Docking-based virtual screening of known drugs against murE of Mycobacterium tuberculosis towards repurposing for TB

Repurposing has gained momentum globally and become an alternative avenue for drug discovery because of its better success rate, and reduced cost, time and issues related to safety than the conventional drug discovery process. Several drugs have already been successfully repurposed for other clinical conditions including drug resistant tuberculosis (DR-TB). Though TB can be cured completely with the use of currently available anti-tubercular drugs, emergence of drug resistant strains of Mycobacterium tuberculosis and the huge death toll globally, together necessitate urgently newer and effective drugs for TB. Therefore, we performed virtual screening of 1554 FDA approved drugs against murE, which is essential for peptidoglycan biosynthesis of M. tuberculosis. We used Glide and AutoDock Vina for virtual screening and applied rigid docking algorithm followed by induced fit docking algorithm in order to enhance the quality of the docking prediction and to prioritize drugs for repurposing. We found 17 drugs binding strongly with murE and three of them, namely, lymecycline, acarbose and desmopressin were consistently present within top 10 ranks by both Glide and AutoDock Vina in the induced fit docking algorithm, which strongly indicates that these three drugs are potential candidates for further studies towards repurposing for TB.


Background:
New drugs for the effective treatment of tuberculosis are greatly needed now than ever due to the increasing trend in the emergence of drug resistant strains of Mycobacterium tuberculosis (M. tuberculosis), especially in India, China and Russian Federation [1]. Poor success rate of the treatment of drug resistant tuberculosis (DR-TB) than drug sensitive TB alarmingly indicates for newer and effective drugs urgently for the successful treatment and control of TB, globally [1]. It has been reported that the current drug discovery pipeline for TB appears to be healthy since a good number of candidate drugs are being evaluated in different phases of clinical and preclinical research [2,3]. However, emergence of resistant mycobacteria even for newer drugs also cannot be ruled out in the future, thus, search for newer and alternative drugs for the treatment of TB should be a continuous process to ensure the sustained control of TB.
The conventional drug discovery process is challenged by several obstacles including expensive budget, lengthy process for discovery of new compounds, and most importantly meagre success rate [4]. Identifying new applications for already FDAapproved drugs for clinical conditions other than intended use is becoming a promising new avenue to pursue for drug discovery, a process defined as repurposing or repositioning [4]. Strategy of repurposing of already approved drugs appears to be a viable approach since it reduces the time spent in establishing pharmacokinetics and safety issues for human use [4]. This approach has been already adopted as a successful strategy in various clinical conditions including, premenstrual dysphoria (fluoxetine, primarily used for depression), multiple myeloma (thalidomide, primarily used for sedation, nausea and insomnia) [4]; the process of repurposing has recently been recognised and well supported by international agencies including National Institute of Health, USA and Medical Research Council, UK for research funding to explore possible repurposing of drugs for various diseases and to understand their mechanisms [5]. Fluoroquinolones, clofazimine, linezolid are few drugs that are successfully repurposed in the treatment of drug resistant tuberculosis [6].
The genome of M. tuberculosis encodes for about 4000 proteins [7]. Therefore, selection of a protein as a drug target is crucial for drug discovery for TB. Extensive research, including bioinformatics based studies has been carried out to identify and prioritize drug targets for TB [8,9]. Essentiality of protein for the pathogen and absence of homolog in eukaryotes are commonly used criteria for drug target selection, which we have used in the present study. The gene murE (Rv2158c) is an essential gene for the Mycobacterium tuberculosis since it encodes for a protein, UDP-N-acetylmuramoyl-L-alanyl-D-glutamate--2,6-diaminopimelate ligase that catalyzes a biochemical reaction essential for the biosynthesis of peptidoglycan of M. tuberculosis and it does not have a homolog in eukaryotes [10]. Therefore, murE has been suggested as a promising target for new drug discovery for tuberculosis and thus, its crystal structure has been determined in the presence of the substrate, UDP-MurNAc-l-Ala-d-Glu (UAG) [11]. Based on these evidences, we selected murE of M. tuberculosis as the drug target in the present study and performed virtual screening for 1554 FDA approved drugs using Glide and AutoDock Vina. We employed rigid docking algorithm and subsequently induced fit docking algorithm in order to enhance the quality of the docking prediction and to prioritize drugs for repurposing for TB. . LigPrep (Schrödinger, USA, 2014) was used to prepare the drug-library since it helps to prepare high quality, single, low-energy, 3D structure with correct chiralities, and apply force field for energy minimization for each entry of the chemical structure provided in the drug-library.

Virtual screening
All the 1554 drugs retrieved from DrugBank were screened against murE independently by using two docking methods namely Glide (Schrödinger) and AutoDock Vina (ADV) through PyRx (0.8) [15,16]. In the initial round of screening, rigid docking was employed; for the selected drugs which were present in top 10% in the rigid docking by both Glide and ADV, induced fit docking was performed for further prioritization. Using Glide, extra precision (Glide XP) method was used to generate grid around the active side residues with default settings and Glide score was used to rank the drugs after docking. AutoDock Vina was used as the second method through PyRx software in order to obtain consistent results, and binding energy of the docked complex was used to rank and shortlist drugs. The procedure applied in the present study has been illustrated in Figure 1. Molecular interactions were also analysed for all of the shortlisted drugs with murE. RMSD value of the docked complex and original crystal structure of the target were also calculated. Binding affinity, molecular interactions and RMSD value were collectively used to prioritize drugs as potential candidates for repurposing for tuberculosis.

Results and Discussion:
Virtual screening of 1554 FDA approved drugs against murE of M. tuberculosis was performed using two different molecular docking tools and by employing rigid docking followed by induced fit docking algorithms. At the end of first round of docking (rigid docking) 53 drugs were ranked within top 10% by both Glide and AutoDock Vina. Further, induced fit docking of the 53 shortlisted drugs with murE enabled us to prioritize 17 drugs as they were found within top 10 ranks either by Glide or ADV (Figure 1). A total of 3 drugs, namely lymecycline (rank 3 by Glide and rank 1 by ADV), acarbose (4 and 2) and desmopressin (7 and 3) were ranked consistently within top 10 ranks by both Glide and ADV. RMSD value for each of the docked complex relative to the crystal structure of the murE was calculated and it was observed that all of the 17 shortlisted drugs had value lesser than 1A o . More details about these 17 prioritized drugs, including binding affinity with murE, RMSD value, and molecular interactions with the target protein, and the primary use of each of the drugs are provided in the Table 1. Based on the crystal structure of the murE, amino acids, T195, S222, R230, E198, L67, S84, Q70, A69, T86, T85 were reported to involve in the molecular interactions with UAG [11, 12]. Molecular interactions between shortlisted drugs and murE were analysed and the interacting amino acid residues are provided in the Table 1 and those residues known to be present in active site of the murE are provided in boldface [11,12]. Lymecycline was found to be the top most drug by ADV ranking and rank 3 by Glide scoring. When the molecular interactions were analysed, it was found that lymecycline interact with seven amino acids of murE ( Table 1) including, S84, T86, E198 and R230 which are reported to be present in the active site of the murE [11, 12]. The RMSD value of the lymecycline and murE docked complex relative to the crystal structure of the murE was found to be 0.922A o . Acarbose was another drug ranked consistently within top 10 ranks by both ADV (2) and Glide (4) and the RMSD value was calculated to be 0.91A o . Acarbose was found to interact with 11 amino acids of murE ( , which led us to form a hypothesis that 'lymecycline would affect M. tuberculosis by blocking its peptidoglycan biosynthesis by targeting murE', which needs to be validated experimentally. Out of the 17 drugs prioritized by multiple rounds of docking, in addition to lymecycline, six more drugs, namely fidaxomicin, indinavir, caspofungin, vancomycin, micafungin and polymyxin B sulfate are also known to be used against bacterial, viral or fungal infections ( . Indinavir was shortlisted as a potential drug for repurposing for TB by another bioinformatics based study [22]. Since four of the 17 drugs are already reported to have anti-mycobacterial activity in vitro, we are encouraged to suggest these 17 prioritized drugs as potential candidates for further exploration towards repurposing for TB. Additionally, the same fact that four of the prioritized drugs have in vitro anti-mycobacterial activity strongly supports our procedure as a valid one and convinces us that it can be used as a generic protocol for prioritization of drugs for repurposing for other clinical conditions. Our protocol has identified 12 existing drugs as potential candidates for repurposing towards TB for the first time based on virtual screening against murE of M. tuberculosis. We started the present study with 1554 FDA approved drugs and subjected them through a series of stringent filtering criteria of multi-level docking which resulted in 17 prioritized drugs as potential candidates for repurposing for the treatment of TB.

Conclusion:
Our study has several intrinsic advantages. Since our starting materials for virtual screening were already FDA approved drugs for other clinical use (but not for TB) and human safety issues are well established, if these drugs are proved to show efficacy for TB in vivo in animal models in future, they can be expedited for clinical research directly in TB patients. Further, drugs have been screened against one of the essential proteins, murE, which is involved in the making of cell wall of the M. tuberculosis; thus, drug-inhibitors identified for this enzyme would potentially have fatal consequence on the bacteria. The bioinformatics protocol we applied in this study has helped us to prioritize 17 drugs including those which were already reported to have in vitro activity against M. tuberculosis; this observation serves as a validation and proof-of-concept for our protocol that we have conceptualised and applied in this study. Encouraged by these results, we have undertaken some of the prioritized drugs for further experimental studies towards repurposing for tuberculosis.