Tobramycin-Peptide Nucleic Acids Synergize with Polymyxin B and Other Bioactive Compounds to Inhibit Growth of Gram-Negative Bacteria

Abstract

Antisense oligomers show promise for the treatment of infectious diseases but face challenges as therapeutic agents due to limited uptake in bacteria. Although cell penetrating peptides can serve as effective delivery vehicles, they may exhibit non-specific effects and can contribute to resistance development. Aminoglycoside-peptide nucleic acid (PNA) conjugates represent a complementary delivery strategy that leverages the intrinsic cationic and membrane-interacting properties of aminoglycosides to enhance cellular uptake. In this work, we synthesized tobramycin-PNA (TOB-PNA) conjugates targeting the translation initiation region of acyl carrier protein gene (acpP) in Enterobacterales. Conjugation of TOB to PNA enhanced target binding stability, improved inhibition of translation in vitro, and enhanced activity against Gram-negative bacteria. TOB-PNA activity was further enhanced in combination with the outer membrane-permeabilizing antibiotic polymyxin B (PMB), leading to improved antibacterial activity against clinically relevant Gram-negative pathogens, including drug-resistant isolates. Resistance studies demonstrate that the TOB-PNA conjugate, both alone and in combination with PMB, maintains antibacterial activity over prolonged exposure, suggesting a higher barrier to resistance evolution than conventional TOB treatment. Mammalian toxiicty studies in Hek293 cells showed IC₅₀ >100 μM. Morphological analyses indicated that TOB-PNA treatment induces cellular stress and division abnormalities consistent with disruption of essential metabolic pathways while PMB increased membrane permeability. In addition, TOB-PNA behaved synergistically with other compounds targeting distinct cellular processes, including fatty acid biosynthesis (triclosan), folate metabolism (trimethoprim), and cell morphology (A22).

Graphical Abstract

Assisted entry of TOB-PNA in Gram-negative pathogens

Introduction

The global rise of multidrug-resistant bacteria, coupled with the slow pace of discovering new classes of antibacterial agents, has created an urgent need for innovative antimicrobial strategies.^{1, 2} Current efforts to combat antibiotic resistance include modifying existing antimicrobials, employing combination therapies, and developing novel antibacterial modalities with mechanisms distinct from conventional antibiotics.^{3, 4} One promising strategy is the use of antisense oligonucleotides (ASOs).^5–7 ASOs are short synthetic oligomers, typically 10 – 20 nucleobases in length, designed to bind complementary messenger RNA (mRNA) sequences and selectively silence target gene expression.⁸ In bacteria, ASOs commonly inhibit protein synthesis by sterically blocking ribosome access to mRNA.⁹ Among ASO chemistries, peptide nucleic acids (PNAs) have emerged as attractive antibacterial modalities due to their high affinity for complementary nucleic acid sequences, enzymatic stability, and sequence-specific mechanism of action.^10–12 PNAs have been used to target diverse bacterial processes, including virulence factors, transcription, protein synthesis, fatty acid biosynthesis via acyl carrier protein (acpP), cell wall synthesis and DNA replication.^{13, 14}

Despite the promise of ASO technology, efficient intracellular delivery remains a challenge in bacteria, particularly in Gram-negative species, where the outer membrane acts as an additional permeability barrier.^{15, 16} Nucleic acid therapeutics do not readily cross bacterial membranes due to their size and physicochemical properties; therefore, carrier molecules such as cell penetrating peptides (CPPs) are frequently employed.^{17, 18} However, CPPs may exhibit non-specific effects, immunogenicity, proteolytic instability, and potential for resistance development.^{19, 20} Consequently, alternative delivery strategies are being explored, including the use of antibiotics with inherent membrane-interacting properties as potential carriers for PNA delivery.^{14, 21, 22}

Aminoglycosides represent a particularly attractive class of antibiotics for oligonucleotide conjugation due to their intrinsic polycationic nature, strong interactions with bacterial membranes, and established uptake by bacteria.²³ Classically, aminoglycosides exert antibacterial activity through binding to the bacterial 16S ribosomal RNA A-site, leading to mistranslation and inhibition of protein synthesis.²⁴ Importantly, structural modification of aminoglycosides has been shown to modulate ribosomal binding affinity and antibacterial activity, including against resistant Gram-negative pathogens, highlighting their chemical tunability.^25–27 The first aminoglycoside-oligonucleotide conjugates employed neomycin and kanamycin, which enhanced the hybridization stability of DNA/RNA●RNA duplexes.^{28, 29} More recently, aminoglycosides have been conjugated to PNAs using a range of bioconjugation strategies, such as nitrile-aminothiol coupling and copper(I)-catalyzed azide-alkyne cycloaddition.^{21, 30} Studies of amikacin (AMK)- and neomycin (NEO)-PNA conjugates targeting the essential acpP transcript demonstrated antibacterial activity in Escherichia coli (E.coli) and Salmonella Typhimurium.²¹ Reporter assays further confirmed antisense-mediated inhibition of acpP expression and indicated that the aminoglycoside component contributes to membrane interactions, supporting a dual mechanism of action involving both sequence-specific gene silencing and enhanced cellular uptake. However, despite evidence of target engagement, these conjugates exhibited only moderate antibacterial activity, with minimum inhibitory concentrations (MICs) of 8 – 32 μM, compared with 2 – 4 μM CPP-PNA systems³¹, highlighting remaining limitations in delivery efficiency.

Combination treatment strategies may improve the delivery and antibacterial activity of aminoglycoside-PNA conjugates by overcoming the permeability barrier of the Gram-negative outer membrane. Indeed, synergistic interactions between the outer membrane-permeabilizing antibiotic polymyxin B (PMB) and PNAs have been shown to enhance antibacterial activity, lowering MICs by several orders of magnitude.^{32, 33} Furthermore, we previously demonstrated that co-treatment with PMB enhances the activity of antisense oligomers, including PNAs targeting the non-essential gene lacZ, resulting in improved gene knockdown efficiency.³⁴ These findings motivated us to investigate whether PMB could similarly potentiate aminoglycoside-PNA conjugates.

To address this question, we synthesized and evaluated tobramycin-PNA (TOB-PNA) conjugates targeting the essential acpP transcript in members of the Enterobacterales. TOB was selected as the aminoglycoside scaffold because, relative to AMK (+4) and NEO (+6), it possesses an intermediate net charge (+5), which we hypothesized may provide a balance between membrane interaction and antibacterial activity while potentially reducing the ototoxicity associated with AMK and NEO.³⁵ TOB-PNA conjugates incorporating various linkers were assessed for antibacterial activity and antisense efficacy both alone and in combination with PMB. Compared to unconjugated PNAs, TOB-PNA conjugates exhibited improved antibacterial activity, which was further enhanced by co-treatment with PMB through synergistic disruption of outer membrane integrity. Additional synergy was observed with agents targeting complementary cellular processes, including triclosan, trimethoprim, and A22. Together with acpP mRNA binding and antimicrobial susceptibility analyses, these findings demonstrate enhanced activity of TOB-PNA conjugates across Enterobacterales, including extensively drug-resistant (XDR) Klebsiella pneumoniae, Salmonella Typhimurium and Shigella flexneri, and support outer membrane-targeting combination strategies for improving aminoglycoside-PNA delivery in Gram-negative bacteria.

Results and Discussion

Conjugation of TOB to PNA enhances PNA binding stability and improves in vitro translation inhibition

The PNAs evaluated in this study target a 10-base sequence within the highly conserved acpP translation initiation region, which shares 100% homology among members of the Enterobacteriaceae and has been widely used in antibacterial PNA research, making it a well-established model target.^{9, 13} Each PNA contained a C-terminal L-lysine residue to improve solubility.³⁶ TOB was conjugated to the N-terminal end of the PNA either directly or via a linker through amide coupling between the PNA or linker amine and a carboxylic acid installed at the 6”-position of TOB (Figure 1, Scheme S1, Table S1). This strategy is compatible with standard solid-phase synthesis, enabling on-resin coupling under routine conditions with simple removal of excess reagents by washing prior to cleavage. It is a high-yielding, robust, and modular approach that accommodates both direct and linker-mediated conjugation without additional modification of either component. The linkers were selected to vary in length and physicochemical properties. Amino-(polyethylene glycol)₂ acid (PEG) was included as a benchmark linker because it is commonly used in CPP-PNA conjugates to improve solubility and reduce steric interference between conjugated components. Additional linkers included γ-aminobutyric acid (GABA) and 6-aminohexanoic acid (Ahx), both shorter than PEG and composed of alkyl chains, with Ahx being longer and more hydrophobic than GABA. Together, these linkers enable an evaluation of the effects of linker length and hydrophobicity on the antibacterial activity of TOB-PNAs.

Figure 1.

Design and synthesis of TOB-PNAs. a. Conceptual design of TOB-PNAs. A 10-base sequence complementary to and spanning the acpP mRNA start codon (underlined) was used (light blue box). TOB was conjugated to the N-terminal end of the PNA either directly or via one of three linkers (yellow box). TOB was modified with a carboxylic acid at the 6”-position to facilitate amide coupling to the PNA (TOB-COOH, green box). In total, four acpP-targeting TOB-PNAs were synthesized (dark blue box). b. Representative synthesis of TOB-PNAs. Standard Fmoc-based solid-phase synthesis procedures were used, including piperidine-mediated Fmoc deprotection, HATU/DIPEA-mediated coupling, and capping with acetic anhydride. TOB-COOH was incorporated by amide coupling, and the final TOB-PNAs were cleaved from the resin by TFA treatment.

The effects of the linkers and TOB conjugation on PNA hybridization to acpP mRNA were evaluated by thermal denaturation. Melting temperatures (T_m) were determined for ten PNAs. These included five unconjugated PNAs: the fully complementary PNA (1), linker-containing PNAs Ahx-PNA (2), GABA-PNA (3), PEG-PNA (4), and an eight base mismatch (MM) PNA (5) and their corresponding TOB-PNA conjugates: TOB-PNA (6), TOB-Ahx-PNA (7), TOB-GABA-PNA (8), TOB-PEG-PNA (9), and the TOB-MM PNA (10) (Table S2, Figure S1–S16). Among the unconjugated PNAs, all three linkers increased duplex stability relative to PNA 1, which lacks a linker (Table 1). The magnitude of stabilization decreased with increasing linker length, with the greatest stabilization observed for GABA-PNA 3 (+3.7 °C) and the smallest for PEG-PNA 4 (+1.0 °C).

Table 1.

Thermal melting temperatures (T_m) for acpP-targeting PNAs with and without TOB.

PNA #	Description	Tm (°C)a	ΔTm (°C)b
1	PNA	40.6 (1.7)	---
2	Ahx-PNA	43.4 (0.7)	+2.8
3	GABA-PNA	44.3 (0.5)	+3.7
4	PEG-PNA	41.6 (4.5)	+1.0
5	MM PNA	nd	---
6	TOB-PNA	47.0 (0.03)	+6.4
7	TOB-Ahx-PNA	50.3 (0.2)	+9.7
8	TOB-GABA-PNA	48.6 (0.5)	+8.0
9	TOB-PEG-PNA	48.3 (0.9)	+7.7
10	TOB-MM PNA	28.0 (1.2)	−12.6

^aValues are reported as mean ± SD from two independent experiments (n = 2).

^bΔT_m values represent the difference between the T_m of each PNA- or TOB-PNA●RNA duplex and that of the PNA 1●RNA duplex. nd, not determined, as T_m was below 20 °C. For PNA seqeunces, see Table S2.

TOB conjugation further stabilized fully complementary PNA●RNA duplexes, increasing T_m values by 6–10 °C relative to PNA 1. The greatest stabilization was observed for TOB-Ahx-PNA 7 (ΔT_m = +9.7 °C), followed by TOB-GABA-PNA 8 (ΔT_m = +8.0 °C), TOB-PEG-PNA 9 (ΔT_m = +7.7 °C), and TOB-PNA 6 (ΔT_m = +6.4 °C). Although the T_m of the mismatch PNA 5 could not be determined (<20 °C), TOB conjugation increased its T_m to 28.0 °C, consistent with electrostatic interactions between TOB and the RNA phosphate backbone; however, the resulting duplex remained substantially less stable than the fully complementary PNA 1●RNA duplex, as expected. These results closely match previous reports of 10–12 °C T_m increases for AMK- and NEO-PNA conjugates, further supporting that aminoglycoside conjugation enhances PNA●RNA duplex stability.²¹

We next evaluated functional activity using a cell-free in vitro translation assay in which PNAs were assessed for their ability to inhibit expression of an acpP-enhanced green fluorescent protein (EGFP) reporter construct (Figure 2a). Because the PNAs target the translation initiation region, binding is expected to sterically block ribosome access and reduce EGFP expression. The complementary unconjugated PNA 1 reduced EGFP expression by 83% relative to the untreated plasmid reporter control. Linker-containing complementary PNAs 2-4 produced similar reductions in expression (78–85%). In contrast, the unconjugated mismatch PNA 5 produced minimal reduction in EGFP expression, supporting sequence-specific antisense activity.

Figure 2.

Evaluation of EGFP reporter expression in a cell-free in vitro translation assay. a. Normalized EGFP expression for unconjugated PNAs (1-5), TOB-PNAs (6-10), TOB alone, and positive (POS) and negative (NEG) controls. All compounds were evaluated at a final concentration of 5 μM. The POS control represents maximal EGFP expression induced by isopropyl β-D-1-thiogalactopyranoside (IPTG) in the absence of PNA or TOB treatment. The NEG control represents baseline EGFP expression in the absence of IPTG induction. Fluorescence values were normalized to the POS control (100% expression) and NEG control (0% expression). B. Raw EGFP fluorescence measurements for the complementary unconjugated PNA 1 and TOB-PNAs 6-9, each tested at a final concentraiton of 5 μM. In vitro experiemnts were performed at 37 °C, Data are presented as mean ± standard deviation from two independent experiments, each performed with two technical replicates.

Conjugation of TOB to complementary PNAs further reduced EGFP expression, with reductions reaching up to 98%. This enhanced inhibition is consistent with the increased thermal stability of the TOB-PNA conjugates 6-9 relative to the unconjugated PNA 1, suggesting that stronger hybridization to the acpP mRNA target contributes to more effective steric blockade of translation. For comparison, TOB alone reduced EGFP expression by 93%, indicating that the TOB moiety itself contributes substantially to reduced reporter expression due to binding and inactivation of ribosomes. This effect was also observed with the mismatch TOB-PNA 10, which produced a 16% reduction in EGFP expression. Comparison of raw EGFP fluorescence measurements for PNA 1 and TOB-PNAs 6-9 suggested linker-dependent differences in activity (Figure 2b). TOB-Ahx-PNA 7 and TOB-PEG-PNA 9 exhibited the lowest fluorescence signals among the conjugates, whereas TOB-PNA 6 (no linker) and TOB-GABA-PNA 8 showed comparatively higher fluorescence values. These trends indicate weaker inhibitory activity for the latter conjugates.

TOB-PNA conjugates exhibit enhanced antibacterial activity, disrupt cell division, and synergize with polymyxin B in Gram-negative pathogens

The biological activity of the TOB-PNA conjugates was evaluated in the presence and absence of PMB against Gram-negative bacteria by determining their minimum inhibitory concentrations (MICs). E. coli MG1655 (K-12), an aminoglycoside-susceptible laboratory strain, was included as a wild-type reference. E. coli SM101, an aminoglycoside-supersusceptible K-12 derivative with a mutation affecting lipopolysaccharide (LPS) biosynthesis and outer membrane integrity, was included to evaluate the effect of enhanced membrane permeability on TOB-PNA activity. K. pneumoniae 0120, an aminoglycoside-resistant clinical isolate (Table S3), was included to assess activity against a clinically relevant Gram-negative pathogen.

Unconjugated complementary PNAs 1-4 exhibited measurable MIC values only in the mutant strain (8–16 μM), whereas activity in E. coli MG1655 and K. pneumoniae 0120 was generally above the highest concentration tested (> 32 μM), consistent with poor cellular uptake of PNA alone (Table 2). Conjugation of TOB to the PNAs improved antibacterial activity. Relative to their unconjugated counterparts, all complementary TOB-PNAs (6-9) exhibited enhanced activity against E. coli MG1655 and SM101, with MIC improvements of up to four-fold (MIC = 4–32 μM). These values are comparable to those reported for acpP-targeting neomycin- and amikacin-PNA conjugates in mutant and wild-type E. coli strains.²¹ In K. pneumoniae 0120, MIC values remained undetectable for TOB-GABA-PNA 8 and TOB-PEG-PNA 9 within the concentration range tested, whereas TOB-PNA 6 and TOB-Ahx-PNA 7 exhibited MIC values of 16 and 32 μM, respectively. The contribution of TOB to antibacterial activity is evident from comparison of mismatched PNA 5 and TOB-PNA 10. TOB-PNA 10 displayed measurable activity against E. coli MG1655 and a two-fold improvement in MIC against SM101, indicating that the TOB moiety contributes to antibacterial activity independent of sequence-specific antisense effects.

Table 2.

Minimum Inhibitory Concentrations (MICs) of acpP-targeting PNAs and TOB-PNA conjugates against E. coli MG1655, E. coli SM101, and K. pneumoniae 0120.^a

		E. coli SM101		E. coli MG1655		K. pneumoniae 0120
PNA #	Description	− PMB	+ PMB	− PMB	+ PMB	− PMB	+ PMB
-	(KFF)3-K	4	nd	8–16	nd	>32	nd
-	(KFF)3-K-PNA	1	nd	1	nd	2	nd
-	(KFF)3-K-MM PNA	8	nd	8	nd	16	nd
-	TOB	0.5	0.25	1	0.5	16	16
-	PMB	---	0.25	---	1	---	2
1	PNA	16	4	32	16	>32	32
2	Ahx-PNA	8	4	32	16	>32	16
3	GABA-PNA	16	4	>32	32	>32	32
4	PEG-PNA	16	8	>32	32	>32	>32
5	MM PNA	32	>32	>32	>32	>32	>32
6	TOB-PNA	4	1	16	4	16	8
7	TOB-Ahx-PNA	4	2	16	4	32	4
8	TOB-GABA-PNA	8	4	32	4	>32	8
9	TOB-PEG-PNA	4	2	32	4	>32	16
10	TOB-MM PNA	16	8	>32	32	>32	>32

^aMIC values are reported in μM. All compounds were tested at concentrations up to 32 μM. Values reported as >32 μM indicate that no MIC was determined within the concentration range tested. Modal MIC values are reported. TOB = tobramycin; PMB = polymyxin B. nd = not determined. The concentration of PMB in the combination studies was set at a sub-inhibitory concentration for each bacterial isolate: 0.063 μM for E. coli SM101, 0.25 μM for E. coli MG1655 and 0.50 μM for K. pneumoniae 0120.

To benchmark TOB-PNAs, we compared them to a literature-standard CPP-PNA in which the same acpP-targeting PNA was conjugated to the CPP (KFF)₃-K.^{31, 37} Although TOB conjugation improved PNA antibacterial activity, TOB-PNAs exhibited lower potency than the CPP-PNA, which displayed MIC values of 1 – 2 μM across all three strains (Table 2). The antibacterial activity of the CPP-PNA arises from contributions of both components, as the CPP alone and the mismatch CPP-PNA exhibited substantially higher MIC values than the complementary CPP-PNA conjugate. These findings indicate that, while TOB enhances PNA antibacterial activity, further optimization of conjugate design will be required to achieve the potency observed for CPP-PNAs.

Synergistic interactions between polymyxin B (PMB) or their derivatives and antisense PNAs have been reported to potentiate antibacterial activity.^{31, 34, 38} We therefore evaluated the activity of PNAs and TOB-PNA conjugates in the presence of PMB. As a reference, PMB co-treatment enhanced the activity of TOB by two-fold in both E. coli MG1655 and SM101; however, no improvement was observed in the extensively drug-resistant K. pneumoniae 0120 strain (MIC = 16 μM) (Table 2). For unconjugated complementary PNAs 1-4, PMB co-treatment generally improved activity across all three strains, yielding MIC values of 4–32 μM. Given that PMB is itself an outer membrane-disrupting antibiotic, a two-fold enhancement in the activity of mismatched TOB-PNA 10 was also observed. More pronounced effects were observed for the TOB-PNA conjugates, with MIC values as low as 1–2 μM in E. coli SM101 for TOB-PNA 6, TOB-Ahx-PNA 7, and TOB-PEG-PNA 9, and 4 μM in E. coli MG1655 for all complementary TOB-PNAs. In K. pneumoniae 0120, a MIC of 4 μM was observed for TOB-Ahx-PNA 7. These correspond to four- to eight-fold improvements upon PMB co-treatment, consistent with synergistic enhancement.

TOB-Ahx-PNA 7 was selected for further evaluation based on its enhanced thermal stability, greater reduction of EGFP expression, and broad antibacterial activity across multiple bacterial strains. Fluorescence microscopy was used to examine morphological effects following TOB-PNA treatment. E. coli MG1655 and K. pneumoniae 0120 cells were treated and stained with DAPI to visualize chromosomes, FM4–64 to visualize cell membranes, and SYTOX Green to indicate membrane permeabilization. In E. coli, TOB-Ahx-PNA 7 induced cell chaining and membrane blebbing compared with untreated cells (Figure 3a–b). These morphological features are consistent with impaired cell division and cell envelope stress, respectively. In K. pneumoniae 0120, pronounced cell elongation was observed, with individual cells extending beyond 50 μm, further supporting disruption of normal cell division by the TOB-PNA conjugate (Figure 3c). Collectively, TOB-Ahx-PNA 7 induced morphological abnormalities without obvious evidence of cell lysis. These findings are consistent with previous reports of acpP-targeting PNAs in Gram-negative bacteria.³⁹

Figure 3.

Fluorescence microscopy images showing the effect s of TOB-Ahx-PNA 7 on bacterial cell morphology. a. Untreated E. coli MG1655, b. E. coli treated with 32 μM TOB-Ahx-PNA 7 showing cell chaining and membrane blebbing (right), c. K. pneumoniae 0120 treated with 32 μM TOB-Ahx-PNA 7 showing pronounced cell elongation without chaining, d. E. coli treated with 5 μM TOB showing toroidal chromosome formation (lower yellow arrow), membrane blebbing and/or cell elongation (upper two yellow arrows), and a cell undergoing normal cell division (light blue arrow), e. E. coli treated with 0.25 μM sub-MIC PMB showing SYTOX Green uptake and cell lysis, f. E. coli treated with 16 μM TOB-Ahx-PNA 7 in combination with 0.125 μM sub-MIC PMB showing cell chaining, membrane blebbing, and cell lysis, g. K. pneumoniae treated with 16 μM TOB-Ahx-PNA 7 in combination with 0.125 μM sub-MIC PMB showing cell elongation and cell lysis. Cells were stained with DAPI (blue) to visualize chromosomes, FM4–64 (red) to visualize cell membranes, and SYTOX Green (green) to indicate membrane permeabilization.

For comparison, TOB-treated E. coli exhibited toroidal chromosome formation and membrane perturbations, including blebbing and/or elongation, consistent with the known effects of aminoglycosides (Figure 3d).⁴⁰ In contrast, PMB treatment resulted in extensive SYTOX Green uptake, indicating compromised cytoplasmic membrane integrity and access of the dye to intracellular nucleic acids (Figure 3e).⁴¹ In both E. coli and K. pneumoniae, combination treatment with TOB-Ahx-PNA 7 and sub-MIC PMB produced cell elongation accompanied by cellular debris indicative of cell lysis (Figure 3 f–g). Membrane blebbing was also observed in E. coli under combination treatment. Together, these findings suggest that TOB-Ahx-PNA 7 primarily disrupts cell division and induces cell envelope stress, whereas PMB contributes membrane damage and lysis, resulting in distinct but complementary morphological effects during combination treatment.

Synergy between TOB-Ahx-PNA 7 and PMB was confirmed using a checkerboard two-way dilution combination assay. In the two strains carried forward from the initial antibacterial screen, E. coli SM101 and K. pneumoniae 0120, fractional inhibitory concentration indices (FICI) of 0.5 were obtained for both strains, meeting the accepted threshold for synergy (Table 3).^{42, 43} The interaction was further evaluated against an expanded panel of Gram-negative pathogens, including the aminoglycoside-susceptible reference strain E. coli 25922 and aminoglycoside-resistant clinical isolates Salmonella enterica serovar Senftenberg 0127 (S. Senftenberg), Shigella flexneri 0421 (S. flexneri), and K. pneumoniae 0851, to assess the breadth and conservation of PMB-mediated potentiation across strains with differing aminoglycoside susceptibilities (Table S3). Synergy was also observed in E. coli 25922 and XDR isolate S. Senftenberg 0127. Although FICI values for XDR isolates S. flexneri 0421 and K. pneumoniae 0851 could only be estimated because the MIC of TOB-Ahx-PNA 7 alone exceeded the highest concentration tested (32 μM), co-treatment consistently reduced the MICs of both TOB-Ahx-PNA 7 and PMB, indicating mutual potentiation. Collectively, these results demonstrate robust, broadly conserved synergy between TOB-Ahx-PNA 7 and PMB across diverse Gram-negative pathogens.

Table 3.

Interaction of TOB-Ahx-PNA 7 with PMB in Enterobacteriaceae.

Strain	TOB-Ahx-PNA 7 MIC	TOB-Ahx-PNA 7 + PMB MIC	PMB MIC	PMB + TOB-Ahx-PNA 7 MIC	FICI
E. coli 25922	16	4	0.5	0.125	0.500
E. coli SM101	4	1	0.25	0.0625	0.500
S. Senftenberg 0127	32	4	1	0.25	0.375
S. flexneri 0421	>32	8	1	0.25	<0.375
K. pneumoniae 0120	32	8	2	0.5	0.500
K. pneumoniae 0851	>32	4	1	0.25	<0.375

MIC values are reported in μM. Compounds were tested at concentrations up to 32 μM. Values reported as >32 μM indicate that no MIC was determined within the concentration range tested. FICI = fractional inhibitory concentration index. For compounds with MIC values >32 μM, FICI values were estimated using 32 μM as the MIC.

TOB-PNA/PMB combination exhibits bactericidal activity, favorable safety, and a high barrier to resistance

Time-kill assays were performed to confirm the synergistic interaction between TOB-Ahx-PNA 7 and PMB and to determine whether the combination is bacteriostatic or bactericidal.⁴⁴ These assays measure antimicrobial activity against planktonic populations in shaking culture, whereas checkerboard assays are endpoint measurements performed in microtiter plates that may also support biofilm-associated growth. Because planktonic cells are generally more susceptible, comparable or enhanced activity relative to checkboard assays is expected. At 1x or 2x the MIC for the combination identified in checkboard assays (Figure 4a), a ≥3 log₁₀ reduction in colony forming units per milliliter (CFU/mL) relative to the initial inoculum was achieved within four hours, indicating a bactericidal effect (Figure 4b).⁴⁵ At 2x MIC, this ≥3 log₁₀ reduction was sustained up to eight hours. Regrowth under both conditions was observed between eight and 24 hours, suggesting that repeated dosing would be required for sustained activity. Neither TOB-Ahx-PNA 7, PMB, nor TOB alone produced sustained bactericidal activity, with only transient reductions in viable counts followed by regrowth. The observation that bactericidal activity only occurs in combination suggests that simultaneous outer membrane permeabilization and intracellular target engagement may be required for effective killing.

Figure 4.

Synergistic antibacterial activity of the TOB-Ahx-PNA 7/PMB combination against planktonic and biofilm-associated bacteria. a. Representative checkerboard assay showing growth inhibition of E. coli 25922 (left) and K. pneumoniae 0120 (right) following treatment with TOB-Ahx-PNA 7 and PMB. b. Time-kill assay of E. coli 25922 treated with Sub -MIC PMB, TOB-Ahx-PNA 7, their combination, or TOB. Sub-MIC PMB 1X = 0.125 μM. Sub-MIC PMB 2X = 0.250 μM. TOB 1X MIC = 1 μM. TOB-Ahx-PNA 7 1X MIC = 16 μM. TOB-Ahx-PNA 7 2X MIC = 32 μM. TOB-Ahx-PNA 7 + PMB 1X = (4 μM and 0.125 μM, respectively). TOB-Ahx-PNA 7 + PMB 2X = (8 μM and 0.250 μM). Error bars represent the standard deviation of three technical replicates. c. Biofilm biomass remaining after 16 h treatment of established E. coli 25922 or K. pneumoniae 0120 biofilms with TOB-Ahx-PNA 7, PMB, or their combination, expressed as a percentage of the untreated control. The initial OD₆₀₀ of established biofilms was 0.10 ± 0.02 for both strains. Error bars represent the standard deviation of three technical replicates.

In contrast to planktonic growth, bacteria commonly form biofilms that are clinically relevant because they can develop on medical devices, such as catheters, and exhibit markedly increased tolerance to antibiotics, often requiring substantially higher concentrations for eradication.⁴⁶ The activity of TOB-Ahx-PNA 7, PMB, and their combination against established biofilms was evaluated in E. coli 25922 and K. pneumoniae 0120. All treatments reduced established biofilm biomass more effectively in E. coli than in K. pneumoniae (Figure 4c), potentially reflecting the protective role of the capsule in the latter.⁴⁷ At 2x the MIC of TOB-Ahx-PNA 7 (32 μM), biofilm biomass was reduced to ~30 and ~60% in E. coli and K. pneumoniae, respectively. In comparison, PMB required 8x the MIC (8 μM) to reduce biofilm biomass to ~10% and ~30%, respectively. Notably, substantially lower concentrations of both TOB-Ahx-PNA 7 (16 μM) and PMB (2 μM) achieved comparable reductions in biofilm biomass when used in combination, demonstrating that the synergistic interaction extends to established biofilms while enabling dose reduction of both agents.

To assess mammalian compatibility and potential off-target membrane toxicity, cytotoxicity in human embryonic kidney cells (HEK293T) and hemolytic activity in sheep red blood cells (SRBCs) were evaluated for the TOB-Ahx-PNA 7 alone and in combination with PMB. The concentration required to reduce HEK293T viability by 50% (IC₅₀) was 145 μM for TOB-Ahx-PNA 7 and 103 μM for the combination (Figure 5a–b). These values were lower than those observed for PMB (171 μM) and TOB alone (>300 μM) (Figure 5c–d). Nevertheless, the therapeutic window for the TOB-PNA/PMB combination remained favorable, given its antibacterial activity at low micromolar concentrations (MIC = 4 μM). The IC₅₀ of the reference (KKF)₃K-PNA conjugate (122 μM) was comparable to that of the TOB-PNA/PMB combination (Figure 5e). Low hemolytic activity was observed for TOB-Ahx-PNA 7 both alone and in combination with PMB, with concentrations of >256 μM and 90 μM, respectively, required to induce 50% lysis of SRBCs (Figure S22). Collectively, these results support a favorable safety profile for TOB-Ahx-PNA 7 in combination with PMB.

Figure 5.

Cytotoxicity of TOB-Ahx-PNA 7, PMB, TOB, and (KFF)₃-K-PNA in HEK293T cells. a. TOB-Ahx-PNA 7, b. TOB-Ahx-PNA 7 in combination with PMB. For the combination treatment, TOB-Ahx-PNA 7 and PMB were added at equimolar concentrations, c. PMB, d. TOB, e. (KFF)₃-K-PNA. HEK293T cell viability was plotted as a function of the log10 compound concentration to determine the concentration that reduced cell viability by 50% (IC₅₀) after 48 hours. Data represent the mean of two independent experiments, each performed with three technical replicates. Error bars indicate standard deviation.

A 14-day resistance development assay was performed to evaluate the propensity for resistance emergence under repeated exposure and to assess whether PMB co-treatment reduces the evolutionary capacity of bacteria to escape TOB-PNA-mediated killing. The tobramycin-susceptible isolate E. coli MG1655 and the XDR and tobramycin-resistant K. pneumoniae 0120 were used. Test compounds included TOB, PMB and TOB-Ahx-PNA 7, and the corresponding combinations TOB + PMB and TOB-Ahx-PNA 7 + PMB. In combination treatments, PMB was used at sub-micromolar concentrations that did not produce significant growth inhibition on its own in either strain. In E. coli, transient increases in apparent resistance were observed for populations exposed to TOB, PMB, and the combination, followed by reversion to baseline MICs after a 72 hour drug-free recovery period (Figure 6a). In contrast, MICs for TOB-Ahx-PNA 7 alone and in combination with PMB remained stable over the 14-day serial passage (Figure 6b).

Figure 6.

Resistance development during 14 days of serial passage with individual drugs and combination treatments in E. coli MG1655 (a, b) and K. pneumoniae 0120 (c, d). MICs were determined on days 1, 4, 7, 10 and 14. The 72h recovery represents day 14 cultures passage into fresh, drug-free medium and incubated for an additional 72 h before MIC determination to assess the stability of resistance. Two technical replicates were performed for each drug or drug combination. PMB was used at a sub-inhibitory concentration in combination treatment (0.25 μM for E. coli MG1655 and 0.50 μM K. pneumoniae 0120). Daily MIC values are provided in Table S4.

In K. pneumoniae, rapid high-level resistance to TOB emerged by days 4–5, with MICs increasing from 16 μM to >128 μM for an 8-fold increase, and this elevated resistance persisted after the 72 hour recovery period (Figure 6c). A similar upward shift was observed in the TOB + PMB condition, where the MIC increased from 8 μM to 64 μM. In contrast, PMB alone produced only transient resistance, with a four-fold increase in MIC observed at days 9–10 followed by reversion to baseline levels. Despite this transient adaptation to PMB, MICs for TOB-Ahx-PNA 7, both alone and in combination with PMB, remained largely stable at 32 μM and 4–8 μM, respectively (Figure 6d). Collectively, these findings demonstrate that the TOB-PNA conjugate, both alone and in combination with PMB, maintains antibacterial activity over prolonged exposure, suggesting a higher barrier to resistance evolution than conventional TOB treatment.

PMB Potentiates TOB-PNA Uptake via Enhanced Outer Membrane Permeabilization

Outer membrane permeability of TOB-Ahx-PNA 7 was assessed to quantitatively evaluate the mechanism by which TOB conjugation and PMB co-treatment enhance PNA uptake. E. coli MG1655 cells were incubated with N-phenyl-1-naphthylamine (NPN) dye, a hydrophobic fluorescent probe used to report changes in outer membrane permeability.⁴⁸ NPN exhibits weak fluorescence in aqueous environments but increases in intensity upon partitioning into hydrophobic regions of the bacterial envelope. In intact cells, the outer membrane limits NPN access to these regions; thus, increased fluorescence reflects outer membrane disruption and dye access to lipid-rich domains.

NPN fluorescence was the highest in cells incubated with PMB (10 μM), which served as a positive control for outer membrane permeabilization (Figure 7a). All samples were therefore normalized to PMB-induced NPN fluorescence. Although used at a sub-MIC concentration, PMB (0.25 μM) still produced ~70% of the maximal signal, indicating substantial outer membrane perturbation. TOB (10 μM) produced ~30% of the PMB-normalized signal, indicating a weaker permeabilizing effect. Neomycin (NEO, ~50%) and amikacin (AMK, ~20%) yielded values comparable to those reported in previous NPN studies.²¹ These differences are consistent with the relative cationic charge of the aminoglycosides (NEO +6, TOB +5, AMK +4), supporting the idea that increased positive charge may enhance interactions with the negatively charged outer membrane. Addition of magnesium (Mg²⁺) reduced NPN fluorescence for all compounds, consistent with competition between these cationic compounds and divalent cations for binding sites on the outer membrane.⁴⁹

Figure 7.

Outer membrane permeabilization of E. coli MG1655 measured by the NPN uptake assay. a. NPN fluorescence of cells treated with polymyxin B (PMB), neomycin (NEO), amikacin (AMK), tobramycin (TOB), Ahx-PNA 2, and TOB-Ahx-PNA 7. Fluorescence values were normalized to PMB (10 μM). Blue bars represent treatment in the absence of Mg²⁺, and orange bars represent treatment in the presence of Mg²⁺ (3 mM). Data are presented as mean ± standard deviation. Sample sizes were n = 2 for NEO, AMK, Ahx-PNA 2 (20 and 40 μM), and TOB-Ahx-PNA 7 (20 and 40 μM), whereas all other conditions were measured in triplicate (n = 3). All test compounds were used at 10 μM unless otherwise stated. b. NPN fluorescence of cells treated with TOB, Ahx-PNA 2, and TOB-Ahx-PNA 7 in combination with sub-MIC PMB (0.25 μM), and compared with (KFF)₃-K and (KFF)₃-K-PNA. Data are presented as mean ± standard deviation from three independent experiments. All test compounds were used at 10 μM.

As expected, unconjugated Ahx-PNA 2 produced negligible NPN fluorescence, indicating little to no outer membrane permeabilization. In contrast, TOB-Ahx-PNA 7 generated a modest PMB-normalized signal (14%), demonstrating that TOB retained some ability to interact with and perturb the outer membrane following conjugation, albeit to a lesser extent than TOB alone (34%). The decrease in fluorescence observed in the presence of Mg²⁺ further supports the involvement of electrostatic interactions in outer membrane perturbation. Increasing the concentration of TOB-Ahx-PNA 7 to 40 μM increased NPN fluorescence to levels comparable to those observed for TOB alone, whereas Ahx-PNA 2 remained inactive at the same concentration. Collectively, these results suggest that the TOB moiety of the conjugate confers modest outer membrane-permeabilizing properties to PNA.

Co-treatment with sub-MIC PMB (0.25 μM) substantially increased NPN fluorescence produced by TOB (+63%), Ahx-PNA 2 (+72%), and TOB-Ahx-PNA 7 (+50%) (Figure 7b). The permeabilization activity of these combinations was compared with that of (KFF)₃-K-PNA. As observed for TOB-Ahx-PNA 7, conjugation of (KFF)₃-K to PNA reduced its permeabilization activity from 99% to 56% of the PMB-normalized signal, indicating that conjugation attenuated the membrane-disrupting properties of the carrier. Notably, TOB-Ahx-PNA 7 in combination with PMB generated a comparable PMB-normalized signal (64%). These results suggest that PMB-mediated outer membrane disruption can compensate for the reduced outer membrane activity associated with PNA conjugation and may facilitate intracellular access of TOB-Ahx-PNA 7.

In addition to outer membrane permeabilization, cytoplasmic membrane damage was assessed using propidium iodide (PI), a nucleic acid stain that reports loss of inner membrane integrity. In untreated E. coli MG1655, negligible PI fluorescence was observed, whereas 0.1% Triton-x-100 and sub-MIC PMB induced uptake. Treatment with PNA 1 resulted in minimal PI fluorescence, while co-treatment with sub-MIC PMB increased PI uptake. In contrast, TOB-PNA 6 induced detectable PI uptake, which was further increased by co-treatment with sub-MIC PMB. Collectively, these results suggest that TOB-PNA exerts modest effects on cytoplasmic membrane integrity that are enhanced by PMB-mediated permeabilization.

TOB-PNA conjugate shows synergy with four mechanistically distinct compounds in Enterobacteriaceae

The interaction of TOB-Ahx-PNA 7 with four mechanistically distinct compounds, comprising antibiotics and chemical probes (Table S5), was evaluated to assess whether synergy extended beyond outer membrane permeabilization. Polymyxin B nonapeptide (PMBN) acts as a benchmark outer membrane permeabilizer that increases Gram-negative membrane permeability with minimal bactericidal activity⁵⁰, enabling discrimination between effects driven primarily by outer membrane disruption and those arising from the additional bactericidal and stress-inducing properties of PMB. Triclosan (TCS) inhibits fatty acid biosynthesis via FabI⁵¹, probing whether perturbation of lipid metabolism influences TOB-PNA activity.

Trimethoprim (TMP) is a dihydrofolate reductase inhibitor used clinically in combination therapy and serves as a representative intracellular metabolic inhibitor.⁵² A22 (S-(3,4-dichlorobenzyl) isothiourea) disrupts MreB-dependent cell shape and cytoskeletal organization^{53, 54}, allowing evaluation of whether perturbation of bacterial morphology influences TOB-PNA potentiation. Collectively, this panel enables systematic dissection of whether TOB-PNA synergy is restricted to outer membrane permeabilization or can also arise from modulation of intracellular metabolic, biosynthetic, or structural pathways.

Synergistic to additive interactions were observed between TOB-Ahx-PNA 7 and PMBN in all strains except K. pneumoniae 0851, for which the interaction was estimated to be indifferent (FICI <2.00) (Figure 9, Table S6). These findings suggest that outer membrane permeabilization alone is insufficient to fully account for the broader potentiation observed with PMB. With TCS, synergistic interactions were observed in all strains except S. flexneri 0421, where the interaction was estimated to be additive (FICI <0.625). The broad synergistic interaction suggests that simultaneous targeting of FabI and acpP enhances antibacterial activity through complementary inhibition of fatty acid biosynthesis. TMP exhibited synergistic interactions in all strains except K. pneumoniae 0120, where the interaction was estimated to be indifferent (FICI <1.25), suggesting that disruption of folate metabolism can potentiate TOB-PNA activity despite targeting a distinct cellular pathway. Lastly, A22 displayed synergistic interactions in all strains, suggesting that perturbation of MreB-dependent cell physiology and morphology can also sensitize bacteria to TOB-PNA treatment. The synergistic interactions observed between TOB-PNA and the potentiating compounds were specific to the conjugate, as interactions TOB alone and the same potentiators were mainly indifferent or additive (Table S7). Collectively, these results demonstrate that TOB-PNA synergy is not limited to outer membrane permeabilization but can also be achieved through modulation of bacterial metabolism, lipid biosynthesis, and cell physiology.

Figure 9.

Fractional inhibitory concentration indices (FICIs) of TOB-Ahx-PNA 7 in combination with five mechanistically distinct compounds. Dotted lines indicate the thresholds for synergy (FICI ≤ 0.5) and additivity (0.5 < FICI ≤ 1.0). PMB = polymyxin B; PMBN = polymyxin B nonapeptide; TCS = triclosan; TMP = trimethoprim; A22 = S-(3,4-dichlorobenzyl) isothiourea. FICIs for S. flexneri 0421, K. pneumoniae 0120, and K. pneumoniae 0851 are provided in Table S6.

Conclusion

Aminoglycoside-nucleic acid conjugates represent an emerging but still underexplored class of antisense antibacterial agents. In particular, their application in Gram-negative bacteria is constrained by limited intracellular delivery across bacterial membranes. Here, we designed and synthesized four TOB-PNA conjugates targeting the essential acpP transcript to evaluate the generalizability of the aminoglycoside-PNA framework and to investigate how antibacterial activity can be enhanced through co-treatment with bioactive compounds targeting distinct cellular pathways.

TOB conjugation enhanced PNA●RNA duplex stability and improved translation inhibition and antibacterial activity, consistent with prior observations for neomycin- and amikacin-PNA conjugtes.²¹ Collectively, these findings support aminoglycosides as a promising and complementary class of PNA delivery carriers, in which the cationic and membrane-interacting properties contribute to both cellular uptake and partial intrinsic antibacterial activity. Importantly, aminoglycoside conjugation provides a modular chemical handle for tuning nucleic acid binding and bacterial envelope interactions, underscoring its value as a generalizable delivery strategy for antisense agents in Gram-negative pathogens.

Comparison of the four TOB-PNA conjugates across thermal denaturation, in vitro translation, and antibacterial assays suggests that linker identity has a modest but context-dependent effect on overall performance. While linker effects were modest within individual assays, TOB-Ahx-PNA 7 emerged as the most consistently performing conjugate across the full dataset. This may reflect an optimal balance of spacer length, flexibility, and hydrophobic character that reduces steric constraints between TOB and PNA while maintaining favorable interactions with bacterial membranes. Given that only a limited set of linker chemistries was evaluated, further systematic exploration will be required to fully define structure-activity relationships governing TOB-PNA performance.

Although outer membrane permeability remains a limitation for aminoglycoside-PNA activity, co-treatment with sub-MIC PMB enhanced antibacterial activity and membrane permeabilization of TOB-Ahx-PNA 7 to levels comparable to those observed for a CPP-delivered PNA. This synergy was broadly conserved across Enterobacterales, including drug-resistant isolates, and extended to clinically relevant biofilms. Importantly, TOB-Ahx-PNA 7, alone and in combination with PMB, also exhibited a favorable safety window in cytotoxicity and hemolysis assays. The complementary properties of aminoglycosides and CPPs could be leveraged in hybrid delivery systems as a strategy to further improve intracellular delivery and antibacterial efficacy of PNAs. Our previous findings that peptide-aminoglycoside conjugates synergize with PMB and exhibit enhanced antibacterial activity against drug-resistant bacteria also provide optimization options for such future works.^{26, 55}

Beyond PMB, TOB-Ahx-PNA 7 showed synergistic or additive interactions with mechanistically distinct compounds, including inhibitors of fatty acid biosynthesis, folate metabolism, and cell morphology. The ability of underexplored potentiators such as A22 to enhance TOB-PNA activity suggests that synergy may extend beyond membrane permeabilization to broader, physiology-dependent effects. These findings highlight the need for future mechanistic studies to define how different combination partners could modulate antisense delivery to enhance the efficacies of antibacterial antisense conjugates.

Material and Methods

Chemicals and bacterial strains

Biological reagents, antibiotics and oligomers.

Mueller Hinton broth without cations, 2-[4-(2-Hydroxyethyl)-1-piperazinyl]-ethanesulfonic acid (HEPES), fluorescent dyes FM4–64, 4′,6-diamidino-2-phenylindole (DAPI), SYTOX Green and propidium iodide, antibiotics tobramycin, polymyxin B sulfate, polymyxin B nonapeptide, triclosan, trimethoprim and 3,4-dichlorobenzyl-isothiourea (A22) were purchased from Fisher Scientific. Oligomers with different backbone modifications (except for PNA synthesized in our laboratory) were designed and purchased as HPLC-purified solids from ChemGenes Corporation (Wilmington, MA). Figures S17–S27 are the HPLC and MS profiles for the oligomers with other backbone modification. PNA and oligomers with other backbone chemistries were prepared as 1 mM stock solutions in deionized sterilized water and heated to 65°C for 15 minutes prior to use.

Bacterial isolates.

Escherichia coli strains 25922 and MG1655 were purchased from the American Type Culture Collection (ATCC, Manassas VA) as glycerol stocks, E. coli strain SM101 was purchased as a freeze-dried solid from the E. coli Genetics Resource Center’s Coli Genetic Stock Center (ECGRC Cheshire, CT), Salmonella enterica serovar Senftenberg strain 0127, Shigella flexneri 0421, and Klebsiella pneumoniae strains 0120 and 0851 were received as glycerol stocks free of charge from the Center for Disease Control and Prevention and the Food and Drug Administration Antimicrobial Resistance Isolate Bank (ARIB, Atlanta, GA). Original bacterial stock solutions were stored at −80°C. Working stocks were prepared from the frozen stock by transferring an inoculation loop of frozen stock to a Luria Bertani agar (LBA) plate and streaking for colony isolation and purity. Bacteria were routinely maintained on LBA plates for up to 2 weeks. Additional glycerol stocks were made from the LBA plates. The antibiotic profile for each strain was routinely checked to check for changes in the antibiotic susceptibility phenotype. For the CDC isolates, the antibiotic profile was compared to that reported on the ARIsolateBank website.

Thermal denaturation of PNA●acpP RNA duplexes

The ten base RNAs were purchased as desalted solids from Integrated DNA Technologies (Portland, OR) The RNAs had the base sequences as follows: 5’- AGAGUAUGAG −3’ which is the exact match to the translation initiation region and a portion of the coding region for acpP. RNA was dissolved in diethyl pyrocarbonate (DEPC) treated sterilized deionized water to a concentration of 500 μM and were stored at −20°C until use. RNA and PNA were mixed 1:1 for a final concentration of 2.5 μM in 10 mM HEPES buffer (pH 7.0), 100 mM NaCl and 0.4 mM EDTA. The RNA and PNA were heated to 95°C for 5 min and slowly cooled to room temperature then incubated at 4°C overnight to ensure proper annealing and base pair alignment. The temperature program for thermal denaturation experiments was as follows: Start at 15°C with a hold for 5 minutes followed by an increase to 95°C at 0.5°C per min, followed by a decrease in temperature to 15°C at 1°C per min. The denatuation temperature was defined as the midpoint of the transition in the melting curve. A Cary UV-VIS spectrophotmeter equipped with a temperature controller (Agilent Technologies, Santa Clara CA) was used for collection of thermal melting data. Data were fitted to a sigmoidal curve using the Boltmann equation with ORIGIN PRO software (OriginLabs Coorporation, Northampton MA).

Screening of PNA conjugates for antibacterial activity

MIC and Checkerboard Assays.

An overnight bacterial culture was diluted 1:100 in Meuller Hinton broth (MHB) without cations and grown on a shaking incubator at 37°C to achieve exponential phase cells ~2 – 3 h. The culture was diluted 1:1000 in MHB to for a final cell concentration of ~ 5 X 10⁵ CFU/mL. After dilution of the cell culture, 90 μL of the bacterial suspension was aliquoted into the wells of 96-well non-binding plate (REF# 655901, Greiner Bio-One, Fisher Scientific) the wells of which contained to 10 μl of 10X solutions of PNA or TOB-PNA for a final concentration range of 1 – 32 μM. One row of wells served as the untreated growth control which contained 10 μL of water and 90 μL cell suspension. After a 16 h incubation at 37°C in a humidified chamber, the OD600 was determined using a SPARK plate reader (TECAN, Morrisville NC). The MIC was determined as the lowest concentration of PNA which inhibited growth in the wells (OD600 < 0.07, ≥ 90 % growth inhibition relative to the untreated growth control). Two way dilution checkerboard plates were prepared as follows: 20X solutions (640 μM) of each test drug (PMB, PMBN, TCS, TMP, A22) were prepared in water and for each test drug to be combined with TOB-PNA, the test drug was added to the top row of a 96 well-plate and serially diluted 2-fold down into an equivalent volume of water and then a 20X solution of TOB-PNA was added in the first column of wells in an equivalent volume and serially diluted 2-fold across into the equivalent volume of drug solution. After the checkerboard plates were prepared, 10 μL from each well was dispensed into their corresponding well positions into non-binding 96 well plates using a multichannel pipettor by an automated liquid dispenser (epMotion 5075, Eppendorf, Enfield CT). The 90 μL of bacterial suspension was then added to each well and plates were incubated and growth measured as given above.

The MIC where ≥ 90% growth inhibition was observed was defined as follows:

$$\%\mathit{\text{Growth}}\mathit{\text{Inhibition}}=100-100\times \frac{A_{\mathit{\text{drug}}}-A_{\mathit{\text{background}}}}{A_{\mathit{\text{control}}}-A_{\mathit{\text{background}}}}$$

The MIC and the MIC in combination where ≥ 90% growth inhibition was observed was used to determine the fractional inhibitory concentration index (FICI) which was defined as follows:

$$\text{FICI}=\frac{\mathit{\text{FIC}}\mathit{\text{DrugA}}\mathit{\text{combination}}}{\mathit{\text{MIC}}\mathit{\text{DrugA}}\mathit{\text{alone}}}+\frac{\mathit{\text{FIC}}\mathit{\text{DrugB}}\mathit{\text{combination}}}{\mathit{\text{MIC}}\mathit{\text{DrugB}}\mathit{\text{alone}}}$$

where the FIC is the fractional inhibitory concentration in combination and MIC is the minimal inhibitory concentration.

Biofilm reduction assay using MTT for viability.

Bacterial biofilms were pre-formed onto 96 well flat bottom high-binding polystyrene microtiter plates as follows: exponential phase cell density was adjusted in MHII broth to an OD₆₀₀ of 0.06 and 100 μL was inoculated to each well and then the plate was then incubated for 4 h at 37 °C to allow attachment of cells, growth and biofilm formation to achieve an OD600 of ~ 0.1. After incubation, the biofilm was treated with 100 μL of 20X test solutions containing MHII (growth control), 8, 16 or 32 μM TOB-PNA, 2, 4, or 8 μM PMB or for the combinations: 0.5 μM PMB + 4 μM TOB-PNA, 1 μM PMB + 8 μM TOB-PNA or 2 μM PMB + 16 μM PMB. Biofilms were incubated for 16 h at 37°C. After incubation, the media in each well was gently aspirated and washed gently with 100 μL / well of phosphate buffered saline (pH 7) followed by recording the OD₆₀₀. After the OD₆₀₀ measurement, 20 μL of a 5 mg/mL MTT solution in PBS (pH 7.0) was added to each well and incubated at 37°C for 1 h. After incubation, PBS containing the MTT was removed and replaced with 100 μL DMSO and incubated at 37°C for 15 min. to lyse the cells and solubilize the formazan from reduction of the MTT. Formazan was quantified by spectroscopy by measurement at OD₅₇₀ using a microplate reader (SPARK, Tecan).

In vitro cell growth inhibition of human embryonic kidney cell line HEK 293T

PMB, TOB, TOB-PNA and combinations were prepared by serial dilution 0.250 – 512 μM in Dulbecco’s Modified Eagle’s Medium (DMEM) amended with 20% FBS. The human embryonic kidney cell line HEK293T was prepared for drug testing by seeding 100 μL culture to 25,000 cells per well in a 96 well plate and returned to culture conditions overnight in a humidified chamber at 37°C and 5% CO₂ to allow for recovery and attachment to the surface of the well bottoms. The next day, the culture medium was removed and replaced with 100 μL each test compound diluted to different concentrations in DMEM without Pen/Strep. Each concentration of compound was tested in triplicate. After 24 h exposure, the media was gently removed by aspiration and treated again with test compounds for an additional 24 h. Compounds were tested alone with concentrations up to 512 μM and in combination up to 256 μM. Cells were incubated with treatments according to cell culture conditions for a total of 48 hours. After 48 hours, 20 μL of a 1 mg mL⁻¹ cell proliferation reagent MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) was added to each well. After 4 hours of incubation at 37°C and 5% CO₂, the media was gently aspirated from the wells and 100 uL of DMSO was added to each well, and the plate was incubated for 15 minute at 37°C to facilitate cell lysis and solubilization of formazan crystals. To determine cell inhibition, the absorbance of formazan was read at OD570 nm using a TECAN SPARK microplate reader. Triplicate readings were averaged for each drug/concentration and control, the average absorbance values for the media only background was subtracted from each drug/concentration average (corrected absorbance). Wells that contained cells in media, but without treatment served as the growth/viability control. Percent cell viability was calculated as follows:

$$\%\text{Cell}\text{viability}:\frac{\left(\mathit{\text{Control}}-\mathit{\text{Sample}}\right)}{\mathit{\text{Control}}}\times 100$$

Hemolysis assay

The hemolysis assay was performed using sheep red blood cells (10% packed, Fisher Scientific). Briefly, the 10% RBCs were diluted to a 2% solution in phosphate buffered saline (pH 7.4) and 100 μL was dispensed to the wells of a sterile 96 well round bottomed polypropylene plate. Test solutions at 2X the concentration were then dispensed to the wells with RBC suspension and mixed gently using a multichannel pipettor by pipetting up and down 2 times with care to avoid bubbles. The final concentration range of test compounds was 0.25 – 256 μM and 10% Triton X-100 served as the positive control (100% RBC lysis) and PBS served as the negative control. Plates were covered with plate sealer tape and incubated at 37oC for 60 min in a humidified chamber. After incubation, plates were centrifuged for 5 min at 1700 × g. After pelleting the RBCs, 100 μL supernatant was gently removed from the RBC pellets with care not to disturb pellet and introduce bubbles and transferred to a transparent 96 well flat bottomed plate. The absorbance at 415 nm was measured using the TECAN SPARK plate reader. The data were normalized to the Triton X-100 control and plotted as the percent hemolysis as a function of the LOG of the drug concentration. Sigmoidal fits to the data were performed using ORIGIN PRO software. Assays were performed in triplicate on two separate occasions and the results were averaged.

Outer Membrane NPN Uptake Assay

Outer membrane permeabilization was assessed using the N-phenyl-1-naphthylamine (NPN) uptake assay. E. coli MG1655 was cultured at 37 °C to mid-log phase in Luria-Bertani broth (LB) (OD₆₀₀ ≈ 0.4), harvested by centrifugation (4500 × g, 5 min), washed twice with 5 mM HEPES buffer (pH 7.0), and resuspended in the same buffer. Cells were then diluted in HEPES buffer to an OD₆₀₀ of 0.7 – 0.8 McFarland standard using a densimeter which corresponded to a plate reader OD600 of 0.07–0.08. Assays were performed in black-walled, clear-bottom 96-well plates in a final volume of 200 μL cell suspension per well. NPN was prepared as a 10 mM stock in ethanol and diluted to 200 μM in HEPES buffer (final concentration 10 μM). MgCl₂ was prepared at 60 mM in buffer (final concentration 3 mM), and test compounds were prepared at 200 μM in buffer (final concentration 10 μM unless otherwise stated; combination treatments included sub-MIC polymyxin B at 0.25 μM). Selected wells received 10 μL of MgCl₂ solution or buffer control, followed by 10 μL of NPN solution, and 170 μL of bacterial suspension. The test compound 10X solutions were added last to initiate the assay (10 μL). Control wells containing NPN with cells (±MgCl₂) were included. Fluorescence was measured using excitation/emission wavelengths of 350/420 nm immediately after compound addition and every minute for up to 4 min. The 2 min time point was used for data analysis. Fluorescence was normalized using (F_x-F₀)/(F_PMB-F₀), where F_x is the signal of the test compound, F₀ is the signal of the untreated control without MgCl₂, and F_PMB is the signal of polymyxin B in the absence of MgCl₂. The assay was performed in two or three independent biological replicates, each with three technical replicates.

Time-Kill Assay

Bacterial strains in exponential growth were diluted to ~5×10⁵ CFU/mL to 10 mL of non-cation adjusted Mueller-Hinton broth (MHB) containing TOB-PNA, PMB and 0.5x, 1x and 2x the MIC of the combination PNA and PMB at 37 ◦C for up to 24 h. Bacterial cultures were harvested at 0, 1, 2, 3, 4, 6, 8 and 24 h. At each time point, 100 uL aliquots from each test tube was removed and transferred to the first column of wells of a sterile 96 well plate in which all other wells contained 270 uL sterile phosphate buffered saline. Tenfold serial dilutions were performed by transfer of 30 uL from the preceding well to the subsequent well using a multichannel pipettor. After the serial dilution, 10 μL aliquots were spot plated onto square MHB agar plates and incubated for 16 h at 37°C to determine the colony forming units per milliliter. Experiments were performed in triplicate and spot plates were performed in duplicate from each serial dilution well.

Cell-free translation assay

TOB-PNAs and oligomers with various backbone modifications were assessed for their effect on acpP gene expression via cell-free coupled transcription – translation assays. The acpP gene target region was positioned within a green fluorescent protein (Egfp) reporter plasmid so that it is under control of an IPTG-inducible promoter (T7) upstream of the Egfp reporter gene. Egfp gene expression was assayed in 96 well microtiter plates and fluorescence imparted by Egfp was measured in each well using a plate reader. Treatment results were normalized to the positive control without oligomer treatment. The reporter plasmid pET-(acpP target site/Egfp) was designed and purchased from VectorBuilder (Chicago, IL). The AcpP/Egfp fusion product was expressed by in vitro transcription/translation using the S30 T7 High-Yield Protein Expression System L1110 (Promega). In vitro translation reactions were performed for 1 – 2 h at 37°C according to manufacturer’s recommendations, using 5 μM PNA or TOB and 1 ug reporter plasmid per 25 μL reaction. The expression of the AcpP-Egfp fusion product was induced with 1 mM IPTG. Controls included mismatch PNA and assays without IPTG induction. Reactions were stopped by chilling on ice for 10 min. Fluorescence was measured using a TECAN INFINITE M1000 PRO multi-well plate reader with 488 nm excitation and 510 nm emission. The Egfp expression inhibition percentage for an oligo was calculated as: inhibition % = (1 – fluorescence with oligo / fluorescence without oligo) × 100%. All reactions were performed in 96-well microtiter plates in triplicate.

Fluorescence microscopy

After treatment of bacteria with TOB-PNA with or without PMB, untreated controls and positive control TRITON X100 the bacterial cells in the wells representing 1 or 2 units below the MIC or at MIC were stained with 10 μL of 10X solutions of the cell membrane stain FM4–64 cell membrane stain, DAPI nuclear material stain and SYTOX Green cell membrane damage indicator dissolved in 100 mM Tris base for a final concentration of each dye at 1 μg/ml FM4–64 and 2 μg/ml DAPI, and incubated at room temperature for 15 min in the dark. After incubation with the dyes the contents of the well were transferred to a microfuge tube and centrifuged for 3 minutes at 7500 RPM to pellet cells. Cells were resuspended in TRIS buffer to achieve a cell density of ~10⁹ cells/mL, vortexed to remove clumps and 5 μL was spotted onto a thin 1% agarose pad containing 20% Luria Bertani broth that was laid onto a microscope slide and a coverslip overlaid the agar pad. Live cells were imaged with 1000X magnification using a Leica DMi8 Thunder Imager fluorescence microscope (Leica Microsystems, Boston, MA). To investigation the effect of TOB-PNA on cell membrane permeability, ~5 × 10⁸ exponential phase E. coli MG1655 cells in 20% MHB were treated with PNA, TOB-PNA with and without PMB for 30 min. Propidium iodide indicator of membrane damage (45 μM) and DAPI (2 μg/mL) were added during the treatment. After treatment bacterial cells were treated as mentioned before and imaged accordingly.

Resistance development assay

A 14-day resistance development assay was performed using one tobramycin-susceptible isolate E. coli MG1655 and tobramycin-resistant K. pneumoniae 0120. To prepare for the drug challenge, bacteria were grown overnight in MHII non-cation-adjusted (MHII) broth. After overnight incubation, the cell densities were adjusted to 5 × 10⁵ CFU mL⁻¹, and 90 μL was dispensed into the wells of a 96-well non-binding microtiter plate containing 10X solutions of the drug or a combination over the dilution range above and below the MICs and incubated at 37°C for 24 h. The cell suspension was quickly mixed with the drug solution using a multichannel pipettor to avoid the localized effects of the drugs. For subsequent challenge days, the wells at the subinhibitory MIC were sampled and diluted to achieve an OD600 reading of 0.08 and then diluted again by mixing 2 μL of the cell suspension with 5 mL MHII broth. After adjusting for cell density, 90 μL of the cell suspension was dispensed into drug challenge plates. Growth was monitored by measuring absorbance at OD600 over a period of 14 days. Purity checks were performed on day 1, day 7, day 14 and after the 72 hour recovery period without drug exposure after the 14^th day to identify whether cross-contamination occurred by spread plating from wells with the most diluted drug exposure. When resistance development was observed, the stability of resistance was determined by taking the day 14 contents at the MIC and suspending them in 2 mL MHII, taking the contents at the sub-MIC, suspending them in 10 mL MHII, and incubating for 3 consecutive 24 h periods. After incubation, 5 × 10⁵ CFU mL⁻¹ was dispensed onto the drug challenge plate and incubated for 24 h.

Supplementary Material

The structures of the oligomers, characterization of compound purity and identification verification as determined by reverse phase chromatography and mass spectral analysis are given for the oligomers along with their chemical stuctures.

Figure 8.

Fluorescence micrographs showing membrane permeation of E. coli MG1655 assessed by propidium iodide (PI) uptake. Exponentially growing cells (1 × 10⁹ cells/mL) were treated for 30 minutes. Panels (left to right): a) no treatment, b) 0.1% Triton X-100 (TX-100, positive control), c) sub-MIC PMB (0.0625 μM), d) 16 μM PNA 1, e) 8 μM PNA 1 + 0.0625 μM PMB, f) 16 μM TOB-PNA 6, and g) 8 μM TOB-PNA 6 + 0.0625 μM PMB. Cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) to visualize DNA and PI to assess membrane permeabilization.