Article pubs.acs.org/biochemistry
CD/MCD/VTVH-MCD Studies of Escherichia coli Bacterioferritin Support a Binuclear Iron Cofactor Site Yeonju Kwak,† Jennifer K. Schwartz,† Victor W. Huang,‡ Emily Boice,‡ Donald M. Kurtz, Jr.,*,‡ and Edward I. Solomon*,† †
Department of Chemistry, Stanford University, Stanford, California 94305, United States Department of Chemistry, University of Texas at San Antonio, One UTSA Circle, San Antonio, Texas 78249, United States
‡
S Supporting Information *
ABSTRACT: Ferritins and bacterioferritins (Bfrs) utilize a binuclear non-heme iron binding site to catalyze oxidation of Fe(II), leading to formation of an iron mineral core within a protein shell. Unlike ferritins, in which the diiron site binds Fe(II) as a substrate, which then autoxidizes and migrates to the mineral core, the diiron site in Bfr has a 2-His/4carboxylate ligand set that is commonly found in diiron cofactor enzymes. Bfrs could, therefore, utilize the diiron site as a cofactor rather than for substrate iron binding. In this study, we applied circular dichroism (CD), magnetic CD (MCD), and variable-temperature, variable-field MCD (VTVH-MCD) spectroscopies to define the geometric and electronic structures of the biferrous active site in Escherichia coli Bfr. For these studies, we used an engineered M52L variant, which is known to eliminate binding of a heme cofactor but to have very minor effects on either iron oxidation or mineral core formation. We also examined an H46A/D50A/M52L Bfr variant, which additionally disrupts a previously observed mononuclear non-heme iron binding site inside the protein shell. The spectral analyses define a binuclear and an additional mononuclear ferrous site. The biferrous site shows two different five-coordinate centers. After O2 oxidation and re-reduction, only the mononuclear ferrous signal is eliminated. The retention of the biferrous but not the mononuclear ferrous site upon O2 cycling supports a mechanism in which the binuclear site acts as a cofactor for the O2 reaction, while the mononuclear site binds the substrate Fe(II) that, after its oxidation to Fe(III), migrates to the mineral core.
N
result in weaker iron binding, thereby facilitating release of the biferric product to form the mineral core (Figure 1a). In bacteria, three types of ferritins have been identified: bacterial ferritin, Dps (DNA protection during starvation protein, also termed mini-ferritin), and bacterioferritin (Bfr).34,35 The structure and function of bacterial ferritin closely resemble those of its eukaryotic counterparts. Bfr, which has so far been found only in bacteria, has the 24-subunit nanocage structure characteristic of ferritins, but unlike ferritins, Bfr contains 12 identical heme binding sites. These sites are located between two subunits with the heme iron axially ligated by methionine sulfurs from each subunit.36 The functional role of this heme is not clear, but it may assist in reductive iron release.37,38 A second distinctive feature is the binuclear active site of Bfr. While this site is found at a position analogous to that in ferritins and catalyzes a ferroxidase reaction, the binuclear site of Bfrs has the 2-His/4-carboxylate ligand set with two bridging carboxylates39−43 commonly found in diiron cofactor sites of O2-activating enzymes (Figure 1b). The kinetics of the ferroxidase and mineralization reactions of
umerous enzymes containing non-heme diiron active sites catalyze several types of O2-activating reactions, including H atom abstraction,1 hydroxylation,2−4 and desaturation.5 These enzymes include ribonucleotide reductase,6 soluble methane monooxygenase,2 and Δ9-desaturase.7 Crystallographic8−10 and spectroscopic11−14 studies have shown that the diiron sites in these enzymes share a common 2-His/4carboxylate ligand set with two bridging carboxylate ligands in their biferrous forms. Although these O2-activating enzymes have been included within the broad “ferritin-like superfamily”,15,16 ferritins function as iron storage proteins, using a diiron site to bind “substrate” Fe(II) and catalyze a ferroxidase reaction rather than O2 activation.17 Ferritins consist of 24 subunits assembled into a protein nanocage that typically stores up to 3000 irons in the form of ferric oxyhydroxide mineral phases.18 A binuclear ferroxidase site within the ferritin subunits catalyzes autoxidation of Fe(II), and the oxo/hydroxo biferric product is released to nucleate and grow the mineral core. Numerous crystallographic and spectroscopic studies19−33 indicate that this ferroxidase activity is related to the diiron coordination sphere of ferritin, which is distinct from those of the O2activating diiron cofactor sites; ferritin binuclear sites have only one bridging carboxylate and only one His ligand, which could © 2015 American Chemical Society
Received: September 19, 2015 Revised: November 9, 2015 Published: November 9, 2015 7010
DOI: 10.1021/acs.biochem.5b01033 Biochemistry 2015, 54, 7010−7018
Article
Biochemistry
Figure 1. Biferrous non-heme iron active sites based on the spectroscopic study of (a) bullfrog M ferritin26 and the X-ray crystal structure of (b) Escherichia coli Bfr.41 While M ferritin has one His and one bridging carboxylate ligand, which could result in weaker iron binding, Bfr has a 2-His/4carboxylate ligand set with two bridging carboxylates, which is commonly found in diiron cofactor sites of O2-activating enzymes.
Escherichia coli Bfr support a cofactor function in which the binuclear site catalyzes oxidation of substrate Fe(II) that has penetrated the protein shell by other pathways.44,45 However, it is not clear whether the binuclear site serves as a cofactor rather than a substrate binding site in all Bfrs.46,47 A significant gap in knowledge involves the geometric and electronic structures of the biferrous site in Bfr and whether these structures are maintained upon redox cycling in solution. In the study presented here, we applied the combination of circular dichroism (CD), magnetic CD (MCD), and variabletemperature, variable-field (VTVH) MCD spectroscopies to define the geometric and electronic structures of the biferrous active site and of an additional ferrous binding site in E. coli Bfr. CD and MCD spectroscopies probe the ligand-field (LF) transitions, which are characteristic of the coordination sphere of each iron in the active site. The VTVH-MCD behavior of the LF transitions determines the zero-field splitting [ZFS, D (axial), E (rhombic)] of each ferrous center and the exchange coupling (J) between the two ferrous centers, the latter of which reflects the bridging ligation. For these studies, we used an engineered heme free variant, M52L Bfr, which eliminates interfering heme spectroscopic signals but showed only minor differences in iron oxidation and mineralization activities compared to those of the wild-type protein.36 We also examined an additional engineered variant H46A/D50A/ M52L Bfr. An X-ray crystal structure showed a mononuclear iron with H46 and D50 ligands on the inner surface of the Bfr protein shell.41 Also, the kinetic study of the H46A/D50A variant showed a mineralization reaction significantly slower than that of the wild-type protein. Comparisons of our spectroscopic results on Bfr with those for the biferrous active site in ferritin provide new insights into the distinctive ferroxidation and iron mineralization mechanisms of E. coli Bfr.
The plasmid used to express M52L Bfr was generated from a wild-type E. coli Bfr expression plasmid 48 using the QuikChange method (Agilent Technologies). The gene encoding the H46A/D50A/M52L Bfr variant was synthesized and inserted into the same parent expression plasmid (pT7-7) by GenScript (Piscataway, NJ).36,41 Nucleotide sequences of the inserted Bfr variant genes were confirmed by sequencing at either the University of Georgia Molecular Genetics and Instrumentation Facility or by GenScript. The Bfr variant plasmids were transformed into E. coli BL21(DE3), and the Bfr variants were expressed from cultures of these transformed E. coli strains in M9 medium. The Bfrs were isolated and purified by published procedures.48 The as-isolated proteins contained no detectable heme but typically contained residual non-heme iron (
