Total synthesis and mechanism of action of the antibiotic armeniaspirol A

Emerging antimicrobial resistance urges the discovery of antibiotics with unexplored, resistance-breaking mechanisms. Armeniaspirols represent a novel class of antibiotics with a unique spiro[4.4]non-8-ene scaffold and potent activities against Gram-positive pathogens. We report a concise total synthesis of (±) armeniaspirol A in six steps with a yield of 20.3% that includes the formation of the spirocycle through a copper-catalyzed radical cross-coupling reaction. In mechanistic biological experiments, armeniaspirol A exerted potent membrane depolarization, accounting for the pH-dependent antibiotic activity. Armeniaspirol A also disrupted the membrane potential and decreased oxygen consumption in mitochondria. In planar lipid bilayers and in unilamellar vesicles, armeniaspirol A transported protons across membranes in a protein-independent manner, demonstrating that armeniaspirol A acted as a protonophore. We provide evidence that this mechanism might account for the antibiotic activity of multiple chloropyrrole-containing natural products isolated from various origins that share a 4-acylphenol moiety coupled to chloropyrrole as a joint pharmacophore. We additionally describe an efflux-mediated mechanism of resistance against armeniaspirols.


Unsuccessful attempts to synthesize armeniaspirol A
To a solution of 7 (0.222 g, 1 mmol, 1 equiv.) in DCM (4 mL) was added SnCl 4 (0.15 mL, 1.3 mmol, 1.3 equiv.) at 0 oC and stirred for 1h at the same temperature. Then dichloromethoxymethane (0.11 mL, 1.2 mmol, 1.2 equiv.) was added and stirred for at the same temperature for another 1h. The reaction mixture was poured into cold water (5 mL) and stirred for 1h at RT. Then the mixture was extracted with DCM (10 mL). The resulting organic layers were washed with brine and dried. Concentrated under vacuo and crude was purified by flash column chromatography (PE: EA= 5:1) resulted the desired aldehyde 11 (0.225 g, 0.9 mmol, 90%) pale yellow oil.
To a solution of 11 (0.175 g, 0.7 mmol, 1.0 equiv) in DCM (4 mL) was added I 2 (0.018 g, 0.07 mmol, 0.1 equiv.) and 1,3-propanedithiol (0.078 mL, 0.77 mmol, 1.1 equiv) and stirred for 1h at RT. The reaction mixture was quenched with 0.1 M Na 2 S 2 O 3 solution and extracted with DCM (5 mL). The resulting organic layers were washed with brine, dried and concentrated under vacuo. The crude was used for the next step without any purifications.  To a solution of 7 (0.25 g, 1 mmol, 1 equiv) in THF (2.5 mL) was added MeLi (1.6 M, 0.9 mL, 1.5 mmol, 1.5 equiv) at 0 o C and stirred for 2h at the same temperature. The mixture was quenched with aq. NH 4 Cl (2 mL) and extracted with Et 2 O (5 mL). The resulting organic layers were washed with brine, dried and concentrated under vacuo gives crude (S3), which was used directly for the next step.
To the above crude in DCM (4 mL) was added Dess-Martin reagent (0.63 g, 1.55 mmol, 1.5 equiv) and stirred for 1h at RT. The reaction was quenched with aq. Na 2 S 2 O 3 solution and extracted with DCM. The resulting organic layers were washed with brine, dried and concentrated under vacuo gives crude (S4), which was used for the next step without any further purifications.
To the above crude in MeOH (1.5 mL) was added TosNHNH 2 (0.186 g, 1 mmol, 1 equiv) at RT and stirred at 60 °C for 2h. The excess solvent was removed under vacuo, the resulting crude was purified by flash column chromatography (PE: EA= 5:1) gives the desired S4 (0.33 g, 0.78 mmol, 77%) as a solid.

Synthesis of (±) armeniaspirol A (5) and derivatives 13 and 15
Synthesis of 2-hexylresorcinol (12)  To a solution of 10 (1.3 mL, 10 mmol) in dry THF (50 mL) at 0 o C was added n-BuLi (2.5 M in hexane, 4.8 mL, 12 mmol, 1.2 equiv). The resulting solution was stirred at the same temperature for 1h and at r.t. for 2h. Again, the mixture was cooled to 0 o C then added 1-bromohexane (1. 6mL, 11 mmol, 1.1 equiv). The mixture was allowed to warm to r.t. and then stirred at the same temperature for 17h. The reaction was quenched with aq. NH 4 Cl (50 mL). The aqueous phase was extracted with Et 2 O (100 mL). The combined organic extracts were dried over Na 2 SO 4 , filtered and concentrated in vacuo. The crude was used for the next step without any further purification.
Synthesis of (±)-armeniaspirol A (5) To a solution of 5 (0.02 g, 0.052 mmol) in dry DMF (0.5 mL) were added MeI (0.037 g, 5 equiv) and K 2 CO 3 (0.014 g, 2 equiv). The mixture was stirred at rt for 24h. The reaction mixture was quenched with water (1 mL) and extracted with EtOAc (1 mL x2). The resulting organic layers were washed with brine and dried. The crude was purified by flash chromatography (PE: EA= 5:1) to give the product in 50 % yield (0.01 g) as brown oil.

Effect of pH on antibacterial activity of Micrococcus luteus in 384 well MTP format
To determine the effect of the pH of the medium on the activity of 1, cultures of Micrococcus luteus DSM1790 were grown in 384 well MTP format as described above. The pH of the MH medium (7.4,Roth) was adjusted to 6.5, 7, 8, 9 and 10 using HCl and NaOH solutions, and ciprofloxacin was included as a positive control. The MICs were determined as described above for two biological replicates with two technical replicates. DMSO controls were included for each pH, DMSO up to 1% did not impede bacterial growth. The final OD 600 of the cultures was strongly dependent on the pH, so that the bacterial growth was normalized to the final OD 600 of respective DMSO control.   Fluorescence was recorded with a Synergy 2 Multi-Mode Reader (BioTek, USA) at λ ex = 485 nm, λ em = 520 nm and λ ex = 485 nm, λ em = 600 nm, respectively. All measurements were done in triplicate. The membrane potential, expressed as the red/green fluorescence ratio, was calculated with respect to the DMSO-treated control.
The kinetic membrane potential assay using methicillin-resistant S. aureus N315 (MRSA) and E. coli ∆tolC followed a similar protocol to the one for MSSA, with the exception that the overnight culture was re-inoculated into fresh media to obtain an OD 600 of 0.05 (approximately 2.5  10 7 CFU/ml), and bacteria were cultivated until they reached exponential phase. the initial values (before treatment) and the mitochondrial membrane potential (m) was plotted over time.

Effects on electron transport chain
To determine whether 1 acts as an uncoupler of oxidative phosphorylation (an effect commonly observed in protonophores) or as an inhibitor of the electron transport chain, the effect on oxygen consumption in

Planar lipid bilayer assays
Planar lipid bilayers (BLM, black lipid membrane) conferring to Montal and Mueller were formed as published. 6 Briefly, an aperture in a Teflon septum with a diameter of 100 μm was pre-painted with hexadecane dissolved in n-hexane at 1-5% (v/v) and the cuvette compartments were dried for 30-35 min, in-order to eliminate the solvent. The bilayers were made with 1,2-diphytanoyl-sn-glycero-phosphatidylcholine at a concentration of 5 mg/ml in n-pentane.
We first measured the conductance of the bilayer membrane alone, which was negligible. After insuring a tight membrane, we added the indicated concentration of 1 or CCCP both dissolved in DMSO. Standard Ag/AgCl electrodes were used to detect the ionic current. Moreover, the cis side electrode of the cell was

Relevance of observed mutations
The observed single point mutations for armeniaspirol B (2) in the intergenic region upstream of cvpA and fdoG as well as plsB can all possibly be linked to the stringent response that is a stress response that can lead to the obtained resistance. Guanosine 5′-(tri)diphosphate 3′-diphosphate [(p)ppGpp] is an alarmone that is produced during the stringent response, and bacterial resistance mechanisms have been reported in several species and against a range of antimicrobials. 8 The cvpA gene is responsible for Colicin V secretion as well as the activation of the stress response pathway, which promotes membrane potential homeostasis. 9 Cho and co-workers (2021) found that cvpA is upregulated during the stringent response in the absence of MDR efflux systems, which is mostly regulated by the RpoS sigma factor. The RpoS sigma factor plays a major role in controlling gene expression during the stationary growth phase. The fdoG gene protects E. coli cells against antimicrobial peptides in the stationary-phase. 10 The protection is RpoS independent, however, mediated by BipA GTPase (BPI-inducible protein A) that is dependent on oxidation by a formate dehydrogenase. The BipA GTPase has been described in several cellular processes including antimicrobial resistance against a variety of antimicrobials. 11 The plsB gene is responsible to catalyze the first step in the phospholipid biosynthesis and can be linked to the stringent response as it is a proposed target of (p)ppGpp. 12 Inhibition occurs during the production of (p)ppGpp which interferes with membrane-associated steps in peptidoglycan biosynthesis, which can allow for resistance. 13 The interaction between these genes and the stringent response should still be further investigated, however, the obtained resistance linked with the mutations of these genes could possibly describe an efflux-independent mode of resistance against armeniaspirols.

Resazurin assay
The cytotoxicity testing was performed by monitoring resazurin reduction.  Figure S5). The effect was about 40x lower than the effect observed after 5d in the MTT test, as both the direct toxicity and loss of cell proliferation account for the long-term decrease of MTT. The highest compound concentration without a solvent effect was 580 µM, which was not high enough to record a complete dose-response curve. The IC 50 of 5 could therefore not be accurately calculated, but was estimated to be at 200-300 µM.
The graphs represent an exemplary result from technical duplicates of 2 independent biological replicates.

Calculation of pK a values and other physicochemical properties of chloropyrrolecontaining natural products
To be considered as a protonophore, a substance should have a pK a value close to the physiological pH as well as good membrane permeability. Three programs were used to calculate the pK a values with different methods. While Schrödinger's Epik 19,20 uses Hammett and Taft methods in conjunction with ionization and tautomerization tools, Jaguar 21 utilizes quantum-chemical methods. ChemAxon's calculation method is based on micro and macro dissociation constants. The obtained values suggest that all compounds except (±)-deoxy-armeniaspirole-A (13) have a functional group with a pK a value of 5.5 -8.5. CCCP and 1- LogD values are another surrogate parameter for the prediction of passive membrane permeability.
Especially for very polar compounds with a logD value < 0, membrane permeability may be limited, which is not the case for all compounds in the scope of this study.
The pK a values were calculated with the following program versions and settings: