Proanthocyanidin Interferes with Intrinsic Antibiotic Resistance Mechanisms of Gram‐Negative Bacteria

Abstract Antibiotic resistance is spreading at an alarming rate among pathogenic bacteria in both medicine and agriculture. Interfering with the intrinsic resistance mechanisms displayed by pathogenic bacteria has the potential to make antibiotics more effective and decrease the spread of acquired antibiotic resistance. Here, it is demonstrated that cranberry proanthocyanidin (cPAC) prevents the evolution of resistance to tetracycline in Escherichia coli and Pseudomonas aeruginosa, rescues antibiotic efficacy against antibiotic‐exposed cells, and represses biofilm formation. It is shown that cPAC has a potentiating effect, both in vitro and in vivo, on a broad range of antibiotic classes against pathogenic E. coli, Proteus mirabilis, and P. aeruginosa. Evidence that cPAC acts by repressing two antibiotic resistance mechanisms, selective membrane permeability and multidrug efflux pumps, is presented. Failure of cPAC to potentiate antibiotics against efflux pump‐defective mutants demonstrates that efflux interference is essential for potentiation. The use of cPAC to potentiate antibiotics and mitigate the development of resistance could improve treatment outcomes and help combat the growing threat of antibiotic resistance.


Introduction
The global spread of antibiotic resistance is undermining decades of progress in fighting bacterial infections. Due to the overuse of antibiotics in medicine and agriculture, we are on the cusp of returning to a pre-antibiotic era in which minor infections can once again become deadly. Countering the fall in antibiotic efficacy by improving the effectiveness of currently available antibiotics is therefore an important goal. Antibiotic efficacy is limited by the expression of intrinsic tolerance mechanisms such as production of antibiotic-tolerant and/or persister can prevent bacterial infections, [20][21][22][23][24][25] and one reports the combined effect of bulk cranberry derivatives (devoid of the proanthocyanidin fraction) and β-lactam antibiotics against a Gram-positive bacterium. [26] It has also been suggested that cPAC can inhibit biofilm formation and potentiate gentamicin against P. aeruginosa; [17] however, the potential for cPAC to interfere with the evolution of resistance to antibiotics or rescue the effectiveness of antibiotics has never been investigated. Furthermore, no studies have explored the activity spectrum or the molecular mode of action of specific cranberry-derived fractions such as proanthocyanidins for the treatment of bacterial infections. Here, we report the ability of cPAC to thwart the evolution of resistance to tetracycline in E. coli and P. aeruginosa, and demonstrate the broad spectrum, antibiotic-potentiating activity of cPAC against various pathogenic Gram-negative bacteria, both in vitro and in vivo. We also show that cPAC is able to repress two important intrinsic antibiotic resistance mechanisms: selective membrane permeability and multidrug efflux pump activity.

cPAC Potentiates a Broad Range of Antibiotics In Vitro
We conducted screening assays with cPAC in combination with several classes of clinically approved antibiotics belonging to the World Health Organization's list of essential medications. [27] Checkerboard microdilution analysis was performed using pathogenic strains of E. coli, P. mirabilis, and P. aeruginosa. We determined whether cPAC was able to potentiate the effectiveness of antibiotics by calculating the fractional inhibitory concentration index (FICI). Figure 1A-C shows positive and/ or negative interactions between cPAC and antibiotics against bacterial pathogens. FICI values of ≤0.5 (indicated by the gray zone) demonstrate that cPAC potentiated the effectiveness of sulfamethoxazole (SMX), nitrofurantoin (NIT), gentamicin (GEN), kanamycin (KAN), tetracycline (TET), and azithromycin (AZT) to inhibit the growth of E. coli CFT073, P. mirabilis HI4320, and P. aeruginosa PA14 using up to 98% less antibiotic than that required in the absence of cPAC. cPAC also potentiated trimethoprim (TMP) and fosfomycin (FOS) activities to inhibit the growth of P. mirabilis HI4320; 81% and 98% less antibiotic were needed, respectively, than in the absence of cPAC. In the case of P. aeruginosa strain PAO1, cPAC enhanced the efficiency of the antibiotics SMX, FOS, NIT, GEN, KAN, and AZT ( Figure S1A, Supporting Information). The fact that cPAC potentiates a given antibiotic against one strain but not another (e.g., cPAC potentiates FOS against P. mirabilis HI4320 but not against P. aeruginosa PA14 or E. coli CFT073) provides evidence that the effect is specific and that cPAC is not simply inactivating the antibiotic. It is not surprising that the FICI values for the two P. aeruginosa strains were different for some antibiotics as they had different MIC antibiotic values. [28] At concentrations required to potentiate antibiotic efficacy, cPAC alone had no detectable growth inhibition activity against all four pathogenic strains (Figures S1B and S2A-C, Supporting Information). Given the ability of cPAC to potentiate TMP or SMX alone, we investigated the interaction of cPAC with co-trimoxazole (the combination of SMX and TMP, commonly used to treat urinary tract infections and bacterial dysentery [27] ). cPAC enhanced the synergistic efficacy of cotrimoxazole, reducing the MIC up to 64-fold against P. mirabilis HI4320. In the case of P. aeruginosa PA14, combination of cPAC with co-trimoxazole decreased the MIC by 32-fold, which is significantly more effective than the potentiating combinations of cPAC with TMP or SMX alone ( Figure S3, Supporting Information). The fact that cPAC potentiates antibiotics, but does not act as an antibiotic on its own, suggests that treatment with cPAC is unlikely to create selective pressure for the evolution of resistance.

cPAC Prevents Re-Activation of Antibiotic-Exposed Cells
To investigate the inhibitory activity of cPAC against antibioticexposed bacteria, a modified disk-diffusion test was performed. Figure 2A shows that following treatment with the bacteriostatic antibiotic tetracycline (Step 1: application of TET antibiotic disk), bacteria in the growth-inhibition zone were able to recover when a new disk impregnated with glucose replaced the TET disk (at Step 2). An analysis of bacterial re-activation based on the occurrence of colonies inside a typical inhibition/clear zone shows that the degree to which cells "re-activate" to form colonies differs depending on the presence or absence of cPAC. Replacement with a glucose-only disk (at Step 2) enhanced the re-activation of cells (i.e., bacterial lawn in previous inhibition zone), while a disk with a combination of cPAC and glucose showed no re-activation of antibiotic-exposed cells. There were no colonies observed close to the cPAC-only disk, and the size of the clear zone with the cPAC-only disk (at Step 2) was similar to the TET disk (Step 1) and the glucose+cPAC disk (Step 2). Since cPAC alone did not inhibit bacterial growth of this, or any other tested strains (Figures S1B,C and S2A-F, Supporting Information), it is probable that the clear zones around the cPAC-only and the glucose+cPAC disks result from synergy of cPAC with TET remnants, which is in agreement with a slightly smaller diameter, since lower concentrations of cPAC and TET should be present at the distal edge of the zone. Similar effects were observed with minocycline (MIN; Figure S4, Supporting Information) against E. coli CFT073. These results suggest that cPAC can prolong the efficacy of remnant antibiotic against antibiotic-exposed cells even after treatment has ceased.

cPAC Thwarts Evolution of Antibiotic Resistance
To understand the role of cPAC in effectively inhibiting the reactivation of antibiotic-exposed cells, we analyzed the ability of cPAC to suppress evolution of resistance in E. coli CFT073 and P. aeruginosa PA14. As shown in Figure 2B,C, sequential passaging on TET alone for 21 days resulted in 128-fold and 32-fold increases in MIC for E. coli CFT073 and P. aeruginosa PA14, respectively, while cPAC prevented the evolution of resistance in both strains when co-administered with TET. cPAC alone did not promote resistance in either strain. This result shows that cPAC can suppress the emergence of antibiotic resistance in bacteria. www.advancedscience.com

cPAC Potentiates In Vivo Activity of SMX
Our next goal was to investigate the potential of cPAC to enhance the efficacy of antibiotics against bacterial infections in vivo. To this end, we used the model host Drosophila melanogaster infected with P. aeruginosa PA14, in which cPAC or SMX was administered alone or in combination. The median survival of the insects following infection was 138 h in the absence of treatment, but more than 225 h when cPAC and SMX were combined ( Figure 3A). Survival of flies following combination therapy was significantly (χ 2 = 3.88, df = 1, P < 0.05) higher than survival with SMX treatment alone. The median survival time of infected D. melanogaster treated with cPAC alone or SMX alone was 202 and 178 h, respectively ( Figure 3A), which is not significantly (P > 0.05) different from the survival of untreated flies infected with P. aeruginosa PA14. The survival of uninfected D. melanogaster was similar to treatment with cPAC or SMX alone ( Figure S5A, Supporting Information), which indicates that cPAC (at 50 µg mL −1 ) or SMX (at 256 µg mL −1 ) alone is safe for this animal.
To confirm that cPAC is able to potentiate antibiotics in more than one host, we also used the greater wax moth (Galleria mellonella) larvae killing model, in which cPAC or SMX was administered alone, or in combination, to larvae infected with a lethal dose of P. aeruginosa PA14. The median survival of infected and untreated larvae was 21 h, but increased to 47 h with the  cPAC and SMX combination treatment. This is a significantly (χ 2 = 14.3, df = 1, P < 0.005) longer survival time than the 36 h median survival time with SMX treatment alone ( Figure 3B). The median survival of infected larvae with cPAC treatment alone was similar to the treatment of SMX alone ( Figure 3B), which is in contrast to cPAC's inability to act as an antibiotic at this concentration in vitro, or in the fly feeding model. As with the D. melanogaster feeding assay, there was no difference in survival curves of uninfected G. mellonella larvae with or without treatment of cPAC or SMX alone ( Figure S5B, Supporting Information). These results confirm that cPAC potentiates the activity of SMX in vivo, at the tested concentration of 50 µg mL −1 cPAC.
Adv. Sci. 2019, 6, 1802333 Figure 2. Synergistic effect of cPAC with TET for the inhibition of growth re-activation of antibiotic-exposed cells and prevention of the evolution of resistance. A) Detection of growth re-activation of antibiotic-exposed E. coli CFT073 cells using a modified disk-diffusion assay.
Step 1: a TET antibiotic disk was placed on top of MHB-II agar. The dashed lines mark the diameter of the clear zone surrounding the TET disk.
Step 2: the TET disks are replaced with a glucose alone, cPAC+glucose, or cPAC alone disk on the MHB-II agar plate. The diameter of the clear zone (CZ) and no colony formation inside the clear zone surrounding the cPAC+glucose disk indicate no re-activation of antibiotic-exposed cells after disk replacement at Step 2 and colonies inside the inhibition zone (indicated by black arrow) surrounding the glucose disk indicate re-activation of antibiotic-exposed cells. The cPAC-only disk prevented re-activation of growth, most likely because of synergy with TET remnants. The images shown are only representative images of three independent experiments. Disk diameter: 6 mm. Emergence of antibiotic resistance in B) E. coli CFT073 and C) P. aeruginosa PA14 during 21 serial passages in the presence of sub-MIC levels of TET compared to 400 µg mL −1 cPAC alone or its combination with 100 µg mL −1 cPAC.

cPAC in Combination with an Antibiotic Impairs Biofilm Formation
To explore whether cPAC can impair the formation of biofilms during antibiotic treatment, monoculture bacterial biofilms were grown in microtiter plates in the presence of different concentrations of cPAC and SMX. As shown in Figure 4A-C, cPAC decreased biofilm formation of E. coli CFT073, P. mirabilis HI4320, and P. aeruginosa PA14. cPAC alone at 100 µg mL −1 showed significant (P < 0.05) inhibition of monoculture biofilms, which is consistent with other studies showing that cPAC or cranberry extracts can decrease bacterial adhesion. [8][9][10][11] Importantly, cPAC in combination with SMX had a significant (P < 0.01) inhibitory effect on the development of biofilms in a dose-dependent manner (Figure 4A-C; Figure S1D, Supporting Information). Untreated control biofilms were composed of viable cells ( Figure 4D, in green) attached to the surface and forming dense microcolonies ( Figure 4Di). SMX alone at tested concentrations had minimal effect on biofilms. However, treatment with cPAC and SMX in combination resulted in a decrease in the total density of attached biomass and viability when compared to untreated biofilms ( Figure 4E,F). Interestingly, the majority of the bacterial biomass was dead when treated with 512 µg mL −1 SMX in combination with cPAC ( Figure 4Dviii,ix). If we presume that the few cells left alive in the presence of 512 µg mL −1 SMX and 50 µg mL −1 cPAC are dormant, capable of reactivating when the antibiotic threat has passed and forming a new biofilm, then increasing the concentration of cPAC to 100 µg mL −1 was sufficient to eradicate these surviving cells. Because a combination of cPAC and SMX is more effective against biofilms than either compound is alone, these observations demonstrate that cPAC acts synergistically with SMX to reduce the ability of bacteria to form biofilms, and supports the hypothesis that cPAC contributes to the impairment of dormant antibiotic-exposed cells.

Mechanisms by Which cPAC Potentiates Antibiotic Activity
To identify the mechanism(s) of action by which cPAC potentiates antibiotic activity, we quantified the changes in bacterial cell outer membrane permeability using 1-N-phenylnapthylamine (NPN) as an indicator, which revealed that cPAC increases outer membrane permeability of bacterial cells (Figure 5A-C). Investigations into efflux pump activity using ethidium bromide (EtBr) as a fluorescent indicator substrate showed that, in contrast to the decay in fluorescence observed in untreated cells ( Figure 5D-F), cells treated with cPAC, or the efflux pump inhibitor carbonyl cyanide m-chlorophenylhydrazone (CCCP), remained fluorescent over time due to a failure to pump out EtBr ( Figure 5D-F). These observations demonstrate that cPAC is able to inhibit multidrug resistance efflux pumps. To further investigate the interaction between cPAC and multidrug resistance efflux pumps, we employed a systematic checkerboard MIC analysis of P. aeruginosa PA14 efflux pump mutants. The overexpression of efflux pump systems causes increased resistance to antibiotics compared to wild-type isolates while disruption of efflux pump protein components is associated with increased intracellular antibiotic accumulation and antibiotic susceptibility. [29] We show here that the ability of cPAC to potentiate TET activity correlates with efflux pump activity. The mutant strains with nonfunctioning efflux pumps were more susceptible to TET compared to the wild-type strain, such that addition of cPAC provided no further benefit for the potency of TET. Therefore, potentiation between cPAC and TET was not observed in efflux pump-nonfunctioning mutants ( Figure 5G). This lack of TET potentiation correlates well with TET potentiation by cPAC in the case of wild type P. aeruginosa PA14, which has functioning efflux pumps ( Figure 1C). Interestingly, cPAC caused no significant damage to bacterial cell membranes ( Figure 5H), compared to significant (P < 0.05) membrane disruption by cetyl trimethylammonium bromide (CTAB), a cell  membrane-disrupting agent, suggesting that outer membrane permeabilization and target specific efflux pump inhibition are achieved by cPAC without altering cell membrane integrity.

cPAC Interacts with Efflux Pump Components In Silico
To understand how cPAC is able to inhibit the activity of efflux pumps, we performed in silico docking analyses using the efflux pump protein complexes AcrAB-TolC of E. coli, and MexAB-OprM of P. aeruginosa, with the A-type dimeric cPAC molecule as the test ligand. Ligand-binding domains of AcrAB-TolC and MexAB-OprM efflux pump components exhibit sufficient space to accommodate the cPAC molecule with an average volume of 497.1 Å 3 (Table S1, Supporting Information). We tested structural in silico interactions of cPAC with the exit duct, adapter, and transporter components of the efflux pumps ( Figure 6A; Figure S6A  . Effect of cPAC alone and in combination with SMX on biofilm formation of A) E. coli CFT073, B) P. mirabilis HI4320, and C) P. aeruginosa PA14. The graphs present normalized biofilm levels (OD 570 /cell OD 600 ) at subinhibitory concentrations of SMX. Statistically significant differences are indicated for each sample treated with cPAC and SMX compared to the control (sample treated with the corresponding concentration of SMX alone) (**, P < 0.01; *, P < 0.05; two-way ANOVA) and for samples treated with cPAC plus SMX compared to the sample treated with the same concentration of cPAC without SMX (*, P < 0.05; two-way ANOVA). D) Confocal microscopy images of P. aeruginosa PA14 biofilms grown with or without cPAC and/ or SMX at sub-MICs. Green color cells represent viable cells with intact membranes and red color cells are dead cells. Each representative image was selected from experiments performed in triplicates. E) Biofilm biomass as a percentage of untreated biomass, and F) percentages of live cells present in each treatment were quantified from confocal microscopy images. In all confocal microscopy images, red and green axis lengths are 100 µm, and the blue z-axis length is 10 µm.  exit duct, and proper functioning of the entire efflux pump assembly ( Figure 6A). cPAC showed the highest affinity for AcrB and MexB efflux pump components, with −9.9 and −9.6 kcal mol −1 binding energies, respectively (Table S1, Supporting Information).
A diverse range of antibiotics are pumped out by AcrAB-TolC. However, only the co-crystal structures of MIN and doxorubicin binding to the active site of AcrB are reported. [30] To compare the binding of cPAC to that of a well-characterized efflux pump substrate, we docked cPAC to the distal binding pocket of the AcrB periplasmic porter domain co-crystallized with MIN, and found that cPAC and MIN bound to the same furrow of the distal substrate binding pocket ( Figure 6B). We separately examined the predicted docking of the AcrB substrates FOS, TMP, NIT, TET, SMX, and GEN in the absence and presence of cPAC ( Figure 6C,D). All tested substrates bound to the AcrB-cPAC complex, specifically, in the same furrow of the distal binding pocket in which MIN binds ( Figure 6C). Interestingly, the binding position of FOS and TMP was unaffected by the presence of cPAC ( Figure 6D), which is in agreement with the inability of cPAC to potentiate FOS or TMP against E. coli in vitro ( Figure 1A). In contrast, we observed different binding conformations of the potentiated antibiotics NIT, TET, SMX, and GEN in the presence of cPAC compared to that in the absence of cPAC ( Figure 6C,D). This suggests that binding of cPAC in the distal binding pocket leads to efflux inhibition by interfering with the preferred binding position of the potentiated antibiotics. Finally, we confirmed that the binding conformation of cPAC in the distal binding pocket of AcrB is similar to that of known efflux pump inhibitors ( Figure S7A-F, Supporting Information), which supports the hypothesis that cPAC acts as a potent efflux pump inhibitor.

Discussion
Tetracycline resistance is a well-documented phenomenon caused by efflux pump activity and the evolution of acquired antibiotic resistance. [31] We discovered that in the presence of cPAC, the acquisition of resistance in E. coli and P. aeruginosa following TET treatment is completely abrogated. cPAC's ability to interfere with intrinsic resistance mechanisms is therefore able to suppress the typically inevitable, long-term evolution of acquired antibiotic resistance.
We have shown that cPAC potentiates the in vitro activity of a range of antibiotic classes against the opportunistic human pathogens E. coli, P. mirabilis, and P. aeruginosa. This potentiation also occurs in vivo, at least with the antibiotic SMX. cPAC is an especially promising antibiotic-potentiating agent because it does not exhibit antimicrobial activity of its own. Agents that do not negatively affect the viability of bacterial pathogens are less likely to promote resistance than conventional antibiotics [32][33][34] and are, therefore, especially well suited for combination treatment with antibiotics. Accordingly, we have not observed any evolution of resistance to cPAC.
Biofilms can lead to chronic bacterial infections and are commonly associated with antibiotic treatment failure. [35,36] The presence of persister and antibiotic-tolerant cells is closely linked to biofilm formation, [5] and also plays an important role in the recalcitrance of chronic infections. The antiquorum sensing activity of a cranberry proanthocyanidin fraction alone [19] and in combination with ciprofloxacin [37] has been reported. As quorum sensing is required for normal biofilm formation, this agrees with our observations that cPAC in combination with antibiotic substantially represses biofilm formation. Furthermore, we propose that it is the ability of cPAC to target not only actively metabolizing cells, but also dormant antibiotic-exposed cells that enable cPAC to potentiate antibiotic efficacy against biofilms.
The intrinsic antibiotic resistance mechanisms of bacterial cells are naturally occurring phenomena found in bacterial species that complicate antibiotic treatments. For instance, P. mirabilis has intrinsic resistance to NIT and TET, and P. aeruginosa has intrinsic resistance to multiple antibiotic classes including aminoglycosides, β-lactams, quinolones, and polymyxins. [38,39] The intrinsic antibiotic resistance mechanisms include selective outer membrane permeability, poor antibiotic transport, and active multidrug efflux. [40,41] Among resistancenodulation-division family efflux pumps, AcrAB-TolC [42,43] and MexAB-OprM [44] are well known for their importance in bacterial survival and intrinsic antibiotic resistance. Our in silico analysis predicts that the A-type dimeric cPAC molecule (which is the most common terminating unit of the cPAC fraction [45] ) can occupy the ligand-binding pocket of the AcrB transporter in E. coli and MexB transporter of P. aeruginosa. [43,44,46] Molecular docking calculations indicate that of all the antibiotics tested against E. coli, the two nonpotentiated antibiotics adopt the same position in the distal binding pocket in the presence and absence of cPAC. In contrast, the potentiated antibiotics are predicted to take a different binding position in the presence of cPAC, which would hinder their efflux from the cell. Exposure to cPAC enhanced membrane permeabilization and decreased efflux pump activity in all tested wild-type bacteria in vitro. The failure of cPAC to potentiate antibiotic activity in efflux pump-defective mutants supports a model where the specific effect cPAC has on efflux is essential for the observed antibiotic potentiating activity. Inactivation of efflux pump activity has previously been shown to have a negative impact on biofilm formation [47] and pathogenicity [48] in Gram-negative bacteria. This is consistent with our observations that in the presence of cPAC, biofilm formation is decreased. Overall, this study provides a proof of concept and a starting point for investigating the molecular mechanism of the reported efflux pump inhibition in bacteria by cPAC.
Insect animal models have a relatively evolved antimicrobial defense system and are thus often used to generate information relevant to the mammalian infection process. [49,50] In both infection models used in this study, cPAC at a dose of 50 µg mL −1 was sufficient to potentiate the activity of the tested antibiotic and significantly increase survival rates of the animals during infection with P. aeruginosa. These in vivo studies provide a promising outlook for the potential future development of cPAC as an antibiotic-potentiating agent in higher organisms; however, a few studies report on the safety of cPAC to human cells [51] or its bioavailability and the rate of clearance in animals. [52] Thus, further work is needed to verify the efficacy and safety of the combination treatment in an in vivo mammalian (e.g., mouse) model. Encouragingly,

www.advancedscience.com
Adv. Sci. 2019, 6, 1802333 our data show that cPAC is able to enhance the efficacy of a broad spectrum of antibiotics. The ability to potentiate the action of antibiotics in a patient could improve treatment outcomes and hinder the emergence of antibiotic-resistant infections.
Determination of MICs: MICs were determined by preparing twofold serial dilutions of cPAC, and antibiotics in MHB-II broth. A range of concentrations of the antibiotics (0.0003-1024 µg mL −1 ) was chosen due to their known potency against all four bacterial strains. Bacterial growth was assessed by i) monitoring the cell growth (observed as a pellet and turbidity) in the wells [58] ii) monitoring the optical density of the cell suspension in each well at 600 nm (OD 600 ), and iii) using the resazurin microtiter plate assay. [59] Additional details can be found in the Supporting Information.
In Vivo Infection Assay Using D. Melanogaster Flies: Fruit flies (D. melanogaster) were infected orally in fly feeding assay in which flies were anesthetized, starved of food and water for several hours, and separated into vials containing filter paper disks inoculated with freshly grown P. aeruginosa PA14, as well as 5% sucrose agar (sterile) with and without 50 µg mL −1 cPAC alone or in combination with 256 µg mL −1 SMX. Post infection mortality of flies was monitored daily for 14 days, with each treatment tested twice in triplicate. Additional details can be found in the Supporting Information.
In Vivo Infection Assay Using Galleria Mellonella Larvae: Twenty G. mellonella larvae were injected with active cultures of P. aeruginosa per treatment. All injected larvae were incubated in Petri dishes at 28 °C under 30% relative humidity in the dark, and the number of dead larvae was scored daily post infection. A larva was considered dead when it displayed no vital signs in response to touch, followed by increased melanization. Additional details are given in the Supporting Information.
Biofilm Assays: Biofilm formation was quantified using the standard microtiter plate model [60] and crystal violet staining, the details of which are described in Appendix S1 (Supporting Information). For biofilm imaging and analysis, biofilms were grown for 16 h at 37 °C under static conditions, after which planktonic cells were removed by rinsing. The cell-membrane impermeable fluorescent stain TOTO-1 (Thermo Fisher) and membrane permeable stain SYTO 60 (Thermo Fisher) were added to final working concentrations of 2 × 10 −6 and 10 × 10 −6 m, respectively. [61] Biofilms were imaged as z-stacks using a 63× objective on a Zeiss 800 confocal laser scanning microscope and rendered such that living cells are colored green, and dead cells are colored red as described in the Supporting Information.
Modified Disk-Diffusion Assay: A modified disk-diffusion assay [62] was used to detect recovery of antibiotic-exposed cells. This modified diskdiffusion assay was conducted in two steps, where Step 1 was similar to classical disk-diffusion assay and Step 2 was slightly modified. Briefly, an overnight bacterial culture (MHB-II broth, 37 °C, 200 rpm) was diluted into fresh MHB-II broth to 10 6 CFU mL −1 and plated on MHB-II agar plate.
Step 1: commercially available TET 30 µg or MIN 30 µg antibiotic disk (Thermo Scientific Oxoid, Fisher Scientific, Canada) was placed on top of the inoculated agar surface and the plate was incubated at 37 °C for 18 h. Custom disks were prepared using sterile blank disks (Thermo Scientific Oxoid, Fisher Scientific, Canada) supplemented with 400 µg glucose or 400 µg cPAC with or without 400 µg glucose and left to dry at room temperature.
Step 2: The antibiotic disk was carefully replaced with custom disks containing glucose or cPAC or their combination, and the plate was incubated at 37 °C for an additional 18 h.
Emergence of Resistance Analysis: For characterization of in vitro resistance evolution by standard sequential passaging technique, [63,64] E. coli CFT073 and P. aeruginosa PA14 cells were grown to exponential phase in MHB-II at 37 °C. MICs were determined by preparing twofold serial dilutions in 96-well microtiter plates. Each well was inoculated with the desired bacterial strain, and the plate was incubated at 37 °C for 18 h under static conditions. Bacterial growth was assessed by i) monitoring the cell growth (observed as a pellet and turbidity) in the wells [58] and ii) monitoring the optical density of the cell suspension in each well at 600 nm (OD 600nm ). Bacterial suspension from sub-MIC (0.5 × MIC) of TET or cPAC or cPAC+TET combination was used to prepare the inoculum for the next day MIC experiment by diluting it to a final concentration of ≈10 6 CFU mL −1 in MHB-II. This was repeated for 21 passages, and the ratio of the MIC obtained during each day relative to the MIC at 0 day (first time exposure) was determined. The data were expressed as relative fold increase in MIC with each day or passage. Experiments were performed with biological replicates.
Membrane Permeabilization and Membrane Integrity Assays: The outer membrane permeabilization activity of cPAC was determined by the NPN (Sigma-Aldrich, Canada) assay, [65] with some modifications. Overnight bacterial cultures were diluted 1:1 in MHB-II medium to a final volume of 10 mL, with or without supplementation of cPAC or GEN (positive control), and grown to an OD 600 of 0.5-0.6 (37 °C, 200 rpm). The cells were harvested, washed, resuspended 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer containing 1 × 10 −3 m N-ethylmaleimide (NEM; Sigma-Aldrich, Canada). Aliquots were mixed with NPN to a final concentration of 10 × 10 −6 m (in cell suspension), and fluorescence was measured using the microplate reader (excitation: 350 nm; emission: 420 nm). The BacLight kit (L-13152; Invitrogen, Life Technologies Inc., Canada) was used to assess cell membrane damage, [66] with fluorescence readings from samples normalized to the values obtained from the untreated control to determine the ratio of membrane-compromised cells to cells with intact membranes. Details can be found in the Supporting Information.
EtBr Efflux Assay: To assess the effect of cPAC on the inhibition of the proton motive force-driven multidrug efflux pump, an EtBr efflux assay was performed. [67] An overnight-grown culture of each strain was diluted 1:100 in MHB-II broth to a final volume of 10 mL and was grown to an OD 600 of 0.8-1.0 (at 37 °C, 150 rpm). Cells were loaded in polystyrene microcentrifuge tubes (2 mL) and mixed with 5 × 10 −6 m EtBr and cPACs at 25% of their MIC or a 100 × 10 −6 m concentration of the proton conductor CCCP (Sigma-Aldrich Canada) as a positive control. Replica tubes that did not receive cPACs or proton conductor served as negative controls. The tubes were incubated for 1 h (37 °C, 150 rpm). The inoculum then was adjusted to an OD 600 of 0.4 with MHB-II broth containing 5 × 10 −6 m EtBr, and 2 mL of aliquot of this mixture was pelleted (5000 × g, 10 min, 4 °C). The pellets were incubated on ice immediately, resuspended in 1 mL of MHB-II, and aliquoted (200 µL) into a polystyrene 96-well, black, clear-bottom plate (Corning, Fisher Scientific Canada). EtBr efflux from the cells was monitored at room temperature using the microplate reader (excitation wavelength 530 nm; emission wavelength, 600 nm). Readings were taken at 5 min intervals for 1 h to monitor efflux pump activity. The background fluorescence of the medium was subtracted from all measurements, and the assay was repeated independently in triplicate.
In Silico Docking Analysis: 3DLigandSite server was used to predict ligand binding sites on targeted protein structures using an automated method by searching a structural library to identify homologous structures with bound ligands. [68] These bound ligands were superimposed onto the targeted protein structure to predict a ligandbinding site. In silico docking was performed using the Autodock Vina tool without the incorporation of water molecules. [69] In silico docking analysis by Autodock Vina was compared with SwissDock web service (http://www.swissdock.ch/) to confirm the accuracy and robustness of predicted docking complexes. The ligand docking methods are described in the Supporting Information.
Statistical Analysis: Where indicated, a two-tailed Student's t-test (P < 0.05) was used to determine whether the presence of cPAC resulted in a significant difference compared to levels for the control. A two-way analysis of variance (ANOVA), followed by Sidak's multiple comparison, was used for biofilm assays to analyze statistical significance of the difference in biomass. Fruit fly survival curves were prepared using GraphPad Prism 6 (GraphPad Software, Inc., San Diego, CA) to perform a statistical log-rank (Mantel-Cox) test. Throughout the text, all of the changes (increase or decrease) reported were statistically significant. Statistically nonsignificant values were not mentioned in the text.
Full Methods: Detailed procedures of all methods are available in the Supporting Information.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.