Crystal structure of <i>Escherichia coli</i> YfhJ protein, a member of the ISC machinery involved in assembly of iron–sulfur clusters

Yoshimitsu Shimomura, Yasuhiro Takahashi, Yoshimitsu Kakuta, Keiichi Fukuyama

Proteins Structure Function and Bioinformatics · 2005 · 29 citations · 29 references

DOIFull text

Open access

Concepts

Abstract

Iron–sulfur (FeS) proteins are ubiquitously distributed in organisms, playing essential roles in various biological processes as electron carriers, enzymes, and sensors of oxygen and iron.1-3 These functions depend on the chemical versatility of their cofactors, FeS clusters, which consist of several iron and sulfur atoms. Recent genetic and biochemical analyses have revealed that several enzymatic systems are involved in vivo in the synthesis of FeS clusters and in their incorporation into apo FeS proteins; three FeS cluster assembly machineries (NIF, ISC, and SUF) have so far been identified. In the NIF system, NifS and NifU are required for the formation of metalloclusters of nitrogenase in Azotobacter vinelandii.4-6 Furthermore, the NIF-like machinery has been shown to work in the maturation of a wide variety of FeS proteins.7 The ISC machinery, discovered from the homology between NifS and IscS,6, 8 has broad specificity and targets general FeS proteins.9, 10 The third SUF machinery was identified as an alternative pathway of FeS cluster assembly in Escherichia coli.11 The ISC machinery is present in α-, β-, and γ-proteobacteria and the mitchondria of eukaryotes, whereas the SUF machinery is widely distributed among archaebacteria, eubacteria, and the plastids of eukaryotes.7 The components of the ISC machinery are encoded in the so-called isc operon (iscRSUA-hscBA-fdx-yfhJ). This machinery was unambiguously demonstrated to be involved in the FeS cluster assembly; the overexpression of the isc operon increases the production of recombinant FeS proteins, whereas mutation of the operon decreases the activity of FeS proteins.9, 10, 12 IscS is a cysteine desulfurase, which releases a sulfur atom from L-cystein and serves as a sulfur donor for the FeS cluster assembly reaction.6, 8, 13 IscU and IscA bind unstable FeS clusters during the in vitro reconstitution process and appear to function as scaffolds in the FeS cluster assembly process.14, 15 HscA is a molecular chaperone16 and HscB is a co-chaperone.17 Fdx is a [2Fe-2S]-type ferredoxin.18 IscR is a transcription factor that regulates expression of the isc genes.19 YfhJ (also called ORF3 or IscX) is a small, acidic protein of 66 residues, and pI = 3.7. The homologs of YfhJ are distributed in proteobacteria categorized in the β- and γ-subdivisions, in which the gene order (iscRSUA-hscBA-fdx-yfhJ) is highly conserved in the respective chromosomes as well as the acidic character of the protein. Two-hybrid and pull-down experiments have shown a specific interaction between YfhJ protein and IscS,20 indicating that YfhJ protein is involved in the ISC function. However, there is limited information concerning how YfhJ protein functions, since the mutation of the yfhJ gene does not show any conspicuous phenotype.12 Here we report a crystal structure of YfhJ from E. coli at 1.75 Å resolution and the characteristic charge distribution on its protein surface. The coding region of YfhJ was amplified by polymerase chain reaction (PCR) with the primers 5′-GGATCCGGACTTAAGTGGACCGATAG-3′ and 5′-GTCGACTATTATTCGGCCTCGTCCAGC-3′, where the underlined sequences are restriction sites of BamHI and SalI, respectively. The PCR product was cloned into the pCR2.1-TOPO vector (Invitrogen) using a TA cloning method, and the sequence was verified. The plasmid was digested with BamHI and SalI, and the fragment was cloned into the corresponding sites of the pQE-30 vector (Qiagen) to construct the pQ-ORF3 plasmid expressing the N-terminal (His)6-tagged YfhJ protein. This plasmid was introduced into E. coli M15 cells harboring pREP4, carrying the lacI gene. The cells were cultured at 32°C for 10 h in Terrific broth supplemented with 50 μg/mL ampicillin, 20 μg/mL kanamycin, and 0.5 mM IPTG. The pQ-ORF3 plasmid was also introduced into E. coli B834(DE3) cells auxotrophic for methionine, to produce the selenomethionine (SeMet)-labeled YfhJ protein. The cells were cultured at 32°C for 30 h in minimal medium containing SeMet instead of methionine. The (His)6-tagged YfhJ protein was purified by Ni-NTA (Qiagen) chromatography followed by gel filtration on a Sephacryl S-200 High Resolution column (Amersham Biosciences). The purified YfhJ protein solution was concentrated to 5.3 mg/mL with Centriprep and Centricon (Millipore). Crystallization conditions were surveyed using the hanging-drop vapor-diffusion method with the crystal screen kits (Hampton Research) at 20°C. The drops, each containing 1 μL of the protein solution and 1 μL of the reservoir solution, were equilibrated against the reservoir solutions. The YfhJ crystal was produced when the reservoir solution of 1.6M sodium citrate (pH 6.3) was used. The SeMet-labeled YfhJ protein crystal was produced under the same condition as the native protein. All X-ray measurements were carried out at 100K. YfhJ crystals were soaked in mineral oil, mounted in cryo-loops, and flash-cooled. The native crystals belong to the space group C2, with unit cell parameters of a = 76.63 Å, b = 27.25 Å, c = 39.68 Å, and β = 111.0°. Diffraction data of the native crystal and the SeMet-labeled crystal were collected with synchrotron radiation at BL40B2 and BL44B2, respectively, at SPring-8. Diffraction data of the native crystal were collected twice in the same oscillation range with different oscillation angles and X-ray dose to cover a wide range of diffraction intensities; oscillation angles and exposure times per frame were 0.75° and 45 s for the first set and 3.0° and 30 s for the second set. Diffraction data for SeMet-labeled YfhJ crystal were collected at different wavelengths (edge, 0.9795 Å; remote, 0.9740 Å). All diffraction images were processed using the HKL2000 program suite.21 The MAD method was applied to determine phase angles of SeMet YfhJ crystal using the CNS program suite.22 One Se atom was located in the Bijvoet difference Patterson map. Density modification was applied to the electron density derived from the MAD phases. The initial molecular model was built using the program O,23 and refined using the remote data. The model was further refined using the native data to a 1.75 Å resolution. Except for the first three residues in the histidine tag and a C-terminal residue, all residues were visible in the 2Fo-Fc map. The final R and Rfree values were 19.4 and 23.1%, respectively. Results of data collection and structure refinement are given in Table I. One YfhJ molecule is present in an asymmetric unit. YfhJ has five α-helices [Fig. 1(a)]; three long helices of α1 (residues 8–18), α3 (residues 28–37), and α5 (residues 51–64), and two short helices of α2 (residues 24–26) and α4 (residues 46–48). The program DALI24 was used to find proteins that have similar foldings with the YfhJ fold. The C-terminal DNA-binding domains of Abp1 and Pax6, DNA-binding proteins with helix-turn-helix motifs, were found to be similar to YfhJ [Fig. 1(b)]. When loop regions are excluded, the root mean square deviation of 36 pairs of Cα atoms is 1.6 Å between Abp1 and YfhJ, and that of 30 pairs of Cα atoms is 2.1 Å between Pax6 and YfhJ. Abp1 is a fission yeast centromere protein,25 whereas Pax6 is a human transcription factor critical in the development of the eye, nose, pancreas, and central nerve systems.26 The recognition helix, which binds to the major groove of DNA, corresponds to the α5 helix in YfhJ. It is unlikely, however, that YfhJ binds to DNA because the surface of YfhJ is negatively charged [Fig. 1(c)]. YfhJ is a highly acidic protein; its pI value is 3.7, and Asp and Glu residues occupy 27.3% of 66 residues. This character is conserved in the bacterial homologs of YfhJ. (a) Ribbon diagrams of the YfhJ protein. The histidine tag region is shown in red. (b) Superimposition of Cα traces of YfhJ (residues 1–65, green) on Abp1 (PDB code 1IUF, residues 81–141, red) and Pax6 (PDB code 6PAX, residues 74–128, blue). (c) Surface charge distribution of YfhJ. View direction of (c) is rotated around the vertical axis by 120° relative to (a) so as that α1 is on the front. The histidine tag is excluded in (b) and (c). All figures were prepared the programs MOLSCRIPT34 and GRASP.35 The acidic residues of YfhJ, Asp7, Glu10, Glu13, Asp17, Glu40, Asp41, Asp63, and Glu64, form a negatively charged patch. This patch is on α1, the C-terminal part of α5 and the loop between α3 and α4. It is known that negatively charged residues are localized in such iron-binding proteins as ferritin and frataxin. The iron-storage protein ferritin, composed of 24 subunits, has an inner cavity in which about 4500 iron atoms are stored.27 The negatively-charged patches are located on the inner surface to promote ferrihydrite nucleation.28 Frataxin, on the other hand, is a mitochondrial protein involved in iron metabolism; lack of frataxin causes iron accumulation in mitochondria and leads to a neurodegenerative disease, Friedreich ataxia.29 It has been reported that frataxin forms a regular multimer to sequester iron atoms like ferritin.30 Crystal structures of human frataxin and E. coli CyaY, a frataxin homolog, also show acidic patches on their molecular surfaces.31, 32 Although the YfhJ fold is similar to the fold of neither ferritin nor frataxin, the common surface character among the three proteins may indicate that YfhJ is an iron donor for the FeS assembly in the ISC system. Further study of how YfhJ is involved in the synthesis of FeS clusters is underway. We thank Keiko Miura of JASRI and Taiji Matsu of the RIKEN Harima Institute for their aid with data collection using the synchrotron radiation of BL40B2 and BL44B2, SPring-8, and Kei Wada for his help with the structure determination.

References

29