2-Aminobenzothiazoles Inhibit Virulence Gene Expression and Block Polymyxin Resistance in Salmonella enterica

Abstract

One promising strategy to combat antibiotic-resistant bacteria is to develop compounds that block bacterial defenses against antibacterial conditions produced by the innate immune system. Salmonella enterica, which causes food-borne gastroenteritis and typhoid fever, requires histidine kinases (HKs) to resist innate immune defenses such as cationic antimicrobial peptides (CAMPs). Herein, we report that 2-aminobenzothiazoles block histidine kinase-dependent phenotypes in Salmonella enterica serotype Typhimurium. We found that 2-aminobenzothiazoles inhibited growth under low Mg²⁺, a stressful condition that requires histidine kinase-mediated responses, and decreased expression of the virulence genes pagC and pagK. Furthermore, we discovered that 2-aminobenzothiazoles weaken Salmonella’s resistance to polymyxin B and polymyxin E, which are last-line antibiotics and models for host defense CAMPs. These findings raise the possibilities that 2-aminobenzothiazoles can block HK-mediated bacterial defenses and can be used in combination with polymyxins to treat infections caused by Salmonella.

Graphical Abstract

For the defense: Salmonella enterica must fight antibacterial conditions encountered during an infection in order to cause food-borne illness or typhoid fever. We discovered a class of small molecules, 2-aminobenzothiazoles, that inhibit the expression of certain virulence genes and weaken Salmonella’s resistance to polymyxin antimicrobial peptides. The findings suggest that these compounds can block Salmonella resistance to host immune defenses.

Bacteria that can resist clinically used antibiotics pose an urgent threat to worldwide public health. One concerning antibiotic-resistant pathogen is the gram-negative bacterial species Salmonella enterica.^[1–3] Non-typhoidal serotypes of Salmonella cause food-borne gastroenteritis and can be transmitted from animals to humans.^[1] Typhoidal serotypes of Salmonella cause typhoid fever and include human-specific serotypes.^[2] The World Health Organization has listed Salmonella as a “high-priority” pathogen for the development of new antibiotics,^[4] and the Centers for Disease Control in the United States has deemed Salmonella as a “serious threat" to public health.^[5] A strategy to combat antibiotic resistant bacteria is to target molecular processes that are required for bacterial virulence.^[6–8] There has been to our knowledge no clinical development of compounds that target Salmonella virulence.^[8] Here, we report the discovery of small molecules that block certain virulence traits in a non-typhoidal serotype of Salmonella enterica.

A key trait for Salmonella virulence is the ability to resist the cationic antimicrobial peptides (CAMPs) that are produced by the innate immune system of the host.^[9,10] Salmonella resistance to CAMPs occurs through remodeling of the Salmonella outer membrane, which prevents CAMPs from binding to the lipopolysaccharide (LPS) and passing through the outer membrane.^[9,11] This mechanism of resistance to CAMPs is conserved throughout the family Enterobacteriaceae, which includes multiple antibiotic-resistant pathogens in addition to Salmonella.^[11] Therefore, targeting bacterial resistance to CAMPs has potential application to a wide variety of antibiotic-resistant bacteria.^[12]

LPS modifications that confer resistance to CAMPs are regulated by two-component signaling systems in Salmonella and other enterobacteria.^[9,11,13] These signaling systems comprise a histidine kinase (HK) embedded in the cytoplasmic membrane and a response regulator in the cytoplasm.^[14] Upon sensing a stressful condition in the periplasmic space, the HK then autophosphorylates a conserved histidine residue in the cytoplasmic domain of the HK from ATP. The phosphorylated HK transfers its phosphoryl group to its cognate response regulator. The phosphorylated response regulator subsequently activates the transcription of genes that are required for the response to the stressful condition. In Salmonella, genes conferring CAMP resistance are regulated by the PhoP/PhoQ and PmrA/PmrB two component systems.^[9,11,13] The PhoP/PhoQ system also activates the transcription of genes that are required for virulence of Salmonella, and Salmonella mutants lacking the PhoP/PhoQ system are avirulent.^[9,15] A variety of small molecule inhibitors of HKs, including PhoQ, have been discovered.^[16–19] Moreover, there have been very recent reports that certain classes of HK inhibitors block HK-dependent gene expression, CAMP resistance, and intramacrophage survival in Salmonella.^[19,20] These recent findings underscore the relevance of HKs as a key target for sensitizing Salmonella to CAMPs.

To identify small molecules that inhibit HK-dependent physiological processes, we focused on commercially available compounds with the 2-aminobenzothiazole scaffold for several reasons: 1) These compounds inhibit a model HK protein (HK853).^{[16, 21]} 2) These compounds inhibit multiple HK-dependent processes in Pseudomonas aeruginosa.^[21] 3) These compounds include riluzole, an FDA-approved drug for treatment of amyotrophic lateral sclerosis, indicating that the 2-aminobenzothiazole scaffold has pharmacological properties suitable for clinical use.^[22] We developed a condition-based screen to identify compounds that prevent growth in a stressful condition that requires HK-dependent responses. Specifically, we found that certain 2-aminobenzothiazoles prevent growth of Salmonella under low Mg²⁺ (Figure 1), which is a condition that requires the PhoQ HK.^[23] None of the 2-aminobenzothiazoles affected growth in high Mg²⁺ (Figure S1 in the Supporting Information), which is a condition that does not require the PhoQ HK.^[23] These results demonstrate that HK inhibitors can block the ability of Salmonella to grow in a stressful condition.

Figure 1.

2-Aminobenzothiazoles inhibit the growth of wild-type Salmonella (14028 s) under low-Mg²⁺ conditions. Compounds were present at a final concentration of 250 μM in 1 % (v/v) DMSO. Data are plotted as mean ± standard error of four independent experiments. Statistical analysis was performed by using one-way ANOVA (α = 0.05). **p<0.01, ****p<0.0001 for the difference between compound-treated sample and DMSO-treated sample.

The 2-aminobenzothiazoles we tested inhibited growth under low Mg²⁺ to various extents when assayed at the same final concentration. Compounds 1 and 2 inhibited growth to the greatest extent (Figure 1). Notably, 2 is the FDA-approved drug riluzole. Compound 3 inhibited to a moderate extent, followed by 4 and 5, which inhibited to the lowest extent (Figure 1). Compounds 6 and 7 did not inhibit growth relative to the DMSO control. Although we tested a small number of compounds, our results suggest that certain structural features are important for efficacy in this growth assay. The effectiveness of 4 but not 6 suggests that having the amino group at the 2 position of the benzothiazole ring is important for its function. The stronger effectiveness of 1–3 suggests that electron-withdrawing substituents at the 6 position are important. This chemotype has been shown previously to not cause protein aggregation,^[16] suggesting that the effects we observe against live bacterial cells are not due to denaturation of cellular proteins. Our additional work with these compounds focused on 2, due to its strong inhibition and its clinical use as a drug, 4, due to its moderate activity, and 6, as a negative control.

Because growth under low Mg²⁺ requires PhoQ-dependent gene expression, we examined the effect of 2 on the expression of several PhoP-activated virulence genes using reverse transcription-quantitative PCR (RT-qPCR). We treated wild-type Salmonella with either 2 or DMSO in PhoP-activating conditions. Bacteria treated with 2 showed seven- to 16-fold lower mRNA levels for two PhoP-activated virulence genes – pagC and pagK – when compared with the DMSO treated bacteria (Figure 2). pagK and its two homologues in the Salmonella genome are required for virulence,^[24] and pagC is required for Salmonella virulence.^[25]

Figure 2.

Compound 2 decreases the expression of certain Salmonella virulence genes. Levels of the indicated mRNAs were determined by RT-qPCR from wild-type Salmonella (14028 s) treated with either DMSO (1 % v/v) or 2 (125 μM) under low-Mg²⁺ conditions. Log₂ of fold-change values are plotted as mean ± standard error of four independent experiments. Statistical analysis was performed by using one-way ANOVA (α = 0.05). Asterisks indicate **p<0.01 for the difference between the compound-treated and DMSO-treated samples for each mRNA.

Bacteria treated with 2 did not have lower levels of mRNAs for other PhoP-activated virulence genes (Figure 2). The genes that were not affected by treatment with 2 are ones that are less sensitive to the level of phosphorylated PhoP.^[26] That is, these genes are expressed at high levels under PhoP-activating conditions even when there are low levels of phosphorylated PhoP. On the other hand, pagC and pagK require high levels of phosphorylated PhoP for full expression.^[26] Our results suggest that 2 is potent enough to knock down the expression of genes that are highly sensitive to levels of phosphorylated PhoP but not to shut down the PhoP/PhoQ signaling system completely. Indeed, wild-type Salmonella treated with 2 showed higher levels of mRNA for pagC and pagK than phoP⁻ Salmonella treated with DMSO (Figure S2), indicating that 2 did not knock down expression to the level observed in bacteria that lack PhoP. We were unable to test higher concentrations of 2 in the gene expression experiments, which used 125 μM 2, because 2 inhibits growth under low Mg²⁺ at 250 μM (Figure 1). Nevertheless, our results indicate that 2 is a promising lead for the development of an inhibitor of virulence gene expression in Salmonella.

Because HKs are required to activate genes that confer resistance to CAMPs, we postulated that 2-aminobenzothiazoles would block the ability of Salmonella to resist CAMPs. Therefore, we tested the sensitivity of wild-type Salmonella to polymyxin B and polymyxin E (colistin) in the presence of a constant concentration of 2, 4, or 6. Polymyxin B and polymyxin E are commercially available, have been used as a model system for studying CAMP resistance in Salmonella,^[27,28] and are used clinically as last-line antibiotics.^[29] We found that the minimum inhibitory concentration (MIC) of polymyxin B and polymyxin E against wild-type Salmonella was two times lower when Salmonella was treated with 2 and 4 than when treated with DMSO (Table S1 and Figure S3). When varying concentrations of 2, 4, or 6 were tested at a constant concentration of polymyxin B or polymyxin E, 2 was the most potent, with an MIC of 62.5 μM, followed by 4, with an MIC of 250 μM (Table S2). This result is consistent with the relative activities of 2 and 4 in preventing growth under low Mg²⁺ (Figure 1). Notably, 2 did not affect growth in the absence of polymyxins (Figure S3), which is consistent with its lack of effect on growth in media with high Mg²⁺ (Figure S1). Taken together, these results demonstrate that 2 does not block processes that are essential for growth.

To test whether the increased sensitivity to polymyxins is caused by the compounds disrupting the outer membrane, we examined the effect of 2 or 4 on sensitivity to vancomycin. Disruption of the outer membrane increases sensitivity to vancomycin, which targets cell-wall biosynthetic precursors in the periplasm.^[30–33] Treatment of wild-type Salmonella with 2 or 4 did not affect sensitivity to vancomycin (Figure S4 and Table S3), whereas treatment with ethylenediaminetetraacetic acid, which disrupts the outer membrane, did (Figure S4). This result suggests that 2-aminobenzothiazoles do not disrupt the permeability barrier of the outer membrane.

Because 2-aminobenzothiazoles had only a twofold effect on the MIC of polymyxins when tested against wild-type Salmonella, we next determined whether 2-aminobenzothiazoles were effective against a polymyxin-resistant strain of Salmonella. We tested the effect of 2, 4, and 6 on a strain that has high intrinsic resistance to polymyxin but still requires a functional PhoP/PhoQ TCS for full resistance to polymyxin.^[34] Strikingly, treatment of this strain with 2 decreased the MIC for polymyxin B and polymyxin E by at least 16-fold (Table 1 and Figure S5). Compound 4 was also effective against this strain, and 6 was ineffective (Table 1 and Figure S5), as seen in other assays. This result demonstrates that 2-aminobenzothiazoles are significantly effective against a polymyxin-resistant strain of Salmonella.

Table 1.

2-Aminobenzothiazoles sensitize polymyxin-resistant Salmonella enterica to polymyxins.

Compound [a]	MIC [μg/mL][b]
	Polymyxin B	Polymyxin E
2	2.0	2.0
4	16	32
6	> 32	> 32
DMSO	> 32	> 32

^[a]Compounds were present at a final concentration of 250 μM in 1 % (v/v) DMSO.

^[b]Strain EG9492. Results are from 3 independent experiments.

In summary, we have discovered that 2-aminobenzothiazoles block polymyxin resistance and inhibit the expression of certain virulence genes in Salmonella. Although we cannot rule out other mechanisms of action, the following findings suggest that 2-aminobenzothiazoles act through inhibition of HK-mediated signaling: 1) The relative effectiveness of these compounds against Salmonella corresponds with the relative IC₅₀ values for their inhibition of a model HK protein, HK853.^[21] 2) Structural alignments of the catalytic domain of HK853 with the catalytic domains of PhoQ from E. coli and S. enterica indicate root mean square deviations of <1.25 Å for Cα atoms (Figure S6 and Table S4), indicating highly similar active sites between the model HK and the targeted HK proteins. 3) These compounds inhibit multiple physiological processes that require HK-mediated signaling. 4) These compounds only block growth in conditions that require HK-mediated signaling. 5) These compounds do not disrupt the permeability of the outer membrane. Our discovery that 2-aminobenzothiazoles weaken Salmonella resistance to polymyxins suggests that these compounds might be effective when used in combination with polymyxins and raises the possibility that these compounds could block Salmonella’s resistance to innate immune defenses.