Metal and redox selectivity of protoporphyrin binding to the heme chaperone CcmE

Abstract

The interaction of heme with the heme chaperone CcmE is central to our understanding of cytochrome c maturation, a complex post-translational process involving at least eight proteins in many Gram-negative bacteria and plant mitochondria. We have shown previously that Escherichia coli CcmE can interact with heme non-covalently in vitro, before forming a novel covalent histidine-heme bond, in a redox-sensitive manner. The function of CcmE is to bind heme in the periplasm before transferring it to apocytochromes c. In the absence of structural information on the complex of CcmE and heme, we have further characterized it by examining the binding of the soluble domain of CcmE (CcmE′) to protoporphyrins containing metals other than Fe, namely Zn-, Sn-, Co- and Mn-protoporphyrin (PPIX). CcmE′ demonstrated no affinity for the Zn- or Sn-containing protoporphyrins and low affinity for Mn(II)-PPIX. High-affinity, reversible binding was, however, observed for Co(III)-PPIX, which was highly sensitive to oxidation state as demonstrated by release of the ligand from the chaperone on reduction; no binding to Co(II)-PPIX was observed. The non-covalent complex of CcmE′ and Co(III)-PPIX was characterized by non-denaturing mass spectrometry. The implications of these observations for the in vivo function of CcmE are discussed.

---

Introduction

Transient heme-binding proteins are employed in nature to sequester heme in order to provide protection against the oxidative damage that it can catalyze (e.g. hemopexin in blood plasma¹) and in bacteria as a means of obtaining extracellular iron (e.g. the hemophore HasA²). In the pathway providing heme to respiratory cytochromes c in many bacteria and plant mitochondria, the heme chaperone CcmE was discovered.³ CcmE binds heme on the p-side of the energy-transducing membrane (i.e. in the periplasm in E. coli) before the cofactor is transferred to the apocytochrome.⁴ An unexpected covalent intermediate between CcmE and heme has been detected in vivo,³ as well as interactions between CcmE and other proteins encoded by the same operon (CcmABCDEFGH), which, in E. coli at least, are all essential for the process of cytochrome c maturation and are located in the cytoplasmic membrane.^5–7 These include CcmC, which is involved in attaching heme to CcmE,⁸ and CcmF,⁹ which is thought to assist in the transfer of heme to the characteristic CXXCH motif of apocytochromes forming two covalent bonds between the cysteine thiols and the heme vinyl groups. We have shown that the soluble periplasmic domain of CcmE (lacking its native single helical membrane anchor; referred to herein as CcmE′) binds heme non-covalently in vitro prior to formation of a covalent bond between a heme vinyl group and an essential histidine residue,^10,11 for which reduced heme is required. Spectroscopic analyses have shown that protein ligation to the heme iron is sensitive to the oxidation state of the metal, in both the non-covalent and covalent complexes.^12,13 The structure of a peptide of CcmE (containing 11 amino acids of the protein) with heme bound to the essential histidine revealed an unexpected bond involving the N^δ1 of the histidine (H130 in E. coli) and the β-carbon of a heme vinyl group.¹⁴ The structure of apo-CcmE is shown in Fig. 1. We and others have shown tyrosine ligation (Y134) to Fe from the protein to the covalently linked heme.^12,15 Why the counter-intuitive covalent bond is formed, and how it is broken in the heme transfer reaction to apocytochromes, remains unknown. The ATP-binding cassette proteins CcmAB, encoded by the Ccm operon, appear to be involved in the heme-CcmE reactions.^16,17

Fig. 1 Structure of the apo-form of the heme chaperone CcmE and protoporphyrin IX (top left; M indicates the different metals, replacing the heme Fe, used in this study). His130, which forms a covalent bond with a heme vinyl group is shown in stick representation on CcmE, as well as Tyr134, which ligates the iron of covalently bound heme. The structure was rendered in PyMol (DeLano, W.L. The PyMOL Molecular Graphics System (2002) at http://www.pymol.org). PDB 1LIZ.¹⁸

The structure of the apo-form of E. coli CcmE (Fig. 1) revealed a β-barrel core with a flexible C-terminal domain¹⁸ that has been studied in vitro and in vivo.^19,20 Mutagenesis and modeling studies have implicated a hydrophobic patch on the surface of the β-barrel in heme binding^18,21 but the structure of the heme-bound form of the protein remains elusive. In order to further characterize the interaction of heme with CcmE, and in view of the redox sensitivity of the interaction with Fe-PPIX, we have explored the effect of metal substitution in the protoporphyrin (Fig. 1) on its binding to the chaperone by studying the binding to CcmE′ of Zn-, Sn-, Co- and Mn-protoporphyrin (PPIX). Recent work has shown that Zn- and Sn-PPIX have an inhibitory effect on the cytochrome c maturation system in vivo, but that these are not incorporated into the resulting cytochromes c.²² Previous in vitro attachment studies to apocytochromes with these metallo-protoporphyrins (M-PPIXs) showed non-covalent complex formation with Co-PPIX and Zn-PPIX, but only the latter was found to form covalent linkages with the apocytochrome.²³ We discuss the implications of our findings on the function of CcmE in vivo.

Results

The apo-forms of CcmE′ used in this work were expressed in the cytoplasm of E. coli with a cleavable His₆-tag attached to their N-termini. After purification and thrombin cleavage of their His₆-tags, the proteins were shown to be pure by SDS-PAGE and of expected mass by ES-MS (not shown).

Interaction of apo-CcmE′ with Zn-, Mn- and Sn-protoporphyrin IX

Spectroscopic analysis of mixtures of apo-CcmE′ plus Zn-PPIX, which would be expected to detect binding, revealed no indication of interaction with the protein. The UV-Vis spectrum of Zn-PPIX did not change significantly when apo-CcmE′ was added, as shown in Fig. 2a. To confirm that a stable complex was not formed, a mixture of Zn-PPIX and CcmE′ was passed through a gel filtration column; the protein and the M-PPIX eluted separately. Fluorescence studies with Zn-PPIX are highly sensitive to its interaction with protein.²⁴ We examined the fluorescence of mixtures of Zn-PPIX and CcmE′ and no change in fluorescence of the Zn-PPIX was observed, indicating that there was no interaction between this M-PPIX and CcmE′ (not shown). Having observed no binding with Zn-PPIX, we then tested Sn-PPIX, which has also been shown (as described above) to inhibit the cytochrome c maturation system in which CcmE functions²², and also detected no binding (Fig. 2a). Small spectral shifts were observed upon addition of Mn(II)-PPIX to CcmE′, as shown in Fig. 2b. These changes are consistent with weak binding of the M-PPIX to the protein. No binding was observed spectroscopically with Mn(III)-PPIX (not shown). Neither of the oxidation states of Mn-PPIX were retained with the protein fraction on a size-exclusion column following incubation with CcmE′.

Fig. 2 (a) Absorption spectra of Zn-PPIX (10 μM) and Sn-PPIX (3 μM), free in solution (solid lines) or in the presence of apo-CcmE′ (1 : 3 M-PPIX to protein ratio; dotted lines). The spectra of the M-PPIX-protein mixtures were recorded after 10 min incubation at RT. No change in the spectra was observed over 24 h. (b) Absorption spectrum of Mn(II)-PPIX (10 μM) in the absence (solid line) or presence (dotted line) of apo-CcmE′ (30 μM). The spectrum of the Mn(II)-PPIX-protein mixture was recorded after 10 min anaerobic incubation at RT and no further change was observed over 24 h.

Interaction of apo-CcmE′ with Co-protoporphyrin IX

When apo-CcmE′ was added to Co(III)-PPIX the UV-Vis spectrum of the mixture shows that a complex forms rapidly, as shown in Fig. 3; significant shifts in the wavelength maxima of the PPIX occurred on the addition of protein. Free Co-PPIX is shown for comparison. The spectrum of the complex is similar to what has been observed previously for a non-covalent Co-PPIX apocytochrome c complex.²³Fig. 4 shows the spectroscopic change that occurs when reductant is added to the Co(III)-PPIX complex and the change with time following this addition. This result indicates that the M-PPIX appears to be released from the protein following the change in its oxidation state to Co(II)-PPIX with the simultaneous appearance of a spectrum characteristic of free Co(II)-PPIX. The protein mixture was then allowed to oxidize in air and the spectra show that the resulting Co(III)-PPIX binds to the protein again (data not shown). The reversibility of this phenomenon is suggestive of non-covalent binding of Co(III)-PPIX to apo-CcmE′. No binding of apo-CcmE′ to Co(II)-PPIX was detected by either UV-Vis spectroscopy or size-exclusion chromatography (not shown), which separated the protein and M-PPIX. A Co(III)-PPIX and CcmE′ mixture eluted from a size-exclusion column in a single peak, indicating the formation of a stable complex.

Fig. 3 Absorption spectrum of Co(III)-PPIX (5 μM) in the absence (solid line) or presence (dotted line) of apo-CcmE′ (5 μM). The spectrum of the Co(III)-PPIX-protein mixture was recorded after 10 min incubation at RT and no further changes were observed over 24 h.

Fig. 4 Evolution of the visible spectrum of the Co(III)-PPIX-CcmE′ complex (∼10 μM) over time upon addition of dithionite. The spectra recorded at t = 0, 0.5, 1, 1.5, 2.5, 4, 8, 14 and 20 min after addition of dithionite are presented. The direction of the peak shifts upon reduction is indicated by the arrows.

Characterization of the Co(III)-PPIX-CcmE′ complex

We attempted to characterize further the complex formed by Co(III)-PPIX and CcmE′ using various techniques. Chromatographic separation of the protein and the M-PPIX occurred with neither size-exclusion nor reverse-phase HPLC. It should be noted that the non-covalent complex between Fe-PPIX and CcmE with a His-tag, which enhances the binding, was also sufficiently stable to undergo RP-HPLC intact (unpublished observation), indicating that the observation we have made with the Co(III)-PPIX-CcmE′ complex is not necessarily due to a covalent interaction. Passing the complex through a size-exclusion column run under denaturing conditions (6M guanidinium chloride) also did not separate the protein from the M-PPIX. Mass spectrometry was also employed to characterize the complex; spectra from denaturing ES-MS of the complex showed only the apo-CcmE′ protein; spectra of holo-CcmE′ (with Fe-PPIX attached in vivo) showed the presence of the covalent complex. This is further evidence against the complex between Co(III)-PPIX and CcmE′ being covalent in nature. Retention of Co(III)-PPIX with the protein at high denaturant concentrations presumably indicates that the protein is not fully denatured. In order to examine further this apparently non-covalent complex we employed non-denaturing mass spectrometry, a technique that is being used increasingly to examine protein-protein interactions as well as complexes of proteins and various ligands.²⁵ A 2 : 1 mixture of apo-CcmE′: Co(III)-PPIX was prepared, desalted and examined by mass spectrometry under conditions selected to maintain the native complex. The spectrum observed is shown in Fig. 5. Three ion series, with decreasing abundance, were observed: series A (mass 14 765 Da) is apo-CcmE′, series B (mass 15 384 Da) is a 1 : 1 complex of CcmE′ : Co(III)-PPIX, and series C (mass 16 003 Da) is a small proportion of a 1 : 2 complex of CcmE′ : Co(III)-PPIX. If the protein-ligand mixture was run on the spectrometer without a desalting step to remove unbound ligand, a higher proportion of 1 : 2 complex was observed, suggesting that interaction can be disrupted by desalting. The latter indicates that the complex of 1 : 2 stoichiometry is unlikely to reflect binding of the M-PPIX to a physiologically relevant site, because the interaction has a low affinity. Conditions, specifically desolvation voltages, which disrupt non-covalent interactions in the mass spectrometer, were applied and resulted in the disappearance of the complex between the chaperone and the Co(III)-PPIX, further confirming its non-covalent character (not shown).

Fig. 5 Non-denaturing ES-MS spectrum of apo-CcmE′ (10 μM) in the presence of Co(III)-PPIX (5 μM) in 10 mM ammonium acetate, pH 7. The A, B and C ion series correspond to apo-CcmE′, apo-CcmE′ + Co(III)-PPIX and apo-CcmE′ + (Co(III)-PPIX)₂.

Co(III)-PPIX binding to variants of CcmE′

Several variants of CcmE′ were tested for their ability to bind the M-PPIXs. These proteins carry amino acid substitutions of H130, H147 or Y134, amino acids that have been shown to be involved in heme iron coordination in either the in vivo produced holo-CcmE′ or the in vitro formed non-covalent heme-CcmE′ complex.^12,13 A CcmE′ mutant lacking its C-terminal domain from and including P136, that was found to be markedly impaired in the ligation of the heme iron but not in the binding with the protoporphyrin ring,²⁰ was also examined. The ligation of this M-PPIX was only altered in the truncated mutant, where the wavelength maximum of the Soret band shifted from 425.5 nm (WT) to 421 nm, but the effect was not as dramatic as in the case of heme.²⁰

Discussion

We have shown previously that the soluble periplasmic domain of CcmE (CcmE′) binds heme non-covalently in vitro and is dependent on the heme iron oxidation state.^10,11 In addition, the protein ligands to the iron are different in the two oxidation states.¹³ It is not yet clear what the physiological significance of the conformational changes involved in switching ligands is, though we have proposed that CcmE might have evolved these properties to prevent transfer of ferric heme to apocytochromes, which could result in incorrectly attached heme to protein, and therefore result in misassembled cytochromes c. The importance of correct heme attachment to respiratory proteins would warrant such protective mechanisms.

We attempted to provide further insight into the interaction between CcmE and its cofactor by examining the binding of M-PPIXs containing metals other than Fe, namely Zn, Sn, Mn and Co. In view of the oxidation state sensitivity of CcmE we sought also to examine whether there would be any difference between M-PPIXs that cannot undergo reduction/oxidation reactions, like Zn-PPIX, and those that can. In addition, we had hoped to identify a diamagnetic heme analogue that interacted with CcmE in the same way as heme to utilize in NMR studies for characterization of the complex. Unexpectedly, we were unable to detect any binding of either Zn- or Sn-PPIX to CcmE′. This was surprising in view of the fact that CcmE′ binds protoporphyrin.²⁰ It appears, however, that the heme-binding site of CcmE is more selective than we initially thought. This could be due to the inability of the porphyrin binding site on CcmE to accommodate the larger metal within the heme. This is consistent with the ionic radii of the metals in the octahedral field: high/low spin Co(III) is similar in size to high spin Fe(III) and high/low spin Fe(II), whereas high/low spin Zn, Sn, Mn(II) and high spin Co(II) are considerably larger. The weak binding that we have observed with Mn(II)-PPIX might be because Mn(II) is isoelectronic with Fe(III); Fe(III)-PPIX binds to CcmE′.¹¹

In this work we found a strong interaction between CcmE′ and Co(III)-PPIX, which appeared to be non-covalent, and was reversible when the complex was chemically reduced. We attempted to monitor the interaction by ¹H-NMR by titrating in Co(III)-PPIX but were unable to detect any changes in the protein spectrum indicative of binding. For reasons that are not clear, though we are confident of strong interaction, binding of this heme analogue was invisible to NMR spectroscopy. We also note that CcmE′ from Shewanella putrefaciens formed a tight complex with Fe(II)-PPIX (detected spectroscopically, not shown) in contrast with previously published observations suggesting that no interaction occurred in this case.²⁶ Attempts to characterize this interaction by NMR were also unsuccessful although UV-Vis spectroscopy clearly indicated complex formation.

The non-denaturing mass spectrometry shown in this work suggests that CcmE′ binds Co(III)-PPIX with higher affinity than heme, as complexes were observed in the gas phase only when Co(III)-PPIX was bound to CcmE′. A mixture of CcmE′ and Fe(II)-PPIX showed only apo-CcmE′ in its mass spectrum (not shown) under conditions which showed the Co(III)-PPIX+CcmE′ complex. Non-denaturing mass spectrometry also provided the opportunity to examine the stoichiometry of complex formation. The major complexed form of CcmE′ with Co(III)-PPIX was a 1 : 1 complex. No evidence for multimeric protein complexes was observed in the mass spectra.

In order to characterize the interaction of CcmE′ with Co(III)-PPIX we studied its binding to variants of CcmE′ that are affected in their interaction with heme: H130A, which is incapable of binding heme covalently¹¹ and has been found to ligate non-covalently bound Fe(III) heme; H147A, which is affected in heme ligation in the non-covalent heme-protein complex;¹³ Y134F, a replacement of the tyrosine ligand to the heme iron in holo-CcmE′;^12,15 ΔC-CcmE′ that lacks ligands to non-covalently bound heme.²⁰ Only the latter variant was affected in its interaction with Co(III)-PPIX, suggesting that the flexible C-terminal domain of CcmE′ contributes to the binding site for this M-PPIX. Our previous observations have indicated that heme ligation in CcmE is dynamic and that removal of a ligand in the non-covalent complex results in the ligation of the Fe by other residues around the heme-binding site.¹³ The observations made here with CcmE′ variants are consistent with this because removal of potential wild-type heme ligands did not disrupt complex formation with Co(III)-PPIX.

There is only one other known example of the formation of a covalent histidine-heme bond in a protein other than CcmE, namely a form of a cyanobacterial hemoglobin from Synechocystis that binds, via a histidine side chain, to the 2-vinyl group of heme.²⁷ Interestingly, this bond, which has been shown to form in vitro with ferrous heme, does not form with Zn-PPIX. We might have expected that, as Co(III)-PPIX interacts with CcmE and because it is isoelectronic with Fe(II)-PPIX, a Co(III)-PPIX-CcmE covalent bond might form. It seems, however, that Fe is required for the bond formation to occur.

Physiological implications

A recent study examined the effect of some alternative M-PPIXs on cytochrome c biogenesis in E. coli.²² It was found that Zn-PPIX inhibited the Ccm system, but it was concluded that this inhibition occurred after heme release from CcmE and interaction with the putative WWD motif of CcmF was suggested as the target of inhibition. The Zn-PPIX was not found to be incorporated into the cytochrome produced. Various in vitro studies have shown that cytochromes c can tolerate metals other than Fe in the porphyrin ring,^28,29 as well as form thioether bonds in some cases.²³ This would suggest that, because the biogenesis proteins are essential for cytochrome c formation, the Zn-PPIX inhibition observed²² was due to interaction with a Ccm protein. The present work indicates that the point of inhibition was not likely to have been CcmE, as it has no affinity for Zn-PPIX. It is striking that the M-PPIXs that do not bind CcmE in vitro appear to inhibit the Ccm system in in vivo experiments. It is not, however, certain that the heme-binding site of CcmE would be exposed to the periplasm. In fact, it has been shown that extracellularly added heme cannot complement for holo-CcmE formation in a CcmC deletion strain,⁸ indicating that CcmE depends on CcmC for heme delivery and cannot spontaneously acquire periplasmic heme. Presumably Zn-PPIX is able to bind to and inhibit a protein that is downstream in the pathway, which is most likely CcmF (though there is no experimental evidence for it being able to bind the heme that is destined for a c-type cytochrome). Notably, Co-PPIX did not appear to inhibit the production of cytochromes c,²² yet we find high affinity binding of this M-PPIX to CcmE′. This could again be because the CcmE binding site is not accessible for PPIX binding, except to delivery from CcmC, which would depend on the ability of the latter to bind Co-PPIX. Reconciliation of the results of in vivo and in vitro experiments is complicated by the involvement of multiple proteins in vivo. The fact that Zn-PPIX was shown to inhibit both System I (the Ccm system) and System II,²² which is a distinct biogenesis system found in chloroplasts and Gram-positive bacteria and lacks a homologue of CcmE, is further evidence that the site of inhibition was not CcmE in the E. coli system.

In summary, our results show that CcmE is able to discriminate between different metallo-protoporphyrins and is finely tuned to the redox state, which is consistent with its role as a specific heme chaperone. The ability to discriminate between different metals is well-documented in the case of metal chelatases, which catalyze the insertion of metal atoms into porphyrins.³⁰ The absence of an obvious heme-binding pocket in the NMR structure of apo-CcmE′ (Fig. 1) makes the high level of selectivity surprising and suggests that the heme-binding site is more structured than previously thought. Specificity of metal binding is not unusual in nature and provides insight into the structure-function relationships in CcmE as a heme chaperone.

Experimental

Protein preparation

The apo-form of the soluble domain of CcmE lacking its N-terminal membrane anchor (CcmE′) was expressed cytoplasmically in E. coli from plasmid pE151¹¹ and purified as described previously.²⁰ All the CcmE′ mutants used in this work were derived from pE151 by site-directed mutagenesis and have been described elsewhere.^13,20

Spectroscopic analysis

Visible absorption spectra were recorded on a Varian Cary 50 Bio spectrophotometer in 50 mM Tris-HCl buffer, pH 7.4, 150 mM NaCl. Stock solutions (1 mM) of the M-PPIXs (Frontier Scientific) were prepared in DMSO and their concentrations were determined by weight or using the following extinction coefficients: Zn-PPIX in phosphate buffer, pH 7.0 ε₄₁₂ = 87.4 mM cm⁻¹;³¹ Mn(III)-PPIX in 0.1 M NaOH ε₄₆₂ = 25 mM cm⁻¹;³² Co(III)-PPIX in 0.1 M NaOH-pyridine ε₄₂₄ = 180 mM cm⁻¹.³³ Mn- and Co-protoporphyrin were reduced by the addition of disodium dithionite. Solutions were thoroughly sparged with humidified argon and reactions were carried out in the dark.

Other methods

Gel filtration chromatography was carried out on Bio-Gel P-6DG columns (Bio-Rad), according to the manufacturer's instructions. HPLC was performed on a Beckman System Gold apparatus, using a C₁₈ reverse-phase column and absorbance detection at 280 and 420 nm. Elution was performed with a 10–90% acetonitrile gradient in 0.1% aqueous trifluoroacetic acid. Denaturing mass spectrometry was carried out on an LCT Premier (Waters UK Ltd.) time-of-flight mass spectrometer equipped with a NanoMate (Advion Inc., USA) nanoflow electrospray ionization interface. Samples were injected at a concentration of 20 μM in 1 : 1 v/v 1% aqueous formic acid: methanol. Non-denaturing electrospray mass spectrometry (ES-MS) samples were prepared in 10 mM ammonium acetate pH 7 and run using an automated chip-based infusion system (Advion Nanomate Triversa) coupled to a quadrupole time-of-flight tandem mass spectrometer (Waters Q-TOF Synapt). Experimental parameters were adjusted for optimum desolvation conditions while maintaining non-covalent interactions (spray voltage 1.6 kV, backing gas 0.2 mbar, cone voltage 120 V, extraction voltage 1 V, combined injection energies 15 V at a backing pressure of 5.5 mbar). Protein-ligand mixtures were desalted on Micro Bio-Spin 6 columns (Biorad) before injection into the spectrometer.

Abbreviations

Ccm	cytochrome c maturation
CcmE′	the soluble periplasmic domain of CcmE (Ser32 at N-terminus)
PPIX	protoporphyrin IX
M-PPIX	metallo-protoporphyrin IX

Acknowledgements

This work was supported by grant C20071 and BBE0048651 from the UK BBSRC to SJF and JMS. EMH acknowledges an Abraham Newton Scholarship in Biological Sciences from Oxford University. OD was a Junior Research Fellow of Christ Church, Oxford. The authors thank R.J.P. Williams for helpful discussion.

References

1. E. Tolosano and F. Altruda, DNA Cell Biol., 2002, 21, 297–306.
2. S. Cescau, H. Cwerman, S. Letoffe, P. Delepelaire, C. Wandersman and F. Biville, BioMetals, 2007, 20, 603–613.
3. H. Schulz, H. Hennecke and L. Thöny-Meyer, Science., 1998, 281, 1197–1200.
4. L. Thöny-Meyer, Biochemistry, 2003, 42, 13099–13105.
5. L. Thöny-Meyer, F. Fischer, P. Kunzler, D. Ritz and H. Hennecke, J. Bacteriol., 1995, 177, 4321–4326.
6. J. M. Stevens, O. Daltrop, J. W. Allen and S. J. Ferguson, Acc. Chem. Res., 2004, 37, 999–1007.
7. J. W. Allen, O. Daltrop, J. M. Stevens and S. J. Ferguson, Philos. Trans. R. Soc. London, Ser. B, 2003, 358, 255–266.
8. Q. Ren and L. Thöny-Meyer, J. Biol. Chem., 2001, 276, 32591–32596.
9. Q. Ren, U. Ahuja and L. Thöny-Meyer, J. Biol. Chem., 2002, 277, 7657–7663.
10. O. Daltrop, J. M. Stevens, C. W. Higham and S. J. Ferguson, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 9703–9708.
11. J. M. Stevens, O. Daltrop, C. W. Higham and S. J. Ferguson, J. Biol. Chem., 2003, 278, 20500–20506.
12. T. Uchida, J. M. Stevens, O. Daltrop, E. M. Harvat, L. Hong, S. J. Ferguson and T. Kitagawa, J. Biol. Chem., 2004, 279, 51981–51988.
13. J. M. Stevens, T. Uchida, O. Daltrop, T. Kitagawa and S. J. Ferguson, J. Biol. Chem., 2006, 281, 6144–6151.
14. D. Lee, K. Pervushin, D. Bischof, M. Braun and L. Thöny-Meyer, J. Am. Chem. Soc., 2005, 127, 3716–3717.
15. I. Garcia-Rubio, M. Braun, I. Gromov, L. Thöny-Meyer and A. Schweiger, Biophys. J., 2007, 92, 1361–1373.
16. O. Christensen, E. M. Harvat, L. Thöny-Meyer, S. J. Ferguson and J. M. Stevens, FEBS J., 2007, 274, 2322–2332.
17. R. E. Feissner, C. L. Richard-Fogal, E. R. Frawley and R. G. Kranz, Mol. Microbiol., 2006, 61, 219–231.
18. E. Enggist, L. Thöny-Meyer, P. Guntert and K. Pervushin, Structure, 2002, 10, 1551–1557.
19. E. Enggist and L. Thöny-Meyer, J. Bacteriol., 2003, 185, 3821–3827.
20. E. M. Harvat, J. M. Stevens, C. Redfield and S. J. Ferguson, J. Biol. Chem., 2005, 280, 36747–36753.
21. E. Enggist, M. J. Schneider, H. Schulz and L. Thöny-Meyer, J. Bacteriol., 2003, 185, 175–183.
22. C. L. Richard-Fogal, E. R. Frawley, R. E. Feissner and R. G. Kranz, J. Bacteriol., 2007, 189, 455–463.
23. O. Daltrop and S. J. Ferguson, J. Biol. Chem., 2004, 279, 45347–45353.
24. J. Falk, Porphyrins and metalloporphyrins, Elsevier, New York, 1964.
25. M. Sharon and C. V. Robinson, Annu. Rev. Biochem., 2007, 76, 167–193.
26. F. Arnesano, L. Banci, P. D. Barker, I. Bertini, A. Rosato, X. C. Su and M. S. Viezzoli, Biochemistry, 2002, 41, 13587–13594.
27. B. C. Vu, A. D. Jones and J. T. Lecomte, J. Am. Chem. Soc., 2002, 124, 8544–8545.
28. L. C. Dickinson and J. C. Chien, Biochemistry, 1975, 14, 3526–3534.
29. J. M. Vanderkooi, F. Adar and M. Erecinska, Eur. J. Biochem., 1976, 64, 381–387.
30. S. Al-Karadaghi, R. Franco, M. Hansson, J. A. Shelnutt, G. Isaya and G. C. Ferreira, Trends Biochem. Sci., 2006, 31, 135–142.
31. J. J. Leonard, T. Yonetani and J. B. Callis, Biochemistry, 1974, 13, 1460–1464.
32. T. Yonetani and T. Asakura, J. Biol. Chem., 1968, 243, 3996–3998.
33. T. Yonetani, H. Yamamoto and G. V. Woodrow, 3rd, J. Biol. Chem., 1974, 249, 682–690.

---

Footnote

† This article is published as part of a themed issue on Cytochromes, Guest Edited by Norbert Jakubowski and Peter Roos.

---