Comparative Studies to Uncover Mechanisms of Action of N-(1,3,4-oxadiazol-2-yl)benzamide Containing Antibacterial Agents

Abstract

Drug-resistant bacterial pathogens still cause high levels mortality annually despite the availability of many antibiotics. Methicillin-resistant Staphylococcus aureus (MRSA) is especially problematic and the rise in resistance to front line treatments like vancomycin and linezolid calls for new chemical modalities to treat chronic and relapsing MRSA infections. Halogenated N-(1,3,4-oxadiazol-2-yl)benzamides are interesting class of antimicrobial agents, which have been described by multiple groups to be effective against different bacterial pathogens. The modes of action of a few N-(1,3,4-oxadiazol-2-yl)benzamides have been elucidated. For example, oxadiazoles KKL-35 and MBX-4132, have been described as inhibitors of trans-translation (a ribosome rescue pathway) while HSGN-94 was shown to inhibit lipoteichoic acid. However other similarly halogenated N-(1,3,4-oxadiazol-2-yl)benzamides neither inhibit trans-translation nor lipoteichoic acid biosynthesis but are potent antimicrobial agents. For example, HSGN-220, −218, and −144 are N-(1,3,4-oxadiazol-2-yl)benzamides that are modified with OCF₃, SCF₃ or SF₅, and have remarkable minimum inhibitory concentrations (MICs) ranging from 1 μg/mL to 0.06 μg/mL against MRSA clinical isolates and show a low propensity to resistance to MRSA over 30 days. The mechanism of action of these highly potent oxadiazoles is however unknown. To provide insights into how these halogenated N-(1,3,4-oxadiazol-2-yl)benzamides inhibit bacterial growth, we performed global proteomics and RNA expression analysis of some essential genes of S. aureus treated with HSGN-220, −218, and −144. These studies revealed that the oxadiazoles HSGN-220, −218, and −144 are multi-targeting antibiotics that regulate menaquinone biosynthesis and other essential proteins like DnaX, Pol IIIC, BirA, LexA, and DnaC. In addition, these halogenated N-(1,3,4-oxadiazol-2-yl)benzamides were able to depolarize bacterial membranes and regulate siderophore biosynthesis and heme regulation. Iron starvation appears to be part of the mechanism of action that led to bacterial killing. This study demonstrates that N-(1,3,4-oxadiazol-2-yl)benzamides are indeed privileged scaffolds for the development of antibacterial agents and that subtle modifications lead to changes to mechanism of action.

Graphical Abstract

Antimicrobial resistance has become a global crisis and it has been estimated that if antimicrobial resistance is not controlled, then by 2050 ten million people each year will die as a result, surpassing annual deaths due to cancer.¹ In the United States, antimicrobial resistance is also a problem. The Centers for Disease Control and Prevention (CDC) recently stated that more than 2.8 million antibiotic-resistant infections occur in the U.S. each year, resulting in more than 35,000 deaths.² Gram-positive bacterial pathogen, Staphylococcus aureus (S. aureus) is one of the leading causes of community- and hospital- acquired bacteremia, surgical site infections, osteomyelitis, pneumonia, and skin infections.³ The CDC has designated methicillin-resistant S. aureus (MRSA) as a serious threat as it causes 323,700 infections per year resulting in 10,600 deaths.² Yet, MRSA’s continued rise in both infection and death rate has been attributed to the shortage of newly approved antibiotics with novel mechanisms of action.⁴ This lack in development of novel antibiotics is probably because only a few pharmaceutical companies have antibiotic research program; the high likelihood of resistance emerging towards any antibiotic, coupled with the fact that for most infections clinicians would likely prescribe cheaper antibiotic alternatives, makes investment in antibiotic drug development a risky one.⁵ The few and newer antibiotics that have been recently approved or are in clinical trials are derivatives of existing drugs so resistance mechanism that affect the old drugs will also likely affect the newer ones as well.⁶ The recent pandemic, while being of viral origin, has reminded us that the development of new compounds that kill infectious pathogens (of any type) with novel mechanisms of action should be elevated to high priority and that we should not wait for pandemics to occur before rushing to find new therapeutics against resistant strains that would inevitably emerge.⁷

Halogenated N-(1,3,4-oxadiazol-2-yl)benzamides have emerged as novel compounds with potent activities against several strains of bacteria.^8–16 We recently reported that the sulfonamide containing N-(1,3,4-oxadiazol-2-yl)benzamide, HSGN-94 (Figure 1), inhibits LTA biosynthesis and is a potent inhibitor of MRSA growth (MIC = 0.25 μg/mL).^15–16 Keiler and co-workers have also reported that a chloro-substituted N-(1,3,4-oxadiazol-2-yl)benzamide (KKL-35, Figure 1) was an inhibitor of trans-translation in bacteria. However, Gillet et al. have argued that trans-translation is not the only target for KKL-35 in vivo.¹⁷ While trans-translation inhibition or LTA biosynthesis inhibition might partly explain the antibacterial action of some oxadiazole-containing compounds, other halogenated N-(1,3,4-oxadiazol-2-yl)benzamides developed by the Gillet group (CT1–115, Figure 1)¹¹ and by our group (HSGN-220, −218, and −144)^12,13 do not inhibit trans-translation or LTA biosynthesis. To provide insights into how various halogenated N-(1,3,4-oxadiazol-2-yl)benzamides inhibit S. aureus, we performed various mechanistic studies, including comparative global proteomics, membrane depolarization and membrane permeation assays, to identify pathways that are impacted by the oxadiazole-containing compounds and to investigate how subtle modifications affect the biological activities of these interesting antibacterial agents.

Figure 1.

Structures of previously reported N-(1,3,4-oxadiazol-2-yl)benzamides as potent antibacterial agents against Gram-positive bacteria.

Results and Discussion

HSGN-220, −218, and −144 demonstrate potent activity against MRSA clinical isolates.

To evaluate HSGN-220, −218, and −144 potential as therapeutics against drug-resistant bacteria, we proceeded to evaluate their antibacterial profile against a panel of MRSA clinical isolates. HSGN-218 was previously reported to have highly potent activities against C. difficile but its activity against MRSA was not explored.¹² However, our analysis demonstrates that HSGN-218 has remarkable activity against MRSA clinical isolates with MICs ranging from 0.06 μg/mL (0.1 μM) to 0.25 μg/mL (0.6 μM) (See supplementary information (SI) for Table S1). Additionally, HSGN-144 was reported to have activity against MRSA as well as other drug-resistant Gram-positive strains.¹³ Here, the compound also demonstrates high potency against MRSA clinical isolates with MICs of 0.5 μg/mL (1.3 μM) (Table S1). Furthermore, HSGN-220 showed potent antibacterial activity against MRSA clinical isolates as well with MICs ranging from 0.25 μg/mL (0.6 μM) to μg/mL 1 (2.4 μM) (Table S1). Overall, the compounds performed similarly or better than vancomycin (especially HSGN-218) but significantly outperformed linezolid against MRSA clinical isolates.

HSGN-220, −218, and −144 have a low propensity to develop resistance to MRSA:

Based on the impressive activity of HSGN-220, −218, and −144 against MRSA, our next step was to determine their mechanism of action against S. aureus. A classical method to do this involves obtaining resistant mutants via serial passaging and then using genomic mapping to identify mutations in the target protein(s).¹⁸ Therefore, we performed the multi-step resistance selection to generate MRSA resistant-mutants for HSGN-220, −218, and −144. However, we failed to generate any resistant mutants after 30 days, indicating that these compounds have a low propensity to develop resistance to MRSA (see SI for Figure S1). In contrast, under identical experimental conditions, the MIC of ciprofloxacin, an antibiotic that targets DNA gyrase, increased 128-fold after the 30 days, agreeing with previous reports.^19–21

Effects of HSGN-220, −218, and −144 on global proteomics in S. aureus:

Since we could not obtain HSGN-220, −218, or −144 resistant-mutants, we decided to use global proteomics with the aim of assessing the proteins and pathways that are affected by treatment with these compounds. Thus, we treated S. aureus with either HSGN-220, −218, or −144 for 2 hours and extracted the total protein for profiling using mass spectrometry. HSGN-220, −218, or −144 treated samples were compared with samples treated with DMSO only (untreated). For each compound, a Venn diagram illustrating proteins found either in the untreated or treated samples is shown in Figure S2. A total of 1,000 proteins were identified after S. aureus treatment with HSGN-220. Of these 1,000 proteins, 781 proteins (78.1%) were observed to be shared by both DMSO control and HSGN-220 whereas 157 proteins (15.7%) were only identified in DMSO control, and 62 proteins (6.2%) were only identified in HSGN-220 (Figure S2A). Additionally, a total of 982 proteins were identified after S. aureus treatment with HSGN-218. Of these 982 proteins, 736 proteins (74.9%) were observed to be shared by both DMSO control and HSGN-218 while 202 proteins (20.6%) were only identified in DMSO control, and 44 proteins (4.5%) were only identified in HSGN-218 (Figure S2B). Furthermore, a total of 1,018 proteins were identified after S. aureus treatment with HSGN-144. Of these 1,018 proteins, 825 proteins (81.0%) were observed to be shared by both DMSO control and HSGN-144 whereas 113 proteins (11.1%) were only identified in DMSO control, and 80 proteins (7.9%) were only identified in HSGN-144 (Figure S2C). Heatmap analysis depicts how HSGN-220, −218, and −144 differentially regulate S. aureus gene expression (Figure 2A).

Figure 2.

Global proteomics analysis of HSGN-220, −218, and −144. (A) Heatmap evaluation of global proteomics data displaying differentially expressed proteins between DMSO- and HSGN-220, −144, or −218-treated S. aureus. (B) Venn diagram for comparison of downregulated proteins identified individually, dually, or amongst all three compounds: HSGN-220, −218, and −144. (C) Venn diagram for comparison of upregulated proteins identified individually, dually, or amongst all three compounds: HSGN-220, −218, and −144.

The proteomics data were stringently filtered using label-free quantitation (LFQ) intensities and MS/MS counts. Proteins in samples that had LFQs for all three replicates and with at least 2 MS/MS counts were included for further analysis. Proteins found in only treated samples were considered to be highly upregulated while those found in only DMSO control were considered to be highly downregulated by compound.

Next, we then sought to compare proteins downregulated by HSGN-220, −218, and −144 via generation of a Venn diagram (Figure 2B). After stringent filtration, 249 proteins were identified in DMSO control only when compared to samples treated with HSGN-220, −218, and −144. Of the 249 proteins, 81 (32.5%) were observed to be shared among all three compounds. Individually, 11 proteins (4.4%) were identified in HSGN-144, 77 proteins (30.9%) were detected in HSGN-218, and 19 proteins (7.6%) were found in HSGN-220. Furthermore, 40 proteins (16.1%) were shared between HSGN-218 and HSGN-220, while 17 proteins (6.8%) were detected in both HSGN-144 and HSGN-220. Only 4 proteins (1.6%) were detected in both HSGN-218 and HSGN-144.

During our assessment, we first began by analyzing the 81 proteins that were downregulated by all three compounds (see SI for Table S2). Specifically, we were interested in identifying proteins that were essential to S. aureus as these are the most important for the bacteria’s growth. We did this by utilizing the comprehensive list of S. aureus essential genes developed by Charles et al.²² We found that all three compounds had a significant impact on DNA synthesis in S. aureus. For instance, all three compounds downregulated DNA polymerase III subunit τ/γ (DnaX), DNA polymerase III subunit α (Pol IIIC), and the replicative DNA helicase (DnaC) (Table 1). Both DnaX and Pol IIIC are required for replicative DNA synthesis while DnaC participates in initiation and elongation during chromosome replication.^23–24 Additionally, we found that HSGN-220, −218, and −144 also downregulate the essential proteins BirA and LexA (Table 1). BirA is a Group II biotin protein ligase (BPL) which is essential for the catalytic attachment of biotin to biotin-dependent enzymes, resulting in the synthesis of important lipids that make up the cell wall.²⁵ Furthermore, LexA is vital for S. aureus to produce the SOS response which modifies transcription in response to environmental stress.²⁶ Also, this SOS response mediated by LexA has been found to play an important role in antibiotic resistance and persistence of S. aureus infections.²⁶

Table 1.

Select essential proteins that were identified to be downregulated by HSGN-220, −218, and −144.

ID	Protein	Description	Essential in S. aureus?
gi|685631628	DnaX	DNA polymerase III τ/γ	Yes
gi|685632359	Pol IIIC	DNA polymerase III α	Yes
gi|685632567	BirA	biotin--acetyl-CoA-carboxylase ligase	Yes
gi|685632450	LexA	XRE family transcriptional regulator	Yes
gi|685631229	DnaC	Replicative DNA helicase	Yes

All three compounds downregulated DNA Pol IIIC, which has been viewed as a viable new target to combat Gram-positive infections.²⁷ Pol IIIC is vital for replication of the bacterial chromosome as it makes up a major portion of the DNA Pol III core (see Figure 3A). Additionally, Pol IIIC also interacts with other essential proteins necessary for DNA replication like DnaC and DnaX (Figure 3B), meaning if Pol IIIC is affected, these other proteins will also be altered, and DNA replication will not take place. Furthermore, Pol IIIC is a highly conserved enzyme and is exclusive to bacteria whose genomes contain less than 50% guanine and cytosine such as: Streptococcus, Enterococcus, Staphylococcus, Bacillus, Clostridioides, Pneumococcus, Listeria, and Lactobacillus.²⁸ Pol IIIC is not found in Gram-negative bacteria and has little homology with mammalian DNA polymerase, making it specific for Gram-positive pathogens.²⁷ Because of the uniqueness of Pol IIIC as well as its ability to be a novel drug target to combat Gram-positive bacteria, inhibitors of this enzyme have been developed. For instance, the earliest known inhibitors of Pol IIIC contained the anilino-uracil moiety which showed potent binding to Pol IIIC in Bacillus subtilis (B. subtilis) but only had moderate activity against Gram-positive bacteria with MICs between 20 μg/mL – 40 μg/mL.^29–30 Ibezapolstat (ACX-362E), which selectively inhibits C. difficile Pol IIIC is currently in Phase III clinical trials for C. difficile infection.³¹ The downregulation of DNA Pol IIIC as well as other proteins essential for DNA replication by HSGN-220, −218, and −144 could explain the potency of these compounds against MRSA.

Figure 3.

(A) DNA replication in S. aureus. The helicase (DnaC) separates the double-stranded DNA into two single strands. The leading strand is synthesized continuously by the DNA Pol III core, while the lagging strand is synthesized in smaller fragments. Highlighted proteins were found to be downregulated by HSGN-220, −144, and −218. (B) Predicted functional protein-protein association networks for Pol IIIC, DnaX, and DnaC. All three proteins interact with one another due to their importance in DNA replication. Line meanings are as follows: gene neighborhood (green lines), gene fusions (red lines), gene co-occurrences (blue lines), textmining evidence (yellow lines), and co-expression evidence (black lines). The STRING v 10.5 software was utilized to construct the figure.³²

To investigate if the observed modulations of the aforementioned proteins occurred at the protein or mRNA level, we performed RT-qPCR analysis for dnaX, pol IIIC, birA, lexA, and dnaC (Figure 4). In line with the observations from the global proteomics analysis, we observed decreased dnaX, pol IIIC, birA, lexA, and dnaC mRNA levels in S. aureus treated with either HSGN-220, −144, or −218 (Figure 4A). These data suggest that HSGN-220, −144, or −218 modulates the target mRNA expression or stability, which leads to differential protein abundance.

Figure 4.

The effect of HSGN-220, −144, or −218 treatment on the transcription of: (A) Effect of HSGN-220, −144, or −218 treatment on transcription of dnaX, pol IIIC, birA, lexA, and dnaC. (B) Effect of HSGN-220 or −218 treatment on the transcription of menA and Pth. (C) The effect of HSGN-220 or −144 treatment on the transcription of mvak1. (D) The effect of HSGN-218 treatment on the transcription of relA and pgsA. (E) The effect of HSGN-220, −144, or −218 treatment on the transcription of HarA and IsdA. Total RNA isolated from S. aureus treated with either DMSO or 1X MIC HSGN-220 (0.5 μg/mL), −144 (0.5 μg/mL), or −218 (0.06 μg/mL) was reversed transcribed and cDNAs were quantified by RT-qPCR using target-specific primers. The data represents the mean ± SD of triplicate experiments normalized with 16S RNA. Statistically significant differences between DMSO-treatment and HSGN-220, −144, or −218 -treatment was determined by Student’s t-test analysis (unpaired, two-tailed) and is represented as *p ≤ 0.05 or **p ≤ 0.01.

After analyzing the 81 proteins that all three compounds commonly modulated, we then went on to evaluate essential proteins downregulated by only two compounds. For example, both HSGN-218 and HSGN-220 have 40 proteins in common (Figure 2B and Table S3). Yet, both compounds downregulate the essential proteins 1,4-dihydroxy-2-naphthoate octaprenyltransferase (MenA) and peptidyl-tRNA hydrolase (Pth) (Table 2). MenA is important in S. aureus for the synthesis of menaquinone.³³ Menaquinone plays an important role in electron transport and ATP generation in Gram-positive and anaerobically respiring Gram-negative bacteria.^34–36 Its been shown that inhibition of MenA, leading to loss of menaquinone, causes cell death, thereby demonstrating the protein’s potential as a possible drug target.^{34, 36} Pth is an essential enzyme in S. aureus as it recycles peptidyl-tRNAs which arise from untimely termination of translation.³⁷ Pth is found in both Gram-positive and Gram-negative bacteria, and inhibition of this protein leads to the buildup of tRNAs that are toxic to the cell due to impairment of protein synthesis.^38–39 To confirm our global proteomics results, we conducted RT-qPCR analysis for menA and Pth expression (Figure 4B). We detected a decrease of menA and Pth mRNA levels in S. aureus treated with either HSGN-220 or −218 (Figure 4B).

Table 2.

Select essential proteins that were identified to be downregulated by both HSGN-220 and HSGN-218 treatments, as well as with HSGN-144 and HSGN-220 treatment.

Essential Proteins Downregulated by HSGN-218 and HSGN-220
ID	Protein	Description	Essential in S. aureus?
gi|685632127	MenA	1,4-dihydroxy-2-naphthoate octaprenyltransferase	Yes
gi|685631655	Pth	peptidyl-tRNA hydrolase	Yes
Essential Protein Downregulated by HSGN-144 and HSGN-220
ID	Protein	Description	Essential in S. aureus?
gi|685631744	mvak1	mevalonate kinase	Yes

Since HSGN-218 and −220 were found to downregulate menA expression (see Figure 4B and Table 2), we hypothesized that treatment of S. aureus with either HSGN-218 or −220 would result in a decrease in menaquinone concentration. Both menaquinone 7 (MK-7) and menaquinone 8 (MK-8) have been found to be the two most abundant menaquinones in S. aureus.^40–44 Therefore, to analyze the impact of HSGN-218 and −220 on menaquinone levels in S. aureus growing cells, the concentration of MK-7 and −8 by treatment of S. aureus with 1× MIC of HSGN-218 or −220 for 2 hours was quantified by LC/MS/MS using commercially available MK-9 as a standard (Figure 5). Interestingly, treatment of S. aureus with either HSGN-218 or −220 resulted in a drastic decrease in both MK-7 and MK-8 levels. For instance, S. aureus treated with 1% DMSO showed an average concentration for MK-7 and MK-8 equal to 407 μg/μL and 1474 μg/μL respectively. Yet, treatment of S. aureus with 1× MIC HSGN-218 demonstrated a reduction in the concentration of MK-7 (38 μg/μL) and MK-8 (105 μg/μL) (see Figure 5). Similarly, treatment of S. aureus with 1× MIC HSGN-220 also showed a reduction in MK-7 and MK-8 concentrations equal to 36 μg/μL and 79 μg/μL respectively (see Figure 5). Therefore, we hypothesize that both HSGN-218’s and −220’s downregulation of MenA is responsible for the compounds’ effect on menaquinone biosynthesis.

Figure 5.

Effect of HSGN-218 and −220 on menaquinone biosynthesis. (A) MK-7 concentration (μg/μL) after treatment of S. aureus with 1% DMSO, 0.06 μg/mL (1X MIC) HSGN-218, and 0.5 μg/mL (1X MIC) HSGN-220. (B) MK-8 concentration (μg/μL) after treatment of S. aureus with 1% DMSO, 0.06 μg/mL (1× MIC) HSGN-218, and 0.5 μg/mL (1×MIC) HSGN-220. Statistically significant differences between DMSO-treatment and HSGN-218 or −220 -treatment was determined by Student’s t-test analysis (unpaired, two-tailed) and is represented as **p ≤ 0.01.

Both HSGN-144 and HSGN-220 downregulate 17 similar proteins (Figure 2B and Table S4). However, both compounds only downregulate one essential enzyme, mevalonate kinase (mvak1) (Table 2). Mvak1 is essential for isoprenoid biosynthesis in S. aureus since it converts coenzyme A to isopentenyl diphosphate.⁴⁵ Although humans and plants also contain mevalonate kinase, the bacterial mvak1 is much different, making it an ideal drug target.⁴⁵ Additionally, HSGN-144 and HSGN-218 both share 4 proteins in common (Figure 2B and Table S5), yet none are essential to S. aureus and will not be discussed in this manuscript. To verify that HSGN-220 and −144 both downregulate mvak1, we did RT-qPCR analysis for mvak1 expression. The results confirmed our proteomics analysis since we witnessed a decrease of mvak1 mRNA levels in S. aureus treated with either HSGN-220 or −144 (Figure 4C).

Next, we moved to assess essential proteins downregulated by the compounds individually. On its own, HSGN-218 downregulates 77 proteins (Figure 2B and Table S6), yet only 2 are essential in S. aureus: GTP pyrophosphokinase (RelA) and CDP-diacylglycerol--glycerol-3-phosphate 3-phosphatidyltransferase (PgsA) (Table 3). RelA controls the activation of guanosine 3′, 5′-bis(diphosphate) when bacteria are under stressful conditions making it essential for bacterial survival.^46–47 Moreover, PgsA is highly important for the synthesis of phosphatidyl glycerol (PG), the most abundant membrane phospholipid in bacteria.⁴⁸ Reduction of PgsA causes a decrease in PG production and subsequent surface charge alterations, leading to bacterial cell death.⁴⁹ To validate that HSGN-218 does indeed downregulate RelA and PgsA, we performed RT-qPCR. Our results demonstrated that S. aureus treated with HSGN-218 showed a decrease in mRNA expression levels of both relA and pgsA, thereby confirming our global proteomics analysis (Figure 4D). Furthermore, both HSGN-144 and HSGN-220 independently downregulate 11 and 19 proteins in S. aureus, respectively, however none were considered to be essential to the organism (see Table S7 and Table S8).

Table 3.

Select essential proteins that were identified to be downregulated individually by HSGN-218 treatment.

Essential Proteins Downregulated by HSGN-218
ID	Protein	Description	Essential in S. aureus?
gi|685632815	RelA	GTP pyrophosphokinase	Yes
gi|685632378	PgsA	CDP-diacylglycerol--glycerol-3-phosphate 3-phosphatidyltransferase	Yes

After examining downregulated proteins, we then wanted to compare proteins upregulated by HSGN-220, −218, and −144 via generation of another Venn diagram (Figure 2C). After stringent filtration, 108 proteins were identified in HSGN-220, −218, and −144 treated samples only and not in DMSO control. Of the 108 proteins compared, 18 (16.7%) were observed to be shared among all three compounds. Individually, 28 proteins (25.9%) were identified in HSGN-144, 14 proteins (13%) were detected in HSGN-218, and 6 proteins (5.6%) were found in HSGN-220. Furthermore, 8 proteins (7.4%) were shared between HSGN-218 and HSGN-220, while 30 proteins (27.8%) were discovered in both HSGN-144 and HSGN-220. Only 4 proteins (3.7%) were detected in both HSGN-218 and HSGN-144.

HSGN-220, −218, and −144 appears to have a significant effect on iron transportation in S. aureus. Particularly, it appears that treatment of S. aureus with our N-(1,3,4-oxadiazol-2-yl)benzamides increases heme transport. For instance, proteins involved in heme-mediated iron acquisition like IsdA, HarA, HrtA, and SirA are all upregulated by HSGN-220, −218, and −144 (Table 4). IsdA and HarA are part of the S. aureus Isd system inside the cell wall, which mediates iron acquisition.⁵⁰ Heme binds to HarA and is passed across the cell wall through the near iron transporter (NEAT) domains of IsdA and IsdC which then results in the import of heme into the cytoplasm, followed by heme degradation to release free iron to satisfy nutrient needs.⁵⁰ HrtA is part of the heme-regulated transport (Hrt) system which regulates the amount of heme being transported into the cell to avoid any toxicity.⁵¹ SirA also controls the amount of free iron throughout the bacterial cell by importing S. aureus siderophores staphyloferrin A and staphyloferrin B.^52–53

Table 4.

Select proteins that were identified to be upregulated by HSGN-220, −218, and −144.

ID	Protein	Description
gi|685632908	HarA	Haptoglobin-binding heme uptake protein
gi|685632217	IsdA	Iron-regulated surface determinant protein A
gi|685631802	SirA	Siderophore compound ABC transporter binding protein
gi|685632219	HrtA	Heme ABC transporter permease

To demonstrate that HSGN-220, −218, and −144 upregulate heme transport in S. aureus, we completed RT-qPCR for HarA and IsdA. We witnessed an increase in mRNA levels for both HarA and IsdA in S. aureus treated with either HSGN-220, −218, or −144 (Figure 4E), further verifying our global proteomics results.

HSGN-220, −218, and −144 cause iron starvation:

The beautiful work by Huigens and co-workers showed that phenazine antibiotics bind directly to iron.^54–55 Inspired by this work and the fact that HSGN-220, −218, or −144 also regulate proteins involved in iron transport, we evaluated whether HSGN-220, −218, or −144 directly bind to iron (II) by measuring UV-Vis absorption of 10 mM solutions of the corresponding compounds in dimethyl sulfoxide from 200 nm to 800 nm (Figure 6A–C), following Huigens’ protocol.⁵⁴ We found that all three N-(1,3,4-oxadiazol-2-yl)benzamides chelate to iron (II) and the chelation intensifies over a 3 hour period.

Figure 6.

HSGN-220, −218, and −144 Chelate to Iron (II). (A) UV–vis spectroscopy of HSGN-218 binding iron (II). (B) UV–vis spectroscopy of HSGN-144 binding iron (II). (C) UV–vis spectroscopy of HSGN-220 binding iron (II).

Additionally, because treatment of S. aureus with HSGN-220, −218, or −144 appeared to upregulate SirA and HrtA, two proteins important for siderophore biosynthesis and heme transport respectively, we further evaluated the effect of HSGN-220, −218, and −144 on siderophore biosynthesis and heme regulation in S. aureus using plate bioassays^56–57 (see Table 5). For these bioassays, we used iron-restricted Tris-minimal succinate (TMS) agar by adding 25 μM of the iron-chelator ethylenediamine-N,N′-bis(2-hydroxyphenylacetic acid) (EDDHA). 1% DMSO (untreated control) or HSGN-220, −218, or −144 at concentrations of 0.25× MIC, 0.5× MIC, 1× MIC was supplemented into the TMS agar. S. aureus was incorporated into plates and 50 μM of siderophore (ferrichrome, defersal, and 2,3-dihydroxy benzoic acid (DHBA)), hemoglobin, or hemin was added onto sterile paper disks and incubated at 37°C for 24 hours. Interestingly, we found that S. aureus treated with 0.5X MIC and 1X MIC of HSGN-220, −218, or −144 showed the inability to utilize siderophores, hemoglobin, or hemin to transport iron and promote growth (see Table 5 and SI). However, in the absence of compound, S. aureus was able to use siderophores, hemoglobin, or hemin to promote growth (see Table 5 and SI). Therefore, we hypothesize that HSGN-220, −218, and −144’s effects on both siderophore biosynthesis and heme regulation, which results in up-regulation of SirA and HrtA, starves S. aureus of iron and causes bacterial death.

Table 5.

Utilization of siderophores (ferrichrome, defersal, and 2,3-DHBA), hemoglobin, or hemin by S. aureus in the presence and absence of HSGN-220, −218, and −144.

Siderophore	Growth Promotion in:
	DMSO	HSGN-220			HSGN-218			HSGN-144
		0.25× MIC	0.5× MIC	1× MIC	0.25× MIC	0.5× MIC	1× MIC	0.25× MIC	0.5× MIC	1× MIC
Ferrichrome	+	−	−	−	+	−	−	−	−	−
Hemoglobin	+	−	−	−	+	−	−	−	−	−
Hemin	+	−	−	−	+	−	−	+	−	−
Desferal	+	−	−	−	+	−	−	+	−	−
2,3-DHBA	+	−	−	−	+	−	−	−	−	−

Note: +, growth; −, no growth.

Effects of HSGN-220, −218, and −144 on membrane depolarization and permeability in S. aureus:

We recently reported that the SCF₃ or SF₅ moiety enhances association with bacterial membranes.⁵⁸ Thus, we wondered if HSGN-220, −218, and −144 were membrane-targeting agents. Targeting the bacterial membrane has been deemed a potential drug target because it can disrupt the membrane’s function and/or physical integrity.⁵⁹ The bacterial membrane is essential because it comprises about one third of the proteins in a cell and is the site for highly important processes like respiration, active transport of nutrients and wastes, and the formation of the proton motive force.⁶⁰ Additionally, the bacterial cell membrane contains an electrical potential difference which acts as a source of free energy.⁶¹ This energy allows for the bacteria to undergo its essential functions.⁶¹ For instance, membrane potential has been shown to regulate pH homeostasis,^62–63 membrane transport,⁶⁴ motility,⁶⁵ antibiotic resistance,⁶⁶ cell division,⁶⁷ and environmental sensing.⁶⁸ Thus, we began by investigating HSGN-220, −218, and −144’s effect on membrane depolarization in S. aureus using a fluorescent based assay with DiSC3(5) as a stain. DiSC3(5) is a cationic membrane-permeable dye which accumulates in polarized cells resulting in quenching of overall fluorescence.⁶⁹ Upon depolarization, DiSC3(5) is rapidly released into the medium resulting in dequenching. This dequenching triggers a large spike fluorescence intensity.⁶⁹ All three compounds showed potent membrane depolarization activity. For instance, at 10× MIC HSGN-220, −218, and −144 showed either equal or greater depolarization when compared to daptomycin (5 μg/mL), a known depolarizer of S. aureus membranes⁷⁰ (Figure 7).

Figure 7.

Effects of HSGN-220, −218, and −220 on membrane depolarization in S. aureus at 10× MIC concentrations using DiSC3(5) dye. Increase in fluorescence indicates depolarization. Daptomycin is used as positive control while 1% DMSO is used as negative control.

Furthermore, disruption of the membrane integrity can lead to leakage of cytosolic content and harmful pleiotropic effects, eventually causing cell death.⁵⁹ Therefore, we proceeded to determine the effects of HSGN-220, −218, and −144 on membrane permeability in S. aureus via a fluorescent-based assay using sytox green as a dye. Sytox green stain was used as it is a high-affinity nucleic acid stain that does not cross the membranes of live cells but easily penetrates cells with compromised membranes, resulting in strong fluorescence.⁷¹ None of the three compounds caused membrane permeability in S. aureus cells at 10× MIC concentrations as their fluorescence was comparative to the negative control (1% DMSO) (see SI for Figure S3). However, bithionol, a known antibacterial that affects membrane integrity,⁷² had a significant influence on membrane permeability (Figure S3).

Conclusion

In conclusion, HSGN-220, −218, and −144 are potent antibacterial agents with MICs greater than or equal to those of vancomycin against MRSA clinical isolates. Additionally, HSGN-220, −218, and −144 did not develop resistance to MRSA over 30 days. Global proteomic analysis demonstrates that HSGN-220, −218, and −144 downregulated essential proteins involved in DNA replication. In particular, the compounds downregulated DNA Pol IIIC, a novel target to combat Gram-positive bacteria. Furthermore, HSGN-218 and −220 downregulated the essential protein MenA and significantly decreased the concentration of menaquinones MK-7 and MK-8 in S. aureus. The reductions in protein levels were due to lowering of the respective mRNA levels. While the levels of some mRNA were reduced, others increased, which indicate that the effects of the compounds are not due to promiscuous reduction in all mRNAs but rather selective transcriptional control and/or mRNA stability. Additionally, HSGN-220, −218, and −144 starve bacteria of iron which we hypothesize is due to their upregulation of proteins involved in heme acquisition and siderophore biosynthesis. We also found that HSGN-220, −218, and −144 had a significant effect on membrane depolarization but did not alter membrane permeability. Therefore, despite their common core, HSGN-220, −218, and −144 showed several similarities and differences in the pathways they affected, demonstrating that subtle changes in a compound’s structure can impact their mechanism of action. N-(1,3,4-oxadiazol-2-yl)benzamides are interesting new antimicrobials, but it appears that their modes of action are more complicated that initially thought.

Compounds that act on multiple pathways in bacteria have a better chance of clinical utility without bacterial resistance. While the compounds described herein also inhibit multiple pathways in bacteria and recent studies have documented that indeed many approved antibiotics also inhibit multiple pathways,⁷³ medicinal chemists are still not at the stage whereby compounds that are multi-targeting without gross toxicity can be developed a priori. Thus far, such compounds are found after the fact.⁷⁴ We hope that the studies described herein adds to the database, which would be needed as training set for future modeling for nodes that when targeted lead to potent antimicrobials that resist bacterial resistance.

Materials and Methods:

Bacterial strains, media, and reagents

Bacterial strains were obtained from various sources as listed in Table S10 and S11. Cation-adjusted Mueller Hinton broth, tryptic soy broth (TSB) and tryptic soy agar (TSA) were purchased from Fisher Scientific (Waltham, MA, USA). Linezolid (Chem-Impex International, Wood Dale, IL, USA), vancomycin hydrochloride (Gold Biotechnology, St. Louis, MO, USA), daptomycin (AK Scientific, Union City, CA, USA), ciprofloxacin (Sigma Aldrich, St. Louis, MO, USA), menaquinone-9 (Cayman Chemical, Ann Arbor, MI, USA), deferoxamine mesylate (Cayman Chemical, Ann Arbor, MI, USA), 2,3-Dihydroxybenzoic acid (Sigma Aldrich, St. Louis, MO, USA), ferrichrome (Sigma Aldrich, St. Louis, MO, USA), hemin (Sigma Aldrich, St. Louis, MO, USA), hemoglobin (Sigma Aldrich, St. Louis, MO, USA), ammonium iron(II) sulfate hexahydrate (Sigma Aldrich, St. Louis, MO, USA), and EDDHA (Arctom Chemical, Westlake Village, CA, USA) were purchased commercially. Tris-minimal succinate (TMS) was prepared as previously described.⁷⁵ Compounds were previously synthesized in our laboratory and used as DMSO stock solutions.

Determination of the MICs against clinically important Gram-positive bacteria

The MICs were determined using the broth microdilution method following the guidelines of the Clinical and Laboratory Standards Institute (CLSI).⁷⁶

Multi-step Resistance Selection:

Resistance formation of HSGN-220, −218, or −144 to MRSA USA300 was performed using the multi-step serial-passaging experiment as described previously.^{14, 77–78} Resistance was considered as a greater than four-fold increase in the MIC as compared to the initial MIC.

Proteomics Analysis:

Exponentially growing S. aureus ATCC 25923 was treated 0.5 μg/mL HSGN-220, 0.0625 μg/mL HSGN-218, 0.5 μg/mL HSGN-144, or an equivalent amount of DMSO for 2 hours. After, the cells were pelleted by centrifugation and washed twice with PBS. Sample preparation for LC/MS/MS and data acquisition was performed as described previously.⁷⁹ Briefly, extraction of proteins from samples was performed by homogenization of the lysate in 100 mM ammonium bicarbonate buffer using a Barocycler NEP2320 (Pressure Biosciences, South Easton, MA) system at 5 °C for 90 cycles. The Bicinchoninic Acid (BCA) assay was used to measure protein concentration and we precipitated 50 μg of lysate by using four volumes of 100% acetone at −20°C overnight. Next, we centrifuged the samples and the pellets formed were re-suspended in 8M urea with 10 mM dithiothreitol (DTT) added. This sample was incubated at 55 °C for 45 min and then underwent cysteine alkylation with 20 mM iodoacetamide at room temperature in the dark for 45 min and an additional 5 mM DTT for 20 min at 37 °C. Reduced and alkylated samples were diluted to bring urea concentrataion down to 1.5 M before digestion overnight at 37°C using Trypsin/Lys-C Mix (Promega, Madison, WI, USA) at 1:25 (w/w) enzyme-protein ratio. Digested peptides were passed through C18 silica micro spin columns (The Nest Group Inc., Southborough, MA, USA) for peptides desaltation. We used the BCA assay as above to determine the peptide concentration and adjusted it to 0.2 μg/μL. 5 μL (~1 μg) was injected for LC-MS/MS analysis. Data acquisition was achieved using a reverse-phase HPLC-ESI-MS/MS system composed of an UltiMate^™ 3000 RSLCnano system attached to a Q-Exactive (QE) High Field (HF) Hybrid Quadrupole-Orbitrap^™ mass spectrometer (Thermo Fisher Scientific, Waltham, MA) and a Nano-spray Flex^™ ion source (Thermo Fisher Scientific) was utilized to evaluate the samples standard data-dependent mode. MS/MS scans were taken at a resolution of 15,000 at m/z 200. To prevent repetitive scanning of duplicate peptides the dynamic exclusion was set at 30 s. Data was analyzed using the MaxQuant software (v. 1.6.0.16),^80–82 while bioinformatics analysis was done using the Perseus software,⁸³ as outlined previously.⁷⁹

Total RNA isolation and RT-PCR:

RNA isolation and RT-PCR was peformed following a previously reported procedure.⁷⁹ Briefly, exponentially growing S. aureus ATCC 25923 was incubated with 0.5 μg/mL HSGN-220, 0.0625 μg/mL HSGN-218, 0.5 μg/mL HSGN-144 or 1% DMSO for 2 hours at 37 °C in triplicates. The cells were then pelleted by centrifugation. RNA isolation, cDNA synthesis, and RT-PCR analysis was performed following a previously reported procedure.⁷⁹ A BioRad CFX96^™ Touch Real-Time PCR Detection System was used. PCR primers were either designed using Primer-BLAST or obtained from the referenced literature (Table S12). The data were normalized against 16S rRNA, as an internal control, and the P-values from student’s t-test showed * ≤ 0.05 or ** ≤ 0.01.

Quantification of MK Levels

Quantification of MK levels was performed as previously outlined.⁸⁴ Briefly, exponentially growing S. aureus ATCC 25923 was treated with 0.5 μg/mL HSGN-220, 0.0625 μg/mL HSGN-218, or an equivalent amount of DMSO for 3 hours at 37 °C. Samples were then normalized to OD₆₀₀ 0.8 using sterile saline. Next, the cells were pelleted via centrifugation at 5,000 rpm for 10 minutes and washed twice with phosphate buffered saline (PBS). Then, the samples were suspended in 0.5 mL of distilled water by vortexing for 30 s. Proteins were denatured by adding 0.75 mL of 2-propanol:hexane (3:2) and vortexing for 3 minutes. Next, the mixture was centrifuged at 4 °C and 1800 g for 5 min, the upper hexane layer was collected and concentrated under nitrogen stream. 200 μL methanol:methylene chloride (3:1) was added to the concentrated sample and then a 5 μL aliquot was injected into an Agilent 1200 HPLC (Agilent Technologies)–AB SCIEX Triple Quad 5500 mass spectrometry system for identification and quantification of MKs. MK-7 and MK-8 were detected as m/z 650 (MK-7) and m/z 718 (MK-8). MK-9 was used as a standard for generating a calibration curve as previously reported.⁴¹

UV–Vis for HSGN-220, −218, and −144 Binding Iron (II):

HSGN-220, −218, or −144–iron(II) complex formation was determined using UV–vis spectrometry as previously reported.⁵⁴ Briefly, ammonium iron (II) sulfate hexahydrate (0.5 eq) was added to a stirring solution of HSGN-220, −218, or −144 (10 mM) in dimethyl sulfoxide. 50 μL aliquots of HSGN-220, −218, or −144 were added to 1 mL of dimethyl sulfoxide in a cuvette. Spectral scanning was performed from 200 to 800 nm in 2 nm increments over a 180 min period.

Plate Bioassays:

Plate bioassays were performed described previously.⁵⁷ S. aureus was streaked onto molten TMS agar containing 25 μM EDDHA as an iron-chelating agent. 0.25× to 1× MIC concentrations of HSGN-220, −218, or −144 were added to agar. Iron sources to be tested (10 μl of a 50 μM solution) were added to sterile 6-mm-diameter paper disks and placed on the surfaces of the plates. Growth promotion was determined after 24 hours of incubation.

Bacterial membrane depolarization assay:

Membrane depolarization activity of HSGN-220, −218, and −144 at 10× MIC was assayed in triplicate following a previously reported procedure⁵⁸ using DiSC3(5) (3,3’-Dipropylthiadicarbocyanine Iodide) (Fisher Scientific, Waltham, MA) as the fluorometric dye. Daptomycin at 10× MIC was used as a positive control. 1% DMSO and DiSC3(5) acted as the negative control.

Bacterial membrane permeability assay:

Membrane permeability activity of HSGN-220, −218, and −144 at 10× MIC was assayed in triplicate following a previously reported procedure⁵⁸ using Sytox green (Invitrogen, Waltham, MA) as the fluorometric dye. Bithionol (4 μg/mL) was used as the positive control.⁷² 1% DMSO and Sytox green served as the negative control.

Data Availability:

All raw LC-MS/MS data can be found in the Mass Spectrometry Interactive Virtual Environment (http://massive.ucsd.edu) with the ID: MSV000088344.