The Crystal Structure of Human Chloride Intracellular Channel Protein 2: A Disulfide Bond with Functional Implications
Wei Mi,Yu-He Liang,Lanfen Li,Xiao-Dong Su
DOI: https://doi.org/10.1002/prot.21922
2008-01-01
Proteins Structure Function and Bioinformatics
Abstract:Human chloride intracellular channel proteins (CLICs) are intracellular ion channels based on sequence homology. Human CLICs include six members (CLIC1-6), and all of them contain a conserved core about 240 amino acids. The most distinct feature of CLICs is that they can exist in two states: the soluble cytosolic state and the integral membrane ion channel state. Purified water-soluble CLICs adopt the glutathione S-transferase (GST) fold and can insert into a lipid membrane to form chloride channel.1, 2 Among the human CLICs, CLIC1 and CLIC4 have been well studied with the crystal structures solved.1, 3, 4 The CLIC1 and CLIC4 have important functions in regulating cell cycle, apoptosis, and so forth, and have become potential drug targets for cancer therapy.5-8 Human CLIC2 has a molecular weight of 28.4 kDa and a calculated isoelectric point of 5.44. The CLIC2 gene locates in the telomeric region of Xq28 and is composed of six coding exons and five introns. Since this region of the X chromosome is closely associated with many hereditary diseases, CLIC2 has thus been proposed as a candidate gene for some genetic disorders linked to Xq28.9 Northern blot has demonstrated that the gene of CLIC2 could be expressed in lung, spleen, heart, and skeletal muscle.10 Compared with the CLIC1 or CLIC4, little functional knowledge is known about CLIC2 so far. The only documented function of CLIC2 is to regulate the cardiac calcium ion channel, ryanodine receptor 2 (RyR2).11 In this work, we present the crystal structure of the soluble form of human CLIC2 and tried to correlate some structural features to its potential function. The gene coding for human CLIC2 was amplified by the polymerase chain reaction (PCR) from a human liver cDNA library. The PCR product was cloned into the pET28a (Novagen) expression vector with an N-terminal His6 tag. The resulting plasmid was transformed into E. coli strain BL21 (DE3) and was cultured at 37°C in 1 L Luria-Bertan (LB) medium containing 50 mg/L kanamycin. Protein expression was induced at 30°C for 6 h. Cells were harvested and resuspended with 20 ml buffer A (20 mM Tris-HCl, pH 7.5, and 500 mM NaCl) and lysed by sonication. The cell lysate was centrifuged at 16,000g and the supernatant was applied to a Ni2+ chelating column (HiTrap™, GE, USA) where CLIC2 protein was eluted with imidazole gradient. 5,5′–Dithio–bis(2-nitrobenzoic acid) (DTNB) was added to the fraction containing CLIC2 protein with a final concentration of 1 mM, and gel-filtration chromatography (Hiload Superdex 75, column volume 120 mL) was chosen for the further purification in buffer C (20 mM Tris-HCl, pH 7.5, 200 mM NaCl). Purified protein was concentrated to 30 mg/mL and crystallization experiments were performed by the hanging-drop vapor-diffusion method at 289K. One microliter of protein was mixed with an equal volume reservoir solution (35% PEG400, 0.1M Bis-Tris pH 7.0, 0.1M CH3COONH4) and equilibrated with 500 μL reservoir solution. The CLIC2 crystals were flash-cooled to 100K in a stream of cold nitrogen gas and X-ray data were collected on a MAR165 CCD at the beamline 3W1A, Beijing Synchrotron Radiation Facilities (BSRF), China. The diffraction data were processed to 2.0 Å with the program HKL2000.12 The structure of CLIC2 was determined by the molecular replacement method, using the program MolRep in the CCP4 package.13 The CLIC4 structure (PDB ID: 2AHE) was used as the searching model. The program ARP/wARP was used for further refinement and automatic model tracing.14 The structure refinement was performed with the program Refmac5 combined with manually rebuilding of the model by the program COOT.15, 16 Stereochemistry qualities of the model were checked by PROCHECK.17 The data collection and structure refinement statistics were listed in Table I. Structure factors and the coordinates have been deposited in the PDB (PDB ID: 2PER). The final model of CLIC2 has been refined to 2.0Å resolution (Table 1). The CLIC2 structure comprises the GST fold, with two domains: an N-terminal glutaredoxin-like domain and an all α-helical C-terminal domain (Fig. 1). The overall structure of CLIC2 is similar to that of CLIC1 and CLIC4 [Fig. 1(C)]. The root mean square deviation (RMSD) among Cα atoms from CLIC2 to CLIC4 is 0.69 Å, and from CLIC2 to CLIC1 is 1.7 Å. The electron density of residues 55–69 could not be observed at the contour level of 1.0 sigma due to the flexibility of this region. Since this region was named helix α2 in the models of CLIC1 and CLIC4, in order to make the secondary structure assignment consistent with previous CLIC structure descriptions, we skipped the helix α2 when naming the secondary structure of CLIC2. We named the loop joining the N- and C-terminal domains between helices α3 and α4A the "joint loop." A long loop between α5 and α6 in C-terminal domain protrudes out of the surface, and was named the "foot loop" region [Fig. 1(A,B)]. Structure of human CLIC2. (A) Structure based sequence alignment of human CLIC1-6 and DHAR2, DHAR3 from A. thaliana. Identical and conserved residues are colored in red and yellow, respectively. CXXC motif is colored in cyan and marked by upwards triangles. The additional N-terminal sequences of CLIC5 and CLIC6 were truncated. (B) Overall structure of CLIC2. N-terminal domain is colored in red; C-terminal domain in green, the "joint loop" in yellow, the "foot loop" in blue, disulfide bond between Cys30 and Cys33 is showed as stick (yellow). (C) Structure superimposition of CLIC1 (yellow), CLIC4 (red) structures on CLIC2 (green). Disulfide bond in CLIC2 is showed in stick (yellow). Although the N-terminal domain of CLIC2 shows no obvious sequence homology with a glutaredoxin, their overall structures are quite similar [Fig. 2(A)]. They both consist of a four-β-strand core and two helices on one side. The RMSD of Cα atoms over the 56 residues between the N-terminal domain of CLIC2 and a pig glutaredoxin (PDB code 1KTE) is about 1.6 Å, and the active-site regions containing strand β1 and helix α1 in both structures superimpose well and the RMSD value of Cα atoms in these regions is only 0.6 Å. Disulfide bond between Cys30 and Cys33 and the electrostatic potential surface of CLIC2. (A) Structure comparison between the N-terminal domain of CLIC2 (green) and pig glutaredoxin (magenta), poxviral glutaredoxin (cyan). The disulfide bond of CLIC2 is showed in stick (yellow). (B) 2Fo-Fc electron density map and stick diagram of the disulfide bond in human CLIC2. (C) Structure comparison of disulfide bonds between CLIC2 (green) with pig glutaredoxin (magenta), poxviral glutaredoxin (cyan). (D) The electrostatic potential surface of CLIC2 and foot loop insertion into the cavity of the neighbor molecule. The most notable novel feature of CLIC2 structure is that a disulfide bond was well-resolved between Cys30 and Cys33 on the N-terminal of helix α1 [Fig. 2(A,B)]. The sequence of Cys-Pro-[Phe/Ser]-[Cys/Ser] is conserved in all human CLICs [Fig. 1(A)]. However, the CXXC signature motif for redox active disulfide bond is only conserved in CLIC2 and CLIC3, the other CLICs only have CXXS motif. The main-chain and the Sγ atoms of CXXC are superimposed well with oxidized CXXC structure of the pig and the poxviral glutaredoxin,18, 19 while the side chains of Pro31 and Phe32 show some deviation [Fig. 2(C)]. The disulfide bridge is located at the N-terminal of helix α1 with Cys30 exposed to the cavity between the N- and C-terminal domains and accessible to other molecules, whereas Cys33 is located one turn after Cys30 and buried inside the N-terminal domain. The partial positive charge at the N-terminal end of helix can increase the nucleophilicity of the thiol group on Cys30, which is a conserved feature for all glutaredoxin family proteins.19 The major difference between the CLIC2 and a glutaredoxin is that, in the pig glutaredoxin for example, the residue next to the CXXC motif is an Arg, whereas it is Gln34 in CLIC2. The Arg in the pig glutaredoxin was proposed to help the active cysteine to form thiolate anion and is hence important for thiol transferase activity,19-21 whereas the Gln34 in CLIC2 is a neutral amino acid, which might increase the pKa value of the hydrosulfide. As a result, the nucleophilicity of Cys30 might be decreased in CLIC2 compared with a typical glutaredoxin. The acidic "foot loop" connecting the helices α5 and α6 is a unique feature of the human CLICs family [Fig. 1(A)]. This loop is flexible and easily affected by crystal packing. The electrostatic potential surface of the CLIC2 structure shows that the loop is negatively charged because of six acidic residues in this region, and the cavity between the N- and C-terminal domains is positively charged [Fig. 2(C)]. In our structure, the negatively charged "foot loop" is observed to insert into the positively charged cavity of another symmetry-related molecule due to crystal packing [Fig. 2(C)]. The Asp161 forms a salt bridge with Lys189 from a neighbor molecule and is important for this protein–protein interaction. The insertion peptide of the loop was stabilized by extensive hydrogen bond network formed between the peptide and the cavity residues near the CXXC motif. These features indicate that the specific protein–protein interaction events could happen by utilizing the cavity and the "foot loop." The structure of CLIC2 has demonstrated the existence of intramolecular disulfide bond between Cys30 and Cys33 which may give some hints for the function of CLIC2 in the cytoplasmic soluble state. It is believed that the N-terminal domain of a GST fold is evolved from the thioredoxin/glutaredoxin family and the appearance of CLICs during evolution was an earlier event than the GSTs.22 Glutaredoxins are well known as general reductants to reduce protein disulfide bonds by the two cysteines in CXXC motif, and the CXXC motif may have thiol transferase activity which needs only the first cysteine.20 Thus, it is reasonable to propose that CLIC2 can reduce disulfide bonds in its specific target proteins. Considering that CLIC2 has a C-terminal helical domain and a "foot loop" that may recruit and interact with other proteins, we propose that the positively charged cavity of CLIC2 may react on some specific targets containing negatively charged residues, in contrast with glutaredoxins as general hydrogen donors. In the CLIC2 structure, the "foot loop" inserts into the cavity near the CXXC motif of a neighbor molecule. This feature shows that the CXXC motif is accessible and the cavity is big enough to accommodate a peptide segment of the target proteins for specific interactions. The "joint loop" between N- and C-terminal domains is a proline-rich region. In the reduced CLIC1 structure, the plasticity of the "joint loop" allows two domains to have relative movements.3 All the structural features described above could provide evidence that the CLICs can interact with target proteins and regulate their activity. Interestingly, recent identified four glutathione-dependent dehydroascorbate reductases (DHAR1-4) in Arabidopsis thaliana have shown about 30% sequence identity with CLIC2 [Fig. 1(A)]. They have been shown both thiol transferase activity23 and ion channel functions.24 DHARs have active site Cys-Pro-Phe-[Ser/Cys/Gly], which is very similar with human CLIC proteins. These facts indicated that CLIC2, or some other human CLICs, might also possess thiol transferase activity, and play important roles in redox-regulated biological events. The authors thank Dr. Rong Gao for critical and helpful reading of the manuscript, and the beamline staff (Drs. Yu-Hui Dong and Peng Liu) at the Beijing Synchrotron Radiation Facilities for help with data collection. X.-D. Su is a recipient of a National Science Fund for Distinguished Young Scholars.