4-Hydroxy-2-pyridone analogs exemplified by PTC-847 and PTC-672 have been synthesized (Gerasyuto et al., 2016; Wang et al., 2016) to further explore the potential of this pharmacophore as a platform to uncover antibacterial agents targeting Ng. The minimum inhibitory concentrations (MIC) of PTC compounds were determined against six Ng strains (WHO strains, Table 1A) that represent a full range of drug resistant phenotypes relevant to current WHO gonorrhea treatment guidelines (Unemo et al., 2016), and ten Gram-negative pathogens and normally occurring gut organisms (Table 1B). Two of these strains, WHO F (strain 13477) and WHO K (13479), were selected for a number of subsequent in vitro and in vivo experiments due to their respective sensitivity and resistance to a quinolone antibiotic, ciprofloxacin, that exhibits antibacterial activity by selectively inhibiting DNA synthesis. PTC-847 exhibited potent antibacterial activity against all six Ng strains with MICs ranging from 0.05 to 0.1 µg/mL. The weakly immunogenic NMSB strain had a MIC of 0.2 µg/mL (Table 1A). PTC-672 exhibited MICs ranging from 0.05 to 0.4 µg/mL against the six Ng strains, and a MIC of 0.1 µg/mL against the NMSB strain (Table 1A). The compound PTC-565 representing the [5,6] fused ring structure (Figure 1A) exhibited antibacterial activity against a broad spectrum of Gram-negative pathogens (Table 1A and B); in contrast, PTC-847 and PTC-672 are inactive against the panel of Gram-negative pathogens and normal gut organisms evidenced by MICs ranging from 12.5 to ≥62.5 µg/mL (Table 1B).

The unique antibacterial activity against Ng strains with PTC-847 and PTC-672 (Table 1) provided the impetus to examine them against a panel of 206 Ng strains with broad resistance phenotypes collected by the GIST Programme at Public Health England (Bolt et al., 2016). Susceptibility testing resulted in MIC₉₀ values for PTC-847 and PTC-672 equal to 0.12 µg/mL (determined from values in Supplementary file 1). This collection includes strains lacking sensitivity to ceftriaxone and azithromycin, the current standard of care, and underscores that PTC-847 and PTC-672 are highly potent against Ng.

At concentrations of 1 X MIC or greater, PTC-847 and PTC-672 exhibit time-dependent cidality against the Ng 13477 strain (Supplementary file 2). In addition, a time-dependent post antibiotic effect was observed with PTC-847. Suppression of bacterial growth persisted even after the removal of PTC-847 (Supplementary file 3).

Our previous studies (Gerasyuto et al., 2018) established that a variety of [5,6]-fused indolyl-containing 4-hydroxy-2-pyridones (Figure 1A) targeted DNA synthesis, more specifically, fluoroquinolone resistant DNA topoisomerases, and inhibited a variety of Gram-negative pathogens. To determine the target(s) of PTC-847, we followed the incorporation of radiolabeled precursors into total nucleic acids, DNA, and proteins in the Ng 13477 strain. The results revealed that inhibition of DNA but not protein synthesis appears to be responsible for growth inhibition by PTC-847 (Supplementary file 4).

We then examined if inhibition of DNA topoisomerases (gyrase and topoisomerase IV) by PTC-847 was responsible for the observed DNA synthesis inhibition. Recombinant Ng 13477 gyrase (subunits A and B) and topoisomerase IV (both subunits) were cloned, expressed, and purified and the assay results compared to our previous similar studies on Ec DNA gyrase and topoisomerase IV. The PTC-847 half maximal inhibitory concentration (IC₅₀) value for Ng gyrase was 19 µM and >32 µM for Ng topoisomerase IV and both Ec topoisomerases (Table 2). Furthermore, PTC-847 at 100 μM did not appreciably inhibit the activity of human topoisomerase II. As a control, the fluoroquinolone antibiotic ciprofloxacin showed IC₅₀ values ranging from 0.4 to 18 µM (Table 2). These findings together suggest that topoisomerases are not the primary targets for PTC-847 in Ng.

In an effort to identify the target of inhibition of PTC-847, we used Ng 13477 (Table 1) to select and isolate a stable PTC-847-resistant isolate (PTC-847^R) (Supplementary file 5). The frequency of resistance for PTC-847^R at 4 x the MIC was 3.6 × 10^–8, compared to ciprofloxacin at 1.2 × 10^–8 and ceftriaxone at 7.7 × 10^–9 under identical conditions. PTC-847^R was tested against a wide variety of antibiotics having different modes of action (Supplementary file 6). The PTC-847^R strain was as sensitive to all classes of antibiotics as the wild-type Ng 13477 strain, except for susceptibility to PTC-847. The MIC for PTC-847 against Ng 13477 was 0.05 µg/mL compared to 15.6 µg/mL for the PTC-847^R isolate.

To identify the specific PTC-847 molecular target(s), the PTC-847^R isolate was subjected to whole genome sequencing using the Ng 13477 WT strain as the reference. A single C to T transition that results in a serine (S) to leucine (L) amino acid substitution at position 41 (S41L) in the α-subunit of the Ng class Ia RNR was identified (Figure 3). The gene sequence change was verified by comparing DNA sequences from two of the resistant PTC-847^R clones against the WT reference strain sequence. The target site conferring resistance to PTC-847 was further confirmed in an isogenic strain (designated PTC-847^S41L) made by transforming the WT 13477 strain with a DNA fragment containing the α-subunit gene mutation, followed by selection on agar plates containing PTC-847. Hundreds of PTC-847 resistant colonies were obtained when the 13477 strain was transformed with the mutated, but not the WT, α-subunit gene sequence.

A similar strategy using PTC-672 resulted in a stable resistant strain PTC-672^R. Sequencing the PTC-672^R class Ia RNR gene revealed a single A to G transition that resulted in a histidine (H) to arginine (R) amino acid substitution at position 25 (H25R) in the α-subunit (Figure 3). Thus, both inhibitors target the large subunit of the Ng Ia RNR that together with its small subunit are essential for DNA synthesis by catalyzing the conversion of NDPs to dNDPs. In support of the mechanism of inhibition of PTC-847, an increased ratio of NTPs to dNTPs in the susceptible WT 13477 strain was observed upon treatment with the inhibitor at 1 x MIC for 1 hr followed by nucleotide extraction and LC/MS analysis (Chen et al., 2009; Supplementary file 7).

The active form of the class Ia RNR is α₂β₂ (Figure 2B). The two subunits, however, have weak affinity with the K_D ~0.2 µM for Ec and human RNRs; therefore, each subunit was cloned, expressed, and purified independently (Greene et al., 2020). Furthermore, assay protocols are optimized for each RNR from each organism. For Ng α, an expressed tagged α subunit protein was purified by affinity chromatography and the tag removed to eliminate potential interference with the α₄β₄ quaternary structure (Figure 2B and C) and enzyme activity. Subunit β was cloned and expressed with an N-terminal (His)₆-tag and purified to homogeneity to give 0.7–0.9 diferric tyrosyl radical cofactors (Materials and methods, Supplementary file 8). The EPR spectrum of the Ng Ia β2 subunit is almost identical to that of the E. coli Ia β2 (Supplementary file 8). The activity of Ng RNR was optimized (Supplementary file 8) so that its inhibition by dATP and PTC-847 and PTC-672 could be determined. The activity at 0.1 µM and 1:1(α:β) ratio of subunits was 1300 nmol min^–1 mg^–1 with GDP/TTP as substrate and effector, similar to that for Ec RNR (2500 nmol min^–1 mg^–1 with CDP/ATP as substrate and effector; Materials and methods, Supplementary file 8). The apparent K_m is 0.03 µM for Ng RNR and as noted above that for Ec and human RNRs is ~0.2 µM. The lower K_m and the basis for the drop off in activity at >0.3 µM (Supplementary file 8) require further investigation.

All characterized class Ia RNRs have N-terminal cone domains within α-subunit (~100 amino acids) that control enzyme activity (A-site Figure 2A, ATP activates and dATP inhibits). Studies of Ng RNR activity as a function of dATP concentration (Supplementary file 9) reveal a profile very similar to Ec RNR, including biphasic kinetics based on dATP binding to both the S-site and A-site (Hofer et al., 2012). This result is expected given the 76% sequence identity for α between Ng and Ec, including the H25 and S41 (Figure 3) found altered in Ng resistant to the PTC compounds. Human α, on the other hand, shares only 30% sequence identity to Ng and does not have residues homologous to either H25 or S41. Inhibition studies with H25R and S41L Ngα mutants showed much lower sensitivity to dATP (Supplementary file 9).

The Ng GDP/TTP (Supplementary file 8) and [³H]-CDP/ATP (Supplementary file 9) assays were used to assess the ability of PTC-847 and PTC-672 to inhibit Ng RNR activity. The results are summarized in Table 3 and Figure 4—figure supplement 1. PTC-847 exhibited 93% inhibition at 4 μM PTC-847, whereas PTC-672 showed 78% inhibition at 2.5 μM. Similar experiments with Ec and human RNRs using the [³H]-CDP/ATP assay were also carried out. As expected, based on sequence similarity, the Ec RNR is also inhibited, whereas the human enzyme, even at 100 μM inhibitor concentration, still retains substantial activity (Table 3).

Inhibition of Ng RNR is enhanced in the presence of allosteric inhibitor dATP (Figure 4 and Figure 4—figure supplement 1), suggesting that in vivo, PTC-847 and PTC-672 could act by stabilizing a dATP-induced inhibited state of Ng RNR.

We have previously shown using negative stain EM analysis with Ec RNR that dATP shifts active RNR into the inhibited α₄β₄ state (Figure 2C), which is readily observed due to the ring structures (Ando et al., 2011). Comparable experiments with Ng RNR were carried out under very similar conditions to the inhibition studies and the results are shown in Figure 5. Panel A shows that, in the presence of 1 mM dATP, ring structures are present. The 2D classifications and 3D reconstruction are shown in Figure 5B and C, and reveal that Ng Ia RNR, like the Ec enzyme, forms the α₄β₄ ring structure inhibited state induced by dATP.

One hypothesis for RNR inhibition by the 4-hydroxy-2-pyridone derivatives, based on the mapped resistance mutations (Figures 3 and 5), is that they potentiate α₄β₄ formation. To determine if the selected resistant mutants affected the extent of α₄β₄ formation, PTC-672^H25R and PTC-847^S41L α subunits were prepared and purified by the same methods as for the WT proteins and examined as described in Figure 5. As shown in Figure 6, in the WT control, α₄β₄ ring structures are clearly visible whereas no such structures are observed with either mutant. An additional control in which ATP (an activator) replaces the dATP inhibitor, also reveals no α₄β₄ structures for either the WT or mutant proteins (data not shown).

The above data suggest that PTC-847 and PTC-672 may act through the formation/stabilization of Ng RNR’s α₄β₄ inactive state, explaining why residue substitutions that prevent α₄β₄ ring formation would lead to drug resistance in vivo. If ring formation were the only mechanism of inhibition, we would expect that PTC-672 and PTC-847 could not inhibit the H25R and S41L Ng RNR variants that cannot form rings. We pursued this idea and found that PTC-672 is far less effective at inhibiting these RNR variants than it is at inhibiting WT Ng RNR (Figure 6—figure supplement 1). However, a significant amount of inhibition is still occurring (Figure 6—figure supplement 1), especially compared to the substantial loss of inhibition for dATP for RNR variants (Supplementary file 9) that cannot form rings. This comparison suggests that PTC-672, and likely PTC-847 (not tested), has more than one mode of RNR inhibition.

A mode of inhibition does appear to be related to the formation/stability of α₄β₄ rings based on the above experiment. Thus, we further pursued whether PTC-672 and PTC-847 act as dATP mimics, that is binding to Ng RNR, and leading to increased α₄β₄ formation. Using mass photometry to evaluate oligomeric states (Figure 6—figure supplement 2, Figure 6—figure supplement 3), we find an increase in the number of α₄β₄ rings in the presence of PTC-672/GDP/TTP compared to GDP/TTP alone, but fewer rings than in the dATP control (Figure 6—figure supplement 2, Figure 6—figure supplement 3). For PTC-847, no increase in rings is observed (Figure 6—figure supplement 2, Figure 6—figure supplement 3). Based on these results and on the finding that PTC compounds potentiate the inhibitory effects of dATP (Figure 4), we postulate that in vivo, PTC-672/PTC-847 are not binding to the dATP site in the cone domain and inducing the conformational changes necessary for α₄β₄ formation but rather acting synergistically with dATP. Given the differences in structure between PTC-672/PTC-847 and dATP, having different binding sites is not surprising. The exact mechanism(s) by which PTC-672/PTC-847 act in concert with dATP, leading to an additionally inhibited state of Ng RNR, awaits further study.

In the following sections an in vivo mouse model for vaginal infection with drug susceptible and resistant strains of Ng was used to monitor infection clearance. Ng 13477 and 13479 (WHO F and K) susceptible and resistant isolates to PTC-672 (see Table 1) were constructed for in vitro and in vivo fitness testing. These isolates were used to monitor the development of resistance to the test compounds in vivo and to assess bacterial survival.

Ng has adapted for the human vaginal tract and will not grow in other species. With administration of high doses of estrogen to ovariectomized female mice and depletion of the normal vaginal bacterial flora, however, mice can be infected with Ng. (Raterman and Jerse, 2019). To better control the levels of estrogen, these mice are administered estrogen by slow-release pellet two days prior to infection. Concomitantly, they are administered vancomycin, streptomycin, and trimethoprim to deplete commensal flora. These studies are accompanied by administration of streptomycin (Strep) once daily for the remainder of the studies to out-compete the normal flora (for details see Materials and methods).

As a starting point for susceptibility testing with PTC-672, the Ng 13477 strain, which lacks antimicrobial resistance elements (Unemo et al., 2016), was tested against the MDR Ng 13479 strain that has a high level of resistance to quinolones and carries mutations associated with decreased susceptibility to cephalosporins (Unemo et al., 2016). Four new isogenic strains were created for the in vivo studies from Ng 13477 and 13479. One set was engineered to have PTC resistance (PTC-672^H25R) and the second to have resistance to both PTC-672 and streptomycin (Strep^R-PTC-672^H25R). They are designated Iso 13477 Strep^R-PTC-672^H25R and Iso 13479 Strep^R-PTC-672^H25R. The strains were created using a PCR product containing the H25R α-subunit gene and/or DNA isolated from the streptomycin resistant Ng FA1090 laboratory strain.

To test if normal intestinal flora are resistant to inhibition by these compounds, PTC-847 and PTC-672 MIC values were determined against a select subset of organisms representing normally occurring intestinal organisms (Thursby and Juge, 2017). PTC-847 and PTC-672 spared the commensal bacteria relative to the control antibiotic solithromycin (Supplementary file 10), a fourth-generation macrolide with enhanced activity against macrolide-resistant bacteria (Farrell et al., 2016). Furthermore, PTC-672’s effect on two commensal Neisseria strains was also assessed. These isolates were inhibited, but to a lesser extent (approximately 10-fold) than that of Ng (Supplementary file 11). The reason for the sensitivity differences observed amongst the Neisseria species is not known at this time. Based on the target, biological selectivity, and narrow anti-bacterial spectrum of PTC-847 and PTC-672, these compounds may be better tolerated compared with current therapies.

The efficacy of PTC-672 was assessed in the mouse model. On Day 0 of the study, mice were infected vaginally with the Ng Iso 13477 Strep^R or Iso 13479 Strep^R strains. Beginning on Day 1, mice were swabbed daily for 1 week to determine the bacterial load. On Day 2, mice were randomized into groups based on Day 1 bacterial load and administered vehicle, ciprofloxacin, ceftriaxone, or PTC-672. Efficacy is defined as complete and sustained clearance of infection at Day 5 post-treatment.

In mice infected with the Iso 13477 Strep^R strain (Figure 7A), those that received vehicle had a robust infection for 7 days, with 5/9 mice remaining infected through Day 7. Treatment with a single oral dose of ciprofloxacin at 30 mg/kg resulted in complete clearance of the infection in 10/10 mice (~4 log drop in bacterial infection) within 24 hr following treatment. PTC-672 administered as a single oral dose (10, 15, 20, 25, or 30 mg/kg) resulted in clearance of infection at all doses in 10/10 mice within 24 hr. The clearance of infection by ciprofloxacin or PTC-672 was sustained through Day 7.

In mice infected with the Iso 13479 Strep^R strain (Figure 7B), those that received vehicle had a robust infection for 7 days of study, with 8/10 mice remaining infected through Day 7 of study. Treatment with a single intraperitoneal dose of ceftriaxone (100 mg/kg) resulted in complete clearance of the infection in 10/10 mice within 24 hr and clearance of infection was sustained until Day 7. PTC-672 administered as a single oral dose (60 mg/kg) resulted in complete clearance of the infection in 10/10 mice within 24 hr of dosing and was also sustained until Day 7. At the lower dose of 30 mg/kg PTC-672 resulted in an average 1.46-log drop in bacterial load (8/9 mice cleared infection) 24 hr post-dose. However, infection in several mice re-bounded on subsequent days of the study suggesting that a dose of 30 mg/kg was sub-efficacious in the Iso 13479 Strep^R model.

The relative fitness cost associated with PTC-672 resistance was determined in vitro by the competition strains described above. Competitive fitness assays measure net growth of two populations of isogenic resistant and susceptible strains and account for differences between these populations across the full growth cycle including lag times, exponential growth rates, and stationary phase dynamics.

In this in vitro study, Iso 13477 Strep^R-PTC-672^H25R or Iso 13479 Strep^R-PTC-672^H25R and Iso 13477 Strep^R or Iso 13479 Strep^R were mixed, inoculated into growth medium, and aliquots taken at intervals over a 24 hr incubation period. At each sampling time point, the number of total (susceptible+ resistant) bacteria and resistant bacteria were determined by plating onto non-selective and PTC-672-supplemented plates, respectively.

Relative fitness is expressed as the competitive index (CI), the ratio of bacterial burdens between the PTC-672 resistant and susceptible strains at each time point divided by the baseline ratio at the beginning of the experiment. CI of ~1 indicates that the fitness of the resistant bacteria is similar to that of the parent bacteria, CI >1 indicates that the resistant bacteria are more fit, and CI <1 suggests that the resistant bacteria are less fit in the absence of selection pressure. In this experiment, the CI for Ng Iso 13477 Strep^R-PTC-672^H25R and Iso 13479 Strep^R-PTC-672^H25R decreased exponentially to a CI of 0.06 at 8 hr (Figure 8), demonstrating that the H25R engineered resistant bacteria are less fit for competitive infection with their respective parent strains while the isolates had similar in vitro doubling times.

For these experiments, mice pre-treated as described above were inoculated vaginally with either the isogenic Strep^R-PTC672^H25R strain or a mixture of the Strep^R parent and Strep^R-PTC672^H25R strains and swabbed daily over a 1-week period. At each sampling time point, the relative fitness expressed as the CI was determined by plating onto non-selective and PTC-672-supplemented plates. CI was calculated as the ratio of bacterial burdens between the PTC-672 resistant and susceptible strains at each time point divided by the baseline ratio at the beginning of the experiment.

When infected with only the Iso 13477 Strep^R-PTC-672^H25R or the Iso 13479 Strep^R-PTC-672^H25R strain, the infection was similar to that elicited with the Strep^R parent strain. A similar number of mice were infected on Day 1 and the infection level was comparable throughout the study. For fitness testing with mixed strains, the mean and median CIs for Ng Iso Strep^R-PTC-672^H25R 13477 and Iso 13479 Strep^R-PTC-672^H25R resistant strains decreased to <0.01 by Day 2 post-infection (Figure 9).

Thus, although the PTC-672-resistant Ng strains infection levels in mice were comparable to those for the isogenic PTC-672 susceptible parent strains, when the two isogenic strains were inoculated in the same anatomical location in vivo, the PTC-672-resistant isolates did not successfully compete with the parent strains. These data suggest that if PTC-672 resistance does arise in Ng strains in situ, the PTC-672 resistant strain would be at a competitive disadvantage and thus less likely to propagate. These results combined with the outcomes of the in vitro fitness study suggest that PTC-672 may lessen the emergence of resistance compared with current therapies.

PTC-565, PTC-847, and PTC-672 susceptibility testing for Ng and Nm and other Gram-negative pathogens and normal gut organisms was performed in accordance with the CLSI M07-A9 guideline (Clinical and Laboratory Standards Institute, 2012). The direct colony suspension procedure was used when testing Ng and Nm. Ng were tested on a gonococcal typing (GC) agar base with 1% defined growth supplement and Nm were tested on Mueller-Hinton Agar (MHA) supplemented with 5% sheep blood. Colonies from an overnight chocolate agar culture plate were suspended in 0.9% phosphate buffered saline (PBS), pH 7.0, to a 0.5 McFarland standard. Compounds serially diluted 2-fold were tested in a microtiter assay where all samples tested in duplicate. MIC determinations performed in 96-well plates using FB media (Takei et al., 2005). Reaction volumes were 200 µL, with cells in FB media mixed with compound. The final DMSO concentration was limited to 2.5%. Plates were incubated at 35 °C in the presence of 5% CO₂ for 48 hr. Turbidity was evaluated by determining the OD₆₀₀ values. The lowest concentration of drug that completely inhibited bacterial growth, in duplicate, was identified as the MIC.

The time-dependent kill and post antibiotic effect kinetic assays were performed in accordance with the CLSI M26-A guideline (Clinical and Laboratory Standards Institute, 1999 M26-A 1999). Aliquots from the cultures at the indicated time points were serially diluted 10-fold in PBS (10^–1 to 10^–6) which were then spotted onto agar plates for 24 hr and the colony forming units per mL (CFU/mL) determined in duplicate. The kill kinetics were represented graphically by plotting the log₁₀ CFU/mL against time at each concentration of the compound.

The DNA and protein synthesis assays were performed as described (Jyssum and Jyssum, 1979). Ng 13477 was grown overnight from a single colony in 20 mL fastidious broth (FB) liquid medium at 37 °C with 5% CO₂. The culture was then again diluted ~10 fold in FB and grown for 3 h to an OD of 0.4.

For protein assays, 23 mL of this culture was sedimented and washed in 50 mL minimal medium without leucine [10 mM NaH₂PO₄ pH 7.0, 12 mM KCl, 6 mM MgCl₂, 16 mM (NH₄)₂SO₄, 24 mM NaCl, 1 x BD BBL IsoVitaleX Enrichment (Thermo Fisher Scientific, Waltham MA), and 100 µM amino acids minus leucine]. The cells were then resuspended in 22 mL of this media. For the DNA synthesis assay, the cell culture was used directly without refreshing the medium.

Aliquots (200 µL/well) of these cultures were transferred to 96 well microplates. To each well, 5 µL of various 40 x concentrations of PTC-847 in 100% DMSO and a control with 100% DMSO were added (2.5% DMSO final). Then the [¹⁴C] leucine or [¹⁴C] uracil (1 µCi/mL), were added and the plates incubated at 35 °C with 5% CO₂ atmosphere for 3 h. At the end of the incubation, 90 µL was transferred to the wells of a 96-well filter plate containing an equal volume of 20% tricholoracetic acid (TCA). Another 90 µL was transferred to the wells of a 96 well polypropylene plate containing 10 µL of 3.5 N KOH at 37 °C.

In the former case, protein and total nucleic acids were TCA precipitated at 4 °C for 30 min and the resulting precipitate was collected on filters by centrifugation at 3000 x g for 6 min. In the latter case, RNA was hydrolyzed by KOH treatment at 37 °C overnight and 90 µL of each hydrolysate was transferred to a 96 well filter plate containing an equal volume of 20% TCA (at 4 °C). The remaining DNA was then TCA precipitated at 4 °C for 30 min and collected on filters by centrifugation.

In both workups, the filters were washed with 200 µL ice cold 10% TCA and then dried. The plates were sealed, scintillation cocktail (50 µL/well) was added, and the radioactivity analyzed by scintillation counting (Perkin Elmer).

Earlier purification attempts of individual subunits lead to solubility issues of gyrase B, and we therefore chose to co-express the subunits for purification. Plasmids (pET-Duet1) containing the S-tagged Ng parC and His-tagged parE genes (cloned into sites NcoI/PstI and XhoI/BglII, respectively) were transformed into the recombination-deficient Ec (recA) expression strain BLR(DE3) (Novagen) to avoid gene rearrangements. The ParC and ParE proteins were induced and purified using a previously published method (Pan and Fisher, 1999).

Single colonies were picked from plates and grown overnight at 37 °C in 50 mL of Luria-Bertani (LB) medium containing the selective antibiotic. A culture (10 mL) of the overnight growth was used to inoculate 450 mL of LB medium containing ampicillin (100 mg/mL) or kanamycin (50 mg/mL). Cells were grown at 30 °C for 3–4 hr, until the optical density at 600 nm reached 0.4–0.6. The culture was transferred to 18 °C and IPTG was added to a final concentration of 0.5 mM, and growth was continued overnight. Bacteria were harvested by centrifugation at 5000 x g for 15 min at 4 °C. The supernatant was discarded, and the bacterial pellet was stored at –80 °C.

The suspension was thawed on ice, and the pellet was resuspended in 20 mL of buffer A (20 mM Tris-HCl [pH 7.9], 500 mM NaCl, 5 mM imidazole, containing EDTA-free complete protease inhibitor (Roche)). Lysozyme (Sigma) and Triton X-100 was added to achieve a final concentration of 0.02% and 0.1% respectively. Incubation was continued on ice for another 30 min and then the mixture was sonicated and centrifuged at 35,000 x g for 60 min.

The supernatant was carefully removed to a 50 mL sterile tube and mixed with 2 mL volume of 50% Ni-NTA resin (Qiagen) which was pre-equilibrated in the buffer A. The tube was shaken gently on a Nutator overnight at 4 °C and then sedimented to remove the Ni-NTA flow through. The Ni-NTA resin was then loaded into a column and washed initially with 20 mL of buffer A, followed by washes with 20 mL of Buffer B (Buffer A plus 50 mM imidazole). The tagged ParC and ParE protein was sequentially eluted with 20 mL buffer C (Buffer A plus 200 mM imidazole) collected as fractions that were stored frozen at –80 °C. The column fractions were examined by western blot. The fractions containing the desired protein were pooled and dialyzed twice for 4 hr against 4 L of 20 mM Tris-HCl (pH 7.9), 250 mM NaCl, 1 mM DTT and 10% glycerol. The dialyzed solution was spun in a microcentrifuge at 16,000 x g for 10 min at 4 °C to remove any precipitate. Purified combined topoisomerase IV was flash frozen in aliquots and stored at –80 °C. The activity of Ng topoisomerase IV was measured using a kDNA decatenation assay and was comparable to commercially available Ec topoisomerase IV.

Ng gyrase isolation was performed using a modification from Gross et al., 2003. S-tagged gyrA and His-tagged gyrB were cloned into NcoI/PstI and XhoI/BglII, sites respectively in pET-Duet1 and transformed into BLR(DE3). Growth and lysozyme treatments were as described above for topoisomerase IV (Pan and Fisher, 1999) and the supernatant was mixed with Ni-NTA resin and shaken overnight at 4 °C. The resin was loaded into a column and the flow through containing S-tagged gyrase A was further purified using an S-protein agarose (Sigma) by elution with 20 mM Tris pH 7.5, 300 mM NaCl, 3 M MgCl₂.

His-tagged gyrase B was loaded into a Ni-NTA column, washed with 20 mM Tris-HCl [pH 8.0], 300 mM NaCl, 10% glycerol, 0.1% Triton X-100, eluted in the same buffer containing 200 mM imidazole. The purified protein judged by SDS PAGE were pooled and diluted 32-fold in buffer D (50 mM Tris HCl [pH 8.0], 2 mM dithiothreitol (DTT), 1 mM EDTA, 10% glycerol).

Further purification of gyrase A and gyrase B used a mono-Q (GE Healthcare) column in buffer D and a linear gradient elution from 0 to 1 M NaCl. Fractions were analyzed by SDS PAGE and Coomassie blue staining.

The fractions containing the proteins were pooled, concentrated and desalted using Ultrafree-15 Biomax-100 membrane centrifugal filter units (Millipore) in 10 mM Tris pH 7.9, 50 mM KCl, 0.1 mM EDTA, 2 mM DTT. Glycerol was added to the purified protein at a final concentration of 25% and flash frozen in liquid nitrogen, and stored at –80 °C. The protein concentrations were determined by A280 nm and calculated extinction coefficients of Ng GyrA (56,400 M^–1 cm^-1) and Ng GyrB (51,520 M^–1 cm^–1). Active gyrase (A₂B₂) was formed prior to the start of an assay using equal molar amounts (15 μM GyrA and GyrB).

Microplate-based supercoiling assays for Ec or Ng DNA gyrase were performed as described (Maxwell et al., 2006). Ec DNA gyrase was purchased (TopoGEN, Buena Vista, CO). Pierce black streptavidin coated 96-well microplates (Thermo Fisher Scientific, Waltham, MA) were rehydrated and washed 3 x in wash buffer (20 mM Tris-HCl, pH 8.0, 137 mM NaCl, 0.01% bovine serum albumin, 0.05% Tween-20). A total of 100 µL of 500 nM biotinylated triplex forming oligonucleotide (TFO1) was immobilized onto the streptavidin plate. Excess oligonucleotide was washed off using the wash buffer. Enzyme reactions (30 µL) containing 1 µg relaxed plasmid pNO1 DNA (Inspiralis Limited, Norwich, UK) in 35 mM Tris-HCl (pH 7.5), 24 mM KCl, 4 mM MgCl₂, 2 mM DTT, 1.8 mM spermidine, 1 mM ATP, 6.5% glycerol, 0.1 mg/ml albumin, and 1 U of Ec DNA gyrase (TopoGEN, Buena Vista, CO) or DNA gyrase purified as described above from the Ng 13477 strain was incubated at 37 °C for 30 min. A total of 100 µL of TF buffer (50 mM sodium acetate, pH 5.5, 50 mM NaCl, 50 mM MgCl₂•6 H₂O) was then added to the reaction and the entire mixture transferred to the microplate wells after the TFO1 immobilization process. The microplate was incubated at room temperature for 30 min to allow triplex formation. Unbound plasmid was washed off with 3 × 200 µL of TF buffer, and 200 µL of 1 X SYBR Gold (Invitrogen, Carlsbad, CA) in 10 mM Tris-HCl (pH 8.0), 1 mM EDTA was added and allowed to stain for 20 min. Fluorescence was read using the EnVision Multimode Plate Reader (Perkin-Elmer, Waltham, MA).

These topoisomerases can convert the large network of concatenated kinetoplast DNA (kDNA) from C. fasciculata into decatenated DNA. Topoisomerase decatenation assays are based on the principle that decatenated DNA can be separated from concatenated DNA by microfiltration and quantified by fluorescence. Decatenation assays using purified human DNA topoisomerase II (ProFoldin, Hudson, MA), Ec DNA topoisomerase IV (ProFoldin, Hudson, MA), and Ng DNA topoisomerase IV purified in-house from the Ng 13477 strain as described were performed according to ProFoldin’s protocols for microplate-based topoisomerase DNA decatenation assays using kDNA (ProFoldin, Hudson, MA) and in-house reagents. The separated decatenated product was stained using the Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA) followed by fluorescence quantitation on an EnVision Multimode Plate Reader.

Ng 13477 and PTC-847^R were grown overnight in FB medium and then subcultured. Two Ng 13477 and one PTC-847^R subcultures were allowed to grow for 4 hr. Then PTC-847 (0.05 µg/ml) was added to one of the Ng 13477 subcultures and incubated an additional 1 hr at 37 °C with CO₂. The bacteria were collected by centrifugation for 15 min at 4000 x g at 4 °C. The pellets were resuspended in 5 mL cold acidified ACN (65% acetonitrile, 35% water, 100 mM formic acid) and extracted on ice for 30 min with periodic mixing. The extractions were sedimented for 20 min at 30,000 x g at 4 °C. The supernatants were collected, frozen on dry ice, lyophilized to dryness, and subjected to LC/MS. ¹³C9,¹⁵N₃-CTP was added for LC/MS analysis and peak areas of NTP and dNTPs were normalized to the peak area of ¹³C9,¹⁵N3─CTP (Chen et al., 2009).

The gene for the α subunit of Ng Ia RNR was optimized for expression in Ec and cloned into a vector with an N-terminal pET(His)₆SUMO-Kan-N. The protein was expressed, purified to homogeneity by Ni-affinity chromatography, and the tag cleanly removed using the SUMO protease (Parker et al., 2018). Subunit β was cloned and expressed with an N-terminal hexahistidine (His)₆- tag that resulted in homogeneous protein with variable amounts of metallo-cofactor. The diferric tyrosyl radical cofactor was self-assembled with Fe²⁺ and O₂ to give 0.7–0.9 Y•s/β2 using the protocol of Yee et al., 2003 except that 10 molar equivalents of O₂ instead of 3.5 was used and the ferrous ammonium sulfate was dissolved in buffer instead of water.

To optimize RNR activity a 1:1 ratio of α₂ and β₂ subunits was examined over a physiological concentration range of 0.01–10 µM. The specific activity (SA) of RNR was determined in a reaction mix of 100 µM Ec TR, 1 µM Ec TRR, and 0.2 mM NADPH at 37 °C. The SA for 0.01–0.12 µM subunits was measured with 1 mM GDP, 0.25 mM TTP, and 0.2 mM NADPH using the continuous spectrophotometric assay. The SA for 0.2–10 µM was measured with 1 mM 5-[³H]-CDP, 3 mM ATP and 2 mM NADPH by the discontinuous radioactive assay. Assay detection methods followed the standard protocols (Ge et al., 2003; Ravichandran et al., 2020).

The stock solution of 0.48 mM PTC-672 was prepared in 50% DMSO aqueous solution and stored at –20 °C. The working solution was a 50-fold dilution of the stock solution to a final concentration of 9.6 μΜ in assay buffer with 1% DMSO. The concentration was determined by A348 nm with an ε = 24.4 mM^–1cm^–1.

The assay for inhibition of Ng RNR by PTC-847 and PTC-672 was performed spectrophotometrically using the GDP/TTP protocol with the addition of various concentrations of PTC-847 (0.1–8 µM) or PTC-647 (0.025–2.5 µM). The inhibition of RNR by dATP was performed spectrophotometrically with the addition of various concentrations of dATP (2 nM to 1 µM). At concentrations of dATP from 50 to 200 µM, [³H]-CDP/ATP and the discontinuous assay was used.

Negative stain EM samples contained 1:1.5 α:β, 15 ng/µL protein. Concentrations of α and β were determined using A280 nm (ε = 89,050 M^–1cm^–1 for α and 61,310 M^–1cm^–1 for β). αs (WT and mutants) and β were first combined on ice at 150 ng/µL and incubated for one min in the presence of specified nucleotides, also at 10 x concentration. The samples were then diluted tenfold in RNR assay buffer (50 mM HEPES pH 7.6, 15 mM MgCl₂, 1 mM EDTA) and 5 μL of the mixture was applied to carbon-coated 300 mesh Cu grids (Electron Microscopy Sciences) after glow discharging for 1 min at –15 mA in an EasiGlo glow discharger (Ted Pella). The protein was adsorbed to the grid for 1 min and excess liquid was removed by blotting. The grid was stained 3 x with 5 μL of 2% uranyl acetate (VWR) with blotting immediately after application for the first two rounds of staining. The final uranyl acetate stain was allowed to sit on the grid for 45 s before blotting. All blotting was completed manually with filter paper (Whatman, grade 40).

Screening images for the Ng RNR negative stain reconstruction were taken on a Tecnai Spirit (FEI) instrument with a XR16 camera (AMT) operated at 120 kV at ×68,000 magnification. Screening images for the H25R and S41L α variant samples were taken on a Tecnai F30 (FEI) operated at 200 kV equipped with an FEI Falcon II direct electron detector. Five to 10 grid squares were examined, and a representative field is shown for each protein sample. For the EM reconstruction of Ng RNR, images were taken on a Tecnai F20 Twin (FEI) operated at 120 kV equipped with a US4000 CCD detector (Gatan) at 1.79 Å/pix.

All processing for negative stain data sets was completed in Relion 3.0.8 (Scheres, 2012). Micrographs were imported and CTF-corrected prior to manual picking of 1000 particles subdivided into five classes to train Relion’s autopicking algorithm. Autopicking yielded ~27,000 particles for the inactivated complex. These particles were 2D classified (100 classes, 25 iterations) and an initial 3D model was created (no symmetry imposed) prior to 3D classification (one class, 100 iterations, D2 symmetry) and 3D auto-refinement. Auto re-finement was completed on 12,469 particles and yielded a structure at 21 Å resolution (0.143 FSC). Following autorefinement, the inactive Ec RNR structure (PDB 5CNS) was docked into the Ng RNR EM map using the Chimera MaptoModel functionality (Pettersen et al., 2004; Zimanyi et al., 2016).

Mass photometry (interferometric scattering mass spectrometry) was performed on a Refeyn instrument using AcquireMP v2.4.1 and DiscoverMP v2.4.2. All movies were taken for a length of 60 s using the default parameters and Gaussian curves were fit to each histogram distribution using the DiscoverMP software. All samples were at a final concentration of 2.5 ng/µL protein with a 1:1 α:β ratio. Samples were incubated for 2 min at a concentration of 50 ng/µL at 37 °C prior to final dilution on the instrument and data recording. Concentrations of additives were as follows: 1 mM GTP, 0.25 mM TTP, 0.2 mM dATP, 3 mM ATP, 25 µM PTC-672 (0.45% DMSO), 8 µM PTC-847 (0.04% DMSO).

All animal studies were performed under IACUC approved protocols at AAALAC-certified animal facilities. Female ovariectomized Balb/c mice (Charles River, Wilmington, MA) were obtained at 3 weeks of age and acclimatized for ~1 week prior to initiation of a fitness or efficacy study. Day 0 was defined as the day mice were inoculated vaginally with ~1 × 10⁹ bacteria suspended in 20 µL saline. A single 17-β-estradiol pellet (0.5 mg, 21day release) was then implanted in mice with a trochar on Day –2. Mice were then treated with three antibiotics to control commensal flora induced by the estradiol. Vancomycin HCl (0.6 mg) and streptomycin sulfate (0.3 mg) were administered intraperitonealy and trimethoprim sulfate (0.8 mg) was administered by oral gavage twice daily from Day –2 to Day 1 of study. On Day 2 of the study the administration of all antibiotics except streptomycin sulfate was stopped. Streptomycin sulfate was then administered daily from Day 2 throughout the remainder of the study.

The relative fitness cost associated with PTC-672 resistance was determined in vivo by competition experiments using the validated mouse model of vaginal Ng infection described above. This study evaluated the relative fitness of the PTC-672 resistant strains Iso 13477 Strep^R-PTC-672^H25R and Iso 13479 Strep^R-PTC-672^H25R in competition with the isogenic PTC-672 susceptible parent strains Iso 13477 Strep^R and Iso 13479 Strep^R. On Day 0 of the study, mice pre-treated as described above were inoculated vaginally with either the Strep^R parent strain, the isogenic Strep^R-PTC672^H25R strain, or a mixture of the Strep^R parent and the isogenic Strep^R-PTC672^H25R strains suspended in saline and swabbed daily over a one week period. At each sampling time point, the numbers of total (susceptible+ resistant bacteria) and resistant bacteria were determined by plating onto non-selective and PTC-672-supplemented plates, respectively. Relative fitness was expressed as the CI, calculated as the ratio of bacterial burdens between the PTC-672 resistant and susceptible strains at each time point divided by the baseline ratio at the beginning of the experiment.

The efficacy of an increasing single oral dose of PTC-672 administered by oral gavage was evaluated in the mouse model described above. The antibiotics, ciprofloxacin and ceftriaxone, were used as positive controls for efficacy. Efficacy was defined as sustained clearance of infection for 5 days post dosing in 10/10 mice in a treatment group.

On Day 0 of the study, mice pre-treated as described above were inoculated vaginally with Iso 13477 Strep^R or Iso 13479 Strep^R bacteria suspended in saline. On Day 2 mice were randomized into treatment groups based on Day 1 CFU counts. All mice were treated on Day 2 with a single oral dose of either vehicle comprising 0.5% hydroxypropyl methylcellulose and 0.1% Tween 80, ciprofloxacin, ceftriaxone, or PTC-672. In the Iso 13477 Strep^R portion of the study, PTC-672 was tested at doses of 10, 15, 20, 25, and 30 mg/kg and ciprofloxacin at 30 mg/kg. In the Iso 13479 Strep^R portion of the study, PTC-672 was tested at doses of 30 and 60 mg/kg and ciprofloxacin at 100 mg/kg. Mice were vaginally swabbed daily using a sterile swab over a one-week period. At each sampling time point, the numbers of resistant bacteria were determined by plating onto PTC-672-supplemented plates and the level of infection was determined for each test mouse in each treatment group.