|Home | About | Journals | Submit | Contact Us | Français|
Francisella tularensis is a highly virulent and contagious gram-negative intracellular bacterium that causes the disease tularemia in mammals. The high infectivity and the ability of the bacterium to survive for weeks in a cool, moist environment have raised the possibility that this organism could be exploited deliberately as a potential biological weapon. Fatty acid biosynthesis (FAS-II) is essential for bacterial viability and has been validated as a target for the discovery of novel antibacterials. The FAS-II enoyl reductase ftuFabI has been cloned and expressed, and a series of diphenyl ethers have been identified that are subnanomolar inhibitors of the enzyme with MIC90 values as low as 0.00018 μg/ml. The existence of a linear correlation between the Ki and MIC values strongly suggests that the antibacterial activity of the diphenyl ethers results from direct inhibition of ftuFabI within the cell. The compounds are slow onset inhibitors of ftuFabI, and the residence time of the inhibitors on the enzyme correlates with their in vivo activity in a mouse model of tularemia infection. Significantly, the rate of breakdown of the enzyme-inhibitor complex is a better predictor of in vivo activity than the overall thermodynamic stability of the complex, a concept that has important implications for the discovery of novel chemotherapeutics that normally rely on equilibrium measurements of potency.
Francisella tularensis is a highly virulent and contagious Gram-negative intracellular bacterium that causes the disease tularemia in mammals (1). The ability of F. tularensis to be aerosolized, coupled with the small number of bacteria required to cause disease and the ability of the bacterium to survive for weeks in a cool, moist environment, have raised the possibility that this organism could be used deliberately as an infectious agent (2). Consequently, NIAID has classified F. tularensis as a Category A priority pathogen. Streptomycin and gentamicin are currently used as chemotherapeutics to treat tularemia, however neither of them can be orally administrated. In addition, despite the availability of drugs such as the aminoglycosides, macrolides, chloramphenicol and fluoroquinolones, infection can result in a mortality as high as 40%. Taken together, there is a pressing need to develop chemotherapeutics with novel mechanisms of action for the treatment of tularemia.
The fatty acid synthesis pathway in F. tularensis is a type II (FAS II) dissociated synthase where individual reactions are carried out by separate proteins. Importantly, eukaryotes utilize the type I fatty acid biosynthesis multienzyme complex (FAS I) which is fundamentally different from the FAS II pathway in which each activity is encoded by a separate polypeptide (3). The NADH-dependent enoyl reductase (FabI) which catalyzes the last reaction in the elongation cycle is known to be an essential component in the FAS-II system (4). Genetic knockout and knockdown experiments together with studies utilizing small molecule FabI inhibitors have demonstrated that FabI is essential for bacterial cell growth, thus making it an attractive target for drug discovery (5–8). Several classes of chemicals have been identified that are picomolar inhibitors of FabI (9–12), including the diphenyl ether triclosan, a broad spectrum chemotherapeutic with activity against a variety of important pathogens including E. coli, methicillin-resistant S. aureus and M. tuberculosis (13–18).
In this study, we expressed and purified the FabI from F. tularensis (ftuFabI), and identified a series of diphenyl ether-based ftuFabI enzyme inhibitors. The most potent alkyl diphenyl ether is a slow onset inhibitor with a Ki value of 0.44 nM and MIC90 value of 0.00018 μg/ml. The existence of a linear correlation between Ki and MIC90 values, supports the conclusion that the compounds target ftuFabI within the cell. A selection of the ftuFabI inhibitors are active in a mouse model of F. tularensis infection, however the increase in mean time to death and %survival caused by these compounds correlates best with the residence time of the inhibitor on the enzyme (19, 20), rather than the overall thermodynamic stability of the enzyme-inhibitor complex (Ki). This observation has important implications for rational drug design which is often driven solely by equilibrium measurements of inhibitor action, such as the determination of Ki or IC50 values, rather than by considerations of parameters such as the residence time of the drug on the target.
The equilibrium dissociation constant of triclosan (1) (Figure 1) from ftuFabI was determined by preincubating ftuFabI and triclosan in the presence of a high concentration of NADH and a low concentration of NAD+ (compared to their Kd values) (21). Apparent inhibition constants (Ki′) were measured at six different NAD+ concentrations (10, 15, 20, 50, 100 and 200 μM) in the presence of 250 μM NADH and the data were fit to equations 2–4 with Km,NAD constrained to 21 mM which was calculated from equation 5 using Km,NADH = 18.8 μM. Equation 2 gave the best fit to the data, demonstrating that triclosan is an uncompetitive inhibitor with respect to NAD+ with a K1 value of 51±3 pM at saturating NAD+. To estimate the affinity of triclosan for the E:NADH complex, the data were reanalyzed using equation 4 to provide a K1 value of 53±1 pM and a poorly determined value for K2 of 5±35 μM, indicating that triclosan binds approximately 90,000-fold more tightly to E:NAD+ than to E:NADH (Scheme 1A).
Evidence from X-ray structural (17, 22–24) and kinetic studies (21, 25) indicate that triclosan forms a very stable ternary complex with the FabI enzyme from E. coli (ecFabI), which is stabilized by a π-π stacking interaction between the hydroxychlorophenyl ring (ring A, Figure 1) and the nicotinamide ring of the NAD+ cofactor, together with an edge-on π interaction between the 5-chloro group and the aromatic ring of F203 (Figure 2A). Structural studies have previously shown that the major conformational change that occurs upon binding of triclosan to ecFabI-NAD+ is the ordering of the “substrate-binding” loop (Figure 2B). This loop (residues 195–200) is disordered in the binary complex (Figure 2B) but adopts an α-helical conformation to cover the active site of the enzyme when triclosan is bound (17, 22, 26, 27). We have previously postulated that the slow-onset inhibition of ecFabI by triclosan is coupled to the ordering of this loop (26). The high similarity between the structure of the ecFabI:NAD+ complex and the 2.9 Å resolution structure of the binary ftuFabI:NAD+ complex reported here (Figure 2C), provide a firm foundation for extending the conclusions on the ecFabI system to ftuFabI. Similar to ecFabI, the substrate binding loop in the ftuFabI:NAD+ complex is also disordered, and we postulate that slow onset inhibition of ftuFabI is also coupled to loop ordering.
To probe the role of specific interactions in the ternary ftuFabI complex, we synthesized a series of diphenyl ether compounds (Figure 1). In ecFabI it was found that ring B points away from the active site and participates in several unfavorable steric interactions with the enzyme based on the observation that 5-chloro-2-phenoxyphenol (4) has a 7-fold higher affinity than triclosan (25). However, in our studies compound 4 showed a 37-fold decrease in affinity indicating that the two chlorine atoms on ring B form important interactions with ftuFabI. In addition, for the other slow-onset inhibitors of ftuFabI that lacked chlorine atoms on ring B (4,5,7–12), binding to both E-NAD+ and E-NADH could be observed (Figure S1). The dependence of Ki′ on [NAD+] was fit to equation 4 to obtain the inhibition constants for both forms of the enzyme (Table 1). These data indicate that the primary effect of removing the B ring chlorines is a specific effect on the affinity of the inhibitor for the E-NAD+ form of the enzyme. Thus, all the slow-onset inhibitors have similar affinities for E-NADH and structural changes to both A and B rings principally modulate the affinity of the analogues for the E-NAD+ product complex. While the precise structural basis for this effect will require additional structural data, one possibility is that the affinity of each inhibitor for the E-NAD+ complex is critically dependent on the precise orientation of the A ring with respect to the oxidized nicotinamide ring and thus the strength of the π-stacking interaction between inhibitor and cofactor.
It has been demonstrated that the electronic and steric properties of ring A play a key role in determining the affinity of the diphenyl ether inhibitors for ecFabI. The importance of the 5-chloro group on ring A for binding to ftuFabI was initially explored by removing the chlorine to generate analogue 2 which binds 1200-fold less tightly to ftuFabI than does analogue 4. In addition, analogue 2 was no longer a slow-onset inhibitor indicating that a substituent on the A ring is essential for triggering loop ordering. To investigate this effect in more detail, we incorporated other halogen atoms into ring A and also replaced the 5-chloro group with alkyl substituents of different lengths (3,5,7–12). All the alkyl substituted diphenyl ethers (7–12) were shown to be slow, tight-binding inhibitors of ftuFabI. Since the 5-chloro substituent points to a hydrophobic pocket for the fatty acid substrate, it is quite possible that replacement of this substituent with a larger and more hydrophobic group will increase the affinity of the inhibitor for the enzyme. However, the data in Table 1 indicate that analogues with smaller alkyl chains generally bound more tightly to the enzyme, with analogue 9 (5-propyl-2-phenoxyphenol) exhibiting the highest affinity for the enzyme.
A remarkable observation is that replacement of the 5-chloro group with a fluorine results in a compound (5-fluoro-2-phenoxyphenol, 3) that is not a slow-onset inhibitor and that exhibits a 150-fold decrease in binding affinity compared to 5-chloro-2-phenoxyphenol (4) and 5-methyl-2-phenoxyphenol (7). The pKa values of 3 and 4 are almost identical, in agreement with electrostatic potential calculations (25), and thus the difference in affinity of 3 and 4 for ftuFabI may arise from the slightly smaller van der Waals radius of fluorine compared to chlorine (0.45 Å), which translates to a ~9 Å3 reduction in molecular volume of 3 compared to 4. If the loss of slow-onset inhibition is primarily due to the smaller size of the fluorine atom, then replacement of this group with a more bulky substituent might be expected to restore loop ordering and indeed the bromo (5) and methyl (7) analogs, which have molecular volumes ~5 Å3 larger than 4, as well as the propyl (9) and pentyl (10) analogs, are all slow-onset inhibitors. These data are in agreement with previous studies on the E. coli FabI (25, 28), where it was concluded that the shape of the inhibitor was the principal determinant in modulating the affinity of diphenyl ether inhibitors for the enzyme. Interestingly, however, incorporation of the larger but more electronegative nitro group (6) actually reduced binding affinity 3-fold compared to 3, and also resulted in a compound that was no longer a slow-onset inhibitor. Since 6 has a molecular volume that is only slightly larger (~3 Å3) than 5 or 7, the reduction in binding affinity of the nitro analog compared to diphenyl ethers bearing, for example, chloro, bromo or methyl substituents, likely results from a decrease in electron density of the A ring and a weakening of the π-π stacking interaction between this ring and the electron deficient positively charged nicotinamide ring of NAD+.
In summary, the SAR studies reflect the subtle interplay of steric and electronic effects that modulate inhibition in the FabI class of enoyl reductases (25, 28). For ftuFabI, a 5-chloro or 5-bromo substituent is essential for slow, tight-binding inhibition, while incorporation of hydrophobic alkyl chains of up to three carbons increases the affinity of the inhibitor for the enzyme, presumably due to favorable hydrophobic interactions in the substrate binding pocket. In addition, the π-π stacking interaction between ring A and the NAD+ nicotinamide ring is also important for slow onset enzyme inhibition, as shown by the dramatic impact of introducing a 5-nitro substituent into the inhibitor.
Minimum inhibitory concentrations (MIC90) of triclosan and the diphenyl ether inhibitors were obtained against F. tularensis LVS. Notably, the diphenyl ethers have more favorable MIC90 values than clinically used drugs (Table 1). Specifically, the best diphenyl ether analog (9) is more than 22,000-fold and 4,000-fold more potent than gentamicin and doxycycline, respectively. Significantly, a positive linear correlation between log Ki and log MIC90 exists for the diphenyl ether derivatives (Figure 3) which strongly suggests that ftuFabI is the primary cellular target for these compounds. Data fitting reveals that the compounds fall into two groups: those with halogen or nitro substituents (1–6) and those that have alkyl substituents on the A ring (7–12). In contrast, for those compounds for which the residence time of the inhibitor on the enzyme was measured (Table 2), no correlation could be observed between the length of time the inhibitor remained bound to the enzyme and the MIC values.
A rapid animal model of infection was used to evaluate the in vivo efficacy of ftuFabI inhibitors against F. tularensis. Initially, the in vitro cytotoxicity of the selected compounds was determined using a Vero cell assay which suggested that the compounds chosen for analysis had high therapeutic indices with LC50/MIC90 ratios ranging from 600 to 106 (Table 2). In addition, all compounds had maximum tolerated doses of >300 mg/kg. Animals in the infection model were treated for 5 consecutive days, starting from the day of infection, and then monitored for death for an additional 5 days. In general, our experiments showed that all treated animals survived longer than untreated control animals indicating that diphenyl ethers have antibacterial activity and are able to reduce the bacterial load. In the 10-day animal model of infection, all selected compounds (1,4,8,9 and 11) demonstrated efficacy, with longer median survival values in comparison to untreated control animals whose mean time to death value was 5.2 days (Figure 4 and Table 2). While all the other compounds reduced bacterial load and delayed animal death, compound 11 demonstrated greatest efficacy. Importantly, 11 prevented the death of all infected animals to the end of the 10 day study period (100% survival). Consequently, the mean time to death for compound 11 is undefined.
In order to provide further insight into the in vivo activity of the compounds, we used progress curve analysis to evaluate the rate constants for formation and breakdown of the final E-I* complex (29, 30). Based on the preincubation experiments, we already knew that the potency of the slow onset inhibitors increased as the concentration of NAD+ increased. Hence NAD+ (200 μM) was added at the beginning of the reaction so that product formation would not affect inhibitor affinity during the reaction. The progress curves showed that the turnover velocity decreased exponentially with time, from an initial velocity vi to a final steady-state velocity vs (Figure 5A). Higher concentrations of inhibitor caused the steady state to be reached more quickly and to give lower values of vs. This behavior demonstrates that the diphenyl ethers are slow onset reversible inhibitors that interact rapidly with the enzyme to form an initial complex, E-NAD+-I and then slowly isomerize to a more stable final inhibited complex, E-NAD+-I* (Scheme 1C) (29), which we speculate is the loop ordering step. Consistent with the mechanism shown in Scheme 1C, the kobs values obtained from fitting the data to equation 7 displayed a hyperbolic dependence on the concentration of inhibitor (Figure 5B), while subsequent nonlinear curve fitting using equation 8 gave pre-steady state constants (k2, k−2, and K1app) for each inhibitor (Table 2). Based on the assumption that the rate limiting step for breakdown of the E-I* complex is isomerization of E-I* to E-I (k−2), k−2 can be used to calculate the residence time of the inhibitor on the enzyme. This analysis provided residence times of 40, 30, 21, 59 and 143 min for 1, 4, 8, 9 and 11, respectively (Table 2). Interestingly, although triclosan has the highest thermodynamic affinity for ftuFabI and is the most potent compound in the in vitro antibacterial assays, compound 11 has the longest residence time on the enzyme. This is significant since 11 is more active in the animal model of infection than triclosan. Indeed, a positive linear correlation can be observed between the residence time of each compound on the enzyme and the in vivo antibacterial activity as gauged by the number of animals that survived to day 10 (Figure 6). This strongly suggests that the enhanced in vivo antibacterial activity of 11 is a direct consequence of the longer residence time of this compound on the enzyme drug target.
Recently Swinney provided a perspective on the importance of slow, tight-binding inhibitors as drug candidates and pointed out that the efficiency and potency of classical competitive inhibitors, which establish rapid equilibrium binding with the target, would be greatly diminished with the accumulation of substrates in the open system (31). In order to avoid this and maintain in vivo efficiency, non-equilibrium based inhibitors such as those that operate through a two-step slow-onset mechanism (Scheme 1C) should be emphasized in drug discovery programs. A slow-onset inhibitor with a long drug-target residence time can effectively create a transition away from rapid equilibrium inhibition and prevent the equilibrium between inhibitor and substrate from being achieved. Recent studies support the notion that longer residence time can improve target selectivity and reduce the side effects of small molecule inhibitors (19). In this context, Schramm and colleagues have noted that DADMe-ImmH, a slow onset inhibitor of purine nucleoside phosphorylase (PNP), is the ultimate physiological inhibitor since it remains bound to PNP for the life-time of the target in the cell (32).
The most potent diphenyl ether based compounds in the present study are slow onset inhibitors of ftuFabI. Based on a series of studies with this and other FabI enzymes, we believe that the slow step in formation of the final enzyme-inhibitor complex involves ordering of the substrate binding loop (11, 26, 33, 34). To better understand the biochemical mechanism of these slow, tight-binding inhibitors and their relationship with in vivo efficacy, we used progress curve analysis to determine the pre-steady state constants for the interaction of triclosan and selected diphenyl ethers with ftuFabI. We show that the kinetics of enzyme inhibition and the residence time of the compounds on the enzyme is a better predictor for in vivo antibacterial activity and that the longer residence time of 11 on the enzyme might provide the molecular explanation for the increase of the in vivo efficacy of this compound compared to the other diphenyl ethers (19). Note that the in vitro MIC values for the other selected compounds are significantly better than 11, yet 11, which has the longest residence time on the enzyme, is the most potent compound in vivo (Figure 6). While we are aware that many factors may contribute to differences in the in vivo activity of structurally related drugs, and that even minor changes in structure can impact properties such as pharmacokinetics, the change in residence time observed between the five compounds evaluated here may have a critical impact on target selectivity and intracellular activity. The data highlight the issues in translating in vitro antibacterial activities, determined at constant drug concentrations, to in vivo systems where the drug concentration will fluctuate between doses. These studies have broad implications for drug discovery programs that are largely driven by the measurement of inhibitor activity under equilibrium conditions, and show the benefit of determining both the kinetic and thermodynamics for the formation of drug-target complexes.
The F. tularensis FabI (ftuFabI; FTT 0782) and crotonyl-CoA (Cr-CoA) were prepared as described previously. All the kinetic experiments were carried out on a Cary 300 Bio (Varian) spectrophotometer at 25°C in 30 mM pipes, 150 mM NaCl and 1.0 mM EDTA pH 8.0 buffer. Kinetic parameters were determined spectrophotometrically by following the oxidation of NADH to NAD+ at 340 nm (ε= 6300 M−1cm−1). Since most of the diphenyl ethers are slow-onset inhibitors and in each case bind preferentially to the E-NAD+ form of the enzyme (Scheme 1), typical progress curve analysis cannot be used to obtain the inhibition constant since the concentration of NAD+ varies during the assay. Instead, ftuFabI (10 nM) was preincubated with a fixed concentration of NAD+, inhibitor (0–1000 μM) and DMSO (1%, V/V) in the presence of 250 μM NADH for 5 hours at 4 °C. The mixture was then warmed to room temperature, and the reactions were initiated by the addition of Cr-CoA (160 μM). Equation 1 was used to estimate the apparent inhibition constant Ki′
where v0 is the rate in the absence of inhibitor and [I] is the triclosan concentration.
This experiment was repeated at varying concentrations of NAD+ (10–200 μM) and the mechanism of inhibition with respect to NAD+ was determined by fitting the data to equations 2–4. K1and K2 are defined in panels A and B in Scheme 1 and represent the equilibrium constants for inhibitor binding to E-NAD+ and E-NADH, respectively.
Inhibitor binds exclusively to the E-NAD+ form
Inhibitor binds exclusively to the E-NADH form
Inhibitor binds both to the E-NAD+ and E-NADH forms
where the Km,NAD value for NAD+ was calculated from equation 5:
and Ki,NAD is the dissociation constant of NAD+. Equation 5 presumes that NAD+ is a competitive inhibitor with respect to NADH as was shown by the product inhibition studies (data not shown).
The mechanism of inhibition with respect to NADH was determined in an analogous fashion by including different concentrations of NADH (100–500 μM) in the inhibition incubations at a fixed concentration of NAD+ (50 μM) and by fitting the data to equations 2–4 using the concentration and Km value for NADH instead of NAD+.
For those diphenyl ethers that were rapid reversible inhibitors, the concentration of inhibitor in each assay was at least 10-fold larger than the enzyme concentration allowing the inhibition constants to be determined using standard Michaelis-Menten kinetics. Initial velocities were measured using a fixed Cr-CoA concentration (160 μM) and by varying the concentration of NADH and inhibitors. All these inhibitors were shown to be competitive with respect to NADH, and inhibition constants (Kis) were calculated by fitting the data to equation 6,
where [S] is the concentration of NADH, Km is the Michaelis-Menten constant for NADH, Vmax is the maximum velocity, [I] is the concentration of inhibitor added and Kis is the inhibition constant.
For the crystallization of the ftuFabI:NAD+ complex, equal amounts of protein (5 mg/ml ftuFabI in 30 mM PIPES pH 8.0, 150 mM NaCl and 1 mM EDTA) and precipitant solution (20 % PEG 3350, 0.2 M Mg acetate, 2 mM NAD+) were mixed and equilibrated against the precipitant solution without NAD+. Crystals grew within two days, were transferred into precipitant solution containing 25 % (v/v) glycerol and frozen in liquid nitrogen. Diffraction data were collected at beam line X26C at the National Synchrotron Light Source at Brookhaven National Laboratory and indexed, integrated and scaled with HKL2000 (data collection statistics in Table S1). The structure was solved by molecular replacement with the program Phaser (36), using one monomer of the apo form of the F. tularensis enoyl reductase (unpublished data) as the search model. Molecular replacement yielded four molecules in the crystallographic asymmetric unit that form a homotetramer. Only one NAD+ molecule was identified in the Fo−Fc difference density map and was built into the electron density using Coot (37). Model building and refinement was carried out using alternating rounds of Coot for manual model building and Refmac (38) for maximum likelihood refinement. Four-fold noncrystallographic symmetry (NCS) restraints were maintained throughout refinement. Five percent of all reflections were omitted throughout the refinement for the calculation of Rfree (refinement statistics in Table S1).
Progress curves were used to further quantitate the inhibition of ftuFabI by those compounds that were slow-onset inhibitors. In order to take into account the change in NAD+ concentration during the assay as mentioned above, assays contained 200 μM NAD+ in addition to ftuFabI (10 nM), Cr-CoA (160 μM), NADH (250 μM), inhibitor and DMSO (1%, v/v). Reactions were allowed to proceed until the progress curve became linear, indicating that the steady-state velocity had been reached. Low enzyme concentrations were used to ensure only a small fraction of Cr-CoA and NADH was consumed during the course of measurement so that the progress curves were essentially linear in the absence of inhibitor. The resulting progress curves were fitted to the integrated rate equation (7)
where γ = [E]*(1−vs/vi)2/[I], vi and vs are the initial velocity and steady-state velocity, and kobs is the observed rate constant. Values for vi, vs and kobs were extracted by fitting the data to equation 7 at each inhibitor concentration. Subsequently, Kiapp was calculated by fitting the values for kobs to the equation for a two-step inhibition mechanism (equation 8) in which the initial rapid binding of the inhibitor to the enzyme is followed by a second slow step resulting in the final enzyme-inhibitor complex (Scheme 1C).
where Kiapp = k−1/k1
F. tularensis LVS was cultured in modified Mueller-Hinton broth medium (0.025% ferric pyrophosphate, 2% IsoVitaleX, and 0.1% glucose) or on modified Mueller-Hinton agar medium (0.025% ferric pyrophosphate, 2% IsoVitaleX, 0.1% glucose, and 0.025% FBS) as previously described (39). MIC values were determined via NCCLS protocols using modified Mueller-Hinton broth as previously described (40). Compounds were tested at 2-fold serial concentrations from 0.000015 to 128 μg/ml in triplicate. MIC values were determined by visual inspection after ~20 hours at 37°C and by Microplate Alamar blue assay (MABA) as described by Franzblau et al. with modifications for Francisella (41).
Six week-old female ICR mice were purchased from Charles River Laboratories. All mice were housed in sterile microisolater cages in the Laboratory Animal Resources facility or in the Biohazard Research Building BSL-3 facility at Colorado State University. All research involving animals was conducted in accordance with animal care and use guidelines and animal protocols were approved by the Animal Care and Use Committee at Colorado State University. Mice were infected with F. tularensis Schu4 via low-dose aerosol infections in a Glascol Inhalation Exposure System (Glas-Col, Inc, Terre Haute, IN). Prior to exposure, the nebulizer was loaded with bacteria diluted in PBS to a concentration of approximately 5 × 106 cfu/ml. Mice were exposed to a total of approximately 4 × 107 bacteria aerosolized into a volume of 5 cubic feet over a period of 30 min, followed by a 20 minute period of cloud decay when airflow was maintained but bacteria were no longer introduced. Drugs were formulated in 5% ethanol-water with the exception of triclosan which was formulated in 8%-Solutol (BASF) and 5% ethanol-water. Either compound or vehicle as control were delivered intraperitoneally once per day for 5 days at 200 mg/kg/day beginning at the time of challenge. Mice were monitored for morbidity and mortality twice daily for a period of 10 days, at which time survivors were euthanized. Efficacy was determined from median survival values and percent survival to day 10.
African green monkey kidney cells (Vero cells) were grown in RPMI 1640 medium supplemented with 1.5 g/l sodium bicarbonate, 10 ml/l 100 mM sodium pyruvate, 140 ml/l100× nonessential amino acids, 100 ml/l penicillin-streptomycin solution (10,000 IU/10,000 μg/ml), and 10% bovine calf serum at 37°C in a 5% CO2 incubator with 75% humidity. Compounds were evaluated at 2-fold serial concentrations from 0.000015 to 128.0 μg/ml in triplicate. Following compound addition, cells were incubated for 72 h at 37°C in a 5%CO2 incubator. The cells were then washed with PBS, CellTiter 96 AQueous One solution was added to each well, and plates were incubated for 4 h at 37°C. Plates were read at 490 nm using a spectrophotometric plate reader, and the absorbance readings were used to calculate the 50% lethal concentration (LC50) using GraFit software as previously described (42)
This work was supported by New Opportunities funding to PJT and RAS from the Northeast Biodefense Center (AI057158) and the Rocky Mountain Regional Center of Excellence (AI065357), as well as NIH grants AI44639 and AI70383 to PJT.
Data Deposition: Coordinates have been deposited with the pdb: 2jjy.pdb