Molecular Characterization of Mycobacterium ulcerans DNA Gyrase and Identification of Mutations Reducing Susceptibility to Quinolones In Vitro

ABSTRACT Buruli ulcer disease is a neglected necrotizing and disabling cutaneous tropical illness caused by Mycobacterium ulcerans. Fluoroquinolone (FQ), used in the treatment of this disease, has been known to act by inhibiting the enzymatic activities of DNA gyrase. However, the detailed molecular basis of these characteristics and the FQ resistance mechanisms in M. ulcerans remains unknown. This study investigated the detailed molecular mechanism of M. ulcerans DNA gyrase and the contribution of FQ resistance in vitro using recombinant proteins from the M. ulcerans subsp. shinshuense and Agy99 strains with reduced sensitivity to FQs. The IC50 of FQs against Ala91Val and Asp95Gly mutants of M. ulcerans shinshuense and Agy99 GyrA subunits were 3.7- to 42.0-fold higher than those against wild-type (WT) enzyme. Similarly, the quinolone concentrations required to induce 25% of the maximum DNA cleavage (CC25) was 10- to 210-fold higher than those for the WT enzyme. Furthermore, the interaction between the amino acid residues of the WT/mutant M. ulcerans DNA gyrase and FQ side chains were assessed by molecular docking studies. This was the first elaborative study demonstrating the contribution of mutations in M. ulcerans DNA GyrA subunit to FQ resistance in vitro.

FQ is effective against M. ulcerans in vitro and in vivo (16,(20)(21)(22). Evidence exists that DNA topoisomerase II is the therapeutic target of the drug. Most eubacteria have two DNA topoisomerases II (DNA gyrase and DNA topoisomerase IV), which are essential for efficient DNA replication and transcription (23,24), and among a few clinically validated targets for antibacterial therapies (25,26). Remarkably, M. ulcerans expresses only DNA gyrase (27,28) from a gyrB linked gyrA contig in the complete genome and this enzyme is the sole target of FQs (26). The catalytically active mycobacterial DNA gyrase has a GyrA 2 GyrB 2 tetrameric structure (29) and is an ATP-dependent enzyme that transiently cleaves and unwinds double-stranded DNA to catalyze DNA negative supercoiling (30,31). However, the detailed molecular mechanism of M. ulcerans DNA gyrase and the mechanisms of FQ resistance were not determined.
This study aimed to determine the functional analysis of M. ulcerans DNA gyrase activities in vitro from M. ulcerans shinshuense and Agy99 strains. DNA gyrase subunits of both strains were expressed and purified as a recombinant protein and its activity was investigated in vitro via supercoiling assays. In addition, specific structural interactions between wild-type (WT)/mutant M. ulcerans DNA gyrase and FQs were identified via molecular docking. Because the FQs tested had limited activity against the FQ-resistant M. ulcerans DNA gyrase, the development and design of novel antibiotics against BU D are recommended.

RESULTS
Expression and purification of recombinant M. ulcerans DNA gyrases. The entire gene sequences of WT gyrA and gyrB from M. ulcerans shinshuense and Agy99 strains and mutant gyrA (Ala91Val and Asp95Gly) were amplified and inserted into expression vector pCold-I on downstream of the cspA promoter to heterologous express N-terminal hexahistidine-tagged gyrase subunits (Table 1). Molecular docking predicted that the his 6 -tag was located away from the FQ binding site of the M. ulcerans DNA GyrA (Fig. 1A) and GyrB (unpublished data) subunit suggesting that it will not interfere with GyrA activity. Expressed WT/mutant GyrA and GyrB subunits were purified to homogeneity using a two-step column chromatographic procedure described in Materials and Methods with the expected molecular masses of GyrA (93 kDa) and GyrB (76 kDa) subunits determined by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1B). All recombinant DNA gyrase subunits were obtained at high purity (.95%) in milligram amounts. Contaminating Escherichia coli (E. coli) topoisomerase activity was denied by the lack of supercoiling activities either only with M. ulcerans GyrA or GyrB subunit ( Fig. 2; lanes 2 and 3 for M. ulcerans shinshuense; lanes 6 and 7 for M. ulcerans Agy99).
Supercoiling activities of WT and mutant M. ulcerans DNA gyrase. The supercoiling activities of each DNA gyrase subunit were investigated by varying the subunit concentration  Mutant GyrA (Ala91Val and Asp95Gly) had DNA supercoiling activity in the presence of WT GyrB (data not shown). Furthermore, we found that the optimum temperature of M. ulcerans DNA gyrase was 30 to 37°C and its activity decreased at 40°C (Fig. 4) and is similar to those of the Mycobacterium leprae DNA gyrase from our previously reported (32). Therefore, all other assays were performed at 30°C.  (Table 2), with the structure of each FQ shown in Table 2A to D. IC 50 s from both strains of DNA gyrase were comparable to that observed ( Table 2). The mutant DNA gyrase was highly resistant to inhibition by CIP and LVX ( Fig. 5 and Table 2) with IC 50 s of .320 mg/mL, whereas the M. ulcerans shinshuense and Agy99 WT gyrase were 11.80 and 7.52 mg/mL, respectively (Table 2). To examine the effects of FQ on cleavage complex formation by M. ulcerans recombinant DNA gyrases, cleavage activities were performed in which supercoiled pBR322 DNA was incubated with WT or mutant DNA gyrases in the presence or absence of increasing concentrations of FQs. The representative results of cleavage activity using LVX against M. ulcerans Agy99 DNA gyrase were shown in Fig. 6, and Table 2 presents the CC 25 of CIP and MOX. The CC 25 of FQs for WT DNA gyrase ranged from 0.038 to 1.53 mg/mL, while those for the mutant DNA gyrases ranged from 1.04 to 67.68 mg/mL (Table 2).
Binding mode between DNA gyrase and FQs. The detailed interaction between WT/mutant M. ulcerans shinshuense DNA gyrase and FQs was determined via molecular docking using molecular operating environment (MOE) software (Fig. 7). The representative molecular docking results of CIP or MOX against M. ulcerans WT shinshuense and mutants DNA gyrase are shown in Fig. 7, whereas the Agy99 DNA gyrase data are not shown because the amino acid sequence has 95% homologous identity and FQ binding site was 100% (Fig. S1). The docking score and root mean square deviation (RMSD) between CIP and WT DNA gyrase, or GyrA mutants (Ala91Val/Asp95Gly) were 26.3081 and 4.02 Å, or 26.3239/26.4195 and 3.48/4.04 Å, respectively (Fig. 7C). Meanwhile, the docking score and RMSD between MOX and WT DNA gyrase, or GyrA mutants (Ala91Val/Asp95Gly) were 28.0615 and 2.15 Å, or  (Fig. 7A). However, the interaction of CIP with the Asp95Gly GyrA mutant was not observed (Fig. 7A) which correlated with the significantly higher IC 50 value of the Asp95Gly GyrA mutant compared with that of the WT and Ala91Val mutants. Similarly, the binding mode of MOX and distance were observed (Fig. 7B). Based on the molecular docking results, these amino acid residues in the GyrA subunit may play an important role in the interaction between M. ulcerans DNA gyrase and FQs.

DISCUSSION
Although the inhibitory effects of FQs against BU D are known in vitro and in vivo (16,(20)(21)(22), the molecular details between M. ulcerans DNA gyrase and FQs interactions are not understood.
Nakanaga et al. (32) reported the drug susceptibility test of M. ulcerans shinshuense and Agy99 strain. Their findings demonstrate the differences in MIC values between M. ulcerans shinshuense (0.25 mg/mL) and Agy99 (8.0 mg/mL) strains for LVX. Although the precise molecular mechanism contributing to this drastic difference in the MIC values of the two M. ulcerans strains has not been elucidated, it is speculated that these differences may be due to the drug efflux pump which may pump out the LVX, thereby raising the increased MIC value (33). In this regard, we are focused on the more detailed mechanism of FQ resistance against two strains in vitro using a bacterial recombinant system. Furthermore, we investigated the amino acid substitutions at positions 91 and 95 on the GyrA subunit from both M. ulcerans strains which is equivalent to positions 90 and 94 in M. tuberculosis known to contribute to FQ resistance (32,34,35). The resistance of mutant DNA gyrase to FQs was demonstrated at the molecular level using purified recombinantly expressed subunits by supercoiling and inhibition assays.
To measure the absolute concentration for DNA gyrase activity, the concentration of purified DNA gyrase subunits was calculated by two steps, because the DNA gyrase activity required the 1:1 ratio of GyrA and GyrB subunit. The first step was a general calculator system from the Qubit assay kit (data not shown), and the second step was a concentration-dependent supercoiling assay with variable concentrations of GyrA or GyrB subunits (Fig. 3). The optimum concentrations of GyrA and GyrB subunits (1:1 ratio) were decided. The WT M. ulcerans Agy99 and mutants (including M. ulcerans shinshuense mutants) were each 3 mM of GyrA and GyrB subunit (Fig. 3). However, WT M. ulcerans shinshuense GyrA and GyrB subunits used to determine the absolute enzyme activity were 3 mM and 1 mM, respectively (Fig. 3A). This estimation was based on the erroneous first step, which yielded an incorrect result for the WT M. ulcerans GyrB subunit. Hence, it was rectified to 3 mM WT M. ulcerans shinshuense GyrA and GyrB (Fig. 3A). The amino acid substitution of GyrA subunits at Ala91Val and Asp95Gly showed reduced sensitivity to all FQs, and more markedly than for WT DNA gyrase (2.19 to 14.65 mg/mL) (Fig. 5  and 6 and Table 2). MOX exhibited the highest inhibitory activity against M. ulcerans GyrA mutants among FQs examined in this study. The IC 50 values of CIP, MOX, and LVX were over 20-fold higher against M. ulcerans shinshuense DNA gyrase mutants than those for WT DNA gyrase ( Fig. 5 and Table 2). Similar results were observed for M. ulcerans Agy99 GyrA mutants ( Fig. 5 and Table 2) and this may be related to its amino acid sequence homology (95%) to the M. ulcerans shinshuense GyrA subunit (unpublished data). Furthermore, to examine the effects of quinolone on cleavage complex formation by M. ulcerans recombinant DNA gyrases were performed that a similar tendency was observed in the DNA cleavage activities, CC 25 was over 20-fold higher than for the WT DNA gyrase (Fig. 6 and Table 2). The observations suggested the contribution of amino acid substitutions Ala91Val and Asp95Gly in the GyrA subunit resulted in reduced sensitivity to FQs.
The structure-activity relationship between WT/mutant DNA gyrases and FQs was analyzed by molecular docking using MOE software (Fig. 7) to predict whether the GyrA subunit is associated with FQ resistance. A prediction of the three-dimensional (3D) structure of M. ulcerans shinshuense GyrA subunit was generated for structure-binding analysis with FQs because the crystal structure has not been reported. The WT or mutant M. ulcerans shinshuense GyrA subunit was modeled by Swiss-Model (https://swissmodel.expasy.org) using PDB 3IFZ (36) and 6RKS (37) as the templates (Fig. 7). The position of amino acid residues Ala91, Asp95, and Thr96 in WT GyrA interacted with that of the R3, R1, and R7 ring of CIP, respectively (Fig. 7A), which is suggested to contribute to potent inhibitory activity against M. ulcerans WT DNA gyrase (IC 50 of 11.80 mg/mL). Modeling showed that CIP tightly binds to the quinolone-binding site on the GyrA subunit through hydrogen bonding interactions CH 3 Table 2). In contrast, the amino acid substitution of Asp95 to Gly on the GyrA subunit leads to a significant reduction in interaction with CIP (IC 50 .320 mg/mL). This is primarily due to the disruption of the hydrogen bond interaction (Fig. 7A and Table 2), which induces extensive conformational changes in the quinolone-binding   98°C for 10 s, annealing at 60°C for 10 s, and extension at 72°C for 3 min 30 s, with a final extension at 72°C for 2 min. The PCR products corresponding to the 2.5-kb gyrA and 2.1-kb gyrB fragments were ligated into the TA cloning plasmid, transformed into E. coli Top 10 and plated onto Luria-Bertani (LB) agar containing ampicillin (100 mg/mL). Colonies were selected, and plasmids were purified using a Miniprep DNA purification kit (Promega Madison, WI, USA), followed by digestion with NdeI and XhoI (for gyrA)/HindIII (for gyrB). The gyrA and gyrB fragments were ligated into the pCold-I expression vector restriction digested with the same restriction endonucleases. Mutant M. ulcerans DNA gyrases (Ala91Val or Asp95Gly) were generated from WT gyrA and gyrB using a QuikChange site-directed mutagenesis kit (Agilent Technologies, Inc., Santa Clara, CA, USA) according to the manufacturer's instructions using the primers stated in Table 1. Plasmids were purified using a Miniprep DNA purification kit. WT and mutant plasmids were confirmed by sequencing (GENEWIZ corp., Tokyo, Japan) and were checked for errors by comparing to their respective WT and mutant sequences using BioEdit software 7.0.5.3.
Overexpression and purification of recombinant M. ulcerans DNA gyrase subunits. Recombinant DNA gyrase subunits were purified as previously described (38)(39)(40) with minor modifications. Briefly, the recombinant WT and mutant M. ulcerans gyrA and gyrB expression vectors were transformed into E. coli BL21(DE3) cells. Single colonies were picked and grown overnight at 37°C in 4 mL of LB medium containing 100 mg/mL ampicillin. Overnight cultures were used to inoculate 400 mL of fresh LB medium with ampicillin. Cells were cultured at 37°C for 7 to 8 h until the optical density (OD) at 600 nm reached 0.6 to 0.8, followed by the addition of 1 mM isopropyl b-D-1-thiogalactopyranoside (Wako Pure Chemicals Ltd.) to induce protein expression and incubated at 14°C for 18 h. Cells were harvested by centrifugation at 13,000 Â g for 20 min at 4°C and stored at 280°C for 12 h. Frozen cell pellets were resuspended in 20 mL ice-cold Talon binding buffer (50 mM sodium phosphate pH 7.4 and 300 mM NaCl) containing an EDTA-free protease inhibitor cocktail, and disrupted with a UP50H sonicator (Hielscher Ultrasonic, Teltow, Germany) using 10 cycles (40 s on/60 s off) at 80% pulsar power on the ice. The lysate was centrifuged at 9,400 Â g at 4°C for 20 min and the supernatant was applied onto a 5 mL His-Trap TALON crude column (GE Healthcare Bioscience, Piscataway, NJ, USA) preequilibrated with deionized water and Talon binding buffer. After sample application, the column was washed with Talon wash buffer (50 mM sodium phosphate pH 7.4, 300 mM NaCl, and 5 mM imidazole) until they reached a steady baseline. Proteins were eluted using Talon elution buffer (50 mM sodium phosphate pH 7.4, 300 mM NaCl, and 150 mM imidazole). The eluted proteins were concentrated using an Amicon Ultra-15 centrifugal filter unit (Millipore, Billerica, MA, USA) at 4,830 Â g at 4°C for 15 min. WT and mutant DNA gyrases were further purified by gel filtration chromatography (ÄKTA pure, GE Healthcare Bioscience) using a Hi-Load 16/600 Superdex 200 prep grade column (GE Healthcare Bioscience) equilibrated with 20 mM Tris-HCl (pH 8.0) to remove imidazole. The eluted peaks of the samples (280 nm) were assayed using supercoiling assays and analyzed using SDS-PAGE. Protein concentrations were determined using a Qubit assay kit (Thermo Fisher Scientific, Waltham, MA, USA).
DNA gyrase activities and inhibition by FQs. DNA supercoiling activity was determined using a combination of purified recombinant M. ulcerans GyrA and GyrB subunits as previously described (38)(39)(40). The reaction mixture (30 ml) consisted of supercoiling assay buffer (35 mM Tris-HCl pH 7.5, 24 mM KCl, 4 mM MgCl 2 , 2 mM DTT, 1.8 mM spermidine, 6 mM ATP, 0.1 mg/mL BSA, 6.5% wt/vol glycerol) and relaxed pBR322 DNA (0.3 mg) as the substrate. Assays were performed for 1 h at 30°C and stopped by the addition of 30 ml chloroform/iso-amyl alcohol (24/1) and 3 ml 10Â stop and loading solution (40% wt/vol sucrose, 100 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5 mg/mL bromophenol blue). The product of the reaction was separated by electrophoresis using a 1% agarose gel in 0.5Â Tris-borate-EDTA (pH 8.3) buffer for 90 min at 30 mA, followed by staining the agarose gel with ethidium bromide (0.7 mg/mL). Supercoiling activity was quantified by measuring the band brightness of supercoiled pBR322 DNA using Image J 1.52a (http://rsbweb.nih.gov/ij). A concentration-dependent supercoiling assay using 1 to 24 mM GyrA or GyrB subunit was performed to determine the optimal concentration of each DNA gyrase subunit. The optimal temperature of WT M. ulcerans DNA gyrase supercoiling activity was measured at 20, 25, 30, 37, 40, and 50°C using the same concentrations as stated above and the optimal DNA gyrase subunit concentrations. Inhibition of M. ulcerans DNA gyrase supercoiling activity by FQs followed previous methods (38)(39)(40) with minor modifications. Briefly, reaction mixtures containing optimal DNA gyrase subunits and increasing FQ concentrations (0 to 320 mg/mL) were assayed as described above. The inhibitory effects of FQs on DNA gyrase activity were assessed by determining the drug concentration required to inhibit the supercoiling activity by 50% (IC 50 ) using R studio free software version 1.4.1717. All assays were carried out at least three times and processed on the same day under identical conditions. To facilitate direct comparison, all incubations with WT and mutant DNA gyrase were carried out and processed in parallel on the same day under identical conditions, and assays were done at least three times, with reproducible results. Furthermore, to determine the more detailed functional role of M. ulcerans DNA gyrases, we performed FQs mediated DNA cleavage assays following previous methods (33,41). Supercoiled pBR322 DNA (0.1 mg) was used as the substrate for DNA cleavage assays, and linearized pBR322 DNA by HindIII digestion was used as a marker for cleaved DNA. The quinolone concentrations required to induce 25% of the maximum DNA cleavage (CC 25 ) were determined for CIP, MOX, and LVX.
Predicted binding mode between DNA gyrase and FQs by molecular docking. Molecular docking studies and visualization were conducted in the Molecular Operating Environment (MOE 2020.09, Chemical Computing Group ULC, Montreal, QC, Canada; https://www.chemcomp.com/index.htm). The docking model was designed using Swiss-Model (https://swissmodel.expasy.org). FQ coordinates were sketched using ChemBioDraw software (PerkinElmer, Waltham, MA, USA), and the amino acid substitutions were generated using the MOE-Protein builder module. Double-stranded oligonucleotides were adopted from the Research Collaboratory for Structural Bioinformatics Protein Data Bank (PDB ID: 6RKS) (37) because our docking model was not included in the double-strand oligonucleotide. Modified docking models were prepared using a flexible docking method with the scores expressed as a sum of five potentials: accessible surface area, Coulomb potential, hydrogen bonds, anisotropy, and van der Waals interactions and refined by MOE through energy minimization. The DNA gyrase and FQ binding energies were estimated using the Amber10: EHT force field and the implicit solvation model of the reaction field was selected. The best binding models were selected for the lowest free energies and optimized RMSD refinement. The distance between amino acid residues on GyrA and the side chain of FQs was calculated using WinCoot-0.9.4.1 (https://bernhardcl.github.io/coot/), and molecular graphics were generated using PyMOL v1.8 (https://pymol.org/2/).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.9 MB.