**2. Materials and methods**

#### **2.1 Protein homology modeling**

The protein sequence of trypanothione reductase of *Leishmania major* was retrieved from the NCBI database with the accession number XP\_001687512.1, having 491 amino acids [19]. Using the Basic Local Alignment Search Tool (BLAST), the query sequence was compared to known structures which generated structures similar to *Lm*TR. Modeller 9.20 [20] was used for modeling the structure of *Lm*TR.

#### **2.2 Active site prediction and quality assessment**

To predict the active site, the protein was submitted to CASTp 3.0 [21, 22]. The predicted active site was corroborated via a blind docking process using AutoDock Vina within PyRx version 0.9.7 [23, 24]. The active region was confirmed by the 'Toggle selection of Spheres' function which highlighted predicted residues from CASTp 3.0. Binding pocket was also viewed with PyMOL v2.0.0 [25]. The quality of the modeled protein was assessed by some quality measure tools. This included PROSA which determines the quality of experimentally solved structures and theoretical models in protein engineering by comparing these to that of experimentally solved protein structures in the PDB database [26]. Verify3D was used to validate the three-dimensional structure of the model [27]. PROCHECK, a quality assessment tool was also used to check the stereochemical properties of the model by generating a Ramachandran plot [28]. ProQ was also used to carry out further validation. ProQ predicted protein quality based on the LGscore and MaxSub scores [29].

#### **2.3 Energy minimization of protein target**

The modeled *Lm*TR was energy minimized using GROMACS 5.1.1 [30, 31]. Simulations were performed with the force field, GROMOS96 43a1. The system was solvated using an equilibrated SPCE216 water model. The charged protein had a net charge of 9 which was neutralized with an equal amount of Na<sup>+</sup> ions. Energy minimization of the protein was then carried out.

#### **2.4 Pharmacophore modeling and screening**

#### *2.4.1 Pharmacophore generation*

Pharmacophore model, virtual screening, and molecular docking studies were performed to find novel *Lm*TR inhibitors. The ligand-based structural design incorporates the absence of the macromolecular structure by generating pharmacophore models from a set of ligands. This method takes advantage of the conformational flexibility of the ligands [32]. The active compounds, CNQB and PNTPC were retrieved with IDs CID1323435 and CHEMBL242165 from PubChem (https://pubchem.ncbi.nlm.nih.gov)and CHEMBL (https://www.ebi.ac.uk/chembl) [33, 34], respectively and were used to generate the pharmacophore. These ligands served as a training set for pharmacophore generation. All customized settings were kept in default.
