Evaluation of Bronopol and Disulfiram as Potential Candidatus Liberibacter asiaticus Inosine 5′-Monophosphate Dehydrogenase Inhibitors by Using Molecular Docking and Enzyme Kinetic

Citrus huanglongbing (HLB) is a destructive disease that causes significant damage to many citrus producing areas worldwide. To date, no strategy against this disease has been established. Inosine 5′-monophosphate dehydrogenase (IMPDH) plays crucial roles in the de novo synthesis of guanine nucleotides. This enzyme is used as a potential target to treat bacterial infection. In this study, the crystal structure of a deletion mutant of CLas IMPDHΔ98-201 in the apo form was determined. Eight known bioactive compounds were used as ligands for molecular docking. The results showed that bronopol and disulfiram bound to CLas IMPDHΔ98-201 with high affinity. These compounds were tested for their inhibition against CLas IMPDHΔ98-201 activity. Bronopol and disulfiram showed high inhibition at nanomolar concentrations, and bronopol was found to be the most potent molecule (Ki = 234 nM). The Ki value of disulfiram was 616 nM. These results suggest that bronopol and disulfiram can be considered potential candidate agents for the development of CLas inhibitors.


Introduction
Huanglongbing (HLB) is one of the most destructive citrus diseases; it affects the citrus industry worldwide. HLB is a phloem-restricted, Gram-negative bacterium and caused by Candidatus Liberibacter asiaticus (CLas), Candidatus Liberibacter africanus, and Candidatus Liberibacter americanus. The pathogen is transmitted by the citrus psyllid [1]. CLas is highly virulent and distributed worldwide. HLB can infect all commercial varieties of cultivated citrus, and has caused enormous economic losses in the past [2]. Although some developments on CLas and plant-Liberibacter interaction have been achieved, no effective management method is presently available to control this disease once the trees are infected [3][4][5][6][7][8][9]. In September 2019, Ha and Beyenal from Washington State University were part of a research team that determined the process of culturing in a laboratory the bacteria that cause citrus greening. However, the culture technology of HLB has not been used in the development of drugs to control HLB [10]. Chemical control is considered to be an effective method to control citrus HLB. Controlling the transmit vector is a critical component which can slow down the spread, but it is not sufficient to eliminate this disease. Additional attempts have focused on the pathogen, and some broad-spectrum antimicrobials have been used against Ca. Liberibacter spp [11,12]. Streptomycin, Molecules 2020, 25, 2313 3 of 14 recombinant CLas IMPDH∆98-201 was soluble and stable. This protein was purified using a Ni-NTA resin affinity chromatograph and a high-resolution gel filtration column (Superdex 200), which showed a main peak (Figure 1b). CLas IMPDH∆98-201 consisted of 390 amino acids with a theoretical molecular mass of 41 kDa. CLas IMPDH∆98-201 appeared as a single band at approximately 40 kDa (Figure 1c).
Molecules 2020, 25, x FOR PEER REVIEW 3 of 14 and the catalytic residue Cys309 was completely conserved (Figure 1a). In the E. coli express system, recombinant CLas IMPDHΔ98-201 was soluble and stable. This protein was purified using a Ni-NTA resin affinity chromatograph and a high-resolution gel filtration column (Superdex 200), which showed a main peak (Figure 1b). CLas IMPDHΔ98-201 consisted of 390 amino acids with a theoretical molecular mass of 41 kDa. CLas IMPDHΔ98-201 appeared as a single band at approximately 40 kDa (Figure 1c). The initial crystallization conditions that were tested from Index, SaltRx, PEG/Ion Screen, Crystal Screen kits (Hampton Research, Aliso Viejo, CA, USA), and the Wizard kit (Emerald BioSystems, Bainbridge Island, WA, USA). After the initial screening, crystals formed under two conditions only. After further optimization, diffraction-quality crystals were obtained by mixing 1 μl of protein solution at 8 mg/mL with 1 μl of reservoir solution (consisting of 30% PEG400, 200 mM sodium chloride, and 100 mM HEPES, pH 7.0) at 20 °C. Long rectangular crystals of approximately 0.2 × 0.1 × 0.05 mm formed (Figure 1d).

Crystal Structure and Loop Refinement of CLas IMPDHΔ98-201
Crystals of CLas IMPDHΔ98-201 appeared after 3 days at 293 K. The resolution of the diffracting crystal was 2.55 Å. Data collection and refinement parameters are summarized in Table 1.  The initial crystallization conditions that were tested from Index, SaltRx, PEG/Ion Screen, Crystal Screen kits (Hampton Research, Aliso Viejo, CA, USA), and the Wizard kit (Emerald BioSystems, Bainbridge Island, WA, USA). After the initial screening, crystals formed under two conditions only. After further optimization, diffraction-quality crystals were obtained by mixing 1 µL of protein solution at 8 mg/mL with 1 µL of reservoir solution (consisting of 30% PEG400, 200 mM sodium chloride, and 100 mM HEPES, pH 7.0) at 20 • C. Long rectangular crystals of approximately 0.2 × 0.1 × 0.05 mm formed ( Figure 1d).

Crystal Structure and Loop Refinement of CLas IMPDH∆98-201
Crystals of CLas IMPDH∆98-201 appeared after 3 days at 293 K. The resolution of the diffracting crystal was 2.55 Å. Data collection and refinement parameters are summarized in Table 1. The structure of CLas IMPDH∆98-201 was determined through molecular replacement using IMPDH from Campylobacter jejuni (PDB entry 4R7J) as a template. Finally, the structure was refined to 2.55 Å resolution by using the PHENIX software. This crystal protein existed as a homotetramer (Figure 2a  The structure of CLas IMPDHΔ98-201 was determined through molecular replacement using IMPDH from Campylobacter jejuni (PDB entry 4R7J) as a template. Finally, the structure was refined to 2.55 Å resolution by using the PHENIX software. This crystal protein existed as a homotetramer (Figure 2a  The flap loop and a C-terminal loop are not visible in the electron density. Thus, the loop refinement of CLas IMPDH∆98-201 (PDB ID: 6KCF) was performed using Modeller 9.23 ( Figure 2c). The nonterminal missing structure was refined ( Figure 2d). Verification of the 3D results showed that 88.27% of the amino acid residues had an average 3D-1D score ≥ 0.2 ( Figure S2). The Ramachandran plot analysis indicated that 82.4% of the residues were in the core region, 13.4% of the residues were in the allowed region, 2.9% of the residues were in the generously allowed region, and 1.3% of the residues were in the disallowed region ( Figure S3).

Molecular Docking
The eight candidate compounds and the refined structure were selected to perform molecular docking. CDOCKER was used to perform a docking study of the selected molecule; the molecular docking binding affinities are shown in Table 2. Three molecules, namely, bronopol, mercaptopurine, and disulfiram, showed the -CDOCKER_ENERGY ≥ 10. Because mercaptopurine is an analog of IMP, it was hypothesized that bronopol and disulfiram would exhibit the best inhibitory effect for CLas IMPDH. The pose with the lowest binding energy was recognized as the most stable conformation for further structural analysis. The 3D and 2D structures of the CLas IMPDH∆98-201 with bronopol and disulfiram are displayed in Figure 3. Nine hydrogen bonds formed between bronopol and the residues ILE189, Gly190, Gly192, ASP228, Gly229, Gly230, Gly251, and Ser252 of CLas IMPDH∆98-201 (Figure 3a,b). Disulfiram formed two hydrogen bonds with CLas IMPDH∆98-201, namely, Ala41 and Ala42; four alkyl hydrophobic interaction with Met43, Pro190, and Met249; and one sulfur-x interaction with Met43 ( Figure 3c,d). The 3D and 2D structures of the CLas IMPDH∆98-201 with the rest of molecules are displayed in Figure S4.

Kinetic Characterization of CLas IMPDHΔ98-201
According to the standard assay conditions, the kinetic properties of CLas IMPDHΔ98-201 were as follows: Kcat = 7.2 ± 0.

Kinetic Characterization of CLas IMPDHΔ98-201
According to the standard assay conditions, the kinetic properties of CLas IMPDHΔ98-201 were as follows: Kcat = 7.2 ± 0.   The steady-state parameters from other bacterial species are listed in due to the fact that the results described here were measured at 30 • C, whereas the other IMPDHs were measured at the lower temperature of 25 • C.

Inhibitory Assay against CLas IMPDH∆98-201 Enzyme Activity
Extending the measurement time, no exponential enzyme decay against CLas IMPDH∆98-201 was observed. Hence, the inhibition of bronopol, disulfiram, and ebselen was treated as a reversible mode ( Figure S5). As shown in Figure S6a, the V max was found to be reduced with an increase in the inhibitor concentration, suggesting that bronopol inhibited CLas IMPDH∆98-201 in a noncompetitive manner against IMP. Disulfiram also inhibited CLas IMPDH in a noncompetitive manner against IMP, where regression lines meet on the X-axis ( Figure S6b). The various types of inhibition by other small molecule inhibitors are summarized in Figure S6.
To study the mechanism of enzyme inhibition, the inhibition constant K i with respect to the IMP substrates was measured at a fixed NAD + concentration. The K i values of these eight compounds are summarized in Table 3. All values ranged from 0.234 µM to 3500 µM. Although the percentage of DMSO and the high concentration of the compound affected the stability of the target protein, the values for mizoribine and ribavirin may have been inaccurate ( Figure S7c,f). Ribavirin is a guanosine analog with broad-spectrum activity against RNA virus [58], and has almost no effect on the CLas IMPDH∆98-201 enzyme activity. Mizoribine is an imidazole nucleoside which is used as an immunosuppressive agent [59]. Mizoribine was a potent inhibitor of IMPDHs, with K i = 307.7 µM for CLas IMPDH∆98-210, whereas the K i value of E. coli IMPDH was 0.5 µM. Mercaptopurine yielded uncompetitive inhibition with K i = 165 µM ( Figure S7b). Mycophenolic acid was shown to be a potent inhibitor of mammalian IMPDHs with K i = 2.43 µM ( Figure S7d). Mycophenolate mofetil is a prodrug of mycophenolic acid [60], yielding K i = 24.42 µM ( Figure S7e). Three compounds, namely, disulfiram, bronopol, and ebselen, have been repurposed as IMPDH inhibitors [61]. Bronopol had the best inhibitory effect with K i = 234 nM (Figure 5a). The K i value of disulfiram was 616 nM (Figure 5b). The K i values of ebselen was 4.13 µM ( Figure S7a).
Molecules 2020, 25, x FOR PEER REVIEW 7 of 14 but was the highest affinity binding NAD + among the tested IMPDHs. The Kcat value may be due to the fact that the results described here were measured at 30 °C, whereas the other IMPDHs were measured at the lower temperature of 25 °C.

Inhibitory Assay Against CLas IMPDHΔ98-201 Enzyme Activity
Extending the measurement time, no exponential enzyme decay against CLas IMPDHΔ98-201 was observed. Hence, the inhibition of bronopol, disulfiram, and ebselen was treated as a reversible mode ( Figure S5). As shown in Figure S6a, the Vmax was found to be reduced with an increase in the inhibitor concentration, suggesting that bronopol inhibited CLas IMPDHΔ98-201 in a noncompetitive manner against IMP. Disulfiram also inhibited CLas IMPDH in a noncompetitive manner against IMP, where regression lines meet on the X-axis ( Figure S6b). The various types of inhibition by other small molecule inhibitors are summarized in Figure S6.
To study the mechanism of enzyme inhibition, the inhibition constant Ki with respect to the IMP substrates was measured at a fixed NAD + concentration. The Ki values of these eight compounds are summarized in Table 3. All values ranged from 0.234 μM to 3500 μM. Although the percentage of DMSO and the high concentration of the compound affected the stability of the target protein, the values for mizoribine and ribavirin may have been inaccurate ( Figure S7c,f). Ribavirin is a guanosine analog with broadspectrum activity against RNA virus [58], and has almost no effect on the CLas IMPDHΔ98-201 enzyme activity. Mizoribine is an imidazole nucleoside which is used as an immunosuppressive agent [59]. Mizoribine was a potent inhibitor of IMPDHs, with Ki = 307.7 μM for CLas IMPDHΔ98-210, whereas the Ki value of E. coli IMPDH was 0.5 μM. Mercaptopurine yielded uncompetitive inhibition with Ki = 165 μM ( Figure S7b). Mycophenolic acid was shown to be a potent inhibitor of mammalian IMPDHs with Ki = 2.43 μM ( Figure S7d). Mycophenolate mofetil is a prodrug of mycophenolic acid [60], yielding Ki = 24.42 μM ( Figure S7e). Three compounds, namely, disulfiram, bronopol, and ebselen, have been repurposed as IMPDH inhibitors [61]. Bronopol had the best inhibitory effect with Ki = 234 nM (Figure 5a). The Ki value of disulfiram was 616 nM (Figure 5b). The Ki values of ebselen was 4.13 μM ( Figure S7a).

Discussion
CLas causes HLB and affects citrus. Although HLB has become a global problem, no effective HLB management strategy is available [11]. IMPDH is a validated target for the design of potent antibacterial agents, and the inhibition of this enzyme depletes cellular guanine nucleotides [36]. The development of inhibitors against bacterial IMPDHs has attracted increasing attention [62]. This study focused on the development of CLas IMPDH inhibitors. The first structure of CLas IMPDH∆98-201 was determined. On the basis of its crystal structure, the refined structure was constructed, and molecular docking was performed to predict the binding energy. Then, we used an inhibition assay against CLas IMPDH∆98-201 to validate the molecular docking predictions.

Purification and Crystallization of CLas IMPDH∆98-201
To overcome the instability of CLas IMPDH, CLas IMPDH mutation was designed and purified. In the solution, recombinant CLas IMPDH∆98-201 was more stable than the wild type. MtbIMPDH2 without the CBS domain displayed higher solubility [51]. The steady-state kinetics parameters of CLas IMPDH∆98-201 were similar to those of other IMPDHs (Table S2), suggesting that deleting the CBS domain would not affect the CLas IMPDH∆98-201 catalytic properties [7]. Crystals of the apo form of CLas IMPDH∆98-201 were obtained in 100 mM HEPES (pH 7.5) and 200 mM NaCl, with 30% (w/v) PEG 4000 as the precipitant. The RMSD of CLas IMPDH∆98-201 and BaIMPDH∆95-200 was 0.929 Å.

Docking Interaction Analysis of CLas IMPDH∆98-201 with Molecules
To find inhibitors of CLas IMPDH∆98-201, molecular docking was performed using Discovery Studio 2018. The docking scores of bronopol and disulfiram binding to CLas IMPDH∆98-201 were −11.19 and −25.03 kcal/mol, respectively. Bronopol was stabilized by nine hydrogen bond interactions with residues ILE189, Gly190, Gly192, ASP228, Gly229, Gly230, Gly251, and Ser252. Additionally, disulfiram was stabilized by hydrophobic and sulfur-x interactions. Given that a flap loop and a C-terminal loop were not visible in the apo form structure of CLas IMPDH∆98-201, the nonterminal missing structure of CLas IMPDH∆98-201 was refined by Modeller. The Ramachandran plot and Verify 3D analysis suggested that the refined CLas IMPDH∆98-201 structure was reliable. Bacterial IMPDHs were similar in sequence and structure ( Figure S1). Homology modeling and in silico docking were performed to study the structure-activity relationship of indole derivatives against Helicobacter pylori IMPDH [63]. The crystal structure of IMPDH from Cricetulus griseus was prepared by using Discovery Studio 2.5 to build a pharmacophore model of IMPDH inhibitors and for the in silico docking analysis [64]. These studies supported the feasibility of molecular docking.

Inhibitory Assay against CLas IMPDH∆98-201 Activity
To explore the inhibition of the eight compounds, an inhibitory assay against CLas IMPDH∆98-201 activity was measured by monitoring the production of NADH. The inhibitions of BaIMPDH92-220, CjIMPDH∆92-195, and ClpIMPDH∆89-215 to a given compound showed significant differences, although the same residues interacted with the inhibitor [7]. A previous study found that a single residue showed mycophenolic acid resistance, although the binding sites were identical [65]. The kinetic mechanism was controlled for the mycophenolic acid resistance of PbIMPDH-A and PbIMPDH-B [66]. These studies showed that virtual screening by simple prediction is fast and low cost, although experimental verification is needed. Many other compounds against HLB have been reported. Five compounds, namely, C16, C17, C18, C19, and C20, were identified against CLas SecA, with IC50 values of 0.25, 0.92, 0.48, 0.64, and 0.44 µM, respectively [16]. ZINC05491830 is one of the most potent inhibitors of CLas Esbp, with an IC50 value of 2.59 µM [19]. ChemDiv C549-0604 is an inhibitors of CLas VisNR, with an IC50 value of 0.7 µM [18]. The inhibition assay suggested that the K i values of bronopol and disulfiram were 234 and 616 nM, respectively. The inhibition of CLas IMPDH∆98-201 Molecules 2020, 25, 2313 9 of 14 suggested that bronopol and disulfiram, unlike the aforementioned other compounds, could be used as CLas IMPDH∆98-201 inhibitors against other CLas genes.

Cloning and Mutant Construction of CLas IMPDH Gene
The coding sequence of IMPDH was amplified by PCR from the chromosomal DNA of CLas (strain psy62). The PCR product was cloned into the pET28at-plus expression vector.
The CBS domain deletion mutant (CLas IMPDH∆98-201) was constructed via splicing overlapping extension polymerase chain reaction (PCR). The ∆S construct involved the deletion of 104 residues from M98 to T201. The CLas IMPDH gene in vector pET28at-plus was used as a template. The F1 and R1 primers were applied to amplify a region of CLas IMPDH ranging from residue M1 to residue M98. The F2 and R2 primers were used to amplify a region of CLas IMPDH ranging from residue T201 to residue I493. Codons for residues M98-T201 were replaced with codons for G. I1 and I2 were used as templates. PCR was performed to amplify the CLas IMPDH CBS domain deletion mutant gene by using the F1 and R2 primers. The CLas IMPDH∆98-201 gene was digested by Bam HI and Xho I and inserted into a pET28a-SUMO vector. Then, pET28a-SUMO-CLas IMPDH∆98-201 was transformed into E. coli BL21(DE3) cells.

Protein Purification and Crystallization of CLas IMPDH∆98-201
Cells carrying pET28a-SUMO-CLas IMPDH∆98-201 plasmid were cultured in LB media supplemented with 50 µg/mL of kanamycin at 37 • C. The culture was induced by adding 0.3 mM of isopropyl-β-D-thiogalactopyranoside when its OD 600 reached 0.8-1.0. After 20 h of incubation at 16 • C, the cells were harvested by centrifugation at 6000 rpm for 6 min at 4 • C, resuspended in lysis buffer [20 mM Tris-HCl (pH 8.0), 500 mM KCl, 40 mM imidazole, 1 mM PMSF, and 10% glycerol], and then sonicated. The lysate was clarified by centrifugation at 16,000 rpm for 50 min at 4 • C. Clarified lysate was subsequently purified on a Ni-NTA agarose column, and the protein was eluted with the same buffer containing 500 mM imidazole. The SUMO tag was subsequently removed with the Ulp1 protease at 16 • C for 1 h. The target protein was additionally purified using a Ni affinity chromatograph to remove the released tag and uncut protein, followed by a size exclusion chromatography step on a Superdex TM 200 (GE Healthcare) column equilibrated with buffer [20 mM Tris-HCl (pH 8.0), 100 mM KCl, and 10% glycerol]. All proteins were purified according to this protocol.
Crystallization screening was set up using the sitting drop vapor diffusion method in 96-well plates. Crystals of the protein appeared after 3 days at 293 K. The best crystals of CLas IMPDH∆98-201 were obtained by mixing 1 µL of protein solution at 8 mg/mL with 1 µL of reservoir solution consisting of 30% (v/v) PEG 400, 100 mM HEPES (pH 7.5), and 200 mM sodium chloride.

Data Collection and Processing
Crystals were mounted on nylon loops and flash-cooled in liquid nitrogen. Diffraction data were collected at 100 K by using the Q315r CCD detector at the BL17U beamline of the Shanghai Synchrotron Radiation Facility. Single wavelength data at 0.9792 Å were obtained, and all data were processed and scaled with HKL3000 [67]. The structure of the CLas IMPDH∆98-201 was solved by molecular replacement using PHENIX [68]. The refined model and structure factors were deposited in the PDB.

Loop Refinement and Molecular Docking
The 3D structure of CLas IMPDH∆98-201 was used for refinement. The Modeller program was used to refine the nonterminal missing structure [69]. This refined structure consisted of 358 amino acids (CLas IMPDH 12-98 and 202-472). The PROCHECK validation server was used to check the quality of the refined model [70]. This structure was also validated by Verify 3D [71].
Molecular docking was performed using CDOCKER, a frequently applied module of Discovery Studio 2018. CDOCKER employs a CHARMm force field to calculate the binding free energy of the ligand to the receptor [72]. The eight filtered molecules used for docking were bronopol, ebselen, mercaptopurine, mizoribine, mycophenolate_mofetil, mycophenolic acid, ribavirin, and disulfiram. In the docking experiment, the refined structure of CLas IMPDH∆98-201 was used as the receptor; the docking parameters are listed in Table S1. The best pose of each molecular binding with a refined structure was estimated according to the binding energy. Interactions between the compound and protein, such as van der Waals force, hydrogen bond, electrostatic, hydrophobic, and halogen, were analyzed.

Steady-State Kinetics
Standard enzyme activity assay was performed in an assay buffer (50 mM Tris-HCl, 100 mM KCl, 1 mM DTT, pH 8.0) and a final CLas IMPDH∆98-201 enzyme concentration of 100 nM at 30 • C. The production of NADH was monitored by the increase in absorbance at 340 nm (E = 6.22 mM −1 × 0007 cm −1 ). Apparent steady-state kinetic parameters were evaluated at varying concentrations of IMP (0.005-1 mM) and a fixed saturating concentration of NAD + (3 mM), or at varying concentrations of NAD + (0.005-5 mM) and a fixed saturation level of IMP (1 mM). Assays were performed in duplicate. The IMPDH enzymes displayed strong substrate inhibition with respect to NAD + under the standard assay conditions. The method described by Kerr et al. was used to determine the kinetic constants [73].

Inhibition Assay against IMPDH∆98-201 of CLas
The eight molecules that were purchased were screened in vitro. The assay was performed in a 200 µL final volume in a 96-well plate with a reaction buffer consisting of 50 mM Tris-HCl (pH 8.0), 100 mM KCl, and 1 mM dithiothreitol. Assays were performed using 100 nM CLas IMPDH∆98-201 in the presence or absence of test compounds. The assay was allowed to proceed at 30 • C for 60 min.
The value of K i for eight molecules was determined at a fixed saturation concentration of NAD + The concentration of mercaptopurine ranged from 50 to 200 µM. The concentration of mizoribine was 500 and 750 µM. The concentration of ribavirin was 500 and 800 µM. Each determination of K i was derived from duplicate measurements.
To determine the K i values, the initial rate data versus substrate concentration at different inhibitor concentrations was fitted using Prism software (GraphPad) to equations for competitive, noncompetitive, or uncompetitive inhibition.

Conclusions
In conclusion, bronopol and disulfiram were confirmed as CLas IMPDH inhibitors. These results indicate that these compounds could be used as the lead scaffold to further design and develop potent CLas IMPDH inhibitors. However, the effect of compounds with activity against CLas was not tested. In future studies, we will focus on the effect of compounds in the treatment of HLB diseases. The apo form structure of CLas IMPDH∆98-201 was solved, providing a means to study the complex structure of cocrystallization with inhibitors. The binding information may be helpful for the further development of antimicrobial compounds against CLas.