Portal de Periódicos da CAPES

Novel cation‐π interaction revealed by crystal structure of thermoalkalophilic lipase

2007; Wiley; Volume: 70; Issue: 2 Linguagem: Inglês

10.1002/prot.21799

1097-0134

Hiroyoshi Matsumura, Takahiko Yamamoto, Thean Chor Leow, Tadashi Mori, Abu Bakar Salleh, Mahiran Basri, Tsuyoshi Inoue, Yasushi Kai, Raja Noor Zaliha Raja Abd Rahman,

Enzyme Catalysis and Immobilization

Cation-π interactions are unique binding motifs that frequently occur between electron-rich aromatic ring and organic and inorganic (metallic) cations. This noncovalent interaction can be strong, as has been confirmed by the solid state studies of small molecule crystal structures,1, 2 and the theoretical and experimental analyses in aqueous media.3 A previous protein database search showed that the cation-π interaction can occur for every 77 amino acid residues in proteins, where the positively charged amino acids (e.g., arginine and lysine) and the aromatic amino acids (e.g., tryptophan, tyrosine, phenylananine) are usually involved.4 Cation-π interactions are therefore considered to be an essential force in generating the tertiary and quaternary structures of proteins that are induced by oligomerization and protein folding. These interactions are also important in biological processes such as protein–ligand1, 5, 6 and protein–DNA7-9 complex formations. Many biological studies on cation-π interactions have considered alkali metal cations, particularly Na+ and K+ ions1, 10, 11 because both the Na+ and the K+ are abundant in living systems. These cations are responsible for various noncovalent interactions such as hydrogen bonds, salt bridges, and hydrophobic interactions, importance of which on physiological roles is increasingly recognized recently in molecular biology. The energy of interaction between alkali metals and π systems is fairly strong; for example, the binding energy of K+ and benzene was estimated to be 19 kcal/mol, the value being as large as that of K+ – water is 18 kcal/mol.1 Nevertheless, such metallic cation-π interaction in protein structures have been observed only in few examples, although cation-π interactions between the positively charged and the aromatic amino acid residues are frequently detected.4 The coordination of Na+ with π-systems has been only described in two examples. The coordination of Na+ with tryptophan ring was reported in the crystal structures of hen egg-white lysozyme12 and with thermophilic triosephosphate isomerase mutant.13 Other examples involves the coordination of Cs+ with aromatic residues in crystal structures of rhodanese,14 glutamine synthase,15 and methylamine dehydrogenase.16 However, the biological role of these coordinated metal ions is not clear. An extracellular lipase (T1 lipase) from Geobacillus zalihae strain T1 is a thermoalkalophilic enzyme that was isolated from palm oil mill effluent (POME) in Malaysia. T1 lipase conserves the classical catalytic triad composed of Ser-His-Asp, which is very common in lipase families.17, 18 This enzyme is a secreted protein which can catalyze the hydrolysis of long-chain triglycerides into fatty acids and glycerol at the interface between water and insoluble substrate at high temperature (∼70°C). Since POME contains a high concentration of alkali metals (e.g., more than 50 mM of K+),19 the T1 lipase has the potential to catalyze the hydrolysis under distinctive conditions20, 21 (e.g., in organic solvent, at high temperatures, and at high concentrations of alkali metals). It is thus intriguing to explore the catalytic mechanism of the T1 lipase in detail, which would be helpful in designing a molecular catalyst for any industrial use. Herein, we have crystallized wild-type and mutant F16L T1 lipase in the presence of alkali metal cations (both Na+ and K+), and have determined both crystal structures at 1.5 and 1.8 Å resolution, respectively. The resolved structures revealed that a unique Na+-π interaction with Phe16 was only observed in the wild type T1 lipase. Therefore, in addition to the electrostatic and induction interaction between cation and lone-pair electrons of nitrogen and oxygen, the cation-π interaction is vital for the coordination of metal ions in the T1 lipase. T1 lipase, lipase from Geobacillus zalihae strain T1; POME, Palm Oil Mill Effluent; L1 lipase, lipase from Bacillus stearothermophilus L1. The structural gene of T1 lipase and its mutant F16L was overexpressed in Escherichia coli BL21(DE3)pLysS harboring recombinant plasmid pGEX/T1S.20 The GST fusion lipase was purified using Glutathione Sepharose HP affinity chromatography. The GST fusion lipase was subjected to thrombin cleavage at 16°C for 20 h. The GST tag and thrombin enzyme were further removed by using Glutathione-Sepharose HP, HiTrap Glutathione-Sepharose 4FF, and HiTrap Benzamidine in sequence after Sephadex G-25 gel filtration chromatography, to obtain pure mature T1 lipase. Crystals of wild-type and mutant T1 lipase were grown by the hanging-drop vapor-diffusion method. An initial screening for crystallization conditions was performed using the sparse-matrix sampling protocol,22 using drops containing 2 μL protein solution and 2 μL precipitating solution equilibrated against 500 μL of reservoir solution at 293 K. Plate-like crystals of wild-type and mutant F16L were formed after 1 week under three conditions of Crystal Screen I (Hampton Research) kit, all of which contained PEG as the primary precipitating agent. Optimization experiments led to the following conditions for crystallization: drops containing 3 μL protein at 2 mg mL−1 in 50 mM Tris-HCl pH 8.0 and 3 μL precipitating buffer were equilibrated against 500 μL of precipitating buffer containing 0.5M NaCl, 0.1M KH2PO4, 0.1M NaH2PO4, and 0.1M MES buffer pH 6.6. The crystals typically grew to a maximum size of 0.3 × 0.3 × 0.1 mm3 over 1 week. X-ray diffraction data of wild-type and mutant T1 lipase were collected at SPring-8 BL41XU and in-house X-ray beam, respectively. Crystals of wild-type T1 lipase diffracted up to a 1.5 Å resolution and belonged to the C2 space group with lattice constants of a = 117.73, b = 81.27, c = 99.91 Å, and β = 97.09°. Crystals of mutant F16L lipase diffracted up to a 1.8 Å resolution and belonged to the C2 space group with lattice constants of a = 117.78, b = 81.11, c = 99.48 Å, and β = 96.90°. Diffraction data was indexed, integrated, and reduced with HKL2000.23 Data processing statistics are given in Table I. The crystal structure of wild-type T1 lipase was solved by molecular replacement using atomic coordinates of lipase from Bacillus stearothermophilus P1 (PDBID: 1JI3) as a search model. Mutant F16L lipase was solved by using atomic coordinates of wild-type T1 lipase. Model building and refinement of T1 lipase structure were done using the programs O24 and CNS.25 Water molecules, zinc ions, calcium ions, and chloride ions were placed according to strict density and distance criteria. At the end of the refinement, 30 additional cycles of occupancy refinement with CNS25 were carried out for the atom interacting with the Phe16 ring. The stereochemical quality of the final structure was assessed with the programs PROCHECK26 and WHATCHECK.27 The obtained refinement statistics are shown in Table I. Figures were generated by MOLSCRIPT28 and RASTER3D.29 The atomic coordinates and the structure factor for wild-type and F16L mutant T1 lipase are available from the Protein Data Bank under accession code 2DSN and 2Z5G, respectively. All calculations were performed on Linux-PCs using the TURBOMOLE 5.9 program suite.30 Geometries of the models were taken from the partial structure (57 atoms) of crystal structure of wild-type T1 lipase, and sodium or potassium cation was tentatively located at the center of the (undetermined) electron density. Single point energy calculations were carried out by the DFT method with the standard B3-LYP functional by employing a Gaussian AO basis set of valence triple-ζ quality augmented with polarization functions on all atoms (TZVP).31 The calculated binding energies of Na+ and K+ (in gas phase) were −70.9 and −51.4 kcal/mol, respectively. The results of the structural analysis of wild-type and mutant F16L T1 lipase are shown in Table I. Both wild-type and mutant enzymes have crystallized with the same crystal lattice, molecular packing, and cell parameters. The structures of wild-type and mutant are also very similar (rmsd of 0.22 for 774 Cα atoms). The final models of wild-type and F16L mutant enzymes both include 776 amino acid residues (for two molecules), two Zn2+, two Ca2+, two Cl− ions (for two molecules). Although the wild-type T1 liase contains 1148 water molecules in addition to two alkali metal cations (probably Na+, vide infra), the mutant F16L contains smaller number of watermolecules (686 H2O) and more notably no metal ions. It was unexpected that Cl− ions were observed in the both structures [Fig. 1(B)], which have never been reported in the structures of lipases. The larger electron density on the spots compared with water supports this assignment of Cl− is correct, which was also supported by the relatively long distances between the coordinating side chains and the Cl−: distances between Cl− and NH1 or NH2 of Arg227, N atom (main chain) of Arg214, or NE of Gln216 are 3.1, 3.6, 3.3, and 3.6 Å, respectively. (A) Stereo diagram of the wild-type T1 lipase structure. Sodium, chloride, calcium, and zinc atoms are represented as cyan, yellow, grey, and brown balls, respectively. Side chains of alkali metal cation binding residues (Phe16, Ser133 and His358) are shown as stick models. The anomalous electron density coming from the anomalous signal of the zinc and chloride atoms is in cyan (4.0 σ). (B) Close-up view of Cl− binding site. The Cl− and Cl− binding residues are shown as stick models with labels. The water molecule is shown as a red sphere labeled “Wat”. The σA weighted Fo − Fc electron density map (4 σ) was calculated after the omission of the relevant moiety from the model at a resolution of 1.5 Å. The asymmetric unit of the crystal contains two copies of the T1 lipase. The overall structure is globular in shape, with a central β-sheet consisting of seven strands surrounded by 13 α-helices and 10 310-helices and loops, which results in an overall topology of a typical α/β hydrolase canonical fold [Fig. 1(A)]. Lipases are generally known to adopt a closed or open conformation. The structure of current T1 lipase showed a closed conformation and the active site was buried inside the molecule. The localization and coordination spheres of Zn2+ and Ca2+ were very similar to those observed in the crystal structure of lipase from Bacillus stearothermophilus L1 (L1 lipase).32 During crystallographic refinement of wild-type T1 lipase, the metal-free model phased (2Fobs − Fcalc) map indicated an existence of additional atom near the aromatic ring of Phe16 [Fig. 2(A)]. The electron density looks spherical, and an aromatic ring of Phe16 faces toward the peak of the electron density. The 1.5 Å resolution map revealed that a specific atom tightly interacts with the aromatic π-system of Phe16. Since the crystallization buffer contains a high concentration of alkali metal cations (600 mM of Na+ and 100 mM of K+, respectively) but with no divalent cations, this atom is likely either alkali metal cation or water molecules. The distance between the peak and the hydrogen bonded oxygen (2.67 Å) is very similar to that between alkali metal cations (Na+ and K+) and the oxygen atom in the Cambridge Structural Database (CSD)33 of small molecule crystal structures. A similar example have been described in the reviewed paper by De Wall and coworkers.34 In the article, they have proposed that the water molecule in the structure of tryptophanase (PDB ID code 1AX4) is probably an alkali metal cation by the same reasons, although a water-π interaction has been found in a 2.1 Å crystal structure of tryptophanase (PDB ID code 1AX4). For the same reasons, the binding species are probably alkali metals, not water molecules. Close-up views of electron density map of wild-type T1-lipase (A) and mutant enzyme F16L (B). The alkali metal cation and the side chains are shown in stick models with labels. The σA weighted 2Fo − Fc electron density map (1 σ) was calculated after omission of the relevant moiety from the model at a resolution of 1.5 Å. (C) The alkali metal cation binding site. The cation and cation binding residues are shown in stick models with labels. The water molecule is shown as red a sphere labeled “Wat”. The oxyanion hole is shown as a transparent sphere. Hydrogen bonds and cation-π interactions are represented by blue and yellow dashed lines, respectively. Next, we estimated the atomic occupancies and temperature factor by the following protocol at the final stage of the refinement. In the protocol, a water molecule and alkali metal cations (Na+, K+, and Rb+) were put at the peak of the electron density, respectively. Then, the temperature factors of the atoms were set at 19.7 Å (the average value of the temperature factors of six aromatic ring carbons), before occupancy and temperature factor refinement by CNS.25 The calculation shows a K+, Rb+, and water molecule bind with low occupancies (estimated occupancies of 0.35, 0.14, and 0.55, respectively), while Na+ has reasonably high occupancy of 0.80. This calculation shows that the binding species are most likely to be Na+, although we can not completely rule out the possibility of coexistence of other species with low occupancies. To further investigate the coordination structures of Na+ and K+ in our T1 lipase, we performed the quantum chemical calculations for the substructure of the binding pocket in the presence and absence of metal ions. The binding energies were compared at the B3-LYP/TZVP level of theory. The calculations show that the stabilization energy with Na+ is much larger (ΔE < 20 kcal/mol) than that with K+, which also supports the assumption that the metals involved in the T1 lipase are Na+ ions. To the best of our knowledge, this is the first direct observation of the Na+ ion coordinated with the π-system of phenylalanine ring in protein structures, while the Na+ ion coordinated with the π-system of tryptophan ring has been observed in HEW lysozyme12 and thermophilic triosephosphate isomerase mutant.13 To further confirm the positive contribution of the side chain of Phe16 for the present cation-π interaction, we have also solved the crystal structure of mutant F16L T1 lipase at 1.8 Å, which were prepared and crystallized similarly to the wild-type enzyme (see Materials and Methods). Although the resolution was slightly lower (1.8 A), we found almost zero electron density in mutant F16L enzyme at the position corresponding to the metal ions in the wild-type enzyme [Fig. 2(B)]. Since the site is located inside the molecule, a possible artifact by the crystal appreciably packing can be excluded. Threrefore, this observation confirmed that the side chain of Phe16 contribute to the binding of the alkali metal ion in the protein structure. Table II collects the several coordination geometries of the alkali metal cation with π-system. The average coordination distance between the metal and ring carbons was found to be 3.34 Å, which is much shorter than those observed in the crystal structures of HEW lysozyme12 and thermophilic triosephosphate isomerase mutant.13 The shorter distance is in good agreement with the experimental values of small molecule crystal structures in CSD (Fig. 3) and the theoretical values reported previously.12 It is also reported that the distances are greatly influenced by the coordination of additional water molecules.35 (A) The coordination geometry of the alkali metal cations. (B) The distances between alkali metal cation and ligands. The cation-bound partial structure are retrieved from the CSD using the program ConQuest.37 The partial structure, the alkali metal cations, the number of reports, the minimum of distance, the maximum of distance, and the mean distance are shown, respectively. The cation also interacts with the side chains of Ser113 and His358 [Fig. 2(C)], both of which participate in the classical catalytic triad. Significantly, the phi and psi values of Ser113 correspond in the distorted region of the Ramachandran plot, which seems to be essential for exhibiting enzymatic activity.18 The coordination of the metal might stabilize the distorted conformation of the catalytically essential Ser113. Additionally, since the cation is positioned next to so-called “oxyanion hole” in the active site, the cation might directly contribute the catalytic activity directly. As mentioned above, due to the fact that the POME contains a relatively high concentration of alkali metals;19 we speculate that the secreted T1 lipase may utilize alkali metals to exhibit its catalytic activity. According to the Lipase Engineering Database at http://www.led.uni-stuttgart.de,36 Phe16 is highly conserved in Bacillus, Geobacillus, and Staphylococcus, but is not completely conserved in the Burkholderia superfamily (abH15.1), suggesting that the cation-π interaction observed in T1 lipase is unique to specific organisms in the Burkholderia superfamily. The analysis of enzymatic activity in the presence of alkali metal cations is highly desired in the future study in this area. We thank Drs. K. Takano and S. Kanaya for their support in crystallization and helpful discussion. We are grateful to Dr. K. Hasegawa and Dr. H. Sakai at BL38B1 (SPring-8) for their kind help in the data collection. The synchrotron radiation experiments were performed at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI). Accession Numbers: Atomic coordinates have been deposited in Protein Data Bank, www.rcsb.org (PDB ID code 2DSN for wild-type and 2Z5G for F16L mutant enzyme).