Skip to main content
NCBI home page
Frontiers in Pharmacology logoLink to Frontiers in Pharmacology
. 2026 Sep 24;17:1895729. doi: 10.3389/fphar.2026.1895729

Targeting the AcrAB-TolC efflux pump by small molecules selectively addressing the outer membrane factor TolC

Janis Veliks 1,†, Cristina D Cruz 2,†, Tania Szal 3,4,†, Marina Madre 1, Irena Leite 1, Laura Paunina 1, Philipp Lewe 5, Maren Hinz 5, Susanne Witt 5, Päivi Tammela 2, Aigars Jirgensons 1, Björn Windshügel 3,4,*
PMCID: PMC13645612  PMID: 42851487

Abstract

Background

Gram-negative bacteria express several efflux pump complexes extruding a broad spectrum of antibiotics. The major efflux pump in E. coli is the AcrAB-TolC complex that expels a broad spectrum of antibiotics from the bacterium, resulting in impaired efficacy of these drugs. Therefore, the inhibition of efflux pump function is an attractive approach to restore the susceptibility of Gram-negative bacteria for antibiotics.

Methods

A virtual screening campaign was utilized to identify potential TolC binders. Virtual hits were investigated using a functional efflux assay, followed by synergy tests with piperacillin and determination of the intrinsic antibacterial potential. Bindings studies with individual efflux pump proteins were carried out. Synthetic chemistry was applied to improve the efficacy of the compound. Microbiological and biophysical tests were used to characterize the compounds with respect to synergistic activity with antibiotics and efflux pump binding.

Results

We identified a small molecule interfering with efflux pump function, showing synergistic activity with piperacillin and binding to TolC which was further optimized at various positions. The most promising compounds revealed TolC-selective binding and shifted the minimum inhibitory concentration of piperacillin and levofloxacin at low micromolar concentrations.

Conclusion

Our study provides evidence that TolC may represent a valid target for the development of an antibiotic adjuvant.

Keywords: AcrAB-TolC, antimicrobial resistance, E. coli, efflux pumps, outer membrane factor, TolC

1. Introduction

Antimicrobial resistance (AMR) is a global threat with tremendous implications for human and animal health (Dadgostar, 2019). Infections with drug-resistant bacteria are estimated to contribute to nearly five million deaths every year (Antimicrobial Resistance Collaborators, 2022) and are predicted to be responsible for 39.1 million deaths between 2025 and 2050, an immense number that could be decreased by approximately one-third if a new Gram-negative targeting antibiotic is developed (Naghavi et al., 2024). Therefore, there is an urgent need to develop new therapeutic approaches to treat infections caused by multidrug-resistant (MDR) Gram-negative bacteria (Theuretzbacher and Piddock, 2019). The increasing prevalence of AMR has stimulated the research on adjuvants that directly address bacterial resistance mechanisms. Such drugs could revive the efficacy of already available antibiotics, for example, membrane permeabilizers, efflux inhibitors, or β-lactamase inhibitors (Wright, 2016; Annunziato, 2019; Laws et al., 2019).

Efflux systems play a major role in AMR of Gram-negative bacteria, several of which are listed on the World Health Organization priority pathogens list for R&D of new antibiotics (Santajit and Indrawattana, 2016; Tacconelli et al., 2018; Sati et al., 2025). In particular, the overexpression of efflux pumps belonging to the resistance-nodulation-cell division (RND) superfamily is a major factor of AMR (Yasufuku et al., 2011), causing sub-inhibitory antibiotic levels, thereby impairing therapeutic efficacy. In addition, this may promote the development of target-based resistance (Cernicchi et al., 2021). Indeed, efflux pump overexpression correlates with the development of other resistance mechanisms to fluoroquinolones due to the coupled lower expression of the DNA mismatch repair gene mutS (El Meouche and Dunlop, 2018). Therefore, the development of efflux pump inhibitors (EPIs) is a promising strategy to overcome AMR and to restore the susceptibility of MDR bacteria to existing and future antibiotics by preventing their extrusion (Bolla et al., 2011). Due to the reported roles of efflux systems in biofilm formation and virulence, EPIs could also be employed as antibiofilm and antivirulence agents (Alav et al., 2018; Reza et al., 2019; Piddock, 2006).

The main efflux system in E. coli is the tripartite complex AcrAB-TolC, which actively transports antibiotics from the bacterium into the external medium (Li et al., 2015). The complex is composed of the RND-type multidrug efflux pump AcrB, the periplasmic adapter protein AcrA, and the outer membrane (OM) factor TolC (Du et al., 2014). TolC is a homotrimer embedded in the OM with a 12-stranded β-barrel and a periplasmic domain composed of 12 α-helices forming an α-barrel that extends approximately 100 Å into the periplasmic space (Koronakis et al., 2000). In the absence of other efflux system components, TolC remains closed, whereas after formation of the tripartite complex TolC opens by an iris-like movement of the helices at the periplasmic tip (Andersen et al., 2002a; Sjuts et al., 2016). In addition to AcrA and AcrB, TolC can be associated with several RND- and non-RND efflux systems (e.g., MdtEF, MacAB, EmrAB) and its deletion has been shown to result in increased susceptibilities of E. coli for several antibiotics (Alav et al., 2021; Sulavik et al., 2001).

Various classes of antimicrobials have been identified as AcrAB-TolC substrates, including chloramphenicol as well as different macrolide and β-lactam antibiotics (Bolla et al., 2011). Therefore, targeting the AcrAB-TolC efflux system represents an attractive strategy for pharmaceutical intervention to reconstitute the effect of antibiotics (Annunziato, 2019). Pyranopyridine derivatives and pyridylpiperazine-based compounds have been reported as AcrB inhibitors, thereby decreasing the minimum inhibitory concentration (MIC) of the antibiotics levofloxacin and piperacillin (Sjuts et al., 2016; Plé et al., 2022). Similarly, several EPIs targeting AcrA have been identified that shift the MIC of different antibiotics in E. coli (Haynes et al., 2017; Abdali et al., 2017). The colicin E1-TR fragment has been demonstrated to plug the TolC channel and to shift the MIC of erythromycin, ciprofloxacin, and benzalkonium chloride (Budiardjo et al., 2022). Inorganic cations such as hexaamminecobalt have been shown to block TolC by interacting with conserved aspartates (Asp371, Asp374) within the TolC lumen close to the periplasmic tip (Andersen et al., 2002b; Higgins et al., 2004), although lacking synergistic activity with antibiotics (Andersen et al., 2002a; Gilardi et al., 2017). Recently, we reported the first small molecule (LP-115) targeting the AcrA-TolC interface resulting in the disassembly of the AcrAB-TolC complex in vitro (Szal et al., 2025). Small organic molecules interfering with TolC function or its homologs in other Gram-negative bacteria might be further developed into adjuvants that restore the sensitivity of bacteria for antibiotics (Phan et al., 2015).

In this study we report the discovery and optimization of compounds that interfere with antimicrobial efflux. Based on a virtual screen against E. coli TolC, several compounds were identified, one of which shifted the MIC of piperacillin and levofloxacin. Using comprehensive medicinal chemistry efforts, microbiological assays, and binding studies, the initial hit was further optimized, resulting in two compounds selectively binding to TolC.

2. Materials and methods

2.1. Molecular modeling

A subset of approximately 5.9 million compounds was downloaded from the ZINC database (Irwin et al., 2020) based on the following selection criteria: (i) molecular weight >375 Da (based on internal data of TolC-binding compounds), (ii) logP −1 to 5 (−1 lowest possible logP in ZINC tranche browser, five based on Lipinski et al., 2001), (iii) reactivity: standard, (iv) purchasability: agent, (v) pH: reference and mid, (vi) formal charge 0 to +2 (negative charge associated with poor membrane permeability, positive charge limited to +2 to limit membrane binding). Using the molecular modeling suite Molecular Operating Environment (MOE) version 2019.01.02 (Chemical Computing Group Inc., Montreal, Canada), the set was further filtered (logS > -4.5, modified from Jorgensen and Duffy, 2002; compounds containing a positively charged nitrogen atom), resulting in a final library of 653,617 compounds.

E. coli TolC X-ray crystal structures (PDB IDs 1TQQ and 1EK9) were downloaded from the Protein Data Bank (Koronakis et al., 2000; Higgins et al., 2004; wwPDB consortium, 2019). All non-protein atoms were removed. Both TolC structures were superposed (Cα atoms) and hydrogen atoms were added using MOE, followed by energy minimization (Amber14:EHT force field with R-Field implicit solvation model) using a three-step procedure (Cα atoms, backbone, all atoms), each step including a tether of 0.5 and an RMS gradient of 5. Compounds were docked into the periplasmic tip area of both structures using GOLD version 1.8 (Cambridge Crystallographic Data Centre, Cambridge, UK). The search space was defined by a sphere of 15 Å radius centered within the small TolC opening at the periplasmic tip. For each compound 10 docking runs were conducted. The early termination option was switched off.

2.2. Bacterial strains and reagents

Clinical control strain E. coli BW25113 (parental strain) and the single-gene knock-out mutant strains E. coli JW5503 (ΔtolC) and E. coli JD17464 (ΔlpxC) were obtained from the E. coli collection of the National BioResource Project at the National Institute of Genetics (Japan) (Baba et al., 2006). Uropathogenic E. coli UMN026, selected as model for a multi-resistant and clinically relevant strain (Manges et al., 2001). was obtained from the Leibniz Institute DSMZ German Collection of Microorganisms and Cell Culture. Piperacillin sodium salt, phenylalanine-arginine β-naphthylamide dihydrochloride (PAβN), bis-benzimide Hoechst 33,342 trihydrochloride (H33342), levofloxacin and mefloquine hydrochloride were purchased from Sigma.

2.3. Hoechst 33342 accumulation assay

The assay was conducted as previously described (Coldham et al., 2010). Compounds were added to the bacterial suspension to achieve a final concentration of 100 µM. PAβN was used as positive control (25 μg/mL). E. coli JW5503 cells can naturally and rapidly accumulate H33342 intracellularly, thus showing a maximum fluorescence. The accumulation obtained in E. coli BW25113 treated with compounds and PAβN was compared to the accumulation observed in JW5503 and calculated as percentage values (E. coli JW5503 was assumed as 100%). A compound was considered a contributor to the increased intracellular H33342 concentration if it reached an equivalent of 80% of that observed in E. coli JW5503. H33342 efflux was monitored using a fluorescence microplate reader (VarioskanLUX, Thermo Fisher Scientific) to measure the fluorescence (excitation 355 nm, emission 460 nm) at 37 °C every 5 min for 30 min. The experiment was carried out once with two technical replicates for each compound. Further details are available in the Supplementary Material.

2.4. Antibiotic potentiation assay and antimicrobial activity tests

The assays were carried out by the broth microdilution method in 96-well plate format according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI, 2015). MIC values were determined according to CLSI guidelines and the definition used in our assay corresponds to ≥90% growth inhibition as determined spectrophotometrically. For the potentiation activity, an equal volume of bacterial suspension and compound solution, diluted into Cation-Adjusted Mueller Hinton Broth (CAMHB) containing piperacillin (final concentrations of 1 μg/mL) was used. The intrinsic MIC of each compound against E. coli BW25113 was determined using a similar assay, with the exception that the serial dilutions of compounds did not contain antibiotics. PAβN at 50 μg/mL was used as positive control in potentiation assays, while piperacillin alone (4 μg/mL) for antibacterial activity of compounds. Compounds were tested at concentrations ranging from 125 to 0.39 µM. Two independent experiments with three technical replicates each were carried out. MIC determinations for PAβN were also carried out to ascertain that the concentration used was not inhibitory in itself (Supplementary Table S3). Further details are available in the Supplementary Material.

For the most promising compounds (MPC4 ≤6.25 μM, and TolC-binding) similar assays as described above, were conducted to test the ability to potentiate the activity of levofloxacin. Similarly, compounds were assessed in clinical multi-resistant uropathogenic E. coli UMN026 strain for the potentiation of piperacillin. Mefloquine hydrochloride at 3.2 μg/mL combined with 128 μg/mL of piperacillin and piperacillin alone at 512 μg/mL were used as a positive control on every assay plate for MPC4 and MIC determinations. Two independent experiments with three technical replicates each were carried out.

2.5. Protein expression and purification

TolC, AcrA, and AcrB expression and purification were conducted according to a previously reported protocol (Szal et al., 2025; Szal et al., 2023). Tobacco, etch virus protease (TEVp) was kindly provided by the Protein Production Core Facility at CSSB. The pET24a based plasmid for the expression of a C-terminally His-tagged full-length TolC was kindly provided by Mathias Winterhalter, Constructor University, Bremen, Germany, and E. coli C43 (DE3) ΔacrAB was kindly provided by Martin Pos, Goethe University, Frankfurt, Germany.

2.6. Protein-ligand binding studies

Microscale thermophoresis (MST) studies with AcrA, AcrB, or TolC and test compounds were carried out in quintuplicates on a Monolith NT.115 LabelFree device (NanoTemper Technologies) at the Sample Preparation and Characterization (SPC) facility at EMBL Hamburg, as previously described (Szal et al., 2023; Szal et al., 2025). The compounds were titrated in a two-fold (1a) or three-fold (2c, 2f, 4b, 4c, 4d, 4e) dilution series in the respective buffer, and the final DMSO concentration was 4% (2c, 2f, 4b, 4c, 4d, 4e) or 0% (1a) throughout all samples of the dilution series. The highest compound concentrations in the assay were limited by solubility constraints or unspecific quenching of the initial fluorescence (for MST data analysis). Compound binding to TEVp was assessed in quintuplicates for compounds that did not bind TEVp (1a, 4b, 4c, 4d, 4e), and in duplicate (4f) or once (4c) for TEVp-binding compounds.

MST data points that showed irregular capillary shape or differed by more than 20% in the initial fluorescence were excluded (López-Méndez et al., 2021). Initial fluorescence (IF) data points with irregular capillary shape or MST curves were removed. IF values or MST signals at a certain time point (specified at the tables or graphs) were fitted with the ThermoAffinity tool (https://spc.embl-hamburg.de/) (Burastero et al., 2021) using a one-site fitting equation, and plotted against the concentrations of the compound using GraphPad Prism 10.0.2 (GraphPad Software, Inc.). Compounds with curves that could not yield a 95% confidence interval (Paketurytė et al., 2021) for the KD estimation or for which fitting parameters could be estimated were interpreted as not binding.

Samples for nano differential scanning fluorimetry (nDSF) using a Prometheus (NanoTemper Technologies) were prepared in a similar manner to that for the MST studies (identical protein and compound concentration, dilution series and the buffer composition). After incubation, the samples (triplicates) were loaded into Prometheus High Sensitivity Capillaries (NanoTemper Technologies). The samples were heated 1 °C per minute in a temperature range adapted to the melting temperature of TolC (TM = 75 °C–79 °C). For KD determination, the melting curves were analyzed with the FoldAffinity tool (Niebling et al., 2021) using the “Local” equation for parameters (i.e., TM estimation, unfolded fraction). The TM shift was fitted using the One-site model. Isothermal fitting was conducted close to the TM of TolC without compound (78 °C) using the one-site model. Data points that exhibited irregular melting curves were excluded. The data points and curves were plotted against the concentrations of the compound using GraphPad Prism 10.0.2 (GraphPad Software, Inc.).

2.7. Chemistry

Reagents and starting materials were obtained from commercial sources and used as received. Starting materials and dry solvents (1,4-dioxane, MeCN, DMF, DMSO) were purchased from Fluorochem, SigmaAldrich, Acros, or Alfa Aeser and used as received. Dry THF, Et2O and DCM were obtained from dry solvent still. Flash column chromatography was carried out using silica gel (35–70 µm). Thin layer chromatography (TLC) was performed on silica gel using Merck TLC Silica gel 60 F254 Aluminum sheets and was visualized by UV lamp. NMR spectra were recorded on 300 or 400 MHz Bruker spectrometers with chemical shift values (δ) in parts per million using the residual solvent signal as an internal standard. HRMS analyses were performed on a hybrid quadrupole time -of -flight mass spectrometer equipped with an electrospray ion source. The purity of all the tested compounds was estimated by HPLC (Empower three or SHIMADZU LabSolutions, UV detection at λ = 210 nm) analysis on an Adamas C18 column size: 4.6 × 150 mm, mobile phase: MeCN - 0.1% H3PO4, flow rate: 1.0 mL/min, detector: UV 210 nm and UV 254 nm column temperature: 40 °C, sample concentration: 0.5 mg/mL. All compounds are >95% pure by HPLC analysis unless otherwise stated. Further information on the synthesis of the compounds is included in the Supplementary Material.

3. Results

3.1. Virtual screening

Recently, the periplasmic tip of TolC has been shown to be a potential target for small molecules (Szal et al., 2025). To identify additional compounds addressing TolC, a structure-based virtual screening campaign was conducted. For this purpose, a screening library was designed, containing commercially available small organic compounds with a molecular weight above 375 Da and possessing a positively charged nitrogen atom, which is predicted to facilitate uptake into Gram-negative bacteria (Richter et al., 2017) (Figure 1). Subsequently, the library containing 654,000 compounds was docked into the propeller-shaped cavity at the periplasmic tip of TolC. Based on the docking scores and the normalized docking scores (score divided by the number of heavy atoms), the binding modes of the 250 top-ranked compounds were visually inspected. In the analysis, special attention was paid to the geometric fitting of the docking pose into the putative binding pocket and the presence of polar interactions with TolC, in particular with the conserved aspartates Asp371 and Asp374. Finally, eleven compounds with promising binding modes were identified as potential TolC binders (Supplementary Table S1).

FIGURE 1.

Infographic describes compound selection and docking workflow: compounds are filtered from the ZINC database through criteria like molecular weight, logP, reactivity, and charge, then further refined for docking, culminating in docking into the periplasmic tip of superposed TolC X-ray crystal structures PDB ID 1TQQ and PDB ID 1EK9, with tip regions highlighted in red boxes.

Open in a new tab

Computational workflow illustrating the selection criteria for ZINC database subset download, additional descriptors calculated within MOE to focus on compounds with better solubility and likelihood of bacterial accumulation, and docking of the resulting collection into the periplasmic tip of two TolC X-ray crystal structures.

3.2. Hit validation and characterization

All virtual hits were validated experimentally using a two-step procedure. Initially, we assessed each compound for interference with efflux pump function by quantifying the accumulation of the fluorescent DNA dye H33342, a known AcrAB-TolC substrate. Subsequently, efflux-inhibiting compounds were tested for synergistic activity in combination with the antibiotic piperacillin.

Only one compound (1a) resulted in an intracellular accumulation of H33342 in E. coli BW25113 ≥ 80% relative to the tolC-deficient E. coli strain JW5503 at 100 μM (Supplementary Table S1). All other compounds showed an accumulation ranging between 14.5% and 78.3%. Similarly, 1a inhibited bacterial growth in combination with sub-inhibitory concentration of piperacillin (1 μg/mL, 0.25 MIC). Subsequently, we determined the minimum potentiating concentration (MPC4) that resulted in a 4-fold improvement in the MIC of piperacillin (4 μg/mL), which was 75 µM. To exclude any interference due to an intrinsic antimicrobial activity of 1a, the assay was carried out in the absence of piperacillin. The MIC for 1a was not reached at the highest tested concentration (125 µM) (Table 1; Supplementary Table S5), indicating synergism between 1a at 75 µM and piperacillin.

TABLE 1.

Modification of substructure B at the middle benzene ring.

Chemical structure diagram showing a molecule with a 4-chlorophenoxy and a hydroxyalkylamine side chain. A central aromatic ring, labeled B, is boxed in blue with a variable R group attached via an ether linkage.
Compound Structure, R MPC4 a PIP (µM) MIC b (µM) MIC/MPC4 H33342 accumulation (%, average ± SD) c
1a Me- 75 >125 >1.67 142.1 ± 1.4
2a Et- 50 >125 >2.5 113.4 ± 0.3
2b Structural formula illustration showing an isobutyl group, with three carbon branches in blue connected to a straight chain in black ending with a squiggly line representing an unspecified group. 12.5 25 2 137.4 ± 1.3
2c Chemical structure diagram showing a cyclopentyl ring with a two-carbon ethyl side chain extending from one carbon and ending in a wavy line, indicating a generic or unspecified attachment point. 6.25 12.5 2 146.1 ± 4.4
2f Chemical structure diagram showing a benzyl group with a benzene ring attached to a methylene group, which is connected to a wavy line indicating an unspecified substituent or point of attachment. 6.25 25 4 152.0 ± 0.9
PAβN (25 μg/mL) ​ na na na 113.9 ± 1.6
Open in a new tab
a

MPC4 PIP: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 4 μg/mL to 1 μg/mL) in Escherichia coli BW25113. If MPC4 was not achieved, the value is represented by the highest concentration of compound tested.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited Escherichia coli BW25113 growth by ≥ 90%. If MIC, was not achieved, the value is represented by the highest concentration of compound tested.

c

Compounds were tested at 100 µM. The accumulation of H33342 is expressed as a percentage and standard deviation in relation to tolC-deficient strain Escherichia coli JW5503. Values obtained in relative fluorescence units were converted into percentage and values for Escherichia coli JW5503 were attributed as 100%. The fluorescence intensity was recorded at excitation and emission wavelengths of 355 and 460 nm, respectively, after 30 min incubation period at 37 °C. PAβN (25 μg/mL) was used as a positive control (Lomovskaya et al., 2001). Experiments were performed once in duplicate.na: not applicable.

Next, we assessed 1a binding to individual efflux pump subunits using microscale thermophoresis and fluorescence quenching. MST signals obtained for TolC and AcrA were used for analysis, since no quenching of the fluorescence occurred. We observed quenching for AcrB in the concentration range used for the binding studies. By analyzing the quenching of the initial fluorescence for AcrB, it was ensured that the same compound concentrations were analyzed. Compound 1a demonstrated weak affinity for TolC with a KD of 347 µM (95% confidence interval [127; 1200 µM]) and AcrB (KD = 182 µM [49.5; 690 µM]) (Figures 2a,b). In contrast, the compound did not bind to AcrA (Figure 2c). To further rule out any non-specific protein binding or measurement artefacts, such as an inner filter effect, binding of 1a to the structurally unrelated TEVp was examined. 1a did not interact with TEVp (Figure 2c).

FIGURE 2.

Panel a shows a graph of ΔFnorm of TolC in per mille against the concentration of compound 1a in mol. It shows a decline in ΔFnorm values at higher compound concentrations with the corresponding error bars. The KD is estimated to be at 347 μM with a 95% confidence interval of 127 to 1200 μM. Panel b displays a similar plot for AcrB, showing normalized initial fluorescence percentage decreasing with 1a concentration. Panel c presents ΔFnorm for TEVp and AcrA, with grouped data points and error bars indicating variability but no strong trend. Panel d is the chemical structure of 1a, highlighting three regions: A, B, and C. 1a’s flexible side chain is marked as substructure A, the middle core as substructure B, and the benzylic substituent as substructure C.

Open in a new tab

Dose-response curves for 1a binding to efflux pump proteins TolC (a), AcrB (b) and AcrA (c), as well as tobacco, etch virus protease (TEVp, (c), obtained by analyzing MST signals (TolC, AcrA) or initial fluorescence values (AcrB). Signals recorded at 9–10 s were used for data analysis of TolC, AcrA and TEVp. For AcrB fitting, the signal for each concentration is the mean of the initial fluorescence intensity at 1 to 0 s before heating by laser normalized on the signal of AcrB with DMSO alone. a, b, c Each data point is presented as mean ± SD (n = 5). KD values are presented together with their respective 95% confidence interval (Paketurytė et al., 2021). Data points were derived from Supplementary Figures S1–S4. (d) Chemical structure of the initial hit 1a and the investigated substructures A, B, and C.

3.3. Structure-activity relationship (SAR) studies

To identify the essential parts of 1a for synergistic activity and to improve efficacy and selectivity, an optimization campaign was initiated, which comprised simple modifications through synthesizing analogues with subtle changes to the flexible side chain (A), the middle core (B), and the benzylic substituent (C) (Figure 2d). All synthesized analogues were investigated experimentally. Compounds displaying a MPC4 of ≤125 μM in combination with piperacillin were subjected to MIC determination. All compounds demonstrated H33342 accumulation in E. coli BW25113 of at least 80% relative to the tolC-deficient E. coli strain JW5503 at a concentration of 100 μM, irrespective of the piperacillin-potentiating activity (Table 1). Data for compounds without improved synergistic activity compared to 1a are presented in Supplementary Tables S2-4.

In a first step, the flexible side chain consisting of a 2-({3-[(2-hydroxyethyl)amino]propyl}amino)methyl substituent (substructure A) was investigated with respect to its relevance for the synergistic effect of 1a with piperacillin. Two analogues containing a shortened side chain were synthesized. This substructure was found to be critical for the biological activity, as a reduction in the length by a methyl (1b) or aminoethyl group (1c) abolished the synergistic effect with piperacillin (Supplementary Table S2).

Next, we focused on altering the substituents at the middle benzene ring (substructure B). Replacing the methyl group of the methoxy substituent with larger nonpolar aliphatic groups (2a–2e, Table 1; Supplementary Table S3) substantially decreased the MPC4, thereby improving the synergy with piperacillin. However, these compounds also demonstrated low MICs (6.25–25 μM, Table 1; Supplementary Table S3), resulting in a poor ratio between intrinsic antibacterial activity and synergy for compounds 2d and 2e. Larger aliphatic substituents at substructure B such as tert-butyl (2d) or cyclohexane (2e) led to a loss of synergistic activity. However, the benzylic substituent (2f) demonstrated a low MPC4 (6.25 µM) and a 4-fold MIC/MPC4 ratio.

Subsequently, we investigated the SAR of substituents at the aromatic ring in substructure C. Removal of chlorine (3a) resulted in the loss of synergistic properties (Supplementary Table S4). Changing the chlorine substituent of 1a to ortho (3b) or para position (3c) at the benzene ring slightly impaired the synergistic activity. Likewise, the exchange of chlorine for fluorine (3d) or bromine (3f) in the meta-position resulted in loss or no change in synergism with piperacillin (Supplementary Table S3). However, increasing the size of the halogen in para position was beneficial for the synergistic effect, as compounds containing a bromine (3e) or iodine substituent (3g) were slightly more potent than 1a (Table 2). None of the two monohalobenzyl analogues exhibited an intrinsic antibacterial activity (MIC >125 µM), resulting in a >1.67 to >2.5-fold ratio between MPC4 and MIC.

TABLE 2.

Substructure C modifications connected via an oxymethylene group to substructure B.

Chemical structure diagram showing a molecule with a benzene ring attached to side chains including ethylene glycol and amine groups, highlighted with a red box around an ether-linked group labeled R, denoted as C.
Compound Structure, R MPC4 a PIP (µM) MIC b (µM) MIC/MPC4 H33342 accumulation (%, average ± SD) c
1a Chemical structure diagram showing a benzene ring with a chlorine atom substituent at the first position and a wavy line at the fourth position indicating a generic attachment point. 75 >125 >1.67 142.1 ± 1.4
3e Chemical structure diagram showing a benzene ring with a bromine atom bonded to one carbon and a wavy line on the opposite carbon, indicating a para-substituted bromobenzene derivative. 50 >125 >2.5 99.9 ± 2.6
3g Chemical structure diagram showing a benzene ring with an iodine atom attached to one carbon and a wavy line on the opposite carbon, indicating a substituent or continuation. 50 >125 >2.5 103.8 ± 3.0
3k Chemical structure diagram showing a benzene ring with a trifluoromethyl group, labeled as CF subscript 3, attached to one carbon and a wavy line indicating a variable group attached to the opposite carbon. 50 >125 >2.5 92.8 ± 1.4
3L Chemical structure diagram showing a benzene ring with a tert-butyl group attached on the left and a wavy line on the right indicating a variable substituent position. 25 125 5 121.4 ± 0.1
3m Chemical structure diagram showing a benzene ring with a phenyl group abbreviated as "Ph" attached to one carbon and a wavy line extending from the opposite carbon, indicating unspecified substitution. 25 75 3 148.1 ± 1.5
3o Chemical structure diagram showing a benzene ring with two chlorine atoms attached at positions one and three, indicating a 1,3-dichlorobenzene molecule, with a wavy line representing an unspecified substituent. 25 >125 >5 113.2 ± 1.0
3p Chemical structure diagram showing a benzene ring with two chlorine atoms substituted at positions one and two, labeled as Cl in blue text, with a wavy line attached to the third position. 25 >125 >5 113.0 ± 1.6
3q Chemical structure diagram showing a benzene ring with chlorine atoms bonded at the one and three positions, representing meta-dichlorobenzene; the remaining bond is indicated by a wavy line. 25 >125 >5 118.3 ± 0.6
3r Structural formula diagram of a benzene ring with three substituents: a chlorine atom at the first position, a trifluoromethyl group at the second position, and an unspecified group at the fourth position. 25 100 4 127.2 ± 1.7
3s Structural formula diagram showing a benzene ring with two CF3 groups attached at the 2 and 4 positions and a wavy line extending from the 1 position, indicating a variable substituent. 12.5 75 6 127.3 ± 0.6
3u Structural diagram of a benzothiophene molecule featuring a fused benzene and thiophene ring with a sulfur atom, and a single bond extending from the thiophene ring indicating a substitution point. 50 >125 >2.5 107.5 ± 1.7
PAβN (25 μg/mL) ​ na na na 117.5 ± 0.5
Open in a new tab
a

MPC4 PIP: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 4 μg/mL to 1 μg/mL) in Escherichia coli BW25113. If MPC4 was not achieved, the value is represented by the highest concentration of compound tested.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited Escherichia coli BW25113 growth by ≥ 90%. If MIC, was not achieved, the value is represented by the highest concentration of compound 5tested.

c

Compounds were tested at 100 µM. The accumulation of H33342 is expressed as a percentage and standard deviation in relation to tolC-deficient strain Escherichia coli JW5503. Values obtained in relative fluorescence units were converted into percentage and values for Escherichia coli JW5503 were attributed as 100%. The fluorescence intensity was recorded at excitation and emission wavelengths of 355 and 460 nm, respectively, after 30 min incubation period at 37 °C. PAβN (25 μg/mL) was used as a positive control (Lomovskaya et al., 2001). Experiments were performed once in duplicate. na: not applicable.

Next, we investigated the impact of hydrophobic substituents at the benzylic group (substructure C). Replacement of the nonpolar chlorine (1a, lipophilicity contribution ΔlogDradius=0: 0.6) with polar, hydrophilic groups (3h: cyano, ΔlogDradius=0: −0.28; 3i: MeO ΔlogDradius=0: −0.05; Supplementary Table S4) resulted in loss of activity. A methyl substituent in the same position as the chlorine (3j, ΔlogDradius=0: 0.3; Supplementary Table S4) impaired the MPC4. Conversely, lipophilic substituents improved the MPC4. The trifluoromethyl group (3k, ΔlogDradius=0: 0.9) increased the ratio between MPC4 and MIC by > 2.5-fold, and the tert-butyl-containing analogue 3L (ΔlogDradius=0: 1.3) resulted in a 5-fold ratio (Table 2). However, we observed an intrinsic antibacterial activity for a sterically demanding and strongly lipophilic biphenyl analogue 3m (ΔlogDradius=0: 1.4).

Furthermore, a series of Cl- and CF3- substituted analogues 3n–3s were synthesized to systematically probe different double substitutions at the benzene ring of substructure C. The 2,6-substitution with chlorine (3n, Supplementary Table S3) resulted in a slightly higher synergistic effect than the mono-chlorine substitution in ortho position (3b, Supplementary Table S3) but showed no improvement compared to 1a. The 3,4-, 2,3-, and 2,4-substitution patterns of chlorine (3o, 3p, and 3q; Table 2) improved the synergistic activity by a factor of three compared to 1a, without increasing the intrinsic antibacterial activity. The exchange of chlorine in the two-position in 3o with a trifluoromethyl-substituent (3r, Table 2) resulted in a higher intrinsic antibacterial activity. The 3,5-substitution with trifluoromethyl was most beneficial for synergism with piperacillin in E. coli BW25113 (3s) with a comparably low MPC4 (12.5 µM, Table 2) and an MIC/MPC4 ratio of 6.

Most of the heterocyclic analogues 3t–3w did not show an improved MPC4 compared to compound 1a (Table 2; Supplementary Table S4). The bioisosteric replacement of the aryl group with a thiophene ring in 3t resulted in an activity comparable to that of 1a (Supplementary Table S4), while the benzothiophene-containing compound was slightly more potent (3u, Table 2). However, the inclusion of an isoquinoline (3v) or 2-methyl pyridine (3w) abolished any synergistic activity.

In a final step, we combined a side chain length as present in 1a at substructure A, a 3,5-trifluoromethylsubstituted benzylic ring as in 3s at substructure C, and a benzyl ring at substructure B, containing additional ring substitutions. To this end, several 3,5-disubstituted benzyl analogues 4b–4e were synthesized (Table 3). A compound missing the benzylic substituent and ether (4a, Table 3) was synthesized for comparison with the other analogues. Indeed, all compounds exhibited a synergistic effect with piperacillin at lower concentrations than 1a with varying MICs. Compound 4a demonstrated the highest MPC4 (12.5 µM, Table 3) of this series. The addition of a large steric substituent in para-position slightly improved the MPC4 (6.25 µM) and showed antibacterial activity at 12.5 µM (4b). The inclusion of a trifluoro-benzylic ring in compound 4c, revealed the same synergistic activity with piperacillin (MPC4 = 6.25 µM) but a reduced intrinsic antibacterial effect, resulting in an 8-fold MIC/MPC4 ratio. Substitution with a dichloro-benzylic ring in 4d and 4e resulted in the lowest MPC4 values of all tested compounds in this study (3.13 µM, Table 3) but also showed a strong intrinsic antibacterial activity.

TABLE 3.

Modification of substructure B with 3s-pattern at substructure C.

Chemical structure diagram showing a molecule with a hydroxyethyl, diamine, and ether linkages connecting three aromatic rings. The central ring is marked with a blue rectangle and labeled "B" with a substituent "R superscript 1". The rightmost ring displays two trifluoromethyl groups.
Compound Structure, R MPC4 a PIP (µM) MIC b (µM) MIC/MPC4 H33342 accumulation (%, average ± SD) c
4a Hydrogen atom symbol in blue connected by a bond to a solid wedge with a wavy line, representing a stereochemical or undefined bond often used in chemical structure diagrams. 12.5 50 4 117.6 ± 1.3
4b Structural diagram of a chemical compound showing a benzene ring with an isopropyl group attached to one carbon and a formaldehyde group attached via a single bond to another carbon. 6.25 12.5 2 211.7 ± 8.6
4c Chemical structure diagram showing a benzene ring with two trifluoromethyl groups in blue attached at the second and fifth positions, and a methoxy group connected via an ethyl chain at the fourth position. 6.25 50 8 223.4 ± 11.0
4d Chemical structure diagram showing a benzene ring with chlorine atoms attached to positions one and three, and a side chain containing an ether group extending from the fourth position. 3.13 6.25 2 142.8 ± 7.2
4e Chemical structure diagram showing a benzene ring with chlorine atoms attached at two adjacent positions and a side chain consisting of an ethoxy group. Wavy line indicates an unspecified group or stereochemistry. 3.13 6.25 2 207.4 ± 16.4
Open in a new tab
a

MPC4 PIP: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 4 μg/mL to 1 μg/mL) in Escherichia coli BW25113.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited Escherichia coli BW25113 growth by ≥ 90%.

c

Compounds were tested at 100 µM. The accumulation of H33342 is expressed as a percentage and standard deviation in relation to tolC-deficient strain Escherichia coli JW5503. Values obtained in relative fluorescence units were converted into percentage and values for Escherichia coli JW5503 were attributed as 100%. The fluorescence intensity was recorded at excitation and emission wavelengths of 355 and 460 nm, respectively, after 30 min incubation period at 37 °C. PAβN (25 μg/mL) was used as a positive control (Lomovskaya et al., 2001). Experiments were performed once in duplicate.

3.4. Prioritization of analogues

Altogether six compounds demonstrated a MPC4 ≤6.25 μM (2c, 2f, 4b, 4c, 4d, and 4e, Tables 1, 3). To prioritize compounds for further studies, the interaction of each compound with TEVp was examined. Only 2c (KD = 298 µM CI [36.6, 1910]) and 2f (2494 µM CI [690, 7060]) showed weak TEVp binding and were therefore excluded from further investigations (Supplementary Figure S6).

To ensure that the observed potentiation of piperacillin in E. coli BW25113 was specific to AcrAB-TolC inhibition, compounds with off-target activity were excluded by conducting antibiotic potentiation assays with efflux-deficient E. coli JW5503 (ΔtolC) in presence of piperacillin or levofloxacin (Table 4). The MIC/MPC4 ratio of EPIs in the mutant strain is expected to be lower than in wildtype bacteria, ideally approaching 1. Except for 4b, all selected compounds showed a lower MIC/MPC4 ratio in the mutant strain. 4d and 4e exhibited an MIC/MPC4 ratio of one together with piperacillin and levofloxacin and 4c displayed an MIC/MPC4 ratio of one in combination with levofloxacin. Interestingly, compound 4c demonstrated increased antimicrobial activity in efflux-deficient E. coli, indicating that it is a substrate of the AcrAB-TolC efflux pump complex. Furthermore, the activity of levofloxacin was not potentiated by PAβN, suggesting that it is not affected by the membrane-permeabilizing off-target effect.

TABLE 4.

MPC4 and MIC of compounds tested in combination with piperacillin or levofloxacin in Escherichia coli JW5503 (ΔtolC).

Compound MPC4 PIP a (µM) MIC b (µM) MIC/MPC4 PIP MPC4 LEV c (µM) MIC/MPC4 LEV
4b 6.25 12.5 2 6.25 2
4c 6.25 12.5 2 12.5 1
4d 6.25 6.25 1 6.25 1
4e 6.25 6.25 1 6.25 1
PAβN 12.5 μg/mL 50 μg/mL 4 50 μg/mL 1
Open in a new tab
a

MPC4 PIP: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 0.25 μg/mL to 0.0625 μg/mL) in Escherichia coli JW5503.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited JW5503 growth by ≥ 90%.

c

MPC4 LEV: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of levofloxacin by 4-fold (i.e., from 0.01 μg/mL to 0.0025 μg/mL) in Escherichia coli JW5503.

Compounds 4c, 4d, and 4e were assessed for potentiation of levofloxacin in wildtype E. coli and E. coli JD17464 (ΔlpxC) possessing an impaired OM (Table 5) to investigate whether the synergistic effect of the compounds is due to permeabilizing the OM which may lead to increased antibiotic permeation into the cells (Mood et al., 2021). EPIs are expected to retain their activity in OM-compromised bacteria. We observed no change in MPC4 for 4c and 4e with levofloxacin in BW25113 and JD17464. Only 4d lost its potentiating activity in E. coli JD17464, attributing its potentiating effect with antibiotics to off-target mechanisms, and therefore was excluded from further investigations.

TABLE 5.

MPC4 and MIC of compounds 4c to 4e in combination with piperacillin or levofloxacin in Escherichia coli BW25113 (wildtype) and JD17464 (ΔlpxC).

​ E. coli BW25113 E. coli JD17464 (lpxC)
Compound MPC4 LEV a (µM) MIC b (µM) MIC/MPC4 LEV MPC4 LEVc (µM) MIC (µM) MIC/MPC4 LEV
4c 6.25 50 8 6.25 50 8
4d 3.13 6.25 2 6.25 6.25 1
4e 3.13 6.25 2 3.13 6.25 2
Open in a new tab
a

MPC4 LEV: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 4 μg/mL to 1 μg/mL) in Escherichia coli BW25113.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited Escherichia coli BW25113 or JD17464 growth by ≥ 90%.

c

MPC4 LEV: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of levofloxacin by 4-fold (i.e., from 0.0625 μg/mL to 0.0156 μg/mL) in Escherichia coli JD17464.

3.5. Characterization of selected analogues

Uropathogenic E. coli is a leading cause of urinary tract infections and frequently exhibits elevated expression of the AcrAB-TolC efflux system, a major determinant of intrinsic and acquired MDR (Sulavik et al., 2001). In this context, the UPEC reference strain UMN026 was selected in this study as a clinically relevant MDR isolate model (Lescat et al., 2009) To assess the antimicrobial potentiating activity of 4c and 4e in clinically relevant settings, the compounds were tested in combination with piperacillin against E. coli UMN026. 4e retained its MIC/MPC4 ratio, whereas this ratio was slightly lower for 4c compared to BW25113 (Table 6). Notably, the intrinsic antibacterial activity of 4c and 4e increased against the MDR strain compared to E. coli BW25113.

TABLE 6.

MPC4 and MIC of 4c and 4e in combination with piperacillin in uropathogenic Escherichia coli UMN026.

Compound MPC4 a PIP (µM) MIC b (µM) MIC/MPC4
4c 1.56 6.25 4
4e 1.56 3.13 2
Open in a new tab
a

MPC4: Minimum Potentiating Concentration was defined as the concentration of compound that decreased MIC, of piperacillin by 4-fold (i.e., from 512 μg/mL to 128 μg/mL) in uropathogenic Escherichia coli UMN026.

b

MIC: Minimum Inhibitory Concentration was defined as the concentration of compound that inhibited Escherichia coli UMN026 growth by ≥ 90%.

Finally, binding studies with individual efflux pump subunits were conducted to identify the bacterial target protein of compounds 4c and 4e. Both compounds showed binding to TolC in a similar concentration range to 1a, as determined by fluorescence quenching, (Figure 3; Table 7). Since all proteins’ initial fluorescence was quenched by the compounds, the quenching was used for analysis, so higher concentrations could be analyzed. In contrast to 1a, both compounds demonstrated a very weak affinity for AcrA. Notably, the binding affinity of 4c and 4e for TolC was 20-fold higher than that for AcrB, whereas 1a exhibited similar binding affinities for TolC and AcrB.

FIGURE 3.

Four-panel chart showing the effects of compounds 4c and 4e on proteins AcrA, TolC, and AcrB. Panel a: Binding curve displaying normalized initial fluorescence versus 4c concentration for AcrA (red squares), TolC (green triangles), and AcrB (blue circles), showing decreased fluorescence for TolC and AcrB with increasing concentration. Panel b: Similar graph for 4e, with a similar behaviour for AcrB and TolC and a notable fluorescence increase for AcrA at the highest concentration. Panel c: Binding curve showing observed melting temperature of TolC increases with compound concentration for both 4c and 4e. Panel d: Curve showing unfolded fraction of TolC decreases as compound concentration increases for both compounds. Each data series includes error bars.

Open in a new tab

Binding curves obtained by label-free MST (a,b) and nano differential scanning fluorimetry (c,d). Interaction of 4c (a) and 4e (b) with AcrA (red squares), TolC (green triangles), and AcrB (blue dots) based on fluorescence quenching signals, normalized to the fluorescence intensity of the proteins with DMSO only. (c) Dose-response curves for 4c (turquoise triangles) and 4e (orange dots) with TolC using nDSF and analyzed by fitting the observed melting temperatures of TolC with and without compound treatment at different concentrations. (d) Dose-response curves for 4c and 4e analyzed by fitting the unfolded fractions of TolC at 78 °C (close to the non-treated melting temperature) with and without compound treatment at different concentrations. Data points were derived from Supplementary Figures S12–S19. Each data point is presented as mean ± SD (n = 5 for MST, n = 3 for nDSF). The corresponding KD values are shown in Table 7.

TABLE 7.

Summary of binding affinities obtained by MST (quenching and MST signals at 14–15 s) and nano differential scanning fluorimetry (TM shift and unfolded fractions at 78 °C).

Compound a TolC (quenching) TolC (TM shift) TolC (unfolded fractions at 78 °C) AcrA (quenching) AcrB (quenching)
KD [µM] 95% CI KD [µM] 95% CI KD [µM] 95% CI KD [µM] 95% CI KD [µM] 95% CI
1a 347 b 127, 1200 no shift no shift n.b 247 90, 692
4c 224 90.4, 563 69 39, 120 12 4.2, 27 1587 201, 6580 4263 1060, 7360
4e 274 99.9, 699 12 2.7, 43 9.6 2.1, 27 >8000 ​ 6526 967, 9930
Open in a new tab
a

Compounds were used as diacetate salts.

b

MST, signals at 9–10 s were used for analysis of the 1a-TolC interaction.

n.b.: not binding; KD: equilibrium dissociation constant; CI: confidence interval.

As an orthogonal assay nano differential scanning fluorimetry was used for TolC and analyzed either by fitting the TM shift (Schellman, 1975) or the unfolded fractions close to the melting temperature of TolC (Niebling et al., 2021). The observed TM shift of TolC and in the unfolded fractions upon compound treatment confirmed binding of 4c and 4e to TolC in the micromolar range (Figures 3C,D; Table 7), whereas no shift was observed for TolC in presence of 1a (Table 7; Supplementary Figure S5).

4. Discussion

The ever-increasing demand for novel antibiotics requires novel approaches in drug discovery (Miethke et al., 2021). EPIs are considered as a solution to tackle the AMR burden as they may restore the efficacy of antibiotics. However, to date, no such compound has reached clinical trials. Most EPI drug discovery programmes have focused on the efflux pump components AcrA and AcrB as primary targets and have predominantly been conducted using E. coli K12-derived laboratory strains or engineered efflux backgrounds (Lomovskaya et al., 2001; Bohnert et al., 2005; Opperman et al., 2014). While a limited number of studies have evaluated EPIs in MDR clinically relevant isolates, including Enterobacteriaceae, Acinetobacter baumannii, and Klebsiella pneumoniae, these investigations remain relatively scarce (Schumacher et al., 2006; Coban et al., 2011; Vera-Leiva et al., 2018). The fact that TolC deletion increases the susceptibility of E. coli to several antibiotics and a recent study reporting the discovery of peptides sterically blocking TolC and shifting the MIC of various antibiotics suggests that TolC is indeed a potential drug target (Sulavik et al., 2001; Budiardjo et al., 2022). Recently, we described the first small molecule (LP-115) interfering with efflux pump assembly by binding to the AcrA-TolC interface (Szal et al., 2025). The identification of LP-115 reassured TolC as a potential target for developing EPIs. TolC offers a significant advantage over AcrA and AcrB, as compounds targeting TolC do not necessarily need to cross the OM, depending on the binding site. The OM largely restricts permeation of large and hydrophobic compounds; therefore, the chemical space of drugs targeting bacterial proteins is limited and has led to the development of rules to prioritize compounds likely to penetrate the OM (Richter et al., 2017; Acosta-Gutiérrez et al., 2018). Targeting TolC might have the advantage of inhibiting not only a crucial protein for efflux, but also a factor for other physiological functions. TolC is associated with OM homeostasis (Zhu et al., 2024), virulence factor transport (Zgurskaya et al., 2011), and iron metabolism (Alav et al., 2021).

The pyranopyridine inhibitor MBX2319 is a potent EPI in laboratory E. coli strains, where it shifted the MIC of fluoroquinolones and β-lactams between 4- and 8-fold, with no activity in ΔacrB or ΔtolC mutants (Opperman et al., 2014). Complementary computational and functional analyses showed that MBX2319 binds within the distal binding pocket of AcrB (Vargiu et al., 2014). Related AcrB-directed inhibitors, including D13-9001 (Nakashima et al., 2013), WK2 (Wang et al., 2021), and BDM88855 (Plé et al., 2022) displayed pronounced antibiotic-dependent activity profiles. WK2, evaluated in E. coli K12 strains, potentiated chloramphenicol, erythromycin, and levofloxacin with ∼2–4-fold MIC reductions, but showed minimal activity against rifampicin. In contrast, BDM88855, studied in K12-derived systems, resulted in ∼4–8-fold MIC reductions for chloramphenicol, novobiocin, fusidic acid, and linezolid, while exhibiting little or no effect in combination with aminoglycosides and several β-lactams such as ceftazidime. Collectively, these inhibitors act by stabilizing non-productive AcrB conformational states or blocking substrate access within the distal pocket, but their efficacy remains strongly dependent on the antibiotic class and is largely derived from laboratory K12-based efflux models.

In this study, we report the identification of a compound (1a) that interferes with TolC and AcrB and potentiates the activity of piperacillin in E. coli. A comprehensive SAR analysis of 1a resulted in analogues with higher potentiating activity. The subsequent exclusion of non-specific compounds identified analogues 4c and 4e with an additional aromatic ring at substructure B as promising candidates for further development. Both compounds feature a lower MPC4 and a slightly higher binding affinity to TolC compared to 1a, but show a much lower binding affinity to AcrB (factor of 20), indicating increased selectivity for TolC. Binding to TolC was confirmed to be in the micromolar range, whereas the affinity to AcrB is in the millimolar range. Compared to previously identified EPIs, the determined binding affinities of 4c and 4e for TolC are about 2-fold weaker compared to the AcrA-binding NSC60339 and about one order of magnitude weaker than AcrB-targeting inhibitors (Haynes et al., 2017; Grimsey et al., 2020). Compared with the recently reported LP-115, these compounds showed higher binding affinity and selectivity for TolC and displayed an antimicrobial-potentiating effect in wild-type E. coli that is reduced or absent in the tolC-deficient strain. LP-115 is mechanistically intriguing because it was shown to act through a novel mode of action, namely, disruption of the AcrA-TolC interface. In contrast, the mechanism of action of compounds 4c and 4e has not yet been identified. Nevertheless, 4c and 4e appear to be more promising starting points for hit optimization, owing to their antimicrobial-potentiating effect and TolC selectivity, both of which are lacking for LP-115.

The observed differences in TolC binding affinity for compounds 4c and 4e, obtained by MST and nDSF, can be explained by the underlying measurement principles. MST data are acquired at room temperature, whereas nDSF data are obtained in the range of the TM, which is about 77 °C for TolC. The KD is influenced by several factors, for example, solvent viscosity and temperature (Niebling et al., 2021; Papaneophytou et al., 2014). While moderate temperature changes have only a marginal effect on the KD, larger temperature differences are expected to yield more divergent results. Considering this, the absence of a TM-shift of compound 1a does not necessarily mean that binding does not occur, but that the KD may be shifted to another concentration range.

A direct comparison of biophysical (KD) and microbiological data (MPC4) is challenging as both types of results can be influenced by many factors (Holdgate, et al., 2019). In particular, the environment of the target protein differs substantially between assays, especially when studying an isolated membrane protein like TolC. The observed increased cellular potency may be influenced by off-target effects, the mechanism of action, or compound partitioning (Schwaid and Cornella-Taracido, 2018; Mateus, et al., 2017). The possibility of off-target effects was carefully addressed and several common non-specific effects were investigated. Importantly, compounds 4c and 4e showed reduced or no activity in a tolC-knockout strain, which strongly indicates that the mechanism of action is at least partially TolC-dependent. Testing the compounds in membrane-compromised E. coli JD17464 (ΔlpxC) further excluded the compounds to act through a common off-target effect. Furthermore, the compounds inhibited efflux and/or potentiated three AcrAB-TolC substrates in two different assays. Together, these experiments support the conclusion that 4c and 4e mainly act through a TolC-dependent mechanism. Nevertheless, additional off-target effects may contribute in part, particularly for 4c, since the erythromycin-potentiating activity is not completely lost in the tolC-knockout strain. Another reason for the increased intracellular potency compared to the biophysical affinity may be partitioning of the compounds into lipophilic regions. In cellular assays, partitioning into the lipid-bilayer may potentiate activity through a non-specific (Rusinova et al., 2021) or specific effect (Aguayo-Ortiz et al., 2021), depending on the mechanism of action and whether the binding site is situated at a transmembrane region or at another membrane-distant site of the protein. Given the number of factors influencing the intracellular potency and the substantial differences in target protein environments between the assays, the affinity data mainly act as confirmation of TolC interaction and selectivity (Holdgate et al., 2019) and as a useful indicator for future compound optimization, but these values should not be directly equated with cellular potency.

Compound 4c demonstrated a MIC/MPC4 ratio of eight for both piperacillin and levofloxacin in E. coli BW25113, the highest ratio in this set of compounds. The compound continued to potentiate the activity of piperacillin in the tolC-deficient bacteria to a minor extent, which could be caused by off-target effects, something to consider in further optimization. However, the activity of levofloxacin was not potentiated by 4c, a behavior similar to EPI PAβN, which is known to have a membrane-permeabilizing activity as off-target effect (Lamers et al., 2013). In uropathogenic E. coli UMN026, the compound was even more potent with an MPC4 of 1.56 µM, translating laboratory conditions to clinically relevant ones. However, the intrinsic antimicrobial effect was substantially lower for both compounds. Similarly, in BW25113, the MICs of 4c and 4e were low compared to other EPIs, for which no antimicrobial effect at concentrations of more than 500 µM have been reported (Plé et al., 2022). The increased antimicrobial activity for the TolC-selective compounds may indicate that inhibition of TolC increases the antimicrobial activity, presumably caused by interference with crucial physiological processes. EPIs and antibiotic adjuvants in general are expected to lack any intrinsic antibacterial properties (Wright, 2016), leading to the question of whether TolC-targeting compounds could be categorized as antibiotic adjuvants. To answer this question, more research concerning TolC inhibitors is needed.

The compounds in the present study provide insights into the chemical structure of compounds capable of binding to TolC and may guide the future search for compounds with higher affinity and better potentiation of antibiotics. This study further emphasizes that TolC represents a potential target for antibiotic adjuvants.

Acknowledgments

We acknowledge technical support by the SPC facility at EMBL Hamburg and by the PP facility, University Medical Center Hamburg-Eppendorf, Hamburg, Germany at CSSB Hamburg. We thank Daina Gustina from LIOS for synthesis of compound 3r. The DDCB unit at the Faculty of Pharmacy, University of Helsinki, Finland, supported by HiLIFE and Biocenter Finland, is gratefully acknowledged for providing access to microbiology facilities and instruments.

Funding Statement

The author(s) declared that financial support was received for this work and/or its publication. This work was supported under the framework of the JPIAMR–Joint Programming Initiative on Antimicrobial Resistance. BW was supported by the German Federal Ministry for Education and Research (01KI1827A). AJ was supported by the state education development agency (1.1.1.5/ERANET/19/03), CDC and PT were supported by the Research Council of Finland (Grant no. 326261).

Footnotes

Edited by: Hendrik W. Van Veen, University of Cambridge, United Kingdom

Reviewed by: Jean Michel Bolla, Aix-Marseille Université, France

Shainaba A Saadhali, National Institute of Research in Tuberculosis (ICMR), India

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

JV: Methodology, Investigation, Visualization, Writing – original draft, Formal Analysis, Writing – review and editing. CDC: Formal Analysis, Writing – original draft, Investigation, Writing – review and editing. TS: Investigation, Writing – original draft, Methodology, Formal Analysis, Visualization, Writing – review and editing. MM: Investigation, Writing – review and editing, Formal Analysis. IL: Writing – review and editing, Investigation, Formal Analysis. LP: Investigation, Writing – review and editing, Formal Analysis. PL: Investigation, Methodology, Writing – review and editing. MH: Methodology, Investigation, Writing – review and editing. SW: Writing – review and editing, Methodology, Supervision, Resources. PT: Resources, Funding acquisition, Supervision, Writing – review and editing. AJ: Funding acquisition, Resources, Supervision, Writing – review and editing. BW: Visualization, Writing – review and editing, Conceptualization, Investigation, Writing – original draft, Funding acquisition, Supervision.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphar.2026.1895729/full#supplementary-material

DataSheet1.pdf (8.1MB, pdf)

References

  1. Abdali N., Parks J. M., Haynes K. M., Chaney J. L., Green A. T., Wolloscheck D., et al. (2017). Reviving antibiotics: efflux pump inhibitors that interact with AcrA, a membrane fusion protein of the AcrAB-TolC multidrug efflux pump. ACS Infect. Dis. 3 (1), 89–98. 10.1021/acsinfecdis.6b00167 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Acosta-Gutiérrez S., Ferrara L., Pathania M., Masi M., Wang J., Bodrenko I., et al. (2018). Getting drugs into gram-negative bacteria: rational rules for permeation through general porins. ACS Infect. Dis. 4 (10), 1487–1498. 10.1021/acsinfecdis.8b00108 [DOI] [PubMed] [Google Scholar]
  3. Aguayo-Ortiz R., Creech J., Jiménez-Vázquez E. N., Guerrero-Serna G., Wang N., da Rocha A. M., et al. (2021). A multiscale approach for bridging the gap between potency, efficacy, and safety of small molecules directed at membrane proteins. Sci. Rep. 11 (1), 16580. 10.1038/s41598-021-96217-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Alav I., Sutton J. M., Rahman K. M. (2018). Role of bacterial efflux pumps in biofilm Formation. J. Antimicrob. Chemother. 73 (8), 2003–2020. 10.1093/jac/dky042 [DOI] [PubMed] [Google Scholar]
  5. Alav I., Kobylka J., Kuth M. S., Pos K. M., Picard M., Blair J. M. A., et al. (2021). Structure, assembly, and function of tripartite efflux and type 1 secretion systems in gram-negative bacteria. Chem. Rev. 121 (9), 5479–5596. 10.1021/acs.chemrev.1c00055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Andersen C., Koronakis E., Bokma E., Eswaran J., Humphreys D., Hughes C., et al. (2002a). Transition to the open State of the TolC periplasmic tunnel entrance. Proc. Natl. Acad. Sci. U. S. A. 99 (17), 11103–11108. 10.1073/pnas.162039399 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Andersen C., Koronakis E., Hughes C., Koronakis V. (2002b). An aspartate ring at the TolC tunnel entrance determines ion selectivity and presents a target for blocking by large cations. Mol. Microbiol. 44 (5), 1131–1139. 10.1046/j.1365-2958.2002.02898.x [DOI] [PubMed] [Google Scholar]
  8. Annunziato G. (2019). Strategies to overcome antimicrobial resistance (AMR) making use of non-essential target inhibitors: a review. Int. J. Mol. Sci. 20 (23), 5844. 10.3390/ijms20235844 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Antimicrobial Resistance Collaborators (2022). Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet lond. Engl. 399 (10325), 629–655. 10.1016/S0140-6736(21)02724-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Baba T., Ara T., Hasegawa M., Takai Y., Okumura Y., Baba M., et al. (2006). Construction of Escherichia coli K-12 in-Frame, single-gene knockout mutants: the keio collection. Mol. Syst. Biol. 2, 2006.0008. 10.1038/msb4100050 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bohnert J. A., Kern W. V. (2005). Selected arylpiperazines are capable of reversing multidrug resistance in Escherichia coli overexpressing RND efflux pumps. J. Antimicrob. Chemother. 49 (2), 849–852. 10.1093/jac/dki445 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bolla J. M., Alibert-Franco J. M., Handzlik J., Chevalier J., Mahamoud A., Boyer G., et al. (2011). Strategies for bypassing the membrane barrier in multidrug resistant gram-negative bacteria. FEBS Lett. 585 (11), 1682–1690. 10.1016/j.febslet.2011.04.054 [DOI] [PubMed] [Google Scholar]
  13. Budiardjo S. J., Stevens J. J., Calkins A. L., Ikujuni A. P., Wimalasena V. K., Firlar E., et al. (2022). Colicin E1 opens its Hinge to plug TolC. eLife 11, e73297. 10.7554/eLife.73297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Burastero O., Niebling S., Defelipe L. A., Günther C., Struve A., Garcia Alai M. M. (2021). eSPC: an online data-analysis platform for molecular biophysics. Acta Crystallogr. Struct. Biol. 77, 1241–1250. 10.1107/S2059798321008998 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cernicchi G., Felicetti T., Sabatini S. (2021). Microbial efflux pump inhibitors: a journey around quinoline and indole derivatives. Molecules 26 (22), 6996. 10.3390/molecules26226996 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. CLSI (2015). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically. Tenth Edition (CLSI Document M07-A10). Wayne, PA, United States: Clinical and Laboratory Standards Institute (CLSI). [Google Scholar]
  17. Coban A. Y., Guney A. K., Tanriverdi Cayci Y., Durupinar B. (2011). Effect of 1-(1-Naphtylmethyl)-piperazine, an efflux pump inhibitor, on antimicrobial drug susceptibilities of clinical Acinetobacter baumannii isolates. Curr. Microbiol. 62 (2), 508–511. 10.1007/s00284-010-9736-9 [DOI] [PubMed] [Google Scholar]
  18. Coldham N. G., Webber M., Woodward M. J., Piddock L. J. V. (2010). A 96-Well plate fluorescence assay for assessment of cellular permeability and active efflux in Salmonella enterica Serovar Typhimurium and Escherichia coli . J. Antimicrob. Chemother. 65 (8), 1655–1663. 10.1093/jac/dkq169 [DOI] [PubMed] [Google Scholar]
  19. Dadgostar P. (2019). Antimicrobial resistance: implications and costs. Infect. Drug Resist. 12, 3903–3910. 10.2147/IDR.S234610 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Du D., Wang Z., James N. R., Voss J. E., Klimont E., Ohene-Agyei T., et al. (2014). Structure of the AcrAB–TolC multidrug efflux pump. Nature 509 (7501), 512–515. 10.1038/nature13205 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. El Meouche I., Dunlop M. J. (2018). Heterogeneity in efflux pump expression predisposes antibiotic-resistant cells to mutation. Science 362 (6415), 686–690. 10.1126/science.aar7981 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gilardi A., Bhamidimarri S. P., Brönstrup M., Bilitewski U., Marreddy R. K. R., Pos K. M., et al. (2017). Biophysical characterization of E. coli TolC interaction with the known blocker hexaamminecobalt. Biochim. Biophys. Acta BBA - Gen. Subj. 1861 (11), 2702–2709. 10.1016/j.bbagen.2017.07.014 [DOI] [PubMed] [Google Scholar]
  23. Grimsey E. M., Fais C., Marshall R. L., Ricci V., Ciusa M. L., Stone J. W., et al. (2020). Chlorpromazine and amitriptyline are substrates and inhibitors of the AcrB multidrug efflux pump. mBio 11 (3), e00465. 10.1128/mBio.00465-20 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Haynes K. M., Abdali N., Jhawar V., Zgurskaya H. I., Parks J. M., Green A. T., et al. (2017). Identification and structure-activity relationships of novel compounds that potentiate the activities of antibiotics in Escherichia coli . J. Med. Chem. 60 (14), 6205–6219. 10.1021/acs.jmedchem.7b00453 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Higgins M. K., Eswaran J., Edwards P., Schertler G. F. X., Hughes C., Koronakis V. (2004). Structure of the ligand-blocked periplasmic entrance of the bacterial multidrug efflux protein TolC. J. Mol. Biol. 342 (3), 697–702. 10.1016/j.jmb.2004.07.088 [DOI] [PubMed] [Google Scholar]
  26. Holdgate G., Embrey K., Milbradt A., Davies G. (2019). Biophysical methods in early drug discovery. ADMET DMPK 7 (4), 222–241. 10.5599/admet.733 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Irwin J. J., Tang K. G., Young J., Dandarchuluun C., Wong B. R., Khurelbaatar M., et al. (2020). ZINC20-A free ultralarge-scale chemical database for ligand discovery. J. Chem. Inf. Model. 60 (12), 6065–6073. 10.1021/acs.jcim.0c00675 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jorgensen W. L., Duffy E. M. (2002). Prediction of drug solubility from structure. Adv. Drug Deliv. Rev. 54 (3), 355–366. 10.1016/s0169-409x(02)00008-x [DOI] [PubMed] [Google Scholar]
  29. Koronakis V., Sharff A., Koronakis E., Luisi B., Hughes C. (2000). Crystal structure of the bacterial membrane protein TolC central to multidrug efflux and protein export. Nature 405 (6789), 914–919. 10.1038/35016007 [DOI] [PubMed] [Google Scholar]
  30. Lamers R. P., Cavallari J. F., Burrows L. L. (2013). The efflux inhibitor phenylalanine-arginine beta-naphthylamide (PAβN) permeabilizes the outer membrane of gram-negative bacteria. PLoS One 8 (3), e60666. 10.1371/journal.pone.0060666 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Laws M., Shaaban A., Rahman K. M. (2019). Antibiotic resistance breakers: current approaches and future directions. FEMS Microbiol. Rev. 43 (5), 490–516. 10.1093/femsre/fuz014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Lescat M., Calteau A., Hoede C., Barbe V., Touchon M., Rocha E., et al. (2009). A module located at a chromosomal integration hot spot is responsible for the multidrug resistance of a reference strain from Escherichia coli clonal group A. Antimicrob. Agents Chemother. 53 (6), 2283–2288. 10.1128/AAC.00123-09 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Li X.-Z., Plésiat P., Nikaido H. (2015). The challenge of efflux-mediated antibiotic resistance in gram-negative bacteria. Clin. Microbiol. Rev. 28 (2), 337–418. 10.1128/CMR.00117-14 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lipinski C. A., Lombardo F., Dominy B. W., Feeney P. J. (2001). Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings (1997). Adv. Drug Deliv. Rev. 46 (1-3), 3–26. 10.1016/s0169-409x(00)00129-0 [DOI] [PubMed] [Google Scholar]
  35. Lomovskaya O., Warren M. S., Lee A., Galazzo J., Fronko R., Lee M., et al. (2001). Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. Agents Chemother. 45 (1), 105–116. 10.1128/AAC.45.1.105-116.2001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. López-Méndez B., Baron B., Brautigam C. A., Jowitt T. A., Knauer S. H., Uebel S., et al. (2021). Reproducibility and accuracy of microscale thermophoresis in the NanoTemper Monolith: a multi laboratory benchmark Study. Eur. Biophys. J. 50 (3), 411–427. 10.1007/s00249-021-01532-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Manges A. R., Johnson J. R., Foxman B., O’Bryan T. T., Fullerton K. E., Riley L. W. (2001). Widespread distribution of urinary tract infections caused by a Multidrug-Resistant Escherichia coli Clonal Group. N. Engl. J. Med. 345 (14), 1007–1013. 10.1056/NEJMoa011265 [DOI] [PubMed] [Google Scholar]
  38. Mateus A., Gordon L. J., Wayne G. J., Almqvist H., Axelsson H., Seashore-Ludlow B., et al. (2017). Prediction of intracellular exposure bridges the gap between target- and cell-based drug discovery. Proc. Natl. Acad. Sci. U. S. A. 114 (30), E6231–E6239. 10.1073/pnas.1701848114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Miethke M., Pieroni M., Weber T., Brönstrup M., Hammann P., Halby L., et al. (2021). Towards the sustainable discovery and development of new antibiotics. Nat. Rev. Chem. 5 (10), 726–749. 10.1038/s41570-021-00313-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mood E. H., Goltermann L., Brolin C., Cavaco L. M., Nejad A. J., Yavari N., et al. (2021). Antibiotic potentiation in multidrug-resistant gram-negative pathogenic bacteria by a synthetic peptidomimetic. ACS Infect. Dis. 7 (8), 2152–2163. 10.1021/acsinfecdis.1c00147 [DOI] [PubMed] [Google Scholar]
  41. Naghavi M., Vollset S. E., Ikuta K. S., Swetschinski L. R., Gray A. P., Wool E. E., et al. (2024). Global burden of bacterial antimicrobial resistance 1990–2021: a systematic analysis with forecasts to 2050. Lancet 404 (10459), 1199–1226. 10.1016/S0140-6736(24)01867-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nakashima R., Sakurai K., Yamasaki S., Hayashi K., Nagata C., Hoshino K., et al. (2013). Structural basis for the inhibition of bacterial multidrug exporters. Nature 500 (7460), 102–106. 10.1038/nature12300 [DOI] [PubMed] [Google Scholar]
  43. Niebling S., Burastero O., Bürgi J., Günther C., Defelipe L. A., Sander S., et al. (2021). FoldAffinity: binding affinities from nDSF experiments. Sci. Rep. 11 (1), 9572. 10.1038/s41598-021-88985-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Opperman T. J., Kwasny S. M., Kim H. S., Nguyen S. T., Houseweart C., D'Souza S., et al. (2014). Characterization of a novel pyranopyridine inhibitor of the AcrAB efflux pump of Escherichia coli . Antimicrob. Agents Chemother. 58 (2), 722–733. 10.1128/AAC.01866-13 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Paketurytė V., Petrauskas V., Zubrienė A., Abian O., Bastos M., Chen W. Y., et al. (2021). Uncertainty in protein-ligand binding constants: asymmetric confidence intervals versus standard errors. Eur. Biophys. J. 50, 661–670. 10.1007/s00249-021-01518-4 [DOI] [PubMed] [Google Scholar]
  46. Papaneophytou C. P., Grigoroudis A. I., McInnes C., Kontopidis G. (2014). Quantification of the effects of ionic strength, viscosity, and hydrophobicity on protein-ligand binding affinity. ACS Med. Chem. Lett. 5 (8), 931–936. 10.1021/ml500204e [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Phan G., Picard M., Broutin I. (2015). Focus on the outer membrane factor OprM, the forgotten player from efflux pumps assemblies. Antibiot. (Basel) 4 (4), 544–566. 10.3390/antibiotics4040544 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Piddock L. J. V. (2006). Multidrug-Resistance efflux pumps - not just for resistance. Nat. Rev. Microbiol. 4 (8), 629–636. 10.1038/nrmicro1464 [DOI] [PubMed] [Google Scholar]
  49. Plé C., Tam H. K., Vieira Da Cruz A., Compagne N., Jiménez-Castellanos J. C., Müller R. T., et al. (2022). Pyridylpiperazine-Based allosteric inhibitors of RND-Type multidrug efflux pumps. Nat. Commun. 13 (1), 115. 10.1038/s41467-021-27726-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Reza A., Sutton J. M., Rahman K. M. (2019). Effectiveness of efflux pump inhibitors as biofilm disruptors and resistance breakers in Gram-Negative (ESKAPEE) bacteria. Antibiot. (Basel) 8 (4), 229. 10.3390/antibiotics8040229 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Richter M. F., Drown B. S., Riley A. P., Garcia A., Shirai T., Svec R. L., et al. (2017). Predictive compound accumulation rules yield a broad-spectrum antibiotic. Nature 545 (7654), 299–304. 10.1038/nature22308 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Rusinova R., He C., Andersen O. S. (2021). Mechanisms underlying drug-mediated regulation of membrane protein function. Proc. Natl. Acad. Sci. U. S. A. 118 (46), e2113229118. 10.1073/pnas.2113229118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Santajit S., Indrawattana N. (2016). Mechanisms of antimicrobial resistance in ESKAPE pathogens. Biomed. Res. Int. 2016, 2475067. 10.1155/2016/2475067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sati H., Carrara E., Savoldi A., Hansen P., Garlasco J., Campagnaro E., et al. (2025). The WHO Bacterial Priority Pathogens List 2024: a prioritisation study to guide research, development, and public health strategies against antimicrobial resistance. Lancet Infect. Dis. 25 (9), 1033–1043. 10.1016/S1473-3099(25)00118-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Schellman J. A. (1975). Macromolecular binding. Biopolymers 14 (5), 999–1018. 10.1002/bip.1975.360140509 [DOI] [Google Scholar]
  56. Schumacher A., Steinke P., Bohnert J. A., Akova M., Jonas D., Kern W. V. (2006). Effect of 1-(1-naphthylmethyl)-piperazine, a novel putative efflux pump inhibitor, on antimicrobial drug susceptibility in clinical isolates of Enterobacteriaceae other than Escherichia coli . J. Antimicrob. Chemother. 57 (2), 344–348. 10.1093/jac/dki446 [DOI] [PubMed] [Google Scholar]
  57. Schwaid A. G., Cornella-Taracido I. (2018). Causes and significance of increased compound potency in cellular or physiological contexts. J. Med. Chem. 61 (5), 1767–1773. 10.1021/acs.jmedchem.7b00762 [DOI] [PubMed] [Google Scholar]
  58. Sjuts H., Vargiu A. V., Kwasny S. M., Nguyen S. T., Kim H.-S., Ding X., et al. (2016). Molecular basis for inhibition of AcrB multidrug efflux pump by novel and powerful pyranopyridine derivatives. Proc. Natl. Acad. Sci. U. S. A. 113 (13), 3509–3514. 10.1073/pnas.1602472113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Sulavik M. C., Houseweart C., Cramer C., Jiwani N., Murgolo N., Greene J., et al. (2001). Antibiotic susceptibility profiles of Escherichia coli strains lacking multidrug efflux pump genes. Antimicrob. Agents Chemother. 45 (4), 1126–1136. 10.1128/AAC.45.4.1126-1136.2001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Szal T., Chauhan S. S., Lewe P., Rachad F. Z., Madre M., Paunina L., et al. (2023). Efflux pump-binding 4(3-Aminocyclobutyl)Pyrimidin-2-Amines are colloidal aggregators. Biomolecules 13 (6), 1000. 10.3390/biom13061000 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Szal T., Petsolari E., Veliks J., Cruz C. D., Paunina L., Madre M., et al. (2025). Disassembly of the Escherichia coli AcrABZ-TolC efflux pump by ligand-mediated disruption of TolC-AcrA interfacial contacts. bioRxiv 2025.02.18.638599. 10.1101/2025.02.18.638599 [DOI] [Google Scholar]
  62. Tacconelli E., Carrara E., Savoldi A., Harbarth S., Mendelson M., Monnet D. L., et al. (2018). WHO pathogens priority list working group. Discovery, research, and development of new antibiotics: the WHO priority list of antibiotic-resistant bacteria and tuberculosis. Lancet Infect. Dis. 18 (3), 318–327. 10.1016/S1473-3099(17)30753-3 [DOI] [PubMed] [Google Scholar]
  63. Theuretzbacher U., Piddock L. J. V. (2019). Non-Traditional antibacterial therapeutic options and challenges. Cell Host Microbe 26 (1), 61–72. 10.1016/j.chom.2019.06.004 [DOI] [PubMed] [Google Scholar]
  64. Vargiu A. V., Ruggerone P., Opperman T. J., Nguyen S. T., Nikaido H. (2014). Molecular mechanism of MBX2319 inhibition of Escherichia coli AcrB multidrug efflux pump and comparison with other inhibitors. Antimicrob. Agents Chemother. 58 (10), 6224–6234. 10.1128/AAC.03283-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Vera-Leiva A., Carrasco-Anabalón S., Lima C. A., Villagra N., Domínguez M., Bello-Toledo H., et al. (2018). The efflux pump inhibitor phenylalanine-arginine β-naphthylamide (PAβN) increases resistance to carbapenems in Chilean clinical isolates of KPC-producing Klebsiella pneumoniae . J. Glob. Antimicrob. Resist. 12, 73–76. 10.1016/j.jgar.2017.12.003 [DOI] [PubMed] [Google Scholar]
  66. Wang Y., Alenazy R., Gu X., Polyak S. W., Zhang P., Sykes M. J., et al. (2021). Design and structural optimization of novel 2H-benzo[h]chromene derivatives that target AcrB and reverse bacterial multidrug resistance. Eur. J. Med. Chem. 213, 113049. 10.1016/j.ejmech.2020.113049 [DOI] [PubMed] [Google Scholar]
  67. Wright G. D. (2016). Antibiotic adjuvants: rescuing antibiotics from resistance. Trends Microbiol. 24 (11), 862–871. 10.1016/j.tim.2016.06.009 [DOI] [PubMed] [Google Scholar]
  68. wwPDB consortium (2019). Protein data bank: the single global archive for 3D macromolecular structure data. Nucleic Acids Res. 47 (D1), D520–D528. 10.1093/nar/gky949 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Yasufuku T., Shigemura K., Shirakawa T., Matsumoto M., Nakano Y., Tanaka K., et al. (2011). Correlation of overexpression of efflux pump genes with antibiotic resistance in Escherichia coli strains clinically isolated from urinary tract infection patients. J. Clin. Microbiol. 49 (1), 189–194. 10.1128/JCM.00827-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Zgurskaya H. I., Krishnamoorthy G., Ntreh A., Lu S. (2011). Mechanism and function of the outer membrane channel TolC in multidrug resistance and physiology of enterobacteria. Front. Microbiol. 2, 189. 10.3389/fmicb.2011.00189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Zhu S., Alexander M. K., Paiva T. O., Rachwalski K., Miu A., Xu Y., et al. (2024). The inactivation of tolC sensitizes Escherichia coli to perturbations in lipopolysaccharide transport. iScience 27 (5), 109592. 10.1016/j.isci.2024.109592 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

DataSheet1.pdf (8.1MB, pdf)

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Articles from Frontiers in Pharmacology are provided here courtesy of Frontiers Media SA

RESOURCES