A novel protein kinase inhibitor IMB-YH-8 with anti-tuberculosis activity

Protein kinase B (PknB) is one of the Mycobacterium tuberculosis serine/threonine protein kinases and has an essential role in sustaining mycobacterial growth. Here, we identified and characterized a novel small molecule compound named IMB-YH-8 that inhibited PknB and served as anti-mycobacteria lead compound. IMB-YH-8 inhibited PknB auto-phosphorylation and the phosphorylation of GarA by PknB in a dose-dependent manner. The compound did not inhibit human Akt1 or other serine/threonine kinases in M. tuberculosis except for the highly homologous PknA. IMB-YH-8 bound to PknB with a moderate affinity. Molecular docking revealed that IMB-YH-8 interacts with the catalytic domain of PknB. Observations of electron microscopy showed that IMB-YH-8 changed the morphology of H37Rv and disrupted the cell wall. The differential transcriptional response of M. tuberculosis to IMB-YH-8 revealed changes in SigH regulatory pathways modulated by PknB. Notably IMB-YH-8 not only potently inhibited drug-sensitive and multidrug-resistant clinical isolates but also exhibited a dose dependent inhibition of intracellular M. tuberculosis. Taken together, these in vitro data demonstrate that IMB-YH-8 is a novel inhibitor of PknB, which potently prevents growth of M. tuberculosis. It is as yet unclear whether inhibition of PknA contributes to the anti-tubercular action of IMB-YH-8.

IMB-YH-8 inhibits auto-phosphorylation and substrate phosphorylation of PknB with good specificity. The interaction between IMB-YH-8 and PknB was determined in the binding affinity experiments and docking study. Gene expression profiles show that IMB-YH-8 modulates PknB and SigH regulatory pathways.

IMB-YH-8 inhibits auto-phosphorylation and substrate phosphorylation activities of
PknB. Results of our high throughput screening assay revealed IMB-YH-8 (C 12 H 12 O 4 , MW: 220.07) as an inhibitor of M. tuberculosis PknB 13 . To confirm IMB-YH-8 inhibition of PknB, we measured the effect of IMB-YH-8 on PknB auto-phosphorylation and PknB-mediated phosphorylation of GarA. PknB is activated by auto-phosphorylation of Ser and Thr residues to control phospho-signaling pathways [14][15][16] . In the non-radioactive assay, IMB-YH-8 inhibited auto-phosphorylation of PknB with an IC 50 of 20.2 μM (Table 1). In order to assess the effect of IMB-YH-8 on substrate phosphorylation of PknB, a Forkhead-associated (FHA) domain-containing protein (GarA) was utilized as the substrate of PknB in the phosphorylation assay. It is known that PknB efficiently phosphorylates GarA at a single N-terminal threonine residue Thr22 17 . Phosphorylated and non-phosphorylated GarA proteins were separated using SDS-PAGE containing Phos-tag acrylamide . Phos-tag acrylamide provides a phosphate affinity SDS-PAGE for mobility shift detection of phosphorylated proteins. In the presence of MnCl 2 , phosphorylated proteins migrate slower than the non-phosphorylated form due to phosphate trapping by the Phos-tag chemical 18 . While use of SDS-PAGE alone produced the equal amount of GarA protein, samples separated on SDS-PAGE containing 20 μM Phos-tag showed two discernible bands (Fig. 1). In lane 1, the non-phosphorylated GarA forms a lower band, while in lane 2, ATPtreated GarA was phosphorylated and showed as an upper band (Fig. 1). Treatment with IMB-YH-8 increased the signal of the non-phosphorylated GarA band, indicating that IMB-YH-8 inhibited the PknB-catalyzed phosphorylation of GarA. The level of non-phosphorylated GarA increased in a dose-dependent manner (Fig. 1, lanes 3-5). Taken together, these results demonstrate that IMB-YH-8 could inhibit both auto-phosphorylation (Table 1) and substrate phosphorylation by PknB.

IMB-YH-8 selectively inhibits
PknB and PknA without affecting other STPKs. The STPKs share a well-conserved catalytic scaffold 12,16 . In order to investigate the specificity of IMB-YH-8 in inhibiting serine/ threonine kinases, we tested the effect of IMB-YH-8 on the auto-phosphorylation activity of 5 M. tuberculosis STPKs (PknA, PknB, PknG, PknF and PknH) and human kinase Akt1. It was observed that IMB-YH-8 did not affect PknG, PknH, PknF or Akt1 activity (Table 2). Not surprisingly, IMB-YH-8 inhibited PknA at an IC 50 of 44.3 μM which is two-fold higher than the IC 50 of 20.2 μM measured for PknB (Table 2). This is because the kinase domains of PknA and PknB are highly conserved in sequence and structure 6,19 . These results demonstrate that IMB-YH-8 selectively inhibits PknB and its homolog PknA of M. tuberculosis, but does not inhibit other M. tuberculosis STPKs or human serine/threonine kinase Akt1.   ITC provides the thermodynamic signature of a small molecule when it interacts with its macromolecular target in terms of enthalpy change (ΔH), entropy change (ΔS), and Gibbs free energy change (ΔG) along with the binding affinity constant (Ka) and the stoichiometry. Results of the ITC experiments showed that the binding of IMB-YH-8 to PknB was a single binding event and an exothermic process (Fig. 2b). The K d value of IMB-YH-8 (13.5 μM) was calculated from observed K a value, which is consistent with the data derived from SPR measurements and the IC 50 value. Binding of IMB-YH-8 to PknB was enthalpy-driven (ΔH = −5.7 kcal/mol) together with a favorable entropic contribution (−TΔS = −0.94 kcal/mol), which results in a ΔG of −6.64 kcal/mol. These data indicate that IMB-YH-8 interacts with PknB with a moderate binding affinity.

Molecular docking reveals that IMB-YH-8 interacts with the catalytic domain of PknB.
We next utilized molecular docking to illustrate IMB-YH-8 binding to PknB. During the docking studies, the original ligand in the crystal structure mitoxantrone (MIX) was selected as the reference compound. The predicted binding mode by molecular docking is quite close to that in crystal ribbon structure with the RMSD value of 1.0 Å (Fig. 3a), which illustrates the reliability of the docking method. We next docked IMB-YH-8 into the binding site at the catalytic domain of PknB using the same protocols. The results ( Fig. 3b and c) showed that IMB-YH-8 formed a hydrogen bond with Val95 and engaged in hydrophobic interactions with Leu17, Phe19, Val25, Ala38, Met92, Val95, Met145 and Met155. By further alignment with MIX in the binding site, we observed that the benzene ring of IMB-YH-8 matched well with the "A" ring of MIX (Fig. 3c).   Inactivation of PknB changes cell morphology 6 . Therefore, we next measured the effect of IMB-YH-8 on cell morphology and assessed its inhibition of PknB in vivo. Cells were treated with IMB-YH-8 and then examined by scanning electron microscopy. While the untreated control cells displayed a rod-like shape and a smooth cell surface (Fig. 4a), treatment with IMB-YH-8 at 1 μg/ml for 5 days led to a weakening in the overall cell structure (Fig. 4b). We observed significant proportions (61.5%) of ruptured cells in 1 μg/ml IMB-YH-8 treated M. tuberculosis H37Rv (Fig. 4b).Given the role of PknB in maintaining cell morphology and keeping cell wall integrity, these data suggest that IMB-YH-8 disrupts the cell wall of mycobacteria by inhibiting PknB functions in vivo.

IMB-YH-8 modulates PknB and SigH regulatory pathways. To monitor the IMB-YH-8-induced
gene response profile at an early stage in the inhibitory effect of the drug, we treated logarithmically growing M. tuberculosis with IMB-YH-8 at the concentration of 2.5 μg/ml throughout the first 4 hr. Gene transcripts showing a mean change of at least twofold in the normalized intensity values were considered as differentially regulated genes following induction by IMB-YH-8. Five genes related to PknB including trxB2 (rv3913), trxC (rv3914), rv2466c, rshA (rv3221A) and sigH (rv3223c) were found to be overexpressed during IMB-YH-8 exposure ( Table 3). The whole transcriptional profiling data has been deposited to the Gene Expression Omnibus (GEO) database under the accession number GSE90858. Microarray results were validated by real-time RT-PCR analysis of the above genes from IMB-YH-8-treated and untreated H37Rv. The analysis of the up-regulated genes for the IMB-YH-8 stress condition showed a similar trend (Table 3). Among these five up-regulated genes, rshA showed the highest induction (>260-fold) followed by rv2466c (about 80-fold).

IMB-YH-8 inhibits both drug-sensitive and drug-resistant M. tuberculosis. IMB-YH-8 inhibited
the growth of the M. tuberculosis strain H37Rv (MIC = 0.25 μg/ml) and has acceptable cytotoxicity against THP-1 cells with a selectivity index of 30.4 (Table 1). We further tested a panel of 14 M. tuberculosis clinical isolates with varying drug resistance profiles including those resistant to isoniazid, rifampin, streptomycin, ethambutol and ofloxacin by determining MICs using a microplate Alamar blue assay (MABA). The MICs of IMB-YH-8 for drug-resistant isolates ranged from 0.25 to 1 μg/ml (Table 4). These results indicate the potential utility of this compound to treat drug-resistant strains.  (Table 5). As shown in Table 5, with increasing concentrations (0.5, 1 and 2 μg/ml) of IMB-YH-8, the intracellular activity is significantly increased (P < 0.05), though the inhibitory effect of 0.5 μg/ml of IMB-YH-8 on CFU is not significant (p > 0.1). These results further demonstrate that IMB-YH-8 can inhibit both cell-free and intracellular mycobacteria.

Systematic name Gene Description Regulation
Fold induction in microarray (mean ± SD)

Discussion
The essential role of PknB in sustaining the growth of M. tuberculosis makes it a potential drug target for developing TB inhibitors 9 . We discovered a new compound IMB-YH-8 that inhibits the activities of PknB to auto-phosphorylate itself or to phosphorylate its substrate GarA. As a result, IMB-YH-8 potently inhibits both cell-free and intracellular M. tuberculosis in a dose-dependent manner. IMB-YH-8 does not inhibit the activity of human serine/threonine kinase Akt1 and M. tuberculosis STPKs including PknG, PknF and PknH but does inhibit PknA. On the basis of sequence homology of their kinase domains, nine M. tuberculosis STPKs are classified into three distinct groups, including membrane-bound PknA/PknB/PknL, PknF/PknI/PknJ and PknD/ PknE/PknH, whereas the two soluble enzymes PknG and PknK are separated from the others 5,20 . PknA and PknB are highly conserved in M. tuberculosis and share more than 85% amino acid identity, which suggests that they act by a conserved mechanism and modify similar substrates 6 . PknB exhibits less than 30% homology with eukaryotic kinases 5,12 . Although several Akt1 inhibitors were reported to diminish intracellular growth of MDR-TB 21 , they cause potentially serious off-target toxicities because Akt1 is involved in numerous different biological processes 22 . For example, Akt1 inhibitors increase insulin secretion and cause abnormal glucose metabolism 23 . Our study demonstrates that IMB-YH-8 is more specific for bacterial targets and may have limited off-target effects on human cells. Docking IMB-YH-8 on the catalytic domain of PknB revealed that IMB-YH-8 formed a hydrogen bond with Val95 of PknB and that the aromatic ring of IMB-YH-8 enhanced the hydrophobic interaction with PknB. Both types of interactions are expected to increase the binding affinity of IMB-YH-8 to PknB. It is noteworthy that Val95 also forms a hydrogen bond with one hydroxyl group of PknB inhibitor mitoxantrone and also hydrogen bonds the N1 atom of adenosine in the PknB-AMPPCP complex 24,25 . To further validate the essentiality, residues such as Val95 need to be mutated to show that their loss disrupts IMB-YH-8 binding and activity.
IMB-YH-8 causes damage in the cell wall of mycobacteria as shown by the results of electron microscopy. IMB-YH-8 treatment also leads to changes in morphology of M. tuberculosis including disruption of the cell wall. Similar observations were previously reported for M. tuberculosis that was treated with isoniazid or other compounds that target mycolic acids 26,27 . These effects of IMB-YH-8 on cell wall and cell morphology are in agreement with the function of its target PknB in controlling cell shape and regulating cell wall synthesis that requires mycolic acids 6,28,29 .
PknB phosphorylates SigH and RshA in vitro and in vivo 30 . SigH plays a central role in an extensive transcriptional network that regulates oxidative and heat-stress responses in M. tuberculosis 31 . As a stress sensor and redox switch, RshA provides a direct mechanism for sensing stress and activating transcription and regulates sigH negatively 30 . Under oxidative stress, SigH induces the expression of a thioredoxin-like gene rv2466c and the thioredoxin reductase/thioredoxin genes trxB2/trxC 31 . Our transcriptional work supports that IMB-YH-8 inhibits PknB phosphorylation and modulates SigH regulatory pathways. We did attempt to determine the effect of IMB-YH-8 on an M. tuberculosis pknB deletion mutant using pMIND vector. However, this deletion proved unsuccessful. It may be attributed to stringent requiremnet of PknB expression for M. tuberculosis survival. Depletion of PknB eventually leads M. tuberculosis to cell death 10  Both pknA and pknB are part of an operon that encodes genes for cell shape control and cell wall synthesis in M. tuberculosis 6 . Another possibility is that IMB-YH-8 inhibits PknB phosphorylation and then activates the enzyme cascade such as the SigH regulatory pathways in vivo. While Lougheed described that their PknB inhibitors may not target PknB in cells or PknB may not be bound by the inhibitors 32 . We also cannot exclude the possibility that IMB-YH-8 also targets other M. tuberculosis enzymes. Further studies are needed to confirm whether IMB-YH-8 targets PknB inside the mycobacterial cell. Nevertheless, we can conclude that IMB-YH-8 inhibits PknB and PknA with a moderate binding affinity in in vitro and displays anti-tuberculosis activity based on our data and discussion.
It is very likely that the methyl ester of IMB-YH-8 is hydrolyzed by mycobacterial and mammalian esterases. So we synthesized the carboxylate form of IMB-YH-8. However, it has no activity against M. tuberculosis. Recently, Zhai et al. found that serum C max values of IMB-YH-8 following oral delivery of 200 mg/kg only reach sub-microgram/ml concentrations and fall below the therapeutic range within less than 1 hour 34 . It does not seem likely that IMB-YH-8 itself can be considered an attractive drug candidate. Despite these points, it is worth describing IMB-YH-8 as a lead compound for drug discovery. 2.57 ± 0.14 3.14 ± 0.06 3.24 ± 0.05 Table 5. Activity of IMB-YH-8 against intracellular M. tuberculosis H37Rv. Values represent mean ± SD. **P = 0.008 compared to the 1 μg/ml treatment, *P = 0.011 compared to the 0.5 μg/ml treatment.
Protein expression and purification. The 279-residue kinase domain of PknB was expressed in E. coli BL21 (DE3) pLysS and purified as described 13 . The kinases PknA, PknG, PknF, PknH from M. tuberculosis H37Rv and human Akt1 were expressed with an N-terminal His6-tag in E.coli BL21 (DE3) pLysS. Recombinant GarA with His6-tag was purified by affinity chromatography and used as a kinase substrate. Briefly, the kinase expression vectors were first constructed. The transformed bacteria were grown in Luria-Bertani liquid medium containing 50 μg/ml kanamycin at 37 °C. Expression of recombinant protein was induced at OD 600 nm = 0.6 with 0.5 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) at 20 °C for overnight. Cells were lysed using a cell disrupter (one shot, Constant Systems, UK) and the recombinant protein was purified from the soluble extract. The protein was first captured by affinity chromatography on a Ni-NTA column, eluted with an imidazole gradient (0-500 mM). Protein preparations were dialyzed against 40% glycerol in 25 mM Tris-buffered saline (pH = 7.5) and analyzed for purity using SDS/PAGE. Protein concentrations were determined by A 280 nm , and preparations were stored at −80 °C for further use.
In vitro protein kinase assay. Docking studies. The mitoxantrone bound crystal structure (PDB-ID: 2FUM) was selected for docking studies 24 . The binding mode for IMB-YH-8 towards the binding site of PknB was generated through molecular docking using Molecular Operating Environment (MOE) version 2009.10 35 . In general, the docking was performed through "DOCK" module in MOE using the alpha triangle placement method. Refinement of the docked poses was carried out using the force field refinement scheme and scored using both the affinity dG and London dG scoring system. The pose with the highest docking score was returned for further analysis.
scanning electron microscopy (SEM) was performed. H37Rv was grown to early log phase (optical density at 600 nm, ~0.2). Then 1 μg/ml IMB-YH-8 was added to treat cells, followed by incubation at 37 °C for 5 days. Cells were harvested by low-speed centrifugation and washed in 0. Minimal Inhibitory Concentration (MIC) assay. The M. tuberculosis strains used in these studies included the laboratory strain H37Rv (ATCC 27294; American Type Culture Collection, Rockville, MD) and drug-resistant clinical isolates. All clinical isolates were obtained from the State Laboratory of Tuberculosis Reference of China. M. tuberculosis H37Rv or a clinical isolate was grown at 37 °C in Middlebrook 7H9 Broth (Difco Laboratories, USA) supplemented with 0.2% (vol/vol) glycerol, 0.05% (vol/vol) Tween 80, and 10% (vol/ vol) oleic acid albumin dextrose catalase (OADC) (BD and Company, USA). The activities of compound against H37Rv and clinical isolates were determined using the microplate alamar blue assay (MABA), as described previously 37,38 . The MIC was defined as the lowest concentration eliciting a reduction in fluorescence of ≥90% relative to the mean of replicate bacterium-only controls.
Cytotoxicity assay. Human monocytic leukemia THP-1 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum at 37 °C under 5% CO 2 . Cells were seeded into 96-well plates (4 × 10 4 cells/ well in 100 μl culture medium) in duplicate, and differentiated with 10 ng/ml Phorbol-12-myristate-13-acetate (PMA) for 24 h. The contents of the wells were replaced with fresh RPMI 1640 medium containing 10% FBS. Three-fold serial dilutions of the stock solutions resulted in final concentrations of 64 to 0.26 μg/ml with a volume of 100 μl. After incubation at 37 °C for 48 h, 10 μl of 5 mg/ml methyl-thiazolyldiphenyl-tetrazolium bromide (MTT) were added to each well. Then the plates were incubated for 4 h, after which the absorbance was read at 570 nm. Intracellular activity assay. THP-1 cells (4 × 10 5 cells/well in 1 ml culture medium) were differentiated with 10 ng/ml PMA, and grown overnight in RPMI 1640 medium containing 10% FBS in 24-well plates. M. tuberculosis H37Rv cultures were passed through an 8 μm-pore-size filter to remove clumps, and diluted to infect macrophages at a multiplicity of infection of 10 bacteria per cell. Infection was carried out for 4 hours, followed by three times washing with fresh media to remove the extracellular mycobacteria. The medium was replaced daily with different concentrations of compounds (2, 1, and 0.5 μg/ml). After 3 days of incubation, the medium was removed and the macrophages were lysed with 200 μl of 0.1% sodium dodecyl sulfate. Then the lysates were diluted with fresh media and plated onto 7H11 plates supplemented with 10% OADC to measure colony forming units (CFU). The results from the trend analysis were analyzed using Analysis of Variance (ANOVA) followed by the Fisher's Least Significant Difference (LSD) post hoc test, and analysis over time was performed using repeated-measures ANOVA followed by the Fisher's LSD post hoc test. Differences were considered statistically significant at p < 0.05. SPSS 23.0 (StatSoft Inc., Tulsa, OK) was used for the statistical analysis.