Approximating Protein Flexibility without Molecular Dynamics 

Tutorial Created with Software Release: 2025-3
Topics: Hit Discovery, Hit-to-Lead & Lead Optimization, Small Molecule Drug Discovery, Structure Prediction & Target Enablement
Products Used: Glide, Prime

Tutorial files

26 MB
This tutorial is written for use with a 3-button mouse with a scroll wheel.
Words found in the Glossary of Terms are shown like this: Workspacethe 3D display area in the center of the main window, where molecular structures are displayed

 

Tip: You can hover over a glossary term to display its definition. You can click on an image to expand it in the page.
Abstract:

 

In this tutorial, we demonstrate two techniques for modeling the flexibility of a protein in response to a binding event without the use of molecular dynamics simulations. First, we will soften the potentials in Glide to reduce the rigidity of the receptor in the context of a virtual screen. Then, we will use the Induced-Fit Docking (IFD) protocol to facilitate docking of ligands for which the adjustment to Glide potentials is insufficient.

This tutorial does not cover the full IFD-MD protocol which is able to more reliably predict induced-fit binding modes. For an introduction to IFD-MD, see the Cross-docking with IFD-MD tutorial.

 

Tutorial Content

  1. Introduction

  1. Creating Projects and Importing Structures

  1. Softening the Potential in Glide

  1. Induced Fit Docking for Side Chain Conformational Changes

  1. Conclusion and References

  1. Glossary of Terms

1. Introduction

Proteins are much more flexible than the static structures we obtain from experiment tempt us to believe. A binding pocket or interface may be very rigid, with very similar apo- and bound conformations, or quite permissive, allowing ligands of various sizes and shapes to bind. You can determine where your system of interest falls on this scale from knowledge of how it behaves in vivo, or by exploring the flexibility of the system computationally (see the Understanding and Visualizing Target Flexibility tutorial).

Accounting for this flexibility in modeling is crucial, as many computational methods operate on the assumption of a rigid receptor for the sake of computational efficiency. As a prominent example, Glide docking is used in both virtual screening and lead optimization to predict bound poses of (putative) ligands of various sizes. By default, the repulsive potentials for the receptor atoms in Glide are scaled down by a factor of 0.8 to simulate some “give”. Further softening of the potentials can also be applied to specific parts of the receptor (e.g. a specific residue that can move out of the way) to allow more permissive docking. This approach can be especially useful in a virtual screen when only one receptor structure is available (but the pocket is also known to bind larger ligands).

While selectively softening the receptor potential can allow ligands to dock without worrying about clashes with the problematic parts of the receptor, it is not able to predict the structure of the receptor-ligand complex as the receptor geometry is not adjusted. Additionally, this selective scaling is not useful when the entire pocket can grow to accommodate far larger compounds than the cocrystal ligand. In these cases, the induced-fit effect must be explicitly modeled.

In the Schrödinger suite, two methods are available for predicting induced-fit structures. The IFD protocol uses Glide to sample ligand binding modes and Prime to sample protein flexibility. The IFD-MD protocol additionally uses explicit solvent molecular dynamics simulations, water thermodynamics, and empirical scoring functions to explore and reliably predict induced-fit binding modes. Often, the amount of resources (time, CPUs, information) available to the project dictate the method(s) that will be employed.

In this tutorial, you will first use Glide with softened potentials to dock a set of differently-sized ligands into the binding pocket of human Cyclin-dependent Kinase 2 (Cdk2). Then, you will use IFD to additionally predict induced-fit poses for one of the ligands from that data set.

For tutorials showcasing the use of IFD-MD in a variety of applications, see the Further Learning section in the Conclusion of this tutorial.

2. Creating Projects and Importing Structures

At the start of the session, change the file path to your chosen Working Directorythe location where files are saved in Maestro to make file navigation easier. Each session in Maestro begins with a default Scratch Projecta temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project, which is not saved. A Maestro project stores all your data and has a .prj extension. A project may contain numerous entries corresponding to imported structures, as well as the output of modeling-related tasks. Once a project is saved, the project is automatically saved each time a change is made.

Structures can be built in Maestro or can be imported using File > Import Structures (or drag-and-dropped), and are added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data. The Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion is located to the left of the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed. The Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data can be accessed by Ctrl+T (Cmd+T) or Window > Project Table if you would like to see an expanded view of your project data.

  1. Double-click the Maestro icon.

Figure 2-1. Change Working Directory option.

  1. Go to File > Change Working Directory.
  2. Find your directory, and click Choose.
  3. Pre-generated input and results files are included for running jobs or examining output. Download the zip file here: https://www.schrodinger.com/sites/default/files/s3/release/current/Tutorials/zip/ifd_tutorial.zip
  1. After downloading the zip file, unzip the contents in your Working Directorythe location where files are saved for ease of access throughout the tutorial.

Figure 2-2. Saving the project in the Save Project panel.

  1. Go to File > Open Project.
  2. Choose IFD_tutorial_project.prjzip in your Working Directorythe location where files are saved and click Open.
  3. In the Save scratch project warning box, click OK.
  4. Go to File > Save Project As.
  5. Change the File name to IFD_tutorial, click Save.
  • The project is now named IFD_tutorial.prj.

3. Softening the Potential in Glide

Glide allows residue potentials to be softened based on specific suggestions or property ranges, such as B-factor. For a general introduction to setting up docking calculations with Glide, see the Structure-Based Virtual Screening with Glide tutorial.

In this section, we will use Glide to predict poses for a set of six diverse ligands for human Cdk2. The ligands we provide represent a small selection of Cdk2 inhibitors for which crystallographic data is available in the PDB and have already been prepared. We will identify three flexible residues in the binding site and generate a receptor grid where the potentials corresponding to these residues are scaled down. Then, we will compare the results of docking the ligands with and without the scaling.

3.1 Generate a receptor grid

Figure 3-1. The prepared structure of Cdk2 (PDB ID 1PXJ), colored by B-factors.

  1. Confirm that 1PXJ_prepared is includedthe entry is represented in the Workspace, the circle in the In column is blue in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • The protein atoms are colored by the B-factor information from the PDB file.  

Note: This receptor structure has already been prepared.

The atoms in the protein structure are colored by the Atom PDB B-factor (Temperature) color scheme. The higher the B-factor, the more uncertain this atom’s position is in the crystal structure, and the redder the coloring. While high B-factors can also result from a bad fit of a structure, here we can use them as a guide to estimate the flexibility of the associated residues.

 

There are three flexible residues with high B-factors pointing into the binding pocket (ILE 10, LYS 33, and LYS 89). As the ligands we’d like to dock are a bit larger than the 1PJX cocrystal ligand these residues would likely be in the way, so we will soften their potentials a bit.

Figure 3-2. The Receptor Grid Generation panel.

  1. Go to Tasks > Browse > Glide > Receptor Grid Generation.
    • The Receptor Grid Generation panel opens in the Receptor tab.
  2. Under Define receptor, confirm that the boxes for Pick to Identify the ligand (Molecule) and Show Markers are checked.
    • A banner in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed prompts you to click on an atom in the ligand.

Figure 3-3. Excluding the ligand from grid generation.

  1. Click on any ligand atom in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • This creates axes and grid boxes in the Workspace that originate from the location of the ligand.
  2. Click the Fit-all button to see the box in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 3-4. The Receptor Grid Generation Advanced settings.

  1. In the Receptor tab, click the Advanced Settings button.
    • The Receptor - Advanced Settings dialog box opens.

Figure 3-5. Scaling specific residues in the Advanced Settings.

  1. Under Per-atom scale factors, choose Specify for selected atoms.
    • The controls below it are enabled.
  2. Change the VdW radius scale factor to 0.80 (80% scaling).
  3. Check the box for Pick and choose Residues from the menu.
    • A banner appears prompting you to pick an atom to adjust its radius and charge scaling. 
  4. In the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed, click any atoms in the ILE 10, LYS 33, and LYS 89 residues.
    • The table in the Advanced Settings panel updates to include the selected residues.

 

Note: You can search for the residues using the search toggle in the Hierarchy.

 

  1. Click OK

 

Note: There are additional residues with high B-factors near the binding site (GLY 11 through TYR 15), but scaling this entire region would be a very drastic change.

You can now finish the receptor grid generation as usual. For the example in this tutorial, all additional settings are left at the defaults. The pre-generated files are included in the zip archive.

You have now completed the setup for a softened potential grid and if you want you can use the Ligand Docking panel to dock the prepared ligands from the ligprep_cdk2_ligs-out group in the project. In the next section, you will view the pre-generated results where the softened potential grid was used to dock these ligands.

3.2 Analyzing the docking results

For simplicity, the pre-generated results (Glide_results.maegz) can be used to view results of this docking run using the grid with scaled potentials and the results of a Glide docking of the same ligands without softening the potentials. Please note that your results may vary slightly if you run your job with a different version of the software.

Figure 3-6. The Workspace after importing the docking results.

  1. Go to File > Import Structures.
  2. Select the file Glide_results.maegz and click Open.
    • A banner appears and a group containing the following sub-groups is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion:
      1. CDK structures for comparison – prepared 1PXJ structure (faded orange carbons) as well as crystallographic references for all docked ligands (pink carbons).
      2. glide-dock_Cdk2_soft_pv – Results from docking with softened potentials for the ILE 10, LYS 33 and LYS 89 residues (green ligand carbons)
      3. glide-dock_Cdk2_default_pv – Results from docking with default scaling (blue ligand carbons)
    • If ribbons are not shown, click the Ribbons button in the Workspace Configuration Toolbar (bottom right).

 

Note: Only the 1PXJ structure has been prepared.

Figure 3-7. 1PJX receptor structure (orange) overlaid with 6GUK crystal structure (pink) and the docked poses for the 6GUK ligand with softened potentials (green) and Glide default settings (blue).

  1. Fixdouble-click the circle in the In column (or right-click the circle and click “Fix in Workspace”) the 1PXJ_prepared entry from the glide-dock_Cdk2_soft_pv group in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  2. Includethe entry is represented in the Workspace, the circle in the In column is blue the entries for each ligand from each of the Glide results groups as well as the crystallographic reference to analyze the results.  

Note: Multiple poses are returned for some ligands, as ligand preparation resulted in several relevant states. The docking score should not be used to compare these poses quantitatively, for this application, consider them equally valid.

Feel free to fully explore the results for all ligands. A few things should be evident from the comparison of the docked structures with and without scaling for this small ligand data set:

  • For the 2R3I and 1GZ8 ligands, both docked poses are reasonably close to the reference, as the ligand doesn’t come too close to the scaled residues.
  • For the 6GUK and 6Q4H ligands, scaling back the potentials of the three residues allows us to recover the correct bound pose (in the case of 6Q4H it seems that softening the potential for LYS 89 actually pulls the acid group out of the pocket).
  • For the 2FVD and 1KE8 ligands, both docked poses are not at all similar to the reference, as the pocket needs to rearrange substantially to properly accommodate them.

 

It should be clear that softening the potential will create more room by somewhat scaling back the VdW radii. This alleviates some problems of working with a rigid receptor grid, as seen in this case, but scaling back does not physically move residue coordinates. In the case of the longer 2FVD and 1KE8 ligands, scaling alone cannot facilitate docking, and we would need to turn to a full induced fit docking protocol.

4. Induced Fit Docking for Side Chain Conformational Changes

To account for movements in the protein structure, side chains and beyond, we can employ the Induced Fit Docking (IFD) protocol. IFD generates binding poses for targets where there is suspected conformational flexibility, in both the ligand and receptor, which is crucial for accurate docking of the ligand. This protocol uses Glide to sample ligand binding modes and Prime to sample protein flexibility and is a useful tool to predict active site geometries with minimal expense.

The following example uses the same 1PXJ structure of human Cdk2 and uses IFD to obtain a pose for the ligand from the 1KE8 structure.

4.1 Importing Starting Structures

Figure 4-1. The Crystal structure of 1PXJ (Human Cdk2 complexed with CK2 ligand).

  1. Go to Workspace > Clear Workspace.
  2. Go to File > Import Structures.
  3. Choose Cdk2_IFD.maegz.
  4. Click Open.
    • The IFD group is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and contains two pre-generated groups for the inputs and results of the IFD calculations.

Using our starting structures, we can overlay both binding sites in order to understand how the cavity and ligand would need to move as part of the Induced Fit strategy. This gives us a good feeling for the starting conformation of the protein and the final conformation of the protein-ligand complex. Of course, this is only possible because we have a crystallographic reference for the ligand for which we want to predict a pose.

Figure 4-2. Two superimposed crystal structures of 1PXJ (cyan) and 1KE8 (brown) and their native ligands.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue both Input_Receptor_Cdk2–1PXJ and Target_Receptor_Cdk2–1KE8 in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • These have been pre-aligned using Structure Alignment tools so that we see them in a common reference frame.
  2. Click Fit view to Ligand or press L on the keyboard to zoom to the ligand.

Figure 4-3. Selecting the binding site residues.

  1. Under Quick Select, click the Choose item button (three horizontal dots) and choose Binding Sites.

 

Figure 4-3. 1PXJ (cyan) and 1KE8 (brown) shown superimposed. 1PXJ Lys 33 and Lys 89 are clearly obstructing the binding site in order for the 1KE8 ligand to bind successfully (shown in the red circled regions).

  1. Go to Style Toolbox and click Display selected atoms (open-eye icon).
  2. Press the L key to zoom to the ligands.

 

Note: Two residues in particular, LYS 33 and LYS 89 in 1PXJ, are in close proximity to the 1KE8 ligand, and hinder its ability to dock well even with the potentials softened as seen in the previous section. The same residues are in a different conformation in the crystal structure of the target receptor 1KE8, and are located away from the ligand, as shown opposite. Additionally, the binding site reorganizes beyond these two residues to accommodate the larger ligand.

It is worth noting that the initial part of the IFD protocol uses a side-chain trimming procedure (as well as VDW scaling) in order to superficially create more room in the active site so that the non-native ligand can be sampled thoroughly. In this case, LYS 33 and LYS 89 will be candidates for such side chain trimming, and we will explore this in the next section. The ILE 10 residue whose potential we had also softened in the previous section does not clash with this ligand, so we will not include it here.

4.2 Setting up the IFD Job

This section will illustrate the use of IFD to achieve docking where Glide was not able to produce a pose.

Figure 4-4. The Induced Fit Docking in Tasks.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue only Input_Receptor_Cdk2–1PXJ in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  2. Select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries only Input_Ligand_from_Cdk2–1KE8.
  3. Go to Tasks > Browse > Receptor-Based Virtual Screening > Induced Fit Docking.
    • Choose IFD Application dialog box appears.
  4. Click IFD in the dialog box.
    • The Induced Fit Docking panel opens.

Figure 4-5. The Receptor tab of the Induced Fit Docking panel.

  1. For Ligands to be docked, choose Project Table (1 selected entry).
  2. In the Receptor tab, in the Box center section, click Centroid of Workspace ligand.
    • Centroid of Workspace ligand is used to define the grid box.
  3. Check the Pick option.
    • A banner appears prompting you to pick any atom of the ligand in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 4-6. The native ligand of 1PXJ is used to define the docking region, denoted by the green and purple grid boxes.

  1. Click on any ligand atom.
    • The ligand atoms are highlighted and grid boxes are shown in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  2. Click the Fit-all button to see the grid boxes.

 

Note: The native ligand of 1PXJ is being used to define the docking region.

Figure 4-7. The Glide Docking tab of the Induced Fit Docking panel.

  1. Go to the Glide Docking tab.
  2. Check Trim Side chains and choose Manual.
  3. Click Specify Residues.
    • The Select Trim Sidechains Residues dialog box opens.
    • A banner appears prompting you to pick residue to include in the active site definition.

Figure 4-8. Selecting residues for side chain trimming.

  1. In the dialog box, click Select.
    • The Atom Selection dialog box opens.
  2. In the Atom Selection dialog box, for Residue number type 33, 89.
  3. Select Add.
  4. Click OK.
    • The selected residues populate the table.
  5. Click OK to close the Select Trim Sidechains Residues dialog box.

Figure 4-9. Running the Induced Fit Docking job.

  1. Change the Job name to InducedFit_Cdk2 and click Run.
    • If prompted to prepare the structure with the protein preparation workflow, please choose Continue as this structure has already been prepared for you.
    • This job takes ~ 1 hour to complete on 4 CPUs.
    • You can find the pre-generated results for this job in the Cdk2_IFDOutputs group.

4.3 Analyzing the IFD Results

IFD scores poses by combining the GlideScore and the Prime energy. For more details on how IFD finds and scores poses and customization options, see the Induced Fit Docking Panel documentation. Beyond the IFDScore, there is no built-in method to validate poses. A critical eye on the ligand pose, interactions with the receptor, and running molecular dynamics to assess the stability of the pose are helpful to determine promising poses.

Figure 4-10. Setting up the Workspace for visualizing the results.

 

  1. Expand the Cdk2_IFDOutputs group.
    • A total of 18 poses are captured from running IFD.
    • These are shown in order of increasing IFD scores.
  2. Fixdouble-click the circle in the In column (or right-click the circle and click “Fix in Workspace”) the Target_Receptor_Cdk2-1KE8 in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  3. Click the Change table settings (three vertical dots) in the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and choose Show Property.
    • The Show properties in table dialog box opens.

Figure 4-11. Adding IFDScore property for analyzing the results.

  1. In the dialog box, click Choose.
  2. Search and select the IFDScore from the list.
  3. Click OK.
    • The IFDScore property is added to the table.

Figure 4-12. An IFD-generated pose for the input ligand (maroon) with Cdk2 (cyan). The corresponding crystal structure (brown) of this complex is also shown for reference.

  1. Select the Cdk2_IFDOutputs group by clicking on the group heading.
  2. Includethe entry is represented in the Workspace, the circle in the In column is blue the predicted poses one at a time in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  3. Use the right and left arrow keys to step through the poses.

The top scored ligand poses obtained from IFD are all quite similar to the crystallographic reference, and the positions of the two lysine residues are also well predicted. However, the top-ranked pose is not necessarily the one that is most similar to the experimental structure. This is an inherent consequence of the fact that both terms in the IFDScore (Glide gscore and Prime energy) cannot capture the dynamics and entropic contributions to the protein-ligand binding. Similar to how you shouldn’t expect the Glide score to correlate with experimental binding affinities, you should not rely on the relative ranking of poses by IFD score without further validation.

Further analysis of these results, e.g. by using molecular dynamics, would be needed to determine the correct pose.

5. Conclusion and References

In this tutorial, you learned how to use softened potentials in Glide and IFD to predict bound poses allowing for some level of receptor flexibility.

Softening the potentials in Glide is an efficient way to account for known flexibility of specific residues and allow docking-based screening and pose prediction. However, this method cannot capture a larger-scale range of motion of the residues (short of entirely deleting them) or reorganization of the pocket.

The Induced Fit Docking protocol can allow the pocket to re-organize around a ligand and give better ligand poses in cases where a simple scaling of the potentials is insufficient. However, good knowledge of the system and additional validation is required to obtain reliable poses.

The IFD-MD protocol extends IFD by including implicit and explicit solvent molecular dynamics to more systematically explore the range of motion of the target and give more reliable determinations of the relative energetics of bound poses.

For further reading:
  • See the help documentation on the IFD and IFD-MD panels.

6. Glossary of Terms

Entries- a simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion

Fix - double-click the circle in the In column (or right-click the circle and click “Fix in Workspace”)

Included - the entry is represented in the Workspace, the circle in the In column is blue

Project Table - displays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data

Recent actions - This is a list of your recent actions, which you can use to reopen a panel, displayed below the Browse row. (Right-click to delete.)

Scratch Project - a temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project

Selected - (1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries

Working Directory - the location where files are saved

Workspace - the 3D display area in the center of the main window, where molecular structures are displayed