Proteins Structure Function and Bioinformatics · 2007 · 32 citations · 16 references
Crystal StructureChemokine BiologyLymphocyte DevelopmentImmune RegulationImmunologyMolecular BiologyImmunologic MechanismSdf-1α GeneInflammationSdf-1α MoleculeSignaling PathwayCell RegulationCell DevelopmentCell SignalingMolecular SignalingImmune SurveillanceCell EngineeringCell BiologyCytokineHuman CellDevelopmental BiologyImmune Cell DevelopmentStem Cell ResearchMedicineCxcr4 Gene
Chemokines are a superfamily of cytokine molecules that mediate several cellular functions.1, 2 The chemokines superfamily has been classified into three species, CC, CXC, and CX3C, based on the presence of an additional one or three amino acids between the first two cysteine residues.3 Stromal cell-derived factor-1α (SDF-1α) is an unusual member of the CXC family that was first isolated from a mouse bone marrow stromal cell line.4, 5 SDF-1α has been implicated in a variety of physical functions, including hematopoiesis, cardiogenesis, B cell lymphopoiesis, and migration and cellular colonization of primordial germ cells in mammals.6-8 SDF-1α interacts specifically with the physiological receptor CXCR4 for multicellular functions. A knockout experiment of either the sdf-1α or cxcr4 gene supported the claim that the SDF-1α/CXCR4 complex has an important role in embryonic development and proliferation.5, 9 Mice lacking the sdf-1α gene died in the perinatal period due to a severe cardiac ventricular system septum defect.5 Although various functions of SDF-1α have been reported, biochemical assays of its activity have been limited because the protein is not over-produced in a recombinant form. To overcome this obstacle, we expressed human SDF-1α in E. coli and obtained the refolded protein. The three-dimensional crystal structure of refolded SDF-1α displays some structural differences from previously reported structures. Also, the chemotaxis assay shows that refolded SDF-1α binds to receptors in the CCRF-CEM cell line, indicating that the refolded protein is an active form. The sdf-1α gene was amplified from normal human fibroblast cDNA by polymerase chain reaction (PCR). The sdf-1α gene was inserted into pET28a containing a (His)6-tag and a TEV protease site. This construction was transformed into E. coli strain BL21(DE3) for overexpression. The mutant K1A was constructed by site-directed mutagenesis using PCR. The transformed SDF-1α E. coli cell was cultured in Luria Bertani (LB) broth at 37°C until the culture reached an absorbance of 0.6 at 600 nm. Expression of SDF-1α was induced by addition of isopropyl-thio-β-D-galactopyranoside (IPTG) to a final concentration of 0.4 mM. After 4 h, the cultured cells were harvested by centrifuging. The pellet was resuspended in a PBS buffer (10 mM sodium phosphate, 150 mM NaCl, pH 7.4) and was disrupted by sonication. Inclusion bodies were solubilized in a denaturation buffer (8M urea, 20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 0.5 mM β-mercaptoethanol, 3% glycerol) and allowed to stand for 2 h at room temperature. The denaturatant was centrifuged, and the supernatant was applied to a Ni-NTA column, which was preequilibrated with the denaturation buffer. On-column refolding10 of the protein was induced by decreasing the urea concentration in a linear gradient from 8.0 to 0.0M over 6 h. After refolding, the protein was eluted by increasing the concentration of imidazole from 0 to 0.3M with elution buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 0.5 mM β-mercaptoethanol, 3% glycerol, 300 mM imidazole). The refolded protein was pooled and concentrated. To remove the (His)6-tag, the TEV protease was treated in pooled protein for 6 h at room temperature. After digestion, the protein was applied to a desalting column, and a second Ni-NTA column was used to separate the target protein from the (His)6-tag and undigested target protein. The crystallization and chemotaxis assays with refolded SDF-1α were performed after further purification through a Superdex peptide gel filtration column (Amersham Biosciences, GE Healthcare). The purified SDF-1α was concentrated to 5 mg/mL. Crystals of refolded SDF-1α were grown using the hanging-drop vapor diffusion method at 20°C. The protein solution was mixed with an equal volume of reservoir solution (0.1M Hepes, pH 7.5, 1.4M Sodium citrate). The crystals were then transferred to the mother liquor, containing 15% ethylene glycol as a cryoprotectant of the fresh cooled system (100 K), at beamline 4A in the Pohang Acceleration Laboratory (Pohang, Korea). The diffraction data were processed, scaled, and reduced using the HKL2000 package.11 The crystals belong to the orthorhombic space group P212121, with cell dimensions a = 35.22 Å, b = 56.68 Å, and c = 71.63 Å and contained two molecules in the asymmetric unit. The crystal structure of the refolded SDF-1α was solved by the molecular replacement, using the Molrep program.12 The previously reported crystal structure of SDF-1α (PDB code: 1QG7) was used as the search model.13 The complete model was built manually using the Coot program.14 Crystallographic refinement was carried out using CNS,15 including simulated annealing, minimizing, and B-factor refinement. The geometries of the final models were checked with the program PROCHEK,16 no residues were found in disallowed regions. Table I sets out data collection and refinement statistics. The CCRF-CEM cells, which naturally express CXCR4, were obtained from the KCLB (Seoul, Korea). The cells were grown and maintained in an RPMI 1640 medium with 25 mM Hepes and 10% FBS.17 On the day of the experiment, the cells were harvested and washed once with migration assay buffer (RPMI 1640 with 25 mM Hepes, pH 7.5, 0.1% BSA). Also, both SDF-1α and its mutants were diluted into the same buffer. The migration assay was performed in a 48-well ChemoTx chamber (Neuroprobe, Gaithersburg, MD). A total of 27 μL of SDF-1α was added to the bottom well of the chamber. The wells were covered with a polycarbonate nucleopore filter (5-μm pore size), and 50 μL of a suspension of CCRF-CEM cells (4 × 106 cells/mL) was then added to the upper wells. After assembly, the chamber was incubated for 3 h at 37°C with 5% CO2. The cells that had migrated into the bottom wells were then counted. Measurement of the activity of mutant K1A was also performed, as above. This experiment was performed in triplicate. The crystal structure of refolded SDF-1α has a chemokine topology (β-β-β-α) very similar to the previously reported solution18 and crystal structures13, 19 [Fig. 1(a)], consisting of a three-stranded antiparallel β-sheet followed by an α-helix. The structure of the refolded SDF-1α monomer looks like a forefinger protruding from a closed fist. Refolded SDF-1α exists in three antiparallel β-strands (β1: residues 24–28, β2: residues 38–42, and β3: residues 46–50) in both the A and B chains. Also, the α-helix (residues 56–66) in the B chain is one residue longer (Lys56) than the α-helix (residues 57–66) in the A chain. A 310 helix defined in previous crystal structures is not found. The dimer of the refolded SDF-1α structure is stabilized through interaction between the antiparallel β1 strands of the A and B chains. (a) Refolded SDF-1α overall structure is a β-β-β-α topology with a three stranded antiparallel β-sheet, upon which lies a C-terminal α-helix. The A chain is shown in yellow and the B chain in blue. A 310 helix region, found in other crystal structure, is shown to the long loop region (residues 20–22) in red. The disulfide bonds of two chains are shown in orange. (b) Disulfide bridges of refolded SDF-1α shown in purple. Disulfide bonds (orange) have 2.04 Å (Cys9–Cys34) and 2.05 Å (Cys11–Cys50) of length, respectively. It is well known that the extended N-terminal region plays an important role in interactions with receptor CXCR4.18, 20 The N-terminal region of refolded SDF-1α adopted an extended conformation. In the monomeric structure based on NMR spectroscopy,18 SDF-1α was observed with a disordered N-terminal region (residues 1–8). Also, in the dimeric structure based on X-ray crystallography,13, 19 the side-chains of each residue in the N-terminal region of SDF-1α were not exactly assigned. However, in the present dimeric refolded SDF-1α structure, the side-chains of the entire residues 1–68 are clearly observed with corrected mapping through continuous electron density. In comparison between the present structure and previously reported structures, the N-terminal regions of chain A and B have different conformations. In case of the previous structure, the N-terminal regions of chain A and B are commonly interacting with neighboring SDF-1α molecule. In refolding SDF-1α, the N-terminal residues of chain A have also interaction with neighboring molecule. Especially, the Lys1 of chain A interacts with the neighboring His17 and Phe13 residues in crystal packing. However, the N-terminal region of chain B is snugly laid in the solvent layer in the crystal lattice without any interactions with neighboring molecules. The superposition between the A and B chains shows that the N-terminal region of the A chain has a perpendicularly bent loop at the Cys9 residue and rest regions are very conserved. The Cys residue is involved in the formation of disulfide bridges [Fig. 1(b)]. Two disulfide bridges (Cys9–Cys34 and Cys11–Cys50) is a common structural feature of CXC chemokines.1 These disulfide bridges provide strong geometric constraints on the surrounding residues.13 The disulfide bridges provide good evidence that SDF-1α has refolded correctly. The structure of the refolded SDF-1α refined in this work has a similar morphology to the mammalian SDF-1α (native SDF-1α expressed in mammalian cell), as described above. However, comparison of the refolded SDF-1α structure (A and B chains) with the mammalian SDF-1α structure (A′ and B′ chains) revealed several differences in local conformation. The differences in the lengths of the strand and the helix between the two structures are specified in Figure 2(a–d). β1 and β2 in the A chain are shorter by four residues and three residues, respectively, than β1′ (residues 23–28) and β2′ (residues 35–42) in the A′ chain. But β3 in the A chain is two residues longer than β3′ (residues 48–50) in the A′ chain [Fig. 2(a)]. The β1 in the A chain is shorter by one residue than the β1′ strand (residues 23–28) in the B′ chain, but β2 is the same length in the two structures, and β3 in the A chain is two residues longer than β3′ (residues 48–50) in the B′ chain [Fig. 2(b)]. The lengths of β1 and β2 in the B chain are shorter by four and three residues, respectively, than β1′ (residues 23–31) and β2′ (residues 35–42) in the A′ chain, but β3 in the B chain is two residues longer than β3′ (residues 48–50) in the A′ chain [Fig. 2(c)]. The length of β1 in the B chain is one residue shorter than β1′ (residues 23–28) in the B′ chain, but β2 is the same length in the two structures, and β3 in the B chain is longer by two residues than β3′ (residues 48–50) in the B′ chain [Fig. 2(d)]. The β-strand differences are affected by the length and direction of the non-overlapped region between the β-strands. In addition to differences in the β-strands, the α-helix of the refolded structure also shows some differences compared with the mammalian structure. Comparison of the α-helix between the B chain and the A′ chain shows the same length (residues 56–65), but the α-helix (residues 57–65) of the A chain is one residue longer than the α′-helix of the A′ chain [Fig. 2(a,c)]. There are two α′-helices (residues 56–61 and 63–65) in the B′ chain, but the α-helix of the refolded SDF-1α has a single α-helix [Fig. 2(b,d)]. As a result of the conformational differences of the secondary structures, the C-terminal ends of the A and B chains in refolded SDF-1α run in the opposite direction. Other structural differences are found on in residue 20–22 region, which forms a single 310 helix. This region of the refolded SDF-1α is composed of a very long loop formation. The differences in the lengths of the strands and the helices among mammalian,13 NMR,18 synthetic,19 and refolded SDF-1α structures are summarized in Table II. The each two chains of refolded SDF-1α and mammalian SDF-1α are called A, B, A′, and B′ chain, respectively. Superposition of the two molecules was effected by optimizing the overlap of the Cα atom. (a) Superposition between A and A′, (b) superposition between A and B′, (c) superposition between B and A′, and (d) superposition between B and B′. Color code is: yellow (A), blue (B), red (A′), and green (B′). (e) According to the Cα atom of AA′, AB, BA′, and BB′, the RMS deviations calculated in (a) to (d) are 1.32, 4.45, 1.73, and 3.31, respectively. The B-factor profiles of β-strands are quite low, whereas the helix region appears higher than the strands; this indicates that the helix is more flexible than the strands. The loop regions of refolded SDF-1α and mammalian structure are more flexible than the secondary structures. Both the N- and C-terminal regions have an extremely high B factor value. (f) Chemotactic activity of refolded SDF-1α and its mutant K1A. Concentration-dependent chemotaxis of CCRF-CEM cells expressing CXCR4 is shown in response to refolded SDF-1α. The right side column of the graph shows that 250 nM of K1A mutant has no activity for chemotaxis to receptor CXCR4. As shown in Figure 2(e), the B-factor profile of the refolded SDF-1α has uniformly higher values than those of the mammalian SDF-1α.13 The N- and C-terminal regions of the refolded SDF-1α are very flexible, like mammalian SDF-1α.13 In the B-factor profile for the refolded SDF-1α, the B-factor of the loop region between β1 and β2 is relatively higher than that of the other secondary structures. Because the β1′ of the A′ chain is five residues longer than that of the others, the length of the loop between β1′ and β2′ in the A′ chain is short and stable. The loop region (residues 20–22), composed of the 310 helix of the mammalian SDF-1α, is also highly flexible. Based on several differences in the lengths of the secondary structures, it can be suggested that subtle differences have significant effects on the conformation and flexibility of the overall structure of SDF-1α. Previous NMR studies18 have demonstrated that the N-terminal residue of SDF-1α participates in the binding and activation of the target receptor. In particular, Lys1 and Pro2 are directly involved in interaction with CXCR4 receptor.20 Also, the RFFESH motif (residues 12–17) is considered another region for CXCR4 receptor binding and activation.18 To validate the activity of the refolded SDF-1α, we investigated the interactions of SDF-1α with the receptor using a CCRF-CEM cell line, which naturally expresses the CXCR4 [Fig. 2(f)]. The activity of the refolded SDF-1α (wild type) occurred with the same concentration-dependent pattern. A K1A mutant was used as a negative control. The chemotactic activity of the K1A mutant is similar to the basal level [Fig. 2(f)]. Thus, the structural differences between the refolded SDF-1α and mammalian SDF-1α are irrelevant to the chemotactic activity taking place through interaction between the ligand SDF-1α and the receptor CXCR4. For the first time, a recombinant SDF-1α in E. coli has been obtained as a refolded form using on-column refolding. Structural differences between refolded SDF-1α and previously reported SDF-1α are found in various positions. The chemotactic properties of refolded SDF-1α interacting with CXCR4 receptor are similar to previous results. On the basis of the structural and functional analysis, the refolded SDF-1α proteins should provide a cornerstone for studying the structural and functional features of SDF-1α and its interaction with CXCR4. The coordinates have been deposited in the PDB (ID: 2J7Z).
16