Compound synthesis

RUP2 and RUP4 were synthesized as described previously [27]. RUP7 was synthesized as detailed in the Supporting Information (S1 Appendix).

Bacterial strains and reagents

The following strains were obtained from the American Type Culture Collections (ATCC): A. baumannii 19606 and 17978, K. pneumoniae 13883, Pseudomonas aeruginosa 27853, Escherichia coli 25922, Listeria monocytogenes 19115, Bacillus subtilis 23857, and Staphylococcus saprophyticus 15305. A. baumannii clinical isolates (NR-13374, NR-13376, NR-13377, NR-13378, NR-13379, NR-13380, NR-13382, and NR-13385) and the Staphylococcus aureus (MRSA) clinical isolate (NR-46234) were obtained from BEI Resources. Mutant strains of S. aureus NR46234 expressing either G196S or N263K mutant FtsZ were generated as described previously [22]. Unless otherwise indicated, A. baumannii strains were grown in modified M9 media containing 6.8 g/L Na₂HPO₄, 3 g/L KH₂PO₄, 0.5 g/L NaCl, 1 g/L NH₄Cl, 2 mM MgSO₄, 0.1 mM CaCl₂, 0.4% glucose, 0.2% low-iron casein amino acids, 16.5 µg/mL thiamine hydrochloride, and 0.2 µM FeCl₃. Low-iron casein amino acids, tryptic soy agar (TSA), cation-adjusted Mueller Hinton (CAMH) media, brain-heart infusion (BHI) media, and agar were obtained from Difco. Iron (III) chloride (FeCl₃), thiamine hydrochloride, D-(+)-glucose, piperacillin, tazobactam, n-dodecyl-β-D-maltoside (DDM), GTP disodium salt, and GDP sodium salt were obtained from Sigma. Luria-Bertani (LB) media was obtained from Millipore. Ciprofloxacin was obtained from Fluka, cefsulodin sodium salt and mecillinam were obtained from RPI, aztreonam was obtained from Alfa Aesar, and cefepime hydrochloride was obtained from Toku-E. Tris-acetate-EDTA (TAE) buffer was obtained from Thermo Scientific, dimethyl sulfoxide (DMSO) was obtained from Fisher, and phosphate buffered saline (PBS) was obtained from Lonza.

Antibacterial assays

Broth microdilution assays of RUP2 antibacterial activity against wild-type (WT) S. aureus NR-46234 (MRSA), mutant S. aureus NR-46234 expressing either G196S or N263K mutant FtsZ, B. subtilis 23857, S. saprophyticus 15305, L. monocytogenes 19115, A. baumannii 19606, K. pneumoniae 13883, P. aeruginosa 27853, and E. coli 25922 were conducted by preparing two-fold serial dilutions of RUP2 on microtiter plates containing either CAMH media (for all strains except L. monocytogenes) or BHI media (for L. monocytogenes). Concentrations of RUP2 ranged from 0.16 to 80 µM, with each test concentration being present in triplicate. Log-phase bacteria were added to the plates at a final inoculum of 5 x 10⁵ CFU/mL and the final volume in each well was 100 µL. All plates were incubated with shaking at 37 °C for 18 h, whereupon the bacterial growth in each well was determined by measuring the optical density at 600 nm (OD₆₀₀). The percent bacterial growth in each well was determined by normalizing the corresponding OD₆₀₀ relative to that observed in the absence of test compound (100% growth) and the background value observed for media alone (0% growth).

Broth microdilution assays comparing the antibacterial activities of RUP7, RUP4, and RUP2 against A. baumannii 19606 in M9 media were conducted and percent growth determined as described above. Concentrations of each compound ranged from 0.16 to 80 µM, with each test concentration being present in triplicate.

For assays measuring the comparative antibacterial activity of RUP7, RUP4, and RUP2 against a library of A. baumannii clinical isolates, six microtiter wells each of DMSO vehicle, 20 µM RUP7, 20 µM RUP4, and 20 µM RUP2 were prepared in M9 media for each isolate tested. The final volume in each well was 100 µL. To all experimental wells, 5 x 10⁵ CFU/mL of exponential bacterial cells were added. The microtiter plates were then incubated with shaking at 37 °C for 18 h. All experimental wells were then serially diluted 10-fold in PBS, and 100 µL of each serial dilution was spread onto duplicate TSA plates. The plates were incubated overnight at 37 °C, and the resulting colonies were counted to determine the bacterial growth (CFU/mL) in each vehicle and compound sample.

To determine the impact of added exogenous Fe³⁺ on the antibacterial activity of RUP7 against A. baumannii 19606, serial dilutions of RUP7 were prepared in triplicate as described above in M9 media supplemented with 0, 0.5, 1, 5, or 10 µM of FeCl₃. The experimental and positive control wells were then inoculated with 5 x 10⁵ CFU/mL of exponential A. baumannii 19606 cells grown in the corresponding concentration of FeCl₃. Plates were incubated and percent growth was determined as described above.

Cellular uptake assays

For assays comparing the cellular uptake of RUP2 into the Gram-negative strains A. baumannii 19606 and K. pneumoniae 13883 versus the Gram-positive strains S. aureus NR-46234 (MRSA) and S. saprophyticus 15305, overnight cultures were diluted 1:100 into 100 mL of CAMH media and incubated with shaking at 37 °C until the OD₆₀₀ reached 0.4–0.8 (approximately 3–4 hours). Each 100 mL culture was centrifuged at 4,000 x g for 5 minutes, and the resulting pellets were resuspended in 5 mL of CAMH media and transferred to 14 mL culture tubes. 20 µM of RUP2 was added to the cultures, each being run in quintuplicate. All tubes were incubated with shaking at 37 °C for 30 min. 100 µL of each sample was then serially diluted 10-fold in PBS and 100 µL of each serial dilution was plated on duplicate TSA plates. The plates were incubated overnight at 37 °C, and the colonies on each plate were counted to determine the total CFUs in each sample culture. The remaining 4.9 mL of each culture was centrifuged at 16,000 x g for 1–5 min and the resulting pellets were washed with 4 x 1 mL of PBS. After the final centrifugation, each pellet was resuspended in 500 µL of PBS for A. baumannii and K. pneumoniae cells or PBS supplemented with 1 mg/mL lysostaphin for S. aureus and S. saprophyticus cells. The S. aureus and S. saprophyticus cells were then incubated for 1 h at 37 °C. All cells were then ultrasonicated in a 0 °C ice bath for 20 minutes with a 10 second on/off cycle at 30% amplitude using a Qsonica Q500 sonicator equipped with a 0.5-inch probe. The resulting lysates were centrifuged at 16,000 x g for 1 minute, and 200 µL of the resulting supernatants were transferred into microfuge tubes. 20 µL of 10x TURBO DNase buffer and 3 µL of TURBO DNase (Invitrogen) were then added to each lysate. The mixtures were then vortexed and incubated at 37 °C for 30 min. After incubation, 220 µL of acetonitrile acidified with 0.1% TFA was added to each sample, and the samples were vortexed at maximum speed for 20 sec. The samples were then centrifuged at 16,000 x g for 1 minute, and the concentration of RUP2 in each of the resulting supernatants was characterized by reverse-phase HPLC using standard curves generated for 0.5–4 µM and 4–20 µM RUP2 in DMSO. Differences in the cellular uptake of RUP2 in the Gram-positive versus the Gram-negative strains were compared statistically via one-way ANOVA.

For assays comparing the uptake of RUP7, RUP4, and RUP2 into A. baumannii 19606 cells, overnight cultures were diluted 1:100 into 100 mL of M9 media and incubated with shaking at 37 °C until the OD₆₀₀ reached 0.4–0.8 (approximately 4 hours). Each 100 mL culture was centrifuged at 4,000 x g for 5 minutes, and the resulting pellets were resuspended in 5 mL of M9 media and transferred to 14 mL culture tubes. 20 µM of RUP7, RUP4, or RUP2 was added to the cultures, with each being run in quintuplicate. All tubes were incubated with shaking at 37 °C for 30 min. The samples were then treated as described above for the A. baumannii cultures, with the concentrations of each compound in the final supernatants being characterized by reverse-phase HPLC using standard curves generated for 10–50 µM RUP7, 1–8 µM RUP4, and 0.5–4 µM RUP2 in DMSO. HPLC chromatograms of representative lysate replicates after treatment with each compound are shown in S1–S3 Figs. Reference HPLC chromatograms and associated standard curves used for determination of compound concentrations in lysate replicates are also shown in S1–S3 Figs. Differences in the cellular uptake of RUP7, RUP4, and RUP2 were compared statistically via one-way ANOVA.

HPLC measurements for the cellular uptake assays were performed on a Shimadzu LC-AT liquid chromatograph equipped with a Shimadzu SPD-20AV UV/Vis detector that was set at 260 nm. The samples were run through a 150 x 4.6 mm² SPHER-100 C18 column with a particle size of 5 µm and a pore size of 100 Å (Princeton Chromatography). The samples were injected into the column at a volume of 1 µL and a flow rate of 1 mL/min was applied. The mobile phase was a 10–90% gradient of acidified acetonitrile (0.1% TFA) and water. The total run time for each acquisition was 23 minutes, and the sampling frequency was 2 Hz with a response time of 1 sec. All peak areas were measured using the Shimadzu LabSolutions Lite software package (v5.82).

Fe³⁺ coordination assay

A solution of 10 µM RUP7 was prepared in methanol and transferred to a quartz cuvette (Hellma) with a 1-cm pathlength in both the excitation and emission directions. The fluorescence emission spectrum of the solution was acquired at 25 °C from 540 to 330 nm, with the excitation wavelength set at 318 nm. FeCl₃ was then titrated into the solution at concentrations ranging from 1 to 14 µM. After each addition, the sample was allowed to equilibrate at 25 °C for 3 min, whereupon the emission spectrum was reacquired. Each emission spectrum was corrected by subtracting the corresponding emission spectrum of methanol alone run as a background control. The fluorescence measurements were acquired on an AVIV Model ATF105 spectrofluorometer equipped with a thermoelectrically controlled cell holder. The bandwidths in both the excitation and emission directions were set at 5 nm. All spectra were acquired in 1 nm increments, with a 1 sec averaging time at each wavelength reading.

Values of fluorescence emission intensity (I) at 418 nm were plotted as a function of Fe³⁺ concentration, with the resulting plot being fit by non-linear least squares regression using the following dimeric binding formalism:

(1)

In this relationship, I₀ is the fluorescence emission intensity in the absence of Fe³⁺, I_∞ is the fluorescence emission intensity in the presence of an infinite concentration of Fe³⁺, [RUP7₂]_tot is the total RUP7 dimer concentration, [Fe³⁺]_tot is the total Fe³⁺ concentration, and K_d is the dissociation constant for the coordination reaction.

Assays for expression of acinetobactin uptake transporter, exporter, and biosynthesis genes in A. baumannii

We determined the expression of acinetobactin uptake transporter genes bauA, bauB, bauC, bauD, and bauE, the acinetobactin exporter genes barA and barB, and the acinetobactin biosynthesis genes basB and basD in A. baumannii 19606 cells grown in CAMH media, M9 media, or M9 media supplemented with 10 µM FeCl₃ by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) using a protocol we have described previously [27]. The primers used for each gene are listed in S1 Table.

Cloning, expression, and purification of the A. baumannii BauA and FtsZ proteins

Cloning, expression, and purification of A. baumannii BauA were performed as previously described [51]. Briefly, isolated outer membrane pellets were solubilized with 7% n-octylpolyoxyethylene (octyl-POE). The detergent was then exchanged to n-dodecyl-β-D-maltoside (DDM) at 3 x the critical micelle concentration (CMC) during the wash step of the Ni²⁺-NTA chromatography. The His-tag was subsequently cleaved by tobacco etch virus (TEV) protease, and the protein was further purified by a second immobilized metal affinity chromatography (IMAC) column. Lastly, the protein was then separated via gel filtration chromatography using Superdex S200 resin in 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.24 mM DDM. A. baumannii FtsZ (AbFtsZ) was cloned, expressed, and purified as detailed previously [42].

Fluorescence anisotropy assays of compound binding to BauA and AbFtsZ

Fluorescence anisotropy measurements were conducted at 25 °C on an AVIV model ATF105 spectrofluorometer equipped with a thermoelectrically controlled cell holder and computer controlled Glan-Thompson polarizers in both the excitation and emission directions. In all experiments, excitation and emission wavelengths were specifically chosen to maximize compound fluorescence and minimize any potential contributions from intrinsic protein fluorescence.

Binding of the RUP7-ferric iron coordination complex to BauA.

A solution of 1 µM RUP7 and 1 µM FeCl₃ (an FeCl₃ concentration chosen to ensure an excess of Fe³⁺ given the observed 2RUP7:1Fe³⁺ coordination stoichiometry) was prepared in buffer containing 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 0.24 mM DDM. The solution was transferred to a quartz ultra-micro cuvette (Hellma) with a 2 x 5 mm aperture and a 15 mm center height. The pathlengths in the excitation and emissions directions were 1 and 0.2 cm, respectively. The fluorescence anisotropy was then recorded as an average of 5 readings, with the excitation and emission wavelengths set at 341 and 457 nm, respectively. The bandwidths were set at 8 nm in both the excitation and emission directions. BauA was then titrated into the solution at concentrations ranging from 0.25 to 2.75 µM. After each addition, the sample was allowed to equilibrate for 5 min, whereupon the fluorescence anisotropy was reacquired in quintuplicate.

Values of fluorescence anisotropy (r) were plotted as a function of BauA concentration, with the resulting plot being fit by non-linear least squares regression using the following 1:1 binding formalism:

(2)

In this relationship, r₀ is the fluorescence anisotropy intensity in the absence of BauA, r_∞ is the fluorescence anisotropy in the presence of an infinite concentration of BauA, [RUP7₂-Fe³⁺]_tot is the total concentration of RUP7₂-Fe³⁺ coordination complex, [BauA]_tot is the total BauA concentration, and K_d is the dissociation constant for the binding reaction.

Binding of RUP7 and RUP2 to AbFtsZ.

All the anisotropy studies of AbFtsZ binding were conducted in buffer containing 50 mM Tris-HCl (pH 7.6) and 50 mM KCl. In these characterizations, AbFtsZ was titrated into a solution of compound. After each protein addition, samples were allowed to equilibrate for 5 min, whereupon the fluorescence anisotropy was acquired in quintuplicate.

For our characterization of the RUP7-AbFtsZ binding interaction, the protein was titrated into a solution of 5 µM RUP7, with the final AbFtsZ concentrations ranging from 0 to 62 µM. These anisotropy measurements were conducted using the same ultra-micro cuvette described above in the BauA binding studies. The bandwidths were set to 5 nm in both the excitation and emission directions, and the excitation and emission wavelengths set to 325 and 517 nm, respectively. For our characterization of the RUP2-AbFtsZ binding interaction, the protein was titrated into a solution of 1 µM RUP2, with the final AbFtsZ concentrations ranging from 0 to 3 µM. The bandwidths were set to 7 nm in both the excitation and emission directions, and the excitation and emission wavelengths were set to 300 and 370 nm, respectively. These measurements were conducted using a quartz cuvette with a pathlength of 1 cm in both the excitation and emission directions.

Values of r were plotted as a function of AbFtsZ concentration, with the resulting plot being fit by non-linear least squares regression using the following 1:1 binding formalism:

(3)

In this relationship, r₀ is the fluorescence anisotropy intensity in the absence of AbFtsZ, r_∞ is the fluorescence anisotropy in the presence of an infinite concentration of AbFtsZ, [Comp]_tot is the total compound concentration, [AbFtsZ]_tot is the total AbFtsZ concentration, and K_d is the dissociation constant for the binding reaction.

AbFtsZ polymerization assay

AbFtsZ polymerization was monitored using a 90°-angle light scattering assay conducted at 25 °C on an AVIV Model ATF105 spectrofluorometer equipped with a thermoelectrically controlled cell holder. In this assay, second-order light scattering was measured at an emission wavelength of 680 nm, with the excitation wavelength set at 340 nm. The bandwidth was set at 2 nm in both the excitation and emission directions. Samples containing 250 µM GTP and either DMSO vehicle, 20 µM RUP7, or 40 µM RUP7 were prepared in buffer consisting of 50 mM Tris-HCl (pH 7.6), 50 mM KCl, and 5 mM Mg(CH₃COO)₂. The polymerization reaction in each sample was initiated by subsequent addition 10 µM AbFtsZ. The second-order light scattering of each sample was then recorded once per second over a period of 480 seconds, with a time constant of 0.5 seconds. As a negative control, a DMSO vehicle sample was prepared and monitored as described above except that 250 µM GDP was used in place of the GTP. All the light scattering measurements were conducted using the same ultra-micro cuvette described above in the BauA binding studies.

Antibacterial synergy assays

Studies measuring the antibacterial activity of sub-inhibitory concentrations of RUP7 against A. baumannii 19606 when combined with sub-inhibitory concentrations of different clinical antibiotics were performed in microtiter plates. Three replicates each of DMSO vehicle, 80 µM RUP7, 2.5 µM ciprofloxacin, 20 µM cefepime, 20 µM TXH11106, 2560 µM mecillinam, 10 µM ceftazidime, 20 µM piperacillin:tazobactam (8:1), 80 µM aztreonam, 80 µM cefsulodin, or the combination of each antibiotic with RUP7 were prepared in 100 µL of M9 media. Log-phase A. baumannii 19606 cells were added to all experimental wells at an inoculum of 1 x 10⁶ CFU/mL. The microtiter plates were then incubated with shaking at 37 °C for 24 hours. Experimental wells were then serially diluted 10-fold in PBS, and 100 µL of each serial dilution was plated on duplicate TSA plates. The plates were incubated at 37 °C overnight, and colonies were counted to quantify bacterial growth (CFU/mL).

The observed logs of cell kill resulting from each combination treatment that exceeded the additive effect from treatment with each agent alone was determined using the following relationship:

(4)

In this relationship, Δlog(CFU/mL)_Antib, Δlog(CFU/mL)_RUP7, and Δlog(CFU/mL)_Comb, were derived by subtraction of the log(CFU/mL) value for the vehicle-treated culture from the log(CFU/mL) values for the cultures treated with antibiotic alone, RUP7 alone, and the RUP7-antibiotic combination, respectively. An LKEA >2 was viewed as reflecting synergy.

Cytotoxicity assay

Cytotoxicity studies of RUP7 against mammalian opossum Kidney (OK) cells (ATCC CRL-1840) were performed using a 4-day MTT assay as described previously [52]. OK cells were treated with six replicates each of DMSO vehicle or RUP7 at a concentration of 5, 10, 20, or 40 µM for 4 days and cell viability was then measured. 4 µM cycloheximide (CHX) was used as a positive control. Statistical comparisons of each treatment relative to DMSO vehicle were performed via one-way ANOVA.

Non-conjugated RUP2 exhibits potent activity against Gram-positive bacteria but poor activity against Gram-negative pathogens

We initially sought to delineate the general specificity of antibacterial action associated with non-conjugated RUP2. To this end, we evaluated the antibacterial activity of RUP2 against a broad range of both Gram-positive and Gram-negative bacterial species. Gram-positive species included S. aureus NR-46234 (MRSA), S. saprophyticus 15305, L. monocytogenes 19115, and B. subtilis 23857, while Gram-negative species included A. baumannii 19606, K. pneumoniae 13883, P. aeruginosa 27853, and E. coli 25922. As revealed by the percent growth profiles shown in Fig 2A (acquired in the presence of RUP2 concentrations ranging from 0.16 to 80 µM), RUP2 exhibits potent activity versus the Gram-positive bacterial species but poor activity versus the Gram-negative species. Control studies showing reduced activity of RUP2 against S. aureus NR-46234 expressing G196S or N263K mutant versus wild-type FtsZ are consistent with a FtsZ-targeting mechanism of antibacterial action (S4 Fig). The Gram-positive specificity of antibacterial action exhibited by RUP2 is commonly observed among benzamide-based FtsZ inhibitors [9–23].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Antibacterial activity and cellular uptake of RUP2 in Gram-positive versus Gram-negative strains.

(A) Antibacterial activity of RUP2 against the Gram-positive bacterial strains S. aureus NR-46234 (MRSA), B. subtilis 23857, S. saprophyticus 15305, and L. monocytogenes 19115 as well as against the Gram-negative strains A. baumannii 19606, K. pneumoniae 13883, P. aeruginosa 27853, and E. coli 25922. All bacterial strains were assayed in CAMH media, except for L. monocytogenes, which was assayed in BHI media. Each experimental datapoint represents the average of three replicates, with the error bars reflecting the standard deviations from the mean. (B) Cellular uptake of RUP2 into S. aureus NR-46234 (MRSA), S. saprophyticus 15305, A. baumannii 19606, and K. pneumoniae 13883 cells. Each bar represents an average of five replicates, with the error bars reflecting the standard errors from the mean. Statistical comparisons of RUP2 uptake into the Gram-positive relative to the Gram-negative pathogens were conducted via one-way ANOVA. ***, P ≤ 0.001; **, 0.01 ≥ P > 0.001.

https://doi.org/10.1371/journal.pone.0334409.g002

The comparatively poor activity of RUP2 against Gram-negative versus Gram-positive species is directly correlated with correspondingly diminished cellular uptake

To determine whether differential influx may be a contributing factor in the antibacterial specificity of action exhibited by RUP2, we used an HPLC-based assay to quantify the cellular uptake of RUP2 into a large inoculum (~10¹¹) of the Gram-negative strains A. baumannii 19606 and K. pneumoniae 13883 versus the Gram-positive strains S. aureus NR-46234 and S. saprophyticus 15305 after 30 minutes of treatment with 20 µM compound. Notably, the uptake of RUP2 was significantly greater into the Gram-positive S. aureus and S. saprophyticus cells relative to the Gram-negative A. baumannii and K. pneumoniae cells, with measured intracellular quantities of 9.4 ± 1.2 x 10¹⁰ nmol/CFU for S. aureus and 5.0 ± 1.5 x 10¹⁰ nmol/CFU for S. saprophyticus relative to 1.4 ± 0.1 x 10¹⁰ nmol/CFU for A. baumannii and 0.17 ± 0.03 x 10¹⁰ nmol/CFU for K. pneumoniae (Fig 2B). The diminished uptake of RUP2 into the Gram-negative versus Gram-positive bacterial cells correlates well with the comparatively poor activity of the compound against the Gram-negative versus Gram-positive species. In a broader sense, the correlation we observe in the antibacterial and cellular uptake properties of RUP2 highlights differential cellular influx as a contributing factor in the Gram-positive specificity of antibacterial action typically associated with benzamide-based FtsZ inhibitors.

RUP7 exhibits enhanced activity against A. baumannii relative to that of the synthetic chlorocatechol conjugate RUP4 and the non-conjugated RUP2

The rational design of RUP7 was motivated by the hypothesis that conjugating the native siderophore acinetobactin to the oxazole-benzamide FtsZ inhibitor would enhance the activity against A. baumannii relative to the non-conjugated RUP2 as well as the synthetic chlorocatechol siderophore conjugate RUP4. We first tested this hypothesis against A. baumannii 19606 by measuring percent growth in Fe³⁺-limiting M9 media at RUP7, RUP4, and RUP2 concentrations ranging from 0.16 to 80 µM. Consistent with our hypothesis, the activity of RUP7 was enhanced over a broad concentration range relative to both RUP4 and RUP2 (Fig 3A). RUP4 had modest activity at higher concentration levels, with RUP2 exhibiting minimal activity even at the highest concentration tested. Thus, conjugation of the native acinetobactin siderophore to the FtsZ inhibitor markedly enhanced activity against A. baumannii, with this enhancement being significantly greater than that produced by conjugation to the synthetic chlorocatechol siderophore.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Antibacterial activity and cellular uptake of RUP7, RUP4, and RUP2 in A. baumannii 19606.

Each experimental datapoint represents the average of three replicates, with the error bars reflecting the standard deviations from the mean. (B) Cellular uptake of RUP7, RUP4, and RUP2 into A. baumannii 19606 cells. Each bar represents the average of five replicates, with the error bars reflecting the standard errors from the mean. Statistical comparisons of compound uptake were conducted via one-way ANOVA. ***, P ≤ 0.001.

https://doi.org/10.1371/journal.pone.0334409.g003

The enhanced activity of RUP7 against A. baumannii relative to RUP4 and RUP2 is correlated with a correspondingly enhanced cellular uptake

To explore the basis for the enhanced activity of RUP7 against A. baumannii relative to RUP4 and RUP2, we compared the uptake of the three compounds into A. baumannii 19606 cells using the same HPLC-based assay described above. Significantly, the relative cellular uptake of the compounds correlated with their relative antibacterial activities. In this connection, RUP7 exhibited the greatest extent of cellular uptake at 29.9 ± 15.8 x 10¹⁰ nmol/CFU, followed by RUP4 at 4.3 ± 2.1 x 10¹⁰ nmol/CFU, and RUP2 at 0.44 ± 0.14 x 10¹⁰ nmol/CFU (Fig 3B). These findings provide a clear demonstration that siderophore conjugation enhances the cellular uptake of the FtsZ inhibitor into A. baumannii cells, with conjugation to acinetobactin yielding the greatest degree of uptake.

The enhanced activity of RUP7 relative to RUP4 and RUP2 extends to a library of A. baumannii clinical isolates

Our initial characterization of the relative activities of RUP7, RUP4, and RUP2 was focused on a single A. baumannii strain (19606). We also extended these characterizations to include a library of eight additional A. baumannii clinical isolates (NR-13374, NR-13375, NR-13377, NR-13378, NR-13379, NR-13380, NR-13382, and NR-13385) as well as reference ATCC strain 17978. In these assays, we compared the activity of RUP7, RUP4, and RUP2 against the different isolates by measuring the number of CFUs after 18 hours of treatment at 20 µM, a concentration at which RUP7 inhibits the growth of A. baumannii 19606 by approximately 75% (see Fig 3A). Significantly, the CFU-based results (shown in Fig 4) revealed a similar pattern of relative activity to that observed against 19606, with RUP7 consistently exhibiting the greatest extent of growth inhibition relative to vehicle, followed by RUP4, which had a modest inhibitory effect against most isolates. RUP2 was completely inactive against all but one isolate. Note that two of the isolates tested (NR-13375 and NR-13377) are resistant to the current standard-of-care antibiotics piperacillin:tazobactam and ceftazidime. Thus, the activity of RUP7 also extends to resistant isolates of A. baumannii.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Antibacterial activity of RUP7, RUP4, and RUP2 compared to DMSO vehicle against A. baumannii reference strains 19606 (A) and 17978 (B) as well as clinical isolates NR-13374 (C), NR-13375 (D), NR-13377 (E), NR-13378 (F), NR-13379 (G), NR-13380 (H), NR-13382 (I), and NR-13385 (J).

Each box plot represents 6 replicates. Statistical comparisons of RUP7, RUP4, or RUP2 relative to DMSO vehicle were conducted via one-way ANOVA. ***, P ≤ 0.001; **, 0.01 ≥ P > 0.001; *, 0.05 ≥ P > 0.01; ns, not significant.

https://doi.org/10.1371/journal.pone.0334409.g004

RUP7 coordinates ferric iron (Fe³⁺) with a 2:1 stoichiometry

As Fe³⁺ coordination is critical for siderophore recognition and intracellular uptake via endogenous siderophore transport pathways, we next endeavored to establish the ability of RUP7 to coordinate Fe³⁺ in a manner analogous to free unconjugated acinetobactin. To this end, we used a fluorescence-based approach to detect and characterize the interaction of RUP7 with Fe³⁺ by leveraging the intrinsic fluorescence properties of the conjugated acinetobactin functionality, which were similar to those previously defined by Kim and coworkers for unconjugated acinetobactin [50]. Addition of Fe³⁺ induced a profound concentration-dependent decrease in the magnitude of the RUP7 fluorescence emission spectrum (Fig 5A), indicative of RUP7 complexation with Fe³⁺. The concentration dependence of the Fe³⁺-induced decrease in RUP7 fluorescence emission intensity at 418 nm (I₄₁₈) could not be fit by a 1:1 binding model but was well fit by a dimeric (2RUP7:1Fe³⁺) binding model, which yielded an apparent K_d of 1.7 ± 0.2 µM for the RUP7₂-Fe³⁺ coordination reaction (Fig 5B). This observed stoichiometry for Fe³⁺ coordination is consistent with that previously reported by Kim and coworkers for unconjugated acinetobactin [50], a gratifying concordance indicating that conjugation of the FtsZ inhibitor to the imidazole moiety of acinetobactin does not interfere with the ability of the siderophore to coordinate Fe³⁺. Previously reported crystallographic and computational studies with acinetobactin have shown that coordination of Fe³⁺ involves both the imidazole and catechol functionalities of two acinetobactin molecules to form a hexadentate coordination complex [49,51,53]. It is therefore likely that two RUP7 molecules form a similar hexadentate coordination complex with Fe³⁺ (as schematically depicted in the inset of Fig 5B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Coordination of ferric iron (Fe³⁺) by RUP7.

(B) Values of fluorescence emission intensity at 418 nm (I₄₁₈) plotted as a function of Fe³⁺ concentration. The solid line represents a non-linear least squares fit of the I₄₁₈ values using Equation (1), with the indicated K_d value for the RUP7₂-Fe³⁺ coordination reaction being derived from this fit. The inset depicts a proposed hexadentate model for the RUP7₂-Fe³⁺ coordination complex.

https://doi.org/10.1371/journal.pone.0334409.g005

The activity of RUP7 against A. baumannii is acutely dependent on the concentration of exogenous Fe³⁺

In addition to demonstrating the coordination of Fe³⁺ by RUP7, we wanted to determine if the antibacterial activity of the conjugate depends on the concentration of exogenous Fe³⁺. To this end, we measured the activity of RUP7 (at concentrations ranging from 0.16 to 80 µM) against A. baumannii 19606 cells grown in Fe³⁺-limiting M9 media supplemented with either 0, 0.5, 1, 5, or 10 µM exogenous Fe³⁺. RUP7 activity was significantly attenuated in a Fe³⁺-dependent manner, with activity being abolished in the presence of 5 and 10 µM added Fe³⁺, even at the highest tested concentrations of RUP7 (Fig 6A). Thus, the activity of RUP7 is highly dependent on Fe³⁺-limiting environmental conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Impact of environmental Fe³⁺ on the activity of RUP7 against A. baumannii 19606 and the expression of acinetobactin uptake transporter, exporter, and biosynthesis genes.

The concentrations of added exogenous Fe³⁺ are indicated. Each experimental datapoint represents the average of three replicates, with the error bars reflecting the standard deviations from the mean. (B) Expression of acinetobactin uptake transporter, exporter, and biosynthesis genes in A. baumannii 19606 cells grown in CAMH media (black), M9 media (red), or M9 media supplemented with 10 µM Fe³⁺ (blue). Each bar represents an average of 5 replicates, with the error bars reflecting the standard errors of the mean. Statistical comparisons were conducted via one-way ANOVA. ***, P ≤ 0.001.

https://doi.org/10.1371/journal.pone.0334409.g006

Expression of acinetobactin biosynthesis, exporter, and uptake transporter genes in A. baumannii is highly upregulated in Fe³⁺-limiting versus Fe³⁺-rich conditions

To better understand the basis for the Fe³⁺ dependence of RUP7 activity, we used RT-qPCR to evaluate the expression of acinetobactin biosynthesis, exporter, and uptake transporter genes by A. baumannii 19606 cells in Fe³⁺-rich versus Fe³⁺-limiting growth environments. In these characterizations, we investigated the following acinetobactin uptake transporter genes: the outer membrane transporter bauA, the periplasmic chaperone bauB, the heterodimeric inner membrane transporter components bauC and bauD, and the inner membrane transporter ATPase bauE. We also investigated the acinetobactin exporter genes barA and barB as well as the acinetobactin biosynthesis genes basB and basD, which encode critical enzymes for late stage acinetobactin assembly from precursors. Expression of RNA polymerase component genes rpoB and rpoD was used as endogenous controls for all comparisons.

All the tested acinetobactin biosynthesis, exporter, and uptake transporter genes were highly upregulated in Fe³⁺-limiting M9 media compared to Fe³⁺-rich CAMH media (Fig 6B). Among the uptake transporter genes, expression of bauC was the most upregulated at 163-fold, followed by bauD, bauE, bauA, and bauB at 139-, 90.3-, 71.0-, and 32.0-fold, respectively. Expression of the exporter genes barA and barB, as well as the biosynthesis genes basB and basD, was upregulated to an extent ranging from 29.7- to 96.2-fold. The upregulated expression under Fe³⁺-limiting conditions we observe for the biosynthesis genes basB and basD and the inner membrane transporter component gene bauD is consistent with that previously reported by Sheldon and Skaar in A. baumannii 17978 cells [44]. As expected, the addition of exogenous 10 µM Fe³⁺ to M9 media repressed the expression of all tested genes to a level comparable with that observed in CAMH (Fig 6B). Our gene expression results, coupled with the Fe³⁺ dependence of RUP7 activity, point to upregulated acinetobactin uptake transport systems as being important determinants of RUP7 activity against A. baumannii. Importantly, such upregulation would be manifest in the Fe³⁺-limiting conditions typically present at common sites of A. baumannii infection, including the respiratory tract, bloodstream, urinary tract, skin, and soft tissue [44,54–56]. It is also important to note that upregulated biosynthesis of endogenous acinetobactin under Fe³⁺-limiting conditions introduces potential competition with RUP7 for transporter-mediated cellular uptake. That said, the activity of RUP7 we observe provides an indication that such potential competition can be overcome.

The RUP7₂-Fe³⁺ coordination complex directly interacts with the acinetobactin outer membrane uptake transporter BauA

Given that RUP7 coordinates Fe³⁺ in a similar manner to unconjugated acinetobactin, we hypothesized that RUP7 can enter A. baumannii cells via the acinetobactin uptake transport system BauABCDEF, which is highly upregulated under the Fe³⁺-limiting conditions that promote RUP7 activity. In this connection, we probed for a direct interaction of the RUP7₂-Fe³⁺ coordination complex with the acinetobactin outer membrane uptake transporter BauA. To this end, we leveraged the intrinsic fluorescence properties of RUP7 by monitoring the fluorescence anisotropy of the RUP7₂-Fe³⁺ coordination complex as a function of added BauA. Addition of BauA induced a protein-dependent increase in fluorescence anisotropy (Fig 7A) consistent with a direct binding interaction. The anisotropy increase was well fit by the 1:1 binding formalism described by Equation (2), with this analysis yielding an apparent K_d of 1.8 ± 0.4 µM for the binding of RUP7₂-Fe³⁺ to BauA. This direct binding interaction suggests that BauA may indeed serve as the outer membrane transporter for uptake of RUP7₂-Fe³⁺ into A. baumannii, as schematically depicted in Fig 7B. Following BauA-mediated transport across the outer membrane, Fe³⁺-charged acinetobactin is then transported across the periplasm and inner membrane via the BauBCDE transporter system [44]. It is therefore possible that RUP7₂-Fe³⁺ is transported via the same system (Fig 7B). Additional studies with specific Bau mutants will be required to fully validate this proposed uptake pathway.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Binding of the RUP7₂-Fe³⁺ coordination complex to A. baumannii BauA and proposed model for the uptake of RUP7₂-Fe³⁺ via the BauABCDEF acinetobactin uptake transport system.

Each experimental datapoint represents the average of five replicates, with the error bars reflecting the standard deviations from the mean. The solid line represents a non-linear least squares fit of the r values using Equation (2), with the indicated K_d value for the binding interaction being derived from this fit. (B) Proposed model for the uptake of RUP7₂-Fe³⁺ via the BauABCDEF acinetobactin uptake transport system.

https://doi.org/10.1371/journal.pone.0334409.g007

RUP7 binds AbFtsZ and inhibits GTP-dependent AbFtsZ polymerization

We next wanted to confirm that the RUP7-acinetobactin conjugate can target AbFtsZ. To this end, we probed for the direct binding of RUP7 to AbFtsZ using fluorescence anisotropy. Addition of AbFtsZ induced an increase in the fluorescence anisotropy of RUP7 in a concentration-dependent manner (Fig 8A), indicative of a direct binding interaction. Analysis of this anisotropy profile with Equation (3) yielded an affinity constant (K_d) of 26.5 ± 3.7 µM. Corresponding anisotropy measurements with non-conjugated RUP2 yielded a K_d of 0.23 ± 0.09 µM (S5 Fig), a binding affinity two orders of magnitude greater than that of RUP7. Thus, conjugation of the bulky acinetobactin moiety markedly reduced affinity for AbFtsZ. Yet, despite this reduced target affinity, the RUP7-acinetobactin conjugate is associated with significantly enhanced activity against A. baumannii relative to non-conjugated RUP2, due in large part to its enhanced intracellular uptake. These collective results highlight the contributions of intracellular uptake to antibacterial activity, perhaps even more so than target affinity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Binding of RUP7 to AbFtsZ and the impact of binding on GTP-dependent AbFtsZ polymerization.

Each experimental datapoint represents the average of five replicates, with the error bars reflecting the standard deviations from the mean. The solid line represents a non-linear least squares fit of the r values using Equation (3), with the indicated K_d value being derived from this fit. (B) Time-dependent 90°-angle light scattering profiles of 10 µM AbFtsZ in the presence of 0 µM RUP7 + 250 µM GTP (black), 20 µM RUP7 + 250 µM GTP (red), 40 µM RUP7 + 250 µM GTP (blue), or 0 µM RUP7 + 250 µM GDP (green).

https://doi.org/10.1371/journal.pone.0334409.g008

We also explored the impact RUP7 on the GTP-dependent self-polymerization of AbFtsZ using a 90°-angle light scattering assay analogous to that previously established by Mukherjee and Lutkenhaus [57]. In this assay, 250 µM GTP was combined with DMSO vehicle (0 µM RUP7), 20 µM RUP7, or 40 µM RUP7. In addition, a 0 µM RUP7 sample with 250 µM GDP in place of GTP was also prepared as a negative control. 10 µM AbFtsZ was then added to each sample, thereby initiating polymerization in the GTP-containing samples. Immediately following addition of AbFtsZ, 2^nd-order light scattering was monitored as a function time, with the resulting light scattering profiles being shown in Fig 8B. Note that the GTP-containing samples each exhibit a time-dependent transition from a higher to a lower level of light scattering, reflecting a transition from a polymerized to a depolymerized state as the GTP is exhausted due to hydrolysis [57]. By contrast, no such transition is observed in the GDP-containing sample, which exhibits a steady low level of light scattering indicative of a lack of induced polymerization. Significantly, the presence of 20 and 40 µM RUP7 systematically decreases the maximal light scattering levels in a concentration-dependent manner, indicating correspondingly reduced extents of GTP-dependent polymerization in the presence of increasing compound concentrations. These observations are consistent with RUP7 acting as an inhibitor of AbFtsZ polymerization. Scoffone et al. observed a similar behavior associated with the benzothiadiazole AbFtsZ inhibitor C109 [24].

RUP7 exhibits bactericidal synergy against A. baumannii when used in combination with antibiotics that target PBP3

Synergistic combinations of antibiotics are an appealing strategy to increase antibacterial potency, reduce cytotoxicity, and mitigate the potential for the emergence of resistance [58–60]. We probed for antibiotics that can act synergistically in combination with RUP7. Toward this end, we combined RUP7 and representative members of different classes of antibiotics, including the DNA gyrase inhibitor ciprofloxacin (CIP), the PBP1a/1b inhibitor cefepime (CFP), the MreB inhibitor TXH11106 (TXH), the PBP2 inhibitor mecillinam (MEC), and the PBP3 inhibitor ceftazidime (CFZ). In each analysis, we tested DMSO vehicle, RUP7 alone, antibiotic alone, or a combination of RUP7 and antibiotic. Upon assessment of the number of CFUs after 24 hours of treatment, the combination of RUP7 with all antibiotics except CIP exhibited notably reduced CFUs relative to DMSO vehicle or test agent alone (Fig 9A). By contrast, all test agents alone exhibited CFUs comparable to that of vehicle. To assess synergy, we further quantified the log kill exceeding additivity (LKEA) for each antibiotic+RUP7 combination relative to both agents alone using Equation (4), with an LKEA of >2-logs being reflective of synergy. Note that the combinations of CFZ+RUP7, MEC+RUP7, and TXH+RUP7 were all associated with bactericidal synergy, with CFZ+RUP7 exhibiting the greatest degree of synergy (LKEA = 4.2 ± 0.4), followed by MEC+RUP7 (LKEA = 3.6 ± 0.2) and TXH+RUP7 (LKEA = 2.6 ± 0.2) (Fig 9B). Neither the combination of CIP+RUP7 (LKEA = 0 ± 0.1) nor CFP+RUP7 (LKEA = 1.8 ± 0.5) was associated with any significant synergistic activity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Assessment of bactericidal synergy associated with different antibiotic+RUP7 combinations.

(A) Screen for antibacterial synergy of RUP7 in combination with the antibiotics ciprofloxacin (CIP), cefepime (CFP), TXH11106 (TXH), mecillinam (MEC), or ceftazidime (CFZ) against A. baumannii 19606 in M9 media. (B) Logs of kill exceeding additivity (LKEA) for each antibiotic+RUP7 combination shown in panel (A), as derived using Equation (4). (C) Antibacterial synergy of RUP7 in combination with the antibiotics ceftazidime (CFZ), piperacillin:tazobactam (8:1) (P:T), aztreonam (AZN), or cefsulodin (CFS) against A. baumannii 19606 in M9 media. (D) LKEA for each antibiotic+RUP7 combination shown in panel (C), derived as above. In panels (A) and (C), the bacteria were treated with DMSO vehicle (black), RUP7 alone (red), antibiotic alone (blue), or RUP7 in combination with antibiotic (green). The red dashed line in panels (B) and (D) denotes the 2-log threshold above which LKEA is reflective of synergy. All data bars represent the average of three replicates, with the associated error bars reflecting the standard deviations from the mean. Statistical comparisons of each antibiotic+RUP7 combination relative to either agent alone were conducted via one-way ANOVA. ***, P ≤ 0.001; **, 0.01 ≥ P > 0.001; *, 0.05 ≥ P > 0.01.

https://doi.org/10.1371/journal.pone.0334409.g009

Our initial screen revealed that the greatest degree of bactericidal synergy was associated with combination of RUP7 and the PBP3 inhibitor CFZ. We further explored this behavior by performing a second screen with a panel of four PBP3-targeting antibiotics that included piperacillin:tazobactam at a ratio of 8:1 (P:T), aztreonam (AZN), and cefsulodin (CFS) in addition to CFZ. All four combinations were associated with a significant reduction in CFUs relative to vehicle or either test agent alone, with no detectable colonies being observed for either the AZN+RUP7 or CFS+RUP7 combination (Fig 9C). All four antibiotic + RUP7 combinations exhibited significant bactericidal synergy, with CFS+RUP7 and AZN+RUP7 exhibiting the greatest degrees of synergy (LKEA = 7.1 ± 0.2 and 6.5 ± 0.4, respectively), followed by P:T+RUP7 (LKEA = 4.4 ± 0.1) and CFZ+RUP7 (LKEA = 4.4 ± 0.5) (Fig 9D). The synergy was sufficiently significant for the AZN+RUP7 and CFS+RUP7 combinations that we observed complete kill of the entire bacterial inoculum after 24 hours of treatment (Fig 9C).

PBP3 plays an important role in bacterial cell division, serving as a critical component of the bacterial divisome [61,62]. Significantly, the localization of PBP3 to the midcell of a dividing bacterium depends on a functional FtsZ protein [63,64]. In addition, constriction of the FtsZ Z-ring during bacterial cytokinesis relies on cell wall synthesis at midcell, a process for which PBP3 is indispensable [65]. This complementarity of function between PBP3 and FtsZ for cell division could account for the significant synergy we observe with the combination of RUP7 and PBP3 inhibitors. In the aggregate, our synergy results highlight combinations of FtsZ and PBP3 inhibitors as an appealing pharmacologic approach for the treatment of A. baumannii infections.

RUP7 exhibits minimal cytotoxicity against mammalian cells

The potential utility of any antibacterial agent requires a specificity of action against bacterial versus host cells. We therefore sought to evaluate RUP7 for any potential cytotoxicity against mammalian cells. To this end, we used a 4-day MTT colorimetric assay to monitor the viability of opossum kidney (OK) cells (ATCC CRL-1840) upon exposure to RUP7 at concentrations ranging from 5 to 40 µM. Additionally, we included DMSO vehicle as a negative control and 4 µM cycloheximide (CHX) as a positive control. No significant reduction in A₅₇₀ was observed at an RUP7 concentration of 5, 10, or 20 µM when compared to vehicle treatment, with 40 µM RUP7 resulting in only a 16% reduction in A₅₇₀ (Fig 10). As expected, 4 µM CHX treatment resulted in a 44% decline in A₅₇₀. While these encouraging initial results reveal minimal mammalian cytotoxicity on the part of RUP7, additional MTT and clonogenic (colony formation) studies against a broader panel of both mammalian and human cells will be required to fully confirm this behavior.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. MTT-based assay for cytotoxicity against mammalian opossum kidney (OK) cells associated with exposure to RUP7.

Cells were treated with DMSO vehicle (black), 5 µM RUP7 (red), 10 µM RUP7 (blue), 20 µM RUP7 (green), 40 µM RUP7 (cyan), or 4 µM cycloheximide (CHX) (purple). Each box plot represents 4 replicates, and statistical comparisons of each treatment relative to DMSO vehicle were performed via one-way ANOVA. A₅₇₀, absorbance at 570 nm; ***, P ≤ 0.001; **, 0.01 ≥ P ≥ 0.001; ns, not significant.

https://doi.org/10.1371/journal.pone.0334409.g010