Mechanism and Inhibition of saFabI, the Enoyl Reductase from Staphylococcus aureus

Abstract

About one third of the world’s population carries Staphylococcus aureus. The recent emergence of extreme drug resistant strains that are resistant to the “antibiotic of last resort”, vancomycin, has caused a further increase in the pressing need to discover new drugs against this organism. The S. aureus enoyl reductase, saFabI, is a validated target for drug discovery. In order to drive the development of potent and selective saFabI inhibitors, we have studied the mechanism of the enzyme and analyzed the interaction of saFabI with triclosan and two related diphenyl ether inhibitors. Results from kinetic assays reveal that saFabI is NADPH-dependent, and prefers acyl carrier protein substrates carrying fatty acids with a long acyl chains. Based on product inhibition studies, we propose that the reaction proceeds via an ordered sequential ternary complex, with the ACP substrate binding first, followed by NADPH. The interaction of NADPH with the enzyme has been further explored by site-directed mutagenesis, and residues R40 and K41 have been shown to be involved in determining the specificity of enzyme for NADPH compared to NADH. Finally, in preliminary inhibition studies, we have shown that triclosan, 5-ethyl-2-phenoxyphenol (EPP) and 5-chloro-2-phenoxyphenol (CPP) are all nanomolar slow-onset inhibitors of saFabI. These compounds inhibit the growth of S. aureus with MIC values of 0.03-0.06 µg/mL. Upon selection for resistance, three novel safabI mutations, A95V, I193S and F204S, were identified. Strains containing these mutations had MIC values ~100 fold larger than the wild-type strain whereas the purified mutant enzymes had K_i values 5 to 3000 fold larger than the wild-type saFabI. The increase in both MIC and K_i values caused by the mutations supports the proposal that saFabI is the intracellular target for the diphenyl ether-based inhibitors.

Staphylococcus aureus is one of the most common causative agents of hospital-acquired infections (1, 2). It is estimated that 30% of healthy people carry S. aureus, usually in the anterior nares, providing a ready reservoir for future infection (3). S. aureus is able to acquire resistance to antibiotics rapidly, and penicillin-resistant strains developed after only a few months of clinical trials (4). After methicillin was introduced to treat penicillin-resistant S. aureus in 1960, methicillin-resistant S. aureus (MRSA) was isolated within one year (5). The percentage of MRSA has been increasing continuously in the last few decades. In the USA, MRSA increased from 2.4% in 1975 (6) to 35% in 1996 (7). Currently, glycopeptide antibiotics, such as vancomycin and teicoplanin, are considered to be the only antibiotics for treating MRSA infections. However, the recent emergence of clinical isolates of vancomycin-resistant S. aureus (VRSA) (8, 9) reaffirm the pressing need to continuously discover new antibiotics against this bacterium.

Fatty acid biosynthesis is divided into two types, FASI and FASII, based on whether the reactions are performed by a single polypeptide or by individual enzymes. The FASI pathway usually exists in vertebrates and certain bacteria, whereas the FASII pathway, as shown in Scheme 1, is usually found in plants and bacteria. The inhibition of the FASII pathway in bacteria causes the breakdown of the cell wall and the disruption of the cell membrane (10, 11), demonstrating its importance for bacterial survival. A number of studies suggest that the enoyl reductase (FabI), which catalyzes the final and rate-limiting step in each cycle, is a regulator of the FASII pathway, and is essential for the viability of bacteria (12). Due to its necessity and low sequence homology to the mammalian FASI reductase, FabI is an attractive target for novel antibiotic discovery. Isoniazid, which has been used in tuberculosis (TB) chemotherapy for decades, is now known to target InhA, the FabI homologue in Mycobacterium tuberculosis (13). A number of inhibitors against FabI have been reported recently (14-19). In the present study we have examined the mechanism of saFabI, as a prelude to the development of potent inhibitors of this enzyme.

Although kinetic studies on saFabI have been conducted previously, a detailed mechanistic analysis of the enzyme has so far not been reported. In addition, existing studies are based on N-acetylcysteamine (NAC) or CoA based-substrates, rather than the natural acyl carrier protein (ACP) substrates (20, 21). In the present work we have compared the kinetic properties of NAC, CoA and ACP-based substrates with saFabI, explored the molecular basis for cofactor specificity, and also studied the mechanism of the saFabI-catalyzed reaction. In addition, we have also analyzed the inhibition of saFabI by triclosan and two related diphenyl ethers, 5-ethyl-2-phenoxyphenol (EPP) and 5-chloro-2-phenoxyphenol (CPP) (Scheme 2). Triclosan is a broad-spectrum antimicrobial that is present in a wide variety of consumer products, such as toothpaste, mouthwashes and hand soaps (22). It has been reported that triclosan can inhibit the growth of S. aureus (23), and this compound is recommended as a method of controlling MRSA in hospitals (24, 25). In order to extend previous studies concerning the mode of action of triclosan (20, 21, 26, 27), we performed selection experiments and identified several saFabI mutations that correlated with an increase in resistance to the three inhibitors. Analysis of the impact of these mutations on cell growth and enzyme inhibition strongly suggests that the diphenyl ether based inhibitors target saFabI within the bacterium.

MATERIALS AND METHODS

Materials

trans-2-Dodecenoic acid was purchased from TCI. His-bind Ni²⁺-NTA resin was obtained from Invitrogen, and centriplus units were from Millipore. QuikChange site-directed mutagenesis kit was obtained from Stratagene. Triclosan was a gift from Ciba whereas 5-ethyl-2-phenoxyphenol (EPP) and 5-chloro-2-phenoxyphenol (CPP) were available from a previous study (28, 29). All other chemical reagents were purchased from Sigma-Aldrich.

Synthesis of trans-2-dodecenoyl CoA and trans-2-dodecenoyl N-acetylcysteamine

trans-2-Dodecenoyl CoA (DDCoA) was synthesized by using the mixed anhydride method as described previously (30). This method was also used to synthesize trans-2-dodecenoyl N-acetylcysteamine (DDNAC) with minor modifications. Briefly, 4 mmol N-acetylcysteamine was dissolved in 10 mL anhydrous THF, and then 4 mmol of the mixed anhydride was added. After stirring at room temperature for 2 hours, DDNAC was purified by silica gel chromatography using 1:1 ethyl acetate:hexane as the eluant. The identity of the product was confirmed by NMR spectroscopy and ESI mass spectrometry. ¹H NMR: δ 0.880 (3H, t), 1.265 (12H, m), 1.964 (3H, s), 2.165-2.237 (2H, q), 3.090 (2H, t), 3.435-3.496 (2H, q), 6.094-6.156 (1H, m), 6.886-6.984 (1H, m). Mass: m/z 300.2 (M+H)⁺.

Expression and Purification of saFabI

The fabI gene from the methicillin-sensitive S. aureus strain NCTC 8325 was amplified using the KOD DNA polymerase (Novagen) and the primers listed in Table S1. The 0.77 kb PCR product was digested with NdeI and BamHI, and cloned into the same restriction sites in pET-16b. The sequence of the construct was confirmed by DNA sequencing (Sequetech) and the plasmid was then transformed into Escherichia coli strain BL21(DE3)plysS. A single colony was used to inoculate 10 mL of LB medium containing 50 µg/mL ampicillin (LB/AMP) and the culture was grown at 37 ºC overnight. The cells in the starter culture were collected by centrifugation, resuspended in fresh media, transferred to 500 mL LB/Amp medium, and grown at 37 ºC until an optical density of 0.8 at 600 nm had been reached. Subsequently, 0.5 mM IPTG was added to induce the expression of saFabI and the culture was shaken overnight at 25 ºC. The cells were harvested, and resuspended in 30 mL His binding buffer (20 mM Tris HCl, 500 mM NaCl, and 5 mM imidazole, pH 7.9). After cell disruption by three passages through a French press cell (1000 psi), the cell lysate was centrifuged at 120,000 x g for 1 h, and the supernatant was loaded onto a 5 mL Ni²⁺-NTA His-bind column, which had been preincubated with 20 mL binding buffer. The His-bind column was washed with 40 mL binding buffer, followed by 30 mL of wash buffer (20 mM Tris HCl, 500 mM NaCl, and 60 mM imidazole, pH 7.9), and saFabI was then eluted using 30 mL of elute buffer (20 mM Tris HCl, 500 mM NaCl, and 500 mM imidazole, pH 7.9). SDS-PAGE was used to identify those fractions containing saFabI, and these fractions were combined and subjected to size exclusion chromatography on a G25 column using 30 mM PIPES, 150 mM NaCl, and 100 mM EDTA, pH 7.8 buffer as the mobile phase. Fractions containing saFabI were pooled and concentrated using a centriplus 30.

Site-directed mutagenesis, expression and purification of FabI mutants

Site-directed mutagenesis was performed using the QuikChange mutagenesis kit from Stratagene using the primers listed in Table S1. The sequence of each mutant plasmid was confirmed by ABI DNA sequencing, and the expression and purification of each saFabI mutant followed the same protocol as that described above for the wild type saFabI protein.

Cloning, expression and purification of saACP

The S. aureus acp gene was PCR amplified from NCTC 8325 genomic DNA using the primers listed in Table S1, digested with NdeI and XhoI, then ligated into the pET23b expression vector so that the coding region of the gene was in-frame with a C-terminal His tag. After DNA sequencing, saACP was expressed and purified using the procedure described above for saFabI. After purification, saACP was dialyzed into 75 mM Tris-HCl, 10 mM MgCl₂, pH 7.5 and the protein concentration determined using the Bradford method (31).

Enzymatic preparation of crotonyl saACP and dodecenoyl saACP

Conformationally sensitive SDS PAGE and ESI-MS revealed that saACP expressed and purified from E. coli was obtained primarily in the apo form. This protein was converted to the required acyl-saACP using sfp, a phosphopantetheinyl transferase from Bacillus subtilis (32). Briefly, 25 mg/L apo-saACP was incubated with 80 μM crotonyl CoA or dodecenoyl CoA in the presence of 1 μM sfp for 2 h at 37 ºC. The reaction mixture was then loaded onto a Q-sepharose column, and eluted with a gradient consisting of buffer A (20 mM Bis-Tris, pH 7.0) and buffer B (20 mM Bis-Tris, 800 mM NaCl, pH 7.0). Fractions were analyzed by SDS-PAGE and those containing crotonyl-saACP or DDsaACP were pooled and dialyzed into 20 mM Tris-HCl, pH 7.0 buffer. The final products were characterized by ESI mass spectrometry.

Steady-state kinetic assays

Steady-state kinetics were performed at 25 °C in 100 mM Na₂HPO₄ buffer, pH 7.8. Initial velocities were measured after addition of 50 nM saFabI to give a final assay volume of 500 μL.

For DDsaACP, the reactions were carried out by varying the concentration of DDsaACP at several fixed concentrations of NADPH (50, 100 and 250 μM) or by varying the NADPH concentration at several fixed concentrations of DDsaACP (4.4, 8.2 and 15.0 μM). Kinetic parameters were obtained by fitting the data to equation 1 for a sequential mechanism with GraFit 4.0.

$$v=V_{\max }\left[\mathrm{A}\right]\left[\mathrm{B}\right]∕(K_{\mathrm{ia}}{K}_{\mathrm{mb}}+K_{\mathrm{ma}}\left[\mathrm{B}\right]+K_{\mathrm{mb}}\left[\mathrm{A}\right]+\left[\mathrm{A}\right]\left[\mathrm{B}\right])$$ (1)

K_ma and K_mb, are the K_m values of DDsaACP and NADPH, respectively, whereas K_ia is the dissociation constant for DDsaACP.

For crotonyl-saACP, as well as the NAC and CoA-based substrates, the kinetic studies employed a concentration range for one substrate (0.3 K_m – 4 K_m) at near saturating concentration of the second substrate. Kinetic parameters were calculated by fitting the data to the Michaelis-Menten equation (2) using GraFit 4.0.

$$v=V_{\max }\left[\mathrm{S}\right]∕(K\mathrm{m}+\left[\mathrm{S}\right])$$ (2)

k_cat values were obtained by using the relationship between k_cat and V_max: V_max = k_cat [E].

Product inhibition assays

Product inhibition studies were conducted by varying the concentration of DDsaACP at a fixed concentration of NADPH (250 μM) and at different fixed concentrations of NADP⁺ (0, 1.4 and 2.7 mM) or by varying the concentration of NADPH at a fixed concentration of DDsaACP (20 μM) and at different fixed concentrations of NADP⁺ (0, 0.7 and 1.4 mM).

The type of inhibition was determined using Lineweaver-Burk plot analysis. Inhibition constants for competitive, uncompetitive, and non-competitive inhibitor were determined by globally fitting all the data points to the equations 3-5, respectively, by using GraFit 4.0.

$$v=V_{\max }\left[\mathrm{S}\right]∕(K_{\mathrm{m}}(1+\left[\mathrm{I}\right]∕K_{\mathrm{is}})+\left[\mathrm{S}\right])$$ (3)

$$v=V_{\max }\left[\mathrm{S}\right]∕(K_{\mathrm{m}}+\left[\mathrm{S}\right](1+\left[\mathrm{I}\right]∕K_{\mathrm{ii}}))$$ (4)

$$v=V_{\max }\left[\mathrm{S}\right]∕(K_{\mathrm{m}}(1+\left[\mathrm{I}\right]∕K_{\mathrm{is}})+\left[\mathrm{S}\right](1+\left[\mathrm{I}\right]∕K_{\mathrm{ii}}))$$ (5)

The abbreviations K_is and K_ii represent the K_i slope and intercept in the double-reciprocal plot according to the nomenclature of Cleland (33).

Fluorescence titration

Equilibrium fluorescence titrations were performed using a Fluorolog-3-21 fluorimeter (Spex) at 25°C by making microliter additions of ligand to a solution containing 1 µM saFabI. In the fluorescence titration of saFabI with ACP or the diphenyl ethers, the excitation wavelength was 290 nm and the emission wavelength was 335 nm. When titrating saFabI with NADPH, the excitation wavelength was 350 nm, and the emission wavelength was 460 nm. A control experiment was performed using an identical procedure except that enzyme was omitted from the cuvette. K_d values were calculated by fitting the data to equation 6.

$$\Delta ={\Delta F}_{\max }\{(K_{\mathrm{d}}+\mathrm{E}+\mathrm{L})-{[{(K_{\mathrm{d}}+\mathrm{E}+\mathrm{L})}^2-4K_{\mathrm{d}}\mathrm{L}]}^{1∕2}\}∕\left(2\mathrm{E}\right)$$ (6)

Selection for resistance

Cells from the methicillin-resistant S. aureus strain N315 were incubated at 37°C in Mueller-Hinton broth (Becton-Dickinson) containing 0.02 µM triclosan, 0.08 µM EPP, or 0.08 µM CPP, respectively. After 48 h, 100 µl of each culture was plated onto Mueller-Hinton agar medium containing 0.08 µM triclosan, 0.15 µM EPP, and 0.15 µM CPP, respectively. The plates were incubated at 37 °C for 24 h. Resistant colonies were picked and their phenotype confirmed by regrowth on the same media containing selective concentrations of triclosan, EPP, and CPP, respectively. The S. aureus fabI gene from the diphenyl ether-resistant mutants was characterized by double-strand nucleotide sequencing of PCR products amplified with Ex-taq polymerase (Takara Bio, Shiga, Japan) and the following primers: SafabI-F (5'-GTCATCATGGGAATCGCTAA-3') and SafabI-R (5'-GCGTGGAATCCGCTATCTAC-3'). Sequencing reactions were performed using the above SafabI-F and SafabI-R primers with the ABI PRISM^TM BigDye^TM Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, Calif), and sequencing data were obtained using an Applied Biosystems 3100 DNA sequencer (Applied Biosystems).

MIC measurement

The MIC values were determined by the microbroth dilution assay according to the Clinical and Laboratory Standards Institute methods for antimicrobial susceptibility tests for aerobically growing bacteria (34).

Preincubation inhibition assays for slow-binding inhibitors

Kinetic assays were performed essentially as described previously (28). Briefly, a 500 μL reaction mixture containing 50 nM saFabI, 250 μM NADPH, and various concentrations of diphenyl ether inhibitors (0 – 2 μM) at different fixed concentrations of NADP⁺ (20 – 800 μM) were preincubated at 4 ºC for 3 hours. The reaction mixture was then warmed to room temperature, and the reaction was initiated by the addition of 20 μM DDsaACP. Apparent inhibition constants (K_i,app) were calculated by fitting the data to equation 7 at each concentration of NADP⁺

$$v_{\mathrm{i}}∕v_0=1∕(1+\left[\mathrm{I}\right]∕K_{i,\mathrm{app}}$$ (7)

Subsequently, the series of K_i,app values were fit to equation 8 to obtain K₁ and K₂, which represent the inhibition constants for inhibitor binding to NADP⁺ bound or NADPH bound forms of the enzyme, respectively.

$$K_{i,\mathrm{app}}=K_2(1+\left[{\mathrm{NADP}}^{+}\right]∕K_{\mathrm{mNADP}})∕(1+\left[{\mathrm{NADP}}^{+}\right]∕(K_{\mathrm{mNADP}}{K}_1∕K_2))$$ (8)

Steady-state inhibition studies of diphenyl ethers

Steady-state inhibition experiments were performed by varying the concentration of NADPH at a fixed concentration of DDsaACP (20 μM) and at different fixed concentrations of triclosan, EPP or CPP. Reactions were initiated by the addition of the enzyme. Unless noted, the enzyme concentration was 50 nM. The data analysis was performed as described for the product inhibition assays.

RESULTS AND DISCUSSION

Expression and purification of wild-type and mutant saFabI proteins

The protein was purified using His-tag affinity chromatography, providing homogeneously pure recombinant protein with the predicted molecular mass (~30 kDa) as determined by SDS-PAGE. As previously reported (20), the protein precipitated if 100 mM EDTA was omitted from the G25 buffer. In addition, we also observed that saFabI would precipitate even in 100 mM EDTA when the enzyme concentration was higher than 40 μM.

Preparation of ACP substrates

saACP was purified by His-tag affinity chromatography, and only one form of ACP was observed on a 15% Tris-glycine gel, unlike the ACP (AcpM) from Mycobacterium tuberculosis which is obtained in three forms (apo, holo and acyl form) (35). The ESI mass of saACP after deconvolution was 9711.6 ± 0.4, which is close to the calculated mass of apo-saACP (9707.7), suggesting that only the apo form was expressed. Subsequently, apo-saACP was converted into crotonyl saACP or dodecenoyl saACP (DDsaACP) by sfp, the phosphopantetheine transferase from Bacillus subtilis. According to analysis by SDS-PAGE, the apparent molecular mass of DDsaACP is smaller than apo-ACP, probably due to favorable hydrophobic interactions between the dodecenoyl acyl chain in DDsaACP and the SDS. The MW of DDsaACP was confirmed as correct by ESI mass analysis.

Substrate specificity

As shown in Table 1, saFabI has a 180-fold higher k_cat/K_m with NADPH than with NADH, suggesting that saFabI is an NADPH-dependent enzyme, which is consistent with previous results when trans-2-octenoyl N-acetylcysteamine was used as the substrate (20). As expected, saFabI also has a preference for ACP-linked substrates compared to those attached to NAC or CoA, indicating that saFabI is specific for its natural substrate. While we were able to detect binding of apo-saACP to saFabI (K_d = 18.2 µM) by fluorescence titration (Figure S1), no change in fluorescence could be detected upon addition of NAC or CoA, suggesting that these carriers do not bind strongly to the enzyme in agreement with the kinetic data. In addition, the enzyme also had a preference for a long chain (C12) substrate compared to the shorter (C4) acyl group, with DDsaACP having a k_cat/K_m value ~30 fold larger than for the corresponding crotonyl-ACP substrate. This effect is also observed for the NAC or CoA-based substrates (Table 1). Interestingly, the effect of the acyl chain length on k_cat/K_m is primarily an effect on catalytic efficiency (k_cat) rather than K_m, suggesting that remote interactions between the longer acyl chain and the enzyme modulate the precise orientation of the catalytic groups in the active site. Finally, both the NAC and CoA-based substrates suffered from some disadvantages, the NAC substrates possessing limited aqueous solubility even in the presence of 5% DMSO, whereas DDCoA exhibited substrate inhibition at high concentrations (60 μM). Substrate inhibition was not observed with DDNAC or DDsaACP, which leads to the hypothesis that the adenosine portion of CoA can bind non-productively in the NADPH binding-site.

Table 1.

Kinetic Parameters for the Reduction of Different Substrates with saFabI

Substrates	kcat (min−1)	Km (μM)	kcat/Km (min−1 μM−1)
Crotonyl NACa	1.5 ± 0.5	8.0 ± 0.3	0.20 ± 0.06
Crotonyl CoAb,c	-	-	-
Crotonyl saACPc	11.6 ± 1.1	11.5 ± 2.0	1.1 ± 0.2
Dodecenoyl NACc	30.4 ± 1.6	23.3 ± 2.9	1.3 ± 0.2
Dodocenoyl CoA c	18.0 ± 0.8	24.1 ± 2.5	0.8 ± 0.1
Dodecenoyl saACP d	130.2 ± 9.1	4.5 ± 0.5	29.5 ± 5.0
NADHe	-	-	0.010 ± 0.001
NADPH d,e	130.2 ± 9.1	70.8 ± 6.0	1.840 ± 0.311

^aParameters were measured by Heath et al. (20);

^bno enzymatic activity was observed;

^creactions were carried out with 300 μM NADPH;

^dK_m and k_cat values were obtained by fitting data into the sequential mechanism equation;

^ereactions were carried out with 20 μM DDsaACP.

It is interesting that saFabI is NADPH-dependent, whereas the FabI homologues from E. coli, B. subtilis, Haemophilus influenzae and M. tuberculosis, are all reported to be NADH-dependent ACP reductases (36-38). According to the ecFabI-NAD⁺ crystal structure (Figure 1, pdb code 1DFI), Q40 is very close (2.8 Å) to the 2'-hydroxyl group of the NAD⁺ adenosine moiety. Sequence alignment of the FabI proteins from different organisms (Figure 2), indicates that two positively-charged residues, R40 and K41, appear close to the position of Q40 in saFabI, whereas nonpolar residues (L, F), or a negatively charged residue (E) is present in the cases of the FabIs from B. subtilis, H. influenzae and M. tuberculosis, respectively. Basic residues, such as R or K, in this position are expected to interact strongly with the 2’-phosphate of NADPH. In order to investigate the importance of R40 and K41 in the interaction of NADPH with saFabI, we constructed two single-mutants, R40Q and K41N, and found that both mutants show at least a 50-fold decrease in k_cat/K_m for NADPH_, whereas k_cat/K_m for NADH increases by 5 to 7 fold (Table 2). The double mutation R40Q/K41N further decreases k_cat/K_m for NADPH by 10-fold, proving that both R40 and K41 are involved in interactions with 2’-phosphate of NADPH. Since all the residues apart from Q40 which have polar interactions with NAD⁺ in the ecFabI-NAD⁺ crystal structure are highly conserved in saFabI, it might be expected that the cofactor specificity would be completely reversed in the saFabI double mutant. However, the NADH-specificity of the double mutant remains almost the same as the single mutants, only 7 fold higher than that of the wild type saFabI, indicating that other factors are critical for optimal binding of NADH.

Figure 1. Structure of the ecFabI:NAD⁺ Complex.

Structure of ecFabI complexed with NAD⁺ (pdb code 1DFI) showing interactions between the protein and the NAD⁺ ribose. ecFabI is colored yellow and polar interactions between the protein residues (green) and NAD⁺ (cyan) are indicated with red dashed lines. Q40 interacts with the 2’-hydroxyl group in the adenosine moity of NAD⁺. The Figure was made using pymol (44).

Figure 2. FabI Sequence Alignment.

Sequence alignment of FabIs from E. coli, S. aureus, H. influenzae, M. tuberculosis and B. subtilis performed using ClustalW. The red-colored amino acids are proposed to interact with the cofactor, according to the X-ray structure of ecFabI:NAD⁺ complex.

Table 2.

Effect of Mutagenesis on the Specificity of saFabI for NADH and NADPH

Enzyme	kcat/Km (min−1 μM−1)
	NADPH	NADH
WT	1.840 ± 0.311	0.010 ± 0.001
R40Q	0.020 ± 0.001	0.053 ± 0.003
K41N	0.036 ± 0.002	0.072 ± 0.002
R40Q/K41N	0.004 ± 0.001	0.076 ± 0.005

Enzymatic mechanism of the saFabI-catalyzed reaction

As expected, double-reciprocal plots of the enzyme kinetic data generated intersecting lines (Figure S2A and S2B), which is a characteristic of a ternary complex mechanism in which NADPH transfers a hydride directly to the substrate. In order to determine the order of substrate binding, the effect of the product NADP⁺ on the reaction was analyzed. Initial velocities were measured at various concentrations of DDsaACP and a fixed concentration of NADPH and at different fixed concentrations of NADP⁺. The parallel lines in the double reciprocal plot that were obtained (Figure S3A) suggest that NADP⁺ is an uncompetitive inhibitor with respect to DDsaACP. However, the assays were also performed with NADPH as the varied substrate. As shown in Figure S3B, the lines intersect to the left of the y-axis, indicating that the NADP⁺ is a non-competitive inhibitor with respect to NADPH. The inhibition patterns of NADP⁺ against the two substrates, are only consistent with an ordered bi-bi mechanism, in which DDsaACP binds first, followed by NADPH.

In fluorescence titration experiments, no fluorescence change was observed when titrating saFabI with NADPH, suggesting that NADPH does not bind to saFabI, or that binding does not affect the fluorescence of the NADPH fluorophore. In contrast, a K_d value of 2.3 μM was obtained when saFabI was titrated with DDsaACP (Figure S4). These results provide further support for the proposed ordered bi-bi mechanism.

Antimicrobial activity of the diphenyl ether saFabI inhibitors

The molecular basis for the antibacterial activity of triclosan against S. aureus has attracted much attention. It was proposed that triclosan inhibits S. aureus by targeting the enoyl reductase in this organism (20, 21, 26), since it is known that triclosan inhibits the FabI homologue in E. coli (39-41), and that saFabI shows 43% sequence identity to ecFabI. However, the observation that small-colony variants of S. aureus have increased resistance to triclosan suggests more complex underlying mode of action of this antimicrobial agent (27).

As a first step in developing novel inhibitors against S. aureus, we synthesized two diphenyl ether analogues of triclosan: 5-ethyl-2-phenoxy-phenol (EPP) and 5-chloro-2-phenoxy-phenol (CPP). Both compounds have similar MIC values to triclosan against different strains of S. aureus (Table S2). It is very encouraging that the MIC values are similar for both methicillin-sensitive (ATCC 29213) and methicillin-resistant S. aureus strains (N315, Mu50), suggesting that the diphenyl ethers are orthogonal to current antibiotics. In order to probe the mechanism of action of the diphenyl ether inhibitors, selection experiments resulted in the identification of three novel mutations in the fabI gene, A95V, I193S and F204S (Table 3), suggesting that saFabI is the intracellular target for these compounds. Based on the ecFabI/NAD/triclosan crystal structure (42), the three residues are thought to lie in the cofactor binding pocket of saFabI (Figure 3). Interestingly, the residues corresponding to A95 and F204 in ecFabI (G93 and F203), were also found to be mutated in triclosan-resistant E. coli. In addition, the residue corresponding to I193 in ecFabI (I192) interacts with the cofactor, as shown in Figure 3. As expected, the three saFabI mutations affected the kinetic parameters associated with NADPH more significantly than the parameters for the ACP substrate (Table 4). In particular, the k_cat/K_m value for NADPH decreased by 240 fold for the A95V mutant, while the k_cat/K_m value for DDsaACP decreased only 18 fold, confirming that the mutation has a major impact on the interaction of the cofactor with the enzyme.

Table 3.

Selection for Resistance to the Diphenyl Ether saFabI Inhibitors

Mutants derived from S. aureus N315	Selections for mutants resistant to compound	fabI gene
		Nucleotide changes a	Amino acid changes
N315Tr.4	Triclosan	GCA → GTA	Ala95 → Val
N315Tr.5	Triclosan	GCA → GTA	Ala95 → Val
N315PT01.1	EPP	GCA → GTA	Ala95 → Val
N315PT01.2	EPP	GCA → GTA	Ala95 → Val
N315PT01.3	EPP	GCA → GTA	Ala95 → Val
N315PT01.4	EPP	GCA → GTA	Ala95 → Val
N315PT52.1	CPP	ATC → AGC	Ile 193 → Ser
N315PT52.2	CPP	TTC → TCC	Phe204 → Ser
N315PT52.3	CPP	TTC → TCC	Phe204 → Ser
N315PT52.4	CPP	TTC → TCC	Phe204 → Ser

^aMutation sites are shown in underline.

Figure 3. Structure of ecFabI Complexed with NAD⁺ and Triclosan.

Structure of triclosan bound to ecFabI (pdb code 1D8A) showing the proximity of residues G93, I192 and F203 to the inhibitor binding site. The corresponding residues in saFabI were found to be mutated in the diphenyl ether resistant S. aureus strains. ecFabI is colored yellow, while the polar interactions between the residues (green) in ecFabI and NAD/triclosan (cyan), are indicated with red dashed lines. The Figure was made using pymol (44).

Table 4.

Kinetic Parameters for the Reduction of DDsaACP by Wild Type and Mutant FabIs

Enzyme	kcat (min−1)	NADPH		DDsaACP
		Km (μM)a	kcat/Km (min−1 μM−1)a	Km (μMb	kcat/Km (min−1 μM−1)b
WT	130.2 ± 9.1	70.8 ± 6.0	1.8 ± 0.3	4.5 ± 0.5	29.5 ± 5.0
A95V	18.4 ± 0.1	269.4 ± 29.2	0.007 ± 0.001	11.4 ± 1.6	1.7 ± 0.2
I193S	-	> 1000	0.13 ± 0.02	-	5.1 ± 0.4
F204S	157.3 ± 7.2	429.6 ± 45.0	0.17 ± 0.02	18.4 ± 1.3	7.8 ± 0.7

^aReactions were carried out with 20 μM DDsaACP;

^breactions were carried out with 300 μM NADPH.

To analyze further the affect of the mutations on enzyme inhibition, we quantified the interaction of the three diphenyl ether inhibitors with both wild-type and mutant saFabIs. In the case of wild-type saFabI, all three compounds were slow-onset inhibitors, binding preferentially to the enzyme-NADP⁺ product complex. The slow step in the formation of the final enzyme-inhibitor complex is believed to result from the ordering of a flexible loop that covers the active site, as seen for the inhibition of ecFabI by triclosan. Since the loop in ecFabI (residues 193-202) is highly conserved in saFabI, we propose that the slow step during the inhibition of saFabI by the diphenyl ethers is also caused by the loop ordering. We used the preincubation method to measure K₁ and K₂, which represent the equilibrium dissociation constants of the compounds from the E-NADP⁺ and E-NADPH forms of the enzyme, respectively. According to the inhibition results (Figure 4A-D), the compounds are uncompetitive inhibitors of NADP⁺, and competitive inhibitors of NADPH. K₁ is 20 to 40 fold smaller than K₂, suggesting that the diphenyl ethers bind preferentially to E-NADP⁺. In addition, the K₁ values correlate well with the MIC data, suggesting that saFabI is the target of this class of inhibitor. The use of equation 6 to analyze the inhibition data is based on the assumption that the inhibitors do not bind tightly to the free enzyme. In agreement with this assumption, fluorescence titrations of saFabI with the three inhibitors (Figure S5A-S5C) gave K_d values between 50 and 200 µM, at least 200-fold higher than either K₁ or K₂.

Figure 4. Inhibition of Wild Type saFabI by the Diphenyl Ether Inhibitors.

(A) Dependence of K_i,app on [NADP⁺] for inhibition by triclosan. (B) Dependence of K_i,app on [NADPH] for inhibition by triclosan. (C) Dependence of K_i,app on [NADP⁺] for inhibition by EPP. (D) Dependence of K_i,app on [NADP⁺] for inhibition by CPP.

In contrast to the wild-type enzyme, the diphenyl ether compounds are classical reversible inhibitors of the saFabI mutant enzymes (Figures S6-S8). Triclosan, CPP and EPP are uncompetitive or noncompetitive inhibitors of the mutant saFabIs with respect to NADPH, with K_i values 5 to 3000 fold higher than for the wild-type enzyme. This increase in K_i correlates with the increase in the MIC values observed for the S. aureus strains harboring the safabI mutations (Table 5). This further substantiates the hypothesis that saFabI is the target of diphenyl ethers within this bacterium. Interestingly, for the mutant I193S, K_i values of triclosan, EPP and CPP increase only by 5 to 25 fold, but the MICs increase by more than 60 fold. This suggests that additional mechanisms of resistance are operative, such as the overexpression of the mutant saFabI as reported by Fan et al. (26). We also observed that the frequency with which each mutation occurred also correlated with the effect of that mutation on enzyme inhibition. Thus, A95V, the most frequent mutation in the resistant strains, had the largest impact on inhibition of saFabI by the diphenyl ether inhibitors.

Table 5.

Inhibition of Wild Type and Mutant FabIs by the Diphenyl Ether Inhibitors

Enzyme	Compound	K1 (nM)	K2 (nM)	Inhibition pattern	MIC against S. aureus N315 μg/mL (μM)
WT	Triclosan	4.9 ± 0.6a	375.1 ± 20.3a	UC with respect to NADP+.	0.03 (0.10)
		5.6 ± 0.2b	216.4 ± 7.5b	C with respect to NADPH.
	EPPa	8.3 ± 0.8	273.1 ± 13.2	UC with respect to NADP+.	0.03 (0.14)
	CPPa	2.0 ± 0.2	48.7 ± 1.6	UC with respect to NADP+.	0.06 (0.27)
	Triclosanc	8100 ± 1190	-	UC with respect to NADPH.	2.80 (10)
A95V	EPPc	1660 ± 50	-	UC with respect to NADPH.	4.40 (20)
	CPPc	6150 ± 160	-	UC with respect to NADPH.	8.80 (40)
	Triclosand	35 ± 18	473 ± 196	NC with respect to NADPH.	2.80 (10)
I193S	EPPd	41 ± 5	783 ± 281	NC with respect to NADPH.	8.80 (40)
	CPPd	52 ± 2	690 ± 320	NC with respect to NADPH.	17.60 (80)
	Triclosanc	210 ± 20	-	UC with respect to NADPH.	0.70 (2.5)
F204S	EPPd	160 ± 60	300 ± 110	NC with respect to NADPH.	2.20 (10)
	CPPc	570 ± 30	-	UC with respect to NADPH.	8.80 (40)

^aNADP⁺ was varied;

^bNADPH was varied;

^cUncompetitive. K_ii has been listed under K₁. Enzyme concentration was 200 nM;

^dnoncompetitive inhibition. K_ii has been listed under K₁, and K_is has been listed under K₂. K_ii and K_is were defined by Cleland previously (33).

Finally, we also analyzed the ability of the three compounds to inhibit the growth of Enterococcus faecalis. This organism contains both a FabI enoyl reductase as well as a FabK homologue that is not sensitive to triclosan (43). In Table S3 it can be seen that the three diphenyl ethers have MIC values against E. faecalis that are ~1000-fold higher than for S. aureus again supporting the contention that the antibacterial action of these compounds is directed against FabI.

CONCLUSION

In summary, we have studied the mechanism of the reaction catalyzed by saFabI using substrates based on the natural ACP carrier molecule. From the results of the kinetic assays, we propose that the binding of substrates in the ternary complex mechanism is ordered, with DDsaACP binding first followed by NADPH. Two residues, R40 and K41 are shown to be critical for cofactor specificity. In addition, we investigated the mechanism of action of three diphenyl ether-based saFabI inhibitors. All three compounds are slow-onset inhibitors of saFabI in vitro, binding preferentially to the E-NADP⁺ product complex with K₁ values of 2-8 nM. These compounds have potent antibacterial activity, with MIC values against S. aureus of 0.03-0.06 µg/mL. Selection experiments led to the identification of three mutations in the safabI gene, A95V, I193S and F204S, which correlate with resistance to the diphenyl ethers and cause a significant reduction in the affinity of the inhibitors for the enzyme. These experiments confirm the hypothesis that the diphenyl-ether-based inhibitors target saFabI in live cells, and substantiate the view that saFabI is a novel target for antibacterial drug discovery.

Scheme 1.

The Type II Fatty Acid Biosynthesis Pathway.

Scheme 2.

The Diphenyl Ether saFabI Inhibitors