ABSTRACT
The type II fatty acid synthesis (FASII) pathway is essential for bacterial lipid biosynthesis and continues to be a promising target for novel antibacterial compounds. Recently, it has been demonstrated that Chlamydia is capable of FASII and this pathway is indispensable for Chlamydia growth. Previously, a high-content screen with Chlamydia trachomatis-infected cells was performed, and acylated sulfonamides were identified to be potent growth inhibitors of the bacteria. C. trachomatis strains resistant to acylated sulfonamides were isolated by serial passage of a wild-type strain in the presence of low compound concentrations. Results from whole-genome sequencing of 10 isolates from two independent drug-resistant populations revealed that mutations that accumulated in fabF were predominant. Studies of the interaction between the FabF protein and small molecules showed that acylated sulfonamides directly bind to recombinant FabF in vitro and treatment of C. trachomatis-infected HeLa cells with the compounds leads to a decrease in the synthesis of Chlamydia fatty acids. This work demonstrates the importance of FASII for Chlamydia development and may lead to the development of new antimicrobials.
INTRODUCTION
Chlamydia trachomatis is the most common sexually transmitted bacterial pathogen, with over 100 million new cases of infection occurring each year (1). Infection of the genital tract epithelium with C. trachomatis is associated with a variety of diseases, including pelvic inflammatory disease, salpingitis, ectopic pregnancy, and infertility (1). C. trachomatis is also an ocular pathogen and the leading cause of infectious blindness worldwide (2, 3).
Options for the treatment of infections caused by C. trachomatis are limited to broad-spectrum antibiotics like doxycycline and macrolides. Resistance to antibiotics readily develops in vitro, and cases of treatment failure have been described (4, 5). Targeted anti-Chlamydia treatments would be beneficial, as they would reduce the effect on the microbiota and lower the possibility of selection of pathobionts and other sexually transmitted bacteria that become resistant to important broad-spectrum antibiotics (6). Drugs with anti-Chlamydia activity can also be utilized as investigative tools to gain more information on Chlamydia biology (7, 8).
Chlamydia spp. undergo a developmental cycle that transitions between two distinct bacterial forms; the elementary body (EB) is spore-like and infectious but metabolically limited and upon internalization differentiates into the noninfectious but metabolically active reticulate body (RB). Chlamydia replication and development occur within a specialized membrane-bound vacuole called the Chlamydia inclusion, a compartment that is unique to Chlamydia among all intracellular pathogens. The inclusion is nonfusogenic with endocytic organelles and instead appears to intercept components of an exocytic vesicle-trafficking pathway to acquire host-derived glycerophospholipids, cholesterol, and sphingolipids (9, 10). Despite its obligate intracellular developmental cycle and a genome whose size is reduced compared to the sizes of the genomes of free-living bacteria, C. trachomatis possesses all the genes necessary for the de novo biosynthesis of fatty acids via the type II fatty acid synthesis (FASII) pathway, and this pathway has recently been demonstrated to be essential for the growth of Chlamydia (11). In bacteria, FASII is responsible for the synthesis of fatty acids required for both membrane phospholipids and lipopolysaccharides (12).
Previous work identified acylated sulfonamides to be potent inhibitors of C. trachomatis growth, with 50% maximal effective concentrations (EC50) being 3 to 12 μM. These compounds are bactericidal, and their mechanism of action is different from that of sulfonamide antibiotics (8). Here, we report that acylated sulfonamides target the Chlamydia FASII pathway and that the compounds bind directly to the 3-oxoacyl-(acyl carrier protein [ACP]) synthase II (FabF) enzyme in vitro. These data validate previous work showing the importance of FASII to C. trachomatis growth and further establish chlamydial fatty acid synthesis to be a suitable target for therapeutic intervention.
RESULTS
ME0640 is active throughout the infectious cycle of C. trachomatis.To pinpoint when the acylated sulfonamide ME0640 (Fig. 1) (8) is active during C. trachomatis infection, a time course study was performed by adding 6 μM ME0640 at different time points during infection. Following drug treatment, the number of inclusions and the average area of the inclusions were assessed at 48 h postinfection (hpi) and compared to those for untreated control infections. ME0640 treatment at 19 hpi led to a reduction in inclusion number of over 50%, and reductions of over 90% were obtained when treatment began at 3 hpi or earlier (Fig. 2A). Similarly, reductions in the average area of the inclusions of over 50% were observed when ME0640 treatment began at up to 26 h postinfection, demonstrating that ME0640 treatment affected all stages of Chlamydia development (Fig. 2B).
Structures of the compounds used in this study.
Acylated sulfonamides are active throughout the developmental cycle. HeLa cells were infected with WT C. trachomatis serovar L2 strain 434/Bu and subsequently treated with 6 μM ME0640 at 0, 1, 3, 19, and 26 hpi. The number of inclusions (A) and the average inclusion area (B) were assessed at each concentration and compared to those for the DMSO-treated control wells at 48 hpi. Bars indicate standard deviations. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ns, not significant.
Acylated sulfonamide-resistant mutants accumulate single nucleotide mutations in fabF.Two resistant populations of C. trachomatis resistant to acylated sulfonamides, populations M1 and M2, were independently generated by serial passage of wild-type (WT) serovar L2 strain 434/Bu in the presence of low concentrations of ME0640. Fourteen isolates, six from population M1 and eight from population M2, were obtained by plaque purification. Whole-genome sequencing of the isolates in the two independently isolated resistant populations and of 10 plaque-purified isolates was performed. All sequenced resistant strains harbored single nucleotide substitutions in fabF leading to an amino acid change of glycine to serine at position 200 (G200S) or glutamic acid to lysine at position 122 (E122K) in the lipid synthesis enzyme FabF [3-oxoacyl-(acyl carrier protein) synthase II], which functions as part of the Chlamydia FASII pathway (Table 1). Sanger sequencing of four additional isolates from the second population demonstrated that the G200S substitution was present in all isolates. Preliminary analyses of all 14 isolates and the isolates from both drug-resistant populations indicated that their growth rates were significantly higher than the growth rate of the WT strain in the presence of ME0640 (Table 1; see Fig. S1 in the supplemental material). One isolate from each population with the G200S substitution, isolate M1-2 and isolate M2-17, and one isolate with the E122K substitution, isolate M1-4, were selected for further analysis. The EC50s of ME0640 for the three selected isolates were significantly higher than the EC50 for the wild type, and the strains were also resistant to another acylated sulfonamide, ME0619 (Fig. 3; Table 2). M1-2 and M2-17 had mutations only in fabF, while M1-4 had a low-frequency (31.6%) mutation leading to an amino acid substitution in GroEL. To rule out the possibility that the GroEL substitution contributed to the resistance phenotype, M1-4 was further subcloned by plaque assay until the mutation was no longer detected by Sanger sequencing (subclone 3-1). This subclone had a level of susceptibility to ME0640 equal to that of the wild-type strain (Fig. S2). The isolation of independently generated resistant strains harboring substitutions in only the fabF gene suggested that FabF is a target of acylated sulfonamides.
All acylated sulfonamide-resistant isolates have fabf mutationsa
Mutant isolates are more resistant to acylated sulfonamides than the WT. HeLa cells were infected with the three plaque-purified resistant isolates, M1-4 (E122K), M1-2 (G200S), and M2-17 (G200S), and with WT C. trachomatis and were then treated with a serial dilution (25 to 0.195 μM) of ME0640 (top) or ME0619 (bottom) in 96-well plates. The number of inclusions (A) and the average inclusion area (B) were assessed at each concentration and compared to those for the DMSO-treated control wells at 48 hpi. Bars indicate standard deviations. ***, P ≤ 0.001.
EC50s for WT C. trachomatis and the three selected mutant isolates
fabF mutant isolates are differentially susceptible to cerulenin.FASII is ubiquitous and essential for growth in both Gram-positive and Gram-negative bacteria (13, 14). Consequently, a multitude of natural products targeting key regulatory steps of this fatty acid synthesis pathway have been identified from a wide array of microorganisms (12, 15). To assess the impact of the fabF mutations on susceptibility to known FASII-targeting drugs, we tested the growth of wild-type C. trachomatis L2/434/Bu and the M1-4 mutant isolate containing the FabF E122K amino acid substitution in the presence of the FabF-inhibiting drugs cerulenin (16, 17) and platensimycin (18), as well as the FabI-inhibiting compound triclosan (19). All three drugs have been reported to be active against many Gram-positive and Gram-negative bacteria (15, 20). Platensimycin and triclosan were equally active against both strains (Table 2; Fig. 4A and B). However, the M1-4 strain with the E122K substitution was more resistant to cerulenin than the wild-type strain, with EC50s of 4.7 μM and 19.6 μM for the wild-type and M1-4 strains, respectively (Table 2; Fig. 4C). We then sought to test the impact of the G200S amino acid substitution in FabF, which accumulated in both the M1 and M2 mutant populations at higher frequencies than other substitutions (Table 1). Thus, we selected the M2-17 isolate containing the mutation leading to the G200S substitution in FabF and compared its susceptibility to cerulenin to that of the wild-type strain. Interestingly, this isolate was slightly more susceptible to cerulenin than the wild type, with an EC50 of 2.4 μM (Table 2; Fig. 4D). A homology model of the Chlamydia FabF structure was created to visualize the locations of the mutations G200S and E122K in relation to the active site and to the binding site of cerulenin (Fig. S3 to S6). The G200S mutation was located in close vicinity to the catalytic site, and Bacillus subtilis has the same substitution, while the E122K mutation was situated on an α-helix 23 Å from the center of the catalytic site (measured from the centroid of cerulenin) (Fig. S4B). Despite the distance from the catalytic site, this substitution led to resistance to ME0640, ME0619, and cerulenin (Fig. 3 and 4C).
Isolates M1-4 and M2-17 are differentially susceptible to the FabF inhibitor cerulenin. (A to C) The M1-4 isolate (E122K) and WT C. trachomatis were treated with serial dilutions of triclosan (6 to 0.375 μM) (A), platensimycin (25 to 3.125 μM) (B), and cerulenin (50 to 0.1953 μM) (C) in 96-well plates. (D) The M2-17 isolate (G200S) and the WT were treated with the same serial dilutions of cerulenin described in the legend to panel C in a 96-well plate. The number of inclusions in 12 fields per well at each concentration was assessed and compared to the number for the DMSO-treated control wells at 48 hpi. Bars indicate standard deviations.
Acylated sulfonamides protect rFabF from proteolysis.To investigate possible interactions with acylated sulfonamides and FabF, two in vitro techniques were employed. Drug binding to a target protein often stabilizes a folded state or conformation, leading to increased resistance to proteolytic digestion. The drug affinity response target stability (DARTS) assay provides a simple label-free method to assess if the protease susceptibility of recombinant FabF (rFabF) is altered in the presence of acylated sulfonamides (21). rFabF with both His and ZZ tags was incubated with ME0640, ME0619, and ME0518 (Fig. 1), followed by digestion with pronase. Compound ME0518 has the central benzene ring in ME0640 replaced with a pyridine moiety that increases aqueous solubility. In all cases, acylated sulfonamide-dependent protection of rFabF was observed when the compounds were present, but no increased protection was observed for green fluorescent protein (GFP) with both the His and ZZ tags (Fig. 5). These data suggest that rFabF is a molecular target for acylated sulfonamides.
Acylated sulfonamides protect recombinant FabF from proteolysis in the drug affinity response target stability (DARTS) assay. (A) Representative gels with a typical result by the DARTS assay obtained using either rGFP (top) or rFabF (bottom) fused to His and ZZ tags. P, pronase; NP, no pronase. (B) DARTS assay results from three independent experiments at the indicated concentrations. Bands from each gel were quantified using ImageJ software, followed by calculation of the average percent protection compared to that for the pronase control for each sample treated with compound. Bars indicate standard errors. (C) A representative DARTS assay gel showing the results for 2 serial dilutions of ME0619 from 4 mM. *, P ≤ 0.05; ns, not significant.
Acylated sulfonamide interaction with rFabF is detectable by 1H NMR spectroscopy.Nuclear magnetic resonance (NMR) spectroscopy provides another label-free method to assess the interactions of proteins and small molecules (22, 23). We employed 1H NMR to detect changes in the relaxation time for the acylated sulfonamide that was associated with binding to the target protein (24). Due to the poor solubility of the compounds, we were unable to conduct NMR experiments with ME0640 or ME0518. Instead, we used the related compound ME0619, which demonstrated potency similar to that of ME0640 and against which the isolates containing amino acid substitutions in FabF displayed elevated EC50s compared to the EC50 for the WT (Fig. 3). Comparison of the spectra for 100 μM ME0619 in the absence and the presence of 10 μM rFabF shows that the signals from ME0619 disappear in the presence of rFabF in this relaxation-edited 1H NMR experiment, indicating binding (Fig. 6B). Sulfamethoxazole, which shares the same structural core with acylated sulfonamides, was also utilized as a control compound in a separate 1H NMR experiment with rFabF. In this case, sulfamethoxazole signals were not affected by the presence of rFabF in solution, indicating that the sulfamethoxazole structural core is not sufficient for binding to rFabF (Fig. 6A).
Detection and specificity of binding of acylated sulfonamides to FabF using 1H NMR spectroscopy. (A and B) Relaxation-edited spectra for 100 μM sulfamethoxazole (A) and 100 μM ME0619 (B) in the absence and presence of 10 μM rFabF. (C) 1H NMR spectra of 100 μM ME0619 in the absence of protein and in the presence of 10 μM rGFP and 10 μM rFabF. Line broadening in the presence of rFabF but not in the presence of rGFP shows specific binding to rFabF. (D) Relaxation-edited spectra for 100 μM ME0619 alone or with 1 μM rFabF (WT rFabF and rFabF G200S and E122K mutants).
To further confirm that ME0619 specifically binds to FabF, 1H NMR experiments that were not relaxation edited were performed using both recombinant GFP (rGFP) and rFabF containing the same His and ZZ tags. The results showed that the signals from ME0619 were unaffected by the presence of rGFP but exhibited significant line broadening in the presence of rFabF (Fig. 6C and S7). A separate set of relaxation-edited NMR experiments was also performed, where binding to rFabF variants containing the E122K and G200S substitutions alongside WT rFabF was investigated (Fig. 6D). In these experiments, the rFabF concentrations were reduced to 1 μM (compared to the 10 μM used in the experiment whose results are presented in Fig. 6B) to compare the binding to the three rFabF variants. To our surprise, all three FabF variants exhibited nearly equal binding to the compound, as seen by the similar signal intensities of ME0619 in the presence of the proteins. Taken together, these data suggest that ME0619 binds specifically to rFabF and that this binding is dependent on the acyl group of the compound. Further, these data also suggest that the E122K and G200S amino acid substitutions have no effect on ME0619 binding to FabF.
Acylated sulfonamides inhibit C. trachomatis fatty acid synthesis. C. trachomatis is known to synthesize its own branched-chain fatty acids commencing from a branched-chain acyl coenzyme A (acyl-CoA), followed by the addition of carbon units from malonyl-acyl carrier protein (ACP) ultimately derived from acetyl-CoA (11, 13). The amino acids valine, leucine, and isoleucine are precursors to isobutyryl-CoA, isovaleryl-CoA, and 2-methylbutyryl-CoA, which are generated by the bacterial enzyme α-keto acid dehydrogenase and incorporated into Chlamydia-specific branched-chain fatty acids by 3-oxoacyl-ACP synthase III (FabH). Thus, lipids synthesized by C. trachomatis can be detected by incubation of Chlamydia-infected cells with radiolabeled valine, leucine, or isoleucine followed by total lipid extraction and monitoring of the increase in radioactivity in this fraction compared to that for control uninfected cells (11, 25, 26). Of these three amino acids, isoleucine has previously been reported to show the most extensive incorporation into Chlamydia branched-chain fatty acids (25).
To determine the effect of acylated sulfonamides on FASII, infected HeLa cells were incubated with 35 μM ME0640, ME0619, or dimethyl sulfoxide (DMSO) only at 20 hpi, followed by the addition of [3H]isoleucine at 22 hpi. Cerulenin (50 μM) and ofloxacin (30 μM) were also included as control drugs for lipid synthesis and DNA synthesis, respectively. Total lipids were extracted 4 h later, and the levels of radioactivity were assessed by liquid scintillation. Uninfected cells similarly treated with DMSO and [3H]isoleucine were included as a baseline control for any host cell metabolism of the [3H]isoleucine, as isoleucine can be broken down into acetyl-CoA by the mammalian host cell, which can then be incorporated into both mammalian cell- and bacterium-derived fatty acids (11). Lipid fractions from infected HeLa cells incubated with DMSO only or with ofloxacin showed a significant increase in radioactivity compared to uninfected control cells, indicating incorporation of [3H]isoleucine into phospholipids, while cerulenin, ME0640, and ME0619 treatment showed no incorporation of [3H]isoleucine (Fig. 7A). As a control experiment, incorporation of [3H]guanine into DNA was tested after incubation with 25 μM ME0640 or 20 μM ofloxacin. As expected, ofloxacin treatment significantly inhibited DNA synthesis, as measured by determination of the level of [3H]guanine incorporation into DNA, but ME0640 treatment had no effect (Fig. 7B). Taken together, these data provide compelling evidence that ME0640 and ME0619 specifically block Chlamydia fatty acid synthesis.
Acylated sulfonamides block [3H]isoleucine incorporation into Chlamydia fatty acids but do not block [3H]guanine incorporation into DNA. (A) Percent inhibition of [3H]isoleucine incorporation into lipid extracts of infected HeLa cells grown in 24-well plates and treated at 20 hpi with 35 μM ME0640, 35 μM ME0619, 50 μM cerulenin, or 30 μM ofloxacin, followed by the addition of [3H]isoleucine at 22 hpi. Total lipids were extracted 5 h later, and the levels of radioactivity were assessed by liquid scintillation. (B) Percent inhibition of [3H]guanine incorporation into DNA extracts of infected HeLa cells in 24-well plates treated at 20 hpi with 25 μM ME640 or 20 μM ofloxacin, followed by the addition of [3H]guanine at 22 hpi. DNA was purified 5 h later, and the levels of radioactivity were assessed by liquid scintillation. Bars indicate standard deviations. **, P ≤ 0.01.
C. trachomatis FabF can functionally replace Escherichia coli FabF in vivo. Escherichia coli has two distinct 3-oxoacyl-synthase enzymes (I and II, also known as FabB and FabF, respectively) that participate in the fatty acid condensation steps of FASII. FabB appears to catalyze a key reaction in unsaturated fatty acid synthesis in E. coli, a process that does not seem to be required in Chlamydia due to the lack of any unsaturated fatty acid synthesis genes (27, 28). To test if the Chlamydia FabF enzyme is functional in vivo, we inserted the WT and the two mutant variants of the ChlamydiafabF gene into the arabinose-inducible vector pBAD322. As a positive-control plasmid, we also inserted the fabF gene from E. coli strain K-12. The resulting plasmids, along with the empty vector, were then transformed into temperature-sensitive (ts) strain E. coli fabB(ts) fabF CY244. This mutant strain lacks the activity of both the FabB and FabF enzymes at the nonpermissive temperature (42°C) and thus is unable to grow even when supplied with oleic acid as a source of unsaturated fatty acid (29). The growth of all four transformants containing plasmids with inserted fabF genes at 42°C indicated complementation of the fabF mutation (Fig. 8). Thus, the WT and the E122K and G200S variants of the Chlamydia FabF enzyme can catalyze all the elongation reactions required for saturated fatty acid synthesis.
The Chlamydia FabF enzyme complements an E. coli strain lacking FabF. E. coli fabB(ts) fabF strain CY244 transformants carrying pBAD322 plasmids containing C. trachomatis fabF (fabFCt; the WT or the E122K or G200S variants) or E. coli fabF (fabFEc) were grown at the indicated permissive (30°C) and nonpermissive (42°C) temperatures on rich broth (RB) plates supplemented with 0.1% potassium oleate, 100 μM carbenicillin, and 0.2% arabinose.
DISCUSSION
Our experiments provide compelling evidence that acylated sulfonamides block Chlamydia replication via the inhibition of FASII. First, two different acylated sulfonamide-resistant mutant populations harbored mutations in the fabF gene, leading to amino acid substitutions in the FabF protein, indicating the importance of this protein in resistance to acylated sulfonamides. Second, an interaction between acylated sulfonamides and recombinant FabF protein was independently detected by two in vitro techniques, NMR spectroscopy and the DARTS assay, strongly suggesting that this interaction occurs with native FabF. Third, Chlamydia fatty acid synthesis, as measured by determination of the level of incorporation of radiolabeled isoleucine, was blocked by acylated sulfonamides, directly demonstrating that these compounds inhibit the synthesis of Chlamydia-specific branched-chain fatty acids.
As an obligate intracellular pathogen, Chlamydia has a genome of a reduced size compared to the sizes of the genomes of free-living Gram-negative bacteria (30, 31). However, bioinformatic analyses of whole-genome sequences have revealed that Chlamydia spp. nonetheless possess the complete set of genes typically present in free-living Gram-negative bacteria required for saturated fatty acid synthesis via FASII (11, 31). The genes for unsaturated fatty acid synthesis are absent from the Chlamydia genome (11, 32). Although host lipid metabolism and trafficking are important for the development and maintenance of the specialized chlamydial inclusion membrane (10, 33, 34), biochemical studies have also confirmed that Chlamydia produces specific molecular species of phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and cardiolipin (CL) by utilizing fatty acids produced by its own FASII pathway and that this process is essential for chlamydial replication (11, 32, 35).
Membrane biosynthesis is a critical component of bacterial physiology and is frequently a target of interest for the development of novel antibacterial agents (13–15). Natural products from diverse collections of microorganisms have been identified to be inhibitors of FASII, reflecting how microorganisms themselves have learned to target FASII as an effective antimicrobial strategy through evolution (12). Here, we show that the FabF-inhibiting natural compounds cerulenin and platensimycin block the growth of Chlamydia, confirming that FabF performs an essential function. This is consistent with FabF being the only elongation-condensing enzyme present in the C. trachomatis genome. Indeed, FabF is the most common essential condensing enzyme among bacteria, including Gram-positive pathogens (36). Accordingly, Chlamydia FabF complements FabF-deficient E. coli, indicating that the Chlamydia FabF enzyme can catalyze all the elongation reactions required for saturated fatty acid synthesis.
Cerulenin has a 12-carbon acyl chain that forms a covalent bond with the catalytic cysteine located in the active site of the E. coli enzyme (37–39). Thus, cerulenin inhibits FabF activity by blocking access to the hydrophobic pocket of the enzyme that normally binds its natural substrate. Our experimental data show the differential susceptibility of fabF mutants to cerulenin, in contrast to their consistent resistance to acylated sulfonamides, suggesting that the binding and modes of action between these two drugs may be distinct. According to molecular modeling (see Fig. S3 to S6 in the supplemental material), the glycine and the corresponding mutation to serine at position 200 are conserved with the serine at position 119 in the FabF from B. subtilis. On the basis of the B. subtilis FabF structure, the susceptibility of Chlamydia with FabF with the G220 mutation (FabFG200S) to cerulenin inhibition can be explained by the observation that this amino acid does not seem to hinder cerulenin binding, even though it is situated in the cerulenin binding site only 3.5 Å from the inhibitor. The E122K mutation, in contrast, is located far from the active site yet increases resistance to both cerulenin and acylated sulfonamides, suggesting that it may cause compensatory conformational changes in the protein structure. According to this hypothesis, acylated sulfonamides may bind to another site distinct from where cerulenin binds, thereby leading to a conformational change that interferes with enzyme catalysis and/or substrate binding to FabF, while the mutant variants may retain their functional conformation upon binding to acylated sulfonamides. Alternatively, the possibility that these mutations are compensating for a deficiency elsewhere in the pathway cannot be discounted. These hypotheses await confirmation from structural studies.
Previous data indicated that E. coli and Staphylococcus aureus growth was not effectively inhibited by acylated sulfonamides (8), in contrast to the growth inhibition caused by triclosan, platensimycin, and cerulenin, all of which target a broader range of Gram-positive and Gram-negative bacteria (8, 15, 20). Thus, acylated sulfonamides may be selective inhibitors of Chlamydia fatty acid synthesis, presenting a possible starting point for the development of a narrow-spectrum antibiotic. A selective inhibitor of Chlamydia FASII may also be advantageous to future studies attempting to ascertain the precise role of this pathway in determining the lipid composition of the chlamydial cell envelope and inclusion membranes during the developmental cycle. Characterization of the precise interaction of acylated sulfonamides with the chlamydial FabF enzyme and elucidation of whether this compound class inhibits FabF catalytic activity are key focuses for future work.
MATERIALS AND METHODS
Compounds.Platensimycin, triclosan, and cerulenin were obtained from Sigma-Aldrich (St. Louis, MO, USA), and DMSO stock solutions (50 mM) were prepared and stored at −20°C until use. The sulfonamide compounds ME0640 and ME0619 were prepared according to published procedures (8). ME0640 and ME0619 are identical to compounds 18 and 23, respectively, described by Marwaha et al. (8). The synthetic procedures and analytical data for ME0518 are provided in Method S7 in the supplemental material. The structures of all the acylated sulfonamide compounds used in this study are shown in Fig. 1. All compounds were >95% pure according to reversed-phase high-performance liquid chromatography UV tracing at 214 nm. Stock solutions (10 mM) were prepared and stored at −20°C.
Cell culture, Chlamydia infection, and plaque purification.HeLa 229 cells (CCL-2.1; ATCC, Manassas, VA, USA) were propagated in T75 flasks at 37°C with 5% CO2 in RPMI 1640 medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) and 2 mM l-glutamate. C. trachomatis serovar L2 strain 434/Bu (VR-902B; ATCC, Manassas, VA, USA) was propagated in HeLa cells, and elementary bodies were purified as previously described (40) and stored in SPG buffer (0.25 M sucrose, 10 mM sodium phosphate, 5 mM l-glutamic acid) at −80°C. To infect cells with Chlamydia, the inoculum was diluted in Hanks balanced salt solution (HBSS; Life Technologies, Carlsbad, CA, USA) and incubated with the cells at 37°C with 5% CO2 at the desired multiplicity of infection (MOI) for 1 h. HBSS was subsequently aspirated and replaced with RPMI 1640 medium that was supplemented as described above and that contained dilutions of the desired compounds. Clonal populations were collected by plaque assay as previously described (41). Briefly, Vero cells (CCL-81; ATCC, Manassas, VA, USA) were infected with 100 to 10 inclusion-forming units (IFUs) for 2 h, and infected cells were overlaid with a 0.54% agarose-Dulbecco modified Eagle medium overlay (Life Technologies, Carlsbad, CA, USA). Plaques were collected 10 to 20 days after infection and propagated in HeLa cells.
Selection of acylated sulfonamide-resistant mutants.HeLa cells were grown in large flasks and infected at an MOI of 5. Each passage was treated with a subinhibitory concentration of ME0640 (2 μM for passages 1 to 7, 2.5 μM for passage 8, 3 μM for passage 9, 3.5 μM for passage 10, and 2.5 μM for passage 11). From each mutant passage, infectious EB progeny were collected at 44 to 48 h postinfection (hpi) and used for reinfection. During the first 4 to 6 passages, the treated bacteria were grown at an MOI of 2 or higher; thereafter, the selection proceeded with an MOI of 2 or lower (preferentially an MOI of less than 1) (42).
EC50 determination and inclusion area analysis.HeLa cells were inoculated into 96-well plates (Corning, Corning, NY, USA) at a density of 15,000 cells per well and cultured overnight at 37°C with 5% CO2 in RPMI 1640 medium supplemented with 10% FBS and 2 mM l-glutamine. On the following day, the cells were infected with C. trachomatis serovar L2 strain 434/Bu at an MOI of approximately 0.3 in HBSS. At 1 hpi, the inoculum was replaced with RPMI 1640 medium containing 0.5% DMSO and serial 1:1 dilutions of the tested compounds. The infection was allowed to proceed for 44 to 48 h before fixation by aspiration of the medium and addition of methanol for 5 min. Chlamydial inclusions were stained with an in-house-generated primary rabbit anti-Chlamydia antibody (8) and a secondary donkey anti-rabbit fluorescein isothiocyanate-labeled antibody (Jackson ImmunoResearch, West Grove, PA, USA). The DNA of the cells and Chlamydia was stained with DAPI (4′,6-diamidino-2-phenylindole). The numbers of inclusions exceeding 50 μm2 (for the 19-h time point) or 130 μm2 (for the 48-h time point) in size were counted by use of an ArrayScan VTI HCS automated scanner (Thermo Scientific, Waltham, PA, USA). Inhibition of Chlamydia growth was evaluated as the number of inclusions in compound-treated infections (in 12 fields per well evaluated at a ×20 magnification) compared to the numbers of inclusions in DMSO-treated control infections (given as a percentage of that for the control). EC50s are representative of those from at least three independent experiments. For inclusion area analysis, one-way analysis of variance with the Bonferroni-Dunn posttest was used to compare the WT and mutant groups. Data analysis was performed using nonlinear regression (curve fit) in GraphPad Prism (v5) software.
Sanger and whole-genome sequencing.Sanger sequencing was performed by amplifying the desired sequence by PCR, purifying the products using a GeneJet purification kit (Thermo Scientific, Waltham, PA, USA), and then sending the products to GATC Biotech for sequencing (Cologne, Germany). To amplify the fabf gene, forward primer 5′-ATGAACAAAAAACGTGTAGTCGTTACAG-3′ and reverse primer 5′-TCATAATGAAGGTTCATACCTCG-3′ were used. To check the amplicon sequence, the forward primer was used. For whole-genome sequencing, 1 μg of Chlamydia-enriched DNA isolated from the following strains was used for sequencing library preparation: the wild-type C. trachomatis L2/434/Bu parental strain (GenBank accession number NC_010287 [47]), the M1 population, the M2 population, 6 plaque purified M1 single isolates (isolates M1-2, M1-3, M1-4, M1-5, M1-6, and M1-9), and 4 plaque purified M2 single isolates (isolates M2-3, M2-5, M2-6, and M2-17). DNA was sheared by ultrasound with an Adaptive Focused Acoustics S220 instrument (Covaris, Inc., Woburn, MA). Sequencing libraries were prepared from sheared DNA and barcoded by use of a TrueSeq DNA sample preparation kit (Illumina Inc., San Diego, CA) according to the manufacturer's instructions. Barcoded sequencing libraries prepared from the M1 population and M1 population-derived plaque-purified individual strains (6 genomes) were pooled and sequenced in one lane of a HiSeq2000 sequencing platform (Illumina Inc.). Barcoded sequencing libraries prepared from the M2 population, M2 population-derived plaque-purified individual strains (5 genomes), and the wild-type parental strain were pooled and sequenced in one lane of a HiSeq2000 sequencing platform (Illumina Inc.). Sequencing was performed at the Duke University Sequencing and Genomic Technologies Core Facility. Sequencing data generated by the Illumina platform were separated by barcode sequence. The 50-mer single end reads were aligned to the C. trachomatis serovar L2 strain 434/Bu reference sequence, and single nucleotide variants (SNV) were identified with Geneious software (Biomatters Limited). An SNV was defined as a nucleotide variant that was present at a frequency of greater than 15% and that had a strand bias of less than 70%.
Recombinant protein expression and purification. C. trachomatis fabF (NCBI CTL0139) was amplified from the WT and mutant Chlamydia strains using forward primer 5′-GTACTCATGAACAAAAAACGTGTAGTC-3′ and reverse primer 5′-GTACGGTACCTCATAATGAAGGTTCATACCTC-3′, cloned into pETZZ1a (43) using restriction enzymes BspHI/NcoI and Acc65I, and then transformed into E. coli Rosetta (Millipore, Billerica, MA, USA) for overexpression. The ZZ tag was needed to improve the solubility of the FabF proteins. The proteins were purified on Ni-nitrilotriacetic acid agarose columns (Qiagen, Venlo, Netherlands) followed by gel filtration using HiLoad GF200 16/60 columns (GE Healthcare, Chicago, IL, USA).
Drug affinity response target stability.The DARTS assay was performed as previously described with modifications (21). Briefly, recombinant proteins (1 mg/ml) were incubated with 0.5 mM ME0640, 1.0 mM ME0518, 4.0 mM ME0619, or 10% the DMSO solvent as a control for 2 h at room temperature in phosphate-buffered saline (PBS) containing 5 mM 2-mercaptoethanol and 15% polyethylene glycol (PEG) 400, followed by digestion with pronase (1:100; Roche, Basel, Switzerland) for 1 h. Digestion was halted by addition of SDS-PAGE buffer and boiling. The samples were loaded and separated on 4 to 12% bis-Tris gels, and the proteins were visualized by Coomassie blue staining. Experiments with each compound were repeated thrice, and bands were quantified with ImageJ (v1.50) software as previously described (44). The data from all three experiments was pooled, and t tests were used to compare the protection of the two proteins within the different drug-treated groups.
Binding measurement using 1H NMR spectroscopy.Reference samples with 100 μM ME0619 and sulfamethoxazole in PBS (pH 7.4) were prepared from 10 mM and 100 mM DMSO stock solutions, respectively. D2O (10%) was added before the solutions were transferred to 5-mm NMR tubes. Samples containing 100 μM ME0619 and either 5 μM rFabF or 5 μM rGFP were prepared separately. The relaxation-edited experiments contained a 200-ms Carr-Purcell-Meiboom-Gill spin lock and excitation sculpting for water suppression. Data from 256 scans were accumulated with a relaxation delay of 1 s (d1). The relaxation-edited spectra shown in Fig. 6A and B were recorded on a Bruker 850 MHz Avance III HD spectrometer equipped with a cryoprobe. The spectra shown in Fig. 6C were recorded on a Bruker 600 MHz Avance III HD spectrometer equipped with a cryoprobe. Data from all experiments were acquired at 298 K.
Metabolic radiolabeling experiments.HeLa cells were inoculated into 24-well plates (Corning, Corning, NY, USA) and cultured to confluence at 37°C with 5% CO2 in RPMI 1640 medium supplemented with 10% FBS and 2 mM l-glutamine. The cells were infected with C. trachomatis serovar L2 strain 434/Bu at an MOI of approximately 3.0 in HBSS or mock infected. At 1 hpi, the inoculum was replaced with RPMI 1640 medium containing 1 μg/ml cycloheximide (500 μl per well). At 20 hpi, the medium in each well was aspirated and replaced with medium (250 μl per well) containing the test compounds at the concentrations indicated above in 0.5% DMSO. An equal volume of DMSO was added to the control wells containing no compounds. At 22 hpi, radiolabeled isoleucine or guanine (American Radiolabeled Chemicals, St. Louis, MO, USA) was added, without dilution, directly to the culture medium to achieve a final concentration of 3 μC/ml or 1 μC/ml, respectively. Labeling was allowed to proceed for 5 h at 37°C with 5% CO2. For the experiments with [3H]guanine, the DNA from each well was extracted using a blood and tissue DNA extraction kit (Qiagen, Venlo, Netherlands) according to the manufacturer's instructions. For the experiments with [3H]isoleucine, lipids were extracted from the cells using the protocol of Bligh and Dyer (45). Briefly, cells were lysed in 0.2 ml of water for 10 min. The lysates from each well was collected in 2-ml Eppendorf tubes, followed by the addition of 750 μl of a 1:1 (vol/vol) mixture of methanol and chloroform to each tube. Chloroform (250 μl) and water (250 μl) were then added to each tube, with vortexing being performed after each addition. The solutions were then allowed to separate into two phases, and the lower phases containing lipid were collected. For all radiolabeling experiments, the amount of radioactivity incorporated was measured by mixing the extracts with Optiphase HiSafe 3 liquid scintillation cocktail (at a 1:4 extract volume/cocktail volume ratio) within clear flexible polyethylene terephthalate microplates (PerkinElmer, Waltham, MA, USA). The plates were sealed with adhesive covers, and the amount of radioactivity per well was measured using a 1450 Microbeta TriLux scintillation counter (GMI, Ramsey, MN, USA). Inhibition of radioisotope incorporation was evaluated as the number of counts per minute in compound-treated infected cells compared to the number of counts per minute in DMSO-treated control infected cells (percent inhibition), with cells mock infected with the radioisotope and treated with DMSO being used to determined the 0% inhibition baseline. Measurements were made in triplicate, and averages with standard deviations are reported. One-way analysis of variance with the Bonferroni-Dunn posttest was used to compare the different drug-treated groups.
E. coli CY244 transformation and propagation.The CY244 strain of E. coli and the pBAD322 vector were generous gifts from John Cronan (29, 46). C. trachomatis isolates with WT and mutant (G200S and E122K) fabF (NCBI CTL0139) were amplified using forward primer 5′-GTACGGTCTCCCATGCACCATCATCACCACCATATGAACAAAAAACGTGTAGTC-3′ and reverse primer 5′-GTACTCTAGATCATAATGAAGGTTCATACCTC-3′. WT fabF from Escherichia coli (NCBI JW1081) was amplified using forward primer 5′-GTACGGTCTCCCATGCACCATCATCACCACCATATGTCTAAGCGTCGTGTAG-3′ and reverse primer 5′-GTACTCTAGATTAGATCTTTTTAAAGATCAAAGAAC-3′. The amplicons were cloned into pBAD322 using restriction enzymes NcoI and XbaI and then transformed into CY244 via electroporation. The growth phenotypes of CY244 containing the different plasmids were evaluated on rich broth (RB) medium (29) supplemented with 0.1% potassium oleate solubilized by the addition of 0.15% Brij 58 detergent, 100 μg/ml carbenicillin, as well as 0.2% arabinose for FabF induction.
Accession number(s).All raw sequencing reads used for analysis of each isolate genome have been deposited at the NCBI Sequence Read Archive under the accession number SRP115651 .
ACKNOWLEDGMENTS
We gratefully acknowledge Mikael Lindberg of the Protein Expertise Platform at Umeå University for cloning and protein production and Anna Kauppi for technical assistance.
This work was supported by the Swedish Government Fund for Clinical Research, the Scandinavian Society for Antimicrobial Chemotherapy Foundation, Molecular Infection Medicine Sweden, and Umeå University (for Å.G.), the Swedish Research Council (for M.E.), and the National Institutes of Health (grant R01AI100759 for R.H.V.).
FOOTNOTES
- Received 6 April 2017.
- Returned for modification 13 May 2017.
- Accepted 28 July 2017.
- Accepted manuscript posted online 7 August 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.00716-17 .
- Copyright © 2017 American Society for Microbiology.
