Antibacterial small molecules targeting the conserved TOPRIM domain of DNA gyrase

To combat the threat of antibiotic-resistant Gram-negative bacteria, novel agents that circumvent established resistance mechanisms are urgently needed. Our approach was to focus first on identifying bioactive small molecules followed by chemical lead prioritization and target identification. Within this annotated library of bioactives, we identified a small molecule with activity against efflux-deficient Escherichia coli and other sensitized Gram-negatives. Further studies suggested that this compound inhibited DNA replication and selection for resistance identified mutations in a subunit of E. coli DNA gyrase, a type II topoisomerase. Our initial compound demonstrated weak inhibition of DNA gyrase activity while optimized compounds demonstrated significantly improved inhibition of E. coli and Pseudomonas aeruginosa DNA gyrase and caused cleaved complex stabilization, a hallmark of certain bactericidal DNA gyrase inhibitors. Amino acid substitutions conferring resistance to this new class of DNA gyrase inhibitors reside exclusively in the TOPRIM domain of GyrB and are not associated with resistance to the fluoroquinolones, suggesting a novel binding site for a gyrase inhibitor.


Introduction
Antibiotic resistance is an immediate and growing danger to human health. The morbidity and mortality wrought by these resistant bacteria know no international or socioeconomic boundaries. The so-called ESKAPE pathogens typify the risks and experience with controlling and combating bacterial infections in the community and in the clinic [1]. Amongst the ESKAPE pathogens are the Gram-negatives Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and Escherichia coli. Historically, these Gram-negative pathogens have been controlled with many of the currently approved antimicrobials. With the emergence and transmissibilityof a number of resistance mechanisms, such as β-lactamases, efflux pumps, or drug-modifying enzymes, treatment options can be very limited or non-existent [2]. These life-threatening, difficult-to-treat infections have prompted a call for the identification and a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 development of novel antibiotics addressing new modalities or new chemistry that circumvents resistance to current therapies [3].
The fluoroquinolones (FQs, e.g, ciprofloxacin) are an essential class of antibiotics encompassing several generations of modification and improvement. Ciprofloxacin, for example carries a large number of approved therapeutic indications including urinary tract, respiratory tract, skin and skin structure infections, among others (fda.gov). There are however significant limitations to the use of this important class of antibiotics [4,5]. Bacterial DNA gyrase, the target of the FQs, is a heterotetrameric complex of approximately 380 kDa containing two A (GyrA) and two B (GyrB) subunits that controls the torsional strain induced by the positive supercoils resulting from transcription and DNA replication [6,7]. By creating a double strand break (gate segment) and passing another strand of DNA (transfer segment) through the break and resealing it in an ATP-dependent reaction, gyrase introduces negative supercoils, thereby reducing positive supercoiling [8]. That reaction, the structure of gyrase, and the effects of inhibitors such as the FQs have been studied and reviewed in considerable detail (e.g., [9][10][11]). Gyrase is a clinically validated, structurally tractable target for antibacterial drug development, and in recent years a considerable number of non-FQ gyrase inhibitors have been described and reviewed [12,13]. Bacterial topoisomerase II inhibitors can be divided into two broad categories; those that inhibit the essential ATPase function of the B subunit and those that inhibit gyrase in a different manner. The founding member of the ATPase inhibitor class is novobiocin, an aminocoumarin natural product, while a number of other inhibitors have also been developed that target the ATPase domain [13]. Gyrase inhibitors can also act through interactions with other regions of the tetrameric complex. Examples of an inhibitor class that differs from ATPase inhibitors and FQs are NXL101 and GSK299423 [14,15]. An important region in the GyrB subunit is the TOPRIM domain; a conserved structural element found in a number of DNA-interacting enzymes such as topoisomerases (types IA/II), DnaGlike primases, archael nucleases, and RecR-like DNA repair proteins [16]. AZD0914, a novel gyrase and topoisomerase IV inhibitor appears to interact with the TOPRIM domain [17,18].
As an alternative to target-focused approaches to drug discovery [19,20] we began with a collection of proprietary small molecules that had demonstrated activity against E. coli lacking efflux capability (ΔtolC) and having reduced barrier function (lipopolysaccharide) due to a mutation in lpxC. This strain enabled the identification of molecules that would likely otherwise not be expected to inhibit growth of wild-type bacteria, but may yield novel target/compound pairs for further development [21]. The collection of bioactives was then counterscreened for activity against a model eukaryote (Saccharomyces cerevisiae) to eliminate generally bioactive molecules and examined for suitability for synthesis and derivatization as lead molecules for drug development. Further triage of this set encompassed testing for cell lysis in bacteria and erythrocytes [22,23], excluding those with activity close to the antibacterial activity. As the annotated library of bioactives became more refined, we sought to identify the pathway involved through inhibition of macromolecule synthesis. Finally, we attempted target identification through selection for resistance and whole genome sequencing and, where possible, biochemical assays. From this screening and evaluation effort, we detail here the discovery of a novel class of Gram-negative bacterial DNA gyrase inhibitors targeting the TOPRIM domain.

Macromolecule labeling
Macromolecule synthesis profiling was conducted as described by Montgomery, et al. [32] with minor modifications. E. coli strain JL553 (tolC, lysE) [32] was used to measure inhibition of DNA, RNA, protein, peptidoglycan and phospholipid biosynthesis while E. coli MB5500 (lpxC [envA1], galE; Merck & Co., Inc., Kenilworth, NJ, USA, strain collection) was used to measure inhibition of lipopolysaccharide biosynthesis. Radioisotope labeled precursors were used at the following final concentrations stopped with trichloroacetic acid, filtered and incorporation was measured as described [32]. Percent inhibition of incorporation was calculated relative to an untreated control culture for each labeled precursor and plotted with GraphPad Prism.

Agarose gel DNA gyrase assays
Conditions to measure the supercoiling of a relaxed, closed-circular plasmid substrate and cleavage complex stabilization in agarose gels were adapted from the literature and carried out as described below [33]. Plasmid supercoiling reactions were assembled by combining 0.4 μL compound (in 100% DMSO), 12.5 μL H 2 O, 4 μL 5Xbuffer, 0.4 μL relaxed pUC19 DNA substrate (0.5 μg), 1 mL dilute DNA gyrase (buffer, substrate, and enzyme from New England Biolabs, Beverly, MA). Alternatively the substrate pHOT-1 (TopoGEN, Buena Vista, CO) was used as noted (the substrate from New England Biolabs has been discontinued). P. aeruginosa DNA gyrase was supplied by Inspiralis, Norwich, UK. Reactions were allowed to proceed for 30 min. at 37˚C after which 5 μL of stop buffer (5% sarkosyl, 25% glycerol, 1X Blue Juice [Invitrogen, Carlsbad, CA]) were added. Reactions were then extracted with 20 μL chloroform/isoa-mylOH (24:1) and a portion of the aqueous phase was loaded into a well in a 1% agarose/Trisborate EDTA gel. Electrophoresis was conducted at 35 VDC until the loading dye migrated about 6.5 cm down the gel. DNA in the gel was then stained in 0.5 μg/mL ethidium bromide in water for 30 min. and rinsed (3X) in H 2 O. The gel was illuminated with UV light, photographed, and DNA quantitated with an Alpha Imager and accompanying software (Protein-Simple, San Jose, CA). Following background subtraction, image data were used to estimate the concentration to achieve 50% inhibition (IC 50 ) on a four-parameter curve fit in GraphPad Prism software (IC 50 derived from 2-3 independent experiments).

Broth microdilution minimum inhibitory concentration determination
Compound doubling dilutions were prepared in 100% DMSO (Sigma Aldrich) and 2 μL of each dilution was transferred to a 96-well plate with lid (Thermo Scientific). Inoculum for all MIC testing was prepared using the BBL Prompt Inoculation System (Becton Dickinson, Franklin Lakes, NJ USA) as described by the manufacturer to produce a suspension of cells at approximately 1.5 x 10 8 colony forming units (CFU)/mL [34][35][36] and a further dilution (1:1000) was made into cation adjusted Mueller-Hinton broth (CAMH broth, Becton-Dickinson). One hundred μL of this inoculum were then transferred to each well of the 96-well plate containing diluted compound (above). Plates were covered and incubated overnight (18 hrs.) at 37˚C. The MIC for the compound was determined as the minimum concentration required to completely inhibit visible growth of the cells.

In vitro time-dependent bactericidal studies
A culture of E. coli HS151 in CAMH broth was grown overnight (18 hrs.) at 37˚C, diluted 1:10,000 in 200 mL of fresh CAMH broth and further incubated for 1 hour at 37˚C with shaking. A time-zero sample was then removed and 20 mL portions of the remaining culture were transferred into 125 mL vented flasks to which were added MRL-1082 or control antibiotics, at the indicated concentrations. These flasks were then incubated at 37˚C with shaking (70 rpm). Samples were withdrawn at the indicated times, plated on blood agar (trypticase soy agar with 5% sheep blood, Becton-Dickinson) and, following overnight growth at 37˚C, colonies were counted to determine CFU/mL.

Resistance selection and whole genome sequencing
The MRL-770 or MRL-423 agar-dilution MIC was determined prior to selection for resistance by first serially diluting the compound in 100% DMSO (Sigma Aldrich) followed by the addition of the diluted compound to 4 mL of molten CAMH agar in each chamber of an 8-well dish (Thermo Scientific). After the agar solidified, 5 μL of a saturated bacterial culture, grown in CAMH broth, were dispensed on the agar surface and allowed to dry at room temperature. The plate was then incubated overnight at 37˚C and the agar MIC was identified as the lowest dilution of compound that completely inhibited visible growth. To select for resistance, a 100 mm petri dish was filled with 25 mL CAMH agar containing 8 μg/mL MRL-770 or 0.5 μg/mL MRL-423 (Fig 1). The cells in 1 mL of an overnight culture of E. coli HS151 (~10 9 CFU/mL) were concentrated by centrifugation and spread onto the surface of an MRL-770 or MRL-423 containing plate. After the surface dried at room temperature the plate was incubated at 37˚C overnight. Any colonies that arose were counted and an approximate frequency of resistance was calculated after confirming resistance by restreaking on media containg compound as well as measuring MICs (see Results). Genomic DNA from the parent strain (E. coli HS151) and purified (restreaked) resistant isolates was obtained with a GeneJET genomic DNA purification kit according to the manufacturer's instructions (Fisher Scientific). Assembled whole genome mutant sequences were then compared to the parent strain and polymorphisms were identified as described [37][38][39][40].

yacG deletion
The yacG deletion allele from the Keio collection [26] was amplified by PCR (primers: GGACATAAGAGCCGTTTTGCCATTCG and CAGGCCTAAAGGGATTTCTAACTCATTGAT) and moved to BW25113 ΔtolC as described below. The E. coli strain BW25113 ΔtolC containing pKD46 [25] was used to inoculate SOB (Invitrogen)/0.2% w/v arabinose (Sigma Aldrich)/ ampicillin (100 μg/mL) and grown overnight at 30˚C. The overnight culture was then diluted 1/40 into 20 mL fresh SOB/0.2% w/v arabinose/ampicillin (100 μg/mL) and grown at 30˚C while shaking for~3 hrs until OD 600 = 0.5. Cells were pelleted by centrifugation at 4000 rpm for 10 min, washed with 40 mL ice cold sterile, distilled H 2 O, then 3 x 1 mL ice cold 10% glycerol followed by centrifugation at 5000 g for 3 minutes. The pellet was resuspended in 100 μL of 10% glycerol (1/200 of starting culture). For transformation, 5 μL of gel purified PCR product was added to 50 μL cells, mixed, and immediately electroporated in a 0.2 cm gap cuvette at 2.5kV, 25 μF, 200 O (BioRad, Hercules, CA USA). Electroporated cells were recovered in 1 mL SOC (Invitrogen)/0.2% arabinose at 37˚C for 2 hours with shaking. Cells in amounts of 50 and 950 μL were then plated on LB agar supplemented with kanamycin (25 μg/mL) and incubated overnight at 37˚C. Correct gene replacements were verified by PCR.

Results
Beginning with a collection of compounds active against a sensitized E. coli strain containing mutations affecting the integrity of the outer membrane and a xenobiotic efflux system, we sought to categorize these bioactives by inhibition of macromolecule synthesis. One such compound (MRL-770, Fig 1) exhibited selective and reproducible inhibition of DNA synthesis in E. coli JL553 (Fig 2). This inhibition, observed following a brief exposure to the compound (25 min), began at concentrations near the minimum inhibitory concentration (MIC; 4.5 μM; 2 μg/mL) and produced a dose-response profile similar to the control antibiotic ciprofloxacin. Additional in vitro microbial susceptibility testing shows that MRL-770 is active against sensitized strains of bacteria, but as Table 1 indicates efflux or cell permeability may be a major impediment to broader Gram-negative antibacterial potency.
To understand the mechanism of growth inhibition by our initial hit, MRL-770, spontaneous resistance was selected for in the efflux-deficient E. coli strain HS151. Following 24-48 hours on agar plates containing 8 μg/mL (4XMIC) rare resistant isolates were found at a rate of approximately 5X10 -9 . Comparison of the genome sequences of resistant mutants with that of the wild-type parent strain identified missense mutations and a single codon insertion mutation in gyrB (DNA gyrase) that were associated with selective resistance to MRL-770 ( Fig  1, Table 2). The more potent analogs MRL-423 and MRL-1082 (Fig 1, Table 2) were designed and synthesized as part of a compound optimization program based on our initial discovery of MRL-770 and confirmation of the target. The relative location of each mutation, all of which reside within the conserved TOPRIM domain of GyrB [16], is depicted graphically in Fig 2C. Interestingly the single amino acid substitution mutants are not cross-resistant to other DNA gyrase inhibitors such as fluorquinolones and novobiocin, nor do they show any changes in susceptibility to other antibiotics ( Table 2). The unusual codon insert mutant, GyrB (E466EV) has a 4-8-fold higher MIC to FQs. In an X-ray crystal structure of S. aureus DNA gyrase, the  equivalent residue in GyrB (E477) is adjacent to a QRDR loop [14] and a codon insert at that site could affect positioning of the QRDR loop and thereby affect FQ binding (see Discussion). A mutation residing in the hotspot region of GyrA conferring moderate resistance to FQs does not affect sensitivity to our novel compounds (Fig 1 and Table 2). Selection for spontaneous resistance to MRL-423 identified a single isolate (frequency of resistance = 3.4 x 10 −10 ) with an amino acid substitution mutation in GyrB (T469K) that was not resistant to MRL-770. Aside from GyrB (E466EV), resistance to MRL-423 was 4 to 16-fold higher than the parent strain E. coli HS151, suggesting that selection for higher level resistance to MRL-423 may provide additional insights into the cellular target(s). Reselection for resistance to MRL-423 in either the GyrB (F513L) or GyrB (L470I) mutant background identified additional rare spontaneous mutations in the same region of GyrB (Table 2, Fig 2C). The macromolecule synthesis profile and the identification of resistance mutations in a subunit of DNA gyrase strongly suggested that the initial hit molecule MRL-770 may be a DNA gyrase inhibitor. While the inhibition of E. coli DNA gyrase by MRL-770 is limited (S1 Fig) it did inspire us to make further modifications resulting in the synthesis of MRL-423 followed by MRL-1082. Fig 3A shows the results of a gel-based assay demonstrating inhibition of E. coli DNA gyrase by MRL-423 (IC 50~7 .5 μM). Consistent with the growth inhibitory activity of MRL-423 and MRL-1082 against efflux-defective P. aeruginosa (Table 1), MRL-1082 is active against P. aeruginosa DNA gyrase (Fig 3B). Our compounds did not exhibit any activity at the maximum achievable concentration (200 μM) against E. coli topoisomerase IV or S. aureus DNA gyrase (S2 Fig). With the improved antibacterial potency of MRL-423 and MRL-1082 we did observe relatively weak activity against mammalian cells in culture [41] at levels approximately 50-140-fold, respectively, above the MIC for E. coli HS151 (S1 Table). The mechanism of this cytological activity is unknown, but improvement in antibacterial activity of this series may expand that gap still further.
Certain gyrase inhibitors, such as the FQs, poison gyrase in such a way that the double strand break in the substrate DNA is stabilized (cleavage complex) [42]. To explore the mechanism of inhibition of MRL-423 we tested whether the observed gyrase inhibition is accompanied by a similar blockage in the gyrase reaction cycle resulting in cleavage complex stabilization. As depicted in Fig 3C, like ciprofloxacin, MRL-423 stabilizes cleavage complex formation in a dose-dependent fashion. Cleavage complex stabilization by the FQs may be associated with the bactericidal action of this class [43]. Similar to the FQs and consistent with the notion that cleavage complex stabilization is a lethal event, this novel class of gyrase inhibitors shows potent bactericidal activity causing a greater than 3-log reduction in culture viability at MRL-1082 concentrations of 2XMIC or greater following 2-4 hours of treatment (Fig 4).
An important prediction and test of the proposed mechanism of growth inhibition by these novel compounds is that DNA gyrase bearing a mutation conferring resistance in culture should also be resistant in vitro. To test the relationship between gyrase inhibition and bacterial growth inhibition DNA gyrase was reconstituted with either recombinant wild-type GyrB or GyrB (F513L) and inhibition by MRL-1082 was measured using the gel-based assay. We Plasmid-encoded resistance to the FQs (qnr) takes the form of a small peptide that interacts with gyrase and protects it from inhibitor binding [44,45]. Using a qnr determinant expressed from an E. coli plasmid (qnrVC3) [27], we tested whether this peptide can confer resistance to MRL-423. As shown in Table 3, qnrVC3 expression conferred approximately 8-fold resistance to CIP in E. coli HS151 but had no effect on MRL-423 potency. Another protein known to interact with DNA gyrase is the chromosomally-encoded, Zn-binding YacG protein [46]. By binding to a specific region of DNA gyrase, remodeling a region of the complex, and blocking DNA binding, YacG selectively inhibits enzyme function [47]. An important residue in YacG that contributes to that binding and inhibition is YacG F40. The region of GyrB that YacG F40 interacts with is formed, in part by GyrB F513, a residue that has a role in sensitivity to our novel gyrase inhibitors. To test the contribution of YacG to the activity of MRL-423, we deleted yacG from the chromosome of E. coli BW25113 (ΔtolC) by allele replacement. As shown in Table 3, deletion of yacG does not affect the potency of MRL-423.

Discussion
The work described here originated from the systematic evaluation of a large set of small molecules with antibacterial activity. We applied established, classical methods such as macromolecule synthesis and resistant mutant selection and newer technologies likewholegenome sequencing. As we refined this privileged set of bioactives we used these increasingly involved analyses to arrive at compounds with associated targets. One example is the novel gyrase inhibitor MRL-770. Taken together, the observations presented here demonstrate that  we have discovered and improved upon a novel class of DNA gyrase inhibitors. These bactericidal compounds are active against efflux and permeability defective Gram-negatives and resistance arises at a very low frequency. While presently limited to Gram-negatives with impaired efflux or a permeability barrier defect, this class of compounds holds promise to treat isolates with FQ resistance arising from target site mutations and plasmid encoded qnr factors. The X-ray structures of a number of bacterial type II topoisomerases have enabled a better understanding of the functioning of these molecular machines and aided in the discovery and refinement of enzyme inhibitors. The most elaborate structures comprise ternary complexes of gyrase, DNA substrate, and an inhibitor (e.g., [14,[48][49][50]). While we were not successful in obtaining a structure of our novel inhibitors in DNA gyrase, mapping of the amino acid residues in GyrB associated with resistance onto a model of DNA gyrase does provide some insights into the novelty and selectivity of these small molecules. To compare the sites of resistance with the binding of FQs and the relationship with the GyrA QRDR, we overlaid the E. coli DNA gyrase X-ray structure (PDB ID: 3NUH) [51] with a ciprofloxacin-bound S. aureus DNA gyrase X-ray structure (PDB ID: 2XCT) [14] and mapped the MRL-770/MRL-423 spontaneous resistance mutations onto GyrB (Fig 6). Most of the primary and reselected mutations reside adjacent to the ciprofloxacin binding site, but none appear to be in direct contact with ciprofloxacin. Glutamic acid 466 is the closest to the ciprofloxacin binding site (E477 in S. aureus DNA gyrase) at approximately 5 Å away. Interestingly, the only mutation we isolated at that site was a codon insertion conferring low level resistance to the FQs suggesting that an additional residue inserted at that site causes significant distortion of that region affecting binding of both FQs and our gyrase inhibitors. Additionally, cross-linking studies with FQ derivatives and E. coli DNA gyrase indicate a close juxtaposition of E466 and the C-7 pendant ring of ciprofloxacin [52]. GyrB F513 is the farthest away from the FQ/DNA binding region of any of the resistance mutations we isolated, but confers complete resistance to MRL-1082 at the cellular and enzyme level. This residue, F513, has been implicated in the binding of the gyrase regulator protein YacG (PDB ID: 4TMA) [47], but cells lacking YacG do not have altered sensitivity to our inhibitors. Two related compounds, QPT-1 and AZD0914 interact with the TOPRIM domain of S. aureus and Neisseria gonorrhoeae GyrB [17,49], but mutations at sites conferring resistance to those inhibitors (E.coli GyrB G425, D426, and K447) are distinct from the sites conferring resistance to the inhibitors described here. With the close spatial association of the amino acid residues and the selective resistance observed with the primary and secondary mutations we isolated, the gyrase inhibitors we describe here appear to interact with DNA gyrase in a novel way.
The two analogs of the initial hit, MRL-770, were derived from an extensive structure activity relationship exploration. Medicinal chemistry efforts were focused on improving the overall physiochemical properties, including a reduction in clogP, in an attempt to improve cell potency. Details of that medicinal chemistry effort and the subsequent in vitro results will be the subject of a future publication. Further enhancement of antibacterial potency and mitigation of any mammalian toxicity will likely come from a better understanding of the interaction of this series with DNA gyrase and balancing that binding with enhancing the cell permeability characteristics of the molecule. Application of biophysical methods to quantitate compound uptake may allow a clearer understanding of the properties of this series that govern bacterial cell permeability [53,54].