TD001. Epidermal growth factor receptor inhibitors
The Epidermal Growth Factor Receptor (EGFR) is crucial for cell growth and division, and its malfunction is implicated in many cancers. Targeting EGFR with FDA-approved inhibitors offers a strategic approach to cancer therapy. This MolModa tutorial demonstrates how to dock compounds into the EGFR structure.
Tip
You can open the final .molmoda file and see the results of this tutorial.
Learning objectives
At the end of this tutorial, you should be able to
- Load a structure from the Protein Data Bank;
- Process and prepare a protein for docking;
- Identify binding pockets;
- Load in multiple compounds from a single SMILES file;
- Protonate several compounds at a pH of 7.4;
- Dock multiple ligands into a single protein and analyze the results.
Protein
Load PDB
This tutorial guides you through loading the PDB structure 3W32 into MolModa.
The 3W32 file provides a detailed representation of the EGFR kinase domain, a key target in cancer therapy.
To load the structure, select File PDB ID, then type in 3W32 and click Load.
MolModa will load in the PDB file as shown below.
Remove water molecules
A key step in preparing for a protein-ligand docking simulation involves the removal of water molecules from the protein structure. This preprocessing step is essential for several reasons.
Firstly, water molecules in the protein structure can introduce noise and computational complexity to the docking process. Water molecules can occupy potential binding sites or interact with the ligand in ways not representative of the dry environment within which the docking simulations typically aim to approximate. Secondly, excluding water molecules simplifies the computational model, reducing the number of variables that must be considered during the docking process. This simplification can lead to a more focused and efficient exploration of the potential binding modes of the ligand to the protein.
Furthermore, removing water molecules allows for specialized scoring functions and algorithms designed for 'dry' docking scenarios, which may not account for the dynamic and complex nature of water-protein and water-ligand interactions. However, while removing water molecules is a common practice, the decision to do so should be informed by the specific objectives of the simulation and the characteristics of the protein-ligand system under study. In some cases, the inclusion of key water molecules, especially those that play a critical role in the binding process or in maintaining protein structure, may be necessary for a more accurate representation of the biological system.
This tutorial will ignore all crystallographic water molecules from our structure. We can ignore the water molecules in two ways: hide or delete.
Here, we see the original structure with the red oxygen atoms from our water molecules.
Tip
X-ray crystallography cannot determine the positions of hydrogen atoms; thus, we only have lone oxygen atoms for our water molecules.
One option is to hide water molecules by clicking the eye next to Solvent.
In most calculations in MolModa, there is an option only to use visible or selected molecules.
Hiding water molecules is generally the recommended way to proceed since we never lose information.
The other option is to delete the water molecules by clicking the x next to Solvent.
This accomplishes the same results, except we cannot see the water molecules again unless we load in the PDB structure again.
The result of option 2a would look like this.
Protonation
As mentioned previously, no hydrogen atoms exist in the original PDB structure.
Since these are crucial to describe intermolecular interactions accurately, we have to add these in.
MolModa provides an automated way to protonate the protein by selecting Proteins Protonation.
Here is what our protein looks like with the "Sticks" style. We can see that there are no hydrogen atoms.
Here is where you can find the protein protonation menu.
This menu will pop up and give you options to select what you want to protonate.
After hitting Protonate, we can see that our structure is now protonated with hydrogen atoms.
Notice that MolModa creates a new group for our protonated protein instead of overwriting the original protein structure.
We will delete our unprotonated structure to clean up our Navigation.
Pockets
Selecting the appropriate binding pocket for docking molecules is critical in the protein-ligand docking process. The identification and selection of the binding pocket begin with a comprehensive analysis of the protein structure to pinpoint potential sites that can accommodate the ligand. This analysis often involves a combination of computational methods and biological insights.
Understanding the protein's function and the biological context within which it operates is critical to this selection process. Knowledge of the protein's active sites, obtained from literature reviews, experimental data, and databases of protein structures, provides invaluable guidance. For proteins with known ligands, these active sites serve as primary targets for docking. A more exploratory approach is required in cases where the binding site is unknown, or the protein may interact with multiple ligands. Here, algorithms that predict potential binding sites based on geometric, energetic, and chemical criteria are beneficial.
Selecting the right pocket for docking also involves evaluating the compatibility of the pocket's characteristics with the ligand's physicochemical properties. The pocket's shape, size, and charge distribution should complement the ligands to facilitate a stable and biologically relevant interaction. Additionally, the pocket's accessibility, flexibility, and the presence of critical amino acids that can form key interactions with the ligand are considered to ensure that the selected site can realistically accommodate the ligand under physiological conditions.
Once a potential binding pocket is identified, it may be refined and optimized through further computational analysis to enhance the accuracy of the docking simulation. This refinement can include adjusting the pocket's dimensions, optimizing the alignment of pocket residues, and incorporating flexibility into the pocket structure to allow for a more dynamic interaction with the ligand.
For this tutorial, we will use the pocket containing the co-crystallized ligand.
Select Proteins Detect Pockets to get to this option menu.
After hitting Detect, MolModa will detect potential binding pockets and show them as rectangular positions in space.
MolModa identified 21 potential pockets, but we want to select the one that best captures our ligand.
If you do not have a ligand you are trying to capture with your pocket, you can see pocket properties under the Data tab of the main window.
The higher the score and druggability, the more likely the pocket will be an effective ligand binding site.
As shown below, we deleted all pockets that did not contain the ligand.
Upon closer inspection, the original pocket does not entirely capture the ligand. This can cause issues once we dock, so we will manually change the pocket's origin and dimensions.
We will delete the original compounds now that we have our defined region.
Note
You may have noticed that sulfate ions were excluded in our region. This is irrelevant to our docking procedure, so we also removed it.
Compounds
We first must load the compounds of interest to conduct the docking simulations. The compounds we'll be working with can be imported from a SMILES file, which encodes each compound's structure in text format. This file can be downloaded here. The steps below guide you through loading these compounds into MolModa for docking.
To begin, navigate to File Open and select the SMILES file you've downloaded (td001.smiles).
This action prompts MolModa to load the molecular structures contained within the file.
Upon loading, the molecules from the SMILES file will appear in the Navigation pane. Each molecule is listed separately, allowing for individual selection and manipulation.
Here's an example of how a single molecule from the SMILES file is represented in MolModa.
Protonate
Before docking, it's crucial to protonate the compounds at a physiological pH (7.4) to ensure their ionization states accurately reflect biological conditions. MolModa facilitates this process through an automated protonation tool.
Note
You can skip this step if you load compounds with the desired protonation state.
Initially, the loaded molecules may lack hydrogen atoms for the specified pH. We will protonate the molecules regardless to ensure consistency in our protocol.
Access the protonation options by selecting Compounds Protonation.
Ensure that the pH is set to 7.4 to reflect physiological conditions.
After protonation, the compounds are updated to include hydrogen atoms, better representing their reactive form under physiological pH.
Docking
With the protein pocket selected and compounds prepared, we move on to the docking stage. Docking is a crucial step in the drug discovery process, as it allows us to predict how the prepared compounds interact with the target protein and assess their potential as inhibitors.
The docking process involves several key aspects:
- Docking algorithms generate multiple possible orientations and conformations (poses) of each compound within the selected protein pocket. These poses represent potential binding modes of the compound to the target site. Each generated pose is evaluated and assigned a score based on the predicted strength and favorability of the compound-protein interaction. Scoring functions consider shape complementarity, hydrogen bonding, hydrophobic interactions, and electrostatic forces. The poses are then ranked according to their scores, with lower scores indicating more favorable binding.
- The top-ranked poses for each compound are visually inspected to assess their plausibility and interactions with key residues in the binding pocket. This analysis helps to identify the most promising compounds and provides insights into the molecular basis of their potential inhibitory activity.
The docking results provide valuable information for guiding the selection and prioritization of compounds for further experimental testing. Compounds with high-scoring poses that exhibit favorable interactions with the target protein are considered promising candidates for potential inhibitors.
Begin docking by navigating to Docking Compound.
This opens the docking menu, where you can select your target pocket and the compounds to dock.
In the docking options, you can specify parameters such as the pocket, exhaustiveness of the search, and a limit on rotatable bonds. A higher exhaustiveness value increases the thoroughness of the search but also requires more computational time.
Once you start the docking process, MolModa will display the progress on the Jobs tab.
This process can take some time, depending on the number of compounds and the exhaustiveness setting.
However, it's important to note that docking is a computational prediction with limitations. The accuracy of docking results depends on factors such as the protein structure's quality, the binding site's flexibility, and the scoring function's ability to capture the relevant interactions. Therefore, docking results should be interpreted cautiously and validated through experimental assays to confirm the predicted binding and inhibitory activity.
Despite its limitations, docking is a powerful tool in the early stages of drug discovery. It allows researchers to virtually screen large libraries of compounds and identify those with the highest likelihood of binding to the target protein. This virtual screening approach saves time and resources by prioritizing compounds for experimental testing and reducing the number of compounds that need to be synthesized and assayed.
Exhaustiveness
8
Setting the exhaustiveness to 8 provides a balance between speed and thoroughness.
The default setting of 8 is generally sufficient for most screenings.
Tip
Docking all 11 compounds took around 2 minutes on an i9-12900K CPU.
After docking, MolModa displays each compound's different poses within the binding pocket. These poses are ranked based on their docking scores.
The docking scores are indicative of the binding affinity. Lower scores typically suggest a stronger binding interaction.
The pose with the lowest docking score is considered the top hit, implying it is the most favorable binding orientation.
32
Increasing the exhaustiveness to 32 provides a more comprehensive search, potentially uncovering better-fitting poses.
Tip
Docking all 11 compounds with this exhaustiveness took around 10 minutes on an i9-12900K CPU.
With a more exhaustive search, you might observe changes in the docking scores, indicating different binding affinities.
We see that the order changes slightly, but within 1.0 kcal/mol of the lowest score, we have the same molecules 2, 5, 7, and 10 for both exhaustiveness.
The top hit at this exhaustiveness level may differ from the one at a lower level, reflecting a more accurate prediction of the ligand binding.
Pose refinement
After identifying promising poses through the initial docking process, further refinement can be performed to improve the accuracy of these predicted binding modes. Pose refinement is a crucial step that aims to adjust the ligand's conformation and orientation within the binding pocket to achieve a more precise and energetically favorable fit. During the pose refinement process, the ligand's position is fine-tuned while considering the protein environment's influence. This involves exploring subtle rotations and translations of the ligand within the pocket and adjusting the ligand's internal torsion angles. The goal is to optimize the interactions between the ligand and the surrounding amino acid residues, such as hydrogen bonding, hydrophobic contacts, and electrostatic interactions.
Pose refinement algorithms often employ more sophisticated scoring functions for various factors, including the protein's flexibility, solvation effects, and entropic contributions. These scoring functions evaluate the quality of the refined poses and guide the optimization process toward more energetically favorable and physically realistic binding modes. The refinement process can be computationally intensive, especially when considering multiple ligand conformations and protein flexibility.
The exhaustiveness parameter in MolModa allows users to control the extent of the refinement search.
Increasing the exhaustiveness value dedicates more computational resources to exploring a larger conformational space and potentially discovering better-refined poses.
However, this comes at the cost of increased computational time.
In this example, we increased the exhaustiveness to 400 to demonstrate the computational tradeoffs associated with more extensive pose refinement.
This higher exhaustiveness value enables a more thorough exploration of the ligand's conformational space within the binding pocket, leading to potentially better-refined poses.
Tip
Docking this compound with this exhaustiveness took around 8 minutes on an i9-12900K CPU.
Refinement operations can adjust the ligand's orientation and conformation, seeking to optimize interactions with the pocket.
The refined docking scores provide a more accurate binding affinity estimation after the ligand's pose is adjusted.
According to the refined scores, the top poses exhibit the best fit and strongest interactions with the binding pocket after refinement. We see only some slight rotations between the two exhaustiveness parameters.
