Portal de Periódicos da CAPES

Structural insights into Staphylococcus aureus enoyl‐ACP reductase (FabI), in complex with NADP and triclosan

2009; Wiley; Volume: 78; Issue: 2 Linguagem: Inglês

10.1002/prot.22581

1097-0134

Amit Priyadarshi, Eunice EunKyeong Kim, Kwang Yeon Hwang,

Computational Drug Discovery Methods

Staphylococcus aureus, a gram-positive bacterium, is responsible for the wound infections and staphylococcal scalded skin syndrome, a cutaneous reaction to a staphylococcal exotoxin that is absorbed into the bloodstream. 1 The emergence of antibiotic-resistant pathogens is a serious health problem worldwide, and S. aureus has become resistant to many commonly used antibiotics such as penicillins. Because of the existence of methicillin-resistant S. aureus (MRSA), the bacterium is a popular source of target proteins for use in drug design.2 Most bacteria and plants synthesize fatty acids using a discrete and highly conserved group of enzymes called the type II fatty-acid synthase (FAS II) system, whereas yeasts and animals utilize type I fatty-acid synthase (FAS I). The differences in prokaryotic and eukaryotic fatty acid biosynthesis offer an attractive opportunity for selective FAS II inhibition. 3-5 Enoyl-[acyl-carrier-protein] reductase (ENR), also known as FabI (EC 1.3.1.9), is one of the key components of the FAS II system. ENR completes the fatty acid chain elongation cycle by catalyzing the stereospecific reduction of the double bond between positions C2 and C3 of a growing fatty-acid chain.3, 4 Two clinically significant classes of ENR inhibitors, Isoniazid (INH) and Triclosan (TCL), have been used to form structural complexes with ENR proteins from various sources. TCL (2,4,4′-trichlolo-2′-hydroxy diphenyl ether) is a bactericidal agent and is effective against a variety of microorganisms.5 TCL is reported to be a noncompetitive inhibitor of purified SaFabI (FabI from S. aureus), demonstrates potency against S. aureus cultured in vitro as well as in vivo, and is safe to administer to humans.6 Kinetic assay of SaFabI mutant, R40Q and K41N, exhibit an at least 50-fold decrease in kcat/Km for NADPH.6 Herein, we describe biochemical affinity assay using surface plasmon resonance (SPR) and report the structures of the apo-form of SaFabI, solved to 2.70 Å and the ternary complex of SaFabI + NADP + TCL, solved to 2.3 Å. This is the first reported structure of FabI in complex with NADP. The residues Ala95 to Leu116 and Ser197 to Phe204 of the loop region are in an open conformation in the apo-SaFabI structure, but adopt a closed conformation in the ternary structure. This information will be of use in the identification of novel features of the enzyme that could be utilized in a program of rational structure-based drug design. The FabI gene (GenBank Accession Number BX571856, 771 bp) was amplified by PCR from S. aureus strain MRSA252, digested with NotI and XhoI, ligated into the pET28a vector and transformed into the E. coli BL21 (DE3) strain. Cells were grown at 291 K for 18 h after induction, harvested, and disrupted in 50 mM of ice-cold lysis buffer (Tris-HCl pH 8.0, 200 mM NaCl, 2 mM 2-mercaptoethanol, 5% glycerol). The supernatant was loaded onto a nickel-chelated His-Trap column (GE healthcare) and eluted with a linear gradient of 40 to 250 mM imidazole in Tris-HCl pH 8.0, 200 mM NaCl, 2 mM 2-mercaptoethanol, and 5% glycerol. The final purification step involved gel filtration on a prep-grade Superdex-200 26/60 column (GE healthcare), which was previously equilibrated with buffer A containing 25 mM (Tris-HCl pH 8.0), 200 mM NaCl, and 1 mM DTT. SaFabI was soluble when overexpressed and purified to over 95% purity. A yield of ∼60 mg of homogeneous protein from 1.0 L of culture concentrated to 15 mg/mL using an Amicon concentrator. After reiterated optimization, diffraction-quality crystals of SaFabI, the rod-shaped (Form I) and hexagonal shaped (Form II) crystals were obtained with a reservoir solution consisting of 10–15% PEG 400, 0.1M Tris pH 8.3–8.5 and 4% glycerol at 295 K. The ternary SaFabI complex (Form III) with NADP + TCL was obtained after co-crystallization with 5 μM NADP and 10 μM TCL using 10% PEG 400, 0.1M Tris, pH 8.3 and 4% glycerol. X-ray diffraction data of SaFabI cryoprotected crystals with 20% (w/v) glycerol were collected at 100 K using an ADSC Quantum CCD 210 detector (Madison, WI) at beamline 6C and 4A of Pohang Light Source (PLS), South Korea. The collected data were processed and scaled with the programs DENZO and SCALEPACK from the HKL 2000 suite. 7 The structure of the ternary complex (form III) holoenzyme, SaFabI-NADP- TCL was determined by molecular replacement using the program MOLREP with CCP4 8 and CNS9 suite programs simultaneously with the H. pylori ENR structure (PDB:2PD3) as a search model.5 The resulting rotation/translation solution gave a clear solution in space group P212121 with four molecules in the asymmetric unit (AU). Manual rebuilding was performed in COOT10 and refined with REFMAC5.11 The holoenzyme structure of SaFabI was used as a model for the apo-form; form II, C2 space group (two molecules in AU) and form I, I222 space group (one molecule in AU) with a corresponding Matthews coefficient (VM) of 2.4 Å3 for both forms.12 Analysis using the program PROCHECK13 showed that no residues lie in the disallowed regions of the Ramachandran plot, with 92% and 8% being in the most favored and additionally allowed regions, respectively. The diffraction data and final refinement statistics are given in Table I. Biospecific-interaction analysis was performed using a BIAcore 2000 biosensor system (Amersham Pharmacia Biotech, Uppsala, Sweden). Approximately 1700 resonance units (RUs) were coupled after the immobilization of SaFabI (0.1 mg/mL) on a CM5 (carboxymethylated)-certified grade sensor chip in 10 mM sodium acetate buffer (pH 4.5). All measurements were carried out in 10 mM HEPES, pH 7.5/8.0, 150 mM NaCl and 2 mM EDTA (pH 7.5). To determine the association rate constant for the binding of the various compounds to the immobilized SaFabI, NADH, NADP, NAD+, INH and TCL were passed over the surfaces at various concentrations in triplicate from at 25°C. The association (k1) and dissociation (k−1) rate constants were obtained by non-linear fitting of the primary sensor gram data using the BIA evaluation software version 3.1 and were fitted globally to the 1:1 Langmuir model. The χ2 and residual values were used to evaluate the quality of fit between the experimental data and the binding model. The SaFabI subunit has ∼65% (49% α-helix and 16% β-sheet) of its residues in elements of secondary structure. The X-ray structure revealed four molecules in an AU, where the monomer structure comprised of seven β strands, 11 helices, plus a number of loops [Fig. 1(A)]. This fold is highly reminiscent of the Rossmann fold commonly found in dinucleotide-binding enzymes. The details of the extent of the secondary structure and tetramer packing are given in Figure 1(B). Gel-filtration experiments suggested that SaFabI is a tetramer and this was supported by the packing in the crystal, in which an obvious tetramer occurred with the approximate dimensions 89 Å × 82 Å × 76 Å [Fig. 1(B)]. DLS also confirmed that, when in solution, this protein behaved as a tetramer (data not shown). Each monomer in the tetramer makes contact with its symmetry-related partners. The solvent-accessible surface areas of the SaFabI tetramer was calculated using the program AREAIMOL of CCP4 (probe radius of 1.4 Å), 8 and was ∼12218 Å2 for the holoenzyme and 10270 Å2 for the apo-form. Thus, on formation of the tetramer, ∼3023 Å2 of the solvent-accessible surface is buried per monomer of the SaFabI holoenzyme (2560 Å2 for the apo-form). FabI structure and its crystallographic packing. A: Monomer structure of SaFabI. The fold is highly reminiscent of the Rossmann fold, composed of eleven α helices (α 1–11) and seven β strands (β 1–7), which create a parallel β sheet. The first β strand of the monomer is located at the extreme N-terminal end of the protein. B: Overall structure of SaFabI in an AU. The four subunits are shown in cyan, pink, orange, and green. Tetramer packing of SaFabI occurred with the approximate dimensions 89 Å × 82 Å × 76 Å. NADP shown in green and Triclosan in yellow bindind with each individual subunit of SaFabI. C: Superposition of nine FabI structures. FabI from S. aureus (SaFabI) is shown in blue, H. pylori FabI (PDB ID 2PD3) in yellow, A. aeolicus FabI (PDB ID 2P91) in green, M. tuberculosis FabI (PDB ID 3FNH) in light green, B. napus FabI (PDB ID 1D7O) in cyan, E. coli FabI (PDB ID 2FHS) in pink, P. falciparum FabI (PDB ID 2OP1) in red, B. pseudomallei FabI (PDB ID 3EK2) in orange, and B. anthracis FabI (PDB ID 2QIO) in green cyan. The C-terminal region (red circle) shows the extended structural loop region for H. Pylori FabI compared with SaFabI. In the crystal structure of the apo-form of SaFabI with both I222 (Form I) and C2 SG (Form II) there is no clear electron density for cofactor binding via the flexible loop region. In the apo-form of the SaFabI structure, residues Ile193 to Leu196 form part of a flexible loop adjacent to the active site. This loop was also reported as disordered in the ENR-NAD binary complex, 3 and adopts an open conformation with an average B-factor of 48Å2. In the open conformation, the loop is on the surface of the protein, and does not form any secondary structure or contribute to inhibitor or cofactor binding [Fig. 2 (A)]. When we superposed one molecule of apo-SaFabI (Form II) with apo-SaFabI (Form I) for 219 Cα, the overlay with an RMSD of 0.66 Å suggest that they were overall very similar, however, minor conformational changes observed between α5 and α6. In the apo-form of the SaFabI structure, residues Arg194 to Leu198 and Asn205 to Glu210 have high temperature factors and do not interact with either the coenzyme or the inhibitor. The regions Ala95 to Ala117, Gly149 to Asn156, and Ser197 to Phe204 were disordered in the apo-form of SaFabI. A previous report supports our suggestion that the Ser197 to Phe204 region is important for the more ordered flipping loop14,15. In apo-SaFabI structure, the substrate binding loop (residues 196 to 219) was shifted 3.5 Å compared to ternary SaFabI structure [Fig. 2(A)]. This observation, along with a rotation in the side chain of Tyr157 (3.3 Å distant from ternary SaFabI's Tyr157), causes the substrate-binding cavity to widen and flex, thus demonstrating the flexibility of this loop. Superposition of apo-SaFabI and ternary complex SaFabI. A: NADP is shown in green and Triclosan in yellow. Cartoon of apo-SaFabI is shown in green and the ternary complex SaFabI in cyan. B: The alpha helix (α8), considered to be a flipping loop, is shown in the red circle. Active site region moved 3.5 Å towards solvent. When we superposed one molecule of apo-SaFabI with holo-SaFabI for 219 Cα (except Ala95 to Ala117, Gly149 to Asn156, and Ser197 to Phe204), the overlay with an RMSD of 1.16 Å suggested a conformational change. Superposition of the apo-SaFabI and ternary complex structures revealed subtle conformational changes upon inhibitor binding, with a slight shift of the α6 helix (α7-8 in SaFabI ternary complex) by 3.45 Å towards the solvent. Of the 427.4 Å2 accessible hydrophobic surface area of TCL, only 72 Å2 remains accessible in the ternary complex. Indeed, the inhibitor buries 63% of its accessible surface area upon binding to the enzyme in the ternary complex. In addition, 612 Å2 of the hydrophobic surface area of NADP is buried in the ternary complex. These results provide further support for the formation of a high-affinity ternary complex between the enzyme, NADP and TCL. Molecular interactions involving conformational changes in the interacting molecules are more versatile in apo-SaFabI. The observed shift in the position of the loop in the ternary complex, as compared with its apo-form, suggests that the presence of TCL and nicotinamide may cause this change in conformation. In the SaFabI holoenzyme structure, the hydrophobic pocket, which is important for nicotinamide binding, is mostly formed by hydrophobic residues (A95-F96-A97, M99, L102, G104, F106, G113-F114-L115-L116, G149-G150, F152-A153-V154, A198, and G200-V201-G202-G203-F204) which are fully ordered and forming the loop, however, this region was disordered in the apo-form of SaFabI. It was reported that Ser197 (Ala in HP), Val201 (Ile in HP), and Phe204 play a pivotal role in the so-called substrate-binding loop of ENRs from various sources) 5, 14-18. The substrate-binding loop of holo-SaFabI containing Ser197, Val201 and Phe204 has a "closed" conformation, which is similar to that of H. pylori, E. coli and P. falciparum ENRs; however, the corresponding loop is missing in apo-SaFabI. The most significant difference observed was the disorder of the residues in the adjacent loop (region β6 to α6) in apo-SaFabI [Fig. 2(B)], which form an ordered loop in other ENR structures and in our ternary complex SaFabI structure. Thus, a close examination and comparison of the binding site interactions in ENRs reveals that the binding affinities of TCL to the enzyme are related to the involvement and positioning of the substrate-binding loop. The nicotinamide ring of NADP, as observed in the ternary complex SaFabI, interacts via five specific hydrogen bonds formed by the oxygen and nitrogen of the carboxamide moiety and the pyrophosphate moiety within the substrate-binding loop (residues 194 to 204) shown in Figure 3(A). Two positively charged residues namely Arg40 and Lys41 lie very close to 2′-phosphate of NADP and play important roles in catalytic activity as shown in Figure 3(B). The electrostatic potential of the interacting residues, Arg40 and Lys41 with phosphate moiety of NADP cofactor shown in Figure 3(C) reveals its importance in catalytic activity as described elsewhere. 6 Residues of SaFabI interacting with NADP and Triclosan. A: Interacting residues of SaFabI with NADP and Triclosan depicted with distance less than 3.2 Å. Diagrammatic view of σA-weighted simulated annealing omit maps of NADP (in green) and Triclosan (in yellow). Maps are shown at 1.5 σ for the NADP and Triclosan adduct. This figure was generated using CNS Xtal and CCP4 mapcover. B: Positive charged residues, Arg40 and Lys41 interacting with phosphate moiety of NADP. C: Electrostatic surface map around phosphate moiety of NADP interacting with Arg40 and Lys41. To investigate the similarities and differences between members of the ENR family, the structures from H. pylori, E. coli, A. aeolicus, B. napus, P. falciparum, B. anthracis, B. pseudomallei, and M. tuberculosis were compared with SaFabI 3, 5, 14-20 [Fig. 1(C)]. As might be expected on the basis of their sequence similarity, SaFabI is more closely related to the bacteria H. pylori, E. coli, A. aeolicus, and B. anthracis. It is presumed that the ENRs from all eight sources exist as homotetramers. The major difference was observed at a loop in the binding region, called the substrate-binding loop, which has been shown to be flexible in earlier studies as well as in our apo SaFabI structure. Analysis with DALI showed Z scores of 37.8 (HP), 38.8 (AA), 32.1 (BN), 34.3 (PF), 38.9 (EC), 36.3 (BP), 41.8 (BA) and 34.6 (MT), respectively, that reflects their lower similarity. The binding of SaFabI with its cofactors and inhibitors was analyzed by SPR. The binding of the cofactors NADH, NADP and NAD+ to SaFabI was monitored in real time by observing changes in the response units shown in Table II. The overall dissociation constant Kd (M) observed for NADP to SaFabI was 1.86 × 10−5M at pH 8.0, and 1.3 × 10−5M at pH 7.5 reveal the effect of pH. A lower affinity of 3.46 × 10−4M was observed for the NAD+ and no binding was observed for the products of the reaction, namely, NADH and INH. TCL bound to the enzyme with a binding constant of 3.43–5.47 × 108M−1. In conclusion, we have used a combination of X-ray crystallography and a biological assay to provide the first insights into the structure of SaFabI's mobile and flexible loop (which is mainly comprised of hydrophobic residues), and its dependency on nicotinamides for structural stability. The far-UV CD spectra of SaFabI showed that it retained a well-folded structure in presence of NADP, however, ellipticity was decreased with TCL (data not shown). SaFabI, unlike its H. pylori counterpart, interacts with its substrate, and cofactor NADP, independently. Consequently, several new contacts between SaFabI, NADP and TCL are formed in addition to the enhancement of many existing ones in apo-SaFabI. Indeed, the formation of the ternary complex is aided by a conformational transition. SPR analysis confirmed that SaFabI has a higher affinity for NADP than for NAD+. The binding experiment with TCL revealed a picomolar affinity, which should provide further insights into a selective approach for structural-based drug design. Coordinates have been deposited in the Protein Data Bank (accession codes PDB 3GNS and PDB 3GNT for apo-SaFabI and PDB 3GR6 for ternary complex SaFabI). The authors thank Dr. H. S. Lee and his staff for assistance during data collection at beamline 4A and 6C of Pohang Light Source, Korea.