Dimeric 2-Aminoimidazoles are Highly Active Adjuvants for Gram-positive Selective Antibiotics against Acinetobacter baumannii

Abstract

The Centers for Disease Control and Prevention (CDC) reports that hospital acquired infections have increased by 65% since 2019. One of the main contributors is the gram-negative bacterium Acinetobacter baumannii. Previously, we reported aryl 2-aminoimidazole (2-AI) adjuvants that potentiate macrolide antibiotics against A. baumannii. Macrolide antibiotics are typically used to treat infections caused by gram-positive bacteria, but are ineffective against most gram-negative bacteria. We describe a new class of dimeric 2-AIs that are highly active macrolide adjuvants, with lead compounds lowering the minimum inhibitory concentration (MIC) to or below the grampositive breakpoint level against A. baumannii. The parent dimer lowers the clarithromycin (CLR) MIC against A. baumannii 5075 from 32 μg/mL to 1 μg/mL at 7.5 μM (3.4 μg/mL), and a subsequent structure activity relationship (SAR) study identified several compounds with increased activity. The lead compound lowers the CLR MIC to 2 μg/mL at 1.5 μM (0.72 μg/mL), far exceeding the activity of both the parent dimer and the previous lead aryl 2-AI. Furthermore, these dimeric 2-AIs exhibit considerably reduced mammalian cell toxicity compared to aryl-2AI adjuvants, with IC₅₀s of the two lead compounds against HepG2 cells of >200 μg/mL, giving therapeutic indices of >250.

Graphical Abstract

2-aminoimidazole dimers in combination with the gram-positive selective antibiotic clarithromycin provide a new approach towards combating antibiotic resistant Gram-negative infections. Macrolides typically have limited efficacy against Gram-negative bacteria due to poor permeation across the Gram-negative outer membrane. However, in the presence of 2-aminoimidazole dimers, clarithromycin becomes efficacious against several Gram-negative species including Acinetobacter baumannii.

Introduction

Antibiotic resistant bacteria remain one of the greatest threats to global human health. The Centers for Disease Control and Prevention (CDC) have estimated that over 2 million people annually in the United States alone are infected with an antibiotic resistant bacterium, resulting in approximately 35,000 deaths.¹ Increased use of antibiotics during the COVID-19 pandemic, particularly the macrolide azithromycin (AZM), will likely further contribute to the antibiotic resistance crisis. The World Health Organization (WHO) has stated that the antibiotic resistance crisis is so serious that it threatens many advances of modern medicine, and that we are on the precipice of a post-antibiotic era, one in which common infections and/or minor injuries could routinely result in increased patient mortality.² Of significant concern are the ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, A cinetobacter baumannii, P seudomonas aeruginosa, and E nterobacter species), which are notorious for their ability to “escape” the action of numerous antibiotics.³ A. baumannii is a gram-negative bacterium that causes life-threatening infections, and is now a leading cause of antibiotic-resistant infections worldwide.^4,5 A. baumannii is especially problematic for patients in intensive care units, causes millions of infections per year across the globe, and exhibits mortality rates as high as 70%.^5,6 Recent reports have indicated that A. baumannii is one of the most common coinfections with severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2), which has the potential to exacerbate the already considerable burden of A. baumannii infection.⁷ Although concerted efforts to identify novel antibiotics remain crucial, progress has been slow, especially in the context of gram-negative bacteria, and resistance can develop rapidly. A critical need therefore exists for the exploration of alternative therapeutic approaches.

One alternative approach to combat antibiotic resistance is the development of antibiotic adjuvants. The adjuvant approach consists of the combination of an antibiotic with a nonantibiotic compound that either inhibits resistance mechanisms directly, or otherwise alters the physiology of an antibiotic-resistant bacterium to render it more susceptible to the antibiotic.⁸ Typically, macrolide antibiotics are prescribed to treat bacterial infections caused by gram-positive bacteria, and have limited efficacy against gram-negative pathogens due to low permeability across the gram-negative outer membrane (OM).⁹ Accordingly, it has been shown that if the OM is compromised then macrolide antibiotics can become efficacious against various gram-negative bacteria, including A. baumannii. For example, compounds that physically disrupt the OM, such as polymyxin derivatives or pentamidine, sensitize gram-negative bacteria to certain gram-positive selective antibiotics.¹⁰ Additionally, compounds that impact synthesis, transport or assembly of OM components can also sensitize gram-negative bacteria to gram-positive selective antibiotics.¹¹

We previously investigated a series of aryl 2-aminoimidazole (2-AI) compounds for potentiation of macrolide antibiotics against multi-drug resistant (MDR) strains of A. baumannii, and identified adjuvants 1 and 2 (Figure 1)^{12, 13} as the most active compounds. Mechanistic studies with the parent 2-AI (Figure S1) indicated that activity is underpinned by altered lipooligosaccharide (LOS) (A. baumannii does not contain O-antigen, thus the outer leaflet of the OM is referred to as LOS)¹⁴ biosynthesis or assembly, and not direct disruption of the OM.¹³ At 7.5 μM (3.0 μg/mL), compound 1 reduces the clarithromycin (CLR) minimum inhibitory concentration (MIC) by 32-fold, from 32 μg/mL to 1 μg/mL against the highly virulent MDR A. baumannii strain 5075 (AB5075), while at 5.0 μM (2.0 μg/mL), it reduces the CLR MIC to 8 μg/mL (Table 1). Compound 2, at 7.5 μM (3.7 μg/mL), reduces the CLR MIC 128-fold to 0.25 μg/mL and at 5.0 μM (2.4 μg/mL), reduces the CLR MIC to 4 μg/mL.¹² Subsequently however, compounds 1 and 2 were determined to have IC₅₀ values of 6.8 μg/mL (16.9 μM) and 23.2 μg/mL (47.6 μM) respectively against HepG2 cells (Table 1, Figure S2). The therapeutic index (TI), in the case of conventional antibiotic development, is typically defined as (mammalian cell IC₅₀)/(antibiotic MIC), with a TI of ≥50 desirable for further development. Since adjuvants are typically non-toxic to bacteria by themselves, the aforementioned definition of TI is not directly applicable. For adjuvant development, we define TI as (mammalian cell IC₅₀)/(adjuvant concentration that reduces the antibiotic MIC to breakpoint level). As macrolides are typically viewed as gram-positive selective antibiotics, there is no clinical breakpoint established for CLR against A. baumannii. For the few gram-negative bacterial species where macrolide therapy is employed, breakpoint MICs are 1–8 μg/mL.¹⁵ In the absence of direct guidance, we have elected to use the CLR breakpoint for S. aureus as a guide. The Clinical Laboratory and Standards Institute (CLSI) currently defines the resistance breakpoint for CLR in S. aureus as 2 μg/mL,¹⁵ thus TIs for compounds 1 and 2 are 2.3–3.4 and 6.3–9.5 respectively. As these TIs are significantly less than 50, we precluded 1 and 2 from further development.

Figure 1.

Aryl-2AI and dimeric 2-AI compounds that potentiate CLR against AB5075.

Table 1.

CLR potentiation against AB5075, and toxicity data for 1–5.

Compound	Compound MIC (μM)	Concentration Tested (μM)	CLR MIC (μg/mL)	Fold Reduction	HepG2 IC50 (μg/mL)	TI
-	-	-	32	-	-	-
1	>200 (>80.3 μg/μL)	30 (12.0 μg/mL)	0.5	64	6.8 ± 0.8	2.3–3.4
		7.5 (3.0 μg/mL)	1	32
		5.0 (2.0 μg/mL)	8	4
2	50 (24.4 μg/mL)	15 (7.3 μg/mL)	0.125	256	23.2 ± 0.4	6.3–9.5
		7.5 (3.7 μg/mL)	0.25	128
		5.0(2.4 μg/mL)	4	8
3	>200 (>89.5 μg/mL)	30 (13.4 μg/mL)	0.125	256	N.D.a	N.D.
		20 (8.9 μg/mL)	0.5	64
		10 (4.5 μg/mL)	4	8
		7.5 (3.4 μg/mL)	8	4
4	>200 (>89.5 μg/mL)	30 (13.4 μg/mL)	0.125	256	N.D.	N.D.
		20 (8.9 μg/mL)	1	32
		10 (4.5 μg/mL)	16	2
5	200 (89.5 μg/mL)	30 (13.4 μg/mL)	≤0.0625	≥512	380 ± 1	113–170
		20 (8.9 μg/mL)	≤0.0625	≥512
		10 (4.5 μg/mL)	0.5	64
		7.5 (3.4 μg/mL)	1	32
		5.0 (2.2 μg/mL)	4	8

^aN.D. not determined

Based upon these undesirable TIs, we initiated a search for additional scaffolds that display high macrolide potentiation activity and significantly reduced cytotoxicity. Herein, we describe the identification and subsequent initial optimization of a series of dimeric 2-AI derivatives for macrolide potentiation, evaluation of leads for cytotoxicity, and mechanistic studies that establish that, similar to the aryl 2-AI series, activity is not dependent upon direct physical disruption of the OM in contrast to pentamidine and polymyxin derivatives.

Results and Discussion

We initiated this study by performing an screen for macrolide potentiation using compounds that have been added to our internal library since the discovery of our first macrolide adjuvants.^{13, 16} Compounds were screened at 30 μM, using AB5075 as our representative bacterial strain and CLR as our representative macrolide. From this screen, we identified dimeric compounds 3-5 (Figure 1) as highly active CLR adjuvants. At 30 μM (13.4 μg/mL, Table 1) both the para-para linked dimer 3 and the para-meta linked dimer 4 suppress the CLR MIC to 0.125 μg/mL, a 256-fold reduction in comparison to CLR alone. Compound 5, the meta-meta linked dimer, was the most active macrolide adjuvant we had discovered to date, returning a CLR MIC of ≤0.0625 μg/mL (≥512-fold reduction) at 30 μM. All three compounds were resynthesized (Scheme S1) to confirm structure and activity. We next measured the inherent anti-bacterial activity of each dimer against AB5075, with compounds 3 and 4 registering MICs of >200 μM (>89.5 μg/mL), and compound 5 an MIC of 200 μM (89.5 μg/mL), indicating that all three compounds potentiate CLR through a non-microbicidal mechanism. We also performed a dose-response study with each dimer (Table 1) and noted that compound 5 is the most potent, returning CLR MICs of 1 and 4 μg/mL at 7.5 μM (3.4 μg/mL) and 5.0 μM (2.2 μg/mL) respectively.

Following identification of the highly active dimeric 2-AI scaffold, we next evaluated the cytotoxicity of lead dimer 5 against HepG2 cells using an XTT assay to assess its potential for further development. Compound 5 exhibits significantly lower cytotoxicity against this cell line than either 1 or 2, with an IC₅₀ of 849.5 μM (380 μg/mL), corresponding to a TI of 113–170 (Table 1, Figure S2), thus we elected to further explore the dimer scaffold. Seeking to augment activity through analog synthesis, we utilized structure activity relationship (SAR) data obtained from our previous studies on the aryl 2-AI scaffold (i.e. compounds 1 and 2).¹² These data indicated that adjuvant activity could be optimized through strategic functional group placement (specifically fluorine) within the central aromatic ring, and thus we focused on modification of the corresponding ring in the dimer scaffold. Based upon this plan, we designed dimers that were accessed by dimerization of various 2-AI-derivatized anilines, with each 2-AI-derivatized aniline assembled through one of two synthetic routes depending on the commercial availability of the starting material (Scheme 1). Synthetic route A began by conversion of commercially available 4- or 3-nitrobenzoic acid derivatives 6a-g to the corresponding acid chloride by treatment with oxalyl chloride, followed by subsequent reaction with diazomethane and quenching with hydrobromic acid to form the target α-bromo-ketones. Cyclization of each α-bromo-ketone with Boc-guanidine delivered 2-AIs 7a-g. Exhaustive Boc-protection of the exocyclic amino group, followed by reduction of the nitro groups using 10% Pd/C and H₂ yielded anilines 8a-g for subsequent dimerization. Route B began by following our previously reported procedure using commercially available 4-aminobenzoic acid derivatives 10a-c¹². Briefly, each amino group was protected using allyl chloride, and the carboxylic acid was then transformed into the Boc-protected 2-AIs 12a-c using an identical four-step approach to that described above. Protection of the exocyclic amino group of the 2-AI with Boc anhydride, followed by removal of the alloc protecting group using palladium tetrakis(triphenylphosphine) and sodium borohydride yielded aniline derivatives 13a-c. For simplicity, our first library focused on evaluation of the corresponding homodimers, thus each 2-AI-aniline was subjected to dimerization using triphosgene, which followed by Boc-deprotection with trifluoroacetic acid and counterion exchange, delivered urea linked 2-AI dimers 9a-g and 14a-c (Figure 2).

Scheme 1.

A Synthesis of 2-AI dimers.

Figure 2.

Dimeric 2-AI analogs.

Dimers were first analyzed for standalone activity against AB5075. Each dimeric 2-AI derivative is essentially inactive alone, with MICs of ≥200 μM (Table 2). The MIC of CLR was then determined against AB5075 in both the absence and presence of each compound at an initial concentration of 30 μM (Table 2). To begin the SAR study, symmetrical dimers 9a and 9b (Figure 2) with fluorine substitution at the 2,2’ or 6,6’ positions on the phenyl core were synthesized (numbering scheme outlined in Figure 1). Both 9a and 9b display high CLR potentiation activity against AB5075 at 30 μM (14.5 μg/mL) (Table 2). The para-para fluorinated dimer 9a is more potent than the meta-meta fluorinated dimer 9b, lowering the CLR MIC ≥512-fold at 30 μM (14.5 μg/mL), compared to 256-fold, and so we elected to first evaluate additional para-para analogues. We found that incorporation of substituents larger than a fluorine, including trifluoromethyl- (9c), methoxy- (9d) chloro- (14b), or methyl- (14c), at either the 2- or 3- positions of the phenyl rings reduces/abolishes all activity, suggesting that steric effects are the likely cause for loss of activity, since neither electron donating nor electron withdrawing groups (outside of fluorine) are highly active. Further, moving the fluorine substituents from the 2,2’ to the 3,3’ positions (14a) also lead to a decrease in activity, returning a CLR MIC of 4 μg/mL at 30 μM (14.5 μg/mL).

Table 2.

Activity of homodimers 9a-g and 14a-c against AB5075.

Compound	Compound MIC (μM)	Concentration Tested (μM)	CLR MIC (μg/mL)	Fold Reduction
-	-	-	32	-
9a	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	≤0.0625	≥512
9b	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	0.125	256
9c	>200 (>116.7 μg/mL)	30 (17.5 μg/mL)	2	16
9d	200 (101.5 μg/mL)	30 (15.2 μg/mL)	8	4
9e	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	32	1
9f	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	0.25	128
9g	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	≤0.0625	≥512
14a	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	4	8
14b	>200 (>103.2 μg/mL)	30 (15.5 μg/mL)	32	1
14c	>200 (>95.1 μg/mL)	30 (14.3 μg/mL)	32	1

Since fluorination appeared to be a promising substitution choice for augmenting adjuvant activity, the next set of analogs synthesized were the meta-meta dimers where we systematically incorporated fluorine to access all possible fluorinated homodimers (9e-g, Figure 2). Compound 9g (5,5’-difluoro) displays greater activity at 30 μM (14.5 μg/mL) than dimers that contain fluorine substituents at either the 2,2’- (9f), 4,6’-(9e) or 6,4’-positions (9b) of the ring and has comparable activity to the parent meta-meta-dimer 5 at 30 μM (13.4 μg/mL), reducing the CLR MIC ≥512-fold (Tables 1 and 2).

The last set of analogs we evaluated was a series of unsymmetrical dimers. The design of the unsymmetrical dimers 9h-m (Figure 2) was based on the activity seen with the homodimers, were strategic fluorine incorporation enhanced activity. With the exception of compound 9i, all heterodimers suppress the CLR MIC below 2 μg/mL, with compounds 9j-9m all giving CLR MICs of ≤0.0625 μg/mL (Table 3).

Table 3.

Activity of heterodimers 9h-m against AB5075.

Compound	Compound MIC (μM)	Concentration Tested (μM)	CLR MIC (μg/mL)	Fold Reduction
-	-	-	32	-
9h	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	0.5	64
9i	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	8	4
9j	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	≤0.0625	≥512
9k	>200 (>96.7 μg/mL)	30 (14.5 μg/mL)	≤0.0625	≥512
9l	>200 (>93.1 μg/mL)	30 (14.0 μg/mL)	≤0.0625	≥512
9m	>200 (>93.1 μg/mL)	30 (14.0 μg/mL)	≤0.0625	≥512

After this initial survey of activity at a fixed concentration (30 μM), a dose response study was conducted with the nine most active compounds (9a-b, 9f-h, 9j-m, Table 4 and Table S1), where we sought to define the lowest concentration at which the adjuvant suppresses the CLR MIC to 2 μg/mL. Eight of the nine compounds tested gave CLR MICs of ≤2 μg/mL at 7.5 μM, with 9h delivering a CLR MIC of 4 μg/mL at 20 μM (9.7 μg/mL). Compounds 9f, 9l, and 9m achieved the 2 μg/mL metric at 7.5 μM, while compounds 9a, 9b, and 9j did so at 5.0 μM. The two most active compounds, 9g and 9k, suppress the CLR MIC to 2.0 mg/mL at 2.5 μM (1.2 μg/mL) and 1.5 μM (0.72 μg/mL) respectively. To the best of our knowledge, 9g and 9k are the most potent A. baumannii macrolide adjuvants disclosed to date.

Table 4.

Dose response activity of lead dimers against AB5075

Compound	Concentration Tested (μM)	CLR MIC (μg/mL)	Fold Reduction
-	-	32	-
9a	20 (9.7 μg/mL)	≤0.0625	≥512
	10 (4.8 μg/mL)	0.125	256
	7.5 (3.6 μg/mL)	0.5	64
	5.0 (2.4 μg/mL)	2	16
9b	20 (9.7 μg/mL)	0.25	128
	10 (4.8 μg/mL)	0.5	64
	7.5 (3.6 μg/mL)	1	32
	5.0 (2.4 μg/mL)	2	16
9f	20 (9.7 μg/mL)	0.5	64
	10 (4.8 μg/mL)	1	32
	7.5 (3.6 μg/mL)	2	16
9g	20 (9.7 μg/mL)	≤0.0625	≥512
	10 (4.8 μg/mL)	0.125	256
	7.5 (3.6 μg/mL)	0.25	128
	5.0 (2.4 μg/mL)	0.5	64
	2.5 (1.2 μg/mL)	1	32
	1.5 (0.72 μg/mL)	2	16
9h	20 (9.7 μg/mL)	4	8
9j	20 (9.7 μg/mL)	0.125	256
	10 (4.8 μg/mL)	0.125	256
	7.5 (3.6 μg/mL)	0.25	128
	5.0 (2.4 μg/mL)	1	32
	2.5 (1.2 μg/mL)	32	1
9k	20 (9.7 μg/mL)	≤0.0625	≥512
	10 (4.8 μg/mL)	≤0.0625	≥512
	7.5 (3.6 μg/mL)	≤0.0625	≥512
	5 (2.4 μg/mL)	≤0.0625	≥512
	2.5 (1.2 μg/mL)	0.125	256
	1.5 (0.72 μg/mL)	2	16
9l	20 (9.3 μg/mL)	≤0.0625	≥512
	10 (4.7 μg/mL)	0.5	64
	7.5 (3.5 μg/mL)	1	32
	5 (2.3 μg/mL)	16	2
9m	20 (9.3 μg/mL)	0.5	64
	10 (4.7 μg/mL)	1	32
	7.5 (3.5 μg/mL)	2	16

Following determination of the dose response activity, the most active dimers from each class (9a, 9g, and 9k) were evaluated against a panel of A. baumannii clinical isolates that encompasses all major and most minor clinically relevant clades^{13, 17, 18} (Table 5, Table S2 and Table S3). The standalone CLR MIC for each of the strains was measured first, with one strain returning a CLR MIC of 1024 μg/mL, nine strains returning a CLR MIC of 64 μg/mL, two strains registering a CLR MIC of 16 μg/mL, and the rest displaying a CLR MIC of 32 μg/mL (MIC₉₀ of 64 μg/mL) (Tables S3 and S4). At 10 μM (4.8μg/mL) compound 9a reduces the CLR MIC₉₀ to 8 μg/mL against this panel, while compounds 9g and 9k surpass breakpoint levels with CLR MIC₉₀ values of 0.5 μg/mL and 0.0625 μg/mL respectively (Table S3). Table 5 summarizes the activity seen with these compounds among several representative AB isolates where compound 9a lowered the CLR MIC by at least 64-fold, and compounds 9g and 9k lowered it at least 256-fold. Compounds 9a and 9g alone displayed no activity toward any of the panel of isolates, however compound 9k displayed moderate standalone activity toward several of the AB isolates, exhibiting MICs of 25–100 μM, (Table S2).

Table 5.

Potentiation of CLR against select AB clinical isolates by lead 2-AI dimers 2-AI 9a, 9g and 9k.

A. baumannii strain	CLR MIC (μg/mL)	CLR MIC (μg/mL) + 9aa	CLR MIC (μg/mL) + 9g	CLR MIC (μg/mL) + 9k
967	16	0.25	≤0.0625	≤0.0625
3560	32	0.5	≤0.0625	≤0.0625
3785	64	1	≤0.0625	≤0.0625
4498	64	1	≤0.0625	≤0.0625
5674	64	0.5	≤0.0625	≤0.0625
5711	64	1	0.125	≤0.0625

^aAll compounds were tested at 10 μM (4.8 μg/mL), <30% standalone MIC

Next, we examined the activity of compounds 9a, 9g, and 9k, as well as the three original dimers 3, 4 and 5, in combination with two additional macrolides: AZM and erythromycin (ERY) as well as rifampin (RIF) against AB5075. At ≤30% their MIC, all six dimers exhibit activity with all three additional antibiotics. Compounds 9a, 9g and 9k display greater potentiation activity at lower concentrations than their corresponding parent dimer, decreasing both the AZM and ERY MICs by at least 16-fold, and the RIF MIC by 32-fold at 10 μM (Table 6). Additionally, compound 9k remained the most active adjuvant in combination with these additional antibiotics, decreasing the AZM and ERY MICs by at least 64-fold at a concentration of 7.5 μM (3.6 μg/mL), and the RIF MIC by 32-fold at a concentration of 5.0 μM (2.4 μg/mL).

Table 6.

Potentiation of other antibiotics against AB 5075 by select compounds.

Compound	Concentration Tested (μM)	AZM MIC (μg/mL)	ERY MIC (μg/mL)	RIF MIC (μg/mL)
-	-	32	32	4
3	30 (13.4 μg/mL)	0.5	0.5	0.0078
	20 (8.9 μg/mL)	1	2	0.125
	10 (4.5 μg/mL)	8	8	2
4	30 (13.4 μg/mL)	1	1	0.03125
	20 (8.9 μg/mL)	4	2	0.25
	10 (4.5 μg/mL)	16	16	2
5	30 (13.4 μg/mL)	0.5	0.25	0.0039
	20 (8.9 μg/mL)	0.5	0.5	0.0156
	10 (4.5 μg/mL)	4	4	0.125
	7.5 (3.4 μg/mL)	-	-	1
9a	30 (14.5 μg/mL)	0.5	0.125	0.125
	20 (9.7 μg/mL)	1	0.5	0.25
	10 (4.8 μg/mL)	2	1	1
	7.5 (3.6 μg/mL)	16	8	-
9g	30 (14.5 μg/mL)	0.5	0.125	0.0039
	20 (9.7 μg/mL)	1	0.25	0.0078
	10 (4.8 μg/mL)	1	0.25	0.03125
	7.5 (3.6 μg/mL)	2	1	0.0625
	5.0 (2.4 μg/mL)	4	2	0.25
9k	30 (14.5 μg/mL)	≤0.0625	≤0.0625	0.0019
	20 (9.7 μg/mL)	≤0.0625	≤0.0625	0.0019
	10 (4.8 μg/mL)	0.125	≤0.0625	0.0078
	7.5 (3.6 μg/mL)	0.5	0.25	0.015
	5.0 (2.4 μg/mL)	1	0.5	0.125
	2.5 (1.2 μg/mL)	32	32	-

To probe the spectrum of activity of these compounds outside A. baumannii, we tested compound 9k in combination with CLR against three other gram-negative bacterial species: K. pneumoniae ATCC BAA-2146 (KP2146), P. aeruginosa PAO1, and Escherichia coli ATCC 25922 (EC25922). Compound 9k displays no standalone activity against either KP2146 or PAO1 (MICs > 200 μM), however it did exhibit activity against EC25922 (MIC 12.5 μM). Compound 9k potentiates CLR against all three gram-negative species at least 256-fold (Table 7). We also examined the activity of 9a, 9g, and 9k in combination with RIF against KP2146 (Table S5). Again, we see greater activity at lower concentrations with the fluorinated compounds than with the parent dimers, with compounds 9g and 9k reducing the RIF MIC from 32 μg/mL to 1 and 2 μg/mL respectively at 7.5 μM.

Table 7.

Activity of compound 9k against other gram-negative species.

Strain	CLR MIC (μg/mL)	Compound MIC (μM)	Concentration Tested (μM)	CLR MIC+ 9k (μg/mL)
KP 2146	512	>200 (>96.7 μg/mL)	30 (14.5 μg/mL) 10 (4.8 μg/mL) 7.5 (3.6 μg/mL)	1 2 16
PAO1	256	>200 (>96.7 μg/mL)	30 (14.5 μg/mL) 20 (9.7 μg/mL)	0. 5 16
EC 25922	32	12.5 (6.0 μg/mL)	3.75 (1.8 μg/mL) 3.0 (1.4 μg/mL) 2.5 (1.2 μg/mL)	0.1 25 4 32

We next probed whether the CLR adjuvant activity of the lead dimers was dependent upon OM disruption. First, we examined the impact that addition of MgCl₂ had on activity (Table S6). Divalent cations, such as Mg²⁺ stabilize the OM of gram-negative bacteria and the activity of membrane disruptors is muted/abrogated by the addition of such cations.¹⁰ When the media was supplemented with MgCl₂ (20 mM), activity was essentially abolished, with compound 9k only reducing the CLR MIC two-fold at 30 μM (vs. 512-fold in non-supplemented media). As suppression of activity could conceivably originate from either directly blocking the compound’s mechanism of action, or simply reducing the permeability of the OM and thereby reducing uptake of the compound, we next studied the impact that the addition of exogenous LOS had upon activity. If these compounds are indeed acting as general membrane disrupters via direct binding to the membrane, the addition of LOS should compete for compound binding and lead to abrogation of activity. LOS was isolated from AB5075 and when added to the media at either 0.2 mg/mL or 0.5 mg/mL, no loss of adjuvant activity was noted for 9k (Table S7), indicating that LOS binding is not involved in the mechanism by which these dimers sensitize A. baumannii to CLR. Finally, we quantified how compound 9k impacts membrane integrity using the BacLight assay (Table S8). Compound 9k had essentially no impact on the membrane in comparison to untreated control (−1.2 ± 2.2 % change in membrane permeability). Taken together, these data strongly suggest that the mechanism by which these dimers potentiate CLR is not dependent upon binding and physically disrupting the OM.

To establish the TI of these dimers, we determined the toxicity of several analogs towards HepG2 cells. Focusing on the most active fluorinated dimers and comparing them to the lead aryl-2AI analogs, half-maximal inhibition activity (IC₅₀) values were measured. As delineated above, compounds 1 and 2 displayed moderate to high toxicity, exhibiting IC₅₀ values of 23.2 ± 0.4 μg/mL (47.6 μM) and 6.8 ± 0.8 μg/mL (16.9 μM) respectively (Figure S2). Lead compounds 9g and 9k however, displayed reduced cytotoxicity against HepG2 cells, giving IC₅₀ values of >256 μg/mL (>530 μM) and 208.1 ± 15.6 μg/mL (430.6 μM) (Table 8, Figure S2), and TIs of >353 and 287 respectively. Compound 9k was further evaluated for hemolytic potential against defibrinated sheep’s blood and displayed little to no lytic activity at concentrations up to 400 μM (highest concentration tested), nor did it exhibit any notable hemolytic activity (3.21 ± 0.71%) at 10 μM in combination with 0.0625 μg/mL CLR (Figure S3, Table S9).

Table 8.

Toxicity and stability data for select compounds.

Compound	Plasma stability % remaining 4 h*	HLM stability % remaining 1 h	CLint (mL/min/kg)	HepG2 IC50 (μg/mL)	TI
1	>95%	57%	10.9	6.8 ± 0.8	2.3–3.4
2	89 ± 1%	54%	<10.4	23.2 ± 0.4	6.3–9.5
5	88 ±10%	93%	<20.8	380 ± 1	113–170
9a	83 ±1%	82%	<20.8	127 ± 25	52.9
9g	> 95%	86%	<20.8	>256	>353
9k	76 ± 9%	87%	<20.8	208 ± 16	287

^*t_1/2 for all compounds is >4 hours.

Finally, we evaluated select dimers for metabolic and plasma stability (Tables S10 and S11). Compounds 5, 9a, 9g, and 9k displayed favorable metabolic and plasma stability, with >80% compound remaining over a 60-minute time-period for microsomal stability (incubation longer than 60 min is usually not done to provide optimal conditions for enzymatic activity¹⁹). The stability of each compound in human liver microsomes as a function of time is shown in Table S12. Furthermore, these dimers displayed favorable plasma stability with ≥76% over four hours (Table 8). Intrinsic clearance (CL_int) was estimated for compounds 5, 9a, 9g, and 9k to be <20.8 mL/min/kg. Low clearance drugs are defined as having CL_int of <15 mL/min/kg, moderate clearance drugs of 15–45 mL/min/kg, and high clearance drugs of >45 mL/min/kg,²⁰ meaning compounds 5, 9a, 9g, and 9k have low to moderate clearance.

Conclusion

In conclusion, it has been well documented that MDR A. baumannii infections cause a serious burden on the health care system, and novel approaches to eradicate such infections are greatly needed. Here we have identified and synthesized a series of novel 2-AI dimeric analogs for potentiation of CLR against a highly virulent strain of A. baumannii, 5075, as well as 25 additional AB isolates. An SAR study revealed dimeric 2-AI compounds 9g and 9k to be our most active adjuvants disclosed to date, enhancing CLR activity by 16-fold at a concentration as low as 1.5 μM (0.72 μg/mL). Compounds 9g and 9k also sensitized AB5075 to several additional antibiotics including, AZM, ERY, and RIF. Mechanistic studies have revealed that these adjuvants do not directly bind to LOS and have no impact on membrane integrity; further mechanistic studies are currently ongoing to address whether these compounds are affecting assembly or biosynthesis of LOS. Lead 2-AI dimer 9k also potentiates CLR against several other gram-negative pathogens, highlighting its potential as a broad-spectrum adjuvant. Finally, lead dimeric 2-AI analogs exhibited low mammalian toxicity, with IC₅₀ values >200 μg/mL against HepG2 cells (which corresponds to TIs of >200), showed no lytic activity towards red blood cells, and had favorable metabolic and plasma stability profiles, indicating that 2-AI dimers such as 9k, have the potential for in vivo evaluation.

Experimental Section

Bacterial strains and antimicrobial agents:

Acinetobacter baumannii clinical isolate 5075 and Pseudomonas aeruginosa PAO1 were obtained from Dr. Colin Manoil at the University of Washington. The panel of A. baumannii clinical isolates was obtained from the Walter Reed Army Institute for Research (WRAIR). Klebsiella pneumoniae BAA-2146 and E. coli 25922 were obtained from ATCC. Colonies were grown on lysogeny broth (LB) Lennox agar purchased from Thermo Scientific (catalog number H26760.36). Cation-adjusted Mueller-Hinton Broth (CAMHB) (catalog number 212322) was purchased from BD Diagnostics. LB broth was purchased from Fisher Scientific (catalog number BP9722–2). Clarithromycin (catalog number C2220), azithromycin (catalog number A2076) and erythromycin (catalog number E0751) were purchased from TCI. Rifampicin (catalog number R7382–1G) and colistin (catalog number C4461–1G) were purchased from Sigma. All assays were completed in duplicate on at least two occasions. All compounds were dissolved as their HCl salts in molecular biology grade DMSO as 1 mM, 5 mM, 10 mM, or 100 mM stock solutions and stored at −20 °C. Both clarithromycin, azithromycin, erythromycin, and rifampicin were dissolved in molecular biology grade DMSO. Colistin was dissolved in water.

Broth microdilution method for MIC determination:

Day cultures (6 h) of each bacterial strain in cation adjusted CAMHB were subcultured to 5×10⁵ CFU/mL in CAMHB. Aliquots (1 mL) were placed in culture tubes and compound from the 10 mM DMSO stock samples was added, where compound concentration equaled the highest concentration tested (200 μM). Samples were then aliquoted (200 μL) into the first row of wells of a 96- well plate, with wells in rows 2–11 being filled with 100 μL of initial bacterial subculture. Row 12 was filled with CAMB as a blank and sterility control. The wells in row-one were mixed eight times, before 100 μL was transferred into row two. Subsequently the wells in row two were mixed eight times, and then 100 μL was transferred into row three. This process was repeated until row 11 had been mixed. The plates were then covered with GLAD Press n’ Seal and incubated under stationary conditions at 37 °C for 16 h. A BioTek Synergy HTX Multimode plate reader was then used to determine the optical density at 600 nm (OD₆₀₀) of cell growth in each well. The lowest dilution with <10% growth as compared to a positive control was recorded as the minimum inhibitory concentration.

Broth microdilution method for antibiotic potentiation:

Day cultures (6 h) of bacteria in CAMHB were subcultured to 5×10⁵ CFU/mL in CAMHB. Aliquots (4 mL) were placed in culture tubes and dosed with compound from 10 mM DMSO stock samples to give the desired concentration of the compound to be tested. 1 mL of the resulting solution was then placed in a separate culture and dosed with antibiotic at the highest concentration to be tested. Bacteria treated with antibiotic alone served as the control. Row one of a 96-well plate was filled with 200 μL of the antibiotic/2-AI dimer solution, and rows 2–11 were filled with 100 μL each of the remaining 4 mL of bacterial subculture containing adjuvant at the desired concentration, except for the control lane which contained only bacterial subculture. Row 12 was filled with CAMB as a blank and sterility control. Row one was then mixed eight times, and 100 μL was transferred into row two, which was then mixed eight times before being transferred into row three. This process was repeated to row ten. Row eleven which would only contain 2-AI dimer to serve as a control to monitor potential toxicity was not mixed, nor was row twelve which contained only CAMHB which served as a control for media contamination. The 96-well plate was then covered with Glad Press n’ Seal and incubated under stationary conditions at 37 °C for 16 h. MIC values were determined as the lowest concentration at which no bacterial growth was observed, and fold reductions were determined by comparison to antibiotic control lane.

BacLight^™ assay:

The BacLight^™ assay (Invitrogen) was used to determine membrane permeability. A. baumannii 5075 was grown overnight in CAMHB at 37 °C with shaking at 200 rpm. The culture was diluted 1:40 in fresh CAMHB and incubated at 37 °C with shaking at 200 rpm for 4 h. Cultures were centrifuged at 4,000 rpm for 20 minutes at 7 °C, the supernatant discarded, and the cell pellet washed with 1 mL sterile water. Cell pellets were resuspended to 1/10 the original volume and diluted 1:20 in sterile water or water containing tested compounds at desired concentrations (10 μM). Suspensions were incubated at 37 °C with shaking for 1 hour and centrifuged at 10,000 rpm for 15 minutes. Cell pellets were washed with sterile water and resuspended in water. A mixture of 1:1 SYTO-9 and propidium iodide was added to each suspension to give a final concentration of 3 μg/mL each. The suspension was aliquoted (100 μL) into each well of a black-walled, clear-bottom 96-well plate and incubated at room temperature in the dark for 15 minutes. Green fluorescence (SYTO-9) was read at 530 nm and red fluorescence (propidium iodide) was read at 645 nm. Both dyes have an excitation of 485 nm (BioTek Synergy HTX). The ratio of green to red fluorescence was expressed as a percentage of the control to determine relative membrane permeability.

Red blood cell hemolysis:

Hemolysis assays were performed on mechanically difibrinated sheep blood (Hemostat Labs: DSB100). Difibrinated blood (1 mL) was placed into a microcentrifuge tube and centrifuged for 10 min at 10,000 rpm. The supernatant was then removed and then the cells were resuspended in 1 mL of phosphate-buffered saline (PBS). The suspension was centrifuged, the supernatant was removed, and cells were resuspended two additional times. The final cell suspension was diluted 10-fold. Compound and antibiotic were added at varying concentrations from DMSO or water stock solutions. PBS was used as a negative control and a zero-hemolysis marker. Triton X 100 (1%) was used as a positive control and as the 100% lysis marker. Samples were then placed in an incubator at 37 °C while being shaken at 200 rpm for one hour. After incubation, samples were centrifuged for 10 min at 10,000 rpm. The resulting supernatant was aliquoted (100 μL) into the wells of a 96-well plate. A BioTek Synergy HTX Multimode plate reader was used to determine the absorbance of the supernatant at a 540 nm wavelength.

Isolation of LOS:

AB5075 was grown for 16 h in CAMHB (500 mL) then cells collected by centrifugation (4000g, 1 h, 4 °C). Pellets were stored at −80 °C then shipped on dry ice for LOS extraction. A. baumannii LOS was extracted using a hot phenol/water extraction as previously described.^{21, 22} To a 50 mL conical tube containing 500 mL of cell culture pellet, 20 mL endotoxin free water was added and vortexed. To this conical tube, 20 mL 65 °C 90% phenol was added and then incubated for 1 hr at 65 °C with intermittent vortexing. After 1 hr, the sample was cooled on ice for 5 minutes and then centrifuged at 3,000 × g for 20 min at room temperature. The aqueous layer was removed and placed into a separate 50 mL conical tube. Endotoxin free water (20 mL) was added to the bottom layer and this process was repeated. Both aqueous layers were combined and dialyzed overnight using 1KD MWCO tubing [Spectrum Laboratories] in at least 4 mL of ddH₂O for at least 24 hr. After dialysis, aqueous volume was split and centrifuged at 5,000 ×g for 20 min at room temperature. Samples were lyophilized in a pre-weighted 50 mL conical tube. The lyophilized material was resuspended in 10 mL of 20 mM Tris-HCl, 2mM MgCl₂ (pH 8.4), and then 2 μL benzonanse [Sigma-Aldrich] and 10 μL DNase [Roche] were added and the solution was incubated for 2 hrs at 37 °C. The pH of the solution was lowered to 7.4 using HCl. Once the pH was 7.4, 100 μL of Proteinase K [Invitrogen] was added and the solution was incubated for 2 hrs at 37 °C. After incubation, 5 mL of water-saturated phenol was added and then the sample was centrifuged for 20 min at 3,000 ×g. The aqueous solution fraction was then dialyzed using 1KD MWCO tubing for at least 12 hr against ddH₂O. After dialysis, the aqueous solution was centrifuged at 5,000 ×g for 20 min and then lyophilized in a pre-weighed 50 mL conical tube. The lyophilized sample was resuspended in 10 mL 2:1 v/v chloroform: methanol, vortexed and centrifuged at 5,000 ×g at 4 °C for 10 min, and then the supernatant was carefully removed making sure to not disturb the pellet. This process was repeated three times. The sample was resuspended in 0.2% TEA and 10% DOC was added to give a final concentration of 0.5%. Water-saturated phenol was added and vortexed intermittently for 5 min and then centrifuged at 5,000 ×g for 10 min. The top layer was transferred to a new tube and 0.2% TEA, and 10% DOC was added to the lower layer and the process was repeated. The aqueous layers were combined, and the water-saturated phenol extraction was repeated ending with the transfer of the aqueous solution into a new 50 mL conical tubes. To this, 37.5 mL cold 100% ethanol and 500 μL of 3 M sodium acetate was added and then placed at −20 °C to precipitate. The precipitate was centrifuged at 5,000 ×g for 10 min and supernatant was removed. The pellet was washed with 24 mL 100% ethanol and centrifuged. Supernatant was removed and the pellet was suspended in 4 mL endotoxin free H₂O and lyophilized in a pre-weighed 50 mL conical tube. For every 5 mg of LPS, 1 mL of 10 mM sodium acetate (pH 4.5) was added and heated for 1 hr at 100 °C. Sample was frozen and lyophilized overnight. For every 5 mg of LPS, sample was washed with 170 μL of endotoxin free H₂O and 850 μL acidified ethanol and then centrifuged for 5 min at 5,000 ×g. The pellet was resuspended in 1 mL endotoxin-free H₂O and then frozen and lyophilized. Purified LOS was confirmed via mass-spectrometry.

XTT assay with HepG2 cells:

HepG2 (ATCC HB-8065) cells were incubated at 37 °C with 5% CO₂. The medium was changed every 2–3 days. The cells were seeded in 96-well plates at a concentration of 5×10⁴ cells/well and were placed in 100 μL of culture medium and incubated for 24 hours at 37 °C with a 5% CO₂ atmosphere. The media was removed from the wells containing the cells using a multichannel pipetter with sterile pipette tips. The test compound (200 μL of a 256 μg/mL solution in medium) was added, followed by serial dilution. The plates were then incubated for 24 hours at 37 °C under an atmosphere of 5% CO₂. The total number of viable cells was estimated by an XTT Cell Proliferation kit (ATCC, Manassas, VA). The activated solution (50 μL of 1.0 mL of Activation Reagent mixed with 5.0 mL of XTT Reagent) was added, followed by incubation for 1–2 hours at 37 °C under an atmosphere of 5% CO₂. The plate was then shaken for 3 min before using the microplate reader (Epoch Microplate Spectrophotometer, BioTek Instruments, Inc., Winooski, VT). The absorbance of the wells was measured at both 475 nm and 660 nm. The specific absorbance was calculated by: Specific absorbance = A_475nm (Test) – average A_475nm (Blank) – A_660nm (Test). The % viability was calculated by: % Viability = Specific Absorbance (test well)/Average specific absorbance untreated cells × 100. The % viability (y-axis) was plotted versus compound concentration (x-axis, log scale) using Graph Pad Prizm and the IC₅₀ was calculated. The assay was done in triplicate.

Metabolic stability:

Metabolic stability was investigated at a concentration of 5 μM in human liver microsomes (Xenotech, 1 mg/mL, pooled from 50 male and female donors) in 1 mL volume of 50 mM potassium phosphate buffer (pH 7.4) containing 0.5 mM NADPH, 0.5 mg of microsomal protein, and 0.1% DMSO at 37 °C. A 100-μL aliquot was drawn at 0, 5, 10, 20, 40, and 60 min and quenched with 200 μL of internal standard (0.02 μM in 50% ACN/50% MeOH) to precipitate protein. After centrifugation, a 5 μL-aliquot was analyzed by ultraperformance liquid chromatography with UV and electrospray ionization mass spectrometry detection in the positive ion mode. The chromatographic conditions consisted of a Waters Acquity UPLC System equipped with a binary solvent manager, an autosampler, a column heater, a photodiode array detector, and a TQD tandem quadrupole detector. A Phenomenex Kinetex 2.6 μ C18 column (2.1 i.d. ×100 mm) column was used. The mobile phase consisted of elution at 0.4 mL/min with 5% A/95% B for 1.5 min, 7-min linear gradient to 100% B, where A = 0.1% formic acid/water and B = 0.1% formic acid/ACN. Transitions were: 452 → 175 and 452 → 201 for compound 1, 365 → 152 and 365 →173 for compound 2, 375 →175 and 275 → 201 for compound 5, 411 → 177 and 411 → 219 for compounds 9a, 9g, and 9k, and 319 → 182 for the internal standard. Assays were carried out in duplicate. Intrinsic clearance was calculated by: CL_int = (0.693/t_1/2) ×(mL incubation volume/mg of microsomal protein) ×(45 mg microsomal protein/gram of liver) ×(20 g of liver/kg body weight).²³

Plasma stability:

For each compound tested a stock at a final a concentration of 20 μM was prepared by adding 998 μL of human plasma and 2 μL of 10 mM compound in DMSO, followed by incubation at 37 °C. A 100-μL aliquot of the incubation was taken at 0, 30, 60, 120, and 240 minutes and mixed with 900 μL of internal standard dissolved in 1:1 ratio of ACN/MeOH. The mixture was centrifuged (14000xg for 10 min) to pellet protein. A 250-μL aliquot of the supernatant was transferred to an autosampler vial and a 5-μL aliquot was analyzed by UPLC-MRM. The chromatographic conditions were the same as those for metabolic stability. The peak areas of the compound and internal standard were integrated, and the ratios calculated (where time 0 = 100% of parent compound) and the percentages of the parent drug remaining at the different time points relative to 0 min were determined, followed by calculation of the half-life. The assay was done in duplicate.

Chemistry Experimental:

All reagents used for chemical synthesis were purchased from commercially available sources (VWR U.S., Fisher Scientific U.S., or Sigma Aldrich U.S.) and used 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 and ¹³C NMR were recorded at 25°C on Bruker AVANCE III HD spectrometers (400, 500, or 800 MHz). 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; bs, broad singlet; d, doublet; dd, doublet of doublets; t, triplet; dt, doublet of triplets; tt, triplet of triplets; m, multiplet. High- resolution mass spectrometry measurements were obtained at the Notre Dame Department of Chemistry Mass Spectrometry and Proteomics Facility. Infrared spectra were obtained on a Bruker Alpha II FTIR spectrophotometer (νmax in cm⁻¹). UV absorbance was recorded on a Genesys 10 scanning UV/visible spectrophotometer (λmax in nm). The purities of the tested compounds were all verified to be ≥ 95% by LC-MS analysis on an Advion LC-MS 2020 with Kinetex, 2.6 mm, C₁₈ 50 × 2.10 mm, using 20–100% acetonitrile/water with 0.1% formic acid for either 5 or 3 minutes.

General procedure for alloc protection:

The corresponding aniline (17 mmol, 1 eq) was dissolved in a 1:1 ratio of H2O:1,4-dioxane (45 mL:45 mL) at room temperature. DIEA (33 mmol, 2 eq) and NaHCO3 (50 mmol, 3 eq) were added to the reaction mixture. The reaction mixture was cooled to 0°C and allyl chloroformate (17 mmol, 1 eq) was added dropwise to the reaction mixture, and the reaction was allowed to stir for 16 h. Upon completion, the reaction mixture was cooled to 0° C and quenched with 1N HCl (70 mL), at which point a white solid began to precipitate out of solution. The precipitate was vacuum filtered and collected as the desired product.

General procedure for α-bromo-ketone formation:

In a flame dried round bottom under N₂ atmosphere was added either the corresponding alloc-protected benzoic acid or the nitro benzoic acid (6.0 mmol, 1 eq) dissolved in anhydrous DCM (20 mL). The reaction mixture was cooled to 0 °C and oxalyl chloride (6.6 mmol, 1.1 eq) was added dropwise followed by the addition of a catalytic amount of DMF (6.5 μmol, 0.0001 eq). The reaction mixture was allowed to stir for 2 h, at which point the solution was concentrated under reduced pressure to yield corresponding desired acid chloride. The acid chloride product (10 mmol, 1 eq) was dissolved in anhydrous DCM (20 mL) and added dropwise to a distilled solution of diazomethane (50 mmol, 5 eq) in diethyl ether at 0 °C and stirred for 1.5 h. TLC analysis was used to verify consumption of acid chloride starting material, at which point concentrated HBr (50 mmol, 5 eq) was added dropwise at 0 °C and stirred an additional 30 min. Upon reaction completion via TLC analysis, the reaction was quenched with NaHCO₃ (200 mL), and the product was extracted with DCM (3 × 300 mL). The combined organic layers were rinsed with brine (1 × 30 mL) and dried with Na₂SO₄. The solvent was removed under reduced pressure, and the residue was rinsed three times with Et₂O to remove impurities. The remaining residue was collected and dried on high vacuum as the corresponding desired α-bromo-ketone and was used without further purification in the general procedure for Boc-guanidine cyclization.

General procedure for Boc-guanidine cyclization:

The corresponding α-bromo-ketone (2 mmol, 1 eq) and Boc-guanidine (6 mmol, 3 eq) were dissolved in anhydrous THF (20 mL) and stirred for 3 h at 56 °C. The solvent was removed under reduced pressure to yield a crude residue. The crude residue was purified by flash chromatography (5 – 100% EtOAc/hexanes) to the afford desired 2-aminoimidazole compounds.

General Procedure for Boc-protection:

To a flame dried round bottom was added the corresponding cyclized 2-AI derivative (4.6 mmol, 1 eq), Boc-anhydride (28 mmol, 6 eq), triethyl amine (14 mmol, 3 eq), and DMAP (0.46 mmol, 0.1 eq). The mixture was dissolved in anhydrous THF (80 mL) and stirred for several hours until TLC analysis indicated reaction completion. Upon completion the mixture was partitioned between EtOAc and water, and the aqueous layer extracted 3x times with EtOAc. The combined organic layers were then washed with 1 N HCl (1x), saturated NaHCO₃ (3x), brine (1x), and then dried with Na₂SO₄, followed by concentration under reduced pressure. After concentration under reduced pressure the product was rinsed with hexanes to remove any impurities. The remaining residue was collected and dried on high vacuum as the corresponding desired tri-Boc-protected 2-AI and was used without further purification in the general procedure for nitro group reduction.

General procedure for nitro group reduction:

The corresponding tri-Boc-protected 2-AI nitro analog was added to a solution of anhydrous methanol and 10 % Pd/C. Air was removed from the system, and the reaction was back flushed with hydrogen. This process was repeated three times before the reaction was placed under hydrogen at room temperature for 4 hours. Upon completion the reaction mixture was filtered and then concentrated down. The remaining residue was purified using flash chromatography (30% EtOAc in hexanes) to yield the corresponding aniline product.

General procedure for alloc deprotection:

The corresponding alloc-protected compound (1.5 mmol, 1 eq) was dissolved in 200-proof EtOH (20 mL). Next, the reaction solution was cooled to 0 °C and tetrakis(triphenylphosphine)palladium (0) (0.035 mmol, 0.02 eq) and sodium borohydride (3.0 mmol, 2 eq) were added to the reaction mixture and allowed to stir for 4 h. After completion, ethanol was removed in vacuo and the remaining residue was then extracted with EtOAc (2 × 30 mL) and washed with deionized water (10 mL). The combined organic layers were rinsed with brine (10 mL) and dried with Na₂SO₄, then concentrated under reduced pressure. The residue was then purified using flash chromatography (30% EtOAc in hexanes) to yield the corresponding aniline product.

General procedure for urea formation using triphosgene:

The corresponding tri-Boc-protected 2-AI aniline (0.421 mmol, 1 eq.) was dissolved in 30 mL DCM. Sodium carbonate (0.674 mmol, 1.6 eq.) which was dissolved in 10 mL H₂O was added to the reaction mixture and was left to stir for 5 minutes at room temperature. Triphosgene (0.139 mmol, 0.33 eq.) was then dissolved in 5 mL DCM and added in dropwise to the reaction mixture and was allowed to stir until TLC showed complete isocyanate formation. Upon formation of the isocyanate, the second tri-Boc-protected aniline (0.421mmol, 1 eq.) dissolved in DCM was added to the mixture and was left to stir until TLC showed complete consumption. The reaction mixture was then extracted with DCM (3 × 10 mL) and washed with deionized water (10mL). The combined organic layers were rinsed with brine (10 mL) and dried with Na₂SO₄, followed by concentration under reduced pressure. The residue was then purified using flash chromatography (30% EtOAc in hexanes) to yield the corresponding dimer product.

General procedure for urea formation using carbonyl diimidazole:

The corresponding tri-Boc-protected 2-AI aniline (0.55 mmol, 1 eq) was dissolved in 30 mL of DCM. Triethyl amine (TEA) (1.1 mmol, 2 eq.) was added to the reaction mixture and was left to stir for 5 minutes at room temperature. Carbonyl diimidazole (0.55 mmol, 1 eq.) was then dissolved in 5mL of DCM and added in dropwise to the reaction mixture and was allowed to stir until TLC showed complete isocyanate formation. Upon formation of the isocyanate, the second tri-Boc-protected aniline (0.55 mmol, 1 eq.) dissolved in DCM was added to the mixture and was left to stir until TLC showed complete consumption. The reaction mixture was then extracted with DCM (3 × 10 mL) and washed with deionized water (10 mL). The combined organic layers were rinsed with brine (10mL) and dried with Na₂SO₄, followed by concentration under reduced pressure. The residue was then purified using flash chromatography (30% EtOAc in hexanes) to yield the corresponding dimer product.

General procedure for Boc-deprotection:

The corresponding Boc-protected 2-AI dimer (1.3 mmol, 1 eq) was dissolved in DCM (1 mL) and TFA (2 mL) was added dropwise into the reaction vessel. The reaction was stirred at room temperature for 3 h. Once reaction reached completion via TLC analysis, the solvent was removed under reduced pressure. The resulting TFA salt was dissolved in MeOH (10 mL) then spiked with 6 M HCl (0.5 mL) and solvent was removed to yield the corresponding HCl salts.

• Small molecule adjuvants overcome intrinsic antibiotic resistance in gram-negative bacteria

• 2-aminoimidazole dimers are highly active macrolide adjuvants against A. baumannii

• Structure activity relationship study leads to higher potency

• High therapeutic indices will allow for exploration of in vivo potential