Benzimidazole Isosteres of Salicylanilides Are Highly Active Colistin Adjuvants

Abstract

Multidrug-resistant bacterial infections have become a global threat. We recently disclosed that the known IKK-β inhibitor IMD-0354 and subsequent analogues abrogate colistin resistance in several Gram-negative strains. Herein, we report the activity of a second-generation library of IMD-0354 analogues incorporating a benzimidazole moiety as an amide isostere. We identified several analogues that show increased colistin potentiation activity against Gram-negative bacteria.

Graphical Abstract

The increasing prevalence of antimicrobial resistance among pathogenic bacteria is one of the largest threats to human health.¹ Among all bacterial pathogens, the ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) are responsible for many of the most difficult to treat infections as a result of high levels of multidrug resistance.² Given the absence of new classes of antibiotics that are active against Gram-negative bacteria approved in recent decades, clinicians have turned to the use of polymyxins, which are typically viewed as the antibiotics of last resort for the treatment of multidrug-resistant (MDR) Gram-negative infections.^3,4 Polymyxins exert their antibiotic effect through interaction with the outer membrane of Gram-negative bacteria and are believed to displace divalent cations from the lipopolysaccharide (LPS) component of the membrane, leading to a permeabilized outer membrane. This permeabilization allows polymyxin access to the inner bacterial membrane, where the polymyxin causes disruption of the inner membrane, leading to cell death.^5,6 The use of polymyxins in clinical settings, however, is hampered by toxicity and increasing resistance.⁷ Polymyxin resistance is mediated by the PmrAB and/or PhoPQ two-component systems in several bacteria,⁸ resulting in modification of lipid A with cationic residues such as phosphoethanolamine or aminoarabinose.^8,9 Such modifications reduce the overall negative charge of the LPS and temper the affinity for cationic polymyxin antibiotics. Further compounding this predicament is the emergence of the plasmid-borne mobile colistin resistance genes (mcr-1-10),¹⁰ which has facilitated more rapid dissemination of colistin resistance into the general pathogen pool.

Orthogonal to the development of new conventional antibiotics is the use of non-microbicidal small molecule adjuvants that disrupt bacterial resistance mechanisms.¹¹ Mechanisms of action of such adjuvants include inhibition of antibiotic modification or target modification, inhibition of efflux or enhancement of antibiotic uptake, inhibition of biofilm formation, or inhibition of signaling pathways that mediate antibiotic resistance.¹¹ In the context of colistin resistance, we and others have identified a number of adjuvants that suppress resistance in bacterial strains that harbor either chromosomally or plasmid encoded resistance determinants. These include the 2-aminoimidzole 1,¹² the tryptamine derivative 2,¹³ the meridianin D analogue 3,¹⁴ salicylanilides IMD-0354 4,¹⁵ niclosamide 5,¹⁶ the natural products kuwanon G 6,¹⁷ prodigiosin 7,¹⁸ the antifungal drug econazole 8¹⁹ (Figure 1), and others.^20-22

Figure 1.

Previously reported colistin adjuvants.

IMD-0354 4 is an inhibitor of the NF-Kβ inhibitor kinase (IKKβ),²³ and we recently reported the discovery that it suppresses colistin resistance against several highly resistant strains of Gram-negative bacteria while showing no standalone antibacterial activity.¹⁵ In a subsequent report, we disclosed the structure activity relationship of a first-generation library based upon the IMD-0354 scaffold, from which we identified several analogues with equal or improved activity.²⁴ To further explore the structure activity relationship of this scaffold, we designed an IMD-0354 analogue in which the phenyl amide moiety was replaced by a benzimidazole ring, compound 9 (Figure 2). Herein, we report the synthesis and activity of 21 benzimidazoles based upon IMD-0354 and our first-generation analogues.²⁴ Of these analogues, 17 exhibit equal or increased colistin potentiation compared to IMD-0354. At 5 μM, the most active compounds reduce the colistin MIC from 2048 to 0.25 μg/mL (8192-fold) and from 512 to 0.125 μg/mL (4096-fold) against the highly colistin-resistant strains A. baumannii 4106 (AB4106) and K. pneumoniae B9 (KPB9), respectively.

Figure 2.

Design of benzimidazole-based IMD-0354 analogue 9.

RESULTS AND DISCUSSION

Synthesis and Evaluation of Benzimidazole Library.

We initiated this study by synthesizing the IMD-0354 benzimidazole analogue 9 via a condensation reaction between 3,5-bis (trifluoromethyl)benzene-1,2-diamine and 5-chloro-2-hydroxybenzaldehyde in the presence of sodium metabisulfite in DMF (Scheme 1). We then assessed the activity of compound 9 in the context of breaking colistin resistance in AB4106 and KPB9 and compared this activity to IMD-0354 (Table 1). The MICs of colistin alone against these two strains are 2048 μg/mL (AB4106) and 512 μg/mL (KPB9). IMD-0354 4, when dosed at 5 μM, reduces the colistin MIC against AB4106 to 2 μg/mL (1024-fold) and against KPB9 to 0.5 μg/ mL (1024-fold). Compound 9 reduces the colistin MICs to 0.5 and 0.25 μg/mL against AB4106 and KPB9, respectively, when dosed at 5 μM (lower than the susceptibility/intermediate breakpoint of colistin of 2 μg/mL for these species).²⁵ Interestingly, compound 9 exhibits reduced standalone toxicity compared to IMD-0354 4, with standalone MICs of greater than 200 μM against both strains, compared to MICs of 50 μM for IMD-0354 against both strains (Table S1).

Scheme 1.

Synthesis of Benzimidazole 9

Table 1.

Colistin MICs in the Presence of Compounds 4 and 9

compound	KPB9 MIC [μg/mL] (fold reduction)	AB4106 MIC [μg/mL] (fold reduction)
	512	2048
IMD-0354 4 5 μM (1.9 μg/mL)	0.5 (1024)	2 (1024)
9 5 μM (1.9 μg/mL)	0.25 (2048)	0.5 (4096)

In light of the activity of compound 9, we designed an additional 20 benzimidazole-containing analogues possessing varied substitution patterns on both the phenyl and benzimidazole rings (Figure 3). All compounds were synthesized via the same procedure used to synthesize compound 9 from the appropriately substituted 1,2-phenyl-endiamines and benzaldehydes, which, with the exception of 3,5-dibromo-1,2-benzenediamine, were commercially available. 3,5-Dibromo-1,2-benzenediamine was accessed from 2,4-dibromo-6-nitroaniline via a tin chloride-mediated reduction.

Figure 3.

Structures of benzimidazole derivatives 10–29. Yield in parentheses.

The library was first screened to quantify standalone toxicity (Table S1) and then assayed for colistin potentiation activity against AB4106 and KPB9 at a concentration of 5 μM (Table 2). All the derivatives exhibited standalone MICs of greater than 200 μM, with the exception of compounds 13, 18, 27, and 28, which exhibited MICs of 25 μM against both strains (Table S1).

Table 2.

Colistin MICs against KPB9 and AB4106 in the Presence of Benzimidazole Analogues

compounda	KPB9 MIC [μg/mL] (fold reduction corresponding to compound 4)	AB4106 MIC [μg/mL] (fold reduction corresponding to compound 4)
	512	2048
10 (1.8 μg/mL)	0.5 (1)	2 (1)
11 (2.1 μg/mL)	0.25 (2)	0.5 (4)
12 (2.4 μg/mL)	0.25 (2)	1 (2)
13 (2.1 μg/mL)	0.125 (4)	0.25 (8)
14 (1.4 μg/mL)	16	>16
15 (1.6 μg/mL)	0.25 (2)	2 (1)
16 (2.0 μg/mL)	0.25 (2)	2 (1)
17 (1.6 μg/mL)	>16b	>16
18 (1.6 μg/mL)	0.25 (2)	2 (1)
19 (1.7 μg/mL)	0.5 (1)	0.5 (4)
20 (1.5 μg/mL)	0.125 (4)	2 (1)
21 (1.8 μg/mL)	0.5 (1)	2 (1)
22 (2.0 μg/mL)	0.5 (1)	0.5 (4)
23 (2.5 μg/mL)	0.25 (2)	4 (0.5)
24 (1.9 μg/mL)	>16	>16
25 (1.6 μg/mL)	0.5 (1)	1 (2)
26 (1.8 μg/mL)	0.5 (1)	1 (2)
27 (1.7 μg/mL)	0.125 (4)	1 (2)
28 (2.1 μg/mL)	0.25 (2)	0.5 (4)
29 (1.8 μg/mL)	>16	>16

^aAll compounds screened at 5 μM.

^bHighest concentration tested.

We first focused on evaluating the impact of substituents on the phenol moiety of compound 9 and observed that the most potent compound of this series was compound 13, which contained a 5-trifluoromethylphenol moiety in place of the 4-chlorophenol moiety of compound 9, which lowered the colistin MIC against KPB9 and AB4106 to 0.125 and 0.25 μg/mL, respectively (both two-fold lower than compound 9). Substitution of the benzimidazole ring with various halo-genated and trifluoromethyl substituents did not result in the identification of any compounds that improved upon the activity of compound 9, though compound 19, which possessed a 6-fluoro-5-trifluoromethyl benzimidazole, did exhibit equipotent activity to compound 9 against AB4106. Combining the derivatizations made to each region of the molecule resulted in the identification of 4,6-dichlorobenzimidazole analogues 20–22, which exhibited comparable activity to compound 9 against KPB9 (and for compound 22 against AB4106). Exchanging the chloro-substituents of compound 22 for other halogens resulted in reduced or abrogated activity, while replacing 5-iodophenol with 4-chlorophenol or 4-bromophenol resulted in comparable activity. The most potent activity for this series against KPB9 was found upon installing a 4-trifluoromethylphenol moiety in compound 27, which lowered the colistin MIC to 0.125 μg/mL, equivalent to that of compound 13. Combining 4-trifluoromethylphenol with a 4,6-dibromobenzimidazole in compound 28 resulted in AB4106 activity equivalent to compound 9 (colistin MIC of 0.5 μg/mL). Finally, we confirmed the necessity of the hydroxyl moiety by preparing the methoxy derivative 29 and observed that methylation of the hydroxyl moiety results in a complete loss of activity against both AB4106 and KPB9 strains. This observation is consistent with that observed for methylation of the IMD-0354 hydroxyl.²⁴

Dose–response activity of lead compounds was then determined down to a concentration of 1 μM (Figure 4 and Table S2 and S3). At 1 μM, nine members of this benzimidazole library exhibited comparable activity to IMD-0354 4 against KPB9, with compounds 13, 16, and 25–27 exhibiting two-fold increased activity compared to IMD-0354. Additionally, compounds 9, 13, 16, and 25–28 reduced the colistin MIC to or below the breakpoint at 3 μM. At the same concentration, against AB4106, 14 of the compounds exhibited comparable or increased activity compared to IMD-0354. Compound 13 effected a four-fold lower colistin MIC at 3 μM and two-fold lower at 1 μM, while compounds 25 and 26 effected a four-fold lower colistin MIC at 1 μM, meaning that these two compounds maintain the colistin MIC below the breakpoint at 1 μM. We also noted that compound 13 had a moderate MIC (50 μM), as opposed to a majority of the benzimidazoles that returned MICs of >200 μM. Based on this, a checkerboard assay was conducted using compound 13 and colistin against AB4106. An FICI of 0.033 was determined, indicating that this combination was synergistic (Figure S4).

Figure 4.

(A) Dose–response for lead compounds against KPB9. (B) Dose–response for lead compounds against AB4106.

Time–kill curves were then constructed for compounds 13 and 27 in the presence and absence of colistin against both AB4106 and KPB9 (Figure 5). From these, we confirmed that at 5 μM, neither compound alone exhibits any effect on bacterial growth, and we observed a considerable reduction in CFUs at early time points for the combination of either compound and colistin.

Figure 5.

Time–kill curves for compounds 13 and 27. (A) Compound 13 against AB4106. (B) Compound 27 against AB4106. (C) Compound 13 against KPB9. (D) Compound 27 against KPB9.

We then determined the activity of compounds 13 and 27 against a panel of additional colistin-resistant strains of K. pneumoniae and A. baumannii, including strains harboring the mcr-1 gene, and also determined activity against colistin-resistant strains of P. aeruginosa. Select strains are shown in Table 3, and full data can be found in Tables S4-S8. The compounds retained activity against all strains tested, including mcr-1-harboring strains, and also displayed potent colistin potentiation against two highly resistant P. aeruginosa strains, TRPA161 and TRPA162, with compound 13 exhibiting notably high activity, with fold reductions in the colistin MIC of 4096-fold, bringing the colistin MIC well below the breakpoint.

Table 3.

Potentiation of Colistin by Compounds 13 and 27 against Select Members of a Panel of Colistin-Resistant Gram-Negative Strains^a

	colistin MIC (μg/mL) (fold reduction)
strain		+5 μM 13 (2.1 μg/mL)	+5 μM 27 (1.7 μg/mL)
KP C3	128	0.25 (512)	0.25 (512)
KP F3	128	1 (128)	2 (64)
KP F9	256	0.5 (512)	1 (256)
KP E2	512	4 (128)	2 (256)
KP I2	512	2 (256)	2 (256)
KP H2	256	2 (128)	1 (256)
KP F2210291mcr-1	16	0.25 (64)	0.5 (32)
AB 3941	1024	0.5 (2048)	0.5 (2048)
AB 4112	2048	0.5 (4096)	1 (2048)
AB 1E4	>512	0.5 (>1024)	1 (>512)
AB 2A8	>512	0.125 (>4096)	0.125 (>4096)
AB 1G2	>512	>16	0.25 (>2048)
AB 1F9	>512	0.5 (>1024)	0.25 (>2048)
AB 17978mcr-1	64	0.5 (128)	0.5 (128)
TRPA 161	1024	0.25 (4096)	2 (512)
TRPA 162	1024	0.25 (4096)	2 (512)

^aNote: KP: K. pneumoniae; AB: A. baumannii; and TRPA: P. aeruginosa.

To determine the effect of any non-specific aggregation on activity, the assay was then repeated in the presence of 0.001% Triton X-100, a detergent that inhibits aggregation. For 11 of the compounds tested, activity was replicated under these conditions (Table S9), while against AB4106, the activity of compounds 11, 13, and 15 activity was moderately reduced (four-fold) and that of compound 19 was reduced eight-fold. These results indicate that aggregation does not play a large role in the activity of these adjuvants.

Benzimidazoles Do Not Affect Modification of Lipid A.

To investigate the mechanism of action of these benzimidazole adjuvants, we first determined any effect upon lipid A modification by MALDI-TOF analysis of lipid A from cells grown in the presence and absence of compounds 13 and 27. AB4106 exhibited phosphoethanolamine additions associated with colistin resistance, but there were no significant changes upon treatment with a 5 μM compound. We conducted the same analysis for KPB9 and TRPA161 cells treated with compounds 13 and 27 and again observed no significant changes (Figure S1). These results were further verified via RT-qPCR evaluating changes in expression of pmrCAB when treated with 5 μM compound 27 (Figure S2). With AB4106, we observed slight, though statistically insignificant, decreases in expression of pmrC and pmrA (0.65- and 0.71-fold, respectively). The 0.33-fold reduction in expression of pmrB, however, was statistically significant. These results together indicate that it is likely that the compound is not breaking colistin resistance by reversing modification of lipid A.

Combination of Benzimidazoles and Colistin Shows a Significant Increase in Reactive Oxygen Species.

Finally, given that it has been reported that polymyxin antibiotics induce cell death by generating reactive oxygen species (ROS),^26,27 we examined the effect of lead benzimidazole adjuvants on ROS generation using 2′,7′-dichlorofluorescin diacetate (DCFH-DA) as previously reported²⁸ (Figure 6). Compared to untreated cells, ROS levels in both AB4106 and KPB9 increased upon treatment with active compounds, by between 78% and 934%, with the exception of compound 20, for which ROS generation was decreased to 64% for AB4106 and 51% for KPB9. Upon treatment with the combination of an active adjuvant and colistin (at 5 μM and MIC), a much more substantial increase in ROS, up to 1788% for AB4106 and 1327% for KPB9, was observed. Inactive compounds 17 and 29 did not effect a substantial increase in ROS levels against KPB9 surprisingly; however, these compounds did cause increases of 491 and 535%, respectively, against AB4106 when co-dosed with colistin. These data suggest that ROS generation may play a role in killing cells by the combination of adjuvants and colistin, although determination of the exact mechanism of action and specific target requires further investigation.

Figure 6.

ROS levels following treatment by select compounds.

Benzimidazoles Do Not Lyse Red Blood Cells at Their Active Adjuvant Concentrations.

In order to gain a preliminary insight into the eukaryotic toxicity of these benzimidazole derivatives, we quantified the degree of hemolysis effected by the lead compounds upon defibrinated sheep blood. Seven active compounds were tested at concentrations ranging from 25 μM to 400 μM. Compounds 11, 12, 19, 20, 27, and 28 exhibited less than 50% lysis compared to 1% Triton-X at 400 μM (highest concentration tested, full data in Table S12), while compound 13 exhibited the greatest hemolytic activity, returning an HD₅₀ of 252 μM (Figure S3A). The lowest concentration of 13 that lowers colistin resistance down to the breakpoint against both AB4106 and KPB9 is 3 μM, giving an (HD₅₀)/(colistin potentiation concentration) index of 84, indicating a large therapeutic window for these compounds. We also determined the hemolytic activity of the combination of active compounds with 2 μg/mL colistin and observed only minor changes (Table S12). Again, compound 13 was the only compound that effected greater than 50% lysis at concentrations up to 400 μM, returning an HD₅₀ of 260 μM (compared to 252 μM in the absence of colistin) (Figure S3).

Benzimidazoles Are Acting as the Adjuvant Not the Antibiotic.

As colistin is a membrane-permeabilizing agent, we wanted to confirm that these benzimidazoles are acting as the adjuvant of the combination (as opposed to colistin potentiating the antibiotic activity of the benzimidazoles). To do this, we determined the activity of four lead benzimidazoles in the presence of polymyxin B nonapeptide, a polymyxin that retains membrane-permeabilizing activity but does not exhibit antibiotic activity (Table S13).²⁹ PMBN itself shows no toxicity at 10 μg/mL against AB4106; the MICs of compounds 13, 19, 20, and 27 did not change in the presence of up to 10 μg/mL PMBN, indicating that these compounds are acting as the adjuvant.

Benzimidazole Adjuvant Activity Is Specific for Colistin.

To determine whether the adjuvant activity of this series of compounds is specific for colistin, or more a general phenomenon, we measured the effect of compound 13 at its active colistin potentiation concentration (5 μM) on the MICs of three mechanistically dissimilar antibiotics, clarithromycin, penicillin G, and tobramycin against AB4106, and observed no change in MIC for any of the antibiotics (Table S14).

Benzimidazoles Show Reduced Inhibition of NF-κB Compared to IMD-0354.

Finally, we wanted to determine if it is possible to decouple colistin adjuvant activity from eukaryotic kinase activity for the IMD-0354 scaffold. IMD- 0354 4 inhibits the IKK-induced phosphorylation of IkBα, resulting in inhibition of NF-κB activation and induction of apoptosis.²³ HEK293 recombinant cells containing an NF-κB reporter construct were utilized to quantify inhibition of NF-κB by IMD-0354 4 and several of our benzimidazole analogues (Table 4). IMD-0354 4 exhibited an IC₅₀ of 1.073 ± 0.06 μM in this assay, while several benzimidazole analogues exhibited similarly potent inhibition, namely, compounds 9, 11, and 22. Compound 13 exhibited considerably increased inhibition, returning an IC₅₀ of 0.209 ± 0.11 μM, while compounds 15 and 18, in which the 4,6-bistrifluoromethyl substituents on the benzimidazole moiety of 9 are replaced with 4,6-dichloro or 5-trifluoromethyl groups, respectively, exhibited moderately reduced inhibition (IC₅₀ values of 10.2 ± 3.97 μM and 4.82 ± 1.05 μM, respectively). Compound 19, which possesses a 4-fluoro-5-trifluoromethyl benzimidazole moiety, and compound 20, which has a 4,6-dichloro substitution pattern on the benzimidazole and a fluoro substituent in place of the chloro substituent on the phenyl ring, exhibited considerably reduced inhibition of NF-κB, returning IC₅₀ values of 17.3 ± 7.46 μM and 27.6 ± 13.46 μM, respectively. Both 19 and 20 are highly active colistin potentiators (MIC reductions of 1024–4096-fold in the initial screen), indicating that it is possible to fine tune selectivity via modification of this scaffold.

Table 4.

Inhibition of NF-κB Select Compounds

compound	IC50 (μM)
IMD-0354 4	1.073 ± 0.06
9	2.29 ± 1.55
11	1.46 ± 0.33
12	0.776 ± 0.13
13	0.209 ± 0.11
15	10.2 ± 3.97
18	4.82 ± 1.05
19	17.3 ± 7.46
20	27.6 ± 13.46
22	3.55 ± 0.92

CONCLUSIONS

In conclusion, following previous reports describing the colistin potentiation activity of IMD-0354 and related analogues, we report the activity of a second-generation library of analogues in which the phenyl amide moiety of the first-generation analogues was replaced by a benzimidazole ring. Several benzimidazole analogues exhibit improved colistin potentiation activity against K. pneumoniae and A. baumannii compared to the parent compound IMD-0354, with MIC reductions of up to 8192-fold observed at 5 μM, while several compounds retain activity at 1 μM against highly resistant strains, marking them as some of the most potent colistin adjuvants identified to date. Furthermore, we have shown that it is possible to decouple adjuvant activity from eukaryotic kinase inhibition, identifying a compound that reduces the colistin MIC by 4096-fold against K. pneumoniae B9 (four-fold more active than IMD-0354) and exhibits 26-fold decreased potency of NF-κB inhibition as measured using a luciferase reporter strain assay. These compounds do not appear to potentiate colistin by affecting modification of lipid A, but they do effect an increase in ROS generation. Future work will focus on identifying the mechanism of action of this series, determining in vivo activity, and further structural refinement to augment both potentiation activity and selectivity over NF-κB inhibition.

METHODS

General Biological Methods.

Bacteria Strains and Growth Conditions.

Bacterial strains and their minimum inhibitory concentrations with regard to tested antibiotics used in this study can be found in the corresponding table in the Supporting Information. A. baumannii and K. pneumoniae strains were routinely grown on LB Lennox 1.5% (w/v) agar plates from glycerol stocks every 3 weeks. For use in biological assays, these strains were grown in cation-adjusted Mueller Hinton broth (CAMHB, BD). A. baumannii strains 4106, 4112, 4119, 3941, and 3942 were obtained from Walter Reed Army Institute for Research (WRAIR). A. baumannii strain 17978^mcr-1 and K. pneumoniae strains B9, A5, C3, and F2210291^mcr-1 were obtained from Professor Robert Ernst at The University of Maryland, Baltimore. A. baumannii strains 1E4, 1H7, 1A7, 1A3, 1E6, 2A8, 3A4, 2B6, 1G2, and 1F9, K. pneumoniae strains F1, F3, F9, E6, E2, I6, I4, I2, and H2, and P. aeruginosa strains TRPA 161 and TRPA 162 were obtained from Professor Yohei Doi at the University of Pittsburgh. Stock cultures were stored in 25% glycerol and maintained at −80 °C. CAMHB was purchased from BD Diagnostics. Colistin sulfate salt was purchased from Sigma Aldrich (Cat# C4461). All assays were run in duplicate and repeated at least two separate times.

Broth Microdilution Methods for the Determination of Minimum Inhibitory Concentrations.

Bacteria were cultured for 6 h in CAMHB and then subcultured to 5 × 10⁵ CFU/mL in fresh CAMHB. To aliquots (1 mL) were added compounds from stock solutions in DMSO, such that the compound concentration equaled the highest concentration tested. (For colistin, penicillin G, and tobramycin, samples were dosed from stock dissolved in water.) Samples were then dispensed (200 μL) into the first row of a 96-well microtiter plate in which all but the final row of subsequent wells were prefilled with 100 μL of the untreated bacterial subculture. The final row was filled with media to act as a sterility control and blank. Row one wells were mixed 6–7 times, and then, 100 μL was withdrawn and transferred to row two. Row two wells were mixed 6–7 times followed by a 100 μL transfer from row two to row three. This procedure was used to serially dilute the rest of the rows of the microtiter plate, excluding the last prefilled row, which was used to measure growth in the absence of compounds. Plates were then sealed with GLAD Press’n Seal and incubated under stationary conditions at 37 °C. After 16 h, the plates were removed, and MIC values were measured by recording the OD₆₀₀ of each well. MIC values were determined as the minimum concentration required to achieve 90% growth inhibition compared to growth in untreated wells.

Broth Microdilution Methods for Measurement of Colistin Potentiation.

Bacteria were cultured for 6 h in CAMHB and diluted to 5 × 10⁵ CFU/mL in fresh CAMHB. (In triton potentiation experiments, triton X was added to the CAMHB from a 1% stock to make the final concentration of 0.001%.) To aliquots (4 mL) were added compounds from stock solutions in DMSO. One aliquot was not dosed to allow measurement of the colistin MIC in the absence of compounds. A 1 mL aliquot of each sample was dosed with colistin, and from this, 200 μL was dispensed into the first row of a 96-well microtiter plate in which all but the final row of subsequent wells was prefilled with 100 μL of the corresponding compound-dosed bacterial suspension The final row was filled with media to act as a sterility control and blank. Row one wells were mixed 6–7 times, and then, 100 μL was withdrawn and transferred to row two. Row two wells were mixed 6–7 times followed by a 100 μL transfer from row two to row three. This procedure was used to serially dilute the rest of the rows of the microtiter plate, excluding the last prefilled row, which was used to measure growth in the presence of a compound alone. Plates were then sealed with GLAD Press’n Seal and incubated under stationary conditions at 37 °C. After 16 h, the plates were removed, and MIC values were measured by recording the OD₆₀₀ of each well. MIC values were determined as the minimum concentration required to achieve 90% growth inhibition compared to growth in untreated wells.

Hemolysis.

Defibrinated sheep blood (Hemostat Labs: DSB30) (1.5 mL) was placed into a microcentrifuge tube and centrifuged for 10 min at 10,000 rpm. The supernatant was then removed, and the cells were resuspended in 1 mL of phosphate-buffered saline (PBS). The suspension was centrifuged, then the supernatant was removed, and cells were resuspended two additional times. After the final time, suspension was then diluted 10-fold with PBS. Test compound solutions were made in PBS in small culture tubes and then added to aliquots of the 10-fold suspension dilution of sheep blood. PBS was used as a negative control and a zero-hemolysis marker. Colistin (2 μg/mL) was used as a colistin control. Triton X (a 1% sample) was used as a positive control serving as the 100% lysis marker. Samples were then placed in an incubator at 37 °C while being shaken at 200 rpm for 1 h. After 1 h, the samples were transferred to microcentrifuge tubes and centrifuged for 10 min at 10,000 rpm. The resulting supernatant was diluted by a factor of 40 in distilled water. The absorbance of the supernatant was then measured with a UV spectrometer at a 540 nm wavelength.

Time–Kill Curves.

Strains were cultured overnight in CAMHB and subcultured to 5 × 10⁵ CFU/mL in fresh CAMHB. The subculture was then transferred to culture tubes in 4 mL aliquots, which were dosed with an adjuvant, colistin, or an adjuvant plus colistin. One aliquot was not dosed to serve as a control. All samples were then incubated at 37 °C with shaking. At 2, 4, 6, 8, and 24 h time points, 100 μL was taken from each sample and 10-fold diluted in CAMHB up to 7 times. The diluted culture (100 μL) was plated on LB (Lennox) agar and incubated at 37 °C overnight. The total number of bacterial colonies on each plate was determined using a SphereFlash colony counter (NEUTEC Group Inc.)

Luciferase Reporter Assay.

HEK293 Recombinant Cells containing the NF-κB reporter construct (HEK293-NF-κB-Luc (BPS Biosciences, San Diego, CA)) were thawed from frozen stocks into minimum essential media (MEM) (Gibco, Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS) (Gibco), 1 mM non-essential amino acids (NEAA) (Gibco), 1 mM sodium pyruvate, 2 mM GlutaMAX (Gibco), and 1× Pen-Strep. After 48 h of incubation at 37 °C, cells were trypsonized, counted, and plated on a 96-well black-wall clear-bottom tissue culture to a final concentration of 30,000 cells/well and incubated overnight. In fresh media, analogues were dosed from DMSO stocks and serially diluted, adjusting each dilution point to a final concentration of 0.2% DMSO. Media was aspirated from the culture plate, and cells were treated with 100 μL of media spiked with compounds and incubated for an additional 30 min. A total of 11 μL of a 100 ng/μL solution of human TNFα (Millipore-Sigma, St. Louis, MO) was then added to each well and incubated for an additional 6 h after which 110 μL of One-Step Luciferase Assay System (BPS Biosciences, San Diego, CA) was then added to each condition and the plate was incubated with shaking at 37 °C for 15 min. The 96-well plate was then read on a FLUOstar Optima (BMG Labtech Cary, NC) microplate reader. Each compound was assayed in triplicate.

ROS Fluorescent Dye Assay for Evaluating Oxidative Stress in Bacterial Cells.

In the ROS assay, 2′,7′-dichloro-fluorescin diacetate fluorescent dye (DCFH-DA, Calbiochem) was adopted to measure intracellular ROS upon treatment with benzimidazole derivatives, colistin, and combinations. The 20 mM stocks of DCFH-DA were prepared in DMSO and stored, protected from light, at −20 °C in 1000 μL aliquots, and freeze–thaw cycle was avoided. Bacterial strains were incubated in 11 mL of CAMHB at 37 °C and 200 rpm overnight. Cultures were split and diluted 1:2 in fresh media. Cells were either left untreated or dosed with one of the following conditions: untreated, 2 μg/mL colistin, 0.5 μg/mL colistin, 5 μM benzimidazole compound, and 5 μM benzimidazole compound with colistin at their active concentration. Cultures were incubated in the dark at 37 °C and 240 rpm for 4 h. The treated culture (2 mL) was centrifuged (15,000 rcf, 6 min) and washed with 1 mL 1× PBS. The pellets were suspended in 1 mL of 100 μM DCFH-DA in 1× PBS and incubated in the dark at 37 °C for 1.5 h. Cells were pelleted (15,000 rcf, 6 min) and washed with 1× PBS before resuspension in fresh CAMHB. Cells were recovered for 1 h at 37 °C. A cell suspension (100 μL) was aliquoted into black-walled, clear-bottom 96-well plates. Fluorescence was measured on a HTX Synergy Multimode plate reader with excitation at 485/20 nm and emission at 528/20 nm. Each treatment was measured 5 times in six separate biological samples. Each compound was repeated as triplicate. CAMHB without cells was measured for potential background fluorescence, and 100 μM DCFH-DA in PBS was measured to track spontaneous fluorescence of the dye.

Lipid A Analysis Procedure.

Bacteria was cultured in 5 mL of CAMHB for 18 h at 37 °C and then subcultured to 1 × 10⁶ CFU in 120 mL of fresh CAMHB. Subcultures were split and left untreated or dosed with compound at the minimum effective concentration (5 μM) and incubated for 8 h. Treatment was completed with biological duplicates. Cells were pelleted for 15 min at 4000 rpm and 4 °C. The supernatant was discarded, and the pellet was washed with 5 mL of endotoxin free water. The cells were pelleted again, and the supernatant was discarded. Pellets were stored at −80 °C until ready to ship and were shipped on dry ice.

Flat.

Processing and analysis: 1 μL of sample pellets was scraped and plated directly to a steel re-usable MALDI plate. A total of 1 μL of 70% citric acid extraction buffer was spotted on top of the plated bacteria. The steel MALDI plate was added to the chamber with water on bottom and placed in 110 °C oven for 30 min. After, the plate was rinsed with water and air-dried. A total of 1 μL of Norharmane matrix was spotted on each sample. Samples were analyzed in negative ion mode on a Bruker Microflex. Data were processed with flexAnalysis software.

Caroff Protocol.

Processing and analysis: 250 μL of 70% isobutyric acid + 150 μL of 1 M NH₄OH were added to the tubes, and the sample was resuspended and transferred to a caroff tube. Samples were incubated at 100 °C for 1 h. After centrifugation, the supernatant was transferred to 400 μL of endotoxin-free H₂O, flash-frozen, and lyophilized overnight. Samples were washed 1× with 1 mL of MeOH and reconstituted with 3:1.5:0.25 C:M:H₂O for 5 min. The sample (1 μL) was spotted on a steel re-usable MALDI plate with 1 μL of Norharmane matrix (10 mg/mL in 2:1 C:M). Samples were analyzed in negative ion mode on a Bruker Microflex. Data were processed with flexAnalysis software.

Reverse-Transcriptase Quantitative PCR (RT-qPCR) Analysis.

Bacterial strains were incubated at 37 °C overnight in 5 mL of CAMHB and subcultured 1:20 in fresh media in 15 mL conical tubes. The subcultures were incubated for 2 h at 37 °C, at which point 5 μM compound 27 was added to the subculture. Cultures were incubated for another 2 h. Cells were centrifuged (4000 rpm, 20 min, 4 °C), and the supernatant was discarded. RNA was isolated from the bacterial pellet with TRIzol reagent (ThermoFisher) following manufacturer’s instructions with an additional wash with 75% ethanol before solubilizing the RNA pellet. Solubilized RNA was treated with DNase (Turbo DNA-free Kit, Invitrogen) following manufacturer instructions for undiluted samples. Digestion was completed in two steps: 2 μL of DNase was added to the sample and incubated at 37 °C for 30 min, and another 2 μL of DNase was added before an additional 30 min incubation. DNase treated samples were transferred to new 1.5 mL centrifuge tubes. RNA was quantified using a Take3 plate on a BioTek Synergy HTX multimode plate reader. Samples were stored at −20 °C. Changes in expression for genes associated with colistin resistance in A. baumannii (pmrCAB) were evaluated by RT-qPCR. Primers used in this study can be found in Table S10. RT-qPCR was conducted with the Power SYBR Green RNA-to-CT 1-Step Kit (ThermoFisher) according to the manufacturer’s instructions using an Applied Biosystems StepOne instrument. A β subunit of a bacterial RNA polymerase gene rpoB was used as the reference gene in all experiments. Relative standard curves were conducted to ensure linear amplification before completing comparative Ct experiments. Reactions were performed in triplicate on three independent biological samples. Standard error and Student’s t-test were conducted for delta Ct values across all three samples to determine the statisical significance of the fold change for each gene.

Checkerboard Assay for Measurement of Synergy with Colistin.

CAMHB was inoculated with the AB4106 strain (5 × 10⁵ CFU/mL), and 100 μL aliquots were distributed to all wells of a 96-well plate except for well 1A. Inoculated CAMHB (200 μL) containing the selected compound (at a concentration for 2× the highest concentration being tested) was added to well 1a, and 100 μL of the same sample was added to wells 2A–12A. Row A cells were mixed 6–8 times, and then 100 μL was withdrawn and transferred to row B. This process was repeated up to row G (row H was not mixed to determine the MIC of the antibiotic alone). Inoculated media (100 mL) containing colistin at 2× the highest concentration being tested was placed in wells 1A–1H and serially diluted, all the way until column 11 (column 12 was not mixed to determine the MIC of the compound alone). The plates were covered and sealed with Glad Press’n Seal and incubated under stationary- conditions at 37 °C for 16 h. After 16 h, the MIC values of both the compound and antibiotic were recorded as well as the combination. The ΣFIC values were calculated as follows: ΣFIC = FIC (compound) + FIC (antibiotic), where FIC (compound) is the MIC of the compound in the combination/MIC of the compound alone and FIC (antibiotic) is the MIC of the antibiotic in the combination/MIC of the antibiotic.

General Methods for Chemistry.

General Procedure of Characterization and Purification.

General: All reagents were purchased from commercially available sources without further purification. Flash chromatography was performed using 60 Å mesh standard grade silica gel from Sorbetch. NMR solvents were obtained from Cambridge Isotope Labs and used as is. All ¹H NMR (400 MHz) spectra were recorded at 25 °C on a Bruker Avance spectrometer. All ¹³C NMR (101 MHz, 126 MHz and 201 MHz) spectra were also recorded at 25 °C on Bruker Avance spectrometers. Chemical shifts (δ) are given in parts per million (ppm) relative to the respective NMR solvent; coupling constants (J) are in hertz (Hz). Abbreviations used are s, singlet; d, doublet; dd, doublet of doublets; td, triplet of doublets; and m, multiplet. All high-resolution mass spectrometry measurements were made in the Mass Spectrometry and Proteomics Facility at the University of Notre Dame. Infrared spectra were obtained on a FT/IR-4100 spectrophotometer (ν_max in cm⁻¹). UV absorbance was recorded on a Genesys 10 scanning UV/visible spectrophotometer (λ_max in nm). HPLC data was obtained from an Advion A-2030 Scientific Instruments LC system.

HPLC General Procedure.

Samples were analyzed using reverse phase liquid chromatography coupled with low-resolution tandem mass spectrometry (LC–MS/MS). An Advion A-2030 Scientific Instruments LC system was used. Separation was achieved using a Phenomenex C¹⁸ reversed-phase column (5 μm, 150 × 4.6 mm). For analysis, 20 pL of a 1 mg/mL sample was injected. The method used a solvent flow of 1.000 mL/min. Initial gradient conditions (95% water:5% acetonitrile with 0.1% formic acid) were held for 1 min. From 1 to 20 min, the mobile phase composition increased linearly to 5% water:95% acetonitrile with 0.1% formic acid. From 20.1 to 25 min, the mobile phase composition was increased linearly to 1% water:99% acetonitrile with 0.1% formic acid. At 25.1 to 30 min, the column was re-equilibrated with 95% water:5% acetonitrile with 0.1% formic acid. HPLC data acquisition and analysis were performed using Advion LC Data Express software. All compounds showed >95% purity.