Potent synergistic efficacy of 2-methoxy-1,4-naphthoquinone derived from quinones against drug-resistant bacteria

The emergence and worldwide dissemination of mobile tigecycline resistance genes tet (X3)/ tet (X4) posed an enormous threat to the public health. Urgently, feasible strategies must be implemented to restore the clinical efficacy of tetracyclines and prolong the lifespan of existing drugs to address the emerging global antimicrobial resistance threat. Herein, versatile structural scaffolds of quinones for antibiotic adjuvants discovery enlightened a promising and underappreciated reservoir to circumvent the antibiotic resistance. 2-methoxy-1,4-naphthoquinone (MNQ) exhibited the potent potentiation (4 to 32-fold) with tetracyclines, along with effective inhibition on biofilm formation. Mechanistic studies revealed that MNQ synergistically operates with tetracyclines by inhibiting the enzymatic activity of Tet(X3)/Tet(X4) proteins through interaction with their active residues. Furthermore, exposure to MNQ significantly dissipate the proton motive force, leading to a cascade of membrane structural damage and metabolic homeostasis imbalance. Encouragingly, the MNQ-tetracyclines combination showcased substantial therapeutic benefits in two in vivo infection models, as evidenced by the reduced bacterial burden and mitigated pathological injury. Our findings propose a potential therapeutic option and a novel tetracyclines’ adjuvant against drug-resistant pathogens carrying Tet(X3)/Tet(X4).


Introduction
Over the past 100 years [1], antibiotics were broadly used in bacterial infection therapeutics and drastically changed modern medicine.Since 1980s, a gradual decline in antibiotics discovery and the development of multidrug resistance have led to a surge of drug-resistant bacterial infections and seriously threatened the therapeutic options, with a potentially estimated 700,000 deaths attributed to antimicrobial resistance (AMR) annually [2].Urgently, alternative strategies to tackle the crisis were absolutely imperative.
Alarmingly, with the emergence of metallo-β-lactamases (MBLs)-mediated carbapenems resistance [3] and MCRmediated colistin-resistance [4], tigecycline was conferred as a last-resort antibiotic to address serious infections.Nevertheless, the potency of tetracyclines is compromised by multiple resistance mechanisms, including acquirement of catalyzing enzyme, overexpression of multidrug efflux pumps, bacterial membrane permeability changes and alteration of target sites [5,6].Notably, recent investigations implied that the evolution of transferable high-level tetracyclines resistance by the novel tet(X) orthologs confers an enormous threat to the efficacy of next-generation tetracyclines (tigecycline, omadacycline, and eravacycline) [7].The emergence of novel plasmid-mediated tigecycline resistance genes tet(X3)/tet(X4) in Enterobacteriaceae constitutes a serious threat for therapeutic options and public health [8].Additionally, more than 30 tetracycline-specific efflux pump genes and nine MDR efflux pump genes of the resistance-nodulation-cell division (RND) family have been found to impair tetracyclines susceptibility [9].Scientifically, focusing on these resistant mechanisms is the most effective and feasible strategy to prolong the lifespan of conventional antibiotics and tackle the crisis of antibiotic resistances.
It is an emerging model of antibiotic-mediated death in bacteria involving dysregulation of central metabolism and energetic pathways [10].Accumulating evidences support the involvement of multifarious bacterial metabolic pathway, for instance, the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OxPhos) pathway, thereby inducing in bacterial death [11].Hence, the metabolic reprogramming might be a novel promising strategy for the development of antibiotic adjuvants.
To date, the progress of novel antibiotics discovery was lagging far behind the rapid dissemination of MDR pathogens globally.Therefore, extending the antibiotic adjuvants and the combination therapy exhibited tremendous potential for infection therapeutics [12].Given that previous studies found the potentiation potency of plumbagin with tetracyclines [13], we further explored the potential mechanisms underlying the synergistic effects of quinones.In the current study, 2-methoxy-1,4-naphthoquinone (MNQ) with negligible cytotoxicity displayed potentiation activity with tetracyclines in Tet(X3)/Tet(X4)-producing pathogens in vivo/ in vitro.Collectively, our study confirms that MNQ is a promising candidate for the development of 1,4-naphthoquinones as a novel and safe tetracyclines adjuvant.Our discovery provided an effective therapeutic regimen for combating MDR pathogens infections and the distinctive mechanisms displayed potent promise for the drug discovery and development.

Screening of quinones-tetracyclines combinations identify MNQ and tetracyclines synergism against Tet(X3)/ Tet(X4)-producing E. coli
To date, numerous quinone-derived compounds have attracted increasing interests due to the abundant biological activities, such as antimalarial [14], antidiabetic [15], anti-inflammatory [16], and antibacterial [17] potentials.The emergence of resistance to the last-resort antibiotic tigecycline (due to large scale use) and therapeutic failure are the alarming obstacles to combat MDR pathogens infections globally [8].Nowadays, an adjuvant therapy could become an effective first-line-treatment strategy against MDR pathogens until new antibacterial drugs are available.Our recent evidence has emerged indicating that plumbagin, as a natural naphthoquinone, was identified as a potent broad-spectrum inhibitor of Tet(X3)/ Tet(X4) with the antibacterial synergistic activity plus tetracyclines [13].Hence, the largely unexplored novel property of quinones, remains an intriguing potential target against multiple drug-resistant bacterial infections.
Quinones, as a key source for future medicine development, are ubiquitous in nature and formed through diverse mechanisms [14].Notably, 1,4-Naphthoquinone (1,4-NQ), a class of natural compounds derived from naphthalene, contain a 1,4-naphthalenoid ring and easily susceptible to reduction, oxidation, and addition or substitutions of one or more methyl, hydroxyl and/or methoxy.Together, the high reactivity and the mature chemical modification technology of quinones make these natural products potential for the profound development of novel types of substances with abundant biochemical properties, and provide new targets and strategies for addressing multiple challenges.
Therefore, 28 widely used quinones, as listed in Supplementary Table 3, were screened using antimicrobial susceptibilities assay (Fig. 1a).Dramatically, plumbagin, 2-methoxy-1,4-naphthoquinone (MNQ), 2,3-dichloro-1,4-naphthoquinone (DNQ) and p-benzoquinone restored the susceptibility of tetracyclines (including tetracycline and tigecycline) against tet(X4)-positive bacteria (Supplementary Table 3).The results suggested that compounds which contain the cyclohexadienedione structure exhibited a distinct activity range depending on the electronic nature of different active fragments and their various substituents.Remarkably, MNQ along with high potency, selectivity and negligible toxicity manifested a broad-spectrum synergistic activity against drug-resistant bacteria harboring Tet(X3)/Tet(X4) (Fig. 1b  and c), which could be explored as a potential tetracyclines synergist with a novel mechanism of action.Taken together, MNQ was selected as a representative to analyse the mechanisms underlying the recovery of bacterial sensitivity to tetracycline antibiotics.
To demonstrate the broad-spectrum synergistic effect of MNQ, we additionally evaluated the potency of the combination of MNQ with various antibiotics using chequerboard microdilution assays.Interestingly, MNQ at sub-MIC levels potentiated multiple antibiotics activity, such as lincomycin and erythromycin, against diverse Tet(X3)/Tet(X4)-producing pathogens infections (Supplementary Fig. 2), suggesting that MNQ plus multiple antibiotics, especially tetracyclines, is a potential antibacterial and anti-infection strategy against bacterial infections.These findings demonstrated that MNQ was a promising broad-spectrum antibiotics adjuvant for fighting MDR bacteria infections.To further illustrate the potentiation of MNQ plus tetracyclines, the time-killing assay were performed to investigate their synergistic bactericidal activity.Excitingly, the combination of MNQ plus tetracycline (Fig. 2a  and b) exhibited potent synergistic effects against E. coli DH 5α + pAM401-tet(X3)/tet(X4), despite difficult to abrogate exponentially growing cells thoroughly.
To explore the ability of MNQ on the evolution of tigecycline resistance, serial passages of tet(X3)/tet(X4)positive E. coli supplementing with tigecycline in the presence and absence of MNQ for 25 days were performed.Excitingly, MNQ co-administrated with tigecycline significantly suppressed the evolution of tigecycline resistance (Fig. 2c and d).By contrast, MIC values of tigecycline treatment group were increased by 16-fold (E. coli DH 5α + pAM401-tet(X3)) and 32-fold (E. coli DH 5α + pAM401-tet(X4)) (Fig. 2c and d).These findings suggested that combined therapy with MNQ could hardly bring the selective pressure for the evolution of resistance.

MNQ in combination with tetracycline impairs biofilm formation and maturation
Biofilms play an important role in the bacterial survival, colonization and pathopoiesis [18].Therefore, biofilm formation and inhibition assays were further employed to probe the efficacy of MNQ on the inhibition of the bacterial biofilms formation.As indicated in Fig. 3a and b, MNQ treatment showed a significant inhibition on biofilm formation of E. coli DH 5α + pAM401-tet(X4) and a decreased biomass compared to the untreated control group.Moreover, the inhibitory effect on the development of E. coli DH 5α + pAM401-tet(X4) biofilms were also observed while MNQ in combination with tetracycline (Fig. 3a and b).Consistently, confocal laser scanning microscope images showed that MNQ monotherapy, or in combination with tetracycline resulted in a thinner and structurally incomplete layer of cells (Fig. 3c).Additionally, the combinations effectively decreased the biofilm formation of E. coli DH 5α + pAM401-tet(X3) (Supplementary Fig. 3a).
We further investigated whether MNQ also contributed to the disruption of pre-existing biofilm structures.As depicted in Fig. 3d and e, with the addition of MNQ, or the combinations with tetracycline to the developing E. coli biofilms, the biofilms development and maturation were significantly inhibited, as evidenced by a decreased crystal violet (CV) quantification and bacterial biomass.Consistent with these results, a rare density of attached biomass and fragmentary biofilm were observed in the MNQ monotherapy and MNQ + TET groups (Fig. 3f ), suggesting that MNQ addition could disrupt the development of biofilm and deactivate bacteria.Such inhibitions were also observed in the biofilms maturation against E. coli DH 5α + pAM401-tet(X3) (Supplementary Fig. 3b).Collectively, these observations concluded that MNQ works synergistically with tetracycline to impair the biofilms formation and maturation, which underscored the

Determination of the interaction between MNQ and Tet(X3)/Tet(X4)
Considerable evidence revealed that Tet(X3)/Tet(X4) are tetracycline-inactivating monooxygenase, the presence of which severely compromises the efficacy of tetracyclines.
And the catalytic activity of Tet(X3)/Tet(X4) could be evaluated by monitoring the changes in absorbance values at 400 nm [19].Accordingly, as shown in Fig. 4a and b, MNQ exhibited an obvious inhibitory efficacy on the catalytic activity of Tet(X3)/Tet(X4), suggesting that MNQ represents an effective Tet(X3)/Tet(X4) inhibitor that reverses tetracyclines activity against tet(X3)/tet(X4)positive pathogens.To initially explore the molecular basis for the suppression of Tet(X3)/Tet(X4) activity by MNQ, the secondary structures of Tet(X3)/Tet(X4) supplemented with or without MNQ were determined using a CD spectroscopy analysis.Dramatically, coadministrated with MNQ resulted in obvious alterations in the secondary structures of Tet(X3) (Fig. 4c and d) and Tet(X4) (Fig. 4e and f ), as evidenced by significant decrease in α-helix1, α-helix2 and turn conformations and increase in anti-1 and anti-2 conformations (Fig. 4c  and f ), suggesting that a direct engagement of MNQ with Tet(X3)/Tet(X4) accompanying the alternation of second structure may contribute to the inhibitory potent of MNQ.
The potential binding mode between MNQ and Tet(X3)/Tet(X4) was further determined using molecular docking and molecular dynamics (MD) simulations.Initially, the rootmean-square deviation (RMSD) value of the complex was calculated for the subsequent MD simulations.As revealed, MNQ could localize to the catalytic pocket of Tet(X3)/Tet(X4) and the active residues of Tet(X3) (LEU222, PHE319, PHE224 and MET215) and Tet(X4) (LEU219, MET212, ILE234, PHE221 and PHE 316) could form strong interactions with MNQ as shown in the binding modes (Fig. 4g and h).Specifically, the residues and their contribution to the system were calculated using the MM-PBSA approach.As indicated in Fig. 4i, PHE224 and PHE319 of Tet(X3) with a E vdw of less than − 5 kcal/mol, which showed a strong van der Waals interaction with MNQ.Similarly, PHE224 and PHE319 were observed to exhibit prominent solvation contribution (E solv ≥ 1.0 kcal/mol).Additionally, the residues PHE221, PHE316 and ILE234 of Tet(X4) also exhibited appreciable van der Waals interactions with MNQ (Fig. 4j).
On the basis of above results, site-specific mutagenesis and fluorescence quenching assays were further performed to confirm the binding mode of MNQ and Tet(X3)/Tet(X4).As expected, the binding constants for the interaction between MNQ and Tet(X3)/Tet(X4) mutants were significantly decreased (Fig. 4k and l), suggesting that the binding of Tet(X3)/Tet(X4)-WT with MNQ is evidently stronger than mutants.As shown in Fig. 4m and n, the mutants were more sensitive to tetracycline in comparison with the Tet(X3)/Tet(X4)-WT and the synergistic effects of MNQ combined with tetracycline against the mutants were significantly decreased compared to the wild-type E. coli DH5α-pGEX-4 T-1-Tet(X3)/Tet(X4). Collectively, these results suggested that MNQ is a potent Tet(X3)/Tet(X4) inhibitor by occupying residues L222, M215, F224 of Tet(X3) and L219, I234 of Tet(X4).

Mechanisms of the synergy of MNQ with tetracyclines
Given the broad-spectrum potentiation of MNQ to multiple antibiotics, we next sought to explore the potential synergistic mechanisms.Considerable evidences revealed that bacterial resistance might be associated with an increase of intracellular reactive oxidative species (ROS), which in turn, could potentially undermine the membrane integrity [20].We therefore inferred that MNQ, in combination with tetracyclines, might destruct the cell membrane integrity.As shown in Fig. 5a, addition of MNQ combined with tetracycline to tet(X4)-positive E. coli led obvious swelling or wrinkling on the surface of bacteria and eventual lysis of cells, while monotreatment exhibited insignificant effects on the bacterial morphological damage.These results indicated that MNQ could potently restore the antimicrobial effect of tetracyclines by triggering membrane damage.Further, the antimicrobial activity of MNQ in combination with tetracycline was further confirmed using the LIVE/DEAD BacLight viability assay, as shown by increased numbers of PI-labeled dead bacteria (red fluorescence) by combination treatment, whereas relatively weaker red fluorescence and stronger green fluorescence were observed by the monotreatment in both tet(X3)/tet(X4)-positive E. coli (Fig. 5b and c; Supplementary Fig. 4a and b).Furthermore, We assessed the membrane permeability using a fluorescent probe propidium iodide (PI).The data revealed that membrane permeability was dramatically increased by the addition of MNQ, confirmed by an enhanced fluorescence intensity (Fig. 5d and e), agreeing with the bacterial viability observation.Collectively, these results suggested that MNQ manifested an obvious impact on the inner membrane permeability.
Considerable evidences revealed that bacterial membrane damage was closely relevant to the proton motive force (PMF) required for ATP synthesis, which is a form of potential energy consisting of ∆ψ (membrane potential) and ∆pH (proton gradient) [21].As observed in Fig. 5f-i, MNQ administration dissipated the proton motive force (PMF), as evidenced by the disruption of ∆ψ (membrane potential; Fig. 5f and g) and ∆pH (proton gradient; Fig. 5h and i).Because PMF directly drives ATP production, the intracellular ATP was sequentially decreased by the co-administrated with MNQ (Fig. 5j  and k).Cumulatively, with the membrane permeability increasing, the monotherapy with MNQ or combination therapy resulted in a notably increased levels of cellular contents, such as β-galactosidase (Fig. 5l and m).These results indicated that MNQ triggered membranes damage by the disruption of proton motive force.
Microbes encode a diverse range of multidrug transporters, most of which are rely on the proton motive force (PMF) [21].We further evaluated the efflux activity of bacteria using rhodamine as an indicator, as expected, exposure to MNQ markedly decreased the efflux of rhodamine in tet(X4)-positive E. coli (Fig. 6a and b).

Multivariate data analysis of bacterial metabolites
Metabolomics was employed to further analyse the effect of MNQ on the bacterial metabolic homeostasis.As for the PCA model, the points were distinctly separated between the control and MNQ treatment groups in the NEG mode (Supplementary Fig. 5a) and POS mode (Supplementary Fig. 5b).Further, the OPLS-DA model was established to distinguish the differential metabolites between the two groups (Supplementary Fig. 5c-f ).Specifically, the R2Y and Q2 were 0.996 and 0.908 in the NEG mode and 0.991 and 0.81 in the POS mode, respectively (Supplementary Fig. 5e and f ).As shown in Fig. 7a and b, differentially enriched metabolites were identified and the heat maps analysis indicated that 114 metabolites were down-regulated and 83 metabolites were up-regulated in the NEG mode and moreover, 62 metabolites were down-regulated and 92 metabolites were up-regulated in the POS mode.Furthermore, KEGG and MetaboAnalyst analysis were employed to identify the significantly enriched metabolic pathways.As indicated in Fig. 7c and e, purine metabolism, TCA cycle, arginine biosynthesis, and alanine, aspartate and glutamate metabolism exhibited significant difference (P < 0.05) in the NEG mode.And beta-alanine metabolism, E. coli biofilm formation and ATP-binding cassette (ABC) transporter pathway were prominently enriched in the POS mode (Fig. 7d and f ).Collectively, bacterial metabolites differences indicated that MNQ addition significantly interfered with bacterial metabolism, which prompts us to focus on the above mentioned pathways.And an imbalance in metabolic homeostasis induced by MNQ might eventually lead to an advancement in cell death.The existing results illustrated that the multiple mechanisms underlying its mode of action involve directly inhibiting Tet(X3)/Tet(X4) catalytic activity and proteins synthesis in tet(X3)/tet(X4)-positive pathogens, undermining the functions of PMF-driven efflux pump, and accelerating bacterial metabolic homeostasis imbalance.

MNQ reverses tetracyclines resistance in vivo
To extend clinical application of MNQ, further work is imperative to evaluate the efficacy and cytotoxicity of candidate drugs.Encouragingly, we found that MNQ at the indicated effectively concentrations, or combined with tetracyclines exhibited negligible cytotoxicity in Caco2, HEK-293, and HeLa cells (Supplementary Fig. 6a and b).Additionally, the combination of tetracycline and MNQ (2-128 μg/mL) displayed no significant hemolytic activity to RBCs (Supplementary Fig. 6c).Moreover, for the toxicity test of MNQ, no mortality was observed in 10-day studies.And no treatment-related adverse effects were observed in body weight or adverse biological responses (Supplementary Fig. 6d and e).In addition, safety assessments manifested that compared with the control group, the histopathological examination consisted of hearts, livers, spleens, lungs and kidneys showed no abnormal signs (Supplementary Fig. 6f and g).These results confirmed satisfactory pharmaceutical properties and safety profile of MNQ with no overt toxicity up to 5 and 15 mg/kg body weight.
Then, the synergistic efficacy of MNQ with tetracyclines in vivo was further evaluated using two mice infection models.

Mouse thigh infection model
A schematic of the combinatorial therapy protocol was shown in Fig. 8a.Compared with the monotherapy, the combination therapy of tetracycline and MNQ (5 + 5 or 10 + 5 mg/kg) significantly decreased the bacterial burden in mouse thigh muscle (Fig. 8b).Dramatically, haemorrhage, hyperemia, substantial inflammatory cells infiltration in muscle tissue, abscess formation and tissue necrosis were observed by E. coli infection, as evidenced by H&E staining and gross pathological observation (Fig. 8c).Whereas, the combination of MNQ plus tetracycline prominently alleviated thigh abscess lesion (Fig. 8c), indicating that MNQ could potentiate tetracycline activity against E. coli 47EC infection.potential of MNQ to enhance the efficacy of methacycline in vivo (Fig. 8d).As shown in Fig. 8e, the bacterial load in livers and spleens were significantly decreased by the combination therapy of methacycline and MNQ, which further confirmed by the alleviated pathological changes (Fig. 8f ).Hematoxylin and eosin (H&E) staining of major organs in the E. coli-infection mice showed severe morphological damages and fibrosis of myocardial cells, necrosis and congestion in germinal center of spleen, severe cellular degeneration and spotty necrosis in livers, thickened alveolar septum and accumulating inflammatory cell in lungs, renal tubular damage in kidneys.Whereas, pathological damage were obviously alleviated by the combination therapy (Fig. 8f ).Collectively, these data convincingly confirmed the adjuvant potential of MNQ with methacycline to tackle MDR bacterial infections.

Discussion
The emergence and evolution of tet(X) orthologs paralysed the tetracycline antibiotic therapeutic efficacy and posed a severe threat to public health [8,9].Given that combination approaches potentiate antibiotics activity to address this crisis, we evaluated synergetic activity of quinones with tetracyclines derived from over 200 candidate compounds, and in light of the structural determinants and functional importance of quinones, our findings suggested that the introduction of lipid soluble groups, such as methyl (plumbagin) or substituted methoxyl (2-methoxy-1,4-naphthoquinone) to the 2-position, or chlorine (2,3-dichloro-1,4-naphthoquinone) into the C-2,3 position effectively restored tetracyclines activity against tet(X3)/tet(X4)-positive E. coli.Remarkably, among the active substances, MNQ with negligible cytotoxicity exhibited the potent potentiation (4 to 32-fold) with tetracyclines against MDR pathogens.
MNQ, a diverse class of 1,4-naphthoquinones, was isolated from Impatiens glandulifera and its potential pharmacoactive and applications have not been extensively explored, especially for the infection therapeutics.In the current study, MNQ as a novel tetracyclines adjuvant significantly potentiates tetracyclines activity.The synergetic mechanisms in tet(X3)/tet(X4)-positive pathogens revealed that MNQ potently inhibited the catalytic activity and production of Tet(X3)/Tet(X4) by engaging with the catalytic pocket of Tet(X3)/Tet(X4).Additionally, exposure to MNQ enhanced membrane permeability and dramatically undermined efflux pump activity.Moreover, MNQ works synergistically with tetracyclines by perturbation of bacterial metabolism homeostasis, which brought about the disturbance of amino acid metabolism and multidrug resistance ATP-binding cassette (ABC) transporters were also perturbed.Inspiredly, understanding how the combination therapy impact bacterial metabolism may provide insight into the concrete mechanisms of action and support evidences to potentiate antibiotics activity may be an effective means [11].
Moreover, we found that MNQ works synergistically with tetracyclines to destroy the biofilms formation and maturation, which provide a promising outlook for the potential application of MNQ as an antibiotic-potentiating adjuvant to treat chronic infection or prevent secondary bacterial infections.Encouragingly, in the two infection models used in our study, MNQ exhibited a robust potentiation with the tested antibiotics against MDR pathogens infections in vivo, as evidenced by decreased bacterial counts.Worryingly, poor water solubility and low bioavailability always limited the clinical utility of MNQ.To tackle these problems, improving the dosage form and developing other drivers for effective delivery, for instance, ointments, polymeric microparticles and nanoparticles, were apparently promising, highefficacy, low-cost and low-toxicity strategies.
In conclusion, we found that MNQ derived from quinones exhibited robust synergistic activities with tetrayclines against MDR pathogens, which provided versatile structural scaffolds for antibiotic adjuvants discovery.Importantly, the discovery of MNQ with distinctive mechanisms highlights great promise for infection therapeutics.

Reagents
Unless otherwise noted, commercial reagents and solvents were purchased from commercial suppliers and applied without further purification.Purities of all tested compounds were confirmed to be > 95% by high-performance liquid chromatography (HPLC) analysis.

Cell lines and cell culture
Human epithelial colon carcinoma (Caco2) cells line, human embryonic kidney (HEK293) cells line and human cervical carcinoma (HeLa) cells line were cultured in high glucose Dulbecco's Modified Eagles Medium (DMEM) with 10% foetal bovine serum (FBS, Biological Industries) and 1% (w:v) penicillin-streptomycin at 37 °C.All cell lines were obtained from the American Type Culture Collection (ATCC).

Mice
Female BALB/c mice (6-8 weeks) were purchased from Liaoning Changsheng Technology Industrial Co., Ltd (Liaoning, China).Mice were acclimated in a standardized environment (constant temperature of 23 ± 2 °C and humidity of 55%) for 1 week.All experimental procedures were strictly adhered to the guidelines of the Jilin University Institutional Animal Care Committee.

Antimicrobial susceptibilities
Bacteria used in this study were listed in the Supplementary Table 1.Minimum inhibitory concentrations (MICs) of potential candidate compounds and antibiotics were determined by the broth micro-dilution susceptibility test in Mueller−Hinton broth (MHB, Qingdao Hope Bio-Technogy Co., Ltd.(Qingdao, China)), according to the Clinical and Laboratory Standards Institute (CLSI) 2015 guidelines [22].After 16-24 h incubation at 37 °C, MIC values against the bacterial isolates were defined as the lowest concentrations of antimicrobial agents with no visible growth of bacteria.Synergistic activity of antibiotics and candidate compounds was evaluated by checkerboard assays as described previously [23].Antibiotics and candidate compounds were serially diluted in a sterile 96-well microtiter plate.After 16-24 h co-incubation with bacterial suspension (5 × 10 5 CFUs/well), the MIC values of each combination were monitored and then used for FIC index (FICI) calculation according to the following formula [24]: FIC index = FICI A + FICI B = MIC AB / MIC A + MIC BA /MIC B .MIC A and MIC B are the MIC of compound A and B, respectively; MIC AB and MIC BA are the MIC of one compound in combination with another.And FICI ≤0.5 was defined as synergistic.

Time-dependent killing assay
Overnight cultures of E. coli DH5α + pAM401-tet(X3) and E. coli DH 5α + pAM401-tet(X4) were respectively diluted 1:1000 in MHB medium and incubated at 37 °C with shaking at 180 rpm.At the initial exponential phase of growth, different concentrations of MNQ, tetracycline or the combinations were added and incubated for the microbiological plating at the indicated time point.And the colony-forming units (CFUs) were calculated after incubation at 37 °C for 24 h.

Resistance development analysis
Sequential culturing of E. coli DH5α + pAM401tet(X3)/tet(X4) supplemented with subinhibitory levels of MNQ, TIG, or the combinations was repeated for 25 days.After 24 h co-incubation, the MICs of cultures were determined by broth micro dilutions in 96-well microtiter plates.

Biofilm formation and inhibition assays Biofilm inhibition assays
Overnight bacterial cultures were diluted to 1 × 10 6 CFUs/mL with fresh MHB broth in the presence or absence of tetracycline, MNQ or the combinations in a 24-well plate.The formation of biofilms were developed for 24 h at 37 °C.Then, the supernatant per well was decanted and the washed biofilm was stained with 0.1% crystal violet (CV).The CV-staining biomass in each well was solubilized with 33% glacial acetic acid and quantified by measuring the absorption values at OD 570 nm using a microplate reader (BioTek SLXFA, Winooski, VT, USA).

Biofilm disruption assays
For the biofilm destruction assays, biofilms were initially grown for 12 h in the condition of static cultures at 37 °C, as described above.Subsequently, the spent supernatant was removed and the indicated concentrations of tetracycline, MNQ or the combinations (diluted in MHB medium) was supplemented into the wells respectively and incubated for an additional 12 h to analyse the disruption of biofilm via crystal violet (CV) staining.

Confocal microscopy images of biofilms
For the imaging and analysis, confocal laser scanning microscopy (CLSM) was further employed.Briefly, the biofilm formation and inhibition assays were performed as described above and the media was removed, the underlying biofilms were labeled with SYTO 9 and PI (LIVE/DEAD BacLight Bacterial Viability Kit, Invitrogen, Carlsbad, CA, USA) in the dark environment.The samples were visualized using an inverted confocal laser scanning microscope.And the biofilm thickness of each groups was evaluated using ImageJ software (version number 1.8.0).

Bacterial viability determination
Overnight bacterial cultures were diluted and incubated with tetracycline, MNQ or the combinations for 5 h with the shaking of 200 rpm at 37 °C.The bacterial cells were collected, washed and resuspended in PBS.For CLSM observation, the cells (OD 600 nm = 0.8) were stained using the LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen).And the fluorescent intensity was quantified using ImageJ software.

Proton gradient determination
E. coli DH5α + pAM401-tet(X4) cultures were washed, collected and resuspended with HEPES (5 mM, pH 7.0) to OD 600 nm of 0.5.And the proton motive force was determined using pH-sensitive fluorescence probe BCECF-AM (Merck, Billerica, USA), to recapitulate, the cells were incubated with BCECF-AM (10 μM) at 37 °C for 1 h and incubated with the indicated concentrations of tetracycline, MNQ or the combinations for an additional 1 h.Then, the fluorescent intensity was determined at the excitation and emission wavelengths of 488/535 nm.

Membrane depolarization analysis
Bacterial cultures were washed and resuspended to 5 × 10 8 CFUs/mL.Then, 3, 3-dipropylthiadicarbocyanine iodide DiSC 3 (5) (2 μM; Aladdin, Shanghai, China) was added and measured with excitation wavelength at 622 nm and emission wavelength at 670 nm continuously for 10 min.Whereas, the indicated concentrations of MNQ, tetracycline or the combinations were supplemented and measured for an additional 30 min.

Intracellular ATP analysis
The bacterial lysis supernatant co-incubated with MNQ, tetracycline or the combinations was collected for the determination of Intracellular ATP level, using an Enhanced ATP Assay Kit (Beyotime, Shanghai, China).

Inner membrane (IM) integrity assay
Bacterial cells were washed, resuspended and cultured with MNQ, tetracycline or the combinations for 1 h, followed by the addition of 10 nM propidium iodide (PI; Sigma-Aldrich, St Louis, MO, USA).After incubation for 30 min, membrane integrity was determined by measuring with the excitation wavelength at 535 nm and emission wavelength at 615 nm.

Efflux pump assay Rhodamine B efflux analysis
To probe whether efflux pump plays a vital role in the recovery of antibacterial activity of tetracyclines plus MNQ, a specific fluorescent dye Rhodamine B (Sigma-Aldrich, St Louis, MO, USA) was applied to analyse the effect of MNQ on the efflux pumps [25].Bacterial cultures were washed and resuspended to obtain an OD 600 nm of 0.5, then cultured with Rhodamine B (5 μM) for 30 min at 37 °C.After washing for three times, the bacterial suspension was harvested and co-incubated with the indicated concentrations of tetracycline, MNQ or the combinations for 1 h, respectively.The fluorescence intensity of bacterial co-culture supernatant was measured at the excitation/emission wavelength at 553 nm/627 nm.

Ethidium bromide (EtBr) efflux assay
To explore the effect of MNQ on the inhibition of multidrug efflux pump, an EtBr efflux assay was performed as described previously [26].In brief, overnight bacterial culture adjusted to an OD 600 nm of 0.5 was cultured with 5 μM EtBr and indicated concentrations of tetracycline, MNQ or the combinations at 37 °C for 1 h.After centrifuged at 5000×g at 4 °C, the bacterial pellets were resuspended with MH-broth.And efflux pump activity was determined using a microplate reader at the excitation/ emission wavelength at 526 nm/605 nm.

Recombinant Tet(X3)/Tet(X4) construction, expression and purification
For the production of recombinant pGEX-4 T-1-Tet(X3)/ Tet(X4) and the specific mutants, prokaryotic expression systems were employed and performed according to our previous study [13].And the specific primers used in the current study are listed in Supplementary Table 2.

Enzyme inhibition assay
The purified protein Tet(X3)/Tet(X4) exposed to the indicated concentrations of MNQ in the presence of NADPH, Mg 2+ and O 2 were incubated at 37 °C for 30 min.Subsequently, the cultures were mixed with tetracycline for an additional 30 min.And the inhibition of catalytic activity was determined by the absorbance at 400 nm and expressed as Inhibition (%) = (1−(∆OD 400 compounds /∆OD 400 Vehicle control )) × 100%.
The molecular mechanics/Poisson-Boltzmann surface area (MM-PBSA) approach supplied with the Amber 10 package were further employed for the calculation of the interaction energy of MNQ-Tet(X3)/Tet(X4) complexes, including the van der Waals contribution (∆E vdw ), the electrostatic contribution (∆E ele ) and the solvation contribution (∆E sol ) and the total contribution (∆E total ) [27,28].

Circular dichroism (CD) spectroscopy
Tet(X3)/Tet(X4) in the presence of MNQ (128 μg/ mL) or DMSO was prepared and CD spectroscopy was employed to explore the effects of MNQ on the Tet(X3)/ Tet(X4) secondary structure.Briefly, CD spectra with the scanning wavelengths ranged from 190 to 250 nm at a 0.5 nm interval were recorded using a CD spectrometer and analysed via the BeStSel web server (https:// bests el.elte.hu/ index.php).

Western blotting analysis
Western blotting assay was employed to explore the effect of MNQ on the production of Tet(X3)/Tet(X4).Tet(X3)/Tet(X4)-positive bacteria co-incubated with different concentrations of MNQ (0, 16, 32, 64 μg/mL) were collected to analyze the expression of Tet(X3)/Tet(X4).Briefly, the samples were separated on 10% SDS-PAGE gels, and transferred to polyvinylidene difluoride (PVDF) membranes.After blocking with 5% (w/v) skimmed milk for 2 h, the membranes were incubated with the indicated primary antibodies at room temperature for 2 h and HRP-conjugated goat anti-mouse secondary antibodies for an additional 2 h.The blots were visualized using an enhanced chemiluminescence ECL kit (Biosharp, Beijing, China).

Metabolomic analysis
Bacterial culture of E. coli DH 5α + pAM401-tet(X4) in the presence or absence of MNQ (128 μg/mL) was prepared for subsequent experiments.Bacterial samples were analysed based on the Ultra High Performance Liquid Chromatography Q Exactive Mass Spectrometer (UHPLC-QE-MS, Thermo, Waltham, MA, USA) following the procedures described previously [30].The ionization source of LC-MS is electrospray ionization and both positive ion (POS) mode and negative ion (NEG) mode were employed to achieve maximal coverage for bacterial metabolites.Briefly, MS raw data involving the peak number, sample name and normalized peak area were processed with R package metaX and analysed for principal component analysis (PCA) and orthogonal partial least square-discriminate analysis (OPLS-DA).PCA indicated the distribution of origin data.Additionally, supervised OPLS-DA was further applied to detail group separation and facilitate classification.The results of screening differential metabolites were visualized using a volcano plot.By means of the Kyoto Encyclopedia of Genes and Genomes (KEGG, http:// www.kegg.jp) [31] and MetaboAnalyst (http:// www.metab oanal yst.ca/) [32], we further explored the key metabolic pathways represented by the differential metabolites and analysed the biological functional alterations.

In vitro cytotoxicity evaluation
Fresh washed rabbit erythrocytes in the presence or absence of MNQ, or the combination of MNQ and tetracycline were incubated at 37 °C for 1 h.And the culture supernates were harvested and measured at OD 570 nm using a microplate reader (BioTek SLXFA).
Additionally, Caco2 cells, HeLa cells and HEK-293 cells were incubated with the indicated concentrations of MNQ, either alone or in combination with tetracyclines for 5 h and the cell viability was detected by a Cytotoxicity Detection Kit (Roche, Basel, Switzerland) according to the manufacturer's instructions.

Animal study In vivo-toxicity evaluation
To evaluate the toxic effects of MNQ in vivo, BALB/c mice (6-8 weeks, n = 6 per group) were orally administrated with MNQ at 5 or 15 mg/kg body weight at 12-h intervals for continuously 5 days.The clinical characterizations including animal behaviors and body weight were observed and monitored daily for 10 days.At experimental end points, all mice were euthanized, followed by the histological observation of tissues and organs, then fixed in 4% formaldehyde and stained with haematoxylin and eosin (H&E) for pathological analysis.

Mice thigh infection model
Given that tetracyclines [33] and quinones [34,35] exhibited the similar pharmacokinetic parameters, such as the terminal elimination half-life (t 1/2 ) and time to maximal drug concentration (t max ) in the previous study, combination regimens to maximize the synergistic antibacterial activity in vivo were employed in the current study.
E. coli 47EC (1.0 × 10 8 CFU) was injected into the left thigh muscle of mice, which caused localized bacterial infection and abscess formation.MNQ (resuspended in 0.5% CMC-Na, 5 mg/kg or 10 mg/kg), tetracycline (TET, dissolved in water, 5 mg/kg or 10 mg/kg), and the combinations was administered intragastrically in the indicated concentrations, respectively.The vehicle-treated group received the equal amount of solvent control (0.5% CMC-Na) by oral administration.At 72 h post-infection, the infected thigh muscles of mice were homogenized, diluted and plated on MHB agar plates for CFU counting.Further, the infected muscles in each groups were fixed and preserved in 4% formaldehyde for pathological observation using haematoxylin and eosin (H&E) staining.

Mice systemic infection model
Exponentially growing E. coli DH5α + pAM401tet(X4) was collected and adjusted to 1.0 × 10 8 CFU.All mice were infected by intraperitoneal injection of bacterial cultures.And the mice were randomly divided into 4 groups: E. coli infection group, 5 mg/ kg MNQ (resuspended in 0.5% CMC-Na) treatment, 10 mg/kg methacycline (MET, dissolved in water) treatment, combined therapy of MNQ + MET (5 mg/ kg + 10 mg/kg).The therapeutic doses selection for E. coli DH5α + pAM401-tet(X4) systemic infection were referred by the thigh infection model described above.All the drugs were orally administrated at 12-h intervals, and at 48 h post-infection, all mice were euthanised for the histopathological lesion assessment.Briefly, the histopathological changes in the hearts, livers, spleens, lungs and kidneys in each group were assessed using H&E staining for observation.Additionally, livers and spleens were collected, homogenized and diluted with sterilized PBS for microbiological plating to evaluate the bacterial burdens.

Statistical analysis
Statistical analysis was performed using Graph-Pad Prism 8.0 software.All data are expressed as the means ± SEM.Sample size (n) was mentioned in the figure legends and the statistical significance was analyzed by the unpaired student's t-test method between two groups or one-way ANOVA (analysis of variance) among multiple groups to calculate P values (*P < 0.05; **P < 0.01; ns, P ≥ 0.05).Differences with P < 0.05 were regarded as significant.

Fig. 3
Fig. 3 Effect of MNQ, or in combination with tetracycline on the biofilm formation and maturation.E. coli DH5α + pAM401-tet(X4) biofilms were grown for 24 h with the presence or absence of MNQ and/or tetracycline (TET) (a-c).Additionally, the indicated concentrations of MNQ and/or TET were subsequently added at 12 h post-inoculation for an additional 12 h-incubation to assess the disruption of preformed E. coli biofilms by MNQ or the combinations (d-f).The graphs presented the OD 570 nm values of crystal violet (CV)-staining biofilms (a, d) and biomass (CFUs) quantified by the microbiological plating (b, e).c, f Confocal microscopy images of E. coli biofilms grown with or without MNQ and/or TET.Fluorescence intensity of each sample represents biofilms thickness.The representative image was chosen from three independent replicates.All data were repeated three times independently.**P < 0.01 and *P < 0.05 vs. the control group; ## P < 0.01 vs. MNQ monotherapy group

Fig. 5
Fig. 5 Synergistic mechanisms of MNQ-tetracyclines combination.a MNQ, or in combination with tetracycline (TET) triggered membrane damage, as evidenced by SEM observation.b E. coli DH5α + pAM401-tet(X4) cells treated with or without MNQ and/or TET.Live cells were stained green by SYTO9 and dead cells were stained red by PI.Further, percentages of live/dead cells present in each samples were quantified using ImageJ software.d-e Increased membrane permeability of propidium iodide (PI)-labeled E.coli cells with the addition of MNQ.MNQ (f) or in combination with TET (g) dissipated membrane potential of E. coli DH5α + pAM401-tet(X4).The drugs were supplemented into the cultures following DiSC 3 (5) dye incubation for 10 min, and the fluorescence intensity was continuously monitored for additional 30 min.h-i The effect of MNQ or the combinations on intracellular ∆PH was determined using fluorescence dye BCECF-AM-probed E. coli cells.j-k Accordingly, decreased level of ATP production in E. coli cells were exhibited in a MNQ-dependent manner.l-m β-galactosidase activity in the culture media of E. coli with the presence or absence of MNQ and/or tetracycline.All data were repeated three times independently, and the representative data were shown.**P < 0.01; *P < 0.05

Fig. 6
Fig. 6 MNQ deprives the function of multidrug-resistant efflux pumps.Efflux pump activities of E. coli DH5α + pAM401-tet(X4) with the treatment of MNQ or in combination with tetracycline were determined using the fluorescence dye Rhodamine (a-b) or Ethidium bromide (EtBr) (c-d).All data were repeated three times independently.**P < 0.01; *P < 0.05

Fig. 7
Fig. 7 Multivariate data analysis of bacterial metabolites.Volcano plot for the differentially expressed metabolites under the treatment of MNQ in the NEG mode (a) and POS mode (b).Red circles represent up-regulated metabolites.Green circles represent down-regulated metabolites.Grey circles represent non-significant metabolites.The enriched pathways related to the differential metabolites in the NEG mode (c, e) and POS mode (d, f) by KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis.The data are representative of six biological replicates

Fig. 8
Fig. 8 MNQ rescues tetracyclines activity in vivo.Scheme of the experimental protocols for the thigh infection model induced by E. coli 47EC (a) and E. coli DH5α + pAM401-tet(X4)-evoked systemic infection model (d).b In the mouse thigh infection model, bacterial load in infected thigh muscle were markedly decreased by combination therapy (n = 8 per group).c Histopathology and macroscopic observation of the infected thigh muscle.e Additionally, in the systemic infection model, the bacterial burden in the livers and spleens of mice challenged with a intraperitoneal inoculation of E. coli DH5α + pAM401-tet(X4) (n = 10 per group).f H&E staining for the histopathological changes of mice heart, liver, spleen, lung and kidney at 48 h post-infection.And the typical pathological lesions were marked using white