Abstract
Antibiotic-resistant infections present significant challenges to patients. As a result, there is considerable need for new antibacterial therapies that eradicate pathogenic bacteria through non-conventional mechanisms. Our group has identified a series of halogenated phenazine (HP) agents that induce rapid iron starvation that leads to potent killing of methicillin-resistant Staphylococcus aureus biofilms. Here, we report the design, chemical synthesis and microbiological assessment of a HP-quinone ether prodrug model aimed to (1) eliminate general (off-target) iron chelation, and (2) release an active HP agent through the bioreduction of a quinone trigger. Here, we demonstrate prodrug analogue HP-29-Q to have a stable ether linkage that enables HP release and moderate to good antibacterial activities against lab strains and multi-drug resistant clinical isolates.
Bacterial pathogens are able to evade the action of frontline therapies during infection due to antibiotic resistance and tolerance.1–5 As a result, human pathogens cause a multitude of severe and life-threatening infections that lead to numerous deaths worldwide.2,3 Bacteria acquire resistance to antibiotics through multiple mechanisms, including: target modification (mutation), antibiotic inactivation (via enzyme action), efflux pump action (to expel antibiotics), and changes in membrane chemistry to decrease antibiotic penetration.6–8 In addition, bacteria form surface-attached biofilm communities that contain enriched populations of metabolically-dormant persister cells.2,4,9–11 As our antibiotic arsenal currently consists of agents that operate through growth-dependent mechanisms, non-replicating persister cells are innately tolerant to each class of antibiotics.2,11
We have identified a series of halogenated phenazine (HP) compounds that target resistant and tolerant bacteria, and display potent antibacterial activities against multiple pathogens (Fig. 1A).12–19 The discovery of HP antibacterial agents was initially inspired by phenazine antibiotics that are utilized by the Gram-negative pathogen Pseudomonas aeruginosa to eradicate established Staphylococcus aureus infections in the lungs of young cystic fibrosis (CF) patients.20–22 We hypothesized the initial S. aureus infections in CF patients to be biofilm-associated and set out to evaluate a series of phenazine antibiotics (e.g., pyocyanin, 1-hydroxyphenazine) and non-natural phenazines in planktonic growth inhibition (to determine minimum inhibitory concentration, MIC) and biofilm eradication assays (to determine minimum biofilm eradication concentration, MBEC).
Fig. 1.
(A) Discovery and developmental pathway regarding novel HP antibacterial agents. (B) Design of an ether-linked HP-quinone prodrug model system for investigations.
From these efforts, we discovered a diverse series of HP analogues that demonstrate potent antibacterial activities (MIC ≤ 1 μM) and biofilm-killing activities (MBEC ≤ 10 μM) against multiple Gram-positive pathogens, including: S. aureus, S. epidermidis, Enterococcus faecalis.13–16,19 Additionally, select HPs have shown good to excellent antibacterial activities against Mycobacterium tuberculosis and Streptococcus pneumoniae.16,19 HPs have excellent cytotoxicity profiles and do not lyse red blood cells. Transcript profiling experiments indicate that active HPs induce rapid iron starvation (via RNA-seq and RT-qPCR),17 which is dependent on the metal-binding moiety in the HP scaffold that includes the 1-hydroxyl and adjacent nitrogen atom.14–19 Recently, HP-29 demonstrated good efficacy in wound infection models against S. aureus and E. faecalis in mice (Fig. 1A).19
One of our primary aims at this stage is to design and synthesize prodrug analogues through functionalization of the hydroxyl group of the parent HP compound to mitigate off-target iron binding, prevent phase II conjugation (metabolic liability), improve pharmacokinetic (PK) profiles and reduce toxicity.16,23,24 A critical goal of our prodrug design element is to install molecular triggers that will enable a bacteria-promoted reaction and subsequent liberation of an active HP (e.g., HP-87, Fig. 1A). Our previous efforts related to HP prodrug development have included: (1) carbonate-linked prodrugs that improve water solubility (with the addition of polyethylene, or PEG, moieties)15,16 and form HP-cephalosporin23 or HP–erythromycin conjugates,24 and (2) AOCOM-linked prodrugs, such as HP-87. Despite encouraging progress, we are motivated to develop non-carbonate based prodrug linkers that would be stable to general (bacterial or mammalian) esterase-promoted hydrolysis and set our sights on an ether-linked model system for this work. Here, we incorporate an electron-poor quinone trigger that is conjugated to HP-29 through an ether linkage, which is designed to undergo a bioreduction in the cytosol of a bacterial cell to generate a “spring-loaded” (electron-rich) hydroquinone intermediate that spontaneously expels the antibacterial payload through ether cleavage (Fig. 1B).
We designed and synthesized a HP-quinone model to investigate the action of an ether-linked prodrug (Scheme 1). A short synthetic sequence began upon treatment of hydroquinone 1 with sodium hydride, then the addition of excess methyl iodide to generate compound 2 (29% yield). Following this, 2 was subjected to paraformaldehyde in hydrobromic acid/acetic acid to afford benzyl bromide 3 (69% yield). Ceric ammonium nitrate (CAN) oxidation of 3 led to the synthesis of quinone 4 (68% yield), which was then used to alkylate HP-29 following treatment with potassium carbonate in acetone under microwave irradiation to generate target compound HP-29-Q in 38% yield.
Scheme 1.
Chemical synthesis of ether-linked HP-quinone prodrug HP-29-Q.
Our initial goals were to assess the ability of HP-29-Q to (1) bind iron(II), and (2) undergo a quinone reduction-promoted release of HP-29. Each of these items were investigated via UV-vis spectroscopy and initial experiments demonstrated that HP-29-Q does not chelate iron(II), as designed. The elimination of iron-binding in the prodrug (HP-29-Q) compared to the parent HP is a result of functionalizing the 1-hydroxyl group of HP-29, which disrupts the structural moiety critical for chelation (Fig. 2A; note: based on the disappearance of HP-29’s peak at ∼390 nm and the appearance of a new peak at ∼560 nm, HP-29 binds iron to form a HP : iron complex in 1 minute when evaluated alongside HP-29-Q). These results align with our group’s previous UV-vis studies related to HPs binding iron while prodrug HP versions do not directly bind iron.14–16,19
Fig. 2.
UV-vis experiments demonstrate HP-29-Q (A) does not directly bind iron(II), and (B) glutathione (GSH) causes direct release of HP-29. Note: HP-29 was tested alongside HP-29-Q in these experiments as a reference/control.
In addition, HP-29-Q was subjected to glutathione (GSH; reducing agent) to investigate the quinone trigger25 and ether linkage for release of HP-29 in UV-vis experiments. These results were insightful as 1 equivalent GSH led to the slow release of the parent HP from HP-29-Q, based on the appearance of the HP-29 peak at ∼390 nm (required 18.6 hours to observe significant release/formation of HP-29; Fig. 2B). A second experiment showed 25 equivalents of GSH released a significant amount of HP-29 from HP-29-Q in 4.5 hours (see ESI†). Based on these UV-vis experiments, GSH-promoted release of HP-29 from HP-29-Q indicates that the quinone trigger is indeed susceptible to reduction; however, it appears the ether linkage of HP-29-Q is more stable than initially anticipated and we were curious to evaluate this prodrug molecule in bacteria.
Following chemical synthesis and UV-vis spectroscopy, HP-29-Q was evaluated in microbiological assays against multiple pathogens, including S. aureus and S. epidermidis clinical isolates. Initial MIC assays were performed with HP-1, HP-29, and HP-29-Q against methicillin-resistant S. aureus (MRSA) BAA-1707 in the presence and absence of GSH. This assay was designed to probe the potential for bioreductive activation of HP-29-Q’s quinone prodrug moiety in a biological setting. In the absence of GSH, HP-29-Q reported an MIC of 3.75 μM (2.1 μg mL−1) against MRSA BAA-1707, which was 47-fold higher than parent HP-29 (MIC = 0.08 μM). However, in the presence of GSH (200 μM, a non-growth inhibitory concentration), prodrug HP-29-Q reported a highly potent MIC of 0.16 μM, which was only 2-fold higher than parent HP-29 (MIC = 0.08 μM; Fig. 3). These findings demonstrate that GSH leads to the bioreduction of the quinone prodrug moiety of HP-29-Q to release the potent antibacterial agent HP-29. In addition, these assays also show that MRSA BAA-1707 alone can cause a similar bioreduction-promoted release of HP-29 to a lesser degree.
Fig. 3.
Image of MIC assay with HP-1, HP-29, and HP-29-Q against MRSA BAA-1707 in the presence and absence of glutathione (GSH, 200 μM; note: this is a single replicate and HP-29-Q also demonstrated MIC values at 2.5 μM without GSH against MRSA BAA-1707).
HP-29-Q was evaluated against a panel of Gram-positive pathogenic bacteria (Table 1). When treating MRSA strains BAA-1707 and BAA-44, HP-29-Q reported MIC values of 3.75 μM (2.1 μg mL−1; previously mentioned) and 7.5 μM (4.2 μg mL−1), respectively. HP-29-Q reported an MIC of 7.5 μM against methicillin-resistant S. epidermidis (MRSE) strain ATCC 35984. Enterococcus strains (vancomycin-resistant E. faecium ATCC 700221 and E. faecalis OG1RF) were not susceptible within the test concentrations of HP-29-Q (MIC > 10 μM, highest test concentration in these assays due to compound solubility issues); however, S. pneumoniae ATCC 6303 proved to be the most sensitive pathogen/strain in our panel as HP-29-Q reported an MIC of 2.5 μM (1.4 μg mL−1). In all instances, the MIC values of prodrug HP-29-Q were significantly higher than the parent HP-29 due to the time required for bioreductive activation and release of the active HP; however, HP-29-Q demonstrated good antibacterial activities comparable to HP-1 against multiple Gram-positive bacterial strains. In addition, HP-29-Q demonstrated low levels of cytotoxicity in mammalian cells as the IC50 value against HeLa cells was found to be >100 μM in 24 hours lactate dehydrogenase release assays, similar to HP-29 (Table 1).
Table 1.
Antibacterial summary for HP-1, HP-29 and HP-29-Q as MIC values against a panel of Gram-positive bacterial pathogens. All biological results in this table are reported in micromolar (μM) concentrations and do not include MIC values from the GSH co-treatment assays
| Compound | MRSA BAA-1707 | MRSA BAA-44 | MRSE 35984 | VRE 700221 | E. faecalis OG1RF | S. pneumoniae 6303 | HeLa cytotox. IC50 |
|---|---|---|---|---|---|---|---|
|
| |||||||
| HP-1 | 2.5 | 2.5 | 3.75a | 5 | 25 | 1.88a | >100 |
| HP-29 | 0.08 | 0.12a | 0.16 | 0.31 | 0.94a | 0.08 | >100 |
| HP-29-Q | 3.75a | 7.5a | 7.5a | >10 | >10 | 2.5 | >100 |
| Vancomycin | 0.39 | 0.39 | 0.78 | >100 | 0.78 | — | — |
| Daptomycin | 3.13 | 18.8a | 12.5 | 125 | — | — | — |
| Linezolid | 12.5 | 1.56 | 3.13 | 3.13 | — | — | — |
Midpoint value for 2-fold range in MIC values observed. Each data point is the result of three independent experiments.
Following initial assessment in lab strains, HP-29-Q was then evaluated for antibacterial properties against a panel of clinical isolates obtained from patients treated at Shands Hospital (Gainesville, FL; Table 2). HP-29-Q demonstrated a similar antibacterial profile against the panel of clinical isolates (MIC values 3.75 to >10 μM), all of which were shown to be multi-drug resistant during these investigations. S. aureus 156 was found to be resistant to three of the six antibiotics tested (methicillin, MIC = 25 μM; erythromycin, MIC > 100 μM; tobramycin, MIC = 12.5 μM), yet HP-29-Q demonstrated an MIC = 5 μM (2.8 μg mL−1) against this clinical isolate. Overall, we were encouraged by the findings with our prodrug model compound.
Table 2.
Summary of HP-1, HP-29 and HP-29-Q reported as MIC values against S. aureus and S. epidermidis clinical isolates. All MICs are reported in micromolar (μM) concentrations
| Compound | MRSA-1 | MRSA-2 | S. aureus 129 | S. aureus 147 | S. aureus 138 | S. aureus 156 | S. epidermidis 1 |
|---|---|---|---|---|---|---|---|
|
| |||||||
| HP-1 | 2.35a | 3.13 | 1.56 | 2.35a | 3.13 | 2.35a | 1.56 |
| HP-29 | 0.06a | 0.12a | 0.12a | 0.12a | 0.08 | 0.08 | 0.06a |
| HP-29-Q | 3.75a | >10 | >10 | 5 | 5 | 5 | 10 |
| Vancomycin | 0.39 | 0.39 | 0.39 | 0.59a | 0.39 | 0.39 | 0.59a |
| Methicillin | 37.5a | >100 | 37.5 | 6.25 | 37.5a | 25 | 12.5 |
| Ciprofloxacin | 0.78 | >100 | >100 | 1.17a | >100 | 0.78 | 0.30a |
| Tetracycline | 0.10 | 0.10 | 0.10 | 0.10 | 0.15a | 0.15a | 25 |
| Erythromycin | 100 | 18.8a | >100 | 37.5a | >100 | >100 | >100 |
| Tobramycin | 6.25 | 6.25 | >100 | 6.25 | 12.5 | 12.5 | 1.56 |
Midpoint value for 2-fold range in MIC values observed. Each data point is the result of three independent experiments.
In addition to MIC assessment, we performed zone of inhibition experiments in Petri dishes16,24 for HP-29 and HP-29-Q against S. aureus and S. epidermidis (Fig. 4 and ESI†). Our results were encouraging as all strains/clinical isolates tested in these experiments were found to be susceptible to HP-29-Q, further demonstrating the potential for bacterial-promoted bioreductive activation of the quinone prodrug moiety and HP release. Against MRSA BAA-1707, the zone of inhibition for parent HP-29 was 104.4 ± 2.2 mm2 while prodrug HP-29-Q gave a zone of inhibition of 64.9 ± 18.5 mm2 (DMSO served as the vehicle control and gave no zone of inhibition). Likewise, against clinical isolate MRSA-1, HP-29 reported a zone of inhibition at 105.9 ± 10.7 mm2, while the zone for HP-29-Q was 58.2 ± 10.4 mm2. Similar activity profiles were demonstrated from all strains tested as HP-29-Q showed a smaller zone of inhibition in these experiments likely as a result of the time needed for (inactive) quinone prodrug to be processed for HP release.
Fig. 4.
(A) Agar diffusion assay to determine zone of inhibition against MRSA BAA-1707 regarding (A.) DMSO (vehicle control, zone of inhibition: 0 mm2), (B.) HP-29 (zone: 104.4 ± 2.2 mm2), and (C) HP-29-Q (64.9 ± 18.5 mm2). (B) Agar diffusion assay to determine zone of inhibition against clinical isolate MRSA-1 regarding (A.) DMSO (zone of inhibition: 0 mm2), (B.) HP-29 (zone: 105.9 ± 10.7 mm2), and (C) HP-29-Q (58.2 ± 10.4 mm2). Note: Either 3 μL of DMSO alone or DMSO stock solution of HP agent (1 mM) was added directly to a lawn of bacteria spread over agar in these Petri dish assays.
In conclusion, we have designed, synthesized and evaluated an ether-linked HP-quinone prodrug molecule (HP-29-Q) that undergoes reductive release of a potent HP-29 molecule that induces iron starvation in bacteria. Our findings support the notion that the ether linkage in HP-29-Q is quite stable as a slow loss of HP-29 under reductive conditions results in significantly elevated MIC values when comparing prodrug to parent HP. However, despite losses in potency, this quinone prodrug demonstrated moderate to good antibacterial activities against multiple Gram-positive strains and multi-drug resistant clinical isolates. Additional studies are required to further develop ether-linked HP-quinone prodrug analogues and will be focused on (1) tuning HP-release profiles, (2) improving water solubility properties, and (3) investigating the quinone’s ether linkage to determine if the methylene unit is a metabolic “hot spot” that could lead to undesired HP release. These initial results will inform future prodrug design for HPs and potentially other antibacterial agents that can address a multitude of problems associated with antibiotic-resistant infections.
Supplementary Material
Acknowledgements
We acknowledge the University of Florida and National Institute of General Medical Sciences (R35GM128621 to R. W. H.) for providing financial support for this work.
Footnotes
Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ob01107c
Conflicts of interest
The authors have no conflicts of interest to declare.
References
- 1.Lewis K, Cell, 2020, 181, 29. [DOI] [PubMed] [Google Scholar]
- 2.Abouelhassan Y, Garrison AT, Yang H, Chávez-Riveros A, Burch GM and Huigens RW III, J. Med. Chem, 2019, 62, 7618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fitzpatrick MA, Lancet Infect. Dis, 2020, 20, 1108.32505229 [Google Scholar]
- 4.Liu K. and Huigens RW III, ACS Infect. Dis, 2020, 6, 541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Brown ED and Wright GD, Nature, 2016, 529, 336. [DOI] [PubMed] [Google Scholar]
- 6.Wright GD, Chem. Commun, 2011, 47, 4055. [DOI] [PubMed] [Google Scholar]
- 7.Blair JMA, Webber MA, Baylay AJ, Ogbolu DO and Piddock LJV, Nat. Rev. Microbiol, 2015, 13, 42. [DOI] [PubMed] [Google Scholar]
- 8.Munita JM and Arias CA, Microbiol. Spectrum, 2016, 4, DOI: 10.1128/microbiolspec.VMBF-0016-2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lewis K, Handb. Exp. Pharmacol, 2012, 211, 121. [DOI] [PubMed] [Google Scholar]
- 10.Yan J. and Bassler BL, Cell Host Microbe, 2019, 26, 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wood TK, Biotechnol. Bioeng, 2016, 113, 476. [DOI] [PubMed] [Google Scholar]
- 12.Borrero NV, Bai F, Perez C, Duong BQ, Rocca JR, Jin S. and Huigens RW III, Org. Biomol. Chem, 2014, 12, 881. [DOI] [PubMed] [Google Scholar]
- 13.Garrison AT, Abouelhassan Y, Kallifidas D, Bai F, Ukhanova M, Mai V, Jin S, Luesch H. and Huigens RW III, Angew. Chem., Int. Ed, 2015, 54, 14819. [DOI] [PubMed] [Google Scholar]
- 14.Garrison AT, Abouelhassan Y, Norwood VM, Kallifidas D, Bai F, Nguyen MT, Rolfe M, Burch GM, Jin S, Luesch H. and Huigens RW III, J. Med. Chem, 2016, 59, 3808. [DOI] [PubMed] [Google Scholar]
- 15.Yang H, Abouelhassan Y, Burch GM, Kallifdas D, Huang G, Yousaf H, Jin S, Luesch H. and Huigens RW III, Sci. Rep, 2017, 7, 2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Garrison AT, Abouelhassan Y, Kallifidas D, Tan H, Kim YS, Jin S, Luesch H. and Huigens RW III, J. Med. Chem, 2018, 61, 3962. [DOI] [PubMed] [Google Scholar]
- 17.Abouelhassan Y, Zhang Y, Jin S. and Huigens RW III, Angew. Chem., Int. Ed, 2018, 57, 15523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huigens RW III, Abouelhassan Y. and Yang H, ChemBioChem, 2019, 20, 2885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yang H, Kundra S, Chojnacki M, Liu K, Fuse MA, Abouelhassan Y, Kallifidas D, Zhang P, Huang G, Jin S, Ding Y, Luesch H, Rohde KH, Dunman PM, Lemos JA and Huigens RW III, J. Med. Chem, 2021, 64, 7275–7295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Machan ZA, Pitt TL, White W, Watson D, Taylor GW, Cole PJ and Wilson R, J. Med. Microbiol, 1991, 34, 213. [DOI] [PubMed] [Google Scholar]
- 21.Laursen JB and Nielsen J, Chem. Rev, 2004, 104, 1663. [DOI] [PubMed] [Google Scholar]
- 22.Price-Whelan A, Dietrich LEP and Newman DK, Nat. Chem. Biol, 2006, 2, 71. [DOI] [PubMed] [Google Scholar]
- 23.Xiao T, Liu K. and Huigens RW III, Bioorg. Med. Chem. Lett, 2020, 30, 127515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yang H, Liu K, Jin S. and Huigens RW III, Org. Biomol. Chem, 2021, 19, 1483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Daniel KB, Sullivan ED, Chen Y, Chan JC, Jennings PA, Fierke CA and Cohen SM, J. Med. Chem, 2015, 58, 4812. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.