|Year : 2021 | Volume
| Issue : 3 | Page : 307-311
Computer-assisted screening of mycobacterial growth inhibitors: Exclusion of frequent hitters with the assistance of the multiple target screening method
Kohei Kuriki, Junichi Taira, Masato Kuroki, Hiroshi Sakamoto, Shunsuke Aoki
Department of Bioscience and Bioinformatics, Graduate School of Computer Science and Systems Engineering, Kyushu Institute of Technology, Iizuka, Japan
|Date of Submission||06-Jul-2021|
|Date of Acceptance||17-Aug-2021|
|Date of Web Publication||03-Sep-2021|
Department of Bioscience and Bioinformatics, Graduate School of Computer Science and Systems Engineering, Kyushu Institute of Technology, Iizuka 820-8502
Source of Support: None, Conflict of Interest: None
Background: The emergence of frequent hitters (FHs) remains a challenge in drug discovery. We have previously used in silico structure-based drug screening (SBDS) to identify antimycobacterial candidates. However, excluding FHs has not been integrated into the SBDS system. Methods: A dataset comprising 15,000 docking score (protein–compound affinity matrix) was constructed by multiple target screening (MTS): DOCK–GOLD two-step docking simulations with 154,118 compounds versus the 30 target proteins essential for mycobacterial survival. After extraction of 141 compounds from the protein–compound affinity matrix, compounds determined to be FHs or false positives were excluded. Antimycobacterial properties of the top nine compounds selected through SBDS were experimentally evaluated. Results: Nine compounds designated KS1–KS9 were selected for experimental evaluation. Among the selected compounds, KS3, identified as adenosylhomocysteinase inhibitor, showed a potent inhibitory effect on antimycobacterial growth (inhibitory concentration [IC]50 = 1.2 M). However, the compound also showed potent cytotoxicity. Conclusion: The MTS method is applicable in SBDS for the identification of enzyme-specific inhibitors.
Keywords: Adenosylhomocysteinase, DOCK program, GOLD program, multiple-target screening, Mycobacterium vanbaalenii, structure-based drug screening
|How to cite this article:|
Kuriki K, Taira J, Kuroki M, Sakamoto H, Aoki S. Computer-assisted screening of mycobacterial growth inhibitors: Exclusion of frequent hitters with the assistance of the multiple target screening method. Int J Mycobacteriol 2021;10:307-11
|How to cite this URL:|
Kuriki K, Taira J, Kuroki M, Sakamoto H, Aoki S. Computer-assisted screening of mycobacterial growth inhibitors: Exclusion of frequent hitters with the assistance of the multiple target screening method. Int J Mycobacteriol [serial online] 2021 [cited 2021 Dec 8];10:307-11. Available from: https://www.ijmyco.org/text.asp?2021/10/3/307/325499
| Introduction|| |
Recent developments in computational chemistry and three-dimensional protein structure determination have allowed for in silico screening of drug candidates on the basis of precise docking simulations, referred to as in silico structure-based drug screening (SBDS)., SBDS saves cost, time, and labor in drug candidate identification relative to traditional, high-throughput screening approaches that require biological assays. However, the prevalence of false-positive compounds, especially compounds called frequent hitters (FHs) that have been identified as hits for a range of target proteins, poses a significant challenge to SBDS. In general, FHs refer to false positives or otherwise nuisance compounds that have been identified as binders in many structurally unrelated proteins.
Previously, we demonstrated that SBDS can be used to identify candidates for antitubercular agents with novel mechanisms of action. This approach succeeded in identifying inhibitors for essential enzymes in mycobacterial survival: enoyl-acyl carrier protein reductase,,,,, mycobacterial cyclopropane mycolic acid synthase 1 (cmaA1), 7,8-diaminopelargonic acid synthase (bioA), and thymidylate kinase (tmk). Although our previous SBDS strategy successfully identified inhibitors targeting individual antimycobacterial enzymes, a system to exclude FHs has not been integrated with SBDS.
In the present study, we applied a multiple target screening (MTS) method,, developed by Fukunishi et al., to our SBDS. A schematic representation of the present SBDS strategy is shown in [Figure 1]. Antimycobacterial properties of the top nine compounds were validated experimentally.
|Figure 1: Schematic representation of in silico structure-based drug screening strategy in the present study. The asterisk denotes exclusion process of frequent hitters: among the protein–compound affinity matrix, the compounds with the potential to bind multiple target proteins were regarded as frequent hitters|
Click here to view
| Methods|| |
Construction of the protein–compound affinity matrix and exclusion of frequent hitters
The three-dimensional structural data of chemicals, comprising 154,118 compounds supplied by ChemBridge (San Diego, CA, USA), were obtained from Ressource Parisienne en Bioinformatique Structurale (Paris, France). Structural data of 30 potential target proteins were obtained from the Protein Data Bank Japan (PDBj, Osaka, Japan). Adding hydrogen atoms and assigning partial charges on the structures for docking simulations were carried out by the Dock Prep tool in UCSF's Chimera software (San Francisco, CA, USA).
First, the target proteins were subjected to primary screening by UCSF DOCK program, based on grid-docking. The top 500 compounds with high grid scores against each target protein were then screened by the GOLD program (Cambridge Crystallographic Data Centre, Cambridge, UK). Hereafter, the dataset comprising 15,000 docking scores is referred to as the protein–compound affinity matrix (or simply, affinity matrix). The 141 candidates with high GOLD score (≥85) were extracted from the affinity matrix. Among the affinity matrix, the compounds with the potential to bind multiple target proteins were regarded as FHs. The 48 compounds with the potential to bind multiple target proteins were excluded as FHs. The 37 compounds that archived as false positive (i.e. nonantimycobacterial activity confirmed compounds) in the PubChem database (Bethesda, MD, USA) were also excluded. Eleven multiple conformations of the remaining 56 compounds were generated using the LowModeMD in the molecular operating environment (Chemical Computing Group, Montreal, Canada) and were then rescreened with the GOLD program. The nine compounds, summarized in [Table 1], were eventually selected on the basis of averaged GOLD scores and were purchased from ChemBridge. The series of the compounds were designated as KS1–KS9. The chemical properties of the compounds are available on the manufacturer's website (http://www.chembridge.com).
|Table 1: The candidate compounds identified by in silico structure-based drug screening (KS1–KS9)|
Click here to view
Antimycobacterial activity measurement
Mycobacterium vanbaalenii (NRRL B-24157 strain, RIKEN BioResource Center, Tsukuba, Japan) was grown overnight at 28°C in 3.7% brain–heart infusion broth (Sigma, St. Louise, MO, USA). Cultures were then diluted 6-fold with broth containing the candidate compounds in 96-well, flat-bottom clear plates. Each well was inoculated with 0.200 mL of culture. Isoniazid (LKT laboratories, St. Paul, MN, USA) and 0.3% dimethyl sulfoxide (DMSO) were used as positive and negative controls, respectively. The plates were incubated at 28°C for 18 h, after which cell cultures were subjected to growth inhibition assays. Inhibition of bacterial growth was determined by measuring OD595 with a Bio-Rad Model 680 microplate reader (Bio-Rad, Hercules, CA, USA).
Cytotoxic effects were evaluated as previously described. Briefly, SH-SY5Y cells (1.5 × 104 cells/well) were seeded into 96-well plates and maintained in high-glucose Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/mL of penicillin, and 100 μg/mL of streptomycin and cultured at 37°C for 24 h in a humidified atmosphere containing 5% CO2. Following 48 h of serum starvation, KS3, 0.3% DMSO, and 30 μM isoniazid were added to cultures. The cultures were grown for an additional 12 h before cytotoxicity was analyzed using a Cell Counting Kit-8 according to the manufacturer's instructions (Dojin, Kumamoto, Japan).
| Results|| |
Multiple target screening method applied in silico drug screening
The 30 target proteins known as essential for Mycobacterial tuberculosis survival are listed in [Table S1]. In silico SBDS identified nine compounds (KS1–KS9) as possible antimycobacterial candidates. These compounds were predicted to bind to the following targets in M. tuberculosis: KS1, KS4, KS6, KS7, and KS9 to bioA (PDB ID 3LV2), KS2 to tmk (PDB ID 1MRN), KS3 and KS5 to adenosylhomocysteinase (ahcY, PDB ID 3CE6), and KS8 to cma1 (PDB ID 1KPH).
Antimycobacterial activity of the selected compounds
Inhibitory effects on mycobacterial growth by the nine candidate compounds were evaluated with nonpathogenic M. vanbaalenii (biosafety level 1). As shown in [Figure 2], KS3 exhibited potent growth inhibition, while KS1, KS4, and KS8 showed moderate inhibitory activity. Dose dependence of the inhibitory effect of KS3 was further examined [Figure 3], and the KS3 IC50 values (the concentration at which 50% of mycobacterial growth is inhibited) was 1.2 μM.
|Figure 2: Antimycobacterial activity of the selected compounds (KS1–KS9) on the growth of Mycobacterium vanbaalenii. Inhibition of bacterial growth was monitored at OD595 after 0 and 18 h of treatment with the candidate compounds (50 M). Isoniazid (INH, 50 M) and 0.3% DMSO were used as positive and negative controls, respectively. All experiments were performed in quadruplicate, and data with error bars are expressed as mean ± standard deviation significance was evaluated by Dunnett's test: nS: Not significant; *P < 0.05; **P < 0.01; ***P < 0.001|
Click here to view
|Figure 3: Dose dependence of KS3 on the growth of Mycobacterium vanbaalenii. The inhibitory effect on mycobacterial growth was monitored by optical density595 at 24 h after treatment with the compound. All experiments were performed in quadruplicate, and data with error bars are expressed as mean ± standard deviation. The 50% inhibitory concentration (IC50) value was determined based on a nonlinear regression method|
Click here to view
Cytotoxicity of KS3
The cytotoxicity of KS3 in a SH-SY5Y human neuroblastoma cell line was measured to evaluate any damaging effects of the compound on mammalian cells. As shown in [Figure 4], the growth of SH-SY5Y cells was completely inhibited by KS3, suggesting the compound has potent cytotoxicity in mammalian cells.
|Figure 4: Cytotoxicity of KS3 on SH-SY5Y human neuroblastoma cell lines. Here, 0.3% DMSO (light gray) and 30 M isoniazid (white) were used as negative and positive controls, respectively. All experiments were performed in quadruplicate, and data with error bars were expressed as mean ± standard deviation. Cell survival rates were compared using Dunnett's test for significance: n.s.: Not significant, ***P < 0.001|
Click here to view
| Discussion|| |
We have demonstrated the success of using in silico SBDS for the identification of antimycobacterial drugs in the past decade.,,,,,,, Here, we constructed a protein–compound affinity matrix based on DOCK–GOLD two-step docking simulations with 30 target proteins in M. tuberculosis versus 154,118 commercially available compounds. Compounds showing high affinity to multiple target proteins have been excluded as FHs. The broad-scale screening allowed for the identification of three compounds, KS1, KS3, and KS8, possessing antimycobacterial activity. Among them, KS3 showed antimycobacterial activity comparable to isoniazid, a first-line therapeutic drug for tuberculosis. KS3 was identified as a compound that can associate with the active site in M. tuberculosis ahcY. The chemical structure of KS3 and the predicted interaction of KS3 with the active site in M. tuberculosis ahcY are shown in [Figure 5]a. Growth inhibition of Escherichia More Details coli B and the K12 derivative (BL21 and JM109, respectively), in which ahcY is absent, was not observed in 50 μM KS3 treatment [Figure S1]. This observation supports the specific inhibitory action of KS3 on ahcY activity and demonstrates that the MTS method can be applicable for SBDS in the effective identification of enzyme-specific inhibitors.
|Figure 5: Interaction of KS3 with Mycobacterial tuberculosis ahcY. (a) upper panel, chemical structure of KS3; lower panel, predicted interaction of KS3 with the active site in Mycobacterial tuberculosis ahcY. (b) comparison of binding between KS3 with the Mycobacterial tuberculosis ahcY and human S-adenosyl-L-homocysteine hydrolase|
Click here to view
However, potent cytotoxicity of KS3 was observed on the growth of SH-SY5Y cells, suggesting KS3 inhibited not only M. tuberculosis ahcY but also human S-adenosyl-L-homocysteine hydrolase. The sequence homology of both proteins is highly conserved (61% homology), both proteins share structural similarity in substrate binding pockets [Figure 5]b. In addition, the results of KS3 binding simulations also showed similar binding mechanisms of KS3 to these targets. Indeed, although antiviral, antitumor, antiparasitic, antiarthritic, and immunosuppressive agents targeting ahcY have been designed, it has been difficult to distinguish between ahcY derived from different organisms due to the above-mentioned similarity., The present study provides an effective seed compound for an ahcY inhibitor. Acquiring interorganism ahcY selectivity remains a goal for future research.
| Conclusion|| |
The exclusion of FHs is important issue to obtain enzyme-specific inhibitors through in silico SBDS. In the present stage, experimental validation of enzyme specificity still remains, though the MTS method seems to be applicable in identification of an inhibitor targeting specific enzyme.
Financial support and sponsorship
This work was supported in part by a Grant-in-Aid for Scientific Research (C) (26460145) to SA and (18K05358) to HS from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Cheng T, Li Q, Zhou Z, Wang Y, Bryant SH. Structure-based virtual screening for drug discovery: A problem-centric review. AAPS J 2012;14:133-41.
Ghosh S, Nie A, An J, Huang Z. Structure-based virtual screening of chemical libraries for drug discovery. Curr Opin Chem Biol 2006;10:194-202.
Yang ZY, He JH, Lu AP, Hou TJ, Cao DS. Frequent hitters: Nuisance artifacts in high-throughput screening. Drug Discov Today 2020;25:657-67.
Izumizono Y, Arevalo S, Koseki Y, Kuroki M, Aoki S. Identification of novel potential antibiotics for tuberculosis by in silico
structure-based drug screening. Eur J Med Chem 2011;46:1849-56.
Kinjo T, Koseki Y, Kobayashi M, Yamada A, Morita K, Yamaguchi K, et al.
Identification of compounds with potential antibacterial activity against Mycobacterium through structure-based drug screening. J Chem Inf Model 2013;53:1200-12.
Kanetaka H, Koseki Y, Taira J, Umei T, Komatsu H, Sakamoto H, et al.
Discovery of InhA inhibitors with anti-mycobacterial activity through a matched molecular pair approach. Eur J Med Chem 2015;94:378-85.
Taira J, Umei T, Inoue K, Kitamura M, Berenger F, Sacchettini JC, et al.
Improvement of the novel inhibitor for Mycobacterium enoyl-acyl carrier protein reductase (InhA): A structure-activity relationship study of KES4 assisted by in silico
structure-based drug screening. J Antibiot (Tokyo) 2020;73:372-81.
Taira J, Nagano T, Kitamura M, Yamaguchi M, Sakamoto H, Aoki S. Structural modification of a novel inhibitor for mycobacterium enoyl-acyl carrier protein reductase assisted by In silico
structure-based drug screening. Int J Mycobacteriol 2020;9:12-7.
] [Full text]
Taira J, Morita K, Kawashima S, Umei T, Baba H, Maruoka T, et al.
Identification of a novel class of small compounds with anti-tuberculosis activity by in silico
structure-based drug screening. J Antibiot (Tokyo) 2017;70:1057-64.
Taira J, Ito T, Nakatani H, Umei T, Baba H, Kawashima S, et al. In silico
structure-based drug screening of novel antimycobacterial pharmacophores by DOCK-GOLD tandem screening. Int J Mycobacteriol 2017;6:142-8.
] [Full text]
Koseki Y, Kinjo T, Kobayashi M, Aoki S. Identification of novel antimycobacterial chemical agents through the in silico
multi-conformational structure-based drug screening of a large-scale chemical library. Eur J Med Chem 2013;60:333-9.
Fukunishi Y, Mikami Y, Kubota S, Nakamura H. Multiple target screening method for robust and accurate in silico
ligand screening. J Mol Graph Model 2006;25:61-70.
Omagari K, Mitomo D, Kubota S, Nakamura H, Fukunishi Y. A method to enhance the hit ratio by a combination of structure-based drug screening and ligand-based screening. Adv Appl Bioinform Chem 2008;1:19-28.
Ewing TJ, Makino S, Skillman AG, Kuntz ID. DOCK 4.0: search strategies for automated molecular docking of flexible molecule databases. J Comput Aided Mol Des 2001;15:411-28.
Verdonk ML, Cole JC, Hartshorn MJ, Murray CW, Taylor RD. Improved protein-ligand docking using GOLD. Proteins 2003;52:609-23.
Tsunosaki J, Mieno Y, Hamade Y, Aoki S. Prediction of antimicrobial action mechanism on chemical compounds of essential oils using chemoinformatics methods. Jpn J Aromather 2016;16:25-36.
Reddy MC, Kuppan G, Shetty ND, Owen JL, Ioerger TR, Sacchettini JC. Crystal structures of Mycobacterium tuberculosis
S-adenosyl-L-homocysteine hydrolase in ternary complex with substrate and inhibitors. Protein Sci 2008;17:2134-44.
Stepkowski T, Brzeziński K, Legocki AB, Jaskólski M, Béna G. Bayesian phylogenetic analysis reveals two-domain topology of S-adenosylhomocysteine hydrolase protein sequences. Mol Phylogenet Evol 2005;34:15-28.
Nakanishi M. S-adenosyl-L-homocysteine hydrolase as an attractive target for antimicrobial drugs. Yakugaku Zasshi 2007;127:977-82.
Kitade Y, Kojima H, Zulfiqur F, Kim HS, Wataya Y. Synthesis of 2-fluoronoraristeromycin and its inhibitory activity against Plasmodium falciparum
S-adenosyl-L-homocysteine hydrolase. Bioorg Med Chem Lett 2003;13:3963-5.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]