Escherichia coli (25922), Pseudomonas aeruginosa (15159), and Staphylococcus aureus (29213) were purchased from the American Type Culture Collection (ATCC).

LPS from Escherichia coli (serotype 0111:B4, cat#L3024) was purchased from Sigma-Aldrich.

It was reported that the conjugated β-diketone causes instability and metabolic liabilities in natural curcumins; therefore, this substructure was replaced by a single ketone in the studied derivates (26). Two curcuminoids were designed as (2-pyridyl)methylidene and (4-hydroxy-3-methoxyphenyl)methylidene cyclopentanones (Figure 1). Curcuminoids (Py-cPen and V-cPen) were synthesized in the laboratory of the medicinal chemistry group of the First Medical Faculty of Charles University BIOCEV, Vestec, Czech Republic, following a prescribed synthesis (27, 28).

Cathelicidin antimicrobial peptide LL37 was purchased from Innovagen (United States) (29, 30).

THP-1-XBlue-CD14 reporter monocytes (InvivoGen, United States) were cultured in RPMI 1640-GlutaMAX-1 (Gibco, Life Technology Ltd., Renfrew, UK), and the media was supplemented with 10% (v/v) heat-inactivated FBS (FBSi, Invitrogen, Waltham, MA, United States) and 1% (v/v) antibiotic-antimycotic solution (AA, Invitrogen) at 37°C in 5% CO₂.

RAW 264.7 (InvivoGen) cells were cultured in DMEM (HyClone, GE Healthcare Life Science, United States) supplemented with 10% (v/v) heat-inactivated FBS (FBSi; Invitrogen, United States) and 1% (v/v) antibiotic-antimycotic solution (AA; Invitrogen) at 37°C in 5% CO₂.

The potential antibacterial activity of curcuminoids (Py-cPen and V-cPen) on E. coli, P. aeruginosa, and S. aureus strains was explored by incubating one colony overnight in 5 mL of Todd-Hewitt (TH) medium. The next morning, the bacterial culture was refreshed and grown to a mid-logarithmic phase (OD_620nm = 0.4). The bacteria were then centrifuged, washed, and diluted 1:1,000 in 10 mM Tris buffer at pH 7.4 to obtain an approximate concentration of bacteria amounting to 2 × 10⁶ CFU/mL. Next, 50 μL of bacterial suspension was incubated with curcuminoids (Py-cPen 5, 10, and 30 μM and V-cPen 1, 3, and 10 μM), LL37 (1 μM) (used as a positive control), or a buffer control (10 mM Tris buffer at pH 7.4) for 2 h at 37°C. After 2 h, serial dilutions of the samples were plated on TH agar plates, incubated overnight at 37°C, and followed by colony counting the next day (31, 32).

The viability of E. coli, P. aeruginosa, and S. aureus was assessed using the LIVE/DEAD^® BacLight^™ Bacterial Viability Kit (Invitrogen, Molecular Probes, Carlsbad, CA, United States). Bacterial suspensions were prepared for VCA as described above. Bacterial strains were treated with curcuminoids (Py-cPen 5, 10, and 30 μM and V-cPen 1, 3, and 10 μM) and LL37 (1 μM). After 2 h of incubation at 37°C, the samples were mixed 1:1 with the dye mixture, followed by incubation for 15 min in the dark at room temperature. The dye mixture was prepared according to the manufacturer’s protocol, i.e., 1.5 μL of component A (SYTO-9 green-fluorescent nucleic acid stain) and 1.5 μL of B (red-fluorescent nucleic acid stain propidium iodide) were dissolved in 1 mL of 10 mM Tris at pH 7.4. A measure of 5 μL of stained bacterial suspension was trapped between a slide and an 18-mm square coverslip. Overall, 10 view fields (1 × 1 mm) were examined from three independent sample preparations using a Zeiss AxioScope A.1 fluorescence microscope (objectives: Zeiss EC Plan-Neofluar 20×; camera: Zeiss AxioCam MRm; and acquisition software: Zeiss Zen 2.6 [blue edition]) (31, 32).

Molecular docking was performed using AutoDock Vina software. Only docking poses with a free energy of binding lower or approximately −5.5 kcal/mol were considered, which corresponds to an approximate K_i value of 0.1 mM or lower. The calculations were carried out using the parameters recommended in the user manual. AutoDock tools were used to find and determine the center and size of the grid box for the docking calculations.

The 3D structure of LPS was obtained by modeling two parts of a molecule from the PDB Data Bank using COOT software from the CCP4 package. The base part (lipid A) was obtained with ID 6BPP, and the O-antigen part was obtained with ID 7ML5. PDB IDs were downloaded, repaired, and saved, using the UCSF Chimera software, as .pdb file. The UCSF Chimera software was also used to remove unfavorable compounds such as water, solvents, and ligands so that only the molecule of LPS remained.

After that, both structures were joined using COOT software and edited using Avogadro software for the complete creation of bonds between O-antigen and lipid A parts.

The 3D structures of drugs (ligands) were obtained from the Jakubek’s Lab internal database. Atoms with double conformations were checked and repaired using a self-written script in Python programming language.

The predicted binding affinity (kcal/mol) was calculated by using AutoDock Vina software. To visualize molecules and analyze the docking results, two pieces of software were used. The PyMOL 3.0 software (by Schrödinger) was used to verify the positioning of the ligand on the receptor surface and to create a 3D structure of the complex. To generate overall views of the docking outcomes, UCSF Chimera was used, while BIOVIA Discovery Studio Visualizer (by Dassault Systèmes) was used to detect and illustrate the type of interactions of ligands (Py-cPen and V-cPen) with LPS.

Curcuminoids were visualized using TEM (Jeol Jem 1,230; Jeol, Tokyo, Japan) in combination with negative staining after incubation with LPS or buffer. Images of endotoxin LPS (10 μM) in the presence or absence of Py-cPen (10 μM) or V-cPen (3 μM) were taken after incubation for 30 min at RT. For the mounted samples, 10 view fields were examined on the grid (pitch 62 μm) from three independent sample preparations. The samples were adsorbed onto carbon-coated grids (copper mesh, 400) for 60 s and stained with 7 μL of 2% uranyl acetate for 30 s. The grids were rendered hydrophilic via glow discharge at low air pressure. The size of LPS was analyzed as the mean gray value/μm² ± SEM using an external software ImageJ 1.52 k. All images were converted to 8-bit (grayscale), and the threshold was adjusted accordingly (31).

NF-κB/AP-1 activation in THP-1-XBlue-CD14 reporter monocytes was determined after 20–24 h of incubation according to the manufacturer’s protocol (InvivoGen). Briefly, 1 × 10⁶ cells/mL in RPMI were seeded in 96-well plates (180 μL) and treated and incubated with curcuminoids (Py-cPen 10, 15, 20, and 30 μM and V-cPen 3, 5, 8, and 10 μM) or LL37 (1 μM), LPS (10 ng/mL) or both overnight at 37°C, 5% CO₂ in a total volume of 200 μL. The following day, the activation of NF-κB/AP-1 was analyzed as the secretion of embryonic alkaline phosphatase (SEAP). The supernatant (20 μL) from the cells was transferred to 96-well plates, and 180 μL of Quanti-Blue was added. The plates were incubated for 2 h at 37°C, and the absorbance was measured at 600 nm using a VICTOR3 Multilabel Plate Counter spectrofluorometer (17, 31).

A volume of 20 μL of sterile filtered MTT solution (3-(4,5-dimethylthiazolyl)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich) at 5 mg/mL in PBS was added to the remaining overnight culture of THP-1-XBlue-CD14 reporter monocytes from the above NF-κB activity assay in 96-well plates, which were incubated at 37°C (see above). After 2 h of incubation at 37°C, the supernatant was removed and the blue formazan product generated in cells was dissolved by adding 100 μL of DMSO to each well. The plates were then gently shaken for 10 min at room temperature to dissolve the precipitates. The absorbance was measured at 550 nm using a VICTOR3 Multilabel Plate Counter spectrofluorometer.

The macrophage cell line RAW 264.7 (passage 4–7) in DMEM was seeded in 96-well tissue culture plates (8 × 10⁴ cells per well) overnight at 37°C in a 5% CO₂ atmosphere. Curcuminoids (Py-cPen 1, 5, 10, and 15 μM and V-cPen 1, 1.5, 3, and 5 μM) or LL37 (1 μM) were mixed with and without FITC-LPS (100 ng/mL). The pre-mixed samples (20 μL) were added to the adherent RAW 264.7 cells. To measure phagocytosis, the samples were incubated with the cells for 1 h at 37°C and washed twice using DMEM media. Then, fluorescence was measured using a VICTOR3 Multilabel Plate Counter spectrofluorometer (PerkinElmer, United States) at excitation/emission wavelengths of 530/570 nm. The baseline uptake (of only media) was subtracted from the signal of each sample. For analysis using fluorescence microscopy, RAW 264.7 (passage 4–7) cells were seeded in 24-well tissue culture plates with inserted coverslips (2 × 105 cells per well) overnight at 37°C in a 5% CO₂ atmosphere. Then, the cells were treated as described above. After 1-h incubation at 37°C, the cells were stained with 300 ng/mL DAPI (Sigma-Aldrich, United States) for 3 min at room temperature, washed with PBS, and mounted onto glass slides using a mounting medium (Mountant, PermaFluor, Thermo Fisher Scientific, KU). In total, 10 view fields (1 × 1 mm) were examined from three independent sample preparations using a Zeiss AxioScope A.1 fluorescence microscope (objectives: Zeiss EC Plan-Neofluar 20×; camera: Zeiss AxioCam MRm; acquisition software: Zeiss Zen 2.6 [blue edition]) (17).

The graphs of VCA, TEM analysis, NF-κB activity, and phagocytosis are presented as the mean ± SEM from at least four independent experiments. Differences in these assays were assessed using a one-way ANOVA with Dunnett’s multiple comparison tests. All data were analyzed using GraphPad Prism 10 (GraphPad Software, Inc., La Jolla, CA, United States). In addition, p-values less than 0.05 were considered to be statistically significant (ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

The antibacterial activity of curcuminoids Py-cPen and V-cPen against bacterial strains of E. coli (25922), P. aeruginosa (15159), and S. aureus (29213) was analyzed using VCA (Figure 2).

A bacterial growth reduction was observed with the application of both curcuminoid compounds. Py-cPen significantly decreased growth in all bacterial strains at concentrations of 10 and 30 μM. For Gram-negative strains such as E. coli and P. aeruginosa, Py-cPen at 10 μM caused approximately a 20% reduction, while at 30 μM, it led to a 30% reduction in growth for both strains.

V-cPen at 3 μM resulted in approximately a 30% reduction in both Gram-negative strains. At 10 μM, V-cPen significantly reduced growth by approximately 40% in E. coli and 60% in P. aeruginosa.

For the Gram-positive strain S. aureus, both curcuminoids caused a notable decrease in growth. Py-cPen led to nearly a 50% reduction at 10 μM, while V-cPen achieved an approximately 80% reduction at just 3 μM. TH agar plates illustrated a decline in colony numbers after treatment with Py-cPen and V-cPen across all bacterial strains (Supplementary Figures S1A–C). Cathelicidin antimicrobial peptide LL37 was used as a positive control to compare the effects of bacterial killing efficiency.

The VCA data were confirmed using fluorescence microscopy with LIVE/DEAD staining (Figure 3). Dead bacteria with compromised membrane integrity were stained red, and live bacteria with intact membranes were stained green. The results indicated the presence of dead bacteria following treatment with 10 μM of Py-cPen and 3 μM of V-cPen. Additional experiments were conducted with Py-cPen at concentrations of 5 and 30 μM and using V-cPen at 1 and 10 μM. The data revealed a similar trend to that observed with VCA, showing an increase in killing ability with higher concentrations of both compounds (Supplementary Figures S2A,B).

Molecular Docking with AutoDock Vina revealed that both curcuminoids (Py-cPen and V-cPen) exhibited favorable and comparable docking scores (−5.913 kcal/mol and −6.823 kcal/mol, respectively). This suggests their potential binding affinity for the aliphatic side chain of lipid A, which is a hydrophobic anchor and the most bioactive component of LPS (Figures 4A,B). It should be noted that V-cPen also showed high binding affinity in the part of LPS, where core and O-antigen are joined (−6.711 kcal/mol) (Figure 4C). This suggests that both curcuminoids have the potential to modulate LPS activity through their high binding activity. Thus, the results suggest that both Py-cPen and V-cPen exhibit a potential binding affinity to LPS, indicating their potential to modulate LPS activity.

Using BIOVIA Discovery Studio Visualizer, key interactions in the complexes of LPS with Py-cPen and V-cPen were detected. In the figures, the left side (generated using UCSF Chimera) displays the overall molecular structure, emphasizing the positioning of Py-cPen or V-cPen within the LPS molecule. Meanwhile, the right side (generated using BIOVIA Discovery Studio) provides a close-up view of these specific interactions.

As shown in Figure 4A, the visualization of Py-cPen reveals its interaction with the aliphatic side chain of the lipid A component of LPS. The primary interaction identified is a π-alkyl interaction, where the π-electron system of the pyridine group in Py-cPen engages with the aliphatic side chains of the glyceride. This hydrophobic interaction stabilizes the complex, while weak electrostatic forces between the π-cloud of the aromatic side group of curcumin and the glyceride alkyl chains contribute to the overall binding affinity (−5.913 kcal/mol). Notably, the π-alkyl interactions between the pyridine group of Py-cPen and the aliphatic chain of LPS are more centrally positioned within the glyceride region compared to V-cPen.

Figure 4B illustrates the interaction between the aromatic rings of guaiacol (ortho-methoxyphenol) in V-cPen and the aliphatic side chain of the lipid A portion of LPS, similar to the interactions observed with Py-cPen. However, V-cPen establishes additional interaction types with the lipid A component of LPS. In addition to π-alkyl interactions, V-cPen engages in π-sigma and alkyl interactions, enhancing its binding affinity to the hydrophobic region of lipid A (−6.823 kcal/mol) compared to Py-cPen. The π-sigma interaction occurs between the π-electron system of the guaiacol group in V-cPen and the σ-electron system of the aliphatic chain of lipid A. The π-alkyl interactions are situated more toward the periphery of the aliphatic region of the molecule, in contrast to the more central positioning observed with Py-cPen. The third type of interaction, alkyl, represents a hydrophobic interaction between the methoxy group of the ortho-methoxyphenol in V-cPen and the aliphatic side chain of lipid A.

Finally, Figure 4C depicts the affinity (−6.711 kcal/mol) between V-cPen and the polysaccharide part (O-antigen) of LPS. This visualization highlights three key interactions. A conventional hydrogen bond is formed between the oxygen atom of the methoxy group in the guaiacol side group of V-cPen and a hydrogen atom within the polysaccharide region of LPS. In addition, a carbon–hydrogen bond is observed, where the hydroxy group from the guaiacol group of V-cPen bonds with a carbon atom in the polysaccharide region. Finally, a π-donor hydrogen bond is established in which the π-electron cloud of the guaiacol acts as a donor for a hydrogen bond with the hydroxy group from the polysaccharide components of LPS.

One possible explanation for why Py-cPen and V-cPen are active at different concentrations is the interaction of V-cPen with two parts of the molecule LPS (lipid A and part where core and O-antigen are joined), whereas Py-cPen interacts only with the lipid A part of LPS.

TEM analysis was performed to further examine the ability of curcuminoids to bind to LPS and alter its structural properties (Figure 5). The TEM images confirmed that the ribbon-like structure of LPS was significantly diminished upon the addition of Py-cPen (10 μM) and V-cPen (3 μM).

We investigated the effect of curcuminoids (Py-cPen and V-cPen) on LPS signaling to determine NF-κB /AP-1 activation in THP-1XBlue-CD14 reporter monocytes (Figures 6A,B). Py-cPen (10–30 μM) inhibited NF-κB activation triggered by E. coli LPS. In contrast, V-cPen (3–10 μM) enhanced LPS-induced NF-κB activation. Neither of the curcuminoids showed a significant cytotoxic effect on THP-1XBlue-CD14, suggesting that the alteration of NF-κB activation was not due to toxic effects on the cells but to the effect of the curcuminoids themselves. LL37 peptide was used as a positive control.

Furthermore, a phagocytosis assay using fluorescence detection was conducted to investigate the effect of curcuminoids on uptake by RAW 264.7 cells (Figure 7B). RAW 264.7 cells were treated with Py-cPen (1–15 μM) and V-cPen (1–5 μM) or their mixtures with LPS (100 ng/mL). Both Py-cPen (1–15 μM) and V-cPen (1.5–5 μM) significantly increased FITC-LPS uptake by RAW 264.7 cells. LL37 was used as a positive control (reference). To confirm the uptake of LPS by RAW 264.7 cells, fluorescence microscopy imaging of cells stained with DAPI and FITC-LPS was performed (Figure 7A). RAW 264.7 cells were treated as described above but used only one concentration for Py-cPen (10 μM) and V-cPen (3 μM). This technique enabled the visualization of cells stained with DAPI and LPS labeled with FITC.