Combining Anion Transport and Phospholipid Binding for Improved Antibacterial Activity of Diamidocarbazoles

Abstract

Increasing bacterial resistance necessitates focusing on new antibiotic targets, such as bacterial membranes. Herein, heteroditopic diamidocarbazole-based receptors are shown to display significant antibacterial activity against Gram-positive bacteria by combining transmembrane anion transport with phosphatidylethanolamine binding, while remaining nontoxic for human red blood cells. Liposome-based and bacterial studies consistently support this dual mode of action, highlighting the potential of combining anion transport and lipid-binding properties as a strategy for the development of novel antibiotics.

Introduction

The rise in bacterial resistance to antibiotics poses a severe global health risk, intensified by the stagnation in the development of new antibiotic classes. Traditional antibiotics often target bacterial proteins and nucleic acids (ribosomes), leading to the rapid emergence of resistant strains. Consequently, there is an urgent need to explore alternative therapeutic targets. One promising option is the bacterial phospholipid membrane, which is crucial for maintaining cell integrity and function. The membrane can be targeted by either binding to the lipids in the membrane (e.g., antimicrobial peptides), or by increasing membrane permeability via pore formation or ionophore activity (e.g., valinomycin). Either way, selectivity can be achieved due to the different membrane composition, smaller size (and thus larger surface-area-to-volume ratio) and higher membrane potential of bacteria compared to mammalian cells.

Bacterial membranes primarily consist of phospholipids such as phosphatidylethanolamine (PE), phosphatidylglycerol (PG), or cardiolipin (CL). In particular, PE is a major component of bacterial membranes of Gram-negative and certain Gram-positive strains. In contrast, phosphatidylcholine (PC) is the dominant component of mammalian cell membranes, and differs from PE in the degree of methylation of the ammonium group (Figure ). This difference in composition creates an opportunity to develop synthetic molecules that can selectively bind to bacterial phospholipids and thereby exhibit antibacterial activity without being toxic to human cells. Busschaert and co-workers have applied this strategy to develop neutral urea-based compounds that target bacterial PE , or PG over mammalian PC, and other research groups have also used similar lipid-targeting strategies. − As an alternative to lipid binding, a variety of synthetic anion receptors and anion transporters (anionophores) have also been shown to possess promising antibacterial activity. − Lipid binding has the advantage that a clear strategy for selectivity toward bacterial cells can be envisioned, but high concentrations are needed before lipid headgroup binding can lead to detrimental effects on the bacterial membrane. On the other hand, a single anionophore can transport multiple anions across the membrane, and antibacterial activity should theoretically be possible at low concentrations. However, the basis of selectivity for bacterial cells is less clear in this case, although it is likely related to the fact that fewer ions need to be transported in bacteria to induce a change in membrane potential compared with mammalian cells. We therefore wondered whether both strategies could be combined to create bifunctional molecules capable of anion transport and phospholipid binding. This could allow combining the advantages of both strategies (good selectivity and high potency) and achieving more effective antibacterial agents. While many antimicrobial peptides are thought to form pores upon lipid binding, this effect only occurs at high concentrations. The combination of lipid binding for selectivity with mobile carrier type ionophore activity (which theoretically can function at the single molecule level) has not yet been reported. In this manuscript, we report a crown ether-containing diamidocarbazole derivative that is able to function as both an ionophore and as a receptor for PE lipids in liposomes, and exerts antibacterial activity against a variety of Gram-positive bacteria with minimum inhibitory concentrations (MICs) as low as ∼4 μM and no observable hemolysis of human red blood cells.

1.

Chemical structures of urea-based receptors U1–U3, carbazole-based receptors R1–R4 and natural phospholipids POPC and POPE, which were used as examples of PC and PE lipids in this work.

Results and Discussion

Design and Synthesis

In our previous work on PE-binding antibacterial compounds, we used a single urea as the anion-binding moiety. , Monoureas are generally modest anion transporters, , and within the membrane environment anion transport and lipid binding are two competing processes. Given this competition, monoureas functionalized with a crown ether, U1–U3, preferentially interact with lipid headgroups and do not promote transmembrane ion transport. To overcome this limitation and develop compounds capable of both lipid binding and efficient anion transport, we selected a more effective anionophore scaffold. Diamidocarbazoles have previously been shown to function as potent transmembrane anion transporters with antibacterial activity. Furthermore, they also exhibit high affinity for oxyanions, − including phosphates, making them good candidates for combined ion transport and phospholipid headgroup binding. However, selective recognition of PE over PC will depend on the ability of the receptors to bind the cationic moiety of these phospholipids as well. Natural and synthetic oligoethers are known to bind ammonium cations, and we therefore designed and synthesized the first heteroditopic diamidocarbazole-based receptors R2 and R4 with oligoether domains for the simultaneous binding of the phosphate and ammonium moieties of the lipids, and compared their properties with model monotopic anion receptors R1 and R3 (Figure ).

Two different design strategies were employed to obtain the heteroditopic receptors. For R2, we incorporated flexible oligoether chains at the ester groups of carbazole-based receptor R1, while for R4 a benzo-18-crown-6 moiety replaced one neopentyl group of receptor R3. Crown ether derivatives are notable for their ability to selectively bind primary ammonium cations over substituted ammonium cations. Furthermore, macrocyclic compounds often show better binding than flexible compounds and a stronger effect of the oligoether was therefore expected for R4 compared to R2. The monotopic analogues R1 and R3 were used as comparisons throughout this manuscript, especially because R3 has previously been shown to exhibit antibacterial activity, likely arising from its anion transport properties. Synthetic details and characterization of R2 and R4 are provided in the Supporting Information.

Phospholipid Binding Studies

Carbazole-based receptors have a strong chromophore that allows the use of UV–vis titrations to assess the affinity of R1–R4 to PE and PC lipids. Titrations were performed in CHCl₃ using palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanol-amine (POPE) or palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) as model guests (Figure A and Supporting Information). These phospholipids do not form membranes in chloroform and remain as free molecules, allowing accurate determination of association constants. The titration data were fitted using BindFit, and the obtained binding constants are summarized in Table . The results show that all compounds except R4 bind POPC with similar strength, suggesting comparable affinity to the phosphate moiety in POPC. A modest effect of the oligoether domain was observed for receptor R2 vs R1 (3.4-fold increase in binding), whereas a much larger effect was seen for R4 vs R3 (19-fold increase). Interactions between 18-crown-6 and quaternary ammonium groups have previously been observed (even though they are not as strong as the interactions with primary ammonium cations), , and this likely accounts for the enhanced POPC binding by R4. Whereas all compounds showed binding to POPC, R1–R3 did not show any measurable binding to POPE, while a high affinity toward POPE was observed for heteroditopic receptor R4 (Table ). Unlike POPC, POPE is known to adopt a “closed” conformation in both nonpolar solvents and membranes due to intramolecular hydrogen bonding (Figure ). In this conformation, the phosphate anion is in close proximity to the ammonium cation, which makes it difficult for monotopic or flexible receptors to bind. Among all tested receptors, only R4 binds to the POPE headgroup. This unique behavior arises from its two preorganized binding domains, which enable simultaneous binding of phosphate by the diamidocarbazole moiety and ammonium by the crown ether moiety, as supported by DFT calculations (Figure B).

2.

(A) UV–vis titration curves of R1–R4 (5 × 10^–5 M) with POPE in CHCl₃ at 298 K. (B) The geometry-optimized structure of R4 and methyl phosphoethanolamine, as model of PE. (C) Stern–Volmer plots from the fluorescence titrations of 100 nm POPE:POPC (9:1) liposomes containing 1 mol % of 18:1–06:0 NBD-PE with receptors R1–R4 at 530 nm.

1. Overview of the Phospholipid Binding Properties and Biological Activity of Receptors R1–R4 .

	log  K		K SV , M –1			MIC, μM
	POPC	POPE	NBD-PC	NBD-PE	PE/PC selectivity	B . cereus	B . subtilis	S . aureus	E . faecalis	HC50, μM
R1	3.68 ± 0.02	weak	2.47 × 104	1.29 × 104	0.52	>128	2–4	32	16	885 ± 30
R2	4.21 ± 0.13	weak	2.49 × 104	1.47 × 104	0.59	>128	>128	>128	>128	524 ± 8
R3	4.50 ± 0.28	weak	2.54 × 104	1.75 × 104	0.69	>128	2–4	32	16	602 ± 12
R4	5.77 ± 0.25	>7	9.06 × 104	1.71 × 105	1.89	4	4–8	2	4	>950

^aAssociation constant obtained through UV–vis titrations in CHCl₃ at 298 K.

^bStern–Volmer constant obtained through titrations of R1–R4 into POPC or 9:1 POPE:POPC liposomes containing NBD-labeled lipids.

^cRatio of the Stern–Volmer constants obtained with PE and PC to indicate lipid selectivity.

^dMinimum inhibitory concentration (MIC) obtained using broth microdilution methods.

^eHemolytic activity of the compounds was determined by the concentration of host needed to achieve 50% hemolysis in washed single-donor human red blood cells (HC₅₀).

^fNo significant changes in absorbance spectra were observed.

^gThe percentage hemolysis was still less than 50% at the highest concentration measured.

Studies on the Interactions with Lipid Bilayers

The ability of R1–R4 to interact with phospholipids within the lipid bilayers of liposomes was investigated using our previously reported NBD-based fluorescence assay (see Supporting Information for experimental details). In this assay, Stern–Volmer-type fluorescence quenching of nitrobenzoxadiazole (NBD)-labeled lipids by putative lipid binders provides information on headgroup binding. The NBD group is attached to one of the lipid tails but is known to curl up toward the lipid–water interface. As a result, differences in Stern–Volmer constant (K _SV ) between compounds reflect variations in their local concentrations at the interface, which are influenced by differences in membrane partitioning caused by lipophilicity and/or headgroup binding. The K _SV values were determined using 100 nm POPC liposomes containing 1 mol % of 18:1–06:0 NBD-PC (1-oleoyl-2-(6-[(7-nitro-2–1,3-benzoxadiazol-4-yl)­amino]­hexanoyl)-sn-glycero-3-phosphocholine) to estimate PC binding, or 9:1 POPE:POPC liposomes containing 1 mol % of 18:1–06:0 NBD-PE (1-oleoyl-2-(6-[(7-nitro-2–1,3-benzoxadiazol-4-yl)­amino]­hexanoyl)-sn-glycero-3-phosphoethanolamine) to estimate PE binding. The Stern–Volmer plots obtained using POPC liposomes containing NBD-PC are provided in the Supporting Information, while the results obtained using 9:1 POPE:POPC liposomes containing NBD-PE are shown in Figure C. All Stern–Volmer constants are summarized in Table .

Receptors R1–R3 showed comparable Stern–Volmer constants for NBD-PC, while the Stern–Volmer constant of R4 was 3.6 times higher, consistent with the logK values obtained from UV–vis titrations (Table ). Similarly, R1–R3 had only a slight effect on NBD-PE fluorescence, suggesting weak interactions with PE, again in agreement with the titration results. In contrast, addition of R4 to POPE-containing liposomes caused pronounced quenching of the NBD fluorophore. The calculated Stern–Volmer constants reveal almost 2-fold selectivity of R4 for POPE over POPC, whereas R1–R3 display the opposite trend, favoring the more exposed phosphate group of POPC (PE/PC selectivity < 1, Table ).

Anion Transport Studies

Model receptor R3 was previously shown to transport various biologically important anions via a mobile carrier mechanism, , and to possess antibacterial activity. To evaluate the ability of the other receptors to mediate anion transport and to examine the potential of R4 to act as a dual lipid binder and anion transporter, we used 200 nm liposomes as a model lipid bilayer system. POPC:cholesterol (7:3) liposomes were loaded with the halide-sensitive dye lucigenin and suspended in a sodium nitrate buffer (Figure A; see Supporting Information for experimental details). A pulse of NaCl was added to initiate chloride transport in the presence of receptors R1–R4 (1 mol % with respect to lipids, Figure B).

3.

(A) Schematic representation of the lucigenin assay in POPC:cholesterol (7:3) liposomes. (B) Change in the normalized fluorescence due to the transport of Cl^– into 200 nm liposomes loaded with lucigenin by postinserted receptors R1–R4 at 1 mol % with respect to lipids.

To quantify the Cl^– transport rates, we fitted the experimental fluorescence decay curves with exponential functions and compared their half-time values (t _1/2, see Supporting Information). The highest transport activity was observed for R3 (t _1/2 = 89 s, Table ), followed by R4 (t _1/2 = 176 s). The reduced activity of R4 relative to R3 may result from its lower lipophilicity (Table ) and/or competition between lipid binding and anion transport. These results set R4 apart from our previously reported urea/crown ether-based PE receptors (e.g., U1–U3), which did not show any significant anion transport activity even at 6 mol % concentration, and make it the first receptor to combine both lipid binding and anion transport functions. In contrast, compounds R1 and R2 exhibited only low but measurable transport rates (t _1/2 = 453 and 441 s, respectively), suggesting that ester substituents are less favorable than chlorine substituents for promoting transmembrane anion transport.

2. Half-Time Value of Anion Transport by R1–R4, and Calculated log P Values.

	Half-time values, t 1/2 , s
Receptor	Lucigenin assay	HPTS assay	clog P *
R1	453	58	6.27
R2	441	72	5.76
R3	89	12	7.28
R4	176	31	6.04

^aCalculated half-time values from Cl^– transport experiments with postinserted receptors R1–R4 at 1 mol % with respect to lipids.

^bCalculated logP values predicted using ChemDraw software.

Because R2 and R4 were designed as heteroditopic receptors, they could potentially function as M⁺/Cl^– symporters, where M⁺ binds to the crown ether domain and Cl^– binds to the diamidocarbazole unit. To test this hypothesis, the lucigenin assay was repeated in sodium sulfate buffer with pulses of NaCl, KCl, RbCl, or CsCl. Sulfate (SO₄ ^2–) is a highly hydrophilic anion that is difficult to transport; therefore, any observed Cl^– flux under these conditions is more likely due to M⁺/Cl^– symport rather than Cl^–/SO₄ ^2– antiport. This experiment revealed that some M⁺/Cl^– symport occurs with R4 at 1 mol %, although to a smaller extent than Cl^–/NO₃ ^– antiport, whereas R2 remained inactive (see Supporting Information).

Complementary transport studies using an HPTS-based assay provided additional evidence for anion transport by R1–R4. In this assay, POPC:cholesterol (7:3) liposomes were loaded with the pH-responsive fluorophore 8-hydroxypyrene-1,3,6-trisulfonate (HPTS), and a transmembrane pH gradient was generated by adding a NaOH pulse (see Supporting Information). The ability of receptors R1–R4 (1 mol % with respect to lipids) to dissipate the pH gradient through H⁺/Cl^– symport (or the functionally equivalent OH^–/Cl^– antiport) was assessed by monitoring the change in fluorescence emission ratio between the protonated and deprotonated forms of HPTS (for R4, Na⁺/H⁺ symport or Na⁺/OH^– antiport are also possible mechanisms). The fastest equilibration of pH gradient was observed for receptor R3 (t _1/2 = 12 s), which has previously been reported as an exceptionally efficient H⁺/Cl^– transporter under similar conditions. Importantly, our novel receptor R4 also displayed notably rapid transport (t _1/2 = 31 s), while the ester-containing R1 and R2 were less active (t _1/2 = 58 and 72 s, respectively).

Taken together, the lucigenin and HPTS assays consistently demonstrate the superior activity of R3, while also revealing the unique versatility of R4, which can engage in multiple transport pathways. Compound R4 is therefore a promiscuous membrane-active receptor that binds lipid headgroups (especially PE) and transports both anions and cations, with a preference for anion transport.

Antibacterial Properties

To assess the antibacterial properties of receptors R1–R4, the minimum inhibitory concentration (MIC) was determined against a panel of bacteria strains (B. cereus, B. subtilis, S. aureus, E. faecalis, and E. coli) using standard broth microdilution methods. None of the compounds showed any activity against the Gram-negative bacterium E. coli. Although E. coli membranes contain approximately 75% PE, this lipid is predominantly located in the inner membrane, rendering access to this target challenging.

More interesting are the results obtained with Gram-positive bacteria (Table ). For example, B. cereus is a bacterium whose membrane contains large amounts of PE lipids (up to 60%), and as a result only the PE-binding receptor R4 displayed measurable antibacterial activity against this strain. Compound R4 completely inhibited B. cereus growth with an impressive MIC of 4 μM, which is much lower than that of the known PE-targeting antibiotic duramycin (MIC ∼ 32 μM). Considering that B. cereus is the closest relative to B. anthracis (anthrax), these results are very promising.

In contrast, both anion transporters R1 and R3 and the dual anion transporter and lipid binder R4 display potent antibacterial activity against the other Gram-positive bacteria: B. subtilis and the potentially pathogenic , S. aureus and E. faecalis (Table ), which all have much lower PE concentrations than B. cereus (10–20% for B. subtilis, and essentially 0% for S. aureus and E. faecalis, see Table S1). This suggests that the high PE content in B. cereus hinders transmembrane anion transport, as was recently shown for PE-rich organelle-mimicking vesicles. As a result, the anion transporters R1–R3 do not show antibacterial activity against B. cereus, but can still inhibit the growth of other Gram-positive bacteria where transmembrane anion transport remains feasible. In contrast, R4, which combines ionophoric activity with PE binding, has a broader antibacterial spectrum and is capable of inhibiting the growth of B. cereus as well as other Gram-positive bacteria.

To further confirm this hypothesis, we conducted mechanistic studies on the antibacterial activity of R3 (anionophore only) and R4 (dual PE binder and anionophore). These studies were performed using B. subtilis, because this bacterium contains intermediate levels of PE lipids, making both PE binding and anion transport plausible modes of action. A Sytox Green assay confirmed that detergent-like cell lysis or pore formation is not responsible for antibacterial activity (see Supporting Information). We next investigated the effect of the receptors on the bacterial membrane potential using the DiSC₃(5) assay. DiSC₃(5) is a membrane-permeable fluorescent dye that accumulates and self-quenches inside polarized cells. When the cells are depolarized, the dye leaks out into the external solution, causing an increase in fluorescence that can be measured spectroscopically. As expected, the lipid-binding receptor R4 clearly shows membrane depolarization of B. subtilis, whereas the chloride transporter R3 does not (Figure A). The membrane potential depends on many factors and ions, and chloride transport alone might be insufficient to cause substantial depolarization. Indeed, an influx of Cl^– would be expected to hyperpolarize rather than depolarize the membrane. In contrast, lipid headgroup binding can increase membrane permeability to a variety of ions, producing a more pronounced effect on membrane potential, as we have previously observed. ,

4.

(A) Fluorescence intensity of DiSC₃(5) in B. subtilis after 10 min incubation with gramicidin (10 μM, positive control), R3–R4 at various concentrations, clindamycin (2 μg/mL, negative control), and DMSO (1%, blank). (B) Percentage fluorescence of MQAE in B. subtilis in the absence (blank) and presence of R3–R4 at various concentrations after 5 min incubation. Results are the average of at least 2 technical × 2 biological repeats and error bars represent standard deviations.

To determine whether the anion transport activity of R3 and R4 alters cytosolic chloride concentrations, we employed the recently reported MQAE chloride influx assay, in which Cl^– transport into bacterial cells quenches MQAE fluorescence. The presumed anionophore R3 exhibited high Cl^– transport activity in B. subtilis, whereas PE receptor R4 showed significantly lower activity, even at high concentrations (Figure B). This is consistent with the liposome-based studies, which showed greater transport activity for R3 compared to R4. It also supports our hypothesis that the antibacterial activity of R3 is due to Cl^– transport only and therefore only works against bacteria that are sensitive to Cl^– transport, whereas R4 can function as both a lipid headgroup binder and moderate anionophore and therefore has activity against a broader range of bacteria.

Hemolysis Studies

One of the most common side effects of membrane-active compounds is the lysis of red blood cells. Red blood cells, like most mammalian cells, contain around 25% PE lipids, but this lipid is mostly located on the inner leaflet and thus not easily accessible. , We therefore determined the concentration of R1–R4 required to induce 50% hemolysis (HC₅₀, Table ) in single-donor human red blood cells, using a standard protocol. All receptors showed low toxicity at concentrations below 0.5 mM (see Supporting Information). Notably, even at concentrations approaching 1 mM, receptor R4 caused less than 50% hemolysis, highlighting its potential as a promising antibiotic lead.

Conclusions

In summary, we have developed the first heteroditopic molecular receptor based on a diamidocarbazole scaffold incorporating a crown ether domain, R4. This receptor is able to function both as an anionophore and as a selective binder for the bacterial lipid PE over the mammalian lipid PC in both solution and liposomes. Furthermore, R4 functions as an antibacterial agent against a variety of Gram-positive bacteria, including PE-rich B. cereus, with MIC values below 8 μM. In contrast, receptors R1–R3, which do not interact with PE lipids, function solely as anionophores and display antibacterial activity only against strains with low PE content. Overall, our findings demonstrate that combining transmembrane anion transport and lipid headgroup binding offers a promising approach for developing next-generation membrane-targeting antibiotics, although further studies will be required to fully elucidate the mechanism of action of these antibacterial agents.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c09348.

• Synthesis and characterization of compounds R1–R4, UV–vis titration, fluorescence titration, anion transport data and analysis, biological and in silico studies (PDF)

K.M.-J.: conceptualization, investigation (synthesis, UV–vis titrations, fluorescence titrations, anion transport studies); data analysis; visualization; writingoriginal draft; writingreview and editing. E.O.O.: investigation (anion transport and biological studies). M.L.K.: investigation (synthesis). M.J.C.: supervision; funding acquisition; writingreview and editing. N.B.: conceptualization; supervision; funding acquisition; writingreview and editing. All authors have read and agreed to the published version of the manuscript.

The authors declare no competing financial interest.