant. All variants were functional, but the mutations at G169 had a significant influence on the orientation of the porphyrin ring, with notable influence on the kinetic rate, active site stability, and conformation of the active site loop. We also identify residues at the back of the active site that provide additional selectivity through hydrogen bonding, allowing seamless bond switching as the MMD intermediate rotates.

These findings provide evidence that the positioning/conformation of the porphyrin and active site loop are indeed coupled. These data add new information to the structural framework for understanding the mechanism of ChdC enzymes. Our model describes key structural elements involved in substrate coordination, MMD rotation, and product release, offering insights into porphyrin selectivity and potential drug targets for pathogenic gram-positive bacteria.

Experimental

Enzyme expression and protein purification.

The gene encoding the wild-type ScChdC enzyme was cloned into the pTrcHisA (Invitrogen, Carlsbad CA) constitutive expression vector, as previously described.(28) The carbenicillin-resistant plasmid utilizes the tryptophan repressor to regulate gene transcription and was transformed into E. coli JM109. A 20 mL starter culture of LB medium was inoculated with a single colony and incubated overnight at 37 °C with shaking at 200 rpm. The overnight culture was then used to inoculate 1L of terrific broth in 2.8 L baffled flasks. The cultures were grown at 37 °C until reaching an optical density at 600 nm (OD₆₀₀) of 0.6. At this point, isopropyl B-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.2mM, the temperature was lowered to 18 °C, and the incubation was continued for 12-16 hours. Cells were harvested by centrifugation, flash-frozen and stored at −80 °C. Frozen cell pellets were resuspended at a ratio of 1g/3mL in 50 mM HEPES, pH 7.5, supplemented with 10 mg each of DNAase and lysozyme. A Pierce Protease Inhibitor tablet (Thermo Scientific, Rockford IL) was also added. Cells were lysed using a French Press, and lysates were cleared by ultracentrifugation at 100,000 x g for one hour. Resulting lysates were applied to an equilibrated cobalt chelation resin (G-Biosciences), and the column was washed with three column volumes of 50 mM HEPES, pH 7.5. Purified protein was eluted using an imidazole step gradient (50, 100, 150, 200 mM) and eluted fractions were analyzed by SDS-PAGE to confirm the presence and purity of the protein. Fractions containing the enzyme were concentrated to 10 mL using Amicon Ultra-15 centrifugal filters and further purified by size-exclusion chromatography on a Sephadex S200 column (2.5x150 cm) equilibrated with 50 mM HEPES, 150 mM KCl, pH 7.5. Purified protein was dialyzed overnight in 50 mM HEPES, pH 7.5 then concentrated again before being frozen and stored at −80 °C.

Enzyme assays and stopped-flow spectroscopy.

ScChdC was assayed as previously described (18), and the relative amounts of coproheme, MMD, and heme b were determined using high-performance liquid chromatography (HPLC). Coproheme binding to the wild-type and variant ScChdC’s was assessed using UV-Visible spectroscopy (Figure S1). For the activity assay, a 1 mL reaction mixture containing equimolar amounts of protein and coproheme was prepared (10 μM), followed by the addition of three equivalents of H₂O₂ to initiate the reaction. UV-Visible spectral data was collected for each sample over a 15-minute period. For HPLC analysis, turnover assays were allowed to proceed for 5 minutes before being quenched with concentrated HCl. Porphyrins were extracted using ethyl acetate, resolubilized in triethylamine (TEA), and applied to a C18 reverse-phase column. This approach enabled visualization of the relative amounts of coproheme, MMD, and heme b present after each 5-minute turnover reaction. No additional intermediate or product was observed if the reactions were allowed to continue beyond 5 minutes. Stopped-flow spectroscopy was conducted in 0.1 M triethanolamine-HCl (pH 8) using either an OLIS RSM 1000 rapid-scanning stopped-flow spectrophotometer for scanning experiments or an Applied Photophysics SX-20 instrument for single-wavelength experiments. The multiwavelength experiment to assess the Soret hypochromicity, associated with rapid compound I formation and decay, is included in the supplementary information (Figure S2). The Applied Photophysics instrument was connected to a refrigerated water bath for temperature control and, unless otherwise stated, experiments were performed at 25 ° C. Single-wavelength kinetic data were collected at 390 nm, with each dataset containing 1000 time points. Data were analyzed using IgorPRO software and fit to one of two equations that reflect the relaxation kinetics for two rapidly occurring processes. For Equations 1 & 2 below, Y represents the absorbance at 390 nm and tau (τ) represents the relaxation time for each process. Therefore, the rate constant for each process (k) is equivalent to the reciprocal of tau (k = 1/τ). Equation 1 describes two exponential processes where one process is increasing the absorbance, and the other is decreasing the observed absorbance. Equation 2 accounts for two exponential processes coming to a plateau with no decrease in the absorption over the observed time course. To better compare transient kinetic traces, data were adjusted such that at time zero the absorption at 390 nm was set to zero. Raw kinetic traces are shown in the Supplementary Information, Figure S3.

$$Y={\text{Abs}}_{\text{o}(390\mathrm{nm})}+A\cdot \left\{-\mathit{exp}\left[\frac{-\left(x-x_{\mathrm{o}}\right)}{{\tau }_1}\right]+\mathit{exp}\left[\frac{-\left(x-x_0\right)}{{\tau }_2}\right]\right\}$$ Equation 1.

$$Y={\mathrm{Abs}}_{\text{o}(390\mathrm{nm})}+A_1\mathit{exp}\left\{\frac{-\left(x-x_{\text{o}}\right)}{{\tau }_1}\right\}+A_2\mathit{exp}\left\{\frac{-\left(x-x_{\text{o}}\right)}{{\tau }_2}\right\}$$ Equation 2.

Rate constants are therefore equal to the reciprocal of tau, specifically, k₁ = 1/τ₁ and k₂ = 1/τ₂.

Site-directed mutagenesis.

Variants of the ScChdC enzyme were prepared as previously described (28). Briefly, site-directed mutagenesis was performed using the New England Biolabs (NEB) Q5 site directed mutagenesis protocol. The primers and corresponding melting temperatures for each variant are listed below.

Variant	Primer		Tm
H168C	Forward	GCTCGCCGACtgcGGCAAGATGGCCC	72
	Reverse	ATGCGGCGGCGGTCCTCG
G169V	Forward	CGCCGACCACgtgAAGATGGCCCGCG	72
	Reverse	AGCATGCGGCGGCGGTCC
G169L	Forward	CGCCGACCACctgAAGATGGCCCGCGG	72
	Reverse	AGCATGCGGCGGCGGTCC

Wild-type and variant ScChdC DNA was subsequently transformed into BL21 (DE3) competent E. coli cells and expressed as described above and in Dailey et al. (28).

EPR spectroscopy.

X-Band electron paramagnetic resonance (EPR) spectroscopy was used to investigate differences in the electronic environment of the heme iron atom and to detect the presence of a tyrosyl radical in the active site. To generate pre-turnover samples, purified proteins (500 uM) were incubated at a 1:1 ratio with coproheme. The coproheme solution was prepared by dissolving in a small volume of 0.5 M NaOH, followed by dilution to the desired concentration in 20 mM Tris buffer (pH 8). EPR samples were prepared under strictly anaerobic conditions in a controlled atmosphere of 5% hydrogen (balance nitrogen). Buffer, protein samples, and ligands were degassed with oxygen-free argon on a vacuum manifold before use. Sample mixtures were transferred to 4 mm quartz EPR tubes (Wilmad, Vineland NJ) and immediately flash-frozen in liquid nitrogen upon removal from the anaerobic chamber. EPR spectra were acquired using a Bruker Elexys system. Data visualization and analysis were performed using IgorPRO software. Pre-turnover samples were measured at 12, 14, and 15K with a 3800G sweep width, centered at 2100 G, using 9.35 GHz microwave frequency, 2 mW power, 20 dB power attenuation, a field modulation amplitude of 4 G, modulation frequency of 100 kHz, time constant of 0.01 ms, and a 60 second sweep time.

To capture intermediates during turnover, samples were prepared as described above, this time with 250 uM of each protein and a 1:1 molar ratio with coproheme. Immediately before freezing, two equivalents (500 μM) of H2O2 were injected directly into the EPR tubes containing pre-formed complexes, mixed for ~0.5 seconds, and frozen using an isopentane slurry. Data for the iron in the initiated complexes was collected at 12K and 15K, with the same parameters as listed above. To capture formation of tyrosyl radical, the same parameters were maintained, except samples were recorded at 77K, magnetic field range of 200 Gauss, center at 3350 G, and modulation amplitude of 1G. EPR data was analyzed with IgorPRO, which was used to plot traces as well as calculate g-values.

X-ray crystallography, Data collection, and Refinement.

Initial sitting-drop crystallization screening was performed at 18 ° C using a 96-well Intelli-Plate (Hampton Research, Aliso Viejo CA). Each well contained 100 μL of reservoir solution, and crystallization drops were prepared by mixing 2 μL protein with 2 μL of reservoir solution before sealing the tray with clear tape. Optimization of crystallization conditions was performed using hanging drop vapor-diffusion on siliconized cover slips with 800 μL of well solution. The ratio and volume of protein to well solution remained the same as the initial screening. Once promising conditions were identified, the wild-type ScChdC crystallization was further optimized, yielding the highest diffraction-quality crystals in a mother liquor containing 0.075-0.1M betaine, 0.1-0.15 M KCl, 0.1 M HEPES (pH 7.5), and 10% PEG 3350. Crystals of the G169L variant were obtained under conditions of 12% PEG 3350 and 100 mM sodium malonate (pH 5) and exhibited morphological similarity to the wild-type crystals. Crystals of the G169V variant were obtained using protein (500 uM) prepared with coproheme (15 mg/mL) in 100 mM HEPES/100 mM KCl pH7.5 buffer and 10 mM of KCN. Batch crystallization was performed against a solution of 0.2 ammonium phosphate pH 7.0, 20% w/v polyethylene glycol 3,350, and 200mM betaine. Diffraction quality and space group determination were initially assessed using an in-house diffractometer before being sent to the synchrotron (Table 1). Structural determination was performed using molecular replacement with MolRep, employing the Corynebacterium diptheria CdChdC (PDB ID 6XUB) structure from the protein data bank (PDB) as a search model to generate an initial electron density map. Subsequent rounds of model building and refinement were conducted using COOT and PHENIX.

Table 1.

X-ray data collection and refinement statistics.

	Wild-type ScChdC	G169V ScChdC	G169L ScChdC
Data Collection
Beamline	APS 22-ID	APS 22-ID	NSLS-II NYX
Space Group	P4 1 2 1 2	P6 3	P1
a, b, c (Å)	184.7, 184.7, 241.6	132.7, 132.7, 122.2	67.2, 77.1, 77.3
angles (°)	90, 90, 90	90, 90, 120	114, 98, 110
Wavelength (Å)	0.97	0.97	0.98
Resolution Range (Å)	50.0-2.5	50.0-1.97	50.0-1.50
Outer Shell	2.59-2.5	2.04-1.97	1.56-1.50
Unique Observations	144,172	86,577	221,562
Completeness (%)	99.7(99.9)a	97.9(99.9)	96.5(94.3)
Rsym (%)b	5.2(52.1)	7.2(56.4)	6.0(26.6)
CC* (CC ½) outer shell	0.993(0.592)	0.990(0.564)	0.987(0.950)
Redundancy	6.8(7.0)	4.5(3.5)	1.8(1.6)
I/σ	17.3(1.1)	14.4(1.1)	28.5(4.5)
Refinement
Protein Atoms	18,280	8,850	8,989
Solvent Atoms	101	231	853
Resolution Limits (Å)	50.0-2.5	50.0-1.97	50.0-1.5
Rcryst (%)	20.1	19.3	17.0
Rfree (%)	24.9	22.5	18.9
rmsd bonds (Å)	0.011	0.008	0.006
rmsd angles (°)	1.4	1.06	0.88
average B factor (Å2)	36.4	31.9	13.3
PDB ID code	9MIQ	9MHQ	9MHR

^aNumbers in parentheses denote values for the outermost resolution shell.

^b$R_{\mathrm{sym}}=\sum _{\mathrm{hkl}}[\sum _{\text{I}}(\mid I_{\mathrm{hkl},\text{I}}-<I_{\mathrm{hk}l}>\mid )]/\sum _{\mathrm{hkl},\text{I}}<I_{\mathrm{hkl}}>$, where $I_{\mathrm{hkl}}$ is the intensity of an individual measurement of the reflection with indices hkl and $<I_{\mathrm{hkl}}>$ is the mean intensity of that reflection.

Results and Discussion

Structure of the wild-type ScChdC.

Crystals of the wild-type ScChdC, pre-incubated with coproheme, diffracted to 2.5 Å resolution. The overall fold and oligomeric state were consistent with previous ChdC structures, with monomers assembling into a circular pentamer (Figure 3, Panels A & B). For the wild-type enzyme, two pentamers were present in the asymmetric unit. Structural alignment of all ten monomers showed high similarity (RMSD < 0.5 Å), with only minor differences in the N- and C-termini as well as several loop regions. Although crystallization experiments were performed with coproheme-loaded enzyme, diffraction data revealed the presence of the reaction intermediate, monovinyl monopropionate deuteroheme (MMD), in the active site (Figure 3, Panel C) of the wild-type enzyme. MMD was also observed in the active site of the G169V variant structure, despite the addition of 10 mM KCN, which has prevented turnover in past crystal structures (10). It is unclear whether this conversion is something that occurred during crystallization or data collection but given the prevalence of peroxides in PEG precipitants during storage and the time required to obtain crystals, we posit that turnover occurred during crystallization. In addition to MMD, clear electron density was observed for several conserved residues that are critical to the proposed reaction mechanism. Among the amino acids key to the mechanism are H128, which deprotonates hydrogen peroxide and facilitates compound I formation; H168, the proximal ligand to substrate; and Y145, the catalytic tyrosine (Figure 3, Panel C). The binding mode of the MMD intermediate confirms that a 90° rotation has occurred, positioning the second propionate for oxidative decarboxylation next to Y145. In addition, the active site loop (residues 119-129) has adopted a closed conformation in all the monomers, with the oxygen atoms of propionate 7 (P7) of MMD within 3.0 angstroms of the backbone of the active site loop, forming hydrogen bonds with the amide protons of residues 124 and 125. These interactions help position the intermediate for catalysis, as there would not be an equivalent interaction when coproheme is bound. In fact, based on previous work, p7 of coproheme is positioned to interact with R179 prior to turnover. The guanidinium group of R179 is sandwiched between the propionate of the substrate and the side chain of E136, which is part of the active site loop, and in our MMD-bound structure, the latter salt bridge is intact. Structurally speaking, the active site loop links the N- and C-terminal ferredoxin-like domains of Actinobacterial ChdCs and is shorter than the corresponding loop in Firmicutes (6). H128 is part of this loop, is conserved in Actinobacterial ChdCs (Figure 2), and positioned to act as a distal base (Figure 3). Similarly, this loop has been proposed to trap the substrate as well as prevent MMD dissociation (21), essentially playing a critical role in orienting both the substrate and the intermediate as well as positioning the general base for catalysis. The ScChdC structure reported here (Figure 3, Panel C) supports this hypothesis. Moreover, because H128 is essential for compound I formation it makes mechanistic sense that the active site loop, containing H128, correctly orients the substrate, or the MMD intermediate, prior to positioning this catalytic residue ahead of hydrogen peroxide binding. If the hypothesis that the active site loop positions the porphyrin is correct then, conversely, the orientation of porphyrin should influence the loop structure.

Figure 3. Cartoon representation (Panel A) and side view (Panel B) of the ScChdC pentamer highlighting the position of one active site (red dashed box) and a wall-eyed stereo view (Panel C) showing a stick representation of the model and 1F_O-F_C composite omit map (green cage) for key amino acid side chains and the MMD intermediate observed in the wild-type ScChdC data.

The composite omit map is contoured at 3 σ and was generated using the simulated annealing protocol. Interactions with key conserved residues are shown in the active site in panel C.

Enzymatic activity of wild-type ScChdC and helix variants.

To probe the role of conserved residues in catalysis and address the hypotheses above, we focused on the active site α-helix and generated the ScChdC variants H168C, G169V and G169L. These mutations target conserved residues important for catalysis: H168C alters iron ligation while preserving alpha-helical structure, and substitutions at G169 modulate helix flexibility by destabilizing (V) or stabilizing (L) the α-helix. In addition, the larger side chain will crowd the active site and probe how the orientation of the porphyrin plane influences structure and function. To address the activity of the enzymes, we employed a fixed time point assay to assess enzyme activity and intermediate accumulation, as others have previously described (18, 29) (Figure 4). Reactions were allowed to proceed for 5 minutes, the time point at which no further spectroscopic changes are observed, and quenched with concentrated HCl to ensure the termination of the reaction. Interestingly, all variants were active, producing some heme b, while the wild-type ScChdC exhibited the highest efficiency, converting ~90% of coproheme to heme. In contrast, all variants displayed impaired conversion: G169V achieved ~10% conversion, H168C achieved ~22% conversion, while G169L achieved ~26% conversion of coproheme to heme. Like the wild-type enzyme, all the variants also accumulated some (10-15% conversion from coproheme) of the intermediate MMD at the end of the fixed time point assay. However, in contrast to the wild-type enzyme, turnover was incomplete for all the variants with as much as 80% of the substrate, coproheme, remaining in the assay solution for the G169V variant. Given these data, two questions can be raised. First, why are the ScChdC variants failing to convert as much coproheme to heme b, when compared to the wild-type enzyme, and second, why is the bulkier group (G169L) producing more heme b, when compared to the G169V variant?

Figure 4. HPLC analysis of the reaction mixtures following a 5-minute fixed time point turnover assay.

The assay was performed as described by Michlits et al.(15), and HPLC analysis was utilized to determine the relative amounts of coproheme, MMD, and heme b for the wild-type ScChdC and variants. Error intrinsic to baseline correction and peak integration is typically less than 2% of the total relative amount (total integrated peak area).

Stopped-flow spectroscopy.

To investigate any mechanistic differences in the initial steps of the reaction, we used stopped-flow spectroscopy to monitor compound I formation and degradation at 390 nm (Figure 5). These experiments were performed with 10 uM ScChdC equilibrated in a 1:1 molar ratio with coproheme, followed by the addition of sub-stoichiometric equivalents of hydrogen peroxide (7 uM - 0.35 eq, 10 uM - 0.5 eq, 13 uM - 0.65 eq, and 16 uM - 0.8 eq) to limit the reaction to a single oxidative decarboxylation event as well as prevent heme bleaching. Similar to peroxidases, an exceptionally fast rate of compound I formation has been predicted for ChdC enzymes and therefore previous stopped-flow investigations utilized a variant that removed the catalytic tyrosine (Y147A) of the Listeria monocytogenes enzyme (LmChdC)(10). Our investigation is the first to investigate compound I formation with a wild-type enzyme which showed showed Soret hypochromicity consistent with compound I formation (Figure S1). However, we found that absorption changes observed were best modelled using relaxation kinetics that were consistent with at least two chemical processes occurring (Figure 5). However, given the exceptionally fast rate of compound I formation and the timeframe of the stopped-flow experiment, subsequent steps, including formation of the tyrosyl radical as well as the first oxidative decarboxylation, could be occurring. In fact, we address the formation of the tyrosyl radical in the follow section. The use of sub-stoichiometric amounts of hydrogen peroxide in the experiments will prevent reaction progression past the first oxidative decarboxylation and/or prevent the accidental measurement of heme bleaching. Curve fitting (Figure 5, red lines) was consistent with at least two reactions occurring at exponential rates (Table 2). The rate constants obtained from these fits most likely represent multiple reaction steps and are therefore being investigated further. However, a couple observations can be made that are relevant to this investigation. First, the wild-type and H168C variant were best fit by Equation #1, which accounts for increasing and decreasing absorption at 390 nm over the time course of the experiment. Essentially, corresponding to the formation, then decomposition of compound I as the radical is transferred to the conserved tyrosine residue (See Figure 1). In this case (wild-type enzyme and H168C), the plane of the porphyrin should be similar, this isn’t the case for the G169V and G169L variants. According to equation #1, k₂ corresponds to the rate associated with an increase in absorption (compound I formation), while k₁ represents the rate associated with the decrease in absorption (decomposition of compound I). Given the exceptionally fast rate of compound I formation, we suspect that our stopped flow experiments are missing a substantial amount of compound I formation, and this will be investigated further. However, for the wild-type data, the observation that k₁ remains constant as the hydrogen peroxide concentration is increased, supports the hypothesis that the rate of hydrogen atom abstraction, or radical transfer (as compound I is converted to tyrosyl radical) does not change as hydrogen peroxide is increased. The observation that the G169V and G169L are better fit by Equation 2 may simply be due to the slower rate of compound I formation, as the adsorption only increases over the time course of these experiments. Future experiments will increase the time course of the stopped-flow experiment to better capture the decomposition of compound I and the rate of tyrosyl radical formation.

Figure 5. Stopped-flow spectroscopy monitoring absorption changes at 390 nm for the wild-type ScChdC (Panel A), as well as the H168C (Panel B), the G169V (Panel C), and the G169L (Panel D) variant of ScChdC.

Purified ScChdC (10 μM) was equilibrated in a 1:1 molar ratio with coproheme for five minutes prior to being mixed with buffer containing 0.35 (trace 1), 0.5 (trace 2), 0.65 (trace 3), and 0.8 (trace 4) equivalents of hydrogen peroxide. Each trace represents the average absorption change for five experiments and is fit (red line) to the equations described in the text. Data were normalized for better comparison by adjusting the absorption at 390 nm to zero at time zero. All fitting parameters for the normalized data are shown in Supplementary Tables S1 and S2. Raw data is shown in Supplementary Figure S3.

Table 2.

Rate constants for relaxation kinetics observed for the wild-type and variants of ScChdC.

Equivalents of hydrogen peroxide	k1 (s−1)a	k2 (s−1)
Wild-type
0.35	106.0 ± 1.4	21.6 + 0.3
0.5	104.3 ± 0.6	15.9 ± 0.2
0.65	107.9 ± 0.5	8.5 ± 0.2
0.8	111.6 ± 0.7	1.3 ± 0.6
H168C
0.35	92.7 ± 0.5	6.2 ± 0.1
0.5	90.9 ± 0.3	5.3 ± 0.1
0.65	98.5 ± 0.4	5.2 ± 0.1
0.8	105.6 ± 0.5	5.2 ± 0.6
G169V
0.35	60.7 ± 0.4	18.7 ± 0.3
0.5	82.0 ± 0.3	26.5 ± 0.1
0.65	80.1± 0.1	21.5 ± 0.1
0.8	89.8 ± 0.2	29.9 ± 0.2
G169L
0.35	33.0 ± 1.4	8.1 ± 1.9
0.5	37.8 ± 0.6	10.5 ± 0.4
0.65	46.4 ± 0.4	13.1 ± 0.1
0.8	35.7 ± 0.2	11.0 ± 0.1

^aThe change in absorbance at 390 nm was fit to one of two equations, as described in the Materials and Methods. Specifically, the wild-type and H168C variant were fit to Equation #1, while the G169V and G169L variants fit to Equation #2.

EPR spectroscopy.

To address the mode of coproheme binding as well as generation of the tyrosyl radical during the time course of the stopped-flow experiment described above, we recorded EPR spectra prior to, and ~0.5 s following the addition of hydrogen peroxide to coproheme-bound enzyme. In the presence of the substrate coproheme (Figure 6, Panel A), all the samples exhibited some mixture of high- and low-spin species with the wild-type enzyme displaying predominantly six-coordinate, low-spin, ferrous iron. Specifically, the low-field g-values of 2.3 and 2.9 have been observed in complexes with two axial heterocyclic ligand rings, such as histidine, that are oriented orthogonally to one another (30). This agrees with the previous proposal that the distal histidine on the loop may be acting as the 6^th low spin ligand in the WT Actinobacterial enzyme (31). However, both the G169V and G169L variants displayed characteristic spectra of five-coordinate, high-spin, ferric heme species, indicated by g-values of 5.9 and 1.98 (32). This shift in coordination environment indicates that the conformation of the porphyrin has been altered compared with the coproheme bound to the wild-type enzyme. The wild-type data is consistent with six coordinate, majority low-spin, ferrous coproheme and the shift to greater high-spin character suggests that the G169V and G169L variants have distorted the porphyrin ring and ligation geometry of the iron atom, making the porphyrin ring more planar, and interrupting interactions with this sixth ligand (the distal histidine). However, if turnover is initiated and samples are freeze-quenched within one second, we clearly see a radical signal in all samples (Figure 6, Panel B), consistent with the observation that all enzymes are catalytically active. Moreover, if the EPR spectra are recorded at 77 K, the signal from the iron center is lost and the line shape and intensity of the radical signal provide us with additional information (Figure 6, Panel C). Specifically, revealing a radical signal that is similar to tyrosyl radical signals that have been previously reported for other ChdCs (16) as well as the homologous enzymes such as the dye-decoloring peroxidases (33). In addition, it is notable that the radical signal for the wild-type enzyme has significantly less intensity, consistent with the turnover (Figure 4) and stopped-flow data (Figure 5). Concentrations of protein, substrate, and hydrogen peroxide remained constant between samples. Hence, turnover for the wild-type enzyme is much faster and consequently less of the tyrosyl radical is remaining. At the present time, we are unable to explain the EPR spectra observed for the H168C variant, with the exception that this variant is still active and capable of generating the tyrosyl radical (Figure 6, Lower Panel, Trace 4). In addition, this signal is present in the purified enzyme and is under further investigation.

Figure 6. Electron paramagnetic resonance spectra for coproheme bound wild-type ScChdC (trace 1), G169V (trace 2), G169L (trace 3), and H168C (trace 4) variants.

Panel A; Spectra recorded at 10 K for coproheme-loaded samples prior to turnover. Panel B; Turnover samples frozen one second after addition of one equivalent of hydrogen peroxide and recorded at 10 K. Panel C; The same samples recorded at 77 K confirming the formation of a tyrosyl radical in all samples.

Structures of the G169V and G169L variants.

To further investigate the coordination environment implied by the EPR experiments we set out to obtain structural data for the ScChdC variants. Diffraction quality crystals were obtained for the G169V and G169L ScChdC variants to 1.97 and 1.5 Å, respectively (Table 1). Like the wild-type ScChdC, despite beginning the crystallization trials with coproheme-bound enzyme, density was seen in the active site of each monomer, for both variants, that was consistent with the presence of MMD (Figure 7).

Figure 7. Wall-eyed stereoview showing stick models and 2F_o-F_c composite omit maps (green cage) for the G169V (Panel A) and G169L (Panel B) ScChdC variants.

The 2F_o-F_c composite omit maps were generated using the simulated annealing protocol and are contoured at 1.2 σ. The key catalytic amino acids H128 and Y145, as well as the mutations (V169 and L169), are labelled along with propionate #7 (p7). The next propionate to be oxidatively decarboxylated, propionate #4, on MMD is positioned next to Y145 in both models. The monomer for the G169L variant is in an open conformation

Interestingly, for at least two of the monomers in the G169L ScChdC model, the active site loop is clearly in an “open” conformation, with the side chain of H128 flipped out of the active site (Figure 7, compare Panels A & B). Consistent with the high-spin EPR signal, the porphyrin is now planar (Figure 7, Panel B) with a density for a water molecule observed within 2.4 angstroms of the iron atom. A comparison of the MMD binding mode in the wild-type ScChdC as well as the G169V and the G169L variants is shown in Figure 8. The porphyrin ring of MMD is considerably more ruffled in the structure of the wild-type enzyme. The porphyrin ring of MMD becomes increasingly more planar with the increasing size of the side chain introduced at position 169, with the G169L variant having a mostly planar porphyrin (Figure 8). In addition, propionate #7 adopts two different conformations in the G169L ScChdC model (Figure 8, compare Panels A & B). The conformation shown in Figure 8, Panel B, is seen in monomers with an open conformation. In addition to the large movement of the catalytic residue H128, there is a cluster of charged amino acids, both positive and negative, that also undergo substantial conformational changes (Supplemental Movie). In the closed conformation, E136 forms a salt bridge with R179. This salt bridge effectively locks down the C-terminal end of the active site loop. When the loop adopts the open conformation, this salt bridge is broken, allowing the side chain of E136 to flip over and move eight angstroms away. Compensating for the loss of the salt bridge, D136 flips over and moves to within four angstroms of R179, which has formed hydrogen bonds with the backbone carbonyl of T137 and the side chain hydroxyl oxygen (See supplemental movie).

Figure 8. Structural overlay of the wild-type ScChdC (red), as well as the G169V (yellow) and G169L (green) variants showing the relative conformation and orientation of MMD in the A monomer (Panel A).

Panel B shows a similar comparison of the D monomers, wherein the active site loop of the G169L ScChdC has adopted an open conformation and propionate #7 (p7) has flipped. All the monomers in the wild-type and G169V pentamers adopted a closed conformation. The RMSD for backbone alpha carbon atoms was less than 1.5 Å when pentamers were aligned.

Conclusions

The biochemical and structural data for a ChdC from the Actinobacteria Streptomyces coelicolor (ScChdC) presented here provides new insight into the architecture and structural dynamics that occur during catalysis. Through a few simple mutations we probe the interplay between the active site helix, the porphyrin intermediate, and the active site loop. Mutations at a conserved glycine show that the enzyme remains catalytic despite changes in iron ligation and the conformation of the substrate coproheme. Structural analysis of these variants provides evidence to support the hypothesis that the active site loop facilitates porphyrin selectivity and repositioning of the porphyrin influences the conformation of the active site loop.

Active site loop, enzyme selectivity, and catalysis.

The mechanism of ChdC is complicated in so much as the active site must perform the same chemistry on two different substrates, coproheme and MMD. In addition, the product, heme, can also react with hydrogen peroxide to generate reactive oxygen intermediates, something that the enzyme must avoid. A considerable amount of work has interrogated the specifics of substrate binding, generation of compound I and the tyrosyl radical, as well as rotation of the intermediate (5, 7, 18, 23, 29, 34, 35), and therefore we focus on how the data presented here advance the current catalytic model. First, H168 and G169 are strictly conserved residues across all ChdCs and part of an active site helix. H168 ligates the iron atom of coproheme and MMD, while G169 forms a hydrophobic “notch” in the helix (Figure 9). How this structural feature influences the mechanism has not been addressed, but structural data from several enzymes show that the edge of the porphyrin, either coproheme or MMD, sit in this notch (Figure 9). Rotation of the porphyrin by 90 ° between decarboxylation steps requires reorientation of the newly formed vinyl group and remaining propionate groups, requiring movement of the porphyrin through this notch. At the beginning of the catalytic mechanism coproheme is bound with p2 and p4 initially positioned at the back of the active site, with p2 next to the catalytic Y145 (Streptomyces coelicolor numbering). P6 and p7 are facing the front of the active site with the edge of the porphyrin ring between them sitting in the groove formed by G169 (Figure 9, PDB IDs - 7Q4F, 7Q4G, & 6XUC). P4 is forming hydrogen bonds with an arginine and the N-H proton of the indole ring of W153. This interaction will reform with p6 after the first round of catalysis occurs and the MMD rotates 90°. Prior to the first oxidative decarboxylation and rotation, the propionate of p6 is on the same side of the porphyrin as the iron ligand (H168 in ScChdC) forming a hydrogen bond with the side chain of N125, while the propionate of p7 is on the opposite side of the porphyrin and interacts with the catalytic histidine (H128 in ScChdC, Figure 9). N125 and H128 are part of the active site loop and conserved in Actinobacteria (Figure 2). In actinobacteria, this histidine (H128) has a critical role as a distal base in the formation of compound I (Figure 1)(18). Immediately following turnover, the newly formed vinyl from p2 is free to rotate past a cluster of hydrophobic residues at the back of the active site. P7 is repositioned to form hydrogen bonds with two backbone amide protons of residues 124 and 125. Again, both amino acids are part of the active site loop and precede the critical distal base (H128 in ScChdC). Essentially, these hydrogen bonds reposition the distal histidine for the next round of catalysis. P6 rotates to assume the same orientation and interactions held by p4 before the first turnover. Specifically, p6 forms hydrogen bonds with an arginine (R149) and the N-H proton of the indole ring of W153. Recent work has underscored the importance of the hydrogen bonding network in the active site and the evidence that the intermediate rotates within, but does not leave, the active site(16, 34, 36). All available data support the hypothesis that porphyrin-loop interactions keeps the MMD in the active site during rotation and reposition H128 for the next round of catalysis, as we see for the wild-type ScChdC (Figure 3) and the G169V variant (Figure 7). The alternate conformation we observe for the active site loop in the MMD-bound G169L structure confirms that repositioning the porphyrin influences these interactions. More specifically, the larger side chain has resulted in repositioning the porphyrin (Figure 8), which does not appear to affect compound I formation but may disrupt the interaction between p6 of coproheme (or p7 of MMD) and the backbone amides of residues 124 and 125. Hence, for the G169L variant the thermodynamic landscape has shifted in favor of the open conformation of the active site loop, breaking the interaction of p7 with the backbone (Figure 9). Confirmation of this can be seen in the fixed time point assay, where G169L produces nearly three times more heme b when compared with the G169V variant. A possible explanation for this observation is discussed below.

Figure 9. Structural overlay showing a ribbon diagram for three coproheme-bound ChdC enzymes (PDB IDs 7Q4F, 7Q4G, and 6XUC) as well as the MMD-bound ChdC enzyme from Corynebacterium diphtheriae (PDB ID 6XUB) and the G169L ScChdC variant described in this work.

Structures are color coded by their carbon atoms as denoted by the legend in the figure.

Significance of MMD rotation and two different hydrogen bonding networks.

The significance of the hydrogen bonding network in MMD rotation and the general ChdC mechanism has been discussed in several papers(16, 34). In addition, the current model for the ChdC mechanism posits that the second oxidative decarboxylation occurs slower than the first(29). ChdC from Corynebacterium diphtheria (CdChdC) is another Actinobacterial enzyme that has been investigated thoroughly. Like ScChdC, a conserved histidine in the active site loop functions as a distal base (H118 in CdChdC). Interestingly, mutation of this histidine to phenylalanine (H118F CdChdC) resulted in an enzyme that could catalyze the first oxidative decarboxylation, but not the second(36). The investigators explained this observation by using molecular dynamics to evaluate the different binding modes for each “ES complex”. More specifically, they considered the molecular dynamics of coproheme bound with p2 near the catalytic tyrosine, compared to MMD bound with p4 near the catalytic tyrosine. Based on these calculations, and additional observations, they concluded that hydrogen peroxide had limited accessibility to the active site in the coproheme-bound H118F CdChdC. Compound I formation was slow and, due to the bulky phenylalanine group, once the oxidative decarboxylation had occurred, rotation of the MMD intermediate was essentially prohibited, halting catalysis. It is important to note that the phenyl ring could participate in ring stacking or pi-pi interactions with the porphyrin. Regardless, while no explanation was provided as to what replaces H118 as the distal base in the reaction mechanism for the first oxidative decarboxylation, the work provided unequivocal evidence for rotation of the intermediate and not re-binding. The introduction of the hydrophobic phenyl group may not require too much additional space, when compared to an imidazole group, but it will increase the hydrophobic character of the active site, consistent with the slower binding of molecules like hydrogen peroxide and cyanide. When these data are considered in the context of the observations reported here, we can make a converse argument to explain the greater production of heme b by the G169L ScChdC compared to the G169V variant. Specifically, that access to the active site and the hydrogen bonding network for each “ES complex” (coproheme versus MMD) must be considered independently. Applying the accessibility argument discussed above(36), the greater production of heme b, relative to the G169V variant, could be explained by greater accessibility, consistent with the observation of the open conformation for the active site loop seen in the G169L structure. Both variants (G169V and G169L ScChdC) accumulate MMD, but the G169V variant does not favor the open conformation. In essence, both variants (G169V and G169L) displace the porphyrin moiety relative to the wild-type enzyme (Figure 8), but the degree of displacement in the G169L variant is large enough to disrupt the hydrogen bond network that keeps the active site loop closed, through the disruption of the hydrogen bonds between p7 of MMD and the backbone amide protons of residues 124 and 125 (residues 113 and 114 in CdChdC). While this disrupts the proper positioning of the distal histidine, resulting in slower compound I formation, shifting the thermodynamic landscape towards the open conformation facilitates MMD re-binding and another oxidative decarboxylation for the G169L ScChdC.

In summary, this work provides evidence to support the hypothesis that the active site loop of ChdC enzymes plays a key role in substrate selectivity and intermediate rotation. We have confirmed an open conformation for the active site loop and the interconnected nature of the loop structure and porphyrin for ScChdC. While these observations are for a ChdC from an Actinobacteria, similar structural dynamics between the α-helix and porphyrin are most likely at work in enzymes from Firmicutes, although further investigation into the impact on the loop is necessary, considering the differences in this structural element between the two phyla.

Supplementary Material

A PYMOL file with a movie showing the conformational changes between the open and closed conformation of the active site loop is included. MMD in the closed and open conformation is shown with carbon atoms colored cyan and magenta, respectively. The reader may manipulate the PYMOL movie freely, but we have also included a movie file (loop.mp4) that highlights structural interactions discussed in the text.

• Captured a novel conformation of the active site loop of coproheme decarboxylase

• There is a relationship between loop and porphyrin conformation

• An active site glycine is key for porphyrin rotation, orientation, and selectivity

Data availability.

The accession number for the atomic coordinates reported in this paper are 9MIQ, 9MHQ, and 9MHR for the wild-type ScChdC, G169V ScChdC, and G169L ScChdC models, respectively.