|Home | About | Journals | Submit | Contact Us | Français|
Feo is a transport system commonly used by bacteria to acquire environmental Fe2+. It consists of three proteins: FeoA, FeoB, and FeoC. FeoB is a large protein with a cytosolic N-terminal domain (NFeoB) that contains a regulatory G protein domain and a helical S domain. The C-terminal region of FeoB is a transmembrane domain that likely acts as the Fe2+ permease. NFeoB has been shown to form a trimer pore that may function as an Fe2+ gate. FeoC is a small winged-helix protein that possesses four conserved cysteine residues with a consensus sequence that likely provides binding sites for the [Fe-S] cluster. Therefore, FeoC is presumed to be an [Fe-S] cluster-dependent regulator that directly controls transcription of the feo operon. Despite the apparent significance of the Feo system, however, the function of FeoC has not been experimentally demonstrated. Here, we show that Klebsiella pneumoniae FeoC (KpFeoC) forms a tight complex with the intracellular N-terminal domain of FeoB (KpNFeoB). The crystal structure of the complex reveals that KpFeoC binds to KpNFeoB between the switch II region of the G protein domain and the effector S domain and that the long KpFeoC W1 loop lies above the KpNFeoB nucleotide-binding site. These interactions suggest that KpFeoC modulates the guanine nucleotide-mediated signal transduction process. Moreover, we showed that binding of KpFeoC disrupts pore formation by interfering with KpNFeoB trimerization. These results provide strong evidence suggesting that KpFeoC plays a crucial role in regulating Fe2+ transport in Klebsiella pneumonia in addition to the presumed gene regulator role.
The Feo system is a transporter commonly used by bacteria to acquire environmental Fe2+ (5, 12, 17). The importance of the Feo system has been confirmed in several bacterial systems. Both Escherichia coli and Salmonella feoB mutants are less able to colonize the mouse intestine, presumably because they cannot transport Fe2+ from the mouse intestine (32, 34). In Helicobacter pylori, FeoB appears to be the main Fe2+ transporter and is required for H. pylori colonization of mouse gastric mucosa, for normal growth, and for Fe2+ uptake under Fe2+-restricted conditions (35). FeoB is also required for intracellular growth of Legionella pneumophila (29) and for virulence of Porphyromonas gingivalis (8). Therefore, various studies have clearly established a role for the Feo system in bacterial colonization of the gut and in virulence.
The feo operon encodes three proteins: FeoA, FeoB, and FeoC. FeoA is a small SH3-like protein predicted to act as a GTPase-activating protein and/or an Fe2+-dependent repressor (5). Recently, FeoA was shown to interact with FeoB and is required for Fe2+ uptake in Salmonella enterica in vivo (18). FeoB is a large protein (773 residues) containing a cytosolic N-terminal domain (NFeoB) that can be divided into a G domain (residues 1 to 170) and an S domain (residues 171 to 270) (9, 22). The C-terminal region of FeoB is a helical transmembrane domain that likely acts as the Fe2+ permease. FeoC is a small (78 residues), hydrophilic, winged-helix protein found only in gammaproteobacteria. In the feo operon, feoC is downstream of feoB. A multiple alignment of various bacterial FeoC amino acid sequences shows that they possess four conserved cysteine residues with a consensus sequence, CxxGxCKxCPx4-7C, that likely provides binding sites for iron in the form of an iron-sulfur ([Fe-S]) cluster (5). Therefore, FeoC is presumed to be an iron-sulfur-dependent regulator that directly controls transcription of the feo operon (5).
Given its importance and novelty, delineating how the Feo system functions is of great interest. Notably, the structures of NFeoB from several species have been solved (3, 11, 13, 14, 19, 27). We previously determined the Klebsiella pneumoniae (KpNFeoB) and Pyrococcus furiosus NFeoB crystal structures with and without bound ligands (14). The NFeoB structures are quite similar between species, and each contains a Ras-like G domain and a helical S domain. The nucleotide-free (apo) protomer and the protomer bound with the nonhydrolyzable GTP analog, GMPPNP, of KpNFeoB and EcNFeoB (E. coli NFeoB) pack as a funnel-shaped trimer in the crystal, with the center containing a cation-binding site (11, 14). The trimer was proposed to be the gateway for Fe2+ transport into the bacterial cell, and the S domain may act as an open/close switch (11).
We also solved the solution structure of K. pneumoniae FeoC (KpFeoC) (15), which consists of a three-stranded β-sheet and three α-helices that fold into a winged helix structure (10). However, the function of FeoC is unknown, and nothing, until this report, has been known concerning how the Feo proteins interact. Here, we show that KpFeoC forms a tight 1:1 complex with KpNFeoB. We also report the crystal structure of the KpNFeoB/KpFeoC complex and show that KpFeoC interacts with residues near switch II, the nucleotide-binding site, and the S domain of KpNFeoB. Furthermore, binding of KpFeoC interferes with the formation of the KpNFeoB trimer. These results suggest that KpFeoC possesses regulatory roles in modulating Fe2+ transport.
KpNFeoB was expressed and purified as described previously (14). Briefly, E. coli cells harboring the gene encoding KpNFeoB in a pET24a- or a pTAC-type vector were grown at 37°C in Luria broth in the presence of 100 μg/ml ampicillin. Protein expression was induced by addition of 1.0 mM isopropyl-β-d-thiogalactopyranoside (IPTG; final concentration) when a culture reached an optical density at 600 nm (OD600) of ~0.6. Each culture was incubated for an additional 4 h and harvested by centrifugation. Each cell pellet was suspended, lysed with a microfluidizer Model 110S (Microfluidics Corp.), and centrifuged at 16,000 rpm (Beckman rotor JA25.5) at 4°C for 20 min. The supernatants were loaded onto a DEAE-Sepharose open column (bed volume, 10 ml; GE Healthcare). After washing the column with 20 mM Tris, pH 7.5, 100 mM NaCl, protein was eluted in 20 mM Tris, pH 7.5, 300 mM NaCl. The eluted protein solution was concentrated, and the protein was subsequently denatured by the addition of a solution containing 20 mM NaPi, 50 mM NaCl, and 8 M urea (pH adjusted to 11). The solution was then dialyzed against 20 mM Tris, pH 7.5, 50 mM NaCl to promote refolding of KpNFeoB, which was then further purified using size-exclusion chromatography (SEC; 120-ml Superdex 75; GE Healthcare.) The purity of KpNFeoB was verified by SDS-PAGE (Fig. 1, inset) and mass spectrometry (Voyager-DE STR; PerSeptive Biosystems, MA).
KpFeoC was expressed in E. coli as a glutathione S-transferase-tagged protein and purified through a glutathione Sepharose 4B column (GE Healthcare) as described previously (15). The tag was removed by incubation of the loaded resin with PreScission Protease (GE Healthcare), after which KpFeoC was eluted with pH 7.5 buffer containing 20 mM Tris and 50 mM NaCl from the column. The protein was further purified using SEC through a HiLoad 16/60 Superdex 75 column. The purified protein contained five additional N-terminal residues (GPLGS) (83 residues in total). Protein purity was verified by SDS-PAGE and mass spectrometry. 15N/13C-labeled KpFeoC proteins were prepared by growing the cells at 37°C in M9 minimal medium supplemented with 15NH4Cl and either 13C6-d-glucose or 12C6-d-glucose. IPTG (1.0 mM) was utilized for the induction of protein expression when the OD600 of culture reached 0.6 to ~0.8. The culture was grown for an additional 20 h at 15°C before harvesting. For preparing 2H-, 15N-KpFeoC, E. coli cells were grown in D2O supplemented with perdeuterated glucose and 15NH4Cl.
The KpNFeoB mutants, including R100A, D238A, D239A, and A158C/A254C, were generated by using a previously reported protocol (36). The primers containing desired mutations were utilized for PCRs with 6.5 min of extension time using KOD DNA polymerase (Novagen). The PCR products were treated with DpnI (Fermentas) at 37°C for 1 h to destroy the template DNA. The authenticities of all constructed mutants were further confirmed by nucleotide sequencing.
Size-exclusion chromatography experiments were performed at 4°C on a GE AKTA fast protein liquid chromatography (FPLC) purifier 10 equipped with a Superdex-75 HR 10/30 column (GE Healthcare). One hundred μM KpNFeoBA158C,A254C monomer was incubated with or without 100 μM KpFeoC at 4°C for 30 min in 20 mM Tris buffer (pH 7.5) containing 50 mM NaCl. For those Cu2+-catalyzed experiments, 0.5 mM Cu2+ was next added to both samples with an additional 30-min incubation at 25°C before being loaded onto a column (26). Proteins were eluted with the same buffer at 4°C at a flow rate of 0.5 ml/min and detected at a UV wavelength of 280 nm. Protein mass was calibrated with protein standards of molecular masses ranging of 6.5 to 75 kDa. The elution with molecular sizes corresponding to KpNFeoBA158C,A254C trimer and KpNFeoBA158C,A254C-KpFeoC complex was further verified by SDS-PAGE analysis with or without 100 mM β-mercaptoethanol (BME) (see Fig. 6 and and77).
Nuclear magnetic resonance (NMR) spectra were acquired at 15°C on Bruker Avance 500- and 600-MHz spectrometers each equipped with a triple resonance cryogenic probe. Samples were 0.2 to 1.0 mM protein in 20 mM sodium phosphate, pH 6.5, 50 mM NaCl, 10% D2O, with or without 10 mM dithiothreitol (DTT). 1H chemical shifts were externally referenced to the methyl resonance of 2,2-dimethyl-2-silapentane-5-sulfonate (0 ppm), and 13C and 15N chemical shifts were indirectly referenced according to IUPAC recommendations (21). NMR data were processed using Topspin software, and crosspeaks were analyzed using SPARKY 3 (T. D. Goddard and D. G. Kneller, University of California, San Francisco, CA). The 1H, 15N, and 13C resonances of the DTT-reduced KpFeoC at pH 6.5 and 15°C were assigned using standard multidimensional heteronuclear NMR techniques as previously reported (6, 33) (resonance assignments were deposited in the BioMagResBank under accession number 15634 ).
Isothermal titration calorimetry (ITC) was performed with an ITC200 calorimeter (MicroCal Inc.) at 15°C. Samples were centrifuged prior to titration. The titrations were carried out by injecting 1 μl (first injection) or 2.5 μl (injections 2 to 19) of a KpFeoC solution (500 μM) into the sample cell filled with a 50 μM KpNFeoB solution. The protein solutions contained 20 mM Tris, pH 7.5, 50 mM NaCl, with or without 1 mM tris(2-carboxyethyl)phosphine (TCEP). After an initial delay of 120 s, injections were made every 180 s. ITC binding curves were fit to a single-site binding equation using Microcal Origin software.
The hanging-drop vapor-diffusion method (23) was used for crystallization. ApoKpNFeoB was mixed with apoKpFeoC at a 1:1 molar ratio in 20 mM Tris-HCl, pH 7.5, 50 mM NaCl. The final concentration of the KpNFeoB/KpFeoC complex was 10 mg/ml. The KpNFeoB/KpFeoC solution (1 μl) and the reservoir (1 μl) solution were mixed and equilibrated against 500 μl of the reservoir solution (100 mM Tris-HCl, pH 8.5, 1 M Li2SO4, 10 mM NiCl2) in wells of a Linbro plate. Crystals appeared 1 week later and were held in a reservoir solution that also contained 12% (vol/vol) glycerol before data collection. Data were collected at beamline BL13C1 at the National Synchrotron Radiation Research Center in Taiwan and processed using the program HKL2000 (25). The KpNFeoB/KpFeoC crystals belong to the tetragonal space group P43 with unit cell parameters of a = b = 59.69 Å and c = 141.14 Å and were diffracted to 2.3-Å resolution. The specific volume (VM) (28) is 3.19 Å3 Da−1 with a solvent content of 61.5%. The asymmetric unit contains one KpNFeoB/KpFeoC complex (Table 1).
The initial phase of KpNFeoB/KpFeoC was obtained using the molecular replacement software AutoMR in PHENIX (1, 2) and the KpNFeoB (PDB code 2WIA) structure as the search model. XtalView (24) was used to manually model the KpNFeoB coordinates to the KpNFeoB electron density in the KpNFeoB/KpFeoC map. After initial model refinement using the program CNS (4), three distinguishable KpFeoC helices were located in the |Fo|−|Fc| and 2|Fo|−|Fc| electron density maps. Following several cycles of refinement and model building, the structure of the complex was refined to an Rwork of 21.3% for all reflections of >2σ and between 25.52- and 2.30-Å resolution. An Rfree of 22.1% was calculated using 5.0% random reflections that had not been used in the refinement. The final structure of the complex includes Gln2 to Ile66, Thr70 to Ala125, and Val130 to Ser260 for KpNFeoB and Met1 to Glu51, Gln73 to Leu80, and the N-terminal tag residues Pro−3 to Ser0 for KpFeoC. The complex contains 132 water molecules, 3 Ni ions, and 1 sulfate ion. The Ramachandran plots (28) for the KpNFeoB and KpFeoC molecules in complex structures did not violate allowed backbone torsion angles. Refinement statistics are summarized in Table 1.
Coordinates and structure factors of the KpNFeoB/KpFeoC complex have been deposited in the Protein Data Bank under accession number 4AWX.
We first studied KpFeoC/KpNFeoB complex formation by size-exclusion chromatography (SEC) (Fig. 1). The elution profiles of KpNFeoB and KpFeoC indicated that, when isolated in solution, these two proteins each exist as monomers of ~30 and ~10 kDa, respectively, and that a 1:1 molar mixture of KpNFeoB and KpFeoC forms a tight binary complex with a molecular mass of ~40 kDa.
We next employed ITC to characterize the binding isotherm of complex formation (Fig. 2). The thermogram was fit to a single-binding-site model with an enthalpy change (ΔH) of 3.87 ± 0.08 kcal/mol, an entropy change (ΔS) of 12.17 kcal/mol, and ΔG of −8.3 kcal/mol, corresponding to a dissociation constant (Kd) of 0.5 ± 0.2 μM (Table 2). Thus, binding of KpFeoC to KpNFeoB is strongly entropy driven. Reduction of the cysteines in the W1 loop did not affect the binding isotherm (data not shown), in agreement with our observation that the W1 loop is not involved in KpNFeoB recognition.
We previously assigned the NMRs and determined the solution structure of KpFeoC (15). Here, we employed NMR chemical shift perturbation data to identify the KpFeoC residues that reside at the complex interface. In the presence of unlabeled KpNFeoB, many of the resonances associated with free 2H-, 15N-labeled KpFeoC disappeared (blue peaks in Fig. 3A), and many new, weak resonances appeared (red peaks in Fig. 3B), which probably are the resonances of residues affected by KpNFeoB binding. The spectral change upon KpNFeoB binding indicates that the KpNFeoB dissociation from the binary complex is slow on the NMR time scale and the binding stoichiometry was 1:1, since no spectral change was observed at a molar ratio greater than 1:1. The observation that the KpFeoC resonances are relatively weak is probably the consequence of faster transverse relaxation times (T2) for the amide resonances in the larger binary complex, which would cause signal loss prior to detection. The chemical shift perturbations for all resonances have yet to be determined, because we have not fully assigned all resonances of the complex. Nonetheless, the resonances of certain residues could be grouped as strongly perturbed (Δσ of >0.05 ppm for M1, A2, S3, L4, M5, E6, V7, D9, M10, A12, Q14, Q21, Q27, L32, I33, D34, A35, R39, M40, A42, M43, V47, and R79) and as weakly perturbed (Δσ of <0.05 ppm for R8, L11, G15, M17, E18, A19, S23, R25, L26, M36, G44, K45, I49, W75, W76, and L78). Mapping these chemical shift perturbations onto the KpFeoC structure showed that the strongly perturbed residues are located primarily in the α1 helix at the N terminus and the α3 helix (colored red in Fig. 3B). Notably, the W1 loop (Thr52-Arg72) was not affected by KpNFeoB.
The X-ray structure of the apoKpNFeoB/KpFeoC complex (referred to as KpNFeoB/KpFeoC throughout the remainder of the report) was determined at 2.3-Å resolution (Table 1). The structure of isolated apoKpNFeoB (14) was used as the template for phase determination of KpNFeoB in the complex. Figure 4A depicts a ribbon representation of the structure of the KpNFeoB/KpFeoC complex in stereo. One KpNFeoB/KpFeoC heterodimer, one sulfate ion (red spheres), and three Ni2+ ions (magenta spheres) are present per asymmetric unit. The protein molecules are packed so that one KpNFeoB molecule is surrounded by three KpFeoC molecules (Fig. 4A). The surface areas of the three interfaces, corresponding to complexes 1, 2, and 3, respectively, as calculated by PISA (20), are 866, 344, and 258 Å2. Both gel filtration and NMR results indicated that the two proteins form a 1:1 complex. Comparison to the NMR chemical shift perturbation data indicated that complex 1, which has the largest contact interface area, is the correct complex (colored red in Fig. 4A). Specifically, NMR results showed that the strongly perturbed residues are located at helices α1 and α3 of KpFeoC. Only complex 1 has both helices α1 and α3 in contact with KpNFeoB, while in complexes 2 and 3 the α3 helix is far away from the interface. The results clearly establish complex 1 as the correct structure for the KpNFeoB/KpFeoC complex.
To further corroborate the observation, we mutated the three interfacial residues in KpNFeoB involved in intermolecular hydrogen bonding to generate three alanine mutants, KpNFeoBR100A, KpNFeoBD238A, and KpNFeoBD239A. The KpFeoC binding affinities of these three mutants were determined by ITC, and the thermodynamic parameters extracted from these data are reported in Table 2. Surprisingly, only KpNFeoBD239A produced a 5-fold decrease in the equilibrium dissociation constant (Kd), while the binding affinities of the other two mutants were not altered. However, closer inspection of the thermodynamic data indicated that mutation at Arg100, Asp238, or Asp239 generated marked ΔH of −1.6, 2.91, and 1.68 kcal/mol and TΔS of 1.42, −2.78, and −0.67 kcal/mol for KpNFeoBR100A, KpNFeoBD238A, and KpNFeoBD239A, respectively. The compensatory changes in enthalpy and entropy of R100A and D238A mutants resulted in no change in binding affinity. Nonetheless, the observed large changes in enthalpy and entropy of these three mutants support their roles as important interfacial residues, and complex 1 is the correct KpNFeoB/KpFeoC complex.
The structure of KpNFeoB in the complex is similar to that of free apoKpNFeoB (Fig. 4B). The N-terminal ~170 residues with six helices (H1 to H6) and a seven-strand β-sheet (B1 to B7) (G domain) assumes a G protein fold, which was highly homologous to the structure of the canonical G protein. The C-terminal five helices (H7 to H11; residues 171 to 270) (S domain) assumed a compact hammer-shaped helix-bundle structure, with the long H11 helix positioned as the handle. The S domain interacted extensively with G domain residues opposite the nucleotide-binding pocket and near the switch II region (14). The root mean square deviations (RMSD) of the α carbons for KpFeoB in the complex and apoKpNFeoB, GDP/KpNFeoB, or GMPPNP/KpNFeoB are 0.786, 0.564, and 0.744 Å, respectively (Fig. 4C). In particular, the functionally important G1, G3, and G4 motifs and switch I are not perturbed by KpFeoC binding. As with all other KpNFeoB structures, except that of the GDP-bound form, the switch II residues are not observed in the KpNFeoB/KpFeoC binary complex. Nevertheless, there are several distinctive features for the KpFeoC-complexed KpNFeoB. (i) The KpNFeoB G5 motif in the complex is in the in conformation (bent toward the G1 motif), similar to that observed for the nucleotide-bound form, such as in GDP/KpNFeoB and GMPPNP/KpNFeoB complexes, whereas the motif is in the out conformation (flipped away from the G1 motif) in apoKpNFeoB. Since the G5 motif coordinates with the guanine nucleotide base, it will be interesting to investigate whether binding to KpFeoC affects the nucleotide-binding affinity/specificity of KpNFeoB. (ii) In the complex, the S and G domains move closer together, with changes in the positions of H9 and H10 and the C terminus of H11 in the S domain most affected (Fig. 4C). Switch II is positioned between the α1 helix of KpFeoC and switch I. This positioning is consistent with a proposal made by Guilfoyle and colleagues concerning the mechanism by which the G protein affects Fe2+ transport (11), i.e., GTP-Mg2+ binding in the nucleotide-binding pocket of the G domain causes a structural change in switch I, and/or switch II is transmitted to the S domain, which directly modulates gating of the intracellular Fe2+ pore. (iii) H4, adjacent to the G5 motif, is not observed in the complex, which means that this helix was not formed or was very mobile. Notably, H4 is present in all other KpNFeoB structures (its usual position is indicated by the dashed red circle in Fig. 4B). (iv) Finally, three Ni2+ ions, which are present in the crystallization solvent, coordinating His50, His79, and His243 (magenta spheres), are present. The biological significance of the aforementioned distinctive features is unclear at present.
The crystal structure of KpFeoC in the complex contains a three-strand β sheet and a three-helix bundle and folds in the winged-helix-type topology, α1-β1-α2-α3-β2-β3, similar to that observed in solution (15). The 20-residue W1 loop, which connects the two β strands, is not observed. Superpositioning of the crystal and solution structures shows that they are rather similar, with an RMSD of 1.543 Å for the Cα atoms (Fig. 4D) (15). Nonetheless, there are several notable differences between the two structures. First, the orientation of the β2-β3 sheet in the crystal structure is significantly tilted away from that of free KpFeoC by ~30°, which would put the W1 loop closer to the nucleotide-binding site. We speculate that the tilting could be caused by interactions between residues in the W1 loop and those in the switch II and/or nucleotide-binding pocket. However, this could not be confirmed, since the W1 loop and several switch II residues could not be observed in the crystal. Second, the orientations and positions of α2 differ substantially in the two structures. Third, in the solution structure, residues Arg16-Glu18 form the short β1 strand, which is absent from the crystal structure. Intriguingly, superpositioning the KpFeoC solution structure onto the KpFeoC structure in the complex shows that the flexible W1 loop residues can reach and interact with residues in the nucleotide-binding site, thereby sensing the structural information induced by nucleotide binding to KpNFeoB (Fig. 4E). The presence of four cysteine residues with the consensus sequence CxxGxCKxCPx4-7C and their structural arrangement (Fig. 3B) suggest that the W1 loop provides favorable binding sites of metal ions and/or [Fe-S] clusters. Such metal complexes and/or iron-sulfur clusters, if they indeed exist, could affect the interaction of the W1 loop with surrounding residues and provide additional layers of regulation in the G protein signal transduction process.
KpFeoC binds to KpNFeoB between the G domain, the S domain, and switch II, with the flexible W1 loop residues potentially being capable of reaching and interacting with residues in the nucleotide-binding site (Fig. 4E). Specifically, helix α3 of KpFeoC extensively interacts with KpNFeoB H3, H11, and switch II (Fig. 5A) by forming 10 hydrogen bonds, 7 salt bridges, and several hydrophobic interactions. Hydrogen bonds are formed by Ser3, Leu4, Arg8, Ala35, Glu38, and Arg39 of KpFeoC and Ser62, Arg100, Asp238, Asp239, Asp246, and Tyr249 of KpNFeoB, respectively. Arg8, Glu38, and Lys45 of KpFeoC form salt bridges with Arg100, Asp238, and Asp239 of KpNFeoB, respectively (Fig. 5A and andB).B). A hydrophobic interaction is observed between α3 in KpFeoC and H11 in KpNFeoB (Fig. 5B). The locations of many of the residues in switch II are not seen in the crystal structure, therefore no interactions between switch II and KpFeoC residues could be identified.
apoEcNFeoB and 2′(3′)-O-(N-methylanthraniloyl) (mant)-GMPPNP-bound EcNFeoB crystallize as a funnel-like trimeric assembly (11), which forms a cytoplasmic accessible pore ~20 Å in length and ~1.2 Å in diameter at its narrowest point, with the Glu133 residues from the three subunits coordinating Mg2+. This pore structure may facilitate the passage of Fe2+ into the cytoplasm with the Glu133 residues forming the gate. In G proteins, e.g., Ras, conformational changes in switch I and switch II caused by nucleotide binding are believed to be transmitted to an effector site that then exerts the biological effect (30, 31). In FeoB, H11 in the S domain is situated next to switch II and may act as the effector (11, 14). We also found that crystals of apoKpNFeoB and its GMPPNP complex form trimeric structures (14). Superpositioning of the KpNFeoB/KpFeoC structure onto that of the KpNFeoB trimer shows that KpFeoC would bind to a site near the trimer interface (Fig. 5C), which raised the possibility that binding of KpFeoC affects KpNFeoB trimer formation.
To further corroborate the biological role(s) for KpNFeoB/KpFeoC complex formation, we investigated the effect of KpFeoC binding on KpNFeoB trimer formation using a published approach (11, 14). We mutated Ala158 and Ala254, which are intermolecularly near each other (3.7 Å between the two Cβ atoms) at the interfaces of the KpNFeoB trimer, to cysteines (KpNFeoBA158C,A254C), which can be oxidized to form a disulfide bond. Notably, Ala158 and Ala254 are spatially distant (26.9 Å between the two Cβ atoms) in the monomer and cannot be intramolecularly cross-linked. KpNFeoBA158C,A254C formed a covalently linked trimeric structure when incubated at 4°C for a week (Fig. 6A and andC).C). We then determined if KpFeoC could bind trimeric KpNFeoBA158C,A254. When KpFeoC was added to preformed KpNFeoBA158C,A254C trimer, SEC of the mixture revealed two peaks that correspond to the KpNFeoBA158C,A254C trimer and KpFeoC, indicating that KpFeoC and the KpNFeoBA158C,A254C trimer did not associate (Fig. 6B and andC).C). This is consistent with the observation that the KpFeoC binding site overlaps with the trimerization site of KpNFeoB (Fig. 5C). The Cys variant of NFeoB allows NFeoB to form a disulfide-linked covalent trimer complex, and the formation of the NFeoB trimer completely blocks the FeoC binding site. Addition of Cu2+, which facilitates cross-linking of cysteines, caused formation of the KpNFeoBA158C,A254C trimer within 30 min (Fig. 7). However, when Cu2+ was added to a solution containing a mixture of KpNFeoBA158C,A254C and KpFeoC, the amount of trimer present was reduced and free KpFeoC was detected (Fig. 7B and andC).C). Therefore, KpFeoC appears able to retard KpNFeoB trimer formation, most likely by binding to monomeric KpNFeoB, whereas KpFeoC cannot bind trimeric KpNFeoB (solid line in Fig. 7B).
FeoC contains four conserved cysteines in the W1 loop with a consensus sequence and spatial arrangement that likely provide a binding site for an iron-sulfur cluster (Fig. 2B) (5). Consequently, FeoC has been assumed to be an iron-sulfur cluster-dependent regulator of transcription that directly controls the expression of the feo operon (5). However, such an iron-sulfur cluster in FeoC has not been experimentally identified. Nonetheless, iron-sulfur clusters often are not stable under aerobic conditions, and their formation may require biosynthetic machinery (7, 16). Therefore, the possible role of FeoC as an iron-sulfur cluster-dependent transcriptional regulator remains to be confirmed. Our present observations, that KpFeoC tightly binds to monomeric KpNFeoB at a strategic site that permits KpFeoC to interact with the nucleotide-binding site, switch II, and S domain helices, which may act as an effector, raise the possibility that KpFeoC modulates signal transduction initiated by nucleotide binding and results in the regulation of Fe2+ transport. More interestingly, our finding that KpFeoC does not bind to the trimeric KpNFeoB and that the KpNFeoB/KpFeoC complex does not form a trimer further indicates that KpFeoC can also regulate pore formation. Our present results provide strong evidence that KpFeoC plays multiple roles in mediating signal transduction, pore formation, and perhaps transcriptional regulation of the feo gene. However, much more needs to be done to substantiate the current findings and to further explore the functional consequences.
This work was supported by grants (NSC 100-2311-B-001-023 to T.-H.H. and NSC98-2311-B-001-009-MY3 to C.-D.H.) from the National Science Council of the Republic of China.
NMR experiments were carried out at the High-Field Nuclear Magnetic Resonance Center (Taipei, Taiwan), supported by the National Research Program for Genomic Medicine of the National Science Council of the Republic of China. We are grateful for access to the synchrotron radiation beamline 13C1 at the National Synchrotron Radiation Research Center, Hsinchu, Taiwan.
Published ahead of print 28 September 2012