Computational Exploration of a Protein Receptor Binding Space with Student Proposed Peptide Ligands

Abstract

Computational molecular docking is a fast and effective in silico method for the analysis of binding between a protein receptor model and a ligand. The visualization and manipulation of protein to ligand binding in three-dimensional space represents a powerful tool in the biochemistry curriculum to enhance student learning. The DockoMatic tutorial described herein provides a framework by which instructors can guide students through a drug screening exercise. Using receptor models derived from readily available protein crystal structures, docking programs have the ability to predict ligand binding properties, such as preferential binding orientations and binding affinities. The use of computational studies can significantly enhance complimentary wet chemical experimentation by providing insight into the important molecular interactions within the system of interest, as well as guide the design of new candidate ligands based on observed binding motifs and energetics. In this laboratory tutorial, the graphical user interface, DockoMatic, facilitates docking job submissions to the docking engine, AutoDock 4.2. The purpose of this exercise is to successfully dock a 17-amino acid peptide, α-conotoxin TxIA, to the acetylcholine binding protein from Aplysia californica-AChBP to determine the most stable binding configuration. Each student will then propose two specific amino acid substitutions of α-conotoxin TxIA to enhance peptide binding affinity, create the mutant in DockoMatic, and perform docking calculations to compare their results with the class. Students will also compare intermolecular forces, binding energy, and geometric orientation of their prepared analog to their initial α-conotoxin TxIA docking results.

Background and Introduction

Computational molecular docking studies have become increasingly widespread for use in drug design, ligand screening across a single or multiple receptors, and analysis of ligand-receptor binding interactions [1, 2]. DockoMatic is a free and open source graphical user interface (GUI) that streamlines the use of an array of programs applicable to molecular docking studies and for the generation of protein homology models [3–6]. Built into the interface is the capability to construct ligand or receptor models using MODELLER [7] without needing to be familiar with python script. Timely integrated modeler (TIM), a GUI wizard, easily guides users in the step-by-step creation of both ligand and receptor models [6]. The DockoMatic GUI facilitates the submission of the desired receptor-ligand system to a molecular docking engine, either AutoDock 4.2 [8] or AutoDock Vina [9]. The relative ease of job submission using the DockoMatic GUI makes computational docking studies readily accessible to students and researchers without the need for prior computer programming experience or proficiencies in computational modeling. Additionally, the DockoMatic program and the supplemental software requirements can be efficiently run on a local installation using a standard desktop computer, or alternatively run on parallel computing clusters to better facilitate High throughput virtual screening (HTVS) applications.

In this laboratory exercise, students will utilize DockoMatic 2.0 [6] in a docking study of a small neurotoxic peptide, α-conotoxin (α-CTx) TxIA, to an acetylcholine binding protein (AChBP) isolated from the marine mollusk, Aplysia californica [10]. α-CTxs are neurotoxins produced by marine snails in the genus, Conus [11]. The high affinity of these peptides for binding to AChBPs is of medicinal interest due to the similarity of AChBPs to mammalian nicotinic acetylcholine receptors (nAChR). α-CTxs are antagonists of nAChRs; certain isoforms of which are have been shown to degrade in individuals with neurodegenerative disorders, such as Parkinson’s disease [12]. This tutorial uses α-CTx TxIA, with amino acid sequence GCCSRPPCILNNPDLC [10], and the AChBP isolated from the mollusk, A. californica, with the PDB ID: 2XNV [13], to demonstrate the process by which computational chemistry can be used to probe ligand to receptor binding paradigms.

Students are directed how to use DockoMatic to investigate the binding of α-CTx TxIA to the known active site of Ac-AChBP. While α-CTxs are known to be highly selective towards Ac-AChBP, side chain substitutions may result in a peptide with higher or lower binding affinity than that of native peptide when bound to the active site of Ac-AChBP [14, 15]. Based upon analysis of their docking results, students will propose two amino acid substitutions they hypothesize will bind with higher affinity to the active site of the receptor. DockoMatic will generate the structures of the proposed α-CTx TxIA mutants, and subsequently dock each peptide to the active site of Ac-AChBP.

Recent studies have demonstrated the effectiveness of in silico prediction of binding affinities between ligands and protein receptors by screening compound libraries [16–18], and peptide mutant structures that were generated by homology modeling [19]. Integration of such HTVS techniques provides a rapid and cost effective way to identify drug leads that can be validated by wet lab biochemical assays. The automated process of molecular docking using Dockomatic 2.0 facilitates HTVS; the approach described in this student laboratory experiment may be easily applied to a wide range of research questions focused on understanding ligand-receptor and protein–protein interactions. HTVS has become increasingly prevalent with advances in computational bandwidth, and the ease by which software has evolved to predict increasingly sophisticated binding interactions. This tutorial demonstrates only an example of the far-reaching capabilities and utility of computational molecular docking, with the broader aim of using these techniques to gain better understanding of biomolecular interactions and functions, and in the design and discovery of potential new drug candidates. Exposure to these methods of investigation at the undergraduate level constitutes a valuable tool for aspiring scientists to integrate modern approaches to testing hypotheses.

Methods

Uploading PDB and GPF Files

The instructor will have previously prepared protein database (PDB) files of the necessary structures available in a folder accessible to all students (see “Instructor Notes” section). To upload the ligand file, the students will click on the “Ligand” dialog box and open the alphaCTxTxIA.pdb file, or they may alternatively type the directory path to said file in the space next to the dialog box (see Supporting Information: student instructions). The same procedure is used in the PDB selection for the “Receptor” dialog box, using the directory to the receptor PDB file, TxIaReceptor.pdb. For the “Box Coordinates” dialog box, the grid parameter file (.gpf) located in the folder along with the PDB files will be used. This file defines the spatial coordinates restricting the region of the receptor available for ligand binding. To upload, the students will click the Box Coordinates dialog box and select the 2BR8.gpf file, or indicate the directory path as before.

Selection of Parameters for Docking

In an effort to save time, it is best to change the number of cycles that AutoDock 4.2 will perform from 100 (default) to five cycles. The output directory must also be indicated via the “Output Directory” dialog box. Each student should create a new folder for their output directory of generated files for ease of access later. After confirming that all of the files and parameters are correct, the students will then submit the docking job by clicking the “New Job” dialog box. A new job should appear under the “Output” tab located at the bottom of the DockoMatic screen. The students should then highlight and right click on this file, and click “Start” to initiate the docking job submission.

It should be noted that the decrease in the number of docking cycles recommended for this tutorial is for the purpose of reducing the time requirement for this educational exercise. In pursuing further molecular docking studies, larger numbers of cycles would be desirable. To increase the probability of achieving the correct ligand binding pose and energy rankings for a given ligand-receptor system, it is advantageous to sample all conformational space for ligand binding, thus requiring more docking cycles. Also, the number of cycles necessary to produce reliable results will vary between ligand-receptor systems based on factors such as system complexity, ligand structure, binding mode, and the types of primary stabilizing interactions (i.e. hydrogen bonding, polar, hydrophobic).

Analysis of α-CTx TxIA Bound to Ac-AChBP; Selection of Amino Acid Substitutions and Docking of α-CTx TxIA Analog

The binding interactions between α-CTx TxIA and Ac-AChBP may be analyzed using UCSF Chimera, a PDB file viewing program [20]. After ensuring that all AutoDock cycles have completed, students will open the PDB file of their 1st ranked structure, as determined from the histogram found in the AutoDock log file, and the receptor PDB file in Chimera to view the interactions between the ligand and receptor (Fig. 1). Students will label the amino acids on both the receptor and ligand that are suspected of interacting, and they should further indicate the distances between hydrogen bond partners. Based upon the interactions between the amino acid side chains of α-CTx TxIA (GCCSRPPCILNNPDLC) and the active site of Ac-AChBP, students will propose two amino acid substitutions that they hypothesize to result in a α-CTx TxIA analog with higher binding affinity for the Ac-AChBP receptor. The ligand analog will be created as described in the following section, and the docking study will be repeated using the α-CTx analog in place of α-CTx TxIA. After the successful docking of the analog and Ac-AChBP, students will analyze and compare the binding free energy results of both systems. Class results may be compiled to see which two amino acid substitutions resulted in the most energetically favorable ligand to receptor binding complex.

FIG 1.

Image of α-CTx TxIA (green) bound to Ac-AChBP (blue) generated with Chimera.

Creation of α-CTx TxIA Analogs

Creation of the α-CTx TxIA analogs can be done by modification of the ligand PDB file name. Students will again upload the same pdb files for ligand and receptor, and set the output directory and number of docking cycles to five as was done initially. However, this time in the text box next to the Ligand dialog box, where the file directory is located, the students will edit the PDB file name to indicate their choice of mutations. This is accomplished by putting a colon after the .pdb portion of the file, and with no spaces, indicate the amino acid and its number in the sequence, followed by the desired amino acid to be substituted, with all mutations of the same peptide separated by additional colons. An example of the data entry for the native peptide (aCTxTXIA.pdb), the glycine in position 1 to cysteine mutation (aCTxTXIA.pdb:G1C), and the glycine to cysteine and aspartic acid in position 14 to glutamic acid mutation (aCTxTxIA.pdb:G1C:D14E) is as follows:

• /home/username/DockomaticTutorial/aCTxTxIA.pdb

• /home/username/DockomaticTutorial/aCTxTxIA.pdb:G1C

• /home/username/DockomaticTutorial/aCTxTxIA.pdb:G1C: D14E

The mutations are accomplished using TreePack, a utility in DockoMatic that allows for combinatorial HTVS of peptide ligands [5]. The process of peptide analog creation between DockoMatic and TreePack is briefly described here. First, the target residue for point mutation and the two surrounding amino acids are placed in a new pdf file. This is followed by removal of the side chain atoms of the tripeptide from the pdb file, leaving only the backbone atoms and beta carbon atom. The amino acid at the point mutation is then replaced to create the peptide analog. The new analog tripeptide file is submitted to TreePack, which performs the amino acid side chain addition and subsequent energy minimization for atom geometry optimization. The modified side chains are extracted from the newly generated analog pdb file and placed into the original ligand pdb file, which is adjusted for residue and atom numbering. When submitting the point mutations as described above, DockoMatic will produce the new analog pdb files, and the results can be analyzed via Chimera for direct comparison of the original ligand and student proposed analogs.

Results and Discussion

This student laboratory experiment has been individually evaluated by ten undergraduate chemistry and/or biology students prior to being presented to an upper division undergraduate Advanced Chemistry Laboratory class at Boise State University. The classroom presentation consisted of a lecture on background material, video tutorials of DockoMatic and Chimera, and a brief discussion covering the utility and limitations of the software. The presentation with embedded video clips have been posted on Sourceforge along with the DockoMatic software, and is recommended for use as a prelaboratory activity by instructors wishing to implement this experiment in their laboratory curriculum.

Upon completion of the prelaboratory presentation, students were asked to provide feedback related to their opinion of the DockoMatic software, and the usefulness of the presentation to help them understand how to practically perform HTVS (Table I). Based on the evaluation results, the greatest strength of the presentation was the embedded video tutorials and the instructor’s level of preparation and helpfulness. While useful, almost every student thought the videos lacked a desired level of enthusiasm and charisma, which led to a lower than optimal audience engagement. In response to this, the video clips have been redone by an individual with heightened levels of enthusiasm, clarity, and pleasantry. Students identified the greatest weakness of the presentation as a lack of compelling evidence that the software could be meaningfully incorporated into their own independent research projects. To address this perceived weakness, it is recommended that the instructor emphasize specific instances of HTVS from the published literature to more prominently communicate its significance for application to address problems that the students are dealing with in the laboratory course or their own research projects. It may be effective to additionally challenge the students to imagine how in silico molecular docking could be integrated into their thesis research.

TABLE I.

Results from student evaluation of DockoMatic prelaboratory presentation

Survey questions	Participant response: poor (1) to excellent (5)					Number of responses	Average response
	1	2	3	4	5
Well informed of objectives		1		8	8	17	4.4
Expectations were met			3	8	5	16	4.1
Objectives were clear			2	8	7	17	4.3
Appropriate level of difficulty			5	5	7	17	4.1
Appropriate pace of presentation		2		8	7	17	4.2
Instructor was prepared			1	6	10	17	4.5
Instructor was helpful			1	6	9	16	4.5
Utility of what was learned	1	3	5	1	7	17	3.6
Presentation succeeded in teaching content	1	1	1	8	6	17	4.0
Overall Rating		1	2	5	9	17	4.3

Conclusions

This computational experimentation process provides instructors with a tool to broaden the curriculum offered in biology, chemistry, or pharmacology programs. The estimated duration of this tutorial is approximately one hour for the prelaboratory presentation available on Sourceforge, and two hours for the students to predict binding paradigms, perform molecular docking to test their hypotheses, and analyze the results of the docking experiments. The goal of this computational experiment is to add another tool for instructor use in guiding student inquiry in the laboratory. In a very short time, each student can perform their own molecular docking study and visualize whether their own prediction of ligand structure modification led to the desired result; all within a single laboratory session. In the process of performing this experiment, students will develop a firm understanding of the importance of peptide primary sequence in the determination of binding affinities of ligand-receptor complexes. This simplified laboratory experiment will introduce students to technological applications in biology, chemistry, and pharmacology, providing them with an engaging learning experience and exposing them to new and exciting research methodologies.

Instructor Notes

In order to properly set up the computers for this study, each computer must have the updated version of Dockomatic 2.0 [21], as well as Modeller, UCSF Chimera, AutoDock, and MGL Tools downloaded and installed. The PDB and GPF files necessary for this tutorial are available from SourceForge.net along with the prelaboratory presentation. The files required for this experiment must be copied to the same folder, and be accessible to all students participating in the tutorial. It is suggested that a special folder be created for use with this experiment with a name such as Dockomatic_Tutorial or Conotoxin_Tutorial to minimize confusion.