LEGO-lipophosphonoxins: length of hydrophobic module affects permeabilizing activity in target membranes of different phospholipid composition

In the past few decades, society has faced rapid development and spreading of antimicrobial resistance due to antibiotic misuse and overuse and the immense adaptability of bacteria. Difficulties in obtaining effective antimicrobial molecules from natural sources challenged scientists to develop synthetic molecules with antimicrobial effect. We developed modular molecules named LEGO-Lipophosphonoxins (LEGO-LPPO) capable of inducing cytoplasmic membrane perforation. In this structure–activity relationship study we focused on the role of the LEGO-LPPO hydrophobic module directing the molecule insertion into the cytoplasmic membrane. We selected three LEGO-LPPO molecules named C9, C8 and C7 differing in the length of their hydrophobic chain and consisting of an alkenyl group containing one double bond. The molecule with the long hydrophobic chain (C9) was shown to be the most effective with the lowest MIC and highest perforation rate both in vivo and in vitro. We observed high antimicrobial activity against both G+ and G− bacteria with significant differences in LEGO-LPPOs mechanism of action on these two cell types. We observed a highly cooperative mechanism of LEGO-LPPO action on G− bacteria as well as on liposomes resembling G− bacteria. LEGO-LPPO action on G− bacteria was significantly slower compared to G+ bacteria suggesting the role of the outer membrane in affecting the LEGO-LPPOs perforation rate. This notion was supported by the higher sensitivity of the E. coli strain with a compromised outer membrane. Finally, we noted that the composition of the cytoplasmic membrane affects the activity of LEGO-LPPOs since the presence of phosphatidylethanolamine increases their membrane disrupting activity.


Supplementary Figure S3.
Representative dynamic light scattering (DLS) spectra for the two sizes of PE:PG liposomes (10 µM).A) Size distribution of LUV100 (a 100 nm filter used for extrusion) where a single peak was observed, suggesting quite narrow range of liposome diameter of about 120 nm, with polydispersity index (PDI) under 0.1 indicating reliable evaluation of the data.B) The size distribution analysis of the larger LUV1000 liposomes (when a 1000 nm filter was used for extrusion) showed broader distribution of intensities but still with a main significant peak around 600 nm diameter (PDI increased to about 0.347).The Z-average (340 nm) is then less reliable.

DLS measurement
Size distribution of liposomes was measured by DLS using a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, UK) operating with a He-Ne ion laser (λ = 633 nm).The detection angle was 173 degrees.The measurements were performed in disposable polystyrene cuvettes at 25°C using buffer composed of 100 mM NaCl, 0.5 mM Na₂EDTA and 5 mM HEPES, pH 7.4.Three runs of measurements were performed for each sample.Malvern Zetasizer software was used for collecting data and their analysis.The size distribution, the hydrodynamic diameter, and the polydispersity index (PDI) were acquired from the autocorrelation fit of the data.Additionally, the volume distributions were converted from the size distributions by the software.in the range up to 200 pS (Supplementary Figure S6, A).Higher concentrations (2.5 mg/L and 5 mg/L) induce the occurrence of high conductance pores: ~600 pS, 1000 pS and 1500 pS, respectively (Supplementary Figure S6, B and C).In line with the membrane permeabilization assay our results indicate that C9 molecules act in a highly cooperative manner and could form narrow pores with low conductance.However, C9 can probably assemble into large oligomers causing membrane instability of high conductance.

Planar Lipid Membrane Experiments -Method
Experiments were performed in a Teflon chamber divided into two compartments by a diaphragm with a circular aperture of approximately 0.5 mm in diameter.Planar lipid bilayers were formed across the aperture with a solution of 3% E. coli lipids (E. coli Polar Lipid Extract, Avanti Polar Lipids) in n-decane/butanol (9:1, v/v).Both chamber compartments contained 1.5 mL of 1 M KCl, 10 mM Tris, pH 7.4.The temperature was kept at 25 °C.LEGO-LPPO was added to the trans compartment of the chamber in the concentration of 1.25, 2.5 or 5.0 mg/L.Membrane current was registered with Ag/AgCl electrodes with applied membrane voltage of 50 mV (trans negative), amplified by an LCA-200-100G amplifier (Femto), and digitized by a KPCI-3108 card (Keithley).
Recorded signal was processed with QuB software 6 .represents non-selective "graded" leakage, while the other values usually signify preferential graded leakage of DPX + or ANTS -molecules.
In our experiment, we tested selectivity of the most active LEGO-LPPO molecule C9 on PE:PG liposomes in a wide concentration range (0.1-6 mg/L).Our results presented in Supplementary Figure S5 show graded DPX + preferential leakage (α~2).Such a value is usually explained by formation of transient membrane pores that allow DPX + flux after its accumulation in the proximity of negatively charged membranes 8 .

Mechanism of dye leakage from liposomes -Requenching Method
The requenching method was used to differentiate the mode of liposome leakage caused by LEGO-LPPOs.Briefly, using the method described in detail in Ladokhin et   al., (1995)  9 one measures the dependence of ANTS -quenching inside the vesicles (Q in ) as a function of the fraction of ANTS -that has leaked out of the vesicles (f out ) after incubation with membrane-active compounds.The suspension of LUV 100 diluted to 10 μM final phospholipid concentration and loaded with fluorophore/quencher pair ANTS -/DPX + (2 mL in quartz cuvettes) were treated with the range of LEGO-LPPO concentrations.Tested substance (C9) was added to suspension from a diluted stock (0.1 mg/L) in large volumes (~600 μL), to avoid localized artifacts (immediate leakage induced by locally high concentration) that could affect the data interpretation.The buffer was used to compensate for the differences in volumes of individual additions.
Each addition was followed by fast mixing and the samples were then incubated for an hour to reach the plateau level of fluorescence.Then, 30 µL of concentrated DPX + was added into the individual samples (158 mM final concentration) for determination of total quenching followed by adding 5 µL of 10% Triton X-100.Fluorescence intensities were recorded using FluoroMax-3 spectrofluorometer (Jobin Yvon, Horiba) as follows: Emission spectra (in the range of 420-600 nm, excitation at 360 nm) were measured using excitation and emission filters (365WB50 and 3RD410LP, Omega Optical) with bandwidth of 8 nm.Intensities in the range 500-510 nm were integrated for data analysis after background subtraction.Obtained intensity values were used for calculation of quenching inside of liposomes (Q in ), fraction of ANTS -outside (f out ) and a value of selectivity parameter , according to Ladokhin et al., (1995)  9 .
Figure S4.Leakage of CF from liposomes induced by LEGO-LPPOs.Curves show concentration dependent percentage of leakage from liposomes over time (30 min).Liposomes were composed of phospholipids in 2:1 ratio-PE:PG, PG:CL, PG:PE.Maximum leakage (100%) was achieved using 0.1% Triton X-100.Representative kinetics from two independent liposome preparations are shown.The data indicate that the activity of the LEGO-LPPOs depends dramatically on the liposome composition and to a certain extent on the length of the hydrophobic chain.