In our day to day experiments, we regularly use agar plates containing lawns of pathogenic bacteria. During these experiments, we often find random environmental bacteria growing on these plates. On one lawn of N. gonorrhoeae we observed an environmental contaminant that was surrounded by a zone of growth inhibition suggesting that it produced an anti-N. gonorrhoeae metabolite (Figure 1a). When we screened this contaminant for antibacterial activity against lawns of other Gram-negative bacteria it did not produce a zone of growth of inhibition against any of the bacteria we tested (e.g., Escherichia coli, Vibrio cholerae, Caulobacter crescentus). Since antibiotics that preferentially inhibit the growth N. gonorrhoeae are rare, we looked at this contaminant in more detail. Sequencing of the contaminant’s genome and genome clustering analysis (Appendix 1—figure 1a) revealed that it was most closely related to Bacillus amyloliquefaciens, which is a root-colonizing bacterium that is used as a biocontrol agent (Ngalimat et al., 2021). We named the anti-N. gonorrhoeae contaminant B. amyloliquefaciens BK.

To identify the biosynthetic gene cluster (BGC) responsible for the observed antibiosis we screened B. amyloliquefaciens BK transposon mutants for strains that no longer produced anti-N. gonorrhoeae activity. The sequencing of the non-producer strain revealed that it surprisingly contained four transposon insertions and one frame shift mutation (Appendix 1—figure 1b). The frame shift mutation and one transposon insertion were predicted to each disrupt unique BGCs. The transposon inserted into the bacillomycin (mycosubtilin) BGC, while the frame shift mutation was predicted to disrupt the (oxy)difficidin BGC. To determine which of these two BGCs was responsible for the N. gonorrhoeae activity we tested B. amyloliquefaciens strains with targeted disruptions of each BGC for activity against N. gonorrhoeae (Koumoutsi et al., 2004). Only disruption of the (oxy)difficidin BGC eliminated the anti-N. gonorrhoeae activity (Appendix 1—figure 1c). To confirm the identity of the N. gonorrhoeae active antibiotic we carried out a bioassay guided fractionation of B. amyloliquefaciens BK culture broth (Wilson et al., 1987; Zimmerman et al., 1987). High-resolution mass spectrometry (HRMS) and nuclear magnetic resonance (NMR) data from the major active peak we isolated were consistent with its being an oxydifficidin isomer (Figure 1b, Appendix 1—figures 5–12, and Supplementary file 2). Oxydifficidin contains a 27-carbon polyketide backbone that is cyclized through a terminal carboxylic acid and an oxidation at position 21. This hydrophobic core is phosphorylated at C16. Oxydifficidin occurs naturally as a collection of interconverting thermal isomers (Figure 1b). As reported previously, we observed an interconversion of isomers with the compound we purified from B. amyloliquefaciens BK cultures (Wilson et al., 1987; Zimmerman et al., 1987). All assays were performed using the mixture of interconverting oxydifficidin isomers we obtained from B. amyloliquefaciens BK cultures.

We tested oxydifficidin for activity against diverse bacterial pathogens. Oxydifficidin showed only weak activity against most pathogens, however we observed potent activity against N. gonorrhoeae (Figure 1c). When we examined other Neisseria species, we found that oxydifficidin was consistently more active against Neisseria than any of the other bacteria we tested. Among Neisseria spp., oxydifficidin was most active against N. gonorrhoeae underscoring a unique narrow spectrum of potent activity. A key issue with the current treatment of N. gonorrhoeae infections is the development of resistance to existing therapeutics. Resistance to the standard of care cephalosporins is particularly problematic. Oxydifficidin was more potent against N. gonorrhoeae MS11 than most other antibiotics we tested. Notably, unlike clinically used antibiotics such as ceftriaxone, azithromycin, and ciprofloxacin, oxydifficidin retained activity against all multidrug-resistant clinical isolates we examined (Table 1).

Oxydifficidin’s structure is interesting as phosphorylated antibiotics, and moreover natural products in general, remain rare (Petkowski et al., 2019; Cao et al., 2019). Its structure together with its specific and potent activity against drug resistant N. gonorrhoeae suggested it might have a unique mode of action (MOA). This is appealing from the perspective of developing therapeutics that are capable of circumventing clinically problematic resistance mechanisms and therefore we focused on characterizing the mechanism of oxydifficidin’s potent anti-N. gonorrhoeae activity.

As a first step to understanding the MOA of oxydifficidin, we raised resistant mutants by directly plating N. gonorrhoeae cultures on antibiotic containing plates (1 μg/ml, 4× minimum inhibitory concentration [MIC]) (Figure 2a). Out of the >1.5 × 10¹⁰ cells we screened, only one resistant mutant appeared. Oxydifficidin’s MIC against this mutant increased by eightfold (2 μg/ml). No increase in MIC was observed for any other antibiotic we tested (Figure 2a). Sequencing and comparison of this mutant’s genome to the sensitive parent genome revealed a single mutation that introduced a frame shift in one of the three predicted dedA genes found in the N. gonorrhoeae MS11 (NCBI:txid528354 dedA NGFG_RS04905; Jen et al., 2021). To confirm that DedA was necessary for oxydifficidin’s potent activity, we created a dedA deletion mutant in a clean N. gonorrhoeae background. This mutant showed the same eightfold increase in oxydifficidin’s MIC, confirming that the deletion of dedA is sufficient to reduce oxydifficidin’s potent activity. We also generated deletion mutants for two other predicted dedA-like genes, and the MIC of oxydifficidin for these mutants remained the same as for the N. gonorrhoeae MS11 wild-type (WT) strain. Interestingly, not only was dedA deficient N. gonorrhoeae less susceptible to oxydifficidin, oxydifficidin also kills this mutant more slowly (Figure 2b) than WT N. gonorrhoeae MS11. The dedA deletion mutant also showed an altered cell morphology with reduced membrane integrity and lower formation of microcolonies (Appendix 1—figures 4). A survey of 220 N. gonorrhoeae strains with high-quality assemblies in NCBI found no mutations in the DedA protein.

The DedA protein superfamily is highly conserved, with examples in almost every sequenced genome across all domains of life (Doerrler et al., 2013). DedA family members are predicted to be transmembrane proteins with still largely, poorly defined, functions. However, a few recent studies indicate that DedA homologs are flippases. They have been reported to flip phospholipids (phosphatidylethanolamine, phosphatidylserine) or phospholipid-like structures (C55-isoprenyl pyrophosphate) across prokaryotic and eukaryotic lipid bilayers (Sit et al., 2023; Roney and Rudner, 2023; Huang et al., 2021; Li et al., 2021). Although oxydifficidin is not a phospholipid, its overall structure resembles that of reported DedA protein substrates, especially C55-isoprenyl pyrophosphate (Figure 3a). Interestingly, among the two characterized bacterial family members, the DedA protein associated with oxydifficidin potency is most closely related to the C55-isoprenyl pyrophosphate flippase YghB from V. cholerae (Appendix 1—figure 2; Sit et al., 2023). As oxydifficidin’s activity was not completely abrogated in the dedA knockout we postulated that DedA was not the direct target of oxydifficidin, but it instead acted to increase oxydifficidin’s potency in N. gonorrhoeae. The structural similarity between oxydifficidin and the known substrates of DedA homologs led us to explore the possibility that DedA was responsible for assisting with oxydifficidin uptake into N. gonorrhoeae.

We examined the effect of DedA on antibiotic accumulation by comparing the amount of compound found in the cell pellet collected from antibiotic treated cultures of WT and dedA knockout N. gonorrhoeae strains (Figure 3b). In the case of oxydifficidin we saw six times more antibiotic in cells from WT cultures than from dedA deletion strain cultures. For the other antibiotics we tested (tetracycline and chloramphenicol) this ratio was less than two. Based on DedA homologs flipping phospholipid-like structures across a lipid bilayer our data is consistent with DedA flipping oxydifficidin across the inner membrane to increase its cytoplasmic concentration and in turn increase its potency against N. gonorrhoeae. A DedA-assisted uptake mechanism could also explain the slower rate of killing we observed for oxydifficidin against dedA deficient N. gonorrhoeae compared to WT N. gonorrhoeae. While we cannot definitely rule out the possibility that DedA accumulation of oxydifficidin in the membrane could also have a direct toxicity effect, we did not detect any cell lysis or membrane depolarization at even 100 times oxydifficidin’s MIC (Appendix 1—figure 3).

To look for an intracellular target of oxydifficidin we carried out a second round of resistant mutant screening. In this case, we used the N. gonorrhoeae dedA deletion strain and searched for colonies with an even higher tolerance to oxydifficidin. From ~1 × 10¹⁰ cells plated on 8 μg/ml oxydifficidin (4× MIC for N. gonorrhoeae dedA deletion strain) we identified 12 resistant mutants. In each case, the oxydifficidin MIC increased to 16 μg/ml. Sequencing of these strains revealed that each contained a point mutation in the gene encoding for the large ribosomal protein(s) L7/L12 (rplL). Eleven strains contained the same R76C mutation and one contained a K84E mutation (Supplementary file 1). These two mutations were not found in the survey of the same collection of N. gonorrhoeae strains used to look for DedA mutations.

To determine if mutations in the rplL gene alone were sufficient to confer oxydifficidin resistance, we created an RplL R76C mutant in a WT N. gonorrhoeae background (i.e., non dedA deletion). This mutant exhibited an eightfold increase in oxydifficidin’s MIC (2 μg/ml) compared to the parent strain, confirming that the R76C mutation in L7/L12 alone was sufficient to increase the MIC of oxydifficidin. Ribosomal proteins L7 and L12 have the same sequence, however L12 has an N-terminal acetylation (Diaconu et al., 2005). L7/L12 is part of the L10/L7 stalk of the large (50S) subunit of the bacterial ribosome and is critical to a number of processes including GTP hydrolysis (Diaconu et al., 2005; Carlson et al., 2017).

The appearance of resistance mutations in rplL suggested that oxydifficidin inhibited protein synthesis, which would be consistent with isotope feeding studies performed with difficidin and E. coli (Zweerink and Edison, 1987). We initially tested this hypothesis in vitro using a coupled transcription/translation system. While this reaction mixture contained all the components necessary to produce a protein from DNA, no protein was produced in the presence of oxydifficidin (Figure 3c). Using a coupled system, it was not possible to distinguish between inhibition of RNA or protein synthesis, and therefore, to rule out inhibition of transcription, we next looked directly at RNA synthesis in vitro. In this case we saw no inhibition of RNA synthesis, even at the highest oxydifficidin concentration we tested (13.4 μg/ml) (Figure 3d). Taken together these two experiments indicate that oxydifficidin inhibits translation but not transcription and are consistent with rplL mutations providing resistance to oxydifficidin.

To the best of our knowledge rplL mutations have not been previously associated with antibiotic resistance and no characterized antibiotics have been found to bind L7/L12. When we screened for cross resistance, the L7/L12 R76C mutation did not confer resistance to any other antibiotics we tested (Table 2). These included ribosome targeting antibiotics with diverse binding sites. Antibiotics that bind different regions of the 30S decoding center, including tetracycline, gentamicin, and spectinomycin showed no increase in MIC with the L7/L12 R76C mutation (Lin et al., 2018; Paternoga et al., 2023). Similarly, the L7/L12 R76C mutation did not confer resistance to chloramphenicol or erythromycin, which interact with distinct regions of the 50S peptidyl transferase center (Lin et al., 2018; Paternoga et al., 2023). Antibiotics that bind away from these two hot spots also showed no cross resistance (Lin et al., 2018; Paternoga et al., 2023; Polikanov et al., 2018). The activity of avilamycin, which binds in a unique site in the 50S subunit at the entrance to the A-site tRNA accommodating corridor was also not affected by the L7/L12 R76C mutation (Arenz et al., 2016). In the case of thiostrepton A (Harms et al., 2008), which targets the L11 protein GTPase-associated center and is part of the only class of antibiotics known to bind the ribosome in the vicinity of L10/L7 stalk, we also observed no change in MIC for N. gonorrhoeae with the L7/L12 R76C mutation. The inability of the L7/L12 R76C mutation to confer cross resistance to known antibiotics suggests that oxydifficidin binds to a different site on the ribosome and that L7/L12 is uniquely associated with oxydifficidin’s MOA.

To determine whether the DedA and RplL mutations we observed were specific to oxydifficidin resistance in N. gonorrhoeae, we selected for resistant mutants using two other Neisseria species: Neisseria subflava and Neisseria cinerea. Both strains were natively less susceptible to oxydifficidin than N. gonorrhoeae. In the case of N. cinerea all resistance mutants contained the RplL K84E mutation that we first observed in N. gonorrhoeae, and in the case of N. subflava all resistance mutants contained mutations in dedA (Supplementary file 1). Other Neisseria species likely show higher MICs than N. gonorrhoeae because of either the absence of DedA-assisted uptake or reduced RplL protein oxydifficidin sensitivity. The model that arises from our studies is that high oxydifficidin sensitivity results from a combination of DedA-assisted oxydifficidin uptake into the cytoplasm and the presence of an oxydifficidin-sensitive RplL protein (Figure 3e). N. gonorrhoeae is uniquely sensitive to oxydifficidin because it contains both. DedA is critical to oxydifficidin’s potent activity as it not only sensitizes N. gonorrhoeae to oxydifficidin but also decreases its kill time, both of which are appealing from a clinical development perspective.

Neisseria strains were grown at 37°C with 5% CO₂ in GCB medium (Freitag et al., 1995) (VWR CA90002-016). B. amyloliquefaciens BK was isolated from a lab agar plate, B. amyloliquefaciens FZB42 WT and various mutant strains were provided by Bacillus Genetic Stock Center. All Bacillus strains were grown at 30°C in Luria-Bertani (LB) medium. E. coli DH5α, Klebsiella pneumoniae ATCC 10031, Enterobacter cloacae ATCC 13047, Acinetobacter baumannii ATCC 17978, Pseudomonas aeruginosa PAO1, Morganella morganii ATCC 25830, and Staphylococcus aureus USA300 were grown at 37°C in Tryptic Soy Broth (TSB).

Bacillus strains were grown overnight at 30°C on LB agar. 3 ml of methanol was added to the top of the plate and spread uniformly by gentle shaking. The supernatant was then collected, centrifuged at 15,000 rpm for 2 min and filtered through a 0.2-μm cellulose acetate membrane (VWR). Then, 10 μl of the crude extract was added onto a paper disc (VWR) and after 5 min drying the paper disc was transferred on top of GCB agar plate lawned (swabbed from a single colony) with N. gonorrhoeae cells. The plate was incubated at 37°C with 5% CO₂ overnight and the inhibition zone was visually inspected.

A single colony of B. amyloliquefaciens BK was grown in 250 ml LB in a 1-l flask at 30°C on a rotary shaker (200 rpm) overnight, then 50 ml of the overnight culture was transferred to 1 l modified Landy medium (20 g glucose, 5 g glutamic acid, 1 g yeast extract, 1 g K₂HPO₄, 0.5 g MgSO₄, 0.5 g KCl, 1.6 mg CuSO₄, 1.2 mg Fe₂(SO₄)₃, 0.4 mg MnSO₄ per 1 l deionized H₂O, pH adjusted to 6.5 before autoclaving) in a 2.8-l triple baffled flask and fermented at 30°C on a rotary shaker (200 rpm) for 3 days.

Genomic DNA from B. amyloliquefaciens BK WT and transposon mutant Tn5-3 was isolated using PureLink Microbiome DNA purification kit (Invitrogen) according to the manufacturer’s instructions. The B. amyloliquefaciens BK WT genome was assembled by mapping its sequencing data onto the annotated genome of B. amyloliquefaciens FZB42 using Geneious Prime. Differences in the mutant strain Tn5-3 were identified by mapping its sequencing data onto the assembled B. amyloliquefaciens BK WT genome. The mutated genes were then annotated using NCBI BLAST. The oxydifficidin BGC was annotated using the antiSMASH online server.

The isolation process followed the published protocol (Wilson et al., 1987) with adjustments. The fermentation culture (5 l) was centrifuged at 4000 rpm for 10 min and the supernatant was acidified to pH 3.0 before extraction with 5 l of ethyl acetate. The crude extract was concentrated under reduced pressure (60 mBar) on a rotary evaporator to a final volume of about 20 ml. The concentrated sample was then loaded onto Diaion HP-20 resin (Sigma-Aldrich) and the resin was then washed by 500 ml of water, 500 ml of methanol–water (3:7) and 500 ml of methanol–water (7:3). The oxydifficidin rich cut was then obtained by eluting the resin with 500 ml of methanol. This rich cut was concentrated and absorbed onto DEAE-Sephadex A-25 (Cl⁻) resin (Sigma-Aldrich). The resin was washed with 250 ml of methanol–water (2:8), 500 ml of methanol–water (9:1), and a rich cut was then obtained by eluting the resin with 100 ml methanol–water (2:8) with 3% ammonium chloride. The rich cut from A-25 resin was then diluted with 300 ml of water and adsorbed onto Amberlite XAD-16N resin (Sigma-Aldrich) and washed successively with 500 ml of water, 250 ml of 0.075 M K₂HPO₄ (pH 7), and 250 ml of methanol–water (2:8), followed by elution with 100 ml of methanol. This oxydifficidin-containing eluate was diluted with 100 ml of 0.075 M K₂HPO₄ (pH 7) and applied to a RediSep Rf C18 column (Teledyne ISCO, 100 Å, 20–40 µm, 300 ± 50 m²/g) with a reverse-phase linear gradient system (5–100% 0.075 M K₂HPO₄/MeOH, 45 min). Fractions containing oxydifficidin was monitored by anti-N. gonorrhoeae activity and further purified by semipreparative HPLC using a reverse-phase C18 column (Waters,130 Å, 5 µm, 10 mm × 150 mm) with an isocratic system (0.075 M K₂HPO₄/MeOH (32:68)). The purified oxydifficidin was then desalted by Amberlite XAD-16N resin (Sigma-Aldrich).

LC–HRMS data were acquired using a SCIEX ExionLC HPLC system coupled to an X500R QTOF mass spectrometer (The Rockefeller University). The system was equipped with a Phenomenex Kinetex PS C18 100 Å column (150 mm × 2.1 mm, 2.6 µm) and operated with SCIEX OS v.2.1 software. The following chromatographic conditions were used for LC–HRMS: 5% B to 2.0 min; 5–95% B from 2.0 to 30.0 min; 95% B from 30.0 to 37.0 min; 95–5% B from 37.0 to 38.0 min; and 5% B from 38.0 to 45.0 min (flow rate of 0.25 ml/min, 3 µl injection volume, A/B: water/acetonitrile, supplemented with 0.1% (vol/vol) formic acid). Both electrospray ionization (ESI) modes Full HRMS spectra were acquired in the range m/z 100–1000, applying a declustering potential of 80 V, collision energy of 5 V, source temperature of 500°C and a spray voltage of 5500 V. A maximum of seven candidate ions from every Full HRMS event were subjected to Q2-MS/MS experiments in the range m/z 60–1000, applying a collision energy of 35 ± 10 V for both ESI modes. Full HRMS and the most intense MS/MS spectra were analyzed with MestReNova software (14.3.0). NMR spectra were acquired on a Bruker Avance DMX 800MHz spectrometer equipped with cryogenic probes (New York Structural Biology Center). All spectra were recorded at 298 K in CD₃OD–D₂O (1:1). Chemical shift values are reported in parts per million (ppm) and referenced to residual solvent signals: 3.31 ppm (1H) and 49.0 ppm (C). Spectra analysis and visualization were carried out in TopSpin (3.6.0) and MestReNova (14.3.0).

Neisseria strains were grown on GCB agar plate overnight, cells were collected by a polyester swab tip (Puritan) and resuspended in 1 ml GCB medium followed by a 1 in 5000-fold dilution. The stock solution of antibiotics was serial diluted twofold in a 96-well plate (Thermo Fisher Scientific) containing GCB liquid medium, then an equal amount of diluted bacterial culture was added to each well and mixed by pipetting. The plates were incubated at 37°C with 5% CO₂ for 16 hr. E. coli, K. pneumoniae, and M. morganii strains were grown at 37°C in TSB medium overnight. The stock solution of antibiotics was serial diluted twofold in a 96-well plate containing TSB agar medium, then 10⁴ cells were spotted on each well. The plates were incubated at 37°C for 16 hr. The top and bottom rows of plates contained an equal volume of media alone to prevent edge effect. The last column did not contain antibiotic to serve as a control for cell viability. The MIC of antibiotics was determined by visual inspection as the concentration in the well that prevents bacterial growth compared to control wells. The MIC assays were performed in duplicate.

The phylogenetic trees for bacterial genomes were built using Genome Clustering function with default parameters in MicroScope (Vallenet et al., 2020). The bacterial DedA family protein sequences were downloaded from a previous study (Todor et al., 2023) and aligned using MUSCLE v5 (Edgar, 2022) with the ‘Super5’ algorithm. The phylogenetic tree was built using FastTree (Price et al., 2010) with the Jones–Taylor–Thornton (default) mode of amino acid evolution. All trees were visualized and edited in iTOL.

Transposon mutagenesis was performed by electroporation using the EZ-Tn5 <KAN-2> Tnp Transposome kit (Lucigen). Briefly, an overnight culture of B. amyloliquefaciens BK was diluted 100 fold in NCM medium (Ito and Nagane, 2001) and grown at 37°C on a rotary shaker (200 rpm) until an OD600 nm of 0.5. The cell culture was then supplemented with glycine, DL-threonine and Tween 80 at a final concentration of 3.89%, 1.06%, and 0.03%, respectively, and grown at 37°C on a rotary shaker (200 rpm) for 1 hr. The culture was then cooled on ice for 30 min and collected by centrifugation (4000 × g, 10 min) at 4°C. The cell pellet was washed three times with ETM (Zhang et al., 2011) buffer. The washed cell pellet was resuspended in 100 μl of ETM buffer containing 0.25 mM KH₂PO₄ and 0.25 mM K₂HPO₄ and 1 μl of the EZ-Tn5 transposome was added. The cell mixture was transferred to a 1-mm electroporation cuvette and pulsed using a Gene Pulser Xcell Electroporation System (Bio-Rad) using 2.1 kV/cm, 150 Ω, and 36 μF. The mixture was immediately and gently mixed with 1 ml NCM medium containing 0.38 M mannitol and recovered in a round-bottom tube (VWR) at 37°C on a rotary shaker (120 rpm) for 3 hr. After recovery, the cells were concentrated by centrifugation and spread on an LB plate with 50 μg/ml kanamycin and incubated at 37°C overnight and single colonies collected for subsequent analysis. A library containing 50 transposon mutants was obtained. In the mutants examined, each strain contained ≥4 transposon insertions.

The transposon mutants of B. amyloliquefaciens BK were grown overnight in LB medium at 30°C. Each overnight culture was then diluted 1:5000, and 1 μl of the diluted culture was spotted onto a GCB agar plate swabbed with N. gonorrhoeae cells. The plate was then incubated overnight at 37°C with 5% CO₂. The mutant strain (Tn5-3) lacking anti-N. gonorrhoeae activity was identified due to its failure to produce a zone of growth inhibition in the resulting N. gonorrhoeae lawn.

For natural mutagenesis, single colonies of N. gonorrhoeae cells were swabbed on GCB agar plates and grown at 37°C with 5% CO₂ for 16 hr. The grown cells were collected by polyester swab tip (Puritan) in GCB medium and cell number was calculated by measuring optical density (OD, estimated value of OD 0.7 = 5 × 10⁸ CFU (colony-forming unit)/ml). Cells were then uniformly spread out on GCB agar plate containing 4× MIC of oxydifficidin (1 μg/ml for MS11 and 8 μg/ml for MS11 ∆dedA) and grown at 37°C with 5% CO₂ until mutant colonies are observed.

For dedA gene deletion, the 3′ (Primer: dedA 3′ overhang F/R) and 5′ (Primer: dedA 5′ overhang F/R) overhang regions of N. gonorrhoeae dedA gene and a kanamycin resistance cassette (Primer: kan^R (kanamycin resistance) cassette F/R) were amplified using GoTaq master mix (Promega). The fragments were then assembled using NEBuilder HiFi DNA assembly master mix (NEB) to place the cassette sequence positioned in the middle. Transformation of the construct to N. gonorrhoeae was done by using spot transformation protocol (Dillard, 2011). For mutant rplL_R76C, the mutated rplL gene including the 3′ and 5′ overhangs (Primer: rplL_R76C F/R) was amplified using N. gonorrhoeae MS11 ∆dedA rplL_R76C as the PCR template. Another construct incorporating a kanamycin resistance cassette with the 3′ (Primer: trpB-lga 3′ overhang F/R) and 5′ (Primer: trpB-lga 5′ overhang F/R) overhangs from locus 272,353 bp (trpB-lga) of N. gonorrhoeae MS11 genome was assembled using HiFi assembly. These two constructs were co-transformed (Dalia, 2018) into N. gonorrhoeae MS11 cells. The sequence verification was carried out using Sanger sequencing services provided by Genewiz.

Genomic DNA from Neisseria parent strains and resistance mutants was isolated using PureLink Microbiome DNA purification kit (Invitrogen) according to the manufacturer’s instructions. The whole-genome sequencing was performed by a MiSeq Reagent Kit v3 using Illumina MiSeq system following the manufacturer’s instructions. Genomes were assembled and mapped using corresponding reference genomes. The single-nucleotide polymorphism was detected by aligning the mutant’s sequencing reads to the genomic sequence of the parent strains.

N. gonorrhoeae cells were grown on a GCB agar plate at 37°C with 5% CO₂ for 16 hr and the overnight cells were normalized to ~10⁵ CFU/ml with GCB medium in a 14-ml round-bottom tube (VWR). The normalized cell culture was supplemented with 8× MIC of oxydifficidin (2 μg/ml for MS11 and 16 μg/ml for MS11 ∆dedA) and incubated at 37°C with 5% on a rotary shaker (200 rpm). 100 μl of the culture was collected and mixed thoroughly into 10 ml of GCB medium at 5, 10, and 30 min, respectively, then 200 μl of the diluted culture was spread on a GCB agar plate and grown at 37°C with 5% CO₂ overnight. Colonies were counted the following day. The killing assays were done in triplicate.

The cell lysis assay was conducted using SYTOX green (Invitrogen) following the manufacturer’s instructions. Briefly, 1 ml of cell suspension in Live Cell Imaging Solution at OD 0.7 was stained by adding 0.4 μl of 5 mM SYTOX solution and incubated in the dark at RT for 5 min. Subsequently, 30 μl of the stained culture was added into a 384-well Flat Clear Bottom Black plate (Corning) and the fluorescence signal was recorded by an Infinite M NANO⁺ (TECAN) with Excitation/Emission = 488/523 nm. The interval time was set to 7 s. Once the signal reached equilibrium, 30 μl of antibiotics diluted in Live Cell Imaging Solution at 16× MIC was added to the culture and mixed thoroughly. The fluorescence signal monitoring was then continued for 1 hr. The cell lysis assays were done in triplicate.

The cell depolarization assay was conducted using DiSC₃(5) dye (Invitrogen) following the manufacturer’s instructions. Briefly, 1 ml of cell suspension in Live Cell Imaging Solution at OD 0.7 was stained by adding 1 μl of 2 mM SYTOX solution and incubated in the dark at RT for 15 min. Then, 30 μl of the stained culture was added into a 384-well Flat Clear Bottom Black plate (Corning) and the fluorescence signal was recorded by an Infinite M NANO⁺ (TECAN) with Excitation/Emission = 622/675 nm. The interval time was set to 7 s. Once the signal reached equilibrium, 30 μl of antibiotics diluted in Live Cell Imaging Solution at 16× MIC was added to the culture and mixed thoroughly. The fluorescence signal monitoring was then continued for 1 hr. The cell depolarization assays were done in triplicate.

N. gonorrhoeae cells were grown on a GCB agar plate at 37°C with 5% CO₂ for 9 hr and the overnight cells were normalized to ~10⁸ CFU/ml with GCB medium in a 14-ml round-bottom tube (VWR). Cells were incubated with 0.125 μg/ml of oxydifficidin at 37°C with 5% CO₂ on a rotary shaker (200 rpm) for 3 hr. Cell pellets were collected by centrifugation at RT, washed twice with fresh GCB medium, and resuspended in an equal volume of GCB medium to that of the supernatant. The cell suspension and supernatant were then separately extracted by ethyl acetate in a 1:1 ratio. The extracts were evaporated to dryness under vacuum, resuspended in 100 μl of methanol and then 3 μl of the resuspension was injected to LC–HRMS for the quantification of oxydifficidin using target peak area. Experiment 1 (oxydifficidin and tetracycline) was done in duplicate, and experiment 2 (oxydifficidin and chloramphenicol) was done in triplicate.

The in vitro transcription assay was conducted using the HiScribe T7 High Yield RNA Synthesis Kit (NEB) following the manufacturer’s instructions. 24 μM of oxydifficidin was added to the transcription mixture, and 20 mM of EDTA was added to provide a positive control for inhibition. The yielded RNA was detected by a BioAnalyzer (Agilent). The in vitro translation assay was performed using PURExpress In Vitro Protein Synthesis Kit (NEB). 24 μM of antibiotics were added individually to the translation mixture and incubated at 37°C for 4 hr, the samples were then analyzed by SDS–PAGE in a 5–20% gradient gel.

The membrane integrity assay was conducted using BacLight Bacterial Viability Kit (Invitrogen) following the manufacturer’s instructions. Briefly, 1 ml of cell suspension at OD 0.7 was stained by adding 3 μl of the dye mixture and incubated in the dark at RT for 15 min. Subsequently, 1 μl of the stained culture was added onto the center of a round cover glass (Warner Instruments), a 1 mm × 1 mm GCB agar pad was overlayed on the culture. The sample was then imaged by an Eclipse Ti Microscope with a 60× objective (Nikon). DIC, GFP, and Texas Red fluorescence channels were applied to the imaging. ImageJ was used to analyze all images. Membrane integrity was assessed by calculating the ratio of the number of green cells to red cells. Boiled cells were served as the negative control. The killing assays were done in 10 replicates.

100 μl of cell suspension at OD 0.7 were added to 1 ml of fresh GCB medium in a 6-well plate (VWR) and incubated at 37°C with 5% CO₂ for 3 hr. Pictures were then taken using a 20× objective (Nikon).

30 μl of cell suspension at OD 0.7 was spotted onto 12 mm glass or Aclar film round coverslips coated with 0.1% Poly-lysin. A drop of fixative (100 μl of 2% glutaraldehyde, 4% formaldehyde in 0.1 M sodium cacodylate buffer pH 7.2) was added on top of the coverslip for 30 min. Additional fixative was then added to the dish containing the coverslips and left it in the fridge overnight. Samples were gently washed three times with buffer for 5–10 min each time and postfixed with osmium tetroxide 1% in 0.1 M sodium cacodylate buffer pH 7.2 for 1 hr at RT. After rinsing three times with buffer, samples were dehydrated in a graded series of ethanol concentrations 30%, 50%, 70%, and 90% for 10 min each and three times in 100% ethanol with molecular sieves for 15 min each and critical point dried in a Tousimis Autosamdri 931. Samples were coated with 10 nm of iridium using a Leica ACE600 sputter coater. Imaging was done in a JEOL JSM-IT500HR at 5.0 kV.

Statistical analysis was performed using Prism 10 (GraphPad). Group data are presented as means with SEM. The significance was determined using one-way ANOVA (and nonparametric or mixed). p values less than 0.05 were considered significant.