New Broad-Spectrum Antibiotics Containing a Pyrrolobenzodiazepine Ring with Activity against Multidrug-Resistant Gram-Negative Bacteria

: It is urgent to ﬁ nd new antibiotic classes with activity against multidrug-resistant (MDR) Gram-negative pathogens as the pipeline of antibiotics is essentially empty. Modi ﬁ ed pyrrolobenzodiazepines with a C8-linked aliphatic heterocycle provide a new class of broad-spectrum antibacterial agents with activity against MDR Gram-negative bacteria, including WHO priority pathogens. The structure − activity relationship established that the third ring was particularly important for Gram-negative activity. Minimum inhibitory concentrations for the lead compounds ranged from 0.125 to 2 mg/L for MDR Gram-negative, excluding Pseudomonas aeruginosa , and between 0.03 and 1 mg/L for MDR Gram-positive species. The lead compounds were rapidly bactericidal with >5 log reduction in viable count within 4 h for Acinetobacter baumannii and Klebsiella pneumoniae . The lead compound inhibited DNA gyrase in gel-based assays, with an IC 50 of 3.16 ± 1.36 mg/L. This study provides a new chemical sca ﬀ old for developing novel broad-spectrum antibiotics which can help replenish the pipeline of antibiotics.


■ INTRODUCTION
The fight against bacterial infections is rapidly being lost as microbes develop multiple mechanisms to evade antibiotics.Antimicrobial resistance is one of the most significant health concerns worldwide.It is estimated that 25,000 people die each year in Europe and 23,000 people die each year in the United States because of resistant bacterial infections, with more than 2 million infections caused by drug-resistant bacteria. 1These values are expected to reach 10 million deaths globally by 2050, overtaking cancer and heart disease as the principal cause of death if no effective measures can be developed. 2Of particular concern is the rapid emergence of multidrug-resistant (MDR) Gram-negative pathogens, defined as having resistance to at least three frontline antibiotic classes, and there are increasing reports of essentially pan-drugresistant (PDR) isolates in the clinic.In the last 30 years, no major classes of broad-spectrum antibiotics have been introduced to the market, and recently, approved agents such as linezolid (2000), daptomycin (2003), and retapamutilin (2007) are active only against Gram-positive pathogens. 3This makes it imperative that we discover new antibacterial drugs with a broad spectrum of activity or specificity for Gramnegative species, in order to help solve the crisis and provide new treatment options for MDR and PDR strains.
Pyrrolobenzodiazepines (PBDs) are naturally occurring molecules produced by Streptomyces bacteria whose family members include anthramycin and tomaymycin. 4,5PBDs have a soft N10−C11 imine electrophile that can covalently bond to guanine bases. 6They have been extensively studied as anticancer agents 7−11 and, more recently, a large number of PBDs are being clinically evaluated as payloads for antibody drug conjugates (ADCs) demonstrating the broad therapeutic utility of this chemical class. 12,13We recently described a series of C8-linked PBD monomers that showed activity against Gram-positive MDR strains, 14,15 which were able to inhibit Staphylococcus gyrase in a biochemical assay.It has been reported in the literature that the reason PBDs are only active against Gram-positive bacteria is due to their inability to cross Gram-negative membranes. 16We designed and synthesized a new generation of C8-PBD monomers with an aliphatic third ring that showed, for the first time, notable activity against Gram-negative bacteria.The introduction of the aliphatic third ring improved the prokaryotic selectivity of the molecules and reduced eukaryotic toxicity of the molecules, as the third ring interfered with the DNA binding ability of these molecules.
The synthesized compounds showed a broad-spectrum activity against both MDR and PDR clinical ESKAPE strains with minimum inhibitory concentrations (MICs) ranging from 0.03 to 1 mg/L against Gram-positive species and from 0.125 to 32 mg/L against Gram-negative species.The compounds demonstrated a rapid bactericidal mode of action against both Gram-positive and Gram-negative species.The mechanism of action of this type of compound was explored both in silico and through a range of microbiological techniques.These methods suggest that the inhibition of DNA gyrase is one of the main mechanisms of action, although the possibility of activity through other mechanisms cannot be completely ruled out.The efflux liability and entry into Gram-negative species of the chemical series and the likelihood of the emergence of resistance was also studied.Importantly, the relative lack of DNA binding, as observed by fluorescence resonance energy transfer (FRET)-based DNA melting, combined with the absence of eukaryotic toxicity, highlights the potential therapeutic value of this new type of C8-linked PBD monomers as antibacterial compounds.

■ RESULTS AND DISCUSSION
Design and Synthesis of PBD Monomers.Three C8linked biaryl-PBD monomers with reported activity against Gram-positive species were selected as the starting point for the medicinal chemistry modification (Figure 1).These monomers contain C8-substituents with two heteroaromatic or biaryl building blocks connected via amide bonds.The side chains were elongated with a third aliphatic heterocycle, resulting in an unconventional class of PBD derivatives that  Journal of Medicinal Chemistry have never been investigated before.Initially, three nitrogencontaining aliphatic heterocycles, thiomorpholine, morpholine, and piperidine, were introduced, and this was followed by coupling the heteroaromatic ring containing PBD monomers with diethyl and dimethylamine.The structural modification aimed to modify the physicochemical properties of the molecules and occupy a different chemical space, as polar compounds have traditionally shown better activity against Gram-negative bacteria. 17The secondary aim was to reduce the DNA binding affinity of the compounds to minimize eukaryotic toxicity, as the replacement of an aromatic group with an aliphatic ring was expected to reduce both interaction and fit of the molecules within the DNA minor groove.A total of 15 target compounds were designed and successfully synthesized to explore the structure−activity relationship (SAR) (Figure 2) and understand the role of the third rings in extending the antibacterial activity of the compounds.C8linked monomers 1 (GWL-78) and 2, in which the heteroaromatic pyrrole ring replaced the third aliphatic ring, were used as the control to compare their activity against Gram-negative bacteria.
The target compounds were synthesized using a convergent strategy based on literature-reported procedures.The synthetic process is based on the C8 derivatization of an alloc, tetrahydropyran (THP)-protected PBD core with different polyamide side chains.The protected PBD core 26 was synthesized in nine steps starting from vanillin accordingly to a reported procedure from our group (Figure 3).The synthetic methods and characterization of intermediates that lead to 26 are included in the Supporting Information file (chemistry section).The three-component side chains were synthesized by the sequential addition of the different constituents via amide coupling.Alternative strategies based on either the use of protecting groups or functional groups interconversion were employed to avoid side reactions.For the synthesis of benzofused containing lateral chains (Scheme 1), the initial step involved the formation of an amide bond between Bocprotected N-methyl pyrrole carboxylic acid and the corresponding 5-amino benzofused methyl ester to give inter- mediates 27 and 28.The ester derivative underwent basic hydrolysis to give the carboxylic acids that were successfully coupled with diethylamine, piperidine, morpholine, or thiomorpholine to give the complete C8-side chains 31−38.The coupling system EDCI/DMAP was used as the reagent for the formation of the amide bonds with good yield with the only exception of diethylamine products.In this case, the coupling between the carboxylic acid derivatives of 27 and 28 and the secondary amine was performed using a propylphosphonic anhydride (T 3 P)-sustained dehydration reaction in the presence of N,N-diisopropylethylamine (DIPEA).The synthesis of dimethylamine derivatives 29 and 30 was accomplished with direct coupling between Boc-protected Nmethyl pyrrole carboxylic acid and corresponding commercially available 5-amino benzofused dimethyl carboxamides.The coupling system EDCI/DMAP was used for the formation of the amide bond similar to previous examples.
Two different approaches were employed, depending on the aliphatic substitution of the products, for the synthesis of methylpyrrole benzenamine (MPB) containing lateral chains (Scheme 2).In the case of piperidine, morpholine, and thiomorpholine substitution, the MPB central building block 39 was synthesized via Suzuki coupling between Br-methyl pyrrole carboxylic acid and the para-Boc-protected benzenamine boronic acid.After acid-catalyzed Boc-deprotection, the aniline derivative of 39 underwent the amide coupling reaction with Boc-protected N-methyl pyrrole carboxylic acid.The reaction was performed with an EDCI/DMAP coupling system to give intermediate 40.The synthetic process was completed by basic hydrolysis of the methyl ester moiety, and successive EDCI/DMAP sustained amide coupling reaction with piperidine, morpholine, or thiomorpholine to give complete side chains 41−43, respectively.This synthetic approach failed in the formation of dimethyl and diethyl amide derivatives because of the poor yield of the amide coupling reaction of the considered secondary amines using traditional coupling reagents.
For this reason, a reverse-direction synthetic process, with aliphatic derivatization as the first step, was applied for the synthesis of dimethyl and diethyl-substituted MPB containing side chains.Specifically, the synthetic route was modified with the initial synthesis of dimethyl and diethyl N-methyl Br pyrrole intermediates 44 and 45 via amide bond formation based on acyl chloride formation.Commercially available 4-Br-N-methyl pyrrole methyl ester was hydrolyzed, and the corresponding carboxylic acid was activated to acyl chloride using oxalyl chloride in the presence of catalytic dimethylformamide (DMF).The reaction mixture was then treated with the corresponding secondary amine to give the desired product.
Intermediates 44 and 45 underwent the Suzuki coupling reaction with para-Boc-protected benzenamine boronic acid to give MPB-modified derivatives 46 and 47.The obtained intermediates were deprotected via acid treatment and then successfully coupled with Boc-protected N-methyl pyrrole carboxylic acid to give final complete lateral chains 48−49.

Journal of Medicinal Chemistry
Finally, the convergent synthetic strategy was completed with the coupling of C8-side chains 29−38, 41−43, and 48− 49 to the PBD core 26 using EDCI/DMAP-mediated amide coupling.This reaction was followed by the palladiumcatalyzed deprotection of allyl carbamate moiety and concerted THP deprotection of the alcohol moiety to generate the reactive N10−C11 imine bond, giving the final products 3−17 (Scheme 3).
Screening of PBDs To Define SAR.Bacterial panels containing drug-sensitive and MDR strains of Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Staphylococcus aureus, and Enterococcus spp.were used as an initial screen to determine the efficacy of the PBD compounds in this series and establish a SAR (Tables 1 and 2) (resistance profiles for all strains are in Table S1).All of the designed PBD molecules had a third aliphatic ring or a ring fragment and contained pyrrole as the first unit followed by MPB, benzofuran, or benzothiophene as the second unit.Against the Gram-negative species tested, the control PBD molecule 2 did not show any notable activity, with MICs ranging from 64 to >128 mg/L (Table 1).However, the aliphatic third ring containing C8-linked PBDs (compounds 3−17) showed notable improvement in activity against Gram-negative bacteria, except for P. aeruginosa, with MICs as low as 0.5 mg/L against MDR A. baumannii and K. pneumoniae strains.
Compounds bearing GC-targeting methylphenylbenzenamine (MPB) substituent as a central group (compounds 13, 15 to 17) showed good efficacy against A. baumannii and K. pneumoniae but were inactive against P. aeruginosa.Compounds containing benzofuran as the second ring were found to be more active than the corresponding benzothiophene series.The structural modification of the third aliphatic ring showed a significant influence on the activity of the compounds.In all cases, compounds with thiomorpholine with the third ring (example compounds 7, 12, and 17) had MICs of ≤2 mg/L against the A. baumannii strains tested.The presence of a second heteroatom in the aliphatic third ring improved activity, but intriguingly, shortening the third ring to a dimethyl or diethyl side chain (example compounds 3, 4, 8, 9, and 13) retained activity, with dimethyl analogues showing better activity than their diethyl counterparts, except for compounds containing the MPB ring in which diethyl substitution, compound 14, resulted in loss of activity.
Against the Gram-positive species tested, all compounds, including the control PBD molecule, showed excellent activity, with MICs ranging from ≤0.03 to 2 mg/L (Table 2).
Two compounds (7 and 8) with promising activity against both Gram-negative and Gram-positive strains were selected for further screening against a wider panel of bacteria including MDR Gram-negative strains of Escherichia coli, A. baumannii, K. pneumoniae, and Burkholderia cepacia/cenocepacia complex.The K. pneumoniae strains selected for the study were resistant  Journal of Medicinal Chemistry against most marketed antibiotics (Tables 3 and S1) including one essentially PDR strain (51851).Similarly, the A. baumannii strains and E. coli strains containing extended spectrum β lactamase (ESBL) and carbapenemase genes were resistant to six to eight clinically used antibiotics (Table S1).Both compounds were found to be active against almost all strains tested except for Burkholderia multivorans C1962 against which compound 7 was found to be inactive (defined here as an MIC > 32 mg/L).The levels of activity observed, considering the high levels of intrinsic or acquired resistance in many of these strains, are exceptional, with low and in some cases sub μM MICs.Notably, compound 8 exhibited a range of MICs from 0.25 to 2 mg/L for K. pneumoniae, 0.125 to 1 mg/L for A. baumannii/baylyi, 0.5 to 1 mg/L for E. coli, and 0.125 to 2 mg/ L for B. cepacia complex.These values are exciting and suggest that this represents a promising new chemical scaffold that can be further developed as broad-spectrum antibiotics.
Defining the Mode of Action of PBDs through Time-Kill Assays.The mode of action of compounds 7 and 8 was explored using time-kill assays.Three K. pneumoniae (NCTC 13368, 16 and NCTC 13438) and three A. baumannii (AYE, ATCC 17978 and NCTC 13424) strains were selected for time-kill analysis (Figure 4).These strains were selected to provide a range of drug resistance profiles from a relatively drug-sensitive strain (ATCC 17978), through MDR strains with multiple resistance mechanisms (NCTC 13438, 16 and NCTC 13424) for each species.Compound 8 was rapidly bactericidal in all strains tested; the viable count was below the limit of detection (LOD) within 4 h and continued to be below the LOD for the full assay for all K. pneumoniae strains challenged with 8, indicating that there was no resistance to this compound within the 24 h period of the assay.A small number of viable cells were observed at the end of 24 h for A. baumannii strains AYE and NCTC 13424 (Figure 4).Subsequent testing with 8 on these populations showed no increase in MIC above wild-type levels, and these populations were not considered to be resistant (results not shown).Compound 7 was also bactericidal against the majority of  Journal of Medicinal Chemistry strains tested, although the speed of killing was slower than compound 8 in a number of cases (NCTC 13438, AYE, NCTC 13424).In the case of K. pneumoniae NCTC 13368, which had a much higher intrinsic resistance to compound 7, only a small reduction in viable count was observed at 6 h, with bacterial numbers returning to the levels of the untreated control by 24 h, in two out of three replicate experiments.When these bacteria were isolated, passaged 10× in the absence of selection and re-assayed, the MICs increased for both compound 7 (32 to ≥128 mg/L) and compound 8 (2 to >32 mg/L).No increases in MIC were observed for other antibiotics, including aminoglycosides, fluoroquinolones, and β-lactams, except for an increase in colistin resistance of 4−32 mg/L in one of the two isolates (Table S2).
Evaluation of the Influence of Efflux and Influx on the Activity of the PBD Compounds.Two efflux pump inhibitors (EPIs), PAβN (used in the presence of Mg 2+ to prevent nonspecific effects on membrane permeability) 18 which competitively inhibits specific resistance nodulation division (RND) efflux pumps and carbonyl cyanide m-chlorophenyl hydrazone (CCCP), which inhibits the proton motive force, were used to assess the impact of efflux on the activity of the lead compounds against wild-type bacterial strains, together with the resistant mutants isolated from the time-kill experiments.
The result (Table 4) suggests that both compounds are substrates for RND-family efflux pumps, but the presence and/ or expression of specific efflux pumps differs between the strains tested.To investigate this further, specifically in the P. aeruginosa strain PAO1, transposon mutants in various RND efflux pumps were obtained from the Manoil collection (Univ.Washington; Table S3).Transposon mutants in any component of the tripartite RND efflux system MexAB-OprM reduced the MIC by ≥8-fold for both compounds 7 and 8.The MIC of compound 8 was also decreased ≥4-fold by transposons in MexCD but not its outer membrane component OprJ; no such effect was observed for compound 7 with transposon mutants in any component of this pump.
A membrane permeabilizer, polymyxin B nonapeptide (PMBN), used at a concentration which did not affect

Journal of Medicinal Chemistry
PMBN also reduced the MIC of compound 7 against an extended panel of P. aeruginosa strains from >32 to 1−8 mg/L (8-to ≥64-fold).The activity of these compounds in P. aeruginosa strains in the presence of a permeabilizer shows that these compounds are effective at killing strains once they are able to cross the outer membrane barrier.This offers the opportunity to optimize the active compounds using medicinal-chemistry modifications or to explore active delivery mechanisms to allow the compounds to permeate the Pseudomonas membrane.
Understanding the Mechanism of Action.Effect of the Aliphatic Third Ring Substitution on DNA Binding and Eukaryotic Toxicity of the Synthesized Compounds.The cytotoxicity of C8-linked PBD monomers is linked with their ability to stabilize DNA sequences.A FRET-based DNA melting assay 7 was used to explore the effect of the aliphatic third ring on the DNA stabilization property of the newly synthesized molecules.Two fluorophore labeled oligonucleotides (sequence F1: 5′-FAM-TAT-ATA-TAG-ATA-TTT-TTT-TAT-CTA-TAT-ATA-3′-TAMRA and sequence-F2: 5′-FAM-TAT-AGA-TAT-AGA-TAT-TTT-ATA-TCT-ATA-TCT-ATA-3′-TAMRA) with preferred PBD binding sites were used in the assay, and the stabilization was compared with a previously reported C8-PBD monomer, GWL-78 19 (compound 1) and the control compound 2 with the third heteroaromatic ring.In all cases, compounds with aliphatic third moiety showed between 3-and 4-fold lower ΔT m values at both 5:1 and 1:1 ligand−DNA ratios compared to the control compounds, showing that the introduction of the aliphatic third moiety reduced the ability of these compounds to interact with and stabilize the DNA sequences (Table S4).The eukaryotic toxicity of the synthesized PBD derivatives was tested at a single concentration (20 μM) using the MTT viability assay on WI-38 fibroblast cell line.The control compounds, 1 and 2, showed potent cytotoxicity with only 35 and 32% of cells viable after 24 h.However, the newly synthesized compounds showed notably less toxicity, with >85% of cells viable after 24 h of treatment with the molecules (Figure S1).Considering an average molecular weight of 700 Da for the synthesized PBD monomers, the concentration used for the test (20 μM) corresponds to a concentration of 14 mg/ L, which is approximately 100 times the average MIC reported for Gram-positive strains and almost 10 times the average MIC reported for Gram-negative strains for active compounds.The obtained results confirmed the presence of a selective toxicity profile toward prokaryotic cells for the PBD-C8 conjugates, confirming the possibility of further development of this class of compounds as antibacterial agents.
Inhibition of DNA Gyrase.The C8-linked PBDs have been recently reported as DNA gyrase inhibitors. 14This was further explored by docking compounds 7 and 8 with the bacterial gyrase from S. aureus (PDB ID 2XCT) and E. coli (PDB ID 6RKS) (Figures 5 and S2).Both compounds bound to the ligand binding domain of DNA gyrase A and showed interaction with several amino acid residues through hydrogen bonds and hydrophobic interactions.The best pose of compound 7 gave a ChemScore of 27.62 and a binding affinity of −31.67 kcal/mol, and the values for compound 8 were comparable with a ChemScore of 27.35 and binding affinity of−30.70 kcal/mol in the case of DNA gyrase A from S. aureus.For the DNA gyrase A from E. coli, the ChemScores were 25.16 and 22.01 while the free energy of binding values were 30.46 and −25.51 kcal/mol for compounds 7 and 8, respectively.The 2D models shown in Figure S3 suggest that compound 7 forms six hydrogen bonds within the ligand binding domain of DNA gyrase A from S. aureus and five hydrogen bonds within the ligand binding domain of DNA gyrase A from E. coli.Similarly, compound 8 forms five hydrogen bonds within the ligand binding domain of DNA gyrase A from S. aureus and four hydrogen bonds within the ligand binding domain of DNA gyrase A from E. coli.Both compounds show notable hydrophobic interaction within the binding site (Tables S5−S8).The interaction of compounds 7 and 8 with the ligand binding sites of gyrase A suggests that the inhibition of gyrase A plays a crucial role in the antimicrobial activity of this class of compound.
We have recently shown that PBD monomers caused the inhibition of DNA gyrase in a rapid assay format. 14A gel-based assay was used here to evaluate whether compounds 7 and 8 were also able to inhibit DNA gyrase in vitro.In assays using purified S. aureus DNA gyrase, compound 8 showed an IC 50 of 3.16 ± 1.36 mg/L compared to ciprofloxacin IC 50 of 13.05 ± 1.36 mg/L (Figure 6).It was not possible to determine an IC 50 for compound 7, as solubility issues prevented the use of high enough concentrations of compound to achieve >50% inhibition (at a concentration of 32 mg/L).Against the E. coli DNA gyrase, it was not possible to reach a concentration high enough to achieve greater than 50% inhibition for either compound 7 or 8, again because of solubility issues.The relative inhibition of the Gram-positive versus Gram-negative gyrase correlates with the relatively higher MICs of Gramnegative bacteria species, and this supports the idea that the novel PBDs described here work through the inhibition of DNA gyrase.
Although we have evidence of the inhibition of DNA gyrase activity, for both Gram-positive and Gram-negative DNA gyrase, this does not preclude the possibility that the PBDs cause inhibition via binding to DNA and/or a DNA−gyrase complex.To further define the mechanism of action, a panel of fluorescent E. coli reporter strains was used to understand the profile of cellular responses to PBDs and compare these with molecules with known mechanisms of action.A panel of strains derived from a comprehensive library of E. coli promoter fusions was used. 20Specific reporters for compounds affecting the DNA metabolism were selected, in line with previous studies, 21 and validated with DNA gyrase inhibitors (fluoroquinolones, novobiocin), compounds known to bind DNA (mitomycin C, PBD control) and control compounds affecting other cellular targets (doxycycline, ampicillin, rifampicin).The four promoter fusions showed the expected induction with fluoroquinolones (specifically levofloxacin, ciprofloxacin, norfloxacin, and moxifloxacin) and the quinolone antibiotic nalidixic acid, associated with the activation of the bacterial SOS response as a result of the poisoning of the DNA gyrase.The level of fold induction was similar between the known gyrase inhibitors, with fold inductions ranging from 20-to 35-fold for recA, 5-to 13-fold for lexA, 14-to 32-fold for recN, and 17-to 41-fold for sulA (Figure 7).Although compounds 7 and 8 both showed the activation of the same four promoter fusions, the levels were significantly lower for each of the genes (8−10-fold recA, 5−9-fold lexA, 6−11-fold recN, and 6−7-fold sulA).By comparison, the three antibiotics known not to induce the SOS response, doxycycline, rifampicin, and ampicillin, showed no induction of any of the four promoters above the level of the control strain.Interestingly, three compounds showed similar levels of activation of the promoters to the two PBDs.Two of these are known to be DNA binding molecules, mitomycin C and the control PBD dimer DSB-120, 22 whereas the third, novobiocin, inhibits DNA gyrase by interaction with the ATP-binding site on the GyrB subunit.These results suggest that the PBDs act by interfering with DNA metabolism, in a way that is clearly different from fluoroquinolones.This is consistent with MIC data which show no significant differences between fluoroquinolone-sensitive and -resistant strains.Further studies will be required to delineate the specific interactions with DNA and/or individual gyrase subunits and the specific mechanisms of inhibition.
Emergence of ResistanceMutation Frequency.The mutation frequency of the most active compound 8 was tested against representative strains of the three Gram-negative pathogens in comparison with ciprofloxacin.At 4× MIC, mutation frequencies of <4.3 × 10 −8 and <1.5 × 10 −8 were measured across three replicate experiments for NCTC 12923 and ATCC 17978, respectively.In both cases, the values were significantly lower than the control antibiotic, ciprofloxacin (4.4 × 10 −7 for NCTC 12923 and 8.4 × 10 −7 for ATCC17978).This was not the case in the K. pneumoniae isolate M6, where the mutation frequency was 1.0 × 10 −6 for compound 8, compared to 8.9 × 10 −7 for ciprofloxacin.This higher mutation frequency is reflected in the time-kill experiments, where resistance was only seen in a single K. pneumoniae strain (NCTC 13368), following exposure to concentrations above the MIC and not in A. baumannii.
The mutations associated with resistance in NCTC 13368 were identified by whole genome sequence analysis of stable resistant isolates.In one of the time-kill repeats, the resistant strains showed the deletion of a single base in position 50 of a gene encoding a predicted nucleoside-specific channel-forming protein (KPN_RS01945), termed Tsx.The mutation caused a frameshift in the tsx gene and was predicted to result in a truncated 42 amino acid protein.A second SNP was also identified in another uncharacterized permease gene, leading to an amino acid change M137K.A clean transposon mutant (tnkp1_lr150124p14q189) in tsx, in K. pneumoniae strain MKP103, also showed elevated resistance to compounds 7 and 8, with respect to their parental strain (compound 7 from 32 mg/L in wild-type to >128 mg/L for tsx transposon, and for compound 8 from 16 to >32 mg/L) (Table S3).The Tsx protein has previously been described as mediating resistance to the antibiotic albicidin in Klebsiella oxytoca. 23A second resistant NCTC 13368 isolate showed a SNP, leading to an amino acid change H50N in a predicted MerR family regulator (KPN_RS12075).

■ CONCLUSIONS
The lack of new antibiotics, especially for rapidly emerging MDR Gram-negative pathogens, is a significant cause for concern.Several studies have suggested that there may be no new antimicrobials, with distinct modes of action, available for treatment of these pathogens for a considerable time. 24Against this background, we report the identification of a new scaffold with significant activity against both Gram-positive and MDR Gram-negative pathogens, which do not appear to be impacted by common resistance mechanisms to other broad-spectrum antibiotics, and with a low mutation frequency, leading to resistance in key species such as A. baumannii and E. coli.
Mechanistic evaluation shows a sub-micromolar IC 50 in a gel-based S. aureus DNA gyrase assay, significantly lower than ciprofloxacin and not affected by GyrA mutations associated with fluoroquinolone resistance.Using a panel of fluorescent reporters, it was possible to show the activation of genes which are associated with the SOS response in Gram-negative bacteria, but again the magnitude of promoter activation differed from fluoroquinolones and was consistent with both known DNA binders (e.g., mitomycin C) and DNA gyrase inhibitors that act via alternative binding sites to fluoroquinolones (e.g., novobiocin).
The compounds show excellent activity against WHO priority pathogens with an in vitro selective index which would allow their continued development.The developability assessment of the lead compounds will be carried out to fully explore the clinical utility of this chemical class.There is renewed interest in developing bacteria-specific ADCs.We intend to exploit well-understood and clinically validated routes to develop ADCs, which has been successfully employed for PBD-type compounds in oncology, to further

Journal of Medicinal Chemistry
improve the prokaryotic selectivity of these compounds and to provide selective targeting of such compounds to bacteria.This will form the basis of broad-spectrum PBD antibiotics which may have significant utility in the clinic.

■ EXPERIMENTAL SECTION
General Chemistry.All reagents and solvents employed in the synthetic processes were obtained from commercially available sources including, among others, Sigma-Aldrich, Fisher Scientific, Fluorochem, and Alfa Aesar. 1 H and 13 C nuclear magnetic resonance (NMR) analyses were performed on a Bruker Spectro Spin 400 Hz spectrometer.Liquid chromatography−mass spectrometry (LC−MS) analyses were performed on a Waters Alliance 2695 system, eluting in gradient.The analyses were performed on a Monolithic C18 50 × 4.60 mm column by Phenomenex.UV detection was performed on a diode array detector.Mass spectra were registered in both ESI + and ESI − modes.Melting points were determined using a Stuart SMP30 melting point apparatus.All of the compounds tested for their biological activity are >95% pure, that is confirmed with two different high-performance LC analysis methods.The high-resolution MS (HRMS) analyses were performed on a Thermo Scientific Exactive HCD Orbitrap Mass Spectrometer.The hydrogenation reaction was conducted using a Parr hydrogenation system.Synthetic protocols and compound characterization for the previously reported PBD core are reported in the Supporting Information file.
Synthesis of Pyrrole Benzofused Intermediates 27 and 28.Commercially available 4-((tert-butoxycarbonyl)amino)-1-methyl-1H-pyrrole-2-carboxylic acid (200 mg, 0.8 mmol, 1.2 equiv) was dissolved in DMF (5 mL).EDCI (2.5 equiv) and DMAP (3 equiv) were added to the solution that was left under a magnetic stirrer in a N 2 atmosphere for 20 min.At that point, the corresponding 5-amino-2-methyl ester benzofused (1 equiv) was added to the reaction mixture and left under a magnetic stirrer overnight.The reaction did not go to completion.The reaction was quenched by the addition of water (15 mL) that was then extracted with ethyl acetate (3 × 10 mL).The organic phase was then sequentially washed with brine (15 mL), NaHCO 3 -saturated aqueous solution (15 mL), and citric acid aqueous solution 0.1 M (15 mL).The collected organic phase was dried on MgSO 4 and subsequently evaporated using a rotary evaporator giving crude of reaction that was subsequently purified by column chromatography [mobile phase: from 100 dichloromethane (DCM) to 85/15, v/v, DCM/EtOAc] to give the final products 27 and 28.
6-(4-((tert-Butoxycarbonyl)amino)-1-methyl-1H-pyrrole-2carboxamido)benzofuran-2-carboxylate (27).Obtained 0.200 g (reaction yield: 70%) as a yellow solid. 1 H NMR (400 MHz, chloroform-d): δ 8.09 (d, J = 2.01 Hz, 1H), 7.85 (s, 1H), 7.51 (d, J = 9.06 Hz, 1H), 7.47 (s, 1H), 7.40 (dd, J = 9.06 Hz, 2.27 Hz, 1H), 6.86 (br s, 1H), 6.70 (br s, 1H), 6.43 (br s, 1H), 3.97 (s, 3 H), 3.91 (s, 3 H), 1.51 (s, 9H). 13  (8 mL), and an excess of NaOH 1 M aqueous solution was added to the solution.The reaction mixture was left under a magnetic stirrer overnight at room temperature until thinlayer chromatography (TLC) showed total disappearance of the starting material.The solvent was evaporated under vacuum using a rotary evaporator, and citric acid 1 M aqueous solution was added to the crude of the reaction until acid pH was reached causing the simultaneous precipitation of a white solid.The precipitate was collected by filtration under vacuum and dissolved in DMF (4 mL).EDCI (2.5 equiv) and DMAP (3 equiv) were added to the solution, and the reaction mixture was left under a magnetic stirrer for 20 min at room temperature under a N 2 atmosphere.Change in color of the solution was observed from light yellow to dark brown.The correspondent amine (1.5 equiv) was added to the solution and left under a magnetic stirrer overnight.The reaction did not go to completion.The reaction was quenched by the addition of H 2 O (10 mL) that was subsequently extracted with EtOAc (3 × 8 mL).The collected organic phases were washed sequentially with brine (10 mL), NaHCO 3 -saturated aqueous solution (10 mL), and citric acid aqueous solution 0.1 M (10 mL).The organic phase was dried on MgSO 4 and evaporated under vacuum using a rotary evaporator, giving the correspondent crude of reaction that was purified by column chromatography on silica gel (mobile phase: from 100 DCM to 70/30, v/v, DCM/EA, depending on the substrate) to give the final products 33−38.
In the case of diethyl derivatives 31 and 32, a different procedure was applied.After the hydrolysis of the methyl ester benzofused intermediates 27 and 28, the obtained acid (1 equiv) was dissolved in DMF (4 mL), and diethyl amine (2 equiv) and DIPEA (2 equiv) were added to the solution that was kept at 0 °C in an ice bath.At that point, T 3 P (50% solution in DMF, 2 equiv) was added to the reaction mixture that was allowed to reach room temperature and left under a magnetic stirrer overnight.The reaction did not go to completion and was quenched by the addition of water (10 mL) and then extracted with ethyl acetate (3 × 8 mL).The organic phase was washed with brine (10 mL) and NaHCO 3 -saturated aqueous solution (10 mL).The collected organic phase was dried on MgSO 4 and subsequently evaporated under vacuum using a rotary evaporator.The crude of the reaction was subsequently purified by column chromatography (mobile phase: from 100 DCM to 70/30, v/v, DCM/EA) to give the final products 31 and 32.
Commercially available 4-((tert-butoxycarbonyl)amino)-1-methyl-1H-pyrrole-2-carboxylic acid (0.659 g, 2.7 mmol, 1 equiv) was dissolved in DMF (7 mL), and EDCI (2.4 equiv) and DMAP (3 equiv) were added to the solution that was left to stir in a N 2 atmosphere for 30 min.The deprotected compound 39 was added to the reaction mixture and left to stir overnight at room temperature in a N 2 atmosphere.The reaction did not go to completion and was quenched by the addition of H 2 O (10 mL).The aqueous phase was then extracted with AcOEt (3 × 10 mL).The organic phase was sequentially washed with citric acid 0.1 M aqueous solution (10 mL), saturated NaHCO 3 aqueous solution (10 mL), and brine (10 mL).The collected organic phase was dried over MgSO 4 and evaporated under vacuum using a rotary evaporator.The crude of the reaction was purified by column chromatography on silica gel (mobile phase: DCM/AcOEt, 60/40, v/v), giving pure compound 40 (0.621 g, reaction yield: 50%) as a white solid. 1    Synthesis of Bromo-Pyrrole Intermediates 44 and 45.Starting from commercially available N-methyl bromo-pyrrole methyl ester, the same procedure used for the synthesis and purification of compound 21 (after basic hydrolysis of the methyl ester) was applied to give intermediates 44−45.
The crude of reaction was purified by column chromatography on silica gel (mobile phase: DCM/AcOEt, 80/20, v/v), giving pure compounds 44 and 45.
Susceptibility Testing.The MIC was determined using the microdilution broth method. 27Compounds were initially dissolved in dimethyl sulfoxide (DMSO) prior to dilution in broth.Equivalent concentrations of solvent had no effect on bacterial growth.The MIC was defined as the lowest concentration of compound which resulted in no visible growth at an optical density of 600 nm.Experiments were performed in triplicate.
Time-Kill Curve Analysis.Three independent repeats were performed for each strain with each compound.Fresh TSB media (10 mL) was inoculated with ∼1 × 10 7 cfu/mL of the test organism.Bacteria were challenged at 4× MIC for each compound and incubated at 37 °C in a rotary shaker at 200 rpm.One hundred microliter aliquots were taken at 0, 1, 2, 4, 6, and 24 h post inoculation and serial dilutions performed in sterile phosphatebuffered saline.Total viable counts were determined by the Miles− Misra dilutions method.The compound is considered bactericidal if the inoculum was reduced >3 log10 cfu/mL and bacteriostatic if inoculum was decreased by 0−3 log10 cfu/mL.
Efflux and Influx Assays.An adapted microdilution broth method was performed to evaluate the efflux potential of the compounds.In this assay, 50 μL/well of EPI at 4× final concentration and 50 μL/well of the test organism at 2 × 10 5 cfu/mL were added to wells containing 100 μL/well of a dilution series of compound.CCCP was added at a final concentration of 10 mg/L, and phenylalanine− arginine β-naphthylamide PAβN was added at 25 mg/L.TSB media was supplemented with MgSO 4 (40 μM) to prevent the permeabilization of the outer membrane of Gram-negative bacteria by PAβN. 18A decrease in MIC of the compound of at least fourfold was defined as significant for efflux activity. 28he influx assay was performed exactly as described for the efflux assay, with the addition of PMBN at a final concentration of 30 mg/L.
Gyrase Inhibition Assay.The effect of the antimicrobial agents on the gyrase activity was assessed using the S. aureus and E. coli gyrase supercoiling gel-based assays obtained from Inspiralis (Norwich, UK). 30 Methods were conducted as per the manufacturer's instructions.
Whole Genome Sequence Analysis and Transposon Mutants.NCTC 13368 isolates resistant to compound 7 were analyzed to identify the site of mutations.Isolates were passaged 10 times in the absence of selection and checked to confirm that the resistance phenotype was maintained.DNA was isolated and the genome was sequenced by PHE-GSDU on an Illumina (HiSeq 2500) as previously described. 31Potential individual mutations were identified using Galaxy. 32ell Culture and MTT Assay.WI-38 cell line was obtained from the American Type Culture Collection.The cells were grown in normal conditions in an incubator at 37 °C in a humidified atmosphere containing 5% CO 2 .The appropriate medium [Eagle's minimum essential medium (MEM) or MEM, Gibco] supplemented with fetal bovine serum (10%, v/v, Sigma-Aldrich) was used for culturing the cells.The viability assay was conducted in a 96-well plate, and the plates were left under continuous incubation with the Journal of Medicinal Chemistry drug for 24 h.The media was removed, the MTT reagent was added, and finally, the absorbance of the formazan crystals was read using a plate reader (Envision Plate Reader, PerkinElmer).The values of absorbance obtained were normalized with the blank and then used for the determination of the % of viability compared to the control.For single-point toxicity screening, the number of repeats (n) was equal to 6, and the result is reported as average values.
FRET-Based DNA Melting Assay.The modified, fluorophoretagged hairpin oligonucleotides were purchased from Eurogentec.The assay was carried out using a previously published procedure 7 with 5:1 ligand oligonucleotide ratio.The working solution of 400 nM was made in FRET buffer (50 mM potassium cacodylate, pH 7.4) and annealed by heating the working solution at 90 °C for 6 min followed by cooling to room temperature and storage at this temperature for 5 h before the assay.The compound and oligonucleotide were combined in a 1:1 ratio (25 μL of each) in a 96-well plate and were heated in the range of 30−100 °C after 3 h incubation at 25 °C.The fluorescence readings (excitation 490 nm, emission 520 nm) were taken at fixed intervals of 0.5 °C, over the cited range.The data obtained by the instrument were processed through normalization of the curve using Origin 7.0 (OriginLab Corp., USA).The DNA melting temperature values were then determined using a preset script that determined the values as the point of inflection of the first derivative of the curve.Each experiment was conducted in triplicate, and the values reported are the ΔT m average along with the standard deviation of the mean.
Mutation Frequency.Mutation frequencies were determined following the method of Evans and Titlow (1998).Briefly, 100 μL of bacterial culture, grown to an OD 600 of approximately 0.5−0.6 was used to inoculate the surface of TSB agar containing compounds at a range of concentrations above and below the MIC.The plates were then incubated at 37 °C and the plates assessed for the presence of colonies following 48 h incubation.Miles−Misra was also performed on drug free TSA plates to determine the exact number of cfu/mL in the initial culture.Mutation frequencies were then calculated by dividing the number of colonies on plates containing the specific agents by the total number of colony-forming units (cfu) that were plated.
Reporter Strain Assays.Strains from the E. coli K12 MG1655 promoter library created by Zaslaver et al. 20 were purchased from Dharmacon (GE Life Sciences).Strains were maintained on 25 mg/L kanamycin TSA plates and cultured in M9 media supplemented with 1% glucose, 0.2% casamino acids, 0.5 mg/L thiamine, 100 μM CaCl 2 , 2 mM MgSO 4 , and 25 mg/L kanamycin. 21Twofold serial dilutions of compounds were made up in DMSO to 50× the required concentration, with a previously determined inhibitory concentration in the middle of the series.The dilution series (2 μL) was added per well, with 2 μL of DMSO in "untreated" wells, and then, 98 μL of overnight culture back diluted to an OD 600 of 0.2 was added to each well.Strain U66 or U139, containing the promoterless plasmids pUA66 or pUA139, respectively, was included for every compound tested as a negative control.Plates were incubated at 37 °C with 200 rpm in a CLARIOstar microplate reader (BMG Labtech).Fluorescence (RFU) and cell growth (OD 600 ) were measured every 30 min for 20 h.Fluorescence data were normalized by cell density (RFU/OD 600 ) at 9 h, where fluorescence became stable, and at the highest concentration of compound for which there was growth.Fold induction of the promoter was calculated by dividing the fluorescence of the treated sample by fluorescence of the untreated sample.
Molecular Docking of the Compounds to Gyrase A. AutoDock SMINA was used for molecular docking of compounds 7 and 8 to the minimized crystal structure of gyrase A from S. aureus (PDB ID 2XCT) and cryo EM structure of gyrase A from E. coli (PDB ID 6RKS) for finding the best binding pocket by exploring all probable binding cavities in the enzyme.All of the parameters were kept in their default values.Then, GOLD molecular docking was used for molecular docking of the compounds into the SMINA-located binding site for performing flexible molecular docking and determining more precise and evaluated energies and scores. 33,34sed on the fitness function score and ligand binding position, the best-docked pose for each compound was selected.

Accession Codes
The PDB ID code for DNA gyrase in complex with 7 and 8 is 2XCT.

Figure 1 .
Figure 1.Reported antibacterial PBDs and structural modification employed to obtain compounds with activity against Gram-negative bacteria.

Figure 2 .
Figure 2. Structure of control compounds 1 and 2 and structures of final compounds 3−17.

Figure 4 .
Figure 4. Rapid bactericidal kill by synthesized compounds 7 and 8, with occasional breakthrough resistance in some drug−strain combinations.The data are the average of three independent experiments, and error bars represent the standard deviation from the mean.
in MIC in the presence of efflux inhibitors or membrane permeabilizers was defined as significant (highlighted by shading).All values are given in mg/L.These data are representative of three or more independent experiments.ND = Not done.

Figure 5 .
Figure 5. Molecular model showing the location of the binding pocket and the interaction of compound 7 with (A) DNA gyrase A from S. aureus; (B) DNA gyrase A from E. coli.

Figure 7 .
Figure 7. Reporter assay to explore the mechanism of action of compounds 7 and 8. Data are the mean of three independent experiments, and error bars represent the standard deviation of the mean.Two-way ANOVA and Dunnett's multiple comparison tests were performed and included here for reporter strain recA, all compounds were compared to ciprofloxacin ****p < 0.0001.

Table 1 .
Activity of the Synthesized Compounds againstGram-Negative Bacteria (mg/L)

Table 2 .
Activity of the Synthesized Compounds against MDR Gram-Positive Bacteria a a MICs are expressed in mg/L.