Development Of Small Molecules That Work Cooperatively With Ciprofloxacin To Clear Salmonella Biofilms In A Chronic Gallbladder Carriage Model

Abstract

Salmonella enterica serovars cause millions of infections each year that result either in typhoid fever or salmonellosis. Among those serovars that cause typhoid fever, Salmonella enterica subspecies Typhi can form biofilms on gallstones in the gallbladders of acutely-infected patients, leading to chronic carriage of the bacterium. These biofilms are recalcitrant to antibiotic-mediated eradication, leading to chronic fecal shedding of the bacteria, which results in further disease transmission. Herein, we report the synthesis and anti-biofilm activity of a 55-member library of small molecules based upon a previously identified hit that both inhibits and disrupts S. Typhi and S. Typhimurium (a nontyphoidal model serovar for S. Typhi) biofilms. Lead compounds inhibit S. Typhimurium biofilm formation in vitro at sub-micromolar concentrations, and disperse biofilms with five-fold greater potentency than the parent compound. Three of the most promising compounds demonstrated synergy with ciprofloxacin in a murine model of chronic Salmonella carriage. This work furthers the development of effective anti-biofilm agents as a promising therapeutic avenue for the eradication of typhoidal Salmonella.

Graphical Abstract

Introduction

Salmonella enterica subspecies cause an estimated 93 million infections globally every year.¹ Salmonella subspecies are categorized as typhoidal or nontyphoidal. Typhoidal subspecies include Salmonella enterica serovar Typhi (S. Typhi), and infection by this serovar results in typhoid fever. Nontyphoidal subspecies include all those that do not cause typhoid fever, and instead typically cause salmonellosis. S. Typhi transmission is typically fecal-oral, and while incidences of typhoid fever are low in the United States and Europe, there is a high burden of disease in developing regions of Sub-Saharan Africa and Southeast Asia.² In these areas, sanitation is lacking and wastewater treatment is underdeveloped, and an estimated 14 million new infections and 136,000 deaths occur each year.³

While treatment with fluoroquinolone antibiotics, such as ciprofloxacin, is typically successful in clearing the infection, about 3–5% of acutely-infected patients are estimated to develop chronic S. Typhi infections due to biofilm colonization in the gallbladder, leading to fecal shedding of bacteria even after resolution of symptoms.^1,4 Humans are the only known reservoir for typhoidal serovars of Salmonella, and this fecal shedding drives re-infection of water sources in areas with poor sanitation, and perpetuates the disease in these populations. In the gallbladder, the bacteria establish and maintain infection by forming biofilms on the surface of gallstones, which are present in up to 90% of chronic carriers.⁵ Biofilms are defined as a surface-attached community of bacteria encased in an extracellular matrix of biomolecules.⁶ The biofilm confers inherent tolerance that allows Salmonella bacteria to survive harsh environments, such as bile within the gallbladder, as well as host immune responses and antibiotic treatment. It has been shown that Salmonella biofilms can be up to 1000-fold more resistant to antibiotic treatment than planktonic Salmonella.⁷ When a patient develops a chronic S. Typhi infection, antibiotic treatment is only moderately successful, necessitating expensive, invasive methods such as cholecystectomy to clear chronic carriage.⁸

Given the importance of biofilms in chronic carriage of S. Typhi infections and the spread of typhoid fever in endemic regions, anti-biofilm treatments represent a promising potential strategy to reduce the spread of infection. Small molecules, such as 2-aminoimidazoles, 2-aminobenzimidazoles, furanones, and N-acyl homoserine lactone derivatives, have previously been used to inhibit biofilm formation and disrupt preformed biofilms of many pathogens.^7,9–12 We recently reported the identification of compound 1 from a screen of 4000 small molecules for Salmonella biofilm inhibition using S. Typhimurium as a surrogate for S. Typhi biofilms.¹³ Compound 1 also disrupts pre-formed biofilms in vitro, an effect that was enhanced upon combination with ciprofloxacin. In a murine model of chronic gallbladder Salmonella carriage, the combination of compound 1 (10 mg/kg/day) and ciprofloxacin (1.0 mg/kg/day) effected a 3–4.5 log reduction in the bacterial burden in the gallbladder, without concomitant bacterial dissemination to peripheral organs, indicating the potential of a dual therapy approach to the clearance of chronic Salmonella carriage in the gallbladder.¹³

Based upon these initial results, we initiated a structure-activity relationship (SAR) study of compound 1 with the aim of augmenting both biofilm inhibition and biofilm dispersion activity. From an analog development standpoint, compound 1 can be divided into three sections: the fluorophenyl alkyl tail (green), the aminopiperidine core (blue), and the thiophene head group (red) (Figure 1). Herein, we report our preliminary SAR findings, focusing on modulating the structure of the fluorophenyl tail and the length of the linker between the tail and the aminopiperidine core. A library of 55 derivatives was constructed and screened for in vitro activity against S. Typhimurium biofilms, leading to the identification of 31 compounds with increased inhibition activity, and 38 compounds with increased dispersion activity in comparison to compound 1. Three derivatives were then evaluated in combination with ciprofloxacin in a murine model of chronic S. Typhimurium gallbladder carriage, of which all three showed synergy with ciprofloxacin towards eliminating bacterial burden, with compound 7d demonstrating the greatest disruptive capabilities.

Figure 1.

Compound 1 divided into tail (green), core (blue), and head (red) sections.

Results and Discussion

We first generated a library of derivatives designed to probe the impact halide identity, number, and position had upon activity. Specifically, this library comprised modifications to the tail group that involved substitution of different halogens in place of the fluorine, as well as incorporation of the fluorine at different positions on the ring. Additionally, as previous SAR studies on anti-biofilm scaffolds have identified 3,5-dihalogenated phenyl motifs as effective in combatting biofilm formation,¹⁴ the 3,5-difluoro motif was also investigated. A derivative with a methoxy group was synthesized to test the effects of an electron-donating group on the activity of the compound, and a control compound was also generated that lacked the aromatic tail completely. The synthetic approach to this first set of analogs is outlined in Scheme 1. The Boc-protected aminopiperidine core (2) was first alkylated to generate Boc-protected piperidines 3a-g. The Boc group of the resulting compounds was removed and the resulting amine alkylated with 3-(bromomethyl)thiophene to generate the eight target compounds 4a-h and compound 5. In addition, an analog of compound 1 lacking the thiophene head group (4h) was generated to determine the necessity of this moiety.

Scheme 1.

Synthetic route to initial library 4a-h and 5. Reagents and conditions: (a) K₂CO₃, RBr, ACN, 82 °C, 24 h; (b) TFA, DCM, rt, 1 h; (c) K₂CO₃, 3-(bromomethyl)thiophene, ACN, 82 °C, 1 h.

These nine compounds were tested for inhibition and disperseion of S. Typhimurium ATCC 14028 (JSG210) biofilms using a procedure that mimics growth in vivo.¹² Inhibition of biofilm formation is assessed by growing bacteria in a 96-well plate in the absence (control) or presence of test compounds. After 24 hours, the wells are washed to remove planktonic bacteria, and the remaining surface attached bacteria (biofilm) are stained with crystal violet (CV). The wells are then washed again to remove excess CV, and the remaining CV is solubilized with a solution of acetic acid and then quantified by spectrophotometry at 570 nm. A dose response curve is then constructed from which an IC₅₀ value is determined. Dispersion of established biofilms is quantified using a similar experimental approach with the exception that biofilms are established first over 24 hrs and then treated with compound. From the dose response curve, 50% dispersion of established biofilms can then be determined, which we refer to as EC₅₀ values.

Under these conditions, the parent compound returned an IC₅₀ of 5.7 μM and an EC₅₀ of 829 μM. The activity of the first library of analogs is summarized in Table 1. Of the compounds containing fluoro substituents (1, 4c-e), the most active biofilm inhibitor remained the orginal lead 1. The 4-fluoro derivative 4d was the most effective dispersion agent, returning an EC₅₀ of 235 μM (ca. 3.5 times more effective than 1). Interestingly, the 3,5-difluoro derivative 4e was inactive. Replacing the fluoro substituent with either a 2-chloro (4a) or 2-bromo (4b) derivative resulted in increased activity in terms of both inhibition and dispersion, with 2-chloro derivative 4a being the most active of these first analogs. Replacement of the fluoro substituent with a methoxy in 4f resulted in decreased inhibition and dispersion activity, as did complete removal of the halogens in 4g. Removal of the thiophene head (5) abrogated activity while removal of the tail (4h), had a significant detrimental impact on inhibition but interestingly led to a modest increase in dispersion activity.

Table 1.

IC₅₀ and EC₅₀ values for compounds 1, 4a-h, and 5. All values are in μM and are presented as the mean ± the standard deviation.

Compound	IC50 (μM)	EC50 (μM)
1	5.7 ± 0.9	829 ± 43
4a	6.0 ± 1.4	232 ± 6
4b	7.2 ± 1.6	308 ± 11
4c	6.9 ± 0.1	388 ± 19
4d	16.8 ± 3.4	235 ± 18
4e	>200a	>12000
4f	16.9 ± 5.2	11763 ± 77
4g	21.2 ± 2.6	>12000
4h	138.1b ± 8.6	442 ± 38
5	>200b	>12000

^aGrowth inhibition at ≥50 μM;

^bGrowth inhibition at ≥100 μM

Leveraging this preliminary information, the next set of compounds included incorporation of chlorine and bromine at the 3- and 4- positions (6a-d) to further determine if there was a correlation between activity and halogen identity (Figure 2). Compounds containing 3,4, 3,5, and 2,6 halogen substitution patterns (6e-h) were synthesized to probe whether di-halogenation in general was detrimental to activity. Lastly, the length of the linker between the tail and the piperidine core was both increased and decreased by one methylene unit to determine if this had any effect on activity. The activities of compounds 6a-n are summarized in Table 2.

Figure 2.

Second library of compounds 6a-n.

Table 2.

IC₅₀ and EC₅₀ values for compounds 6a-n. All values are in μM compound and are presented as the mean ± the standard deviation.

Compound	IC50 (μM)	EC50 (μM)
6a	2.0a ± 1.2	135 ± 12
6b	3.0 ± 1.4	69 ± 4
6c	3.6a ± 0.9	83 ± 6
6d	3.0 ± 0.7	57 ± 4
6e	1.6b ± 0.5	45 ± 2
6f	3.4a ± 1.4	42 ± 3
6g	3.0 ± 0.2	87 ± 13
6h	6.3 ± 0.6	138 ± 18
6i	8.8 ± 0.8	125 ± 6
6j	2.7 ± 0.7	1810 ±176
6k	26.7 ± 1.2	>12000
6l	11.5 ± 2.7	221 ± 17
6m	5.4 ± 1.2	59 ± 4
6n	78.1 ± 3.7	>12000

^aGrowth inhibition at ≥100 μM;

^bGrowth inhibition at ≥50 μM

Following the trend observed with the first set of analogs, replacement of fluorine with either a chlorine or a bromine increased activity in the context of both biofilm inhibition and dispersion. However, while incorporation of chlorine in the 3-position resulted in an improved IC₅₀ of over two-fold (6a vs 4c), and in the 4-position of over four-fold (6b vs 4d), bromine incorporation had a larger effect on dispersion activity (6b vs 6d). Placing two fluorine atoms on the tail in positions 3 and 4 (6g) reduced both the IC₅₀ and EC₅₀ values by over two-fold compared to compounds 4c and 4d, indicating that these positions warranted further exploration via both chlorination and bromination. Strategic placement of chlorines at two positions on the ring as in 6e resulted in the highest activity for inhibition and second-highest for dispersion (1.6 μM/45 μM), while the bis-bromo analog 6f returned the lowest EC₅₀ value (42 μM). Decreasing the length of the linker by one methylene unit (6i-j) resulted in a decrease in IC₅₀ values and notably, a decrease in EC₅₀ values. Lengthening the linker by one methylene unit (6l-n), however, resulted in reduced activity for both inhibition and dispersion.

The final two libraries in this study focused on exploring the addition of multiple Cl/Br substituents on the tail, as well as tails connected via contracted or elongated linkers (Figure 3). Additionally, six iodo derivatives were also synthesized (8a-f) to investigate the effects of a larger, less-electronegative halogen on activity (Figure 4). Finally, five derivatives (8g-k) were generated as steric isosteres of chlorine, bromine and iodine to probe the effects of sterics versus electronics (Figure 4). The activities of these compounds are reported together in Table 3.

Figure 3.

Third library of compounds 7a-u.

Figure 4.

Fourth library of compounds 8a-k.

Table 3.

IC₅₀ and EC₅₀ values for compounds 7a-u and 8a-k. All values are in μM and are presented as the mean ± the standard deviation..

Compound	IC50 (μM)	EC50 (μM)
7a	1.4a ± 0.9	45 ± 2
7b	0.49 ± 0.4	37 ± 4
7c	2.5 ± 0.5	22 ± 2
7d	1.7b ± 0.1	25 ± 6
7e	12.7 ± 4.1	60 ± 2
7f	4.9 ± 0.8	79 ± 3
7g	8.7 ± 1.3	1051 ± 77
7h	5.2 ± 1.3	299 ± 21
7i	6.5 ± 2.9	142 ± 9
7j	12.9 ± 1.4	443 ± 52
7k	5.9 ± 2.8	934 ± 26
7l	8.5 ± 1.8	127 ± 13
7m	14.2 ± 1.7	>12000
7n	>200c	>12000
7o	7.2 ± 1.8	3803 ± 204
7p	15.8 ± 1.4	196 ± 27
7q	3.9 ± 1.2	55 ± 8
7r	2.7 ± 1.4	70 ± 8
7s	11.9 ± 1.7	32 ± 1
7t	145.7c ± 18.3	870 ± 79
7u	20.9c ± 6.1	140 ± 39
8a	4.3a ± 1.8	60 ± 2
8b	4.0 ± 0.7	44 ± 2
8c	2.6a ± 0.4	89 ± 10
8d	5.4 ± 0.5	79 ± 5
8e	10.0 ± 4.5	92 ± 3
8f	5.7 ± 2.9	77 ± 3
8g	6.3 ± 2.3	247 ± 11
8h	2.3b ± 0.6	123 ± 7
8i	1.8b ± 0.4	31 ± 4
8j	3.2b ± 0.3	16 ± 1
8k	2.5a ± 0.8	28 ± 5

^aGrowth inhibition at ≥50 μM;

^bGrowth inhibition at ≥12.5 μM;

^cGrowth inhibition at ≥100 μM

The data for these compounds revealed that while compounds with shorter linkers were generally more active than parent compound 1 for both inhibition and dispersion of biofilms, they were outperformed by several compounds with two methylene units between the halobenzene and the core. Notably compound 7b returned an IC₅₀ of 490 nM, making it, to the best of our knowledge, one of the most potent inhibitors of Salmonella biofilms reported to date. Additionally, compounds that contained steric isosteres of the various halogens proved to be more potent than the corresponding halogenated compounds. Notably, this was seen with the iodinated derivatives and the biphenyl compounds; 8i was two-fold more active at inhibiting biofilm formation than its iodinated counterpart 8b. A number of compounds also exhibited improved dispersion activity, including 7c, 7d, and 8j.

Based upon this in vitro data, compound 1 and three analogs from this study were selected for in vivo studies in a murine model of gallbladder carriage. Biphenyl compound 8j exhibited the highest dispersion activity with an EC₅₀ of 16 μM. Compound 7b demonstrated the most potent inhibition activity with an IC₅₀ of 490 nM. Lastly, 7d was selected due to its lowest combined IC₅₀ and EC₅₀ values of 1.7 μM and 25 μM, respectively. 129X1/SvJ NRAMP^+/+ mice were fed a lithogenic diet for eight weeks in order to induce gallstone formation, thereby mimicking human carriers and allowing biofilm growth within the gallbladder following infection with S. Typhimurium. Mice were randomized to one of five 10-day treatment regimens: vehicle control (DMSO), 5 mg/kg/day compound 1 + 1 mg/kg/day ciprofloxacin (cipro), 5 mg/kg/day 7b + 1 mg/kg/day cipro, 5 mg/kg/day 7d + 1 mg/kg/day cipro, or 5 mg/kg/day 8j + 1 mg/kg/day cipro. A dose of 5 mg/kg/day compound was chosen because previous experiments in mice treated with compound 1 at a dose of 10 mg/kg/day reduced bacterial burden in the gallbladder by a factor of several logs, nearing the limit of detection; thus, a lower dose was used in order to more precisely compare compound activities.¹³ Similarly, compounds were dosed in combination with cipro in order to prevent dissemination and accumulation of released bacteria in distal organs such as the liver and spleen, which we have previously observed when administering anti-biofilm compounds without concomitant antibiotics.¹³ An initial dose of 1 mg/kg/day cipro was chosen as we have previously shown that treating with cipro alone at this concentration was able to significantly reduce gallbladder-borne Salmonella in a model of acute infection (i.e. when gallbladder Salmonella are predominately planktonic), but not in a chronic model of infection (i.e. when gallbladder Salmonella exist primarily within a biofilm.)^{13, 15}

Compounds 7d and 8j effected the greatest reduction in bacterial burden in the gallblader (4.5–5 and 3.5–4 log-fold respectively). Compounds 1 and 7b, which were less active biofilm dispersing agents in vitro than 7d and 8j, were also less active in vivo (~2 and ~1.5 log-fold reduction respectively) (Figure 5A). While concomitant administration of cipro at a dose of 1 mg/kg/day prevented further accumulation of bacteria within the liver and the spleen¹³, mice still harbored a significant number of bacteria within these organs following treatment (Figure 5B–C). In order to determine if the bacterial burden in these organs could be further reduced by increasing the concentration of cipro, we tested cipro doses of 2 and 4 mg/kg/day – either alone or in combination with 7d – over the same 10-day treatment period. Similar to treatment with 1 mg/kg/day cipro alone,^{13, 15} treatment with 2 mg/kg/day cipro alone had no statistically significant effect on bacterial burden within the gallbladder of infected mice, though it did reduce the number of bacteria recovered from the liver and spleen (Figure 5E–F). Further increasing dose of cipro administered in combination with 7d to 4 mg/kg/day further reduced bacterial burden in all organs such that the number of CFUs recovered neared the limit of detection (Figure 5D–F). However, while treatment with 4 mg/kg/day cipro alone significantly reduced bacterial gallbladder burden by a factor of 3–4 logs (Figure 5D), treatment with 7d and 4 mg/kg/day cipro resulted in an additional significant reduction of gallbladder bacteria.

Figure 5.

Enumeration of Salmonella in the gallbladder, liver, and spleen of infected mice at 15 dpi. 129X1/SvJ mice were fed a lithogenic diet for 8 wks. prior to intraperitoneal (I.P.) infection with 10³ S. Typhimurium. In panels A-C, mice were administered I.P. a vehicle control (DMSO) or one of four combination treatments consisting of 1 mg/kg/day ciprofloxacin (cipro) + 5 mg/kg/day compound 1, 7b, 7d, or 8j from days 5–15 post-infection. In panels D-F, mice were administered I.P. a vehicle control, 2 mg/kg/day cipro alone, 4 mg/kg/day cipro alone, 5 mg/kg/day 7d + 2 mg/kg/day cipro, or 5 mg/kg/day 7d + 4 mg/kg/day cipro from days 5–15 post-infection. Dotted lines represent the limit of detection where applicable; significant differences among the bacterial burden of treatment groups were determined via one-way ANOVA with the Tukey correction for multiple comparisons; ns non-significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Conclusions

In conclusion, a structure-activity relationship study of compound 1 involving the halophenyl alkyl tail and the linker between the tail and the core was conducted. The initial library synthesized (Figure 2) revealed that compounds with chlorine (4a) and bromine (4b) substitutions on the tail returned higher activity in terms of biofilm dispersion than compound 1. The subsequent libraries determined that the optimum length of the linker between the piperidine core and the tail was two methylene units (Table 2 and Table 3). Additionally, incorporation of chlorine in the 3-position resulted in an improved IC₅₀ of over two-fold (6a vs 4c) and in the 4-position of over four-fold (6b vs 4d), while bromine incorporation at these positions had a larger effect on dispersion activity (6b vs 6d). Three compounds (7b, 7d, and 8j) were selected for their superior IC₅₀ and EC₅₀ values and were tested alongside compound 1 in vivo in a murine model of gallbladder Salmonella chronic carriage with S. Typhimurium. These compounds dispersed Salmonella biofilms within the gallbladder to varying degrees, with compound 7d proving the most effective, reducing the bacterial burden in the gallbladder by 4.5–5 logs. Combination of compounds with ciprofloxacin effectively prevents the accumulation of bacteria released from the gallbladder to distal organs in an antibiotic dose-dependent manner. Further modifications of the lead compounds from this study are ongoing, as well as mechanism of action studies to determine the target of these molecules.

Experimental Section

All commercial solvents and reagents were purchased from VWR, Sigma-Aldrich, Oakwood Chemical, Matrix Scientific, or Enamine Ltd and used without any further purification. Reactions were monitored via thin layer chromatography (TLC) using glass-backed pre-coated silica gel plates from VWR (TLC Silica Gel 60 Sheets, MilliporeSigma, F254, 60Å pore, 230–400 mesh) using UV visualization, ninhydrin stain, p-anisaldehyde stain, and/or vanillin stain as visualizing agent. Column chromatography was performed using silica gel (60Å, particle size 40–60 μm, VWR). Solvent system for compound purification was a mixture of ammonia-saturated methanol in DCM with an initial DCM column flush unless otherwise indicated. Ammonia-saturated methanol was prepared by bubbling ammonia (Airgas) into methanol over the course of 15 minutes. Deuterated solvents for NMR characterization were purchased from MilliporeSigma via VWR and used as is, with the exception of chloroform-d, to which molecular sieves were added (4Å, grade 514, mesh 8–12, Macron Fine Chemicals). All NMR spectra were performed at room temperature and recorded on either a Bruker AVANCE III HD 400 Nanobay spectrometer or a Bruker AVANCE III HD 500 without the use of signal suppression function and calibrated using the residual undeuterated solvent peak (CDCl₃: δ 7.26 ppm ¹H NMR, 77.16 ppm ¹³C NMR; CD₃OD: δ 3.31 ppm ¹H NMR, 49.00 ppm ¹³C NMR). Proton (¹H) NMR are reported as follows: chemical shift in ppm (multiplicity, coupling constant(s) in Hz, relative integration). Abbreviations used are s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad. High resolution mass spectra (HRMS) were recorded on a Bruker micrOTOF II by electrospray ionization (ESI) time of flight (TOF) experiments using direct infusion in 9:1 acetonitrile:water. Analyses were performed by the mass spectrometry and proteomics facility at University of Notre Dame and reported as m/z. All compounds were characterized and tested at ≥95% purity as determined by liquid chromatography on either a Bruker micrOTOF-Q II by the University of Notre Dame mass spectrometry and proteomics facility or an Advion LC-MS 2020 with Kinetex, 2.6 mm, C18 50×2.10mm.

General Synthetic Procedure for Carboxylic Acid Reduction.

To an oven-dried 250 mL 3-necked round bottom flask was added 50 mL anhydrous tetrahydrofuran (THF) under argon and cooled to 0° C. LiAlH₄ (9.0 mmol, 9.0 mL 1M solution in THF) was added and allowed to cool. Carboxylic acid (3.0 mmol) was dissolved in 15 mL anhydrous THF and added dropwise to the solution over 30 minutes. Reaction was monitored by TLC and allowed to stir for an additional 1.5 hours. Reaction mixture was then quenched with a saturated solution of Rochelle’s salt (20 mL) and allowed to stir for 5 minutes. Organic layer was removed and evaporated under reduced pressure. Crude oil was dissolved in ethyl acetate (100 mL), washed with brine (25 mL), dried over sodium sulfate, filtered, and concentrated in vacuo to yield an oil.

General Synthetic Procedure for Alcohol Mesylation.

The alcohol (3.0 mmol), either purchased or crude reduced carboxylic acid, was dissolved in 50 mL anhydrous dichloromethane (DCM) and added to an oven-dried round bottom flask under argon. Triethylamine (6.0 mmol) was added and the solution was cooled to 0° C. Methanesulfonyl chloride (9.0 mmol) was added in one portion and the solution was allowed to stir for 30 minutes at 0° C followed by 2 hours as it warmed to room temperature. The mixture was quenched with 20 mL di-H₂O and extracted with DCM (3×30 mL). Organic layers were combined and washed with brine, dried over sodium sulfate, and concentrated in vacuo. Crude mixture was then carried forward without further purification.

General Synthetic Procedure for Piperidine Alkylation.

To an oven-dried two-necked 100 mL round bottom flask under argon was added anhydrous acetonitrile (30 mL) and tert-butyl methyl(piperidin-4-ylmethyl)carbamate (0.50 mmol). Oven-dried K₂CO₃ (1.5 mmol) was added to the flask in one portion, and the resulting mixture was heated to reflux. Mesylated intermediate or brominated starting material was added neat in one portion (1.25 mmol) and the reaction mixture was stirred under reflux for 24 hours. The reaction was then cooled and transferred to a single-necked 250 mL round bottom to be concentrated under reduced pressure. The crude was then dissolved in di-H₂O (50 mL) and extracted with DCM (3×30 mL). The organic layers were collected and washed with brine (30 mL), dried over anhydrous sodium sulfate, filtered, evaporated in vacuo and purified via flash column chromatography using a gradient of 2–5% methanol saturated with ammonia in DCM.

General Synthetic Procedure for Boc Deprotection.

Boc-protected intermediate (0.40 mmol) was dissolved in 1 mL DCM. 2 mL trifluoroacetic acid (30 mmol) was added and the reaction was allowed to stir at room temperature open to atmosphere for 1 hour. Methanol (2 mL) was then added and mixture was evaporated in vacuo to dryness. Addition of methanol was repeated four times until no further vapors evolved upon addition of solvent. Crude intermediate was dried for 18 hours under vacuum and no further purification was performed.

General Synthetic Procedure for 3-(bromomethyl)thiophene Addition.

Deprotected intermediate (0.40 mmol) was dissolved in 23 mL anhydrous acetonitrile and was transferred to an oven-dried 100 mL three-necked flask under argon. K₂CO₃ (1.50 mmol) was added and mixture was heated to reflux while stirring. 3-(bromomethyl)thiophene (0.45 mmol) was dissolved in 2 mL anhydrous acetonitrile and added to the reaction dropwise over one hour. The reaction was checked for completion by TLC after full addition of 3-(bromomethyl)thiophene, then cooled and transferred to a single-necked 250 mL round bottom flask and evaporated in vacuo. The crude was then dissolved in di-H₂O (50 mL) and extracted with DCM (3×30 mL). The organic layers were collected and washed with brine (30 mL), dried over anhydrous sodium sulfate, filtered, and evaporated under reduced pressure. The crude mixture was purified via column chromatography using 2–5% methanol saturated with ammonia in DCM.

General Synthetic Procedure for Making HCl Salts.

Pure product was dissolved in 1 mL methanol and glacial hydrochloric acid (0.1 mL) was added to make a salt. The product was evaporated in vacuo. Methanol addition and subsequent evaporation was repeated six times until no further fumes evolved upon solvent addition. Solid obtained was dried under vacuum for 24–48 hours to yield the salt of the pure product.

Synthetic Procedure for Carboxylic Acid Reduction and Mesylation with Iodinated Compounds.

To an oven-dried 2-necked round bottom flask under argon was added anhydrous THF (50 mL) and iodophenylacetic acid (1.9 mmol). The reaction mixture was cooled to 0° C and NaBH₄ (5.7 mmol) was added in three portions and allowed to stir for 20 minutes. BF₃·Et₂O (3.8 mmol) was added via syringe pump over 15 minutes and the reaction was allowed to warm to room temperature and stirred for 16 hours. The reaction was quenched slowly with 10 mL cold methanol and evaporated in vacuo. The residue was dissolved in ethyl acetate (75 mL), washed with 1N HCl (50 mL), dried over sodium sulfate, filtered, and evaporated in vacuo. The intermediate was carried forward with no further purification. The general synthetic procedure for mesylation was followed to yield an oil.

Synthetic Procedure for Piperidine Alkylation with Iodinated Compounds.

To an oven-dried two-necked 100 mL round bottom flask under argon was added anhydrous acetonitrile (30 mL) and tert-butyl methyl(piperidin-4-ylmethyl)carbamate (0.50 mmol). Oven-dried K₂CO₃ (1.5 mmol) was added to the flask in one portion, and the resulting mixture was heated to 55° C. Mesylated intermediate or brominated starting material was added neat in one portion (1.25 mmol) and the reaction mixture was stirred at 55° C for 18 hours. The reaction was then cooled and transferred to a single-necked 250 mL round bottom to be concentrated under reduced pressure. The crude was then dissolved in di-H₂O (50 mL) and extracted with DCM (3×30 mL). The organic layers were collected and washed with brine (30 mL), dried over anhydrous sodium sulfate, filtered, evaporated in vacuo and purified via flash column chromatography using a gradient of 2–5% methanol saturated with ammonia in DCM. Subsequent Boc-deprotection of the product followed the general synthetic procedure for Boc-deprotection.

Synthetic Procedure for 3-(bromomethyl)thiophene Addition with Iodinated Compounds.

Deprotected intermediate (0.40 mmol) was dissolved in 23 mL anhydrous acetonitrile and was transferred to an oven-dried 100 mL three-necked flask under argon. K₂CO₃ (1.50 mmol) was added and mixture was heated to 55° C while stirring. 3-(Bromomethyl)thiophene (0.45 mmol) was dissolved in 2 mL anhydrous acetonitrile and added to the reaction dropwise over one hour. The reaction was checked for completion by TLC after full addition of 3-(bromomethyl)thiophene, then cooled and transferred to a single-necked 250 mL round bottom flask and evaporated in vacuo. The crude was then dissolved in di-H₂O (50 mL) and extracted with DCM (3 × 30 mL). The organic layers were collected and washed with brine (30 mL), dried over anhydrous sodium sulfate, filtered, and evaporated under reduced pressure. The crude mixture was purified via column chromatography using 2–5% methanol saturated with ammonia in DCM to obtain product. Subsequent creation of the HCl salt proceeded using the general synthetic procedure for making HCl salts.

Synthetic Procedure for Piperidine Alkylation with Starting Material Containing One or Three Methylene Linkers.

To an oven-dried two-necked 100 mL round bottom flask under argon was added anhydrous acetonitrile (30 mL) and tert-butyl methyl(piperidin-4-ylmethyl)carbamate (0.50 mmol). Oven-dried K₂CO₃ (1.5 mmol) was added to the flask in one portion, and the resulting mixture was heated to reflux. Brominated starting material (0.75 mmol) was added neat in one portion and the reaction mixture was stirred at reflux for 14–18 hours, checking via TLC for completion. The reaction was then cooled, transferred to a single-necked 250 mL round bottom, and concentrated under reduced pressure. The crude was then dissolved in di-H₂O (50 mL) and extracted with DCM (3 × 30 mL). The organic layers were collected and washed with brine (30 mL), dried over anhydrous sodium sulfate, filtered, evaporated in vacuo and purified via flash column chromatography using a gradient of 2–5% methanol saturated with ammonia in DCM. Subsequent synthesis followed general procedures outlined above.

Salmonella biofilm growth and crystal violet staining.

S. Typhimurium ATCC 14028 was streaked onto Luria-Bertani (LB) (Thermo Fisher Scientific, Catalog No #BP1426) agar plates and incubated at 37 °C overnight. For biofilm assays, individual bacterial colonies were picked and used to inoculate LB broth for overnight liquid cultures, which were grown at 37 °C with aeration using a rotating drum. Overnight cultures were normalized to an OD₆₀₀ of 0.8 (~6.4 × 10⁸ colony forming units (CFU)/mL, further diluted 1:100 into minimal media (TSB (Thermo Fisher Scientific, Catalog No #DF0370-007-5) diluted 1:20 in ddH₂O), and added to non-treated, flat-bottom polystyrene 96-well plates (Corning, Kennebunkport, ME) at a volume of 100 μL/well. Biofilm plates were incubated at 30 °C on a Fisherbrand^™ nutating mixer (Thermo Fisher Scientific, Waltham, MA; 20° fixed angle, 24rpm) for a total of 24 or 48 h for inhibition and dispersion assays, respectively. Biofilm growth was measured using a semi-quantitative method via CV staining. Biofilm plates were then submerged in dH₂O to wash away any remaining non-adherent bacteria and heat fixed (1h, 60 °C). Biofilms were then stained with a 33% crystal violet solution (6 mL PBS, 3.3 mL crystal violet, 333 μL methanol, 333 μL isopropanol) for 5 min. and washed twice by submerging in dH₂O before releasing the bound dye with 33% glacial acetic acid. Biofilm growth was then quantified by measuring the optical density of the solubilized dye at 570 nm using a spectrophotometer (Molecular Devices, SpectraMax M5).

Biofilm Inhibition and Dispersion Assays and IC₅₀/EC₅₀ Determination.

For biofilm inhibition (IC₅₀) assays, S. Typhimurium biofilms were grown as described above but with various concentrations (0.2–100 μM) of compound (diluted in media from 100 mM stock solutions) or vehicle (DMSO) supplied in the media at the time of inoculation. Biofilms were then grown as described above for a total of 24 h prior to CV staining. For biofilm dispersion (EC₅₀) assays, S. Typhimurium biofilms were grown in media only for a total of 24 h as described above. Spent media was then removed and replaced with fresh media containing various concentrations (1.56–200 μM) of compound (diluted from 100 mM stock solutions) or vehicle. Biofilms were then incubated for an additional 24 h in the presence of compounds prior to CV staining. IC₅₀/EC₅₀ values were calculated by plotting normalized compound activity (percent biofilm formed and percent remaining, respectively) as a function of log₁₀ compound concentration and fitting a dose response curve (log[inhibitor] vs. normalized response, variable slope).

Murine model of chronic Salmonella gallbladder carriage and evaluation of compound therapeutic efficacy in combination with ciprofloxacin.

A total of 46 adult 129X1/SvJ NRAMP^+/+ mice (The Jackson Laboratory, Bar Harbour, ME) were used in this study. As described previously⁴, mice were fed a lithogenic diet (conventional mouse chow supplemented with 1% cholesterol and 0.5% cholic acid; Envigo, Indianapolis, IN) for 8 weeks prior to infection in order to promote the formation of gallstones. Liquid cultures of S. Typhimurium ATCC 14028 were diluted in sterile PBS to a final inoculum density of ~ 5 × 10³ CFU/mL, and mice were infected with ~10³ CFU via injection of 200 μL inoculum into the intraperitoneal (I.P.) cavity. Mice were randomly assigned to treatment groups in two separate experiments. In the first experiment, treatment groups were as follows: vehicle (5% [v/v] DMSO in PBS), 5 mg/kg/day compound 1 + 1 mg/kg/day ciprofloxacin (cipro, Fluka cat. Number R1678), 5 mg/kg/day 7b + 1 mg/kg/day cipro, 5 mg/kg/day 7d + 1 mg/kg/day cipro, and 5 mg/kg/day 8j + 1 mg/kg/day cipro. In the second experiment, treatment groups were: vehicle (same as above), 2 mg/kg/day cipro alone, 4 mg/kg/day cipro alone, 5 mg/kg/day 7d + 2 mg/kg/day cipro, and 5 mg/kg/day 7d + 4 mg/kg/day cipro. Treatments were administered daily via I.P. injection from days 5–15 post-infection. On day 15 post-infection, mice were euthanized and gallbladders, livers, and spleens were removed and homogenized in a volume of 1 mL sterile PBS using a TissueLyser LT bead mill (Qiagen, Valencia, CA). Tissue homogenates were serially diluted in PBS, plated onto LB agar, and incubated at 37 °C for 16h in order to quantify bacterial burden via CFU enumeration.

Statistical Information.

For IC₅₀/EC₅₀ experiments, data represents the average ± the standard deviation for three biological replicates. In graphs of animal data, CFU values for individual animals were plotted as data points and treatment group averages are represented by horizontal lines; statistical analyses of CFU values were conducted using log-transformed values. All data transformations and statistical analyses were performed using GraphPad Prism 8, and p values < 0.05 were considered significant unless otherwise specified (i.e. when correcting for multiple comparisons).

• Salmonella enterica infects 93 million individuals around the world each year

• S. Typhi causes typhoid fever and can be a chronic infection caused by biofilms

• A structure-activity relationship study was undertaken with a hit compound

• The lead compounds in the paper both inhibit and disperse Salmonella biofilms

• In combination with ciprofloxacin, they lower the bacterial burden in treated mice