Novel small molecules affecting cell membrane as potential therapeutics for avian pathogenic Escherichia coli

Avian pathogenic Escherichia coli (APEC), a most common bacterial pathogen of poultry, causes multiple extra-intestinal diseases in poultry which results in significant economic losses to the poultry industry worldwide. In addition, APEC are a subgroup of extra-intestinal pathogenic E. coli (ExPEC), and APEC contaminated poultry products are a potential source of foodborne ExPEC infections to humans and transfer of antimicrobial resistant genes. The emergence of multi-drug resistant APEC strains and the limited efficacy of vaccines necessitate novel APEC control approaches. Here, we screened a small molecule (SM) library and identified 11 SMs bactericidal to APEC. The identified SMs were effective against multiple APEC serotypes, biofilm embedded APEC, antimicrobials resistant APECs, and other pathogenic E. coli strains. Microscopy revealed that these SMs affect the APEC cell membrane. Exposure of SMs to APEC revealed no resistance. Most SMs showed low toxicity towards chicken and human cells and reduced the intracellular APEC load. Treatment with most SMs extended the wax moth larval survival and reduced the intra-larval APEC load. Our studies could facilitate the development of antimicrobial therapeutics for the effective management of APEC infections in poultry as well as other E. coli related foodborne zoonosis, including APEC related ExPEC infections in humans.

Primary screening. To identify the APEC growth inhibitors, SM library was screened against APEC O78 which is one of the most frequently isolated APEC serotypes from avian colibacillosis cases 2 . One microlitre SMs (final concentration of 100 µM) were added using a slotted pin tool (V and P Scientific, San Diego, CA, USA) to the wells of the 96-well plate containing 100 µL of overnight grown 0.05 OD 600 (7 × 10 7 CFU/mL) adjusted APEC culture. Controls (four replicates/plate) containing 1 µL of 100% DMSO (final concentration of 1%), 1 µL chloramphenicol (CHL, #C0378 Sigma-Aldrich) (20 µg/mL), 1 µL kanamycin (KAN, #60615 Sigma-Aldrich) (50 µg/mL), and 100 µL of M63 media were included. To determine the effect of DMSO on APEC growth, antimicrobial activity of DMSO was determined at different concentrations (1% to 32%) in a separate experiment. Plate was then incubated at 37 °C for 12 h in Sunrise -Absorbance microplate reader (Tecan Group Ltd. San Jose, CA) with kinetic OD 600 measurement every 30 mins. The quality of screening was assessed by calculating the Z′-score as described previously 18 . The growth inhibition of APEC was calculated by using the formula as previously described 13,14 . The SMs inhibiting at least 80% of the APEC growth were selected as primary hits. Culture from wells considered as hits were subsequently subcultured on LB agar plate to determine the bactericidal effect (no APEC recovered on plating following exposure to SM); these cidal SMs were selected for further studies.
MIC and MBC determination. SMs were two-fold serially diluted from 200 µM to 6.25 µM to determine their MIC and MBC as described previously 19 . One microlitre SM of each concentration was transferred to each well of a 96-well plate containing 100 µL of the 0.05 OD 600 adjusted APEC O78 culture in M63 media. Growth was monitored in Sunrise -Absorbance microplate reader as described above. MIC was indicated by lowest concentration of SM with non-elevated OD 600 measurement. MBC was determined by absence of APEC growth on LB agar plate following subculture. In addition, MIC and MBC of cidal SMs were also determined as described above against multiple APEC serotypes (O1, O2, O8, O15, O18, O35, O109, and O115) that are commonly associated with colibacillosis cases 2 to determine their spectrum of activity. Two independent experiments were conducted. The activity of cidal SMs were also tested at 100 µM in M63 media against Shiga toxin-producing (STEC) O157 and O26 strains (Table S2).
Effect against beneficial microbes. SMs were screened against different beneficial microbes to determine their specificity as described previously 13 . The beneficial microbes used in this study along with their culture requirements are listed in Table S2. SMs were added at 100 µM to 100 µL of 0.05 OD 600 adjusted bacterial cultures in specific growth media in 96-well plate, and plate was incubated under indicated conditions. The specific growth media and conditions required for beneficial microbes limited the use of minimal media. Following incubation, endpoint OD 600 was measured and cultures from the wells with non-elevated OD 600 were plated on selective agar plates to determine the bactericidal effect.
Effect against biofilm embedded APEC. The effect of cidal SMs against biofilm embedded APEC was determined using MBEC High-throughput (HTP) assay (Innovotech Inc., AB, Canada) 22 . Briefly, 150 µL of 0.05 OD 600 adjusted APEC O78 culture was aliquoted into each well of the MBEC device containing polystyrene pegs and incubated at 37 °C for 36 h in LB media under stationary condition. After biofilm formation, the pegs were washed to remove loosely adherent planktonic bacteria, transferred to new 96-well plate, and challenged with different concentrations of SMs (0.5X, 1X, 2X, 4X, and 8X MIC) in 200 µL M63 media. The plate was incubated in the dark for 18 h at 37 °C with rotation at 110 rpm. The DMSO (1%) and M63 media were used as positive and negative controls, respectively. Following incubation, MIC of SMs in challenged plate was recorded. The SMs exposed pegs were then transferred to a new 96-well plate containing PBS and sonicated for 30 mins (Aquasonic ultrasonic cleaner, VWR) to disrupt the biofilm. The sonicated suspensions were ten-fold serially diluted and plated on LB agar plate. Biofilm embedded APEC bacteria were enumerated and minimum biofilm eradication concentration (MBEC) of SMs were determined as described previously 22 . Two independent experiments were conducted.
Antimicrobial resistance studies. To evaluate APEC O78 potential to acquire resistance against cidal SMs, single step (lethal dose) and sequential passage (sub-lethal dose) resistance assays were performed in M63 media as described previously 13,14,19 . Briefly, for single step resistance assay, SMs were mixed with 1.5 mL of molten M63 agar at a final concentration of 2X MBC and transferred to wells of a sterile 24-well plate. Concentration of 2X MBC was used since it has been previously reported that the MIC/MBC of antimicrobials are higher in solid media compared to liquid media 23,24 . Fifty microlitres of overnight grown APEC O78 (~10 9 CFU) culture was plated over the solidified SM amended M63 agar. The plate was incubated for 15 days in the dark at 37 °C. After 15 days, any colonies that grew on the agar were assessed for resistance by determining the MIC and MBC as described above.
For sequential passage resistance assay, SMs were added at a final concentration of 0.75X MIC (concentration that allows at least 70% growth inhibition) to the 100 µL of the 0.05 OD 600 adjusted APEC O78 culture in M63 media in a 96-well plate. The plate was then incubated in the dark at 37 °C with shaking at 150 rpm for 18 h. After the first incubation, bacterial pellet was resuspended in a fresh M63 media amended with 0.75X MIC of each SM and grown as above. This procedure was repeated 14 times. Following 15 passages, susceptibility (MIC and MBC) of APEC to SMs was determined as described above. DMSO (1%), 20 µg/mL CHL, 50 µg/mL KAN, and M63 media were included as controls in both the assays. Experiments were conducted in duplicate wells.
Confocal and scanning electron microscopy. Confocal microscopy was used for bacterial cytological profiling (BCP) to identify the cellular pathways targeted by SMs as described previously 25 . Briefly, 100 µL of logarithmic-phase APEC O78 culture grown in M63 media was treated with 2X MBC of SMs and incubated at 37 °C for 2 h with shaking at 200 rpm. After incubation, treated cultures were centrifuged, washed, and resuspended in 100 µL PBS. FM4-64 (#T13320 Molecular Probes/Invitrogen) (1 µg/mL) and SYTO-9 (#S34854 Molecular Probes/Invitrogen) (5 µM) were added to the bacterial cultures and incubated for 45 mins at room temperature with shaking at 150 rpm. Cultures were then centrifuged, washed, and resuspended in PBS to 1/10 th volume of the original cultures. Three microlitres of concentrated bacterial cultures were transferred onto an agarose pad containing 1.2% agarose and 20% LB medium. Microscopy was performed using Leica TCS SP6 confocal scanning microscope (Excitation/emission (nm); FM4-64 (515/640), SYTO-9 (485/498) and images were analyzed using ImageJV1.50.
The SMs treated APEC O78 cultures prepared above were also processed for scanning electron microscopy (SEM) as described previously 26 . SEM was performed for representative SMs (possessing similar structure and BCP). Briefly, one volume of bacterial culture was mixed with one volume of fixative (3% glutaraldehyde, 1% paraformaldehyde in 0.1 M potassium phosphate buffer, pH 7.2), and incubated at 4 °C overnight. The fixed bacterial cells were then centrifuged for 5 mins at 1,200 × g, washed twice with PBS, and resuspended in 1% osmium tetroxide for 1 h at room temperature in the dark, followed by serial dehydration of the sample in ethanol and platinum splatter-coating. Visualization and imaging of the sample was performed using a Hitachi S-4700 scanning electron microscope.

Cytotoxicity of SMs to chicken and human cells.
The cytotoxicity of cidal SMs to human Caco-2 and chicken HD11 cells were evaluated using Pierce Lactate Dehydrogenase (LDH) Cytotoxicity Assay Kit (Pierce, Thermo Scientific, Rockford, IL, USA) as previously described 13,14 . Cytotoxicity was measured at OD 680 nm and 490 nm after exposing cultured epithelial and macrophage cells to 200 µM of SMs for 24 h. 10X LDH provided in the kit was used as positive control. Two independent experiments with triplicate wells in each experiment were conducted.
Hemolytic activity of SMs to chicken RBCs. The hemolytic activity of cidal SMs to chicken RBCs was evaluated as previously described 13 . Hemolysis was determined at OD 540 nm after exposing 10% RBCs suspension to 200 µM of SMs for 1 h. 0.1% Triton X-100 (#BP151-100, Fisher Scientific) was used as positive control. Two independent experiments with triplicate wells in each experiment were performed.
Effect of the SMs on intracellular survival of APEC in phagocytic and non-phagocytic cells. Intracellular survival assay was conducted as described previously 28,29 to determine the effect of cidal SMs on APEC survival in phagocytic (HD11, THP-1) and non-phagocytic (Caco-2) cells. Briefly, mid-logarithmic phase grown APEC O78, O1, and O2 were washed and adjusted to 1 × 10 7 CFU/mL in cell culture incomplete media (no FBS and antibiotics). One-hundred microlitres adjusted APEC suspension was added at multiplicity of infection (MOI) 10 to wells of 96-well cell culture plate containing cultured macrophage (HD11, THP-1) and epithelial cells (Caco-2) and incubated for 1 h and 3 h, respectively. For APEC O1, invasion time was reduced by 3 times in all cell types as APEC O1 was found with significantly (P < 0.01) higher invasiveness compared to O78 and O2 (Fig. S1). After incubation, cells were washed and treated with 150 µg/mL gentamicin (#157100164, Fisher Scientific) for 1 h to kill extracellular APEC. The cells were then washed, replenished with incomplete media containing different concentrations (0.5X, 1X, 2X, and 4X MIC) of SMs, and incubated for 6 h. The cells were then lysed with 100 µL of 0.1% Triton X-100 for 5 mins, serially diluted, and plated on LB agar plate to enumerate viable bacteria. The intracellular bacteria in SMs treated wells were compared with DMSO (1%) treated wells. Two independent experiments in duplicate wells for each concentration of SMs were conducted.

Toxicity and efficacy of SMs in wax moth (Galleria mellonella) larvae. For toxicity evaluation, G.
mellonella larvae (fifth instar) were inoculated with 12.5 µg of SMs (50 mg/kg body wt.) through last pro-leg using PB600-1 repeating dispenser (Hamilton, Reno, NV) attached to insulin syringe (31 gauge, 8 mm needle length) (ReliOn ® , Bentonville, AR). For the inoculation, SMs were diluted in buffer mix containing 30% DMSO and 10 mM MgSO 4 (#r-375-25, CQ Concepts) as described previously 30 . Post-inoculation, larvae were placed inside sterile petri dishes and incubated up to 72 h in the dark at 37 °C and larval survival was monitored every 12 h. Non-treated larvae, larvae treated with the buffer mix, and larvae treated with CHL (75 mg/kg body wt.; dose sufficient to clear APEC infection in larvae) were used as controls.
For SMs efficacy testing, larvae were first injected with SMs mixed in buffer through the left hind pro-leg at dose rate as described above and incubated for 2 h at 37 °C. Then, larvae were infected with 6 × 10 4 CFU of Rif r APEC O78 in 10 mM MgSO 4 on the right hind pro-leg. Rif r APEC O78 was generated by plating APEC on LB agar plate containing 50 µg/mL rifampicin for specific monitoring of APEC population inside the larvae. SMs displayed identical MIC and MBC to Rif r APEC O78 as the wild-type (Fig. S2). Infection dose of Rif r APEC O78 to larvae was identified based on preliminary study (Table S3). Infected larvae inoculated with buffer mix were used as positive control whereas larvae inoculated with CHL were used as negative control. Post-inoculation, larval survival was monitored as above. For the quantification of APEC load inside the dead and live larvae, larvae from SMs treated and control groups were surface sterilized with 70% ethanol and homogenized in PBS. The suspension was ten-fold serially diluted and plated on MacConkey agar (#R453802, Fisher Scientific) plates supplemented with 50 µg/mL of rifampicin. The plates were then incubated overnight at 37 °C and APEC load was enumerated. Each experiment was repeated twice using larvae (n = 20) obtained in different batches.
Statistical analysis. The statistical significance of the effect of SMs in reducing biofilm embedded and intracellular APEC was determined by one-tailed Student's t-test (P < 0.01). The significance of CV uptake and increase of OD 260 and 280 nm absorbing bacterial supernatants in SMs treated samples was statistically analyzed by one-tailed Student's t-test (P < 0.05). Kaplan-Meir survival curves were generated using GraphPad Prism V.5 and were statistically analyzed by log-rank test (P < 0.05). APEC load inside the SMs treated and control larvae were analyzed by one-way ANOVA tukey's test using GraphPad Prism V.5 (P < 0.05). APEC load inside the live and dead larvae were statistically compared using one-tailed Student's t-test (P < 0.05). Correlation (r) between the larval survivability and APEC load was calculated using Microsoft Excel 2010.
Six SMs affected limited number of commensal/probiotic bacteria. The use of non-specific and broad spectrum antimicrobials have effect on beneficial microbes leading to the disturbance of microbiota which renders host susceptible to infections by pathogens 34 . In addition, use of broad spectrum antimicrobials also enriches the abundance of resistant microorganisms and resistant genes in the microbiota which in turn could foster the antimicrobial resistance problem 35 . Six of these SMs (SM1-SM3, SM8, SM9, and SM11) exerted least effect on beneficial microbes; having cidal activity against one to three of the 12 commensal/probiotic bacteria tested at 100 µM (Fig. 2D). Whereas, three SMs (SM4-SM6) belonging to imidazoles group were bactericidal to most of the tested probiotics/commensals. Even though piperidines (SM1, SM11) and pyrrolidinyls (SM2, SM3) group of SMs possessed higher MICs than other SMs, they displayed more specific activity against APEC. Interestingly, most of the SMs (SM1-SM3, SM8, SM9, and SM10) did not have effect on E. coli Nissle 1917 and E. coli G58-1 and none of the SMs exerted effect on Bifidobacterium lactis Bb12 (Fig. 2D). Overall, Lactobacillus brevis, Lactobacillus rhamnosus GG, Bifidobacterium lactis Bb12, and Bacteroides thetaiotaomicron are the microbes least affected by these SMs.
No resistance was detected in APEC O78 to SMs. Identical MBCs were observed when APEC O78 was grown in sub-lethal (0.75X MIC) doses of SMs in liquid media for 15 overnight passages (90 generations) (Fig. S4). After 15 days of incubation of APEC O78 on solid media amended with a 2X MBC of SMs, no resistant colonies were observed. These results suggest that the 11 SMs were less likely to induce resistance in APEC O78; however, more in-depth characterization of resistance is needed for future development and application of these SMs in the field. SMs exhibited antimicrobial activity by affecting APEC cell membrane. BCP is regarded as a rapid and powerful approach to identify the cellular pathways affected by different antibacterials based on the cytological changes induced by SMs 25 . Our study revealed that the 11 cidal SMs are likely to functions by either disrupting cell membranes or producing membrane defects or inhibiting cell wall peptidoglycan (PG) synthesis (Fig. 3). DMSO treated APEC bacteria showed stained membrane and nucleic acid ( Fig. 3A;I). Imidazoles SMs (SM4-SM6) are likely to disrupt the cell membrane which is similar to polymixins mechanism of action (MOA) 37 and is evident by the absence of FM4-64 stained bacterial cell membrane ( Fig. 3A;II-IV). Pyrrolidinyls SMs (SM2, SM3,  (Table 2). Intracellular constituents such as DNA, RNA, proteins can be leaked through permeable membrane 41 .
SEM results further supported the cell membrane affecting mode of action of SMs. SEM images suggest that SMs treatment produced membrane wrinkling, blebbing/vesicle-like structures, and pores ( Fig. 4) which are the characteristics morphology induced by membranes acting antibiotics and several other antimicrobial agents [42][43][44][45][46][47] . DMSO treated APEC bacteria showed very few wrinkled smooth surfaced cells measuring 1-2 µM (Fig. 4A). The frequency and sites of blebbing and severity of wrinkling differed between the SM treatments. SM6 (Fig. 4B) produced more severe wrinkling and multiple blebbing throughout the cells. SM3 and SM7 formed blebbing and pore at single cell pole, respectively (Fig. 4C,D). SM8 and SM10 produced distinct morphology than other SMs with shortened cells (~0.5 µM) and blebbing at both cell poles (Fig. 4E,F) which is similar to ampicillin induced cells morphology 40 . Even though, these studies showed SMs inducing APEC membrane alterations; however, the observed changes in membrane morphology could be either direct or indirect effects of the SMs.

SMs reduced intracellular APEC in phagocytic and non-phagocytic cells. The fimbria mediated
initial APEC adhesion and OmpA/IbeA mediated invasion into the cells facilitate APEC to survive intracellularly in phagocytic and non-phagocytic cells of the host and is an important aspect of APEC pathogenesis 48 . Therefore, the administered antimicrobial therapeutics must be able to permeate and act inside the APEC infected cells. After 6 h of treatment, SMs significantly (P < 0.01, Student's t-test, see Table S6 for t values) reduced intracellular APEC O78, O2, and O1 in infected Caco-2, HD11, and THP1 cells at varying concentrations (0.5 × -2X MIC) with maximal reduction (3-5 log; 100% clearance) of intracellular APEC O78, O2 and O1 at concentration less than or equal to 4X MIC (  49 . Among 11 SMs, SM4-SM10 were effective in clearing intracellular APEC O78, O2, and O1 at concentration less than or equal to 100 µM for most of the cases; whereas, SM1-SM3 and SM11 were effective only at concentration equal or above 100 µM (Table 1). Interestingly, higher concentrations of SMs were needed to clear intracellular APEC O1 followed by O2 and O78 (   O1 serotype inside the cells (Fig. S1). The serotype O1 is reported to carry IbeA (invasin) and Iss (increased serum survival) genes more frequently compared to O78 and O2 50,51 which might contribute for better invasion and survival. SM8 was the most effective SM in clearing intracellular APEC with complete clearance at concentration less than or equal to 50 µM (Table 1). Overall, SM4, SM7, SM8, SM9, and SM10 were the most effective SMs in clearing intracellular APEC serotypes in all tested cells.
SMs showed low toxicity toward wax moth larvae, extended the larval survival, and reduced the APEC load inside the larvae. The wax moth larval model is increasingly used in the recent years as an alternative to mammalian model to study bacterial pathogenesis and antimicrobial drug testing 52 . Except SM1, rest of the SMs showed less toxicity (<10%) to larvae (Fig. 6A).

Structure-activity relationship analysis. Structural clustering of identified hits based on their
2D-Tanimoto similarity showed imidazole (SM4-SM6) and quinoline (SM8, SM9) SMs structurally more close with nitrogen-containing aromatic ring in common which could contribute for their lower MIC and MBC in comparison to pyrrolidinyl (SM2, SM3, SM7) and piperidine (SM1, SM11) SMs (Figs 1C, 2A,B and 7A). Among the pyrrolidinyl hits, hits with additional benzene ring in the pyrrolidine scaffold showed bactericidal activity (Fig. 1C) whereas hits without benzene ring showed only bacteriostatic activity (not shown). In addition, among the bactericidal pyrrolidinyl SMs (SM2, SM3, SM7), SM7 possesses trifluoro group which could contribute to its bactericidal activity at lower concentration in comparison to SM2 and SM3 (Fig. 1C). The hydroxy-methoxybenzyl group is absent in bactericidal piperidine SMs (SM1, SM11) (Fig. 1C) in comparison to bacteriostatic piperidine hits (not shown). SM10, which belongs to miscellaneous group contains trihalogen (trichloro) group (Fig. 1C) as similar to SM7 which could contribute to its bactericidal activity at lower concentration (Fig. 2B).

Discussion
APEC is responsible for severe economic losses to the poultry industry worldwide 1,2 and is also regarded as a potential source of human ExPECs 1 . Effective novel control methods are needed because of the limitations associated with current control methods 5,9 . Anti-APEC SMs identified in our study are diverse in their structures with three major clusters based on structural similarity (Fig. 7A) and contained pyrrolidinyl, piperidine, imidazole, and quinoline scaffolds (Fig. 7B). In previous studies, chemical compounds having the piperidine and pyrrolidine rings were reported with antimicrobial activity against several infectious pathogens 53,54 . Likewise, imidazoles and the 2,4,5 -trisubstituted imidazole derivatives were regarded as versatile and promising antimicrobials as well as sensitizers of MDR pathogens [55][56][57] . Quinolinedione derivatives were also reported as potent antimicrobial agents against Gram negative and positive bacteria 58 . Though compounds belonging to same chemical groups were reported with antimicrobial activity against several pathogenic bacteria including E. coli in earlier studies, there is no previous report of the antimicrobial activity of the compounds identified in our study except, SM9. SM9 was reported to inhibit E. coli BW25113 growth (57%) at 100 µM (https://pubchem.ncbi.nlm.nih.gov/bioassay/710). These identified anti-APEC scaffolds could facilitate the development of antimicrobial therapeutics to control APEC infections in poultry.
Anti-APEC SMs identified in our study affect the APEC cell membrane. Bacterial cell membranes are regarded as promising targets for discovery of new antimicrobial therapeutics and to combat antimicrobial resistance 59 . Membrane affecting antimicrobials are most likely to act by disrupting membrane architecture and functional integrity 59 which is supported by our confocal and SEM images and membrane permeability assays (Figs 3 and 4). Under confocal microscopy, SMs treated APEC bacteria showed membrane disrupted morphology along with formation of membrane defects throughout the cell (Fig. 3A,B,C). The disruption of the cell membrane and formation of membrane defects could subsequently leads to leakage of cell contents, loss of membrane potential, and eventual cell death 59 . Further, SEM analysis revealed that SMs treatment induced membrane wrinkling, blebbing/vesicle-like structures, and pores (Fig. 4) which consequently could impair the cell membrane integrity leading to cell death 38 . The membrane defects caused by SMs resembles to those caused by already known membrane acting antibiotics such as polymixins (SM4-SM6) 37 , daptomycin (SM2, SM3, SM7, and SM11) 38 , ampicillin/cephalexin (SM1, SM8-SM10) 25,40 , and several other antimicrobial peptides 32,[42][43][44][45][46][47] . Polymixins disrupt the outer membrane integrity of Gram negative bacteria by forming the blebs on the surface of the bacterium 31 . Daptomycin induces holes in the membrane leading to a breach in the cell membrane and subsequent cell death by forming membrane blebs 38 . Multiple antimicrobial peptides such as Human α-defensin 5 (HD5), gramicidin S, peptidyl-glycylleucine-carboxyamide (PGLa), cathelicidins, lactoferricin, and human epididymis 2 (HE2) protein isoforms damage the bacterial cell membrane by forming the blebs 32,[43][44][45][46] . Peptoids, an alternative to antimicrobial peptides, damage the membrane of E. coli by forming membrane blebs 47 . Sericin, a soluble silk glue protein exhibits antibacterial activity against E. coli by inducing blebbing of the membrane 42 . Furthermore, previous studies have also shown that quinoline, imidazole, piperidine, and pyrrolidine compounds possessing antibacterial activity by affecting the bacterial cell membranes [60][61][62][63] which is consistent with our findings.
The SMs identified in our study are effective against multiple APEC strains, STEC strains as well as antimicrobials resistant strains (Fig. 2) which might be explained by their membrane affecting mode of action. Antimicrobials that target the cell membranes exhibit broad spectrum of activity and are being used to control MDR bacteria such as ESKAPE pathogens 31,32 , methicillin resistant Staphylococcus aureus (MRSA) 60 ; therefore, SMs identified in our study could be used to treat APEC infections caused by antimicrobial resistant strains. Membrane affecting antimicrobials also have a low potential for development of resistance mostly due to their effect on multiple targets 59 . Consistent with low resistance acquisition of membrane affecting antimicrobials, no resistant APECs were isolated in vitro in our study which could makes these SMs as emergency antimicrobials in APEC outbreaks situation. Permeability of APEC cell membrane is also impaired following SMs treatment (Fig. 3D). Thus, the incorporation of these SMs in therapy could enhance the uptake or penetration of antibiotics that have intracellular targets 59 or could interact synergistically with other membrane affecting antibiotics 64 . In fact, several SMs significantly decreased the MBC of TET, CST, and CIP that are commonly used to treated APEC infection in poultry (unpublished data). As a result, combining these SMs could increase the activity of antibiotics or reduce the amount of antibiotics needed, and by consequence, could attenuate the development of antimicrobial resistance associated with APEC in poultry.
Most of the identified SMs, especially imidazoles (SM4-SM6) and pyrrolidinyls (SM2, SM3, SM7), eradicated biofilm embedded APEC even at 0.5X to 2X MIC (Table 1) which could be due to low molecular wt. of SMs allowing better penetration inside the biofilms 36 or could be due to inherent biofilm dispersal/disruption activity of imidazoles 65 or anti-biofilm activity of pyrrolidinyls 66 . Membrane affecting antimicrobials have capacity to act against slow-growing or dormant bacteria as well as on biofilms 59 . APEC can form biofilms in poultry facilities such as in water lines and drinker systems 67 and are difficult to eradicate by common disinfectants and antimicrobials. Therefore, the SMs identified in our study could be used to eradicate biofilm embedded APEC in poultry facilities; thereby reducing the incidence and occurrence of APEC infections in poultry farms. SM8 and SM10, which are effective against planktonic and intracellular bacteria even at low concentration (Fig. 2, Table 1) showed decreased effectivity towards biofilm embedded APEC which could be due to restricted penetration of SMs inside the biofilm or could be due to binding with biofilm matrix 68 . Additionally, these SMs contain chlorine atoms in common; bacterial biofilms are increasingly resistant to chlorine treatment 69 .
We used the "yactives" library under the premise that compounds with this property may be enriched for bioactivity against non-yeast-based chemical screens as was originally shown by Wallace et al. 15 . Previous screen in yeast was conducted using higher concentration (200 µM) of SMs as compared to our screening (100 µM) (Fig. 1). At 200 µM, the SMs were considered "yactives" if they inhibited the yeast growth at least by 30%; however, the most potent SMs identified in our study showed almost 100% inhibition of APEC at lower concentration (8-16x less; 12.5-25 µM) (Fig. 2); therefore, we anticipated least effect of these SMs on eukaryotic cells. Indeed, most of the identified SMs (SM1-SM3, SM7-SM10) showed less toxicity towards chicken and human cells (Fig. 5). The toxicity of the membrane affecting antimicrobials depends upon the membrane organization and its lipids composition and proportion 59 . Both epithelial and macrophage cells membrane contain phosphatidylcholine (PC) as major phospholipid 59 ; however, bacterial cell membrane is rich in phospholipids such as phosphatidylglycerol (PG), phosphatidylethanolamine (PE) and cardiolipin (CL) which makes membrane affecting antimicrobials selectively toxic to bacterial cells 70 . RBCs also contain phosphatidylethanolamine (PE) phospholipid in their membrane similar to bacterial cell membrane 71 ; this similarity could attribute toxicity of some of the SMs (SM4-SM6, and SM11) to RBCs. The presence of cyclohexyl and/or benzodioxol groups in SM4 and SM11 could contribute for their relatively high toxicity (Fig. 1C) 72,73 . These SMs were however not toxic to wax moth larvae which could be due to cellular analogy of wax moth larva to mammals (epithelial cells of larva gut similar to intestinal cells of mammals) 74 . Consistent with wax moth studies, no negative impact of SM5 and SM6 on chicken health and performance was observed in our pilot experiment (data not shown). Further, most of the SMs identified exerted no effect on tested Gram positive bacteria such as Lactobacillus and Bifidobacterium (Fig. 2D). The use of rich media however could attribute to lesser effect of SMs to beneficial microbes. The Gram positive and Gram negative bacteria also have different composition and relative amounts of lipids in their membranes 75 . Gram positive bacterial genera such as Clostridium, Lactobacillus, and Bifidobacterium are the predominant commensals of the poultry gut microbiota 76 . Thus, we expect lesser impact on the microbiota of the chickens treated with these SMs 34 . Further, LGG and Bb12, widely used probiotics 77 are unaffected by these SMs, they could be combined with these SMs to enhance the probiotics control of APEC infections in poultry.
The treatment with most of the identified SMs cleared the intracellular APEC in the infected phagocytic and non-phagocytic cells (Table 1); similar effect within the host cells could help to ameliorate APEC pathogenicity 28,48 . Consistent with the SMs intracellular clearance of APEC, SM1-SM3, SM4-SM6, SM7, SM10, and SM11 treatment significantly reduced the APEC load inside the wax moth larvae. The lesser efficacies of SM8 and SM9 in wax moth larvae in comparison to cultured epithelial and macrophage cells could be due to interaction with host immune components of wax moth larvae such as antimicrobial peptides or due to production of drug degradative enzymes 78 . Wax moth larvae possess complex innate immune system similar to mammals and several studies including studies in ExPEC have reported the similar results between wax moth and mammalian models 30,79 . Besides, wax moth larval model has been frequently used to evaluate the efficacy and toxicity of antimicrobial agents 52 . Therefore, the efficacy of these SMs in cultured infected cells and wax moth larvae may suggest their therapeutic efficacy in chickens.
In conclusion, this study had identified seven novel SMs (SM3, SM5-SM10) (Fig. S5) as potentially effective and safe (two foremost parameters of any therapeutic drug) anti-APEC therapeutics for poultry. These SMs function through affecting APEC cell membrane and can also be combined with other anti-APEC strategies such as antibiotics and probiotics. Our future studies will focus on testing SMs efficacy in chickens, identifying SMs molecular targets to define their modes of action, and also to develop these SMs to control E. coli related foodborne zoonosis including APEC related ExPEC infections in humans.

Data Availability
All data generated or analysed during this study are included in this published article (and its Supplementary  Information files).