Isoamphipathic antibacterial molecules regulating activity and toxicity through positional isomerism

Peptidomimetic antimicrobials exhibit a selective interaction with bacterial cells over mammalian cells once they have achieved an optimum amphiphilic balance (hydrophobicity/hydrophilicity) in the molecular architecture. To date, hydrophobicity and cationic charge have been considered the crucial parameters to attain such amphiphilic balance. However, optimization of these properties is not enough to circumvent unwanted toxicity towards mammalian cells. Hence, herein, we report new isoamphipathic antibacterial molecules (IAMs: 1–3) where positional isomerism was introduced as one of the guiding factors for molecular design. This class of molecules displayed good (MIC = 1–8 μg mL−1 or μM) to moderate [MIC = 32–64 μg mL−1 (32.2–64.4 μM)] antibacterial activity against multiple Gram-positive and Gram-negative bacteria. Positional isomerism showed a strong influence on regulating antibacterial activity and toxicity for ortho [IAM-1: MIC = 1–32 μg mL−1 (1–32.2 μM), HC50 = 650 μg mL−1 (654.6 μM)], meta [IAM-2: MIC = 1–16 μg mL−1 (1–16.1 μM), HC50 = 98 μg mL−1 (98.7 μM)] and para [IAM-3: MIC = 1–16 μg mL−1 (1–16.1 μM), HC50 = 160 μg mL−1 (161.1 μM)] isomers. Co-culture studies and investigation of membrane dynamics indicated that ortho isomer, IAM-1 exerted more selective activity towards bacterial over mammalian membranes, compared to meta and para isomers. Furthermore, the mechanism of action of the lead molecule (IAM-1) has been characterized through detailed molecular dynamics simulations. In addition, the lead molecule displayed substantial efficacy against dormant bacteria and mature biofilms, unlike conventional antibiotics. Importantly, IAM-1 exhibited moderate in vivo activity against MRSA wound infection in a murine model with no detectable dermal toxicity. Altogether, the report explored the design and development of isoamphipathic antibacterial molecules to establish the role of positional isomerism in achieving selective and potential antibacterial agents.


Synthetic protocol of N,N,N',N'-tetramethyl diaminohexyloxy benzenes (1b, 2b and 3b).
NHMe 2 gas was collected into 20 mL chloroform solution of individual dibromohexyloxy benzene (1a, 2a and 3a) (1 g) in a sealed tube at 0 °C till the volume of the resulting solution was roughly doubled. Then, the reaction mixture was allowed to stir for 24 h at room temperature. Next, the reaction mixture was cooled and transferred into a RB. The solution was then kept in water bath to remove excess NHMe 2 followed by solvent evaporation by using rotary evaporator. Reaction mixture was then washed with 2 M aqueous KOH solution after dissolving it in CHCl 3 . Finally, CHCl 3 layer was collected and evaporated to dryness to get a yellowish gummy liquid, 1b, 2b and 3b with quantitative yield. 6,6'-(1,4-phenylenebis(oxy))bis(N,N-dimethylhexan-1-amine) (3b). 1  Synthetic protocol of ethyl ester bromide of phenylalanine (1c). Thionyl chloride (0.65 mL, 9 mmol) was added dropwise at 0-5 ˚C to 20 mL suspension of L-Phe (0.5 g, 3 mmol) in ethanol and the entire reaction mixture was refluxed for 12 h. Next, the excess ethanol and thionyl chloride were evaporated by rotary evaporator. The solid residue was washed with dry diethyl ether, resulted a white solid crude product. This crude white solid was dissolved in 10 mL of dichloromethane and potassium carbonate (1 g, 7.5 mmol) was added to the organic solution after dissolving it in 10 mL of distilled water. A solution of bromoacetyl bromide (0.4 mL, 4.5 mmol) in dichloromethane (10 mL) was then added dropwise to the reaction mixture at 5 °C for 1 h. The reaction mixture was stirred at room temperature for another 12 h. The aqueous solution was separated and the organic solution was washed with water and passed over the anhydrous Na 2 SO 4 and concentrated to yield a white or yellowish white solid product with 80-94% yield.

General synthesis procedure of isoamphipathic antibacterial molecules (IAMs 1-3).
Individually, ethyl ester bromide intermediates, 1c (2.8 equiv.) were reacted with N, N, N',N'-tetramethyl diaminohexyloxy benzenes (1b, 2b and 3b) (1 equiv.) in dry CHCl 3 (8-10 mL) at 65 ˚C. At the end of 24 h, reaction mixture was evaporated by using rotaryevaporator and the residue was dissolved in minimum amount of CHCl 3 . The product was then precipitated by adding excess dry diethyl ether and the white residue was washed repeatedly with dry diethyl ether to remove the excess amount of activated ethyl ester bromides. The entire exercise resulted the generation of isoamphipathic antibacterial molecules (IAMs: 1-3) with a quantitative yield. All the final molecules were characterized through 1 H-NMR, 13   Antibacterial assay. 1,2 At first bacteria from the frozen stock (at -80 ˚C) were streaked either on nutrient broth (for Gram-positive bacteria) or MacConkey agar plate (for Gram-negative bacteria). The streaked plates were then incubated overnight at 37 ˚C for bacterial growth. A single bacterial colony was next inoculated for 6 h (midlog phase) in 3 mL of nutrient broth to produce about 10 8 to 10 9 CFU/mL cells depending upon the nature of the bacteria. The 6 h grown culture was diluted to ~10 5 CFU/mL which was then used for antibacterial assay determination. Compounds were 2-fold serially diluted in a 96-well plate from the starting concentration using sterile Millipore water. Afterward, 180 μL of ~10 5 CFU/mL bacterial solution was added in each well containing 20 µL aqueous solution of the test compound. The plates were then incubated for 16-18 h at 37 ˚C in shaking condition. The O.D at 600 nm was recorded by using TECAN (Infinite series, M200 PRO) plate reader. Each concentration was triplicate and the experiment was performed twice and the antibacterial activity (MIC) was evaluated based on visual turbidity.
Antibacterial activity upon human blood plasma and serum pre-incubation. 2 Human blood was centrifuged at 3500 rpm for 5 min and the plasma was collected carefully from the supernatant. Similarly, human blood was collected into BD Vacutainer® Serum Tubes and further centrifuged at 3500 rpm to obtain human blood serum. Next, 500 µL of 512 µg/mL (515.6 µM) solution of IAM-1 in 1×PBS was mixed with equal volume of blood plasma/serum and the individual mixtures were incubated for different time points (1 h, 2 h, and 3 h) at 37 ˚C. At the end of specific incubation time, each mixture was subjected for 2 fold serial S6 dilutions and MIC was conducted with them by following the aforementioned protocol against MRSA.
Hemolytic activity. 2,3 Aqueous solution of individual compounds (IAMs: 1-3) were serially diluted by two fold in triplicate in a 96-well plate. Freshly collected human blood (heparinized) was then centrifuged down and supernatant was thrown away to collect the human red blood cells (hRBCs). Later, collected hRBCs (5 vol %) were slowly suspended using 1×PBS (pH = 7.4). Next, 150μL of this suspension was added to the wells of 96 well plates containing 50 μL compound's solution and plate was allowed to incubate at 37 ˚C for 1 h. Then centrifugation at 3500 rpm for 5 min was performed and the supernatant (100 μL) was then transferred to another 96-well plate for recording the absorbance at 540 nm by using Tecan Infinite M200 PRO microplate reader. In this study, same volume of 1×PBS without compound served as a negative control whereas same volume of Triton X-100 (1 vol% solution in 1×PBS) was used as a positive control. The percentage of hemolysis was determined by using the following formula: (A treat -A nontret ) /(A TX-treat -A nontret ) ×100, where A treat corresponds to the absorbance of the compound-treated well, A nontret stands for the absorbance of the negative controls (without compound), and A TX-treat isthe absorbance of the triton-X-100 treated well. Each concentration had triplicate values and the HC 50 was determined by considering the average of triplicate O.D.

Co-culture study
Co-culture with MRSA and human erythrocytes. 4 MRSA and human erythrocytes were mixed with each other in 1×PBS in such way so that their effective concentration remained as ~10 7 CFU/mL and ~10 8 hRBCs/mL. Next, 180 µL of the co-culture suspension was treated individually with 20 µL of IAM 1-3 at 2560 µg/mL (2578 µM) [working concentration: 256 µg/mL (257.8 µM)]. After 15 min incubation at 37 ˚C in a stationary condition, 20 µL of the suspension was withdrawn and 10-fold serially diluted followed by spot plating each dilutions on nutrient agar plates for counting viable MRSA cells. Finally, at the end of 24 h incubation, the viable bacterial colonies were counted and the results were expressed in percentage with respect to the untreated MRSA count in co-culture condition. On the other hand, the reaming volume of MRSA and hRBC suspension was centrifuged at 3500 rpm for 5 min and the absorbance of the supernatant (100 μL) was recorded at 540 nm by using Tecan Infinite M200 PRO microplate reader. In this study, same volume (20 µL) of triton-X (1 vol% solution in 1×PBS) instead of test compound was used as a positive control for determination of hRBC lysis. The percentage of hemolysis was determined by using the following formula: (A tret -A nontret ) /(A TX-tret -A nontret ) ×100, where A tret corresponds to the absorbance of hRBCs in co-culture condition upon compound treatment , A nontret stands for the absorbance of hRBCs in co-culture condition without compound treatment, and A TX-tret is the absorbance of hRBCs in co-culture condition upon triton-X treatment.
Co-culture with MRSA and RAW cells. 5 RAW 264.7 cells (~10 5 cells/well) were seeded onto the wells of a 96-well plate in DMEM media (supplemented with 10% FBS and 5% penicillin-S7 streptomycin) at 37 ˚C with 5 % CO 2 atmosphere for 12 h. Afterward, cell culture medium was discarded and cells were treated with 100 µL of freshly prepared MRSA (~ 10 5 CFU/mL) supplemented with 256 µg/mL (257.8 µM) compound solution (IAM-1 or IAM-2 or IAM-3) in antibiotic free DMEM media with 10% FBS. Two controls were included in the study, in one case the MRSA cells were left untreated with compound in presence of RAW cells and in other case, RAW cells were left untreated with both compound as well as MRSA where same volume of antibiotic-free DMEM (supplemented with 10% FBS) was added. The plate was then incubated for 3 h at 37 ⁰C under 5% CO 2 atmosphere. To determine the bacterial cell viability, 20 µL of the suspension was withdrawn and 10-fold serially diluted followed by spot plating each dilution on nutrient agar plates for counting viable MRSA cells. Finally, at the end of 24 h incubation, the viable bacterial colonies were counted and the results were expressed in percentage with respect to the untreated MRSA count in co-culture condition. On the other side, RAW cells viability was also determined by performing Alamar blue assay (chapter 2, section 2.4.12) with the reaming supernatant and results were represented in terms of percentage by considering 100% cell viability in case of RAW cells devoid of MRSA and compound treatment.
Membrane fluidity study. 6,7 0.5 mM of DPPG : DPPE (88 : 12) and 0.5 mM of DPPC lipids along with laurdan dye (5 mM) were dissolved in minimal volume of analytical grade chloroform in a 10 mL RB. Next, a thin layer films with the lipids were made within the walls of RB by applying argon flow. Further, the thin films were dried under vacuum for 1 h. Then, 10 mL of 1×PBS (pH = 7.4) was added on the lipid films and incubated for 12 h at 4-8 ˚C for hydration. Further, 10 freeze-thaw cycles (from 70 ˚C to 4 ˚C with intermittent vortexing) were executed with the hydrated films. These multilamellar vesicles were finally sonicated at 70 ˚C for 15 min to obtain unilamellar vesicles. The freshly prepared liposome solution (2 mL) was taken into 3 mL fluorescence cuvette and the emission intensity was measured at 440 nm upon excitation at 350 nm. Similarly, the emission intensity was measured upon IAMs 1-3 treatment at 256 µg/mL (257.8 µM) in case of both bacterial and mammalian membrane mimicking liposomes.
Membrane leakage study. 6,7 To investigate dye leakage upon IAM 1-3 treatment through fluorescence spectroscopy, vesicles entrapped with 30 mM 5(6)-Carboxyfluorescein (CF) dye was prepared in 1×PBS buffer (pH = 7.4). In high concentrations (30 mM), CF does not emit due to self-quenching and displays significant emission at diluted condition of 3 mM concentration. Bacterial membrane mimicking vesicles were prepared using 0.5 mM DPPG and DPPE lipid (88 : 12) in CF solution (30 mM) in 1×PBS buffer by following the previously mentioned protocol. Mammalian membrane mimicking vesicles were also prepared in the similar way using 0.5 mM DPPC lipid. However, herein, each freeze-thaw cycle was accompanied with sonication for 30 s. Further, vesicle untapped dye was removed from the solution using size exclusion chromatography. Sephadex G-50 was used as the stationary phase and 1×PBS was used as the eluent. Finally, 200 µL of vesicles were equilibrated with 1.8 mL 1×PBS and emission intensity was recorded at 517 nm using an S8 excitation at 495 nm. In control case, instead of test compound solution, water was added at 50 s and fluorescence intensity was measured upto 200 s. Likewise, fluorescence intensity was recorded upon treating bacterial and mammalian membrane mimicking vesicles individually with IAM 1-3 [160 µg/mL (161 µM)] at 50 s followed by addition of 1% triton-X at 200 s. The fluorescence intensity obtained after triton-X treatment in case of control was used for normalization. Further, rate constant of dye leakage was calculated by fitting each emission spectrum from t = 50 s (after addition of IAM) to t = 200 s (before addition of triton-X) using exponential rise equation (y = y 0 + A 1 e -x/t 1 ). The rate constant was considered ask = 1/t 1 .

Water-membrane partition experiments
Liposomes were prepared by extrusion, as previously described, with lipid compositions: POPE/POPG (7:3, mol/mol) and POPC/Cholesterol (1:1, mol/mol), as mimics of bacterial and eukaryotic membranes, respectively. The lipid film was hydrated with a 10 mM phosphate buffer (pH=7.4), prepared by using ultrapure water and containing 140 mM NaCl and 0.1 mM EDTA (buffer A henceforth). Liposomes were separated from unencapsulated dye by gel filtration on a 40 cm Sephadex G-50 column. The final lipid concentration was determined according to the Stewart phospholipid assay. 8 A fixed IAM concentration (10 µM) was titrated with increasing concentrations of liposomes. After each addition, the fluorescence spectra were recorded repeatedly until no further changes were observed (about 5-10 minutes). Control experiments to check for possible effects of sample turbidity on the fluorescence signal were performed by repeating the same titration with tryptophan.
The fraction of IAM associated to membranes ( ) was calculated from the fluorescence intensity (F, at the wavelength of 318 nm), according to the following equation: where F b and F w refer to the fluorescence signal of bound and free IAM, respectively. F b was determined by extrapolating the titration with the following equation: where K p represents the lipid/water partition constant, [L] is the molar concentration of lipids in the sample and [W] is the molar concentration of pure water at 25 °C (55.5 M). In Equation 2, the lipid concentration dependence of the fraction of bound IAM was described by the following equation, corresponding to an ideal partition equilibrium model. 8,9 S9 The partition curves of the POPC/Cholesterol neutral liposomes were too far from the plateau to allow extrapolation. Therefore, the value of the bound state fluorescence F b determined for the POPE/POPG liposomes was used for the POPC/Cholesterol vesicles, too, assuming that the membrane composition does not affect significantly the fluorescence in the bound state. In principle, the turbidity of the liposome solution might artifactually cause variations in the fluorescence signal. However, negative control experiments with tyrosine (which does not bind to vesicles) showed no significant changes in the emission spectra, demonstrating the absence of any relevant scattering-related artefacts.

Molecular dynamics simulations
The force field parameters for IAM compounds were obtained by starting from the ATB server. 10 The charges were slightly modified to preserve the symmetry of the molecules and in analogy with those of analogous groups in the GROMOS 54A7 force field. 11 The hydrophobic chains in the compounds were modeled with the Berger parameters used for lipids (see below).
In simulations of the IAM molecules in water, a single copy of the compound was placed at the center of a box (44 nm 3 ) and hydrated with approximately 1400 preequilibrated simple point charge (SPC) water molecules. 6 Cland 4 Na + ions were added to ensure electroneutrality and a ionic strength roughly corresponding to 0.150 M. MD simulations were carried out with the GROMACS 2020.6 software package. 12 Each system was energy-minimized and then equilibrated during a 100 ps MD, where the positions of the IAM atoms were restrained. Production simulations were performed at least in triplicate, for a total amount of almost 12 µs of simulation time at a temperature of 300 K, with the same parameters and settings used in the membrane simulations (see below).
The most representative structures of each simulation were defined by cluster analysis conducted with GROMACS (cluster tool), using a 0.25 nm root mean square deviation cutoff.
MD simulations of IAM-1 in presence of membrane lipids were performed with the "minimum bias" approach, which minimizes the effect of the initial configuration on the final results. In this method, the simulation is started from a random mixture of the membrane-active molecule, lipids and water, and the bilayer forms spontaneously, usually in 50-100 ns. During this self-assembly process, the system is very fluid, particularly in the first stages of the simulation, thus ensuring that the bioactive molecule can sample different environments in a relatively short time, and, as a consequence, it can find its minimum free energy configuration. We demonstrated that the simulative results obtained with this approach are consistent with the depth of membrane insertion and the orientation S10 determined experimentally by fluorescence, ATR-FTIR and solid-state NMR spectroscopies. [13][14][15] Membranes of POPE/POPG (90:38) were used to parallel the conditions of the experimental studies on lipid vesicles and to mimic bacterial membranes. Briefly, a single copy of the compound was placed at the center of a 9 × 9 × 9 nm box. 128 lipid molecules and 7500 water molecules were randomly added into the box. 38 Na + atoms were introduced in replacement of water molecules, as counterions of the negative charges on the lipids, and 2 Clatoms as counterions of the +2 charge on the compounds. MD simulations were carried out with the GROMACS 2020.6 software package. The parameters for the lipids were taken from the literature. 16 Temperature was controlled using a velocityrescaling thermostat. 17 Pressure coupling was applied using the Parrinello-Rahman barostat, with a time constant of 1.0 ps and a reference pressure of 1 bar. 18 All bond lengths were constrained with the LINCS algorithm. 19 Short-range electrostatic interactions were cut-off at 1.4 nm and long-range electrostatic interactions were calculated using the particle mesh Ewald (PME) algorithm. 20 Simulations were run with a 2 fs time step. Each system was energy minimized and then equilibrated using a 100 ps MD, where the positions of the IAM atoms were restrained. At first, simulations were performed for 200 ns, at a temperature of 310 K, with anisotropic pressure coupling. In cases where this time was not sufficient to attain a defect-free bilayer, the system was annealed by cycling the temperature between 310 K and 375 K, for further 120 ns, followed by a production run with semi-isotropic pressure coupling. Analyses were conducted on the last 20 ns of the simulations. The density profiles along the bilayer normal were determined by means of the "gmx density" tool in GROMACS.
Molecular graphics images were produced using the UCSF Chimera package from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco (supported by NIH P41 RR-01081). 21 Membrane active mechanism of action. [1][2][3] Cytoplasmic Membrane Depolarization Assay. Midlog phase (working concentration: ~10 8 CFU/mL) MRSA cells were collected separately (centrifugation at 3500 rpm for 5 min), and washed with 1:1 ratio of 5 mM glucose and HEPES buffer (pH = 7.4). Next, the bacterial plate was resuspended in 1:1:1 ratio of 5 mM HEPES buffer, 100 mM KCl solution supplemented with 0.2 mM EDTA and 5 mM glucose. For this study EDTA was used to allow the dye uptake by permeabilizing outer membrane of A. baumannii. This study was performed in a Corning 96 black well plate with clear bottom containing 2 μM of 3,3′-dipropylthiadicarbocyanine iodide [DiSC 3 (5)] and 190 μL of bacterial suspension. After 60 min incubation of the plate, fluorescence intensity was measured at 622 nm excitation wavelength and 670 nm emission wavelength for 4 min. After that, 10 μL of IAM-1 (at working concentration of 16 μg/mL and 32 μg/mL) was mixed with the suspension of bacteria and dye of each well. Same volume of water without compound was used as the control for this experiment. Increment in S11 fluorescence intensity was measured for another 25 min using Tecan Infinite M200 PRO microplate reader.
Outer Membrane (OM) Permeabilization Assay. Midlog phase E. coli cells were independently harvested, washed with 1:1 mixture of 5 mM HEPES buffer and glucose and resuspended with the same. The working concentration of midlog phase and stationary phase bacteria were ~10 8 CFU/mL. This study was performed in a Corning 96-black well plate with clear bottom containing 10 μM of N-Phenyl naphthylamine (NPN) dye and 190 μL of bacterial suspension. Then, the fluorescence was monitored for first 4 min at excitation wavelength of 350 nm and emission wavelength of 420 nm. After that, bacterial suspension with dye at each well was treated with 10 μL of test compound, IAM-1 at working concentrations of 16 μg/mL and 32 μg/mL. The same volume of water without a compound was used as the control for this experiment. An increase in fluorescence intensity was monitored for another 25 min with Tecan Infinite M200 PRO microplate reader.
Bacterial live/dead assay. 1 IAM-1 [16 µg/mL (16 µM)] was added to 1 mL of ~10 8 CFU/mL midlog phase MRSA suspension in normal saline and was incubated at 37 ˚C. After 6 h of incubation, bacterial suspension was centrifuged to remove the compound completely. Then, the bacterial palate was resuspended with normal saline followed by the addition of 5 µL of Syto-9 (3 µM) and PI (15 µM) mixture and incubated for half an hour. Further, the dye containing bacterial suspension was centrifuged to remove the excess unbound dye and the palate was resuspended with 50 µL of normal saline. Finally, 5 µL of bacterial suspension was taken into a glass slide and processed for confocal microscopy at 63 X resolution.

Cytotoxicity assay. 2
Alamar blue Assay. Cytotoxicity of IAMs: 1-3 was examined against Raw 264.7 cell line by Alamar blue assay. Briefly, cells (~10 4 cells/well) were seeded onto the wells of a 96-well plate in DMEM media supplemented with 10% fetal bovine serum (FBS) and 5% penicillinstreptomycin. Then 100 µL of serially diluted compound solution in DMEM media was added to the each well of the plates containing the cells. The same volume of media (untreated cells) and the cells treated with 0.1% (v/v) Triton-X solution was taken as positive and negative control respectively. The plates were then kept for incubation at 37 ˚C for 24 h maintaining 5% CO 2 atmosphere. Afterward, 10 µL of 10 x Alamarblue solution was added to each well followed by 4 h of further incubation at the same condition. Then, the absorbance was recorded at 570 nm wavelength and 600 nm wavelength was used as the reference. The percentage of cell viability was calculated using the following equation: cell viability (%) = (A c -A t )/(A 0 -A t ) × 100, where A c indicates the absorbance for cells treated with compound, A t is the absorbance for the cells treated with 0.1% (v/v) Triton-X and A 0 is the absorbance of the untreated cells, all at 570 nm. Each concentration had triplicate values, and the average of triplicate absorbance values was plotted against concentration followed by fitting with a sigmoidal plot.

S12
Fluorescence Microscopy. ∼10 4 cells (HEK 293) were seeded into the individual wells of a 96well plate. Then 100 µL of 256 μg/mL of IAMs: 1-3 was added over the seeded cells. 0.1% Triton-X treated and untreated cells were considered as positive and negative controls respectively. After single time washing with 1×PBS the untreated and treated cells were then stained with 50 μL of 1:1 calcein-AM (2 μM) and propidium iodide (PI) (4.5 μM) for 15 min under 5% CO 2 atmosphere at 37 ˚C. Finally, the excess dyes were removed by washing the cells with 1×PBS, and images were captured at 40×objective with the help of a Leica DM2500 fluorescence microscope. During imaging, a band-pass filter for calcein-AM (at 500−550 nm) and a long-pass filter for PI (at 590−800 nm) were used.
Bactericidal kinetics. 2 A single colony of MRSA and E. coli was inoculated separately in nutrient broth for 6 h at 37 ˚C to produce 10 8 CFU/mL to 10 9 CFU/mL cells. Next, this midlog phase bacterial solution was further diluted to ~5× 10 5 CFU/mL and 180 µL of this diluted bacterial solution in Mueller Hinton broth was added to the 20 µL aqueous solution of test compound, IAM-1 with the concentration of 16 µg/mL (16.1 µM), 32 µg/mL (32.2 µM) for MRSA and 32 µg/mL (32.2 µM) and 64 µg/mL (64.4 µM) for E. coli. The same volume of autoclaved water without the test compound was used as a control. Afterward, 20 μL of aliquots from the individual mixture of bacteria and compound were serially diluted by 10fold in sterile saline at different time points (0 h, 1 h, 2 h, 4 h, 6 h, and 12 h). Then, spot plating on agar plates was executed with 20 μL solution from each dilution and allowed to incubate for 24 h at 37 ˚C. Finally, the number of bacterial colonies was counted and results were presented in logarithmic scale, i.e., Log (CFU/mL) vs time.
Activity against stationary and persister cells. 2,3 A midlog phase (6 h grown culture) MRSA and E. coli culture was diluted to 1:1000 ratio in nutrient broth and incubated at 37 ˚C for 16 h in shaking condition to achieve stationary phase cells. Later on, the bacterial suspension was centrifuged (9000 rpm, 2 min) and resuspended in 1xPBS. On the other hand, the persister cells were generated from the stationary phase cells upon specific antibiotic exposure for 3 h. 1 mL of stationary phase culture was treated with 100 μg/mL (for S. aureus) and 300 μg/mL (for E. coli) of ampicillin sodium for 3 h at 37 ˚C. Next, the bacteria were centrifuged, washed 3-4 times, and resuspended in 1xPBS to remove the traces of the antibiotic. Finally, 180 μL of the stationary and persister phase bacteria (~5× 10 5 CFU/mL) was added to 20 μL of IAM-1 solution with the concentrations of 16 µg/mL (16.1 µM), 32 µg/mL (32.2 μM) for stationary phase MRSA and persister phase S. aureus and 32 µg/mL (32.2 µM), 64 µg/mL (64.4 μM) for both stationary and persister phase E. coli. Similarly, for vancomycin (stationary phase MRSA and persister phase S. aureus) and colistin (for both stationary and persister phase E. coli) antibiotic concentrations were 64 µg/mL (64.4 µM) and 32 µg/mL (27.7 µM) and same volume of water without compound was considered as an untreated control. After 6 h time, 20 μL aliquots from that solution were serially diluted 10fold in sterile saline. Then 20 μL solution from each dilution was spot plated on MacConkey agar plates and after 24 h of incubation at 37 ˚C, the number of bacterial colonies was Figure S2. 1 H-NMR of IAM-1. The NMR was taken in DMSO-d6 and the solvent peak was calibrated at the δ value of 2.5 ppm. . Figure S1. HRMS of IAM-1. S17 Figure S3. 13 C-NMR of IAM-1. The NMR was taken in DMSO-d 6 and the solvent peak was calibrated at the δ value of 39.52 ppm.
. Figure S4. HRMS of IAM-2. Figure S5. 1 H-NMR of IAM-2. The NMR was taken in DMSO-d6 and the solvent peak was calibrated at the δ value of 2.5 ppm. . Figure S6. 13 C-NMR of IAM-2. The NMR was taken in DMSO-d 6 and the solvent peak was calibrated at the δ value of 39.52 ppm.

S18
. S19 Figure S7. HRMS of IAM-3. Figure S8. 1 H-NMR of IAM-3. The NMR was taken in DMSO-d6 and the solvent peak was calibrated at the δ value of 2.5 ppm. . Figure S9. 13 C-NMR of IAM-3. The NMR was taken in DMSO-d 6    Laurdan dye is well known to detect the changes in membrane fluidity. When membrane disorder allows water percolation to the region just below the head groups, where the probe is located, a spectral shift and change in intensity is observed. When the liposomes mimicking bacterial membranes were treated with IAMs 1-3 [160 µg/mL (161.1 µM)] (ortho-, meta-and paraisomers) individually, a reduction of fluorescence intensity was noticed in all the cases. This suggested that bacterial model membrane was slightly perturbed by all the positional isomers upon allowing the penetration of surrounding water molecules below the lipid head groups. Interestingly, ortho-isomer, IAM-1 did not alter the emission spectrum of laurdan dye in case of mammalian model membrane, although it had a mild effect on bacterial model membrane. However, along with the perturbation of bacterial model membrane, IAM-2 (meta-isomer) and IAM-3 (para-isomer) upon interacting with mammalian model membrane decreased the fluorescence intensity.